Research Status of Electrolytic Preparation of Rare Earth Metals and Alloys in Fluoride Molten Salt System: A Mini Review of China

Abstract

China’s rare earth reserves and consumption are the highest in the world. Rare earth metals and alloys play a pivotal role in the domains of permanent magnetic materials, hydrogen storage materials, luminescent materials, abrasive materials, etc. The molten salt electrolysis process is the most widely used method for producing light rare earth metals and alloys in China, with distinct advantages such as continuous production and short process flow. This article focuses on the process technology of preparing rare earth metals and alloys by electrolyzing rare earth oxides in fluoride systems. This article summarizes the effects of process parameters such as cathode and anode structures, electrolysis temperature, and current density on the direct recovery and current efficiency of the preparation of light rare earth metals (La, Ce, Pr, Nd), RE–Mg (RE for rare earth) alloys, RE–Al alloys, RE–Ni alloys, and other rare earth alloys. Meanwhile, the disadvantages of the electrolytic cells and electrode configurations that are currently used in industrial production are discussed. Accordingly, the future prospects of molten salt electrolysis technology in the preparation of rare earth metals and alloys are clarified.

1. Introduction

Rare earth metals have unique characteristics, such as their distinctive 4f subshell electronic structure, large atomic spin magnetic moment, and strong spin–orbit coupling, which play an important role in permanent magnet materials, hydrogen storage materials, luminescent materials, and abrasive materials. The development of rare earth separation and purification technology has significant strategic importance [1,2,3]. According to USGS data [4,5,6,7,8,9], in 2022, the global reserves of rare earth elements contained approximately 130 million tons. Among them, China had reserves of 44 million tons, Vietnam had reserves of 22 million tons, and Brazil and Russia both had reserves of around 21 million tons, accounting for approximately 34%, 17%, 16%, and 16% of the global rare earth reserves, respectively (as shown in Figure 1a). With the increasing demand for rare earth functional materials, China’s rare earth consumption in the past two decades (Figure 1b) [10,11] has continuously surpassed its previous peak. According to the data on global rare earth mineral production from 2016 to 2021 (Figure 1c [4,5,6,7,8,9]), China’s rare earth mineral production has been increasing year by year. In recent years, countries such as the United States and Russia have also resumed rare earth mining, and their production has been increasing year by year. A total of 41% of China’s rare earth production is used for magnetic functional materials, and the global application of rare earth metals in permanent magnetic materials accounts for about 60%. Based on information from the China Rare Earth Association [12] (Figure 1d), the price trajectory of praseodymium–neodymium alloy from 2020 to the end of 2023 shows a clear similarity to a normal distribution pattern. Early in 2022, the price reached a peak, followed by a fall soon after. Furthermore, the price trend of praseodymium metal is shown to be in harmony with that of holmium–iron metal and other rare earth metals.

The principal classifications of rare earth deposits include fluorocarbon cerium ore, monazite, mixed rare earth ore, and ion adsorption ore. Figure 2 [13,14] illustrates the distribution of major rare earth deposits worldwide. China’s rare earth mineral resources are predominantly situated in two regions: the Bayan Obo rare earth mine in the Inner Mongolia Autonomous Region in the north, and the southern ion adsorption rare earth mine operated by Jiangxi Province. The distribution of rare earth mineral resources exhibits a pattern of “light rare earth in the north and heavy rare earth in the south”. Notably, light rare earth mineral resources constitute a significant proportion, comprising approximately 90.37% of the total rare earth reserves, as depicted in Figure 3. The content of rare earth oxides in the associated rare earth minerals in the Bayan Obo carbonatite-type deposit is only about 5.60%. However, a rare earth concentrate with a rare earth oxide content of more than 60% can be obtained after treatment using the beneficiation enrichment process. The rare earth concentrate is further refined into highly pure isolated rare earth compounds through hydrometallurgical processes such as acid leaching, solvent extraction, and ion exchange. Finally, the rare earth compounds are reduced into rare earth metals or rare earth alloys through pyrometallurgical techniques [15,16].

Rare earth minerals are non-renewable strategic mineral resources. Rare earth metals, often referred to as the “MSG of the industry”, are critical raw materials for high-end and end-use companies. As the price of rare earth metals declines, production profits undergo a precipitous fall. Concurrently, companies face increasing demands for product quality from downstream companies, stagnant sales of low-end products, and market shrinkage. Enterprises are urgently seeking technical solutions that can enhance current efficiency and product quality, while exploring innovative production processes aimed at reducing manufacturing costs. This paper mainly reviews the production process of light rare earths (La, Ce, Pr, Nd) and their alloys prepared by fluoride molten salt electrolysis, which is currently the dominant production method in China. The advantages and disadvantages of existing production processes are first summarized for reference by technical workers engaged in rare earth and alloy production, and then future perspectives are discussed.

2. Comparison of Pyrometallurgical Methods for Rare Earth Metals

Rare earth elements are composed of the 15 lanthanide elements and Sc and Y, which have similar physical and chemical properties to the lanthanide elements. Among them, Pm is a radioactive element and Sc is a typical dispersed metal. Therefore, these two special elements are generally not contained in the industrial process of producing rare earths from rare earth ores. The selection of the method for preparing rare earth elements is determined by their physical and chemical properties. Table 1 shows the fundamental physical properties and main preparation methods of 15 rare earth metals [17,18].

The primary preparation methods of rare earth metals mainly include metal or carbon thermal reduction and molten salt electrolysis. During the production process, calcium, lithium, carbon, and the light rare earth metal lanthanum are often used as reducing agents. Depending on the type of molten salt system, molten salt electrolysis can be categorized into the chloride molten salt system and the fluoride molten salt system. A comparative analysis of these six methods is presented in Table 2.

Heavy rare earth metals such as Gd, Tb, Dy, Ho, Lu, Y, etc. have low boiling points and high melting points. They are mostly reduced by calcium in the form of their fluorides through the calcium thermal reduction reaction. The calcium thermal reduction process generates a significant amount of reduction slag, and typically, there is a residual of 5 wt.% to 18 wt.% of rare earth fluorides in the calcium thermal reduction slag [19]. This poses a challenge for the secondary recovery of rare earth resources.

3 Ca(l) + 2 REF₃(s) = 3 CaF₂(s) + 2 RE(l)

(1)

The lithium thermal reduction method is also adopted to produce isolated heavy rare earth metals, such as Y and high-purity Pr and Nd using their chlorides. The quality of rare earth metals prepared by the lithium thermal reduction method is better than that obtained by the calcium thermal reduction method using fluorides. However, its application is limited due to its higher cost. The reaction equation is shown in Equation (2).

3 Li(l) + RECl₃(s) = 3 LiCl(l) + RE(s)

(2)

Active rare earth metals like lanthanum or other mixed light rare earth metals or carbon are commonly used as reducing agents to directly reduce rare earth oxides in a vacuum induction furnace and then to produce high-purity rare earths with relatively high vapor pressure (Sm, Eu, Tm, Yb) that are, finally, collected in a condenser. The reaction equations are shown in Equations (3) and (4). The reduction–vacuum distillation method allows for the production of high-quality rare earth metal products with high metal yields. However, it has the disadvantages of discontinuous production and relatively high processing costs.

2 La(l) + RE₂O₃(s) = La₂O₃(s) + 2 RE(g)

(3)

7 C(s) + RE₂O₃(s) = 2 REC₂(s) + 3CO(g)
REC₂(s) = RE(s) + 2C(s)

(4)

Table 1. Basic physical properties and main preparation methods of rare earth metals. Adapted from ref. [17,18].

Metal	Melting Point/K	Boiling Point/K	The Temperature at 133.3 Pa Vapor Pressure/K	Evaporation Rate at 133.3 Pa Vapor Pressure/[g/(cm2·h)]	REF3 Melting Point/K	RE2O3 Melting Point/K	Mineral Characteristics Group	Main Preparation Method
La	1193	3743	2490	53	1763	2490	Light rare earth	Molten salt electrolysis
Ce	1071	3743	2420	53	1710	2415
Pr	1204	3403	2241	56	1668	2400
Nd	1283	3303	2032	60	1647	2484
Sm	1345	2173	1237	83	1579	2603		Lanthanum or carbonthermal reduction
Eu	1095	1713	1110	90	1549	2668
Gd	1584	3273	2295	59	1504	2663	Heavy rare earth	Calcium thermal reduction
Tb	1633	3073	2212	60	1445	2663
Dy	1682	2873	1712	71	1427	2664
Ho	1743	2873	1799	69	1416	2673
Er	1795	3173	1879	68	1413	-
Tm	1818	2003	1368	83	1431	2684		Lanthanum or carbonthermal reduction
Yb	1097	1703	924	108	1430	-
Lu	1929	3603	2371	61	1455	-		Calcium thermal reduction
Y	1796	3203	1355	43	1425	-

Note: “-” indicates no data available.

Table 1. Basic physical properties and main preparation methods of rare earth metals. Adapted from ref. [17,18].

Metal	Melting Point/K	Boiling Point/K	The Temperature at 133.3 Pa Vapor Pressure/K	Evaporation Rate at 133.3 Pa Vapor Pressure/[g/(cm2·h)]	REF3 Melting Point/K	RE2O3 Melting Point/K	Mineral Characteristics Group	Main Preparation Method
La	1193	3743	2490	53	1763	2490	Light rare earth	Molten salt electrolysis
Ce	1071	3743	2420	53	1710	2415
Pr	1204	3403	2241	56	1668	2400
Nd	1283	3303	2032	60	1647	2484
Sm	1345	2173	1237	83	1579	2603		Lanthanum or carbonthermal reduction
Eu	1095	1713	1110	90	1549	2668
Gd	1584	3273	2295	59	1504	2663	Heavy rare earth	Calcium thermal reduction
Tb	1633	3073	2212	60	1445	2663
Dy	1682	2873	1712	71	1427	2664
Ho	1743	2873	1799	69	1416	2673
Er	1795	3173	1879	68	1413	-
Tm	1818	2003	1368	83	1431	2684		Lanthanum or carbonthermal reduction
Yb	1097	1703	924	108	1430	-
Lu	1929	3603	2371	61	1455	-		Calcium thermal reduction
Y	1796	3203	1355	43	1425	-

Note: “-” indicates no data available.

Table 2. Comparison of indicators for the preparation of rare earth metals and alloys by the thermal reduction of metals and molten salt electrolysis in different systems. Adapted from ref. [20,21,22,23,24,25,26,27,28,29,30].

Method	Main Raw Materials	Product Components	Production Efficiency	Production Process
Calcium thermal reduction method [21,22]	REF3, Ca	The purity of rare earth metals is between 95.5% and 99.95%.	The yield of rare earth metals is about 95–98%.	Low investment, short process, and a large amount of slag generated from calcium thermal reduction.
Lithium thermal reduction [23]	RECl3, Li	The purity of rare earth metals can reach more than 99.9%.	The yield of rare earth metals reaches more than 95%.	The quality of the prepared rare earth metals is better than that of calcium thermal reduction. Rare earth chlorides are very hygroscopic compounds.
Carbonaceous reduction [24,25]	RE2O3, C	Not indicated	The yield of metals reaches 59.37%	High processing cost and discontinuous production.
Lanthanum thermal reduction [26,27,28]	RE2O3, La	The purity of rare earth metals reaches over 99.9%.	The recovery of rare earth metals reaches over 90%.	High temperature and pressure conditions require high-quality materials and performance of the equipment.
Chloride system molten salt electrolysis method [29,30]	RECl3-MCl(M = Li, K)	Electrolytic rare earth metals with a purity of 95–98%	The yield of rare earth metals is about 85–90%. Single rare earth current efficiency is around 70%.	Generates toxic chlorine gas, low metal recovery.
Fluoride system molten salt electrolysis method [30]	REF3-RE2O3-LiF	Electrolytic rare earth metals with a purity of 95–98%	The yield of rare earth metals is about 90–95%. Current efficiency is about 70–80%.	The current efficiency and rare earth metal yield are high; the electrolysis temperature is high, and the anode produces greenhouse gases such as CO2, CF4, and C2F6.

Table 2. Comparison of indicators for the preparation of rare earth metals and alloys by the thermal reduction of metals and molten salt electrolysis in different systems. Adapted from ref. [20,21,22,23,24,25,26,27,28,29,30].

Method	Main Raw Materials	Product Components	Production Efficiency	Production Process
Calcium thermal reduction method [21,22]	REF3, Ca	The purity of rare earth metals is between 95.5% and 99.95%.	The yield of rare earth metals is about 95–98%.	Low investment, short process, and a large amount of slag generated from calcium thermal reduction.
Lithium thermal reduction [23]	RECl3, Li	The purity of rare earth metals can reach more than 99.9%.	The yield of rare earth metals reaches more than 95%.	The quality of the prepared rare earth metals is better than that of calcium thermal reduction. Rare earth chlorides are very hygroscopic compounds.
Carbonaceous reduction [24,25]	RE2O3, C	Not indicated	The yield of metals reaches 59.37%	High processing cost and discontinuous production.
Lanthanum thermal reduction [26,27,28]	RE2O3, La	The purity of rare earth metals reaches over 99.9%.	The recovery of rare earth metals reaches over 90%.	High temperature and pressure conditions require high-quality materials and performance of the equipment.
Chloride system molten salt electrolysis method [29,30]	RECl3-MCl(M = Li, K)	Electrolytic rare earth metals with a purity of 95–98%	The yield of rare earth metals is about 85–90%. Single rare earth current efficiency is around 70%.	Generates toxic chlorine gas, low metal recovery.
Fluoride system molten salt electrolysis method [30]	REF3-RE2O3-LiF	Electrolytic rare earth metals with a purity of 95–98%	The yield of rare earth metals is about 90–95%. Current efficiency is about 70–80%.	The current efficiency and rare earth metal yield are high; the electrolysis temperature is high, and the anode produces greenhouse gases such as CO2, CF4, and C2F6.

The molten salt electrolysis method has several advantages, including continuous production, less expensive processing, and simple process equipment for the extraction of rare earth metals. The purity of the prepared rare earth metals can reach more than 95%, and the composition of the prepared rare earth alloys is uniform and not segregated. The composition of the alloy can be controlled by adjusting the electrolysis parameters [31]. Zhao et al. [32] found that the current efficiency, metal yield and energy consumption of rare earth metals La, Ce, and Pr prepared by electrolysis in a fluoride molten salt system were significantly improved compared with those in a chloride molten salt system. The current efficiency increased by 15–30%, the metal yield had an improvement of nearly 15%, and the energy consumption decreased by about 42%. In 2011, China issued and implemented the “Discharge Standards for Pollutants of Rare Earth Industry” [33]. Chloride molten salt electrolysis has been limited in the industry due to many issues, including the generation of harmful fumes, the raw materials’ susceptibility to moisture, and low current efficiency. The fluoride system has better chemical stability and a wider electrochemical window. At present, more than 95% of rare earth metals and their alloys in China are produced by electrolytic preparation using the fluoride molten salt system. This study will, thus, systematically review the process of preparing rare earth metals and alloys using the fluoride molten salt electrolysis system [34].

3. Research Status of Fluoride Molten Salt System

3.1. Preparation of Individual Rare Earth Metal

In recent years, the production of rare earth functional materials has maintained a stable growth trend. The Ministry of Industry and Information Technology and the Ministry of Natural Resources in China have issued rare earth mining quotas in response to the ongoing increase in market demand. These quotas have grown annually over the last two years, with an annual growth rate of roughly 20%, mainly for the exploitation of light rare earths. Therefore, this section mainly describes the research status of the molten salt electrolysis preparation process of light rare earth metals La, Ce, Pr, and Nd in the fluoride salt system. Figure 4 shows the general process flow of preparing rare earth metals through fluoride salt electrolysis. The existing processes mainly improve the quality and electrolysis current efficiency of the products by changing the materials and structures of the anode and cathode, the composition of the electrolyte, the electrolysis temperature, and other process parameters.

3.1.1. Lanthanum Metal

La is an important rare earth element with a low melting point and vapor pressure. It is commonly prepared by the molten salt electrolysis method. A certain concentration of La element has a positive effect on the growth of animals and plants [35]. La₂O₃ nanoparticles can be used to promote the growth and development of plants during the skotomorphogenic stage in agriculture [36]. LaF₃ can inhibit the activity of Streptococcus mutans glucosyltransferase and effectively suppress bacterial growth and caries formation [37]. The lanthanum element is employed in diverse applications such as medical health, livestock breeding, and agriculture. Lanthanum is added to high-refractive-index alkali-resistant glass in camera lenses and oil refineries.

L. Massot et al. [38] and Ryan Chesser et al. [39] applied electrochemical analysis methods to study the oxidation–reduction reaction processes of La(III) in the LiF–CaF₂–LaF₃ molten salt system and the LiF–NaF–KF–LaF₃ molten salt system. Standard potentials were determined through the calculation of reaction equilibrium potentials, and the corresponding Gibbs free energies were also calculated. Liang et al. [40] investigated the solubility of La₂O₃ in the LiF–LaF₃–La₂O₃ molten salt system using the isothermal saturation method. They obtained a regression equation for the solubility using the least squares method:

S (wt.%) = 1 × 10⁻⁴ ω·T − 0.085 ω − 0.003 T + 3.60

(5)

This equation is applicable within the range of dissolution time of 0–90 min, a temperature (T) of 1173–1543 K, and a LaF₃ content (ω) of 65–75 wt.% in the molten salt with an absolute error of less than 5%. These solubility studies and electrochemical mechanisms directly influence the feeding rate of lanthanum oxide, offering a theoretical foundation for modifying the parameters of the production process.

Zhao et al. [41] investigated the effect of different electrolytic raw materials (La₂O₃, LaOF) on current efficiency in the fluorine salt system. In the experiment, graphite crucible was used as the anode and molybdenum rod was used as the inert cathode. It was concluded that tetragonal LaOF was the optimal electrolytic raw material, as detailed in

Table 3. Comparison of current efficiency and metal yield for the preparation of metallic lanthanum under different electrolysis parameters. Adapted from ref. [41,42,43,44].

Molten Salt System	Electrolysis Temperature	Electrolytic Raw Materials	Time of Electrolysis	Cathodic Current Density/(A/cm2)	Metal Quantity	Current Efficiency
65 wt.% LaF3–23 wt.% LiF–12 wt.% BaF2 [41]	1223 K	La2O3	60 min	7	36.8 g	85.21%
	1223 K	Tetragonal crystal system LaOF	60 min	7	38.1 g	88.23%
	1223 K	Rhombic hexagonal crystal system LaOF	90 min	7	46.7 g	86.51%
70 wt.% LaF3–LiF–BaF2 [43]	1223 K	La2O3	70 min	8	45.3 g	93.38%
LiF–LaF3 [44]	1243–1253 K	Fluorescence-level purity La2O3	20 min	5	71.8 g	81.57%

. Although from the perspective of current efficiency, LaOF can replace La₂O₃ as the electrolytic material, the addition of LaOF, according to the law of mass conservation, leads to an excessive amount of fluoride ions in the molten salt. These fluoride ions further react with the graphite anode, resulting in the formation of greenhouse gases such as CF₄ and C₂F₆ fluorocarbons, which severely pollute the air [42]. Therefore, rare earth oxides are still primarily used as electrolytic raw materials in industry. Zhang et al. [43] studied the process parameters of electrolytic rare earth metal lanthanum by orthogonal regression experimental research methods. It was found that the electrolysis temperature and the feeding rate of lanthanum oxide had a great influence on the current efficiency and the direct metal yield. Lin et al. [44] employed fluorescence-grade La₂O₃ as the electrolytic raw material in order to increase the purity of La metal. The anode and electrolytic cell used in the experiment were high-purity graphite with an ash content of less than 0.01%. The cathode and metal receivers were tungsten rods and molybdenum crucibles, respectively. The content of rare earth in the cathode product was more than 99.8%, in which the content of lanthanum was more than 99.99%, and the average contents of carbon and iron were 0.022 wt.% and 0.006 wt.%, respectively. Wang et al. [45] used lanthanum metal or La–Ni alloy as the anode and a high-purity Ta rod as the cathode. Lanthanum metal with an absolute purity of 99.87% was refined and extracted in the LiF–LaF₃ molten salt system. This exploratory research expands the purification process of light rare earth metals and lays the foundation for the subsequent recovery of rare earth elements from rare earth waste.

3.1.2. Cerium Metal

In the crust of the Earth, cerium is the rarest element with the largest abundance. It is frequently added to functional materials such as polishing powder, glass and ceramic colorant, and exhaust gas purification catalysts, and is also used in plant growth regulators.

The electrochemical behavior of Ce(III) on inert electrodes in LiF–CaF₂, LiF–BaF₂ [46], and LiF–NaF–NaCeF₄ [47] molten salts was examined by M. Chandra et al. [48,49]. The findings show that the reduction of Ce(III) to Ce(0) metal is a one-step, reversible process using three electrons. Furthermore, the Ce(III) diffusion coefficient in the molten salt was computed. In a separate study, Lin et al. [50] conducted small-scale experimental research on the electrolytic oxidation of cerium in fluorine salt systems. The electrolyte was a melt of 63 wt.% CeF₃, 21 wt.%LiF, and 16 wt.% BaF₂. The working temperature of electrolysis was equal to 1173 K. Well-aggregated metallic cerium products were formed after an hour at a constant cathodic current density of 6.25 A/cm². The obtained metallic cerium had a purity that was above 98%, and its current efficiency and metal recovery were 53% and 94.7%, respectively. Furthermore, Li et al. [48] further investigated the electrochemical behavior of Ce(III) on a molybdenum cathode in CeF₃–LiF–BaF₂–CeO₂ molten salt and obtained metallic cerium with a purity greater than 99% after electrolysis for two hours at a constant potential of −1.25 V (vs. C/CO_x). Similarly, V. Constantin et al. [49] also measured the rate of electrolytes reaching the equilibrium state using a polarization curve and found that the rate of reaching the equilibrium state in LiF–NaF–NaCeF₄ molten salt without CaF₂ was very fast, and the solvent ions in the molten salt would not interfere with each other. Through a series of tests, Zhang et al. [51] investigated the ideal composition ratio of rare earth cerium prepared by molten salt electrolysis. Based on the comprehensive analysis of the initial crystallization temperature, volatilization loss, and CeO₂ solubility, it was determined that 85.5 wt.% CeF₃–14.5 wt.% LiF was the optimal binary system. It was possible to dissolve 5 wt.% CeO₂ and lower the initial crystallization temperature to 1183 K under the optimal ratio of electrolyte conditions, which significantly minimized energy consumption. Zhongxing Liu et al. [52] simulated and analyzed the temperature field variation inside the electrolytic cell in actual process production using COMSOL. It was found that in a 6 kA new cerium electrolytic cell, the optimal depths of the anode and cathode insertion into the molten salt were 290 mm (the inner diameter and outer diameter of the anode were 218 mm and 410 mm, respectively) and 250 mm (the cathode diameter was 80 mm), respectively. If the depth of the electrode insertion was too deep, the temperature inside the cell would rise, leading to an increase in electrolyte volatilization, which was not beneficial for production. Conversely, when the depth of electrode insertion was insufficient, there were nodulations in the cell, which shortened the life of the electrolytic cell and reduced the current efficiency.

The electrolytic reduction process of cerium, a variable valence rare earth metal, is relatively complex, and it is easy to experience current waste issues [53]. In the case of cerium production, the current efficiency is often low. As a result, in industrial production, cerium is frequently extracted using active electrodes as cathodes in a non-co-deposition manner, and it is then further processed to create cerium alloys.

3.1.3. Praseodymium Metal

Praseodymium is usually manufactured in the form of praseodymium–neodymium alloy, and there is limited research on the electrolytic preparation of pure praseodymium as a rare earth metal. M. Straka et al. [54] studied the electrochemical processing of Pr(III) on a tungsten electrode in LiF–CaF₂–PrF₃ molten salt and found that the reduction step of Pr(III)/Pr(0) is an one-step reduction, but it is not feasible to electrolytically obtain praseodymium metal on inert electrodes. Sichuan Jinjiang Rare Earth Metal Co., Ltd. (China, Chengdu) [55] conducted an experimental study on the electrolysis of praseodymium and the production of praseodymium metal in the PrF₃–LiF molten salt system. Through rigorous experiments in a conventional 3000 A electrolytic cell, the optimal process parameters were determined: a voltage of 9 V, an electrolysis temperature of 1223 K, and a cathode current density of 7.28 A/cm². Continuous electrolysis produced a current efficiency of roughly 70–80% and a metal product with a total rare earth content of over 99.5% and a carbon content of less than 0.01 wt.% under these process settings.

3.1.4. Neodymium Metal

The Baotou Steel Metallurgy Research Institute [56] conducted a systematic study on factors such as the composition ratio of the molten salt electrolyte, cathode current density, electrolyte temperature, and oxide addition rate. It was concluded that the NdF₃–LiF electrolyte without the addition of BaF₂ is more conducive to electrolytic production based on the visual effects of metal product agglomeration and current efficiency. Through experiments, it was found that rare earth metal products could be successfully prepared by continuous electrolysis for 3 h under specific process conditions; that is, the electrolyte composition was 89 wt.% NdF₃–11 wt.% LiF, the cathode current density was set to 7 A/cm², and the electrolysis temperature was controlled at 1308 K. In the experiments, the metal yield was as high as 88.5%, the current efficiency reached 62%, and the total rare earth content in the product reached 98.76%. According to Bao [57], achieving a product qualification rate of 97.2% required managing the NdF₃ concentration in the molten salt to be above 87%. This finding is in line with the Baotou Steel Metallurgy Research Institute’s conclusion that a ratio of 89 wt.% NdF₃–11 wt.% LiF is more appropriate. In a real industrial production environment, Mao [58] conducted a 13 day continuous production study in the plant, examining how various electrolyte compositions affected the carbon content and current efficiency of neodymium metal [58]. From Figure 5, it can be seen that when the electrolyte composition ratio was LiF:NdF₃ = 10:100 (mass ratio), the carbon content in the prepared metal neodymium was the lowest, and the current efficiency could reach above 80%. Therefore, the conclusion was made that LiF:NdF₃ = 10:100 was the optimal ratio, and 1323 K was the optimal electrolyte temperature.

Zhang et al. [59] found that the current efficiency of electrolytic rare earth metal neodymium can be as high as 87.1% under low current density electrolysis conditions with cathode and anode current densities of 1.39 A/cm² and 0.28 A/cm², respectively. Low current density electrolysis can not only reduce the secondary dissolution of rare earth metal neodymium in molten salt but also greatly reduce the unit power consumption. This research conclusion provided a reliable basis for the design of a buried cathode rare earth electrolysis cell. The preparation of rare earth metals using a traditional anode and cathode up plug rare earth cell has the problem that the electrolytic products are easily oxidized again by the anode or the molten salt, resulting in low current efficiency and high carbon impurity in the product [60,61]. However, the buried cathode electrolyzer can effectively improve these problems. Due to factors such as current distribution and cathode and anode structure, the cathode current density of a buried cathode electrolyzer is generally considered to be set below 2 A/cm². Chen et al. [62,63] systematically studied and designed a 3000 A liquid cathode electrolysis cell. The design size of the electrolytic cell is 650 mm × 540 mm × 300 mm, and the cell body is composed of tungsten plates. The bottom of the tungsten plate cell is used as a buried cathode, and the formed liquid metal neodymium is used as the true liquid cathode; eight blocks of graphite are inserted above the electrolytic cell as anodes, and the anode and cathode current densities are set at 1.95 A/cm² and 1.1 A/cm² respectively. After 15 days of the continuous electrolysis of metal neodymium, a current efficiency of 75.36–85.75% was achieved, and the recovery of rare earth metals reached 95.2%. According to the comparison of the current efficiency of the two types of cell bodies, the new type of buried cathode-structure electrolysis cell can effectively prevent the secondary oxidation of the neodymium metal, reduce power consumption, and increase current efficiency.

Table 4. Process parameters for the preparation of rare earth metal neodymium. Adapted from refs. [56,59,62,64].

Electrolyte System	38 wt.% NdF3–35 wt.% LiF–27 wt.% CaF2 [64]	89 wt.% NdF3–11 wt.% LiF [56]	90 wt.% NdF3–10 wt.% LiF [59]	(88.9–90 wt.%) NdF3–LiF [62]
Electrode	Anode: graphite Cathode: tungsten rod	Anode: graphite electrolyzer Cathode: 6–8 mm diameter molybdenum rod	Anode: graphite Cathode: 28 mm diameter tungsten rod	Anode: graphite block Cathode: tungsten plate
Electrolysis temperature	1303 K	1308 K	1343–1353 K	1293 K
Cathode current density	10.97 A/cm2	7 A/cm2	0.85–1.39 A/cm2	1.95 A/cm2
Tank voltage	not indicated	4.50–5.50 V	4–6.50 V	4.80–6 V
Electrolyzer structure	Anode and cathode up plug rare earth cell	Anode and cathode up plug rare earth cell	Anode and cathode up plug rare earth cell	Bottom cathode rare earth cell
Current efficiency	52.7%	62%	57.60–87.10%	75.36–85.75%
Rare earth direct yield	38.60%	88.50%	Not indicated	95.20%

summarizes the process parameters for the preparation of neodymium metal.

As can be seen from the above table, the fluorine salt systems in industrial production are mainly based on the LiF–NdF₃ binary electrolyte system, with a current efficiency of about 80%. In the production of molten salt electrolysis, graphite crucible containers or graphite are often used as conductive anodes, so the appearance of non-metallic impurities such as carbon in the product is inevitable. However, the presence of carbon impurities will reduce the purity of the product and, thus, weaken its performance. Wei et al. [65] analyzed the morphology of non-metallic-impurity carbon in 99% pure rare earth metal neodymium and discovered that the carbon impurities were amorphous carbon, primarily distributed near the metal neodymium’s grain boundaries. They placed the prepared neodymium metal with Mg metal into a vacuum melting furnace for decarburization treatment. Consequently, the carbon content in purified neodymium metal was reduced from 1300 ppm to 700 ppm. This research offers novel ideas and techniques for the purification of neodymium metal. In fact, several researchers have attempted to address the issue at its core by directly minimizing the introduction of carbon impurities. Xue et al. [66] used borate to fill the gaps in the graphite, enabling the graphite anode to isolate from the corrosive molten salts in a short time, thereby decreasing carbon impurities. However, the effective duration of this “protective layer” is limited.

With the increasing demand for rare earth in countries around the world, rare earth resources are listed as strategic resources. As rare earth mines are non-renewable resources, in order to alleviate the pressure on mining, developed countries have begun to actively conduct research on recovering rare earth elements from waste materials. A. Abbasalizadeh et al. [67] proposed an electrolytic method for extracting neodymium from NdFeB magnets. It was found that Na₃AlF₆ serves as a fluorinating agent for Nd in NdFeB magnets, enabling the replacement of Nd ions in the NdFeB magnet by Na ions from the fluorite system. Yang et al. [68] directly utilized rare earth magnet waste as the anode. By controlling the oxidation rate of rare earth ions within the LiF–CaF₂ melt, they achieved the direct reduction of rare earth metals at the cathode. AlF₃–NaF was used as a solvent to extract neodymium metal from neodymium iron boron waste [69]. The rare earth recovery rose from 71.1% to 94.2% with a temperature increase from 1072 K to 1373 K.

3.2. Preparation of Rare Earth Alloys

The electrodeposition method broadly divides the process of manufacturing rare earth alloys using molten salt electrolysis into two categories: co-deposition and non-co-deposition. The reduction potentials of rare earth ions and other metal ions are often similar in rare earth alloys that are created by co-deposition. This makes it possible to reduce and alloy these two kinds of ions on an inert cathode at the same time. Depending on the various cathode types, non-co-deposition techniques can be further separated into liquid and self-consumed cathode approaches. Rare earth alloy compositions prepared using co-deposition methods are easier to control, but the types of alloys that can be prepared are limited [70,71].

3.2.1. RE–Mg Alloys

Magnesium-based alloys have the advantages of high specific strength, low density, good heat dissipation, and strong corrosion resistance. Adding an appropriate amount of rare earth elements can optimize the metallographic structure of magnesium-based alloys, refine the grain size, and further improve the alloy’s comprehensive performance.

Li et al. [72] analyzed the GdF₃–LiF–Gd₂O₃–MgO molten salt system by electrochemical scanning and found that Gd (III) was reduced by a one-step three-electron process on the molybdenum electrode. Gd (III) and Mg (II) in the molten salt were co-electrodeposited on the molybdenum cathode to form a Gd–Mg alloy with uniform composition [72,73], and the current efficiency can reach 80%.

Yang et al. [74] used NdF₃–LiF–BaF₂–Na₃AlF₆ electrolyte and neodymium oxide and magnesium powder as electrolyte raw materials and obtained a uniform Mg–Nd master alloy with current efficiency of 72.6% by non-co-electrodeposition at 1323 K and a cathode current density of 6–7 A/cm². V. Soare et.al [75,76] also adopted the liquid magnesium cathode deposition method, using 28.60 wt.% LiF–64.72 wt.% NdF₃–6.68 wt.% MgF₂ as the electrolyte and Nd₂(CO₃)₃ and MgF₂ as the electrolysis materials. Under the conditions of an electrolysis temperature of 1123 K and anodic and cathodic current densities of 3.9 A/cm² and 0.3 A/cm², respectively, the current efficiency reached 87% and the rare earth recovery reached 83.17%. It was observed that the melting point of the molten salt electrolytes without the addition of BaF₂ and Na₃LiF₆ was lower. Although Nd₂(CO₃)₃ exhibited a higher current efficiency, REO was still widely used as an electrolytic raw material in industry applications because carbon could enter the product. Deng et al. [77] utilized liquid magnesium as the cathode in the electrolysis of YF₃–LiF–Y₂O₃ molten salt. This resulted in the production of a Y–Mg alloy containing more than 55 wt.% rare earth Y content. The alloy was further refined to achieve a purity of 99% sponge yttrium through vacuum smelting at 1293 K. He et al. [78] studied the process of the co-electrodeposition of Y–Mg alloys in the fluoride salt system and found that Mg²⁺ and Y³⁺ in the molten salt were reduced in a one-step reduction, which was controlled by diffusion. When the ratio of Y₂O₃ to MgO was 4:1, Mg²⁺ and Y³⁺ in the molten salt were co-electrolyzed at −0.9 V potential. Yang [79] also believes that electrolysis of this ratio of electrolyte oxides at 1323 K is feasible, and the formed Mg–Y alloy has a uniform composition.

3.2.2. RE–Al Alloys

Aluminum exhibits excellent ductility, while copper possesses exceptional thermal and electrical conductivity. An Al–Cu alloy is mainly composed of 3 wt.% copper and 97 wt.% aluminum. The alloy is well suited for use in sheet metal, forgings, and structural aerospace components due to its exceptional heat resistance, strong mechanical qualities, and superior machinability. However, the potential difference between the Al₂Cu phase and the Al phase matrix in the alloy is large, resulting in the low corrosion resistance of the alloy. Adding an appropriate amount of heavy metal Y can effectively improve the alloy’s high-temperature performance, conductivity, and corrosion resistance. The addition of rare earth elements Nd and Sc also enhances the hardness and tensile strength of Al alloys.

Viscosity is a crucial physical attribute of melts that significantly affects the electrolysis of molten salt. Leng et al. [80] conducted a study on the viscosity of the fluoride salt system for the electrolytic preparation of Al–Cu–Y alloys. They derived a regression equation for the melt viscosity η in the temperature range of 1173–1273 K and with varying oxide additions:

$$\eta \left(Pa·s\right)=0.075-6.49\times {10}^{-5}T+3.7\times {10}^{-4}{W}_{A{l}_2{O}_3}+5.73\times {10}^{-4}{W}_{Y_2{O}_3}+6.78\times {10}^{-4}{W}_{CuO}$$

(6)

Wang et al. [81,82] determined the ratio of AlF₃ to NaF in the AlF₃–NaF–5 wt.% LiF–5 wt.% MgF₂ molten salt electrolyte based on the viscosity data. Using Al₂O₃, CuO, and Y₂O₃ as the raw materials and with both the anode and cathode made of graphite in the traditional anode and cathode up plug rare earth cell, they successfully prepared the Al–Cu–Y ternary alloy via electrolysis. Furthermore, through electrochemical scanning analysis, they provided a detailed understanding of the reduction process of the various ions in the molten salt, and Y(III) was reduced to form the Al₃Y phase in the Al–Cu–Y ternary alloy.

In the experiment of the electrolysis of a Nd–Al alloy, Liao et al. [83,84,85] used graphite crucible as an inert cathode to prepare Al₁₁Nd₃ phase alloy. Through the aluminum thermoreduction of Nd₂O₃ test and cyclic voltammetry analysis, it was concluded that the Al–Nd alloy was not prepared by a co-electro-precipitation method but directly formed by aluminum thermoreduction. Sun et al. [86] used pure aluminum as the liquid cathode and determined that the suitable molten salt composition was 90 wt.% Na₃AlF₆–3 wt.% LiF–7 wt.% Sc₂O₃ by orthogonal experiments. The Al–Sc alloy with current efficiency of 71.1% and Sc content of 7.5 wt.% was obtained by electrolysis at 1253 K and a cathode current density of 0.8 A/cm² for 2 h. When adding 4 wt.% MgF₂ and 2 wt.% CaF₂ [87,88] in molten salt, an aluminum–scandium alloy with a scandium content of 0.5 wt.% was obtained by electrolysis. It was found through detection that the Al₃Sc phase [89] in the alloy had fine grains and high dispersion, which contributed to the improvement of the alloy’s hardness and tensile strength [90]. Liu et al. [91] used Al–7Si alloy as the liquid cathode to prepare an Al–Si–Sc–Ce alloy, and by controlling the cooling rate of the alloy, a low-cerium, high-scandium Al–7Si–0.8Sc–0.6Ce alloy was obtained.

Zhu et al. [92] compared a series of cathodes for preparing Al–Sc alloys and found significant differences in the rare earth Sc content in the Al–Sc alloys obtained from different cathode substrates. From Figure 6, it can be seen that the Al–Sc alloy obtained by electrolysis with Al–5Li alloy as the cathode had the highest rare earth Sc content. According to performance tests, the alloy produced by this cathode had superior hardness and tensile strength performance compared to alloys created by other cathodes. It was primarily composed of the Al₃Sc phase, with a hardness of 289.9 kg/mm² and a tensile strength of 789 MPa. M. Gibilaro et al. [49,93,94] also studied the electrochemical behavior of Ce(III), Nd (III), Sm(III), and Eu(III) in the 55.56 wt.% LiF–44.44 wt.% CaF₂–AlF₃ system and the co-electrodeposition process of rare earth ions with Al. In their experiments, the extraction rates of various rare earth ions were 99.5%, 95%, 99.4%, and 99.6%, respectively. Gao et al. [95] investigated the ion conversion process during the electrolysis of Sm₂O₃ to form Sm–Al phase alloys in a molten salt system consisting of 55.56 wt.% LiF and 44.44 wt.% CaF₂. Specific reduction information can be found in Table 5.

3.2.3. RE–Ni Alloys

Nickel-based alloys have oxidation and corrosion resistance at high temperatures. La–Ni alloy is an excellent representative of hydrogen storage materials because of its high energy density. The electrochemical capacities of La–Mg–Ni hydrogen storage alloy [96] and La–Y–Ni hydrogen storage alloy [97] are greater than that of La–Ni binary alloy materials, but its cycle stability and practicality still need further improvement [98].

The electrochemical behavior of La(III) and underpotential deposition of La–Mg–Ni alloys were investigated at a Ni active electrode in 85.47 wt.% LaF₃–14.53 wt.% LiF–La₂O₃–MgO melts [99]. Cyclic voltammetry and square wave voltammetry scanning curves showed that at least three types of La–Mg–Ni alloys were formed on the nickel cathode in the voltage range of −0.4 V to −1 V (vs. Pt quasi-reference electrode). Wang et al. [100,101] studied the solubility of Yb₂O₃ in the LiF–YbF₃–Yb₂O₃ molten salt system and the ionic structure of the molten salt. Under the conditions of constant voltage electrolysis at 1523 K for 4 h, a Yb–Ni alloy consisting mainly of Ni₅Yb and Ni₁₇Yb₂ phases was obtained. Wang et al. [102,103] also used nickel as a self-consuming cathode to electrolytically obtain a Y–Ni alloy with a Y content of 52.6 wt.%. The alloy is mainly composed of YNi₂ and YNi phases, with a low impurity content.

Beyond the incorporation of rare earth elements such as La, Yb, and Y, there is a lot of research on other RE–Ni alloys. L. Massot et al. [104] drew the E–PO²⁻ diagram of Sm(III) in 52.66 wt.% LiF–47.34 wt.% CaF₂ molten salt, pointing out that the reduction process of Sm(III) is a two-step reduction. The Sm₃Ni alloy is electrolyzed by the non-co-deposition method. P. Chamelot et al. [105] studied the electrochemical process of rare earth ions such as NdF₃, GdF₃, and SmF₃ on the nickel electrode in 52.66 wt.% LiF–47.34 wt.% CaF₂ molten salt, and detected NdNi₂, NdNi₃, GdNi₂, GdNi, SmNi₂, SmNi_3, and other alloy phases through SEM. A. Saïla et al. [106] conducted a study on the oxidation and reduction process of Dy(III) in the medium of a 55.56 wt.% LiF–44.44 wt.% CaF₂ molten salt, as well as the alloying mechanism with Ni. T. Nohira et al. [107,108] determined the formation potentials of NdNi₂, NdNi₃, and NdNi₅ phases in Nd–Ni alloys. Seitaro Kobayashi et al. [109,110,111] summarized the equilibrium potentials of each phase formed in RE–Ni alloys. These RE–Ni alloys are currently in the experimental exploration stage and have not yet been industrialized. Detailed electrochemical information regarding the alloys can be found in the table below (Table 5).

Table 5. Electrochemical formation of RE–Mg alloy, RE–Al alloy, RE–Ni alloy, and RE–Fe alloy in fluoride salt system. Adapted from ref. [40,76,77,78,97,98,99,100,112].

RE	Molten Salt	Electrode	Electrolysis Time	Temperature	Applied Current Density/Current/ Potential (vs. Pt)	RE Alloys
Ce [40]	52.66 wt.% LiF-47.34 wt.% CaF2-CeF3	Ni	60 min	1093 K	−1.5 V	CeNi5
Eu [76]	55.56 wt.% LiF-44.44 wt.% CaF2-EuF3-AlF3	W	20 min	1113 K	−1.1 V	EuAl4
Nd [76]	55.56 wt.% LiF-44.44 wt.% CaF2-NdF3-AlF3	W	20 min	1133 K	−1.33 V, −1.49 V, −1.69 V, −1.88 V	Nd3Al11, NdAl3, NdAl2, Nd3Al
Ce [77]	55.56 wt.% LiF-44.44 wt.% CaF2-CeF3-AlF3	W	20 min	1113 K	−1.25 V, −1.43 V, −1.79 V	Ce3Al11, CeAl3, (Ce3Al, CeAl)
Sm [77]	55.56 wt.% LiF-44.44 wt.% CaF2-NdF3-AlF3	W	20 min	1113 K	−1.45 V	SmAl3
Sm [78]	LiF-CaF2-Sm2O3-AlF3	W	180 min	1223 K	−731.4 mA/cm2	SmAl2, SmAl3
Sm [96]	52.66 wt.% LiF-47.34 wt.% CaF2	Ni	60 min	1120 K	−200 mA/cm2	Sm3Ni, SmNi2
Nd [97]	52.66 wt.% LiF-47.34 wt.% CaF2-NdF3	Ni	120 min	1113 K	35 mA/cm2, 0.65 V	NdNi2, NdNi3
Gd [97]	52.66 wt.% LiF-47.34 wt.% CaF2-GdF3	Ni	120 min	1113 K	35 mA/cm2, 0.65 V	GdNi2, GdNi
Sm [97]	52.66 wt.% LiF-47.34 wt.% CaF2-SmF3	Ni	120 min	1113 K	35 mA/cm2, 0.7 V	SmNi2, SmNi3
Dy [98]	55.56 wt.% LiF-44.44 wt.% CaF2-DyF3	Ni	60 min	1113 K	−15 mA/cm2	Dy2Ni17, DyNi4
Dy [98]	55.56 wt.% LiF-44.44 wt.% CaF2-DyF3	Ni	60 min	1113 K	−15 mA/cm2	Dy2Ni17, DyNi4
Nd [99]	57.84 wt.% LiF-42.16 wt.% CaF2-NdF3	Ni	75 min	1123 K	0.15 V,0.35 V (vs. Li+/Li)	NdNi2, NdNi3
Nd [99]	57.84 wt.% LiF-42.16 wt.% CaF2-NdF3	Ni	50 min	1123 K	0.60 V (vs. Li+/Li)	NdNi5
Pr [102]	57.84 wt.% LiF-42.16 wt.% CaF2-PrF3	Ni	90 min /180 min	1123 K	0.2 V, 0.36 V, 0.6 V(vs. Li+/Li)	PrNi2, PrNi3, PrNi5
Dy [113]	57.84 wt.% LiF-42.16 wt.% CaF2-DyF3	Fe	30 min	1123 K	0.1 V (vs. Li+/Li)	DyFe2

Table 5. Electrochemical formation of RE–Mg alloy, RE–Al alloy, RE–Ni alloy, and RE–Fe alloy in fluoride salt system. Adapted from ref. [40,76,77,78,97,98,99,100,112].

RE	Molten Salt	Electrode	Electrolysis Time	Temperature	Applied Current Density/Current/ Potential (vs. Pt)	RE Alloys
Ce [40]	52.66 wt.% LiF-47.34 wt.% CaF2-CeF3	Ni	60 min	1093 K	−1.5 V	CeNi5
Eu [76]	55.56 wt.% LiF-44.44 wt.% CaF2-EuF3-AlF3	W	20 min	1113 K	−1.1 V	EuAl4
Nd [76]	55.56 wt.% LiF-44.44 wt.% CaF2-NdF3-AlF3	W	20 min	1133 K	−1.33 V, −1.49 V, −1.69 V, −1.88 V	Nd3Al11, NdAl3, NdAl2, Nd3Al
Ce [77]	55.56 wt.% LiF-44.44 wt.% CaF2-CeF3-AlF3	W	20 min	1113 K	−1.25 V, −1.43 V, −1.79 V	Ce3Al11, CeAl3, (Ce3Al, CeAl)
Sm [77]	55.56 wt.% LiF-44.44 wt.% CaF2-NdF3-AlF3	W	20 min	1113 K	−1.45 V	SmAl3
Sm [78]	LiF-CaF2-Sm2O3-AlF3	W	180 min	1223 K	−731.4 mA/cm2	SmAl2, SmAl3
Sm [96]	52.66 wt.% LiF-47.34 wt.% CaF2	Ni	60 min	1120 K	−200 mA/cm2	Sm3Ni, SmNi2
Nd [97]	52.66 wt.% LiF-47.34 wt.% CaF2-NdF3	Ni	120 min	1113 K	35 mA/cm2, 0.65 V	NdNi2, NdNi3
Gd [97]	52.66 wt.% LiF-47.34 wt.% CaF2-GdF3	Ni	120 min	1113 K	35 mA/cm2, 0.65 V	GdNi2, GdNi
Sm [97]	52.66 wt.% LiF-47.34 wt.% CaF2-SmF3	Ni	120 min	1113 K	35 mA/cm2, 0.7 V	SmNi2, SmNi3
Dy [98]	55.56 wt.% LiF-44.44 wt.% CaF2-DyF3	Ni	60 min	1113 K	−15 mA/cm2	Dy2Ni17, DyNi4
Dy [98]	55.56 wt.% LiF-44.44 wt.% CaF2-DyF3	Ni	60 min	1113 K	−15 mA/cm2	Dy2Ni17, DyNi4
Nd [99]	57.84 wt.% LiF-42.16 wt.% CaF2-NdF3	Ni	75 min	1123 K	0.15 V,0.35 V (vs. Li+/Li)	NdNi2, NdNi3
Nd [99]	57.84 wt.% LiF-42.16 wt.% CaF2-NdF3	Ni	50 min	1123 K	0.60 V (vs. Li+/Li)	NdNi5
Pr [102]	57.84 wt.% LiF-42.16 wt.% CaF2-PrF3	Ni	90 min /180 min	1123 K	0.2 V, 0.36 V, 0.6 V(vs. Li+/Li)	PrNi2, PrNi3, PrNi5
Dy [113]	57.84 wt.% LiF-42.16 wt.% CaF2-DyF3	Fe	30 min	1123 K	0.1 V (vs. Li+/Li)	DyFe2

3.2.4. Nd Alloys

Rare earth Pr–Nd alloys and Nd–Fe alloys are key raw materials for the preparation of NdFeB magnetic materials. Rare earths such as Ce, Dy, and Gd are often added to the NdFeB magnet, effectively improving the overall performance of the magnet [112].

Zhou et al. [114] studied the production process of Pr–Nd alloys in 10 kA-level rare earth electrolytic cells. The experimental electrolysis raw materials contained 0.92 wt.%–1.2 wt.% SiO₂. After continuous production for 24 days, it was found that the Si content in the Pr–Nd alloy was around 0.05 wt.%, and the current efficiency was 70%. In order to effectively control the content of non-rare-earth impurities in the Pr–Nd alloy, Lu et al. [115] changed the arrangement of traditional plug-in anodes and cathodes to a semi-hanging arrangement of six tungsten–molybdenum alloy cathodes and changed the dimensions of the electrolytic cell to 1.8 m × 1.0 m × 0.6 m. They established the optimal process parameters for the production of Pr–Nd alloys in a 25 kA molten salt electrolytic cell as follows: within a molten salt system comprising 90.91 wt.% NdF₃–9.09 wt.% LiF–REO (RE: 25 wt.% Pr, 75 wt.% Nd), the electrolysis is conducted at a temperature of 1323 K and a cathode current density ranging between 7 and 12 A/cm². The resulting Pr–Nd alloy exhibits a non-rare-earth impurity content of less than 0.05 wt.%, with a carbon content specifically below 0.04 wt.%, which meets industrial production standards.

Du et al. [116] used rich praseodymium neodymium oxide as the electrolytic raw material. In the process of electrowinning Pr–Nd alloy, the current efficiency reached 57%, and the recovery of rare earth reached as high as 90%. Researchers investigated the solubility of a variety of rare earth oxides in BaF₂–LiF–REF₃ electrolytes at a temperature of 1273 K. It is evident from Figure 7 that rare earth oxide solubility in LiF–BaF₂–MgF₂ fluoride electrolytes is generally low [116]. Therefore, the feeding speed of oxide raw materials is a key factor affecting current efficiency in electrolytic production [117]. K. Kawaguchi et al. [113] determined the electrolysis potentials of various Dy–Fe alloy phases by utilizing electrochemical techniques to analyze the reactivity of Dy (III) with Fe electrodes in the LiF–CaF₂ molten salt system. The Japanese company Showa Denko Company [118] produced Nd–Fe alloy by molten salt electrolysis in an oxidizing atmosphere using NdF₃ as the electrolysis raw material. Although the experimental results are quite perfect, it is difficult to promote this experimental method in an industrial environment. Yu Makaseev et al. [119] used Fe rods as self-consuming cathodes to electrolytically produce Nd–Fe cast alloys in an inert atmosphere. The experiment resulted in a current efficiency of up to 90%, significantly higher than the actual current efficiency in the industry. Combined with the need to introduce inert gas during the experiment, this provides new ideas for improving the molten salt electrolysis process.

The three rare earth elements of Gd, Dy, and Ce are essential additives in the production of NdFeB. They can enhance the stability and magnetic properties of the magnets. These three elements are usually added to the magnet in the form of a single rare earth metal, Dy–Fe alloy, or Gd–Fe alloy, making the production process lengthy. The complete process of creating NdFeB magnets can be shortened by directly preparing Pr–Nd–Ce [120], Pr–Nd–Dy [121], and Pr–Nd–Gd [122] alloys as raw materials via the molten salt electrolysis method.

In the process of preparing Pr–Nd–Ce [120] by molten salt electrolysis, Yu et al. utilized a graphite crucible as the anode and a tungsten rod as the inert cathode. Within a temperature range of 1253 K to 1333 K, appropriate amounts of Pr₆O₁₁, Nd₂O_3, and CeO₂ were added to the LiF–Pr(Nd)F₃–CeF₃ system and the Pr–Nd–Ce ternary alloy was prepared using the co-electrodeposition method. It was discovered that by steadily increasing the CeF₃ content in the electrolyte and keeping the electrolysis current constant for one hour, the ternary alloy’s cerium content rose, and its homogeneity improved with time. Using the manufacturing of Ce₈₅PrNd₁₅ alloy as an example, Figure 8 shows that the Ce content of the alloy and the CeO₂ content of the electrolyte after electrolysis both continually increase as the electrolysis temperature rises. Apparently, the increase in temperature leads to higher solubility of REO, thus providing the higher Ce content in the final product. The effects of electrolyte ratio, electrolysis temperature, current density, and other process factors on the composition of each element in Pr–Nd–Dy and Pr–Nd–Gd ternary alloys, metal direct yield, and current efficiency were investigated experimentally on an industrial scale by Chen et al. [121,122]. With precise process parameters as illustrated in Figure 9, the experiment produced 15 tons of alloy in a 4000 A electrolysis tank using a tungsten rod as the cathode and a graphite rod as the anode. The impact of process parameters on the Pr–Nd–Gd ternary alloy’s metal such as direct yield and current efficiency are illustrated in Figure 9a. The results of chemical analysis for produced alloys from different baths in a continuous regime are displayed in Figure 9b. The metal yield and current efficiency of the two alloys were compared in Figure 9c,d, which demonstrates how temperature had a significant impact on both parameters. Under the same temperature conditions, the metal yield of the Pr–Nd–Dy alloy was generally greater than that of the Pr–Nd–Gd alloy, while the current efficiency was less than that of the Pr–Nd–Gd ternary alloy.

3.3. Discussion and Summary

3.3.1. Electrolysis Parameters

The technological parameters for the preparation of light rare earth metals and their alloys by the fluoride salt system electrolytic preparation of rare earth oxides are shown in

Table 6. Summary of parameters for the preparation of light rare earth metals and alloys by fluoride salt system electrolytic rare earth oxides. Adapted from ref. [41,121].

Rare Earth Metals/Rare Earth Alloys	Electrolyte	Raw Material	Electrolysis Temperature	Cathode Current Density	Current Efficiency	Metal Yield
La [41]	65 wt.% LaF3–23 wt.% LiF–12 wt.% BaF2	La2O3	1233 K	7 A/cm2	85.21%	99.29%
La [43]	70 wt.% LaF3–LiF–BaF2	La2O3	1223 K	8 A/cm2	93.38%	99.29%
La [44]	LaF3–LiF	La2O3	1263 K	5 A/cm2	81.57%	not indicated
Ce [50]	63 wt.% CeF3–21 wt.% LiF–16 wt.% BaF2	CeO2	1173 K	6.25 A/cm2	53 %	94.70%
Pr [55]	PrF3–LiF	Pr6O11	1213–1233 K	6 - 8 A/cm2	70-80%	90%
Nd [56]	89 wt.% NdF3–11 wt.% LiF	Nd2O3	1308 K	7 A/cm2	62%	88.50%
Nd [64]	38 wt.% NdF3–35 wt.% LiF–27 wt.% CaF2	Nd2O3	1303 K	10.97 A/cm2	52.70%	38.60%
Nd [59]	90 wt.% NdF3–10 wt.% LiF	Nd2O3	1343–1353 K	0.85–3.79 A/cm2	57.60–87.10%	not indicated
Nd [62]	88.9 wt.% NdF3–11.1 wt.% LiF	Nd2O3	1233–1293 K	1.95 A/cm2	74.30–85.75%	93.10–95.20%
Nd–Mg [74]	NdF3–LiF–BaF2–Na3AlF6	Nd2O3	1323 K	6–7 A/cm2	60–72.60%	84.2–95.2%
Y–Ni [102]	85 wt.% YF3–15 wt.% LiF	Y2O3	1273 K	10 A/cm2	72.80%	not indicated
Pr–Nd [114]	90 wt.% (Pr, Nd)F3–10 wt.% LiF	REO (RE = Nd, Pr)	1273–1373 K	4.5–6.5 A/cm2	57.40-85%	not indicated
Pr–Nd [116]	65 wt.% NdF3–15 wt.% BaF2–20 wt.% LiF	REO (RE = Nd, Pr)	1303 K	8–13 A/cm2	57%	90%
Pr–Nd–(6–8 wt.%) Dy [121]	(87.50–90.91 wt.%)REF3–LiF	REO (RE = Pr, Nd, Dy)	1263–1353 K	4–8 A/cm2	76.59%	93.20–98.12%
Pr–Nd–(10–15 wt.%) Gd [122]	17.85 wt.% GdF3–69.03 wt.% (Pr, Nd)F3–13.12 wt.% LiF	REO (RE = Pr, Nd, Dy)	1283–1373 K	5 - 9 A/cm2	75.50–79.60%	93.20–98.00%

. Based on the summary of the parameters in the table, it can be seen that REF₃–LiF is the primary electrolyte composition utilized in China’s industrial manufacturing of the fluorine salt–oxide electrolytic system. Due to the stable valence states and diverse applications of La and Nd metals, numerous factories have opted for direct electrolysis to produce these metals, aiming to synthesize individual light rare earth elements. Cerium, a rare earth element, is a variable valence metal, and the current efficiency of direct electrolysis to prepare isolated metal cerium is relatively low. The method of directly electrolyzing the target cerium alloy on the active electrode not only shortens the process flow but also reduces energy consumption and improves current efficiency. The decomposition potential of rare earth oxides is similar to that of MgO and Al₂O₃, so in the preparation of RE–Mg alloy and RE–Al alloy, co-electrodeposition can be used to form uniformly alloyed rare earth alloys on the inert electrode. Due to the significant difference in decomposition potentials between NiO and RE₂O₃, specifically with the decomposition potential of RE₂O₃ being approximately −2.5 V and that of NiO being −0.25 V (vs. H⁺/H₂), Ni-based alloys are primarily produced through non-co-electrodeposition methods. By changing the ratio of electrolyte composition, finding a suitable electrolysis temperature and oxide feeding speed, and improving the structure of the electrolytic cell and the material and pole distance of the anode and cathode, the quality and current efficiency of the product are continuously improved, the production cost is continuously reduced, and the production environment is improved.

3.3.2. Treatment of Anode Gases

While the most common method of producing light rare earth metals is electrolytic rare earth oxides using the fluoride salt system, the industrial graphite anode releases greenhouse gases such as CO₂ [123], CF₄, and C₂F₆, which seriously pollute the climate [124]. Compared with aluminum electrolysis, the preparation of rare earth metals by fluoride–oxide salt electrolysis is carried out at higher temperatures, and fluoride rare earths have stronger erosion properties. Therefore, it is more difficult to study the anode reaction process of the fluoride–oxide system [125,126].

Taking the electrolysis of metallic neodymium as an example, Liu et al. [127] found that the generation of PFCs (perfluorinated compounds) was related to higher anodic current density. At lower current densities, carbon was only oxidized into CO₂ gas. When the current density exceeded a critical value, fluorocarbon compounds such as CF₄ and C₂F₆ were generated in addition to CO₂. M. Gibilaro et al. [128] studied the gas generated at the graphite anode in a molten salt system consisting of 30 wt.% LiF–70 wt.% NdF₃–1.4 wt.% Nd₂O₃ at 1173 K using infrared spectroscopy and electrochemical techniques. They found that the critical current density for the generation of C_xF_y gases was 220 mA/cm², and the production of C_xF_y gases was related to the diffusion of NdOF₅⁴⁻ in the molten salt. Zhang et al. [129] found that the generation of PFCs was related to the content of Nd₂O₃. When the content of Nd₂O₃ in the molten salt was low, the NdF₃ participated in the discharge, resulting in the anode effect. The appropriate adjustment of the cell voltage could alleviate the occurrence of the anode effect, thereby reducing the generation of gases such as CF₄ and C₂F₆ [130]. H. Vogel et al. [131,132] studied the relationship between different Nd₂O₃ concentrations in molten salt and the current density limit, and found that the content of Nd₂O₃ in the molten salt should be controlled between 0.5 wt.% and 1.5 wt.%. Du et al. [133] replaced the traditional graphite anode with an inert anode that was sintered from powders of neodymium oxide and copper oxide. This modification not only significantly improved the wettability and penetration of the electrolyte into the anode, thus reducing the occurrence of anode effects, but also effectively reduced the carbon content in the final product of neodymium metal. Zhu et al. [134] synthesized three kinds of ultra-microporous Al–MOFs materials for adsorbing CF₄ and C₂F₆ gases generated at the anode. Among them, Al–Fum has a superior CF₄ adsorption capacity (2.10 mmol/g) and CF₄/N₂ selectivity (16.5) and shows a record-high C₂F₆ uptake (3.30 mmol/g) and C₂F₆/N₂ selectivity (299.6).

C_xF_y gases have the potential to be harmful to human health in addition to contributing to the impact of the greenhouse effect on the environment. For example, C₂F₆ can cause suffocation and C₃F₆ can cause dizziness [135]. Under high-temperature conditions, the CF₄ and C₂F₆ gases produced at the anode may react with ambient oxygen to form corrosive and poisonous gasses like COF₂, which directly endangers human health [136]. The anode current density and the concentration of neodymium oxide in the molten salt system are the main causes of the production of C_xF_y greenhouse gases. Consequently, it is possible to minimize the production of C_xF_y greenhouse gases by carefully regulating the Nd₂O₃ content and electrolysis temperature as well as by utilizing creative optimizations in the design of the electrolysis cell and the anode materials.

3.3.3. Design of Rare Earth Cell

According to the configuration of cathode and anode, the rare earth molten salt electrolytic baths can be divided into the anode and cathode up plug rare earth cell [61,137,138,139,140,141,142] and the bottom cathode rare earth cell [137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156]. The structure of the anode and cathode up plug rare earth cell type (as shown in Figure 10a,b) is a vertical arrangement of parallel electrodes on the column surface, with the top of the electrolyzer open, a graphite arc as the anode, and a cathode arranged in the center. At present, this cell type is widely used in domestic rare earth smelting plants in China. With the progression of electrolysis, the graphite arc anode and inert cathode are gradually consumed, resulting in significant changes in the anode current density and electrode gap, which leads to significant variations in cell voltage and electrolyte temperature. The furnace condition is not easy to stabilize, resulting in unstable product quality and a low product qualification rate, with carbon impurities exceeding the standard [61]. Among the cathodic corrosion prevention methods, it was a common practice to install a protective casing on the cathode above the molten salt surface to isolate the tungsten cathode from external gases. However, the use of the cathodic protective casing could lead to an increase in the cell voltage and an increase in energy consumption [137]. The water-cooled anode conductor plates were cooled at the upper part of the electrolysis cell, causing the formation of a crust of molten salt on the upper part of the cell body, thereby prolonging the lifespan of the furnace [138,139]. Researchers at Northeastern University [140,141,142] also designed a covered upper-inserted cathode rare earth electrolysis cell, which adopted a concave arcuate anode at the bottom and a convex inert cathode at the top. This design aimed to achieve low anode current density and high cathode current density, reduce the high-temperature corrosion of the lining at the bottom of the electrolysis cell, and also minimize the formation of furnace accretions. Considering the production of neodymium metal as an example, traditional electrolytic cell production consumes about 7–8 kWh/kg of electricity to produce one ton of neodymium metal, and the energy consumption is relatively high [29].

The schematic diagram of the bottom cathode rare earth cell is shown in Figure 10c, where the cathode is located at the bottom of the electrolysis cell and can be either a solid cathode or a liquid cathode made of rare earth metals or rare earth alloys. The use of the bottom-mounted cathode for the preparation of rare earth metals can significantly reduce the cell voltage, achieve higher electrical energy efficiency, and lower power consumption. Additionally, it can significantly increase the anode utilization rate, reduce the consumption of graphite, lower processing costs, and have significant potential for energy conservation and emission reduction [60,61,143,144,145,146,147]. Wang et al. [148,149,150] conducted a calculation of the thermal balance within a 10 kA bottom cathode structure electrolysis cell and used finite element ANSYS software (https://www.ansys.com/) to numerically simulate the electrical field within the cell. The dimensions of the cell were as follows: an inner diameter of 1280 mm, an outer wall of 1580 mm, and a height of 450 mm, with a graphite anode radius of 1160 mm. Through the calculation, they observed that the electric field distribution within the rare earth electrolysis cell with a bottom cathode structure was uniform, and the calculated maximum cell voltage was 5.41 V, which was nearly halved compared to the current rare earth electrolysis cell voltages. Similarly, numerical simulations [151,152] were used to analyze the influence of the cathode–anode distance and anode inclination angle on metal yield and molten salt flow within a 60 kA bottom cathode rare earth electrolysis cell. The optimal setting for the distance between the bottom cathode and the anode was determined to be 110 mm [153], and the optimal anode inclination angle was determined to be 11 [154]. Narrow slots with a width range of 20 mm [152] were symmetrically opened at the center of the bottom surface of the anode, which could promote the discharge of anode gases, improve current efficiency, and enhance product quality. However, challenges such as the secondary oxidation of rare earth metals and the collection of liquid cathodes at the bottom were also observed in the bottom cathode rare earth electrolysis cells. Experimental results showed that the production of rare earth metals in the bottom cathode rare earth electrolysis cells could be increased by 1.3 times compared to the anode and cathode up plug rare earth cell, and the current efficiency could be increased from 65% to 85% [155].

At present, the industrial-grade purity of rare earth metals obtained from the molten salt electrolysis technique is approximately 99.50%, with the primary contaminants found in the products being C, Fe, and Si. In samples, the Fe impurities content does not exceed 0.2 wt.%, and the C and Si contents do not exceed 0.03 wt.% [156]. Graphite tanks and graphite anodes have a major impact on carbon pollution; small amounts of SiO₂ in raw materials and the peeling of refractory materials are the primary sources of Si impurities, and iron chips falling into cathode conductive rods are the main source of metal Fe introduction. To improve the purity of the product, optimization can be made from the following two aspects: first, the purity of the raw material should be improved to eliminate the introduction of impurities at the source; second, the production process should be optimized through measures such as appropriately reducing the electrolysis temperature, adjusting the distance between the anode and cathode, and introducing a siphon device [157,158,159,160,161], which can reduce the carbon content in the product and the corrosion and spalling of the cathode guide rod iron bar. In order to increase the solubility of rare earth oxides and the yield of rare earths, it is necessary to conduct systematic theoretical research on the rare earth molten salt electrolysis process. Additionally, choosing an appropriate electrolysis temperature can reduce energy consumption as well as the amount of carbon in the final product. Lastly, improving the structure of the electrolysis tank can increase current efficiency. At the end of 2024, the facility recently constructed by Ganzhou Chenguang Rare Earth New Materials Co., Ltd. will be operational, and the factory area will be equipped with a cathode automatic lifting device and siphon equipment. This has important guiding significance for the improvement of industrial rare earth molten salt electrolytic cells, and the electrolytic process using liquid metal as the cathode will be the mainstream technology for rare earth molten salt electrolysis in the future.

4. Prospects

Rare earth metals are very important strategic resources, with very wide applications in defense, industrial manufacturing, new energy, information technology, and other fields. For preparing rare earth metals and alloys, molten salt electrolysis is an important technology. With the improvement in scientific and technological levels and the increasing demand for rare earth metals and alloys, prospects for the application of molten salt electrolysis in the production of rare earth metals becomes broader, mainly reflected in the following: (1) More measures should be taken to improve production efficiency and lower costs with molten salt electrolysis. The development of new electrolytic cells and electrolysis processes can further improve production efficiency and reduce energy consumption. (2) Molten salt electrolysis also has a broader application prospects in the preparation of rare earth alloys due to the growing demand for new rare earth alloys. By adjusting electrolysis conditions and adding appropriate alloy elements, rare earth alloys with specific properties can be prepared. Research on the production of rare earth metals and their alloys by molten salt electrolysis has made important progress but still faces challenges and difficulties. It is believed that with the continuous improvement in scientific and technological levels and the deepening understanding of these processes, the production of rare earth metals by molten salt electrolysis will become more and more efficient, environmentally friendly, and economical, making a greater contribution to the rapid development of the rare earth metals industry.