The TRUEX process separates the lanthanides and transuranic actinides from the other components in the CCD-PEG, FPEX or NPEX raffinates, notably zirconium and noble metals. Two TRUEX flowsheets have been tested in the demonstrations. The first TRUEX flowsheet is shown in Fig. 7.8 (Pereira, 2007a). Americium, curium and the lanthanide elements are extracted by the TRUEX solvent, which consists of CMPO and TBP diluted by n-dodecane. When the NPEX step is omitted, plutonium and neptunium are also extracted. Lesser amounts of other fission products are also extracted and must be scrubbed from the solvent. The first TRUEX flow­sheet is unique to the UREX+ process by having three scrub sections. In the first scrub section the impurities are removed from the solvent using oxalic acid. Moderately concentrated nitric acid scrubs oxalic acid from the solvent in the second scrub section. In the third scrub section, addition of dilute nitric acid lowers the acid concentration in the solvent to allow effective stripping. The transuranic actinides and lanthanide elements are stripped from the solvent using a lactate buffer containing a complexant which becomes the feed to TALSPEAK. An acetate solution was used when the Cyanex 301 based process followed. In the simplified TRUEX flowsheet shown in Fig. 7.9 (Pereira, 2007b), the three scrub sections were reduced to a single scrub section by adding oxalic acid to the feed directly. The single

scrub section is used to adjust the acidity of the organic phase prior to entering the strip section.

The raffinate of the NPEX process can be fed directly to the extraction section of the TRUEX segment with no feed adjustment. To serve as the TRUEX feed, the CCD-PEG or FPEX raffinate compositions were adjusted by addition of concentrated nitric acid. Just prior to introduction of the feed, a reductant was added to reduce Np(V) to extractable Np(IV).

To confirm the effectiveness of the UNEX-extractant under operating con­ditions, special tests of a prolonged continuous process duration (66 hours) were performed at RI and the Idaho National Laboratory. Over this period of time, the extractant made 89 recycles. The tests demonstrated that:

• extractable components are not accumulated in the recycled extractant;

• concentration of the most water-soluble component (PEG-400) can be maintained by its addition to the fed strip agent;

• increase in the degree of radionuclide concentration in the strip product is attained due to a temperature rise to 60°C at the stripping operation stage;

• rates of recovery of radionuclides (Cs — 99.78%, Sr — 99.984%, Eu — 99.997%) allow the transfer of the raffinate into the LLW category;

• elimination of the extractant regeneration operation from the improved flowsheet did not result in accumulation of extractable components in the recyclable organic phase;

• as far as metal macroimpurities were concerned, Ba and Pb were almost completely extracted during the tests, whilst Ca, Fe, K, Mo were partially extracted.

The results presented have already been partly published [26-29]. All the UNEX tests up to 2001 year were reviewed in [30] (see Table 9.17). The UNEX process was patented in Russia and in the US. [31]

In order to verify the UNEX process within the framework of DOE Project EM-50/JCCEM under contract with Sandia National Laboratories’ “Testing and demonstration of UNEX-process technology”, an extraction setup was created in hot cells at the Central Plant Laboratory of the Radiochemical Plant (RChP) at Mining Chemical Combine (MCC), Zheleznogorsk, Russia. The EZ-33 centrifugal contactors developed and manufactured by NIKIMT were the main extraction process equipment used. The UNEX-process was tested in two stages — on simulated and actual waste feeds [32]. The experimental objective was to: determine the techno­logical parameters of the UNEX process when treating simulated and actual liquid HLW; to investigate the distribution of stable and radioactive

Flowsheet test Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Solvent wash 3 M HN03 5 M HN03 2 M HN03 None None None Radionuclide Cs 99.4% 137Cs 99.95% 137Cs 99.4% Cs > 97.5% Cs 99.95% 137Cs 99.99% removal Sr 99.97% soSr 99.985% soSr 99.995% Sr > 99.993% Sr > 99.999% soSr 99.73% efficiencies Eu 4.3% to >99.92% Alpha 95.2% Alpha 99.96% Eu 17.2% to 34.1% Nd > 98.3% Ce > 99.6% Alpha > 99.9% Matrix metal Zr 52% Zr >97.7% Zr 87% Zr < 6.4% Zr 3% Zr 0.7% Removal Mo <3.1% Mo 19% Mo 32% Mo < 19.2% Mo > 2% Mo 12% efficiencies Fe 10% Fe 6.9% Fe 8% Fe < 13.2% Fe 9% Fe 2% Ba 99.5% Ba >87% Ba > 99% Ba > 99.6% Ba 99.7% Ba, Pb 100% Pb 99.8% Pb >98.5% Pb > 98.8% Pb > 99.94% К 50% Mn 23% К 20% К 17% Ca 10% К 28% Notes Two tests with 4 hour test with 3 hour test with 66 hr run time w/ 4 hour test with 3 hour test light phase solvent recycle. solvent recycle. solvent recycle. solvent recycle. with solvent solvent; flooding Flooding was No No precipitation No precipitation recycle. No and precipitation observed in the precipitation or or flooding was or flooding was precipitation observed in strip actinide strip flooding was observed. observed. or flooding section of initial section. observed. was test. observed.

aGN = guanidine nitrate, bGC = guanidine carbonate, CAHA = acetohydroxamic acid, dDTPA = diethylenetriamine pentaacetic acid, eANN = aluminium nitrate.

elements over the various stages of the extraction units; to check the opera­tional reliability of the EZ-33 centrifugal contactors with high radiation loadings under prolonged operational conditions.

Initially a simulated solution, corresponding to real process waste but composed of stable macrocomponents, was treated. The duration of the experiments was 48 hours and the volume of treated solution was 29 l. The process flowsheet of the setup involved the following operations: extraction of target elements (18 stages); extract scrubbing (2 stages); combined strip­ping of target elements (16 stages).

In the second part of the experiment, actual high-level waste from a radiochemical plant (Pr. 501) was used, with the following composition: Pu

— 152 mg/l; U — 102 mg/l; Np — 0.38 mg/l; Am — 68 mg/l; Th — < 0.5 mg/l; Fe

— 645 mg/l; Cr — 340 mg/l; Mn — 1,2 g/l; Ni — 510 mg/l; Pb — 6 mg/l; Al — 220 mg/l; Ba — 12 mg/l; Ca — 230 mg/l; F-ion — 2,7 g/l; Si — 100 mg/l; HNO3

— 69 g/l; NaNO3 — 47 g/l; 90Sr — 0.3 Ci/l; gross a-activity — 1,4 Ci/l.

The specific volumetric activities of the radionuclides in Bq/l were:-125Sb

— 4.89*107; 134Cs — 1.02*107; 137Cs — 5.3*108; 155Eu — 5.82*108; 154Eu — 3.37*107; 144Ce — 9.31*109; 60Co — 2.67*107; 106Ru — 2.07*109. In the second test 77 l of actual waste were treated. The time operating on the actual solution was 128 hours, and the solvent made 51 recycles. Results of checks on the extrac­tion properties of the recycled solvent and its chemical analysis during the test did not reveal any considerable deviations from initial characteristics. Some decrease of Sr and Cs distribution coefficients were noted — by 10 and 18.5% of their initial values, respectively; however, this fact did not affect the process parameters as a whole.

On treatment of actual waste, the recovery rates of target radionuclides, as calculated from their content in the raffinate, was: Pu — 99.76%; Am — >99.02%; Sr-90 — 99.99%; U — >99.0%; Eu-155 — >99.91%; Eu-154 — >98.75%; Cs-137 — 99.95%; Ce-144 — >99.98%. Al, Cr, Ru and Sb are not recovered totally (their content in the strip product is below detection levels), while Pb and Ba are extracted and are removed with the strip product. Up to 20% Ca, 25% Fe, 23% Mn, 5% Co and 3% Ni (measured as a percentage of their content in the feed solution) are also withdrawn with the strip product.

The potential applicability of electrochemical methods such as normal pulse voltammetry to the on-line monitoring of actinide concentration in molten chlorides was reported by CRIEPI (Iizuka, 2001a). The device used consists of wire electrodes (W, Ta) and a reference electrode connected to a poten — tiostat. As shown by the linear relationship between the reduction current and concentration in Fig. 10.27, the applicability of electrochemical methods was demonstrated for concentrations of U or Pu in LiQ-KQ-UQ3-PuQ3 molten salt of up to 1.7 wt%.

A recently developed tridentate ligand, N, N,N,,N’-tetraoctyl diglycolamide (TODGA) is a promising extractant being considered for the separation of An(III) and Ln(III) from HLLW solutions (Sasaki et al., 2001, Narita et al., 1998, Sugo et al., 2002, Tachimori et al., 2002). The tridentate functionality provided by the additional etheric oxygen has led a number of researchers to consider TODGA over the commonly used neutral diamides with biden­tate properties (Ansari et al., 2005). The TODGA ligand structure is shown in Fig. 13.11.

Hoshi et al., (2004) evaluated the use of TODGA impregnated SiO2-P solid-phase extraction resin for performing group separation of MA and Ln elements from HLLW. This is one step in the proposed ERIX (Electrolytic Reduction and Ion exchange) process for reprocessing spent fast breeder reactor mixed oxide fuel (FBR-MOX). Column tests using simulated HLLW showed that group separation of Am and some Ln ele­ments relative to other fission products was satisfactory. Zirconium, Ru, Sr, and Y were also extracted; however, all were effectively separated from the An/Ln elements using appropriate elution schemes, i. e. using different concentrations of HNO3 and oxalic acid for Zr. The authors found the uptake and elution kinetics of Ru to be slow compared to the other ele­ments and were only able to recover 85% of the Ru fed to the column. The loss of TODGA to the aqueous phase was also investigated and the researchers concluded that <0.02% (w/w) of the extractant was lost during one separation cycle.

A solid-phase extraction resin consisting of 30% (w/w) TODGA and 10%(w/w) TBP adsorbed onto Amberchrom CG-71 was evaluated for the separation of An and Ln elements from a simulated PUREX raffinate solu­tion (Modolo et al., 2007). The simulant represented a PUREX raffinate (5000 L/t HM) from fuel having an initial 235U enrichment of 3.5%, 33 000 MWd/tHM burn-up, and a 3-year cooling period. Small column (22.9 cm3) studies indicated that the An and Ln elements could be effec­tively separated from the light fission products using four nitric acid scrub­bing steps followed by a strip with 0.01 M HNO3 and a H2O wash. Oxalic acid and HEDTA were also included in the first and second scrub steps to effect the complete removal of Zr and Pd from the column. Similar to comparable investigations (Hoshi et al., 2004), a portion of the Ru remained on the resin, as well as a small fraction of Cf and Y. It was noted that this
issue needs further investigation and that a more radiolysis resistant sub­strate (e. g., silica) may also be needed for large scale applications.

Various flow sheets using CMPO and TODGA sorbed onto SiO2-P have also been proposed and tested for partitioning An(III) and Ln(III) ele­ments as part of the proposed Minor Actinides RECovery (MAREC) process based on solid phase extraction. (Zhang et al., 2004, Wei et al., 2004b, Zhang et al., 2005c, Zhang et al., 2008). The stability of the TOGDA-SiO2-P resin as a function of nitric acid concentration, temperature and y-radiation was investigated by Zhang et al., (2005a). The authors evaluated the effects of these parameters by comparing the batch adsorption of Nd onto the resin. Nitric acid concentrations up to 3M did not decrease the adsorption properties of the resin at 25°C. The capacity of the resin was, however, reduced by approximately a factor of three in 3 M HNO3 at 80°C, presum­ably due to losses of the extractant to the aqueous phase as evidenced by an increased total carbon concentration in the aqueous fraction. Irradiation of the resin from approximately one to four MGy resulted in a linear decrease in extraction capacity. The highest dose (4.1 MGy) resulted in a 70% reduction of capacity, which was also determined to be from extractant losses. Studies to quantify these losses as intact TODGA molecules or its radiolysis products were inconclusive. Similar investigations were con­ducted by these researchers (Zhang et al., 2005b) to assess the effects of HNO3, temperature and y-radiation on CMPO-SiO2-P solid phase extrac­tion resin. Nitric acid (up to 3 M) and increased temperature (80 °C) did not cause a significant decrease in capacity for Nd. The y-radiation experi­ments indicated a linear decrease in capacity, with an approximate decrease of 50% at the maximum dose of 4 MGy.

The generic solution of the above system is provided by a diffusion-reaction model in which metal removal from the bulk liquid may be mass transport limited during high biomass output and reaction rate limited during inhib­ited growth. The mass balance across the bulk liquid compartment (Fig. 15.11a) is given by the following equation:

15.8

15.9

where rfl = dissolved species removal rate (ML^T1); ju = Dwidujdz, dissolved species flux rate (ML^T-1); u = {P, U, C}, is the vector for dissolved species concentration, where P = phenol concentration (ML-3), U = metabolites concentration (ML-3), and C = metal pollutant concentration (ML-3); rXB = cell growth rate in the bulk liquid (ML-2^1); x = {Xp, XE, X/}, is the vector of biomass concentrations, where Xp = viable cell concentration (ML-3) of organic compound degraders, XE = viable cell concentration (ML-3) of toxic metal reducers, and Xj = inert cell concentration (ML-3); bx = {bxp, bxE}, is the vector of cell death coefficients (T^1) for the respective cell types; and X = {V, XE, X/}, is a vector of cell detachment rate coefficients (7"1). The reaction rate and cell growth terms can be formulated based on the enzy­matic removal of substrates as shown in the derivation of Equation 15.5 and the Monod kinetics, respectively.

Biofilm mass balance

The removal of the dissolved species and cell growth is represented by a set of diffusion-reaction partial differential equations (PDEs). The PDEs represent a mass balance across an infinitesimal biofilm section (Sz) parallel to the substratum surface (Fig. 15.11b):

15.10

15.6 where: jx = Dwx3(x)/3z, mass flux rate of biomass (ML-2^1), ruf = the vector of removal rates of dissolved species in the biofilm (ML-3^1), rxf = the vector of biomass production rates in the biofilm zone (ML-3^1), and є = is a biofilm porosity constant (Vfvods/Vftotal). The movement of cells across the biofilm is induced by physical displacement due to growth whereas dis­solved species are transported by diffusion. Thus, the values of the j terms for cells are expected to be lower (by orders of magnitude) than the j terms for dissolved species in the biofilm.

The oxidation state of actinides is a primary determinant of their behavior in solutions. As an example, Table 2.3 and 2.4 summarize their color and stability in aqueous solutions.

The trivalent oxidation state of f elements, which is so markedly stable in the lanthanides and heavier actinides, is not typical for Pu, Np and U. The ability of the light actinides to exhibit multiple oxidation states leads to their rich and complex redox chemistry. Comparison of the stability of their oxidation states (Table 2.4) which are supported by data on their standard reduction potentials at pH = 0 in different acids (Figs 2.1-2.2), can lead to interesting conclusions. For example, while uranium (III) can reduce water, U(IV) and Np(IV) can be stabilized in aqueous solution only under anaerobic conditions. U(III) exists at -0.631 V vs. saturated hydrogen elec­trode, indicating that U(III) liberates H2 from water. Np(III) (0.155 V) is

still a strong reducing agent, but less so than hydrogen. Plutonium Pu(III) is stable at 0.982 V.

The reduction potentials for the four common oxidation states of pluto­nium (III-VI) under acidic conditions are all near 1 V. As a result, all four oxidation states can coexist in aqueous solutions (Newton, 1975). The

HNO2 094 HNO3

2.1 The scheme of standard redox potentials (in volts) for U, Pu, Np and 1 M HNO2/HNO3 (Miles, 1990; Drake, 1990).

equilibrium concentrations of plutonium species existing simultaneously will be determined by Equations 2.1 and 2.2.:

Pu4+ + PuO2+ о Pu3+ + PuO22+ 2.1

Pu4++ 2H2O о Pu3+ + PuO2+ + 4H+ 2.2

The equilibrium constant for Equation 2.2 is dependent on [H+]4, hence the position of this disproportionation equilibrium changes significantly with acidity.

The couples in which only an electron is transferred, e. g. Pu3+/Pu4+ or NpO2+/NpO22+, are electrochemically reversible and the redox reactions are rapid. Redox reactions that involve forming or rupturing of the actinide — oxygen bond, e. g. Np4+/NpO2+ and Pu4+/PuO22+, are not electrochemically reversible and have a slower reaction rate because of the barrier introduced by the subsequent reorganization of the solvent shell and also because some of these are two-electron reductions (Edelstein, 2006).

Disproportionation of neptunium is appreciable only for the Np(V) oxi­dation state, and the reaction is favored by high concentration of acid:

2NpO2+ + 4H+ о Np4+ + NpO22+ + 2H2O 2.3

Table 2.5 summarizes the activation parameters of the An(V) disproportio­nation reaction in perchlorate medium [H+] = 1.0 M at 25oC.

(8)

(9)

Both the tetravalent and hexavalent cations, having higher effective cationic charge, are more strongly complexed by ligands, thus the dispro­portionation reaction will be accelerated toward completion by an addition of complexing agents. Obviously, the reproportionation reaction will be promoted by lower acidity of solution:

Np4+ + NpO22+ + 2H2O о 2NpO2+ + 4H+ 2.4

Examination of the redox potential diagrams of U, Pu and Np, compared with part of the nitrogen-potential diagram, leads to the conclusion that in

HNO3 solution containing NO2-, only U(VI), Pu(IV), Pu(VI) and Np(VI) should be present; trivalent species are not stable (Miles, 1990; Drake, 1990). The oxidation of Np(V) by nitrate ion is favored by high HNO3 and low HNO2 concentrations, and conversely, at high concentrations of HNO2, Np(VI) is rapidly reduced to Np(V) (Drake, 1990). Moulin found the rate of oxidation of Np(V) to be dependent on the [HNO2]/[Np(V)] ratio and the nitrate concentration; the mechanism of oxidation starts with the proton activation of Np(V), followed by oxidation of activated species by nitrate (Moulin, 1978). The rate-limiting step of the overall oxidation of Np(V) is the formation of the activated species, except at very low concentrations of HNO2, when oxidation of the activated species becomes comparatively slow (Siddall, Dukes, 1959).

Obviously, the role of the proton in redox processes is paramount, which is confirmed by the fact that the most dramatic changes in reduction poten­tials are observed with increase or decrease of acidity of aqueous solutions. For example, a change of medium from 1 molar acid to 10 molar base causes a change in redox potential of 2 V and makes possible the oxidation of Np(VI) to Np(VII), as seen in Equation 6 in Scheme 2 (Fig. 2.2). Np(VII) can in fact only be prepared in strongly basic solution; a similar effect is seen for Pu, though higher concentrations of base are required.

Generating unusual oxidation states opens new separation opportunities. For example, the typically trivalent Am, if oxidized to upper oxidation states, can be separated from lanthanides either as Am(VI) in a manner similar to other hexavalent actinides (e. g., PuO22+ or UO22+ by extraction with TBP) or, similarly to poorly extracted Np(V), pentavalent Am should be poorly extracted by most solvents, while Ln will be extracted by many extractants. Numerous candidates, including TRPO, CMPO’s and diamides (malonamides, diglycolamides and even picolinamides) are all potential reagents (Nash, 2009).

The equilibria and variety of actinide species discussed above confirm the necessity of careful control of the conditions and oxidation state of the actinides in separation processes.

The uranium extraction (UREX) solvent system was selected as a model flowsheet for testing Raman and absorbance spectroscopic techniques for process monitoring and safeguarding purposes and served as a guide for selecting the various feed concentrations and solvent system bounding the choice of spectroscopic measurements. The initial effort was directed toward evaluation of current UREX flowsheet specifications received from Argonne National Laboratory (ANL) (Vandegrift 2004). Based on this report evaluation, specifications were developed for preparing solutions that simulate the UREX process flowsheet streams. These solutions were used to evaluate capabilities for testing in-house instrumentation applicable for on-line process monitoring. To demonstrate the feasibility of using spec­troscopic techniques for the process monitoring and control of the UREX flowsheet, a 0.8 M HNO3 solution matrix was selected. The simple baseline feed solution contained 1.33 M UO2(NO3)2 in 0.8 M HNO3 (labeled Simple Feed). The Simple Feed simulant served as a matrix for the Pu(IV, VI) and Np(V) measurements at variable concentrations in the 0-10 mM range. The organic solvent extraction solution containing 30 vol% TBP in n-dodecane was prepared and loaded with actinide nitrates by the batch contact equili­bration with the UREX feed simulant solutions.

Raman spectra were collected on an InPhotonics, Inc. RS2000 echelle spectrograph. The system was equipped with a stabilized 670 nm, 150 mW visible diode laser as the excitation source. Data were collected at 1 cm-1 spectral resolution over a range of 200-4000 cm-1 stokes shift (Raman shift from 670 nm). Samples were measured with an InPhotonics focused fiber-optic probe (RamanProbeTM) with a thermoelectrically cooled CCD detector, normal operating temperature -55°C. The laser beam was coupled to the sample through a 10m fiber-optic cable and probe assembly which focused the excitation beam directly into the sample; this also precluded any air gap between the laser source and the sample. The focal point of the laser beam was 5 mm beyond the end of the laser probe tip, the measured laser intensity at the sample was typically 50 mW, and the excitation laser beam diameter at the sample was measured as 3 mm. Molecue® acquisition software with GRAMS 32® data manipulation software was used to process the Raman data. Typically, an integration time of 2 to 20 seconds was used for each acquisition. Vis-NIR measurements were performed using a 400 series dual source reflectance fiber-optic probe (SI Photonics) coupled with an LS-1 Tungsten Halogen Light Source (Ocean Optics) and a USB2000- vis-NIR spectrophotometer (Ocean Optics). SpectraSuite (Ocean Optics) software was used for collection and processing of the spectrophotometric data.

The adaptations of the PUREX process, whether it is Advanced PUREX, COEXTM or UREX, all use the TBP/kerosene solvent and extract actinides from nitric acid media. As previously discussed, neptunium will be present in several oxidation states where only the pentavalent is completely unex­tractable into the TBP containing organic solvent. Depending on the ulti­mate fate of the neptunium (and plutonium) in the process, several adaptations to the PUREX process are possible (Taylor 1997, Fox 2006).

One option is to route the neptunium to the high active raffinate in the codecontamination cycle of Fig. 6.2 and dispose of it as waste. This requires that neptunium is held at the pentavalent state as NpO2+ while not interfer­ing with the extraction of plutonium, i. e. any reducing agent to reduce Np(VI) to Np(V) cannot reduce Pu(IV) to Pu(III). This criterion makes the single decontamination step of neptunium in PUREX very difficult. Furthermore, this option is also not advantageous from the standpoint of advanced fuel cycles in that the Np and the minor actinides (Am and Cm) would still need to be separated in subsequent steps.

A second option would be to let neptunium extract together with U and Pu in the first cycle and selectively strip neptunium in a separate contactor before the partitioning cycle of Fig. 6.2. This has the same inherent problems as the first option in that the reduction of neptunium to Np(IV) may affect the Pu oxidation state, and therefore Pu behavior.

The final options are all based on the simultaneous extraction and separa­tion of neptunium and plutonium together, creating a relatively pure uranium product and a Np/Pu product. This could be accomplished in the Pu reduction and scrub sections of the partitioning cycle (Fig. 6.2), which is the preferred route in the UK (Fox 2006) and one option of the French COEXTM process. Alternatively, this could be accomplished in the codecon­tamination cycle, routing both neptunium and plutonium together with the fission products and minor actinides to the raffinate; which is a preferred route in the US and provides the reasoning behind the UREX process. In either case, the technique and redox reactions chosen must also take into consideration the possibility to use centrifugal contactors, i. e. the require­ments of fast kinetics, and the aggressive environment, i. e. stable chemical compounds or degradation products that will not interfere with the process.

Several methods for controlling the Pu and Np oxidation states have been investigated. The French have largely focused on established redox tech­niques such as HAN and HNO3 concentration. In the UK and US, the emphasis has been on the use of simple hydroxamic acids for the combina­tion of reduction together with aqueous phase complexation of neptunium and plutonium, effectively decontaminating the uranium product. The current hydroxamic acid candidates are either acetohydroxamic acid

H3C

(a)

O

H

(b)

6.4 Chemical structures of (a) formohydroxamic acid and (b) acetohydroxamate coordinating to neptunium(IV).

(AHA) or formohydroxamic acid (FHA), both having been proven to reduce and strip Np(VI) from the organic solution and to complex Np(IV) and Pu(IV) in the aqueous phase (Taylor 1998, May 1999).

Fuel-cladding separation

The generation of the finely divided fuel powder by means of the voloxida — tion process enables separation of the oxide fuel from the segmented or punctured cladding by means of screening. Under adequate agitation pro­vided by a rotary or vibrated oxidation reactor, over 99% of the fuel can be cleanly separated. A small amount (0.1 to 1 wt %) of the finely powdered fuel clings to the surface of the cladding and can be recovered by acid washing. Tests with a variety of sources of commercial used fuel, ranging from low-burnup, long-decayed fuel to very high burnup, short-time — decayed fuel have demonstrated that an efficient separation (>99%) of fuel from cladding is possible.

After the deposits are mechanically removed from the solid cathode rod or from the cathode crucible used at the electrorefining process, each deposit is heated in a ‘cathode processing’ step to separate the actinide metals by distillation of adhering salt and cadmium. As shown in Fig. 10.4, the vapour pressures of cadmium and chlorides are sufficiently higher than those of the actinide elements to be separated from them by distillation.