L'étude des mécanismes de sorption de l'ion uranyle sur le substrat SrTiO$_3$ en fonction de la température a fait l'objet de cette étude. Tout d'abord, une caractérisation physico¬chimique a été réalisée à l'aide de plusieurs techniques structurales (DRX, FTIR) et morphologique (MEB). La spectroscopie XPS a permis d'identifier deux sites de surface (Ti¬OH et Sr-OH). En utilisant les titrages potentiométriques de SrTiO$_3$ à différentes températures, les caractéristiques acido-basiques ont été déterminées. Ensuite, la simulation des titrages potentiométriques, entre 25 et 90°C, a été réalisée à l'aide du code FITEQL, les constantes d'équilibre ainsi obtenues montrent une nette variation avec la température: la protonation du site $\equiv Sr – OH$ suit un processus endothermique tandis que la déprotonation du site $\equiv Ti – OH$ implique un processus exothermique. A partir de ces constantes d'équilibre, les grandeurs thermodynamiques, enthalpie et entropie de protonation/déprotonation ont été calculées en utilisant la relation de van't Hoff. Les études de sorption de l'ion uranyle sur le substrat SrTiO$_3$ ont été réalisées dans un intervalle de pH de 0.5 à 5. Les sauts de sorption ainsi obtenus montrent une nette augmentation du pourcentage de sorption avec l'augmentation de la température, traduisant un phénomène globalement endothermique. Deux sites de sorption différents ont été identifiés à la surface du solide par SLRTIF. Ils sont associés aux temps de vie de fluorescence de l'uranyle sorbé de 12 $\pm$ 2 et 60 $\pm$ 5 $\mu$s. Les sauts de sorption ont été modélisés à l'aide du code FITEQL en utilisant le modèle à capacitance constante. Cette simulation des sauts de sorption a été réalisée en tenant compte des résultats de l'étude structurale (deux sites de surface $\equiv Sr - OH$ et $\equiv Ti – OH$ et formation de complexe surfacique de sphère interne bidendate, mononucléaire) et des données obtenues dans la modélisation des titrages potentiométriques. Les équilibres de sorption modélisés ont confirmé la formation de deux complexes de surface de caractère bidendate : [($\equiv SrOH)($\equivTiOH)UO_2]^{2+}$ et [($\equiv TiOH)($\equivTiO)UO_2]^{2+}$. Suite à l'obtention des constantes thermodynamiques obtenues par cette simulation, la relation van't Hoff a été appliquée pour déterminer les variations d'enthalpie et d'entropie associés au processus de sorption. Finalement, une étude sur les transferts d'énergie a été présentée entre deux ions sorbés sur le solide SrTiO$_3$. Ainsi, le transfert d'énergie non-radiatif des ions Tb$^{3+}$ vers les ions EU$^{3+}$ a été étudié. L'application du modèle de Inokuti-Hirayama et Dexter a conduit à l'évaluation du rayon de la sphère d'interaction (2,7-3,4 Å) entre les deux ions (Tb$^{3+}$ et EU$^{3+}$)

The long-lived fission product 126Sn (105 years) (Weast (1972)) is of substantial interest in the context of nuclear waste disposal in deep underground repositories. However, the prevalent redox state, the aqueous speciation as well as the reactions at the mineral-water interface under the expected anoxic conditions are a matter of debate. Therefore, in this PhD thesis I present work on the reactions of SnII with three Fe-bearing minerals as a function of pH, time, and SnII loading under anoxic condition with O2 level < 2 ppmv. The first mineral, goethite, contains only trivalent Fe (FeIIIOOH), the second, magnetite, contains both FeII and FeIII (FeIIFeIII2O4), and the third, mackinawite (FeIIS), contains only divalent Fe. The uptake behavior of the three mineral surfaces was investigated by batch sorption studies. Tin redox state was investigated by Sn-K X-ray absorption near-edge structure (XANES) spectroscopy, and the local, molecular structure of the expected Sn surface complexes and precipitates was studied by extended X-ray absorption fine-structure (EXAFS) spectroscopy. Selected samples were also investigated by transmission electron microscopy (TEM) to elucidate the existence and nature of secondary, Fe- and /or Sn containing solids, and by Mössbauer spectroscopy to study FeII and FeIII in the minerals. Based on the such-obtained molecular-level information, surface complexation models (SCM) were fitted to the batch sorption data to derive surface complexation constants. In the presence of the FeIII-bearing minerals magnetite and goethite, I observed a rapid uptake and oxidation of SnII to SnIV. The local structure determined by EXAFS showed two Sn-Fe distances of about 3.15 and 3.60 Å in line with edge and corner sharing arrangements between octahedrally coordinated SnIV and the Fe(O,OH)6 octahedra at the magnetite and goethite surfaces. While the respective coordination numbers suggested formation of tetradentate inner-sphere complexes between pH 3 and 9 for magnetite, bidentate inner-sphere complexes (single edge-sharing (1E) and corner-sharing (2C)) prevail at the goethite surface at pH > 3, with the relative amount of 2C increasing with Sn loading. The interfacial electron transfer between sorbed SnII and structural FeIII potentially leads to dissolution of FeII and transformation to secondary FeII/FeIII oxide minerals. There is no clear evidence to confirm the reductive dissolution in the Sn/ magnetite system, Rietveld refinement of XRD patterns, however, indicates an increase of FeII/FeIII ratio in the magnetite structure. For the Sn/goethite system, dissolved FeII increased with SnII loading at the lowest pH investigated, indicative of reductive dissolution. At pH >5, spherical and cubic particles of magnetite were observed by TEM, and their number increased with SnII loading. Based on previous finding, this secondary mineral transformation of goethite should proceed via dissolution and recrystallization. The molecular structure and oxidation state of sorbed Sn were then used to fit the batch sorption data of magnetite and goethite with SCM. The sorption data on magnetite were fit with the diffuse double layer model (DLM) employing two different complexes, the first ( = -14.97±0.35) prevailing from pH 2 to 9, and the second ( = -17.72±0.50), which forms at pH > 9 by co-adsorption of FeII, thereby increasing sorption at this high pH. The sorption data on goethite were fitted with the charge distribution–multisite complexation model (CD-MUSIC). Based on the EXAFS-derived presence of two different bidentate inner-sphere complexes ((≡FeOH)(≡Fe3O)Sn(OH)3 (1E) and (≡FeOH)2Sn(OH)3) (2C)), sorption affinity constants of 15.5 ±1.4 for the 1E complex and of 19.2 ±0.6 for the 2C complex were obtained. The model is not only able to predict sorption across the observed pH range, but also the transition from a roughly 50/50 distribution of the two complexes at 12.5 µmol/g Sn loading, to the prevalence of the 2C complex at higher loading, in line with the EXAFS data. The retention mechanism of SnII by mackinawite is significantly dependent on the solution pH, reflecting the transient changes of the mackinawite surface in the sorption process. At pH <7, SnII is retained in its original oxidation state. It forms a surface complex, which is characterized by two short (2.38 Å) Sn-S bonds, which can be interpreted as the bonds towards the S-terminated surface of mackinawite, and two longer Sn-S bonds (2.59 Å), which point most likely towards the solution phase, completing the tetragonal SnS4 innersphere sorption complex. Precipitation of SnS or formation of a solid solution with mackinawite could be excluded. At pH > 9, SnII is completely oxidized by an FeII/FeIII (hydr)oxide, most likely green rust, forming on the surface of mackinawite. Six O atoms at 2.04 Å and 6 Fe atoms at 3.29 Å demonstrate a structural incorporation by green rust, where SnIV substitutes for Fe in the crystal structure. The transition between SnII and SnIV and between sulfur and oxygen coordination takes place between pH 7 and 8, in accordance with the transition from the mackinawite stability field to more oxidized Fe-bearing minerals. The uptake processes of SnII by mackinawite are largely in line with the uptake processes of divalent cations of other soft Lewis-acid metals like Cd, Hg and Pb. Very different Sn retention mechanisms were hence active, including oxidation to SnIV and formation of tetradentate and bidentate surface complexes of the SnIV hydroxo moieties on goethite and magnetite, and in the case of mackinawite a SnII sulfide species forming a bidentate surface complex at low pH, and structural incorporation of SnIV by an oxidation product, green rust, at high pH. In all three mineral systems and largely independent on the retention mechanisms, inorganic SnII was strongly retained, with Rd values always exceeding 5, across the relatively wide pH range relevant for the near and far-field of nuclear waste respositories. For the goethite and magnetite systems, the retention could be well modeled with surface complexation models based on the molecular structural data. This is an important contribution to the safety case for future nuclear waste repositories, since such SCMs provide reliable means for predicting the radioactive dose released by 126Sn from nuclear waste into the biosphere across a wide range of physicochemical conditions typical for the engineered as well as natural barriers.

Canada is in the process of implementing a Deep Geologic Repository (DGR) to dispose of used nuclear waste. Adsorption behaviour of both Se(-II) and Tc(IV) onto granite, shale, limestone, illite, and MX-80 bentonite has been elucidated. Se(-II) adsorption onto granite and MX-80 bentonite displays a decrease in Rd with an increase in solution pH. Se(-II) adsorption onto granite decreases with an increase in solution ionic strength. Se(-II) adsorption onto MX-80 bentonite does not return evidence which supports an apparent effect due to the ionic strength. Tc(IV) adsorption onto shale, limestone, illite, and MX-80 bentonite remains constant as the solution pH increases. Ionic strength does not affect the magnitude of Tc(IV) adsorption across the adsorbents, however an increase in ionic strength accelerates Tc(IV) adsorption. Se(-II) surface complexation models are best simulated with the following surface complexes: ≡Feldspar_sSe-, ≡Biotite_sOH2HSe, ≡Albite_sSe-, ≡Montmorillonite_sSe-, and ≡Montmorillonite_sOH2HSe. Tc(IV) adsorption is best simulated with: ≡Biotite_sOTcO(OH), ≡Quartz_sOTcO(OH), (≡Feldspar_sOH)2TcO(OH)-, ≡Montmorillonite_sOTcO(OH), (≡Albite_sOH)2TcO(OH)-, ≡Illite_sOTcO(OH), and ≡Chlorite_sOTcO(OH). Se(-II) adsorption onto granite and MX-80 bentonite in CR-10 solution returns Rd values of (1.80 ± 0.10) m3∙kg-1 and (0.47 ± 0.38) m3∙kg-1, respectively. Tc(IV) adsorption onto granite and MX-80 bentonite in CR-10 solution returned Rd values of (1.47 ± 0.25) m3∙kg-1 and (2.19 ± 0.33) m3∙kg-1, respectively. Tc(IV) adsorption onto shale, limestone, illite, and MX-80 bentonite in SR-270-PW solution returned Rd values of (0.16 ± 0.10) m3∙kg-1, (0.44 ± 0.21) m3∙kg-1, (1.86 ± 0.44) m3∙kg-1, and (0.23 ± 0.10) m3∙kg-1, respectively. This thesis will further deepen the understanding of Se(-II) and Tc(IV) adsorption. / Thesis / Doctor of Philosophy (PhD) / Determining the adsorption of Se(-II) and Tc(IV) onto granite, shale, limestone, illite, and MX-80 bentonite is beneficial to choosing a location within Canada to locate a used nuclear fuel repository. This thesis aims to quantify the adsorption behaviour of Se(-II) and Tc(IV) in Ca-Na-Cl and Na-Ca-Cl solutions with respect to a varying solution ionic strength and pH. Quantification of the adsorption was accomplished with adsorption experiments used in conjunction with geochemical simulations. New simulated surfaces specific to granite, shale, and MX-80 bentonite have been developed to complete these simulations. A final achievement was quantifying the adsorption of Se(-II) and Tc(IV) in groundwater representative solutions specific to locations considered for the used nuclear fuel repository.

The objective of the work described in this thesis was to understand sorption reactions of uranium occurring at the water-clay mineral interfaces in the presence and absence of arsenic and other inorganic ligands. Uranium(VI) removal by clay minerals is influenced by a large number of factors including: type of clay mineral, pH, ionic strength, partial pressure of CO2, load of the sorbent, total amount of U present, and the presence of arsenate and other inorganic ligands such as sulfate, carbonate, and phosphate. Both sulfate and carbonate reduced uranium sorption onto IBECO bentonite due to the competition between SO42- or CO32- ions and the uranyl ion for sorption sites, or the formation of uranyl-sulfate or uranyl-carbonate complexes. Phosphate is a successful ligand to promote U(VI) removal from the aqueous solution through formation of ternary surface complexes with a surface site of bentonite. In terms of the type of clay mineral used, KGa-1b and KGa-2 kaolinites showed much greater uranium sorption than the other clay minerals (STx-1b, SWy-2, and IBECO montmorillonites) due to more aluminol sites available, which have higher affinity toward uranium than silanol sites. Sorption of uranium on montmorillonites showed a distinct dependency on sodium concentrations because of the effective competition between uranyl and sodium ions, whereas less significant differences in sorption were found for kaolinite. A multisite layer surface complexation model was able to account for U uptake on different clay minerals under a wide range of experimental conditions. The model involved eight surface reactions binding to aluminol and silanol edge sites of montmorillonite and to aluminol and titanol surface sites of kaolinite, respectively. The sorption constants were determined from the experimental data by using the parameter estimation code PEST together with PHREEQC. The PEST- PHREEQC approach indicated an extremely powerful tool compared to FITEQL. In column experiments, U(VI) was also significantly retarded due to adsorptive interaction with the porous media, requiring hundreds of pore volumes to achieve breakthrough. Concerning the U(VI) desorption, columns packed with STx-1b and SWy-2 exhibited irreversible sorption, whereas columns packed with KGa-1b and KGa-2 demonstrated slow, but complete desorption. Furthermore, most phenomena observed in batch experiments were recognized in the column experiments, too. The affinity of uranium to clay minerals was higher than that of arsenate. In systems containing uranium and arsenate, the period required to achieve the breakthrough in all columns was significantly longer when the solution was adjusted to pH 6, due to the formation of the uranyl-arsenate complex. In contrast, when pH was adjusted to 3, competitive sorption for U(VI) and As(V) accelerated the breakthrough for both elements. Finally, experiments without sorbing material conducted for higher concentrations of uranium and arsenic showed no loss of total arsenic and uranium in non-filtered samples. In contrast, significant loss was observed after filtration probably indicating the precipitation of a U/As 1:1 phase.

The long-lived fission product 126Sn (105 years) (Weast (1972)) is of substantial interest in the context of nuclear waste disposal in deep underground repositories. However, the prevalent redox state, the aqueous speciation as well as the reactions at the mineral-water interface under the expected anoxic conditions are a matter of debate. Therefore, in this PhD thesis I present work on the reactions of SnII with three Fe-bearing minerals as a function of pH, time, and SnII loading under anoxic condition with O2 level < 2 ppmv. The first mineral, goethite, contains only trivalent Fe (FeIIIOOH), the second, magnetite, contains both FeII and FeIII (FeIIFeIII2O4), and the third, mackinawite (FeIIS), contains only divalent Fe. The uptake behavior of the three mineral surfaces was investigated by batch sorption studies. Tin redox state was investigated by Sn-K X-ray absorption near-edge structure (XANES) spectroscopy, and the local, molecular structure of the expected Sn surface complexes and precipitates was studied by extended X-ray absorption fine-structure (EXAFS) spectroscopy. Selected samples were also investigated by transmission electron microscopy (TEM) to elucidate the existence and nature of secondary, Fe- and /or Sn containing solids, and by Mössbauer spectroscopy to study FeII and FeIII in the minerals. Based on the such-obtained molecular-level information, surface complexation models (SCM) were fitted to the batch sorption data to derive surface complexation constants. In the presence of the FeIII-bearing minerals magnetite and goethite, I observed a rapid uptake and oxidation of SnII to SnIV. The local structure determined by EXAFS showed two Sn-Fe distances of about 3.15 and 3.60 Å in line with edge and corner sharing arrangements between octahedrally coordinated SnIV and the Fe(O,OH)6 octahedra at the magnetite and goethite surfaces. While the respective coordination numbers suggested formation of tetradentate inner-sphere complexes between pH 3 and 9 for magnetite, bidentate inner-sphere complexes (single edge-sharing (1E) and corner-sharing (2C)) prevail at the goethite surface at pH > 3, with the relative amount of 2C increasing with Sn loading. The interfacial electron transfer between sorbed SnII and structural FeIII potentially leads to dissolution of FeII and transformation to secondary FeII/FeIII oxide minerals. There is no clear evidence to confirm the reductive dissolution in the Sn/ magnetite system, Rietveld refinement of XRD patterns, however, indicates an increase of FeII/FeIII ratio in the magnetite structure. For the Sn/goethite system, dissolved FeII increased with SnII loading at the lowest pH investigated, indicative of reductive dissolution. At pH >5, spherical and cubic particles of magnetite were observed by TEM, and their number increased with SnII loading. Based on previous finding, this secondary mineral transformation of goethite should proceed via dissolution and recrystallization. The molecular structure and oxidation state of sorbed Sn were then used to fit the batch sorption data of magnetite and goethite with SCM. The sorption data on magnetite were fit with the diffuse double layer model (DLM) employing two different complexes, the first ( = -14.97±0.35) prevailing from pH 2 to 9, and the second ( = -17.72±0.50), which forms at pH > 9 by co-adsorption of FeII, thereby increasing sorption at this high pH. The sorption data on goethite were fitted with the charge distribution–multisite complexation model (CD-MUSIC). Based on the EXAFS-derived presence of two different bidentate inner-sphere complexes ((≡FeOH)(≡Fe3O)Sn(OH)3 (1E) and (≡FeOH)2Sn(OH)3) (2C)), sorption affinity constants of 15.5 ±1.4 for the 1E complex and of 19.2 ±0.6 for the 2C complex were obtained. The model is not only able to predict sorption across the observed pH range, but also the transition from a roughly 50/50 distribution of the two complexes at 12.5 µmol/g Sn loading, to the prevalence of the 2C complex at higher loading, in line with the EXAFS data. The retention mechanism of SnII by mackinawite is significantly dependent on the solution pH, reflecting the transient changes of the mackinawite surface in the sorption process. At pH <7, SnII is retained in its original oxidation state. It forms a surface complex, which is characterized by two short (2.38 Å) Sn-S bonds, which can be interpreted as the bonds towards the S-terminated surface of mackinawite, and two longer Sn-S bonds (2.59 Å), which point most likely towards the solution phase, completing the tetragonal SnS4 innersphere sorption complex. Precipitation of SnS or formation of a solid solution with mackinawite could be excluded. At pH > 9, SnII is completely oxidized by an FeII/FeIII (hydr)oxide, most likely green rust, forming on the surface of mackinawite. Six O atoms at 2.04 Å and 6 Fe atoms at 3.29 Å demonstrate a structural incorporation by green rust, where SnIV substitutes for Fe in the crystal structure. The transition between SnII and SnIV and between sulfur and oxygen coordination takes place between pH 7 and 8, in accordance with the transition from the mackinawite stability field to more oxidized Fe-bearing minerals. The uptake processes of SnII by mackinawite are largely in line with the uptake processes of divalent cations of other soft Lewis-acid metals like Cd, Hg and Pb. Very different Sn retention mechanisms were hence active, including oxidation to SnIV and formation of tetradentate and bidentate surface complexes of the SnIV hydroxo moieties on goethite and magnetite, and in the case of mackinawite a SnII sulfide species forming a bidentate surface complex at low pH, and structural incorporation of SnIV by an oxidation product, green rust, at high pH. In all three mineral systems and largely independent on the retention mechanisms, inorganic SnII was strongly retained, with Rd values always exceeding 5, across the relatively wide pH range relevant for the near and far-field of nuclear waste respositories. For the goethite and magnetite systems, the retention could be well modeled with surface complexation models based on the molecular structural data. This is an important contribution to the safety case for future nuclear waste repositories, since such SCMs provide reliable means for predicting the radioactive dose released by 126Sn from nuclear waste into the biosphere across a wide range of physicochemical conditions typical for the engineered as well as natural barriers.

The objective of the work described in this thesis was to understand sorption reactions of uranium occurring at the water-clay mineral interfaces in the presence and absence of arsenic and other inorganic ligands. Uranium(VI) removal by clay minerals is influenced by a large number of factors including: type of clay mineral, pH, ionic strength, partial pressure of CO2, load of the sorbent, total amount of U present, and the presence of arsenate and other inorganic ligands such as sulfate, carbonate, and phosphate. Both sulfate and carbonate reduced uranium sorption onto IBECO bentonite due to the competition between SO42- or CO32- ions and the uranyl ion for sorption sites, or the formation of uranyl-sulfate or uranyl-carbonate complexes. Phosphate is a successful ligand to promote U(VI) removal from the aqueous solution through formation of ternary surface complexes with a surface site of bentonite. In terms of the type of clay mineral used, KGa-1b and KGa-2 kaolinites showed much greater uranium sorption than the other clay minerals (STx-1b, SWy-2, and IBECO montmorillonites) due to more aluminol sites available, which have higher affinity toward uranium than silanol sites. Sorption of uranium on montmorillonites showed a distinct dependency on sodium concentrations because of the effective competition between uranyl and sodium ions, whereas less significant differences in sorption were found for kaolinite. A multisite layer surface complexation model was able to account for U uptake on different clay minerals under a wide range of experimental conditions. The model involved eight surface reactions binding to aluminol and silanol edge sites of montmorillonite and to aluminol and titanol surface sites of kaolinite, respectively. The sorption constants were determined from the experimental data by using the parameter estimation code PEST together with PHREEQC. The PEST- PHREEQC approach indicated an extremely powerful tool compared to FITEQL. In column experiments, U(VI) was also significantly retarded due to adsorptive interaction with the porous media, requiring hundreds of pore volumes to achieve breakthrough. Concerning the U(VI) desorption, columns packed with STx-1b and SWy-2 exhibited irreversible sorption, whereas columns packed with KGa-1b and KGa-2 demonstrated slow, but complete desorption. Furthermore, most phenomena observed in batch experiments were recognized in the column experiments, too. The affinity of uranium to clay minerals was higher than that of arsenate. In systems containing uranium and arsenate, the period required to achieve the breakthrough in all columns was significantly longer when the solution was adjusted to pH 6, due to the formation of the uranyl-arsenate complex. In contrast, when pH was adjusted to 3, competitive sorption for U(VI) and As(V) accelerated the breakthrough for both elements. Finally, experiments without sorbing material conducted for higher concentrations of uranium and arsenic showed no loss of total arsenic and uranium in non-filtered samples. In contrast, significant loss was observed after filtration probably indicating the precipitation of a U/As 1:1 phase.