The colloid growth due to irradiation in a rock salt formation is calculated with an extended version of the Jain-Lidiard model. The extensions of the model comprise a description of the nucleation stage of the colloids and the role of impurities on the formation of defect centers. Results of model calculations are shown for a representative design for a high-level radioactive waste repository in a rock salt formation. It is concluded that it is unlikely that, near the waste containers, the fraction of NaCl that will be converted to metallic Na and molecular Cl centers will exceed a few mole percent.

The present study focused on the retention capability of the different local backfill materials and on horizontal and vertical radionuclide migration in simulated repository conditions of a saturated static humid environment, using single or combined components of the near-field. The results obtained from semi-field experiment show that no migration of cesium radionuclide was detected outside the backfill zone within the time interval of the experiment. This reflects the possible efficiency, of the backfill materials used, for the confinement of radioactivity to the disposal site. On the other hand laboratory experiments show the effect of simulated repository condition on the sorption and desorption properties of backfill materials. It is clear from the results obtained that the presence of cement waste forms in equilibrium with underground water affect the retention capability of the backfill materials. The motivation of the work was a desire to provide a basis for minimizing radioactive waste processing by improving nonradioactive engineering barriers.

Carbon steel is one of the candidate overpack materials for high-level waste disposal. The corrosion rate of carbon steel is reduced by the presence of buffer materials such as bentonite and seems to be affected by the diffusion of corrosive materials and corrosion products through the buffer material.

The apparent diffusivities of corrosion product of iron were measured in some bentonite specimens in contact with carbon steel. The apparent diffusivities of iron were also measured without carbon steel for comparison. The apparent diffusivities of corrosion product were on the order of 10−12 m2/s and showed a tendency to decrease with increasing density of the bentonite specimen. There was no significant effect of silica sand on the apparent diffusivities. The apparent diffusivities of iron in the system without carbon steel were in the range of 10−14 m2/s and showed a tendency to increase with increasing silica sand content. The difference of the diffusivities between corrosion product and iron without carbon steel seems to be due to the difference of diffusing species. The color of the corrosion product was dark-green during contact with bentonite specimens and became red on exposure to air in a few minutes. Gas bubbles were also observed in the corrosion product. This suggests hydrogen generation during corrosion of the carbon steel. Thus the diffusing species seems to be in a reduced state, probably ferrous ion. On the other hand, the diffusing species of iron without carbon steel was probably a ferric hydroxide complex that was negatively charged. This suggests that ferrous ion could diffuse in the surface water adsorbed on bentonite, while ferric complex was excluded.

Shaft sealing in a high-level waste(HLW) disposal system functions to minimize the water flow passage, and retard the radionuclide transport from the repository to the accessible environment. It is important to estimate the radionuclide migration along the sealed shaft from the viewpoint of the design and the performance assessment of the sealing system.

This study presents the results of sensitivity analyses on the radionuclide migration in the vicinity of the access shaft of a repository in order to evaluate the effects of the length of a plug, as well as the number of plugs, and curtain grouts.

In this study, the upward hydraulic gradient of the groundwater flow along shafts was used, based on transient coupled thermo-hydraulic analyses around a repository. Hydraulic conductivities of the backfill material and the disturbed zones around the shaft tunnels were also assumed to be one order and two orders of magnitude higher than that of the host rock, respectively.

The results show that the velocity of the groundwater within the shaft and the disturbed zone is reduced by a factor of one third by installing a few plugs into the shaft filled with backfill material. The curtain grouts have the effect of retarding the radionuclide migration from the repository to the ground surface at a factor of approximately five. A few plug installations have the same effect. The sealing system properly constituted with backfill, plugs, and grouts can provide the same performance as the original host rock.

Multi-year, simulated remote-handled transuranic waste (RH TRU, nonradioactive) experiments are being conducted underground in the Waste Isolation Pilot Plant (WIPP) facility. These experiments involve the near-reference (thermal and geometrical) testing of eight full-size RH TRU test containers emplaced into horizontal, unlined rock salt boreholes. Half of the test emplacements are partially filled with bentonite/silica-sand backfill material. All test containers were electrically heated at about 115 W/each for three years, then raised to about 300 W/each for the remaining time. Each test borehole was instrumented with a selection of remote-reading thermocouples, pressure gages, borehole vertical-closure gages, and vertical and horizontal borehole-diameter closure gages. Each test emplacements was also periodically opened for visual inspections of brine intrusions and any interactions with waste package materials, materials sampling, manual closure measurements, and observations of borehole changes. Effects of heat on borehole closure rates and near-field materials (metals, backfill, rock salt, and intruding brine) interactions were closely monitored as a function of time. This paper summarizes results for the first five years of in situ test operation with supporting instrumentation and laboratory data and interpretations. Some details of RH TRU waste package materials, designs, and assorted underground test observations are also discussed. Based on the results, the tested RH TRU waste packages, materials, and emplacement geometry in unlined salt boreholes appear to be quite adequate for initial WIPP repository-phase operations.

Radioactive disequilibria, between U-238, U-234, and Th-230, in crystalline rock adjacent to a fracture, indicates mass transfer of U and Th between water in the fracture and the rock. The matrix diffusion theory was used to interpret the observed profiles of mobilized nuclides around the natural fracture. The interpretation of the profiles was based on the use of uranium series disequilibrium code, URSE, and migration code FTRANS. The model system was characterized using all available site-specific data, and the system evolution was outlined using the geology of the Palmottu site. The simulated concentration profiles, as a function of depth from the fracture surface, indicate that measured profiles can be modelled by matrix diffusion, assuming realistic initial and boundary conditions and diffusion times of 300,000 years.

The redox-potential, pH, chemical composition of fracture waters, and uraninite alteration associated with the Palmottu uranium mineralization, have been studied. The data have been interpreted by means of thermodynamic calculations. The results indicate equilibrium between uraninite, ferric hydroxide and groundwater in the bedrock of the study site. Partially oxidized uraninite (UO2.33) and ferric hydroxide are in equilibrium with the fresh, slightly acidic and oxidized water type, while primary uraninite is stable with deeper waters that have a higher pH and lower Eh. Measured Eh-pH values of groundwater cluster within a relatively narrow range indicating buffering by heterogenous redox-processes. A good consistency between measured Eh and analyzed uranium oxidation states was observed.

The Nopal I uranium deposit at Pefla Blanca, Mexico is being studied as a natural analog of the proposed high-level nuclear waste repository at Yucca Mountain. Identification of secondary uranium phases at Nopal I, and the sequence of their formation after uraninite oxidation, provides insight into the source term for uranium, and suggests that uranophane may control uranium release and transport in a silicic, tuffaceous, chemically oxidizing, and hydrologically unsaturated environment. Possible constraints on contaminant transport at Nopal I are derived from the spatial distribution of uranium and from measurements of 238U decay-series isotopes. The analyses indicate that flow of U-bearing fluids was influenced strongly by fracture density, but that the flow of these fluids was not restricted to fractures. Gamma spectroscopic measurements of 238U decay-series isotopes indicate secular equilibrium, which suggests undetectable U transport under present conditions.

The oxidative alteration of uraninite by alkali-bearing, hydrothermal (200-400° C) solutions in granitic rocks produces the rare mineral clarkeite. The general formula for Na-clarkeite is

{(Na,K)6-y-z(M2+PM3+qM4+r)y(□1+z-xPbx)}[UO2)7-xOt0]·nH2O. z ≈ p+2q+3r.

The terms in square brackets designate the sheet structure; the other elements and vacancies (0) occur in interlayer sites. Clarkeite can accommodate Ca, Sr, Ba, Y, Th, and lanthanides. Thus, clarkeite is a potential actinide and fission product host. The replacement of uraninite by clarkeite occurs in the solid state and results in the loss of one-half of the uranium from uraninite. As U decays to Pb, the radiogenic Pb enters interlayer vacancies in clarkeite. This destabilizes the structure and clarkeite eventually decomposes to wölsendorfite, (Pb,Ca)2U2O7·2H2O, or curite, Pb3U8O27·3H2O.

Clarkeite from Spruce Pine, North Carolina is zoned compositionally, with a K-rich core (atomic ratio K:Na:Ca = 1.0 : 0.11 : 0.12) surrounded by Na-clarkeite (Na:K:Ca = 1.0 : 0.03 : 0.03), rimmed and veined by Ca-clarkeite (Ca:Na:K = 1.0 : 1.25 : 0.05). Volumetrically, Na-clarkeite is the most important constituent. The chemical zoning reflects the disparate chemistries of K+, Na+, and Ca2+. Na-clarkeite and Ca-clarkeite are structurally similar, but have different sheet structures. The solubility of K in Na-clarkeite is less than five mole percent and is due to the different sizes of Na+ and K+ ions. The K-phase may not be related structurally to clarkeite.

Exposed to ground water at low temperatures, clarkeite alters to the uranyl silicates, uranophane and kasolite. The alkalis are leached by ground water. The fate of Th, Y, and lanthanides from clarkeite is uncertain.

The time scale of spent fuel dissolution studies is of the order of magnitude of 2 to 10 years, while the performance of a spent fuel repository should be assessed for much longer times (105-106 years). These time scales can be bridged using appropriate natural analogues. Among other important information, the study of natural systems can give insight of which can be the oxidative alteration of spent fuel in granitic environments. However, in studying such systems, thermodynamic and kinetic data of relevant natural solid phases are needed.

In this work we present preliminary results of dissolution experiments carried out under oxidizing conditions with selected and well characterized natural samples of the alteration chain of uraninite (i.e., uraninite, schoepite, uranophane).

The experiments have been performed using a synthetic granitic groundwater as a leachant, in contact with air and at 25 °C.

A modelling study has been carried out to understand the effect of rock alteration on the fractionation of uranium among coexisting minerals (chlorite, vermiculite, kaolinite, amorphous and crystalline iron minerals) at the Koongarra ore deposit, Australia. The model considers the chlorite weathering process, its mechanism and rate, and assumes the presence of reversible and irreversible sorption sites in the secondary minerals. The calculated uranium concentrations at the two different sites in the minerals were compared with the results of sequential extraction experiments. Good agreement between the calculated and observed uranium concentrations was obtained only when an appropriate fraction of uranium is fixed to the irreversible sorption sites of the altered clay minerals. However, a conventional Kd model gave inconsistent uranium concentrations. The calculated results show that the crystalline iron minerals sorb uranium during all stages of weathering, and that the uranium fractionation among the minerals varies with time until the end of the weathering.

A modelling study was carried out to understand the effect of rock alteration on uranium radionuclide concentrations in rocks, in the vicinity of the Koongarra ore deposit, Australia. The one-dimensional, advection-dispersion-sorption model considers two important factors resulting from the process, mechanism and rate of chlorite weathering, one type of rock alteration that has occurred over a million-year period; (a) the changes in the distribution coefficients of uranium and thorium over time, and (b) that in rock porosity. The distribution coefficient and rock porosity at a given time, are expressed as average values of those of the coexisting minerals. By assuming a Darcy velocity of 0.9 m/y and an initial uranium concentration in the groundwater of 0.5 ppm, similar to present-day values, we were able to derive, from the finite-element modelling, uranium concentrations in the rock which are in good agreement with observed values. The calculated values of 230Th/234U activity ratios show a similar trend to those observed, although more experimental data are necessary to confirm the similarity. However, in the absence of the rock alteration, the calculated results did not agree with those observed. The present study suggests that rock alteration should be included in models used to predict uranium migration over long, geologic timescales.

Exposures of altered rock that have been thought to form by recent discharge of water from depth were examined to address a concern that hydrothermal processes could compromise the isolation capability of a potential high-level nuclear waste repository at Yucca Mountain. The suspected hot-spring and hydrothermal-vent deposits are more likely the products of infiltration of meteoric water into newly deposited and still-hot pyroclastic flows >12 Myr ago.

The strategies for waste package (WP) performance validation will be based on site specific geologic and hydrogeochemical information plus models which can be used to predict potential WP lifetimes. The development and application of such models will include the evaluation of natural analogues (NA). These analogues are needed to resolve issues related to the validation of models. Natural analogues have not had extensive use or widespread acceptance in the area of waste package failure prediction. This lack of acceptance is due to the anticipated choice of alloys for waste package containers. Few of these alloys are similar to naturally occurring metals, and the proposed HLW repositories are in general in geologic settings not very similar to those in which naturally occurring metals are generally found.

Natural and archaeological analogues can be used, however, in analysis of possible waste package failures as a means of testing proposed models for failure. In fact, the analogues are the only available mechanisms for testing models of long-term waste package behavior. A strategy is outlined for incorporating natural and archaeological analogue studies into waste package research, and examples are discussed. The natural/archaeological analogues approach which appears most promising is to use archaeological and mineral samples to develop an understanding of the identities and rates of the mineral alteration reactions at or near the surface of the package, improving present capability for estmating the lifetimes of metallic waste package containers.

Natural materials may be used to advantage in the evaluation of the long-term performance of nuclear waste forms. Three case studies are presented: (I) radiation effects in ceramic waste forms; (II) corrosion products of U02 under oxic conditions; (III) corrosion rate of nuclear waste glasses. For each case, a natural phase which is structurally and chemically analogous to the waste form is identified and used to evaluate the long-term behavior of a nuclear waste form. Short-term experimental results are compared to the observations made of analogous natural phases. The three case studies illustrate that results may range between providing fundamental data needed for the long-term evaluation of a waste form to only providing qualitative data of limited use. Although in the most rigorous view the long-term behaviour of a phase cannot be predicted, the correspondence between short-term experimental results and observations made of natural phases provides confidence in the “predicted” behavior of the waste form. The strength of this approach rests with the degree to which a mechanistic understanding of the phenomenon is attained.

Static leach tests similar to the PCT were performed for times up to two years to assess the long-term reaction behavior of high-level nuclear waste glasses similar to those expected to be produced at the Defense Waste Processing Facility. These tests show the reaction rate to decrease with the reaction time from an initially high rate to a low rate, but then to accelerate to a higher rate after reaction times of about one year, depending on the glass surface area/leachant volume ratio (SAN) used. The solution concentrations of soluble glass components increase as the reaction is accelerated, while the release of other glass components into solution is controlled by secondary phases which form during the reaction. The net result is that the transformation of glass to stable phases is accelerated while the solution becomes enriched in soluble components that are not effectively contained in secondary phases. The rate becomes linear in time after the acceleration and may be similar to the initial forward rate. A current model of glass reaction predicts that the glass reaction will be accelerated upon the formation of secondary phases which lower the silicic acid solution concentration. These tests show the total silicon concentration to increase upon acceleration of the reaction, however, which may be due to the slightly higher pH that is attained with the acceleration. The sudden change in the reaction rate is likely due to secondary phase formation.

We simulated the reaction of SRL-131 glass with equilibrated J-13 water in order to investigate the effects of surface area to volume ratio (SAN/) on glass dissolution. We show that glass-fluid ion exchange causes solution pH to rise to progressively higher values as SA/V increases. Because the ion exchange is rapid relative to the duration of the glass dissolution experiment, the pH effect does not scale with (SA/V)*time. Experiments compared at the same (SA/V)*time value therefore have different pHs, with higher pHs at higher SA/V ratios. Both experimental data and our simulation results show similar trends of increasing reaction rate as a function of SAN ratio when scaled to (SA/V)*time. Glasses which react in systems of differing SA/V ratio therefore follow different reaction paths and high SAN ratios cannot be used to generate data which accurately scales to long time periods unless the ion exchange effect is taken into account. We suggest some simple test designs which enable more reliable high SAN accelerated tests.

Comparative studies of two multi-component silicate glasses have confirmed the observation that glasses with a relatively low SiO2 + AlO3/2 content may exhibit temporary increases in leach rate during the initial stages of their exposure to water. SEM studies of the leached glass surfaces strongly support the assumption that this phenomenon is due to cracking of the leached glass and a consequent increase of the exposed surface area.

Long-term tests on the durability of nuclear waste glasses, both under laboratory conditions and in actual burial environments, have shown that after periods of several months or years the rates of glass corrosion can show unexpected increases. In burial tests in brine environments the growth of the thickness of the surface layer during years 3 to 5 as revealed by SIMS and SEM measurements was found to be faster than expected on the basis of findings during the first 2 years. In this paper we report on our SEM-EDXS observations of a 5 year burial sample of SRL-165 glass from the MIIT test, and compare our findings to those of other studies, especially SIMS measurements. Our average thickness measurement for the leached layer at the glass-metal interface is 3.70 μm, but wide variation is observed, as well as evidence of partial loss of the leached layer in some areas. A thicker, better retained layer is observed on the glass-brine interface near the core heater averaging 8 to 12 μm in thickness with occasional local attack producing leached layers of greater depth, in one case nearly 40 μm. These findings show that it is important to extend durability tests to long periods in order to establish a reliable basis for predictive modeling of long-term glass durability.

A case is made for the use of short-term laboratory data in making predictions on the likelihood of significant colloid formation in supersaturated leachates of glass, under long-term repository conditions, using “accelerated tests” with a large ratio of the surface area of the glass to the leachate volume. In the repository conditions in which colloids can form, long-term leaching may be a kinetically-controlled process that involves the continuous formation of colloids. If this kinetic process dominates, it could lead to a significant increase in the predicted rates of radionuclide release. The question is whether or not colloids may form after prolonged times; the delayed formation would make it difficult to use short-term laboratory test results to represent (or predict) the long-term and cumulative effects of radionuclides. In this work, the pertinent long-term kinetic processes are identified in part. Classical nucleation theory for particle formation, as a potential condensation mechanism for colloid formation, is applied to explain pertinent experimental data on colloid formation. The classical theory, which is justified for this discussion, indicates that as supersaturation of a leachate is decreased, the nucleation rate decreases most significantly, while the incubation time increases at a small rate. As a result of this decreased nucleation rate, the significance of colloids tends to vanish, and usefulness of data from “accelerated“ laboratory tests may be applicable to long-term behavior.