M.Ing. (Civil Engineering) / Natural organic matter (NOM) is a complex mixture of organic compounds resulting from the decay of plants and animals. When not properly removed, NOM reacts with disinfectants to form disinfection by-products, of which some are known to be carcinogenic. Regulations have become more and more stringent about the maximum level of NOM in drinking water. Different water treatment processes can be applied to remove NOM. However, because each process targets specific fractions of NOM, this results in the existence of recalcitrant NOM fractions that are not removed, for each treatment process. The main objective of the study was thus to investigate NOM removal by using three advanced water treatment processes alone and in combination. The processes used were enhanced coagulation, granular activated carbon adsorption and ion exchange resin processes. Strong and weak base ion exchange resins were used. Although NOM removal methods have been investigated intensively, little attention is given to a so-called “multi-barrier” approach of NOM removal through enhanced coagulation, granular activated carbon adsorption and ion exchange resin that could be applied in practice. In order to take the seasonal and geographical variability of NOM into account, water samples were collected at eight geographical areas over five to seven seasons from 2010 to 2013. Results showed that different treatment processes remove different NOM fractions with different efficiency. Results also confirmed that EC performs better than the adsorptive treatment processes. But the literature shows that NOM removal by EC is difficult for water with high alkalinity and low SUVA values, which is common with SA waters. The adsorptive processes used for NOM removal revealed that they are not viable options, because even low percentage of NOM removal would require high adsorbent dosages, which removes this approach as a practical option.

D.Phil. (Chemistry) / Elimination of toxic organic compounds from wastewater is currently one of the most important subjects in water-pollution control. Among the many organic pollutants are dyes and emerging pollutants such as natural organic matter (NOM). Dyes such as Eosin Yellow (EY), an anionic xanthene fluorescent dye, can originate from many sources such as textile industrial processes, paper pulp industries and agricultural processes. Most dyes are problematic because they are resistant to conventional chemical or biological water-treatment methods and therefore persist in the environment. NOM consists of a highly variable mixture of products found in water and soils. NOM is formed as a result of the decomposition of plant and animal material and is a precursor to the formation of disinfection by-products (DBP) during water disinfection. These organic compounds cause undesirable colour, taste and odour in water. NOM affects the capacity of other treatment processes to effectively remove organic micro-pollutants or inorganic species that may be present in the water. Its removal also uses up chemicals and energy and so it is expensive to treat. Titanium dioxide (TiO2) has emerged as one of the most fascinating materials in the modern era due to its semiconducting and catalytic properties. TiO2 is a large band-gap semiconductor that exists mainly in the anatase (band gap 3.2 eV) and rutile (band gap 3.0 eV) phases. Its response to UV light has led to increased interest in its application in the photocatalysis research field. It has been investigated extensively for its super hydrophilicity and use in environmental remediation and solar fuel production. In spite of extensive efforts to apply TiO2 for environmental remediation, photocatalytic activity in the visible region has remained quite low hence the ultimate goal of this research was to fabricate highly photoactive catalysts composed of non-metal, platinum-group metal (PGM) co-doped TiO2 and carbon nanotubes (CNTs) and to apply them for water purification using solar radiation...

D.Phil. (Chemistry) / This thesis describes the improvement of sorption capacity and efficiency of synthetic and biological adsorbents towards selected pollutants by introducing functional groups on the sorbents. Functionalization was achieved by chemically modifying the binding sites of the studied adsorbents. The sorbent materials considered were chemical resins (Amberlite XAD 1180) and biomass (maize tassels and seaweed). The adsorbents were modified with ethylenediamine in order to improve their capacity for extraction of heavy metals, namely lead, copper and cadmium from water. For the removal of phenols from water, maize tassels was modified with polydiallyldimethylammonium chloride (polyDADMAC). The main focus of the proposed study was to formulate cheap and sustainable ways of purifying contaminated water by exploiting the pollutants’ affinity towards the adsorbents. Parameters such as contact time, sorbent quantities, analyte concentrations, desorption solution (for stripping adsorbed metals for sorbent reuse) and pH were optimized. Different isotherms were applied on the experimental data to establish sorption mechanisms and energies involved during the sorption process. The Langmuir isotherm was used to test for monolayer sorption while the Freundlich model tested multilayer adsorption on heterogeneous surfaces...

Three studies were performed to obtain fundamental mechanistic information on the TiO2 catalyzed photooxidations of organic substrates irradiated at 350 nm in dilute aqueous solutions under oxygenated conditions: (a) The photodecomposition of three haloethers, 2-chloroethyl ether, 4-chlorophenyl phenyl ether, and 4-bromophenyl phenyl ether, was investigated in an aqueous media at pH 7.0. (b) A comparative study of structure-reactivity was conducted on para-substituted phenols whose substituents range from electron-withdrawing to electron-donating in an aqueous media at pH 3.0. (c) The initial rates of the TiO2 catalyzed photodegratation of phenol were studied in an aqueous media at pH 1.0, 3.0, 5.0, 7.0, 9.0, 11.0, and 13.7 and a pH effect profile was obtained and compared to the removal efficiency after four hours of irradiation. Controls were carried out throughout the three studies in the absence of light and under anoxic conditions, as well as without the semiconductor to evaluate the role of photolysis. The Langmuir-Hinshelwood model was employed in an attempt to characterize and evaluate differences in reactivity.

The purpose of this research was to conduct a laboratory study to determine the effects of nickel on the completely-mixed activated sludge process. Continuous-flow bench-scale reactors were operated at COD:TKN ratios of approximately 1.0:1 and 0.5:1 by varying the nitrogen concentrations in the feed solutions. Each unit received a COD concentration of 400 mg/l and was dosed continuously with a nickel concentration of 0.77 mg/l. The mean cell residence time was utilized as the operational control parameter to assess the influence of nickel on organic removal efficiency, on the degree of nitrification, and on the maximum yield and the microbial maintenance energy coefficients, Y_max and k_d. The results obtained in this study indicated that the soluble COD removal efficiency of the heterotrophic microorganisms was not affected by the 0.77 mg/l nickel concentration. The low continuous nickel dose to the reactor, however, appeared to stimulate the heterotrophic growth or to cause a replacement of the bacteria with a species of nickel-tolerant microorganisms. At the same time, the nickel appeared to stress the heterotrophs and to cause an increase in their maintenance energy requirement. Nitrification was found to be inhibited by the low nickel concentration, and this inhibition was not reduced by operating the reactors at lower COD:TKN ratios. / Master of Science

The application of packed column air stripping systems to the removal of volatile organic contaminants (VOCs) from drinking water sources was investigated. The crucial element for the design of such systems exists in obtaining accurate predictions of mass transfer rates. The first phase of this study evaluated three semi-empirical correlations available for predicting packed column mass transfer rates. From this initial screening, the Onda model was selected for further investigation. A test data base was established from water treatment pilot study results reported in the literature. Ten separate studies were selected for evaluation, encompassing approximately 450 data points. Eleven different VOCs were encountered in these investigations, and the basic packing types tested included rings, saddles, Tri-Packs, and Tellerettes. Comparison of measured mass transfer rates with the corresponding rates predicted by the Onda correlation yielded a relative standard deviation of 17%. A ± 30% accuracy value was therefore assigned to the model based on 90% confidence limits. This assessment agrees with the observed accuracy of the correlation for the chemical engineering-based system results utilized in the model's original derivation. From the overall evaluation, no severe deficiencies and/or limitations with the Onda correlation were noted. In particular, gas-phase resistance predictions appeared reasonably accurate. However, further investigative studies involving observed column performance with larger packing materials (≥2 inches in nominal size) is encouraged to verify the accuracy of the correlation for such situations. Within the context of the evaluation procedure, several related areas were investigated. First, Henry's constant temperature relations reported in the literature were established within approximately 20% for common VOCs at low concentration levels. Secondly, the transfer unit performance model for calculating observed mass transfer rates was found sensitive to experimental measurement errors below a stripping factor of 1.5. Therefore, measured results obtained under such operating conditions must be viewed with appropriate caution. Finally, economic design boundaries for column operational parameters were established based upon optimization study results reported in the literature. / M.S.

M.S.

The molecular size, charge, and solubility, distributions of natural DOC in raw and treated surface waters were investigated to determine the types of organic compounds removed by coagulation and settling. The distribution of organic precursors that react with chlorine to form organic halide compounds was also determined, along with the reaction rates. DOC removal by coagulation was size dependent. Compounds over 5,000 d were readily removed while compounds under 1,000 d were largely unaffected. Acidic and basic molecules were precipitated much more effectively than neutral ones. Both hydrophilic and hydrophobic molecules were selectively removed compared to compounds of intermediate solubility. Two groups of organic halide precursors were identified. Fulvic acids had high specific organic halide yields, and fast reaction rates with chlorine, but could be effectively removed by coagulation. Nonpolar neutral compounds were lower in molecular weight, slower to react with chlorine, and produced less organic halide, but could not be precipitated by coagulation. / Ph. D. / incomplete_metadata

Submitted in fulfilment of the requirements of the degree of Master of Engineering - Chemical, Durban University of Technology, Durban, South Africa, 2015. / Industrial wastewater containing organic compounds and/or substances is an increasing problem due to its increasing toxic threat to humans and the environment. The removal of organic compounds has become an imperative issue due to stringent measures that are introduced by the Department of Environmental Affairs in South Africa to enforce regulations concerning wastes that emanate from petrochemical industries. Thus, wastewater containing these compounds must be well understood so as to device adequate treatment processes. In this study, the adsorptive capacity of PAD 910 polystyrenic resin originating from China and granular activated carbon (GAC) was evaluated for the removal of benzene, toluene, ethylbenzene and isomers of xylene (BTEX) from an aqueous solution. Batch studies were performed to evaluate the effects of various experimental parameters such as mixing strength, contact time, internal diffusion, adsorbates and initial concentration on the removal of the BTEX compounds. The experiments were conducted at the mixing strength of 180 rpm, in order to comfortably assume negligible external diffusion. The equilibrium isotherms for the adsorption of the adsorbates on the PAD 910 polystyrenic resin were analyzed by the Langmuir, Freundlich and linearized Dubinin-Radushkevich models at a pH of 5.86. The Langmuir model fitted the data adequately; this result was supported by the work done by Site (2001) which concluded that the Langmuir is the most practical model in representing the adsorption of aromatic compounds. The Langmuir model indicated that resin has the highest adsorption capacity of 79.44 mg/g and GAC has 66.2 mg/g. Resin was found to adsorb 98% of benzene, 88% of toluene, 59% of ethylbenzene, 84% m-;p-xylene and 90% o-xylene at an initial concentration of 14.47 mg/l. BTEX adsorption was a two-stage process: a short, fast initial period then followed by a longer, slow period corresponding to the intra-particle diffusion of BTEX molecules in macropores and micropores. The adsorption capacity was determined by total surface area accessible to BTEX and the availability of active surface chemical groups. The dependence of adsorption capacity on the surface of the two adsorbents and temperature was observed, suggesting the chemical nature of the BTEX adsorption. The interaction between BTEX/activated carbon was however weak and energetically similar to that of hydrogen bonds. Generally, BTEX adsorption was an exothermic process that combined physisorption and chemisorption. The PAD 910 polystyrenic resin had a greater specific surface area (SSA) of 1040 m2/g which yielded in higher capacity compared to GAC which had a low SSA of 930 m2/g. The normalized adsorption capacity was found to be higher for PAD 910 polystyrenic resin than GAC (0.66 and 0.27 mg/m2 respectively) which suggests that the resin has a good potential of the adsorbent for removing BTEX compound compared to GAC. Fixed bed columns were used to evaluate the dynamic adsorption behaviour of BTEX/PAD 910 polystyrenic resin through a dynamic column approach. The performance of small-scale fixed bed columns, each containing PAD 910 polystyrenic resin and the other containing GAC were evaluated using 14.47 mg/L of BTEX concentration. The columns with 32 mm diameter, studied bed depths of 40, 80 and 120 mm and flow rate of 6 ml/min were used in order to obtain experimental breakthrough curves. The bed depth service time (BDST) model was used to analyze the experimental data and design parameters like adsorption capacity, adsorption rate and service time at 20% and 60% breakthrough. BDST was also used to predict the service times of columns operated under different influent concentrations and flow rates to produce theoretical values that were compared to the experimental values. Adsorption model by Dubinin and colleagues (Dubinin, 1960), based on the theory of volume filling micropores was used to fit the measured adsorption isotherms. Agreement between the modelled and experimental results for GAC and PAD 910 polystyrenic resin using Dubinin-Radushkevich equation generally improved with increasing the surface area and produced reasonable fits of the adsorption isotherms for both GAC and PAD 910 polystyrenic resin. Granular activated carbon had a lesser performance compared to the PAD 910 polystyrenic resin, in terms of kinetic studies, and this finding was attributed to the pore structure which made accessibility of BTEX molecules more difficult in this study. The results indicate that PAD 910 polystyrenic resin show potential as an adsorbent for removing low concentrations of BTEX from wastewater. It is suggested that necessary treatment of GAC might improve the performance of this adsorbent by creating more mesopore volume and fraction which is essential to enhance adsorption rate. A substantial different SSA could be achieved through high porosity development in GAC by using templating method with a higher potassium hydroxide mixture ratio.

In the following study, the potential role that microorganisms play in the removal of Mn (II) and Fe (II) was assessed using biofilter sand and water samples collected from a biofiltration system (operated by Umgeni Water in KwaZulu-Natal, Nottingham Road, at the Nottingham combined school, South Africa) treating borehole water containing manganese and iron. Initially the presence of Mn (II) and Fe (II) oxidizing bacteria was demonstrated in the biofiltration system. Thereafter, the contribution of individual microorganisms to the overall removal of manganese and iron was assessed in the laboratory by determining the difference in metal oxidation in the presence and absence of active bacteria at neutral pH, simulating conditions in the biofilter. Controls were run to verify the elimination via physiochemical reactions occurring within the biofiltration system. Finally a diversity snapshot of the bacteria present within the biofilter matrix was established via analysis of a clone library. Viable bacterial counts for the biofiltration system were established using MSVP (minimal salts vitamins pyruvate) medium - plus added manganese sulfate or iron sulfate targeting Mn (II) and Fe (II) oxidizing bacteria - and R2A for heterotrophic bacteria. In the first experimental chapter, batch tests using MSVP were employed to determine manganese oxidation, by measuring the pH and ORP (oxidation reduction potential) in experimental flasks and controls over time. There was a clear drop in pH and a concomitant increase in ORP when an isolated manganese oxidizing strain (designated LB1) was grown in MSVP plus added manganese sulfate, indicating manganese oxidation. Based on physiological characteristics established by the VITEK-2 system as well as by 16S rRNA gene sequence analysis and MALDI-TOF (Matrix assisted laser desorption ionization-time of flight mass spectrometry) mass spectrometry of cell extracts, the isolate was identified as a member of the genus Acinetobacter. EDX (energy dispersive X-ray analysis) analysis of crystals formed in batch culture tests, containing MSVP plus either added manganese or iron sulfate, confirmed the ability of the isolate to oxidize both Mn (II) and Fe (II). The leucoberbelin blue colorimetric assay and batch tests using MSVP both demonstrated that in the presence of the isolated strain, Acinetobacter sp. LB1, the rate of Mn (II) oxidation at neutral pH was enhanced as compared to abiotic controls. In the second experimental chapter the difference in Fe (II) oxidation between biological and abiological systems at neutral pH was determined using batch tests run with Acinetobacter sp. LB1 and Fe (II) in saline. In addition, the rate of Fe (II) oxidation was also determined at acidic pH and at alkaline pH in experimental and control flasks. To determine Fe (II) removal under conditions simulating those in the biofiltration system, batch tests were set up using borehole water freshly collected from the biofiltration system. In order to verify the contribution of native microorganisms in the borehole water to Fe (II) oxidation, these flasks were spiked with bacterial strains isolated from the biofiltration system - Acinetobacter sp. LB1 and Burkholderia sp. strain LB2 - and two known iron oxidizing strains Leptothrix mobilis (DSM 10617) and Sphaerotilus natans (DSM 565) were used to determine the contribution of reference iron oxidizers to Fe (II) oxidation. A separate set of the same flasks with the addition of filter sand was used to qualitatively demonstrate iron oxidation as it would occur within the biofiltration system. The ferrozine assay was employed to quantify the amount of Fe (II) in batch tests employing saline medium and in batch tests employing borehole water. EDX analysis was employed to confirm the presence of Fe (II) in oxidation products in the batch test flask with filter sand spiked with Acinetobacter sp. LB1. In the presence of Acinetobacter sp. LB1 at neutral pH in saline medium, the rate of Fe (II) oxidation was very similar to that in the abiological controls thus demonstrating that the presence of metabolically active microorganisms does not per se enhance the oxidation of Fe (II) like in the case of Mn (II) at neutral pH. Surprisingly, in the heat inactivated control, apparently the highest amount of Fe (II) was oxidized. As expected, at acidic pH very little oxidation of Fe (II) took place and at alkaline pH almost all Fe (II) in the flasks was removed and small amounts oxidized as determined by the amount of Fe (III) produced. Batch tests using borehole water proved that native microorganisms within the biofiltration system were more efficient in the oxidative removal of Fe (II) from the system, in comparison to the reference iron oxidizing strains. In the final experimental chapter, the presence of biofilms with actively metabolizing cells was examined on a pooled sample of biofilter matrix from the manganese and iron filter using CLSM (confocal laser scanning microscopy) image analysis. DNA was extracted from the biofilm material associated with biofilter matrix to establish a diversity snapshot of the bacteria present within the biofilter matrix. ARDRA (amplified “rDNA” restriction analysis) analysis of the clone library revealed the presence of 15 unique OTU’s (operational taxonomic unit) based upon restriction patterns of amplified 16S rRNA genes of a total of 100 randomly selected clones. The majority of the clones were closely related to the genera Nitrospira and Lactococcus. Overall, 42% of the clones were assigned to the phylum Proteobacteria, 13% to the phylum Actinobacteria, 24% to the phylum Firmicutes and 21% to the phylum Nitrospirae. Overall, the results demonstrate that bacteria present within an established biofiltration system at neutral pH can contribute to the oxidative removal of Mn (II) and, apparently only to a smaller degree, to that of Fe (II) present in borehole water and that species within the proteobacterial genus Acinetobacter are potentially involved in the geochemical cycling of these two metals. Keywords: Biofiltration, iron and manganese oxidation, Acinetobacter sp. LB1, batch tests, 16S rRNA, MALDI-TOF MS analysis, Mn (II) and Fe (II) colorimetric assays, EDX analysis, biofilm formation, CLSM image analysis, 16S rRNA clone library Abbreviations: MSVP (minimal salts vitamins pyruvate), ORP (oxidation reduction potential), EDX (energy dispersive X-ray analysis), MALDI-TOF MS (Matrix assisted laser desorption ionization-time of flight mass spectrometry), rRNA (ribosomal RNA), ARDRA (amplified “rDNA” restriction analysis), CLSM (confocal laser scanning microscopy), OTU (operational taxonomic unit) / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.