Utilisation of biomass is identified as one of the promising solutions to reduce society’s dependence on fossil fuels and mitigate climate change caused by the exploitation of fossil fuels. By using the concept of biorefinery, biomass can be converted into value-added products such as biofuels, biochemical products and biomaterials in a greener and sustainable way. To enhance the efficiency of biorefinery, the concept of integrated biorefinery which focuses on the integration of various biomass conversion technologies is utilised. To date, various biomass conversion pathways are available to convert biomass into a wide range of products. Due to the substantial amount of potential products and conversion technologies, determining of chemical products and processing routes in an integrated biorefinery have become more challenging. Hence, there is a need for a methodology capable of evaluating the integrated process in order to identify the optimal products as well as the optimal conversion pathways that produce the identified products. This thesis presents a novel approach which integrates process with product design techniques for integrated biorefineries. In the proposed approach, integration between synthesis of integrated biorefinery and computer-aided molecular design (CAMD) techniques is presented. By using CAMD techniques, optimal chemical product in terms of target properties which fulfils the required product needs is designed. On the other hand, in order to identify the conversion pathways that produce the identified optimal chemical product in an integrated biorefinery, chemical reaction pathway map (CRPM) and superstructural mathematical optimisation approach have been utilised. Furthermore, this thesis also presents various chemical product design approaches. In order to solve chemical design problems where multiple product needs are required to be considered and optimised, a novel multi-objective optimisation approach for chemical product design has been presented. By using fuzzy optimisation approach, the developed multi-objective optimisation approach identifies optimal chemical product based on multiple product properties. In addition, fuzzy optimisation approach has been further extended to address chemical product design problems where the accuracy of property prediction model is taken into account. A robust chemical product design approach is developed to design optimal chemical products with consideration of accuracy of property prediction model. Furthermore, together with CAMD techniques and superstructural mathematical optimisation approach, the developed multi-objective optimisation approach has been utilised for the design of mixtures in an integrated biorefinery. For this purpose, a systematic optimisation approach has been developed to identify optimal mixture based on multiple desired product needs as well as the optimal conversion pathways that convert biomass into the optimal mixture. Finally, possible extensions and future opportunities for the realm of the research work have been highlighted in the later part of this thesis.

Reformulation of blended particulate materials has been always a problem for powder industry because formulators have difficulty in measuring, controlling and/or modifying the bulk flow properties of powders. A recently developed powder flow tester (PFT) by The Wolfson Centre for Bulk Solids Handling Technology, University of Greenwich is now available for industry to measure quickly and accurately the flow behaviour of single and blended particulate materials. This new powder flow tester helps to characterise quantitatively the blends which show good flow behaviour in the industrial process lines and define the desirable standards for blends with poor flow behaviour. However, the current process of reformulation is basically a trial and error procedure based on the prior experience of the formulator with other blends reformulated. There is clearly a lack of practical understanding of the links between the particle and bulk scale of powders and how changes in the particle properties and/or blend compositions would affect the flow behaviour of blended powders. The aim of this research work was to develop an “empirical understanding” of the links between the particle properties and bulk flow properties in order to predict the bulk flow properties of blended powders based on changes to the particle properties or blend components. This has been achieved evaluating analytical well-established models found in the literature linking particle and bulk scale for practical purposes in industry, developing new empirical models to predict the flow behaviour of single and blended powders based on the experimental work undertaken in this research and identifying methods or techniques that formulators in industry could use to predict the flow behaviour of their blended powders. The solution provided is a test protocol which in combination with the standard characterisation tests commonly used in industry and the prediction tool called Virtual Powder Blending Laboratory (recently developed by The Wolfson Centre for Bulk Solids Handling Technology using the experimental and modelling work undertaken in this work) will help formulators and process engineers to formulate the composition of blended powders with the desirable flow behaviour in industrial process lines.

This project uses experimental techniques to explore the effects of surface active agent (surfactant) adsorption on a droplet of oil suspended in a flow. The fluorescent surfactant Rhodamine-6G was used to enable the use of optical techniques to visualise the build-up of surfactant at the rear of the droplet and its effects on the internal circulation within the droplet. This was done to enable an exploration of how surfactant accumulates behind a droplet with the aim of predicting the behaviour based on the internal circulation for non- fluorescentt surfactants. An experiment was designed along with calibration procedures in order to utilise the non-intrusive measurement techniques laser-induced fluorescence and particle image velocimetry to measure the volume of surfactant held behind the droplet, the angle of the visible cap caused by the build-up of Rhodamine-6G, and the cap angle of the stagnant region obtained by measuring the velocity of the circulation within the droplet. Laser-induced fluorescence (LIF) was used to visualise the development of a surfactant-rich cap that formed at the rear of the oil droplet as it rose through an aqueous solution, with different bulk concentrations of surfactant. The aqueous solution soluble surfactant, adsorbed to the forwards facing part of the droplet, was transported around to the rear where it accumulated in a surfactant cap before being swept back into the ow behind the droplet. The fluorescent properties of Rhodamine-6G were utilised to measure the size of the visible surfactant cap angle and the volume of surfactant stored within the cap. The results showed that increased concentrations of surfactant caused a larger volume of surfactant to be held behind the droplet with larger droplets resulting in smaller surfactant cap angles. Particle image velocimetry (PIV) was used to explore the effects of surfactant at the interface on the internal circulation within the droplet. As surfactant accumulated an area of very low velocity at the rear of the droplet appeared. This was quantified by measuring the tangential velocity around the droplet with the area of low velocity signifying the presence of a stagnant cap. Measuring the internal velocity around the droplet close to the interface showed that the stagnant cap angle had large growth over the initial region of the tank, with larger droplets resulting in smaller surfactant cap angles. The visible cap angle measured by LIF was related to the stagnant cap angle measured by PIV to compare how the results for each technique developed over the height of the tank. The angles for both techniques exhibited the same trends as the droplet height increased, although they showed different magnitudes revealing that the experimental procedure could be improved in the future to obtain better agreement.

This study is in the area of sustainable chemical technology and is ultimately aimed at developing a novel generic approach towards efficient, clean, safe and scalable synthesis of important intermediates for pharmaceutical applications and catalysis using nonthermal activation of chemical bonds. Current study is focused on molecular activation by light, specifically addressing the class of singlet oxygen reactions. A laboratory-scale annular recirculating photoreactor and a scalable microreactor unit were designed, assembled and tested in oxygenation of α-pinene and stereoselective oxygenation of homoallylic alcohols by the photogenerated singlet oxygen. Efficiencies of light utilization and the quantum yields of the reactions for different sources of light in various lamp-reactor geometries were quantified using actinometry. It was shown that optimal thickness of the reaction layer, high intensity of light and elevated oxygen pressures in continuous gas-liquid segmented flow are crucial for both safe and efficient oxygenation. This study proved the viability of a novel concept of compact reactors with embedded light emitting diodes. Finally, a methodology of efficient oxyfunctionalization of allylic compounds was developed. Following the developed methodology, the applicability of functional nanomaterials based on pure and dye-modified porous silicon as heterogeneous photosensitizers of singlet oxygen was studied in detail. It was demonstrated that the photosensitizing efficiency of all the samples is much lower than that of conventional dye photosensitizers. Low activity of the novel photosensitizers was attributed to quenching of the photogenerated singlet oxygen by the surface quenching groups of porous silicon, low quantum yield of photoluminescence in the case of pure porous silicon and to thermodynamically favorable quenching of the photoexcited porphyrin states due to energy/electron transfer to silicon nanocrystals in the case of dye modified porous silicon.

Lignin is the only plant biomass that contains aromatic groups in its structure and can provide a wide range of low molecular weight aromatic chemicals if its depolymerisation can be achieved successfully. Currently, lignin is mainly produced as a waste by-product by the paper and pulp industry and biorefineries. Therefore, the transformation of the phenolic-rich lignin into value added aromatic platform chemicals can be regarded of primary concern to improve the economic profitability of biorefining. Moreover, being a renewable resource, the consumption of fossil fuels will be reduced if lignin can be utilised efficiently. Between chemical degradation and enzymatic degradation, the latter could be a more sustainable method to break down lignin due to its enhanced substrate specificity and ability to preserve the aromatic ring structure compared with chemical processing. Therefore, laccase from Trametes versicolor (LTV), lignin peroxidase (LiP) and manganese peroxidase (MnP) were studied to determine the scope to depolymerise both water-soluble and insoluble lignins nder mild reaction conditions. The enzymatic activity and stability of all three enzymes was investigated and optimum assay conditions were achieved. LTV was found to be the most stable enzyme as it maintained 55 % of its activity at least for the first 6 h at 30 °C whereas LiP was deactivated after 2 h at 25 °C, and MnP was deactivated after 1 h at 28 °C. However, LTV stability decreased at higher temperatures during the oxidation of 2,2’-azino-bis (3- ethylbenthiazoline-6-sulphonic acid (ABTS)). One of the non-phenolic lignin model compounds, veratryl alcohol, was oxidised by LTV in the presence of ABTS, thus confirming the published data. The enzymatic degradation of Organosolv lignin (OSL) by LTV resulted in the formation of 2,6-dimethoxy-1,4-benzoquinone (DBQ). The OSL degradation by LTV was not improved by ethanol addition as a co-solvent although ethanol could stabilise LTV at 40 % (v/v). LTV catalysed the degradation of Kraft lignin although it indicated little effect on lignosulphonates. Lastly, the effect of varying the concentrations of 92 ionic liquids (ILs) and their equivalent metal salts on LTV activity was investigated to find a suitable co-solvent to improve the poor mass transfer in OSL degradation. The study showed that 62 ILs were laccase compatible at an IL concentration of 6 % (w/v) and more than 50 % laccase activity was retained in 18 ionic liquids up to 10 % (w/v), and 80 % (v/v) of dioctyl sulfosuccinate quaternary ammonium salt, [N4,4,4,4][AOT]. However, there was a progressive loss of activity when the concentrations of the ILs increased. Further study on the enzymatic degradation of ILs-pre-treated OSL is currently ongoing in our research group so that the decomposition of water-insoluble lignin will be understood more comprehensively.

The occurrence of perhydrous (oil-prone) coal deposits within the Firkanten Formation of the Central Tertiary Basin (CTB) in Spitsbergen, is well documented and the oil present is reportedly sourced from the coals (Mokogwu, 2011; Marshall et al., 2015a). This study uses a total of 146 coal samples covering areas of the eastern coalfield (Bassen, Breinosa and Lunckefjellet) and the western coalfield (Colesdalen) of the CTB to investigate the maturity, oil source rock potential, and retorting potential of these perhydrous coals. In addition, the controls on the oil potential of the coals are considered to provide measures that could be used to determine the optimum resource areas in the basin. Samples were provided by Store Norske Spitsbergen Kulkompani AS, and include drill cores, mine sections and outcrop sections from the Svea, Longyear, Svarteper and Askeladden seams (eastern coalfield), and the Sputnik and Verkny seams (western coalfield). The vitrinite reflectance (VR) of the investigated coals are suppressed by bitumen impregnation and hydrogen enrichment of vitrinites; this is indicated by a general decrease in VR towards the top of the Longyear seam, which correlates with increasing Soxhlet yields towards the top of the seam, and a strong negative correlation of VR vs HI (Hydrogen Index) (R square between 0.73 - 0.78 in all areas), which is not maturity induced. Other evidences of VR suppression include a relatively wide range of VR values (between 0.50 - 0.79% Ro) within these seams, which are notably ≤2.1 m thick in all areas investigated, and wide ranges of VR distribution with bi-modal histograms observed in most samples (due to maceral effects rather than mixing of coal seams). Additionally, aliphatic biomarker and aromatic maturity parameters do not decrease towards the top of the Longyear seam (contrary to VR which decreases towards seam top), and indicates that the coals are generally in the maturity range of around 0.70% Ro or higher. Tmax appear suppressed, and the re-arrangements of methylphenanthrene isomers with increasing maturity also appear delayed/suppressed when there is aliphatic enrichment. True (i.e. unsuppressed) VR was estimated using the Lo (1993) method which gives thermal maturities of around 0.68, 0.78, 0.80 and 0.88% Ro in the Bassen, Lunckefjellet, Breinosa and Colesdalen areas respectively. True VR values indicate peak temperatures of around 104 °C in Bassen, 116 °C in Lunckefjellet, 118 °C in Breinosa and 125 °C in Colesdalen. Coalification gradients in the Adventdalen area equate to around 0.37% Ro/km, with an estimated geothermal gradient of approximately 55 °C/km. Peak burial depths in the Adventdalen area range from 1.9 km (up-dip), to 2.2 km (down-dip), indicating an overall overburden erosion estimate of between 0.9 – 1.2 km. In the Lunckefjellet area, peak burial depth is around 2.1 km, which implies a missing overburden of 1.1 km. In Colesdalen, peak burial depths are considerably higher at around 2.3 km, with a missing overburden estimate of 1.4 km. The implications of these results on burial and subsequent uplift and erosion are discussed. The oil potential of the studied coals appears to be mainly due to perhydrous detrovitrinites, although other vitrinites including collotelinite, in addition to some liptinites, may have significantly contributed. Rock-Eval analysis indicates that the coals are enriched in Type II and a mixture of Types II/III kerogens with high HI (150 - 410 mg HC/g TOC), variable TOC contents (44.5 – 89.8 %), and high S2 contents (109 – 368 mg/g), all of which indicate excellent oil potential. The mean S1 contents are 6.8, 11.8, 15.0 and 14.5 mg/g for the Bassen, Lunckefjellet, Breinosa and Colesdalen coals respectively, which reflect the maturity trend across these four localities (i.e. increasing maturity from Bassen, through Lunckefjellet, to Breinosa and Colesdalen). The Bassen coals are at the onset of oil generation, while the Lunckefjellet coals are at peak oil generation/onset of oil expulsion. The Breinosa and Colesdalen coals however, are already expelling oil (meaning they are in the effective oil window), although all samples (i.e. from all four localities) have low production index (PI < 0.10), which suggest that significant expulsion have not occurred. Results indicate that the Lunckefjellet coals will give the best indication of the maturity at which oil expulsion occurs in the CTB. The oil-proneness of the coals resulted from marine influence upon the peatlands, and the consequent marine sulphur enrichment (between 0.4 – 17.7 % S in the coals). The Askeladden, Svarteper and Verkny coals generally contain more sulphur than the Longyear and Sputnik coals, and this trend is consistent with that of oil potential. Between sample localities, oil potential generally increases in the direction towards the inferred palaeocoastline. The varying sulphur contents within seams and between localities are assessed, and the implications of this variation on oil potential are examined and discussed. In addition to the influence of sulphur on oil potential, there are other marine as well as non-marine related factors on oil potential which have been examined and discussed to help in delineating the optimum resource areas in the basin. Results indicate that greatest oil potential is mainly due to the combination of the following factors: a) Thermal maturity (true VR of around 0.78% Ro) b) Relative sea level rise leading to S contents in excess of 0.5 % c) Stable hydrology (i.e. relatively large/stable groundwater catchment) d) Fe/S ratio significantly <0.87 e) Optimum pH levels (alkalinity, which favours high microbial degredation) f) Relative distance to inferred palaeocoastline and local topography g) Ash content ≤30 % Retorting of the CTB coals showed highest bulk yields at Lunckefjellet (160 mg/g on dry whole coal – dwc basis), with sections within seams yielding up to 240 mg/g dwc. At Breinosa and Colesdalen, bulk yields of 140 and 100 mg/g dwc respectively were measured. Lowest bulk yield was measured at Bassen (80 mg/g dwc). Residual semi-coke ranges between 60 – 75 % of starting material in all areas. Retorting yields are notably limited by coal swelling/blocking of the reactor vessel; consequently, further work involving other methods such as the Grey-King assay are required to fully measure the retorting potential in these coals. With a maximum coal resource of 3,300 Mt, of which 600 Mt is recoverable in the CTB (Orheim, 1982), maximum hydrocarbon resource via retorting would range between 3,188 – 5,394 Mbbl in place, with 580 – 981 Mbbl recoverable by mining.

The use of biomass derived solid fuels for electricity generation in combustion, gasification and pyrolysis plant has received increasing levels of interest for commercial operation in recent years. However, there are limited tools available which allow a prediction of the performance of these fuels during thermochemical transformation given an understanding of their original chemical structure. As such, this investigation has concentrated on the derivation of a simply utilised classification system able to predict a series of important fuel combustion characteristics given an understanding of both the organic and inorganic chemical and structural composition of any lignocellulosic biomass fuel. A prediction of volatile matter content and char yields during pyrolysis has been made using correlation with aromatic carbon, potassium and calcium contents using both thermogravimetric slow heating and simulated pulverised fuel (PF) entrained flow rapid heating. Alongside this, investigation of the impact of biomass composition, namely aromaticity and alkali/alkaline earth metal concentrations, on char structure and oxidative char reactivity of simulated PF chars has been conducted. Experimental investigation has involved the pre-treatment of a wide range of commercially available biomass fuels including softwood, hardwood, herbaceous and agricultural waste materials to remove both lignin and ion exchangeable mineral species. In addition to this, torrefaction has been utilised to increase the aromatic character of chosen fuels. This has allowed a quantification of the impact of aromaticity and mineral matter concentration on pyrolysis and char combustion reactions to be derived for a wide range of fuel aromaticity and mineral matter contents. Considerable success has been achieved in the classification of an array of lignocellulosic biomass. Accurate prediction of pyrolysis char and volatile matter yields under both slow and rapid entrained flow drop tube heating conditions have been attained using simple empirical correlations with fuel aromatic carbon and alkali/alkaline earth mineral species concentrations (K+Ca being utilised here). This classification system has relied upon the clear linear correlation observed between aromatic carbon content and char yield in the absence of mineral matter influences (R2 of 0.98 and 0.95 being observed for demineralised biomass under slow and rapid heating pyrolysis respectively). In addition to this, the relative enhancement of char yield due to mineral matter interaction with varying concentration of K and Ca within the fuel has been quantified and is used to calculate total char yields. The empirical relationship derived under slow heating takes the following form: Slow Heating Char Yield=(1×Aromatic Carbon )+(16.1 ×(K+Ca) ) Where slow heating char yield is the char yield wt% on a dry ash free basis (daf), aromatic carbon is the wt% daf aromatic carbon content of the biomass and K+Ca is the wt% K+Ca content of the raw fuel on a dry basis. This relationship applies below K+Ca contents of 0.6 wt% db, beyond this a fixed additional char yield of 9.76 wt% daf can be applied as a quantification of the influence of enhanced char yield due to mineral activity as the second term in the above equation. For rapid heating entrained flow pyrolysis the empirical prediction of char yield is conducted as follows: Rapid Heating Char Yield=(0.58×Aromatic Carbon )+(2.43 ×(K+Ca) ) Strong linear correlations of predicted vs. observed char yield have been derived with correlation coefficient R2 = 0.96 and 0.99 with mean relative errors of 7.8 and 8.4% for slow and rapid heating pyrolysis respectively. Furthermore, the influence of biomass aromaticity and active mineral content on char formation processes, the form of chars generated under PF like devolatilisation conditions and their subsequent oxidation reactivity has been studied in detail. Both alkali/alkaline earth mineral matter content (primarily K and Ca) and aromaticity are instrumental in determining the porosity, morphology and surface area of simulated PF chars. Due to its tendency to soften during heat treatment lignin is shown to produce low surface area, non-porous chars under slow heating and this behaviour drives a reduction in char surface area and combustion reactivity with increasing aromatic carbon content. Although char surface areas have been seen to be negatively correlated with increasing potassium and calcium content this may be due to ash blockage of char pore structures. However, the likelihood of a negative impact of mineral enhanced charring has been discussed. K catalysis of combustion reactions is clearly evident in apparent and inherent char reactivities; however, easy quantitative assessment of this influence has been prevented by the clear complexity of mineral behaviour during the pyrolysis process. The development of char structure and reactivity as a function of char combustion degree has also been investigated under entrained flow combustion conditions. The results of this study indicate that by accurately quantifying aromatic carbon, potassium and calcium contents, all lignocellulosic fuels can be classified in terms of their behaviour during pyrolysis (volatile matter and char yields), the form of char structures generated (surface area and porosity) and char combustion reactivity. It is hoped that this relative classification will shed light on the predicted performance of biomass fuels for use in combustion driven power generation infrastructure, especially in pulverised fuel applications.

Biomass combustion has increasingly been used in pulverised fuel coal fired power stations as a way of addressing a wide range of emissions reduction targets. The reuse of existing equipment such as coal mills is essential to minimise the costs of conversion. However the fundamental fracture mechanics involved in biomass comminution are completely different to coal. Thus a thorough knowledge of the comminution properties of all biomass types in coal and biomass mills is necessary in order to minimise operational issues and to optimise milling and combustion. This thesis provides extensive novel characterisation on densified biomass before and after milling. The study analysed 9 densified biomasses, 2 non-densified biomasses, and a sample coal in five different mills; planetary ball mill, Hardgrove Grindability Index testing mill, Bond Work Index ball mill, cutting mill, and a ring-roller mill. Milling was found to have little impact on particle shape, even when an order of magnitude change in particle size was observed. Particle shape is inherent to the particles which comprise a pellet, and is determined by the pre-densified comminution processes. Milling had little impact on compositional particles of herbaceous or wood pellets. Olive cake had the most spherical of all the materials. Thermal pre-treatments of woody biomass not only saw a significant improvement in grindability in all mills, but also enhanced shape factors. The Hardgrove Grindability Index is a poor indicator of the grindability of biomass. The Bond Work Index can be used to analyse the choking potential of biomass pellets prior to full scale mill trials. To optimise milling in coal mills, biomass pellets should be composed of particles close to the required size so that only the pellet comminution stage occurs. The milling behaviour of densified biomass in a laboratory scale ring-roller mill with dynamic classification was investigated for the first time. The milling studies showed that knowledge of a materials critical particle size for comminution through compression is essential to understand its milling behaviour in different mills. The results presented in this thesis not only provide new insight and addresses significant gaps in knowledge, they also provide useful and practical guidance for addressing operational issues such as mill choking, as well as ways to optimise biomass comminution in laboratory and full scale mills, such as mill classifier optimisation based on real particle characteristics.

Colourimetric indicators were prepared using anthocyanins, poly(vinyl alcohol) and cellulose nano-fibre by the electro-spinning technique. The indicator label containing 2000 ppm anthocyanin was found to be the optimal dye concentration. Fourier-transform infrared spectroscopy and X-ray diffraction results confirmed that the anthocyanin was incorporated into the poly(vinyl alcohol)-cellulose nano-fibre. The anthocyanin solution displayed pH dependent colour changes with solutions below pH 4 appearing rose red; at pH 5 and beyond it turns pink at pH 6 and eventually turns violet at pH 8. Blue-green was observed at pH 10. At highly alkaline pH 14 it turns yellow. Similar results were obtained when the anthocyanin containing indicator label was exposed to solutions of pH’s in the range 1 – 10 with the indicator label colour changing from red to green respectively. The indicator label was subjected to storage stability tests over 40 days by storing under nitrogen atmosphere, refrigeration conditions and at room temperature. The indicator label exhibited a high degree of colour stability with colour change (ΔE<4) under all conditions. The indicator label also showed significant colour changes from blue to violet to purple and ultimately pink when it was placed within the package of fresh cut green pepper for 7 days ambient conditions. These colour changes are a direct consequence of the increased carbon dioxide concentrations in the package resulting from deterioration of pepper at room temperature. Colour changes resulting from accumulating Total volatile basic nitrogen (TVBN) were also observed when the indicator label was placed in the headspace of packaged fresh chicken with the label changing from light pink to grey from day 1 to day 7 at room temperature storage. The results show that the label has great potential to be used as a visual in-package indicator of poultry products. Poultry and green peppers are food products with a potential for in-package quality detection by this indicator label. The results of the current work offer a new possibility to the applicability of water soluble, biodegradable indicators as real time quality detectors of bell pepper/chicken meat; it can be used alongside the best before date. Further work ought to be conducted to evaluate the effectiveness of this in-package freshness indicator label for fresh-cut pepper/chicken meat at different temperature conditions. Application of this freshness indicator for other fresh-cut fruit and vegetables also need to be further investigated Keywords: Colourimetric indicators, anthocyanins, storage stability, TVBN, poultry, carbon dioxide, green pepper, in-package / Dissertation (MSc (Applied Sciences Chemical Technology))--University of Pretoria, 2021. / CSIR / Chemical Engineering / MSc (Applied Sciences Chemical Technology) / Restricted

The quantification of atmospheric emissions from clamp kilns in the clay brick industry has met with limited success globally. The complex configuration of clamp kilns using coal or other carbonaceous fuels, as well as the uncertainty regarding kiln combustion conditions, has proven to be a hurdle in measurement of emissions and standardization of clamp kiln conditions. To enable measurement and quantification of emission and energy metrics, a model kiln was designed to simulate operating conditions and configuration similar to a transverse slice of a typical full-scale clamp kiln, but with a lower capacity (20 000 – 35 000 bricks per firing cycle). The model kiln design ensures the adequate confinement and extraction of flue gases with the aid of a bifurcated fan forcing the draft through a horizontal extraction stack where monitoring occurs. The model kiln design, which comprise two adjacent sealed sides and two partially enclosing and sliding galvanized steel doors, provides adequate spacing for ‘packing’ and ‘un-packing’ of bricks and sufficient oxygen for combustion, while still ensuring minimum losses of emission via the semi-enclosed sides. Concurrent firing and hourly monitoring of flue gases in the flue duct was conducted for fourteen batches of bricks over 8 – 14 days using varying brick products and energy inputs from eleven South African brick factories that utilizes clamp kiln as firing technology. The model kiln was tested for its suitability in firing bricks that are similar to conventional South African clamp kilns, as well as its effectiveness in the capturing and channelling of flue gases through to the stack vent where monitoring of the flue gases took place. Hourly readings are recorded for process parameters, SO2, NOx, NO, NO2, CO and particulate matter (PM) concentrations in the extraction stack. PM size-segregated mass measurement was conducted to produce PM1, PM2.5, PM4, PM10, and PM15 fractions. SO2 monitoring results were also compared to mass balance calculations, using the analysis of sulfur in the coal to indicate that the model kiln design is effective in capturing emissions and standardizing emission factors, as well as providing an effective energy analysis tool for clamp kilns. A statistical mean efficiency for the model kiln emissions capturing and channelling capacity was calculated from sulfur mass balance results of the batches that lie within 95% confidence interval of the assumed true mean (100%) to give 84.2%. Therefore, 15.8% of emissions were considered to escape from underneath the semi-enclosed sides. Final emission factors (mean ± standard deviation) were quantified as 22.5 ± 18.8 g/brick for CO, 0.14 ± 0.1 g/brick for NO, 0.0 g/brick for NO2, 0.14 ± 0.1 g/brick for NOx, 1.07 ± 0.7 g/brick for SO2, 378 ± 223 g/brick for CO2, 0.96 ± 0.5 g/brick for PM10; as well as 1.53 g/brick for hydrocarbons (calibrated to propane emissions) and 0.96 g/brick for PM15, PM4, PM2.5 and PM1. Various kiln technologies were ranked from lowest to highest potential for atmospheric pollution based on available emission metrics as follows: Zig-zag < Vertical shaft < South African Clamps < US coal-fired < Fixed chimney Bull’s trench < Tunnel < Asia Clamps < Down draft < Bull’s trench. Energy analyses indicate that a significant reduction of 0.9 MJ/kg (36%) in energy use could be achieved by the South African clamp kiln industry, thereby reducing cost of input, and significantly reducing the quantity of atmospheric emissions. In addition, chemical reactions and thermodynamic processes occurring in the firing chamber of brick kilns were qualitatively linked to the amount and type of pollutant emissions released at different periods during a firing cycle. The sensitivity of brick kiln emission concentrations and process metrics to these reactions and processes was utilized to proffer emission control measures. These measures are aimed at reducing energy consumption; improving the clay material processing and drying technique; monitoring chemical constituents of input materials in order to eliminate less favourable options; monitoring firing temperature to modify firing process; as well as altering the combustion and firing process in order to favour chemical and thermodynamic processes that will result in the release of lower emissions. Screening dispersion modelling results was additionally employed in recommending the extent of impact zones from the clamp kiln area for small kilns (500 m), medium kilns (1000 m) and large kilns (2000 m). A general reduction in most pollutant emissions was observed when the external fuel (coal) was replaced with a locally available alternative, propane gas. CO, CO2, NOx/NO and PM10 indicated 87%, 7%, 41% and 10% reduction in emissions respectively, during propane gas firing. SO2 emission, however, indicated a 19% increase, which may be attributed to lower energy consumption that alters the complex thermodynamic reactions in the model kiln. Only CO and NOx/NO emissions provided significant reduction in emission rates to support the notion that substituting the external coal with propane gas will result in significant reduction in atmospheric emissions. PM10 and CO2 emission rate do not provide significant reduction to validate this notion, while SO2 emission rate analysis is inconclusive and may require further research. / Thesis (PhD)--University of Pretoria, 2018. / Clay Brick Association of South Africa / Energy Efficient Clay Bricks / National Research Foundation / Chemical Engineering / PhD / Unrestricted