1.Introduction

The calciothermic reduction of PuO₂ generates large volumes of waste because of the need to dissolve the CaO reaction by-product in molten CaCl₂ [1]. An alternative electrochemical method for carrying out the reduction has also been demonstrated [2]. However, it too is challenging to implement as it requires incremental additions of PuO₂ to dissolve in the melt.

In light of the recent demonstration of the processing of TiO₂, there has been considerably renewed interest and investigation of electrochemical reduction methods [3]. The processing of metal oxides in molten salts has been demonstrated to successfully produce a wide variety of metals and alloys. Such processes are of particular interest to the civil nuclear energy industry in pursuit of a closed loop fuel cycle.

Although there are numerous studies of electrochemical reduction of UO₂, there are none of PuO₂ except as a component of mixed oxide nuclear fuel (UO₂, PuO₂). For example, Iizuka et al demonstrated complete reduction of (47wt% U, 35wt% Pu, 4wt% Np,)O₂ pellets to alloy in LiCl-0.5wt% Li₂O [4]. Kurata et al demonstrated electrochemical reduction of (90.55wt% U, 9.45wt% Pu,)O2 in LiCl-1wt% Li₂O[5]. These studies suggest that UO₂ reduction precedes PuO₂ reduction.

The typical configuration for electrochemical reduction requires forming the metal oxide into a pellet which acts as the cathode [3]. In the case of plutonium oxide, preparation of pellets is particularly undesirable as it requires additional processing and leads to greater operator radiological dose. If a free-flowing fine oxide powder were to be immersed in the molten salt of an electrochemical cell, it is anticipated that particle buoyancy and the effect of interfacial tension would tend to insulate the oxide particles.

This communication reports upon the electrochemical reduction of plutonium oxide powder contained within a permeable steel basket cathode in CaCl₂-1wt% CaO at ca. 1123 K. As plutonium forms a dense low temperature eutectic with iron, the container is sacrificed during the process thereby demonstrating the viability of the reduction.

2.Experimental

2.1.Chemicals

Calcium chloride dihydrate (CaCl₂.2H₂O) and calcium oxide (CaO) were supplied by Fisher Scientific Ltd. Each was dried in an oven for five days under reduced pressure (ca. 1.3x10³ Pa) in a stepped programme to 230°C.

Plutonium was obtained from the United Kingdom’s national inventory. Plutonium(IV) oxide was prepared by the oxidation of plutonium metal in oxygen enriched air at 400°C. Chemical analysis was carried out and confirmed as >98.5% purity. Particle size analysis of PuO₂ produced by this method suggests that 1% by weight is smaller than 2μm and approximately 2% by weight is smaller than 5μm [6].

2.2.Apparatus

A permeable steel basket was made from Poroloy^TM tube supplied by Purolator-Facet Ltd. Poroloy is a multilayered composite of 347 stainless steel wire cloth (0.07 mm diameter to 0.13 mm diameter). The nominal pore size of the resulting material is 5μm and should therefore, in an unstirred system, very largely retain PuO₂. A blank steel disc was welded at its base to form a closed end. The basket was 25 mm internal diameter by 60 mm length. Once filled with plutonium oxide, a stainless steel plunger was rested on the surface of the oxide to partially enclose it and provide additional support and electrical contact. This acted as the cathode.

A graphite anode (99.99%, 10 mm diameter by 150 mm length, UF-4S grade) was prepared in-house.

The crucible was Al₂O₃ and supplied by Seagoe Ltd. This crucible had parallel sides and a working volume of ca. 250 cm³.

The study was conducted in a glovebox with high quality dry nitrogen atmosphere (H₂O < 20 ppm, O₂ < 0.5%). The glovebox is equipped with a resistively heated furnace extending below the floor. During heating, the furnace well is purged with dry argon.

An Autolab PGstat 20 potentiostat (Ecochemie BV) in conjunction with a 20Amp Booster was used for electrochemical studies. Bespoke cable assemblies were prepared by Ecochemie (BV) in order that the resistances imposed by the long cable lengths through glovebox feedthroughs could be compensated.

2.3.Procedure

Cast CaCl₂-1wt% CaO (228.9g) was weighed into the crucible and the crucible placed in the furnace. PuO₂ (26.5 g) was placed in the permeable steel basket and assembled onto the electrochemical cell. A graphite rod anode was also attached. The experimental apparatus is shown in Figure 1a.

The furnace was heated to a heating element set point of 950°C. It was estimated to be approximately 1123 K in the melt.

The electrodes were immersed in the molten salt to a depth of approximately 6.5 cm. The upper edge of the cathode compartment remained above the surface of the melt. The cell was allowed to equilibrate for several hours.

A potential of -3.1V, slightly greater than the potential window for CaO, was applied between the electrodes for 45.8 hours. This potential permits for both direct and indirect reduction mechanisms.

The electrodes were raised and the furnace allowed to cool. The furnace contents were then unloaded. The cylindrical cathode was trimmed using snips and subsequently sectioned on a low speed saw with a diamond blade to obtain the contents as six discs.

The analysis of Pu/PuO₂ used an established in-house method which has been demonstrated in the analysis of plutonium and plutonium oxide. It relies upon the differing solubilities of Pu and PuO₂ in HCl and HNO³/HF [Equations 1-3].

A p-type High Resolution Gamma Spectrometer and a certified mixed isotope gamma standard were employed.

Total oxygen was determined using an Eltra ONH 2000 Determinator. Subtraction of %PuO₂ [Equation 3] from total oxygen was used to infer % CaO.

%CaCl₂ analysis was conducted using an Ion Selective Electrode. Total calcium was determined by ICP-AES. Subtraction of %CaO and %CaCl₂ from total calcium was used to infer %Ca metal.

3Results and Discussion

The current-time trace obtained during electrolysis at an applied potential of -3.1V is given in Figure 2. The decrease in current may suggest that the reaction was essentially complete after 15 hours. Given the challenges in conducting such experiments, electrolysis was extended for a further period.

During the electrolysis, the current was periodically interrupted to save data to the PC. The current was interrupted for 60s. The current was higher upon resumption of electrolysis giving discontinuities in the current-time trace. This behaviour is consistent with depletion of charge carrying oxide ions at the anode during electrolysis and replenishment during the interruption. The total charge passed (256082C) was 6.7 equivalents.

Figure 1b shows the electrochemical cell following electrolysis. There was a significant amount of black material on the exterior of the cathode and a quantity of black nodular material near the bottom of the cathode. The material on the exterior of the cathode may have moved, and therefore appear to bridge to the anode, during retrieval of the cell from the molten salt.

The graphite anode lost 1.6 g during electrolysis. The narrowing of the anode suggested the graphite was preferentially consumed near the surface of the melt.

The uppermost (1^st) and lowest (6^th) discs produced by sectioning of the steel basket are shown in Figure 3. The sections became increasingly difficult to cut further down the basket. This would be consistent with production of liquid Pu which would tend to coalesce at the base of the cathode.

Each of the six discs had a dark central region, a pale region around that and a light coloured region near the steel basket wall. The dark central region became wider further down the cathode. Four of the discs were submitted for chemical analysis. Samples in each disc were taken from the dark centre and the results of the analyses are given in Table 1. Separate samples were used to determine %Pu_total and %Pu_metal. In the case of the uppermost section, errors were expected due to difficulty in obtaining samples solely of the smaller dark central region.

The Pu metal content of the permeable basket was found to increase from top to bottom. The mean Pu metal content was 35%. Based upon the contents of the basket only, current efficiency was 12.9%. The remaining charge is presumably associated with reduction of carbonate and formation of Ca. Ca was distributed in the basket with a mean of 8.6% by weight. The PuO₂ was distributed with a mean of 3%.

The black nodular material found on the exterior of the cathode had broken into three parts during handling. A denser part was found to be soluble in HCl and found to be ca. 56% by weight Pu metal. Two less dense and HCl insoluble parts were submitted for XRF analysis that suggested a significant quantity of calcium, iron and carbon were present. It is proposed that the nodular material was plutonium metal formed within the cathode and subsequently attacking the steel. Pu forms a low melting point alloy with Fe. The black colour of the nodular material was due to contamination with carbon arising from reduction of dissolved carbonaceous species at the cathode.

A sample of the electrolyte was submitted for analysis and found to be largely unchanged.

4.Conclusions

The laboratory scale electrochemical reduction of plutonium oxide powder contained in a permeable steel basket is demonstrated in CaCl₂-1wt% CaO. As expected, the Pu metal product reacted with the steel basket and was contaminated with carbon from the anode. The sacrifice of the cathode and extensive carbon contamination of the product are significant challenges to the application of this process and may be the focus of future work. Lessening the charge applied, optimising the potential and alternative cathode materials may address these challenges.