Extraction of Terbium from Terbium–Iron Alloys Using Molten Magnesium

Abstract

Developing a new and energy-saving production process for high-purity terbium that combines molten salt electrolysis, liquid metal extraction, and distillation is necessary. Therefore, we experimentally determined the optimal conditions for the selective extraction of Tb from a Tb–Fe alloy using molten magnesium, which can be obtained via molten salt electrolysis. A eutectic alloy composition of 88 mass%Tb–12 mass%Fe and Mg was maintained at 973 or 1273 K under an Ar atmosphere. The sample was separated into an Fe-based solid phase with a Tb content lower than the eutectic composition and a Tb–Mg-based liquid phase with a small amount of Fe. The obtained results indicate that Tb alloys containing approximately 0.2 mass% Fe can be recovered by liquid Mg extraction at 973 K and subsequent distillation, assuming that only the Mg in the liquid phase is fully volatilized. This concentration of Fe was approximately 2 mass% at 1273 K. Therefore, increasing the extraction temperature significantly increased the concentration of Fe in the Tb obtained by the new process.

This figure corresponds to Fig. 1.

1. Introduction

Rare earth elements (REEs) are essential for high-technology products. Terbium, the subject of this study, is an REE recognized as an important metal in countries and regions, such as China, the United States, and the European Union [1]. Tb is widely used in the production of green lamp phosphors, permanent magnet materials, magneto-strictive materials, magneto-optical materials [1–4]. China is the largest producer of Tb and REEs. Gao et al. analyzed the material flow of Tb in China [1]. They found that in 2011, the use of Tb in Nd–Fe–B permanent magnet applications was 90 t, accounting for 17% of the total, and in 2020, it was 243 t, accounting for 35%. The demand for Tb in magnets has continuously been increasing.

Nd–Fe–B permanent magnets, which are indispensable for realizing a low-carbon society and sustainable development, are used in electronics, battery-powered electric vehicles, and wind turbine generators. Neodymium is mainly used in these magnets, but small amounts of heavy rare-earth elements (HREEs), such as Dy and Tb, are also added to increase the heat resistance of the magnets [5, 6]. Among HREEs, the addition of Tb has been reported to increase the coercivity of magnets compared to Dy [7–9]. Therefore, Tb is essential for Nd–Fe–B permanent magnets, which require high performance at high temperatures. Dy and Tb are more expensive than Nd, and their distribution in the inner (Nd, HREE)₂Fe₁₄B (HREE = Dy and Tb) grains hardly contributes to the coercivity enhancement [6]. Thus, Dy and Tb diffused into the grain boundary region of the matrix phase via the grain boundary diffusion process (GBDP) [5–12]. The GBDP can be roughly divided into three classes based on the characteristics of its diffusion sources. The first class utilizes a Dy/Tb-rich compound powder as the diffusion source; the second class uses Dy/Tb/Pr/Nd-rich alloys with low-melting-point metals, such as Cu and Al, as the diffusion sources; and the third class employs vapors or sputtered thin layers of pure Dy/Tb [5, 7–9, 11–14]. In contrast to Tb, Dy is widely used in pure Dy and Dy–Fe alloys. Because there is a low demand for Tb–Fe alloys, they should be refined to obtain pure Tb. Although molten salt electrolysis is used to obtain Dy–Fe alloys, its application to Tb to produce Tb–Fe alloys remains impractical owing to the low demand for this material. Furthermore, the GBDP characteristics depend on the magnet producer. Hence, the production of Tb requires the refinement of high-purity Tb metal as a raw material for GBDP.

The current production process for high-purity Tb involves thermochemical reduction using metallic Ca [15, 16]. Pure TbF₃ obtained through various processes, such as solvent extraction, is reduced to metallic Tb by metallic Ca at high temperatures, and metallic Tb ingots are obtained through the distillation of unreacted Ca. This process utilizes the highly reactive metallic Ca and involves the reduction of CaF₂ to Ca, which consumes more energy than the reduction of TbF₃ to Tb. Additionally, each process step is performed at a high temperature; thus, a large amount of energy is required. Therefore, an energy-efficient process for the production of metallic Tb is highly desirable.

The metallic Tb produced by thermochemical reduction using metallic Ca is possible owing to the high melting point of Tb (1629 K). In most REEs, metals are recovered from pure oxides or fluorides via molten salt electrolysis at temperatures higher than their melting points [17, 18]. Such elevated temperatures are selected to avoid the electrodeposition of solid REEs, which negatively affects the current efficiency and separation of solid REE metals and molten salt [19]. For high-melting-point REEs, such as Dy, which has a melting point of 1680 K, stable operation on an industrial scale is difficult to implement at temperatures higher than the melting point. In such cases, a consumable cathode, such as Fe, is employed, and REE–Fe liquid alloys are produced at temperatures considerably lower than the melting points of the REEs. Although molten salt electrolysis using a consumable cathode is also a well-established process with high productivity, it is currently not applicable to Tb metal because of the low utilization value of Tb–Fe alloys, as described in this section.

Based on this background, the authors propose a new Tb metal production process that simultaneously achieves high productivity and low energy consumption compared to conventional thermochemical reduction using metallic Ca [20]. Figure 1 shows a conceptual drawing of the new process. The starting materials for this process consist of pure TbF₃, TbCl₃, Tb₂O₃, and Tb₄O₇, which can be obtained via conventional processes, such as solvent extraction, precipitation, and calcination. These chemicals are subjected to molten salt electrolysis using different consumable cathodes, such as Fe, Ni, and Co. Hereafter, only Fe is considered as a consumable cathode material because it has been previously used to produce Dy–Fe alloys and is one of the most promising candidates for our process [21–23]. A Tb–Fe binary phase diagram is shown in Fig. 2 [24]. When electrolysis was conducted at a temperature higher than the eutectic temperature of Tb–TbFe₂ (1115 K), the Tb–Fe liquid alloy with a composition corresponding to the liquid line at the electrolysis temperature was recovered. The formed Tb–Fe liquid alloy was collected in a crucible placed below the cathode. Because the electrochemical properties of Tb are similar to those of Dy, except for the presence of tetravalent Tb ions, the proposed molten salt electrolysis step is presumed to be applicable to Tb. Molten salts include fluoride-based and chloride-based systems; however, a combination of molten fluoride serving as the melt and Tb oxide used as the Tb-ion source may be ideal for commercialization. The obtained Tb–Fe alloy was subjected to liquid metal extraction using various elements, such as Cd, Zn, and Mg, which exhibit high vapor pressures at relatively low temperatures compared to those of most metals. Among these elements, Mg is ideal for our process because it hardly dissolves Fe; thus, Tb can be selectively extracted. Finally, pure metallic Tb ingots were obtained by vacuum distillation and melting. The removed elements, such as Mg, can be reused in the liquid metal extraction step. When established, these new processes are expected to replace thermochemical reduction using metallic Ca to achieve higher productivity, safety, and energy efficiency in the refinement of Tb and other high-melting-point REEs. According to the literature, molten salt electrolysis can be achieved using a consumable cathode and distillation [21–23, 25–28]. Thus, the optimization of liquid metal extraction is essential for the proposed process.

Fig. 1

Conceptual drawing of the new Tb metal production process. (online color)

Fig. 2

Tb–Fe binary phase diagram [24].

Several studies have been conducted on liquid metal extraction to recycle REEs from Nd–Fe–B permanent magnet waste [29, 30]. Some researchers have reported the use of molten Mg for this purpose [31–41]. In addition to Mg, Ishigaki et al., Li et al., and Nakamuta et al. used Bi metal; Takeda et al. utilized Ag metal; and Moore et al. employed Cu metal [42–46]. These studies demonstrate that REEs can be enriched in the liquid metal phase from Nd–Fe–B permanent magnets. However, in these studies, REEs were recovered as a mixture, and a separation process was required to reuse Dy and Tb in high-performance magnets. The present study does not target Nd–Fe–B permanent magnets for extraction, but instead utilizes high-concentration REE-containing alloys obtained via molten salt electrolysis, which has not been reported previously. Furthermore, all the aforementioned studies on liquid metal extraction from Nd–Fe–B permanent magnets focused on Nd, Dy, and Pr, and no studies on Tb extraction have been performed. In addition, no researchers have focused on the low Fe content in the liquid metal phase, which is one of the most important parameters for determining the purity of the obtained metallic Tb. Herein, considering the process for extracting Tb with molten Mg from Tb–Fe alloys obtained via molten salt electrolysis, the temperature, extraction time, and Tb–Fe alloy-to-Mg ratio were varied to optimize the extraction conditions.

2. Experimental Procedure

First, 1.76 g of Tb (purity >99.9%) and 0.24 g of Fe (purity: >99.99 mass%) with particle sizes of less than 3 mm were placed in a MgO crucible (purity: >99.6 mass%) with an outer diameter of 15 mm, an inner diameter of 11 mm, and a height of 50 mm. The sample was vacuum-sealed in a quartz tube with an inner diameter of 18 mm and outer diameter of 20 mm, maintained at 1373 (±3) K for 110 min in a furnace with a SiC heating element, and then quenched with water to obtain the eutectic Tb–TbFe₂ alloy composition with 88 mass%Tb–12 mass%Fe (72 at%Tb–28 at%Fe). The sample was crushed and analyzed quantitatively using the method described below for the alloy after extraction with Mg, and the alloy composition was 88 mass%Tb–12 mass%Fe. The concentration of Mg in the alloy was less than 0.1 mass%, confirming that the alloy hardly reduced the MgO crucible.

A schematic of the experimental setup for the liquid metal extraction process is shown in Fig. 3. The obtained 88 mass%Tb–12 mass%Fe alloy with a grain size of 1–3 mm and Mg (purity: >99.9 mass%) were placed in a MgO crucible of the same type with a total mass of 1.5 g. The mass ratio of the 88 mass%Tb–12 mass%Fe alloy to Mg was varied, and the ratio of Mg to the sum of Tb–Fe alloy and Mg was a minimum of 30 and a maximum of 90 mass%, respectively. The sample and the MgO crucible were placed at the bottom of an Al₂O₃ tube with an outer diameter of 26 mm, inner diameter of 20 mm, and height of 128 mm. The Al₂O₃ tube containing the sample was placed in a SiO₂ reaction tube with an outer diameter of 31 mm, inner diameter of 37 mm, and height of 500 mm. Air was replaced with Ar gas (purity: >99.99 vol%) at 300 (±5) K. Subsequently, they were placed in a heated resistance furnace at 973 or 1273 (±3) K, and then maintained for a predetermined period. Figure 4 shows the temperature profile at the bottom of the MgO crucible when a Type-R thermocouple occupied the place of the metal sample, and the temperature increased. When the set temperatures were 973 and 1273 K, the target temperatures were achieved in approximately 14 and 19 min, respectively. The first 20 min after placing the sample in the resistance furnace were defined as the temperature-heating time, while the remaining time as the maintained time. Maintenance was performed for 20, 110, or 350 min. During these experiments, Ar gas flowed into the SiO₂ reaction tube at a rate of 200 mL/min to suppress sample oxidation. Thereafter, the samples were quenched with water and cut vertically.

Fig. 3

Schematic of the experimental setup. (online color)

Fig. 4

Temperature profile during heating.

The composition of each phase was quantitatively determined by embedding the samples in a phenol-based resin, polishing, and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM–EDX; JEOL, JSM-6510A). SEM–EDX was performed at an accelerating voltage of 20 kV, and the characteristic X-rays were corrected using the ZAF method. In particular, for the sample heated to 973 K for 20 min, a vertical gradient of Mg and Tb concentrations was observed in the Tb–Mg-based phase at high temperatures via SEM–EDX. Furthermore, the concentration of Fe in the Tb-Mg-based phase was low (<1 mass%) in most of the samples. Therefore, the sample was cut at the height of the upper one-third of the sample in the vertical direction and dissolved in nitric acid. The concentrations of Mg, Fe, and Tb in the Tb-Mg-based phase were quantified using inductively coupled plasma optical emission spectrometry (ICP–OES; Agilent Technologies, 5100). The wavelengths of Mg (279.553 nm), Fe (238.204 nm), and Tb (384.873 nm) were used for the concentration determination via ICP–OES.

3. Results and Discussion

Solid phases of Fe or Fe-containing compounds were observed in the lower part of the crucible after extraction. Tb-Mg-based phases, which are liquid phases at high temperatures, were detected entirely in the upper part and around the Fe or Fe-containing compounds in the lower part. Hereafter, the Tb-Mg-based phase is referred to as the liquid phase because molten Mg extracts Tb from the Tb–Fe alloy, and the solid and liquid phases are separated owing to their density difference. Table 1 lists the initial compositions of the studied samples, as well as the compositions of the solid and liquid phases determined via SEM–EDX and ICP–OES, respectively. In addition, it shows the calculated extraction rates of Tb at 973 K. The method for calculating the extraction rate of Tb is described in detail below. Analysis of the solid phases using SEM–EDX involved averaging more than 10 values obtained for each phase (Fe, Tb₂Fe₁₇, Tb₆Fe₂₃, and TbFe₃) with a standard deviation. Table 2 presents the results obtained at 1273 K. Cross-sectional backscattering electron images of the upper and lower parts of samples 4 and 17 are shown in Figs. 5 and 6, respectively. In Figs. 5 and 6, the upper part of the sample is in the liquid phase at high temperature, and a microstructure is formed during the cooling step. In Fig. 5, the liquid and solid phases coexisted in the lower part of the sample, and spherical Tb₂Fe₁₇ and needle-like TbFe₃ solid phases were observed. As shown in Fig. 6, the upper part of the sample was in the liquid phase, and the liquid and solid Tb₂Fe₁₇ phases coexisted in the lower part. Tb oxides were detected in all samples; however, microstructural observations suggested that the amount of Tb oxides was considerably smaller than that of the liquid and Fe-containing solid phases and Tb hardly reduced the MgO crucible. In all experiments conducted at 973 K and for 20 min at 1273 K, the decrease in the sample mass was less than 5%, suggesting that the samples hardly volatilized. At 1273 K, the total masses of the samples decreased from 1.50 to 1.46, 1.12, and 0.593 g after 20, 110, and 350 min, respectively (nos. 19–21). Maintaining the samples at 1273 K increased the vapor pressure of Mg, thereby promoting volatilization. In these experiments, the systems did not reach equilibrium at 973 K because the liquid-phase composition continued to change over time. The Fe-containing solid phases varied depending on the initial composition and heating time; the higher the Tb concentration in the liquid phase, the higher the Tb concentration in the solid phase. During experiment No. 12, in which the Tb–Fe alloy content was 70 mass% and the Mg concentration was 30 mass%, the solid phase was sparsely distributed throughout the liquid phase, and the solid and liquid phases were not vertically separated. Because the densities of liquid Mg at its melting point (923 K), solid Tb at its melting point (293 K), liquid Tb at its melting point (1625 K), and solid Fe at its melting point (293 K) are 1.58 × 10³ kg/m³, 8.23 × 10³ kg/m³, 7.84 × 10³ kg/m³, and 7.87 × 10³ kg/m³, respectively, the density of the liquid phase drastically increases with increasing Tb content [47, 48]. The increased density of the liquid phase is supposed to be the reason for the experimental result that the solid and liquid phases were not separated. These results show the existence of a lower limit for the amount of Mg required for the extraction of Tb–Fe alloys. In this study, an Mg concentration exceeding 40 mass% was necessary for the samples.

Table 1 Initial sample compositions, compositions of the solid phase determined via SEM–EDX, compositions of the liquid phase determined via ICP–OES after extraction, and calculated extraction rates of Tb at 973 K.

Table 2 Initial sample compositions, compositions of the solid phase determined via SEM–EDX, compositions of the liquid phase determined via ICP–OES after extraction, and calculated extraction rates of Tb at 1273 K.

Fig. 5

(a) Cross-sectional photograph, back-scattering electron images of (b) upper and (c) lower part of No. 4. (online color)

Fig. 6

(a) Cross-sectional photograph, back-scattering electron images of (b) upper and (c) lower part of No. 17. (online color)

During the liquid-metal extraction step of the new process, it is desirable to extract a large amount of Tb rather than Fe. The relationship between the Tb and Fe concentrations in the liquid phase is shown in Fig. 7. The legend of the graph shows the temperature, initial Mg concentration, maintenance time, and the Fe-containing solid phases coexisting with each liquid phase. The results obtained at the same temperature and initial composition are indicated by dashed lines. Figure 8 shows the relationship between the maintenance time and Tb concentration in the liquid phase. Figures 7 and 8 show that the lower the initial concentration of Mg and the longer the maintenance time, the higher the concentrations of Tb and Fe in the liquid phase after extraction. Furthermore, by comparing the results obtained at different temperatures, it was found that the concentrations of Fe and Tb in the liquid phase were lower at 973 K, based on the data obtained at the same composition for 20 min. Although there are no reported studies extracting Tb, and a direct comparison between the present results and those in the literature is difficult, the fact that Mg selectively extracts REEs more than Fe is consistent with many previous studies [31–41].

Fig. 7

Relationship between the concentrations of Tb and Fe in the liquid phase (Legend: initial concentration of Mg, maintained time, and Fe-containing solid phases coexisting with each liquid phase). (online color)

Fig. 8

Relationship between the maintained time and the concentration of Tb in the liquid phase (Legend: initial concentration of Mg). (online color)

In this process, the extracted metals are separated from Mg via distillation (Fig. 1). Therefore, for liquid metal extraction, the extraction rate of Tb and the Fe-to-Tb ratio in the liquid phase are evaluation indices. Because the vapor pressures of Tb and Fe are considerably lower than that of Mg, Fe is expected to remain in Tb. It was difficult to calculate the extraction rate of Tb owing to the following reasons: Mg was only slightly soluble in the Fe-containing solid phase, a concentration gradient in the liquid phase was observed in the experiment conducted at 973 K for 20 min, and Mg volatilized over time. Therefore, except for sample Nos. 20 and 21 heated to 1273 K for 110 and 350 min, respectively, the calculated extraction rate can be roughly expressed by the following equation, assuming negligible solubility of solid Mg in the Fe-containing solid phase, concentration gradient in the liquid phase, and volatilization of Mg:

  

\begin{equation} E_{\text{Tb}}\ (\%) = \frac{m_{\text{Tb}}^{\text{e}}}{m_{\text{Tb}}^{\text{i}}} \times 100 = \frac{m_{\text{Mg}}^{\text{i}}}{m_{\text{Tb}}^{\text{i}}} \cdot \frac{c_{\text{Tb}}^{\text{e}}}{c_{\text{Mg}}^{\text{e}}} \times 100 \end{equation} (1)

where E_Tb is the calculated extraction rate of Tb; $m_{\text{Tb}}^{\text{i}}$ is the mass of Tb in the initial sample; $m_{\text{Tb}}^{\text{e}}$ is the mass of Tb in the liquid phase after extraction; $m_{\text{Mg}}^{\text{i}}$ is the mass of Mg in the initial sample (assumed to be the same as the mass of Mg in the liquid phase after extraction); $c_{\text{Mg}}^{\text{e}}$ is the concentration of Mg in the liquid phase after extraction; and $c_{\text{Tb}}^{\text{e}}$ is the concentration of Tb in the liquid phase. The relationship between the calculated extraction rate of Tb and Fe-to-Tb ratio in the liquid phase is shown in Fig. 9. The vertical axis shows the concentration of Fe in the Tb alloy when the liquid phase after extraction has completely volatilized only Mg via distillation (hereinafter referred to as “the concentration of Fe in the distilled alloy”). These results indicate that the calculated extraction rate of Tb at 20 min was higher at 1273 K than at 973 K. Because the eutectic temperature of Tb–TbFe₂ was 1115 K, the Tb–Fe alloy was in the liquid phase at 1273 K, and the mass transfer speed increased at higher temperatures. The Fe concentration in the distilled alloys was approximately 0.2 mass% at 973 K and 2% by mass at 1273 K. Extraction at 923 K is expected to improve the purity of Tb through distillation. Compared with the results obtained at the same temperatures, the conditions corresponding to higher calculated extraction rates of Tb resulted in a slightly higher concentration of Fe in the distilled alloys. Thus, during the liquid-metal extraction process, the extraction rate and speed increased with increasing temperature; similarly, the concentration of Fe in the distilled alloy increased. Because this process was intended to produce high-purity Tb, Experiment No. 5, in which Mg was 75 mass% for the 25 mass% Tb–Fe alloy heated to 973 K for 110 min, was considered ideal for extraction. Under these conditions, 66% of the Tb in the Tb–Fe alloy was expected to be extracted, and the concentration of Fe in the distilled alloy was 0.19 mass%. To obtain purer Tb, it is necessary to change the temperature and metal used for extraction.

Fig. 9

Relationship between the calculated extraction rates of Tb and Fe-to-Tb ratio in the liquid phase (Legend: initial concentration of Mg, maintained time, and Fe-containing solid phases coexisting with each liquid phase). (online color)

4. Conclusion

In this study, a new Tb production process was proposed, and the liquid metal extraction of Tb from Tb–Fe alloys using molten Mg was investigated. A eutectic alloy composition of 88 mass%Tb–12 mass%Fe and Mg was maintained at 973 and 1273 K under an Ar atmosphere. The sample was separated into an Fe-based solid phase at the bottom of the crucible and a Tb–Mg-based liquid phase with a small amount of Fe. The results revealed that the lower the initial concentration of Mg and the longer the heating time, the higher the concentrations of Tb and Fe in the liquid phase after extraction. Thus, the Tb alloy containing approximately 0.2 mass% Fe can be extracted at 973 K, and the alloy containing approximately 2 mass% Fe can be extracted at 1273 K, assuming that only Mg in the liquid phase is volatilized entirely in the subsequent distillation process. The experimental conditions in which 75 mass% Mg and 25 mass% Tb–Fe alloy were heated to 973 K for 110 min were considered ideal for extraction. Under these conditions, 66% of the Tb in the Tb–Fe alloy was expected to be extracted, and the concentration of Fe in the distilled alloy was 0.19 mass%.

Acknowledgments

This work was financially supported by the New Energy and Industrial Technology Development Organization (NEDO) [grant number JPNP23006]. The authors would like to thank the Environmental Safety Center of Waseda University for the use of ICP–OES.

REFERENCES

• 1)   C.  Gao,  Y.  Xu,  Y.  Geng and  S.  Xiao: Uncovering terbium metabolism in China: A dynamic material flow analysis, Resour. Policy 79 (2022) 103017. doi:10.1016/j.resourpol.2022.103017
• 2)   K.  Binnemans and  P.T.  Jones: Perspectives for the recovery of rare earths from end-of-life fluorescent lamps, J. Rare Earths 32 (2014) 195–200. doi:10.1016/S1002-0721(14)60051-X
• 3)   S.  Lee and  J.G.  Kim: Industrial Demand and Integrated Material Flow of Terbium in Korea, International Journal of Precision Engineering and Manufacturing-Green Technology 1 (2014) 145–152. doi:10.1007/s40684-014-0019-y
• 4)   N.A.  Mancheri,  B.  Sprecher,  G.  Bailey,  J.  Ge and  A.  Tukker: Effect of Chinese policies on rare earth supply chain resilience, Resour. Conserv. Recycling 142 (2019) 101–112. doi:10.1016/j.resconrec.2018.11.017
• 5)   F.  Chen: Recent progress of grain boundary diffusion process of Nd–Fe–B magnets, J. Magn. Magn. Mater. 514 (2020) 167227. doi:10.1016/j.jmmm.2020.167227
• 6)   J.  Jin,  T.  Ma,  Y.  Zhang,  G.  Bai and  M.  Yan: Chemically Inhomogeneous RE–Fe–B Permanent Magnets with High Figure of Merit: Solution to Global Rare Earth Criticality, Sci. Rep. 6 (2016) 32200. doi:10.1038/srep32200
• 7)   Y.I.  Lee,  H.W.  Chang,  G.Y.  Huang,  C.W.  Shih and  W.C.  Chang: Comparison on the Coercivity Enhancement of Hot-Deformed Nd₂Fe₁₄B-Type Magnets by Doping R₇₀Cu₃₀ (R = Nd, Dy, and Tb) Alloy Powders, IEEE Trans. Magn. 53 (2017) 2100704. doi:10.1109/TMAG.2017.2703128
• 8)   Y.I.  Lee,  Y.J.  Wong,  H.W.  Chang and  W.C.  Chang: Coercivity enhancement of hot-deformed NdFeB magnets by doping R₈₀Ga₂₀ (R = Pr, Dy, Tb) alloys, J. Magn. Magn. Mater. 478 (2019) 43–47. doi:10.1016/j.jmmm.2019.01.090
• 9)   X.  Xia,  X.  Tang,  Y.  Du,  H.  Wu,  J.  Ju,  L.  Wu,  R.  Chen,  W.  Yin,  L.  Zhang and  A.  Yan: Significant enhancement in the coercivity and thermal stability of bulk hot-deformed Nd-Fe-B magnets by intergranular addition of Nd-Dy/Tb-Cu-Ga alloys, Acta Mater. 232 (2022) 117945. doi:10.1016/j.actamat.2022.117945
• 10)   Y.  Zhang,  T.  Man,  X.  Liu,  P.  Liu,  J.  Jin,  J.  Zou and  M.  Yan: Coercivity enhancement of Nd–Fe–B sintered magnets with intergranular adding (Pr, Dy, Cu)–Hx powders, J. Magn. Magn. Mater. 399 (2016) 159–163. doi:10.1016/j.jmmm.2015.09.071
• 11)   F.  Yang,  S.  Lin,  P.  Liu,  H.  Sun,  Q.  Qin,  Y.  Shao,  Y.  Sui and  Z.  Guo: The effect of Dy₂O₃/MoS₂ co-doping on magnetic performance and microstructure in Nd–Fe–B sintered magnets, J. Magn. Magn. Mater. 498 (2020) 166103. doi:10.1016/j.jmmm.2019.166103
• 12)  K.T. Park, K. Hiraga and M. Sagawa: Proc. the 16th International Workshop on Rare Earth Magnets and Their Applications, (The Japan Inst. Metals, Sendai, 2000) pp. 257–264.
• 13)   H.  Nakamura,  K.  Hirota,  M.  Shimao,  T.  Minowa and  M.  Honshima: Magnetic Properties of Extremely Small Nd–Fe–B Sintered Magnets, IEEE Trans. Magn. 41 (2005) 3844–3846. doi:10.1109/TMAG.2005.854874
• 14)   M.  Tang,  X.  Bao,  K.  Lu,  L.  Sun,  J.  Li and  X.  Gao: Boundary structure modification and magnetic properties enhancement of Nd–Fe–B sintered magnets by diffusing (PrDy)–Cu alloy, Scr. Mater. 117 (2016) 60–63. doi:10.1016/j.scriptamat.2016.02.019
• 15)   G.  Chen,  X.  Bao,  K.  Lu,  X.  Lv,  Y.  Ding,  M.  Zhang,  C.  Wang and  X.  Gao: Microstructure and magnetic properties of Nd–Fe–B sintered magnet by diffusing Pr–Cu–Al and Pr–Tb–Cu–Al alloys, J. Magn. Magn. Mater. 477 (2019) 17–21. doi:10.1016/j.jmmm.2018.12.110
• 16)   W.  Li,  L.  Yang,  Q.  Zhang,  C.  Xu,  Q.  Zhu,  Z.  Song,  B.  Zheng,  F.  Hu and  J.  Jiang: Effect of the grain boundary Tb/Dy diffused microstructure on the magnetic properties of sintered Nd–Fe–B magnets, J. Magn. Magn. Mater. 502 (2020) 166491. doi:10.1016/j.jmmm.2020.166491
• 17)  C.K. Gupta and N. Krishnamurthy: Extractive Metallurgy of Rare Earths, (CRC Press, Florida, 2005).
• 18)   H.  Liu,  Y.  Zhang,  Y.  Luan,  H.  Yu and  D.  Li: Research Progress in Preparation and Purification of Rare Earth Metals, Metals 10 (2020) 1376. doi:10.3390/met10101376
• 19)   S.  Rayaprolu and  D.  Chidambaram: Electrochemical Deposition of Terbium from Molten Salts, ECS Trans. 58 (2014) 51–66. doi:10.1149/05845.0051ecst
• 20)  T. Oishi, Y. Katasho, T. Nohira, K. Kawaguchi and H. Konishi: Methods of recovering rare-earth element, Japanese Patent Disclosure, 2024-095613, (2024).
• 21)   H.  Tamamura: New Industrial Process for Electrowinning of Nd–Fe Alloy, J. Surf. Finish. Soc. Jpn. 60 (2009) 474–479 (in Japanese). doi:10.4139/sfj.60.474
• 22)   B.  Cai,  H.  Liu,  F.  Kou,  Y.  Yang,  B.  Yao,  X.  Chen,  D.S.  Wong,  L.  Zhang,  J.  Li,  G.  Kuang,  L.  Chen,  J.  Zheng,  D.  Guan and  Y.  Shan: Estimating perfluorocarbon emission factors for industrial rare earth metal electrolysis, Resour. Conserv. Recycling 136 (2018) 315–323. doi:10.1016/j.resconrec.2018.04.018
• 23)   K.  Itoh,  E.  Nakamura and  S.  Sasaki: Preparation methods of rare earth metals and their alloys, Sumitomo Light Metal Technical Reports 29 (1988) 127–144 (in Japanese).
• 24)   S.  Landin and  J.  Ågren: Thermodynamic assessment of Fe–Tb and Fe–Dy phase diagrams and prediction of Fe–Tb–Dy phase diagram, J. Alloy. Compd. 207–208 (1994) 449–453. doi:10.1016/0925-8388(94)90261-5
• 25)   K.  Cheng,  J.  Yi,  G.  Zha,  K.  Fan,  Z.  Li,  X.  Kong,  B.  Yang,  H.  Xiong,  B.  Xu and  W.  Jiang: Separation behavior of As, Zn and Cd trace impurities in the deep vacuum purification process of refined lead, Separ. Purif. Tech. 278 (2021) 119531. doi:10.1016/j.seppur.2021.119531
• 26)   H.  Wan,  B.  Xu,  B.  Yang,  J.  Zhao and  Y.  Dai: The impurities distribution and purification efficiency of high-purity aluminum preparation by zone melting in vacuum, Vacuum 171 (2020) 108839. doi:10.1016/j.vacuum.2019.108839
• 27)   G.  Zha,  Y.  Wang,  M.  Cheng,  D.  Huang,  W.  Jiang,  B.  Xu and  B.  Yang: Purification of crude selenium by vacuum distillation and analysis, J. Mater. Res. Technol. 9 (2020) 2926–2933. doi:10.1016/j.jmrt.2020.01.043
• 28)   S.  Park,  D.K.  Kim,  J.  Jeong,  J.H.  Shin,  Y.  Kang,  R.  Liu,  T.S.  Kim and  M.  Song: Separation and recovery Nd and Dy from Mg-REEs alloy by vacuum distillation, J. Alloy. Compd. 967 (2023) 171775. doi:10.1016/j.jallcom.2023.171775
• 29)   M.Z.  Rasheed,  S.  Nam,  J.  Cho,  K.  Park,  B.  Kim and  T.  Kim: Review of the Liquid Metal Extraction Process for the Recovery of Nd and Dy from Permanent Magnets, Metall. Mater. Trans. B 52 (2021) 1213–1227. doi:10.1007/s11663-021-02102-z
• 30)   M.  Firdaus,  M.A.  Rhamdhani,  Y.  Durandet,  W.J.  Rankin and  K.  McGregor: Review of High-Temperature Recovery of Rare Earth (Nd/Dy) from Magnet Waste, J. Sustain. Metall. 2 (2016) 276–295. doi:10.1007/s40831-016-0045-9
• 31)   Y.  Xu,  L.S.  Chumbley and  F.C.  Laabs: Liquid metal extraction of Nd from NdFeB magnet scrap, J. Mater. Res. 15 (2000) 2296–2304. doi:10.1557/JMR.2000.0330
• 32)   T.H.  Okabe,  O.  Takeda,  K.  Fukuda and  Y.  Umetsu: Direct Extraction and Recovery of Neodymium Metal from Magnet Scrap, Mater. Trans. 44 (2003) 798–801. doi:10.2320/matertrans.44.798
• 33)   O.  Takeda,  T.H.  Okabe and  Y.  Umetsu: Phase equilibria of the system Fe–Mg–Nd at 1076K, J. Alloy. Compd. 392 (2005) 206–213. doi:10.1016/j.jallcom.2004.09.020
• 34)   O.  Takeda,  T.H.  Okabe and  Y.  Umetsu: Recovery of neodymium from a mixture of magnet scrap and other scrap, J. Alloy. Compd. 408–412 (2006) 387–390. doi:10.1016/j.jallcom.2005.04.094
• 35)   T.  Akahori,  Y.  Miyamoto,  T.  Saeki,  M.  Okamoto and  T.H.  Okabe: Optimum conditions for extracting rare earth metals from waste magnets by using molten magnesium, J. Alloy. Compd. 703 (2017) 337–343. doi:10.1016/j.jallcom.2017.01.211
• 36)   H.J.  Chae,  Y.D.  Kim,  B.S.  Kim,  J.G.  Kim and  T.  Kim: Experimental investigation of diffusion behavior between molten Mg and Nd–Fe–B magnets, J. Alloy. Compd. 586 (2014) S143–S149. doi:10.1016/j.jallcom.2013.02.156
• 37)   H.W.  Na,  Y.H.  Kim,  H.T.  Son,  I.H.  Jung,  H.S.  Choi and  T.B.  Kim: Effect of Scrap Size on Extraction of Neodymium from Nd-Fe-B Magnet Scrap by Liquid Metal Extraction, Curr. Nanosci. 10 (2014) 128–130. doi:10.2174/1573413709666131109002236
• 38)   M.  Sun,  X.  Hua,  L.  Peng,  P.  Fu,  W.  Ding and  Y.  Peng: On the production of Mg–Nd master alloy from NdFeB magnet scraps, J. Mater. Process. Technol. 218 (2015) 57–61. doi:10.1016/j.jmatprotec.2014.11.036
• 39)   Y.  Kim,  S.  Nam,  K.  Park and  T.  Kim: Extraction Behavior of Rare Earth Element (Nd, Dy) from Rare Earth Magnet by Using Molten Magnesium, Sci. Adv. Mater. 9 (2017) 2166–2172. doi:10.1166/sam.2017.3216
• 40)   M.Z.  Rasheed,  S.W.  Nam,  S.H.  Lee,  S.M.  Park,  J.Y.  Cho and  T.S.  Kim: Phase Evolution During Extraction and Recovery of Pure Nd from Magnet, Arch. Metall. Mater. 66 (2021) 1001–1005. doi:10.24425/amm.2021.136414
• 41)  K. Kawai, R. Sanekata, A. Nakamuta and T. Nagai: Proc. the 2022 MMIJ Annual Meeting, (The Mining and Materials Processing Institute of Japan, Fukuoka, 2022) (in Japanese).
• 42)   A.  Ishigaki and  K.  Okumura: Recovery of Neodymium from Neodymium Magnet Using Bismuth, Tetsu-to-Hagané 104 (2018) 613–619 (in Japanese). doi:10.2355/tetsutohagane.TETSU-2018-037
• 43)   C.  Li,  Z.  Wang and  Z.  Gou: Extraction of rare earth elements from spent NdFeB magnets by supergravity-enhanced bismuth, Nonferrous Metals Science and Engineering 14 (2023) 125–134. doi:10.13264/j.cnki.ysjskx.2023.01.015
• 44)  A. Nakamuta and T. Nagai: Proc. the 2023 JIMM Autumn Meeting, (The Japan Inst. Metals, Toyama, 2023) (in Japanese).
• 45)   O.  Takeda,  T.H.  Okabe and  Y.  Umetsu: Phase equilibrium of the system Ag–Fe–Nd, and Nd extraction from magnet scraps using molten silver, J. Alloy. Compd. 379 (2004) 305–313. doi:10.1016/j.jallcom.2004.02.038
• 46)   M.  Moore,  A.  Gebert,  M.  Stoica,  M.  Uhlemann and  W.  Löser: A route for recycling Nd from Nd–Fe–B magnets using Cu melts, J. Alloy. Compd. 647 (2015) 997–1006. doi:10.1016/j.jallcom.2015.05.238
• 47)  J.W. Arblaster: Selected Values of the Crystallographic Properties of the Elements, (ASM International, Ohio, 2018).
• 48)   P.  Paradis,  T.  Ishikawa,  N.  Koike and  Y.  Watanabe: Physical Properties of Liquid Terbium Measured by Levitation Techniques, J. Rare Earths 25 (2007) 665–669. doi:10.1016/S1002-0721(08)60003-4