Se(VI) and Se(IV), as the two major species of selenium in water, are toxic to aquatic lives and may cause adverse health effects to humans at high levels. Biological reduction of Se(VI) is a two-stage process first from Se(VI) to Se(IV) and then from Se(IV) to Se(0) with potential accumulation of the more toxic Se(IV) due to the slower rate of the second stage. Selenium reduction was first evaluated with batch cultures of Shigella fergusonii strain TB42616 (TB) and Pantoea vagans strain EWB32213-2 (EWB) isolated in our laboratory from sludge and coal slurry sediment samples, respectively. In order to facilitate Se(VI) reduction and reduce Se(IV) accumulation, the Se(VI)-reducing strain TB was co-cultured with a Se(IV)-reducing strain EWB. Although Se(VI) reduction rate was not affected, Se(IV) reduction was significantly enhanced with low Se(IV) accumulation in the defined co-culture. Effects of culture composition as well as nitrate and arsenate on Se(VI) reduction were also investigated. A co-culture composition of 10:1 (EWB:TB) ratio was observed to achieve the best total selenium reduction. In addition, nitrate at 50 mg/L was observed to inhibit Se(IV) reduction but not Se(VI) reduction, while arsenate at 200 mg/L exhibited slight inhibition on both Se(VI) and Se(IV) reduction. Biokinetic parameters were optimized with a Monod-type kinetic model using batch pure culture data through the Robust Global Optimization Algorithm embedded in a computer package. Se(VI) reduction by the defined co-culture was then simulated and verified over a range of culture compositions and initial Se(VI) concentrations, respectively. An inter-species inhibition term was incorporated into the model to illustrate the competition for Se(IV) during Se(VI) reduction in the co-culture. The model showed a significant increase of Se(IV) accumulation with higher initial Se(VI) concentration. However, Se(IV) accumulation can be reduced with increasing population ratio of EWB to TB in the defined co-culture. The relatively high correlation coefficients suggested that the model was robust and applicable in simulating Se(VI) reduction by the defined co-culture. Since activated alumina was reported to be more effective for Se(IV) adsorption than Se(VI), the effect of biological activities on selenium removal was investigated using continuous-flow reactors packed with alum-impregnated activated alumina (AIAA) and cultured with a Se(VI)-reducing strain TB under various influent Se(VI) concentrations and hydraulic retention times (HRTs). A selenium removal efficiency of 92% was achieved in a bioreactor with initial biomass of 2.2×106 cells/g-AIAA after a 70-day operation period. Little improvement was observed by lowering the influent Se(VI) concentration from 50 to 10 mg/L while the removal efficiency was significantly enhanced by either extending the hydraulic retention time from 3.2 to 5.0 days or increasing the attached biomass during the startup. An increase in mass ratios of Se(VI) reduction by immobilized cells to adsorption by AIAA was also observed with increasing cell mass during the operation. Se(VI) reduction using continuous-flow reactors packed with strain TB immobilized Ca2+-alginate beads was investigated under various hydraulic retention times (HRT) and influent Se(VI) concentrations. A high removal efficiency up to 98.7% was achieved under an HRT of 5 days and an influent Se(VI) concentration of 400 mg/L. The results showed that the overall selenium removal was positively correlated to the bed height of the reactor and the HRT but not related to the influent Se(VI) concentration. The steady state was analyzed using a mathematical model based on Monod-type equations with four biokinetic parameters optimized including the half-velocity constants and maximum specific reduction rates. The relatively high correlation coefficients indicate that the model is robust and valid to simulate Se(VI) reduction in the gel-beads-packed continuous-flow system.

Ces travaux de thèse ont débuté par la présentation des propriétés physico-chimiques du chrome, ses principaux domaines d'application. On a aussi indiqué les problèmes posés et les risques associés à l'emploi des dérivés chromiques qui s'expriment essentiellement au niveau cutané, respiratoire et même immunologique. Ensuite, on a cité les différents procédés d'élimination tels que les procédés membranaires, procédés chimiques, procédés biologiques et procédés physiques en précisant les avantages et inconvénients de chacun de ces procédés. Après, on a présenté le procédé d'adsorption sur alumine activée, en rappelant brièvement les différents adsorbants utilisés ainsi que les modèles thermodynamiques et cinétiques utilisés pour décrire ce phénomène. Enfin, on a rappelé quelques généralités sur les membranes échangeuses d'ions et sur le procédé de dialyse ionique croisée. Dans une seconde étape, on a étudié la validation de la technique analytique selon la norme Française XPT 90-120, ensuite une étude préliminaire de l'influence des différents paramètres selon le plan d'expérimentation factorielle afin de montrer l'influence de chaque paramètre indépendamment et leurs dépendances aussi. Ces facteurs sont les suivants : pH, quantité de l'adsorbant, concentration du métal initial et de la température. Cette étude a porté sur l'effet du pH mais aussi l'effet de la masse de l'alumine activée et l'effet de la concentration initiale du chrome afin d'optimiser ces paramètres, les résultats expérimentaux ont montré que l'adsorption du chrome dépend essentiellement du pH et de la concentration du chrome initiale. Dans les conditions optimales, les taux d'élimination du chrome VI peuvent attendre 98 % pour une concentration initiale de 10 mg/L en chrome. Les résultats expérimentaux ont été appliqués aux modèles de Freundlich, Langmuir, Dubinin-Redushkevich et Temkin. Des études cinétique et thermodynamiques suivront pour déterminer l'ordre de la réaction d'adsorption des métaux lourds et nature exothermique du mécanisme. La troisième partie a porté sur la Dialyse Ionique Croisée, dans un premier temps on a développé une étude préliminaire de l'influence des différents paramètres selon le plan d'expérimentation effectué afin d'identifier et d'évaluer l'influence de chaque paramètre sur les taux d'élimination de l'ion métallique. Ces facteurs sont la concentration initiale de l'ion métallique, la concentration initiale de l'ion de substitution, le type de la membrane et l'agitation magnétique. Cette démarche est nécessaire afin de fixer les paramètres optimaux et évaluer leurs interactions mutuelles. Cette étude a permis de montrer que le type de membrane échangeuse d'anions est le plus important paramètre influant sur l'élimination du chrome et ceci peut s'expliquer par les valeurs élevées de la capacité d'échange et de la teneur en eau de la membrane AFN. La modélisation du transfert des ions HCrO4- en dialyse ionique croisée a montré la présence de deux phases : la couche limite de diffusion du côté de l'alimentation (solution à traiter contant les ions HCrO4-) et la membrane. En deuxième étape, nous avons effectué une étude du vieillissement des membranes, qui consiste à faire un suivie de l'épaisseur membranaire, de la conductivité membranaire et de leur morphologie par MEB. Nous avons montré que les membranes AMV et AFN se dégradent peu à priori de la même manière. Finalement, nous avons effectué un nouveau montage qui couple les deux procédés et ceci dans le but d'améliorer le taux d'élimination. Ce couplage a permis d'augmenter l'efficacité et la cinétique d'échange de la dialyse, en plus une réduction considérable de la quantité d'alumine activée à régénérer et l'amélioration de la cinétique de transfert transmembranaire / The removal of chromium (VI) from aqueous solutions by activated alumina has been investigated as a function of solution pH, initial chromium concentration, adsorbent dose of activated alumina and temperature. The pH and the adsorbent dose of activated alumina are the most significant parameters affecting chromium (VI) adsorption. The chromium concentrations were analyzed by reaction with 1,5-diphenylcarbazide. This method has been validated according to the French standard XPT-90-210. In order to optimize the effect of the main parameters and their mutual interaction for the adsorption process, a full factor design of the type nk has been used. In this study the removal of chromium (VI) by adsorption on activated alumina was investigated and results were fitted to Langmuir, Freundlich, Dubinin-Redushkevich and Temkin adsorption models, at different temperatures. The constants of each model were evaluated depending on temperature. Thermodynamic parameters for the adsorption system were determined at 10, 25 and 40°C. The obtained values showed that Cr (VI) adsorption is a spontaneous and exothermic process. The kinetic process was evaluated by first–order, second–order and Elovich kinetic models. In second section, the removal of chromium (VI) from aqueous solutions by Donnan dialysis has been investigated. The tows AEMs: Selemion AMV and Neosepta AFN. The amount of chromium removed was determined on the basis of the following parameters: initial chromium concentration, type of anion exchange membranes, co-ion concentration and magnetic stirring. A 24 full factorial design analysis was performed to screen the parameters affecting Cr (VI) removal efficiency were examined. Using the experimental results, a linear mathematical model representing the influence of the different parameters and their interactions was obtained. Analysis of variance (ANOVA), F-test and student's t- test showed that the type of anion exchange membrane is the most significant parameter affecting chromium (VI) removal. The statistical analysis of experimental data assumes that the data come from normal distribution. Then, a modelisation of chromium removal by Donnan dialysis through anion exchange membrane, these results are compared with the theoretical predictions given by the resolution of the corrected Nernst–Planck's equation in the cases of the homogeneous models of ion-exchange membranes. This confrontation confirms the existence of three domains corresponding respectively to a complete diffusion boundary layers (DBLs) control, a mixed control and a complete membrane control of the inter-diffusion process. And finally, the Donnan dialysis and adsorption onto activated alumina has been coupled. The coupling DIC-AAA was successfully achieved with 90% of chromium removed

This thesis includes both experimental and modelling studies on the treatment of drinking water. Three aspects were studied: (i) removal of fluoride (F– ) by adsorption, (ii) removal of fluoride and other contaminants by solar distillation, (iii) denitrification by a microbial fuel cell. The availability of potable water on earth is about 0.2% of the total available water. This very small quantity is polluted by anthropogenic and natural contaminants. Fluoride is a classic example of a natural contaminant, wherein the dissolution of F– bearing minerals causes the release of F– into the groundwater. Exposure to concentrations > 1 mg/L over ex-tended periods of time results in dental and skeletal fluorosis. Worldwide, about 220 million people are at risk. Nitrate is an example of anthropogenic contaminant, occurring because of addition of high quantities of fertilizers to the soil for better crop yields. The excess fertilizers penetrate the soil and mix with the groundwater, resulting in nitrate contamination. The major effect of nitrate contamination is met haemoglobin , which is caused because of the oxidation of ferrous ion in haemoglobin to ferric ion by the nitrite to form haemoglobin. The effects can be noticed by the change in colour of skin to bluish grey or brownish grey in infants. To counter the drastic effects of these anions, the World Health Organization (WHO) has prescribed permissible limits of 1.5 mg/L and 45 mg/L for F– and NO3 – , respectively. For obtaining contaminant-free water, many methods have been used. Reverse osmosis (RO) is one of the widely-used methods. Even though this process removes most of the contaminants, about 50 - 70% of the inlet water is wasted as a reject stream with higher concentrations of the contaminants. This is a very unsustainable way of using water, particularly in drought-prone areas. So, in the thesis a conceptual strategy with three different methods is developed to treat reject water. In the first part of the thesis, the removal of F– using adsorption was studied. Activated alumina (AA) and a hybrid anion exchange resin embedded with hydrous zirconium oxide nanoparticles (HAIX-Zr) (sample sent by Prof. Arup K SenGupta) were used as the adsorbents. The adsorbents were tested with synthetic water samples and reverse osmosis (RO) reject water. HAIX-Zr had a better adsorption capacity compared to AA when water containing only F– was used. The presence of high concentrations of co-ions affects the uptake of F– drastically, with a decrease of up to 34% and 79% for AA and HAIX-Zr, respectively. With AA, for a synthetic water sample with a small concentration of HCO3 – , there was a two-fold increase in the uptake of F– compared to a water sample containing only F– . There was no removal of NO3 – by AA. HAIX-Zr removes NO3 – , but to a lesser extent than F– . With AA, the pH of the inlet solution affected the adsorption capacity, because of the change in the surface charge of AA. Based on the type of water sample used, the cost of treated water varied from Re. 0.1 - 1.0/L ($ 0.0015 - 0.015/L) for AA and 0.2 - 11.5/L ($ 0.003 - 0.17/L) for HAIX-Zr. A community-level plant was set up to treat the RO reject water using AA. Due to challenges at the field level, the pilot plant had to be stopped after 80 bed volumes of water were treated. From our observations and as also reported by many authors, the adsorption of F– is affected by the presence of many ions. When modelling the adsorption of F– , it is usually taken as a single entity getting adsorbed on the adsorbent. As this is not a proper assumption, a model was developed which takes into account all the speciation reactions that take place during adsorption, and all the species like H+, OH– , Na+, Cl– , and NO3 – present in the solution along with F– . Using the model, the equilibrium constants and rate constants for the reactions were obtained. For one initial concentration of F– , a good fit was obtained to the batch adsorption data, except at short times. Due to uncertainty about the amount of impurity present in the adsorbent, at higher initial concentrations of F– , there was a significant discrepancy between predictions and data. Considering column experiments, the breakthrough curve for F– was simulated using the developed model. For the special case of negligible mass transfer resistances, the predicted break-through volume was within 3% of the observed value. In the second part of the thesis, nitrate removal was investigated using microbial fuel cell (MFC). In a MFC, power is generated by the activity of the microorganisms present in the cell. The organisms present in the anode side release electrons (e– ) by the use of substances that can be oxidized, namely, glucose, acetate, etc. On the cathode side, the organisms have the potential to take in e– and reducible substances, and release reduced products like nitrogen, hydrogen, etc. In the present case, nitrates added to the cathode side were reduced to nitrogen gas by the use of a consortium of micro-organisms taken from seawater. A similar consortium was used in the anode chamber Here, the study was focused on improving the efficiency of MFC for removal of NO3 – , by changing the buffering medium used in the cells. Commonly, phosphate buffer is used, but when using a MFC for treatment, the presence of PO43 – causes water contamination and is not suitable for drinking. There-fore, PO43 – was replaced with HCO3 – on the cathode side of the cell. This resulted in a higher removal of NO3 – and power production compared to the PO43 – buffered solution In the third part of the thesis, contaminant removal using solar distillation was investigated. For this as inclined basin still was used. Investigations were based on the evaporation rate of contaminated water, and the odour and concentrations of ions in the distillate. In order to improve the evaporation rates, different radiation absorbing materials like sand, activated charcoal, and carbon nanoparticles encapsulated in polymer sheets (PCNP) were investigated. It was observed that the evaporation rates were higher with activated carbon than the other materials. Using this technique there was about 99% removal of NO3 – , F– , SO42 – and the concentrations of ions in the distillate were well below the acceptable limits. When sand or PCNP was used as an absorbing/wicking medium, the distillate had an objectionable odour. With the use of AC, the odour could be eliminated because of the adsorption of odour-causing compounds.

Excess fluoride (F-) in drinking water poses a health threat to millions of people around the world. In the present work, activated alumina (AA) has been used as an adsorbent. Data obtained from batch experiments were fitted to the (i) pseudo-first order, (ii) pseudo-second order, and (iii) Langmuir kinetic model. Model (ii) performed better than model (i), and fitted the data well. However, the rate constant for adsorption ka had to be varied as a function of the initial concentration of F- in the liquid phase c0. A more satisfactory approach is provided by Langmuir model, which fitted the data reasonably even though ka was independent of c0. Shreyas (2008) developed a model for the batch adsorption of F- onto porous pellets of AA. Some errors were detected in his computer program were corrected. The parameters of the model were estimated by fitting predictions to data. The parameter values suggest that the adsorption process is likely to be diffusion limited. Column experiments were conducted as follows. The pellets were soaked in deionized water for a time ts before they were loaded into columns. A feed solution having a fluoride concentration cf = 3 mg/L was fed to column and the concentration of F- in the exit stream ce was measured at regular intervals. Breakthrough was deemed to have occurred when ce exceeded the permissible limit (= 1 mg/L). Constant values of the bed height H, and the empty bed contact time tc were used in the experiments. The volume of treated water V, scaled by the volume of the bed Vb, varied strongly with the soaking time ts, with a maximum at ts = 24 h. To understand the possible reasons for this behaviour, XRD, FESEM, and FTIR were used to characterize the surface of AA. Though the concentrations of the surface hydroxyl groups may influence the adsorption of F-, FTIR studies show there is no direct correlation between V/Vb and the concentrations of these groups. The FESEM and XRD studies indicate that fresh AA consists mainly of boehmite, which gradually converted to gibbsite during soaking. For fixed values of H and tc, the dimensionless volume of treated water V/Vb was a maximum at D = 45 mm. This behaviour may be caused by wall effects for small values of D and by occurrence of quasi-static regions near the wall for large values of D. The cost of treated of water was Rs. 0.42/L. It decreased slightly to Rs. 0.37/L after one regeneration cycle, but increased to Rs. 0.41/L after two cycles. The volume of treated water after two regeneration cycles was 595 L/kg. The concentration of Al3+ ions ca in the treated water increased and exceeded the permissible limit of 0.2 mg/L as the number of regeneration cycles increased. The concentration of F- in regeneration effluent cre was in the range 32-70 mg/L. The effluent was subjected to solar distillation, leading to a distillate whose fluoride concentration F- was in the range 9-12 mg/L. The distillate can be discharged into the public sewers, as the permissible limit is 15 mg/L.

Excess Fluoride in drinking water is a cause for concern in several countries in the world. Various techniques have been developed to mitigate the harmful effects of fluoride. In the present work, a combined alum and activated alumina (AA) process has been investigated. Tap water with sodium fluoride dissolved in it to produce a solution having a fluoride concentration of 5 mg/L was used as the feed. It was found that pretreatment with alum extends the time required for the regeneration of the AA column. The volume of water treated increased by 89% compared to AA process alone. Though the regeneration of the AA column has been well documented, subtle issues have ot been reported. The disposal of regeneration effluent is a concern in adsorption-based processes. This study aims to examine some of the issues involved in the regeneration of the AA column such as disposal of effluent, and the quantity of acid and alkali required. The regeneration effluent from the combined process, which had a fluoride concentration of 10-16 mg/L was treated in a solar still. The distillate from the still had a fluoride concentration of 2-3 mg/L, which is much lower than the concentration of the regeneration effluent. The cost of treatment decreased with each regeneration cycle and after four regenerations the cost was Rs 0.5/L of treated water. The volume of water treated after four regenerations was 307 L/Kg of AA. Studies were also done using field water from Banavara, Hassan district, Karnataka, which had a fluoride concentration of 3,0-3.5 mg/L. The combined process successfully produced treated water having an acceptable fluoride concentration. After one regeneration cycle, the operating cost was Rs. 1/L of treated water. Studies have also been conducted on a point-of-use water filter containing a bed of AA pellets. The filter was provided by an organization called TIDE. The present results appear to suggest that a column with a smaller diameter than the TIDE filter has a better removal capacity. Ceramic candles are widely used for water filtration as they are readily available and inexpensive. Hence they are suitable for household water treatment purposes. In the present work, ceramic candles have been impregnated with nano-size alumina and nano-size magnesium oxide and tested for their defluoridation capacity. The nanoparticles were generated in situ in the pores of the candle by solution combustion synthesis. It has been found the candle impregnated with nano-size magnesium oxide has a higher defluoridation capacity than nano-size alumina. Estimation of the particle size in the samples of treated water did not give conclusive evidence for the presence or absence of nanoparticles. The volume of water treated was low and the cost of treatment was high (Rs. 12/L for the candle impregnated with MgO). Hence such candles are unsuitable for defluoridation. Batch adsorption has been employed to measure the adsorption capacity of adsorbents. A model to capture the overall picture of the batch adsorption process, obtaining the kinetic and transport parameters involved has been developed. The mathematical model takes into account external mass transfer resistance, intraparticle diffusion, adsorption, and desorption. The equilibrium adsorption data was fitted using the Langmuir isotherm. The governing equations were solved using a finite difference technique known as the Laasonen method. The parameters were estimated by fitting two sets of data using a MATLAB function. The values estimated suggest that the adsorption process may not be diffusion-limited, in contrast to the assumption commonly used in the literature. The estimated parameter values were used to predict the concentration profiles for the other data sets. It was found that predicted and measured profiles agreed reasonably well.

Submitted in fulfillment of the requirements for the degree of Master of Engineering in Chemical Engineering, Durban University of Technology. Durban. South Africa, 2015. / Several areas in the world such as the United States of America, Sri Lanka, China, Argentina, Canada, Tanzania, Kenya, South Africa and many others have a problem of high fluoride content in drinking water. Generally fluoride levels above 1.5 ppm in water may result in dental and skeletal fluorosis in humans depending on quantity consumed (Fan et al., 2003; Meenakshi, 2004). Remote rural areas where there are no water treatment facilities are more vulnerable to this problem. Adsorbents such as activated alumina and FR-10 resin seem to have a potential for successful application in rural areas. These methods however require pre-treatment if the feed has high turbidity. A membrane based system called woven fabric microfiltration gravity filter (WFMFGF) developed by Durban University of Technology proved to be suitable for turbidity removal. The main objective of this research was to develop a small water treatment system for fluoride removal. The small water treatment system developed in this study consists of WFMFGF for pre-treatment and an adsorption column. The WFMFGF is made up of a 40 L container packed with 15 immersed flat sheet membrane elements. The operation of the WFMFGF is in batch mode, driven by varying static head. The static head variation results in flow rate variation through the system. This in turn result in variation of contact time, velocity as well as pressure drop in the fluoride removal unit. Specific objectives of the study were: (1) to establish the maximum and minimum flow rates through the WFMFGF system, the total run time before cleaning is required and the best cleaning method for this particular membrane system. (2) to evaluate and compare the performance of activated alumina and FR-10 resin on varying contact time, velocity and pressure drop on the fluoride removal unit. The adsorbents were also compared on adsorption capacity, cost and ease of operation. The minimum and maximum flow rates through the WFMFGF were found to be 5 l/hr and 100 l/hr respectively. It was found that the system can be run for more than a month before requiring cleaning. The suitable cleaning method was found to be soaking the membranes in 0.0225 percent sodium hypochlorite solution overnight and brushing them using a plastic brush. The comparison of the performance of FR-10 resin to activated alumina found that the adsorbents gave equal performance based on the given criteria. FR-10 resin had higher adsorption capacity, gave good quality treated water even with shorter contact time and operated at wider velocity range. Activated alumina on the other hand had an advantage of lower costs, lower pressure drop and ease of use. According to Pontius (1990), the performance of activated alumina can be improved by intermittent operation. Point of use (POU) systems are generally operated intermittently. This improves the fluoride removal efficiency of activated alumina giving it more advantage over FR-10 resin. Based on this activated alumina was selected as the best adsorbent for the system. After the adsorbent was selected, the adsorption column was designed. The column operation regime was 3.5 minutes minimum contact time and 1.17 to 7.8 m/hr velocity range. The activated alumina adsorption capacity was 1.53mg/g. The column had an inside diameter of 70 mm. It was packed with activated alumina to a bed height of 400 mm. The column inlet and outlet pipes were made of PVC with a standard pipe size of 20 mm outside diameter. A valve at the column inlet pipe allowed water to flow through the system.