7.2.1 Removal and separation of volatile and semi-volatile components

Voloxidation process

A dry head-end process for oxidation of used fuel has been developed and is being considered for application in advanced reprocessing to remove volatile components and allow selective trapping from the off-gas, prior to performing other separations processes for removal of non-volatile fission products. Among several benefits, the dry head-end treatment removes volatile tritium prior to any aqueous separations that may be subsequently used and, thus, prevents accumulation of tritiated water in the aqueous systems. In addition, there are significant potential benefits for removal and capture of radioiodine.

In the voloxidation process, the oxidation reaction is carried out at tem­peratures above 400°C in an oxygen-containing atmosphere. The UO2 in the fuel reacts with oxygen to form U3O8, and the reaction is accompanied by a volumetric expansion of about 30%. The expansion breaks the ceramic oxide fuel structure and generates a fine powder form of U3O8, which is a significantly less dense material than the ceramic UO2. Powder size distribu­tion depends on the temperature of the oxidation process3,4 and other factors related to the used fuel characteristics, but nearly all oxidized par­ticles are produced in sizes less than 20 pm in diameter, with a good fraction less than 10 pm.5 Higher temperatures increase the rate of oxidation but increase the grain size of the U3O8 powder produced, with a wider distribu­tion of particle sizes.

A drawback of the traditional UNEX process was the use of guanidine carbonate solution + DTPA bearing a large quantity of guanidine carbonate (0.5-1 mole/L) for stripping. Clearly, all of the guanidine carbonate remained in the strip product. Its destruction prior to vitrification is possi­ble, but this is a cumbersome process, especially taking into account the high radioactivity of the strip product.

Alkylamine nitrates have also been proposed for stripping of metals from CCD-based extractants [33]. The effect of methylamine, dimethylamine and trimethylamine nitrates on the stripping of cesium and barium from CCD + PEG solutions was considered. Solutions of these amine nitrates stripped metals from the organic phase efficiently but, they should also be destructed in the resultant strip product before vitrification. All the earlier proposed complexones were found to be inadequate for stripping metals from a UNEX-extractant, with the main difficulties concerning the stripping of cesium.

However, when a solution of methylamine carbonate + DTPA or methyl- amine carbonate + nitrilotriacetic acid (NTA) was proposed instead of guanidine carbonate + DTPA for stripping in the UNEX process [34, 35], all the metals extracted proved to be efficiently stripped. The essential dif­ference in the proposed solution is that it is a regenerable strip agent. Methylamine carbonate is easily regenerated and the strip product pre­pared from regenerated methylamine carbonate exhibits features identical to those of the initial compound. It has therefore been possible to sharply reduce the number of organic compounds in the strip product, easing the problem of its subsequent solidification. The possibility of using methyl — amine carbonate was confirmed in the course of two dynamic experiments on simulated solutions performed concurrently at Idaho National Laboratory and RI. The basic flowsheet for HLW treatment by a regenerable strip agent is presented in Fig. 9.10.

Emphasis should now be placed on reducing HLW volume during the process (~20 times). Radiolysis of methylamine carbonate solutions with

9.11

Tetrabutyldiamide of dipicolinic acid (TBDPA).

complexones was investigated. The results obtained bear witness to the high radiation stability of the complexones and methylamine carbonate by itself, which allows these solutions to be used for the treatment of high-level waste [36]. The radiolysis products at an absorbed dose to 3 MGy do not affect the extraction and stripping processes.

From the early 1990s to the beginning of the 2000s, various extracting systems have been investigated in the framework of European collabora­tive projects (Dozol et al., 1997, Madic and Hudson, 1998, Dozol et al., 2000a, Madic et al., 2000, Dozol et al., 2004, Madic et al., 2004). This thriv­ing exploratory period ended in France with a demonstration phase, between 2000 and 2005, during which the research activities focused on the most mature separation processes, tested on genuine PUREX raffi­nates to assess their scientific feasibility (validation of the partitioning concept) and technical feasibility (global validation of all the separation processes). These tests aimed at consolidating the process flowsheets in order to ensure their robustness by assessing their long-term implementa­tion in laboratory equipment simulating industrial contactors, and by developing efficient spent solvent cleaning treatments to cope with the degradation of the solvents through acidic hydrolysis and radiolysis, as if they were recycled in industrial processes. The remainder of this chapter will describe some of the research programmes deployed in Europe during the past two decades to develop and optimize the design of highly selec­tive ligands for partitioning schemes, in order to limit the long-term nox­iousness of nuclear waste and/or to close future nuclear fuel cycles. Note that this chapter will not review thoroughly the various ligands studied, but focus on three chosen examples: (i) calix[4]arenes for caesium separa­tion, (ii) diamides for the co-extraction of trivalent minor actinides (An(III)) and lanthanides (Ln(III)), and (iii) nitrogen-donor ligands for An(III)/Ln(III) separation.

In the PUREX process, UO2 is first dissolved in a nitric acid solution (3-6 M) to form uranyl ions followed by TBP extraction of the uranyl ions using an organic solvent such as dodecane. Nitric acid oxidizes uranium in

UO2 from the +4 oxidation state to +6 oxidation state in the form of (UO2)2+ which is subsequently extracted into the organic phase with TBP as UO2(NO3)2(TBP)2. Therefore, nitric acid serves two purposes in this process:

(1) it acts as an oxidizing agent converting UO2 to uranyl (UO2)2+ ions and

(2) it provides nitrate ions to neutralize the charges carried by uranyl ions leading to the formation of extractable UO2(NO3)2(TBP)2. In the early sc-CO2 experiments for extracting uranyl from nitric acid solutions it was noted that HNO3 can be carried into the fluid phase in the presence of TBP (Wai et al. 1999c). This is true also in the PUREX process (Chaiko and Vandegrift 1988). After the extraction, the aqueous phase remained sepa­rated from the sc-CO2 phase but the concentration of HNO3 in the aqueous phase was significantly reduced. The possibility of using TBP as a carrier for dissolving HNO3 in sc-CO2 was investigated by several groups including one group at Nagoya University in Japan (Tomioka et al. 2001) and another at the University of Idaho (Samsonov et al. 2001). According to the Lewis acid-base complex formation principle, the P=O group in TBP is an electron donor (Lewis base) which can be bound to protons of Lewis acids such as HNO3 or H2O through hydrogen bonding. When TBP is mixed with con­centrated nitric acid (15.5 M) by rigorous shaking, HNO3 can dissolve in the TBP phase forming a complex of the general form TBP(HNO3)x(H2O) y which is immiscible with water. The values of x and y in the complex can vary depending on the relative amounts of TBP and the concentrated nitric acid used in the preparation of the complex. Experimentally, the values of x and y in the complex can be determined by acid-base titration and by Karl Fischer titration, respectively. As shown in Table 14.1, the complex has the composition TBP(HNO3)1.8(H2O)0.6 if it is prepared by mixing equal volumes of TBP and concentrated nitric acid (15.5 M). If the ratio of TBP to the acid is 5:1 (e. g. 5 mL TBP to 1 mL of the concentrated nitric acid) in the preparation, the TBP phase has the composition TBP(HNO3)0.7(H2O)0.7. Nuclear magnetic resonance (NMR) studies indicate that the protons of HNO3 and H2O in the complex probably undergo rapid exchange resulting in one single peak which shifts upfield with increasing x/y ratio (Enokida

et al. 2003). When the TBP-HNO3 complex is added to a low dielectric constant solvent such as chloroform, the solution becomes cloudy due to formation of very fine droplets of nitric acid. This is attributed to an anti­solvent effect because nitric acid has a very low solubility in chloroform. The same phenomenon occurs when the TBP(HNO3)18(H2O)0.6 complex is added to sc-CO2. This anti-solvent effect provides a mechanism of dispers­ing find droplets of nitric acid in the supercritical fluid phase for dissolution of uranium oxides. The concentration of HNO3 in the small water droplets is probably very high, significantly higher than the 3-6 M nitric acid used in the PUREX process. The small acid droplets perhaps can exist in the sc-CO2 phase like reverse micelles surrounded by TBP. Water-in-sc-CO2 microemulsions (reverse micelles) have been extensively studied using CO2 soluble surfactants (Ji et al. 1998a, Wai 2002). The microemulsions are dynamic in nature and they are able to extract uranium from soil and dis­solve the uranyl ions in the water core of the CO2 microemulsion (Campbell et al. 2001).

The TBP-nitric acid complex is soluble in sc-CO2 and capable of dissolv­ing uranium oxides in contact with the fluid phase. Figure 14.5 shows the amounts of UO2 and UO3 solids dissolved in sc-CO2 containing the TBP(HNO3)o.7(H2O)o.7 complex with respect to time in a continuous flow system (Samsonov et al. 2001). The sc-CO2 solution containing the TBP — nitric acid complex was prepared by bubbling liquid CO2 through a cell containing the complex placed upstream of the dissolution cell. As shown in Fig. 14.5, UO3 is easier to dissolve than UO2 in the sc-CO2 containing the

14.5 Direct dissolution of UO2 and UO3 solids in sc-CO2 with TBP(HNO3)0i7(H2O)0i7. From Samsonov et al. 2001. Reproduced by permission of The Royal Society of Chemistry.

extractant because UO3 is already in the +6 oxidation state. The dissolution of UO3 in sc-CO2 with the TBP-nitric acid complex may be expressed by the following equation:

UO3(solid) + 2 TBP HNO3 ^ UO2(NO3)2(TBP)2 + H2O 14.3

For the dissolution of UO2 in sc-CO2, the reaction may be expressed by Equation (14.4).

3 UO2(solid) + 8 TBP HNO3 ^ 3 UO2(NO3)2(TBP)2

+ 2 TBP + 2 NO + 4 H2O 14.4

If TBP(HNO3)18(H2O)06 is used as the extractant, the rate of dissolution of UO2 in sc-CO2 is more rapid because of a higher HNO3 concentration provided by the complex. After the dissolution, no aqueous phase is formed in the extraction cell because small amounts of water formed during the dissolution process can be carried out of the system by sc-CO2 and TBP. The alkali metals, the alkaline earth metals, and many transition metals cannot be extracted by the TBP-HNO3 complex in sc-CO2.

One advantage of the sc-CO2 dissolution method compared with the conventional PUREX process is that the former process combines dissolu­tion and extraction in one step with no aqueous waste and organic solvent involved. In the PUREX process two steps are involved in the dissolution and extraction of spent fuel, i. e. UO2 is first dissolves in a nitric acid solu­tion followed by extraction of uranyl ions from the acid solution into an organic solvent with TBP under ambient pressure. The possibility of selec­tive dissolution of lanthanides and actinides in the sc-CO2 process has not been explored but is conceivable based on the tunable solvation property of the fluid. Furthermore, the sc-CO2 dissolution process is carried out in a closed system, therefore volatile fission products can be concentrated in the sc-CO2 stream and removed on-line with a suitable sorbent with minimal release into ambient atmosphere. Better control of volatile fission products may be another advantage of the sc-CO2 process for reprocessing spent nuclear fuel. In the conventional PUREX process, volatile fission products including iodine-129 is released in the off-gas after dissolution of the spent fuel.

A conceptual illustration of a dry sc-CO2 dissolution process for repro­cessing spent nuclear fuel is given in Fig. 14.6 (Wai 2002). Using a TBP — HNO3 complex, lanthanides, uranium and transuranic elements would be dissolved in the sc-CO2 phase leaving fission products such as Sr and Cs behind in the residue. After isolation of UO2(NO3)2(TBP)2 from the sc-CO2 phase, the un-used TBP-HNO3 and CO2 can be recycled. Figure 14.6 also includes a hypothetical ligand regeneration step in which UO2(NO3)2(TBP)2 would be converted to UO2 with the regeneration of TBP-HNO3. The pro­posed sc-CO2 reprocessing process presents an idea which requires exten-

sive research to turn it into a reality. Nevertheless, it points to a future direction for reprocessing spent nuclear fuel in an environmentally sustain­able way (Wai 2006b).

Extractions of fission products such as 90Sr and 137Cs by sc-CO2 have also been evaluated using selective ligands such as crown ethers (Wai et al. 1999a, 1999b). Table 14.2 shows the extraction of Sr2+ from water by sc-CO2 with dicyclohexano-18-crown-6 (DC18C6), a CO2 soluble ligand which is known to be selective for Sr2+ in conventional solvent extraction processes. However, the Sr2+-DC18C6 complex is not soluble in sc-CO2. One method of making the Sr2+-DC18C6 soluble in sc-CO2 is to use a fluorinated counter anion which would neutralize the charge of the complex and make the resulting ion-pair soluble in the fluid phase. Fluorinated counter anions such as ammonium or potassium salts of pentadecafluoro-n-octonoic acid and perfluoro-1-octane sulfonic acid are effective for extracting Sr2+ and Cs+
using crown ether ligands in sc-CO2. Selective extraction of Sr2+ over Ca2+ and Mg2+ in water by sc-CO2 containing DC18C6 and pentadecafluoro-n — octonoic acid CF3(CF2)6COOH can be achieved as shown by the results given in Table 14.2 (Wai et al. 1999a). Selective extraction of Cs+ by sc-CO2 with dicyclohexano-21-crown-7 and perfluoro-1-octanesulfonate has also been demonstrated (Wai et al. 1999b). These studies may provide valuable information for designing ligands for selective extraction of fission products in sc-CO2 that may be critical for future development of sc-CO2-based processes for reprocessing spent fuel.

Radiolysis in aqueous solutions has an entirely different effect than in solid materials; water molecules become either excited or ionized. The ionization event occurs on the time scale of an electronic transition (<10-16 s) and the positive ion H2O+ is formed along with an electron; H2O+ reacts with another molecule of water (<10-14 s), forming an •OH radical and H3O+. The elec­tron, if liberated with sufficient kinetic energy, ionizes further water mole­cules until its energy falls below the ionization threshold of water (12.61 eV. Spending the rest of its energy on vibrational and rotational excitation of the water molecules, it becomes solvated (<10-12 s) (Rydberg et al., 2001).

The excited states of irradiated molecules of water dissociate to form radicals •O, H-, •OH and molecular H2; this occurs on the same time scale as a molecular vibration, within 10-14-10-13 s. The physical and physico­chemical (pre-thermal) processes are thus completed within 10-12 s, leaving the species in thermal equilibrium with the water (Rydberg et al., 2001).

The radiolysis products, formed upon excitation and ionization of irradi­ated water molecules, are clustered in “spurs”; i. e. they are inhomoge­neously distributed in the water and proceed to diffuse out of the spur volume. During this “spur diffusion” process, recombination reactions (equation 2.7) take place, leading to the formation of molecular or second­ary radical products.

Alpha particles deposit their kinetic energy in a very short distance (38 ^m in water for 5.3 MeV alphas; de Carvalho, 1952). Because of their higher charge, generally greater kinetic energy and greater mass, their linear energy transfer (LET) is much larger than the LET of photons or high energy electrons. As is shown in Table 2.7, the G-values (radiation product

yields) for the radical products are larger for the radiation with low LET whereas the yields of molecular products (H2, H2O2) are larger for the high LET radiation. Hart estimated that about 88% of the radicals during alpha — radiolysis recombine to give molecular products, the proportion of free radicals in the product being only 12% (Hart, 1954).

In irradiated dilute aqueous solutions, practically all the energy absorbed is deposited in the water molecules and the observed post-irradiation chemical changes are the result of the reactions between the solutes and the products of the water radiolysis. With increasing solute concentrations, the direct radiolysis of the solute gradually can become more important. At higher concentrations, the solute may also be modified through direct inter­action with radiation in the spur.

Self-radiolysis may greatly affect the chemical equilibrium and speciation of actinides in their solutions. For example, in acidic solution of 239Pu, the self-radiolysis results in changes of the oxidation state of PuO22+ which “degrades” by being reduced to Pu4+ at a rate of approximately 1.5% per day (Keogh, 2005). Intermediate Pu(V) is unstable in acidic solutions; it disproportionates to produce Pu4+ and Pu(VI). Equation 2.8 shows the radiolytic reductions in solutions of 239Pu along with the reproportionation and disproportionation reactions.

Pu4+ ~~~> Pu3+

PuO22+—— > PuO2+

Pu4+ + PuO2+ о Pu3+ + PuO22+ .

2PuO2+ + 4H+ о Pu4+ + PuO22+ + 2H2O

Hydrogen peroxide, produced by higher concentrations of 239Pu in a suf­ficient amount, may decrease the concentration of tetravalent plutonium by forming a precipitate of plutonium peroxide (Rydberg et al., 2001). To sta­bilize the concentration of Pu(IV) in processing, nitrite ions are added to the solution to scavenge the hydroxyl radicals formed by the radiolysis and thus eliminate the production of H2O2.

The effect of radiation on actinide containing materials and solutions can be altered by careful selection of different isotopes with varying half-lives
that can be used. In the case of the plutonium example given above, by using the 242Pu isotope, which has a half-life 15.5 times longer than 239Pu, solutions that show diminished radiation effects (for example, increased stability in a given oxidation state) can be created. In contrast, 238Pu, which has an 87.7 year half-life and specific activity about 300 times greater than 239Pu, is well suited for alpha-radiolysis studies.

The radiolytic reduction of Np(VI) to Np(V) by alpha self-radiation (of 237Np) results with a relatively large radiation yield (G = 0.66 mol/microjoul), and does not vary with perchloric acid concentration within the limits investigated (3.1 x 10-9sec-1 in 0.5-1.7 N HClO4) (Cohen, Taylor, 1961; Zielen et al., 1958; Siddall, 1960). This high reduction yield is attributed to the high concentration of Np in irradiated solutions: the investigations of Pu and Am were never performed with concentrations higher than 10-2M (Vladimirova, 1964). Gamma irradiation, typically studied with Co-60 source (with у energy of 1.17 and 1.33 MeV) leads to either reduction of Np(VI) to Np(V) or oxidation of Np(IV) to Np(V) (which has a comparatively high stability to radiolysis) (Burney, Harbour,1974). Studies of beta-radiolysis using accel­erated electron beams show an increasing yield of Np(V) in solutions of low acidities (0.01-0.7 M HClO4); with increasing concentration of acid, the radiation yield of Np(V) drops (Burney, Harbour, 1974).

The most reactive water radiolysis species produced are the oxidizing hydroxyl radical (•OH), hydrogen peroxide (H2O2), reducing aqueous elec­trons (eaq-), and hydrogen atoms (H^), which are produced in equal amounts. The hydrated electron e-eq is a strongly reducing species (E0 = -2.9 V) whereas the hydrogen atom is a less powerful reductant (E0 = -2.3 V) (Rydberg et al., 2001). The hydroxyl radical •OH is a strong oxidant (E0 = 2.7 V in acidic and 1.8 V in basic solution) (Rydberg et al., 2001). Under the acidic, aerated conditions of the solvent extraction process, the aqueous electrons produced would be scavenged according to the fast reactions 2.9 and 2.10 (Mozumder, 1999; Mincher et al., 2009).

Process solutions of used nuclear fuel are quite concentrated. As noted above, with increasing solute concentrations, direct radiolysis of solutes can occur. Nitrate salts and nitric acid predominate in the process solutions; hence, the radiation chemistry of nitric acid in both the water and organic solvent media is very important. HNO3 is distributed in both phases and its radiolytic degradation products have a huge influence on the separation processes.

The most important radiation product in irradiated solutions of nitric acid is nitrous acid (nitrite ion), since it directly affects the redox chemistry

of Np and Pu. At lower acid concentrations, the nitrite yields from alpha irradiation are much lower than the yields from the gamma or beta irradia­tion but steadily increase with increasing HNO3 concentration (Table 2.8, Kazanijan, 1970). The nitrite yields in HNO3 below 1 M exhibit the same trend as has been observed in neutral nitrate solutions since water is absorb­ing most of the radiant energy and the nitrite yields are determined by the reaction of the radicals formed from the radiolysis of water via the follow­ing mechanism:

H -+NO3-^ NO2 + OH — 2.11

2NO2 + H2O ^ NO2- + NO3 — + 2H+ 2.12

OH + NO2-^ NO2 + OH — 2.13

The nitrite yields from the solutions with HNO3 > 1 M are different from those in neutral or low nitrate solutions. Increased concentration of nitrate leads to these reactions:

NO3- ~~~> e-aq + — NO3 2.14

NO3- ~~~> NO2- + O 2.15

NO3- ~~~> NO2 + O + e-aq 2.16

The effect of the solution acidity, expressed by lower nitrite yields in acidic than in neutral solutions is caused by radiolytic decomposition of the undissociated nitric acid molecule (as well as the nitrate ion). Redlich et al (1968) established that the undissociated acid fraction grows from 1% in 1 M to 48% in 10 M nitric acid; hence, the probability of the reaction in Equation 2.14 increases with increased concentration of nitric acid:

HNO3 ^ NO2 + OH 2.17

The different trends for the high (alpha) and low (beta, gamma) linear energy transfer (LET) radiations can be explained by different yields of ionic and molecular radicals upon interaction with aqueous solutions. Radical combination to produce molecular products would be higher for alpha irradiation. The dimer (N2O4) or the H2O2 produced would have no effect on the net yield of nitrite.

Nitrous acid HNO2 is a key redox controlling factor, affecting the specia — tion of neptunium and plutonium in the used nuclear fuel solutions. One of the most significant reactions in this complex system is the reduction of Np(VI) to unextractable Np(V) by nitrous acid; it is reversible, controlled by H+-concentration. Nitrous acid also acts as a catalyst of oxidation of Np(V) by nitric acid back to Np(VI):

2NpO22+ + HNO2 + H2O о 2NpO2 + + 2H+ + HNO3 2.18

Two banks of multiple counter-current 2-cm centrifugal contactors were instrumented with vis-NIR and Raman spectroscopy probes in order to demonstrate the ability to measure solution components in real-time within a solvent extraction system. Typically the centrifugal contactors operated at 3600 rpm, with an aqueous phase feed rate of 12 ml/min, and an organic phase (30% TBP/n-docecane) feed rate of 18 ml/min. The effluent contain­ing the raffinate stream from the extraction bank was passed through a commercial vis-NIR flow cell (2.5 cm path length, Custom Sensors Inc), followed by a custom fabricated flow-through Raman cell containing a commercial 180° backscatter Raman probe (Inphotonics Inc). A bank
consisting of four contactors was located in a non-radiological fume hood for “cold” testing, while a similar bank consisting of 16 contactors was located within a radiologically shielded glovebox allowing for demonstra­tion with feed solutions containing uranium and neptunium.

The joint conversion of uranium and plutonium to the oxide form makes it possible to do away with the complicated step of blending and grinding the two distinct oxide powders, as currently employed for the purposes of MOX fuel fabrication. Over the past few years, a number of co-conversion routes have thus been revisited or devised on the basis of original chemical principles, in order to innovate in the area of uranium and plutonium management. Drawing on 50 years industrial experience and feedback, CEA/DEN, in partnership with Areva NC, are proposing to extend oxalic conversion of Pu(IV) to Pu oxide to the oxalic co-conversion of uranium and plutonium.

Of the various possible variants, the current reference route is the U(IV)- Pu(III) variant, owing to the following advantages (Grandjean 2005):

• Unexpectedly, U(IV) and Pu(III) co-crystallize into a single oxalate structure, thus ensuring homogeneous distribution of the two actinides, at the molecular scale (Arab-Chapelt 2007). This structure is an oxalate solid solution, over a broad range of Pu/(U + Pu) ratios: 0-50% (mol/ mol).

• The mixed oxalate exhibits very low solubility in a nitric acid solution carrying excess oxalic acid: solubility is about the same as that of the plutonium(IV) oxalate precipitate.

• Calcination of this mixed oxalate in an inert atmosphere yields a solid solution of (U, Pu)O2 oxide; the homogeneity of uranium and plutonium distribution, at the molecular level, is thus conserved, from the solution to the oxide.

• Technology similar to that operated in the current plutonium conversion workshops at La Hague may be used.

A major advantage of oxalic co-conversion is to produce mixed oxides exhibiting physicochemical characteristics directly tailored to fuel fabrica­tion, affording, in particular, the ability to use this raw material directly, with no further specific chemical, or mechanical treatment. More broadly speaking, from PUREX to COEXTM, the evolution to an oxalic U-Pu co­conversion process, combined with a process adapted from a MIMAS-type process, affords the following advantages:

• A process yielding a uranium-plutonium mixture, without at any point involving the handling of separated plutonium, thus reducing prolifera­tion risks.

• The (U, Pu)O2 oxide solid solution obtained from co-conversion, as the endproduct of spent fuel treatment by the COEXTM process, and raw material for the fabrication of advanced MOX fuel, results in an enhanced homogeneity of plutonium distribution within the fuel (com­pared with the current upstream grinding operations with mixtures of UO2 and PuO2 powders). This translates to the anticipation of improved in-reactor behavior for MOX fuel fabricated from such raw material, bringing benefits with regard to increased burnups.

• Reduced contamination potential and limited exposure to radiation, during fuel fabrication using COEXTM-produced powder.

• A further outcome would be simplified management of fabrication scrap, and a likely improvement in MOX fuel solubility, subsequent to irradiation (no PuO2 islands).

6.3 Conclusions

PUREX was developed in the US in the late 1940s originally as a separation system for production of pure plutonium for military applications. In a way, PUREX was an industrial revolution and the success of this process led to the development of modern liquid-liquid extraction processes. More than half a decade after its development, PUREX is still the foundation of all modern separation processes for spent nuclear fuel used on industrial scale. The system utilizes tributyl phosphate (TBP) to extract uranium and plu­tonium away from most other elements in the spent nuclear material. The preference of tetravalent and hexavalent (as MO22+) ions is the key to the selective extraction. The stability of UO22+ combined with the relative ease of changes to the oxidation state of plutonium, predominantly between the +3 and +4 state, is the other critical aspect of this process. The current state of the art PUREX processes for commercial reprocessing are used in four countries in the world (France, Japan, the UK, and Russia) and have an annual capacity of more than 5500 tonne heavy metal (including the new

Rokkasho-Mura plant). The reprocessing plants combine extraction (TBP) and redox reagents (e. g. U(IV), hydroxylamine, hydrazine) in several counter current liquid-liquid extraction batteries under different conditions to extract, scrub and strip selected elements to obtain a pure product. PUREX is not entirely without problems and elements such as Np, Tc, Zr, Nb and Ru put high demands on the process such as increased number of stages and the need for parallel decontamination batteries. Improvements and changes made to the PUREX process over the years include more advanced equipment to facilitate easier handling and increase throughput. Although PUREX has been used successfully for a long time, research has been continuous and a number of recent discoveries and ideas are close to being implemented to a future PUREX process, e. g. Advanced PUREX, SuperPUREX, COEX, UREX, etc. These new processes are being devel­oped to meet new or future policies regarding waste and product stream composition. One major issue has been to avoid a pure Pu product by routing either neptunium (UREX and NPEX) or uranium (COEX) into the Pu-product stream, thereby increasing proliferation resistance. The complicated issue of neptunium control have been addressed using simple hydroxamic acids and successful pilot scale processes has been adminis­tered in both U. S. (UREX+), the UK (Adv. PUREX) and France (COEX). All these processes look promising for the future successful use of PUREX — like chemistry in advanced nuclear fuel reprocessing and the adaptability of this process is all but proven to accommodate for future restrictions on waste and product stream composition.

Fission-product separations technology will continue to see progress in the future to meet the needs of advanced reprocessing and waste management in different countries. Addition of dry head-end pyrochemical treatment, such as the voloxidation process, can enable selective separation and capture of the volatile and semi-volatile fission products, while converting ceramic used fuels into finely divided oxide powder that can be separated relatively easily from the fuel cladding and can simplify the dissolution process. Separation and recovery of selected fission products, such as xenon gas, volatile noble metals, and lanthanide elements for reuse may become eco­nomical, thus reducing the waste discharges from used fuel reprocessing. In the advanced aqueous reprocessing arena, a dedicated unit operation for Cs and Sr separation is of uncertain net benefit, which will undoubtedly dampen the pace of applied development in the near term. However, the desirability of reducing the heat load on geologic repositories will likely keep interest alive for improved separations in the long term.

In the long run, there will remain a significant need for improved fission — product separations associated with the nuclear industry. Although declared needs from the industry are often erratic, it will remain true that science — driven advances in technology can always have a transformational impact. This must be true in that existing and developing technologies are not even close to ideal, even as impressive as developments have been in the past two decades. In view of these realities, available technologies will change as well as the way in which technologies are invented and developed. While there can be no substitute for creative inspiration, for example, increasingly the power of computational design of new separation agents that function by molecular recognition can be expected to lead to targeted properties. Already, the first inkling of such capability has occurred, as evidenced by the com­puter-aided design of self-assembled capsules for selective crystallization of sulfate.104 Expansion of applicable techniques beyond solvent extraction, ion exchange, and precipitation can be expected. Electrochemical methods, supercritical fluid extraction, and membrane technologies, for example, are likely to play a role. Finally, the need for, and challenge of, combined separa­tions cannot be overestimated. The ability to target specific combinations of radionuclides for separation simultaneously without adjustment of feed conditions, use of special hold-back reagents, complicated scrubbing, and concentrated stripping reagents will ultimately prove compelling. In summary, the need for advanced fission-product separation technologies to move toward more ideal characteristics will remain strong in the context of new fuel cycles and continued waste cleanup, and in response, both new technologies and the way new technologies are conceived may be expected.

The pyroprocess can be applied to spent oxide fuels if they can be converted into a metal form. Figure 10.8 shows a flowchart of the pyrox process for

10.8

Pyrochemical process flowsheet for oxide fuel treatment.

oxide fuel treatment which employs voloxidation and electroreduction steps at the front-end. ‘Voloxidation’ is a process to facilitate the removal of spent uranium oxide fuel from its cladding. Sections of clad spent nuclear oxide fuels are subjected to high temperature and an oxygen-bearing atmo­sphere to convert UO2 to U3O8 with a 30% volume expansion; the fuel matrix is consequently pulverized and dislodged from its cladding. During the voloxidation, partial removal of volatile fission products such as Cs, Te, Ru, Mo and Tc is expected (Westphal, 2008).

The chemical stability of the BTP ligands was improved by branching the alkyl groups grafted onto the lateral triazines. 2,6-bis-(5,6-di-iso-propyl — 1,2,4-triazine-3-yl)-pyridine (iPr-BTP) successfully passed a once-through hot test performed at the CEA Marcoule (France) in 2001 (Fig. 11.14, Madic et al., 2002). Here, more than 99.9% of Am(III) and more than 99.8% of Cm(III) were recovered from a genuine DIAMEX An(III)-product solu-

tion, with less than 2.5 wt.% contamination of Ln(III), by implementing in laboratory scale centrifugal contactors a solvent consisting of iPr-BTP dis­solved at 0.01 mol. L-1 in n-octanol, with 0.5 mol. L-1 of DMDOHEMA added to increase the mass transfer kinetics. However, the solvent failed when recycled without treatment: iPr-BTP degraded because of alpha/ gamma radiolysis and its extraction performances decreased by 40% after two recycles (Hill et al., 2002). This detrimental observation definitively dumped the BTP ligands as potential extractants for the SANEX process development in France.

Attempts to fully substitute the labile a-benzylic hydrogens of the BTP lateral triazines, by synthesizing, for instance, the tertio-butyl derivate, have all failed so far. However, Hudson et al (2006) managed to synthesize bis — annulated-triazine-pyridines (BATP, Fig. 11.12), with methyl groups in place of a-benzylic hydrogens. As a result, these compounds present much higher chemical stabilities than alkyl-BTP molecules. Because they possess more carbon atoms in their structures than alkyl-BTP molecules, BATP ligands are also assumed to be more lipophilic than alkyl-BTP compounds, and hence to show higher extraction properties, which is actually the case for 2,6-bis(5,5,8,8-tetramethyl-5,6,7,8-tetrahydrobenzo[1,2,4]triazin-3-yl) pyri­dine (CyMe4-BTP). Amazingly, the selectivity observed for CyMe4-BTP is also ten times higher than that of г’Рг-BTP in the same experimental condi­tions (SFAm/Eu > 1500 at [HNO3] = 1 mol. L-1), and the highest ever reported for a nitrogen-donor ligand. This is probably due to the tris-complex forma­tion, in which the extracted metallic cation is completely dehydrated and the three bulky CyMe4-BTP molecules become so rigid that the complex formed is energetically extremely stable.

In fact, the example of CyMe4-BTP demonstrates that radiochemists’ search for ever more efficient and selective extractants could well become detrimental to process development: the drawback of CyMe4-BTP is the almost irreversible extraction of An(III) that makes this compound inap­plicable to counter-current test implementation.