Includes bibliographical references. / The traditional batch brewing process is characterised by serial yeast propagation to build sufficient yeast for pitching. This results in cyclic variations in yeast environment, leading to a slow brewing process. In high gravity brewing the carbohydrate utilisation is inefficient as a result of the Crabtree effect that occurs in the presence of high sugar concentration. When optimising the brewing process the characteristics of conventional batch brewing should be maintained. Fed-batch propagation of yeast is used to improve carbohydrate utilisation and the yeast biomass formation by controlling nutrient supply.

In 2010 the incidence of neo-natal Group B Streptococcus (GBS) disease in South Africa was 3 per 1000 live births, more than twice the global average of 1.21 per 1000 live births. A recent life cycle impact assessment showed that a new vaccine against GBS disease in South Africa could have a potential value of $ 2 million - $ 4 million /kg (R 25 million - R 50 million /kg), as an attractive investment opportunity if a novel process can be successfully synthesised and licensed commercially. In the current global market new biopharmaceutical products require innovative and expedited development pathways. To achieve this, low-cost analytical tools with short turnaround times are needed to assist with process development decision making. Process simulation is one such tool which has been shown to be useful for evaluating process development decisions without the typically expensive investment required for experimental development of a new process. Three technology platforms (stainless steel, single-use, and a hybrid of both) were identified for use in a novel process to manufacture a GBS serotype III polysaccharide-protein conjugate antigen, for formulation into a vaccine against GBS disease. The three technology choices were compared and evaluated for the novel process at two fermentation scales of 20 L and 200 L, with cost of goods (COG) used as a comparison of economic performance for the six different scenarios. It was hypothesised that single use technology would yield the lower COG at both scales compared to stainless steel. Based on a literature survey, single use technology should require lower capital costs for pilot scale processes and should also have lower operating costs due to single use equipment not requiring sterilisation in place (SIP) and cleaning in place (CIP). It was further hypothesised that hybrid technology would yield the lowest COG by combining the best properties of stainless steel and single use technologies. A 3 x 2 factorial experiment design was used to structure the simulation exercise with three technologies at each of the two scales. A GBS serotype III process model was synthesised from literature sources, with fermentation stoichiometry based on an empirical material balance and fermentation kinetics fitted to a two-parameter Monod kinetic model. Equipment, consumables, and raw materials specifications were made using literature and empirical models. A base case simulation model, built for 20 L scale using stainless steel technology, was developed into the five additional scenarios. The best performing scenario in terms COG was then selected for sensitivity analysis using three parameters: fermentation titer, solid-liquid separation efficiency, and electricity dependence on diesel generation. At 20 L scale there was little difference in COG between the three technology options, with COG range across the three platforms of $ 9.7 million - $ 9.8 million /kg. At 200 L scale the best performing technology was stainless steel with a COG of $ 3.7 million /kg, which was $ 600 000 /kg less than the COG for single use of $ 4.3 million/kg. The difference was due to a higher cost of consumables for single use technology, and negligible differences in capital costs for single use over stainless steel. The effect of SIP and CIP costs on operating cost for stainless steel technology was found to be small compared to the greater consumables cost for single use. The 200 L stainless steel process was found to be sensitive to fermentation titer, with an increase in titer to 600 mg/L resulting in the lowest COG of $ 2.2 million /kg. The process was found to be least sensitive to electricity dependence on diesel, with only a $ 60 000 /kg increase in COG when 75% of electricity was derived by diesel generator. The hypothesis was disproved, with single use technology having the higher COG at both 20 L and 200 L scales compared to stainless steel technology. Hybrid technology did not yield the lowest COG either, instead resulting in a COG somewhere between stainless steel and single use. Stainless steel technology outperformed single use and hybrid technologies in COG at both scales, contrary to both parts of the hypothesis. A process to make a GBS vaccine could be profitable at scales of 200 L and above using stainless steel technology. Process simulation modelling was effective for evaluating process technology options without performing costly physical experiments. The simulation exercise provided valuable information on the economic impact of process development decisions as well as context specific information for the South African context. This methodology is therefore recommended for commercial biopharmaceutical process development, particularly for evaluating techno-economic scenarios in different decision pathways during the development process.

The use of laboratory scale Microbial Fuel Cells (MFCs) for the combined generation of electricity and the treatment of wastewater has been well documented in literature. In addition to this the integration of MFCs into wastewater treatment reactors has also been shown to have several benefits. These include the improved treatment of wastewater, reduced solid waste and the potential to offset the energy costs of the process through the generation of electricity (Du et al., 2007). The treatment of sulphate-rich wastewater, and in particular Acid Rock Drainage (ARD), has become of increasing importance in water sparse countries like South Africa where mining is currently and has taken place. A semi-passive method of continuous ARD waste treatment is currently being investigated within the Centre for Bioprocess Engineering Research (CeBER) (van Hille et al., 2015). This research involves the use of a Linear Flow Channel Reactor (LFCR) designed for combined biological sulphide reduction and sulphide oxidation to yield a sulphur product. Sulphate Reducing Bacteria (SRB) mediate the biological sulphide reduction. Chemical and biological sulphide oxidation takes place in a Floating Sulphur Biofilm (FSB) on the surface of the reactor and is mediated by Sulphide Oxidising Bacteria (SOB). Sulphate-rich wastewater can therefore be remediated through total sulphur species removal.

[Page1 is in the PDF] The aim of this study was to investigate and compare various methods to enumerate the number of bacteria in a minerals biooxidation system. In this system most of the bacteria are attached to fine particles of ore and therefore cannot be enumerated by direct cell counting. This has hindered attempts to understand the mechanism by which the bacteria assist in the leaching process. The methods reported in the literature to enumerate both the free and attached bacteria in a biooxidation system can be divided into 2 categories: direct methods and indirect methods. The direct methods involve the quantification of the bacteria by direct observation. It is difficult to enumerate attached bacteria by direct observation but attempts have been made to desorb or dislodge these bacteria. Such experiments have had limited success in achieving dislodgement of all the attached bacteria. However, the results have shown that desorption of the bacteria from the mineral surface is possible. Indirect methods involve the monitoring of a cell component such as protein, nitrogen and carbon. Biomass concentrations have been estimated using its metabolic activity by means of a maximum specific oxygen utilisation rate. The purpose of this study was to compare the various methods and test their suitability to the quantification of biomass in a biooxidation system. In particular the biooxidation system investigated treated an arsenopyrite-pyrite concentrate from Fairview Gold Mine, Barberton, South Africa. The elemental analysis of the concentrate is 5.84% arsenic, 21.71 % sulphur,24.01 % iron and 1.41 % carbon. The dominant bacteria present in the biooxidation system were Leptospirillum ferrooxidans and Thiobacillus thiooxidans as shown by 16S rDNA analysis. The methods investigated are microscopic counting, gravimetric dry weight determination, desorption, determination of chemical oxygen demand, ashing, protein analysis, nitrogen analysis, total organic carbon analysis and measurement of oxygen utilisation rate. The oxygen utilisation rate method differs from the other methods as it uses the metabolic activity of the bacteria to measure the bacterial concentrations.

Bibliography: pages 149-155. / This dissertation presents the results of an investigation into the disruption of microorganisms when agitated in slurries of fine particles in a stirred tank. The most widely used industrial process involving agitation of microorganisms in slurries of particles in stirred tanks is the biooxidation process. Mixed cultures of thiobacilli are used in stirred tank reactors for the biooxidation of sulphide minerals. In addition to operating conditions, the efficiency of biotechnological processes is dependent on the growth and metabolism of the microorganisms. The microorganisms are sensitive to the hydrodynamic conditions generated in the processes. In response to adverse hydrodynamic conditions there may be changes in the growth rate of the microorganisms, the nutrient uptake rate, the product formation rate and morphology of the microorganisms. Under extreme conditions cell damage and disruption may ensue. The presence of particulates in bioprocesses, in the form of solid substrates or support systems for attached growth, further complicate the hydrodynamic conditions. The knowledge of the effect of particulates on microorganisms is an important priority.

This thesis concerns the development of a method to quantify the morphology of the bacterium Bacillus thuringiensis, and to automatically count the bacteria. The need to quantify the bacterial morphology arose out of the possibility of controlling a fermentation based on the morphology of the observed bacteria. Automatic counting of bacteria was considered necessary to reduce the inaccuracies that resulted in manual counts performed by different people. Bacillus thuringiensis is a spore forming, gram-positive bacterium, which produces both intracellular spores and insecticidal protein crystals. The production of the insecticidal protein crystal makes Bacillus thuringiensis important as a producer of biological insecticides. Automatic counting was developed in a Thoma counting chamber (Webber Scientific) at 200x magnification under dark field illumination. It was found that at this magnification the problem of out of focus cells was eliminated. The use of a thick coverslip, which reduces variability in slide preparation, was also possible at 200xmagnification as the focal depth of the 20x objective lens was considerably larger than the 1 00x objective lens and thus the 20x objective lens could focus through the thick coverslip (20x objective lens with 1 Ox magnification in eyepiece = 200xmagnification). An automatic algorithm to acquire images was developed and 5images per sample were acquired. Processing of the images involved automatically thresholding and then counting the number of bright objects in the image. Processing was thus rapid and the processing of the five images took no more than a few seconds. Results showed that the correlation between the automatic and manual counts was good and that the use of a thick coverslip reduced variability in slide preparation. It was shown that the manual -counting procedure, which necessarily used a thin coverslip at 1000x magnification, underestimated the volume of the Thoma counting chamber. This was a result of warping in the thin coverslip.

Bibliography: pages 117-123. / Within the aquaculture industry a potential has been identified for the pigment astaxanthin. Astaxanthin is the carotenoid responsible for the distinctive coloration of salmonids, crustaceans and certain birds. Due to the fact that animals cannot synthesize carotenoids themselves, it is necessary for these pigments to be present in their food source. In the case of farm-raised salmonids and crustaceans, supplementation of their food with astaxanthin is required. The chemical synthesis of astaxanthin is very costly and complicated. As a result natural, microbial sources of astaxanthin are being investigated. Phaffia rhodozyma is the only yeast known to synthesize astaxanthin as its principle carotenoid. The aim of this dissertation is to present a study investigating the growth and pigmentation of P. rhodozyma, with a view to its commercial production. A P. rhodozyma mutant (UCT-1 N-3693) with a 50% increased total carotenoid content was selected after NTG mutagenesis of the wild strain.

The aim of this project is to establish and understand the production of glucose oxidase and gluconic acid using the Aspergillus niger bioprocess and predict its response to operating conditions. Glucose oxidase and gluconic acid are produced by a wide range of microbes and have a variety of applications. In this study Aspergillus niger was chosen as the microorganism as it has the "generally accepted as safe" (gras) status in the U.S.A. It is also the major industrial producer. Glucose oxidase catalyses the conversion of glucose, oxygen and water to hydrogen peroxide and gluconic acid. This enzyme is used as a glucose and oxygen scavenger in the food industry and as a diagnostic tool in medicine for glucose determination. Gluconic acid is an organic acid used as a sequestering agent with a broad spectrum of applications. The world market for gluconic acid and its various salts was 45 000 metric tonnes in 1985 (Bigelis cited by Markwell et al. 1989). Gluconic acid and its derivatives can be produced using three technologies: electrolysis, mild chemical oxidation and bioprocess. The first two technologies have not been proven to be comnercially viable. The bioprocess offers diversity of feed and produces other products such as glucose oxidase. Literature has shown that the production of gluconic acid involves two kinetic areas. Firstly, the glucose oxidase enzyme must be induced. Secondly, glucose is converted to gluconic acid by the enzyme glucose oxidase. The factors affecting the kinetics associated with the induction of glucose oxidase have only been described qualitatively. Glucose, oxygen and pH have been shown to affect the induction of glucose oxidase. The effect of pH has been studied by Roukas and Harvey (1989) who found that induction, is maximal at a pH of between 5 and 6. The effect of glucose and oxygen have not been quantified. The kinetics of glucose oxidase conversion of glucose to gluconic acid have been well described by Atkinson and Lester (1974).

[PAGE 1 missing] The aim of this study was to investigate and compare various methods to enumerate the number of bacteria in a minerals biooxidation system. In this system most of the bacteria are attached to fine particles of ore and therefore cannot be enumerated by direct cell counting. This has hindered attempts to understand the mechanism by which the bacteria assist in the leaching process. The methods reported in the literature to enumerate both the free and attached bacteria in a biooxidation system can be divided into 2 categories: direct methods and indirect methods. The direct methods involve the quantification of the bacteria by direct observation. It is difficult to enumerate attached bacteria by direct observation but attempts have been made to desorb or dislodge these bacteria. Such experiments have had limited success in achieving dislodgement of all the attached bacteria. However, the results have shown that desorption of the bacteria from the mineral surface is possible. Indirect methods involve the monitoring of a cell component such as protein, nitrogen and carbon. Biomass concentrations have been estimated using its metabolic activity by means of a maximum specific oxygen utilisation rate. The purpose of this study was to compare the various methods and test their suitability to the quantification of biomass in a biooxidation system. In particular the biooxidation system investigated treated an arsenopyrite-pyrite concentrate from Fairview Gold Mine, Barberton, South Africa. The elemental analysis of the concentrate is 5.84% arsenic, 21.71 % sulphur,24.01 % iron and 1.41 % carbon. The dominant bacteria present in the biooxidation system were Leptospirillum ferrooxidans and Thiobacillus thiooxidans as shown by 16S rDNA analysis. The methods investigated are microscopic counting, gravimetric dry weight determination, desorption, determination of chemical oxygen demand, ashing, protein analysis, nitrogen analysis, total organic carbon analysis and measurement of oxygen utilisation rate. The oxygen utilisation rate method differs from the other methods as it uses the metabolic activity of the bacteria to measure the bacterial concentrations.

This thesis concerns the topic of wastewater biorefineries (WWBR), in which wastewater is not seen simply as a waste stream to be cleaned but as a valuable material flow to be converted into bioproducts, while still meeting discharge limits at the end. To set the scene, similar developing approaches to valorise wastewaters globally are reviewed. Wastewaters in South Africa are reviewed and categorised with regards to their potential to serve as raw material, in terms of their volume, concentration and complexity. Bioproducts possible from wastewater is reviewed and evaluated. The wastewater biorefinery is conceptualised in the context of current wastewater treatment technologies and a set of evaluation criteria is developed. A multi-reactor setup is suggested in which wastewater is used, in series, as substrate by heterotrophic microbes like bacteria, photo-mixotrophic organisms like algae, macrophytes and fungi. Each reactor group is considered in detail and evaluated with regards to its suitability to the wastewater biorefinery, leading to selection of appropriate reactor designs. Stoichiometric mass balances of all unit operations are established, showing the material value flows, and combined to model this multi-bioreactor approach. Subsequently the model is tested against literature data. Finally, the applicability of the wastewater biorefinery concept for certain waste streams is assessed. The thesis contributes to the current body of knowledge in the following ways: 1. Introduction of the concept of the wastewater biorefinery (WWBR) 2. Provision of a potential preliminary guide for classification of wastewaters for use in the WWBR 3. Development of criteria for reactor evaluation for use in the WWBR 4. Development of an integrated model to interrogate bioproduction from wastewater and determine product yields associated with wastewater treatment 5. Creation of new knowledge through the interpretation of the model on different wastewater systems. The wastewater biorefinery is defined as a bioproduction system that integrates multiple unit operations to deliver compliant water as well as a bioproduct or bioproducts. It is approached through the concepts of industrial metabolism and the circular economy. Wastewater biorefineries are shown in this work to be a viable approach to improving resource efficiency while ensuring the better ecological functioning of humans within “greater than human” systems. The work places emphasis on the recovery of bioproducts that conserve molecular complexity but acknowledges that energy production for use on site and in the immediate surroundings is always an important factor in the WWBR. This thesis introduces the need to include a qualitative way to evaluate the complexity of wastewater, in addition to standard classification of volume and concentration of components. Complexity includes both composition of potentially problematic compounds and how unpredictably it changes over time. In this approach, it is preferable to generate three types of products: products of sufficient value to be economically viable; products of variable value with concomitant assimilation of major contaminants; and clean water as a product, typically through multiple unit operations, allowing multi-criteria optimisation. Through this approach, multiple criteria can be met. Function-based products specific to niche industries, particularly those which produced the wastewater of interest, are of substantive interest owing to their streamlined market uptake. This thesis explores the requirements of the products that can be produced from wastewater in a non-sterile context and suggests product groupings that meet these requirements. Products secreted into the bulk volume are difficult to recover, leading preference to biomass associated and intracellular products. The product needs to offer a selective advantage to the organisms producing it to facilitate enrichment through, ecological selection of the microbial consortium with simultaneous cell retention through reactor design and operation. Four groupings of unit operations were reviewed in detail and evaluated with regards to their suitability to the wastewater biorefinery, using a two-part set of evaluation criteria that was developed in this work, considering the reactor design, and its operation. The four unit operations each contribute a specific role to the functioning of the WWBR as a system. It is acknowledged that not all units are commercially important, and that the concept of diminishing returns should be kept in mind. The heterotrophic microbial bioreactor, of which the bacterial biocatalyst is used as a representative example, is helpful for removing a high proportion of the organic carbon. A wide range of commodity products with market potential is known to be produced through heterotrophic microbial systems. Existing heterotrophic microbial reactor systems like the aerobic granular sludge system (AGS) exist that suit the wastewater biorefinery approach particularly well, while activated sludge along with biological nutrient removal (BNR), the most commonly used reactor system in South Africa, is the least suitable to the WWBR. The photo-mixotrophic reactor represented by the algal bioreactor is helpful to scavenge high proportions of nutrients, particularly nitrogen and phosphorus. The algal bioreactor is also known to produce commodity products. Photo-mixotrophic bioreactor systems complement the heterotrophic systems but are unlikely to be the dominant reactor due to land and energy requirements. The macrophytic bioreactor is targeted for polishing the exiting stream in terms of nitrogen, phosphorus and particulates to ensure compliant, fit for purpose water as a product, with a macrophyte-based byproduct. Macrophyte bioreactors, particularly floating wetlands, are promising tertiary systems that should be viewed in conjunction with water sensitive design principles to overcome potential land availability limitations. The solids bioreactor is an emerging beneficiation technology for biotransformation of bio-slurries and the solid phases recovered during WWBR operation to generate products of value, including biosolids. Solids bioreactors have great potential but require more investigation, with key challenges being mass transfer and separation technologies. Operating waste treatment facilities as net income-producing bioprocesses require a mindset change about investment, risk and associated returns. WWBRs require higher capital investment due to the additional process units and downstream processing required and have higher operating costs due to the greater control required during the process and greater number of operators with advanced skillsets. An identification of the relevant product range and comparison between conventional processing routes and those possible from the wastewater is required on a case by case basis, and an overview is given in this thesis. Waste may need to be re-classified to be used as an intermediate by-product or raw material, requiring legal considerations in terms of both the solid waste as per the National Environmental Management Act (NEMA) and liquid waste as per the National Water Act (NWA). The added complexity of reclassifying waste as raw material needs an acknowledgement of institutional challenges such as speaking across department silo’s. In this thesis, a model of these integrated unit operations was developed to generate material inventories across the system. This can be used to evaluate possible scenarios in an integrated context using a generic flowsheet as well as mass balances generated through the model. Three case studies were examined: municipal, abattoir and pulp and paper wastewater. Municipal wastewater was chosenas it represents a complex, dilute, 'suboptimal’ wastewater stream. Abattoir wastewater was chosen as an example of a complex, nutrient-concentrated stream that may be well suited to biological transformation. Pulp and paper wastewater was chosen as an example where the biorefinery concept is already well established, and is a low complexity, low nutrient, high carbon content stream. In considering the above case studies, a number of key learnings resulted. The impact of solids removal was clear and in keeping with existing bioprocessing and wastewater treatment principles of decoupling the hydraulic and solids residence times. Low nitrogen and phosphorus content in the pulp and paper wastewater as compared to the other two case studies indicated the need to conduct integrative studies of the unit operations to determine the most appropriate unit operations across the system. The effect of improving the product conversion yields and product recovery yields were examined, and a surprising result is the amount of nutrients that remain in compliant effluent, due to the large volumes of liquid involved. This leads to the conclusion that while the WWBR is a valuable way to address resource recovery, separation at source and internal process efficiencies are critical to improve overall resource efficiency and environmental protection. With regards to municipal wastewater, which contributes by far the most in terms of volume and nutrients of wastewaters in South Africa from the perspective of reactor design for waste(water) beneficiation, considering the cleaner production principle of separation at source, along with the need to decouple the solid and hydraulic residence times, dry sanitation presents a clear argument for the best WWBR approach. This approach must acknowledge that the transport of the sanitation raw materials is more difficult if hydro-transportation is not available, and needs to ensure operator equity, health and safety, particularly in the handling of the sanitation raw materials. This thesis was developed in conjunction with the Water Research Commission (WRC) project “Introducing the wastewater biorefinery concept: A scoping study of polyglutamic acid production from a Bacillus-rich mixed culture using municipal waste water” (Verster, et al., 2014) and Water Research Commission (WRC) K5/2380 project titled “Towards Wastewater Biorefineries: integrated bioreactor and process design for combined water treatment and resource productivity” (Harrison, et al., 2017). While the project focused on a global and national review on research on wastewater biorefineries and wastewater as a resource, this thesis explores in greater depth the requirements of each of the reactor units and their integration.