Iron oxyhydroxides play a significant role in radium behavior in uranium mill tailings. Although they are present at trace level (less than 1% in mass), iron oxyhydroxides are responsible of the retention of significant amounts of radium in these residues. A part of these oxyhydroxides is present as amorphous or ill-crystallized forms, the other part corresponds to crystalline iron oxyhydroxides like goethite.

In this paper, we present data concerning radium sorption on a well-characterized goethite (α-FeOOH). Batch experiments have been performed and the influence of different parameters such as contact time, pH and electrolyte concentration of the solution has been examined. The sorption equilibrium was obtained after a contact time of 1 hour. In sodium perchlorate medium, the amount of sorbed radium reaches a maximum for pH higher than 8 and no significant influence of sodium concentration was observed. However, the calcium concentration greatly influences the radium sorption: the greater the calcium concentration, the less radium is sorbed. The kinetics of the radium desorption are slower than those of sorption: more than 30% of radium remains sorbed on goethite after 1 week of contact time.

Using the characteristics of the goethite surface (total number of sites, intrinsic acidity constants, intrinsic constants of calcium), modelling of the experimental data with inner sphere complexes gave satisfactory results. Two surface species involving radium can be formed (SORa+ and SORaOH). These sorption constants determined allow prediction of the radium sorption onto goethite in the chemical conditions of uranium mill tailings.

The stability constants of the hydroxide complexes of zirconium(IV) and uranium(IV) have been measured at 15, 25 and 35°C [in 1.0 mol dm−3 (Na, H)ClO4] using potentiometry. For zirconium(IV), the results indicate the presence of Zr(OH)3+ and the polymeric species Zr3(OH)48+ and Zr4(OH)88+ whereas the results for uranium(IV) indicate the presence of U(OH)3+ and the polymeric species U4(OH)124+. The hydrolysis of both metal ions was studied at three temperatures allowing the determination of ΔH° and ΔS° of reaction for each species. The results were compared with previous results determined for thorium(IV) under the same experimental conditions to ascertain whether thorium should be used as an analogue for other tetravalent metal ions in safety assessment studies of nuclear repositories.

In order to adequately constrain surface complexation models for radionuclide sorption, large ‘multi-dimensional’ data-sets, showing the dependence of sorption on numerous experimental parameters, are required. This paper describes how large sorption databases were used to develop surface complexation models of U(VI) uptake on kaolinite and ferrihydrite. It is shown that the shape of U uptake curves as a function of pH and total U (εU) can be used to distinguish between U adsorption and the precipitation of amorphous uranyl hydroxide. The U sorption data-set obtained in this work also includes the effects of several complexing ligands, including phosphate. The dependence of U uptake on these variables can be used to assess the formation of various types of surface complexes (including ternary surface complexes) and the possibility of precipitation reactions. Adsorption experiments with different sorbing materials were also useful in developing models for U sorption and for excluding some possible surface and precipitation reactions. This study shows how combining multi-dimensional sorption databases with advanced spectroscopic or microscopic techniques (such as EXAFS or AEM) can lead to improved models of interactions at mineral-water interfaces.

When designing an underground repository for spent nuclear fuel it is important to know if and to what extent Np and Pu are reduced. At the moment only Np has been studied, but further investigations with Pu are planned.

Np in the pentavalent state and Th in the tetravalent state (234Th, about 10−9 M) have been used to study the sorption onto TiO2 (no reduction is expected) and onto UO2 with different atmospheres (air, nitrogen and nitrogen and hydrogen at 5 MPa pressure), and with different Np- concentrations (about 10−12M, 239Np and 10−6 M,237Np).

Th is sorbed to a high extent at pH>3 on both TiO2 and UO2. The sorption increases 3 orders of magnitude over one pH-unit when log Ka (the distribution coefficient with respect to surface area) is plotted against pH.

The Np sorption on TiO2 is independent of the Np concentration and shows a slope of one when pH is plotted against log Ka. The higher sorption of Np onto UO2 compared with TiO2 indicates that Np is partially reduced even when it is in contact with air. No large differences can be observed for Np between the nitrogen and hydrogen-atmosphere, indicating that the absence of oxygen is the important factor for the reduction. The different Np concentrations had no great impact on the reduction/sorption on UO2.

Absorption of aqueous solution by soil appears to result in piston-like displacement of the soil solution originally present by the invading solution. Diffusive redistribution of non-reactive solutes, relative to the water, occurs about that front when the solutions are of different chemical constitution. Adsorption of solute by the soil solid retards the solute front relative to the piston- front; it also reduces the effective diffusivity of the solute relative to the water. Such behavior tests prediction of radionuclide transport and retardation during non-steady water flow in unsaturated soil based on measured distribution coefficients.

Experiments, where initially uniform soil columns absorbed tritiated water containing 60Co2+ from a source at constant water potential, illustrate these effects. Columns were sampled at significantly different elapsed times and the transient water content, water soluble salts, tritium and cobalt profiles determined. Theory anticipates, and experiment confirms, self- similarity of these profiles when data is graphed in terms of distance divided by the square root of time. Water, and soil solid-based, material coordinates facilitate data analysis; measurement of absorption isotherms is also discussed.

Batch adsorption experiments with Cs-137 and Co-60 were undertaken using representative samples of geologic materials from the arid region that has been selected for an Australian low-level waste repository. The results indicate that the pH is the main factor affecting the adsorption of Co-60 but has little influence on the sorption of Cs-137. The ionic strength affects Cs sorption, with a decrease in Kd associated with higher ionic strength. Selective sorption sites on mica and illite control uptake of trace Cs, whereas the high total site availability of smectite is significant when the total Cs is higher (1 mmol/L). The effects of mineralogy on Cs sorption which were observed for these complex materials confirmed previous results reported for pure minerals.

The geostatistics of the main properties of transport in crystalline rock at Äspö Hard Rock Laboratory, Sweden have been determined experimentally, using a large number of samples, as a basis for the analysis of the effect of the uncertainty in rock heterogeneity on radionuclide migration in a single fracture. The auto-covariance functions were estimated for the porosity, effective diffusivity and adsorption characteristics, such as partition coefficient, kd and adsorption kinetics. The one-dimensional analytical solutions for the mean values of the temporal moments of the residence time probability density function (PDF) were derived by using the spectral method. The solutions were applied to the evaluation of migration experiments on granitic drill cores from the Äspö Hard Rock Laboratory. Fractures were created in drill cores by sawing and tritium was used as a non-sorbing tracer in several experiments in each drill core. By systematically varying the flow velocity, u, we could verify the non-linear influence of 1/u on the variance of the residence time PDF according to the theory. The effect of the uncertainty in the heterogeneous rock properties on the solute transport observed in the migration experiments corresponds fairly well to the theoretical effect using the auto-covariance functions.

The solubility of metallic selenium was measured in a mixture of 0.1M-NaCl and 0.05M-N2H4under anoxic conditions (O2 < 1 ppm) by both oversaturation and undersaturation methods. Equilibrium was attained in 40 days. The aqueous selenium species identified were HSe at pH between 5 and 8, and Se42at pH between 10 and 13, by UV-Vis absorption spectrometry. The solid phase was identified as Se (cr) by X-ray diffraction. The equilibrium constants of

Se(cr) + H+ + 2e- = HSe- logK0 = -6.5±0.5 and

4Se(cr) + 2e- = Se42- logK0 = -16.8±0.5

were determined. The standard molar free energy of formation of HSe- and Se42-was determined to be (37.1±2.9) and (95.9±2.9) kJ/mol, respectively.

This paper presents the results from an experimental study of the nickel solubility in sulfidic groundwater under anoxic conditions. The waters used were natural groundwater from Olkiluoto and synthetic saline groundwater, to which sulfide had been added. Two sulfide concentrations, 3.1.10−6 M and 9.4.10−5 M, were used. The possible influence of ferrous iron on the nickel concentration was also studied by adding small amounts of iron chloride, which yielded ferrous iron concentrations of 1.8.10−5 M or less.

The initial nickel concentrations in the water samples were 1.0.10−6 M and 1.0.10−3 M. The nickel concentrations were determined using 63Ni and liquid scintillation counting. The duration of the experiments was 177 days.

The results for the waters differed markedly. The nickel concentration in the natural groundwater decreased in all samples to roughly 10% of the initial value.

The nickel concentration in the synthetic groundwater was found to decrease in those samples having initial sulfide and nickel concentrations of 9.4.10−5 M and 1.0.10−6 M, respectively. The nickel concentration was in this case reduced to about 1% of its initial value. For the other synthetic groundwater samples, the nickel concentration remained at its original initial level.

The variation of the ferrous iron concentration did not have any observable effect on the measured nickel concentrations.

Nuclear fuel particles were separated, from two different Chernobyl contaminated soils, by sedimentation in bromoform. Their physico-chemical characteristics were studied by scanning electron microscopy. The two soils show the same types of fuel particles: one part of these particles contains U and O and the other part contains U, Zr and O. The structure and the surface morphology of the studied particles are variable and are characteristic of the conditions of their formation. The source term is thus heterogeneous. This heterogeneity has direct consequences on the dissolution of the fuel particles and should be taken into account in mathematical models.

Migration experiments of 241Am(III) have been performed using a column system, to investigate migration behavior of 241Am through a column packed with porous sedimentary materials: a coastal sandy soil and a reddish soil. Sorption mechanisms of 241Am on the sedimentary materials were examined by a chemical extraction method.

In the case of the reddish soil, most of 241Am sorbed on the influent edge of the column. The 241Am sorbed on the reddish soil was mainly controlled by a reversible ion exchange reaction. The migration of 241Am in the reddish soil could be evaluated by using the Kd concept. In the case of the sandy soil, on the other hand, considerable amount of particulate 241Am passed through the column. The 241Am sorbed on the sandy soil was controlled by irreversible reactions. The migration behavior of particulate 241Am in the sandy soil could be expressed by filtration theory.

The International Mont Terri Project [1, 2] started in 1995 under the patronage of the Swiss National Hydrological and Geological Survey (SNHGS), and has the authorization of the République et Canton du Jura. The underground rock laboratory is located at the northwestern part of Switzerland (Canton Jura), in and beside the reconnaissance gallery of the Mont Terri motorway tunnel, one of the several tunnels of the A16 “Transjurane” motorway. The depth of overburden above the rock laboratory is approximately 300 meters.

The project is aimed to investigate the geological, hydrogeological, geochemical and rock mechanical properties of the Opalinus Clay for assessing the feasibility and safety of a repository for radioactive waste placed in this type of host rock. One of the issues under study is radionuclide migration by diffusion through clays. As a part of this investigation, an interlaboratory comparison on small-scale diffusion experiments was carried out by three research laboratories: AEA Technology (UK), SCK·CEN (Belgium) and CIEMAT (Spain). The radionuclides investigated were tritium and iodine. This paper concerns to the methodological approach and results of the experiments undertaken by CIEMAT.

The effective diffusion coefficients were measured for tritiated water and iodine (as I-), resulting larger for tritium [(1.7±0.4)×10−11 m2/s] than for iodide [(2.7±0.3)×10−12 m2/s]. The porosity available for diffusion was calculated by using the time-lag method, but some results seemed unrealistic and showed a large variability. In general, tritium exhibited higher values of porosity than iodide (17 to 26% and 12 to 17%, respectively), which were consistent with the anion exclusion affecting the distribution of iodide into the clay pores.

Matrix diffusion laboratory experiments in dense porous rock are generally very time consuming and one is limited to rather short diffusion lengths, as well as to a small amount of samples. The large heterogeneity of rock, on the other hand, demands a large quantity of samples that are large enough to exclude effects from e.g. increases in interconnected porosity compared to that of the pristine rock.

Electrical conductivity measurements are very fast and larger samples can be used than is practical in ordinary diffusion experiments. The effective diffusivity of a non-charged molecule is readily evaluated from the measurements, and influences from surface conductivity on diffusion of cations can be studied.

In this study traditional through diffusion experiments as well as electrical conductivity measurements are carried out on the same rock samples. The formation factor is determined by both methods, and the methods are compared and discussed.

The surface conductivity is studied by exchanging the surface sites with Na+, Sr2+ and Cs+. After leaching out the free pore ions the surface conductivity is measured.

With the electrical conductivity method the formation factor is determined directly, whereas it has to be calculated using the bulk liquid diffusion coefficient in the diffusion experiments. This causes some uncertainties in the comparison between the experiments. In estimating the bulk liquid diffusivity, the value for infinitely diluted solutions and in pure water environment is commonly used. The calculated formation factor may therefore be somewhat underestimated.

Matrix diffusion in bedrock is traditionally studied by laboratory liquid diffusion experiments, which are time consuming and expensive. A new way of studying matrix diffusion is to measure the electrical resistivity of the rock. This could be done either in laboratories or in-situ. A fast method of obtaining a formation factor log, later used in matrix diffusion calculations, for an entire borehole is proposed.

It is a standard procedure in geophysical well logging to measure rock resistivity and there are well-developed tools for this in the oil industry. The SKB, Sweden, uses boreholes with a small diameter (56 mm) and this reduces the options in choosing resistivity tools. Therefore they have so far relied on the Normal log that only gives quantitative measurements in special cases, after corrections are made. Modern tools, such as the slimhole Dual-Laterolog, are accurate with a high vertical resolution.

The pore fluid resistivity is required when obtaining a formation factor log. In previous work the borehole fluid resistivity has been used. A new method measures the resistivity in groundwater from local fractures. These values seem more appropriate to use as new in-situ measurements show that there may be local groundwater resistivity differences.

A preliminary study shows that the conductivity could be obtained in saline Swedish groundwater by measuring the chloride concentration only.

Anomalies in the formation factor log have to be sorted out by using supporting core logging and non-electrical in-situ methods.

Distribution coefficients of 226Ra and 227Ac on soils and tuff have been obtained by a batch method. The distribution coefficients of Ra on loam, two kinds of sand and tuff were 28m3/kg, 0.45m3/kg, 0.19m3/kg and 1.9m3/kg, respectively. A linear relationship between the distribution coefficients of Ra on each soil and Cation Exchange Capacity(CEC) of each soil was obtained to suggest cation exchange sorption reaction of Ra2+ on soil surface. Sorption-migration behavior of Ra in sand column has been studied. Most of Ra remained in the column after passing of 5,000 pore volume of the column, showing several times the distribution coefficients of Ra than that by the batch method. The distribution coefficients of 227Ac on the soil samples were in the range from 1m3/kg to 10m3/kg.

In the present study, well-characterized mesoporous silica was used as a reference material for crystalline rock. Saturation-leaching and through-diffusion tracer experiments were systematically carried out to determine the adsorption capacity factors and apparent and effective diffusivities for the radioactive tracers 3H, 22Na and 36Cl. The ionic strength of the supporting NaCl electrolyte, in equilibrium with atmospheric air, was either 0.002 M or 0.1 M. The apparent diffusivities derived from saturation-leaching experiments reveal that equilibrium adsorption on a negatively charged silica surface does not act to retard the mass transfer of sodium ions. The steady-state diffusion results are found to demonstrate the increasing break-through for sodium with decreasing ionic strength of the background electrolyte in comparison to 3H. These observations are consistent with the picture of chloride ion exclusion and of the diffuse double layer surface-excess for sodium ions being transported in the direction of the macroscopic concentration gradient. Furthermore, they are shown to be at odds with the widely adopted macroscopic pore diffusion model, which neglects the mobility of the counter-ion surface- excess.