The reason for making long-term predictions of waste-form performance under repository conditions is to assess the impact of nuclear waste disposal upon the safety of future generations. Assessment of repository safety can be approached in two ways. One method is to define an acceptable release of radioactivity to the accessible environment, and to work back from that point to determine what kind of performance requirements are placed on the waste form. When this is done, it is found that the performance requirements for the waste form are not very restrictive. The waste form can dissolve completely in 100 to 10,000 years without affecting safety [1,2], because the repositories will be 300 meters or more underground. Radionuclide retardation on repository rocks and overburden and dilution during the long flow paths to the accessible environment become the controlling parameters.

The rate of radiation damage from radioactive species in nuclear waste ceramics and the long term stability of the crystalline phases may depend on the crystallographic site of the emitting species. It is therefore of importance to determine the site occupancy of the radioactive elements. Hitherto, no established crystallographic technique has been suitable for such studies due to the small grain size and the many different phases in nuclear waste ceramics, such as Synroc. By taking advantage of the fact that an electron beam can be focussed, and the electron intensity localized, to a particular crystal site by electron diffraction, a new technique capable of locating small concentrations of atoms in small crystal volumes >(10 nm)3 has recently been developed. The application of the technique to determining the site occupancy of solutes, including U, Sr, Th, in the perovskite and ulvo-spinel phases of a Synroc ceramic is described. In the perovskite phase, U, Th, Sr, Zr and Mn, are shown to substitute on the Ca site with Fe partitioning between the Ca and Ti sites.

It has been shown that between 10 and 20% of a simulated PW–4b radwaste composition can be incorporated into a single nhase with the NZP (= ‘MaZr2 P3 o12’) structure. By changing the P/Na and Zr/Na molar ratios (i.e., adjusting the crystal chemical model of where each ion is located in the structure) it has been possible to outline a very ‘forgiving’ compositional regime both at the 10% and the 20% waste loading level within which one obtains one ([NZP]) or two ([NZP] and [monazite]) phases. We have also succeeded in substituting Tio2 for Zro2 in making a TiO2-rich [NTP] waste form analogous to the [NZ]] materials.

Thus we have succeeded in creating monophasic and diphasic ceramic waste forms which can be sintered below 1000° C. Only preliminary leach data have been obtained at 25° and 300°C, and they place this material with good ceramic forms.

Studies of the leachability of waste glass have been in progress at Savannah River Laboratory (SRL) for several years. The principal objective of these studies has been to predict the long-term behavior of Savannah River Plant waste glass when stored in a repository. Such predictions can be made from the results of short-term tests only if the mechanisms of waste glass corrosion are understood. Determining the mechanisms of corrosion and developing a predictive model have therefore been a major thrust of our work.

Preferential dissolution of polyphase nuclear waste materials in short term leach tests can exaggerate radionuclide release rates when extrapolated to the lifetime of the waste form. Possible preferential leach phenomena are associated with the presence of cracks, intergranular phases and readily soluble phases. The rate of dissolution and the microstructural connectivity of the most soluble phase determine the period over which perferential dissolution is observable. The connectivity of phases is amenable to control during processing by altering the starting green density of the precursor powders.

Surface layers are a common feature of leached surfaces of borosilicate waste glasses. Layers are also observed upon weathering of volcanic glasses[l] and of silicate minerals[2]. The question of whether these layers can protect the glass against further attack by decreasing the leach rate is stïll a subject of controversy[3]. Both in geochemical work[4] and in work on waste forms [5,6], surface layers are attributed a protective function, and the stability of leached, million years old volcanic glasses[1] may be due to the presence of palagonite, a thin (≤100 μm) alteration layer, which forms in a few years but does not seem to increase in thickness after this time. The present study investigates the effects of layer formation on leaching kinetics of a borosilicate waste glass containing 20 wt.% LWR-type simulated waste oxides.

Leaching tests of PNL 76–68 glass in deionized water have been performed using the standard MCC-l static leaching procedure but with varied glass surface area to solution volume ratios (SA/V). It was found that leaching could be strongly influenced by the SA/V ratio, due largely to an effect of silicon solubility limitations. The conclusion that solubility and not solid state diffusion is most important in regulating leaching rates is supported by (1) the similarities in depth profiles of all soluble glass components with none depleted to depths greater than that of silicon despite vastly different solid state diffusivities, and (2) the lack of dependence of leaching rates on reaction layer thicknesses. To more directly examine the influence of dissolved silicon on glass leaching rates, leaching tests were performed in silicic acid solutions and in two actual groundwaters. As expected, leaching rates of all soluble glass components were reduced by amounts roughly proportional to the silicon saturation fraction. Since solubility modifies leaching rates in all but very dilute solutions, short-term tests at high SA/V values can be used to predict solution concentrations for long-term tests at low SA/V values, although reaction layers formed are not of the same thickness. Glass leaching data for a range of leaching times and SA/V values can be represented by a single curve when plotted versus the product of SA/V and time. However, the use of SA/V variations may have limited usefulness in accelerated leach testing for multicomponent systems. Events such as silica colloid and certain alteration phase formations modify the above relationship.

Concretes, grouts, clays and/or zeolites are candidate borehole, shaft or tunnel plugging materials for any nuclear waste repository. Interactions between these plugging materials may take place under mild hydrothermal conditions during the life of a repository. Class H cement or mortar (PSU/WES mixture) was reacted with one of two montmorillonites, clinoptilolite or mordenite at 100° and 200°C for different periods under a confining pressure of 30 MPa. The solid reaction products were characterized by x-ray powder diffraction and scanning electron microscopy after the hydrothermal treatments. When zeolites were in contact (not intimate mixture) with class H cement, they did not seem to alter but clinoptilolite altered to analcime, and mordenite became poorly crystalline in the presence of mortar (containing NaCl) at both 100° and 200°C. When cement or mortar was intimately mixed with zeolites or montmorillonites and reacted hydrothermally, the reaction resulted in the formation of Al substituted tobermorite (11Å type) in all cases (this type of reaction is expected at the interface) at both 100° and 200°C. The mechanism of tobermorite formation includes the decomposition of zeolites or montmorillonites in the presence of alkaline (pH ≃ 12) cement or mortar and recrystallization to form Al substituted tobermorite. Cesium sorption measurements in 0.01N CaCl2 on the reaction products revealed that selective Cs sōrption increased in most cases, even though little or none of the original zeolites and montmorillonites remained in the products. For example, Cs sorption Kd (mL/g) increased from 80 in the untreated mortar + Ca montmorillonite mixture to 1700 in the interaction product which is Al substituted tobermorite. Thus, we discover here that Al substituted tobermorite has good selectivity for Cs.

Synroc is a titanate-based ceramic developed for immobilization of high-level nuclear reactor wastes in solid form. Fluid bed Synroc production permits slurry drying, calcining and redox to be carried out in a single unit. We present results of studies from two fluid beds; the Idaho Exxon internally-heated unit and the externally-heated unit constructed at Lawrence Livermore National Laboratory (LLNL). Bed operation over a range of temperature, feed rate, fluidizing rate and redox conditions indicate that high density, uniform particle-size Synroc powders are produced which facilitate the densification step and give HUP parts with dense, well-developed phases and good leaching characteristics.

In the past years a lot of research work has been done in several countries to find a waste form capable of accommodating the wide variety of elements present in commercial high level waste (HLW) and which at the same time, has a higher leaching resistance than the currently favoured high level waste glass. The main problems arising from a product made up of different phases, each of them taking up a certain fraction of the HLW-elements into solid solution, are

1) the necessity of very intimate mixing of the HLW-elements and the inert additives because otherwise undesired highly leachable phases might be formed;

2) the detrimental effect transmutation of the radioactive elements will most probably have on the crystal structure in which they are present. Fracturing and loss of the low solubility of the products are likely to occur.

γ-zirconium phosphate (γ-ZrP), γ-Zr(HPO4)2 2H2O is an extremely selective cesium ion sieve for nuclear waste solutions. Because of its layer structure, the adsorbed Cs can be directly fixed in γ-ZrP by collapsing the layers upon heat treatment. Fixation of 96 and 98% of the adsorbed Cs was achieved by simple heat treatment at 350°C and 600°C respectively. Thus, heat treatment of this γ-ZrP leads to a good radiophase with good leach resistance. Cs sorbed γ-ZrP can also be converted to CsZr2(PO4)3 [;sodium zirconium phosphate (NZP) radiophase] by sintering with ammonium monobasic phosphate and zirconyl nitrate at about 850°C. Thus, γ-ZrP serves as a precursor to CsZr2(PO4)3 radiophase which has comparable leach resistance to other phases under consideration for radio-Cs immobilization.

Aging of simulated nuclear waste glass by contact with a humid atmosphere results in the formation of a double hydration layer penetrating into the glass and in the formation of alteration products on the glass surface. This hydration process has been studied as a function of time, temperature, glass composition and water vapor pressure. A dual stage hydration rate was observed and rate constants were determined at each temperature. An Arrhenius plot for the initial stage alteration rate indicates the reaction mechanism does not change between the temperature limits of the experiment (120-240°C). This conclusion is supported by the sequence of mineral formation on the surface. This hydration process provides a means of accelerating aging reactions while simulating conditions that may exist in a nuclear waste repository.

The current strategy for the immobilization of nuclear wastes is based upon a system of multiply redundant barriers for which geological containment is the final barrier to the migration of radionuclides in the biosphere. The ability of the respository host rock to buffer the oxygen fugacity of the pore waters is a critical concern for the evaluation of the far-field migration of selected elements; notable Tc, U and TRU elements. The buffer capacity of all proposed host rocks, with the exception of salt, is based upon the presence of both ferric and ferrous iron in the host rock phases.

The system U-Fe-O(-H) was selected for study because of its significance on the role of Fe in controlling the oxidation state of the U. It has been established that the Fe2+/Fe3+ couple is the controlling factor in limiting the uranium oxidation to the U4+ state.

The results from this experimental study at 400° C (the maximum worse case temperature for a repository) verified the existence of the compatibility triangles between hematite and UO2.00 and magnetite and between U4O9 and UO2.00 and hematite. These data indicate that in the presence both ferric and ferrous ions, the form of the uranium dioxide is retained as stoichiometric UO2.00 and not as an intermediate member of the UO2+x solid solution series or other higher oxide. These experimental results are in concert with the phase relationships predicted by Freeborn et al. (1980) based upon thermochemical calculations.

Physical properties and placement characteristics of cementitious mortars have been studied for their potential as repository sealing materials. They contained various expansive agents and industrial by-products, and were investigated at curing temperatures up to 250° C, the upper limit of an emplacement site or generally of relevance in accelerated reaction studies. An expansive agent, magnesium oxide, and two industrial by-products, silica fume and granulated blast furnace slag have been used at different percentages in the mixtures. Excellent general performance, including very high strengths up to 240 MPa combined with very low intrinsic permeability <10−8 Darcy (μm2) were generated at 175°C on material having a viscosity of 5000 cP (mPa·s) at 38° C. One 1700 cP(mPa·s) material treated at 250°C had compressive strength >180 MPa and also <10−8 Darcy (μm2) permeability. MgO was found to accelerate formation of tobermorite and generally cause expansion; at 250° C expansion was also related to xonotlite formation.

Cs dissolution kinetics for Cszr2(PO4)3 in a variety of aqueous media at 70°C approximately obeyed a first-order law. However in 0.1 M NaHCO3,at1/2 law was appropriate. The Cs extraction was a minimum under neutral conditions. In the pH range of 1–4, the Cs extraction varied approximately as [H+]0.3 and was a stronger function of pH in alkaline media. The survival of α-Zr phosphate gel in water at 300°C was essentially confirmed. ZrP2O7, reacted sluggishly with water at 200–300°C to produce an apparently fully-protonated analogue of NaZr2(PO4)3

The alkali-borosilicate (ABS) system provides the basis for a wide variety of commercially important products among which are the nuclear waste glasses. Although a large number of investigations have been undertaken in the last five years, the corrosion mechanisms of the ABS glasses have not been characterized nearly as well as for the soda-lime-silicate (NCS) glasses commonly used for containers. It is well known that the corrosion of the latter glasses involves ion exchange, network dissolution, and precipitation mechanisms resulting in the development of one of five types of surface films. In the present paper we compare the corrosion behavior to the ABS and NCS glasses and discuss our current understanding of ABS glass corrosion in terms of mechanisms, kinetics, surface film formation and thermodynamics.

Formula 127 glass has been developed to immobilize ICPP zirconia calcine. This glass has been prepared remotely on a laboratory scale basis with radioactive zirconia calcine retrieved after ten years of storage from Bin Set 2. The aqueous leachability of the glass produced was investigated and compared through application of standard leach tests with that of Formula 127 glass prepared with simulated calcine. Solid state properties of the glasses prepared with actual and simulated calcines were also compared through application of electron spectroscopy for chemical analysis (ESCA) and scanning electron microscopy (SEM) with energy dispersive X-ray (EDS).

The growth of surface layers on simulated waste glass during two different types of leaching has been studied in conjunction with their corrosion kinetics. Static and Soxhlet leach tests were performed in distilled water at a temperature of 100°C. Auger and ESCA analyses of solid samples after leaching showed that the layers consisted of two or three sublayers which were distinguished by their different components. The transition elements Fe and Ni, the rare-earths Nd and La, and Zn were concentrated in the layers, while Si, B and alkali were depleted in most of the layers. Growth kinetics of the layers followed approximately linear relations for the two types of leaching. Growth rates and elemental profiles of the layers depended upon the leaching conditions. Comparison between the leaching rate and the thickness of the layer showed that layers did not work as effective diffusion barriers until a threshold thickness was reached, which depended on the layer structures.

The leaching behavior of two alkali-borosilicate glasses containing 9 wt% simulated fission products and 1.6 wt% uranium oxide has been studied. Samples were exposed to one of eight types of leachants including doubly distilled water, simulated ground silicate water, a brine solution, and solutions containing various concentrations of iron, aluminium or sodium maintained at either 25°C, 40°C or 90°C for up to 182 days. The most aggressive leachants were the solutions containing sodium (excluding brine) and simulated ground silicate water. These solutions increased the extent of leaching by a factor of 2–3 over that for distilled water for one of the glasses. A partially protective surface film rich in magnesium potassium and chlorine was formed on the glasses exposed to the brine solution.

In order to evaluate the effects of atmosphere on leaching, samples were also immersed in doubly distilled water over which the relative concentrations of oxygen, nitrogen and carbon dioxide were varied. Increasing the carbon dioxide concentration from 0 to 50% resulted in a factor of 3 increase in the leaching rate.

For complex glasses such as simulated nuclear wastes, glass dissolution is a complex process, involving selective leaching of cations, reprecipitation reactions, and protective film formation. In order to begin to understand this complex behavior, each of the above phenomena is being studied, one reaction at a time, on simple alkali silicate glasses under controlled environmental conditions. To date, two types of reactions have been investigated. The first reaction type is the selective leaching of cations from the glass, resulting in the formation of a hydrated “gel” layer on the glass surface. The second reaction type is the reaction of ionic or colloidal species in solution with this gel layer. Reactions in the second category include ion exchange reactions and sorption reactions which can result in protective film formation. Studies of these simple reactions have led to the development of new leaching models and to observations which begin to explain the behavior of complex glasses, and predict how glass dissolution should change as a function of the chemical environment.