In-Situ X-ray Diffraction/Fluorescence Analysis of Dy–Cu Electrochemical Alloying and Dealloying in Molten LiCl–KCl–DyCl3

Abstract

Listen

Translate article

1. Introduction Recycling spent Nd magnets is important for guaranteeing a stable rare earth (RE) supply in Japan. As a new recycling process of Nd magnets, the authors have studied a one-step method using molten salt and alloy diaphragms [1-3], where Dy or Nd is selectively permeated by controlling the potential of the alloy diaphragms. This technique is based on the rapid electrochemical formation of RE–iron group (IG) alloys in molten salts. To investigate the mechanism of the rapid alloy formation, the authors recently conducted in situ energy-dispersive X-ray diffraction measurement for the electrochemical alloying and de-alloying of Dy–Ni in molten LiCl–KCl [4,5]. Through these results, new insights into Dy–Ni alloying and de-alloying were obtained. However, the reason why the electrochemical formation of RE–IG alloys proceed so rapidly is still unclear. Therefore, we investigated the mechanism by comparing Ni with the other metal, Cu, which has a similar crystal structure (fcc) to Ni, and its alloy formation rate with RE is relatively high [6]. In this study, in situ X-ray fluorescence (XRF) measurement of Dy–Cu alloying and de-alloying in molten LiCl–KCl was performed and the alloying/de-alloying behaviors of Dy–Cu and Dy–Ni were compared. 2. Experimental The experiments were conducted in a similar manner to the previous studies [4,5] at the BL28B2 beamline of SPring-8. 35.0 g of LiCl–KCl (44 : 56 wt% eutectic composition, 35.0 g) was set in a graphite crucible inside a glovebox with an Ar atmosphere. Pre-electrolysis was conducted before the experiment. An electric furnace with windows was fixed on the transition stage. Electrolysis was conducted in Pyrex vessel at 723 K in a dry Ar atmosphere. After potential calibration by the Li+/Li equilibrium potential using a Mo electrode, DyCl3 was added to the melt up to 0.5 mol%. A Cu plate (5 mm × 15 mm × 0.1 mmt), an Ag+/Ag, and the graphite crucible were used as the working, reference, and counter electrodes, respectively. After the measurement of cyclic voltammetry, the potentiostatic electrolysis was conducted using the same Cu plate for 180 min at 0.5 V vs. Li+/Li and then, potential was shifted to 1.5 V vs. Li+/Li and kept for 30 min. XRF measurements were started simultaneously with electrolysis. The incident beam size was 0.1 mm × 2.0 mm and irradiated to the Cu plate in the molten salt. The take-off angle was set at 5°. Transmitted X-rays were consecutively detected with the integration time of 120 s. 3. Results and Discussion The X-ray fluorescence peaks of Dy Kα1, Kα2, and Kβ1 were confirmed. Fig 1 shows peak area–time data for X-ray fluorescence of Dy Kα1 during Dy–Cu alloying and de-alloying. The peak area increased as the alloying progressed, which seemed to correspond to the formation and growth of Dy–Cu alloy layer. The peak area slightly decreased during de-alloying. Compared to Dy–Ni alloying/de-alloying [5,6], the intensities of fluorescence peaks were weak. These results suggest that the was alloy formation rate is smaller in the case of Cu than that of Ni. Acknowledgement This work was partly supported by the New Energy and Industrial Technology Development Organization (NEDO) of Japan. Synchrotron radiation studies were performed at the BL28B2 beamline of SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI, Proposal No. 2022B1338).