A borosilicate nuclear waste glass was static leached in pure water, silicate water, and brine solution. Three different forms of lead were included in specified corrosion cells to assess the extent to which various lead species alter the leaching behavior of the glass. Weight loss data indicated that Pbm and Pb0 greatly reduce the weight loss of glass when leached in pure water, and similar effects were noted in silicate and brine. Si concentrations, which were substantial in the glass-alone leachate, were reduced to below detection limits in all pure water cells containing a lead form. Lead concentration levels in the leachate were controlled by lead form solubility and appeared to be a significant factor in influencing apparent leaching behavior. Surface analysis revealed surface crystals, which probably formed when soluble lead in the leachate reacted with dissolved or activated silica at the glass surface. The net effect was to reduce the release of some glass constituents to the leachate, although it was not clear whether the actual corrosion of the glass surface was reduced. Significantly different corrosion inhibiting effects were noted among lead metal and two forms of lead oxide.

Three Savannah River Laboratory (SRL) simulated nuclear waste glasses were buried in granite boreholes 345 meters deep. Included in the same boreholes were other potential waste package components including stainless steel and bentonite. Samples were maintained at either ambient mine temperature (8–10°C) or 90°C. Differences in glass leaching performances were observed among the three compositions, with SRL 165 being more durable than SRL 131, both with 29.8% TDS waste. Likewise, the presence of some package components affected the leaching performances. Bentonite resulted in accelerated attack on the glass while the presence of stainless steel did not appear to have much effect. Results obtained through one year of burial are presented in this paper.

The disposal of glass blocks containing fission products in granitic geological formations is one of the options currently investigated in France. This paper discusses the major criteria selected to develop models of such repositories and gives the first results obtained with the corresponding experimental test. Five laboratory mockups corresponding to different alternatives regarding the fill materials were prepared. The following materials were used to encase the French LWR reference glass:

- Test I : Granite, sand, stainless steel;

- Test 2 : Granite, sand, illite, stainless steel;

- Test 3 : Granite, sand, bentonite, stainless steel;

- Test 4 : Granite, sand, Al2O3, ceramic, stainless steel;

- Test 5 : Granite, sand, haitalloy C4.

Samples were taken of the solutions obtained in the preconditioning cell and in contact with the glass to assess the glass corrosion behavior in each of these media. The experiments were completed by surface analyses designed to compare the morphology of the surface layers that develop with and without the environmental materials.

The early stages of surface alteration-layer formation were examined during the leaching of some candidate radioactive waste glasses. The composition-depth profiles of the leached surfaces were obtained with sputter-induced photon spectroscopy (SIPS) and secondary-ion mass spectroscopy (SIMS). It was observed that the formation of multiple surface layers is a general feature of the glass dissolution process. The sequence of the layer compositions depends upon the order in which solubility limits are attained in solution, the adsorption or precipitation of insoluble hydrates at the glass/water interface, and subsequent incorporation of these insoluble hydrates behind the advancing reaction front. The boundary between the surface layer and bulk glass is sharp.

In a joint Japanese, Swedish, Swiss HLW glass characterization programme, interactions of real waste glass specimens with possible burial system components are studied. The report describes the glass surface analytical work performed at the Swiss Federal Institute for Reactor Research (EIR) during the first completed phase of the programme.

Zeta potentials of SRL-131-29.8% TOS simulated nuclear waste glasses leached in D.I. water, Al, Ca, Mg, and Zn chloride solutions at 90°C were measured as a function of leaching time. For short term leaching, the adsorption of Ca, Mg, Zn and Al reverses the glass surface potential from negative to positive. Colloids were found to be stable in D.I. water and AICl3 solutions after leaching, presumably due to the electrostatic repulsion between the glass surface and similarly charged particles. Colloids were not found in Mg, Zn or Ca chloride solutions after leaching; instead, a relatively thick metasilicate surface layer was formed on glass surfaces leached in these solutions. The concentration of Si in solution is reduced by the formation of these surface layers.

A simple kinetic model for the description of the interaction of nuclear waste borosilicate glass with water has been developed. In the case of SRL TDS-131 glass leached in water at 70°C over a broad range of flow rates this model was found useful both in describing the evolution of leachate composition as a function of exposure time at a particular flow rate and in describing the dependence of the steady-state concentrations of the leached elements in solution on flow rate in a series of flow experiments.

This paper examines the effects of solubility in limiting dissolution rates of a number of important radionuclides from spent fuel and high-level waste. Two simple dissolution models were used for calculations that would be characteristic of a Yucca Mountain repository. A saturation-limited dissolution model, in which the water flowing through the repository is assumed to be saturated with each waste element, is very conservative in that it overestimates dissolution rates. A diffusion-limited dissolution model, in which element-dissolution rates are limited by diffusion of waste elements into water flowing past the waste, is more realistic, but it is subject to some uncertainty at this time. Dissolution rates of some elements (Pu, Am, Sn, Th, Zr, Sm) are always limited by solubility. Dissolution rates of other elements (Cs, Tc, Np, Sr, C, I) are never solubility limited; their release would be limited by dissolution of the bulk waste form. Still other elements (U, Cm, Ni, Ra) show solubility-limited dissolution under some conditions.

Samples of SRP glass containing either simulated or actual radioactive waste were leached at 90°C under conditions simulating a saturated tuff repository environment. The leach vessels were fabricated of tuff and actual tuff groundwater was used. Thus, the glass was leached only in the presence of those materials (including the Type 304L stainless steel canister material) that would be in the actual repository. Tests were performed for time periods up to 6 months at a SA/V ratio of 100 m−1. Results with glass containing simulated waste indicated that stainless steel canister material around the glass did not significantly affect the leaching. Based on Li and B (elements not in significant concentrations in the tuff or tuff groundwater), glass containing simulated waste leached identically to glass containing actual radioactive waste. The tuff buffered the pH so that only a slight increase was observed as a result of leaching. Results with glass containing actual radioactive waste indicated that tuff reduced the concentrations of Cs-137, Sr-90, and Pu-238 in the free groundwater in the simulated repository by 10–100X. Also, radiolysis of the groundwater by the glass (approximately 1000 rad/hr) did not significantly affect the pH in the presence of tuff. Measured normalized mass losses in the presence of tuff for the glass based on Cs-137, Sr-90, and Pu-238 in the free groundwater were extremely low, nominally 0.02, 0.02, and 0.005 g/m2, respectively, indicating that the glass-tuff system retained radionuclides well.

The influence of gamma irradiation on the reaction of actinide doped SRL 165 and PNL 76–68 glasses in a saturated tuff environment has been studied in a series of tests lasting up to 56 days. The reaction, and subsequent actinide release, of both glasses depends on the dynamic interaction between radiolysis effects which cause the solution pH to become more acidic and glass reaction which drives the pH more basic. The use of large gamma irradiation dose rates to accelerate reactions that would occur in an actual repository radiation field may affect this dynamic balance by unduly influencing the mechanism of the glass-water reaction. Comparisons are made between the present results and data obtained by reacting the same or similar glasses using MCC-1 and NNWSI rock cup procedures.

Several series of simulated (nonradioactive) defense high-level waste (DHLW) package tests have recently been emplaced in the WIPP, a research and development facility authorized to demonstrate the safe disposal of defense-related wastes. The primary purpose of these 3-to-7 year duration tests is to evaluate the in situ materials performance of waste package barriers (canisters, overpacks, backfills, and nonradioactive DHLW glass waste form) for possible future application to a licensed waste repository in salt. This paper describes all test materials, instrumentation, and emplacement and testing techniques, and discusses progress of the various tests.

These tests are intended to provide information on materials behavior (i.e., corrosion, metallurgical and geochemical alterations, waste form durability, surface interactions, etc.), as well as comparison between several waste package designs, fabrications details, and actual costs.These experiments involve 18 full-size simulated DHLW packages (approximately 3.0 m x 0.6 m diameter) emplaced in vertical boreholes in the salt drift floor. Six of the test packages contain internal electrical heaters (470 W/canister), and were emplaced under approximately reference DHLW repository conditions. Twelve other simulated DHLW packages were emplaced tinder accelerated-aging or overtest conditions, including the artificial introduction of brine, and a thermal loading approximately three to four times higher than reference. Eight of these 12 test packages contain 1500 W/canister electrical heaters; the other four are filled with DHLW glass.

The Basalt Waste Isolation Project (BWIP) is being conducted for the U.S. Department of Energy (DOE) by Rockwell Hanford Operations. The effect of Grande Ronde Basalt groundwater composition and temperature on the corrosion rate of American Iron and Steel Institute (AISI) 1020 steel in basalt-bentonite packing was investigated as part of the Containment Materials Testing Program of the BWIP. The studies were based on a Plackett-Burman statistical screening design, the purpose of which is to identify the important variables affecting a process and to use this information to guide future investigative efforts. For the present studies, four anions (Cl−, F−, SO4−, and CO3−) and temperature were selected as the initial variables affecting corrosion. A minimum and a maximum value was chosen for each variable. The minimum value was 0 mg/L for the anions and 100 °C for the temperature. The maximum values were 780, 76, 576, and 120 mg/L, respectively, for Cl−, F−, SO4−, and CO3−, and 250 °C for the temperature. Analysis of the data revealed that temperature is the only statistically significant variable affecting corrosion of low-carbon steel in basalt groundwater in the composition range of the anions tested. Increasing the temperature from 100 °C to 250 °C decreased the total corrosion by a factor of seven to ∼5O mg/dm2 independent of the water composition. This is attributed to a strongly adherent layer of iron- and silicon-rich clay on the surface of the steel formed by a reaction of the steel surface with the packing material as observed in previous studies.

Mild steels are considered to be strong candidates for waste package structural barrier (e.g., overpack) applications in salt repositories. Corrosion rates of these materials determined in autoclave tests utilizing a simulated intrusion brine based on Permian Basin core samples are low, generally μm (1 mil) per year. When the steels are exposed to moist salts containing simulated inclusion brines, the corrosion rates are found to increase significantly. The magnesium in the inclusion brine component of the environment is believed to be responsible for the increased corrosion rates.

Carbon steels may be used for borehole liners in a potential high-level nuclear waste repository in tuff in Nevada. Borehole liners are needed to facilitate emplacement of the waste packages and to facilitate retrieval of the packages, if required. Corrosion rates of low carbon structural steels AISI 1020 and ASTM A-36 were determined in J-13 well water and in saturated steam at 100°C. J-13 well water is representative of water which has percolated through the tuff horizon where the repository would be located. Tests were conducted in air-sparged J-13 water to attain stronger oxidizing conditions. A limited number of irradiation corrosion and stress corrosion tests were performed. Chromium-molybdenum alloy steels and cast irons were also tested. These materials showed lower general corrosion but were susceptible to stress corrosion cracking when welded.

A test method has been developed to measure the release of radionuclides from the waste package under simulated NNWSI repository conditions, and to provide information concerning materials interactions that may occur in the repository. Data from 13 weeks of unsaturated testing are discussed and compared to that from a 13 week analog test. The data indicate that the waste form test is capable of producing consistent, reproducible results that will be useful in evaluating the role of the waste package in the long-term performance of the repository.

A series of tests have been performed to characterize the chemical stability of a DWPF borosilicate glass sample as part of the Waste Package Task of the NNWSI Project. This material was prepared at the Savannah River Laboratory for the purpose of testing the 165-frit matrix doped with a simulated non-radioactive waste. All tests were conducted at 90°C using deionized water and J-13 water (a tuffaceous formation groundwater). In the deionized water tests, both monoliths and crushed glass were tested at various ratios of surface area of the sample to volume of water in order to compare leach rates for different sample geometries or leaching times. Effects on the leach rates due to the presence of crushed tuff and stainless steel material were also investigated in the tests with J-13 water.

A new analytical solution is presented that predicts the rate of dissolution of species from a waste package surrounded by a wet porous medium. By equating the rate of diffusive mass transfer into the porous rock to the rate of liquid-surface chemical reaction, an analytical solution for the timedependent dissolution rate and the time-dependent concentration of dissolved species at the waste surface is obtained. From these results it is shown that for most of the important species in a package of radioactive waste the surface liquid quickly reaches near-saturation concentrations and the dissolution rate can be predicted by the simpler theory that assumes saturation concentrations in the surface liquid.

Hydrothermal experiments using basalt-repository nuclear waste package components have been carried out at 200°C and 30 MPa using 99Tc-doped PNL 76–68 waste glass. This work was conducted in support of the Basalt Waste Isolation Project. The experiments were carried out in rocking autoclaves that allowed for periodic solution sampling. Preliminary results that illustrated the effect of basalt on 99Tc solution behavior were discussed previously[l]. In this paper, we continue those observations by discussing the additional experiments that investigated the effects of steel. We also include the post-experiment solid phase analyses from all test configurations.

The effect of steel on glass dissolution was observed to be minimal when basalt was not present, i.e., there were no discernible differences in the amount of glass dissolved (based on boron release), with or without steel present.

The 99Tc solution concentration results showed that basalt, steel, and a combination of basalt and steel have an ability to dramatically lower the concentration of 99Tc in the solution, probably through a redox mechanism[2]. Solid run product analyses showed that without basalt present, a gel-like secondary phase consisting of an iron, zinc clay with an apparent smectite structure formed. When basalt was present clinoptilolite formed as the major secondary phase, due primarily to the presence of aluminum in the basalt mesostasis. Separation of the various solid run products and residual initial solid phases has not yet been achieved. Such a separation would facilitate the identification of the phase or phases with which 99Tc was associated. The solution results indicated that 99Tc, a potentially mobile radionuclide, may be incorporated in a relatively insoluble phase in the environment of a basalt repository. In addition, no synergistic effects between waste package components were observed that would increase the concentration of 99Tc in solution.

Sorption isotherms and apparent concentration limits for Tc(VII) and Np(V) for a variety of groundwater/basalt systems were determined using Grande Ronde basalt samples representative of the Hanford Site candidate high-level waste repository. Under oxic redox conditions (air present), little or no sorption of technetium was observed; neptunium exhibited low to moderate sorption ratios. Under anoxic redox conditions (oxygen-free), low to moderate sorption of technetium was often observed, but the extent of sorption was highly dependent upon the groundwater composition and the method of pretreatment (if any) of the basalt. Sorption isotherms for technetium under reducing redox conditions (hydrazine added) indicate an apparent concentration limit of approximately 10−6 mol/L Tc. No apparent concentration limit was found for neptunium for concentrations in groundwater up to.∼10−6 mol/L and 8 × 10−7 mol/L under oxic and reducing (hydrazine added) redox conditions, respectively.

Valence control and valence analysis experiments suggest that the sorption or precipitation of Tc and Np from groundwater in the presence of basalt may result from a heterogeneous reaction occurring on the surface of the basalt. One of the critical factors of this reduction reaction appears to be the accessibility of the reactive ferrous iron component of the basalt. The laboratory simulation of groundwater redox conditions representative of the repository environment through the use of solution phase redox reagents is of questionable validity, and information obtained by such experimental methods may not be defensible for site performance assessment calculations. Anoxic experiments conducted in an argon-filled glove box appear better suited for the laboratory simulation of in situ redox conditions.

A series of hydrothermal experiments were carried out on 237Np- and 239Pu-doped PNL 76–68 glass, synthetic basalt groundwater, basalt, and cast steel. These hydrothermal experiments are part of the Basalt Waste Isolation Project investigation of the interactions of waste package components in a basalt repository. Experiments of three months duration were conducted in Dickson-type rocking autoclaves at 200°C and 30 MPa, with an initial fluid to solid weight ratio of 10:1. All solid materials were ground and sieved to a narrow size fraction. The experiments carried out were: glass and groundwater; glass, basalt, and groundwater; glass, steel, and groundwater; and glass, steel, basalt, and groundwater. Unfiltered, 4000 Â filtered, and 18 Å filtered solutions were analyzed to determine the concentrations of radionuclides in solution and those associated with colloids.

The quantities of 237Np and 239Pu in solution were very small. Worstcase calculations indicate that 0.01% or less of the total radionuclide inventories were present in solution. The highest solution concentrations of the actinide dopants were observed in the experiments with basalt, even though smaller amounts of glass were dissolved than in the experiments without basalt. The observed differences in the solution concentrations of 237Np and 239Pu were probably controlled by differences in the reaction products, which were clinoptilolite in the experiments with basalt, and an Fe-Zn smectite clay in the experiments without basalt.