The two models presented below have been used for physical-chemical treatment of pollutants. The continuous trench allows open access to water in multidirectional flow (Fig. 15.7a). In this system, the user is not interested

in collecting the product of treatment for further processing. The alternative — funnel-and-gate system — offers the user the option of collecting flow from a localized position for further processing (Fig. 15.7b).

The flow in the continuous trench system, which is perpendicular to groundwater flow direction, needs to be slightly larger than the cross sec­tional area of the contaminated groundwater in order to capture the con­taminants in both vertical and horizontal directions (Gavaskar et al., 2000). The funnel-and-gate system is composed of impermeable walls and at least one reactive zone. The funnel structure could be sheet piles or slurry walls where the function of the funnel is to intercept the contaminated ground­water and lead it to the treatment zone. Phillips (2009) has elaborated on the designs of different reactive barriers, including but not limited to: the thickness of the PRB to provide sufficient residence time for the contami­nants within the treatment zone to be completely treated.

Other complex designs have been tried including the multi-sequenced permeable reactive barriers (MS-PRBs) for multiple contaminants. MS-PRBs use multiple reactive materials in more than one reactive zone as shown in Fig. 15.7c (Dries et al., 2004).

For the purpose of treating nuclear and radioactive waste around waste storage sites, the cheapest option could be the inoculated barrier system. All reactive barriers in the context of heavy metal treatment, unless heavily engineered and expensively constructed (which defeats the purpose), are mainly — if not only — suitable as temporary containment barriers to be operated until the pollutant source is completely removed or a more per­manent solution for continuous remediation is found.

Most effective separations are based on the phase transfer of the species of interest away from a diverse mixture. Distillation and precipitation are two examples of practical applications of phase transfer in separations. In solvent extraction, partitioning of the species of interest between two immiscible liquids is employed to accomplish the desired separation. Commonly used in organic chemistry, the difference in solubility of the selected species between the immiscible fluids forms the basis for the separation (Nernst partitioning). For metal ion separations, more complex interactions are typically required. The specific features of the application of both solvent extraction and pyrometallurgy (molten salts/molten metals) to the processing of used nuclear fuel have been discussed in detail elsewhere. [10]

Owing to the substantial differences in the polarity of the aqueous and organic solutions, there is, in general, minimal tendency for metal ions or their electroneutral salts to spontaneously partition into nonpolar media (like kerosene), particularly for polyvalent metal ions, which tend to have substantial hydration energies. For the transfer of metal ions or their polar salts into non-polar medium to occur, the presence of amphiphilic (featur­ing polar and non-polar regions, like surfactants) molecules is necessary. Such molecules can (and often do) aggregate in the bulk organic solutions to improve their compatibility with the typically low dielectric constant medium.

These same molecules tend, at the same time, to arrange themselves at the polar-non-polar fluid interface with the polar end penetrating signifi­cantly into the aqueous solution while the lipophilic portion of the molecule remains firmly in the organic side of the interface. If the interaction of the polar end of the extractant molecule with the cation is strong enough, it promotes dehydration of the metal cation, resulting in a substantial increase in entropy for the biphasic system, which can drive the phase transfer process. The extractant molecule can either simultaneously transfer a more weakly hydrated cation to the organic phase or facilitate the re-solvation of the anion needed to accompany the cation into the organic solution. Typically, once this interfacial transfer reaction is completed, the nascent metal complex completes the transfer to the bulk by completing the dehy­dration through selective solvation (or chelation) by lipophilic molecules present in the organic phase.

Two primary procedures for cation partitioning are possible: cation exchange or solvation of metal salts. The latter process can also exhibit considerable ability to support the partitioning of mineral acids into the organic phase. Under most circumstances in the processing of dissolved used fuel solutions, acidic conditions are maintained (to minimize the possibility for hydrolysis and precipitation of metal hydroxides). In some systems, the application of salting out agents has been required for efficient phase transfer. Under these conditions, the primary cation exchange reaction will involve H+ exchange. A general example of the cation exchange reaction is:

M3+aq + 3 HLorg = ML3org + 3 H+aq

The alternative reaction of partitioning an electroneutral salt could be defined as,

M3+aq + 3 X-aq + 2 Yorg = MX3Y2org

where X is the supporting anion in the aqueous phase and Y is the solvating extractant. A third process related to the ion solvation method arises in extraction systems based on the use of lipophilic tertiary amine or quater­nary ammonium compounds, in which an anionic metal solvate coordina­tion compound is extracted,

M3+aq + 4 X-aq + HY+org = MXYHY+org

The energetic details and kinetics of these systems can be quite different and selectivity can arise from a variety of different interactions.

Selectivity of solvent extraction processes can be altered through the introduction of water-soluble chelating agents. Complex equilibria can control these processes and multiple interactions between species in each phase are possible. Selectivity can arise from the nature of the extracting agent, the anions co-extracted by solvating or anion exchange reagents, the

chelating agents present in the aqueous phase or by changing the oxidation state of the extracted species relative to that of competing matrix ions. In many cases, classical solution chemistry provides useful guidance to the prediction of process efficiency, though supramolecular organization of solute or solvent molecules and the mutual miscibility of aqueous and organic phases can further alter system performance.

Solid-liquid separations based on organic or inorganic ion exchange materials can also be (and have been) employed as an alternative or com­plement to solvent extraction. In some instances, anion exchange has also proven an acceptable mode for conducting separations of nuclear materials. An advantage that accrues from the application of chromatographic methods like ion exchange is the large numbers of theoretical plates that can be achieved in such separations when conducted in column mode. If solvent extraction reagents are immobilized on solid support materials (extraction chromatography), the selectivity advantages and flexibility of solvent extraction can be augmented with multiplicity of re-equilibrations of chromatography. These solid-liquid separation methods have one par­ticular limitation in the context of operating closed-loop fuel cycles in their batch-wise mode of operation, which differs from the continuous (or nearly so) operation that is possible in solvent extraction. Devices that employ easily replaceable ion exchange cartridges have been developed, though their application in a processing canyon (where remote operation and main­tenance is required) might offer unique challenges in operations. In prin­ciple, solvent impregnated membranes could also be operated in a semi-continuous mode, though research to date on such methods applied to metal ion separations processes has revealed issues with the stability for membranes and the kinetics of mass transfer.

Pulsed perforated plate columns are differential type liquid-liquid extrac­tion contactors because they manifest continuous, rather than step, concen­tration profiles. These devices have been used for recycling used nuclear fuel since the Hanford PUREX plant was started up in 1955. All current commercial UNF recycling plants (THORP, UP3 and RRP) employing advanced PUREX separations technology use pulsed columns in three functions:

• Separation of uranium and plutonium from UNF.

• Separation of plutonium from uranium.

• Purification of the plutonium product.

A schematic of a typical industrial pulsed column used for processing used nuclear fuel is illustrated in Fig. 3.13. THORP pulsed columns are cylindrical and are filled with a series of horizontal perforated plates (Fig. 3.14), spaced at 2-inch intervals, and held together by a series of vertical tie-rods (Phillips, 1993a). In contrast, the La Hague plants use annular pulsed columns with inner and outer columns arranged concentrically. The plates, which in this case are rings, are spaced similarly to the ones in the cylindrical columns but are not perforated and alternately attached to the external and internal column walls. Some smaller cylindrical columns are also used and these are filled with alternating rings and discs, with the rings attached to the outer column wall and the disks carried on a central vertical tie-rod (Drain, 1997).

In all column contactors the more dense aqueous phase enters the column near the top while the less dense solvent phase enters through a pulse limb close to the bottom. A liquid-liquid dispersion is created as the phases flow counter-current through the plate perforations under the action of the periodic pulse applied to the solvent. This can be arranged to be a disper­sion of aqueous drops falling through a continuous solvent layer, or it can be a dispersion of solvent drops rising through a continuous aqueous phase. In the former case, the phase separation interface is formed and held in the “bottom settler” at the base of the column while in the latter case this interface is in the top settler. Any solids in the aqueous phase feed will typically congregate at the liquid-liquid interface and be directed with the dispersed phase. Therefore, when extracting uranium and plutonium into solvent from an aqueous UNF feed, the aqueous phase is dispersed so that any solids will be directed to the bottom settler where the aqueous drops coalesce at the solvent/aqueous interface. Thus, contamination of the solvent containing the separated uranium and plutonium is avoided. The top settler provides sufficient freeboard for the solvent to exit the column free of aqueous phase entrainment.

From compressed air reservoir

Pneumacators for density measurement

Top settler

There are two stable hydrodynamic operating regimes for pulsed columns. The mixer-settler operating regime manifests at low pulse frequencies and/ or amplitudes and flow rates when dispersion and coalescence entirely occurs between plates. At higher pulse frequencies and/or amplitudes and flow rates the mixer-settler regime transitions to the dispersive or emulsion regime where a continuous dispersion is established throughout the col­umn’s length. Drop sizes in the dispersion regime are typically 1-3 mm in a nitric acid — 30% tributyl phosphate systems with uranium and plutonium (Phillips, 1993b). At still higher pulse frequencies and/or amplitudes and flow rates, the dispersion becomes unstable, leading to gross wrong-phase entrainment in the products, a condition known as flooding. Pulsed columns are usually operated in the dispersive regime because the higher dispersed phase hold-up and higher interfacial area maximizes mass transfer effi­ciency and minimizes wrong-phase entrainment. However, they also provide good mass transfer performance when operating in the mixer-settler regime because the lower interfacial area is at least partially mitigated by the con­tinuous coalescence and re-dispersion of the drops, which continually exposes a fresh interface for mass transfer.

Detection and control of the position of the bulk interface in the top settler for a continuous aqueous phase, or bottom settler for a continuous organic phase, is critical for stable pulsed column operation and hence good mass transfer performance. Interface detection and control is achieved in the THORP pulsed columns using equipment with no in-cell moving parts so that no in-cell maintenance is required through the life of the facility. Interface detection is achieved using air bubbler tubes (“pneumercators”) to measure the relative densities of the two liquid phases above and below the interface and thus infer the interface position between the bubbler tubes (Phillips, 1993a). In solvent continuous columns, with the interface in the bottom settler, the pneumercator pressure signals are interfered with

by the large pulsing pressure applied. Rapid response pressure transducers and sophisticated signal processing are used to extract a reliable interface position signal from the pulse pressure wave “noise”. Interface position control is accomplished using an air lift system (see below) to control the rate of aqueous phase removal from the bottom settler. The airlift is coupled with a fluidic pump to provide a constant submergence for the air lift air input. When such a constant submergence is provided, liquid flows from airlifts are stable and increase reproducibly with increasing air flow, making them excellent control devices.

Some important chemical engineering attributes of pulsed columns are provided in Table 3.3.

The outgoing aqueous scrub phase, containing a significant fraction (but not necessarily all) of the TcO4-, then undergoes a “complementary” extraction with fresh organic phase to recover plutonium and uranium inadvertently back-extracted with the Tc. The solvent from this complementary extraction operation is recycled to the main extraction section to maintain a high concentration of U and Pu in the outgoing solvent of the codecontamina­tion step and to avoid undesired losses of uranium and plutonium to process wastes. Obviously, the aqueous product from the Tc scrubbing operation cannot be re-combined directly with the previous scrub or feed streams since the higher concentrations of HNO3 would exacerbate the extraction of Zr in the main extraction section and negate the benefits afforded by the dual scrub approach. Consequently, the aqueous raffinate from the comple­mentary scrub operation is sent directly to the waste in combination with the raffinate from the main extraction (Baron 1993, 2003).

In passing, it should be noted that the aqueous streams from the main extraction and Tc scrubbing and complementary extraction operations are subjected to a diluent scrub. The purpose being to wash traces of TBP, lost from the organic via solubility, to the aqueous phase. Such diluent washing has become a standard feature of the PUREX process for aqueous streams that are concentrated in evaporators, and is a standard safety feature in modern reprocessing facilities. The diluent used in these washing steps is recycled to the process to minimize TBP losses. Diluent washing is noted at multiple points in Fig. 6.2 and will not be further discussed.

The diversity of the impurities targeted in the scrub sections of codecon­tamination often involves a compromise, that, when reduced to practice, leads to the use of several different scrub solutions. Other important design requirements of the first cycle include minimization of U and Pu losses to the liquid waste emanating from the cycle, thereby restricting these losses to very low levels (typically <0.1%). Finally, the rather high concentrations of U, typically approaching 300 g/L in the incoming aqueous feed, allows favorable flow ratios of organic and aqueous to be achieved, as the organic can be heavily loaded with uranium. The flowrate of the organic solvent is slightly in excess of that required for the quantity of U and Pu to be extracted; however, a large excess of organic must be avoided in order to restrict excessive extraction of impurities.

In general, there is a high degree of separation of actinides and fission products after the codecontamination step. Typical decontamination factors reported from UP-3 at La Hague are Ru DF ~104, Cs DF > 107, and a Tc DF > 3 was initially targeted (Baron 1993), but is now typically DFTc~10 or higher (Baron 2003).

The process was demonstrated in 2005 and 2006 and yielded high recovery of all products. Excellent TRU/Ln recovery was obtained with TRUEX, and a refined TALSPEAK process for the 2006 test met target specifications. The CCD-PEG separations module was not run as part of the 2005 dem­onstration. Results are given in Table 7.3.

UREX+2

The process was demonstrated in 2004 and yielded high recovery of all products. All primary components reported to their desired product streams.

The Tc and Np fractions found in the streams in which they were targeted were lower than the target values. Zirconium reported to the Np/Pu product, rather than the raffinate, due to other-phase carryover that was ascribed to a mismatch between the contactor design and loaded-solvent density. Results are given in Table 7.4.

Test results on the development and elaboration of the UNEX process have shown that the treatment of 1 volume of feed HLW generates 0.5 volume of strip product containing recovered radionuclides, guanidine carbonate, DTPA and 4-5 g/L of non-radioactive metals. The well-known borosilicate — glass production technique was tested on this product. The main non-vola­tile components of the strip product are sodium (20%), potassium (40%) and iron (10%). Taking these into account, the composition of the flux for glass was determined: it contained silicon dioxide, boric acid, sodium nitrite and zinc oxide.

The glass production process included the following operations:

• drying of the strip product;

• calcination of residue at 700°C;

• preparation of the flux and its drying at 450°C;

• mixing of the calcinated residue with the flux at 1100-1200°C.

The borosilicate glass samples produced were homogeneous, did not contain any non-molten inclusions and did not crack in storage. 1 m3 of strip product produces 15-20 kg (~10 L) of borosilicate glass.

Because operations of pyroprocess are carried out in a batch mode, solid ingots are the main products needed to be transported between different process equipment. To attain higher efficiency, techniques for transporting other forms, such as powders and high-temperature liquids, are being devel­oped for appropriate interoperations.

Transport technologies for high-temperature liquids (molten salt and liquid cadmium) enhance the interoperation of the processes between the electroreduction cell, electrorefiner, counter current contactor and/or cathode processor by decreasing the time required for freezing and remelt­ing. Although the success of large-capacity centrifugal pumps for molten salt reactor systems has been reported (Rosenthal, 1972), the development of small transport systems applicable to pyrochemical processing has not been reported in detail. A molten salt transport test rig (Fig. 10.29) and liquid metal transport test rig (Fig. 10.30) have been installed at CRIEPI to develop transport technologies suitable for pyrochemical treatment (Hijikata 2009a, 2009b). The applicability of a high-temperature centrifugal pump, suction pump and valves has been tested at 500 °C, and the control­lability of the flow rate in a practical range of several L/min demonstrated, as shown in Fig. 10.31. The durability of the high-temperature centrifugal pump was demonstrated by repeating its operation about 500 times during a total of 700 h operation.

Similar to the pellet fabrication system of oxide fuels, handling of the fine fuel powder produced during the voloxidation process needs special care. The voloxidizer installed in KAERI is designed to recover the uniform powder in the bottom vessel after percolation using a vibrating mesh, and the vessel is transported to the electroreducer by a master-slave manipula­tor (Kim, 2005).

The following discussion is not intended to make a case for or against the use of solid-phase extraction technology in nuclear fuel reprocessing. The objective is rather to highlight potential advantages and disadvantages with the aim of stimulating further discussion and developmental work. The appeal of using solid-phase extraction technology for select treatment pro­cesses in the nuclear fuel cycle lies in the potential for achieving the opera­tional advantages inherent to packed bed separations, while maintaining the selectivity provided by ligands historically used in liquid-liquid solvent extraction processes. In theory, the stationary phase functions like a very large number of small, sequential equilibrium contactors, but significantly reduces equipment complexity. A successful implementation of solid-phase extraction for a given separation could potentially eliminate the need for agitated contactors (e. g., mixer-settlers, pulsed columns, centrifugal contac­tors), which in turn reduces the number of moving parts operating in a hot-cell environment. The shortcomings of the technology are, however, not well known at this point due to the lack of engineering and operating expe­rience with any of the solid-phase extraction resins at process scale. There is some debate regarding how well the ligands mimic their liquid phase behavior after they are fixed onto a solid substrate. Several studies have shown deviations to be negligible (Pierce et al., 1963a, Pierce et al., 1963b, Sebesta, 1970, Akaza et al., 1970, Akaza, 1975). Yet other researchers have found some marked differences between ligand performance in certain liquid-liquid and solid-phase extraction systems and suggest that the choice of support may influence ligand behavior (Cortina et al., 1994b, Strikovsky et al., 1996). Horwitz et al. (2006) published a comprehensive study compar­ing the solid-phase and liquid-liquid extraction performance of three acidic organophosphorus extractants for a number of Ln(III) elements. The authors found essentially the same selectivity sequence among the two techniques, but noted that a prediction of solid-phase extraction capacity factors from liquid-liquid distribution data was difficult. This was attributed to potential unavailability of a portion of the extractant in the micropores and/or inexplicable differences in nitric acid dependency between the two systems; also seen with CMPO-PAN studies mentioned previously (Kamenik et al., 2006, Mann et al., 2002). Minor deviation in extractant performance between the two techniques is likely an insignificant factor as long as the design parameters and performance of the solid-phase extraction resin are well characterized prior to scale-up. Nonetheless, any effects due to sub­strate morphology (e. g., pore-size distribution) and extractant impregnation must remain consistent among resin batches and will consequently influ­ence the manufacturing tolerances and economics of the material.

Primary concerns with the application of the resins at industrial scale are the stability of the composite material against radiolysis and acid hydrolysis. It is not expected that degradation of the extractant will be significantly worse than in liquid-liquid extraction for an equivalent amount of absorbed dose. But the potential for longer residence times in the columns, as compared to liquid-liquid contactors, could result in higher doses to the extractant phase. The polymer substrates independently are typically very resistant to acid, and many of the polymers have been shown to remain physically stable up to radiation doses of 106 Gy and cross-linking promot­ers, known as ‘prorads’ can be added to the polystyrene-based polymers to increase the degree of cross-linking under high radiation doses (Drobny, 2003, Sebesta et al., 1995a, 1995b). There are, however, many unknowns regarding how these processes affect the composite material in terms of extractant retention on the substrate and its performance in sequential cycles of loading, washing, stripping and re-conditioning. For instance, can the degradation products be effectively washed from the column, e. g., with caustic wash similar to liquid-liquid extraction systems? Or will they build up on the column over time and reduce stripping efficiency? It is also important that the width of the exchange zones established during the metal loading and stripping cycles remains relatively constant from cycle to cycle. Otherwise, the volume and/or strength of stripping solutions will need to be increased as a reprocessing campaign progresses. A broadening of the exchange zones produces tailing, reducing the purity of separation between metals, and will eventually lead to incomplete column regeneration between cycles. Significant loss of the extractant to the aqueous phase is another key issue from an engineering design and operations perspective. It is desirable that losses are minimal, but also that they are consistent for a given applica­tion. Reproducible performance is essential so that any decrease in perfor­mance with process cycles is known and can be accommodated in design and with manageable adjustments to operating parameters.

As the electrolyte concentration increases, the number of water molecules in the secondary hydration sphere decreases. Consequently, there is a tight­ening of the bond between the metal cation and the hydrate waters in the inner sphere (Choppin, Jensen, 2006). Based on NMR studies of trivalent actinides and lanthanides, Choppin concluded that inner sphere complex — ation by perchlorate ions does not occur below approximately 8-10 M (Choppin, Labonne-Wall, 1997). Multiple equilibria for the uranyl chloride system (UO2Cl2(H2O)2, UO2Cl3(H2O)-, and UO2Cfi2-) have been used for separation of uranium from its progeny or other metals. Since Th4+ does not form anionic chloride complexes, it is retained on cation-exchange resin while anionic chloride complexes of UO22+ pass through the column in the eluate. Alternatively, such anionic complexes can be retained on an anion — exchange column.

The hydration number of Eu(III) remains relatively constant in hydro­chloric acid up to approximately 6-8 M, above which concentration it decreases. The same is true for the hydration number of Cm(III) in HCl, which begins a decline at about 5 M HCl. This difference between (Eu3+ and Cm3+) reflects greater complexation of the actinide trivalent ion by the relatively soft anion Cl-. The difference in chloride complexation has been used to provide efficient separation of trivalent actinides from trivalent actinides in concentrated HCl solutions by passage through columns of cation exchange resin since 1950s (Diamond et al., 1954).

Nitrate complexes for tetravalent actinides, for example, Th4+ and Pu4+, are extremely important in actinide separation and purification processes. Nitrate ions begin to form inner sphere complexes at lower concentrations than chloride anions; this observation is confirmed by the decreased hydra­tion number of the cation even at relatively lower concentrations (Choppin, Jensen, 2006). However, since the oxygen atoms of the nitrate are hard donors, there is no evidence of any covalent enhancement in its bonding as is seen with the chloride anions for the trivalent actinide cations relative to the lanthanide cations (Choppin, Jensen, 2006). In separation and puri­fication processes, the nitrate complexes of actinides are extremely impor­tant. Nitrate-nitric acid solution is the most common aqueous medium in nuclear separation processes. In the case of neutral extractants such as tributylphosphate (TBP), carbamoyl methyl phosphine oxide (CMPO) or dipicolinamides (DPA) it provides nitrate units necessary to compensate the actinide cation charge to enable extraction. Nitrate complexation with hexavalent actinide ions is very weak and the determination of the forma­tion constants for aqueous nitrate solution species is extremely difficult. Under aqueous conditions with high nitric acid concentrations, complexes of the form AnO2(NO3)(H2O)x+, AnO2(NO3)2(H2O)2, and AnO2(NO3)3- (An = U, Np, Pu) are likely to be present. The limiting species in the nitrate series is the hexanitrato complex, An(NO3)62- (Matonic et al, 2002). The complexation of the Pa and Np pentavalent ions by nitrate is known; however, limited thermodynamic and structural data are available. The presumed stoichiometry for the Np(V) species is NpO2(NO3)(H2O)x. For protactinium, which easily hydrolyzes, mixed hydroxo/nitrato or oxo/nitrato complexes have been proposed.

Fluorides and chlorides are the best studied actinide-halide systems, and they are very important for the pyroprocessing and electrorefining processes.

Carboxylic acids are strongly bound to actinide ions. The primary binding mode for simple carboxylic acids is bidentate, while in polycarboxylic acid complexes, carboxylates tend toward monodentate coordination with the metal ion. The affinity of the low-valent actinides for these ligands increases with the denticity of the ligand, for example, ethylenediaminetet — raacetate (EDTA) >>> acetate. For An4+, the EDTA ligand is hexadentate with a twist conformation (a spiral conformation, wrapping around the metal ion, rather than encapsulating the metal ion in a central cavity in the manner of tripodal or macrobicyclic ligands). Diethylenetriamine — N, N,N’,N",N"-pentaacetate (DTPA) has an even higher affinity for both

An3+ and An4+ ions.

The spectroscopic methods discussed above are amenable for measuring components in commercial fuel and fuel simulant feeds. Section 4.3.1 dis­cusses work performed using actual spent commercial fuel, including the preparation and dissolution of the fuel and the spectroscopic measurements of these fuel samples. The spectroscopic measurements were compared with ICP-MS analyses and ORIGEN code calculations. Section 4.3.2 describes real-time spectroscopic monitoring tests using a centrifugal contactor system deployed at PNNL. This section contains measurements resulting from “cold” testing of the contactor system containing HNO3, Nd(NO3)3, and NaNO3 and a “hot” demonstration using solutions containing UO2(NO3)2 and Np(V) in nitric acid.