La contamination du milieu naturel, notamment des aquifères, par des hydrocarbures lourds de type HAP (Hydrocarbure Aromatique Polycyclique) se révèle être une pollution pérenne très difficile à traiter. La dépollution de ces sous-sols par un traitement in situ efficace et peu coûteux constitue encore aujourd’hui un challenge. Cette étude cible une technique de co-injection air-vapeur en zone saturée comme alternative aux techniques thermiques et chimiques en vigueur pour le traitement de cette catégorie d’hydrocarbures. Une première partie expérimentale introductive estime l’influence d’un simple traitement vapeur sur la libération de HAPs par les matrices de sol au sein de l’aquifère, montrant la nécessité de coupler à ce procédé une technique d’oxydation chimique. Si le front de vapeur permet de mobiliser totalement les composés légers, les concentrations des composés les plus lourds augmentent en phase dissoute (augmentation des fractions molaires des composés résiduels). Réalisées en milieu fermé, dans des conditions de température dictées par l’injection de vapeur (120 ̊C), des expériences en laboratoire ont permis de caractériser et de quantifier les réactions d’oxydation chimique mises en jeu. Les résultats démontrent une efficacité modérée de l’oxydation à moyenne température. Des sous-produits de réaction sont identifiables et quantifiables à compter de plusieurs semaines de réaction. A l’inverse l’étude met en évidence l’absence totale de minéralisation, démontrant que l’ana- lyse du CO2 ne peut renseigner sur l’évaluation du phénomène d’oxydation. Hormis la réactivité par oxydation, les conclusions de l’étude montrent que la technique favorise la sorption irréversible des composés organiques en surface des matrices de sol. Ces phénomènes de sorption sont très marqués en présence d’argile ou de matière organique naturelle dans le milieu et augmentent au cours du temps. La technique HPO se pose donc en traitement de soutien à l’injection de vapeur par oxydation mais surtout par stabilisation de la zone de contamination. Enfin nous nous sommes intéressés aux problématiques liées à l’injection d’eau chaude et de vapeur dans un sous-sol au travers d’expériences sur site. La technique d’injection de vapeur étant financièrement et énergétiquement coûteuse, des solutions techniques ont été proposées, appliquées et modélisées. Il en ressort une comparaison en bilans énergétiques pour différents modes d’injection, permettant de mieux appréhender les véritables besoins d’une telle technique. Parmi les paramètres influents, le rayon d’influence est un paramètre essentiel qui conditionne la distance entre chaque puits pour une efficacité optimale. Les résultats de l’étude démontrent que dans certains scénarios d’injection (injection à grande profondeur) le préchauffage de la zone d’étude par injection d’eau chaude couplé à un pompage en profondeur permettra d’accroître significativement le rayon d’influence et d’améliorer le bilan énergétique du traitement global. Les principaux résultats de la thèse sont comparés aux autres études sur l’injection de vapeur pour dégager les meilleures conditions d’application de cette technique, et mettre en évidence les verrous techniques ayant pu être levés au cours de la thèse. / PAHs are the largest, ubiquitous and carcinogenic environmental chemical groups. In a context of polluted soil remediation, today it is still a challenge to reach an effective in situ treatment. This study aims to evaluate as an alternative way the potential of a technology combining a thermal process (injection of steam) and a chemical process (co-injection of air) in the saturated zone. A first introductory experimental section considers the influence of a single steam treatment on the release of PAHs from soil matrices within the aquifer. Results prove the necessity of coupling a chemical oxidation technique with this process. If the steam front allows to recover light com- pounds, concentrations of heavier compounds in the dissolved phase are more important after the steam treatment (molar fractions of residual compounds increase). In order to study the fate of PAHs in polluted soils under medium temperature oxidation, numerous experiments in presence of soil spiked matrices were carried out in closed reactors. The catalytic potential of different mineral surfaces was investigated by studying reactivity of different PAHs. Results demonstrate that the efficiency of oxidation remains moderate. They suggest a lack of mineralization and reactions producing partly polycyclic aromatic compounds (PACs) with equivalent molecular weight as the initial contaminants. Along short term experiments (9 days), the major effect of heating is an increase in irreversible sorption of original compounds. Interestingly, this sorption mostly occurs in presence of natural organic matter and oxygen in the gas phase, suggesting a specific reactivity of the natural organic matter. In long term experiments (6 months), new oxygenated PACs were formed and remained fixed on the solid phase. Consequently, air oxidation catalyzed by minerals and natural organic matter may be a new pathway concerning PAH stabilization in soils. The steam injection technique is still considered as a costly technique. In this last part some technical solutions have been proposed, implemented and modelized. The economic aspect of some injection cases are compared, based on the results of a real field application and some general calculations regarding the costs of wells and energy. Among influential parameters, the radius of influence is a key parameter that determines the distance between each well for an optimum efficiency. The results of this study show that in some scenarios of injection (when the injection takes place in a deeper zone), a preheating phase can be an interesting option to reduce the financial costs of the technique. The main results of the thesis are compared with other steam injection studies to identify the best conditions for the technique application.

This study was carried out to evaluate the performance of Soil Aquifer Treatment (SAT) under different loading regimes, using wastewater of much higher strength than usually encountered in SAT systems, and also to investigate the removal of the endocrine disruptors triclocarban (TCC), estrone (E1), 17β-estradiol (E2) and 17α-ethinylestradiol (EE2). SAT was simulated in the laboratory using a series of soil columns under saturated and unsaturated conditions. Investigation of the removal of Chemical Oxygen Demand (COD), Biochemical Oxygen Demand (BOD), Dissolved Organic Carbon (DOC), nitrogen and phosphate in a 2 meter long saturated soil column under a combination of constant hydraulic loading rates (HLRs) and variable COD concentrations as well as variable HLR under constant COD showed that at fixed HLR, a decrease in the influent concentrations of DOC, BOD, total nitrogen and phosphate improved their removal efficiencies. It was found that COD mass loading applied as low COD wastewater infiltrated over short residence times would provide better effluent quality than the same mass applied as a COD with higher concentration at long residence times. On the other hand relatively high concentrations coupled with long residence time gave better removal efficiency for organic nitrogen. Phosphate removal though poor under all experimental conditions, was better at low HLRs. In 1 meter saturated and unsaturated soil columns, E2 was the most easily removed estrogen, while EE2 was the least removed. Reducing the thickness of the unsaturated zone had a negative impact on removal efficiencies of the estrogens whereas increased DOC improved the removal in the saturated columns. Better removal efficiencies were also obtained at lower HLRs and in the presence of silt and clay. Sorption and biodegradation were found to be responsible for TCC removal in a 300 mm long saturated soil column, the latter mechanism however being unsustainable. TCC removal efficiency was dependent on the applied concentration and decreased over time and increased with column depth. Within the duration of the experimental run, TCC negatively impacted on treatment performance, possibly due to its antibacterial property, as evidenced by a reduction in COD removals in the column. COD in the 2 meter column under saturated conditions was modelled successfully with the advection dispersion equation with coupled Monod kinetics. Empirical models were also developed for the removal of TCC and EE2 under saturated and unsaturated conditions respectively. The empirical models predicted the TCC and EE2 removal profiles well. There is however the need for validation of the models developed

This study was carried out to evaluate the performance of Soil Aquifer Treatment (SAT) under different loading regimes, using wastewater of much higher strength than usually encountered in SAT systems, and also to investigate the removal of the endocrine disruptors triclocarban (TCC), estrone (E1), 17¿-estradiol (E2) and 17¿- ethinylestradiol (EE2). SAT was simulated in the laboratory using a series of soil columns under saturated and unsaturated conditions. Investigation of the removal of Chemical Oxygen Demand (COD), Biochemical Oxygen Demand (BOD), Dissolved Organic Carbon (DOC), nitrogen and phosphate in a 2 meter long saturated soil column under a combination of constant hydraulic loading rates (HLRs) and variable COD concentrations as well as variable HLR under constant COD showed that at fixed HLR, a decrease in the influent concentrations of DOC, BOD, total nitrogen and phosphate improved their removal efficiencies. It was found that COD mass loading applied as low COD wastewater infiltrated over short residence times would provide better effluent quality than the same mass applied as a COD with higher concentration at long residence times. On the other hand relatively high concentrations coupled with long residence time gave better removal efficiency for organic nitrogen. Phosphate removal though poor under all experimental conditions, was better at low HLRs. In 1 meter saturated and unsaturated soil columns, E2 was the most easily removed estrogen, while EE2 was the least removed. Reducing the thickness of the unsaturated zone had a negative impact on removal efficiencies of the estrogens whereas increased DOC improved the removal in the saturated columns. Better removal efficiencies were also obtained at lower HLRs and in the presence of silt and clay. Sorption and biodegradation were found to be responsible for TCC removal in a 300 mm long saturated soil column, the latter mechanism however being unsustainable. TCC removal efficiency was dependent on the applied concentration and decreased over time and increased with column depth. Within the duration of the experimental run, TCC negatively impacted on treatment performance, possibly due to its antibacterial property, as evidenced by a reduction in COD removals in the column. COD in the 2 meter column under saturated conditions was modelled successfully with the advection dispersion equation with coupled Monod kinetics. Empirical models were also developed for the removal of TCC and EE2 under saturated and unsaturated conditions respectively. The empirical models predicted the TCC and EE2 removal profiles well. There is however the need for validation of the models developed / Netherlands Organisation for International Cooperation in Higher Education (Nuffic) / The Appendix files for this thesis are unavailable online via Bradford Scholars.