abstract: Per- and polyfluoroalkyl substances (PFAS) are a group of man-made chemicals that are detected ubiquitously in the aquatic environment, biota, and humans. Human exposure and adverse health of PFAS through consuming impacted drinking water is getting regulatory attention. Adsorption using granular activated carbon (GAC) and ion exchange resin (IX) has proved to be efficient in removing PFAS from water. There is a need to study the effectiveness of commercially available sorbents in PFAS removal at the pilot-scale with real PFAS contaminated water, which would aid in efficient full-scale plant design. Additionally, there is also a need to have validated bench-scale testing techniques to aid municipalities and researchers in selecting or comparing adsorbents to remove PFAS. Rapid Small-Scale Column Tests (RSSCTs) are bench-scale testing to assess media performance and operational life to remove trace organics but have not been validated for PFAS. Different design considerations exist for RSSCTs, which rely upon either proportional diffusivity (PD) or constant diffusivity (CD) dimensionless scaling relationships. This thesis aims to validate the use of RSSCTs to simulate PFAS breakthrough in pilot columns. First, a pilot-scale study using two GACs and an IX was conducted for five months at a wellsite in central Arizona. PFAS adsorption capacity was greatest for a commercial IX, and then two GAC sources exhibited similar performance. Second, RSSCTs scaled using PD or CD relationships, simulated the pilot columns, were designed and performed. For IX and the two types of GAC, the CD–RSSCTs simulated the PFAS breakthrough concentration, shape, and order of C8 to C4 compounds observed pilot columns better than the PD-RSSCTs. Finally, PFAS breakthrough and adsorption capacities for PD- and CD-RSSCTs were performed on multiple groundwaters (GWs) from across Arizona to assess the treatability of PFAS chain length and functional head-group moieties. PFAS breakthrough in GAC and IX was dictated by chain length (C4>C6>C8) and functional group (PFCAs>PFSAs) of the compound. Shorter-chain PFAS broke through earlier than the longer chain, and removal trends were related to the hydrophobicity of PFAS. Overall, single-use IX performed superior to any of the evaluated GACs across a range of water chemistries in Arizona GWs. / Dissertation/Thesis / Masters Thesis Chemical Engineering 2020

The overuse and misuse of antibiotics to treat bacterial infections, the release of unmetabolized residuals into the sewer system, and the incomplete removal antibiotic residues by wastewater treatment plants (WWTPs) pose a severe threat to human health. The accumulation of antibiotic residue induces selective pressure on the bacterial population, resulting in the spread of antibiotic-resistant bacteria (ARB) and antibiotic-resistance genes (ARGs) in water. This study investigated the degradation of different types of ARGs in water matrices using a wide variety of treatment technologies. Real wastewater samples were collected from a WWTP in urban Philadelphia and the presence of single and multidrug-resistant bacteria and resistance genes were investigated using molecular-based techniques. Subsequently, an analytical method was developed and validated for the detection and quantification of the ARGs against a range of antibiotics, such as tetracycline (TCN), ciprofloxacin (CIP), and levofloxacin (LVX). Finally, to remove the ARGs from water matrices, different conventional and advanced oxidation processes were applied. At the very onset, conventional treatment processes such as chlorine treatment was used to inactivate the E.coli resistant strains. It was observed that chlorination can potentially deactivate the ARBs by applying a lower dose and contact time. However, the effectiveness of chlorine treatment in removing all types of ARGs from water matrices was limited. For instance, no significant degradation of extracellular ARGs (e-ARGs) was observed in DI water during chlorine treatment. Subsequently, a peracetic acid (PAA) based treatment process was used to degrade the genomic and plasmid-encoded ARGs from the water matrices. Similar to chlorine treatment, no significant changes were observed in the degradation of extracellular ARGs (e-ARGs) in DI and WW. Then, the degradation kinetics of ARGs across different types (gyrAR, tetAR, qnrSR) and forms (chromosomal, plasmids) were evaluated using the Ultraviolet (UV) disinfection process. Compared to chlorination and PAA, UV treatment showed better removal efficiencies for the degradation of different types of e-ARGs in DI water. The degradation profile of e-ARGs showed 1-4 log reductions at a UV fluence of 900 mj/cm2. The i-ARGs showed similar degradation rates as compared to e-ARGs in phosphate buffer saline (PBS) at the same UV dosage. On the other hand, the regrowth potential of ARBs at low UV dosage (60–180 mJ/cm2) showed the evidence of damage repairment after several hours of exposure to light (photoreactivation) and dark conditions, making it susceptible again to the resistance spread. To resolve this issue, process parameters were optimized, and no regrowth of the ARBs were found from the higher fluence from 300 to 600 mJ/ cm2. Later, UV/ H2O2 based AOP was applied to evaluate the degradation and deactivation of the same resistant genes. The addition of H2O2 during the UV treatment produces strongly reactive •OH radicals during the treatment and showed considerable improvements in e-ARGs degradation (1.2-5 logs) compared to UV treatment alone. However, this AOP showed minimal contribution to i-ARG degradation (1-2.4 logs), possibly due to the scavenging of •OH radicals by the cellular components in PBS. In contrast to PBS, the wastewater matrix moderately enhanced the gene degradation during the treatment. In terms of plasmid degradation, the conformational differences of the supercoiled structures showed 1.2-2.8 times slower degradation rates than chromosomal ARGs. In addition, the degradation kinetics of the free residual ARGs (f-ARGs) were assessed during the treatment to reduce the AMR dissemination risk from the treated sample. This study also examined the potential of ozone (O3) based oxidation process to degrade and deactivate the extracellular and intracellular ARGs, and MGE (plasmid, intl-1) from E.coli ARBs. The degradation kinetics of the ARGs across different sizes (118-454 bps) and types were evaluated in different water matrices (DI water, PBS, and WW), and showed a significantly higher removal for chromosomal, and plasmid encoded ARGs than other treatment technologies. For the e-ARGs in DI water, 3.8-5.2 logs removal was observed at ozone dosage of 2.0 × 10-2 M.s. i-ARGs in PBS and wastewater showed nearly similar degradation (3.8-5 logs) during O3, indicating the elimination of i-ARGs was not dependent on the cellular components and effluent organic matter. Moreover, an analysis of environmental DNA (eDNA) from wastewater was conducted to examine the degradation of DNA and ARGs for different storage periods and temperatures (-20°C, 0°C, 4°C, 22±0.87°C). Result indicated that water samples kept at -20°C and 0°C showed the best performance in preventing the DNA concentration and gene degradation over time. Additionally, the effectiveness of different preservatives (Longmire buffers: LB1 and LB2, benzalkonium chloride at 0.1%, 0.01%) were investigated in preserving the DNA integrity and the gene degradation at an ambient temperature. It was found that the Longmire buffer (LB1) exhibited lowest gene degradation during the three-week storage period. In summary, this research provided a comprehensive assessment on the degradation of e-ARGs, i-ARGs, and free ARGs from water using different treatment technologies (i.e., UV, UV/H2O2, O3, PAA, chlorine). Additionally, this study suggested valuable information on optimizing the process parameters of the selected methods and developed a comparative assessment of removing the ARGs from the water matrix (DI/PBS, WW). The estimation of Electrical Energy per Order (EEO, kWh/m3) during UV and ozone treatments provided a comparison of the energy consumption for ARGs degradation in the water. Overall, the findings of this study can be useful for evaluating different types and forms (chromosomal, plasmid) of ARG degradation from water matrices and can help to reduce the risk of AMR dissemination in the environment. / Civil Engineering

Au vu de l'omniprésence des oxydes de fer dans le milieu naturel, et en particulier la goethite et l'hématite qui sont les formes les plus stables, mais aussi de la prolifération des contaminants émergents dans l'environnement, comme les fluoroquinolones, notre objectif a été d'étudier leur réactivité et de définir les mécanismes de sorption et d'oxydoréduction à l'interface oxyde/solution dans des conditions statiques (batch) et hydrodynamiques contrôlées (colonne) en couplant une étude macroscopique (techniques chromatographiques, LC/MS, LC/UV) avec une approche microscopique (spectroscopie vibrationnelle et XPS) et de modélisation mécanistique (TPM et CD-MUSIC). Ces travaux mettent en évidence les principaux mécanismes responsables de la transformation des molécules organiques à la surface d'un oxyde de fer, et donc fournissent des informations nécessaires à la compréhension du devenir des contaminants émergents dans l'environnement. / Given the ubiquity of iron oxides in environmental settings, particularly goethite and hematite, the most stable forms, but also the proliferation of emerging contaminants, such as fluoroquinolones, in the environment, our goal was to study their reactivity and describe mechanisms of sorption and redox at oxide /solution interfaces in static batch) and hydrodynamic conditions (column) by coupling a macroscopic study (LC/MS, LC/UV) with a microscopic/molecular approach (vibrational spectroscopy and XPS) and mechanistic modeling (TPM and CD-MUSIC).. These works highlight the main mechanisms responsible of the transformation of organic molecules on iron oxide surfaces and thus provide valuable information necessary for the understanding of the fate of emerging contaminants in the environment.

Des études de sorption/désorption ont été effectuées pour neuf contaminants émergents sélectionnés (caféine, sulfaméthoxazole, déséthylatrazine, carbamazépine, atrazine, estradiol, éthinylestradiol, noréthindrone et diclofénac) dans les boues usées provenant de trois systèmes différents. Les contaminants incluent une variété de classes de composés (pesticides, hormones et pharmaceutiques) qui possèdent des propriétés physicochimiques différentes. L’objectif de ces travaux est de modéliser leur comportement dans une station d’épuration, en présence d’une phase particulaire et d’une phase aqueuse, et du même coup, de mieux comprendre leur devenir lors de leur rejet dans l’environnement. Le coefficient octanol-eau (log Kow) permet de bien interpréter les résultats et nous permet de classer les composés selon deux types de comportements observés : les composés avec un log Kow inférieur à 3 montrent peu ou pas de sorption alors que les composés avec un log Kow supérieur à 3 montrent une sorption variant de 30 à 90 % durant les premières minutes, suivi d’une sorption lente durant les heures suivantes. Une augmentation du contenu organique favorise la sorption des composés hydrophobes alors qu’un changement de pH peut modifier la charge à la surface des particules et également la charge des analytes. Les résultats ont montré que seul le diclofénac était sensible aux variations de pH étudiés. Dans une telle situation, il est nécessaire d’utiliser le facteur d’hydrophobicité corrigé en fonction du pH (log Dow). Le coefficient de distribution solide-eau (log Kd) a été déterminé pour chaque composé à la fin de chaque expérience de sorption et se situe entre -0.3 et 2.6. Avec l’augmentation de l’hydrophobicité, la désorption diminue avec le temps et avec l’étape de rinçage. Pour simuler le relargage dans les systèmes aquatiques, les facteurs de rinçage ont été déterminés pour estimer le nombre de rinçages qui seraient nécessaire pour désorber 50 et 99 % de la concentration initialement sorbée. Les bilans de masse ont été effectués après chaque expérience dans le but de ne pas surestimer les capacités de sorption d’un composé et se situent entre 7 et 25 % pour l’estradiol, la noréthindrone et le sulfaméthoxazole et entre 44 et 103 % pour l’éthinylestradiol, l’atrazine, la déséthylatrazine, la carbamazépine, la caféine et le diclofénac. / Sorption/desorption studies were performed for nine selected emerging contaminants (caffeine, sulfamethoxazole, desethylatrazine, carbamazepine, atrazine, estradiol, ethinylestradiol, norethindrone and diclofenac) in sewage sludge from three different systems. Contaminants include a variety of compound classes (pesticides, hormones and pharmaceuticals) with different physicochemical properties. The objective of this work is to model their behavior in a treatment plant in the presence of a particulate phase and an aqueous phase, and at the same time, understanding their fate upon their release into the environment. The octanol-water partition coefficient (log Kow) allowed a good understanding of the results and allowed us to classify the compounds according to the two types of behavior observed: compounds with log Kow below 3 showed little or no sorption while compounds with log Kow over 3 showed a 30 to 90% sorption within the first few minutes, followed by a slow sorption during the next hours. An increase of the organic content promotes the sorption of hydrophobic compounds while a change in the pH can modify the charge on the surface of the particles and also the charge of the analytes. Only diclofenac was found to be sensitive to the different pH studied. In such a situation, it is necessary to use the pH-corrected hydrophobicity factor (log Dow). The solid-water distribution coefficient (log Kd) were determined for each compound at the end of each sorption experiment and ranged from 0.2 to 2.9. With increasing compounds hydrophobicity, desorption decreased with time and rinsing step. To simulate releases into aquatic systems, rinsing factors were determined to estimate the number of rinsing that would be needed to desorb 50 and 100 % of the sorbed concentration. Mass balances were performed after each experiment in order to not overestimate the sorption capacity of the compound and ranged from 7 to 25 % for estradiol, norethindrone and sulfamethoxazole and from 44 to 103 % for ethinylestradiol, atrazine, desethylatrazine, carbamazepine, caffeine and diclofenac.