Crystalline ceramic wasteforms have been fabricated to immobilise a combined fission product and actinide waste stream arising from spent MOx fuel. The fuel is conditioned by a UREX process, as contrasted to PUREX, to produce a waste stream containing fission products and transuranics. Zirconia rich Synroc derivatives have been formulated to minimise formation of perovskite. This ensures that the transuranics are predominantly immobilised in zirconolite. For comparison, a wasteform has also been produced in which transuranics and rare earth element fission products are immobilised in a radiation resistant cubic zirconia solid solution, whilst caesium, strontium and barium are partitioned to an alumina rich magnetoplumbite phase.

The experience of the Laboratory of Applied Mineralogy and Radiogeochemistry of the V.G.Khlopin Radium Institute on synthesis of Pu-Am-doped ceramics is summarized. During the last 5 years, dozens of actinide doped polycrystalline samples and single crystals have been successfully synthesized such as zircon, hafnon, cubic zirconia, monazite, Ti-pyrochlore, perovskite and garnet. Actinide loading has been varied as follows:

-239Pu - from 5–6 wt.% in zircon (polycrystalline and single crystals), hafnon, garnet and perovskite to 10 wt.% in Ti-pyrochlore and up to 37 wt.% in zirconia;

- 238Pu - from 2.5 wt.% in zircon single crystals to 5 wt. % in polycrystalline zircon and 10 wt.% in monazite and cubic zirconia;

- 243Am - 20–23 wt.% in cubic zirconia and monazite.

The weight of each single ceramic pellet varied from 0.2 to 2.0 grams. Special furnaces developed in KRI for ceramic synthesis allowed obtaining up to 7 ceramic pellets simultaneously during the same experiment. The highest amounts of actinides used under glove-box conditions in the same experiment were: 1.5–2.0 g for 239Pu, 0.6 g for 238Pu and 0.3 g for 243Am. Most experiments on synthesis of ceramics and single crystals doped with 239Pu, 238Pu and 243Am carried out at the KRI did not lead to contamination of internal walls of glove boxes. No release of Pu-Am-aerosols was observed as a result of sintering or melting at 1300–1600°C. These results allowed us to conclude that at the present the KRI has developed the experimental basis for transferring laboratory innovations to the industry of actinide immobilization. It is important that adopting ceramic synthesis methods at industrial scale does not require development of new special equipment.

Murataite-based ceramics were recently suggested as promising matrices for immobilization of rare earths and actinides from high-level waste (HLW). Nevertheless, the crystal-chemical formula of the phase has not been accurately determined yet. We have examined structural features of murataite with Mössbauer spectroscopy. Initial batches were prepared from CaO, Al2O3, Fe2O3, MnO2, ZrO2, TiO2, ZrO2, and UO2. Mixtures were melted in platinum crucibles in a resistive furnace in air at 1450°C for 1 h followed by cooling to 1000°C at the rate of 10°C/min. and final cooling down with the furnace switched-off. Study with XRD, SEM, and TEM showed the samples are composed of two co-existing murataite-type phases with five- and eight-fold elementary fluorite cells. To investigate the valence and structural position of iron ions, Mössbauer spectroscopy on 7Fe nuclei in geometry “on absorption” was used. The data obtained let us conclude that in both murataite varieties trivalent iron is distributed almost statistically (3:1) between B octahedra and C five-vertex polyhedra, while tetrahedral sites T are probably not populated. Taking into account these suggestions, the idealized crystal chemical formula of synthetic murataites can be simplified to A3[8]B6[6]C2[5]O22-x/2. Recalculation of chemical analyses to atomic numbers results in the conclusion that average cation valence for murataites is 3.33. Then, calculation for the generalized formula M11O22-x/2 (M=A,B,C) gives a value x/2 = 3.7, i.e., the number of O atoms in the formula is 22–3.7 = 18.3. Final conclusions on the murataite formulae have to be verified by direct structural research on a single crystal.

Ceramic wasteforms designed to be processed by melting in air has been developed to immobilise Cs-rich wastes. Detailed characterisation electron microscopy is presented on versions of these melted materials which are rich in Cr-, Ni-, Zn or Co- substituted titanate hollandites and which have PCT-B normalised Cs leachate concentrations of < 0.2 g/L. To assist in understanding the general crystal chemistry of titanate hollandites, this study also investigates the solubility limits of Cs in single-phase hollandite BaxCsy(M3+)2x+yTi8–2x-yO16 where M = Cr and BaxCsy(M2+)x+y/2Ti8-x-y/2O16 formulations where M = Zn, Co or Ni.

Pyrochemical reprocessing of plutonium generates several ILW waste-streams and a simple process has been developed at AWE to immobilize the chloride and actinide ions present in one of these waste-streams (Type 1) by calcining the waste with calcium phosphate. We have now investigated the possibility of using this process as the basis for treating a more complex waste-stream (Type 2) and have determined that with some minor modification the original process can effectively immobilize the greater range of cations present in Type 2 waste to again produce a free-flowing non-hygroscopic powder. Weight for weight replacement of the Type 1 waste by Type 2 produced a powder which was approximately 20% deficient in PO43- when compared with stoichiometric powder. Addition of phosphorus pentoxide to the simulant waste/calcium phosphate powder mixture prior to calcination to produce stoichiometric powder significantly improved the chloride ion uptake. In order to comply with safety requirements it is necessary to convert the free-flowing powder into a monolithic wasteform suitable for long term storage. Conversion of the powder to a monolithic wasteform by sintering using a sodium aluminophosphate glass as a binder has been investigated and has proved successful, but to achieve high densities it has been necessary to cold press the samples prior to sintering.

A range of potential wasteforms for the immobilisation of iodine-129 have been synthesised and compared. The work used both iodosodalite and iodovanadinite, as identified in previous studies, as baseline iodine host phases. Syntheses of related compounds using alternative precursors had varying degrees of success. In addition it was shown that iodine could potentially be isolated in a simple composite wasteform of silver iodide particles in a rutile or anatase matrix. Preliminary leaching studies suggest that iodosodalite dissolves congruently whilst iodovanadinite shows preferential release of iodine at a rate approximately ten times that of matrix alteration.

Zirconolite is a potential inorganic matrix that is currently investigated in France, in the framework of the 1991 radioactive waste management law, with a view to provide durable containment of the trivalent and tetravalent minor actinides like neptunium, curium, americium and small quantities of unrecyclable plutonium separated from other nuclear waste. To confirm the actinide loading capacity of the zirconolite calcium site and to study the physical and chemical stability of this type of ceramic when subjected to alpha self-irradiation, zirconolite ceramic pellets were fabricated with 10 wt% plutonium oxide (isotope 239 or 238). The 55 pellets are dense (> 93.3% of the theoretical density on average) and free of cracks. They are characterized by a grain size of between 10 and 20 micrometers. X-ray diffraction analyses confirmed the presence of the zirconolite 2M crystalline structure.

Actinide-containing wastes are among the most dangerous for the environment. Such waste streams originate from reprocessing operations with irradiated nuclear fuel and conversion of weapons-grade plutonium metal into dioxide. The long-term toxicity of actinides derives from the presence of isotopes with half-life varying from hundreds of years (Am241) to tens of thousands (Pu239) or even millions of years (Np237). Therefore, these waste fractions need to be incorporated into durable crystalline host phases. The matrices have to incorporate substantial amounts of actinides, and possess chemical durability and resistance to radiation damage. Complex oxides with fluorite-derived and garnet lattices meet these requirements. Self-propagating high-temperature synthesis (SHS) based on exothermic oxidizing-reduction reactions may be used for production of these waste forms. This technology has the following advantages: absence of extrinsic heating sources, low energy requirements for equipment, high reaction velocity, simplicity of design of processing equipment, feasibility of remote-handling the processes, and lack of considerable amounts of facility decommission wastes. The basic features of the SHS technology are as follows: duration of initiation is 0.05–5.0 sec, temperature in a combustion wave is within 1500–3000 °K, and the velocity of advance of the combustion wave is 1–150 mm/s. Two sets of samples composed of pyrochlore and garnet-type phases were produced with SHS. The first of them corresponds to nominal pyrochlore formulation Y2Ti2O7 doped with various amounts of actinides: 10–30 wt.% UO2, 10 wt % PuO2, 10 wt % NpO2, or 9.5 wt % UO2 + 0.5 wt % Am2O3. The precursor was prepared from oxides of the base phases (TiO2, Y2O3, AnO2), an oxidizer (MoO3), and Ti. In the second set of runs, the target phase was garnet (Y2.8Gd0.2)(Al4.7Ga0.3)O12, where Gd3+ was used as a surrogate for Am3+. The initial batches were composed of MoO3, Y2O3, Gd2O3, Al2O3, Ga2O3, and metallic Al. Phase compositions of the samples have been determined by XRD and SEM/EDS. Samples of the first series are composed of major pyrochlore with minor metallic Mo. The samples with 10 wt % actinides do not contain any separate actinide oxide phase. In the samples with 20 and 30 % UO2 a separate uranium oxide phase was observed. SEM/EDS data allows determination of the limit of solid solution of the pyrochlore phase with respect to tetravalent actinide (U) as 12–14 wt. %. Principle phases in the second series were garnet:

(Y2.82–2.88Gd0.13–0.14)(Al4.69–4.74Ga0.17–0.22Mo0.05–0.16)O12 and Mo-Al-Ga alloy. Small amounts of perovskite - (Y0.86Gd0.12)(Al0.93Ga0.05Mo0.03)O3, Mo, and Al oxides were also observed. Gd and Ga mainly entered in the garnet; small amounts of the elements were incorporated into perovskite (Gd), a metallic alloy, and perovskite (Ga).

The dissolution of the thorium analogue of brannerite (ThTi2O6-I) and U(IV)/U(V) doped Th-brannerite (Th0.97U0.03Ti2O6-II and Th0.955U0.03Ca0.015Ti2O6-III) in aqueous media under atmospheric conditions has been studied to elucidate the effects of pH and uranium valence state on the dissolution rate.

The dissolution of I is nearly stoichiometric but slightly preferential release of U occurs for II and preferential release of Ca and U occur for III. The V-shape pH dependence previously observed for U-brannerite only occurs for U (not other matrix elements) for II, indicating that the pH dependence is related to the U oxidation state upon dissolution. The normalised U dissolution rates of III are nearly an order of magnitude higher than those of II for pH values over 3, suggesting brannerite is less durable with U(V) doping. TEM examination of specimens after leaching revealed few surface alteration products, which is consistent with the nearly stoichiometric dissolution of thorium brannerite.

Zirconolite ceramics are candidate hosts for the containment of minor actinides from the reprocessing of spent fuel in France. Although actinide-doped or rare earth-doped zirconolite ceramics have been successfully produced from pure chemical precursors, there has been little focus on the effect of impurities on the zirconolite microstructure. However, this aspect is crucial for technical feasibility to guarantee a microstructure acceptable for nuclear waste management needs.

First, chemical elements from the reprocessing process that could enter into the actinide flux as impurities were listed. We considered rare earths, zirconium, platinoids (Ru, Rh, Pd), molybdenum, barium, cadmium, silver, caesium, iron, and strontium. As we know rare earths and zirconium entering into the zirconolite microstructure they were not included in the study. For each other impurity listed, a zirconolite batch containing 1 wt% oxide impurity, and enriched in Nd simulating minor actinides, was synthesised then examined by SEM. Impurities were then categorised into those digested into the zirconolite microstructure, those partially digested into the zirconolite microstructure but also forming their own phases, and those that did not enter into zirconolite and formed their own separate phases.

A sample was then produced containing all the impurities, to assess any interference effects. Three minor phases were observed: barium molybdenum oxide, noble metal alloy and zirconium titanium dioxide. These minor phases should not decrease the chemical durability of the ceramic, and therefore, we can conclude that zirconolite is not seriously affected by the impurities introduced.

Ceramics compounds based on the pyrochlore, (Ca,Gd,Hf,An)2TiO7; zircon, (Zr,An)SiO4, and garnet, (Y,Gd,An‥)3(Al,Ga,An…)5O12, where An = U, Am, Pu, Np, Cm, structures have been investigated to determine their use as durable nuclear waste forms for the immobilization of weapons-grade plutonium and other actinides. U, Pu, and Am were incorporated into synthetic zircon, pyrochlore, and garnet and were investigated using in situ cathodoluminescence (CL) and compared with Ce3+, Ce4+, Eu2+ and Eu3+ in order to determine the valence state of the actinide dopants. The results obtained allow for the identification of Am3+ ions in synthetic zircon and garnet. The CL emission of Am3+ is characterized by narrow bands in the orange-red spectral region and is similar to CL emission of Eu3+ in zircon and garnet. No characteristic CL emission of Pu4+ was observed in zircon, pyrochlore or garnet. This is similar to CL emission of Ce4+, which does not have characteristic CL emission in the visible region of the spectrum. Although ions of Ce3+ were characterized by intense CL emission, no characteristic CL emission of Pu3+ was identified in zircon, pyrochlore, or garnet.

The CL study of U-doped garnet and pyrochlore allow for the identification of the presence of two forms of uranium: the uranyl ion, (UO2)2+, and tetrahedral complex, (UO4)2−. Ions of U4+ did not have characteristic CL emission in the visible region. The study of pyrochlore samples doped simultaneously with 238Pu and 238U revealed that CL intensity of the uranyl ion increases with increasing dose of self irradiation from the 238Pu. It was assumed that self-irradiation might cause the change of actinide valence state.

Standard enthalpies of formation, ΔH°f (kJ/mol) at 298 K, were derived for three pyrochlore phases: Ca0.93Ce1.00Ti2.035O7.00 (−3656.0±5.6), Ca1.46U4+0.23U6+0.46Ti1.85O7.00 (−3610.6±4.1), Gd2Ti2O7 (−3822.5±4.9) and two zirconolite phases: CaZr1.03Ti1.97O7 (−3719.4±3.9) and CaHf1.02Ti1.98O7 (−3720.5±3.9). Enthalpies of formation with respect to an oxide phase assemblage, ΔH°f-ox (kJ/mol) at 298 K: CaO + MO2 + 2TiO2 = CaMTi2O7 or Gd2O3 + 2TiO2 = Gd2Ti2O7, and an oxide/perovskite phase assemblage: ΔH°f-pv+ox (kJ/mol) at 298 K: CaTiO3 + MO2 + TiO2 = CaMTi2O7, M = Ce, U, Hf, Zr were also calculated. All of the pyrochlore and zirconolite materials studied were stable in enthalpy with respect to their oxides. ΔH°f-ox: Ca0.93Ce1.00Ti2.035O7.00 (−54.1±5.2), Ca1.46U4+0.23U6+0.46Ti1.85O7.00 (−123.1±3.4), Gd2Ti2O7 (−113.4±2.8), CaZr1.03Ti1.97O7 (−89.6±2.8) and CaHf1.02Ti1.98O7 (−74.8±3.1). With respect to a perovskite plus oxide phase assemblage the hafnium zirconolite was marginally metastable in enthalpy (ΔH°f-pv+ox = +6.0±3.5 kJ/mol) while zirconolite was marginally stable in enthalpy (ΔH°f-pv+ox = −8.8±3.3 kJ/mol). The Ce-pyrochlore was not thermodynamically stable as measured by the enthalpy change relative to perovskite plus oxides: ΔH°f-pv+ox = Ca0.93Ce1.00Ti2.035O7.00 (+21.0±5.5). The U-pyrochlore sample was marginally stable in enthalpy relative to a perovskite plus oxide assemblage: ΔH°f-pv+ox = Ca1.46U4+0.23U6+0.46+Ti1.85O7.00 (−5.1±4.0). A significant field where the proposed waste form for excess weapons plutonium is not thermodynamically stable both at room temperature and at high temperature was defined.

ΔH°f (kJ/mol) at 298 K were derived for three brannerite phases: CeTi2O6 (−2948.8±4.3), U0.97Ti2.03O6 (−2977.9±3.5) and ThTi2O6 (−3096.5±4.3). Enthalpies of formation with respect to an oxide phase assemblage, ΔH°f-ox (kJ/mol) at 298 K: MO2 + 2TiO2 = MTi2O6, M = Ce, U, Th were also calculated. ΔH°f-ox: CeTi2O6 (+29.4±3.6), U0.97Ti2.03O6 (−7.7±2.8) and ThTi2O6 (+19.4±1.6). Thus, Ce and Th-brannerite are entropy stabilized and thermodynamically stable only at high temperature.

These thermochemical data are correlated to the observed trends in the radiation “susceptibility” of these phases. Thermodynamic data relevant to waste form processing and optimization are also discussed.

The effective immobilization of Cs+ and/or Sr2+ sorbed on hexagonal tungsten oxide bronze (HTB) adsorbent materials can be achieved by heating in air at temperatures in the range 500 – 1000 °C. Crystalline powdered HTB materials formed by heating at 800 °C show leach characteristics comparable to Cs-containing hot-pressed hollandites in the pH range from 0 to 12. If the Cs-loaded HTB sorbents are pressed into pellets prior to calcination, ceramic monoliths can be prepared. Heating to temperatures in excess of 1250 °C results in the melting of the sorbent to form millimetre-sized crystals of bronzoid phases. Thermal analysis shows that melting of the cation-exchanged tungstate sorbents is initiated at temperatures as low as 850 °C and concludes by 1300 °C. The absence of any significant mass loss immediately after melting, as well as chemical analyses before and after melting, confirm that Cs is not volatilized to any significant extent at the temperatures required to generate durable, coarsely crystalline products. The melted bronzoid product is a multiphase ceramic in which Cs+ remains bound within, and appears to stabilize, the hexagonal bronze phase, even after complete melting at 1300 °C, while elements such as Sr2+ are present within other tungstate phases. The bronzoid chemical system appears capable of accommodating a wide range of other elements. Here we have demonstrated that modification of the sorbent properties by incorporation Mo does not impact severely on the durability of materials prepared below 1000 °C, even when exposed to strong acid (pH=1) and elevated temperature (150 °C). As an example, one-day MCC1 leach rates lower than 1×10−5g/m2/day were measured using demineralized water at 90 °C for Cs-saturated Mo-doped sorbents that had been heated in air at 900 °C, while the fraction of Cs leached from powdered samples in 0.1 M HNO3 solutions at 150 °C for 4 days is only 4×10−3 g/m2/day.

The hydrothermally prepared hexagonal tungsten bronze (HTB) phase displays promising distribution coefficients (KD) for both Cs+ (2 – 100 ppm) and Sr2+ (0.5 – 60 ppm) in acidic (1M HNO3) radioactive waste simulants. The development of an inorganic ion-exchanger that displays such selectivity has previously eluded researchers in this field. The selectivity for Cs+ and Sr2+ can be modulated by isomorphous substitution of molybdenum into the tungstate framework, and is optimum for material of nominal composition, Na0.2Mo0.03W0.97O3·zH2O (Mo-HTB). Both the parent HTB and Mo-HTB phases display fast ion-exchange kinetics for Cs+ and Sr2+ and cation exchange capacities ca. 50% that of the theoretical capacities of 0.9 and 0.45 mmol.g−1, respectively. The Mo-HTB adsorbent has a modest tolerance to alkali metal ions such as Na+ and K+ in acidic solutions with total Cs+ and Sr2+ uptake dropping by 66% as the concentration of Na+ increases from 9 mmol.L−1 to 1200 mmol L−1.

Progress on separating the long-lived fission products has notably implied basic research on specific host matrices, especially for the immobilization of cesium. Barium hollandite (BaAl2Ti6O16) ceramics have received considerable interest because of their high cesium incorporation ability and chemical stability. This study deals with the preparation of hollandite in the BaxCsy(Al,Fe)2x+yTi8–2x-yO16 (x+y<2) compositional range by an oxide route. Different parameters such as the grain size of the precursor or the temperature and duration of sintering were changed in order to optimize ceramics synthesis. To estimate the hollandite radiation resistance, external electron irradiation experiments (simulating the β particles emitted by radioactive cesium) were performed on hollandite of simple composition. The irradiation-induced defects were studied by Electron Paramagnetic Resonance (EPR) spectroscopy and their nature is discussed.

In the early 1980s a synroc variant, SYNROC-D, was developed for immobilisation of high-level defence waste stored at the Savannah River Plant, USA. A key phase in the immobilisation matrix was spinel, used to immobilise the large proportion of iron and alumina in the waste. Here we examine the feasibility of this approach for other alumina-rich wastes, not necessarily containing iron, derived from the dissolution of aluminium fuel cladding. The advantages of using a magnesia spinel, as opposed to hercynite (FeAl2O4), as the primary alumina-bearing phase are discussed in terms of an increase in waste loading and process flexibility. Two options for sodium incorporation, glass and the titanate phase freudenbergite, are considered.

Phase relations in a pseudo-binary system (1-x) CaZrTi2O7- x GdAlO3 suggested for immobilization of a zirconium - rare earth – actinide fraction of high level waste were studied with X-ray diffraction and electron microscopy. Zirconolite and perovskite were found to be major phases in the ceramic samples prepared by cold pressing and sintering at 1400 and 1500 °C. At relatively low perovskite content (x < 0.5) zirconolite is the major host for Gd, which is considered as a trivalent surrogate for Am and Cm. At higher perovskite content, perovskite becomes the major host for Gd. Zirconolite is the major host phase for corrosion products (Al, Fe, Ti, Zr). Leach rates of Gd, 238Pu, and 241Am from the ceramics studied are 10−4–10−5 g/(m2d).

Glass-ceramic matrices containing zirconolite (nominally CaZrTi2O7) as the only crystalline phase in their bulk can be considered as good candidates for actinide-rich nuclear wastes (containing minor actinides or Pu) immobilization. In this study, three different methods are envisaged and compared to prepare such waste forms using neodymium as trivalent actinides surrogate. Independently on the preparation method, zirconolite is shown to be the only crystalline phase to nucleate in the bulk. However, crystallization of silicate phases (titanite CaTiSiO5 + anorthite CaAl2Si2O8) can occur from samples surface and can compete with zirconolite crystallization. The effect of the crystal growth thermal treatment duration (2–300 h) at high temperature (1050–1200°C) on glass-ceramics structure and microstructure is studied. In the oxides system studied here, it appears that zirconolite is not thermodynamically stable in comparison with titanite but, for kinetics reasons, such transformation will not occur during waste forms disposal.

Hollandites with compositions Ba1.2-xCsxMg1.2-x/2Ti6.8+x/2 O16, and Ba1.2-xCsxAl2.4-xTi5.6+xO16 (x=0, 0.1, 0.25) have been synthesised using a modified alkoxide/acetate precursor route. The samples have been sintered using two procedures; hot isostatic pressing and sintering at ambient pressure. X-ray powder diffraction has shown samples from both systems to form tetragonal hollandites, with little change when pressed by both methods. Cs-133 MAS NMR spectra have been recorded showing the chemical shift in Al containing samples to be ∼250ppm, and in Mg hollandites ∼175ppm and 200ppm, with little change when prepared by both methods.

Pu-bearing zirconolite and pyrochlore based ceramics were prepared by melting under oxidizing and reducing conditions at 1550 °C. 239Pu content in the samples ranged between ∼10 and ∼50 wt.%. Phase composition of the ceramics and Pu partitioning were studied using X-ray diffraction (XRD) and scanning electron microscopy with energy dispersive system (SEM/EDS). Major phases in the samples were found to be the target zirconolite and pyrochlore as well as a cubic fluorite structure oxide. Normally the Pu content in the Pu host phases was 10–12 wt.%. This corresponds to the Pu content recommended for matrices for immobilization of excess weapons plutonium. At higher Pu content (up to 50 wt.%) additional phases, such as a PuO2-based cubic fluorite-structured solid solution, perovskite, and rutile were found.