Cyanide (CN) is used in the gold mining industry to dissolve gold from free milling, complex and refractory gold containing ores. Processing sulphide containing refractory ores using biooxidation as a pre-treatment has become increasingly important due to the depletion of free milling ores. The reaction of CN with reduced sulphur species during the cyanidation process results in the formation of thiocyanate (SCN), often at relatively high concentrations (> 5 000 mg/L). The SCN and residual free CN are deported with the tailings as components of the liquid fraction. The concentration of SCN often exceeds the legislated discharge specification, necessitating on-site treatment, while water would also require treatment before on-site recycling and reuse. Biological degradation of CN and particularly SCN in these effluents provides an alternative to the more traditional processes such as SO2 treatment or UV destruction. The traditional destruction processes focus on breaking the chemical bonds, through physical or chemical means, thereby converting the toxic CN and SCN species to less toxic compounds. These processes generally suffer from high reagent cost, incomplete removal of CN and particularly SCN species and the generation of by-products which require further treatment. A number of microorganisms are capable of utilising CN and SCN as a source of sulphur, nitrogen and carbon, as well as generating energy from their oxidation. Additional removal of metal-CN complexes may be achieved by adsorption to the cell surface or extracellular polymeric substances secreted by the cells. The activated sludge tailings effluent remediation (ASTERTM) process was developed for the biological treatment of especially SCN, but also free CN and metal-cyanide complexes, such as CuCN and Zn(CN)2. The basic ASTERTM technology consists of an aerated reactor, in which SCN and CN species are oxidised and a settler to facilitate the recovery of water and potentially biomass. The desire to expand the commercial application of the technology necessitated a more complete, fundamental understanding of the ASTERTM process and required focused, in-depth research. This research aimed to define the viable operating window for SCN destruction, as well as optimising practical SCN and CN destruction process conditions. The ASTERTM process relies on a complex microbial community, so understanding the community structure and metabolic potential for SCN and CN destruction, further enhanced the fundamental and mechanistic understanding of this bioprocess. The research contributed to the fundamental understanding of this technology and enhanced the commercial application thereof. The first step in defining the operating window was to investigate the effect of feed SCN concentration on the SCN destruction ability of the mixed microbial community. Experiments were conducted at feed SCN concentrations ranging from 60- 1 800 mg/L. Complete SCN destruction was achieved across the range at ambient temperature. The maximum SCN destruction rate was 15.7 mg/L.h at an initial SCN concentration of 1 400 mg/L. Temperature was investigated in the range of 10-45°C with an initial SCN concentration range of 60-180 mg/L. A maximum SCN destruction rate of 17.4 mg/L.h was measured at 35°C, with an initial SCN concentration of 180 mg/L. A wide pH range (pH 5.0-10.0) was tolerated, with optimal performance recorded at pH 7.0. This evaluation identified not only the optimum operating pH, but also highlighted the negative impact of a sudden pH change on the efficiency of SCN destruction. Residual SCN concentrations below 1 mg/L were achieved in all cases, which would allow for discharge or recycling of treated water. Floc (sludge) formation was observed in experiments with high initial SCN concentrations and indicated a possible stress response during these batch experiments. Floc (sludge) formation were taken as microbial cells imbedded within extracellular polymeric substances and not only an aggregate of cells. Evaluating the maximum potential for SCN destruction and optimising the operating conditions and system configuration was investigated using continuous reactor experiments. A maximum SCN destruction rate of 87.4 mg/L.h (2 098 mg/L.d) was achieved at a feed SCN concentration of 1 000 mg/L and eight hour hydraulic retention time (HRT) during these experiments. The formation of substantial amounts of sludge was observed, with attachment to the reactor surfaces. The maximum feed SCN concentration, where substantial destruction was measured, was at 2 500 mg/L, achieving a practical SCN destruction rate of 972 mg/L.d. Significant inhibition of microbial inactivity was observed beyond this feed SCN concentration. The microbial community was able recover performance, within six days, after an extended period (54 days) of inactivity when the feed concentration was reduced from 3 500 mg/L SCN to 1 000 mg/L. The nature of the accumulated biofilm did not appear to change during the period of limited SCN destruction activity. Calculation of specific SCN destruction rates was not possible due to the nature of the sludge and heterogeneous dispersion of microbial members. Biomass (cells embedded in the EPS sludge) loading experiments showed SCN destruction rates increased with an increase in biomass loading, but this relationship was not proportional. A 25-fold increased biomass concentration resulted in only a 2-fold increase in destruction rate, suggesting a mass transfer limitation. The sludge most likely offers protection against unfavourable conditions, such as high residual SCN concentrations, by presenting a mass transfer barrier, resulting in an SCN concentration gradient across the sludge matrix. This enhances the robustness of the process and would facilitate rapid recovery in the case of a system upset at commercial scale. This research is the first to demonstrate the effective removal of SCN in the presence of suspended tailing solids, under conditions well suited for commercial application. The maximum SCN destruction rate achieved was 57 mg/L.h in the presence of 5.5% (m/v) solids. Sludge formation was not observed in the reactors containing solids, despite substantial sludge formation under similar operating conditions in the absence of solids, most likely due to shear-related effects. Fluctuations in pH, due to the nature of the solid material, were identified to negatively impact reactor performance and pH control was required. Moreover, the type of solid particle was found to influence the SCN destruction rate showing a need for adaptation not only to the presence of solids but also to various types of solids that are to be treated. Treatment of residual CN in solution is critical to ensure safe disposal or recycling of water. Treatment of SCN and CN was successfully demonstrated at feed concentrations up to 2 000 and 50 mg/L, respectively. The presence of residual CN (0.5 mg/L) prevented complete destruction of SCN, while complete SCN destruction was measured in the absence of CN under identical conditions. A range of reactor configurations were investigated and the optimum system required biomass retention, by means of attached biomass and complete destruction of any residual CN prior to SCN destruction. Conversion of SCN-S to SO4-S was stoichiometrically proportional in solution, while the majority of the liberated nitrogen appeared to be assimilated. Pre-colonisation of the reactor with attached biomass is beneficial and removed the need for a solid-liquid separation unit, reducing the potential footprint of the process. Additional treatment capacity could be created by operation of reactors in series. The diversity of the microbial community responsible for destruction of especially SCN were shown to be far more extensive than initially expected. Initial molecular characterisation of the microbial community selected for 185 representatives of bacterial 16S rRNA genes, of which 106 non-identical genotypes were sequenced. In contrast, for the reactor containing solids, only 48 representatives were selected and 30 genotypes were sequenced. Bacteria implicated in SCN destruction in the reactor containing suspended solids were members from the genera Bosea, Microbacterium and Thiobacillus. In the absence of solids, members capable of SCN destruction were identified from genera including Thiobacillus and Fusarium. High-throughput genome sequencing, followed by sequence assembly confirmed the dominance of Thiobacillus spp. Metabolic predictions indicated the autotrophs, gaining energy from the oxidation of reduced sulphur intermediates produced during SCN destruction were the dominant community members. The potential for ammonium oxidation and denitrification within the microbial community was identified during analysis of the metabolic potential, based on the metagenomic sequence data. These would be required for complete remediation of wastewater. The data generated during the research led to the development of a conceptual model to describe the evolution of system performance. Following inoculation with planktonic culture the SCN destruction is performed by the planktonic microbial community. An increased residual SCN concentration results in floc formation and the colonisation of reactor surfaces by attached biofilm. A concomitant decrease in planktonic cell concentration was observed, while SCN destruction rates increased. The extracellular material provided a matrix for biomass retention, resulting in high cell concentrations, and provided some protection against high SCN concentrations by providing a barrier to mass transfer. The attached biofilm developed to the point where overall SCN degradation rates may become limited by reduced oxygen penetration. The research presented in this thesis has been used to inform the design and operation of the ASTERTM process at commercial scale, specifically with respect to the benefits of attached biomass and the demonstration that the process can be used in the presence of suspended solids. The latter has been particularly important in applications where the available footprint is constrained.

Bioethanol plays a significant role in the world of liquid biofuel. However, majority of bioethanol is produced from edible food crops such as corn and sugarcane that causes an increase in demand for vacant lands for food production and, subsequently, increase in the cost of food manufacturing. Therefore, alternative raw materials for bioethanol production are sought after, such as sugarcane bagasse which is a waste material from the sugar industry. South Africa, a net sugar exporter, has a large potential to produce bioethanol from sugarcane bagasse. This research focuses on the study of the production of bioethanol from glucose and xylose which are the two most abundant sugars in hydrolysed sugarcane bagasse. To date, no suitable wild type organisms can concomitantly ferment both glucose and xylose to ethanol efficiently. Options to address the co-fermentation of glucose and xylose include genetic modification of the selected microorganism to include both pathways - limitation in the understanding of the metabolic pathways regulations - or utilization of two microorganisms in co-culture or sequential culture e.g. Zymomonas mobilis and Pichia stipitis for efficient fermentation of glucose and xylose respectively. In this study, the dual micro-organism route is explored. There are numerous problems associated with co-culturing. Xylose, a non-preferred carbon source is only converted if the glucose concentration is adequately low due to catabolite repression. In order to increase xylose conversion, a low glucose concentration is required. Therefore, two stage sequential fermentation either in one or two reactors was tested. A high inoculum of suspended or immobilized Z. mobilis was inoculated in the first stage to convert the glucose rapidly. Varying reactor configuration, including the continuous fluidized bed, continuous stirred tank reactor (CSTR) and stirred batch reactor were considered. The products and residual substrate from this fermentation was then directed to a second stage, using either a CSTR or stirred batch configuration, with a high inoculum of P. stipitis in suspension culture for conversion of xylose. When immobilized, Z. mobilis was entrapped in calcium alginate beads. On the issue of ethanol tolerance, P. stipitis is generally more easily inhibited by ethanol (threshold ethanol concentration of 35 g L-1) compared to other ethanol producing strains such as Z. mobilis (threshold ethanol concentration of 127 g L-1) and Saccharomyces cerevisiae (threshold ethanol concentration of 118.2 g L-1). In order to overcome this, a continuous bioprocess was investigated to keep ethanol concentrations in Stage II below 35 g L-1 to prevent inhibition of metabolic reactions in P. stipitis. Further, ethanol fermentation by Z. mobilis requires obligate anaerobic conditions while xylose conversion by P. stipitis is optimum under microaerobic conditions. Therefore, oxygen was sparged into the second P. stipitis stage only. The following components were carried out in this project to improve the kinetic model and to find accurate kinetic data in the selected process of the two stage sequential fermentation. Firstly, where kinetic parameters were not available in literature, the kinetic parameter relationships of glucose and xylose utilization between different constructs of the same species were examined, for example, a wild type and engineered strain. This approach was used for glucose conversion using wild type Z. mobilis, owing to the ill-fit of available kinetic parameters with experimental results. In this study, the correction factors on estimated kinetic parameters from linear and non-linear regression when a xylose fermentation route was inserted recombinantly (S. cerevisiae RWB 217) into the native culture (S. cerevisiae CEN.PK 113-7D) were determined. From kinetic parameters of an engineered strain with the xylose-fermenting pathway (Z. mobilis ZM4 (pZB5)) and the correction factors, kinetic parameters of the wild-type Z. mobilis ZM4 were determined. Predicted rates of Z. mobilis ZM4 were then validated with experimental data generated in this study. Then, the optimum initial biomass concentration required to provide a faster volumetric rate of sugar utilisation and ethanol production, as well as the optimum oxygenation level for xylose conversion using P. stipitis achieved through appropriate aeration were investigated through experimental observation and using a MATLAB mathematical model developed through combination of the Andrews and Levenspiel's models, with oxygen, substrate, cell and product terms. Experiments were carried out to validate the kinetic model and data under anaerobic and microaerobic growth conditions in a batch process. The results showed that both increasing the initial biomass concentration (3 g L-1) and operating under optimum oxygenation levels (0.1 vvm) benefitted the ethanol production and yield by P. stipitis from xylose. It was also concluded that the addition of the oxygen effective factors in the developed model allowed for optimization of aeration in the fermentation system. Next, the custom kinetic model for fermentation process of bioethanol production was developed in Aspen Custom Modeller (ACM) and embedded in Aspen Plus. The model includes equations of vapour-liquid equilibrium (VLE), mass balance, and energy balance (e.g. molecular weight, thermodynamic phase equilibria, kinetic equation). The obtained results showed better agreement between industrial data and kinetic model (1% differences) than a stoichiometric model (9% differences). The simulation showed that ACM integrated into Aspen Plus allowed for complex biological processes to be accurately predicted for biomass growth, ethanol production and sugar consumption. Finding suitable microorganisms and process conditions for efficient glucose and xylose conversion is still currently a challenge and requires optimization. Therefore, this research focusses on improving the conversion of glucose and xylose to bioethanol, with specific emphasis on the fermentation systems used to maximize biomass efficiency, and ethanol yields and productivities. Manipulation of process conditions ranging from operation conditions (e.g. batch, fed-batch, continuous), process parameters (aeration, temperature, pH), immobilization technique and type of microorganism initially using kinetic models and thereafter validating with experimental data, therefore, offers a quick and strong foundation in improving bioethanol yields and productivities.

Crystal Violet (CV) decolourising deep-sea actinobacteria could provide a great source of novel redox biocatalysts that can be used in various applications such as removal of triphenylmethane dyes from contaminated wastewater and soil, degradation of aromatic environmental pollutants, biotransformation of antimicrobial agents and degradation of xenobiotics. CV is a triphenylmethane dye that has various applications, including use in medical, research and industrial applications, but its release into the environment poses a threat to aquatic life as it has characteristics of a biocide. Only a limited number of microorganisms are able to decolourise and degrade CV, and one of these proposed mechanisms by which they do so is the catalytic effect of oxidoreductase enzymes, including peroxidases, polyphenol oxidases and laccases. Triphenylmethane reductase has also been reported to be involved in decolourising CV, but the reaction involving this enzyme has not been studied systematically. Eleven deep-sea actinobacteria were investigated and found to decolourise CV by either biodegradation or biosorption. Gordonia sp. JC 51 was selected as a candidate for further study as it could decolourise CV efficiently and could tolerate high concentrations (1mM) of CV. A combination of spectral scan studies, dye decolourisation, biodegradation assays, enzymatic assays, SDS-PAGE, Native PAGE, TLC and LC/MS/MS methods revealed the mechanism involved in the decolourisation of CV. Gordonia sp. JC 51 decolourised CV via enzymatic and non-enzymatic mechanisms. However, true decolourisation of CV was performed via biodegrading enzymes. Triphenylmethane reductase and polyphenol oxidase was confirmed to be the enzymes involved. Leucocrystal Violet was identified as the metabolite produced. CV also was sequentially N-demethylated, oxidised and cleaved into smaller compounds such as Michler’s Ketone. In conclusion, Gordonia sp. JC 51 has potential as a whole cell biocatalyst and should be investigated further.

Acid rock drainage (ARD) is defined as acidic waste-water contaminated with sulphate and heavy metals which is generated through the oxidation of sulphidic ores in the presence of water and oxygen. Mining activities accelerate this process by bringing these ores to the surface where they are further crushed and, eventually end up in waste rock dumps and tailing impoundments where they continue to generate ARD into perpetuity. Active mining operations are mandated to prevent the discharge of ARD into the environment. This ARD is commonly remediated by expensive yet highly effective active treatment strategies such as high-density sludge processes and reverse osmosis. South Africa has an extensive history of gold and coal mining which has left abandoned mine workings with associated waste rock dumps throughout northern and eastern parts of the country. As many of these mines have long been abandoned, the responsibility to mitigate the environmental impact of the generated ARD lies solely with government. Although these diffuse sites often generate smaller volumes of less aggressive ARD compared to that generated through mine water rebound, the sheer number and the continual ARD generation from these sites is a severe threat to South Africa's already poor water security. Biological sulphate reduction (BSR) has long been considered an attractive option for the longterm remediation of these low-volume sources of ARD – but its implementation has shown mixed success. BSR is a process catalysed through the innate metabolism of sulphate-reducing bacteria (SRB) which coexist within complex microbial communities. SRB themselves are a highly diverse group of anaerobic microorganisms which use sulphate as a terminal electron acceptor. The sulphide and bicarbonate produced during BSR can be used to precipitate heavy metals and aid in the neutralisation of the ARD, respectively. The implementation of BSR is, therefore, a comprehensive remediation strategy for diffuse sources of ARD. The study of BSR, using various reactor configurations and operating conditions shows much promise. However, the microbial ecology of the complex communities within BSR systems, and their links to the performance of BSR processes, has received far less attention in published literature. This is not a result of underappreciation of the role microbial communities but rather a historical lack of tools, specifically high-throughput techniques, available to assess complex microbial consortia. It is asserted that the success of a sustainable BSR process developed for the long-term remediation of ARD requires an in-depth understanding the microbial communities associated with this process. The identification of the microorganisms which are key to the process, thosewhich threaten the stability of the community and the optimal growth conditions of these microorganisms, can be used to inform how these bioreactors are designed and operated. This study investigated the performance and microbial ecology of several continuous BSR reactors using culture-independent metagenomic sequencing approaches. The performance and microbial ecology of these reactors were evaluated at a range of hydraulic residence times (HRT) over the course of approximately 1000 days of continuous operation, from five- through to one-day(s). The tested reactor configurations included a continuous stirred tank reactor (CSTR), an up-flow anaerobic packed bed reactor (UAPBR) and a linear flow channel reactor (LFCR) that were each operated in duplicate and supplemented with either lactate or acetate as an electron donor. The different reactor configurations and supplied electron donors, as well as the varied applied HRT, generated a range of microenvironments which were hypothesised to lead to the divergence of the initial microbial community of the inoculum and generate numerous distinct microbial communities throughout and across the reactor systems. 16S rRNA gene amplicon sequencing was used to assess the microbial community structure of the numerous populations across the reactor systems and monitor how these communities responded to the change in the applied HRT. Genome-resolved metagenomics was employed in parallel to recover the genomes of all predominant microorganisms identified through gene amplicon sequencing. This allowed the interrogation of the composition of the respective microbial communities as well as the genetic potential of each microorganism and encompassing the communities represented within specific reactor environments. The CSTRs were selected as these systems are characterised as well-mixed, support solely suspended biomass and kinetic equilibriums are achieved rapidly. This allows the performance of these reactors to be predictable and provides a benchmark to which the LFCRs and UAPBRs could be compared. The lactate-supplemented CSTR performed largely as anticipated based on available literature, demonstrating a maintained sulphate conversion of approximately 55% over the course of the study. The reactor achieved a maximum observed volumetric sulphate reduction rate (VSRR) of 17 mg/ℓ.h at a one-day HRT. The system supported a low SRB diversity, constituted almost entirely by a Desulfomicrobium and two Desulfovibrio operational taxonomic units (OTUs). The acetate-supplemented CSTR was able to maintain sulphate reducing performance at HRT where complete washout of SRB had been predicted based on literature. This reactor exhibited a maximum VSRR of 10.8 mg/ℓ.h at a 1.5-day HRT and was dominated by the same Desulfovibrio and Desulfomicrobium observed in the lactate-supplemented CSTR, along with several other SRB genera at lower abundance. The LFCRs demonstrated an approximately ten-fold greater biomass retention than the corresponding CSTRs. This was facilitated through the incorporation of carbon microfibres, whichfacilitated microbial colonisation and biofilm formation within the reactors. Surprisingly, the lactate-supplemented LFCR, underperformed compared to the lactate-supplemented CSTR, achieving a maximum VSRR of 14.8 mg/ℓ.h at a one-day HRT. This reduced performance, in spite of the enhanced biomass retention, was concluded to result from the out-competition of lactateoxidising SRB in the reactor by Veillonella and Enterobacter OTUs. The acetate-supplemented LFCR exhibited a period of underperformance before recovering and subsequently demonstrated a maximum VSRR of 17.1 mg/ℓ.h at a one-day HRT. Evaluations of the microbial communities of this system during the HRT study revealed a dramatic shift in the SRB communities from being dominated by Desulfatitalea and Desulfovibrio to being dominated predominantly by Desulfomicrobium and Desulfobacter. The UAPBRs are governed by plug-flow which resulted in the generation of gradients of decreasing substrates and increasing products throughout the height of the reactors. This, as hypothesised, resulted in the stratification of the microbial communities throughout the height of these reactors. This allowed many associations to be made between specific microorganisms and their ideal growth environments. Both UAPBRs demonstrated competitive sulphate reducing performance. The lactate-supplemented UAPBR proved especially successful as this system was able to maintain >95% sulphate conversion at one-day HRT, corresponding with a VSRR of 40.1 mg/ℓ.h. The performance of this reactor was attributed to the significant quantity of retained biomass and the successful harbouring of lactate-oxidising SRB towards the inlet zone of the reactor as well as propionate- and acetate-oxidising SRB towards the effluent zones of the reactor. The acetatesupplemented UAPBR exhibited a maximum VSRR of 23.2 mg/ℓ.h at a one-day HRT and a maximum sulphate conversion of 79% at a 2.3-day HRT. The stratification of the microbial communities within the acetate-supplemented UAPBR was less pronounced than the lactatesupplemented UAPBR, as a result of the fewer available volatile fatty acid species. However, the stratification which was observed in this system could be used to postulate the growth kinetics associated with the identified SRB – a Desulfobulbus was associated with rapid acetate oxidation in the inlet zone while a Desulfatitalea and a Desulfosarcina could be implicated in sulphate scavenging in the effluent zone of this reactor. This proved particularly valuable for elucidating the roles of these same SRB in the well-mixed reactor systems. Genome-resolved metagenomics was employed to recover the genomes of the microorganisms identified in these systems and determine the metabolic potential of these microorganisms. Hydrogen-evolving hydrogenase genes were found to be widespread in genomes not capable of sulphate reduction. In contrast, hydrogen-consuming hydrogenases as well as autotrophic gene pathways were common amongst SRB genomes. The ubiquity of hydrogenase genes in these environments indicated that inter-species hydrogen transfer was an important feature within thesemicrobial communities. The dual consumption of both acetate and hydrogen was concluded to have facilitated the maintained sulphate reducing performance of the acetate-supplemented reactor systems at short HRT where system failure had been predicted. Indices of replication (iRep) were used to estimate the instantaneous growth rates of the microorganisms from metagenomic shotgun sequencing datasets. This revealed that, at a four-day HRT, the microorganisms within the biofilms were comparably active to planktonic microorganisms. This, together with the dynamic changes in the composition of these biofilms during the HRT study, suggests these biofilms are even more active and competitive than previously thought. The combined use of next-generation gene amplicon sequencing and genome-resolved metagenomics has given unprecedented insights into the microbial communities of BSR reactor systems. Using this approach, it was possible to uncover a seldom discussed form of hydrogen cycling within BSR systems and has shown that there is no ‘one-size-fits-all' approach when inoculating BSR reactors. The SRB within these systems were often highly specialised to particular environments, specific electron donors and each showed differing growth kinetics. The success of long-term, semi-passive BSR reactor systems would benefit greatly from the tailoring of SRB inoculums informed by the chosen reactor configuration and operating conditions. The outcomes of the kinetic reactor experiments have led to several recommendations for the design and operation of these systems.

Water pollution is recognized as one of the major environmental problems in the mining industry. This has been compounded with an increase in agriculture activities. Water pollution is a major problem on copper and coal mines throughout the world and Zambia, the focus of this study, is no exception. Worldwide freshwater resources, which provide important ecosystem services to humans, are under threat from rapid population growth, urbanization, industrialization and abandonment of wastelands. There is an urgent need to monitor and assess these resources. In this context, the physical, chemical and ecological water quality of the Munkulungwe Stream located on the Copperbelt of Zambia, was assessed with possible contamination from Bwana Mkubwa TSF, agriculture activities and subsequent impact on the surrounding community. The chemical and physical parameters were assessed at four sampling locations. Sampling site S1 was located on the Munkulungwe stream upstream of Bwana Mkubwa TSF, S2, S3 and S4 were on the main stream downstream of Bwana Mkubwa TSF. In addition, a macroinvertebrate composition analysis was performed to estimate the quality of water using the biotic index score. Finally, the relationship between physiochemical parameters and biotic index score was analysed to interrogate their inter-relationship with respect to water quality. The results showed that the average values of dissolved oxygen (DO) of 4.52 mg/l, turbidity (40.96 NTU), Co (0.24 mg/l), Pb (0.25 mg/l), Fe (0.36 mg/l) and Mn (0.22 mg/l) downstream exceeded international standards for drinking water. Upstream, the values of Co, Pb, Fe and Mn were within acceptable standards for drinking water, DO and turbidity were above acceptable standards. The metal concentration and total dissolved solutes were impacted by closeness to the mine tailings deposit with the heavy metal concentration being highest at S2 and S3. Moreover, high turbidity levels revealed that land erosion induced by agriculture activities is a severe problem in the area. Physical parameters were high in the rainy season due erosion escalated by rains while chemical parameters were high post rainy season. During the rainy season, the chemical contaminants are diluted and thus they are not such a big impact, but they tend to concentrate up during the dry MDNLEE001 III season. The stream at sampling points S2 and S3 was dominated by species tolerant (leech, Isopod and Snail: Pouch) and semi tolerant (Blackfly larvae and Amphipod or Scud) to pollution. The change in season influenced the composition of macroinvertebrates, with the number of species increased post rainy season. The average biotic index score (2.5) showed that the stream condition is not good, it is slightly polluted. The results showed that water quality downstream was substantially affected by Bwana Mkubwa TSF, agriculture activities and is likely to affect human health and food security. It is recommended that groundwater surrounding tailings dams should be monitored in both active and abandoned mines. Curtain boreholes around a tailings dam can be drilled and the water extracted and treated so that it doesn't contaminate other water bodies. To improve the environmental management of mining related impacts in Zambia, mining areas should be completely rehabilitated. There is need for remediation strategies for abandoned mine sites. Constructed wetlands, roughing filtration and phytoremediation are highly promising techniques, as they are reliable, cheap, effective and sustainable.

The environmental impacts of coal processing wastes are a challenge in South Africa as large amounts of coal wastes are produced annually, pegged at 60 million tons per year according to Eberhard (2011). Whilst the fossil fuel-based industry is in decline globally, coal is likely to remain the dominant source of power in South Africa. The major environmental impacts reported in several studies are water pollution and soil quality degradation due to acid rock drainage (ARD) and its associated elevated levels of elements and salts. Several studies have shown the environmental performance of the wastes to be dependent on the geochemical properties of the wastes. Owing to the complex nature of coal wastes, their characterisation using tools developed for hard rock ores is associated with inconsistency and uncertainty. As a result, the South African coal processing wastes are poorly characterized and the associated risks not well understood. This study investigates the reliability of relevant characterisation techniques and interpretation of characterisation data in terms of the environmental risk potential of coal wastes. The outcomes of the study address some of the uncertainties and deficiencies arising from the current characterisation tools and evaluate potential environmental risks posed by coal processing wastes. Laboratory-scale characterisation of the physio-chemical properties and of ARD and elemental risk potential of two ultrafine coal waste and one discard waste sample were conducted. Evaluation of accuracy and repeatability of selected analyses was conducted on a certified coal standard. The selected analyses tested for accuracy and repeatability were total sulphur analysis by Leco and Eschka methods in addition to elemental analysis by wavelength dispersive x-ray fluorescence (WDXRF), inductively coupled plasma mass spectrometry (ICP-MS), inductively coupled plasma atomic emission spectroscopy (ICP-AES) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The ISO 157:1996 and ACARP C15034 protocols for assessment of sulphur forms were also compared and evaluated for precision using the coal standard and coal waste samples. Conversions of the sulphur species under static ARD tests were also studied to understand the sulphur species behaviour and implication on ARD potential. The mineralogy of the coal wastes was evaluated from a quantitative evaluation of minerals by scanning electron microscopy (QEMSCAN) and quantitative x-ray diffraction (QXRD) analysis. In addition, conventional net acid generating (NAG) and acid-base accounting (ABA) static tests were enhanced through extended boil NAG tests to assess the organic acids effect on the NAG capacity. The static tests were validated by theoretical ARD calculated from mineralogy as well as biokinetic shake flask tests which gave the timerelated acid generating behaviour of the coal waste samples. Sequential chemical extractions combined with a simple score and ranking protocol were subsequently used to evaluate the potential water and soil-related risks associated with environmentally available elements and salts in the coal wastes. The results showed both the Leco and Eschka methods to be highly precise (±0.01-0.03 % standard error) but the Leco was more accurate (±3.1 % compared to ±12.5 % relative standard error (RSE)). The total sulphur content of the coal processing waste was less than 2 %. The ISO157:1996 and ACARP C15034 protocols gave comparable and slightly different results but the latter was more precise in sulphate analysis. Furthermore, the ACARP protocol could differentiate the acid forming sulphates from the soluble sulphates giving a better theoretical maximum acid producing potential. The sulphur species from the two chemical methods and QEMSCAN mineralogy showed 52-61 %, 12-26 % and 21-43 % to be sulphide, sulphate and organic/low-risk sulphur respectively. The conversion of the sulphur species showed that partial solubilisation of sulphides in ANC and partial conversion of organic/low-risk sulphur under NAG tests can cause an over or underestimation of ARD potential. The static ARD tests has shown the Witbank coal discards sample to be potentially acid forming (PAF) (9.2-25.9 kg H2SO4/Ton), Waterberg coal slurry to be non-acid forming (NAF) (-68.6 to -46.8 kg H2SO4/Ton) and Witbank coal slurry to be uncertain (-12.1 to 9.9 kg H2SO4/Ton). The extended boil NAG tests showed organic acids effect on the Witbank coal slurry likely caused an overestimation of the NAG capacity. Validation of the static tests by biokinetic tests and ARD calculated from mineralogy classified both Witbank samples as PAF and the Waterberg sample as NAF. The results also showed the net acid producing potential of the coal wastes to depend on the mineralogy of the samples. The elemental results showed WDXRF and LA-ICP-MS analysed most of the elements accurately within ±10 % RSE and that a combination of techniques provides more reliable and accurate results. The analyses showed the coal waste to contain significant amounts of environmentally sensitive elements like Cr, As, Mo, Sb, Se. The ranking and scoring of potentially available elements under oxidising leach conditions evaluated Fe in Waterberg coal slurry and Witbank coal discards to pose high risk in drinking water while S (as sulphate), Pb, Sb, Mn, As, Al and Hg in the three samples pose moderate risk. This case study evaluated the accuracy and precision of commonly used analytical techniques and applicability of risk evaluation protocols for coal processing wastes. The research outcomes underlined some factors that cause uncertainty and inconsistency with the evaluation of ARD potential of coal wastes. The findings highlighted the need to validate and complement the characterisation data using various tools and risk evaluation protocols to overcome specific limitations. The results also indicated the coal wastes have the potential to cause environmental impacts from ARD and elevated concentration of elements and salts, thus providing a basis for designing and implementing waste management strategies which minimise these risks. The mineralogy and elemental composition of coal wastes showed enrichment of elements and presence of potentially usable and economically valuable constituencies for future studies on value recovery. Characterisation of coal processing wastes for air pollution impacts is recommended for future studies as well as a study of ARD behaviour under continuous flow systems to more closely represent the conditions in dump disposal scenario.

The South African economy is an energy-driven economy which relies on coal to meet most of its energy demands. Coal mining has resulted in the generation of coal waste over 60 million tonnes, annually. Apart from the huge footprint of this waste, the sulphide minerals contained in the waste have resulted in the generation of acid rock drainage (ARD). A lot of techniques have been developed to prevent and mitigate ARD, however most of these techniques have fallen short in terms of meeting their desired objectives due to the long-term nature of ARD generation which can persist for hundreds of years after mine closure. This has resulted in emphasis being put on long-term prevention techniques that remove ARD risk over treatment techniques. One prevention technique which has shown good technical potential is the two-stage flotation method developed for desulphurisation of hard rock tailings and coal fines, developed at the University of Cape Town. On desulphurising coal, the first stage produces an upgraded coal product that may be sold, with the second stage used to separate the tailings from the first stage into targeted high-sulphide and low-sulphide fractions which may then be appropriately used or disposed of. An economic assessment of the process showed across a wide range of coal wastes the high cost of oleic acid used in the first stage of the process as a collector was a major contributor to the operating costs. The investigation undertaken in this thesis looked at the potential of algal lipids and their derivatives as biocollectors to replace the oleic acid collector in the desulphurisation process at the laboratory scale. A review of cost was carried out for a process that used raw algal lipids (RALs) or fatty acid methyl esters (FAMEs), which are derived from RALs through transesterification. Batch flotation experiments were used to assess the performance of the two bioflotation reagents in comparison to oleic acid and dodecane, an alternative but less successful chemical collector. The algal lipids cost review was a desktop study which was done by adapting literature data from Davis et al. (2014) which focused on economic evaluation of algal lipid biofuels production pathways. Results from laboratory experiments for two different coal waste feed samples showed that the performance of RALs and FAMEs was similar to that of oleic acid for the sample that was high in ash and sulphur, and better than oleic acid for the sample that was low in ash and sulphur. For example, the product from Site 1 discards from Waterberg had 24.37% ash and 2.76% sulphur using FAMEs, 26.13% ash and 2.56% sulphur with RALs, and 23.48% ash and 2.41% using oleic acid, at a reagent dose of 2.8 kg/t for all reagents. For Site 2 waste tailings from the Witbank area, the product had 23.17% ash and 0.72% sulphur when FAMEs were used as collector, 22.75% ash and 0.75% sulphur with RALs, and 20.18% ash and 0.74% sulphur using oleic acid, at the same reagent dose. Discards from Site 1 had an initial ash and sulphur content of 47.61% and 5.71%, respectively. Site 2 waste tailings had 25.56% ash and 0.91% sulphur before flotation. Increasing biocollector dosage resulted in higher yields with a compromise on the upgraded coal quality. The pH tests showed that the performance of the two bioflotation reagents was best at pH 4 in terms of yield. However, increasing the pH of the process from the natural pH of the sample (pH 2.7) to 7 resulted in collection of more ash and sulphur, thus reducing the product quality. The algal lipids cost review showed that RALs and FAMEs were potentially 20 to 21% cheaper than oleic acid, with more room for improvement. Both the laboratory experiments and the technical evaluation showed that algal lipids and their derivatives have the potential to replace oleic acid in the two-stage desulphurisation process for coal waste to obtain a saleable quality coal product while simultaneously decreasing the impact of ARD from coal waste.

Colonisation of mineral surfaces by acidophilic microorganisms during bioleaching is important for accelerating the extraction of valuable metals from mineral sulfide ores of varying grades through biohydrometallurgy. It also influences acid formation and mineral deportment from sulfidic waste rock generated in mining processes and is key to its comprehensive waste rock characterisation for acid forming potential. This study assesses mixed mesophilic microbial interactions with, and colonisation of, pyrite concentrates and pyrite bearing waste rocks. The assessment of these interactions was carried out in this study in a synergistic qualitative as well as quantitative manner, with a particular focus on heap bioleaching for metal extraction and on disposal of waste rock, the latter through the case of characterisation of ARD generation potential. Using the tools developed, both the course of colonisation and development of metabolic activity with time of colonisation, as well as their correlation with leaching performance were studied. Furthermore, specific operating parameters such as ore grade and irrigation rates were explored. Finally, the application of this knowledge in a characterisation study was explored. To achieve the set of tools required for this study, two quantitative techniques were refined to characterise these microbial-mineral interactions. In the first, an isothermal microcalorimetric (IMC) method was developed and optimised to determine microbial colonisation of mineral surfaces quantitatively as a function of surface area (m-2 ). Three IMC configurations were considered: colonised pyrite-coated beads submerged in fresh media; beads submerged in cell free leachate; and beads in an unsaturated bed, each in the IMC vial. The highest heat output was measured in the unsaturated bed (263.3 mW m-2 ). The consistency of heat produced by the colonising microorganisms was determined through reproducibility studies. Using IMC, chemically and microbially facilitated pyrite oxidation rate studies were performed on unsaturated beds with varying surface area loadings, correlating to varying bead number. Results obtained showed similar normalised oxidation rates per surface area across the surface loadings. However, with more microbially colonised surface area loaded, the maximum heat generated was reached more quickly. This suggested that there was reagent (possibly O2) limitation in the system, which restricted microbial activity and its associated heat generation. Reagent limitation in the system was tested and validated through varying the O2 availability in the IMC vial by air displacement with CO2 and N2 gas, with the systems containing less O2 showing limited activity. Collectively the data showed that high activity, facilitated microbially, was achieved in unsaturated systems in a reproducible manner. Secondly, oxidation rates were determined and O2 limitation in the system was overcome. This then fundamentally informed the determination of activity from microbial-mineral interaction, using IMC, as a function of surface area. Secondly, a detachment protocol developed at UCT to recover microbial cells from surfaces of crushed and agglomerated ore to assess microbial growth rates and distribution in the ore bed, including cells in the interstitial phase and those weakly and strongly attached to the ore surface, was refined to assess colonisation of the finely milled pyrite-bearing concentrate or waste rock coated onto glass beads in continuous flow assays. The detachment protocol was assessed quantitatively by measuring initial and residual microbial activity, as a function of wash number, using IMC, thus providing a new level of confidence in the method. Mineral surfaces were visualised using scanning electron microscopy (SEM) following detachment for qualitative assessment. These data, together with microscopic enumeration of detached cells with increased number of washes, allow refinement of the assay and showed that six washes provided reliable estimation of mineral associated microbial cells. Extracellular polymeric substances (EPS) produced in this process were extracted using crown ether and the capsular bound components analysed. The analysed components included lipids (4.2 %), iron (16.4 %), DNA (26.8 %), and total carbohydrates (28.5 %), which are typical components of EPS. The carbohydrate fraction was further resolved to trehalose (26.2 %), fructose (36.5 %) and galactose (37.3 %) sugar monomers. The analysed EPS components confirmed presence of the EPS secreted by cells colonising the mineral ore or waste rock surface in a flow-through system, and visualised via SEM. The microcalorimetric approach developed together with the refined detachment method were applied to samples from a flow-through mini-column system, used to simulate microbe-mineral contacting in a heap. Here, the colonisation of pyrite concentrate by a mixed mesophilic culture of iron and sulfur oxidising microorganisms was assessed progressively over 30 days. The progression of mineral colonisation in the mini-column system was monitored using a combination of IMC, scanning electron microscopy, detachment method and conventional wet chemistry measurements. We observed an increase in the heat output from the colonised surfaces of pyrite mineral concentrate caused by oxidative reactions facilitated by mineral-microbial biofilm. This confirmed that the attached microorganisms were metabolically active and facilitated ongoing mineral leaching through regeneration of lixiviants. Correlation was shown between number of cells detached from the mineral surface and the heat generated, with a constant heat output per cell observed until day 15 of operation. Thereafter, the measured heat generated per cell increased, suggesting reduced efficiency of cell detachment owing to increasing firm attachment, or the lack in separation of single cells embedded within EPS matrix (clumps observed under light microscope after detachment). Using IMC to quantify the activity of the residual microorganisms on the mineral surface following detachment, it was confirmed that >95% of activity was detached through this protocol, hence the lower detached cell numbers determined following EPS formation were attributed to clumping of the detached cells. This correlated to an increased presence of EPS and was supported by SEM observation. Following the study of pyrite concentrate, colonisation of two pyrite bearing waste rock samples was assessed, with simultaneous establishment of the flow-through mini column biokinetic test configuration that resembles open flow in the waste rock dump. The flowthrough configuration was run alongside the refined UCT-developed batch biokinetic test using suspended mineral. In this study, two pyritic waste rock samples, liberated by milling, were characterised using three biokinetic test approaches: the slurry batch test (BT), the batch test using mineral-coated beads (BT-CB) and flow-through column test with mineral-coated beads (FT-CB). Our results have shown through static tests, solution redox potential and pH analysis that both waste rocks were acid forming. Furthermore, it was demonstrated in the FT-CB system that microbial proliferation on the waste rock surfaces progressed with time such that oxidative exothermic reactions facilitated by the increasing microbial presence on the surfaces were demonstrated using Isothermal microcalorimetry. This study presents and informs the on-going refinement of the biokinetic test through establishment of a flow-through test for ARD characterisation while providing insight into the role of the microbial phase in ARD generation. Microbial-mineral association was assessed under various operating conditions, including two solution flow rates (60 and 4 ml h -1 ) and minerals of varying sulfide content, including a pyrite concentrate (96 % pyrite), a high sulfide waste rock (33 % pyrite) and a low sulfide waste rock (14 % pyrite). Mineral grade impacted the activity of mineral associated microorganisms with higher activities observed on a mineral surface with high sulfide content. The activity measured from microorganisms that were associated with the pyrite concentrate was 827 mW m-2 at a 60 ml h -1 flow rate, whereas activity measured on low and high sulfide waste rock (PEL-LS and PEL-HS) were 293 mW m-2 and 157 mW m-2 respectively operated on the same flow rate. On decreasing the flow rate to 4 ml h -1 , the activity of microbial cells on PEL-LS and PEL-HS were 153 mW m-2 and 146 mW m-2 respectively. This study showed that the growth of microbial cell numbers coupled with metabolic activity is important to facilitate accelerated dissolution of sulfidic mineral surfaces. The rate of oxidation increased in the presence of EPS and thus EPS was further analysed, and its composition was confirmed. Overall, this study contributed to the understanding of microbial colonisation of mineral surfaces in a non-destructive quantitative manner. This study thus demonstrates the ability to measure and track both the growth and activity of microorganisms that are associated with mineral surfaces. This is important as it provides an approach to understanding microbe mineral surface interactions and, therefore, potential strategies to increase microbial colonisation of low-grade minerals that house valuable metals, during commercial heap bioleach processes. Furthermore, the ability to monitor progressive growth and activity of mineral associated microbial communities within a flow-through biokinetic test, as successfully demonstrated in this study, has the potential to significantly enhance current management of mine waste materials and ARD mitigation strategies. Therefore, on-going investigations of progressive microbe-mineral interactions will continue to be valuable both in terms of bioleaching for metal recovery and the mitigation of ARD through effective characterisation of mine waste material.

Includes bibliographical references. / The mining and beneficiation of gold generates large tonnages of waste, with up to 99% of mined gold ore discharged as waste. The waste generated contains unoxidized sulfides that when exposed to air and water react to form acid, which results in acid rock drainage (ARD). ARD is usually associated with low pH, high sulfate content and elevated concentrations of toxic elements. The mobility of ARD affects our scarce water resources, land and aquatic species. Methods applied to treat ARD do not provide a walk-away solution and they are either expensive or difficult to maintain. The best solution to completely eradicate ARD is to prevent it from the source. However, the effectiveness of ARD prevention depends on the accuracy of predicting future drainage quality. This can be done by using ARD prediction tests, which are generally classified as either static (acid base accounting, ABA, net acid generation, NAG) or kinetic (column leach, humidity cell, biokinetic test). There is no single test capable enough to accurately predict acid generating potential. It is therefore usual practise to conduct more than one test and cross-check results to ensure that the appropriate conclusions are made. In doing so, the reliability of the tests is improved but in some cases the different test results do not correlate. Mineralogy is an analytical technique that can be used to understand the nature of the errors and to better understand the leaching behaviour of minerals in the different tests. This study uses mineralogy to analyse both static and biokinetic test results of a Witwatersrand gold sample in order to improve the understanding of behaviour of mine wastes under different ARD prediction test conditions. A run-of-mine gold sample from the Witwatersrand region in South Africa was used as a case study to explore the mineral leaching behaviour for different ARD prediction tests.

Low solubility of oxygen has resulted in high bioreactor energy requirements in order to supply sufficient oxygen to aerobic bioprocesses. It is desirable to reduce energy consumption in bioreactors to benefit environmental sustainability as well as economic feasibility. This is particularly important with the resurgence of interest in bio-based commodity products. Some research has suggested that venturi aeration of bioreactors will reduce energy consumption by eliminating the need for air compression, while at the same time maintaining or improving oxygen transfer rates. On the other hand, there are findings that suggest venturi aeration has lower energy efficiency than conventional sparging and oxygen transfer rates achieved are too low sustain aerobic biological processes. A comparison of the aeration of geometrically-similar reactors using the same analytical methods to determine kLa is not available in the literature. This comparison should also address analysis of energy input including energy used for compressing gas sparged into a stirred tank reactor; the investigation of mass transfer rates at higher flow rates (vvm) in venturi-aerated reactors and resulting cell response to these higher flow rates. This is the topic of the dissertation presented. In this laboratory scale study, venturi aerators were characterised and energy consumption as a function of oxygen mass transfer rates compared with a geometrically identical aerated stirred tank reactor by evaluating the volumetric mass transfer coefficients (kLa). The kLa was investigated in varying reactor setups using the dynamic gassing-in method.