Durability and long-term stability of cements in plasters, mortars, and/or concretes utilized as borehole plugging and shaft sealing materials are of present concern in the national effort to isolate nuclear waste within deep geological repositories. The present study consists of an examination of selected ancient building materials and provides insights into the durability of certain ancient structures. These data were combined with knowledge obtained from the behavior of modern portland cements and natural materials to evaluate the potential for longevity of such materials in a borehole environment. Analyses were conducted by petrographic, SEM, chemical, and x-ray diffraction techniques.

Removal of the long-lived radionuclides from high-level waste (HLW) is a potential means not only for making wastes more acceptable in terms of long term hazards, but also for alleviating storage requirement. From these points, the authors are developing a method of partitioning actinides, Sr-90 and Cs-137 from HLW. A chemical flow-sheet has been constructed and experiments with actual HLW were initiated in 1982. Through the partitioning, active elements in HLW can be fractionated into 3 groups. Total volume of the solid materials of the 3 groups was calculated and found to be reduced to less than one-third of the volume of the vitrified material containing 10 wt% of fission products as oxide. Such volume reduction seems to facilitate the long term storage or the deep geological disposal of HLW.

Simulated SRL waste was prepared with compositions varying in iron and aluminum content. Two batches with similar composition were produced with different amounts of reducing agent added. Samples were isothermally heat treated and used to derive time-temperature-transformation diagrams. Supplementary samples were cooled in a manner programmed to simulate the cooling curves of production canisters. Less than 10% total devitrification occurs during normal processing. However, when waste glass 165 was purposely devitrified, un to 30 volume percent total spinel and acmite formed. Formation of these species had minor effect on leachability in MCC-1 and accelerated leach tests.

Simulated PW-4b waste solution along with Na additive was mixed with Zr-P-O gel, dried and then fired to form the desired sodium zirconium phosphate, NaZr2 (PO4) 3 [NZP] ceramic. NZP and monazite were the only phases produced upon firing at 900°C with 10 to 40% of PW-4b mixed with the gel. CsZr2(PO4) 3 which is isostructural with NZP was also identified when fired under reducing conditions. The -200 mesh powders of these waste forms prepared under reducing conditions showed excellent leach resistance under hydrothermal conditions.

Alternatively, PW-4b and Three Mile Island (TMI) wastes were adsorbed on Zr-P-O gel in a column. The gel was dried, pelletized and fired to form the desired [NZP] ceramic. Cesium was found to be selective on the Zr-P-O gel because no breakthrough of Cs was detected up to 38 column volumes of TMI waste. Thus, it is possible to use a tailored gel to sorb Cs and/or Sr from accident waste water and then fired the bed to form the [NZP] ceramic below 1000°C. The main advantages of the gel adsorption process are its simplicity and its enormous compositional flexibility.

Defense waste processing at the Savannah River Plant will include decontamination and disposal of approximately 100 million liters of soluble salts containing primarily NaNO3, NaOH, NaNO2, NaAl(OH)4, and Na2SO4. A cement-based waste form, “saltstone,” has been designed for disposal of Savannah River Plant low-level radioactive salt waste. Bulk properties of this material have been tailored with respect to salt leach rate, permeability, and compressive strength. Microstructure and mineralogy of leached and unleached specimens were characterized by SEM and X-ray diffraction analyses, respectively. It has been concluded that the salt leach rate can be limited so that amounts of salt and radionuclides in the groundwater at the perimeter of the 100 acre disposal site will not exceed EPA drinking water standards.

One of the concepts under development as a nuclear fuel waste isolation container is a thin-wall corrosion-resistant shell, supported internally by a cast metal matrix in which intact used fuel bundles are invested. The integrity of the metal matrix can be influenced by metallurgical factors and by process parameters. Finite element solidification modelling and laboratory experiments with lead as an investment material have shown the influence of heat transfer parameters on matrix integrity. Controlled cooling of the container walls, for example, can be used to reduce the interaction time between the molten matrix, the fuel sheathing and the container wall, and achieve a void-free matrix. The results of the computer simulations have been used to design an improved casting system, based on controlled wall cooling, for investing nuclear fuel waste containers. Ultrasonic testing of bonds between some candidate container and metal matrix materials, in combination with the metallurgical characterization of the interface region, has allowed differentiation between bonded and unbonded regions. Matrix cracking near bonded interfaces was identified as a potential problem, which could limit the use of the ultrasonic scanning technique for matrix inspection. To produce a high quality interface with good chemical bonding, induction skin melting looks promising and is being further evaluated.

The paper presents a method for the immobilization of different waste streams based in the uniaxial in-can hot pressing of a glass frit together with the waste oxides.

The process reported herein offers engineering simplicity in combination with many technological advantages such as significant decrease in temperature, use of small pressure, improved homogeneity of the products, production in easily arrangeable and interchangeable units. Furthermore, the proposed method has also proved to be suitable for treating intermediatelevel waste streams containing Zry-4 cladding fragments.

The distribution of glass constituents within the hydrated layer is investigated through a mathematical representation of the fundamental processes that influence mass transport. A brief description of the models that are developed is presented. The resulting equations are implemented in a computer code named GELOH. The importance of the processes of network hydration and layer growth, alteration product formation, and dissolution is studied through the use of GELOH. The results of these studies are presented and interpreted in the light of the experimental evidence available on the aqueous corrosion of nuclear waste glass.

A simple mathematical model has been set up assuming that the glass leaching is defined by the diffusion and adsorption of the SiO2 in the porous media. It is also assumed that at the contact with the glass, the SiO2 content in the leachant reaches the saturation limit. Experimental weight losses seem to be in good agreement with the calculated values. In order to pass from the leaching rate of the glass to the release of radioactivity, the different radionuclides are subdivided into three different classes following their leaching rate and saturation limit. These three classes are discussed separately. An experiment which allows to measure leaching rate and diffusion of the various elements has been set up. Preliminary results are presented and discussed.

In the selection of a repository for burying vitrified nuclear waste, the leach rate of the waste form in groundwater is of concern. Basaltic groundwaters have inherently low redox potentials which may affect the waste form leach rate. Laboratory simulation of the Eh-pH conditions to be found in a basalt repository can be achieved when crushed basalt is added to deoxygenated deionized water. The effects of other redoxactive waste package components, such as iron, were found to stabilize solution redox potential at different valves under oxic and anoxic conditions. When iron was present, different waste form leach rates were observed with and without oxygen.

Two waste glasses, MCC 76-68 and C31-3, were leached in deionized water and 0.001 M MgCl2 for periods up to 158 days. At 57 days the gel layer was removed from some of the specimens and leaching continued for up to 100 days. Results from leaching in deionized water showed that the gel layer was not protective. Results from leaching in 0.001 M MgCl2 are in good agreement with the predicted results obtained from the use of the PHREEQE geochemical code and with sepiolite [Mg2Si306(OH)4] as the Mg-bearing precipitate. Both B and Si were predicted and observed to increase with increasing glass dissolution while maintaining sepiolite solubility. Both MCC 76-68 and C31-3 glasses showed increased leaching in 0.001 M MgCl2 upon removal of the layer. This suggests that the leaching mechanism is a function of the difference in the silica chemical potential between the glass and the gel layer.

Three leaching tests, devised to distinguish among several proposed nuclear waste glass leaching mechanisms, were carried out for four different waste glasses. In the first test, the influence of a pre-formed reaction layer on elemental release was evaluated. In the second test, glass specimens were replaced with fresh samples halfway through the leaching experiment, to evaluate the influence of the concentration of glass components in leaching. Finally, regular replacement of the leachant at fixed time intervals essentially removed the variable of changing solution concentration, and allowed an assessment of the influence of reaction layer thickness on the leaching rate. Results for all glasses tested indicated that the reaction layer presented little or no barrier to leaching, and that most of the retardation on leaching rates generally observed are attributable to saturation effects.

Leach tests carried out on SRL TDS-131 Defense Waste Class indicate that at high flow rates the controlling mechanism is simple corrosion. The matrix elements (Si, Al) are leached out at rates similar to those of the leaching of the alkalis and of boron, and the leaching process is nearly linear with time. At slow flow rates (below 1 m/yr) leaching becomes controlled by the build-up of a protective layer. Al and most of the Si remain in the leached surface layer. The leach rates decrease in the course of the test before leveling off at constant values which are almost inversely proportional to the contact time, indicating that leachate concentrations have become solubility-limited. The low concentrations observed at this stage indicate the formation of alteration products.

The immobilization of 99Tc in both ceramic and borosilicate glass waste forms and the mechanism of dissolution of technetium from these waste forms are complicated by the range of technetium valence states. Metallic Tc,ionic Tc4+ oxides, and volatile molecular species containing Tc7+ all occur within the stability field of water. Tc4+ will substitute readily for Ti4+ in titanatebased ceramics in such phases as perovskite (CaTiO3) and spinel (Mg2TiO4). Tc4+ can be homogeneously distributed in borosilicate glass by re-melting the glass under closed system conditions to prevent the re-oxidation of technetium. The dissolution of technetium from both glass and ceramic waste forms follow similar kinetic behavior although the dissolution rate from the titanate is two orders of magnitude lower. Dissolution under highly reducing conditions is too slow to measure on the time scale of the experiments.

The chemical durability of SYNROC C has been measured as a function of repository conditions, namely, temperature, surface/volume ratio, groundwater pH and composition. The leach rate of SYNROC C is less dependent on temperature than borosilicate glasses; activation energies range from 15-25 kJ mol-1 compared to 33-74 kJ mol-1 for glass. The rate of release from SYNROC C is very low and independent of solution composition for the solutions studied. There is little variation in leach rate over the pH range 4-9, but at pH 2 the rate is up to 10 times higher. SYNROC C leach rate is independent of flow rate or surface/volume ratio for the mobile waste elements because of the unimportance of solution saturation effects owing to the low concentration of dissolved solids. The differential leach rate decreases rapidly with time because, after initial leaching from grain boundaries and metastable minor phases, the highly insoluble matrix protects the more leachable elements from further attack.

A joint research project with participation from Japan, Switzerland and Sweden is underway at Studsvik (The JSS-project). The project concerns investigations on the leaching of fully radioactive glass (containing 12 wt% fission product oxides and actinide oxides) manufactured by CEA/Marcoule.

So far the glass has been leached in doubly distilled water and in silicate water at 90°C. Some leaching experiments involved the presence of crushed Stripa granite in the same containers as the glass.

Due to strong radiation and the presence of plutonium the leaching was carried out in a specially designed lead cave using gilded stainless steel containers.

Weight losses, pH and elemental mass losses were determined together with infrared reflection spectral changes. These data are compared to those obtained from a simulated nonradioactive glass of nearly the same composition and to similar alkali borosilicate glasses previously investigated at Studsvik.

Experimental radionuclide migration diffusion cell data have been collected as part of the WIPP Waste Package Performance Program. This data was collected under conditions approximating geologic isolation of a backfill barrier in a salt repository. The experiments are designed to aid in the evaluation of engineered backfill barriers.

This paper describes a radionuclide transport model designed to aid interpreting experimental diffusion cell migration data and eventually to simulate the long-term effectiveness of the backfill barrier in a salt repository. The model is designed to test a variety of expressions representative of potential mechanisms for retardation within the backfill for the best-fit with experimental data. From the comparison, the aim is to select the appropriate mechanism from the host of potential mechanisms for retardation. The model employs a novel integral equation approach to the solution of the transport equation with nonlinear retardation terms. The solution technique used in this model is a semi-analytical, iterative method for the general nonlinear problem. It is felt the technique offers improved computational efficiency over comparable finite difference methods.

Comparisons between experimental migration diffusion cell data and the model predictions are presented in this paper. Tentative conclusions concerning the importance of the retardation mechanism to radionuclide transport in the backfill barrier will be drawn.

The effects of the presence of various metals and metal oxides on the leaching of nuclear waste glasses in water can be considerable even under rapid flow conditions. The systems studied here include SRL TDS-131 glass in the presence of lead and stainless steel, and PNL 76-68 glass in the presence of aluminum, alumina and their mixtures. Lead specifically suppresses dissolved silica, but does not inhibit dissolution of defense waste glass as a whole. Stainless steel specifically reduces the levels of dissolved Ca and Sr. Al and Al2 O3 exercise a general reducing effect on the rates of dissolution of commercial waste glass and the presence of alumina may have a beneficial role in enhancing waste package durability.

In order to identify reactions which can occur in systems containing nuclear waste forms, cement, and repository rock in the presence of water, activity diagrams were calculated from free energies for aluminosilicates and calcium silicates. Log (calcium activity)/(hydrogen activity)2 was plotted against log (H4 SiO4 activity) for reactions among calcium aluminum silicates for which free energies are available in the literature. Similar diagrams were made for sodium/hydrogen activity, and potassium/hydrogen activity. For calcium an activity diagram was constructed for the silicates.

Groundwater compositions from candidate repository sites in the Palo Duro Basin of Texas, the Delaware Basin of New Mexico, and the Nevada Test Site were plotted on these diagrams. Essentially all of these are shown to be in the calcium zeolite field as shown on the diagram for calcium in the absence of other cations. When considerable Mg and Fe are present in the water, other phases may replace the calcium zeolite in the pertinent region of the diagram. Chlorite is shown to be stable in this region at the Mg and pH level of the Ogallala if the chlorite is high in iron, and at the Mg and pH level of the Wolfcamp low- or high-Fe chlorites are stable. Potassium and sodium mineralrelationships fall in two categories, dilute waters and saline waters. Of the dilute waters, that from the Ogallala aquifer in the Palo Duro Basin, and most of the Nevada Test Site waters, including that from the proposed radwaste repository horizon, are in equilibrium with Na-beidellite and kaolinite. Boreholes at Yucca Flat and Mercury Valley at the Nevada Test Site, and shallow ground water from the Rolling Plains north and east of the Palo Duro Basin are in equilibrium with kaolinite. Of the saline waters, that from the Castile (deep in the Delaware Basin) is a kaolinite water. The brines from the Salado and Rustler formations are in equilibrium with kaolinite and possibly also with sodium-potassium zeolite and illite.

Leachates of cement and water, and cement, waste, and water were plotted on the calcium silicate activity diagram. These solutions are in equilibrium with calcium silicate hydrate hydrolysis reactions, with grossular and possibly with Ca-zeolites. Among the calcium silicates, calciumsilicate-hydrate gel (C-S-H gel) and tobermorite are the most likely candidates, but the thermodynamic data are not adequate to distinguish all the possibilities. The underlying assumptions are discussed.

A model, based on the relative volumes of surface and interstitial water in soil, has been developed to study the effective mass diffusion coefficient in bentonite and mixtures of bentonite and sand. The results show that, when the surface water layer is fully developed to a thickness of 0.25 nm, the theoretical effective mass diffusion coefficient decreases with increasing dry density andbentonite content in a mixture of bentonite and sand. The development of surface water depends on the initial water content, the conversion of interstitial water to surface water, and the dry density. When the surface water layer is not fully developed, the effective mass diffusion coefficientdoes not necessarily decrease as the dry density increases.

Considering the magnitude of measured mass diffusion coefficients in bentonite compacted to a dry density of 1.75 Mg/m], the model indicates that in this system there is significant conversion of interstitial to surface water. This suggests that mass diffusion coefficients in compacted bentonite are mainly influenced by the relative volumes of interstitial and surface water.