Archaeological artifacts of glass, ceramic, and metal provide examples of long term durability and as such have been surveyed by the nuclear agencies of several countries as a possible guide to choices of materials for the containment of nuclear waste. In the case of metals evaluation is difficult because of the loss of many artifacts to recycling and corrosion processes, as well as by uncertainty as to the environmental history under which the remainder survived. More recently the study of ancient metallurgy has expanded to included other materials associated with metals processing.

It is suggested that an impermeable ceramic composite used in ancient metals processing installations should be reproduced and tested for its resistance to radiation damage. This material was synthesized more than two millennia ago and has a proven record of durability. These installations have had no maintenance but are intact, some still holding water.

Consideration is given to the environmental requirements of environment sensitive cracking (ESC) of ferritic steels, involving hydrogen-induced or dissolution-related mechanisms of cracking. Pitting, with associated local pH changes, may result in hydrogeninduced cracking of simple ferritic steels, as has been observed with high pressure gas pipelines in contact with ground waters, and may constitute a greater potential failure mechanism for waste containers than dissolution-related cracking with its requirement of relatively concentrated solutions. However, the stochastic nature of pit initiation, together with the distributions of crack nucleation and growth rates, suggest that a probabilistic, as opposed to a purely deterministic, approach will need to be applied to life prediction estimates for waste containers.

The selection of materials that are suitable for various high-level waste-packaging designs must reflect the need to meet requirements for long-term performance in repository environments that change with time. With this in mind, we examine how grain boundaries in materials are induced to migrate as a result of solute diffusion even at low temperatures, how the composition of the matrix material is changed significantly by this diffusion-induced grain boundary migration (DIGM), and how the changing microstructures and compositions during DIGM lead to major changes in materials performance, such as corrosion or embrittlement. Methods are discussed for prediction of the long-term behavior of materials affected by DIGM.

Current plans for disposal of low- and intermediate-level radioactive wastes (LLW and ILW) in the UK include enclosing much of the waste in carbon steel or stainless steel containers. Modelling and experimental studies that aim to characterise the period and extent of the localised corrosion in these containers are described. The period, during which localised corrosion can be initiated and sustained in the post-closure phase of the repository is estimated. The likely modes of localised corrosion are identified, based on further consideration of the environmental conditions to which the metals are exposed. Detailed research in progress includes investigation of the rate of pitting corrosion for carbon steel, and the possible occurrence of crevice corrosion of stainless steel.

The effect of pit growth on repassivation potentials (Erp) of type 316L stainless steel (SS) and alloy 825 is investigated using a decreasing potential staircase technique. The Erp decreases initially with increasing pit depth and then attains a value which is relatively independent of pit depth. The Erp also decreases with increasing potential scan rate because of the decreasing time for repassivation with decreasing potential. The Erp, is explained in terms of the effect of applied potential on changes in solution composition inside growing pits and its use is recommended as a bounding parameter for long-term prediction of localized corrosion.

The engineered barrier system (EBS) for geologic disposal of high-level waste may typically consist of the waste solid, a container, and a porous backfill. Whether the surrounding rock is part of the EBS is arguable, but the properties of the surrounding rock strongly affect the EBS performance. The EBS functions and performance issue are addressed.

As part of the U.S. Department of Energy's evaluation of site suitability for a potential high-level radioactive waste repository, long-term interactions between the engineered barrier system and the site must be determined. This requires a waste-package/engineered-system design, a description of the environment around the emplacement zone, and models that simulate operative processes describing these engineered/natural system interactions. Candidate designs are being evaluated, including a more robust, multi-barrier waste package, and a drift emplacement mode. Tools for evaluating designs and emplacement mode are the currently available waste-package/engineered-system performance assessment codes developed for the project. For assessments that support site suitability, environmental impact, or licensing decisions, more capable codes are needed. Code capability requirements are being written, and existing codes are to be evaluated against those requirements. Recommendations are being made to focus waste-package/engineered-system code-development.

In repositories for nuclear waste there are many processes that will be instrumental in damaging the canisters and releasing the nuclides. Based on experiences from studies of the performance of repositories and of an actual design, the major mechanisms influencing the integrity and performance of a repository are described and discussed. The paper addresses only conditions in crystalline rock repositories. The low water flow rate in fractures and channels plays a dominant role in limiting the interaction between water and waste. Molecular diffusion in the backfill and rock matrix, as well as in the mobile water, is an important transport process, but actually limits the exchange rate because diffusive transport is slow. Solubility limits of both waste matrix and of individual nuclides are also important. Complicating processes include alpha-radiolysis, which may change the water chemistry in the near-field. The sizes and locations of water flowpaths and damages in the canisters considerably influence the release rates. Uncertainties in data are large. Nevertheless the system is very robust in the sense that practically no reasonably conceivable assumptions or data will lead to large nuclide releases. Several natural analogues have been found to exhibit similarities with a waste repository and help to validate concepts and to increase our confidence that all major issues have been considered.

The Vault Model for assessing engineered barrier performance has been developed as part of the preparation of an Environmental Impact Statement to be presented to a Federal Environmental Assessment Review Panel reviewing the Canadian nuclear fuel waste disposal concept. The model describes the behaviour of titanium containers, radionuclide release from used fuel, and migration of radionuclides through buffer and backfill materials and into the surrounding geosphere. Vault Model simulations have shown that the release of radionuclides from the engineered barrier system is dominated by the release from the fuel-sheath gap and grain boundaries in used fuel. Sensitivity and uncertainty analyses have illustrated how releases from the vault are affected by both the uncertainty in model parameters and the assumptions made in the development of the models. It is likely that the combined effects of a number of conservatisms in the model result in the releases from the engineered barrier system being overpredicted by several orders of magnitude.

In developing a national strategy for the disposal of radioactive wastes, each country will consider isolation options that are appropriate to the types of waste that are produced. The options are developed in response to specific national regulatory requirements, and thus will focus on different aspects of performance. However, there are a number of technical concerns that are common to all programmes. The major issues concerning the behaviour of the chemical and physical barriers in the near field of a radioactive waste repository are discussed in this paper.

The description of key issues has been divided into the following categories: barrier design, barrier evolution, scientific understanding through modelling, and validation of performance. The near-field barriers are selected and designed to provide appropriate radionuclide containment and control. Factors affecting the evolution of these barriers, such as container degradation and gas generation, determine the subsequent release of radionuclides to the human environment. Modelling repository evolution is therefore an integral feature of performance assessments, and issues such as the treatment of inhomogeneities and non-equilibrium chemistry may need to be addressed. However, the use of mathematical and computer models implies a requirement for validation. The use of demonstration experiments and natural analogues builds confidence in the predictions of repository performance models, and provides a degree of validation for otherwise inaccessible timescales.

A repository for high-level radioactive waste (HLW) will not be in operation in Switzerland (or elsewhere) before the turn of the century. However, extensive investigations for disposal in specific regions or sites are ongoing and formal safety analyses have been performed in many countries. Broadly speaking, these analyses show the feasibility of the chosen options for deep geological disposal. At the present stage, before a licensing application, performance assessments have another important application: to identify further needs to improve system understanding, and to guide the necessary research activities. Performance assessments are thus indispensable tools for focussing in on research requirements and discriminating between necessary, and merely desirable or interesting, research projects.

Based on experience from assessments of HLW disposal in the crystalline of Northern Switzerland (Project Gewähr, KRISTALLIN-I) we consider in detail the chain of models resulting from a scenario analysis. For each model block (e.g., engineered barrier performance, hydrology, radionuclide transport), the adequacy of understanding is addressed and the necessary research needs pointed out. These needs cover a wide span, from a requirement for more reliable input numbers (example: long-term corrosion rate of glass) to a better understanding of important features (example: excavation-damaged-zone) and key mechanisms (example: sorption).

The IAEA initiated, in 1991, a Coordinated Research Programme (CRP), with the aim of promoting the exchange of information on the results obtained by different countries in the performance of high-level waste forms and waste packages under conditions relevant to final repository. These studies are being undertaken to obtain reliable data as input to safety assessments and environmental impact analyses, for final disposal purposes. The CRP includes studies on waste forms that are presently of interest worldwide: borosilicate glass, Synroc and spent fuel.

Ten laboratories leading in investigation of high-level waste form performance have already joined the programme. The results of their studies and plans for future research were presented at the first Research Coordination Meeting, held in Karlsruhe, Germany, in November 1991. The technical contributions concentrated on effecting an understanding of dissolution mechanisms of waste forms under simulated repository conditions. A quantitative interpretation of the chemical processes in the near field is considered a prerequisite for long-term predictions and for the formulation of a "source term" for performance assessment studies.

The near-field radionuclide migration code Tullgarn has been developed for performance assessment purposes. As a part of the PROPER-code package it has been successfully applied in the SKB 91 safety analysis.

The features and processes included in the code are:

• - Radioactive chain decay

• - Different canister failure mechanisms (copper corrosion from sulphide attack, steel corrosion, internal overpressure and initially defective canisters) - Spent fuel dissolution. The model is based on the assumption that the dissolution rate is proportional to the α-dose rate

• - Transport calculations are done with a resistance-network model. Tullgarn calculates the stationary release of radionuclides from a defect in the canister through the buffer and out into a fracture in the rock or up to the damaged zone under the deposition tunnel.

Tullgarn can be used as a stand-alone model for near-field release calculations or as a submodel in an integrated assessment. In the SKB 91 analysis, Tullgarn gave the source term to the far-field model.

A near-field performance assessment (PA) was conducted for the Saltstone Disposal Facility (SDF) at the Savannah River Site in South Carolina. The analysis was conducted in four parts: general screening calculations, degradation calculations, and flow and transport through the fractured and nonfractured facility. Modeling approaches and example sensitivity analysis results from the simulations of the fractured facility are discussed. Design considerations that may not be apparent without addressing flow and transport through fractures and lessons learned during the process are also presented.

The solubility and speciation of Pu(VI) with phosphate is being investigated to determine the ability of phosphate to act as an actinide getter. In the initial studies performed, solubility was approached from oversaturation at an initial pH = 4, 10 and 13.4. Absorption spectra were recorded, the solution filtered, and the filtrate analyzed for Pu content. Absorption spectra were obtained at varying phosphate concentrations, and at pH of 2.7 to 11.9. The effect of complexation on the 833 nm Pu(VI) band was characterized. Evidence for three phosphate complexes was obtained for pH < 10, which have absorption bands at 842, 846 and 849 nm. Evidence for colloid formation was observed, but is not conclusive. The possible presence of colloids prevented an accurate determination of true solubility. A concentration of 10−5 to 10−6M Pu(VI) was measured in solutions at pH ≤ 10 that was filtered with a 50 nm filter. Pu(VI)-phosphate complexes predominated at pH ≤ 11.6. At higher pH, however, only hydrolyzed Pu(VI) was detected. At pH = 12, the concentration of Pu(VI) was as high as 10−4M.

This paper examines Americium behaviour in CI− media at room temperature in connection with environmental and waste disposal programs. Most published values on U, Np, Pu and Am complexation in chloride media have been determined using extraction methods. Spectrophotometric techniques are not sensitive enough to prove actinide complexation by chloride, which is confirmed in this paper for Am(III).

Am(OH)3(s), AmOHCO3(s), Am2(CO3)3(s) or NaAm(CO3)2(s) solid phases can control the Am solubility, depending on the chemical conditions of the aqueous phase (usually PCO2). 241Am solubility is here found to be higher in NaCl 4M media than in NaCl 0.1 M (up to 3 orders of magnitude). Addition of a reducing agent (metallic iron) lowers the solubility. After a week, solubilities in NaCl 0.1 M and 4 M are similar. These results are consistent with Am(III) radiolytic oxidation to Am(V), due to cc radiations. Little evidence of Cl− or mixed Cl−-CO2−3 complexes is found in these conditions. In Na+-OH−-Cl− media, 241Am(III) oxidation had also been proposed. Slow kinetics of precipitation could induce experimental uncertainities.

A number of the hydroxycarboxylic acids generated by the alkaline degradation of cellulosic wastes under reducing conditions in a cementitious repository can significantly increase the solubility of the actinides at high pH, especially plutonium. The solubility of plutonium at pH 12, in the presence of a range of hydroxycarboxylic acids containing a number of hydroxyl groups and between one and three carboxylate groups, has been modelled using the HARPHRQ code. All the plutonium-organic complexes assumed in the model are based on a stable unit in which a central plutonium ion is bound by four oxygen atoms. The oxygen atoms can be provided either by a deprotonated hydroxyl group on one of the ligands, or by hydroxide ions.

Advances in safe management and disposal of radioactive waste have shown that a comprehensive program requires the incorporation of dissolved organics into radwaste and transport effluent models, with respect to their binding of radionuclides. The binding of Eu3+, Th4+ and UO22+ to a well-characterized aquatic fulvic acid has been studied using an ultrafiltration method at a bulk electrolyte concentration of 0.10 M NaClO4, trace amounts of radionuclides and fulvic acid concentrations of 60 and 120 mg/l. The results expressed as the overall complex formation function, βov, versus pH show the following order: Th4+ > Eu3+ > UO22+. The estimated βov, values have been discussed by considering the aqueous chemistry of Eu3+, Th4+ and UO22+ vis-à-vis the solution chemistry of the fulvic acid sample.

Iron exists in aerobic soil and water environments most commonly as insoluble Fe(III). Siderophores are powerful, microbially produced chelating agents that are used to mobilize the insoluble Fe(III) cation. Over 80 siderophores have been isolated and characterized, with some reportedly having iron-binding constants as high as 1052. Fe(III) and Pu(IV) are similar in their charge/ionic radius ratio (4.6 and 4.2, respectively); therefore, Pu(IV) may serve as analog to Fe(III). It is possible that some radioactive wastes could be chelated by naturally occurring siderophores, thereby altering the transport rates of those elements through the subsurface environment. This investigation wn 9 initiated to investigate that possibility. The binding of 239Pu(IV) by four chelating agents is reported in this paper: a siderophore isolated and purified from a Pseudomonas sp.; desferal, a ferrioxamine siderophore commonly used for deferration therapy; EDTA, ethylenediaminetetraacetic acid; and, citrate, trisodium salt.

This study deals with the effects of microorganisms on the transport of radionuclides under deep nuclear-waste disposal conditions. Metabolism of a cellulolytic microorganism is studied. Cellulose, as a carbon source, is representative of nuclear waste.

A pilot device allows the study of the general effect of microrganisms. Bioleaching of radionuclides by a fungal culture is performed on columns of clay and cement used as engineered barriers. Cesium and Uranium had been incorporated into matrices prior to the tests. Operating conditions have been choosen according to realistic conditions of a deep repository. The production of organic acids by microorganisms is qualitatively and quantitatively determined.

In addition, direct effects of microorganisms (biosorption, bioaccumulation) and indirect effects (complexing agents issued from the diodegradation of cellulose) on the transport and/or retardation of radionuclides are studied.