Keywords

Introduction

Neodymium (Nd)–iron (Fe)–boron (B) magnets with addition of small amount of dysprosium (Dy) have excellent properties such as high coercivity even at high temperatures [1]. Because of these properties, Nd magnets can be used in various applications such as hybrid and electric cars. The demand for green cars is expected to increase in the future to meet the need of carbon dioxide (CO2) gas emission reduction; therefore, the utilization of Nd magnets will increase.

Rare earth elements (REEs) are extracted from several mineral resources such as bastnaesite, monazite, and ion adsorption deposits. Among REEs, light rare earth elements (LREEs), such as Nd and praseodymium (Pr), are extracted from bastnaesite and monazite. These mineral resources are abundant in nature and also distributed around the world [2]. However, heavy rare earth elements (HREEs), such as Dy and terbium (Tb), can only be extracted economically from ion adsorption deposits in the southern parts of China [2]. This is the one reason that China monopolizes the world production of REEs.

During the extraction of LREEs from monazite, environmental threats and radiation risks are encountered, because monazite contains radioactive elements such as uranium (U) and thorium (Th) [2]. Therefore, proper treatments of these radioactive elements generated during the extraction are necessary to preserve the environment. In contrast, ion adsorption deposits contain very little amount of radioactive elements because these elements are washed out because of weathering [2]. However, there are still environmental concerns during the extraction of REEs because they are extracted via elution using ammonium sulfate ((NH4)2SO4).

To produce REEs with preservation of environment, recycling of spent Nd magnets is required. However, only Nd scraps generated in magnet production plants are commercially used as feedstock for recycling because these scraps contain a large amount of REEs and processing costs are minimum [2, 3]. Although spent Nd magnets are not recycled at present, it is necessary to develop an efficient and environmentally sound process for the recovery of REEs from spent magnets when the increase of the use of Nd magnets is considered.

Numerous methods have been developed to recover REEs from spent magnets [4,5,6,7,8,9]. Among those, the selective recovery of REEs through the chlorination of spent magnets using metal chloride is promising [8, 9]. When the selective chlorination of spent magnets using ferrous chloride (FeCl2) or magnesium chloride (MgCl2) is utilized, only REEs are recovered as chloride without chlorination of Fe in magnets. Shirayama reported that Nd in a spent magnet was selectively chlorinated to neodymium chloride (NdCl3) by reacting with MgCl2 at 1273 K [8]. In addition, approximately 80% of Nd and Dy was recovered from the initial magnet after the reactions for 12 h [8]. Hua reported that the rate-determining step of the selective chlorination using MgCl2 is the diffusion of MgCl2 into an ash layer consisting of Fe metal remained [10].

In this study, a selective chlorination process using zinc chloride (ZnCl2) was investigated to recover REEs from Nd magnets, as shown in Fig. 1 [11]. When ZnCl2 is used as the chlorinating agent, reaction temperature can be decreased because of its low melting point. In addition, the reaction rate can be increased because low melting temperature of zinc (Zn) metal produced by the chlorination reaction is alloyed with the ash layer of Fe metal. Furthermore, rare earth chlorides such as NdCl3 can be separated from the residue obtained after selective chlorination by utilizing different vapor pressures at high temperatures.

Fig. 1
figure 1

Flowchart of the selective chlorination of Nd magnet using ZnCl2 investigated in this study

Full size image
Fig. 2
figure 2

Chemical potential diagrams of the a Nd–Zn–Cl and b Fe–Zn–Cl systems at 1000 K (the boiling point of ZnCl2 is 999.5 K [12] or 1005 K [14] depending on the reference consulted, Reprinted by permission from Springer Nature: Springer, Journal of Sustainable Metallurgy, Selective Chlorination of Rare Earth Elements from a Nd–Fe–B Magnet Using Zinc Chloride [15]

Full size image

Mechanisms of Selective Chlorination of Nd-Fe-B Magnet Using ZnCl2

Figure 2a, b show the chemical potential diagrams of Nd–Zn–Cl and Fe–Zn–Cl systems, respectively, at 1000 K [12, 13]. ZnCl2 (l,g) equilibrates with Zn (l) and NdCl3 (s) at point a in Fig. 2a while Nd (s) does not coexist with ZnCl2 (l,g) at 1000 K. Therefore, when a sufficient amount of ZnCl2 is used as a chlorinating agent, Nd in the magnets will be chlorinated to NdCl3 by reacting with ZnCl2 at 1000 K, as shown in Eq. (1). In addition, Dy and Pr in the magnets also will be chlorinated to dysprosium chloride (DyCl3) and praseodymium chloride (PrCl3) by reacting with ZnCl2 at 1000 K, as shown in Eqs. (2) and (3), respectively. On the other hand, the chemical equilibrium point for FeClx (x = 2,3, l,g), ZnCl2 (l,g), and Zn (l) does not exist, as shown in Fig. 2b. As a result, even though a sufficient amount of ZnCl2 is used for the reactions, Fe in magnets will not be chlorinated

$$\begin{gathered} {{2 \text{Nd} }}(s,\, {{ \text{in Nd-Fe-B magnet})}} \, + {{ 3 \text{ZnCl}}}_{{2}} \left( {l,g} \right) \, = {{ 2 \text{NdCl}}}_{{3}} \left( s \right) \, + {{ 3 \text{Zn} }}\left( l \right) \hfill \\ \Delta G^\circ_{{{r}}} = - {\text{785.0 kJ at 1000 K}} \,\left[ {{12}} \right] \hfill \\ \end{gathered}$$
(1)
$$\begin{gathered} {{2 \text{Dy} }}( {s, {{ \text{in Nd-Fe-B magnet})}}} \, + {{ 3 \text{ZnCl}}}_{{2}} \left( {l,g} \right) \, = {{ 2 \text{DyCl}}}_{{3}} \left( l \right) \, + {{ 3 \text{Zn} }}\left( l \right) \hfill \\ \Delta G^\circ_{{{r}}} = - { \text {676.0 kJ at 1000K}} \,\left[ {{12}} \right] \hfill \\ \end{gathered}$$
(2)
$$\begin{gathered} {{2 Pr }} \, ( {s,\text{in Nd-Fe-B magnet})} \, + {{ 3 \text{ZnCl}}}_{{2}} \left( {l,g} \right) \, = {{ 2 \text{PrCl}}}_{{3}} \left( s \right) \, + {{ 3 \text{Zn} }}\left( l \right) \hfill \\ \Delta G^\circ_{{{r}}} = - \text{811.6 kJ at 1000 K} \,\left[ {{12}} \right] \hfill \\ \end{gathered}$$
(3)

Experimental

Figure 3 shows a schematic of the experimental apparatus, and Table 1 lists the experimental of the selective chlorination process using ZnCl2 at 1000 K for 1.5 h. A mixture of Nd-Fe-B magnet powder and anhydrous ZnCl2 placed in an alumina (Al2O3) crucibles was positioned in a quartz holder inside a glove box. Afterward, the quartz holder was removed from the glove box and placed at the bottom of a gas-tight quartz reactor. The gas inside the reactor was replaced with Ar gas through evacuation and filling of Ar gas. The pressure inside the reactor was maintained at 1 atm during the experiments by flowing Ar gas. After the completion of atmosphere control, the reactor was placed in an electric furnace, which was preheated to 1000 K. When the reaction time was finished, the reactor was immediately removed from the furnace and cooled to room temperature. The samples obtained after the selective chlorination reaction were subjected to deionized (DI) water leaching at room temperature for 0.5 h, and the residues obtained after the leaching were dried at 343 K for 0.5 h.

Fig. 3
figure 3

Schematic of the experimental apparatus (Reprinted by permission from Springer Nature: Springer, Journal of Sustainable Metallurgy, Selective Chlorination of Rare Earth Elements from a Nd-Fe-B Magnet Using Zinc Chloride, Kyung-Hwan Lim et al. 2021)

Full size image
Table 1 Experimental conditions of the selective chlorination process using ZnCl2 at 1000 K for 1.5 h
Full size table

Results and Discussion

Table 2 lists the composition of the Nd magnet used as feedstock. In addition, Fig. 4 shows the results of XRD analysis of the residues obtained after the selective chlorination reactions without water leaching. As expected from the thermodynamic analysis, NdCl3 and Fe (Zn) were identified after the experiments. DyCl3 and PrCl3 were not identified because the feedstock contains low concentrations of these elements, as shown in Table 2.

Table 2 Composition of the Nd-Fe-B magnet used as feedstock
Full size table

Further, Fig. 5 shows the influence of the particle size of the feedstock on the chlorination efficiencies of Nd, Dy, and Pr after the experiments. When the particle size of the feedstock used for selective chlorination using ZnCl2 increased from <75 μm to 300–600 μm, the chlorination efficiencies of Nd, Dy, and Pr increased from 93.8%, 36.3%, and 96.0% to 96.5%, 57.2%, and 97.6%, respectively. In addition, the total chlorination efficiency of Nd, Dy, and Pr was 92.1% when the particle size of the feedstock was 300–600 μm. As the particle size decreased, the chlorination efficiencies of REEs decreased because the concentration of oxygen in the feedstock increased from 0.188 mass% at 300–600 μm to 0.467 mass% at <75 μm. REEs in the magnet powder feedstock easily react with oxygen during processing because these elements are reactive. However, the oxides of REEs could not be chlorinated to rare earth chlorides by reacting with ZnCl2 at 1000 K. For example, when the oxides of Nd and Dy react with ZnCl2 at 1000 K, neodymium oxychloride (NdOCl) and dysprosium oxychloride (DyOCl) are produced, as shown in Eqs. (4) and (5), respectively [12, 15]. As a result, the chlorination efficiencies of REEs increased when the particle size of the feedstock was increased.

Fig. 4
figure 4

Results of XRD analysis of the residues obtained after the selective chlorination reactions (Reprinted by permission from Springer Nature: Springer, Journal of Sustainable Metallurgy, Selective Chlorination of Rare Earth Elements from a Nd-Fe-B Magnet Using Zinc Chloride, (Kyung-Hwan Lim et al. 2021)

Full size image
Fig. 5
figure 5

Influence of particle size of the feedstock on the chlorination efficiencies of Nd, Dy, and Pr

Full size image
$$\begin{gathered} {{\text{Nd}}}_{{2}} {{\text O}}_{{3}} \left( s \right) \, + {{\text{ZnCl}}}_{{2}} \left( {l,g} \right) \, = {{ 2 \text{NdOCl}}}\left( s \right) \, + {{ \text{ZnO}}}\left( s \right) \hfill \\ \Delta G^\circ_{{{r}}} = - \text{81.3 kJ at 1000K} \,\left[ {{12}} \right] \hfill \\ \end{gathered}$$
(4)
$$\begin{gathered} {{\text{Dy}}}_{{2}} {{\text O}}_{{3}} \left( s \right) \, + {{ \text{ZnCl}}}_{{2}} \left( {l,g} \right) \, = {{ 2 \text{DyOCl}}}\left( s \right) \, + {{ \text{ZnO}}}\left( s \right) \hfill \\ \Delta G^\circ_{{{r}}} = - \text{4.2 kJ at 1000K} \,\left[ {{12},{15}} \right] \hfill \\ \end{gathered}$$
(5)

Conclusions

The selective chlorination process using ZnCl2 to recover Nd, Dy, and Pr from Nd magnets was investigated in this study. When the Nd magnet reacted with ZnCl2 at 1000 K, Nd in the feedstock chlorinated to NdCl3; however, Fe in the feedstock did not. In addition, the chlorination efficiencies of Nd, Dy, and Pr reached 96.5%, 57.2%, and 97.6%, respectively, when the particle size of the feedstock used for the chlorination reaction at 1000 K for 1.5 h was 300–600 μm. As a result, 92.1% of REEs was recovered from the Nd magnets through the selective chlorination process.