In UREX, the Np and Pu are rejected into the aqueous waste stream by the use of AHA. It is then of interest to extract Np and Pu away from the other fission products and minor actinides, especially if the Pu or Pu + Np will be converted to fresh MOX fuel. To accomplish this, the aqueous raffinate from UREX is thermally treated (evaporated) to destroy AHA and increase the nitric acid concentration. The objective is to retain plutonium and neptu­nium in the extractable (IV) and (VI) oxidation states, respectively, and thereby extract them into the organic solvent of TBP in dodecane. In this process, called NPEX, the Np and Pu will behave much like in the PUREX process described earlier. After extraction, the plutonium and neptunium are removed from the solvent using AHA in a stripping stage to obtain a Np/Pu product stream. This was tested at the same time as the UREX test described above (Vandegrift 2004), see Fig. 6.5, and it was observed that the neptunium was not completely extracted due to difficulties maintaining the Np(VI) oxidation state. However, the goal of not creating a pure plutonium product was still fulfilled and the residual Np remaining in the NPEX raf­finate was recovered in the subsequent TRUEX (TRansUranic Extraction) segment of the UREX+ flowsheet (Vandegrift 2004).

The lanthanide fission products are not separated from the high-level waste in established plants and are planned to be disposed to a geologic repository in a vitrified waste form (Chapters 6 and 7). Advanced reprocessing studies are emphasizing development of processes for recovery of the trivalent actinides (Chapter 11). Most of the actinide recovery processes being con­sidered include an initial separation of the combined trivalent actinides and lanthanide fission products, followed by a process to partition the lan­thanides from the actinides. Examples include the combination of TRUEX — TALSPEAK processes (Fig. 8.3) developed in the United States and the DIAMEX-SANEX processes developed in France. Both combinations are dependent on the use of poly-amino-poly-carboxylate complex agents, together with buffering agents that permit precise control of pH in the process solutions. Following the actinide-lanthanide separation, another process is required to remove the complex agents from the product solu­tions. A group separation is obtained but with varying separation factors for the different lanthanide elements, as indicated in Table 8.5.23

If a separation of the individual lanthanide elements is needed, a multi­stage chromatographic separation process using pressurized ion exchange, strong complex agents, and precise pH control would need to be employed.24 These could be operated using either elution chromatography from an ammonium-form cationic resin or displacement cationic-exchange methods

8.3

that utilize a “barrier” metal, such as zinc or nickel. The barrier ion blocks elution of elements that form weaker complexes than the ion being eluted. Large systems of this type have been operated for rare earth recovery.

In the ‘metal waste treatment’ step, the dissolution residues obtained in the anode processing step are melted together with zirconium and NFP (recov­ered from the bottom of the electrorefiner) to synthesize a metal ingot, called a metal waste form (MWF). At the same time, the entrained salt is distilled and separated from the MWF to be recycled in the electrorefiner. Zirconium or stainless steel is added to the metal waste before melting to produce MWF ingots that have a consistent composition, phase assemblage and microstructures. The target composition of the MWF is stainless steel with 15 wt.% zirconium, because the Fe-Zr phase diagram indicates that Fe-15Zr has a relatively low solidus temperature of 1325 °C (Arias, 1993). MWF is a multiphase alloy comprising two Fe rich solid-solution phases and several FeZr2-type intermetallic phases. NFP are distributed in either one or both of the main solid solution and FeZr2-type intermetallic phases, while actinide elements are preferentially incorporated into the intermetal­lic phase. According to extensive examination and testing undertaken by ANL and INL to gain acceptance for the Yucca Mountain geological reposi­tory, MWFs are extremely robust with regard to retention of NFP in repos­itory-like conditions (Ebert, 2005).

After five years of cooling, one ton of a PWR uranium oxide spent fuel (burn-up of 60 GWd/t) contains 785 g of americium (62.4% of Am-241, 37.4% of Am-243, and 0.2% of Am-242m, all three LLRN), 135 g of curium (8% of Cm-245, 1% of Cm-243, 1% of Cm-246, all three LLRN, and 90% of Cm-244, a short-lived radionuclide), but around 60 kg of fission products, one third of which are the lanthanides.

As defined by United States of America’s regulations, transuranic (TRU) waste is without regard to source or form, waste that is contaminated with alpha-emitting transuranium radionuclides with half-lives greater than 20 years, and concentrations greater than 100 nCi/g but not including high level waste. In the US it arises mainly from weapons production, and con­sists of clothing, tools, rags, residues, debris and other such items contami­nated with small amounts of radioactive elements mostly plutonium. These elements have an atomic number greater than uranium thus are transuranic (beyond uranium). Because of the long half-lives of these elements, this waste is not disposed of as either low level or intermediate level waste. It does not have the very high radioactivity of high level waste or its high heat generation. The US currently permanently disposes of transuranic waste of military origin at the Waste Isolation Pilot Plant.

A typical example of waste volumes produced in the power generation industry is shown for the Low Enriched Uranium Once Through (LEU-OT) and Mixed-Oxide Once Through (MOX-OT) fuel processing cycles (Tables 15.1 and 15.2). In the tables, it is shown that the majority of waste in the nuclear power generation industry originates from the uranium mining and milling operations.

The majority of radioactive organic waste is produced in the enrichment and reprocessing operations. All values in the tables are reported in cubic meters per Gigawatt electricity-year (m3/GWe-yr).

3.1.1 Solid-liquid separation

A variety of solid-liquid separation systems have been deployed in the nuclear industry for a range of purposes. Most systems have had their design adapted from standard chemical industry practice so as to facilitate their operation in hot cells or gloveboxes.

Among the best examples of the implementation of IAEA safeguards to commercial nuclear facilities are those found in the Japanese industrial complexes located in the coastal villages of Tokai-mura in the Naka-gun District, Ibaraki-ken Prefecture, and Rokkasho-mura in the Kamikita-gun District, Aomori-ken Prefecture.

The Tokai and Rokkasho industrial complexes each include several of the manufacturing facilities required to support the Japanese nuclear energy industry, which is based primarily on light water reactor (LWR) technolo­gies. The trend is that Japan is to moving toward a nuclear fuel cycle program similar to that practiced by France. These industrial complexes include uranium conversion, uranium enrichment, uranium re-conversion, uranium oxide fuel fabrication, nuclear power generation, spent fuel repro­cessing, mixed oxide (MOX) fuel fabrication, interim spent fuel storage, and radioactive waste storage facilities (IAEA 2009). Our focus here is on the spent fuel reprocessing and MOX fuel fabrication facilities, which standout as closely integrated facilities in an already highly integrated industry.

The basic function of a spent fuel reprocessing plant is to effect separa­tions between uranium, plutonium, and fission products. Fission products and fuel assembly hardware ultimately report to waste, while uranium and plutonium report to separate purified process streams as nitrates. A portion of the uranium stream is converted into oxide that is returned to the process path for uranium oxide fuel production. The remaining portion of the uranium stream is mixed with the plutonium stream and the two metals are co-converted into a mixed oxide that becomes the feedstock for MOX fuel production. There are many technical variants of aqueous reprocessing and only the high-level common goals are described here. From this simple description it is easy to understand why such technologies attract interna­tional attention and require the application of IAEA safeguard principles, practices, and technologies.

In terms of Japan’s national reprocessing capability, the “pilot scale” Tokai Reprocessing Plant (TRP) is the predecessor of the “commercial scale” Rokkasho Reprocessing Plant (RRP). Construction of the TRP began in 1971 and commercial reprocessing operations continued intermit­tently from September 1977 to March 2006, at which time commercial reprocessing operations were ceased. The TRP was operated from its incep­tion until 1988 by the Power Reactor and Nuclear Fuel Development Corporation (PNC), from 1988 to 2005 by the Japan Nuclear Cycle Development Institute (JNC), and from 2005 to present by the Japan Atomic Energy Agency (JAEA). During its nearly 30 years of commercial operation, the TRP processed over 1130 MT spent fuel and achieved a maximum realized annual capacity in 1995 of approximately 95.7 MT spent fuel (Yamamura et al. 2008). The MOX product from the TRP was used to make fuels in the co-located Plutonium Fuel Fabrication Facility (PFFF) and the Plutonium Fuel Production Facility (PFPF). The TRP presently fulfills the role of a research and development facility for the JAEA. Operations include reprocessing MOX fuel from the Advanced Test Reactor Fugen.

As a result of a history of exceptional cooperation between Japan and the IAEA, the TRP served, per se, as a research and development labora­tory for many of the IAEA inspection and verification techniques used at TRP, RRP, and other international sites.

Construction of the RRP began in 1993 and commercial reprocessing operations are scheduled to begin in 2009. The RRP is operated by Japanese Nuclear Fuel Limited (JNFL). The purpose of the RRP is to continue and expand the commercial reprocessing operations formerly conducted at the TRP. The design annual capacity is 800 MT spent fuel containing approxi­mately 8 MT plutonium. The MOX product from the RRP will report to the JNFL MOX Fuel Fabrication Plant (J-MOX), which is scheduled to begin commercial operation in 2012 or later.

The latest safeguards technologies are being incorporated into the design and construction of the RRP (Johnson et al. 2001, Iwamoto et al. 2006, Durst et al. 2007). The continuous measurement and analytical technologies include accurate tank level and volume measurement (Hosoma et al. 1993) and plutonium and uranium concentration measurement via X-ray fluores­cence (XRF) and hybrid k-edge densitometry (Bean 2007). The engineered safeguards systems used to monitor, track, and verify the flow of nuclear materials through the RRP are numerous, complex, and highly integrated. Included are such systems as the Integrated Head-End Verification System (IHVS), Near Real Time Accountancy (NRTA) System, Interim Inventory Verification (IIV) System, Automatic Sampling Authenticated System (ASAS), Spent Fuel Transfer Pool Video (FTPV) System, Solution Measurement and Monitoring System (SMMS), Hull Measurement and Monitoring System (HMMS), and Vitrification Wastes Coincidence Counter (VWCC) (Johnson et al. 2001, Iwamoto et al. 2006, Durst et al. 2007, Yamamura et al. 2008).

From 2003 to 2007 a number of potential flowsheets for the processing of commercial LWR spent fuel were developed and tested with actual spent fuel at laboratory-scale. These processes are outlined in Fig. 7.1. The UREX+1 series was intended for extraction of the TRU elements as a

7.7 UREX+ flowsheets for the treatment of LWR SNF based on the GNEP identified repository benefits given in Table 7.1.

group; there is no separation among the TRUs which are to be burned as fuel in fast spectrum reactors (FR). The UREX+2 and UREX+3 process series were intended for LWR or FR reactor recycle of plutonium and neptunium. The combined product is a recycle fuel feedstock that could be manipulated and fabricated into fuel in a glovebox facility rather than a shielded cell. Because of the large relevant database on fuel performance from prior testing programs, such a fuel may be easier to qualify than the TRU fuel. The UREX+4 process series was selected as it provides the added option of burning americium in specially designed target assemblies. Separation of Cm from the Am reduces shielding requirements for target fabrication. With the isolation of Cs and Sr, the other fission products remaining after any of the UREX+ process strategies have relatively low radiation levels and heat generation rates, and can therefore be immobi­lized at high concentrations in durable ceramic or glass waste forms. The approximately 30-year half-lives of 137Cs and 90Sr isotopes imply that storage is required for 500 years or less.

Many of the head-end operations necessary for an advanced reprocessing plant using the UREX+ strategy are similar to those currently used at the commercial spent fuel reprocessing facilities at LaHague (France), Sellafield (United Kingdom), and Rokkasho (Japan). While there may be qualities unique to the fuel to be processed that require some head-end conditioning differing from that used at these facilities, the major operations and process streams will be similar. As these facilities are designed for PUREX, the major change would be adjusting the dissolver product conditions to meet the requirements of the UREX process.

Off-gases emanating from the fuel-chopper and the dissolver will contain radioactive 129I, tritium, 14C, and 85Kr. In the current reprocessing plants in Europe and Asia, these are not fully captured. If capture and sequestration of volatile radionuclides is mandated, the effectiveness of capture by various adsorbents or scrubbers must be evaluated, and methods to recover the radionuclides and sequester them in waste or storage forms must be developed.

Post-process treatment and solidification of effluents is an area where data are limited for many of the UREX+ separations. There is extensive plant-scale experience in converting uranium and plutonium to solid oxides but not for mixed actinide streams. Conversion of high-level liquid wastes to solid forms by either calcination or vitrification has been demonstrated at industrial scales but not for high activity Cs/Sr and Am/Cm solutions. Extensive data exist for acid-water evaporation, acid recycling, and waste­water treatment. However, information on reagent cleanup and recycle is less extensive. Data on recovery and calcination of trace organics in aqueous streams is likely available, though not for all of the organic species used in a UREX+ processes.

Design of the UREX+ flowsheets developed for the treatment of LWR SNF to achieve the repository benefits listed in Table 7.1 was done by com­bining solvent extraction systems in series. Figure 7.1 shows how several different solvent extraction modules were combined for several of the options. The solvent extraction process modules were tuned to meet the disposition goals of the program, also given in Table 7.1. It should be noted that isolation of any of the products listed in Table 7.2 could be omitted by recombination or exclusion of specific separation modules. In achieving these product goals, the developer is free to select any extraction process that will yield the desired products and can be made compatible with other process modules through intermediate processing.

The PUREX process is fully developed and will yield separate Pu and U streams. Recent modifications to the basic PUREX process are intended to result in incomplete separations, so that Pu is recovered with neptunium and/or uranium. These mixed PUREX separations, have been developed by both DOE and commercial companies, and are referred to generically as co-decontamination processes. Specific examples include, COEX® devel­oped by Areva, and the first separation in Energy Solutions’ NUEX process. The concept has been demonstrated at both Argonne and Oak Ridge National Laboratories with actual spent fuel.

The UREX process was uniquely developed to recover technetium, in addition to uranium. The solvent for the UREX process is the typical PUREX solvent, tributyl phosphate (TBP) dissolved in n-dodecane. Uranium and technetium are extracted from the bulk of the dissolved fuel, but co-extraction of Pu and Np is prevented by introduction of a complex — ant/reductant. Once stripped from the solvent, uranium is separated from Tc by anion exchange. Technetium was targeted because its mobility in the environment, as pertechnetate, contributed to the potential long-term dose to the public from the Yucca Mountain Repository. With UREX, it can be recovered in high yield, >95%. This level of recovery of technetium with a co-decontamination-type separation is difficult.

After the lead UREX separation, there is flexibility in the follow-on separations modules. Using the UREX+3a process, as an example, we can follow a typical separations scheme based solely on aqueous solvent extrac­tion to achieve the desired separations. UREX+3a is a cascade of five solvent separate extraction processes, referred to here as “process modules” and one ion exchange process. Following UREX and ion exchange, the process steps include: (1) recovery of Cs and Sr (CCD-PEG or FPEX),

(2) separation and recovery of a mixed U, Pu and Np product (NPEX),

(3) separation of transition metal fission products from lanthanide fission products and Am and Cm (TRUEX), and (4) separation and recovery of Am and Cm from lanthanides (TALSPEAK). Similar separations can be done for a co-decontamination front end, but in this case, the NPEX segment is omitted as the Pu and Np are recovered in the initial separation segment as shown in Fig. 7.1 (2004).

There is some flexibility in both the sequencing and selecting the process for Cs and Sr recovery. Cs and Sr can be recovered immediately after UREX using either the FPEX or the CCD-PEG process with appropriate feed adjustment. The CCD-PEG solvent is a mixture of chlorinated cobalt dicarbollide (CCD) and polyethylene glycol (PEG); the diluent is phenyl — trifluoromethyl sulfone (FS-13). An alternative process is FPEX, which uses a solvent containing a calixarene, a crown ether, and a modifier, all in Isopar-L® (a refined kerosene) diluent. The FPEX diluent is more compat­ible with the other processes, which are based on n-dodecane, but the CCD-PEG process is more mature. In all of the schemes shown in Fig. 7.1, the raffinate from the UREX segment is the feed to the CCD-PEG or FPEX process segment. However, these processes can be run to recover Cs and Sr from the TRUEX raffinate rather than the UREX raffinate. Reducing the amount of processing prior to TRU recovery should result in a purer product, though shielding requirements are higher. This would remove one process from the main sequence, which may be beneficial for actinide product purity for recycle in reactor fuel.

The NPEX process separates plutonium and neptunium from the other components in the FPEX or CCD-PEG raffinate. Like UREX, the solvent for NPEX is the typical PUREX solvent. Although the process was designed to extract Pu and Np, uranium is added to the extracted Pu and Np in the strip section to yield a U-Pu-Np product with a controllable elemental ratio.

The TRUEX process separates the lanthanide and transuranic actinides from the other components in the NPEX raffinate. The solvent is a mixture of CMPO (octyl(phenyl)-N, N — diisobutylcarbamoylmethyl phosphine oxide) and TBP in n-dodecane. If TRUEX follows NPEX, Am and Cm are recovered with the lanthanides. If NPEX is omitted, Pu and Np are also contained in the “grouped TRU” product.

The actinide:lanthanide separation is achieved using TALSPEAK. TALSPEAK solvent is HDEHP (bis(2-ethylhexyl)phosphoric acid dis­solved in n-dodecane while the aqueous phase is a buffered lactate solution containing DTPA. Extraction of the lanthanides over actinides is regulated by control of pH and operating conditions. The product contains Am and Cm for UREX+2, 3 and 4 series, and all of the transuranics in the case of the UREX+1 series. If a suitable separation is devised, Am can be further separated from Cm, yielding UREX+4 in Table 7.1. At the moment there are no proven candidates for the separation of Am from Cm at an industrial scale.

These separation module sequences (except for Am from Cm separation) are based on proven or at least lab-tested processes that can be run on prototypical equipment. Other aqueous processing concepts have been proposed, but are not currently suitable for bench-scale process tests. As it is attractive to reduce the number of processing steps required to achieve the separations, much current focus is on finding extractants that are highly elective for actinides. One approach under development which has shown promise is a TRUEX-TALSPEAK hybrid, christened TRUSPEAK, which separates actinides from all fission products in a single process module rather than two sequential modules.

A UNEX-extractant (0.08 M CCD, 0.02 M Ph2Bu2 and 0.5% PEG-400 in FS-13) has a density of 1.4 g/cm3 and viscosity of 4 mPa*s at 20°C. Stability of all components of the extraction system is rather high for treatment of

HNO3 solutions of radioactive waste. So, CCD in FS-13 on contact with concentrated HNO3 is stable even at elevated temperatures (>100°C). Diphenyldibutyl-carbamoylphosphineoxide and PEG-400 are also stable in contact with HNO3. The UNEX-extractant and diluent FS -13 do not enter into exothermal reactions with HNO3 even at temperatures above 100°C.

Fire safety of the UNEX-extractant depends primarily on the properties of the diluent FS-13, which is the main component of the mixture. The high flash temperature of FS-13 (93°C) is responsible for the fire-proof nature of the UNEX-extractant; for comparison, it should be noted that the flash tem­perature of dodecane (the diluent used in the PUREX-process) is 70°C. Other components of the UNEX-system (CCD, CMPO, PEG) have a high boiling point and are low fire-hazard compounds. As in the two-phase aque­ous-organic extraction process, a temperature higher than 100°C is practi­cally never reached, and the UNEX-process should be considered fire proof.

Specialist studies have revealed the high radiation resistance of FS-13. The total yield of radiation decomposition for pure sulfone and for sulfone in contact with HNO3 was 4.5-5.0 molecules/100 eV. The low yield of fluo­rine-ions in this (0.15 ions/100 eV) makes the UNEX-extractant relatively corrosion-proof when compared to stainless steel. Radiolysis products do not exert any marked effect on the extraction and hydrodynamic properties of the UNEX-extractant. No interphase films, precipitates and emulsions were detected on exposure.

Since the extractant is used over multiple cycles in the process, an impor­tant characteristic is the prolonged stability of its properties, which may depend on differences in the solubility of individual components in the aqueous phase. Polyethylene glycol is the most soluble component during the stripping operation (up to 250 mg/l). Since PEG concentration in the extractant is critical for Sr extraction, this component should be replenished during the course of the process. With the exception of compensation for PEG losses (for example, by its introduction into the stripping solution), the properties of the UNEX-extractant do not practically change over time. This was confirmed by prolonged testing of the UNEX-process at RI and Idaho National Laboratory.

To process the metal waste forms from the pyrochemical treatment of irra­diated EBR-II fuel, an engineering-scale furnace has been designed (see Fig. 10.22) to separate the adhering salt from the metal waste (consisting of steel cladding, zirconium and NFP) by distillation, and then to melt the metal waste in a crucible assembly while recovering the salt vapour in the condenser located above the crucible assembly (Marsden, 2005). The furnace has a vacuum vessel of approximately 1.2 m diameter and 2 m height, together with convective cooling fins to cool the condenser region and a heat shield to limit thermal flux to the hot-cell window. Induction heating can be operated at 1700 °C and 200 mTorr using a 25 mm square copper tube for the induction coil. After demonstrating the successful pro­duction of surrogate metal waste in a glovebox, a furnace of nearly identical design was installed in the INL hot-cell, as shown in Fig. 10.23, and process testing is ongoing (Goff, 2009).