Electrochemical Formation of Dy-Ni Alloys in a Molten Cesium Chloride System

Abstract

Dy-added Nd-Fe-B magnets are widely used for high-performance motors in battery electric vehicles (BEVs), hybrid electric vehicles (HEVs) and so on. The demand for these magnets is increasing due to the spread of BEVs and HEVs. However, Dy resources are scarce and unevenly located in the earth crust. Thus, it is necessary to develop an inexpensive and effective recovery/separation process of Dy from scraps of the Dy-added Nd-Fe-B magnet. From this background, we have been investigating a new recovery process of Dy and Nd from the magnet scraps using molten salt electrolysis and alloy diaphragms [1-3]. The key step in this process is the alloying and de-alloying of rare earth (RE) elements on both sides of the diaphragm in order to effectively separate them in the molten salt. So far, we have reported that Dy can be selectively alloyed with Ni by potentiostatic electrolysis in a molten LiCl-KCl-DyCl3(0.50 mol%)-NdCl3(0.50 mol%) system at 723 K [2]. The selective permeation of Dy through the RE-Ni alloy diaphragm from an anolyte containing DyCl3 and NdCl3 to a catholyte has been also confirmed [3]. However, the current density was of the order of 10 mA/cm2 or less, which is insufficient for industrial electrolysis. So, we focused on the CsCl system as a candidate melt to achieve a higher current density due to its high operation temperature. Since the alloying and de-alloying of REs in this system is unknown, the electrochemical formation of Dy-Ni alloys was investigated in a molten CsCl-DyCl3 system at 973 K as a first step in the investigation of this system.All experiments were conducted under a dry Ar gas atmosphere at 973 K in a glove box. Mo, Ni wires and Ni plates were used as working electrodes. A glassy carbon rod was used as a counter electrode. The reference electrode was a Ag wire immersed in a molten CsCl containing 1 mol% of AgCl. All the potentials are referred to the Cs+/Cs electrode potential. The alloy samples were prepared by potentiostatic electrolysis. After electrolysis, the samples were analyzed by XRD and were also observed by SEM-EDS. First, cyclic voltammetry was conducted with a Ni wire electrode in a molten CsCl-DyCl3(0.50 mol%) system. During the cathodic potential scan, a large cathodic current was observed from 0.50 V. Since this potential was more positive than that of Dy metal deposition on a Mo wire electrode, the cathodic current was suggested to correspond to the formation of Dy-Ni alloys. Then, open-circuit potentiometry was carried out with a Mo electrode after electrodepositing Dy metal at 0.11 V for 150 s. A potential plateau was observed at 0.26 V. Since Mo does not form any intermetallic compounds with Dy, the observed potential is interpreted as the Dy3+/Dy potential. Also, open-circuit potentiometry was performed with a Ni electrode after depositing Dy metal at 0.18 V for 90 min. As a result, four potential plateaus were observed at 0.29, 0.42 V, 0.52 V, 0.77 V and 0.90 V, which possibly correspond to the potentials of two-phase coexisting states of different Dy–Ni alloys. Based on these results, potentiostatic electrolysis was conducted at 0.38 V for 1 h with a Ni plate electrode. XRD analysis indicated the formation of DyNi2 on the Ni plate electrode. The thickness of DyNi2 layer was estimated to be 50 μm from cross-sectional SEM observations. The growth rate of this DyNi2 layer, 50 µm h-1, was higher than that obtained in a molten LiCl-KCl system at 700 K, 28 µm h-1 [4]. After the formation of DyNi2 on the Ni plate electrode, potentiostatic electrolysis was also conducted at 0.46 V, 0.66 V and 0.82 V for 1 h. These potential values were selected from the observed plateau potentials. The formed DyNi2 were transformed to other phases such as DyNi3 at 0.46 V, Dy2Ni7 at 0.66 V and DyNi5 at 0.82 V due to anodic dissolution of Dy, depending on the applied potentials.