The objective of the work described in this thesis was to improve the scientific basis for modelling the migration of U in the sub-surface environment. The project involved: ?? studying the sorption of U on model minerals (Georgia kaolinite and ferrihydrite) in laboratory experiments ?? carrying out experimental studies of U sorption on complex natural substrates ?? studying the mechanisms influencing U retardation in the natural environment, including transformation processes of iron oxides ?? identifying chemical factors which control U sorption on model and natural substrates ?? developing a mechanistic model for U sorption on ferrihydrite and kaolinite using the surface complexation adsorption model , and ?? assessing and modelling the effect of complexing ligands on uranyl adsorption. Uranium (VI) sorption on geological materials is influenced by a large number of factors including: pH, ionic strength, partial pressure of CO2, adsorbent loading, total amount of U present, and the presence of inorganic and organic ligands. The sorption of UO22+ typically increases with increasing pH (the 'low pH sorption edge') up to about pH 7. In systems equilibrated with air, there is a sharp decrease in sorption above this pH value (the 'high pH edge'), due to strong complexation between uranyl and carbonate. The adsorption model being used for ferrihydrite is a surface complexation model with a diffuse double layer, and both strong and weak sites for U sorption. Based on the analysis of EXAFS data, the U surface complexes were modelled as mononuclear bidentate surface complexes of the form (&gtFeO2)UO20. Ternary surface complexes involving carbonate with the form (&gtFeO2)UO2CO32- were also required for the best simulation of U sorption data. There was a slight decrease in U sorption on ferrihydrite in systems that contained sulfate. It was necessary to consider competition between UO22+ and SO42- for surface sites, as well as complexation between UO22+ and SO42- to model the data. The presence of citrate considerably reduced U sorption and caused dissolution of ferrihydrite. Complexation of citrate with both uranyl and ferric ions was taken into account in modelling this system. The model required the optimisation of the formation constant for a postulated mixed metal (U/Fe/citrate) aqueous complex. Humic acid increased U uptake at pH values below 7, with little effect at higher pH values. In terms of the amount of U sorbed per gram of adsorbent, U uptake on kaolinites KGa-1 and KGa-1B was much weaker than U uptake on ferrihydrite under similar experimental conditions. Electron microscope examination showed that titanium-rich impurity phases played a major role in the sorption of U by these standard kaolinites. A relatively simple model for uranyl sorption on the model kaolinites was able to account for U sorption under a wide range of experimental conditions. The model involved only three surface reactions on two sites (&gtTiOH and &gtAlOH), with a non-electrostatic surface complexation model. The relative amounts of the sites were estimated from AEM results. Precipitation was taken into account in modelling the experimental data obtained with high U concentrations. The effects of sulfate and citrate on U sorption by kaolinite were also assessed and modelled. Sulfate had a small effect on U sorption, which may be explained by aqueous complexation. Citrate had a greater effect, and this was not wholly explained by the formation of aqueous U-citrate complexes. The most likely explanation would also involve competition for surface sites between U and citrate. Uranyl uptake on ferrihydrite was greatly increased by the presence of phosphate. This was not due to precipitation, and was attributed to the formation of a ternary surface complex with a proposed structure of (&gtFeO2)UO2PO43-. The log K value for the formation of this complex was optimised using FITEQL. Phosphate also increased uptake of uranyl on kaolinite, and this was also attributed to the formation of ternary uranyl phosphate surface species. Uranium sorption on weathered schist samples from the vicinity of the Koongarra U deposit in northern Australia was generally similar to the model minerals (in terms of the effects of pH, ionic strength, total U, etc). Many experiments with the natural materials were spiked with an artificial U isotope (236U), which allowed adsorption (of 236U) and desorption (of 238U) to be distinguished, and provided a means of estimating the 'labile' or 'accessible' portion of the natural U content. A significant advantage of this method is that (unlike chemical extractions) it does not rely on the assumptions about the phases extracted by 'selective' reagents. Uranium sorption experiments were also carried out with Koongarra samples which had been treated with citrate / dithionite / bicarbonate (CDB) reagent to remove iron oxides. Uranium sorption was greatly decreased by the CDB extraction, which reduced the surface area of the samples by about 30-40%. To further elucidate the impact of iron minerals on U mobility in the natural environment, the transformation of synthetic ferrihydrite containing adsorbed natural uranium was studied. In these experiments, the ferrihydrite was partially converted to crystalline forms such as hematite and goethite. The uptake of an artificial uranium isotope (236U) and the leaching of 238U from the samples were then studied in adsorption / desorption experiments. The transformation of ferrihydrite to crystalline minerals substantially reduced the ability of the samples to adsorb 236U from solution. Some of the previously adsorbed 238U was irreversibly incorporated within the mineral structure during the transformation process. Therefore, transformation of iron minerals from amorphous to crystalline forms provides a possible mechanism for uranium immobilisation in the groundwater environment. In considering the overall effect on U migration, this must be balanced against the reduced ability of the transformed iron oxide to adsorb U. The experiments with the model and natural substrates demonstrated that trace impurities (such as Ti-oxides) and mineral coatings (such as ferrihydrite) can play a dominant role in U adsorption in both environmental and model systems. Although the various substrates had different affinities for adsorbing U, the effects of chemical factors, including pH, ionic strength, and carbonate complexation were similar for the different materials. This suggests that a mechanistic model for U sorption on model minerals may eventually be incorporated in geochemical transport models, and used to describe U sorption in the natural environment.

The objective of the work described in this thesis was to improve the scientific basis for modelling the migration of U in the sub-surface environment. The project involved: ?? studying the sorption of U on model minerals (Georgia kaolinite and ferrihydrite) in laboratory experiments ?? carrying out experimental studies of U sorption on complex natural substrates ?? studying the mechanisms influencing U retardation in the natural environment, including transformation processes of iron oxides ?? identifying chemical factors which control U sorption on model and natural substrates ?? developing a mechanistic model for U sorption on ferrihydrite and kaolinite using the surface complexation adsorption model , and ?? assessing and modelling the effect of complexing ligands on uranyl adsorption. Uranium (VI) sorption on geological materials is influenced by a large number of factors including: pH, ionic strength, partial pressure of CO2, adsorbent loading, total amount of U present, and the presence of inorganic and organic ligands. The sorption of UO22+ typically increases with increasing pH (the 'low pH sorption edge') up to about pH 7. In systems equilibrated with air, there is a sharp decrease in sorption above this pH value (the 'high pH edge'), due to strong complexation between uranyl and carbonate. The adsorption model being used for ferrihydrite is a surface complexation model with a diffuse double layer, and both strong and weak sites for U sorption. Based on the analysis of EXAFS data, the U surface complexes were modelled as mononuclear bidentate surface complexes of the form (&gtFeO2)UO20. Ternary surface complexes involving carbonate with the form (&gtFeO2)UO2CO32- were also required for the best simulation of U sorption data. There was a slight decrease in U sorption on ferrihydrite in systems that contained sulfate. It was necessary to consider competition between UO22+ and SO42- for surface sites, as well as complexation between UO22+ and SO42- to model the data. The presence of citrate considerably reduced U sorption and caused dissolution of ferrihydrite. Complexation of citrate with both uranyl and ferric ions was taken into account in modelling this system. The model required the optimisation of the formation constant for a postulated mixed metal (U/Fe/citrate) aqueous complex. Humic acid increased U uptake at pH values below 7, with little effect at higher pH values. In terms of the amount of U sorbed per gram of adsorbent, U uptake on kaolinites KGa-1 and KGa-1B was much weaker than U uptake on ferrihydrite under similar experimental conditions. Electron microscope examination showed that titanium-rich impurity phases played a major role in the sorption of U by these standard kaolinites. A relatively simple model for uranyl sorption on the model kaolinites was able to account for U sorption under a wide range of experimental conditions. The model involved only three surface reactions on two sites (&gtTiOH and &gtAlOH), with a non-electrostatic surface complexation model. The relative amounts of the sites were estimated from AEM results. Precipitation was taken into account in modelling the experimental data obtained with high U concentrations. The effects of sulfate and citrate on U sorption by kaolinite were also assessed and modelled. Sulfate had a small effect on U sorption, which may be explained by aqueous complexation. Citrate had a greater effect, and this was not wholly explained by the formation of aqueous U-citrate complexes. The most likely explanation would also involve competition for surface sites between U and citrate. Uranyl uptake on ferrihydrite was greatly increased by the presence of phosphate. This was not due to precipitation, and was attributed to the formation of a ternary surface complex with a proposed structure of (&gtFeO2)UO2PO43-. The log K value for the formation of this complex was optimised using FITEQL. Phosphate also increased uptake of uranyl on kaolinite, and this was also attributed to the formation of ternary uranyl phosphate surface species. Uranium sorption on weathered schist samples from the vicinity of the Koongarra U deposit in northern Australia was generally similar to the model minerals (in terms of the effects of pH, ionic strength, total U, etc). Many experiments with the natural materials were spiked with an artificial U isotope (236U), which allowed adsorption (of 236U) and desorption (of 238U) to be distinguished, and provided a means of estimating the 'labile' or 'accessible' portion of the natural U content. A significant advantage of this method is that (unlike chemical extractions) it does not rely on the assumptions about the phases extracted by 'selective' reagents. Uranium sorption experiments were also carried out with Koongarra samples which had been treated with citrate / dithionite / bicarbonate (CDB) reagent to remove iron oxides. Uranium sorption was greatly decreased by the CDB extraction, which reduced the surface area of the samples by about 30-40%. To further elucidate the impact of iron minerals on U mobility in the natural environment, the transformation of synthetic ferrihydrite containing adsorbed natural uranium was studied. In these experiments, the ferrihydrite was partially converted to crystalline forms such as hematite and goethite. The uptake of an artificial uranium isotope (236U) and the leaching of 238U from the samples were then studied in adsorption / desorption experiments. The transformation of ferrihydrite to crystalline minerals substantially reduced the ability of the samples to adsorb 236U from solution. Some of the previously adsorbed 238U was irreversibly incorporated within the mineral structure during the transformation process. Therefore, transformation of iron minerals from amorphous to crystalline forms provides a possible mechanism for uranium immobilisation in the groundwater environment. In considering the overall effect on U migration, this must be balanced against the reduced ability of the transformed iron oxide to adsorb U. The experiments with the model and natural substrates demonstrated that trace impurities (such as Ti-oxides) and mineral coatings (such as ferrihydrite) can play a dominant role in U adsorption in both environmental and model systems. Although the various substrates had different affinities for adsorbing U, the effects of chemical factors, including pH, ionic strength, and carbonate complexation were similar for the different materials. This suggests that a mechanistic model for U sorption on model minerals may eventually be incorporated in geochemical transport models, and used to describe U sorption in the natural environment.