Selective Dissolution of Nd₂O₃ from the Mixture with Fe₂O₃ and Ga₂O₃ by Using Inorganic Acid Solutions Containing Ethylene Glycol

Abstract

Rare earth elements (REEs) are strategically critical in the manufacture of advanced materials. Red mud and end-of-life NdFeB magnets can be good secondary sources for REEs, but recovery is difficult due to the high iron oxide content and small amount of REEs. Oxide mixtures whose composition of Fe, Nd, and Ga was similar to that in red mud were employed in experiments. In this study, a relatively inexpensive non-aqueous system was used to selectively dissolve Nd₂O₃ in a mixture with Fe₂O₃ and Ga₂O₃. The addition of ethylene glycol (EG) to HCl and H₂SO₄ solution depressed the dissolution of Fe₂O₃ and Ga₂O₃ from the mixtures, and thus selective dissolution of Nd₂O₃ was possible. The optimum conditions were as follows: (a) 1.0 M HCl in EG, 25 °C ± 1 °C, 50 g/L pulp density, 120 min, 200 rpm; and (b) 0.05 M H₂SO₄ in EG, 25 °C ± 1 °C, 50 g/L pulp density, 60 min, 300 rpm. Under these conditions, Nd₂O₃ was completely dissolved, whereas no Fe₂O₃ or Ga₂O₃ was dissolved by the H₂SO₄ system, and the dissolution percentage of these two oxides by the HCl system was less than 1%. Due to the selective dissolution of Nd₂O₃ from the oxide mixtures, it is simple to recover Nd. An efficient process can be developed for the recovery of REEs from red mud and end-of-life NdFeB magnets by applying our results.

1. Introduction

The rare-earth elements (REEs) consist of seventeen metallic elements, namely, scandium, yttrium, and fifteen elements of the lanthanide group [1]. The REEs are critical in manufacturing advanced materials such as fuel cells, mobile phones, displays, hi-capacity batteries, permanent magnets for wind power generation, and green energy devices [2]. REEs have become important raw materials with the highest supply risk due to the increasing consumption of REEs resulting from the growing demand for green technology [3]. However, due to the dominance in the REEs market of China, which accounts for 90% of the world supply of REEs, the global supply of REEs is dependent on Chinese exports of REEs [4]. In order to address the supply challenge of REEs, rare earth recovery from secondary resources such as bauxite residue (as red mud) and end-of-life NdFeB magnets can be considered to be one of the solutions.

REEs in bauxite ores exist as oxides that are adsorbed on the surface of the minerals and are found in red mud after the Bayer process of bauxite ore [5]. Red mud contains a large number of valuable metals including rare earth elements, making it a potential and important resource for these metals in recent times [6].

Table 1. Processes for the recovery of Ga and REE from red mud and NdFeB magnets.

Samples	Processing	Performance	Ref.
Red mud	(1) Leaching in HCl (0.5–6 M; 1:4–50; 25–70 °C; 90–1440 min). (2) Leaching in HNO3 (0.1–0.6 M; 1:10–50; 60–1440 min). (3) Leaching in H2SO4 (0.5–5 M, 20%; 1:10–100; 25–100 °C; 30–1440 min).	- Recovery of about 90% Y, 70% heavy lanthanides (Dy, Er, Yb), 50% middle lanthanides (Nd, Sm, Eu, Gd), and 30% light lanthanides (La, Ce, Pr).	[7]
Greek red mud	- Leaching in HCl, H2SO4, HNO3 (0.5–6 N, T: 25–90 °C, t: 5 min to 24 h, L/S: 50).	- Dissolve about 80% of REEs, 60 % of iron, 30% Sc, and 50%Ti.	[8]
Red mud	(1) Roasting around 700 °C for 1 h with sulfuric acid to bauxite residue mass ratio of 1:1. (2) Leaching for 7 days without agitation or for 2 days with agitation at room temperature.	- Extract 60 wt% of scandium and >80 wt% of other REEs. - Concentration of REEs and other elements in the leach solution (ppm): Y (13), La (19), Ce (62), Nd (17), Dy (3), Sc (15), Na (3859), Al (4638), Ca (441), Ti (82), Fe (450).	[9]
Red mud	(1) Alkali roasting of bauxite residue at 950 °C for 4 h with sodium carbonate followed by water leaching at 80 °C for 60 min can remove about 75 wt% of alumina. (2) Smelter sample at 1500 °C (3) The slag was leached with acids (HCl, H2SO4, or HNO3) at 25 °C and 90 °C.	- Remove about 75 wt% of alumina. - Remove more than 98 % of iron. - Dissolution of about 80% of Sc. - Recover more than 80% of Ti and REEs.	[10]
Indian red mud	(1) Leaching the red mud with 3 M H2SO4 at ambient temperature, S/L = 10 g/L and 200 rpm in 1 h. (2) Leaching in 3 M H2SO4 at 75 °C and S/L = 10 g/L in 1 h time. (3) Extraction of lanthanum, cerium and scandium from the leach liquor by Cyanex 301.	- Extract 99.9% of lanthanum. - Extract 99.9% of cerium. - Recovery of 37% lanthanum at 75 °C, and 44% cerium at ambient temperature (35 °C).	[11]
Red mud	- Dry digestion of bauxite residue with HCl or H2SO4, followed by water leaching.	- Recovery of about 40% of scandium and 25% iron. - Concentration of REEs in the leachate was approximately 6–8 mg L−1 and up to 20 mg L−1 with multi-stage circulation.	[12]
Red mud	(1) Leaching in HCl (159 g/L, L/S: 8 mL/g, T: 55 °C, t: 5 h) (2) Ion exchange by LSD-396 resin under conditions of 45 ± 2 °C, resin dosage of 0.6 g/mL and 2 h.	- Leaching rate of 94.77% Ga and 3.91 mg/L Ga3+ in leachate. - Adsorption rate of 59.84% and desorption rate of 95.32% for Ga.	[13]
Red mud	- Leaching conditions: 130% HCl dosage, L/S = 4 mL/g, T: 75 °C, t: 3 h.	- Leaching efficiencies of metals: Fe (95.9%), Al (82.1%), Ti (68.3%), Sc (93.3%), La (82.3%), Ce (96.9%), Nd (98.3%), and Y (95.6%)	[14]
NdFeB magnet powder	(1) Crushing and milling of the magnet into coarse powder. (2) Roasting to transform the metals into the corresponding oxides (3) The selective leaching with acids (HCl, HNO3). (4) Extracting with the ionic liquid trihexyl(tetradecyl)phosphonium chloride. (5) Precipitating the REEs by the addition of oxalic acid. (6) Calcining the rare-earth oxalates to rare-earth oxides.	- The purity of the REEs in the leachate was higher than 90%	[16]
NdFeB magnets	(1) Grinding conditions: molar ratio of NaOH:Nd = 15:1, 550 rpm, 5 cm3 in water addition, 30 min. (2) Roasting at 400 °C and 2 h. (3) Leaching in 1 kmol·m−3 acetic acid, at 400 rpm, and 90 °C with 1% pulp density.	- Dissolution 94.2% Nd, 93.1% Dy, 1.0% Fe	[17]
NdFeB magnets	(1) The samples were crushed, milled and roasted above 100 °C (2) Leaching with the ionic liquid [Hbet][Tf2N] at 175 °C and at atmospheric pressure.	- Oxidized magnets to oxidize NdFeO3, Nd2O3, Fe2O3 - Recover neodymium, dysprosium, and cobalt.	[18]
NdFeB scrap	(1) Roasting NdFeB scrap at 800 °C (2) Leaching in 10 wt% HCl at 110 °C for 30 min.	- Oxidized to form Fe2O3, NdFeO3, and NdBO3. - Leaching rate of rare earth was 96.27% along with 13.33% Fe	[19]

lists some investigated processes for the recovery of REEs from red mud and end-of-life NdFeB magnets [7,8,9,10,11,12,13,14,15,16,17,18,19]. In most processes, inorganic acid solutions such as HCl, HNO₃, and H₂SO₄ as leaching agents are employed to effectively dissolve iron oxides from red mud [7,8,9]. However, a high concentration of iron in the leaching solutions poses some problems in the recovery of REEs and other valuable metals. Since the weight percentage of REEs in red mud is very low, the recovery of REEs from red mud consists of very complicated processes, such as alkaline roasting, smelting, and then acid leaching of the slag, which result in a very low recovery yield of REEs (<5%) [10].

Another resource for REEs recovery is end-of-life products such as NdFeB magnets. Depending on the application, REEs can make up about 31–32% by weight of NdFeB magnets; these REEs are mainly Nd and Pr, along with small amounts of Dy, Tb, and Gd [15]. Many studies indicate that the roasting treatment of NdFeB magnet powders before dissolution has a favorable effect on the dissolution of REEs from magnets [16]. Some processes to recover the REEs from these magnets, including acid or/and alkali leaching, solvent extraction, ion exchange, or ionic liquid techniques, are shown in Table 1. The crushed magnets are roasted at optimum conditions to obtain Nd₂O₃ and Fe₂O₃ instead of the complex oxide NdFeO₃, which dissolves much more slowly than pure oxide. The leaching of magnet oxides after roasting treatment with ionic liquid [Hbet] [Tf₂N], or/and HCl at high pressure results in a high leaching percentage of both REEs and iron oxides (Table 1) [18].

Eh-pH diagrams of Fe(III), Ga(III), and Nd(III) at 25 °C clearly indicate that there is some difference in the precipitation pH of these metal ions as hydroxides [20,21]. Therefore, it is possible to selectively dissolve Nd₂O₃ over Fe₂O₃ and Ga₂O₃ from the oxide mixtures of the three metals. From the recovery processes presented in Table 1, many methods exist to treat red mud and NdFeB by combining multiple methods including hydrometallurgy and pyrometallurgy. However, the amount of Nd₂O₃ in the red mud is very small and thus it is better to selectively dissolve Nd₂O₃ to yield the highest Nd recovery. In this work, the selective dissolution of Nd₂O₃ from the mixture with Fe₂O₃ and Ga₂O₃ was investigated using inorganic acid solutions containing ethylene glycol. Fe and Nd are two components that both exist in red mud and NdFeB magnets; thus, the oxides, including Fe₂O₃, Nd₂O₃, and Ga₂O₃, were selected as a mixture to simulate the metal composition of interest to conduct selective dissolution studies for REE recovery. Because the composition of the red mud is very complex, with different concentrations of metals and rare earth elements depending on the ore source, it is difficult to choose an exact value for the oxide content to prepare the oxide mixture for the experiment [22]. Based on typical deviations of mineral compositions in

Table 2. The composition of mineral components in the red mud [6,22] and in the oxide mixture in this study.

Mineral Component	Red Mud	Oxide Mixture
Fe2O3	4–55%	98.36%
TiO2	2–17%	-
Al2O3	6–27%	-
SiO2	3–24%	-
Na2O	0–10%	-
CaO	0–40%	-
Rare Earth Elements	500–1700 ppm	-
- La	0–416 ppm
- Ce	0–842 ppm
- Pr	0–95 ppm
- Nd	0–341 ppm
- Sm	0–64 ppm
- Gd	0–56 ppm
- Tb	0–184 ppm
- Dy	0–48 ppm
- Ho	0–25 ppm
- Er	0–28 ppm
- Yb	0–28 ppm
- Y	0–373 ppm
- Sc	0–158 ppm
- Ga	0–570 ppm
Nd2O3	-	1.15%
Ga2O3	-	0.49%

, oxide mixtures with Fe, Nd, and Ga compositions similar to those in red mud were used in the experiments. In general, the dissolution of ionic and covalent solids depends on the dielectric constant of the solution. As the dielectric constant of the solution decreases, the dissolution of solid oxides becomes difficult. By utilizing this relation, ethylene glycol (EG) was employed as a diluent to investigate the effect of a change in the dielectric constant of the solution on the dissolution of the three oxides. EG is a green solvent because it is nonvolatile (boiling point 197.3 °C) and water soluble (polar) [23]. Therefore, EG can be considered to be an efficient substitute for water in leaching experiments. For this purpose, the dissolution behavior of the three oxides was investigated using HCl and H₂SO₄ solutions containing EG. HNO₃ should not be used because it can form explosives such as ethylene glycol dinitrate when mixed with EG [24]. However, a non-aqueous leaching system using EG as a diluent makes the resulting solution more viscous than using H₂O, making it difficult to filter to remove impurities. The most important results were that Nd₂O₃ can be selectively dissolved by adjusting dissolution conditions such as HCl and H₂SO₄ concentration, temperature and time, stirring speed, and pulp density.

2. Materials and Methods

2.1. Materials

Powders of Nd₂O₃ (Strem Chemicals, Newburyport, United States, 99.99%), Ga₂O₃ (Alfa Aesar, Ward Hill, MA, USA, 99.99%), and Fe₂O₃ (Daejung Co., Shiheung, Korea, 99.99%) were mixed at a predetermined weight ratio, and the composition of the oxides in the mixtures is shown in Table 2 [6,22]. Inorganic acid solutions of HCl (Daejung Co., Shiheung, Korea, 35%), and H₂SO₄ (Daejung Co., Shiheung, Korea, 98%) were diluted with ethylene glycol (Daejung Co., Shiheung, Korea, >99.0%) to the desired concentrations. Standard solutions of gallium (Kanto Chemical Co., Inc., Tokyo, Japan), neodymium (Accu Standard, Inc., New Haven, CT, USA), and iron (Kriat Co., Ltd., Daejeon, Korea) were employed in ICP analysis.

2.2. Experimental Procedure and Analytical Methods

All leaching tests were run in a 250 cm³ three-neck round-bottom flask with the determined volume of concentrated HCl and H₂SO₄ in EG. A magnetic stirrer (WiseStir MSH-20D, Daihan Scientific Co., Seoul, Korea) was used to control the stirring speed, time, and temperature throughout the leaching experiments. The powder mixtures of Nd₂O₃, Ga₂O₃, and Fe₂O₃ were put into the reaction flask once the desired temperatures were reached. After leaching experiments, the solution was filtered using filter paper and diluted with doubly distilled water for inductively coupled plasma optical emission spectrometry (ICP-OES) analysis (Spectro Arcos, Cleve, Germany).

The dissolution percentage (%) of the oxides was calculated as:

$$\mathrm{Dissolution}\mathrm{percentage}(\%)=\frac{{\mathrm{m}}_{\mathrm{M}}^{*}}{{\mathrm{m}}_{\mathrm{M}}}\times 100\%$$

(1)

where m^*_M and m_M are the mass of metal ions in the leaching solution and the powder mixtures of oxides before the leaching experiments, respectively. Most of the leaching experiments were repeated two times and the error related to dissolution percentage was within ±5%.

3. Results and Discussion

3.1. Effect of Acid Concentration

The metallic oxides employed in this work are basic and thus dissolve in hydrochloric and sulfuric acid solutions according to the following reactions:

Nd₂O_3(s) + 6H⁺ = 2Nd³⁺ + 3H₂O_(l)

(2)

Fe₂O_3(s) + 6H⁺ = 2Fe³⁺ + 3H₂O_(l)

(3)

Ga₂O_3(s) + 6H⁺ = 2Ga³⁺ + 3H₂O_(l)

(4)

In order to investigate the influence of acid concentration and the role of EG, the concentration of HCl was varied from 0.5 to 3.0 M using either H₂O or EG as a diluent, respectively. When concentrated HCl solution was diluted with EG to prepare the leaching solution, some of the water was inevitably introduced. However, the water content in the leaching solution containing EG was less than 50%, indicating that our system is a kind of non-aqueous system [23]. In these experiments, the conditions of other variables were controlled as follows: pulp density of 50 g/L, 80 °C, 90 min with stirring speed of 400 rpm. Figure 1 shows that there is a large difference in the dissolution behavior among the three oxides. Nd₂O₃ oxide was completely dissolved in the HCl concentration ranges irrespective of the nature of the diluent. Furthermore, the dissolution of Ga₂O₃ and Fe₂O₃ oxides was proportional to the HCl concentration. Moreover, the dissolution percentage of Ga₂O₃ was higher than that of Fe₂O₃. It is noticeable that the dissolution percentage of Ga₂O₃ and Fe₂O₃ oxides from the H₂O solution was higher than that from the EG solution. The most important results of Figure 1 are that Nd₂O₃ can be selectively dissolved by adjusting the HCl concentration and the use of EG as a diluent depressed the dissolution of the Ga₂O₃ and Fe₂O₃ oxides, which would be favorable for the selective dissolution of Nd₂O₃ from the oxide mixtures. A concentration of 1.0 M HCl in EG is the optimal condition to dissolve 100% Nd₂O₃, together with 16% Fe₂O₃ and 24% Ga₂O₃.

According to Figure 1, the low acid concentration is favorable for the selective dissolution of Nd₂O₃ from the oxide mixtures. Therefore, the concentration of H₂SO₄ varied from 0.01 to 0.3 M and the conditions of other variables were the same as those used in the experiments with HCl reported in Figure 1. Concentrated H₂SO₄ solutions diluted with either H₂O or EG were used to investigate the effect of H₂SO₄ concentration on the leaching of the oxide mixtures. When H₂SO₄ concentration was higher than 0.05 M, Nd₂O₃ was completely dissolved from the oxide mixtures irrespective of the nature of the diluent (see Figure 2). In the case of Ga₂O₃ and Fe₂O₃, the dissolution behavior of these two oxides was similar to that of the HCl solution; namely, the dissolution percentage of Ga₂O₃ was higher than that of Fe₂O_3, and the use of EG as a diluent reduced the dissolution percentage of these two oxides. Moreover, when H₂SO₄ concentration was 0.05 M, the dissolution of Ga₂O₃ and Fe₂O₃ was negligible, while most of Nd₂O₃ was dissolved. In particular, Ga₂O₃ and Fe₂O₃ were not dissolved by the H₂SO₄ solution in EG. Due to the simultaneous dissolution of all three oxides, dissolving 100% of Nd₂O₃ required a concentration of HCl/EG acid of 1.0 M, which was higher than the concentration of H₂SO₄/EG of 0.05 M.

Dissolution of metal oxides into solution increases the entropy of the dissolution reaction and thus it is necessary to think about the enthalpy change accompanying the dissolution in order to explain the dissolution of metal oxides. In solution, the dissolution of metal oxides is related to the lattice energy of the oxides and the solvation enthalpy of the resulting cations and oxide ions. In general, the lattice energy of metal oxides has a positive value, whereas the solvation energy of the resulting ions by the solvent is exothermic [25].

Table 3. Ionic radius and solvation enthalpy of ions, lattice energy, and heat of solution of the oxides [25,26].

Ionic	Ionic Radius, pm	Solvation Enthalpy in Water, kJ/mol	Solvation Enthalpy in EG, kJ/mol	Oxides	Lattice Energy (Calculated), kJ/mol	Heat of Solution in Water, kJ/mol	Heat of Solution in EG, kJ/mol
Nd3+	98.3	−1910	−1887	Nd2O3	12736	6234	6310
Fe3+	64.5	−2911	−2877	Fe2O3	14309	5806	5905
Ga3+	62	−3028	−2992	Ga2O3	15590	6852	6954
O2-	140	−894	−884	-	-	-	-

shows that Ga₂O₃ has the highest lattice energy, whereas the lattice energy of Nd₂O₃ is the lowest among the three oxides [25,26]. Solvation enthalpy of ions is related to the ionic radius and charge, and the dielectric constant of the solvent, as represented in Equation (5) [25]:

$$\mathrm{H}=-\frac{{\mathrm{Ze}}^2}{2\mathrm{r}}\left(1-\frac{1}{\mathsf{\epsilon }}\right)$$

(5)

where Z is the charge of the ion, r is its radius, and $\mathsf{\epsilon }$ is the dielectric constant of the solvent ($\mathsf{\epsilon }$ = 78.2 for water and $\mathsf{\epsilon }$ = 41 for EG at 25 °C) [27,28].

The ionic radius and the corresponding solvation enthalpies of the three ions in water and EG, which were calculated by Equation (5), are represented in Table 3. From the lattice energies of the oxides and the solvation enthalpies of the ions, the heat of solution ($\Delta {\mathrm{H}}_{\mathrm{s}}$) of each oxide in H₂O and EG can be obtained by using the following equation [26]:

$$\Delta {\mathrm{H}}_{\mathrm{s}}=\mathrm{U}+2{\mathrm{H}}_{+}+3{\mathrm{H}}_{-}$$

(6)

where H₊ and H₋ are solvation enthalpies of the cations and oxide ion.

The heat of solution of the oxides in H₂O and EG calculated by Equation (6) is listed in Table 3. Although the heat of solution of Fe₂O₃ is the lowest and that of Ga₂O₃ is the highest, the difference is within 1045 kJ/mol. Therefore, the selective dissolution of Nd₂O₃ in low concentrations of HCl and H₂SO₄ may be ascribed to other factors. Since the heat of solution of the oxides in EG is slightly larger than that of water, the reduction in the dissolution percentage of the oxides when EG was employed as a diluent can be explained.

Table 4. The polarizabilities of ions [29].

Ion	αD, Å3
Nd3+	5.01
Fe3+	2.28
Ga3+	1.50
O2-	2.00

shows that Nd(III) has the highest polarizability, whereas the polarizability of Ga(III) is the lowest among the three ions [29]. In general, the ions with high polarizability have a tendency to be dissolved in a medium with a lower dielectric constant. Therefore, Nd(III) has the strongest tendency to be dissolved when EG is employed as a diluent. By contrast, the difference in the polarizability of Fe(III) and Ga(III) is only 0.78, and thus the dissolution behavior of Fe₂O₃ and Ga₂O₃ may be similar in our experiments.

3.2. Effect of Reaction Temperature

According to the previous section’s results, 1.0 M HCl in EG and 0.05 M H₂SO₄ in EG were the optimum conditions to selectively dissolve Nd₂O₃ from the oxide mixtures. Since the heat of solution of the oxides is positive, the dissolution percentage of the oxides increases with reaction temperature. The effect of temperature on the dissolution of the oxides was investigated over a temperature range of 25–100 °C at an acid concentration of 1.0 M HCl in EG and 0.05 M H₂SO₄ in EG, a stirring speed of 400 rpm for 90 min, and a pulp density of 50 g/L. Figure 3 shows that the dissolution percentage of Nd₂O₃ by 1.0 M HCl in EG increased from 85.5% to completeness as the reaction temperature increased from 25 and 100 °C. The dissolution percentage of Fe₂O₃ and Ga₂O₃ was 0.4% at 25 °C and increased sharply when the temperature was higher than 50 °C. When the reaction temperature reached 100 °C, 26.4% of Fe₂O₃ and 48.3% of Ga₂O₃ were dissolved by HCl in EG. The increase in dissolution percentage of the oxides with reaction temperature agreed well with the endothermic nature of the dissolution of the oxides in EG, as represented in Table 3. Figure 3 shows that reaction temperature is important in the selective dissolution of Nd₂O₃ from the oxide mixtures.

Figure 4 shows the variation in the dissolution percentage of the oxides by 0.05 M H₂SO₄ in EG when the temperature was varied from 25 to 10 °C. Unlike the data obtained from 1.0 M HCl in EG, only Nd₂O₃ was dissolved from the oxide mixtures, and thus it is possible to selectively recover Nd(III) from the oxide mixtures using 0.05 M H₂SO₄ in EG. In the case of dissolution by HCl, the lower temperature was effective in the selective dissolution of Nd₂O₃ but the effect of temperature was negligible in dissolution with H₂SO₄ in EG.

3.3. Effect of Reaction Time

The effect of time on the dissolution by both HCl and H₂SO₄ solutions in EG was investigated by varying reaction times at room temperature. In the experiments, pulp density and stirring speed were controlled at 50 g/L and 400 rpm, respectively. Figure 5 shows the data for dissolution by 1.0 M HCl in EG. Nd₂O₃ was completely dissolved after 120 min, whereas the dissolution percentage of Fe₂O₃ and Ga₂O₃ at this reaction time was 0.5% and 1.4%, respectively. Dissolution experiments with 0.05 M H₂SO₄ in EG were carried out in the reaction time range from 15 to 90 min at room temperature. Reaction time did not affect the dissolution of Ga₂O₃ and Fe₂O₃, whereas the dissolution of Nd₂O₃ was affected by reaction time. Figure 5 and Figure 6 clearly show that the selective dissolution of Nd₂O₃ from the oxide mixtures by HCl and H₂SO₄ is not related to the kinetics.

3.4. Effect of Pulp Density

Pulp density is usually expressed as the ratio of the mass of the solid reactant to the volume of the leaching agent (g/L). In general, the use of larger amounts of leaching solution increases the dissolution performance [30,31]. Pulp density was varied from 50 to 150 g/L at room temperature and at a stirring speed of 400 rpm. In experiments with 1.0 M HCl in EG and 0.05 M H₂SO₄ in EG, reaction times were fixed at 120 and 60 min, respectively. Figure 7 and Figure 8 show the same trends in both acids, namely, that lower pulp density enhances the selective dissolution of Nd₂O₃, whereas dissolution percentage decreases at higher pulp density. Nd₂O₃ was completely dissolved at a pulp density of 50 g/L but decreased to 61% as the pulp density increased to 150 g/L (Figure 7). In these experiments, the dissolution percentage of Fe₂O₃ and Ga₂O₃ was only 0.4% irrespective of the pulp density (Figure 7). In the H₂SO₄/EG system, the dissolution of Nd₂O₃ decreased from 100% to 87% when the pulp density increased from 50 to 150 g/L, and there was no dissolution of Fe₂O₃ and Ga₂O₃ (Figure 8). The lower pulp density was beneficial in the selective and complete dissolution of Nd₂O₃. Therefore, it can be said that the optimum pulp density for the complete dissolution of Nd₂O₃ using HCl and H₂SO₄ solution was 50 g/L.

3.5. Effect of Stirring Speed

When the reaction temperature is low, the chemical reaction is more important than the mass transfer of the reactant to the reaction interface. In order to check the importance of mass transfer in the dissolution of the oxides, the stirring speed was changed from 100 to 400 rpm for both HCl and H₂SO₄ solutions at room temperature. In these experiments, the pulp density was fixed at 50 g/L and the concentration of HCl and H₂SO₄ in EG was controlled to 1.0 and 0.05 M, respectively. Figure 9 and Figure 10 show that stirring speed affected little the dissolution of Nd₂O₃ from both HCl and H₂SO₄ solution when stirring speed was higher than 200 rpm. Therefore, it might be said that the chemical reaction is more important than the mass transfer of the leaching solution during the dissolution reaction of Nd₂O₃ from the oxide mixtures by HCl and H₂SO₄ solutions at room temperature.

3.6. Effect of Solution Acidity

When metal oxides are dissolved by inorganic solutions such as HCl and H₂SO₄, the stability of the dissolved ions is related to the acidity of the solution and the nature of the ligands. Since metal ions are Lewis acids, they have a strong tendency to form complexes with ligands. First, the metal ions can be solvated with water and then hydroxides of the metal would be precipitated. In an aqueous solution, the Eh–pH diagram shows the pH at which the hydroxides of metal begin to precipitate thermodynamically.

Table 5. Solubility products of the hydroxides at 25 °C [31].

Formula	Ksp
Nd(OH)3	3.2 × 10−22
Fe(OH)3	2.79 × 10−39
Ga(OH)3	7.28 × 10−36

shows the solubility product of the three hydroxides [31]. From these values, the precipitation pH of the hydroxides was calculated on the assumption that the concentration of the ions was 0.01 M, and the calculated results are listed in

Table 6. The precipitation pH of the metal hydroxides when the concentration of metal ions is 0.01 M [29].

Ion	Concentration, M	Precipitation pH
Nd3+	0.01	6.5
Fe3+	0.01	1.8
Ga3+	0.01	3.0

. Fe(III) and Ga(III) would be precipitated when the solution pH is around 3, whereas Nd(III) would be precipitated at pH 6.5 (Table 6). Second, the metal ions can form complexes with the anions of inorganic acids, which act as Lewis bases. The complex formation constants of the ions with chloride and sulfate ions are listed in

Table 7. Complex formation constants of the ions with sulfate and chloride ions [25,32,33,34,35].

Ligands	Ion	log K1	log K2	log K3	log K4
Sulfate	Nd3+	3.64	-	-	-
Sulfate	Fe3+	2.03	2.98	-	-
Sulfate	Ga3+	2.77	2.77	-	-
Chloride	Nd3+	0.32	0.02	−0.34	−0.74
Chloride	Fe3+	1.48	2.13	1.99	0.01
Chloride	Ga3+	0.01	-	-	-

[25,32,33,34,35]. It is well known that Fe(III) has a strong tendency to form complexes with chloride ions so that the anionic complex, FeCl₄⁻, exists in concentrated HCl solution. Table 7 indicates that sulfate ions have a stronger tendency to form complexes with the metal ions, with the exception of Fe(III) in HCl solution.

Considering the acidity of the 1 M HCl solution employed in the leaching, it can be said that no precipitates of Fe(III) and Ga(III) can be formed in our experimental conditions. However, Fe(III) and Ga(III) can form hydroxides, and then these hydroxides would be precipitated during the leaching with 0.05 M H₂SO₄ in EG. In order to check the formation of Fe(III) hydroxides during the leaching, the XRD data of the leaching residues by HCl and H₂SO₄ solutions are represented in Figure 11 and Figure 12. In both cases, most of the iron existed as Fe₂O₃ and the existence of Fe(III) hydroxides was difficult to detect. Therefore, it can be said that the precipitation of the dissolved Fe(III) and Ga(III) did not occur in our experimental conditions. Although the heat of solution of the oxides cannot explain the selective dissolution of Nd₂O₃ from the oxide mixtures, the combined effect of acid concentration and reaction temperature, together with the highest polarizability of Nd(III), can explain our results.

4. Conclusions

In this work, the dissolution behavior of the oxide mixtures of Nd₂O₃, Fe₂O₃, and Ga₂O₃ was investigated by employing HCl and H₂SO₄ solutions containing ethylene glycol. The addition of ethylene glycol depressed the dissolution of Fe₂O₃ and Ga₂O₃ from the oxide mixtures, and thus selective dissolution of Nd₂O₃ was possible. High polarizability of Nd(III) together with the combined effect of reaction conditions, such as acid concentration, reaction time, and temperature, resulted in the selective dissolution of Nd₂O₃ from the oxide mixtures. The optimum conditions for the selective dissolution of Nd₂O₃ by either HCl or H₂SO₄ solution were as follows: (a) 1.0 M HCl in EG, at 25 °C ± 1 °C, pulp density = 50 g/L, 120 min, 200 rpm; and (b) 0.05 M H₂SO₄ in EG, at 25 °C ± 1 °C, pulp density = 50 g/L, 60 min, 300 rpm. Under these conditions, the dissolution percentage of Fe₂O₃ and Ga₂O₃ was less than 1% by HCl solution, whereas only Nd₂O₃ was dissolved by the H₂SO₄ system. Since H₂SO₄ has a lower cost than HCl and is used at low concentrations, the H₂SO₄/EG process offers better economic value for the rare earth recovery process, specifically that for Nd. In particular, the selective dissolution of Nd₂O₃ over Fe₂O₃ and Ga₂O₃ by H₂SO₄ in EG led to a solution containing Nd(III) having high purity, which can be easily recovered from the solution. Further studies are needed on the application of these results to the treatment of red mud and end-of-life NdFeB magnets.