Challenges of the selective extraction of caesium

After five years of cooling, one tonne of a spent PWR uranium oxide fuel (with a burnup of 60 GWd/t) still contains 4.6 kg of caesium (1.89 kg of stable isotope Cs-133, 50 g of two-year half-life Cs-134, 769 g of 2.3 million- year half-life Cs-135, and 1.93 kg of 30-year half-life Cs-137), which together with strontium is the main short-term heat source in the vitrified waste. In addition to the need for selective caesium extraction from dissolver liquors to decrease this thermal source, selective extraction of caesium from acidic effluents containing large quantities of sodium nitrate arising from the neutralization of nuclear plant technological waste streams by concentrated soda, as well as from basic sludge contaminated with alpha emitters inherited from decades of military research programs, is of primary impor­tance in nuclear waste management.

In aqueous nitric acid solutions, the caesium metallic cation is monova­lent and considered as a ‘hard acid’ in Pearson’s theory. It therefore inter­acts preferentially with ‘hard bases’, such as anions or molecules containing fluoride or oxygen donor atoms and inducing electrostatic interactions. Studies reported in the literature point out the difficulties of extracting caesium selectively from acidic solutions, especially in the presence of other alkali elements (Moyer and Su, 1997, Herbst et al., 2002a).

The first extracting agents envisaged in the 1980s were the crown ethers: macrocyclic polyethers discovered by Pedersen (1967). Parametric studies carried out on the extraction of alkali cations by crown ethers (Danesi et al., 1975, Sadakane et al., 1975, Gerow et al., 1981, Schulz and Bray, 1987, Wood et al., 1995, Dozol et al., 1995, Dietz et al., 1996, Kumar et al., 1998, Kikuchi and Sakamoto, 2000) have shown that the quantity of metallic cation extracted in the organic phase depends on the following factors: the polarity of the organic diluent used to dissolve the crown ether, the nature of the co-extracted anion, the size of the crown ether cavity relative to the alkali cation diameter, the chemical composition of the aqueous solution (acidity, ionic strength), and the type of the chemical functions grafted onto the crown ether. The most evolved design of a crown ether for caesium extraction is that of di-(tert-butyl-benzo)-21-crown-7 (Fig. 11.1) because the cavity size of its crown, containing seven oxygen atoms, matches the coor­dination shell of the caesium cation, and because the tert-butyl-benzo groups grafted onto its skeleton enhance its hydrophobicity.

Although di-(tert-butyl-benzo)-21-crown-7 can extract caesium from nitrate feeds of low acidity, it definitely requires the addition of a synergistic cation exchanger, such as dinonyl-naphthalene sulfonic acid (Kozlowski et al., 2002, 2007), or even better hexabrominated bis(dicarbollide) anion (Gruner et al., 2002), a very strong cation exchanger used in the UNEX process (Herbst et al., 2002a, b, 2003), to extract caesium from PUREX acidic raffinates. Nevertheless, the drawback of synergistic mixtures com­posed of solvating agents (such as crown ethers, which extract metallic cations with increasing efficiency as the ionic strength increases) and cation exchangers (such as cobalt dicarbollides, which extract metallic cations more efficiently at low acidity) is the difficulty encountered when stripping the metallic cations, because of competition between the two extractants. The same problem occurs for crown ethers that have been functionalized with a cobalta bis(dicarbollide) cage (Fig. 11.1) to extract caesium from acidic feeds (Gruner et al., 2002). Their caesium extraction efficiency is comparable with that of the corresponding synergistic mixtures composed of hexabrominated bis(dicarbollide) anion and crown ethers. Their selectiv-

ity with respect to Na+ cation is sometimes even higher (separation factor: SFcs/Na[10] > 100 for trace elements) than that of the corresponding synergistic mixtures, yet too low in the presence of high contents of sodium ions to develop a process for decontaminating technological nuclear waste streams.

When dissolved in organofluorine diluents such as 1,1,7-trihydrodo — decafluoroheptanol, di-(tert-butyl-benzo)-21-crown-7 and other crown ethers appear to extract caesium from acidic feeds ([HNO3] > 3 mol. L-1, Yakshin et al., 2008), but the complexity of their preparation and conse­quently their high cost, as well as their poor Cs+/Na+ selectivity, still hinder their use on an industrial scale.