The immobilization of non-radioactive Na in nuclear waste has not been given much attention. However since Na compounds are normally very soluble, Nabearing phases may render a waste form more susceptiible to leaching of radionuclides. The Na-phases may incorporate radionuclides and also, leaching of soluble Na phases will increase the surface area of the waste form. The crystalline Na-bearing phases which might occur in tailored ceramic formulations for nuclear waste are nepheline [1,2], magnetoplumbite [3] and sodalitetype materials [4]. Other Na phases which might occur in titanate-tailored [2,5] or titanate-encapsulated [6,7] forms are perovskite [8,9], Na2Ti3O7 (especially in titanate-encapsulated materials) and Na2TiSiO5 or Na2Ti2Si2O9, etc., if the waste contained sand or zeolites [10].

A study was carried out to determine the solubility limits of various elements found in waste glasses in demineralized water as a function of temperature. The work was sponsored by the Office of Nuclear Waste Isolation under contract to the Department of Energy. Solubility measurements were carried out at 35°, 65°, 95°, and 150°C using three nonradioactive waste glass compositions. Subsaturation and supersaturation methods were used to determine the solubility limits. The two methods gave markedly different values for most glass components. The results obtained indicate that it is difficult to assign solubility limits to most glass components without thoroughly describing the glass-water system. This includes not only defining the glass type, and system temperature, but also the glass surface area-to-water volume (S/V) ratio of the system and its thermal history.

Previous investigations have shown that the lanthanide orthophosphates (i.e., analogs of the mineral monazite) are promising candidate waste forms for the storage of high-level actinide or TRU wastes [1]. In the present work, the properties of lanthanide orthophosphates loaded with simulated Savannah River defense wastes and simulated reprocessed light water reactor wastes have been investigated with the goal of evaluating the potential of these materials as advanced primary waste forms.

A technique has been developed using 222Rn as a flow path monitor. Its principle advantage over other radiotracers or dyes is that while the Kd of Rn gas in water solution is zero and hence follows the water path, its 210Pb daughter is retained strongly by rock. The immobilized 210Pb is not subject to diffusion, an important consideration in the proposed Nevada Field Test. To evaluate the technique, a series of laboratory scale experiments have been performed. Artificial fissures consisting of glass plates and prepared flat surfaces were coupled to insure well characterized fissures. Water solutions of radiotracer were metered through these fissures and the discharge monitored to produce breakthrough curves as a function of flow rate. The rock surfaces were analysed for radionuclide concentration and the rock sectioned for diffusion profile into the rock.

The formation and sorption properties of colloidal americium in aqueous solutions have been studied with variations of pH (2–12), ionic strength (0.01–1.0 M NaClO4), americium concentration (10−7 −10−11 M) and storage time (6 h–6 months). A large fraction of the americium is sorbed on the container walls or on Al2O3 in the pH-range 7–11. Around pH 5–8 and at pH above 12 centrifugable fractions (particle sizes greater than 100 nm) are obtained. The fraction formed at pH above 12, which seems to be a true hydroxide colloid, migrates through an A12O3-filled column with very little retention. The average charge of americium species in solution changes from positive values at pH below 7–8 to negative values at pH above 10–11, as indicated from electromigration studies.

Radionuclide migration experiments in fractured granite at the Climax Stock, Nevada Test Site, are needed to compare field and laboratory measured retardation factors to determine whether laboratory studies accurately reflect in situ conditions. Initial field activities have concentrated on hydrological investigations to determine whether the fractures in Climax granite are suitable for migration experiments. A critical question was whether we could isolate a single vertical fracture between two boreholes and establish flow along that fracture from an upper to a lower borehole. Of the ten fractures tested, one fracture would not take water at pressures up to 200 psig for 24 hours. Several fractures were so permeable they accepted water at a rate which exceeded the pumping capacity of the equipment. Other fractures failed to show a connection between the two boreholes. In two fractures, we were able to establish a circulating system with up to 95 percent of the injected water being recovered. Constant pressure injection tests were conducted. Intrinsic permeabilities of 33 and 75 (μm)2 were estimated using a radial nonsteady flow model. These values correspond to effective fracture apertures of 20 and 30 μm respectively. Concurrent with the hydraulic testing activities is a study of the Climax ground-water chemistry. Our analyses show the natural water to be very different in composition from the granite equilibrated water used in laboratory sorption studies. This paper includes the results of the hydrogeological and geochemical investigations, and describes the overall experimental design plans for the radionuclide migration experiments.

A project to begin to address the phenomena of flow and element migration in fractured porous rock has recently been started by the Los Alamos National Laboratory, Sandia National Laboratories, and Argonne National Laboratory. The work has three objectives: 1) to develop the experimental, instrumental, and safety techniques necessary to conduct controlled, small-scale, radionuclide migration, field experiments; 2) to use these techniques to define radionuclide migration through rock by performing generic, at-depth experiments under closely controlled conditions in a single fracture in porous rock; and 3) to determine whether available lithologic, geochemical, and hydraulic properties together with existing or developed transport models are sufficient and appropriate to describe real field conditions (i.e., to scale from small-scale laboratory studies to bench-size studies to field studies). The detailed scope of this project and its current status are described.

Knowledge of Pu solid phases present in nuclear wastes is important for predicting the geochemical behavior of Pu. Thermodynamic data and experimental measurements using discrete Pu compounds, Pu-doped borosilicate glasses (simulating a high-level waste form), and Pu contaminated sediments suggest that PuO2(c) is very stable and is expected to be present in the repository. The solubility of the stable phase, such as PuO2(c), can be used to predict the maximum Pu concentration in solutions for long-term safety assessment of nuclear waste repositories.

Slug and pulse tests have been used extensively to measure the hydraulic conductivity and specific storage of granitic rocks. After several hundred tests using an I.G.S. design of straddle packer test equipment, it became clear that the conventional methods of test analysis were inadequate. The straddle packer equipment developed by IGS allows the time scale of a test to be altered without repositioning the packers so that a range of tests can be carried out on the same test interval. The new analysis procedure presented here uses a model of fissured rock which incorporates water movement in both the fissures and the rock matrix. A new variable is introduced which includes the parameters matrix hydraulic conductivity, matrix specific storage fissure specific storage and fissure hydraulic conductivity together with the controllable variables, effective casing radius and packer interval length. Examples of analysis using multiple tests in the same interval are presented together with the apparent relationship between measured hydraulic conductivity and specific storage. It would appear that a small number of fissures endow the rock mass with the bulk of its hydraulic conductivity and that many of the higher hydraulic conductivity tests are probably associated with localised “pipe flow”.

The Bryan-Eldora stock (Colorado) intruded the Precambrian Idaho Springs Formation metamorphic rocks 58 million years ago. Geochronologic-geochemical work by Hart et al. (1) has demonstrated that the heat from the cooling intrusive rocks was sufficient to affect mineral isotopic systematics up to 2,000 m. from the contact, and the nature of these isotopic perturbations can be explained by a simple diffusion model in turn based on various heat flow models. Our new studies are focused on elemental exchange between stock and intruded rock as a function of distance from the contact; the assumption is made that the stock is a very large, high heat source analogous to a waste form emplaced in the metamorphic rocks without benefit of canister or engineered backfill. Data for U, Th and the REE indicate actinide and lanthanide immobility except perhaps in the 0–2m. contact zone where some infiltration of the country rocks by stock-derived fluids occurred. Beyond 4 m. no stock-derived U, Th, REE or *Pb are noted. Further, whole rock Rb-Sr and stable 0 isotopic data indicate conductive cooling as opposed to convective, water-induced cooling. The intruded rocks possess low porosity and permeability; this helped prevent elemental migration during the 105 − 106 years of stock crystallization. The petrographic and geochemical studies show that the Idaho Springs (or equivalent) metamorhpic rocks are well suited for radwaste storage.

Laboratory experiments examining flow and element migration in rocks containing isolated fractures have been initiated at the Los Alamos National Laboratory. Techniques are being developed to establish simple fracture flow systems which are appropriate to models using analytical solutions to the matrix diffusion - flow equations, such as those of I. Neretnieks [1]. These experiments are intended to be intermediate steps toward larger scale field experiments where it may become more difficult to establish and control the parameters important to nuclide migration in fractured media.

Laboratory experiments have been run on fractures ranging in size from 1 to 20 cm in length. The hydraulic flow in these fractures was studied to provide the effective apertures. The flows established in these fracture systems are similar to those in the granite fracture flow experiments of Witherspoon et al. [2]. Traced solutions containing 85Sr and 137Cs were flowed through fractures in Climax Stock granite and welded tuff (Bullfrog and Tram members, Yucca Mountain, Nevada Test Site). The results of the elutions through granite agree with the matrix diffusion calculations based on independent measurements of Kd. The results of the elutions through tuff, however, agree only if the Kd values used in the calculations are lower than the Kd values measured using a batch technique. This trend has been previously observed in chromatographic column experiments with tuff.

Zirconolites, CaZrTi2O7, from Sri Lanka and Pala Bora, South Africa, and a calzirtite, CaZr3 TiO9, from Jacupiranga, Brazil, were examined using the electron microprobe, x-ray diffraction (annealing study), transmission electron microscopy, scanning electron microscopy and optical microscopy. The x-ray data indicate that all three zirconolites are metamict. Both Sri Lanka zirconolites are amorphous to the limits of resolution of the electron microscope (֮10 A). The Pala Bora zirconolite is largely amorphous but contains isolated domains (50–200 A) of crystalline material which may be the result of post—metamict recrystallization and alteration. The only other significant evidence for chemical alteration was the lower Th concentration (1–2 weight percent) and slightly lower analytic totals for the rims of the Sri Lanka zirconolites. Upon annealing at 1130°C for 5 hours, all three zirconolites recrystallized as microcrystalline aggregates. Refined unit cell parameters and volumes are consistent with published data for synthetic zirconolites. Both Sri Lanka zirconolites contain microvoids, spherical in shape, and 200 Angstroms to 2 microns in size. This porosity may be the result of helium accumulation arising from the decay of U and Th. The calzirtite was highly crystalline, exhibited no porosity, and was unchanged by the annealing treatment.

Polyhalite (K2Ca2Mg(SO4).2H2o) is an important mineral present in the bedded evaporites of the WIPP site, southeastern New Mexico. Polyhalite contains two structurally bonded water molecules, and it is thus important to know if this mineral formed at or close to the time of sedimentation or, as proposed by others (1,2) more or less continuously throughout the Mesozoic and Cenozoic. If formed much later than original evaporite formation, then this implies a new source of water in order for polyhalite to form in turn raising questions about water abundance and mobility in the bedded evaporites. Polyhalite is not well suited for Rb-Sr geochronologic study becauseof the large amounts of common Sr (3), but it has been shown to be suitable for study by the K-Ar geochronologic methods(4). Pure polyhalites and those mixed with very small amounts of impurities yield K-Ar dates in the range of 200 – 220 MYBP and indicate little if any, loss of *40Ar. Polyhalites mixed with sylvite, halite, and (Na,K)Cl yield lower ages due to loss of *40Ar from the halides. Apparent age lowering is proportional to halide content and an extrapolated age of formation of near 225 MYBP results. Collectively, the K-Ar data convincingly show that the polyhalites formed at the time of sedimentation (except for isolated occurrences near dike intrusions) in Late Permian time. Finally, thermodynamic calculations, supported by petrographic study, show that polyhalite is often stable with respect to carnallite, sylvite and other minerals formerly thought to be older than polyhalite.

U(VI) bearing species are commonly formed under oxidizing laboratory rock-water-HLW experimental conditions. The experimental stability of artificially produced Na-Cs-Rb weeksite (alkali uranyl hydrosilicate), and other U(VI) species, is uncertain. Naturally occurring U(VI) bearing minerals (uranophane and tyuyamutrite) from the Grants mineral belt, New Mexico, have been studied for their U-Pb systematics as natural analogs to these experimentally produced phases. Samples from highly oxidized rocks and from chemically reducing environments have been studied. The samples from oxidized rocks are mixed with hematite-barite-gypsum-calcite gangue and are close to, but not in contact with, pyritiferous-organic carbon-uraniferous (U(NIV))-chloritic are. The reduced rock uranophane yields nearly concordant238U - 206Pb and 235U - 207Pb ages of 8.32 and 8.33 MYBP (millions of years before the present); the oxidized rock uranophane and tyuyam .nite yield more discordant ages (n=8) between 7.4 and 2.7 MYBP. These data argue for long range stability of U(VI)-rich phases in the geologic environment.

The diffusion of the 235NpO2+ and 95m TCo4− -ions has been measured directly in sample cylinders of two different sediments from the floor of the deep sea. In smectiterich sediment not shielded from contact with atmospheric oxygen the following values were obtained for the effective diffusion coefficients of neptunyl- and pertechnetateions, respectively: Deff(NpO2+) = 1.5 × 10−12 m2 s−1 and Deff(TcO4−) = 3.2 × 10−10 m2 s−1. Under anoxic conditions in sediment with known reducing properties, the pertechnetate ion appears to undergo slow reduction and the effective diffusion coefficient of the reduced species of Deff(Tc,red)= 1.1 × 1014 m2 s−1 reflects a substantial decrease of the mobility of the lower-valent technetium.

Investigation of candidate sites for nuclear waste isolation will require an assessment of their radiogeologic settings. Studies at the Stripa research facility in granitic rock of central Sweden incorporated the distribution and abundance of naturally occurring radioelements in rocks encompassing the underground experiments and in the accompanying fracture-controlled groundwater system. These studies showed that besides defining the natural radioactivity baseline upon which the effects of radioactive waste will be superimposed, radioelement distributions can be used to determine the apparent age of the groundwater and its flow paths. In crystalline rocks, where the groundwater systems are confined to the joints and fractures, the uranium daughter element, radon-222 in the water serves as a natural tracer to locate fractures along which significant flow is occurring and to measure the flow rates. The heat production from radioactive decay of uranium-238, thorium-232 and potassium-40 in the rock, combined with measurements of regional and local geothermal heat flow, permit calculation of the apparent size of the rock mass that will encompass the repository. This method is especially useful in terranes such as at Stripa where the contacts between plutons and older rocks are concealed.

Geochemical research carried out by the U.S. Subseabed Disposal Program is described. Data from studies of high temperature interactions between sediments and porewater (seawater), and of sorption and diffusion of radionuclides in oxidized, deep-sea sediments are used, along with results from heat transfer studies, to predict migration rates of radionuclides in a subseabed repository.

Preliminary results for most radionuclides in oxidized sediments are very encouraging: fission products with moderate values of KD (generally 10 to 10 ml/g), and actinides with high values of KD (generally 103 to 106 ml/g) would not migrate significant distances before decaying to innocous concentrations. 137Cs, 90Sr, and 239Pu are among this group. The results for anionic species are less encouraging, but preliminary work with reduced sediments indicates that Tc can be effectively isolated.

Planning for a field verification of these laboratory and modeling studies is also described.

The diffusion of Cs, Sr, Am and Tc in compacted bentonite clay has been studied, using a diffusion cell where diffusion takes place axially from the center of a cylinder of the clay. The effect of addition of potential radionuclide getters to the clay (apatite for Am, pyrite, olivine, and Fe(s) for Tc) as well as using complexed (acetate) or colloidal Am has been studied.

The diffusivities obtained for Cs and Sr were very high − 9.0×10−13 m2/s and 2.6×10−12 m2/s, considering that if the effect of sorption on the clay 9is 2 substracted the diffusivities would be 1.8×10−9 m2/s and 3.1×10−8 m2 /s. These diffusivities are, however, maximum values as the concentration profiles do not follow the theoretical form for these nuclides. A small amount of the ions seems to migrate very fast, while the rest moves slower.

For Tc, that may be expected I b2 anionic and thus poorly sorbed, a diffusivity of 5.3×10−11 m2/s was obtained. Addition of Fe(s) decreases the diffusivity by one order of magnitude while olivine and pyrite have no measurable influence during the present contact times (2-3 months). Americium is practically immobile.

Rock-salt deposits have been considered as a possible medium for the permanent storage of high-level radioactive wastes and spent fuel. Brine inclusions present in natural salt can migrate toward the waste if the temperature and the temperature gradients in the vicinity of the radioactive waste are large enough. This migration is due to the dissolution of salt at the hot side of the salt-brine interface, ion diffusion through the brine droplet, and the precipitation of salt at the cold side of the salt brine interface.

The corrosion rate and mechanism of construction materials of galleries and casings for boreholes (possibly to be used as consumable overpack of HLW canisters) have been evaluated after a first campaign of corrosion experiments in direct contact with clay, in a humid clay atmosphere and in a gas corrosion chamber. The envisaged construction material is ductile iron (grade 60) and has been tested with and without metallic and organic coatings. The corrosion mechanism is governed by the formation of a rust scale composed of Fe3O4 and C. Preferential attack underneath the surface reaction layer is associated with the progress of a reaction front at the interface between the graphic nodules or flakes and the ferritic or pearlitic structure.

Post-corrosion analyses of old gray iron coal mine shafts exposed for 60 and 90 years in well defined conditions are included to validate the use of the results of the short-time tests for very long exposure times.