Samples of simulated vitrified radioactive waste have been prepared containing, separately, neptunium, plutonium and technetium. These glasses have been leached in static granite groundwater for 150 days. The resulting solutions have been characterised with respect to both the ionic and colloidal species: these solutions have then been contacted with concrete and granite to produce initial measurements of the source term from a typical repository. The results show that (a) the colloidal species arise from detached particles of the so-called gel layer on the surface of the glass, (b) e solution activities were 590 Bq.ml-1 in the case of 99Tc and 20 Bq.ml-1 for 237Np, and (c) the sorption coefficients wIre 1.4 and > 300 ml.g for concrete and 0.4 and 8.3 ml.g-1 for granite respectively. The technetium was only weakly sorbed and the neptunium strongly sorbed in each case.

Segments of fuel and clad have been leached in deionized water and in groundwater. The leachants were centrifuged through membrane filters. Both centrifugate and the filters were analysed for U, Sr-90, α- and γ-emitters. The results are discussed in terms of preferential leaching, solubility limitations and adsorption effects. For U an apparent saturation at about 800 ppb was observed. Pu also appeared to attain saturation at a few ppb. For Sr the leach rate was 3·10-7 /d after ca 400 days. Attempts to impose reducing conditions showed decreased leach rates.

Leaching experiments on borosilicate glass frit and simulated nuclear waste glasses were performed as a preliminary to leaching experiments on glasses incorporating radioactive waste. The experimental design included (i) simulated waste glass, (ii) ASTM Grade-2 titanium container material, (iii) clay buffer material, (iv) Standard Canadian Shield Saline Solution, and (v) granitic rock.

Cumulative fractions of release for boron were determined, as well as the solution concentrations of silicon, iron, strontium and cesium. The leach rates for boron after 28 d were approximately 5x10-6 kg.m-2 .8-1 for experiments in stainless steel vessels, and 2xl0-7 kg.m-2 .s-1 in Hastelloy* vessels. There is an apparently strong relationship between the clay/groundwater ratio, the concentration of iron in the solution, and the concentrations of silicon, strontium, and cesium.

A finite-element simulation has been performed to study the effects on the radionuclide flux, of the failure characteristics of a nuclear fuel disposal waste container, the buffer configuration, and the backfill properties, assuming an intact borehole rock wall. The results show that the total integrated flux from a failed container, through the buffer to a backfill with a diffusion coefficient and hydraulic conductivity higher than that of the buffer, increases as the radial buffer thickness increases, and the vertical buffer thickness decreases. The total flux can be estimated by a one-dimensional diffusion equation. If the backfill has the same diffusion coefficient and hydraulic conductivity as the buffer, the total flux will be reduced by approximately 30% compared with the situation when crushed rock or sand is used as backfill. For localized failure of the waste container, the total flux is mainly dependent on the failure location regardless of the area of the failed surface.

Groundwater from boreholes in granitic rock at six different sites in Sweden has been sampled and analyzed. Sensitive parameters such as redox potential, pH sulphide– and oxygen content have been measured with a field equipment. This is an integrated part of a program of geological, geophysical, geochemical and hydrogeological investigations with the final aim to select a suitable site for a high-level radioactive waste repository. According to present results deep granitic groundwaters are reducing due to the presence of iron(II) ions. The pH is normally ranging from 7 to 9. The total sulphide content is generally less than 0.5 mg/l. The normal alkalinity range is 90-275 mg/l. Copper, which has been suggested as canister material in the present concept for spent nuclear fuel disposal, is stable in this groundwater environment, except for a very slow sulphide corrosion and a limited initial attack of oxygen from the emplacement operations. The dissolution of the spent nuclear fuel matrix, UO2, is ultimately controlled by the total carbonate concentration. Carbonate content, pH and redox conditions will also be decisive for the mobilities of actinides and technetium. The conditions are generally favouring a high degree of retention for these species in the undisturbed deep groundwater rock environment.

The solubility and speciation of U, Np, Pu, Am and Tc in deep groundwaters from crystalline rocks have been calculated. The solution chemistry of these elements is largely determined by the redox conditions (Eh = 0.24 − 0.06pH ± 0.06 V, based on measurements of Fe(II)), pH (usually 7 − 9) and the presence of carbonate (log [CO3 2-] from (pH − 14) to (0.76pH − 10.83), based on measurements of HCO3 -). The oxidation states U(IV)(+VI), Np(IV), Pu(III)(+IV), Am(III) and Tc(IV) dominate, with the tetravalent oxides (U, Np, Pu, Tc) and Am2 (CO3)3 (s) as solubility limiting phases. Maximum concentrations in the range 10-9 − 10-10 M or below would be expected for U, Np and Tc under reducing conditions, somewhat higher for Pu and still higher (10-6 − 10-7 M) for Am.

The fission track method was used to map 235U distribution in thin sections made from twenty-one rock samples. Uranium locations within the studied samples include: veins and veinlets, microcracks, secondary minerals sealing microcracks, rock matrix, grain boundaries and cleavages, and accessory minerals. Two traverses show evidence for nuclide movement by diffusion (−5 and −1 cm, respectively) away from a vein and a veinlet and two other traverses show limited movement along a few microcracks. Three traverses show evidence for no uranium movement. One shows a complex combination of perhaps both bulk diffusion and microcrack transport. Apparent diffusivity (Da) values derived for the rock matrix (10-19 to 10-16 [m2/sec]) are several orders of magnitude less, and distribution constant (Kd) values (1 to 102 [m3/kg]) are several orders of magnitude greater than predicted by laboratory experiments. Values for chlorite phases within the rock matrix were Da = 10-22 to 10-19 [m2/sec] and Kd = 101 to 104 [m3/kg].

An in situ experiment to study migration in a single fracture in granitic rock has been performed. Both sorbing and nonsorbing tracers have been used. Results from a tracer run with a nonsorbing tracer (Uranine) are presented. The breakthrough curves have been explored with a model which takes into account hydrodynamic dispersion as well as diffusion into the rock matrix. Diffusivities and porosities of the rock matrix measured in the laboratory are used to interpret the field data. A short description of the experimental design is given. This investigation is a part of the STRIPA PROJECT phase I.

A migration experiment in the rock matrix is presented. The experiment has been carried out in “undisturbed” rock, that is in rock under its natural stress environment. Since the experiment was performed at the 360 m-level (in the Stripa mine), the rock was subject to nearly the same conditions as the rock surrounding a nuclear waste repository as proposed in the Swedish concept (KBS).

The results show that all three tracers (Cr-EDTA, Uranine and 1-) have passed the zone disturbed by the presence of the injection hole and migrated some distance into “undisturbed” rock.

These results indicate the existence of a connected micro fissure system in undisturbed rock, in which tracers (and therefore radionuclides) can migrate. Diffusivities obtained in this experiment are comparable to those obtained in laboratory experiments.

Hydraulic properties of crystalline rock from four potential repository sites in Sweden were analysed. The hydraulic conductivity of the bedrock was established by means of transient water-injection tests with constant head conducted in 25 m sections. The bedrock at the sites was divided into different hydraulic units. An effective hydraulic conductivity was calculated for the rock mass at each site. This was done on the basis of the frequency distribution of all measured values within this unit. A log-nornal distribution was found to fit the data reasonably well. Regression analysis of hydraulic conductivity as a function of depth indicated similar relationships between the four sites. At a depth of 500 m the effective hydraulic conductivity for three-dimensional flow was about 5.10-11 m/s.

The fracture frequency of the sites was established from existing core-logs. At depths of about 500 m the mean fracture frequency of the rock mass at the four sites was 1.8-2.5 fractures per meter. Of this total fracture frequency only a a certain proportion is considered to be hydraulically conductive. This proportion was established from results of hydraulic tests perforned in 2 or 3 m sections. Results obtained indicated a frequency of hydraulically conductive fractures of 0.1-0.3 fractures per meter in the rock mass at depths below 300 m.

The basic models for predicting the transport of radionuclides in the near field and in the geosphere (far field), used in the Swedish KBS-3 work, are presented. The transport of corrosive substances to the canister is also accounted for.

Good afternoon Ladies and Gentlemen. In leading off this afternoon's session, I would like to establish a framework against which the papers that follow can be referenced. My discussion of the proposed tuff site at Yucca Mountain shall focus on:

Research at Yucca Mountain in southern Nevada (Fig. 1), is being supported by the US Department of Energy to evaluate this site as a possible high-level radioactive waste repository. Yucca Mountain is underlain by a thick sequence of ash-flow and bedded tuffs, with a few silicic lavas.Variations in mode of tuff emplacement and postemplacement alterations have given rise to pyroclastic rocks of quite variable character, ranging from nonwelded to densely welded, vitric to devitrified, and nonzeolitized to completely zeolitized. The proposed repository horizon is in the lower portion of the thick, densely welded Topopah Spring Member of the Paintbrush Tuff in the unsaturated zone. Within the Topopah Spring Member and in the rocks beneath the proposed repository horizon, there are significant variations in mineralogy [1]. Such changes in mineralogy include the localized occurrence of such potentially reactive phases as cristobalite,tridymite, smectite, and volcanic glass. The important sorptive minerals clinoptilolite and mordenite also occur in discrete horizons, and their distribution changes horizontally and vertically. We have undertaken a study of the mineralogy in Yucca Mountain as a function of depth and lateral position to predict the horizontal and vertical distribution of these important potentially reactive and sorptive minerals. This knowledge has aided in locating the repository horizon and will help to put bounds on mineralogic variability within the repository horizon. In addition, studies of the distribution of minerals in Yucca Mountain allow us to deduce the factors that have controlled mineral distributions and to predict mineral assemblages along transport pathways [1,2]. In addition, we are investigating the groundwater chemistry because it and mineralogy are used as input to codes for calculating the transport rate of waste elements from the repository to the accessible environment.

A conceptual design is being developed in support of the Department of Energy's study of a potential site for a nuclear waste repository at Yucca Mountain. This design can be used to support the recommendation of that site for characterization if so designated by the Department of Energy. In the event that the Yucca Mountain site is chosen for development as the first repository, the conceptual design and subsequent preliminary engineering designs could be used to develop design specifications for repository construction. The design addresses the unique surface and subsurface characteristics of the volcanic tuffs at Yucca Mountain. A candidate horizon has been identified approximately 300 meters below the surface and about 100 meters above the water table.

The design activities are being conducted in a manner to attempt to control and reduce to the extent practicable the impact exerted by the repository on the containment and isolation capability of the site. This process consists of careful iteration of design, analysis and interpretation of results to ensure long-term containment and isolation requirements are met and that short-term operations and costs are acceptable.

Experiments have been conducted as part of the Nevada Nuclear Waste Storage Investigations Project to determine the changes in water chemistry due to reaction of the Topopah Spring tuff with natural groundwater at temperatures up to 150°C. The reaction extent has been investigated as a function of rock-to-water ratio, temperature, reaction time, physical state of the samples, and geographic location of the samples within the tuff unit. Results of these experiments will be used to provide information on the water chemistry to be expected if a high level waste repository were to be constructed in the Topopah Spring tuff.

Corrosion rates and the mode of corrosion attack form a most important basis for selection of canister materials and design of a nuclear waste package. Type 304L stainless steel was selected as the reference material for canister fabrication because of its generally excellent corrosion resistance in water, steam and air. However, 304L may be susceptible to localized and stress-assisted forms of corrosion under certain conditions. Alternative alloys are also investigated; these alloys were chosen because of their improved resistance to these forms of corrosion. The fabrication and welding processes, as well as the glass pouring operation for defense and commercial high-level wastes, may influence the susceptibility of the canister to localized and stress forms of corrosion.

A waste form testing program has been developed to ensure that the release rate of radionuclides from the engineered barrier system will meet NRC and EPA regulatory requirements. Waste form performance testing will be done under unsaturated, low water availability conditions which represent the expected repository conditions. Testing will also be done under conditions of total immersion of the waste form in repository-type water to cover the possibility that localized portions of the repository might contain standing water. Testing of reprocessed waste forms for CHLW and DHLW will use reaction vessels fabricated from Topopah Spring tuff. Chemical elements which are expected to show the highest release rates in the mildly oxidizing environment of the Topopah Spring tuff horizon at Yucca Mountain are Np and Tc. To determine the effect of residual canister material and of corrosion products from the canister/overpack, waste form testing will be done in the presence of these materials. The release rate of all radionuclides which are subject to NRC and EPA regulations will be measured, and the interactive effects of the released radionuclides and the rock reaction vessels will be determined. The testing program for spent fuel will determine the release rate from bare spent fuel pellets and from Zircaloy clad spent fuel where the cladding contains minor defects. A metal testing program for Zircaloy will establish the expected lifetime of the cladding material. Estimation of the state of cladding for fuel presently in reactor pool storage will provide baseline data for Zircaloy containment credit.

The most likely mechanism for release of radionuclides from a repository is transport in water from the repository to the accessible environment. The processes that can retard the movement of radionuclides in water are sorption, precipitation, and diffusion. These processes can occur anywhere along flow paths from the repository to the accessible environment.

The Nevada Nuclear Waste Storage Investigations (NNWSI) project includes a Performance Assessment task to evaluate the containment and isolation potential for a nuclear waste repository at Yucca Mountain in southern Nevada. This task includes calculations of the rates and concentrations at which radionuclides might be released and transported from the repository and will predict their consequences if they enter the human environment. Among the major tasks required for these calculations will be the development of models for water flow and nuclide transport under unsaturated conditions and in fractured hard rock. The program must also quantify the uncertainties associated with the results of the calculations. The performance assessment will provide evaluations needed for making major decisions as the U. S. Department of Energy seeks a site for a repository. An evaluation will be part of the environmental assessments prepared to accompany the potential nomination of the site. If the Yucca mountain site is selected for characterization and development as a repository, the assessments will be required for an environmental impact statement, a safety analysis report, and other documents.

This program has been divided into five tasks. Collectively they will provide the performance assessments needed for the NNWSI Project.

Preliminary Exploratory Shaft design studies have been performed for representative sites in all salt basins within the NWTS Program. These studies have been based on data which are not site specific. Earth materials overlying the Richton Dome were characterized by analysis of geotechnical and hydrological data that had been acquired for site selection purposes. Data sets were reorganized, reinterpreted and evaluated in light of published empirical correlations, known constituative relations, experience with other sites, and engineering judgment.

Geotechnical properties were assessed from geophysical logs, lithologic sample descriptions, and limited blowcount, grain size and pump test data. These properties included grain size, plasticity, unit weight, moisture content, bulk density, porosity, shear strength, elasticity, permeability, and saturation. Additionally, chemical and thermal properties were estimated and the local hydrologic flow properties were addressed.

The analyses allowed heretofor unrecognized lithologic material groupings (defineable layers and sublayers) to be identified based on similarities in physical properties. Subsurface conditions, as interpreted, pose no unique excavation problems. However, the analysis identified some potential issues which had not been previously recognized and gave confidence that other previously assumed potential problems may not exist. Future work for subsequent design should focus on shaft wall stability, ground-water inflow, expansive clays, and the impacts of long-term exposure of the materials.