1.Introduction
Pyroprocessing, a nuclear reprocessing technology, has attracted the interest of researchers both in academia and industrial circles due to its advantages: enhancing resourceuse efficiency while reducing the volume and toxicity of nuclear waste with less proliferation risk than the PUREX reprocessing technology. Pyroprocessing generally consists of several unit processes; head-end process, electrolytic reduction, electrorefining, electrowinning, and treatment of waste. During the first head-end process, the spent fuel is recovered from the cladding hull and converted into an oxide feed material for the following electrolytic reduction process. The feed material is then electrochemically reduced to metal form in a molten salt followed by other unit processes to recycle the useful resources in the spent fuel by utilizing them as fuels for sodium-cooled fast reactor [1-4].
For the eventual success in the realization of pyroprocessing, its unit processes must be confirmed strictly including in all aspects: technical stability, environmental sustainability, and economic efficiency in a wide range of scale (from a lab to an enlarged scale). There have been many attempts to study the unit or partially integrated processes of pyroprocessing on lab scales. Based on the lab scale experience, KAERI constructed an engineering scale demonstration facility (PRIDE) for engineering scale pyroprocessing. The pyroprocessing in PRIDE is similar with the general case from head-end to waste treatment. During the first head-end process, a feed material for the following electrolytic reduction process should be prepared. Fragments, granules, and porous pellets are among the available physical forms of the feed material [4-11]. Although each form has respective strengths and drawbacks depending on its physical properties, the porous pellet is considered a superior feed form due to its well-known advantages as described in [12]. In addition to the general advantages, to be a successful practical feed material in the PRIDE facility, the porous pellet should have not only proper bulk and tap density for mass production but also similar chemical characteristics with regard to spent fuel consisting of the UO2 matrix dissolved in other oxides, oxide and metal precipitates. These physical and chemical properties were considered as the criteria for the suitability of the produced UO2 Simfuel porous pellet as a feed material for the electrolytic reduction in the PRIDE. In the head-end process, the chemical state of the starting material can be varied by final pre-treatment of the starting material UO2 after reduction and U3O8 after oxidation. It is therefore necessary to fabricate the porous pellet with both uranium oxides (UO2 and U3O8) as starting material. Previously, we fabricated UO2 porous pellets on an engineering scale with the U3O8 powder as a starting material for pelletizing in the PRIDE facility [11].
The present work deals with the engineering scale fab- rication of UO2 porous pellets (30 kg-U/batch) with UO2 powder as a starting material for the electrolytic reduction process in the PRIDE facility. This includes modifications of the processing and equipment during each unit process and reflection of these changes on the products. The produced porous pellet was characterized physically and chemically to evaluate the processing, equipment and also the product in terms of the suitability for the electrolytic reduction in a molten salt. This work is therefore carried out not only to supply the feed material in the PRIDE facility but also to provide collective information on tailoring the fabrication process and related equipment for the mass production of UO2 porous pellets for pyroprocessing research on an engineering scale.
2.Experimental
The UO2 Simfuel porous pellets were fabricated on an engineering scale via a process that is similar to conventional ceramic processing with some improvements [11]. Table 1 shows the overall fabrication flow including the material and equipment in each unit process for the porous pellet in the PRIDE facility. The fabrication process in this work consists of four sequential unit processes: preparation of starting materials, mixing, pelletizing, and sintering. DU UO2 powder and surrogate oxide powders were prepared as starting materials. Table 2 shows the particle size of the surrogate oxide powders measured by laser diffraction method (Beckman LS-230, USA) including comparison with their specifications. Before mixing with the DU UO2 powder, the surrogate oxide powders were precisely proportioned according to the results calculated by ORIGEN (Oak Ridge Isotope Generation and Depletion) code (v. ORIGEN-ARP 5.1.01). The purity of the surrogate powders were: 99.999% for Y2O3, 99.99% for La2O3, CeO2, and TeO2, 99.95% for MoO3, 99.9% for Pr6O11, Nd2O3, Sm2O3, Eu2O3, Gd2O3, MoO3, RuO2, Rh2O3, and PdO, 99.8% for BaO, 99.5% for SrO, and 99% for ZrO2. The proportioned surrogate powders were mixed in a Turbula mixer and attrition-milled for 4 h in a high purity ethanol medium with 5 mm ZrO2 balls before being mixed with UO2 powder. After drying with a hot plate and a drying oven for 24 h, the milled surrogate oxide powders were mixed with UO2 powder in a tumbling mixer for 4 h with a rotation speed of 38 rpm and then mixed 30 min more with the addition of 0.4wt% of EBS (ethylene bis stearamide, C38H76O2N2, CAS no. 110- 30-5) as a lubricant. The homogeneously mixed powders were pelletized using a rotary press. To improve the filling uniformity of the powder when it is put in a die hole, a mechanical feeder system was installed, as described in our previous report, for better flowability of the mixed powder [11]. The powder filled in the die hole was compacted by upper and lower punches at a range of 200-500 MPa pressure. Thirteen green pellets were produced by the 13 punch/die sets with rotation of the driving part using the rotary press. The green pellets were put into molybdenum crucibles that can hold ~13 kg of pellets and sintered in a double chamber furnace at 1450℃ for 24 h in a 4% H2-Ar balanced atmosphere.
The microstructures of the sintered porous pellet were observed by SEM (Philips XL-30, Netherlands) and EDS (Energy Dispersive Spectroscope) attached on the SEM from the fracture surface of the pellet after vertical crushing. The crystal structure of the pellets was observed by XRD (Rigaku Mini-Flex, Japan) over the two theta range of 20~80° and at a scan speed of 6°·min-1 and a step size of 0.01° after crushing the sintered pellet in an agate mortar and pestle. The bulk density of the pellet was measured by the general water immersion method and was also calculated from the weight and dimensions of the pellet. The average value was determined by measuring 10 randomly sampled pellets.
3.Results and Discussion
The fabrication of the UO2 porous pellet with a Simfuel composition in this study was carried out according to the fabrication process described in Table 1 with a composition of Table 3 that simulates the chemical composition of spent fuel. For the preparation of starting materials, DU UO2 powder was prepared without any pre-treatment while the surrogate oxide powders were mixed and milled due to their large particle sizes as shown in Table 2. The large particles may result in drawbacks during the following pelletizing process. In the case of the DU UO2 powder, its physical properties were analyzed: O/U ratio, surface area, and water content were measured as 2.196, 2.51 (m2·g-1), and 3190 (ppm), respectively. This slightly larger O/U ratio than 2 of normal UO2 indicates the surface of the UO2 particles is slightly oxidized and includes UO2+x and U4O9-y phases. The lower surface area and higher water content compared to the ADU UO2 powder [13] would cause poor compactability and sinterability in the following unit processes.
Fig. 1 shows schematic drawings and photographs of the constructed equipment in the PRIDE facility. The UO2 and surrogate powders were mixed using a tumbling mixer, shown in Fig. 1 (a), which is equipped with a driving motor, rotating shaft, and mixing jar. The mixed powder was compacted into a pellet shape by the rotary press shown in Fig. 1 (b). The concept of the equipment is similar to that of the commercial press machine used in pharmaceutical fields. The driving part contains 13 punch/die sets to uniaxially press the powder with the rotation shaft at a high production yield. The green pellets were then sintered in a furnace with a double chamber structure, as shown in Fig. 1 (c). The schematic drawings show both outer and inner views presenting the mechanical gate structure and refractories with heating element, respectively.
Fig. 2 shows SEM micrographs of UO2 and attrition milled surrogates before mixing them in the tumbling mixer. It is noted that the particle size of the starting materials appears to be up to ~ 30 μm for the attrition milled surrogates and a few microns for UO2. The particle size of the raw surrogate powders measured by the laser diffraction method ranged from a few microns to a few hundreds of microns. The particle size therefore should be reduced to raise the mechanical stability of the green pellet after compaction. As a result of attrition milling of the surrogates as pre-treatment before mixing with UO2, finer size was measured: the largest particle was about ~ 30 μm, as shown in Fig. 2 (b). The surrogate powders were then placed in an inclined mixing jar and mixed with UO2 powder by rotation of the shaft. The powders were homogenized by diffusional mixing with the motion of individual particles when the mixing jar rotates. The continuous rotation provides the inclined plane with fresh powder, leading to intermixing of the powder particles. The mixing process has two different objects: the first mixing is carried out to mix the UO2 powder and surrogates whereas the second is to mix the homogenized UO2 and surrogate powders with EBS, which acts as a lubricant. In this experiment, the amount of added EBS was slightly higher (0.4wt%) than the usual amount (~ 0.2wt%) to enhance the compactibility during the following pelletizing process.
Fig. 3 presents details of the modification of the mechanical feeder and the produced green pellets with application of the modified feeder. The mechanical feeder is attached at the driving part of the rotary press to improve the flowability of powder during the pelletizing process. This helps the powder supplied from the feeder move to the die holes by sweeping motion of the impeller wings. The cross sectional shape of the impeller wings in the mechanical feeder strongly affects the sweeping behavior and therefore the powder filling behavior. Three different impeller wing shapes, two rectangular shapes (with and without inclined plane) and a triangular shape, were compared and it was thereupon concluded that the rectangular shape with an inclined plane shown in Fig. 3 (a) provides superior powder filling behavior compared to the others. This is determined from the amount of the filled powder into the die holes and deviations. In the case of using the rectangular shape with an inclined plane, higher average weight of green pellets and smaller standard deviations were measured as 1.44 g and 0.07 g, respectively, from randomly sampled compacted green pellets. The compacted green pellets generally showed a smooth and glossy surface without any defects as shown in Fig. 3 (b). The green pellets were put into molybdenum crucibles to be sintered.
Fig. 4 presents a thermal profile of the sintering process and a photograph of sintered UO2 porous pellets. High temperature heat treatment at 1450℃ was carried out after de-waxing at 700℃ for 24 hours in a reducing atmosphere using 4% H2-Ar balanced gas. During sintering, the temperature and atmosphere were ensured by a double chamber structure: at inner chamber surrounded by refractories confined the heat while the outer chamber of stainless steel maintained the reducing atmosphere without leakage. As crucibles to stack a large amount of pellets, a molybdenum crucible of 3 mm thickness was designed with loosely assembled bottom and wall. There are also many holes at the bottom plane, allowing the green pellets to contact atmospheric gas by penetration of the gas through the holes, even when the crucibles are piled up vertically, as shown in the inset in Fig. 4 (b). Temperature variations measured by a thermocouple at the inner chamber were consistent with the programmed behavior until the end of isothermal treatment. As a result of grain growth and densification during sintering, sintered UO2 porous pellets with a Simfuel composition were obtained as shown in Fig. 4 (b).
The chemical and physical properties of the final product, i.e., a sintered UO2 Simfuel porous pellet, were analyzed to evaluate its suitability for the following electrolytic reduction process as a feed material. The XRD pattern of the sintered pellet in Fig. 5 indicates a single UO2 phase at the level of the XRD detection limit. It also suggests the reducing atmosphere was maintained without change during sintering. In addition, it is shown in Table 4 that the chemical composition values of three sintered pellets were similar and also similar to the starting material, indicating favorable chemical homogeneity. Although the boiling points of some surrogate oxides such as MoO3, RuO2, Rh2O3, and PdO are lower than the sintering temperature, their fraction appear to be maintained possibly due to increase of boiling points with their reduction to metallic phases before volatilization. In the case of TeO, however, its fraction was significantly decreased after sintering maybe due to preservative oxide phase, which has lower boiling point (~1245℃). The microstructures shown in Fig. 6 are observed at the (a) inside and (b) surface of the sintered pellet by SEM. Similar microstructures in terms of grain size and shape, apparent density, and presence of secondary phases (denoted as arrows) were obtained regardless of the position. These are typical grain and pore structure, grains connected with necks and interconnected pores among grains, suggesting the initial or intermediate stage of sintering. Compared to the initial particle size of UO2 shown in Fig. 2 (a), the grain size slightly increased and the average bulk density also increased from the green pellet of 5.84 g·cm-3 (53.29% TD) to the sintered pellet of 6.89 g·cm-3 (62.94% TD). In this case, the tap density of the sintered pellets was 4.88 g·cm-3, which allows the pellets to be stacked in the cathode basket completely. The volume of the cathode basket has been determined based on the assumption that the tap density of feed materials would be of 4.5 g·cm-3. Therefore the produced Simfuel porous pellets could be loaded into the basket satisfying the intended processing scale of PRIDE facility. The secondary phases were characterized by an EDS attached to a SEM, as shown in Fig. 7. Fig. 7(a) shows the metal precipitates are located at the grain boundaries and junctions as protrusions on the UO2 matrix grains. On the other hand, the oxide precipitate is shown in Fig. 7 (b) as bundles of small grains located between the matrix grains. Both are well distinguished by their morphology and are well-documented, typical characteristics of spent fuel, indicating the suitability of the UO2 Simfuel porous pellet as a feed material for the electrolytic reduction experiment.
4.Conclusions
For the electrolytic reduction experiment in the PRIDE facility, the UO2 Simfuel porous pellets have been fabricated successfully on an engineering scale with modifications of processing and equipment. Processing parameters and equipment were modified and optimized to secure production efficiency and processing stability with controlled product quality. As a final product, the UO2 Simfuel porous pellet was characterized regarding the requirements of the electrolytic reduction process. Macroscopic properties such as size, density, and phase and microscopic grain/ pore structure and distribution of the second phase meet the needs of engineering scale electrolytic reduction experiments in PRIDE.