The potential value of Partitioning/Transmutation (P/T) strategies for high-level nuclear waste management has been investigated (one would be tempted so say “periodically”) since 1970. The basic physics phenomena are well understood, and scenarios for the fuel cycle can be imagined. However, the criteria to judge the impact of P/T are matter for discussion (and controversy). Moreover the relation to disciplines other than reactor physics can be significant (e.g. innovative fuels and Actinide chemistry).

A number of key issues are revisited here and will be discussed in some detail.

A regulatory authority in the field of nuclear waste management will not be responsible for the waste management itself. The regulator will however be responsible for reviewing the safety of the waste management activities proposed and performed by nuclear power plant owners or special organisations set up to manage and dispose of nuclear waste. In this paper the regulator's need for R&D to develop competence in the assessment of nuclear waste management safety is discussed.

First of all I would like to express my appreciation to the MRS for the opportunity to address this symposium. I am especially pleased to be speaking alongside Charles McCombie and Soren Norrby, who have both made me feel welcome whenever we have met in the past. I have been more active in Sweden over the last few years, as Soren will know. Perhaps it is now time I turned my attentions to Switzerland as well.

Is there any need for further scientific and technical development work in the field of waste management? Radically different views on this provocative question have been put forward by opposing sides in the nuclear debate. P. Slovik [1] quotes various leaders on the nuclear community who for years have been claiming that waste management has no unsolved technical problems. At the recent public hearing in England [2], on the other hand, Greenpeace and others convinced the Inspector that, despite a GBP 200 million site characterisation programme with many deep boreholes, UK-Nirex still knew too little about the deep hydrogeology to justify even the sinking of an exploratory shaft at Sellafield. Where does the truth lie?

This paper summarises the last talk given at the conference. Its contents were prepared during the first four days of the symposium with the goal of comparing the R+D needs as specified by three invited lectures with the actual scientific work as given in oral and poster contributions by symposium participants. We first present some general conclusions and then address specific aspects of the individual sessions. Finally, the main conclusions from the six workshops are given.

The short-term dissolution behaviour of a complex borosilicate glass has been investigated by controlled pH leaching, surface titration and leach rate temperature dependence experiments. The results indicate that the rates, congruence and mechanisms of dissolution vary significantly with pH. At low pH, dissolution occurs via a proton-promoted mechanism which results in enhanced release of B and many network modifying elements (relative to Si). At 60°C in high pH media, dissolution is essentially congruent. Here, dissolution is surface reaction controlled and occurs via a hydroxyl-promoted network dissolution process. Selective leaching is favoured at low and near-neutral pH. Congruent dissolution occurs in solutions of pH greater than that at the point of zero net proton charge.

Neodymium-doped zirconolite materials may be synthesized by two melting processes. One involves devitrification of an aluminosilicate parent glass containing titanium, zirconium and neodymium oxides, yielding a glass ceramic comprised of submicron zirconolite needles embedded in a silica-rich glass matrix. The second method consists of melting an oxide mixture with the stoichiometry of a highly Nd-enriched zirconolite, then quickly cooling the melt to produce a ceramic rich in zirconolite crystals several hundred microns long containing a large fraction of the initial neodymium.

The objective of this work was to develop a more representative mathematical formulation of the alteration kinetics of the borosilicate SON68 glass by combining three approaches: (1) Compare extensive prior experimental static leaching results for SON68 glass with the first-order kinetic law in which silica is the predominant element, (2) Assess the exact role of dissolved silica on the alteration rate under conditions near and far from saturation, by means of dynamic leach tests and, (3) Compare the new data with the general kinetic law for silicates in which the reaction affinity, catalysis and inhibition are the three influencing factors.

The local cation environments in a borosilicate glass (containing the main components of the French nuclear waste glass, R7T7) and of basaltic-like glass are presented on the basis of 17A1 and 23Na Multi-Quanta Magic-Angle Spinning Nuclear Magnetic Resonance (MQ-MAS NMR) spectroscopies. The chemical and geometrical environment of each nucleus is characterized. Moreover, qualitative information about the distribution of these parameters, characteristic of the disorder level in the structure, is obtained. These results are presented with the characteristics of the Al and Na dissolution in water at 100°C. The relation between the local environments of these cations and their mass loss in solution is discussed.

The incorporation of cerium, as a simulant for plutonium, has been investigated in both zirconolite and perovskite under oxidising and reducing conditions. The Ce valence in zirconolite was very sensitive to the processing atmosphere, whilst in perovskite the valence of the Ce was determined essentially by the chemical design of the perovskite. Charge compensation in zirconolite and perovskite samples displaying an apparent excess of positive charge was inferred to occur by the formation of cation vacancies, in samples prepared under oxidising conditions, and by trivalent Ti being produced in samples subject to a reducing atmosphere. Oxygen vacancies are inferred to be present in samples exhibiting an apparent excess of negative charge.

Uranyl oxide hydrate phases are known to form during contact of oxide spent nuclear fuel with water under oxidizing conditions; however, less is known about the fate of fission and neutron capture products during this alteration. We describe, for the first time, evidence that neptunium can become incorporated into the uranyl secondary phase, dehydrated schoepite (UO3•0.8H2O). Based on the long-term durability of natural schoepite, the retention of neptunium in this alteration phase may be significant during spent fuel corrosion in an unsaturated geologic repository.

Zirconia containing actinides in solid solution, (Zr, An)Ox, where An = U, Pu, Am, Np, Cm is considered as a starting material for the synthesis of two durable crystalline host-phases for actinide immobilization: zircon (Zr, An)SiO4 and cubic zirconia (Zr,An)O2. The use of this starting material is necessary in order to provide for total actinide incorporation into the structures of the final phases, zircon and cubic zirconia, while avoiding the presence of unincorporated actinides. Difficulties arise when the chemically inert pure oxides, ZrO2 and AnOx are used as precursor materials. Melting techniques, which are commonly used to provide solid solution of inert oxides are not acceptable for ceramic preparation, particularly the synthesis of ceramic Pufuels. This paper presents a new approach utilizing plasma calcination. Solids containing both zirconium and actinides are calcined in a plasma. This allows the use of intermediate, stabilized actinide forms, for example, co-precipitated actinide-zirconium oxalates which have a low solubility level or a solid sol-gel precursor, such as (Zr, An)SiOx.

Calcination was completed using an induction plasma in an argon gas mredium. Two types of material were used: co-precipitated oxalates of Zr and Ce with 10 wt.% Ce and a solidified sol-gel (Zr, Ce)SiOx with 10 wt.% Ce. Weakly milled material with a particle size of 0.34 mm was dropped through the plasma at a speed of 40-60 grams per hour. The final product was examined by high resolution transmission electron microscopy (HRTEM). The calcined oxalate material consisted of aggregates of globular particles (tens to hundreds of nm) of monoclinic and tetragonal (Zr,Ce)O2. The calcined sol-gel consisted of globular particles (50 to 100s nm) of monoclinic and tetragonal (or pseudo-cubic) (Zr, Ce)O2 in an amorphous silica matrix. This paper discusses the application of plasma calcination to the immobilization of excess weapons Pu and other actinides. The plasma calcination of solid materials would be used for ceramic Pu-fuel fabrication, as well as for the final disposal in geological formations.

Zircon (ZrSiO4) is of great interest for the nuclear industry as it is one of the new crystalline waste form considered for the disposal of actinides, for example weapons' plutonium in USA. In this study the effects of displacement cascade due to α-decay has been modelled from an atomistic point of view by molecular dynamics simulation using Born-Mayer-Huggins empirical potential. The numerical values of the parameters of this potential have been fitted on structural equilibrium properties of the crystal and on atomic arrangements.

Displacement cascades are reproduced by accelerating one of the atoms of the cell, thus modelling the effect of the α-decay recoil nucleus. Kinetic energies up to 2 keV have been introduced. The unfolding of the cascades and the final structures have been studied in detail. The centre of the displacement cascade exhibits an amorphous zone where the zircon structure is completely lost. It contains an assembly of distorted SiO4 tetrahedra and disordered zirconium polyhedra. The zirconium ions (originally surrounded by 8 oxygen atoms) exhibit a decrease in their coordination number to 7 or 6 in agreement with what is observed for zirconium ions in amorphous zircon, zirconia or glasses. The size of the amorphous zone and the number of atoms displaced have been estimated for different recoil energies.

The energy stored during the cascade has been calculated. It exhibits an overall good agreement with the available experimental data at complete amorphization.

The existence of an amorphous track in our calculated cascades shows that the correct model for the amorphisation process should take into account the existence of a direct impact amorphous zone.

We have studied the damage caused by an energetic recoil nucleus on the glassy network, from a ballistic viewpoint, using molecular dynamics method. The energies of the displacement cascades performed have been limited to 6keV up to now. This energy represents approximately one tenth of the energy of a recoil nucleus emitted during ca disintegrations in a real nuclear glass. The composition studied comprises the major elements of the real glass: SiO2, B203, Na20, A1203, ZrO2, and the atomic interactions are represented by Born-Mayer-Huggins potentials with formal atomic charges completed by three-body terms. The study described here must be considered in the light of earlier studies in which the authors investigated the complete sequence of a displacement cascade, revealing atom movements occurring when the Coulomb force field is modified in the glass matrix.

As in the ballistic collision sequence, network formers and modifiers exhibit different behaviour. Network modifiers react more quickly and more violently to a modification of the Coulomb field, and relaxation of the alkali metal atoms occurs together with the relaxation of the three-dimensional network of formers. Relaxation of the alkali metals is a key factor ensuring more thorough relaxation of the network formers; the authors have previously shown that preventing relaxation of the alkali metals limits network relaxation.

A more comprehensive description of a displacement cascade can thus be proposed. It is important to distinguish between the ballistic collision sequence per se, which begins first and displaces mainly network formers and oxygen atoms by individual movements, and the network relaxation sequence due to modification of the Coulomb fields, which begins with a slight delay following the collision sequence.

Experiments of spent fuel dissolution performed in presence of iron show, for uranium as well as for some radionuclides, lower final concentrations than measured in similar experiments but in absence of iron. So far, only the final concentrations obtained in these experiments have been published. In this work, we have focused our study on the reduction of uranium (VI) due to the presence of iron and the subsequent precipitation of Uranium (IV) due to its lower solubility. This behavior has been confirmed using a Scanning Electron Microscope by observing an uranium phase precipitated on the iron surface and assuming, under the experimental conditions, that it most likely corresponds to U02(s). We have modeled this process assuming a first-order kinetic.

The release of 99Tc can be used as a reliable marker for the extent of spent oxide fuel reaction under unsaturated high-drip-rate conditions at 90°C. Evidence from leachate data and from scanning and transmission electron microscopy (SEM and TEM) examination of reacted fuel samples is presented for radionuclide release, potential reaction pathways, and the formation of alteration products. In the ATM-103 fuel, 0.03 of the total inventory of 99Tc is released in 3.7 years under unsaturated and oxidizing conditions. Two reaction pathways that have been identified from SEM are 1) through-grain dissolution with subsequent formation of uranyl alteration products, and 2) grain-boundary dissolution. The major alteration product identified by X-ray diffraction (XRD) and SEM, is Na-boltwoodite, Na[(UO2)(SiO3OH)]H20, which is formed from sodium and silicon in the water leachant.

X-ray Absorption Spectroscopy (XAS) together with x-ray diffraction methods were used in inactive analogues of the French nuclear waste glass to analyze Pd-Te precipitates and the peculiarity of the sites occupied by some structural probe elements such as Mo, Zn, Zr and Si in the glassy part of the material. The influence of the precipitates on the structure of the glassy matrix around these structural probes was investigated by comparing noble metal bearing compositions together with noble metal free glasses.

In noble metal bearing glasses, Pd is associated to Te as (Pd90 Te10) precipitates. Pd K-edge EXAFS shows the presence of Te in the Pd coordination shell and is in accordance with a non-metallic character of the Pd-Te bond. Mo K-edge EXAFS shows that Mo occurs as molybdate groups non connected to the glassy matrix, which may be related to the occasional separation of Mo-rich phases in some compositions. Zr is located in an octahedron with d(Zr-O)=2.08Å 0.01Å whereas Zn is 4-fold coordinated with d(Zn-0)=-1.95 ű 0.01Å. Beyond this well defined first coordination shell, XAS detects some degree of medium range order which gives insight on the bonding of the site to the polymeric borosilicate network. The interatomic Zn-Si and Zr-Si EXAFS-determined distances agree with a ZnO4 and Si04 tetrahedra sharing corners geometry and Zr-Si octahedra sharing corner with Si04 tetrahedra. A third shell of neighbors around Zr was evidenced in the noble metal bearing glasses and in a four oxide glass, indicating that Zr is an element useful for detecting subtle changes in the local structure of complex borosilicate glasses. Si K XANES shows modifications in the connection between Si04 tetrahedra when noble metals are present. To obtain accurate and precise structural interpretations, a direct comparison between EXAFS data and Molecular Dynamic (M.D.) calculations on simplified nuclear glass comprising six oxides has been performed. MD calculations show that the distribution of the Zr-O distances is harmonic in the k-range accessible with EXAFS in glasses.

with respect to the state of validation for source term development. Consequences of the various mechanism on mass half lives of the waste forms are calculated with analytical equations. For glass the largest uncertainty stems from the yet unclear dissolution mechanism under silica saturated conditions. Source terms based on silica solubility coupled to Si-mass transfer are probably neither conservative nor realistic. For spent fuel the largest uncertainty is in the extrapolation of radiolytic fuel oxidation for long periods of time. Considering the uncertainties involved, reaction rates cannot yet be extrapolated reliably to values much lower than the lowest reliable experimental measurements.

We have investigated certain aspects of the fundamental chemistry and materials science of the 4fand several 5f-elements in three glass matrices. Two of these matrices were high-temperature (850° and 1450°C melting points) silicate-based glasses and the third was a sol-gel glass. Optical spectroscopy was the principal investigating tool. One aspect of the work here was to ascertain the oxidation state exhibited by these elements in the different glasses, as well as the factors that control and/or may alter this state. An important finding was noting a general correlation between the oxidation states obtained in the two high-temperature glasses and those observed in the oxides of these elements. Of the twenty three f-elements considered here, only three exceptions (Ce, Am and Bk) or examples of variable behavior (Pr, Tb, U and Cf) to this correlation were noted. One explanation offered for the exceptions is based on the stability afforded by the oxides' fluorite lattice. The correlation also applied to the sol-gel glasses after they had been heated. Other oxidation states for some elements could be obtained in the sol-gel glasses prepared at 25°C. Presented here is a summary of the oxidation states observed for these elements in three glass matrices and how these states correlate with those observed in the oxides of these elements and with the relative oxidation potentials for the elements.

This study reports on the use of zirconolite-rich Synroc to demonstrate the safe immobilisation of ‘high-fired’ Pu02. The zirconolite-rich Synroc used in this study was prepared by adding 13 wt% Pu with equimolar amounts of Gd and Hf, relative to Pu, as neutron absorbers. The incorporation of the Pu and neutron absorbers has been studied microstructurally as well as by longer-term leach testing. This work has shown that the sintered ceramic can immobilise 13 wt% of Pu with almost complete incorporation of the Pu (≃ 98%) into the zirconolite phase. Durability studies have shown that under a wide range of leaching conditions there is no major separation of the Pu and neutron absorbers, with the majority of these elements either remaining in the matrix or leaching at low (< 10−4 g m−2 d−1) and comparable rates from the waste form.