Electrochemical Behavior of Dysprosium Ion and Its Co-Electroreduction with Nickel Ions in the Molten KCl-NaCl-CsCl Eutectic

Abstract

The electrochemical behavior of dysprosium ions, as well as dysprosium and nickel ion co-reduction, on inert tungsten electrodes and active nickel electrodes were studied in the eutectic KCl-NaCl-CsCl melt at a temperature of 823 K using the methods of cyclic and square-wave voltammetry and open circuit chronopotentiometry. The process of Dy³⁺ ions electroreduction was found to be reversible and to proceed within a single three-electron stage up to the polarization rate of 0.1 V/s. The increase in the polarization rate indicates a slower rate of the charge transfer, which causes the quasi-reversible character of the charge transfer. It is shown that when the KCl-NaCl-CsCl eutectic melt contains both nickel and dysprosium ions, the voltammetry curves at 823 K have a wave of nickel ion reduction at the potentials of −(0.22–0.28) V and a dysprosium ion reduction at the potentials of −(2.175–2.250) V relative to a chlorine-silver reference electrode. Apart from these waves, the voltammograms have two reduction waves at the potentials of −(1.9–1.95) V and −(2.05–2.1) V. These waves are associated with the reduction of dysprosium ions and their depolarization on metallic nickel, which was preliminary deposited on the tungsten electrode, as well as the formation of the intermetallic phases of dysprosium and nickel of various Dy_xNi_y compositions. The (E-t) dependencies of the open circuit chronopotentiometry elucidate plateaus of the potential delay, which correspond to the dissolution of separate dysprosium and nickel intermetallic phases. Based on the results of the voltammetry changes and the chronopotentiometry of the open circuit, a series of electrochemical syntheses were performed in the potentiostatic regime at the potentials of −(1.7–2.1) V. The intermetallic phases of DyNi₅, DyNi₃ and DyNi₂ were obtained at a definite ratio of the dysprosium and nickel chloride concentrations in the KCl-NaCl-CsCl eutectic melt and at a temperature of 823 K. The synthesized intermetallic samples were characterized by X-ray diffraction and scanning electron microscopy.

1. Introduction

Alloys and intermetallic compounds of rare earth metals are known and distinguished by their functional properties including excellent magnetic, catalytic and fluorescent properties as well as their ability to accumulate and store hydrogen.

To date, rare earth containing alloys and intermetallic compounds are mainly obtained via the methods of the thermal reduction of stoichiometric rare earth metal (REM) oxides or halides by metallic calcium or calcium hydrate [1,2,3,4,5]. At the same time, high temperature electrochemical synthesis appears to be promising for obtaining REM alloys and intermetallic compounds [6,7,8].

High temperature synthesis allows controlling the composition, morphology (powders or coatings) and size (powder size or coating thickness) by electrochemical parameters such as the electrolysis potential and current density. Konishi et al. [9,10,11] illustrate that the formation and phase composition of Dy_xNi_y intermetallic compounds is possible by a regulation of the electrolysis potential in the LiCl-KCl-DyCl₃ melt at 700 K. The authors determined that the potentiostatic electrolysis on the Ni-electrode resulted in the formation of the DyNi₂ film. The anode polarization transformed this film into other phases DyNi₃, Dy₂Ni₇, DyNi₅ and Ni. Later, Yasuda et al. [12] obtained intermetallic compounds of DyNi₂, DyNi₃, Dy₂Ni₇, DyNi₅ by the diffusion saturation of the nickel substrate (cathode) by metallic dysprosium during electrolysis. Su et al. [13] studied the co-reduction of Dy³⁺ and Al³⁺ ions in the eutectic KCl-LiCl melt. Two intermetallic DyAl₃ and DyAl compounds were obtained under potentiostatic and galvanostatic conditions. We have previously studied the mechanism of Dy³⁺ and Ni²⁺ co-reduction, and the possibility of intermetallic Dy_xNi_y compounds’ electrochemical synthesis was verified [14]. Other information on the co-electroreduction of dysprosium and nickel ions in the molten KCl-NaCl-CsCl eutectic is not available in the published literature data.

Therefore, the present paper is focused on the study of the electrochemical behavior of dysprosium ions and their co-electroreduction with nickel ions in the molten KCl-NaCl-CsCl eutectic.

2. Materials and Methods

2.1. Electrochemical Cell and Electrodes

The experiments were performed in a hermetically sealed three-electrode cell in a pure and dried argon atmosphere. To eliminate the traces of oxygen in the argon atmosphere, zirconium straws were put into the electrochemical cell as the oxygen scavenger. A glassy carbon crucible 30 cm³ in volume served as an anode and a melt container. A chlorine-silver electrode Ag|KCl-NaCl-CsCl (eutectic)-AgCl (2.5 mol.%) was used as a reference electrode; it was put into a zirconium oxide tube (zirconium oxide was stabilized by cerium oxide). A tungsten and nickel wire of 1.0 mm in diameter served as a cathode. An alundum tube coated the wire so that 40 mm of wire remained uncoated. The area of the working electrode was calculated depending on the depth of the immersion into the melt (10 ÷ 15 mm). The cell was prepared and assembled in a glove box mBraun Labstar 25 (Munich, Germany) in the purified argon atmosphere. To create a working temperature of 823 K, a shaft resistance furnace with silit rod heating was used. The automatic temperature regulation was achieved using an electron regulator OVEN-TRM-1 (Moscow, Russia) by a chromel-alumel thermocouple, and the accuracy of the temperature maintenance was ±1 °C. The temperature of the melt was additionally controlled by a chromel-alumel thermocouple in an alumooxide sleeve, immersed into the melt. Cyclic chronovoltammograms and open-circuit square-wave voltammograms were obtained using an Autolab PGSTAT 30 complex (Ecochemie, Utrecht, The Netherlands) with the IF-030 interface. Voltammograms were processed using the computer software GPES 4.9.

The Dy–Ni intermetallic samples were obtained by potentiostatic and galvanostatic electrolysis on the tungsten and nickel electrodes. The synthesized product in the majority of cases slipped from the electrode to the bottom of the glassy carbon crucible. To eliminate their dissolution (the glassy carbon crucible served as an anode) and to collect the cathode deposit, a small alumina crucible was put under the tungsten electrode.

2.2. Methods of Analysis and Characterization of Cathode Deposits

To determine phases, the elemental and granulometric composition and microstructure of cathode deposits, we used an X-ray diffractometer D2 PHAZER (Bruker, Bremen, Germany), an X-ray fluorescent spectrometer Spectrscan MAKS GV (Spectron Ltd., Saints-Petersburg, Russia), the scanning electron microscope Vega 3 LMH (TESKAN, Brno, Czech Republic) with the X-ray microanalysis system X-max (Oxford, UK) and the laser dispersion particle analyzer Fritsch Analysette 22 NanoTec («FritschGmbH», Idar-Oberstein, Germany).

2.3. Electrolyte Preparation

The KCl-NaCl-CsCl eutectic melt was chosen as a background electrolyte. Extra pure individual KCl, NaCl and CsCl salts were used. These salts were preliminarily dried in a vacuum drying oven for 10 h and then were calcined in a muffle furnace for 5 h at 450 °C. Ultradry extra pure DyCl₃ and NiCI₂ salts (Chemcraft Ltd., Kaliningrad, Russia) were used as sources of dysprosium and nickel.

3. Results

3.1. Electroreduction of Dy³⁺ in the Eutectic KCl-NaCl-CsCl Melt on the Tungsten Electrode

At first, we reproduced the results on the Dy³⁺ ions’ electroreduction in KCl (24.5 mol.%)-NaCl (30.0 mol.%)-CsCl (45.5 mol.%) obtained in our laboratory [15]. A chloride-silver reference electrode, which is considered to be more stable and resistant considering its potential rather than a glassy carbon reference electrode, was used for these measurements. A glassy carbon reference electrode is a quasi-reference electrode that reacts and records the redox potential of the molten media, and it is resistant to the changes in the molten salt composition. The measurements performed using a chlorine-silver reference electrode are needed to provide a more accurate determination of the electroreduction and dissolution of the different phases of intermetallic nickel and dysprosium compounds. Apart from the method of chonopotentiometry, square-wave voltammetry and open circuit chronopotentiometry were used for a more detailed determination of the electrochemical behavior of Dy³⁺ ions and the mechanism of Dy³⁺ ion electroreduction with nickel ions in the molten KCl-NaCl-CsCl eutectic at 823 K.

Figure 1 illustrates the voltammetry curves of the dysprosium ion electroreduction on the tungsten electrode in the molten KCl-NaCl-CsCl eutectic at 823 K relative to the chlorine-silver reference electrode. The cyclic voltammograms illustrate a clearly observed wave of Dy³⁺ ion reduction up to the potential of the background electrolyte decomposition. The increase in the peak height is characteristic for these waves as the dysprosium trichloride concentration in the melt increases and the polarization rate grows. Generally, the waves of Dy³⁺ ion reduction have more pronounced diffusion peaks than the current reduction waves in the equimolar KCl-NaCl melt [14,16]. A cathode product oxidation wave is observed on the anode branch of the cyclic voltammogram as clearly as the cathode wave. The potential of the Dy³⁺ ion electroreduction in the molten KCl-NaCl-CsCl eutectic at 823 K is shifted to a more negative region than in the equimolar KCl-NaCl melt at 973 K. The latter is likely to be due to the increase in the stability of the neodymium and dysprosium complexes in the molten KCl-NaCl-CsCl eutectic as compared to their stability in the molten KCl-NaCl eutectic.

To determine the character of the electrode process at the Dy³⁺ ion electroreduction, we analyzed the nonstationary voltammetry dependencies according to diagnostic criteria [17,18,19].

Table 1. Values of some parameters of Dy³⁺ ion electroreduction in the KCl-NaCl-CsCl eutectic melt at a temperature of 823 K. C(DyCl₃) = 2.0 × 10⁻⁴ mol/cm³.

υ, V/s	Ip, mA/cm2	ip/υ1/2 mA·s1/2/cm2V1/2	Eкp V	Eкp/2 V	$\Delta$ E V	n (αnα)
0.01	43.8	438.7	−2.221	−2.167	0.054	2.9
0.02	61.4	434.3	−2.262	−2.210	0.052	3.0
0.05	95.95	429.11	−2.274	−2.221	0.053	2.9
0.1	131.6	416.1	−2.295	−2.242	0.053	2.9
0.2	172.1	384.4	−2.304	−2.241	0.063	2.5
0.5	260.5	368.4	−2.292	−2.220	0.072	2.2
1.0	354.1	354.1	−2.378	−2.294	0.084	1.8

illustrates the results of the voltammetry dependency analysis and the calculated values of the Dy³⁺ ion electroreduction on the tungsten electrode in the molten KCl-NaC-CsCl eutectic at a temperature of 823 K. The values of these parameters are in good agreement with the literature data [15,20]. Figure 2 demonstrates the function of the Dy³⁺ ion electroreduction υ^1/2 to i_p/υ^1/2.

At the polarization rates exceeding 0.1 V/s, the i_p/υ^1/2 value decreases as the υ^1/2 value increases. This fact verifies the transfer to a quasi-reversible character of the Dy³⁺ ion electroreduction. At the polarization rates exceeding 0.5 V/s, the i_p/υ^1/2 relation is weakly dependent on the υ^1/2 value, and in this region, the process of electroreduction is controlled by the charge transfer rate. The values of the Dy³⁺ electroreduction wave half peak were used to calculate the number of electrons transported during the electrode process using the following equation:

$$\Delta \mathrm{E}=Ep-E_{\frac{p}{2}}=2. 2 RT/nF$$

(1)

Within the 0.01 ÷ 0.1 V/s interval of polarization rates, the value of n = 3.0 and as the polarization rate increases, the αn_α value decreases (see Table 1).

The values of the i_p/υ^1/2 relation at the polarization rates below 0.1 V/s (Figure 2), when diffusion is the limiting stage and the electrode process is reversible according to the Berzins–Delahay equation, used for the diffusion-controlled process, which includes the deposition of the metallic dysprosium solid phase at the cathode, were used to calculate the coefficients of DyCl₆³⁻ ion diffusion in the molten KCl-NaCl-CsCl eutectic at 823 according to the following equation:

I_p/υ^1/2= 0.446(nF)^3/2 (RT)^−1/2 CD^1/2

(2)

The calculated values of the DyCl₆³⁻ (0.49 × 10⁻⁵ cm²·s⁻¹) ion diffusion coefficients are in good agreement with the values reported in [15,20] for the KCl-LiCl eutectic at 773 K and KCl-NaCl-CsCl at 823 K.

A single stage character of the Dy³⁺ ion electroreduction is verified by the square-wave voltammogram recorded on the tungsten electrode in the molten KCl-NaCl-CsCl eutectic containing DyCl₃ (2 × 10⁻⁴ mol/cm³) (Figure 3).

A clear peak E_c associated with a decrease in the Dy³⁺ ion concentration is observed in the same potential range as at the cyclic voltammogram. The current–potential curve has a bell shape, and the W_1/2 peak width at half height is presented by Equation (3) [21,22].

W_1/2 = 3.52 RT/nF

(3)

The average value of W_1/2 being equal to 81 mV was obtained by three square-wave voltammetry measurements and according to Equation (3), n = 3.1 ± 0.1 at 823 K, which verifies that the Dy³⁺ ion electroreduction in the molten KCl-NaCl-CsCl eutectic is a one-stage process with a transfer of three electrons. Open circuit chronopotentiometry (turn on–off curves) may be used to evaluate the Dy³⁺/Dy electrode potential and to determine the deposition potential of metallic dysprosium on a tungsten electrode in the molten KCl-NaCl-CsCl eutectic at 823 K. Figure 4 shows open circuit chronopotentiometry curves after a galvanostatic pulse at the current densities ranging from 0.05 to 0.08 A/cm² and at a time ranging from 5 to 15 s. When the galvanostatic pulse is turned on, the potential of the tungsten electrode in KCl-NaCl CsCl-DyCl₃ (2 × 10⁻⁴ mol/cm³) immediately shifts from the stationary potential −(0.08–0.1) V to −(2.3–2.4) V. On the turn-off curves, there is a potential delay −(2.098–2.130) V of various time intervals that is attributed to the Dy³⁺/Dy potential. After the plateau, the potential, within a short time interval of 25–40 s, shifts to the initial stationary potential of the tungsten electrode in the KCl-NaCl-CsCl-DyCl₃ (2 × 10⁻⁴ mol/cm³) melt. The turn-off curves in the region of −(1.5–1.55) V potentials have a small band, which is probably associated with the formation of a W-Dy alloy at the tungsten electrode surface.

3.2. Co-Electroreduction of Dy³⁺ and Ni²⁺Ions on the Tungsten Electrode in the Molten KCl-NaCl-CsCl Eutectic at 823 K

To determine the possibility of implementing the co-electroreduction of dysprosium and nickel ions, we carried out voltammetry measurements of the molten KCl-NaCl-CsCl eutectic containing dysprosium and nickel ions. Three different electrochemical measurements were carried out. In the first variant, a certain amount of nickel chloride was added to the KCl-NaCl-CsCl eutectic melt and the voltammetries of the Ni²⁺ nickel ion electroreduction were recorded. Then, dysprosium chloride DyCl₃ was added to this melt, and the voltammetry dependence was recorded in the Ni²⁺ and Dy³⁺ ions containing the KCl-NaCl-CsCl background electrolyte. In the second variant, voltammetry measurements were carried out in the reverse order. According to the third option, where the working electrolyte was prepared in a dry box, the calculated amounts of precursors were weighed: the KCl-NaCl-CsCl background electrolyte, dysprosium chloride and nickel chloride. The weighed portions were mixed in a dry form, poured into a glassy carbon crucible and melted in a sealed three-electrode cell in an argon atmosphere with a constant increase in temperature up to 823 K.

The experiments performed show that the working electrolyte preparation sequence does not affect the nature of the voltametry dependencies of the co-electroreduction of dysprosium and nickel ions, and, therefore, in general, the experiments were carried out according to the third variant. Figure 5 shows cyclic chronovoltammograms of the co-electroreduction of the chloride complexes of dysprosium and nickel. With the combined content of dysprosium and nickel ions in the KCl-NaCl-CsCl melt at 823 K, the voltammetry dependencies show a reduction wave of nickel ions (wave A) at the potentials of −(0.22–0.28) V as well as that of dysprosium ions (wave D) at potentials −(2.175–2.250) V. In addition to these waves, the voltammetry curve has two recovery waves (wave C) at potentials −(2.05–2.1) V and wave B at potentials −(1.9–1.95) V. Wave D corresponds to the process of pure metallic dysprosium extraction. We associate the appearance of the C and B waves on the voltammetry dependence with the reduction of dysprosium ions accompanied with the depolarization on metallic nickel previously deposited on the tungsten electrode (wave A). At these waves, the formation of intermetallic phases of dysprosium and nickel, with different phase compositions of Dy_xNi_y, occurs. At wave E, an alkali metal is released. On the anode branch of the cyclic voltammogram, five waves, A’, B’, C’, D’ and E’, are also observed. The cathode reduction waves correspond to the anode waves of the cathode product’s electro-oxidation, which is verified by the voltammetry curves recorded up to different reverse potentials corresponding to the completion of each reduction wave (Figure 5, curves 2–4).

Figure 6 shows the chronopotentiograms of the molten system KCl-NaCl-CsCl-NiCl₂ (0.5 × 10⁻⁴ mol/cm³) − DyCl₃ (3.0 × 10⁻⁴ mol/cm³) at various current densities at 823 K. The chronopotentiograms showed the presence of two potential delays. The first one, at the potentials of −(0.15–0.25) V, corresponds to the electroreduction of nickel ions, and the second one at the potentials of −(1.875–2.125) V, corresponds to the co-electroreduction of dysprosium and nickel ions.

In contrast to the chronovoltammograms, where three electroreduction waves corresponding to different processes appear, the chronopotentiogram fails to reveal the presence of stages in the process of the dysprosium and nickel ion co-electroreduction. However, the square-wave voltammogram of the KCl-NaCl-CsCl-NiCl₂ (0.5 × 10⁻⁴ mol/cm³)-DyCl₃ (3.0 × 10⁻⁴ mol/cm³) melt at a frequency of 25 Hz, shown in Figure 7, clearly demonstrates the presence of three stages in the overall process of the co-electroreduction of dysprosium and nickel ions. These data are consistent with the results obtained by cyclic voltammetry.

Open circuit chronopotentiometry is also a suitable technique to confirm the formation of intermetallic compounds in electrolysis processes in molten salt systems and to calculate the Gibbs energies of their formation [23,24,25,26]. On the E-t dependence (Figure 8), when the galvanostatic current pulse is turned off, in addition to the delay in the dissolution potential of metallic dysprosium—(2.063 ± 0.02) V, four more potential delays are observed, which are in a more positive region than the equilibrium potential Dy^3+/Dy⁰.

The first delay is −1.854 V, the second delay is −1.733 V, the third delay is −1.612 V and the fourth delay is (0.30–0.25) V. After these delays, the electrode potential gradually shifts to the stationary potential of the tungsten electrode relative to the silver chloride electrode. The first three potential delays correspond to the dissolution of the Dy_xNi_y intermetallic phases of various compositions. Moreover, the more positive the delay potential is, the more enriched with nickel (more electropositive metal) the composition of the intermetallic phase is. At low current densities of a galvanostatic pulse, the polarization of the electrode does not reach the potential for the pure metallic dysprosium phase release (Figure 9, curve 1) and the potential delay on the turn-off curves is not so clearly expressed. At these current densities, the potential drop to the stationary potential of the tungsten electrode occurs during small time intervals of approximately 15 s. With an increase in the density of the polarization current, the time of the potential plateau increases significantly, and their clear separation is observed. Up to a current density of a galvanostatic pulse of 0.1 A/cm², intermetallic phases are formed, while the free dysprosium phase is absent in the cathode deposit. Electrolysis at current densities above 0.1 A/cm² leads to the production of pure metallic dysprosium along with intermetallic phases. An increase in the duration of a galvanostatic pulse at a constant current density increases the duration of the potential delay corresponding to each phase of the intermetallic compound (Figure 9). So, with a galvanostatic pulse duration of 20 s and a current density of 0.2 A/cm², the time to reach the stationary potential of the tungsten electrode increases significantly and amounts to 250–300 s (Figure 9, curve 4). This time also depends on the value and ratio of the concentration of the dysprosium and nickel chlorides in the melt.

3.3. Co-Electroreduction of Dy³⁺ and Ni²⁺ Ions at the Nickel Electrode in the Molten KCl-NaCl-CsCl Eutectic at 823 K

Figure 10 shows the voltammetry dependencies recorded on a nickel electrode in a molten mixture of KCl-NaCl-CsCl-DyCl₃ (3.0 × 10⁻⁴ mol/cm³) − NiCl₂ (0.5 × 10⁻⁴) mol/cm³). As can be seen, the voltammetry dependence has four reduction waves. On wave A at the potentials of −(0.00–0.06) V, Ni²⁺ ions are reduced on the nickel electrode, and on the corresponding wave A’, the metallic nickel is oxidized with the formation of Ni²⁺ ions. Waves B, C and D illustrate the electroreduction of DyCl₆³⁻ ions on a nickel electrode with a certain depolarization with the formation of Dy_xNi_y intermetallic phases.

xDy³⁺ + yNi²⁺ + (3x+2y)e = Dy_xNi_y

(4)

The corresponding waves B’, C’ and D’ on the anode branch of the cyclic voltammogram are associated with the anode dissolution of the more electronegative dysprosium metal from the intermetallic phase.

The square-wave voltammogram of the KCl-NaCl-CsCl-NiCl₂ (0.5 × 10⁻⁴) − DyCl₃ (3 × 10⁻⁴) melt clearly shows the presence of three stages in the overall process of the dysprosium and nickel ion co-electroreduction (Figure 11). Moreover, the potentials of the reduction waves on the chronovoltammogram and the square-wave voltammogram coincide clearly.

Open circuit chronopotentiograms (Figure 12) illustrate that, when the nickel cathode is polarized by a galvanostatic current pulse with a density less than 0.1 A/cm², the electrode polarization does not reach the reduction potential of dysprosium ions. When the current is turned off, the potential of the nickel cathode shifts to the equilibrium potential Ni^2+/Ni in the eutectic melt KCl-NaCl-CsCl within a short period of time (1.0–2.0 s).

As the current density of a galvanostatic pulse (i > 0.1 A/cm²) increases and the pulse time remains unchanged (10 s), three potential delays are observed on the open circuit chronopotentiograms according to the values corresponding to the potentials of the B, C and D waves on the chronovoltammograms. In this case, as the density of the polarizing current increases, the delay time increases (curves 2, 3 and 4, Figure 12). At a polarizing current density of 0.2 A/cm², the turn-off curves show a potential delay in the more negative potential range of −(2.200–2.210) V corresponding to the Dy^3+/Dy electrode potential. The duration of potential delays at a constant current density depends on the galvanostatic pulse time (Figure 13).

Potentiostatic Electrolysis and Deposit Characterization

According to the phase diagram of the Dy–Ni binary metal system [27,28], dysprosium and nickel form a series of intermetallic compounds Dy₃Ni, Dy₃Ni₂, DyNi, DyNi₂, DyNi₃, Dy₂Ni₇, DyNi₄, Dy₄Ni₁₇, DyNi₅ and Dy₂Ni₁₇. Among all the phases, DyNi₅ and DyNi are congruently melting compounds. The remaining phases are incongruently melting compounds formed by a peritectic reaction. Electrochemical formation of the intermetallic phases DyNi₂, DyNi₃, Dy₂Ni₇ and DyNi₅ in the molten KCl-NaCl-DyCl₃ (0.5 mol.%) system at 973 K and the LiCl-KCl-DyCl₃ (0.5 mol.%) system at 673–773 K was confirmed by the authors [9,12,14] using scanning electron microscopy and X-ray diffraction measurements. From the EMF measurements of the intermetallic compounds in the Dy/Ni system, the authors of [29] calculated the standard Gibbs free energies of formation for DyNi₂, DyNi₃, Dy₂Ni₇ and DyNi₅ in the temperature range of 673–773 K.

Based on the results of cyclic voltammetry, square-wave voltammetry and open circuit chronopotentiometry, potentiostatic electrolysis was carried out at the temperatures of 823–973 K on tungsten and nickel electrodes in the molten KCl-NaCl-CsCl-DyCl₃-NiCl₂ system via potentiostatic electrolysis at potentials up to −0.8 V relative to the silver chloride reference electrode for 2 h. According to the X-ray phase analysis, the electrolysis product is pure metallic nickel (Figure 14) with a face-centered cubic structure (Fm-3 m) (PDF in the XRD database: 04–0850). For the electrochemical synthesis of the Dy_xN_iy intermetallic compounds, potentiostatic electrolysis was carried out at more negative potentials, −1.7 V, −1.9 V and −2.1 V, on tungsten and nickel electrodes at temperatures ranging from 823 to 973 K for 2 h in the KCl-NaCl-CsCl-DyCl₃ (3.0 × 10⁻⁴ mol/cm³)-NiCl₂ (0.3–0.5) × 10–10⁻⁴ mol/cm³) melt. As a result of the electrolysis on the tungsten electrode (d = 1.0 mm) and on a nickel plate, volumetric deposits were obtained in the form of a metal-salt “pear”.

According to the X-ray phase analysis, the electrolysis product at the potential of −1.7 V consisted mainly of the DyNi₅ phase, a small amount of metallic nickel and the DyNi₃ phase (Figure 15). The SEM and EDS analysis of the cathode deposit (Figure 16) shows that the Dy element is well dispersed in the intermetallic phase. The results of the quantitative EDS analysis (spectra 14, 15 and 16) sampled at different zones show that the deposits are composed of the Ni and Dy elements. The average content of Ni and Dy atoms (wt.%) in the sample is shown in the EDS spectra. The compositions obtained by EDS differ from the composition of DyNi₅ found by XRD. To study the distribution of Dy and Ni elements in Dy–Ni alloys, an element mapping analysis was used.

The EDS-mapping analysis (Figure 16) of the elements shows that Dy is well dispersed in the Dy–Ni alloy.

With electrolysis at a more negative potential of −1.9 V, as can be seen from X-ray patterns (Figure 17), the cathode deposit consisted mainly of the DyNi₃ and DyNi₂ phases. Electrolysis at a potential of −2.1 V led to the formation of a phase richer in dysprosium content, the DyNi₂ phase and the containment of a small amount of DyOCl and Dy₂O₃ (Figure 18). The formation of DyOCl and Dy₂O₃ is explained by the incorporation of DyCl₃ into the intermetallic pores. When the cathode deposit is leached, DyCl₃ is easily hydrolyzed to form oxychloride and oxide. A similar phenomenon in the preparation of Ce-Al alloys was previously discovered by Liu et al. [30] and Castrillejo et al. [31,32] in the preparation of Gd-Al and Sm-Al alloys. Intermetallic compounds DyNi₃ and DyNi₂ were obtained both with a cubic body-centered Fd-3 m structure and a hexagonal lattice (PDF in the XRD database: 65–5536).

Therefore, based on the results of the cyclic voltammetry, square-wave voltammetry and open circuit chronopotentiometry as well as potentiostatic electrolysis, the potentials of metallic nickel and dysprosium extraction on the inert tungsten and active nickel electrodes in the KCl-NaCl-CsCl-DyCl₃-NiCl₂ molten system differ by 1.5 V. This is why the electrochemical synthesis of the nickel and dysprosium intermetallic compounds is possible in the kinetic regime [6,7,8]. Under such conditions, the composition of the cathode deposit will depend on the electrolyte composition, current density and electrolysis potential. At the electrolysis current densities below the limiting diffusion current of nickel ion electroreduction or at potentials more positive than 1.2 V, the electroreduction of nickel takes place. At potentials more negative than −1.5 V and at current densities exceeding the limiting diffusion current of nickel ion electroreduction, Dy_xNi_y intermetallic compounds will be formed on the nickel electrode or on the preliminary deposited on the tungsten electrode metallic nickel due to the Dy³⁺ ion electroreduction depolarization. The intermetallic compounds are formed according to the reaction diffusion of dysprosium atoms in the metallic nickel structure. In addition, the DyNi₅ phase with a high concentration of nickel forms first, as the free Gibbs energy of formation of this phase has a smaller value among the intermetallic phases in the Dy–Ni system.

Dy³⁺ + 5 Ni + 3e = DyNi₅

(5)

As the electrolysis potential or the electrolysis current density increases in the cathode area, the partial current of metallic dysprosium extraction increases and the phases of intermetallic compounds with a high concentration of metallic dysprosium appear.

7/3 DyNi₅ + Dy³⁺ + 3e = 5/3 Dy₂Ni₇

(6)

3 Dy₂Ni₇ + Dy³⁺ + 3e = 7 DyNi₃

(7)

2 DyNi₃ + Dy³⁺ + 3e = 3 DyNi₂

(8)

The summarized reaction is described by Equation (4).

The number of observed plateaus in the open circuit chronopotentiogram is less than the number of phases in the Dy–Ni state diagram. The formation reactions of some phases are kinetically retarded, and, therefore, the potential plateau corresponding to the formation of these alloys is not clearly observed. So, on the curves of the open circuit chronopotentiometry (Figure 8, Figure 9, Figure 12 and Figure 13), there is a potential plateau at −2.187 ± 0.010 V, corresponding to the Di³⁺/Dy equilibrium, and a number of other potential plateaus (1)–(4), corresponding to the equilibrium of two phases for (DyNi₂/DyNi₃), (DyNi₃/Dy₂N_i7), (Dy₂Ni₇/DyNi₅) and (DyNi₅/Ni), respectively (Figure 13). If the plateau potential (1)–(4) is expressed depending on the potential of Dy^3+/Dy, then this potential directly corresponds to the EMF value of each two-phase coexisting state. To confirm the reproducibility of the experiment, the same measurements were repeated several times in each two-phase region.

Table 2. Transformation relations and obtained EMF values for the two-phase co-existing states of the Dy_xNi_y intermetallic compounds in the KCl-NaCl-CsCl-DyCl₃ (3.0 × 10⁻⁴ mol/cm³) − NiCl₂ (3.0 × 10⁻⁵ mol/cm³) melt at 823 K.

Plateau No. (Figure 13)	E (V, vs. Dy3+/Dy)	Reactions
(1)	0.243 ± 0.01	2 DyNi3 + Dy3+ + 3 e− ⇄ 3 DyNi2
(2)	0.417 ± 0.01	3 Dy2Ni7 + Dy3+ + 3e− ⇄ 7 DyNi3
(3)	0.523 ± 0.012	7/3 DyNi5 + Dy3+ + 3e− ⇄ 5/3 (Dy2Ni7)
(4)	0.712 ± 0.019	5 Ni + Dy3+ + 3e− ⇄ DyNi5

presents the transformation reactions and the EMF values obtained for two-phase coexisting states from several independent measurements of the EMF at 823 K in a KCl-NaCl-CsCl-DyCl₃-NiCl₂ melt.

Based on the results of cyclic and square-wave voltammetry whereby dysprosium exists in the chloride melt in the form of ions with an oxidation state of +3 and according to the values of EMF (E (V, vs. Dy³⁺/Dy), the relative partial molar Gibbs energies and Dy activities in the intermetallic compounds Dy_xNi_y (

Table 3. Thermodynamic properties of dysprosium for intermetallic Dy_xNi_y compounds in two-phase co-existing states at 823 K.

Phase Equillibria	E (V, vs. Dy3+/Dy)	$\Delta {\overline{\mathrm{G}}}_{\mathbf{D}\mathbf{y}}(\mathbf{kJ}/\mathbf{mol} \mathbf{Dy})$	aDy
Two-phase co-existing state between DyNi2/DyNi3	0.243 ± 0.01	−70.35 ± 2.9	(3.43 ± 0.14) × 10−5
Two-phase co-existing state between DyNi3/Dy2Ni7	0.417 ± 0.01	−120.72 ± 2.9	(2.18 ± 0.05) × 10−8
Two-phase co-existing state between Dy2Ni7/DyNi5	0.523 ± 0.012	−151.4 ± 3.4	(2.47 ± 0.06) × 10−10
Two-phase co-existing state between DyNi5/Ni	0.712 ± 0.019	−206.12 ± 5.5	(8.3 ± 0.22) × 10−14

), using the following equations, are calculated:

$$\Delta {\overline{\mathrm{G}}}_{\mathrm{Dy}} \left({\mathrm{Dy}}_{\mathrm{x}}{\mathrm{Ni}}_{\mathrm{y}}\right)=-3\mathrm{FE}$$

(9)

$${\mathrm{a}}_{\mathrm{Dy}}\left({\mathrm{Dy}}_{\mathrm{x}}{\mathrm{Ni}}_{\mathrm{y}}\right)=\exp \frac{\Delta {\overline{\mathrm{G}}}_{\mathrm{Dy}} \left({\mathrm{Dy}}_{\mathrm{x}}{\mathrm{Ni}}_{\mathrm{y}}\right)}{\mathrm{RT}}$$

(10)

where $\Delta {\overline{\mathrm{G}}}_{\mathrm{Dy}}$ is a relative partial free molar Gibbs energy for the Dy component and ${\mathrm{a}}_{\mathrm{Dy}}$ is the Dy component activity.

The relative partial molar Gibbs free energies and Dy activities in Dy_xNi_y intermetallic compounds obtained in this work were compared with the literature data [31]. The existing discrepancies in the values of the relative partial molar free energy of Gibbs and the activity of dysprosium are associated with a temperature higher by 50–150 K in the eutectic molten system KCl-NaCl-CsCl compared to the eutectic KCl-LiCl.

4. Conclusions

The methods of cyclic and square-wave voltammetry, chronopotentiometry, chronovoltammetry and open-circuit chronopotentiometry were used to study the electrochemical behavior of dysprosium ions and their co-electroreduction with nickel ions on an inert tungsten and active nickel electrodes in a eutectic KCl-NaCl-CsCl melt at a temperature of 823 K. The electroreduction of Dy³⁺ ions in the eutectic KCl-NaCl-CsCl melt at a temperature of 823 K proceeds reversibly in a single three-electron stage up to polarization rates of 0.1 V/s. At higher polarization rates, the charge transfer stage is slower.

The open circuit chronopotentiograms allowed for the determination of the electrode potential E(Dy³⁺/Dy) = −(2.098–2.130) V in the KCl-NaCl-CsCl eutectic melt at a temperature of 823 K. The waves of the Dy³⁺ and Ni²⁺ ion co-electroreduction with the formation of corresponding intermetallic phases were recorded by the cyclic and square-wave voltammograms and open circuit chronopotentiograms in the eutectic KCl-NaCl-CsCl melt containing Dy³⁺ and Ni²⁺ ions at a temperature of 823 K on an inert tungsten and active nickel electrodes. The Dy_xNi_y intermetallic compounds were synthesized by the potentiostatic electrolysis of the KCl-NaCl-CsCl melt, containing various concentrations of DyCl₃ and NiCl₂ at the potentials of −1.7 V, −1.9 V and −2.1 V on the tungsten and nickel electrodes at 823–973 K. XRD and SEM-EDS analyses of the cathode deposits show the formation of DyNi₅, DyNi₃ and DyNi₂ phases, respectively. It has been established that the more negative electrolysis potential results in the formation of the cathode deposit containing more intermetallic phases with a high content of dysprosium. EMF was measured for the Dy_xNi_y intermetallic compounds in two-phase coexisting states at the temperature of 823 K. The relative partial free molar Gibbs energy and dysprosium activity in the Dy_xNi_y intermetallic compounds were calculated from the EMF values.