In France and in Europe, the bidentate oxygen-donor solvation extractants that were investigated in the 1990s to co-extract An(III) and Ln(III) were the N, N,N’,N’-tetraalkyl-malonamides (Gasparini and Grossi, 1980, 1986, Musikas, 1987, 1988, Thiollet and Musikas, 1989, Cuillerdier et al., 1991,1993, Nigond et al., 1994a, b, 1995, Madic and Hudson, 1998): RR’N(C=O)-CHR"- (C=O)NRR’, a subgroup of the diamide family bearing a methylene bridge linking the two amide functions (R, R’, and R" representing linear or branched alkyl or phenyl substituents). Malonamides enable the formation of energetically stabilized 6-membered ring complexes when they extract trivalent metallic cations.

Malonamides were designed to compete with carbamoyl-phosphine oxide compounds, such as CMPO (Fig. 11.8), developed in the 1980s at the Argonne National Laboratory (ANL, USA) to decontaminate transuranic waste (Navratil and Thomson, 1979, Horwitz et al., 1981, 1982) arising from the production of military grade plutonium. Although of lower efficiency than CMPO when extracting An(III) from nitric acid solutions, lipophilic

malonamides (containing a suitable number of carbon atoms) present several advantages over CMPO:

• Malonamides are soluble in hydrogenated tetrapropene: concentrations exceeding 1 mol. L-1 can be dissolved in HTP. Therefore, malonamides do not require any phase modifier to cope with the relatively high solvent loading capacities required by PUREX raffinates.

• As malonamides present steeper nitric acid dependence than CMPO, the stripping of extracted elements is facilitated.

• Malonamides are less stable than equivalent neutral organophosphorus compounds, especially versus acidic hydrolysis, but their carboxylic deg­radation products are less detrimental to the back-extraction of the minor actinides in diluted nitric acid.

• As they are made of carbon, hydrogen, oxygen, and nitrogen atoms, malonamides generate only mineral ashes after incineration.

The affinity of malonamides toward An(III) and Ln(III) decreases as the atomic number of the extracted element increases. The stoichiometry of the extracted complexed M(III) cation is assumed to be ML2(NO3)3 (where L = malonamide) at saturation, although higher stoichiometries have been observed owing to malonamide aggregation. The use of small angle neutron/ X-ray scattering techniques and the application of colloidal concepts to malonamide solvents actually proved that these compounds self-organize in small aggregates (Fig. 11.9), consisting of spherical polar cores, mainly

composed of the polar heads of 4 to 10 malonamide molecules and of the extracted solutes (0.5 < фсот<, < 1.2 nm), surrounded by non polar crowns, mainly composed of the alkyl chains of the malonamides and of the diluent (Abecassis et al., 2003, Bauduin et al., 2007, Dozol and Berthon, 2007).

It must be remembered, though, that aggregation can easily become detrimental to process implementation, because the polar core attractions can induce the splitting of the loaded solvent into two phases: an enriched one in metallic complexes and another, lighter (called ‘third phase’), usually composed of almost pure diluent. The stability of an organic phase contain­ing aggregates depends on the equilibrium between the different interac­tions: (i) the attractive van der Waals forces taking place between the aggregate polar cores, (ii) the repulsive forces between the polar cores (assimilated as hard, sticky balls in the Baxter model), and (iii) the repulsive forces between the aggregates, due to the stabilizing repulsive steric interac­tions between the hydrophobic alkyl chains of the extractant and those of the diluent.

Due to the amphiphilic nature of the malonamides, the corresponding solvents behave like reverse micro-emulsions stabilized by surface active compounds. The attractive forces depend on the composition of the polar cores: they increase with the concentration of the extracted solute and depend on its nature. The aggregate repulsive forces depend on the length of the alkyl chains of both the malonamide extractant and the diluent mol­ecules, which act in opposite manner: long alkyl chains for the malonamide and short alkyl chains for the diluent prevent third phase formation, which will occur as soon as the attractive forces between the aggregates exceed their repulsive forces (Berthon et al., 2007).

Development of the DIAMEX process

The structure of the DIAMEX process reference extractant has evolved between 1991 and 2001, from N, N’-DiMethyl-N, N’-DiButyl-TetraDecyl — MalonAmide (DMDBTDMA, Fig. 11.8, Musikas et al., 1991) to N, N’- DiMethyl-N, N’-Octyl-HexylEthoxy-MalonAmide (DMDOHEMA, Fig. 11.8, Madic et al., 2002), in order to:

• increase the total number of carbon atoms, to enhance both the hydro — phobicity of the extractant and the solubility of the extracted metallic complexes in HTP, and hence prevent third phase formation;

• improve the extractant affinity toward trivalent metallic cations, by introducing an ethoxy moiety in the alkyl chain grafted onto the meth­ylene bridge (Spjuth et al., 2000);

• uniformly display the carbon atoms among the alkyl chains grafted onto the malonamide to simplify the elimination of its degradation com­pounds coming from acidic hydrolysis and radiolysis (Berthon et al., 2001): basic washings of the degraded solvent have proved to be efficient in getting rid of the hydrophilic acidic degradation compounds.

The DMDOHEMA flowsheet was directly adapted from that of DMDBTDMA thanks to the PAREX process simulator code, developed at the CEA. Several counter-current tests have been carried out from 1999 to 2005, both at the CEA Marcoule (France), at Forschungszentrum Julich (FZJ, Germany), and at the Institute for Transuranium Elements (ITU, Karlsruhe, Germany) during successive collaborative projects funded by EURATOM (Courson et al., 2000, Madic et al., 2000, 2002, 2004, Christiansen et al., 2004, Warin, 2007). These tests put into operation: [12]

pulsed columns (two for the extraction of An(III) and Ln(III), and one for FP scrubbing), eight mixer-settlers (for the stripping of An(III) and Ln(III)), and four centrifuges (for the spent solvent clean-up) is shown in Fig. 11.10 (Warin, 2007). This test, as well as a long-term hydrolysis/radiolysis endur­ance test, in which the DMDOHEMA solvent was recycled after specific caustic washings to eliminate its acidic degradation compounds (such as carboxylic acids and acid-amide), validated the industrial applicability of the DIAMEX process by demonstrating the possible recovery of more than 99.9% of An(III) and Ln(III) from a genuine highly active PUREX raffi­nate, with high decontamination factors, DF, toward fission products (e. g., DFZr ~ 800).