Rhenium Electrodeposition and Its Electrochemical Behavior in Molten KF-KBF4-B2O3-KReO4

The electrochemical behavior of rhenium ions in the molten KF-KBF4-B2O3 salt was systematically studied, and pure metallic rhenium was obtained at the cathode. The processes of rhenium ions reduction and diffusion in molten KF-KBF4-B2O3 were determined using cyclic voltammetry, stationary galvanostatic and polarization curves analyses. The values of diffusion coefficients were 3.15 × 10−5 cm2/s and 4.61 × 10−5 cm2/s for R1 and R2, respectively. Rhenium electrodeposition was carried out at a constant potential. The process of rhenium cathode reduction in KF-KBF4-B2O3 at 773 K was found to be a one-step reaction Re(VII) → Re, and rhenium electrodeposition presumably occurred from two types of complex rhenium ions (KReO4 and K3ReO5). Both processes are quasi-reversible and controlled by diffusion. The obtained cathode deposit was analyzed by SEM, EDX, ICP-OES and XRD methods. The obtained deposit had a thread structure and rhenium was the main component.


Introduction
Rhenium was discovered in 1925 and it is considered to be an extremely rare metal with unique properties. Despite low rhenium concentration in the earth's crust of about 10 −7 %, it acquired wide industrial application [1,2]. World rhenium consumption from primary sources ranges from about 50 to 60 metric tons per year [3]. About 70% of the excavated rhenium is used for the production of superalloys for the aerospace industry. About 20% of rhenium is used as bimetallic Pt-Re catalysts in the oil-processing industry for the production of high-octane hydrocarbons used for lead-free gasoline production. The remaining 10% of rhenium is used as Re-W and Re-Mo alloys for the production of electrical contact points, light globes, heating elements, vacuum pipes and X-ray tubes [1,4].
To date, only a few methods of metallic rhenium production are known. The method of rhenium oxides and/or potassium and ammonium perrhenates reduction by hydrogen is the most common [5,6]. Rhenium-based coatings may be obtained by the electrolysis of aqueous solutions of perrhenates and chlorrhenates [7][8][9][10][11][12][13]. Despite their drawbacks, the described methods are widely used. For instance, it is impossible to obtain compact rhenium deposits by the hydrogen reduction method. It is possible to obtain coatings composed of compact rhenium by the aqueous solutions electrolysis, but the thickness of the obtained deposits is about 10 µm.
The method of molten salts electrolysis is the most beneficial method of metallic rhenium and rhenium-based alloys production. This method has a number of advantages, as opposed to other known methods of rhenium production; it allows obtaining rhenium in metallic and powder form; compact, plastic, non-porous and coherent rhenium layers may be formed on a complex large-scale substrate. Chloride [14][15][16][17], fluoride [18,19], chloridefluoride [20] and fluoride-borate melts [2,21,22] may be used as electrolytes. We have previously studied the mechanism of rhenium electroreduction on the rhenium substrate in the fluoride-borate melt containing potassium perrhenate. Compact rhenium alloys of 99.98 wt.% purity were obtained. The initial stages of rhenium electrocrystallization form the KF-KBF 4 -B 2 O 3 -KReO 4 melt on a glassy carbon substrate were studied.
The present paper is devoted to the study of the rhenium cathode reduction processes on the glassy-carbon and platinum substrates. The present research elucidates the possibility of forming metallic rhenium nanofibers in the KF-KBF 4 -B 2 O 3 -KReO 4 melt.

Materials and Methods
The KF(37.28 wt.%)-KBF 4 (40.39 wt.%)-B 2 O 3 (22.33 wt.%) melt was prepared in a glassycarbon container by fusing individual salts. Commercial chemically pure individual KF, KBF 4 and B 2 O 3 salts were provided by Vekton Company, Saints-Petersburg, Russia. The obtained melt was heated to 772 K and exposed in molten state for 4 h to remove the fluoric acid. The residual moisture was removed by the galvanostatic electrolysis at the current load of 0.003 A/cm 2 . The electrolysis was performed until hydrogen evolution at the cathode stopped. The visual examination of the melt optical transparency was used to determine the quality of HF and moisture removal. The rhenium concentration (1-6 wt.%) in the melt was set by direct addition of potassium perinate powder. The chemical composition of the prepared melts was controlled by the method of optic emission spectrometry with the inductively coupled plasma using an iCAP 6300 Duo device (Thermo Scientific, Waltham, MA, USA).
Voltammetry measurements were performed in a three-electrode electrochemical cell ( Figure 1) in the ambient air atmosphere. The cell was assembled in a glassy carbon container (GC-2000) of 600 cm 3 in volume. Rhenium rods of 99.9 wt.% purity and 85 cm 2 square were used as auxiliary electrodes. Glassy carbon and platinum rods of 1.9 mm and 1 mm in diameter and 0.8 cm 2 and 0.36 cm 2 in square, respectively, served as working electrodes. A rhenium electrode served as a reference electrode; it was immersed into the KF-KBF 4 -B 2 O 3 -KReO 4 (6 wt.%) melt in the argon atmosphere (99.998 wt.%). The electrolyte of the reference electrode and the studied melt were separated by a diaphragm composed of pyrolytic boron nitride. The measurements were performed using a galvanostate-potentiostate Autolab PGStat 302N (Metrohm, Herisau, Switzerland) with the Nova 2.1.4 software (Metrohm Autolab B.V., Utrecht, The Netherlands). The ohmic losses were controlled and compensated via Autolab. To determine the chemical composition, the obtained cathode deposits were analyzed using the methods of optic emission spectrometry with the inductively coupled plasma iCAP 6300 Duo (Thermo Scientific, USA); X-ray diffraction method using a Rigaku D/MAX 2200VL/PC diffractometer (Rigaku, Tokyo, Japan); scanning electron microscopy using Tescan Vega 4 with the EDX Oxford Xplore 30 system (Tescan, Brno, Czech Republic). The oxygen concentration in the cathode deposit was determined by the method of reductive melting in the bearer gas flow using the nitrogen and oxygen ana-  To determine the chemical composition, the obtained cathode deposits were analyzed using the methods of optic emission spectrometry with the inductively coupled plasma iCAP 6300 Duo (Thermo Scientific, USA); X-ray diffraction method using a Rigaku D/MAX 2200VL/PC diffractometer (Rigaku, Tokyo, Japan); scanning electron microscopy using Tescan Vega 4 with the EDX Oxford Xplore 30 system (Tescan, Brno, Czech Republic). The oxygen concentration in the cathode deposit was determined by the method of reductive melting in the bearer gas flow using the nitrogen and oxygen analyzer Metavack-K (SPO"EKSAN", Izhevsk, Udmurt Republic, Russia).

Electrochemical Behavior of Rhenium Ions in Molten KF-KBF 4 -B 2 O 3
First, we performed cyclic voltammetry measurements using a glassy carbon electrode with a scanning rate of 250 mV/s at 773 K in molten KF-KBF 4 -B 2 O 3 without rhenium ions. This experiment was carried out to identify the electrochemical window. The obtained cycling voltammetry is illustrated in Figure 2. Extremely low current densities were recorded within the interval of potentials of 0.50 to −0.70 V, relative to the rhenium reference electrode. No clear current density peaks were detected, which elucidates that there were not any redox reactions in the analyzed region of potentials. In the region of potentials more positive than 0.5 V, a clear peak is observed, which is characterized by the reaction of carbon dioxide evolution.
To obtain information on the behavior of rhenium ions, we recorded cyclic voltammetry patterns with a scan rate of 100 mV/s on the glassy carbon and platinum wire electrode after the KReO4 addition to the KF-KBF4-B2O3 melt. The voltammetry pattern is illustrated in Figure 3. At the KReO4 concentration of 1 wt.%, two redox peaks appeared. This may illustrate that the rhenium ions electroreduction is a two-stage process. However, as the KReO4 concentration increased up to 6 wt.%, the second peak of rhenium reduction was not recorded at the rate of the potential change of 100 mV/s. Extremely low current densities were recorded within the interval of potentials of 0.50 to −0.70 V, relative to the rhenium reference electrode. No clear current density peaks were detected, which elucidates that there were not any redox reactions in the analyzed region of potentials. In the region of potentials more positive than 0.5 V, a clear peak is observed, which is characterized by the reaction of carbon dioxide evolution.
To obtain information on the behavior of rhenium ions, we recorded cyclic voltammetry patterns with a scan rate of 100 mV/s on the glassy carbon and platinum wire electrode after the KReO 4 addition to the KF-KBF 4 -B 2 O 3 melt. The voltammetry pattern is illustrated in Figure 3. At the KReO 4 concentration of 1 wt.%, two redox peaks appeared. This may illustrate that the rhenium ions electroreduction is a two-stage process. However, as the KReO 4 concentration increased up to 6 wt.%, the second peak of rhenium reduction was not recorded at the rate of the potential change of 100 mV/s.
To study the electrochemical behavior of rhenium ions in the KF-KBF 4 -B 2 O 3 melt, the cyclic voltammetry was performed at different scan rates. The recorded voltammograms are presented in Figure 4. electrode after the KReO4 addition to the KF-KBF4-B2O3 melt. The voltammetry pattern is illustrated in Figure 3. At the KReO4 concentration of 1 wt.%, two redox peaks appeared. This may illustrate that the rhenium ions electroreduction is a two-stage process. However, as the KReO4 concentration increased up to 6 wt.%, the second peak of rhenium reduction was not recorded at the rate of the potential change of 100 mV/s. To study the electrochemical behavior of rhenium ions in the KF-KBF4-B2O3 melt, the cyclic voltammetry was performed at different scan rates. The recorded voltammograms are presented in Figure 4.    As the rate of the potential change increased, the second peak of rhenium reduction became clearer. The reduction potentials R1 and R2 changed insignificantly as the scan rate increased ( Figure 5A), which testified that R1 and R2 denote quasi-reversible processes. To determine the limiting stage of R1 and R2 reduction, the relation between the peak values of the current densities of R1 and R2 and square roots of the potential change rates were studied. Figure 5B demonstrates that R1 and R2 are linearly dependent, which suggests that rhenium reduction is a diffusion-controlled reaction. As the rate of the potential change increased, the second peak of rhenium reduction became clearer. The reduction potentials R 1 and R 2 changed insignificantly as the scan rate increased ( Figure 5A), which testified that R 1 and R 2 denote quasi-reversible processes. To determine the limiting stage of R 1 and R 2 reduction, the relation between the peak values of the current densities of R 1 and R 2 and square roots of the potential change rates were studied. Figure 5B demonstrates that R 1 and R 2 are linearly dependent, which suggests that rhenium reduction is a diffusion-controlled reaction.
The method of stationary galvanostatic polarization curves demonstrated that the process of rhenium reduction was accompanied with the transfer of seven electrons. The stationary polarization curves are presented in Figure 6. A negligible deviation of the potential from the equilibrium value was observed on the initial region of the polarization curve at the increase in the cathode current density up to 0.05 A/cm 2 . The experimental The number of electrons participating in reaction of the cathode reduction was calculated using Equation (1): the calculated values of n were 7.23 and 7.01 for 4 and 6 wt.% of KReO 4 , respectively. Therefore, the process of rhenium cathode reduction is accompanied with the transfer of seven electrons. As the rate of the potential change increased, the second peak of rhenium reduction became clearer. The reduction potentials R1 and R2 changed insignificantly as the scan rate increased (Figure 5A), which testified that R1 and R2 denote quasi-reversible processes. To determine the limiting stage of R1 and R2 reduction, the relation between the peak values of the current densities of R1 and R2 and square roots of the potential change rates were studied. Figure 5B demonstrates that R1 and R2 are linearly dependent, which suggests that rhenium reduction is a diffusion-controlled reaction. As the R 1 and R 2 processes are quasi-reversible, to verify the amount of electrons participating in the reduction reaction determined by the method of stationary polarization curves, we performed additional research by the potentiostatic electrolysis. To determine the number of electrons for R 1 and R 2 , the electrolysis was performed at the cathode overpotential of 400 mV and 690 mV, respectively. In the course of electrolysis, rhenium deposits were obtained in the amount of 0.2283 g and 0.7629 g at 400 and 690 mV, respectively. According to the obtained chronopotentiogramms, presented in Figure 7, the amount of the passed electricity was calculated. The number of electrons in the process was calculated according to the amount of the electricity passed using Faraday's law. It was determined that the reduction reactions during both R 1 and R 2 processes are accompanied with seven electrons.

RT n tg F
the calculated values of n were 7.23 and 7.01 for 4 and 6 wt.% of KReO4, respectively. Therefore, the process of rhenium cathode reduction is accompanied with the transfer of seven electrons. As the R1 and R2 processes are quasi-reversible, to verify the amount of electrons participating in the reduction reaction determined by the method of stationary polarization curves, we performed additional research by the potentiostatic electrolysis. To determine the number of electrons for R1 and R2, the electrolysis was performed at the cathode overpotential of 400 mV and 690 mV, respectively. In the course of electrolysis, rhenium deposits were obtained in the amount of 0.2283 g and 0.7629 g at 400 and 690 mV, respectively. According to the obtained chronopotentiogramms, presented in Figure  7, the amount of the passed electricity was calculated. The number of electrons in the process was calculated according to the amount of the electricity passed using Faraday's law. It was determined that the reduction reactions during both R1 and R2 processes are accompanied with seven electrons. Due to the fact the maximum rhenium oxidation degree is VII, the presence of two reduction peaks, accompanied with the transfer of seven electrons, testifies that two types of Re(VII) ions discharge. We may assume that during the cathode reduction of KReO4, the conditions for K3ReO5 synthesis appear. At the cathode polarization up to 400 mV, rhenium deposits from KReO4, and at the cathode polarization exceeding 400 mV, the process becomes limited by diffusion. The near-electrode layer is depleted of rhenium ions and K3ReO5 forms; the cathode process may be described by Equation (2): At the further electrode polarization, rhenium reduces from both KReO4 (Equation (2)) and K3ReO5 (Equation (3)). 3 2 5 The diffusion coefficient of Re(VII) ions in molten KF-KBF4-B2O3 can be calculated according to Equation (4) [23,24]: where ip is the peak current density (mA/sm 2 ), ν is the scan rate (mV/s), C is the bulk Due to the fact the maximum rhenium oxidation degree is VII, the presence of two reduction peaks, accompanied with the transfer of seven electrons, testifies that two types of Re(VII) ions discharge. We may assume that during the cathode reduction of KReO 4 , the conditions for K 3 ReO 5 synthesis appear. At the cathode polarization up to 400 mV, rhenium deposits from KReO 4 , and at the cathode polarization exceeding 400 mV, the process becomes limited by diffusion. The near-electrode layer is depleted of rhenium ions and K 3 ReO 5 forms; the cathode process may be described by Equation (2): At the further electrode polarization, rhenium reduces from both KReO 4 (Equation (2)) and K 3 ReO 5 (Equation (3)).
The diffusion coefficient of Re(VII) ions in molten KF-KBF 4 -B 2 O 3 can be calculated according to Equation (4) [23,24]: where i p is the peak current density (mA/sm 2 ), ν is the scan rate (mV/s), C is the bulk concentration of the reducible ion (mol/cm 3 ) and D is the diffusion coefficient (cm 2 /s). The calculated diffusion coefficient of Re(VII) ions in molten KF-KBF 4 -B 2 O 3 with 2wt.% KReO 4 at 773 K was 3.15 × 10 −5 cm 2 /s and 4.61 × 10 −5 cm 2 /s for R 1 and R 2 , respectively.

Potentiostatic Electrolysis
During the potentiostatic electrolysis rhenium was formed at the surface of the glassy carbon crucible in the form of nanofibers of 100-600 nm in diameter (Figure 8). The length of the nanofibers was about 150-200 µm. The reduction was performed at the potentials of 400 mV and 690 mV. During the electrolysis, we obtained metallic rhenium nanofibers of 200-600 nm ( Figure 8A) and 100-200 nm ( Figure 8B) in diameter, respectively. The X-ray diffraction analysis of the obtained nanofibers demonstrated that the deposit has a homogeneous rhenium phase (Figure 9). Other types of phases were not determined. The X-ray diffraction analysis of the obtained nanofibers demonstrated that the deposit has a homogeneous rhenium phase (Figure 9). Other types of phases were not determined.
The energy-dispersive X-ray spectroscopy analysis elucidated that rhenium was homogeneously distributed in the samples. The presence of any another metallic inclusions was not detected (Figure 10). The method of reduction melting demonstrated that the obtained rhenium deposits contained 0.5243 ± 0.0026 wt.% of oxygen. The presence of oxygen is probably associated with the removal of the molten salt residues from the surface of the deposit by rinsing in the deionized water. The obtained deposit was analyzed for the concentration of admixtures using the method of optic emission spectrometry with the inductively coupled plasma iCAP 6300 Duo (Table 1). The X-ray diffraction analysis of the obtained nanofibers demonstrated that the deposit has a homogeneous rhenium phase (Figure 9). Other types of phases were not determined. The energy-dispersive X-ray spectroscopy analysis elucidated that rhenium was homogeneously distributed in the samples. The presence of any another metallic inclusions was not detected (Figure 10). The method of reduction melting demonstrated that the obtained rhenium deposits contained 0.5243 ± 0.0026 wt.% of oxygen. The presence of oxygen is probably associated with the removal of the molten salt residues from the surface of the deposit by rinsing in the deionized water. The obtained deposit was analyzed for the concentration of admixtures using the method of optic emission spectrometry with the inductively coupled plasma iCAP 6300 Duo (Table 1).  During the electrolysis, the electricity amounting to 830 C and 2772 C was passed via the electrochemical cell at the potential of 400 mV and 690 mV, respectively. According to the results of the gravimetric analysis, we calculated the actual current efficiency, which value was 100% for the current loads of both 400 mV and 690 mV.
The analysis of the chemical composition of the obtained deposits illustrated that it is possible to obtain metallic rhenium of the 99.98 wt.% purity by the electrolysis of the KF-KBF4-B2O3-KReO4 melts in air. Potassium is the main admixture because it is an integral parts of the melt cation sublattice. It should be mentioned that the boron concentration in rhenium is less than the method determination limit. This may be explained by the high temperatures of rhenium and boron interaction.  During the electrolysis, the electricity amounting to 830 C and 2772 C was passed via the electrochemical cell at the potential of 400 mV and 690 mV, respectively. According to the results of the gravimetric analysis, we calculated the actual current efficiency, which value was 100% for the current loads of both 400 mV and 690 mV.

Conclusions
The analysis of the chemical composition of the obtained deposits illustrated that it is possible to obtain metallic rhenium of the 99.98 wt.% purity by the electrolysis of the KF-KBF 4 -B 2 O 3 -KReO 4 melts in air. Potassium is the main admixture because it is an integral parts of the melt cation sublattice. It should be mentioned that the boron concentration in rhenium is less than the method determination limit. This may be explained by the high temperatures of rhenium and boron interaction.

Conclusions
The electrochemical behavior of Re(VII) in molten KF-KBF 4 -B 2 O 3 was systematically studied in a three-electrode electrochemical cell at 773 K. It was found that the process of rhenium ions reduction is a quasi-reversible reaction controlled by diffusion. Rhenium was found to reduce from two types of seven valency ions. First, rhenium preferably reduced from the potassium perrhenate, and, at the further polarization, the reduction proceeded from the mixture of potassium perrhenate and mesa perrhenate. Diffusion coefficients of Re(VII) corresponded to its reduction from KReO 4 and KReO 4 + K 3 ReO 5 , were analyzed using the method of cyclic voltammetry. Their values were 3.15 × 10 −5 cm 2 /s and 4.61 × 10 −5 cm 2 /s, respectively.
The nano-fibers of metallic rhenium on the glassy carbon substrate were obtained during the potentiostatic electrolysis of the KF-KBF 4 -B 2 O 3 melt. The formed deposits were analyzed by the SEM with EDX and ICP-OES methods. The obtained results testified that the obtained deposits were composed of highly pure metallic rhenium (99.98%).