Les travaux présentés dans ce mémoire sont consacrés à l'étude des processus d'adsorption des anions phosphate par la rouille verte ferrique carbonatée et la ferrihydrite. L'objectif final vise une application au traitement des eaux usées en milieu rural. La synthèse des deux oxydes de fer a été réalisée et leurs propriétés physico-chimiques caractérisées. Un intérêt particulier a été consacré à comparer les propriétés structurales de la rouille verte ferrique en fonction de différents paramètres de synthèse. Une étude approfondie des propriétés physico-chimiques de surface de la ferrihydrite par spectroscopie de photoélectrons X a quant à elle été réalisée. La réactivité de ces deux oxydes a ensuite été abordée en mode discontinu où l'équation cinétique du pseudo-second ordre et le modèle d'isotherme de Freundlich offrent les meilleurs ajustements. L'influence de divers paramètres a été prise en compte comme le pH, la force ionique, etc. Le mode continu a été envisagé sur un matériau de filtration constitué de l'oxyde de fer déposé sur de la pouzzolane. La méthode de fabrication ainsi que les conditions optimales de préparation du matériau de filtration ont été déterminées. Les mécanismes d'adsorption en condition de flux hydrodynamique ont alors mis en évidence des phénomènes advectifs, diffusifs et une régionalisation de l'eau régissant l'adsorption au sein de la colonne. Des informations telles que les capacités d'adsorption ou l'influence du débit sur le processus d'adsorption ont pu être également obtenues. Une expérience préliminaire sur une eau usée prétraitée met finalement en évidence une quantité adsorbée particulièrement intéressante pour une application industrielle potentielle / The Ph.D. work, presented in this manuscript, is devoted to evaluating phosphate adsorption process on carbonate ferric green rust and ferrihydrite. The main objective concerns an application for the treatment of waste water in rural areas. Both iron oxides are first synthesized and their physico-chemical properties characterized. The ferric green rust structural properties differences as a function of synthesis parameters such as aging period and addition of hydrogen peroxide solution is of particular interest. A detailed study of surface physico-chemical properties by X Photoelectron Spectroscopy is carried out in the case of ferrihydrite. The reactivity of these two iron oxides is then evaluated in batch experiments. Adsorption process follows the pseudo-second order kinetic equation and Freundlich isotherm model which give the best adjustments of experimental data. The influence of various parameters such as pH, ionic strength, etc on phosphate adsorption is also taken into account. Column experiments are afterwards carried out by using filtration material constituted of iron oxide deposited onto pozzolana. The optimal conditions to prepare this filtration material are naturally predetermined. Phosphate adsorption in hydrodynamic mode is characterized by advective and diffusive mechanisms and water regionalization which govern the adsorption process in the column. Moreover, phosphate adsorption capacity and flow rate influence on adsorption process are obtained. Finally, a preliminary experiment on a pre-treated waste water finally shows that the filtration material is potentially interesting for an industrial application

Les teneurs en pesticides mesurées dans les eaux de drainage agricole peuvent atteindre 10 µg/L voire 395 µg/L. En Lorraine, des Zones Tampons Végétalisées Agricoles (ZTVA) ont été installées en sortie de drains agricoles afin de limiter le transfert de ces eaux de drainage vers la rivière. L’objectif de ce travail était d’évaluer l’influence de l’hydraulique sur l’épuration des pesticides et métabolites en phases dissoute et particulaire des eaux de drainage, en discriminant les processus associés. Pour cela, une approche multi-échelle intégrant des expérimentations au laboratoire en batch et en bacs pilotes ainsi qu’un monitoring pluriannuel de deux ZTVA sur le terrain (un fossé et une mare) a été mise en place. Les expériences de traçages in-situ ont montré que quel que soit le débit, la ZTVA peut être divisée en différentes zones hydrauliques : un chenal principal et des zones isolées, qui se comportent différemment vis-à-vis de l’épuration. Les efficacités d’épuration dans les deux ZTVA varient de (i) -1176% à 96% pour les pesticides dissous, (ii) -20% à 3% pour les métabolites (chloroacétanilides) à l’état dissous, et (iii) -580% à 79% pour les pesticides sous forme particulaire. L’adsorption sur les sédiments permet l’épuration des pesticides à l’état dissous dont le coefficient d’adsorption (Koc) varie de 364 à 1424 L/g (entre 7 et 65% d’épuration), et est favorisé au sein des zones isolées. Cependant, ce processus est réversible et la désorption peut expliquer les efficacités négatives mesurées sur le terrain. Par ailleurs, les pesticides plus hydrophiles (Koc varie de 54 à 401 L/g) et les métabolites (Koc varie de 0 à 0.77 L/g) sont peu épurés (entre -20 et 8% d’épuration). Enfin, les pesticides entrant dans la ZTVA sous forme particulaire sont épurés via la sédimentation des matières en suspension, plus importante dans les zones isolées que le chenal principal. Ce processus est réversible, pouvant conduire à la resuspension des sédiments. Par ailleurs, les débits entrants vont fluctuer au cours de l’année, pouvant entrainer une variation de l’épuration des pesticides. En effet, les études en batch et en bacs pilotes ont permis de souligner l’influence de l’hydrodynamique (débits, etc) sur l’épuration des pesticides à l’état dissous. Ces ZTVA jouent donc un rôle de puits (adsorption, sédimentation) et de source (désorption, resuspension) vis-à-vis de certains pesticides à l’état dissous ou particulaire, en lien avec l’hydrodynamique de la ZTVA / Pesticides amounts measured in agricultural drained water can reach 10 µg/L up to 395 µg/L. In Lorraine, Constructed Wetlands (CW) were set up between drained fields and the river to limit pesticide release. The aim of this study was to evaluate the influence of hydraulic on the mitigation of pesticides and metabolites in both dissolved and particulate phases of drained water, by discriminating associated processes. To do so, a multi-scale approach was performed by integrating both laboratory experiments, such as batch and dynamic conditions in pilots, and a plurennial monitoring of two different ZTVA (ditch and pond). In-situ tracing experiments highlighted that the volume of CW was not homogeneous, independently of the flow rate. CW are divided in two hydraulic zones: a main channel and isolated areas. Moreover, these two zones behave differently regarding pesticides mitigation. Annual mitigation efficiency in both of the CW studied, vary between (i) -1176 % and 96 % for dissolved pesticides, (ii) -20 % and 3 % for dissolved metabolites (chloroacetanilides), and from (iii) -580 % to 79 % for particulate pesticides. Adsorption on sediments allows the mitigation of dissolved pesticides whose adsorption coefficient (Koc) varied from 364 to 1424 L/g (mitigation ranging from 7 to 65 %), and occurred mainly in isolated areas. However, this process is reversible and desorption can explain negative efficiency measured on the field. Additionally, hydrophilic pesticides (Koc between 54 and 401 L/g) and metabolites (Koc between 0 and 0.77 L/g) are few or not mitigated (mitigation ranging from -20 and 8 %). Finally, pesticides entering CW under particulate phase are mitigated through sedimentation of total suspended solids, higher in isolated areas than in main channel. This process is also reversible, leading to sediments resuspension. Otherwise, inlet flow rates vary throughout the year, which could allow a variation of pesticide mitigation. Indeed, batch and pilots studies highlighted the influence of hydrodynamic (flow rate, etc) on mitigation of dissolved pesticides. CW act as a sink (adsorption and sedimentation) and a source (desorption and resuspension) towards specific dissolved or particulate pesticides and related to hydrodynamic of CW

Currently at the Savannah River Site (SRS), there are fifteen single-shell, 3.6-million liter tanks containing High Level Waste. To close the tanks, the sludge must be removed. Mechanical methods have had limited success. Oxalic acid cleaning is now being considered as a new technology. This research uses sample results and chemical equilibrium software to develop a preferred flowsheet and evaluate the acceptability of the system impacts. Based on modeling and testing, between 246,000 to 511,000 l of 8 wt% oxalic acid were required to dissolve a 9,000 liter Purex sludge heel. For SRS H-Area modified sludge, 322,000 to 511,000 l were required. To restore the pH of the treatment tank slurries, approximately 140,000 to 190,000 l of 50 wt% NaOH or 260,000 to 340,000 l of supernate were required. When developing the flowsheet, there were two primary goals to minimize downstream impacts. The first was to ensure that the Resultant oxalate solids were transferred to DWPF, without being washed. The second was to transfer the remaining soluble sodium oxalates to the evaporator drop tank, so they do not transfer through or precipitate in the evaporator pot. Adiabatic modeling determined the maximum possible temperature to be 73.5°C and the maximum expected temperature to be 64.6°C. At one atmosphere and at 73.5°C, a maximum of 770 l of water vapor was generated, while at 64.6°C a maximum 254 l of carbon dioxide were generated. Although tank wall corrosion was not a concern, because of the large cooling coil surface area, the corrosion induced hydrogen generation rate was calculated to be as high as 10,250 l/hr. Since the minimum tank purge exhaust was assumed to be 5,600 l/hr, the corrosion induced hydrogen generation rate was identified as a potential concern. Excluding corrosion induced hydrogen, trending the behavior of the spiked constituents of concern, and considering conditions necessary for ignition, energetic compounds were shown not to represent an increased risk Based on modeling, about 56,800 l of Resultant oxalates could be added to a washed sludge batch with minimal impact on the number of additional glass canisters produced. For each sludge batch, with 1 to 3 heel dissolutions, about 60,000 kg of sodium oxalate entered the evaporator system, with most collecting in the drop tank, where they will remain until eventual salt heel removal. For each 6,000 kg of sodium oxalate in the drop tank, about 189,000 l of Saltstone feed would eventually be produced. Overall, except for corrosion-induced hydrogen, there were no significant process impacts that would forbid the use of oxalic acid in cleaning High Level Waste tanks. / MATHEMATICAL SCIENCES / M. Tech. (Chemical Engineering)

Currently at the Savannah River Site (SRS), there are fifteen single-shell, 3.6-million liter tanks containing High Level Waste. To close the tanks, the sludge must be removed. Mechanical methods have had limited success. Oxalic acid cleaning is now being considered as a new technology. This research uses sample results and chemical equilibrium software to develop a preferred flowsheet and evaluate the acceptability of the system impacts. Based on modeling and testing, between 246,000 to 511,000 l of 8 wt% oxalic acid were required to dissolve a 9,000 liter Purex sludge heel. For SRS H-Area modified sludge, 322,000 to 511,000 l were required. To restore the pH of the treatment tank slurries, approximately 140,000 to 190,000 l of 50 wt% NaOH or 260,000 to 340,000 l of supernate were required. When developing the flowsheet, there were two primary goals to minimize downstream impacts. The first was to ensure that the Resultant oxalate solids were transferred to DWPF, without being washed. The second was to transfer the remaining soluble sodium oxalates to the evaporator drop tank, so they do not transfer through or precipitate in the evaporator pot. Adiabatic modeling determined the maximum possible temperature to be 73.5°C and the maximum expected temperature to be 64.6°C. At one atmosphere and at 73.5°C, a maximum of 770 l of water vapor was generated, while at 64.6°C a maximum 254 l of carbon dioxide were generated. Although tank wall corrosion was not a concern, because of the large cooling coil surface area, the corrosion induced hydrogen generation rate was calculated to be as high as 10,250 l/hr. Since the minimum tank purge exhaust was assumed to be 5,600 l/hr, the corrosion induced hydrogen generation rate was identified as a potential concern. Excluding corrosion induced hydrogen, trending the behavior of the spiked constituents of concern, and considering conditions necessary for ignition, energetic compounds were shown not to represent an increased risk Based on modeling, about 56,800 l of Resultant oxalates could be added to a washed sludge batch with minimal impact on the number of additional glass canisters produced. For each sludge batch, with 1 to 3 heel dissolutions, about 60,000 kg of sodium oxalate entered the evaporator system, with most collecting in the drop tank, where they will remain until eventual salt heel removal. For each 6,000 kg of sodium oxalate in the drop tank, about 189,000 l of Saltstone feed would eventually be produced. Overall, except for corrosion-induced hydrogen, there were no significant process impacts that would forbid the use of oxalic acid in cleaning High Level Waste tanks. / MATHEMATICAL SCIENCES / M. Tech. (Chemical Engineering)

Excessive release of metals and metalloids into the environment is a major global environmental concern and there is need for environmentally friendly and cost effective methods for their removal. The present study investigated the adsorptive removal of zinc, lead, copper and nickel ions from synthetic aqueous solutions and various metals from three acid mine drainage (AMD) sites using chicken eggshells. Process parameters including the initial metal concentration were evaluated and the optimum conditions obtained were pH 7, adsorbent dose of 7 g and contact time of 360 minutes (for the removal of 100ppm metal ions). Under these conditions the percentage adsorptions were; 97% for lead, 95% for copper, 94% for nickel and 80% for zinc. Aluminium, iron, potassium, nickel and zinc ions all had percentage adsorptions above 75% in AMD Sample 1. Potassium had a 98.78% adsorption while magnesium, strontium and zinc had 72.33; 68.75 and 53.07% adsorption respectively in Sample 2. Arsenic, chromium, copper, iron, antimony and tellurium ions were above 75% for Sample 3. The study demonstrated the efficacy of chicken eggshells and presents it as a viable low-cost adsorbent for bioremediation. / Environmental Science / M. Sc. (Environmental Science)