1 Introduction

A couple of new electrolytic technologies, viz., Fray-Farthing-Chen (FFC) Cambridge and Ono-Suzuki (OS) processes, have been proposed for the preparation of various important metals and alloys directly from their constituent metal oxides [1,2,3,4,5,6,7,8,9]. These direct oxide electrochemical reduction (DOER) processes utilize calcium chloride based melts as the electrolyte because of their high solubility for oxide ions. High-density carbon (graphite) is generally used as the anode in these methods. Solid metal oxide cathodes are reduced to respective metals/alloys either by the electron (FFC Cambridge) or in-situ electro-generated reductant metal (OS process) under suitable electrolysis conditions; the liberated oxide ions transfer through the melt and react with the carbon anode producing a mixture of CO and CO2 gases. The usage of reactive carbon anodes in these processes not only results in generation of greenhouse gases but also leads to decrease in process efficiency due to several parasitic reactions. A significant portion of the CO2 gas gets dissolved in CaCl2 based melts, leading to the formation of carbonate ions, which partake in various parasitic reactions resulting in decreased current and energy efficiencies of the process and, more importantly, a substantial amount of carbon contamination in the metal product [5,6,7, 10,11,12,13]. The redox reactions of carbonate ions can be represented by Eqs. 1–3.

$${\text{Anode}}:\mathrm{ C}+2{{\text{O}}}^{2-}\to {{\text{CO}}}_{2}+4{{\text{e}}}^{-}$$
(1)
$${{\text{CO}}}_{2}+{{\text{O}}}^{2-}\to {{\text{CO}}}_{3}^{2-}(\mathrm{Parasitic\;reaction})$$
(2)
$${\text{Cathode}}:\mathrm{ C}{{\text{O}}}_{3}^{2-}+4{{\text{e}}}^{-}\to {\text{C}}+3{{\text{O}}}^{2-}(\mathrm{Parasitic\;reaction})$$
(3)

In addition, the liberated CO2 can react with elemental calcium and form CaO (Eq. 4), affecting the OS process's current efficiency [2, 6, 10, 11].

$$2{\text{Ca}}+{{\text{CO}}}_{2}\to {\text{C}}+2{\text{CaO}}$$
(4)

Recently, Chen et al. reviewed the mechanism of carbonate cycling and the formation of carbon debris in FFC Cambridge process [12]. A relatively high working temperature (900–1000 °C) and the usage of a sheathing tube around the cathode were recommended for the mitigation of carbon contamination. The reactive nature of graphite anode results in severe physical degradation of the electrode and melt contamination by dispersed carbon particles in long-duration electrolysis runs [5, 7, 12, 13]. Carbon contamination of the reduced metal and alloy products has been reported in the direct electrochemical reduction studies of various metal oxides, e.g., TiO2, Nb2O5, TiO2—Nb2O5—Ta2O5—ZrO2, PuO2, etc. in CaCl2 based melts due to the usage of graphite as anode [5, 8, 10, 11, 14, 15]. Recently, Mukherjee et al. studied the reduction mechanism of ThO2 in CaCl2 and CaCl2–CaO melts at 900 °C [16]. Severe thinning of graphite and carbon contamination in the cathode product (ThC) were observed in CaCl2–CaO melt (OS process). Preparing technologically important metals, e.g., Ti, Nb, etc., especially actinide metals, e.g., Pu, Th, etc., with carbon impurities, is undesirable. The parasitic redox reactions involving CO and CO2 can be avoided if an oxygen-evolving inert electrode is employed as an anode for DOER processes in CaCl2 based melts. The inert anodes interact with oxide ions and liberate molecular oxygen (Eq. 5) instead of CO or CO2, and thus avoid the formation of carbonate ions.

$$Anode: 2{O}^{2-}\to {O}_{2}+4{e}^{-}$$
(5)

Majorly, three kinds of inert anodes viz., ceramic, cermet and metallic, have been reported in the literature. The detailed literature survey on various anode materials studied in the context of FFC Cambridge and OS processes is given in Table 1.

Table 1 Literature survey on usage of different types of inert anode materials in CaCl2 based melts
Full size table

In most of the inert electrodes, a protective coating of the oxide layer is formed around the anode, which generally inhibits corrosion. However, the leaching of anodic material and subsequent product contamination are still the major concerns associated with using inert anodes. For instance, the studies on the electro-deoxidation of Ta2O5 revealed the presence of Ta3Sn2 along with the cathode product (Ta) when a SnO2 based anode is used [19]. Similarly, leaching of nickel and copper is observed in the melt when nickel ferrite cermet-based anodes are used [25, 26]. The process of developing inert anodes becomes futile when the leaching element forms an alloy or intermetallic compound with the cathode product. In preparing strategic elements, especially actinide metals, care must be taken to avoid product contamination. Platinum is usually used as an anode material for the electro-deoxidation of UO2 in LiCl–Li2O based melts [31,32,33]. Iizuka et al. studied the electrochemical reduction of (U,Pu)O2 pellets in both LiCl and CaCl2 melts using platinum anode [34]. The complete reduction of (U,Pu)O2 was observed in both melts without any product contamination. However, the study did not report the amount of platinum lost during the electrochemical reduction experiments. Many studies have reported the electrochemical behavior of Pt anode in LiCl–Li2O melt; the reported studies indicated that Pt anodically reacts in LiCl–Li2O melt and forms Li2PtO3 compound as given in Eq. 6 [35, 36].

$$2{{\text{Li}}}^{+}+\mathrm{ Pt}+3{{\text{O}}}^{2-}\to {{\text{Li}}}_{2 }{{\text{PtO}}}_{3}+4{{\text{e}}}^{-}$$
(6)

The compound lithium platinate is electrically conductive, and thus supports the passage of electrical current [36]. Although the Li2PtO3 is stable and aids oxygen liberation, the compound eventually peels off from the anode during the electrochemical run, thus accounting for the loss of platinum [36]. However, literature regarding the electrochemical behavior of Pt in CaCl2 based melts is limited. Sakamura et al. studied the electrochemical reduction of UO2 in CaCl2–1.2 mol % CaO melt at 820 °C using platinum as the anode [37]. In this context, cyclic voltammetry studies were performed on a Pt electrode in CaCl2 melt with varying concentrations of CaO. The study suggested that undesirable oxidation reaction occurs near the oxygen evolution potential. Recently, Pertuiset et al. studied the electrochemical behavior of platinum, gold and palladium anodes using electrochemical techniques in CaCl2–CaO melt at 850 °C [38]. The study reported the dissolution of gold and palladium as AuClx and PdClx respectively during oxygen evolution. The study concluded that oxygen can be evolved on a platinum anode with a very low degradation under specific electrolytic conditions i.e., low concentration of CaO and current density. However, the study did not report or validate the results by conducting any electrochemical reduction experiments.

The present study aimed to investigate the usage of platinum anode in electro-calciothermic reduction experiments in CaCl2–CaO melt. In this context, electro-calciothermic reduction (OS process) experiments were conducted on ThO2 and NiO using platinum anode in CaCl2–1 wt % CaO melt at 900 °C. ThO2 and NiO were chosen based on high and low thermodynamic stability, respectively. Primarily, the anodic behavior of platinum electrode was studied in CaCl2–1 wt % CaO melt at 850 °C using potentiodynamic polarization, linear sweep voltammetry and electrochemical impedance spectroscopy. The results were compared with nickel, gold and HD graphite electrodes. The study also investigated the anodic behavior of platinum and gold electrodes in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, 3) melts using cyclic voltammetry and potentiostatic electrolysis techniques at 900 °C. The results obtained in the study and their detailed analyses are reported in this article.

2 Experimental

2.1 Preparation of anhydrous electrolyte

Commercially available CaCl2.2H2O (Merck, Emparta, ACS grade, purity 99% min.) was pre-dried at 150 °C under argon atmosphere followed by vacuum drying at 200 °C for the removal of the hydrated water molecules. The detailed drying procedure is mentioned in our previous publications [16, 39]. As received anhydrous CaO (Alfa Aesar, purity 99.95%), was used in the preparation of CaCl2–CaO electrolytes.

2.2 Electrodes

Ni wire (Alfa Aesar, purity 99.95%, 1.6 mm dia.), Au wire (Alfa Aesar, purity 99.95%, 1 mm dia.), Mo wire (Alfa Aesar, purity 99.95%, 2 mm dia.), tungsten (W) wire (1 mm dia., Alfa Aesar, purity 99.95%), and Pt wire (purity 99.95%, 1 mm dia.) were cleaned with emery sheet before use. The platinum and molybdenum coils were prepared from platinum and molybdenum wires, respectively. High-density (HD) graphite rod (6 mm dia., 100 mm long, density 1.85 g cc−1, M/s. Nikunj Group, Mumbai, India), was heat-treated at 800 °C under vacuum for 5 h to remove adsorbed air. A porous nickel oxide preform was prepared by mixing 3 wt % PVA (binder) and 1.5 wt % PEG (plasticizer) with as received NiO powder (Sigma-Aldrich, 637130) and the mixture was made into a pellet (12.5 mm dia.) using uniaxial pellet press. The green pellets were sintered in air at 1200 °C for 2.5 h. The open porosity of the sintered pellet was found to be ~ 30%. ThO2 powder was prepared in-house from thorium nitrate pentahydrate [Th(NO3)4.5H2O] (M/s. Indian Rare Earths Ltd., India) by citrate gel-combustion method [16]. 3 wt % PVA (binder) and 1.5 wt % PEG (plasticizer) was mixed with the ThO2 powder and green ThO2 pellets (12.5 mm dia., ~ 0.8 g mass) were prepared. These pellets were sintered in air for 2.5 h at 1200 °C, and the open porosity of the sintered pellet was determined to be ~ 40%. These electrodes were tied or screwed to SS (1.6 mm dia.) current leads which were electrically insulated from the cell body and each other by using suitable alumina sleeves. A Ni|NiO reference electrode was used in the molten-salt electrochemical experiments [16, 39, 40].

2.3 Experimental setup

Molten salt experiments were carried out in a specially designed leak-tight reactor made of Inconel 600. An alumina crucible was used to contain the electrolyte. Loading of electrodes and salt in the molten salt reactor was carried out inside the high-purity argon atmosphere glove box with oxygen and moisture concentrations < 10 ppm. The properly assembled reactor was then taken out of the glove box. It was placed in a pit-type resistance heating furnace under continuous purging of high-purity argon gas and was slowly (≤ 3° min−1) heated to the required process temperature. For CaCl2–CaO melts, a weighed amount of CaO was added to molten CaCl2 to achieve the desired concentration. The reactor was maintained at the process temperature for 8–10 h to dissolve CaO in the molten salt completely. The temperature of the molten electrolyte was measured with the help of a K-type thermocouple placed inside a one-end closed alumina tube.

2.4 Electrochemical experiments

During each electrochemical experiment, the molten electrolyte (CaCl2 or CaCl2–CaO) was first pre-electrolyzed for complete removal of moisture and redox-active impurities, by applying 2.5 V (in the case of pure CaCl2 electrolyte) or 1.5 V (in the case of CaCl2–CaO electrolytes) between Mo coil cathode and HD graphite rod anode and with the help of a DC power supply (KEYSIGHT TECHNOLOGIES make, E3633A). The pre-electrolysis was carried out until background current densities stabilized at a very low value (> 5 mA cm−2). Linear sweep voltammetry, potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) studies were conducted in CaCl2–1 wt % CaO melt with nickel, HD graphite, gold and platinum electrodes at 850 °C. Electrochemical behavior of gold and platinum electrodes was investigated by using cyclic voltammetry (CV) in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, 3) melts with Mo coil as counter electrode and Ni|NiO as reference electrode at 850 °C and 900 °C, respectively. Potentiostatic electrolysis experiments were carried out in CaCl2–x wt % CaO (x = 1, 2, 3) melts at four different potentials within the oxygen evolution window for the same time durations. All of these electrochemical measurements were carried out with the help of AUTOLAB PGSTAT 302N potentiostat/galvanostat and the data were recorded and analyzed using NOVA 2.1.1 software. DOER experiments (electro-calciothermic reduction) were conducted in CaCl2–1 wt % CaO melt with NiO and ThO2 pellet cathodes and platinum coil anode at 900 °C using constant voltage electrolysis method with the help of a DC power supply (KEYSIGHT TECHNOLOGIES make, E3633A). The half-cell potentials of the cathode and anode were monitored with respect to Ni|NiO reference electrode. The data were recorded using a data acquisition unit (KEYSIGHT TECHNOLOGIES, 34972A). After completion of the experiments, all the electrodes were lifted well above the melt and cooled to room temperature under high pure argon gas flow. The cooled electrolysis cell was taken inside the glove box and was opened there. The platinum coils used in the chronoamperometry and DOER studies were washed thoroughly with distilled water for complete removal of the occluded salt and subsequently were acetone dried. These coils were weighed before and after (thoroughly water-washed and acetone dried) the experiments in an A&D, Japan make analytical balance (GR-200). The micrographs of the fresh and anodically polarized platinum coils were recorded using ZEISS GEMINI CROSSBEAM 340 scanning electron microscope (SEM) and the energy dispersive X-ray (EDX) analyses of the samples were carried out with the help of OXFORD INSTRUMENTS make EDX analysis unit (AZTEC) coupled with the SEM. The electrolyzed pellets were characterized by X-ray diffraction technique (XRD) [PANalytical X’pert Pro, θ/θ diffractometer with Cu k-α (λ = 1.54 Å) radiation source] and SEM. The Ni|NiO reference electrode showed good potential stability during the electrochemical measurements, and all the half-cell potentials were recorded vs Ni|NiO. However, the half-cell potentials mentioned in the present study are rescaled to Ca2+|Ca reference electrode for better understanding.

3 Results and discussions

3.1 Anodic polarization studies in CaCl2–1 wt% CaO melt

The electrochemical stability of Ni, HD graphite, Au and Pt electrodes was investigated in CaCl2–1 wt % CaO melt at 850 °C using linear polarization and Tafel measurements. Linear polarization studies were conducted at a scan rate of 1 mV s−1, and the recorded anodic polarization curves are depicted in Fig. 1. The potentials are plotted with respect to Ca2+|Ca reference. Among the electrodes studied Ni showed the least anodic polarization potential value (B, 2.41 V). At this potential, a steep rise in current density was observed with increasing potential denoting its fast dissolution in the melt. However, at A (~ 2.13 V–2.39 V) current density was almost constant, which might be due to the formation of NiO [27]. In the case of HD graphite, a steep rise in current density was found 2.63 V onwards, followed by a more gradual increase in the potential range of (3–3.41 V) and a moderate rise in current density was perceived from 3.44 V onwards. The increase in current density in C (2.63–3.41 V) was due to CO and/or CO2 formation (Eq. 1) in the CaCl2–CaO melt, whereas 3.44 V (D) was the onset potential of Cl2 evolution on HD graphite anode (Eq. 7) [16].

Fig. 1
figure 1

Polarization curves of Nickel, HD graphite (inset is included to provide better clarity), gold and platinum electrodes recorded in CaCl2–1 wt % CaO melt at 850 °C (Scan rate = 1 mV s−1)

Full size image
$$2{{\text{Cl}}}^{-}-2{{\text{e}}}^{-} \to {{\text{Cl}}}_{2} ({\text{g}})$$
(7)

The standard reduction potentials (vs Ca2+|Ca) of a few related half-cell reactions were calculated from the thermodynamic database, HSC Chemistry (Version 5.11) and FactSage 6.4 [41, 42], considering standard state and unit activity for all reactants and products and are tabulated in Table 2. In case of Au electrode, current density was observed to increase sharply from 2.83 V (E) onwards, with an increase in potential. As evident from Table 2, this might be due to evolution of oxygen at Au electrode (Eq. 5). However, Au dissolution also might be happening simultaneously, as suggested by the very high current density (Eqs. 8 and 9, Table 2). In case of Pt, a slight increase in current density was observed in the potential range (F) (2.95 V–3.22 V) followed by a moderate rise in current density from 3.24 V (G) onwards. The behavior in region F might be due to evolution of oxygen at Pt electrode (Eq. 5), and the dissolution reaction might be prevailing from G onwards (Eq. 11, Table 2).

Table 2 Standard reduction potentials for a few relevant half-cell reactions vs Ca2+|Ca reference electrode [41, 42]
Full size table

The findings suggest that Pt might be the more suitable anode material in CaCl2–CaO melts among the electrodes studied, as it exhibits the most positive anodic polarization potential along with prominent demarcation between oxygen evolution and anodic dissolution potentials.

The electrodes were also subjected to potentiodynamic polarization conditions in CaCl2–1 wt % CaO melt at 850 °C. The corresponding Tafel plots (1 mV s−1 scan rate) are shown in Fig. 2. As perceived from the plots, corrosion potentials for Ni, HD graphite, Au and Pt are 1.54 V, 1.58 V, 1.62 V and 1.7 V, respectively. Platinum exhibits the highest positive corrosion potential (Ecorr), indicating that Pt is the most corrosion-resistant material among the electrodes studied in CaCl2–1 wt % CaO melt.

Fig. 2
figure 2

Tafel plots of nickel, HD graphite, gold and platinum electrodes in CaCl2–1 wt % CaO melt at 850 °C (Scan rate = 1 mV s−1)

Full size image

3.2 Electrochemical impedance spectroscopy studies in CaCl2–1 wt % CaO melt

Electrochemical impedance spectroscopy studies were conducted with HD graphite, gold and platinum electrodes in CaCl2–1 wt % CaO melt at 850 °C at the open circuit potentials (OCP) using a perturbation signal of amplitude 10 mV in the frequency range of 10 Hz to 10,000 Hz for gaining insights about the electrochemical oxidation resistances of the electrodes in CaCl2 –CaO melts. The OCPs of the electrodes were 1.48 V (HD graphite), 1.52 V (gold) and 1.57 V (platinum). The Nyquist plots of the same are depicted in Fig. 3a. Figure 3b shows the equivalent circuit used for fitting the EIS data.

Fig. 3
figure 3

(a) Electrochemical impedance spectroscopy plots (Nyquist) for HD graphite (inset), gold and platinum electrodes recorded at OCP in CaCl2–1 wt % CaO melt at 850 °C, (b) equivalent circuit used for fitting the complex impedance data

Full size image

The components of the equivalent circuit are solution resistance (Rs), charge transfer resistance (Rct) and two constant phase elements (CPE) (Q1 and Q2). The CPEs, Q1 and Q2, are used in place of capacitive and diffusion impedance elements, respectively, to attain a more accurate fit. The Rs and Rct values obtained after analyzing the EIS data of the electrodes are tabulated in Table 3. The solution resistance values obtained from the analysis of complex impedance data are 0.59 Ω (HD graphite), 0.687 Ω (gold) and 0.967 Ω (platinum). The charge transfer resistances of the electrodes are found to be 1.19 Ω (HD graphite), 10.1 Ω (gold) and 146 Ω (platinum). The Rct value of platinum is higher than other electrodes, suggesting that platinum is the most oxidation-resistant material among the electrodes studied. Gold electrode exhibits moderate oxidation resistance, whereas HD graphite shows least oxidation-resistant behavior. Du et al. conducted electrochemical impedance spectroscopy studies in CaCl2–3 wt% CaO melt at 850 °C with platinum electrode and reported the Rct value as 106.20 Ω indicating its superior oxidation resistance properties [24].

Table 3 Various parameters obtained from EIS studies conducted with different anode materials in CaCl2–1 wt % CaO melt at 850 °C
Full size table

3.3 Electrochemical behavior of gold electrode in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, 3) melt under anodic polarization conditions

3.3.1 Cyclic voltammetry studies

Gold electrode exhibited inertness towards the oxidation of oxide ions, especially in carbonate melts, compared to other metal electrodes [43]. However, information on the anodic behavior of gold electrode in CaCl2–CaO melt is very scarce. In the present study, the electrochemical behavior of gold electrode was investigated in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, and 3) melts at 850 °C. Figure 4 depicts the cyclic voltammogram of Au in pure CaCl2 melt. For comparison, the cyclic voltammogram of HD graphite is also included. Graphite is inert to chlorine evolution and a steep increase in current observed from 3.31 V onwards is due to evolution of chlorine (Eq. 7) [16]. The disruption in the current indicates evolution of chlorine gas. The potential value matches the chlorine evolution potential (vs Ca2+|Ca), as shown in Table 2.

Fig. 4
figure 4

Cyclic voltammograms (anodic polarization) of HD graphite rod (area = 1.236 cm2) in pure CaCl2 melt and gold wire (electrode area = 0.322 cm2) electrodes in pure CaCl2 and CaCl2–x wt % CaO (x = 0.5, 1, 2, 3) melts (scan rate = 2 mV s−1, temperature = 850 °C)

Full size image

In case of Au, chlorine evolution is expected to occur before its dissolution, as evident from Table 2. However, the current started increasing linearly at a marginally less positive potential than that of chlorine evolution in graphite electrode (Fig. 4). Therefore, the increase in current might be due to chlorine evolution (Eq. 7) and/or dissolution of Au (Eqs. 8 and 9, Table 2). Figure 4 depicts the cyclic voltammograms of gold electrode in CaCl2–x wt % CaO (x = 0.5, 1, 2, and 3) melts recorded at 2 mV s−1 scan rate. A noticeable increase in current is observed in the region (2.8–3 V) onwards in CaCl2–CaO melts. Cyclic voltammograms of CaCl2–x wt % CaO (x = 0.5, 1) melt showed a gradual increase in current in the region (3.1–3.25 V), and a steeper rise in current was observed upon further anodic polarization. The observed current signature in that region (3.1–3.25 V) could be attributed to oxygen evolution (Eq. 3), and the further increase in current is due to chlorine evolution (Eq. 7) or dissolution of Au (Eqs. 8 and 9, Table 2). Cyclic voltammograms of CaCl2–x wt % CaO (x = 2, 3) melt showed early current onset compared to CaCl2 melt containing less concentration of CaO. The steep increase in current was observed upon polarization without any peak or change in slope associated with evolution of oxygen. A similar observation was noticed by Yao et al. with gold electrode in LiCl–Li2O melt at 650 °C [44]. Similarly, the oxygen evolution potential window was not perceived clearly in cyclic voltammograms recorded at higher scan rates. Determination of the oxygen evolution potential window of an anode material is important because it ensures safe operation during DOER experiments without experiencing much mass loss of the anode. Hence, the difficulty in discerning the oxygen evolution window in the case of gold anode may pose a disadvantage during DOER experiments.

3.3.2 Potentiostatic electrolysis studies

Potentiostatic electrolysis experiment was carried out in CaCl2–1 wt % CaO melt for 30 min at 850 °C with Au coil working electrode. Mo coil and Ni|NiO were used as counter and reference electrodes, respectively. The working electrode's potential was kept at 3.2 V vs Ca2+|Ca during the electrolysis experiment. The potential is within the oxygen evolution window in CaCl2–1 wt % CaO melt, as determined by cyclic voltammetry studies (Fig. 4). During potentiostatic electrolysis experiment, the current was almost constant at ~ 0.9 A. After completion of the experiment, the gold coil electrode was washed with distilled water for removing the occluded salt and it was then dried thoroughly with acetone. No distinct physico-chemical changes, i.e., coating formation, significant thinning due to Au dissolution etc., were observed in the gold coil after electrolysis. Scanning electron micrographs of Au electrode surface before and after electrolysis are shown in Fig. 5. The fresh gold revealed the utmost smooth surface, whereas polarized Au showed a lot of pits and etched surfaces due to dissolution. In the EDX map sum spectra of the gold coils presence of no other element except gold was observed. Around 0.325 g mass loss of the anodically polarized Au coil was discerned upon weighing, which amounted to a corrosion rate of 0.2 g cm−2 h−1. The mass loss might be due to the momentary formation of Au2O3 as per Eq. 10 (Table 2) alongside O2 evolution (Eq. 5). The oxide layer is unstable and dislodges from the electrode surface leading to loss of Au. Additionally, a clear demarcation between O2 evolution (Eq. 5) and Au dissolution (Eqs. 8 and 9, Table 2) was not observed, as revealed by the cyclic voltammetry studies. Hence, mass loss of Au by Eqs. 8 and 9 also cannot be ruled out during the potentiostatic electrolysis experiment.

Fig. 5
figure 5

Scanning electron micrographs of the gold coil electrode (a) before and (b) after potentiostatic electrolysis experiments in CaCl2–1 wt % CaO melt at 850 °C (magnification: 1000×)

Full size image

Despite possessing a relatively high anodic polarization potential and corrosion potential, potentiostatic electrolysis experiment with gold electrode revealed that quite a high amount of corrosion might be associated during a typical DOER experiment which would limit the usage of gold as an anode material in these processes.

3.4 Electrochemical behavior of platinum electrode in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, 3) melt under anodic polarization conditions

3.4.1 Cyclic voltammetry studies

For studying the usability of platinum electrode as an anode during DOER experiments, it is necessary to have profound knowledge about the electrochemical reactions which can take place on platinum under anodic polarization conditions. Cyclic voltammetry is a simple yet excellent technique for probing into the anodic behavior of platinum. Hence, cyclic voltammograms were recorded in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, 3) melts with platinum working electrode.

Figure 6 depicts the cyclic voltammogram of anodically polarized platinum and graphite electrodes in pure CaCl2 melt at 900 °C. In the forward scan, the onset potential of the anodic current was perceived to be 3.24 V (H) at the graphite electrode, and a linear increase in current was observed upon further anodic polarization. Obviously, the linearly increasing current is due to the chlorine evolution reaction on the electrode as per Eq. 7. In the case of platinum, the anodic current was seen to manifest prominently at 3.14 V (I), followed by a steep rise in current upon further anodic polarization. The anodic current on platinum might be arising due to platinum dissolution reaction as per Eq. 11 (Table 2).

Fig. 6
figure 6

Cyclic voltammograms (anodic polarization) of HD graphite rod (area = 1.236 cm2) in pure CaCl2 melt and platinum wire (electrode area = 0.322 cm2) electrodes in pure CaCl2 and CaCl2–x wt % CaO (x = 0.5, 1, 2, 3) melts (scan rate = 25 mV s−1, temperature = 900 °C)

Full size image

The reduction potential for Eq. 11 was calculated to be 3.17 V [41, 42] which was in close accord with that obtained from the CV data. The slight mass loss of platinum observed during the CV studies with platinum anode supports the possibility of platinum dissolution. Further, the platinum dissolution potential of 3.14 V is close to that of Cl2 evolution (Eq. 7) on inert graphite, implying that this reaction might also be taking place along with the Pt dissolution reaction on the electrode. However, platinum cannot be used as an anode in pure CaCl2, as the primary anodic reaction is the dissolution of platinum (Eq. 11, Table 2). Incorporating oxide in the CaCl2 melt might protect the platinum electrode under anodic polarization conditions.

In the cyclic voltammogram of platinum electrode recorded in CaCl2–0.5 wt% CaO melt under anodic polarization conditions, anodic current was noticed to rise gradually from 2.98 V (J) onwards followed by a steeper increase in current from 3.23 V (K) (Fig. 6). The anode current in the potential window of [2.98 V (J) –3.23 V (K)] can be considered to be due to the O2 evolution reaction as per Eq. 5. The anodic current beyond 3.23 V (K) was probably due to the mixed reaction of platinum dissolution (Eq. 11, Table 2) and Cl2 evolution (Eq. 7). Figure 6 also depicts the electrochemical behavior of platinum electrode in CaCl2–x wt % CaO (x = 1, 2, 3) melts. The magnitude of anodic current in the O2 evolution potential window increased with the concentration of CaO in the melt. Furthermore, the O2 evolution potential window was observed to widen more with the rise in CaO content in the electrolyte. It increased from 0.25 V [(J)–(K)] in CaCl2–0.5 wt % CaO melt to 0.47 V [(L)–(M)] in CaCl2–1 wt % CaO melt (Fig. 6). The window further widened to 0.79 V [(N)–(O)] in CaCl2–2 wt % CaO melt and to 0.99 V [(P)–(Q)] in CaCl2–3 wt % CaO melt (Fig. 6). Due to this widening of the potential window, the onset potentials of platinum dissolution were pushed to higher anodic potentials. This condition is conducive for the platinum anode to remain in the oxygen evolution potential region for a large potential range during a typical DOER experiment and, thus, to avoid the damage of the anode by the dissolution reaction discussed above. Hence CaCl2 melts with higher concentrations of CaO are desirable for the electrochemical reduction of metal oxides with the platinum anode. However, a small cathodic current was observed during the reverse scan of the cyclic voltammograms in CaCl2–x wt % CaO (x = 1, 2, 3) melts, and the current magnitude was significantly less than the corresponding anodic current in the forward scan (Fig. 6). The cathodic current was seen to increase slightly with the increase in CaO content in the electrolyte. One more important observation was that the cathodic current was perceived during less anodic polarizations also, i.e., at potentials less than the platinum dissolution (Eq. 11, Table 2) and Cl2 evolution (Eq. 7) potentials. This behavior might be due to reduction of small amount of O2 present in the vicinity of the electrode or reduction of platinum which would have been oxidized to form Pt3O4 as per Eq. 12 (Table 2) during forward scan.

The reduction potential for Eq. 12 (Table 2) was calculated to be 2.98 V [41, 42], which was within the O2 evolution potential windows in these melts. However, the data obtained from CV studies regarding the genesis of the cathodic current in the cyclic voltammograms of the platinum electrode in CaCl2–CaO melt was insufficient. Further investigations would be required to draw a logical conclusion in this regard.

As revealed by the results of the cyclic voltammetry study with platinum electrode in CaCl2–CaO melt, it was possible to ascertain the oxygen evolution potential windows under anodic polarization conditions. Hence it would be safer to use platinum anode in a typical DOER experiment in CaCl2–CaO melt than gold.

3.4.2 Potentiostatic electrolysis studies

After determining the oxygen evolution potential windows with platinum anode in CaCl2–CaO melts using cyclic voltammetry technique, it seemed necessary to carry out potentiostatic electrolysis studies at different potentials within the window with platinum electrode for elucidating physico-chemical changes (if any) of the platinum anode. This would provide useful information required for conducting DOER experiments with solid metal oxide cathode and platinum anode in CaCl2–CaO melts. Chronoamperometry experiments were conducted in CaCl2–1 wt % CaO melt with platinum coil working electrode at four different potentials (2.98 V, 3.08 V, 3.18 V and 3.28 V) in the potential window of O2 evolution (Fig. 6). The duration of each experiment was 15 min. After the experiments were over, the platinum coil was thoroughly washed with distilled water and dried using acetone. The platinum coils were weighed before and after the potentiostatic electrolysis experiments. Small amount of mass loss was observed for the platinum coils subjected to potentiostatic electrolysis conditions. Rate of mass loss (in g cm−2 h−1) was calculated and the same is depicted in Fig. 7. Higher rate of mass loss was observed at higher anodic polarization potentials. The rate of mass loss was 0.023 g cm−2 h−1 at 2.98 V and it was 0.03 g cm−2 h−1 at 3.28 V in CaCl2–1 wt % CaO melt. It was evident from the chronoamperometry experiments that the rate of mass loss was less in case of platinum electrode than that with gold electrode in CaCl2–1 wt % CaO melt. Similar experiments were carried out in CaCl2–x wt % CaO (x = 2, 3) melts. The polarization potentials were 2.93 V, 3.13 V, 3.33 V and 3.53 V in case of CaCl2–2 wt % CaO melt and 2.92 V, 3.17 V, 3.48 V and 3.67 V in case of CaCl2–3 wt % CaO melt. These potentials were within the oxygen evolution potential windows as determined by cyclic voltammetry studies in these melts (Fig. 6). The rate of mass loss of platinum coil working electrodes used in the chronoamperometry experiments in CaCl2–x wt % CaO (x = 2, 3) melts are also shown in Fig. 7. Similar to that observed in CaCl2–1 wt % CaO melt, higher mass loss rate was noticed with platinum coils polarized to higher anodic potentials in CaCl2–x wt % CaO (x = 2, 3) melts. The rate of mass loss was 0.036 g cm−2 h−1 at 2.93 V and it was 0.065 g cm−2 h−1 at 3.53 V in CaCl2–2 wt % CaO melt. In CaCl2–3 wt % CaO melt, the rate of mass loss of platinum electrode was 0.063 g cm−2 h−1 and 0.11 g cm−2 h−1 at 2.92 V and 3.67 V, respectively. However, it was perceived that for any polarization potential, the rate of mass loss was least in CaCl2–1 wt % CaO melt (Fig. 7). The presence of higher CaO content in CaCl2–CaO melt led to a higher rate of mass loss in the platinum coils subjected to potentiostatic electrolysis conditions. Figure 8 depicts the SEM micrographs of platinum coils before and after potentiostatic electrolysis experiments in CaCl2–x wt % CaO (x = 1, 2, 3) melts. The polarization potentials were 3.28 V, 3.53 V and 3.67 V in CaCl2–x wt % CaO (x = 1, 2, 3) melts, respectively. These potentials were very close to the peak potentials of the oxygen evolution potential windows for CaCl2–x wt % CaO (x = 1, 2, 3) melts, as determined by cyclic voltammetry studies (Fig. 6). Localized pitting was observed on the anodically polarized platinum coil in CaCl2–x wt % CaO electrolyte (Fig. 8). In CaCl2–x wt % CaO (x = 2, 3) melts, pitting became more prominent in the polarized platinum coils, respectively (Fig. 8c, d) compared to that in CaCl2–1 wt % CaO melt (Fig. 8b). In the EDX map sum spectra of the polarized platinum coils, presence of no other element except platinum was observed. The small amount of mass loss and pitting of platinum electrodes when polarized to a potential in the oxygen evolution potential window in CaCl2–CaO melts might be due to the side reaction as per Eq. 12 (Table 2) leading to formation of Pt-O compounds, most probably Pt3O4. Sakamura et al. reported blackening of the immersed portion of the platinum anode during constant current electrolysis in CaCl2–1.9 mol % CaO melt due to formation of Pt3O4 [37], but no such black coating was observed on the platinum electrode used in the potentiostatic electrolysis experiments conducted in the present study. The Pt3O4 layer in this case, might not be stable enough to stick onto the electrode surface and could be getting dislodged from it. The chronoamperometry experiments revealed that the presence of higher CaO content in CaCl2–CaO melt led to higher rate of mass loss and more prominent pitting in the platinum coils subjected to potentiostatic electrolysis conditions. However, the rate of mass loss and pitting of platinum coil was least in CaCl2–1 wt % CaO melt. At the same time oxygen evolution window was also found to be moderately wide at platinum anode in this melt (Fig. 6). Hence the study suggests that platinum can be used as anode in DOER experiments conducted in CaCl2–1 wt % CaO melt.

Fig. 7
figure 7

Rate of mass loss (g cm−2 h−1) vs anodic polarization potentials plot for potentiostatic electrolysis experiments with platinum coil electrode in CaCl2–x wt % CaO (x = 1, 2, 3) melts at 900 °C

Full size image
Fig. 8
figure 8

Scanning electron micrographs of the platinum coil electrode (a) before and (b–d) after potentiostatic electrolysis experiments in CaCl2–x wt % CaO (x = 1, 2, 3) melts, respectively, at 900 °C (magnification: 1000×)

Full size image

3.5 Electro-calciothermic reduction experiments with solid metal oxide cathode and platinum anode in CaCl2–1 wt % CaO melt

Two metal oxides viz., ThO2 and NiO were chosen for electro-calciothermic reduction experiments. ThO2 is a more stable metal oxide with standard molar Gibbs energy of formation of –1005.7 kJ mole−1 of O2 whereas the standard molar Gibbs energy of formation of NiO is –268.8 kJ mole−1 of O2 at 900 °C denoting that NiO is a less stable metal oxide [41, 42]. Electro-calciothermic reduction experiments were carried in CaCl2–1 wt % CaO melt with porous solid NiO and ThO2 cathodes and platinum coil anode at 900 °C.

3.5.1 Electro-calciothermic reduction with NiO cathode and Pt anode

The sintered NiO pellet was tied to SS current collector using a tantalum wire (0.5 mm dia.) and the whole assembly acted as the cathode in the electro-calciothermic reduction experiment. Pt coil (prepared out of 1 mm dia. wire) was used as the anode. A constant cell voltage of 3.1 V was applied between NiO cathode and Pt anode for ~ 2.5 h. The current vs time and potential vs time plots are shown in Fig. 9. At the beginning of electrolysis, 1.42 A current was recorded. However, it immediately dropped to ~ 0.7 A and subsequently gradual decrease in current was observed during electrolysis. At the end of electrolysis, current was recorded as 0.44 A. The ratio of total charge passed (Q) and theoretical charge (Qo) required for complete reduction of NiO was 3.6. Figure 9 also shows the cathode potential and anode potential recorded during electrolysis. The anode potential was well within oxygen evolution potential (~ 3 V) and the cathode was at calcium deposition potential. After electro-calciothermic reduction, the product was washed thoroughly with distilled water for complete removal of occluded salt and acetone dried. The product was subjected to XRD and SEM analyses. The X-ray diffraction patterns of the NiO pellet before and after the electrolysis experiment are depicted in Fig. 10. Ni was the only phase in the electrolyzed product. The photographs of NiO pellet and the electro-reduced Ni product are also shown in Fig. 10. Scanning electron micrographs of NiO and electro-reduced Ni are presented in Fig. 11. In the SEM image, nodular morphology typical of an electro-reduced metal was perceived in the whole matrix of the electro-reduced Ni (Fig. 11b). The reduced Ni product was analyzed by energy dispersive X-ray fluorescence (EDXRF) (EX-2600 Genius IF SDD) employing the standardless fundamental parameters (FP) approach. The elemental composition of the reduced product was determined to be Ni (97 wt %), Ca (1 wt %), Cl (1.6 wt %) and Ta (0.4 wt %). The presence of Ca and Cl is most likely due to the presence of a minor amount of residual CaCl2 electrolyte. The presence of Ta could be from the Ta tying wire used for preparing the cathode assembly. The platinum coil was weighed before and after the electro-reduction experiment and the rate of mass loss was minimal (0.022 g cm−2 h−1). Descallar-Arriesgado et al. studied electro-calciothermic reduction of NiO powder using graphite anode in CaCl2–CaO melt at 900 °C [45]. Various concentrations of CaO (0.5 to 5.0 mol %) containing CaCl2 melts were examined to achieve complete reduction of NiO. The ratio of Q/Qo ranged from 1.29 to 1.97. The study suggested that electrolysis in 1 and 3 mol % CaO containing CaCl2 melt yielded a single-phase Ni with less concentration of residual oxygen, whereas both NiO and Ni phases were observed in the XRD pattern of the electrolyzed product in CaCl2–x mol % CaO (x = 0, 0.5, 2, 5) melts. The present study reveals that electro-calciothermic reduction of solid NiO to metallic Ni is possible with platinum anode in CaCl2–1 wt % CaO, i.e., CaCl2–1.94 mol % CaO melt at 900 °C.

Fig. 9
figure 9

The variation of current and half-cell potentials (vs Ca2+|Ca reference electrode) with time during constant voltage electrolysis experiment at 3.1 V with NiO pellet cathode and platinum anode in CaCl2–1 wt % CaO melt; Temperature: 900 °C

Full size image
Fig. 10
figure 10

XRD patterns of the (a) sintered NiO pellet and (b) reduced Ni product after the electro-calciothermic reduction experiment with NiO cathode and platinum anode in CaCl2–1 wt % CaO melt at 900 °C. Inset shows the photographs of NiO pellet and Ni metal

Full size image
Fig. 11
figure 11

Scanning electron micrographs of cross-sections of NiO pellet (a) before and (b) after electro-calciothermic reduction experiments with NiO cathode and platinum anode in CaCl2–1 wt % CaO melt at 900 °C; Magnification: 1000×

Full size image

3.5.2 Electro-calciothermic reduction with ThO2 cathode and Pt anode

Electro-calciothermic reduction experiments of ThO2 were carried out similarly to NiO reduction with platinum anode. The electrochemical reduction experiment with ThO2 pellet cathode was conducted for different time durations (2 h and 8 h) and the cell voltage was fixed at 3.2 V. The current vs time and potential vs time plots for 8 h electrolysis experiment are shown in Fig. 12. The Current was 1.36 A immediately after commencement of electrolysis. It reduced afterwards and a stable current of ~ 0.59 A was recorded. Platinum anode potential was fixed at ~ 3.1 V during the electrolysis run and the potential of ThO2 cathode was at calcium deposition potential. The ratio of total charge passed (Q) and theoretical charge (Qo) required for complete reduction of ThO2 was 16.5 for 8 h electrolysis experiment. The electrolyzed products were analyzed by XRD without removal of occluded salt. The X-ray diffraction patterns are shown in Fig. 13. Both ThO2 (major) and Th (minor) phases were observed in the 2 h electrolyzed product. In the case of 8 h electrolyzed product Th was present as major phase and ThO2 appeared as the minor phase. The photographs of ThO2 pellet before and after the electrolysis (2 h and 8 h) are also shown in Fig. 13. 8 h electrolyzed product was also characterized by SEM. Figure 14a–d depicts the scanning electron micrographs of ThO2 pellet before and after the electro-calciothermic reduction experiments. The cross-sectional morphology of the electrolyzed product was different across the thickness of the pellet. In the region near the edge (area 1) nodular morphology of the electro-reduced Th metal could be observed (Fig. 14c). On the other hand, in the central region (area 2) along with nodular structures of Th metal, agglomerates similar to that in SEM image of sintered ThO2 pellet (Fig. 14a) was also observed. Previous studies have revealed that the electro-reduction of ThO2 in CaCl2 based melts is kinetically sluggish because of the high thermodynamic stability and electrical non-conductivity of ThO2 [16]. This suggests that electrolysis time of more than 8 h might be necessary for the complete conversion of ThO2 to Th. However, it is worth mentioning that platinum can only be used as anode in electro-calciothermic reduction of solid ThO2 in CaCl2 –CaO melt for production of Th metal as usage of graphite anode leads to formation of ThC in the cathode product [16]. This is not desirable as Th metal devoid of impurities is only acceptable for nuclear technology. Previous experiments carried out in our laboratory with ThO2 cathode and graphite anode in CaCl2–0.5 wt % CaO melt at 900 °C conducted for 16 h resulted in black coloration of the electrolyzed product (Fig. 15a) and significant amount of thinning was also noticed in the graphite anode after electrolysis (Fig. 15b). The melt was also black in color after electrolysis due to extensive erosion of graphite anode. XRD analysis of the electrolyzed product revealed that significant amount of ThC was present in the product along with Th and ThO2 [16]. In the present study the platinum coil anode used in the electro-calciothermic reduction experiment with ThO2 cathode was found to be stable and the photograph of the platinum anode after the DOER experiment is shown in Fig. 16a. Minimal amount of mass loss was observed after electrolysis. The rate of mass loss was found to be 0.016 g cm−2 h−1. Figure 16b depicts the scanning electron micrograph and EDX map sum spectrum of the platinum coil after the electro-calciothermic reduction experiment. Presence of no other element except platinum was observed in the EDX map sum spectrum of the polarized platinum coil. The melt was found to be clean after the electrolysis and the same can be reused after adjusting the CaO content.

Fig. 12
figure 12

The variation of current and half-cell potentials (vs Ca2+|Ca reference electrode) with time during constant voltage electrolysis experiment at 3.2 V with ThO2 pellet cathode and platinum anode in CaCl2–1 wt % CaO melt; Temperature: 900 °C

Full size image
Fig. 13
figure 13

XRD patterns of the ThO2 pellet (a) before and after electro-calciothermic reduction experiments for different time durations [(b) 2 h and (c) 8 h] with ThO2 cathode and platinum anode in CaCl2–1 wt % CaO melt at 900 °C. Inset shows the photographs of the ThO2 pellet before and after electrolysis experiments

Full size image
Fig. 14
figure 14

Scanning electron micrographs of cross-sections of ThO2 pellet (a) before (magnification: 1000×) and (b–d) after electro-calciothermic reduction experiments with ThO2 cathode and platinum anode in CaCl2–1 wt % CaO melt at 900 °C conducted for 8 h; (b) denotes the SEM image of the complete cross-section of the electrolyzed product (magnification: 80×); (c) and (d) depict the SEM images (magnification: 1000×) of area 1 and 2, respectively, as shown in (b)

Full size image
Fig. 15
figure 15

Photographs of (a) the ThO2 pellet after the electro-calciothermic reduction experiments in CaCl2–0.5 wt % CaO melt at 900 °C conducted for 16 h with graphite anode; (b) graphite anode before and after the electrolysis experiment

Full size image
Fig. 16
figure 16

(a) Photograph of the platinum coil anode and (b) scanning electron micrograph and EDX map sum spectrum (magnification: 500×) of platinum coil electrode after electro-calciothermic reduction experiments in CaCl2–1 wt % CaO melt at 900 °C

Full size image

Hence the electro-calciothermic reduction experiments conducted in the present study with platinum anode and ThO2 cathode in CaCl2–1 wt % CaO melt revealed that it is feasible to use platinum as anode material for conversion of ThO2 to Th metal without formation of ThC.

4 Conclusions

The electrochemical behavior of nickel, graphite, gold and platinum electrodes was studied in CaCl2–1 wt % CaO melt using linear polarization (LSV) and potentiodynamic polarization (Tafel plots) studies. Platinum exhibited the highest anodic polarization potential and most positive corrosion potential (Ecorr). The EIS studies with graphite, gold and platinum electrodes in the same melt revealed that platinum possessed the highest oxidation resistance. These electrochemical investigations denoted that platinum might be the best anode material among the electrodes studied for DOER experiments in CaCl2–CaO melts. The cyclic voltammetry studies with gold electrode in CaCl2–CaO melts revealed that determination of O2 evolution potential window was difficult and rate of mass loss of gold electrode during the potentiostatic electrolysis studies in CaCl2–1 wt % CaO melt was very high (0.2 g cm−2 h−1). The electrochemical behavior of platinum electrode was investigated in CaCl2–x wt % CaO (x = 0, 0.5, 1, 2, 3) melts by cyclic voltammetry under anodic polarization conditions. Results showed that the O2 evolution potential window could be determined clearly, and the potential window widened with an increase in CaO concentration in the melt. A low rate of mass loss of the platinum electrode was observed during potentiostatic electrolysis experiments at different potentials within the potential window of O2 evolution in CaCl2–x wt % CaO (x = 1, 2, 3) melts, and the corrosion rate was less than that observed with gold electrode. The rate of mass loss of the platinum electrode was least in CaCl2–1 wt % CaO melt. SEM analysis of the platinum electrodes anodically polarized in CaCl2–x wt % CaO (x = 1, 2, 3) melts denoted that the pitting on the electrode surface was least in CaCl2–1 wt % CaO melt, whereas it was highest in CaCl2–3 wt % CaO melt. These results suggest that platinum can be used as an anode material in CaCl2–1 wt % CaO melt with the least rate of mass loss for DOER experiments. To validate the usage of Pt anode, electro-calciothermic reduction experiments were carried out with NiO and ThO2 pellet cathodes and platinum coil anode in CaCl2–1 wt % CaO melt at 900 °C. Complete metallization of the NiO cathode was observed with a low rate of mass loss of platinum (0.022 g cm−2 h−1). In case of ThO2, presence of both Th and ThO2 was perceived in the electrolyzed product because of high thermodynamic stability and electrical resistivity of ThO2 and the rate of mass loss of platinum was found to be 0.016 g cm−2 h−1. Presence of ThC was also not observed in the electrolyzed product. The results indicate the possibility of using platinum as anode in the electrochemical reduction of solid metal oxides to metals in CaCl2–1wt % CaO melt with a minimal rate of mass loss of the platinum electrode.