Vitrification is currently the most widely used technology for the treatment of high level radioactive wastes (HLW) throughout the world. At the Savannah River Site (SRS) actual HLW tank waste has successfully been processed to stringent product and process constraints without any rework into a stable borosilicate glass waste since 1996. A unique “feed forward” statistical process control (SPC) has been used rather than statistical quality control (SQC). In SPC, the feed composition to the melter is controlled prior to vitrification. In SQC, the glass product is sampled after it is vitrified. Individual glass property models form the basis for the “feed forward” SPC. The property models transform constraints on the melt and glass properties into constraints on the feed composition. The property models are mechanistic and depend on glass bonding/structure, thermodynamics, quasicrystalline melt species, and/or electron transfers. The mechanistic models have been validated over composition regions well outside of the regions for which they were developed because they are mechanistic. Mechanistic models allow accurate extension to radioactive and hazardous waste melts well outside the composition boundaries for which they were developed.

The design and development of prototype cold crucible melters for waste vitrification are based on models of the basic physical phenomena, including electromagnetic induction and the thermal and hydraulic properties in natural or forced convection. The complexity of new nuclearized facilities results in significant errors on the results of predictive models based on 2D axisymmetric geometry that can only be resolved by modeling the device in 3D geometry. This document discusses the specification and electromagnetic design of a melter carried out using electromagnetic computation software, FLU3D, developed in 3D geometry by Cedrat. The principles and results of this study are directly applicable to nuclear facilities with allowance for the particular requirements of a nuclearized environment.

Magic-angle-spinning NMR has been used to establish the structural roles of various cations added to the borosilicate glass which is used for the vitrification of high-level nuclear waste (HLW). Representative surrogate oxides with nominal valencies of +1, +2 and +3 have been studied which span the range of oxides from modifier to intermediate and conditional glassformer. NMR has been carried out on those nuclei which are accessible and the species observed have been correlated with the physical and chemical behaviour. The controlling factor is the manner in which the alkali cations partition between the various network groups, changing the distribution of silicon Qn species and the boron N4 ratio. Identifiable superstructural units are also present in these glasses. The aqueous corrosion rate increases with Q3 content, as does the weight loss due to evaporation from the melt. The activation energy for DC conduction scales with N4. Values of N4 obtained for these glasses deviate significantly from those predicted by the currently accepted model (Dell and Bray) and are strongly affected by the modifier or intermediate nature of the surrogate oxide and also by its effect on the distribution of nonbridging oxygens between the silicate and borate polyhedra.

Atomic scale molecular dynamics simulations have been used to predict the location of glass modifying Na, Li and Mg species in a borosilicate Magnox type waste glass adjacent to interfaces with the (100) and (110) surfaces of MgO, CaO and SrO crystals. These simulations show a considerable increase in alkali and alkali earth concentration adjacent to specific interfaces. In particular, there are significant, systematic changes in Na, Li and Mg position and concentration as a function of both the crystal's terminating surface and composition.

57Fe Mössbauer spectroscopy has been used to study the environment and oxidation state of iron in a series of vitrified sewage sludge ash (SSA) wastes, which are broadly similar in composition and variability to some intermediate-level legacy wastes (ILLW). The SSA wastes studied here are incapable of forming an homogeneous glass when melted at 1450 °C although increasing additions of CaO reduce the crystalline content, which consists of Ca3(PO4)2 and Ca3Mg3(PO4)4. Bulk glass transition temperatures of 670–850 °C have been measured, the value decreasing with increasing CaO content. Tetrahedrally coordinated Fe3+ appears to exist in vitrified SSA only at high CaO contents (> ∼ 29 mol. %) and whilst broadly similar to our previous results for simulated vitrified SSA, behavioural differences have been noted between fitted Mossbauer parameters for real and simulated vitrified SSA. We suggest that these could be attributable to the buffering action of carbon and other reducing agents in the real SSA wastes.

The local co-ordination environment of Zr in a model alkali borosilicate glass, of relevance to nuclear waste immobilisation, was studied by in situ X-ray absorption spectroscopy between 25 – 1060°C. Analysis of Zr K-edge XANES spectra, in comparison with those of well characterised standards, demonstrated, for the first time, the reversible transformation of ZrO6 to ZrO7 co-ordination polyhedra at high temperature. This observation was rationalised on the basis of the combined effects of network modifier cation diffusion and thermal expansion.

The U.S. Department of Energy (DOE) is currently processing high-level waste (HLW) through a Joule-heated melter (JHM) at the Savannah River Site (SRS) and plans to vitrify HLW and Low activity waste (LAW) at the Hanford Site. Over the past few years at the Defense Waste Processing Facility (DWPF), work has concentrated on increasing waste throughput. These efforts are continuing with an emphasis on high alumina concentration feeds. High alumina feeds have presented specific challenges for the JHM technology regarding the ability to increase waste loading yet still maintain product quality and adequate throughput. Alternatively, vitrification technology innovations are also being investigated as a means to increase waste throughput. The Cold Crucible Induction Melter (CCIM) technology affords the opportunity for higher vitrification process temperatures as compared to the current reference JHM technology. Higher process temperatures may allow for higher waste loading and higher melt rate.

Glass formulation testing to support melter demonstration testing was recently completed. This testing was specifically aimed at high alumina concentration wastes. Glass composition/property models developed for DWPF were utilized as a guide for formulation development. Both CCIM and JHM testing will be conducted so glass formulation testing was targeted at both technologies with a goal to significantly increase waste loading and maintain melt rate without compromising product quality.

Characterization of glasses and glass nanocomposites using modern transmission electron microscopy techniques is demonstrated. Techniques used include: (i) high-angle-annular dark field STEM for imaging of nanocomposites, (ii) electron tomography for 3D reconstruction and quantification of nanoparticle volume fractions, and (iii) fine structure electron energy loss spectroscopy for evaluation of boron coordination. Precipitation of CeO2 nanoparticles in borosilicate glasses is examined as a function of glass composition and redox partner elements. A large increase in the solubility of Ce is found for compositions where Ce retains +IV valence in the glass. Irradiation experiments with electrons and λ-rays are summarized and the degree of damage is compared by using changes in the boron K-edge fine structure, which allows the gradual transition from BO4 to BO3 coordination to be followed.

Glass composite materials (GCM) are versatile wasteforms for immobilising various types of both radioactive and hazardous wastes. We review current research on the utilisation of GCMs for hazardous and radioactive waste immobilisation. Compared to homogeneous glassy materials GCMs can incorporate larger amounts of waste elements and, in the case of glass matrix composites, they can be produced using lower processing temperatures (by viscous flow sintering) than those of conventional melting.

Pyrochemical and aqueous processing of impure Pu metal and Pu oxide results in a waste stream that contains calcium, plutonium and americium chlorides (CaCl2, PuCl3 and AmCll3) that tend to form insoluble salt layers on top of melts during vitrification. By adding chlorides to ammonium dihydrogen phosphate, ammonium chloride is evolved during vitrification, leaving behind the Ca, Pu and Am cations as oxides. HfO2 and Sm2O3 were used as surrogates for PuO2 and Am2O3 respectively. The effects of waste loading and melt duration on some basic physical properties for different melt durations were investigated. Synthetic batches containing CaO, HfO2 and Sm2O3 in molar ratios 93.34 : 5.91 : 0.74 respectively, reflecting the waste composition, were mixed with a fixed glass composition of P2O5 (75 mol%) and Fe2O3 (25 mol%). NH4H2PO4 was used as a precursor material. The solubility of calcium, hafnium and samarium oxides into phosphate glasses is reported and elemental analysis of the glasses is compared with expected glass compositions.

In the UK, blended high level nuclear waste (HLW) streams from the Magnox and THORP reprocessing plants are currently vitrified using a lithium sodium borosilicate base glass frit. Laboratory and full size non-radioactive simulations (produced on the Vitrification Test Rig at Sellafield [1]) of these compositions have shown that these glasses need to be melted at circa 1050°C to obtain a reasonable viscosity for pouring. Also, at high waste loadings an alkali molybdate phase (termed “yellow phase”) can form in these glasses [e.g. 2, 3]. Vitrification flowsheets are set to avoid yellow phase formation as this phase is highly corrosive to the inconel melter in the molten state and is partially water soluble at ambient temperature and so may challenge product quality.

Ca and Zn additions to the base glass frit have been found to reduce viscosity and allow melt homogeneity and pouring at lower temperatures. It was also theorised that Ca additions could increase the solubility of Mo and thus reduce the likelihood of yellow phase formation. The composition of the phase separated material in as-cast and heat treated specimens of Ca and Zn HLW glasses produced at both laboratory and full scale is examined in this work

Three tests on vitrification of U-bearing DWPF SB2 and SB4 waste surrogates (40-50 wt% waste loading) in a lab-scale (56 mm inner diameter) copper cold crucible with melt surface area of 24.6 cm2 energized from a 5.28 MHz/10 kW generator were conducted. At the feeding of the air-dry batch (3-5 wt% moisture) glass productivity and specific melt productivity achieved under steady-state conditions were 0.64 kg/h and 256 kg/(m2×h) [6150 kg/(m2×day)], respectively. Similar to previously prepared U-free products, the U-bearing vitrified SB2 (highferrous) waste product was composed of U-enriched glassy matrix and high-ferrous spinel structure phase. As expected, the vitreous phase in the vitrified SB2 waste surrogate was enriched with Na, Al, Si, P, S, Ca, Sr, Pb, U oxides and Cl, whereas spinel preferentially accumulated transition metals (iron group elements: Mn, Fe, Ni) as well as Cu and Zn oxides. Iron oxide content in the spinel phase was ∼67–71 wt %. This result was consistent with XRD data and implies that this spinel was close to magnetite with respect to chemical composition and structure. This was consistent with our previous data on characterization of glass crystalline materials containing both U-bearing and U-free SB2 waste surrogate produced in resistive furnace and cold crucibles. The vitrified SB4 (high-alumina) waste products were predominantly amorphous at waste loading of up to 50 wt %. Traces of quartz occurred in the “skull” zone. In all the products, uranium was found to be uniformly distributed within the vitreous phase. Leach rates of Na, Si, Fe, and U from the vitrified SB2 waste surrogate (50 wt % waste loading) determined by a SPFT procedure were found to be 1.38, 1.42, 0.122, and 0.841 (g×m−2×d−1), respectively.

Iodine-129 collected from a reprocessing plant is regarded as the dominant nuclide in terms of safety for TRU wastes disposal in Japan. AgI vitrification (AgI-Ag2O-P2O5) is a potential iodine immobilization technique, which has the advantages of less iodine volatilization (low-temperature vitrification) and high volume reduction efficiency (approx. 1/25 the original waste volume). The iodine immobilization property can be evaluated by examining the surface condition of the AgI glass immersed in water. In this study, immersion tests have been performed on AgI glass in pure water in a 3% H2-N2 atmosphere at room temperature, and the surface characteristics have been examined. The thin layer (<4.3 μm) that is formed has been found to consist of AgI, which may act as a barrier, preventing leaching of glass components. The concepts behind the iodine release model have been proposed based on diffusion and the solubility of the components at the glass surfaces.

In the new Belgian disposal design, the nuclear waste glass will be surrounded by a 3 cm thick carbon steel overpack and a 70 cm thick concrete buffer. An initially high pH is expected after water intrusion in the concrete buffer and this may have an effect on the radionuclide release from the waste glass. This study was performed in order to determine the forward rate of dissolution for SON68 and PAMELA glasses (SM513 LW11 and SM539 HE 540-12), conducting dynamic tests at 30°C in contact with alkaline solutions. In these experiments, the silicon concentration in solution was determined by UV/Visible spectrophotometry according to the blue â-silicomolybdenum method. The forward rates of dissolution were quite similar for the three glasses except at the highest pH for which a slightly higher value was found for SM539 glass. For SON68 glass, a good agreement with the previously established interpolation law was observed until pH 11.5, but at higher pH, the interpolation law slightly overestimates the dissolution rate [1].

Within the scope of the long term behaviour of the R7T7 glass, which is the French nuclear glass, leaching and its coupling with transport mechanisms is studied. Experiments carried out on a SON 68 glass (inactive R7T7 type glass) model cracks in static basic conditions show a strong coupling between solution transport and glass leaching, depending on crack aperture. Moreover, gravity driven convective transport was evidenced for vertical model cracks, whereas only molecular diffusion was detected for horizontal model cracks under the same alteration conditions. In addition, an original device was developed to study the influence of temperature gradients on alteration kinetics as a convective driving force. These experiments show conclusively that thermally- or gravity-induced convective flow must be taken into account, even if such convective effects have not been established experimentally in neutral condition, which is more realistic condition for geological storage. A modeling, based on a porous geochemical software (HYTEC) accounting for both chemistry and transport, has been successfully applied to describe alteration within simple silicate glass cracks. It will be extended to study SON 68 glass model cracks, and more complex fracture networks.

Trying to estimate the long-term durability of cemented wasteforms is a difficult task as the cement matrix is a reactive medium and interactions can occur with the encapsulated waste as well as with the environment. There are few studies of samples that have been stored under controlled conditions for more than 10-15 years. Wasteforms are now being expected to last hundreds of years, much of that likely to be in some form of storage where sample integrity is important. There is also the concern that results from any long-term samples may only be indicative as both formulations and materials change with time.

This paper discusses changes in physical properties that occur in composite cements when some of the short-term accelerated procedures employed in construction testing are applied to encapsulating matrices. Changes after increased temperature of curing, wetting/drying and accelerated carbonation are discussed.

Many of the encapsulating formulations currently used are composite cements where large replacement levels of OPC with supplementary cementing materials (SCMs) such as PFA or BFS are made, primarily to reduce heat output. Accelerating the exposure conditions, either by increasing temperature or through wetting/drying has the effect of changing the hydration pattern of the composite cement by generating more hydration in the SCMs than would normally occur. The large amount of porosity that occurs because of limited hydration allows intrusion of gases and ready movement of water, so the samples subjected to accelerated testing do not appear as durable as expected if stored at ambient.

Storage of the UK's Intermediate Level Wastes (ILW), which comprises Magnox fuel cladding, uranium and small items of equipment exposed to radiation, is currently achieved via encapsulation within cementitious grout housed in 500 litre 316L stainless steel drums. The cements used display a high pH; in such an environment many metals form surface hydroxides or oxides. Magnox reacts with free water at high pH with the liberation of hydrogen whilst undergoing corrosion to form hydroxide species.

Corrosion of Magnox cladding has previously been monitored by measuring the rate of hydrogen evolution and/or weight loss. Recent work by our group has shown impedance techniques may also be useful in monitoring early corrosion behaviour. In this project electrochemical polarisation techniques will be employed to examine the corrosion behaviour of Magnox fuel in situations where it is in electrical contact with other metals, including uranium, and hence determine how galvanic effects influence corrosion behaviour. In this paper we describe the background to such experiments along with some preliminary results.

This paper illustrates the benefits of hot isostatically pressed (HIPed) tailored ceramic waste forms for the immobilisation of Cs and Sr separated from spent nuclear fuel. Experimental data on microstructure and aqueous durability are presented for Cs- and Sr-bearing hollanditerich tailored ceramics prepared with 12-18 wt% waste (on an oxide basis). MCC-1 type leach testing, on the sample containing 12 wt% waste at 90°C for 28 days revealed extremely low normalised 7-28-day Cs and Sr release rates of 0.003 and 0.004 g/m2day respectively.

Zirconolite (nominally CaZrTi2O7) is a constituent phase of potential waste forms for the safe immobilisation of actinide wastes. Structural studies of such materials provide important information about cation ordering, lattice parameters, and strain effects, and provide input into the modelling of alpha decay damage and the development of future wasteform designs. A suite of zirconolites based on the replacement of Ti with Nb and Fe has been studied using high resolution neutron diffraction and resonant X-ray diffraction to determine the degree of disorder across the available cation sites. Resonant X-ray diffraction is a unique method which allows the location of certain cations to be determined accurately by taking advantage of the change in scattering power close to an absorption edge (e.g., Nb-K and Zr-K). Using standard X-ray diffraction alone this is not possible and there is little scattering difference between Nb and Zr.

Raman spectroscopy and measured lattice parameters have shown that the exchange of Ti with Nb and Fe has a non-linear effect on the unit cell dimensions and Raman peak positions while retaining the 2M polytype. Mössbauer spectroscopy has shown that the Fe preferentially fills the Ti split (C2) site. The results from this study provide a more complete picture of the cation order-disorder problem and are generally consistent with the behaviour of lattice parameters across the series.

To immobilize halide and actinide ions present in specific ILW waste a process has been developed that uses mineral phases as the host material. The mechanism of substitution of gallium into these phases will have a large effect on the phase assemblage. This will inevitably affect the total amount of halide that can be immobilized in to total phase mixture.

The full simulated waste stream composition containing varying concentrations (1–40 wt.%) of gallium oxide was studied. Also nominal compositions for gallium doped fluorapatites (Ca10-1.5xGax)F2(PO4)6 (x = 0, 0.25, 0.5, 0.75, 1.0) and gallium doped whitlockites Ca9Gay(PO4)6+y (x = 0.2, 0.4, 0.6, 0.8, 1.0) were prepared at 750–1050 °C.

These were studied by powder x-ray diffraction (XRD) to determine the phase assemblage and solid solution limits of gallium in the apatite and whitlockite phases. It was found that a complete solid solution was formed between whitlockite, Ca3(PO4)2, and Ca9Gay(PO4)6+y. In the nominal apatite compositions it was found that gallium did not substitute into the apatite structure but was instead partitioned over Ca9Gay(PO4)6+y, gallium phosphate, and unreacted gallium oxide. At higher temperatures gallium suppressed the formation of the apatite phase and was largely partitioned into the Ca9Gay(PO4)6+y phase whereas at lower temperature the majority was present as unreacted Ga2O3. In the full DCHP compositions it was found that gallium is likely to be partitioned over a number of phases including apatite, cationdoped whitlockite and gallium phosphate.