The separation strategy for the UREX+ type processes is based on an optimization approach for which key objectives of the fuel cycle are pre­determined. Currently the US operates an “open fuel cycle,” in which all of the spent fuel is scheduled for disposal. Recycling of all of the fissile com­ponents, as is currently done in several countries, yields a partially “closed fuel cycle.” Recycling of both the fissionable and fissile components is referred to as a completely closed fuel cycle, although not all components are fully transmuted. Because a fast reactor is required to transmute the fissionable components, the completely closed cycle is often called a “fast reactor fuel cycle.” All fuel cycles generate high-level waste, but the volume and radiotoxicity is a function of the extent to which the spent fuel is sepa­rated and critical components transmuted. Isolation of specific fission prod­ucts, in addition to the actinides, can have a significant bearing on the performance of the wastes and repositories.

For a closed fuel cycle, the first step is to define the specific target prod­ucts that need to be recovered to meet the requirements of the selected recycle and waste options. The second step is to determine the feedstock to the separations process. The feedstock is a selection of SNF from the current inventory that is to be processed to yield the desired products. The third step is to determine the process flowsheet. Given the large variation within the SNF inventory, the process flowsheets are designed to accommodate the wide variation in composition. The flowsheet selection methodology identifies what combination of solvent extraction systems accomplishes the segregation of the specified target products. Processing options need not be limited to solvent extraction; ion exchange, crystallization, electrochemical and other processes can be used in combination. However, the integration requires that the processes be compatible chemically. In summary the UREX+ approach requires:

1. Definition of the specific target products that need to be separated.

2. Determination of the feedstock for the separations process; historically it is SNF, but targets and other forms of nuclear waste may be used.

3. Selection or development of technologies that can accomplish the desired separations so that the final products and waste forms meet either the product specifications or waste acceptance criteria.

Czech scientists proposed the mixture of CCD with bifunctional neutral organophosphorus compounds (BNOPC) for the extraction of actinides, lanthanides and cesium (strontium was inadequately recovered in this case) [23]. Studies at the Radium Institute have shown [24,25] that the synergistic mixtures of CCD with long-chained (n > 5) phosphorylated polyethylene glycols (PGP — Ph2P(O)CH2O(CH2CH2O)nCH2P(O)Ph2) can simultane­ously recover cesium, strontium, actinides and lanthanides from HNO3 media.

None of the then known extraction systems had such unique extraction properties, and these new developments therefore permitted a universal extraction mixture (UEM) to be patented in 1990. This mixture was suc­cessfully proven to work for simulated and real HLW. Tests were conducted on the recovery of radionuclides from a solution simulating the PUREX process raffinate, containing 1 M HNO3 and the following metals (g/l): REE — 2.1; cesium — 0.3; strontium — 0.25; sodium — 1.5; iron — 1.0 with a total salt content of ~10 g/l. From this solution, the UEM (containing 0.14 M CCD and 0.08 M PGP-300 in metanitrobenzotrifluoride (F-3)) recovered radio­nuclides with distribution ratios of: U(VI) — 6.0; Pu(lV) — 5.1; Np(V) — 4.0; Np(VI) — 6.4; Am — 1.2; Cs — 2.6; Sr — 16.

Extraction of several radionuclides from HLW was proven in a continu­ous regime in the centrifugal extraction bench EZ-40-18 at Mayak PA (Fig. 9.1) by using a UEM composition of 0.14 M CCD and 0.08 M PGP-300 in F-3, with density of 1426 g/cm3 and viscosity 5,4 mPa[8]s at 20°C.

More than 20 L of simulated HLW were treated within 20 hours, with the organic phase making over 15 recycles. Cesium, strontium and REE were

and cumbersome process, which is difficult to achieve even on a laboratory scale.

These circumstances have encouraged scientists at the Radium Institute (RI) and the Idaho National Laboratory (INL), USA, to collaborate on the modification of the above extraction system, which has resulted in the elaboration of the so-called UNEX process (UNiversal EXtraction process). In the UNEX process, a mixture of the commercial reagents CCD, PEG-400 and diphenyl-N, N-dibutylcarbamoyl-phosphinoxide is used in the diluent as the extraction system.

10.3.1 Design characteristics of pyroprocess equipment

As for aqueous reprocessing, an economically feasible facility should have a throughput as much as several hundred tons HM/y. On the other hand, a pyrochemical reprocessing facility is expected to be economically feasible with a throughput of several tens of tons HM/y, because the processing equipment can handle a larger amount of actinides compared with aqueous reprocessing, owing to the difference in the nuclear criticality issue. Once

the pyrochemical process equipment has been demonstrated for several tens of tons HM/y, commercial facilities of several hundred tons HM/y can be realized by multiplexing the process equipment. Hence, the first phase of development is focused on demonstrating the feasibility of engineering — scale equipment capable of dealing with several tens of kg HM/batch (several tons HM/y), and the second phase is to increase throughput by design improvements, since an increase in throughput will directly improve the economies of processing. In this section, development of the main process equipment (described in the flowcharts in the previous section) is reported.

Within the EUROPART collaborative project, a new class of N-donor polyazines was designed and synthesized to better conform to the require­ments of the SANEX process development: the bis-triazinyl-bipyridines (BTBPs, Fig. 11.12, Foreman et al., 2005, 2006). The extraction properties of these tetradentate ligands were investigated and they appeared to differ from those of BTP ligands, in that M : L complexes were identified instead of M : L3 complexes (Drew et al., 2005, Nilsson et al., 2006a, b, Retegan et al., 2007a, b). The selectivity of alkyl-BTBP ligands toward An(III) and their kinetics of extraction are similar to those of alkyl-BTP ligands: the bulkier the alkyl groups are, the slower the mass transfer is, thus requiring a phase transfer catalyst, such as a diamide (Geist et al., 2006).

In the particular case of the bis-annulated-triazine-bipyridines, CyMe4- BTBP appeared less selective toward An(III) (SFAm/Eu > 100) than CyMe4- BTP (SFAm/Eu > 1000), certainly due to the formation of M : L2 complexes as opposed to the rigidified M : L3 complexes observed with CyMe4-BTP. In the M : L2 complexes, two tetradentate CyMe4-BTBP molecules coordinate with the extracted trivalent f cations, hence leaving an additional free access to their inner coordination spheres for an extra ligand (water molecule?). This difference in the mass action law of complex formation is somehow beneficial for the development of partitioning processes based on CyMe4- BTBP, in that the impact of the solvent hydrolytic/radiolytic degradation on the extraction performances is reduced: the apparent decrease of extraction efficiency, resulting from the destruction of the extractant, will be smaller for CyMe4-BTBP than for CyMe4-BTP (because DM <x [CyMe4-BTBP]2).

The formulation of the CyMe4-BTBP solvent was optimized, based on the iPr-BTP solvent formulation (i. e., CyMe4-BTBP and DMDOHEMA, respectively dissolved at 0.015 and 0.25 mol. L-1 in n-octanol), in order to elaborate a SANEX partitioning flowsheet (Geist et al., 2006). Although slow, the kinetics of extraction and stripping allowed a counter-current hot test to be performed in laboratory centrifuges at the ITU (Karlsruhe, Germany), implementing the highly active ‘An(III)+Ln(III)’ product coming from the TODGA hot test (see Fig. 11.15). Excellent feed decon­tamination factors for Am (7000) and Cm (1000) were obtained and the recoveries of these elements were higher than 99.9%. More than 99.9% of the lanthanides were directed to the raffinate except Gd for which 0.32% was recovered in the product (Fig. 11.16, Magnusson et al., 2009c). Nevertheless, the radiolytic stability of CyMe4-BTBP is still weaker than those of TBP, CMPO or diamides, and the possibility of recycling CyMe4- BTBP solvents has not been demonstrated yet.

11.5 Conclusions

In order to ensure the closure of future nuclear fuel cycles and minimize the long-term radiotoxicity of discarded nuclear waste, the separation of all the valuable and/or hazardous radionuclides, such as long-lived fission prod­ucts and minor actinides contained in the spent nuclear fuels will undeni­ably be necessary. The recycling of the minor actinides will probably be managed by applying a partitioning and transmutation policy. Furthermore, separating out the hazardous unrecyclable radionuclides will simplify their future route. The attractiveness of hydrometallurgical partitioning proc­esses to treat spent nuclear fuels or nuclear waste continuously, with high recovery and purification yields but only low energy inputs, largely benefits from the successes of industrial implementations such as the PUREX process.

However, the development of an efficient partitioning process is based on the design of a highly selective hydrophilic ligand or lipophilic extract­ant, that must fulfil specific requirements of solvent extraction chemistry, such as: a high affinity and selectivity toward the target element(s) to be separated, fast mass transfer kinetics, high chemical resistance, etc.

Thousands of creative ideas have emerged from radiochemists’ imagina­tions, but very few compounds have been developed up to optimized solvent formulations and tested on genuine spent fuel dissolver solutions.

Calix[4]arenes-crown-6 are pleasing examples of macrocyclic extractants, designed by functionalizing a calix[4]arene platform with a crown ether: they are perfectly suited for the selective extraction of caesium from acidic as well as basic nuclear waste. Bis-triazinyl-(bi)pyridines, although not very stable chemically, are another good example of linear nitrogen-donor ligands that are highly selective toward An(III), thanks to their particular mode of polydentate complexation. There are, however, single-step proc­esses currently in developement throughout the world that allow the sepa­ration of trivalent minor actinides directly from PUREX raffinates. They are based on the selective stripping of An(III) by hydrophilic ligands, such polyaminocarboxylic acids, in buffered solutions.

Most radioactive species in the actinide family are released in the waste streams as oxyions of the elements. The oxyionic forms are seen when the high oxidation states are present. Metals in the pentavalent and hexavalent states (V to VI, respectively) have high affinity for oxygen. The waste streams can also contain the oxyanions of the transition metals such as chromium and cobalt. The most common of these is hexavalent chromium Cr(VI) in the oxygen combined form of chromate (CrO42-) and dichromate (Cr2O72-).

15.1.2 Biological reduction of actinides

Recent studies have shown that certain species of bacteria are capable of reducing the toxic oxyions of the actinide and transition metals to less toxic and less mobile forms. For example, U(VI) can be reduced to U(IV) by sulfate reducing bacteria such as Desulfovibrio vulgaris (ATCC 29579), Desulfovibrio desulfuricans (ATCC 29577), and Geobacter metallireducens (Lovley et al., 1991; Lovley et al, 1993; Payne et al., 2004). U(VI) reduction in these species was demonstrated to be dissimilatory respiratory with the bacteria deriving metabolic energy through the U(VI) reduction pathway (Lovley et al., 1993). Notably, U(VI) reduction by the above species of bacteria required incubation of the cultures under strictly anaerobic condi­tions. This is because the Desulfovibrio species favour reducing environ­ments [ORP = -300 to -0.400 V] (Boonchayaanant et al., 2007).

Lately, U(VI) reduction by bacteria under micro-aerobic conditions has been demonstrated (Chabalala and Chirwa, 2010a). In this case, a non- purified consortium from the mine soil was used to reduce U(VI) [UO22+] to the less toxic and less mobile tertravalent state [U(IV)]. Three pure isolates were used, i. e., Pantoea sp, Pseudomonas sp. and Enterobacter sp., to reduce U(VI) under a pH range of 5 to 6 (Chabalala and Chirwa, 2010b). Figure 15.3 shows an example of the action of U(VI) reduction by the three species of bacteria while acting as pure cultures.

The significance of the above results is that the new cultures reduced uranium under conditions supporting the facultative range of bacteria. Such a culture could be less expensive to maintain. Additionally, by avoiding the formation of toxic sulfur precipitates, the culture could achieve more sus­tainable long-term operation and the separation of the product for reuse could be less expensive.

The lanthanides from La through about Er are found in measurable amounts in used nuclear fuel (plus the chemically analogous yttrium), with maximum yields (by mass) of neodymium and cerium. Yttrium behaves chemically very much like a lanthanide and is in fact found in nature in association with lanthanide minerals. These elements exist in solution and in the solid state in the trivalent oxidation state, with the minor exceptions of tetrava — lent cerium (Ce4+) and divalent europium (Eu2+). Because the valence elec­trons are placed in relatively constricted 4f valence orbitals, which shield the increasing nuclear charge across the series poorly, the cations decrease in size (by about 20%) from La3+ to Lu3+. [4] The bonding in coordination complexes appears to involve minimal covalent interactions. As a result, the coordination geometry of the complexes formed by the ions is usually dic­tated more by ligand steric constraints, by packing factors (in the solid state), and by rearrangement of the donor atoms in the ligands. In most situations, comparatively simple electrostatic models can be applied to cor­relate thermodynamic data describing lanthanide interactions with com — plexing agents or solvent molecules. This aspect of lanthanide chemistry has been discussed in detail by Choppin. [5] These cations are strongly hydrated, but comparatively weakly hydrolyzed, in part because of their compara­tively large size (cation radii average about 100 pm). The pKa values for the Ln3+ cations range from about 10 for La3+ to 7.8 for Lu3+, with the corre­sponding pH for precipitation of Ln(OH)3(s) ranging from 7.8 for La3+ to about 6.7 for Lu3+. The average primary hydration numbers for La3+ through Sm3+ are 9, for Gd3+ to Lu3+ the cations are octahydrates. Europium(III) represents the transition from hydration numbers of 9 to 8, exhibiting an average hydration number of 8.6. These metal ions form moderately stable soluble complexes with a wide variety of complexing agents. Complex sta­bility typically increases across the series, but only occasionally in a linear fashion all of the way across the series. Lanthanide ions form insoluble oxides/hydroxides, fluorides, sulfides, carbonates, and phosphates, though most of these are readily dissolved in acidic solutions.

In a once-though ion exchange process, the solution to be decontaminated is passed continuously through the ion exchange bed until the bed becomes saturated with the sorbed contaminant. At this point the ion exchange material is replaced. A good example of an operating once through ion exchange plant is the Site Ion Exchange Effluent Plant (SIXEP) at the Sellafield site in the UK. SIXEP (Fig. 3.9) has been in successful hot opera­tion since 1985 and is designed to remove radioactive cesium and strontium from water, routinely purged from the spent nuclear fuel storage basin.

The ion exchange material used is clinoptilolite, a naturally occurring zeolite that is quite specific for cesium and strontium. These species are sufficiently removed by the SIXEP process for the treated basin purge water to meet regulatory requirements for discharge to the ocean. The basin purge water is first filtered in sand filters before the ion exchange process so as to remove suspended solids and sludges that may otherwise block the ion exchange bed. The pressure vessels for both sand filter and ion exchange duties were produced to a common design to maximize operational flexibility.

The basin purge water feed is introduced to the ion exchange vessel top through a diffuser plate to prevent it channeling through the bed. The decontaminated product is collected through a number of stainless steel wedge wire strainers welded into the dished false bottom. Loaded ion exchange material is periodically discharged from the vessel by fluidizing it with water and then discharging it under pressure through the emptying pipe. A lower fluidizing ring is fitted with tangential spray nozzles used to produce a vortex action to assist emptying of the vessel contents, particu­larly the final residue. Two higher rings are fitted with radial spray nozzles to assist with breaking up the bed. It is important to discharge as much loaded ion exchange material as possible since any remaining in the column

will contaminate the product during its subsequent operation after re-filling. Development work on the first SIXEP production vessels enabled the residue remaining in the vessel to be reduced to 0.1%, which was consid­ered well within process requirements.

In nitric acid solutions, zirconium exists only as Zr(IV), albeit in many chemical forms depending primarily on acidity, Zr concentration, and tem­perature. Hydrolysis and polymerization both increase with increasing tem­perature and decreasing acidity and change the Zr(IV) speciation in solution. Increasing nitric acid concentration and/or temperature increase the extraction of Zr. The ever-present radiolytic and hydrolytic degradation products of TBP and diluent in the solvent are yet another principal reason for the extraction of Zr. Though these degradation products cannot be eliminated, their concentration can be minimized by reducing the radiation exposure to the solvent and consequently the radiolytic degradation products.

In UREX+3 process series, the NPEX process follows either CCD-PEG or FPEX. However, prior to NPEX (for recovery of a pure Pu/Np product) there is a significant feed adjustment step. Feed adjust is required to: (1) thermally destroy the reductant/complexant added in the UREX process to suppress extraction of plutonium and neptunium, (2) increase the con­centration of nitric acid, and (3) convert and maintain plutonium and nep­tunium in the extractable (IV) oxidation state.

The CCD-PEG or FPEX raffinate is fed to the NPEX process following feed adjustment as shown in Fig. 7.7 (Pereira, 2007b). The NPEX solvent composition is the same as for UREX, typical PUREX solvent. Impurities are removed from the solvent in the scrub section, and plutonium and neptunium are stripped using the same reductant/complexant that was fed to the scrub section of the UREX process. Uranyl nitrate may be added to

the strip solution to yield a product with a controllable U : Pu ratio, suitable for mixed oxide or metallic fuel fabrication.

When testing the UNEX process for recovery of the Cs+Sr and An+REE fractions, the following results for recovery of radionuclides were attained: Cs — 99.7%, Sr > 99.98%, Eu > 99.92%. These data permit the bulk HLW being treated to be transferred into the low-level waste (LLW) category, suitable for cementing and subsequent near-surface storage. Barium and lead were almost completely co-extracted with long-lived radionuclides in this process; zirconium and molybdenum were also partly co-extracted. Sodium, potassium iron and mercury remained in the low-level raffinate. In this UNEX process variant, the recovery of HLW components is shown in Table 9.14.

Results of UNEX process tests with the combined recovery of Cs, Sr, An and REE (flowsheet in Fig. 9.6)

When testing the UNEX process, prime attention was given to the work flow which included the combined recovery of Cs, Sr, An and REE into one high-level product to be solidified. It was necessary to reach a degree of

radionuclide recovery such that the UNEX-process raffinate conforms to standards established for LLW. The distribution of radionuclides obtained from tests on real HLW in the Idaho National Laboratory is presented in Table 9.15. The distribution of HLW macrocomponents over the workflow products is shown in Table 9.16.

The results of testing show that the recovery rates for Cs, Sr and An are equal to 99.4%, 99.995% and 99.96%, respectively. These recovery rates are sufficient to move the raffinate (the bulk of the wastes being treated) into the LLW category.

The combined stripping of radionuclides was carried out effectively by a solution containing 1 M guanidine carbonate and 20 g/l DTPA. The HLW macrocomponents Ba and Pb were extracted almost completely, whilst Zr, Mo, K, Ca and Fe were only partly extracted (by 87%, 32%, 28%, 10% and 8%, respectively).