Carbon dioxide is widely used for supercritical fluid extraction studies because of its moderate critical constants (Pc = 73 atm and Tc = 31 °C), chemical inertness, and low cost (Phelps et al. 1996). Above the critical point, CO2 becomes a fluid that has both gas-like and liquid-like properties as illustrated in Fig. 14.1. Supercritical fluid CO2 has mechanical properties

14.1 Phase diagram of CO2 and some properties of sc-CO2.

like a gas and yet has solvation strength like a liquid. Therefore, it is capable of penetrating into small pores of solid materials and dissolving organic compounds from the solid matrix. After extraction, the fluid phase can be vented as CO2 gas by reduction of pressure causing precipitation of the extracted solutes. In principle, no liquid waste is produced by this extraction technique according to the idealized operation.

One important factor which determines the efficiency of supercritical fluid extraction is the solubility of the target compound in the fluid phase (Darr and Poliakoff, 1999). Because CO2 is a linear molecule with no dipole moment, sc-CO2 is a good solvent for dissolving non-polar organic com­pounds, but is ineffective for dissolving highly polar compounds or ions. Metal ions are not soluble in sc-CO2. Searching for methods of dissolving metal ions in sc-CO2 was the focus of the initial research started in the author’s laboratory at the University of Idaho two decades ago. In 1991, Laintz et al. (1991) noticed that complexes formed by transition metal ions with a fluorinated dithiocarbamate chelating agent bis(trifluoroethyl) dithiocarbamate (FDDC) exhibited unusually high solubilities (by 2-3 orders of magnitude) relative to their non-fluorinated analogs. Based on this information, the authors demonstrated that copper ions (Cu2+) in aqueous solutions can be effectively extracted into sc-CO2 with the aid of FDDC (Laintz et al. 1992). The idea of using fluorinated chelating agents for dissolution of metal ions in sc-CO2 was actually inspired by the fact that perfluorinated alkanes were considered at that time as blood substitutes because of their high solubilities for oxygen and carbon dioxide. The reverse is apparently true for dissolution of fluorinated compounds in sc-CO2.

Today, fluorine-containing compounds are generally referred to as “CO2- philic” because of their high solubilities in sc-CO2. The successful demon­stration of Cu2+ extraction immediately led to the investigation of extracting lanthanides and uranium in sc-CO2 using fluorinated ligands because of its potential relevance to nuclear waste management. Several reports appeared after 1992 demonstrating that trivalent lanthanide ions and uranyl ions dis­solved in water or spiked in solid materials could be effectively extracted by sc-CO2 using fluorinated P-diketonates as extractants (Lin et al. 1993, Lin and Wai, 1994, Lin et al. 1994). Among a number of fluorinated P-diketonates tested for sc-CO2 extractions, thenoyltrifluoroacetone (Htta) was more often used than the others because it is a solid at room tempera­ture and is easier to handle experimentally (Wai and Wang 1997). After these initial reports, dissolution of uranium oxides directly in sc-CO2 with fluorinated P-diketones was investigated (Wai and Waller, 2000). Direct dissolution of solid UO3 in sc-CO2 with Htta occurs according to the fol­lowing equation:

UOs(s) + 2 Htta ^ UO2(tta)2-H2O 14.1

The dissolution of UO3 proceeds with a much higher efficiency if tri-n- butylphosphate (TBP) is present with Htta in sc-CO2. The enhance effi­ciency is attributed to the fact that TBP is a stronger Lewis base which can replace water in UO2(tta)2-H2O to form a more soluble adduct complex UO2(tta)2-TBP as shown in equation (14.2).

UO3(s) + 2 Htta + TBP ^ UO2(tta)2-TBP + H2O 14.2

The solubility of UO2(tta)2-TBP in sc-CO2 is reasonably high to be poten­tially useful for reprocessing consideration. For example, at a fixed tempera­ture of 40 °C its solubility in sc-CO2 is 7.5 x 10-3 mol/L at 200 atm and increases to about 1.5 x 10-2 mol/L at 300 atm. Reaction (14.2) occurs effi­ciently for UO3 because uranium is in the +6 oxidation state, which leads to the formation of the uranyl-tta-TBP adduct complex with no need of an oxidation step. However, this reaction does not occur with UO2 because uranium is in the +4 oxidation state, which does not form a stable complex with tta and TBP. A CO2-soluble oxidizing agent such as H2O2 has been shown to promote the dissolution of UO2 in sc-CO2 with Htta and TBP but the efficiency is limited (Trofimov et al. 2001). Research in extraction of actinides in sc-CO2 using fluorinated P-diketones and organophosphorus reagents may still find applications in nuclear waste treatments. For example, a report in 2003 showed that plutonium and americium in soil could be effectively removed by sc-CO2 augmented with Htta and TBP (Fox and Mincher 2003). The solubilities of uranium, plutonium, neptunium, and americium P-diketonates and their adducts with organophosphorus reagents in sc-CO2 were also reported (Murzin et al. 2002).

A significant development in supercritical fluid extraction of uranium was made in 1995 by Lin et al. (1995) who reported that uranyl ions in nitric acid solutions could be extracted into sc-CO2 with the aid of TBP, a well — known ligand for uranium extraction in the conventional PUREX (Plutonium Uranium Extraction) process. TBP happens to be highly soluble in sc-CO2. An earlier report showed that about 10% by mole of TBP could be dissolved in sc-CO2 under normal extraction conditions (Page et al. 1993). Actually, TBP becomes miscible with sc-CO2 above a certain pressure at a given temperature according to a later report (Joung et al. 1999). The efficiency of extracting uranium by the sc-CO2/TBP process depends on the nitric acid concentration and follows closely the same trend as the tradi­tional solvent extraction process using dodecane and TBP as shown in Fig. 14.2 (Lin et al. 1995). The extracted uranium species in sc-CO2 was identified as UO2(NO3)2(TBP)2 similar to the uranyl complex extracted from nitric

14.3 Solubility of UO2(NO3)2(TBP)2 in sc-CO2. From Carrott et al. 1998. Reproduced by permission of The Royal Society of Chemistry.

acid into dodecane with TBP (Carrott et al. 1998). To evaluate the solubility of the uranyl complex in sc-CO2, a high-pressure fiber-optic cell was designed to measure the absorption spectrum of the complex in situ in the fluid phase (Carrott and Wai 1998). The solubility of UO2(NO3)2(TBP)2 in sc-CO2 mea­sured by the spectroscopic method is surprisingly high. At 40 °C and 200 atm, the solubility of the uranyl complex in the fluid phase reaches 0.45 moles per liter, a concentration comparable to that used in the conventional PUREX process (Wai 2001). When the solubility of the uranyl complex in sc-CO2 is plotted against the density of the fluid phase, a linear relationship is observed as shown in Fig. 14.3. The results given in Fig. 14.3 demonstrate the solvation strength of the supercritical fluid is tunable with respect to density. Consequently, the solubility of UO2(NO3)2(TBP)2 in sc-CO2 can be predicted based on the relationship log S = a log ф + b, where S is the solu­bility in g/L, ф is the density also in g/L, and a and b are constants. The tunable solvation property is unique for sc-CO2 and suggests a possibility of selective dissolution or separation of metal complexes in sc-CO2 by manipulation of density (i. e., temperature and pressure) of the fluid phase. Conventional liquid solvents are basically not compressible and the solva­tion strength cannot be altered at a given temperature by varying pressure except perhaps at extremely high pressures. Other studies of uranium extraction using sc-CO2 and TBP were reported by Meguro et al. (1998a, 1998b). In another study, Iso et al. (2000) showed that the distribution of Pu(IV) between 3 M nitric acid and sc-CO2 containing TBP depends on pressure at a given temperature (60 °C) and follows the equation log D = a log ф + b, where D is the distribution coefficient, ф is the density of the

fluid and a and b are constants. The D values as well as the constants a and b for U(VI) in the same system are different from those of Pu(IV) suggest­ing some degree of separation of Pu(IV) and U(VI) can be achieved by adjusting the pressure of the sc-CO2 extraction process. These studies provide a basis for using sc-CO2 to replace the organic solvents used in the PUREX process.

Based on the information available in 1998, Smart et al. proposed a supercritical fluid-PUREX process for reprocessing spent nuclear fuel as illustrated in Fig. 14.4 (Smart et al. 1998). In this process, sc-CO2 is employed to replace the organic solvent conventionally used in the PUREX process. The advantages of the supercritical fluid-PUREX process include conceiv­ably higher mass transport properties, faster extraction rates, and tunable distribution coefficients of uranium and other actinides. However, since nitric acid is used to dissolve spent fuel in the supercritical fluid-PUREX process, it would still produce acidic liquid waste like the PUREX process. After the proposed application by Smart et al. (1998), research in finding a dry process for dissolution of uranium dioxide in sc-CO2 intensified (Wai 2006a).