Browsing by Author "Kotsiopoulos, Athanasios"

Now showing 1 - 7 of 7
  • No Thumbnail Available
    Item
    Open Access
    Cytochrome P450 whole cell biohydroxylation of alkanes
    (2024) Seeger, Danielle; Harrison, Susan; Kotsiopoulos, Athanasios
    Production of higher-value fine chemicals via bio-based catalytic alkane activation are becoming increasingly relevant in the context of sustainability. Linear alkanes, sourced from fossil fuel reserves and as by-products in gas-to-liquid technology and CO2-to-power strategies, both through Fischer Tropsch activities, can serve not only as a low-cost fuel or solvent, but also as a precursor to produce valuable long chain alcohols. Effective use, or resource efficiency, with respect to alkanes is important in sustainable processing. Alkanes contain unreactive hydrocarbon bonds and cracking processes apply high temperatures and pressures to split these inert molecules into smaller and reactive molecules. Although widely applied, cracking suffers from poor selectivity and is resource heavy and environmentally burdening. Biocatalysis has shown potential for highly regio-selective alkane oxidation under mild conditions. Enzyme catalytic mechanisms evolved to catabolise fatty molecules in an organism’s metabolic pathway. Biocatalysis harnesses these evolved pathways to produce value added chemicals. Cytochrome P450 (CYP) monooxygenases such as CYP153A6 have previously shown excellent 95% selectivity for hydroxylation at the terminal carbon on n-octane and CYP153A13 has shown promising activity on n-octane and n-decane. In the present study, whole cell biocatalysts were prepared by expressing CYPs in E. coli BL21(DE3) and suspended in aqueous media required to maintain whole cell health. Alkanes applied were immiscible with the aqueous phase, resulting in a ‘2-Liquid Phase System’ (2LPS). Alkane accessibility to the enzyme in the whole cells significantly constrains scalability of biocatalysis due to transfer across the aqueous-organic interface and the cellular membrane to reach CYPs. The extent to which alkanes are transferred is determined by variables such as solubility, thermodynamic equilibrium, cell membrane permeability, temperature etc. The kinetics of the biocatalytic reaction are partly dependent on physicochemical variables, but also on biological parameters such as cellular health, enzyme expression etc. It was shown that although CYP153A13 showed superior coupling efficiencies compared to multicomponent enzyme systems, fusion proteins suffer from low expression levels and cytoplasmic proteolytic truncation, lowering its concentration and activity compared to CYP153A6. High Cell Density Cultures (HCDCs) were used to circumvent CYP153A13’s low expression. HCDCs of both CYP153A6 and CYP153A13 resulted in larger 1-octanol volumetric production rates by 10- and four-fold respectively compared to the LCDC counterparts. The turnover number of CYP153A6 and -A13 LCDCs, however, were respectively 2- and 7- fold higher than their HCDC counterparts, illustrating the oxygen and nutrient limitation of HCDCs. The multicomponent CYP153A6 system was expressed with high functionality and showed descending activity with: n-octane > n-decane > n-nonane, caused by a combination of alkane solubility and transmembrane transport. Although n-decane is less soluble in water than n-nonane, the ‘odd-even’ effect of alkanes resulted in high in vivo activity on n-decane. UNIF-LL predictions illustrated that BEHP reduces the amount of alkane in the aqueous phase than in the absence of BEHP. Therefore, it was illustrated that BEHP regulates the amount of substrate in the aqueous phase, which decreases the toxicity of the alkane substrates towards the whole cells. Therefore, while substrate solubility plays a role in availability to the whole cell biocatalyst which would impact the rate of biocatalytic reaction, cellular health maintenance is also important. This proved that factors related to aqueous substrate solubility play a larger role in the outcomes of 2LPS whole cell biocatalytic reaction kinetics as well as factors such as cellular health. The study showed that although substrate solubility plays an important role in biocatalytic performance, substrate solubility could not be used as the sole indicator of predicting biocatalytic performance and factors such as enzyme expression, membrane transport, enzyme-substrate affinity and physical substrate properties should be considered. To investigate the enzyme-substrate affinity, apparent Michaelis Menten (MM) constants (𝐾𝑚 𝑐 ) of each enzyme-substrate pair were determined from experimental data, of product concentration, accumulated over time using least squares nonlinear regression. 𝐾𝑚 𝑐 constants, however, represented both enzyme-substrate (ES) and substrate-solvent interactions. To determine the true affinity of each ES pair without the interference of substrate-solvent interactions on ES interactions, thermodynamic activity-based constants (𝐾𝑚 𝑎 ) were determined. This approach paves the way for the future development of ‘one-pot’ biocatalytic systems containing multiple different enzymes and substrates. A process flowsheet using the UNIF-LL model was developed and a theoretical calculation were used to determine activity coefficients, which were in turn used to determine the 𝐾𝑚 𝑎 constants. The 𝐾𝑚 𝑐 constants of CYP153A6 illustrated a 17-fold and two-fold larger affinity for n decane and n-nonane compared to n-octane, respectively. By excluding the effects of substrate-solvent interactions, CYP153A6 had an affinity twice as large for n-nonane compared to both n-octane and n decane. Therefore, while the 𝐾𝑚 𝑐 constants depicted that n-decane had the highest affinity with CYP153A6, this resulted from unfavourable molecular interactions between n-decane and the aqueous solvent, rather than favourable ES affinity. The 𝐾𝑚 𝑐 constants suggested that n-decane is the preferred substrate of CYP153A13 by two-fold compared to n-octane, but 𝐾𝑚 𝑎 values revealed that CYP153A13 had a six-fold larger affinity for n-octane than n-decane. The UNIF-LL model predicted substrate available in the aqueous phase at phase equilibrium for each alkane in the 2LPS. These predicted aqueous phase substrate concentrations were compared to the enzyme concentration and the MM constants, respectively. From these comparisons, it was determined that the reactions catalysed by CYP153A6 on n-octane and n-nonane were not limited by substrate concentration, whilst the reactions catalysed by CYP153A13 on n-decane were limited by substrate concentration. These results, together with enzyme expression characterisation mentioned earlier, means that CYP153A6 processes can be improved using enzyme engineering techniques, whilst CYP153A13 processes can be improved using enzyme engineering to improve enzyme expression levels as well as substrate delivery enhancement techniques. These findings highlighted the impact of substrate solubility on substrate partitioning into the aqueous phase to be available to the biocatalyst. However, once the substrate reaches the aqueous phase, effective bioconversion relies on adequate transmembrane transport, sufficient enzyme expression levels, ES affinity, and the effectiveness of the reaction mechanism. Therefore, both physiochemical and biological factors play a role in the demand-based delivery system. This demand-based delivery system is governed by the rate of substrate transfer across the liquid-liquid and membrane barriers, as well as the amount of substrate converted into product by the catalytic mechanism of the enzyme, thereby allowing more substrate to partition into the aqueous phase. Overall, the approach developed could be used to enhance the kinetic studies of CYP153 whole cell alkane primary hydroxylation in two-liquid-phase-systems, a field that receives little attention but shows importance in the goal of scaling up such systems. Globally, researchers and industry share interest in the biohydroxylation of linear chain alkanes due to their high-end properties and value, but poor solubility limits techno-economic feasibility and large-scale application. This study showed promising compatibility of CYP enzymes with medium chain alkanes (C8 C9 C10), but showed the challenges presented in molecular biology such as enzyme expression. Moreover, substrate solubility of linear alkanes impacts the location of the reaction within a 2LPS, accessibility and the apparent ES affinity. Overall, the study developed an integrated approach to investigate the reaction kinetics of an alkane whole cell hydroxylation 2LPS by considering the limitations arising from both molecular biology and biochemistry as well as thermodynamics. From this, the interdisciplinarity of biocatalysis was enhanced by integrating aspects of molecular biology and biochemistry into the greater context of chemical engineering. The findings can be used to facilitate research in alkane bio-activation process control, optimisation and scaling up by implementing the use of multiple enzymes and substrates simultaneously.
  • Thumbnail Image
    Item
    Open Access
    Development of an unsteady state model for the tank bioleaching of sulphide mineral concentrates in flow reactor systems
    (2012) Kotsiopoulos, Athanasios
    In this thesis, it is hypothesized that in bioleaching flow reactor systems, high reaction rate regions exist that can be maintained by application of biological stress trajectories. Reactor models are developed for the purpose of optimising plant operation, understood here as maximising the production rate. Complicating this attempt are a) the non-linear dynamics associated with the kinetics and b) the primary reaction's being multiphase. Mathematical models are developed to establish which particle parameters are necessary to describe reactor performance using the method of segregation. The models are distinguished by the combination of either particle residence time or age and/or particle size distributions. The models evaluated at steady state are validated against pilot plant data obtained from the Fairview Mine in South Africa and were found to be in good agreement with the data. As the model was developed using a segregation approach and thus incorporates age distributions in the model formulation, the model could be extended to unsteady state operation.
  • Thumbnail Image
    Item
    Open Access
    Development of co-disposal methods for coal discards and fine waste for the prevention of acid mine drainage
    (2019) Mjonono, Donald; Kotsiopoulos, Athanasios; Harrison, Susan
    The dependence on coal ores for energy supply has led to the considerable increase in coal discards (CD) and fine waste (FW) arising from mining and processing operations. These wastes typically contain sulphide minerals, which when oxidised may lead to the generation of acidic and toxic discharge. A deficit of naturally occurring neutralising minerals to counteract this acidic discharge results in acid rock drainage (ARD). Far reaching consequences on water systems, vegetation, people and wildlife ensue as a result. To minimise the environmental burden, the acidic water resulting from the oxidation of sulphide minerals present in wastes from both active and abandoned mines is often treated with alkaline materials and is further processed to remove metals. Indefinite maintenance and operational activities emanate from these treatment processes. Further, accumulating sludge from processing streams presents post-closure liabilities. To reduce the environmental footprint, mine waste management strategies have been developed to minimise the risk of ARD formation and proliferation. In this study, the co-disposal of CD and FW was investigated as a means to prevent the initiation of oxidation reactions at source. The CD fraction is sulphide rich with high acid producing potential but can be effectively utilised to construct structurally stable beds. In these beds, large voids are formed between the particles that facilitate the transport of oxygen and water to the sulphide mineral surfaces. Co-packing FW with sulphide-rich CD provides a sustainable approach to ARD prevention. The FW has a high-water retention capacity and can be used to encapsulate, seal or cap the sulphide bearing mineral surfaces. Apart from providing a physical barrier and decreasing voids, FW typically have low sulphide content and high specific surface area that result in increased release rates of any acid neutralising minerals present in these waste materials. Co-disposal techniques thereby provide a longterm end-of-pipe approach to ARD mitigation that may offset indefinite, resource intensive, treatment options. The co-disposal of CD with FW, however, is challenging particularly at large bed cross-sectional areas, as the incidence of high percolation rates increases. This is attributed to decreased inter-particle contact that emerge in packed beds with high void ratios, decreased packing density and increased susceptibility to deformation. This undesirable packing behaviour impacts negatively on bed stability culminating in particle displacement and increased likelihood for sulphide mineral oxidation. Fine wastes conceal these sulphide minerals by either filling voids between coarse particles or forming covers with capillary barriers and acid-neutralising effects. Consequently, the generation of ARD is inhibited. At increased scale, however, the ARD prevention efficiency of covers is enhanced by increasing the CD to FW proportion to result in a structure with high load and acid buffering capacity. The approach adopted in this study entails developing packing arrangements of co-mingled CD and FW in dry-mass ratios of 3:2 and 2:3, respectively, to improve bed stability and hence prevent ARD formation with scale up. In addition to mixture ratios, improved co-packing of CD and FW is contingent on the material geochemical properties and geotechnical parameters of the resulting packed structure. As such, geochemical analyses were performed to determine the acid producing and neutralising potential of the CD and FW through acid base accounting, net acid generating and biokinetic tests followed by geotechnical assessments. The static test results indicated that the high sulphur CD (2.16% S) was potentially acid forming and the low-sulphur FW (0.84% S) was non-acid forming with high acid neutralising capacity. The co-mingled CD and FW samples (ca. 1.5% S) were deemed uncertain as the net acid producing potential was near zero and the NAG pH was less than 4.5. Accordingly, biokinetic tests were conducted over 120 days to fully understand the acid generating and neutralising rates of the inoculated and uninoculated co-mingled samples. Near-neutral conditions were sustained for prolonged periods (> 90 days) in FW dominant samples (2CD:3FW) after which a transition to acidic conditions ensued. This highlighted the limited role of acid neutralising minerals in sustaining near neutral conditions. As ARD mitigation is contingent on preventing the rapid percolation of water and exclusion of oxygen from sulphide mineral surfaces, means to prevent the rapid depletion of neutralising minerals by either dilution or washout are essential in flow through systems. This can be achieved by decreasing voids to result in increased packing density and improved bed stability. Bed stability was shown to be dependent on several interrelated factors that included the degree of saturation of the particles (water to solid ratio, W/S), CD to FW ratio, packing configuration (layers or blends), and the extent of material compaction (assisted versus unassisted packing approach). These factors were integrated to produce 16 packing arrangements. The efficiency of these configurations was compared using packing density, slump and compressibility tests. Packing densities of ca. 0.8 m3solids. m-3mould coupled with low slump spread values (< 0.391 m) were obtained for wet, unassisted packings of CD-dominant layered and blended arrangements. Comparable packing densities and slump results were obtained for assisted packings of CD-dominant layered and blend configurations under dry conditions. Dry assisted packings of either CD or FW dominant layered systems resulted in ore beds with low compressive strain (< 7%), while in wetted structures, extensions of up to 30% over extended periods (>600 s) were noted. The large compressive extensions and the delay to achieving maximum compressive strains signalled the low particle consolidation and decreased bed stability of unassisted wet packings. As engineered co-disposal approaches are associated with long-term bed stability and hence prolonged ARD prevention, select packings were further analysed to validate their efficacy using kinetic column tests of increasing scale. An acidic feed of pH 2 was continuously introduced to the test columns at a flow rate of 3.5 L.m-2.h-1 to expedite the oxidation process and to assess the efficiency of the packing arrangements for ARD mitigation. Segregated disposal of CD in small scale columns (D = 0.19 m, H/D = 1.12) with inherent large voids allowed unrestricted access of the aqueous oxidants to the exposed sulphide minerals leading to rapid discharge of highly acidic effluent (ca. pH 2). For the wet, unassisted co-packed systems, structural instability was observed with the wash out of FW and subsequent fast effluent discharge rates. With the loss of the neutralising and reactive barrier due to migration, acidic conditions presented earlier in these wet packed beds (after 30 days) than in dry packed beds (after 90 days). The loss in geotechnical stability was more prevalent in blended systems than in layered configurations, with a rapid loss of geochemical stability following soon thereafter, despite similar neutralising characteristics in both packing configurations. In these blended arrangements, non-functional migration of the fine waste particles transpired to result in unhindered access of the oxidants with the acid generating minerals. With dilution and wash out of the neutralising components, acidic reactions dominated. In multi-layered systems, a cascading effect prevailed despite breakthrough in some layers such that a fail-safe condition resulted. Consequently, near-neutral effluent discharge at low flow rates transpired. This further emphasised the importance in preventing the displacement of particles to maintain bed stability in co-disposal prevention strategies. Assisted dry packings of blends and layers were anticipated to result in improved bed stability at large scale. As such, CD dominant blends (3CD:2FW) and FW dominant blend and layered (2CD:3FW) systems were investigated in large scale columns (D = 0.32 m, H/D = 1.12). These columns were similarly exposed to aggressive leach conditions over 120 days. As with the smaller scale columns, the packing efficiency in multi-layered arrangements were higher than for the blends. In the blended systems, evolving geochemical and geotechnical conditions were similar regardless of the CD:FW ratio demonstrating the complexity in achieving homogenously packed matrices at large scale. In multi-layered configuration, bed structural stability was sustained for extended periods as the stress imposed on the packed bed was uniformly distributed across the moisture retaining FW layers and dissipated within the matrix. Correspondingly, particle displacement was minimised, and with the cascading phenomena, ARD was successfully prevented over extended periods. A dry cover system composed of multi-layers of CD and FW is therefore recommended for pilot scale studies. Dry cover systems can be easily constructed and present a cost-effective approach to sustainable mine waste management. Further evaluation of the structural stability of multi-layers at large scale is required as changes in bed geometry, particle size and environmental conditions can alter the dump geotechnical properties and hence geochemical stability.
  • No Thumbnail Available
    Item
    Open Access
    Exploring the impact of hydrogen peroxide delivery methods on PaDa-I enzyme deactivation during styrene biotransformation within tandem catalysis framework
    (2024) Mukwenya, Rufaro; Kotsiopoulos, Athanasios; Wilbers, Derik
    Biocatalysis has become an attractive method for hydrocarbon activation due to its alignment with the principles of green chemistry. Enzymes are advantageous over traditional metal-based catalysts because they are obtained from renewable sources, typically require mild reaction conditions, and are highly selective. A few classes of enzymes have been identified as active in these transformations including oxygenases, oxidases, hydroxylases, peroxidases and peroxygenases. One of the most promising classes of enzymes are unspecific peroxygenases (UPOs). These enzymes are able to use hydrogen peroxide (H2O2) as the oxidant and electron donor without the need for expensive co-factors. The main challenge is the deactivation of UPOs in reaction systems containing excess H2O2. A potential approach to circumvent this deactivation is to produce the hydrogen peroxide in-situ in a one-pot tandem catalytic system. However, this system is poorly understood with very little kinetic data reported in the literature. In this study, different hydrogen peroxide delivery methods and concentrations were investigated to minimise enzyme deactivation. The delivery methods were once-off addition where the hydrogen peroxide was added all at once, stepwise addition where it was added at regular intervals and continuous addition where it was fed gradually at a controlled rate using a syringe pump. In addition, the influence of reaction temperature and pH on the enzyme's stability were investigated. Thereafter, the obtained optimised process conditions were evaluated in a hydrocarbon biotransformation reaction, using styrene as the model substrate. The PaDa-I enzyme (UPO variant) lost approximately 90% of its initial activity during the early stages of the reaction. This was significantly higher than the 50% and 40% activity lost when the enzyme was exposed to hydrogen peroxide and styrene as sole substrates. Despite the significant activity loss, a turnover number of 3 070 molproduct molprotein -1 was achieved employing semi-continuous addition which was determined to be the best H2O2 delivery method. A full kinetic study investigating the deactivation of the PaDa-I enzyme was also performed. Three enzyme deactivation models were tested to model the experimental data by nonlinear regression, these kinetic models were first order, two-parameter and four-parameter models. It was established that the two-parameter model best fit the experimental data. This kinetic model described a deactivation pathway where the enzyme is not completely deactivated, but rather transformed to less reactive intermediates over time upon exposure to excess hydrogen peroxide. The crucial kinetic parameter obtained from this model was the ratio of the final (less active form) and initial specific enzyme activity α1. The α1 value of 0.07 obtained for the styrene biotransformation was approximately an order of magnitude lower compared to the studies using hydrogen peroxide and styrene as sole substrates. Further insights on the kinetics of this biotransformation revealed that hydrogen peroxide behaved as a non-competitive inhibitor. This suggested that hydrogen peroxide was bound to an allosteric site on the enzyme, which distorted the structure of the enzyme making it less effective for styrene oxidation. The H2O2 concentration or delivery rate should therefore be reduced to further minimise this inhibition. The findings in this study will form the basis for future tandem systems that combine in-situ H2O2 production with hydrocarbon biotransformation using UPOs.
  • Thumbnail Image
    Item
    Open Access
    The hydrocracking of Fischer-Tropsch wax : using n-tetradecane as a model compound
    (2005) Kotsiopoulos, Athanasios; Fletcher, Jack; Böhringer, Walter
    Increasingly stringent legislation has been applied to transportation fuels to minimise or eliminate aromatics and sulphur compounds in diesel fuel. This has led to manufacturers determining alternative production methods to comply to legislation. Part of the current diesel fuel is being produced by hydrocracking heavier fractions derived from crude oil. These hydrocracking processes utilise bi-functional catalysts which have a metal (hydrogenating/dehydrogenating) function and an acid (cracking) function. The most common of these hydrocracking catalysts are combinations of either noble metals and acid zeolites, such as Pt/ HY, or combined sulphides of group VIA and VIIIA metals on amorphous acidic supports, such as CoMo/SiO2-Al2O3. For good quality diesel, the fuel should have a high cetane number and the aromatics and sulphur content should also be kept to a. minimum (e.g. EU legislation: sulphur content must be below 10 ppm (wt) by 2008). Fischer-Tropsch wax is made up predominantly of long-chain linear paraffins with exceptionally low aromatics and heteroatom content (sulphur and nitrogen-containing compounds) and therefore a good source for very 'clean', good quality diesel. The objective of this study was therefore to investigate the suitability of a conventional bi-functional hydrocracking catalyst namely, CoMo/SiO2-Al2O3 in unsulphided form for the hydrocracking of Fischer-Tropsch wax using n-tetradecane as a model compound. The purpose of using the catalyst in unsulphided form was not to introduce any sulphur to the already sulphur-free feedstock.
  • Thumbnail Image
    Item
    Open Access
    Techno-economic evaluation of integrated process flowsheets for vinasse management with value addition for decision making
    (2018) Azegele, Rony Mung'asia; Harrison, Susan; Kotsiopoulos, Athanasios
    Bioethanol production through fermentation of sugarcane juice and its derivatives such as molasses is gaining popularity worldwide as focus shifts towards renewable energy production. However, ethanol fermentation results in the production of large volumes of a dark brown and low pH liquid waste termed vinasse. At a vinasse production rate of 12-15 liters per liter of ethanol, sustainability of this bioprocess is impacted as effluent handling costs are high. If disposed onto the land, breakdown of the organic matter within may lead to the release of greenhouse gases into the atmosphere. Additionally, disposal into water bodies results in eutrophication due to the overload of plant nutrients (N, P and K). Further, owing to the high potassium content, the use of dewatered vinasse as animal feed supplements has been shown to cause digestive tract problems in ruminants depending on the supplementation rates (>10%). To increase sustainability of bioethanol fermentation processes through combined treatment and resource recovery from vinasse, biological and physico-chemical processes have been developed and implemented in industry. Conventionally, raw vinasse is dewatered through evaporation processes (MEE) as a means of volume reduction. Membrane processes such as reverse osmosis (RO) have in the recent past become popularized as water recovery options from vinasse due to process simplicity and lower costs of equipment. Resulting concentrates from RO and MEE can be used as fertilizer. Due to the high organic content, vinasse is a suitable candidate for anaerobic digestion (AD) where the organic matter is broken down to biogas and an effluent that can be safely used as fertilizer. Additionally, the biogas from AD may be harnessed for electricity generation through combined heat and power processes or upgraded to biomethane to be used as a substitute for natural gas. For high moisture content substrates such as vinasse, up flow anaerobic sludge blanket reactors are best suited as sludge residence time is prolonged thereby increasing contact time with substrate which leading to higher methane yields. AD is often sensitive to changes in temperature, substrate composition, loading rate and pH. The presence of inhibitory components such as potassium salt ions (>11.6 g/L) in the vinasse feed result in a reduction of methanogenic activity manifested through reduced biogas and methane yields. Salt recovery processes including electrodialysis and ion-exchange have been investigated in literature on a pilot scale for the removal of K+ ions from raw vinasse. To improve resource productivity, integration of vinasse treatment processes has been implemented in industry. Integration combines biological and physico-chemical processes which results in performance optimization and energy efficiency thereby improving economic feasibility of the projects. During the project conceptualization phase, process modelling is a vital tool that can be used to predict outcomes such as substrate utilization rates, product yields and optimal operating conditions of integrated processes in a timely and cost effective manner. In addition, techno-economic analyses can be used to determine cost sensitive areas and overall feasibility of the integrated processes. Having reviewed the current industrial practices, this project sought to develop integrated flowsheets consisting of biological and physical processes for the combined vinasse treatment and value creation. Value creation was demonstrated through the recovery of valuable products including energy, salts and water from the raw vinasse. Due to its simplicity and cost effectiveness, AD was selected as the primary technology for vinasse treatment and biogas production. This was coupled with a combined heat and power system for electricity generation to form the base case flowsheet. It was hypothesized that incorporation of pre- and post-treatment as well as alternative biogas utilization processes to the base case flowsheet for recovery of salts and water would generate additional revenue and cost savings. Profitability of the base case process was expected to increase with the additional pre- and post-treatments. To fulfil the objective set out and prove the hypothesis, a three step research approach was taken. The first step involved simulation and benchmarking of the base case flowsheet (AD and CHP). Using techno-economic analyses, the effect of individual addition of pre- and post-treatment options to the base case flowsheet on profitability was investigated. A framework was then developed to investigate the incorporation of combined pre- and posttreatment options to the base case flowsheet. Thereafter, a decision support tool that in comparing various combinations of vinasse treatment routes in terms of process performance and profitability was developed to aid in the synthesis of vinasse treatment processes in industry. As bioprocess modelling is complex, it was important to select an appropriate simulation platform. Given the availability of a dedicated bioprocess compound database, sensitivity and optimization features and flexible customization options within Aspen Plus, it was preferred as the primary simulation platform over SuperPro Designer and high performance programming languages (C++, Java). In developing the base case AD flowsheet, several frameworks in the literature were considered. These included ADM1 (Batstone et al., 2002), ADM-3P (Ikumi et al., 2011) and a comprehensive model by Angelidaki et al. (1993). The presence of a well defined stoichiometric framework motivated the decision to adopt the comprehensive model by Angelidaki et al. (1993). Using a combination of in-built unit operations as well as customized user models (calculator blocks), the AD model by Angelidaki et al. (1993) was implemented on Aspen Plus. As ADM1 was considered an extension of the comprehensive model (Angelidaki et al., 1993) with several similarities, kinetic constants describing substrate uptake and microbial growth were adapted from ADM1. To ascertain the predictive quality of the built AD model, four case studies in the literature concerning the AD of manure (cow and swine) and municipal solid waste were simulated and the predicted simulation results compared to the experimental results. The developed AD model accurately predicted the methane yields of the four case studies as evidenced by the average difference of 10% between simulation and experimental results. A regression analysis between experimental and predicted data yielded a value of 0.74. Given the assumptions made in simplifying the developed model, the R2 value was deemed acceptable and further affirmed the agreement between the model and experimental results. To investigate the robustness of the developed AD model, sensitivity analyses on the feed composition as well as organic loading were conducted. Increasing inhibitory compound concentrations above certain thresholds was shown to negatively impact methanogenic activity as evidenced by the decreasing methane yields. Although ammonia is inhibitory at concentrations above 0.22 g/L, it is an important nitrogen source for biomass growth. Similarly, while acetic acid is inhibitory to acetogenic microbes, it is a crucial substrate for the growth of methanogenic archaea and methane production. Inorganic salt inhibition on the other hand may be reduced through extraction of K2SO4 through pre-treatment processes. The compositional sensitivity analyses as well as the benchmarking study showed that the built AD model had a solid core framework which accurately predicted experimental data for a range of substrates. Combined with a simplified CHP model of a Jenbacher spark ignition engine (General Electric, 2008) to form the base case flowsheet, the built AD model was used for all further simulations in this work. To determine the financial standing of the base case, simulation and subsequent techno-economic analyses were conducted. At an industrial reactor capacity of 2000 m3 and a loading rate of 25 kgCOD/m3 .day, simulation of the base case process resulted in a methane yield of 45 L-CH4/kgVSadded and an electrical production capacity of 410 kW. Discounted cash flow analyses (USD, 2016) showed that the base case was not profitable within a 20-year project lifetime as evidenced by the low return on investment and internal rate of return. However, a further sensitivity on profitability of the base case showed that decreasing potassium ion concentrations in the feed would result in higher profitability higher methane yields because of decreased K+ inhibition. Despite the positive effect of on AD performance, further analyses were required to validate feasibility of K2SO4 recovery processes as well as water recovery processes aimed at further value creation from vinasse. To investigate the effect of pre-treatment on base case flowsheet economics, an ion exchange process adapted from Zhang et al. (2012) was incorporated based on the comparatively higher degree of selectivity to K+ ions exhibited by the ion exchange process than ozonation and electrodialysis. As expected, improved CH4 yields (14%), electrical production and consequently, increases (>100%) in profitability indicators were observed. However, the pretreated base case (IEX-AD-CHP) remained unprofitable which was an indication that the marginal revenue from increased electrical production and K2SO4 sales did not match the additional capital costs. To increase profitability of the base case, biogas upgrading using a HPWS system was used in place of the CHP. Due to the comparatively low cost of HPWS equipment coupled with the increased revenue from biomethane sales, the AD-HPWS process exhibited higher profitability (ROI: 19.6%) than the base case (ROI: 0%). As evidenced by the IRR (16.3%) that was greater than the cost of capital (15%), the AD-HPWS option was profitable over a 20 year lifetime. Resource recovery from the AD effluent was sought through incorporation of RO and MEE to form the AD-CHP-RO and AD-CHP-MEE routes. Most notably, there was a significant (170%) increase in cost savings with the use of RO and MEE concentrates as fertilizer compared to the raw AD effluent from the base case. Additional cost savings of up to $27 700 were achieved with upstream reintegration of RO permeate or MEE condensate water. This savings was based on the municipal water tariff of R5/kL. The combined cost savings led to increased profitability of the base case as evidenced by the increase in ROI from 0% to 3%. Potential knock-on effects of pre-treatments on efficiency of post-treatment or biogas utilization processes were noted. These were investigated through the simultaneous addition of pre- and post-treatment combinations to the base case AD process to form a decision making framework. Through techno-economic comparisons drawn between the 12 distinct vinasse treatment routes resulting from various combinations of pre- and post-treatment options in the decision making framework, three major decision criteria were established. Despite the improved performance and methane yields observed with pre-treatment addition, there was a decline in profitability of the AD-HPWS-RO/MEE processes owing to increased capital costs that remain unrecovered by marginal revenue obtained from biomethane sales. The contrary is observed with the AD-CHP-RO/MEE processes as evidenced by the 20 to 30% increase in profitability indicators upon addition of pre-treatment. This is attributed to the marginal revenues from increased electrical output as well as the cost savings from water reuse and RO/MEE concentrates. Due to the contrasting effect of pre-treatment on CHP and HPWS affiliated processes and profitability, the presence of inhibitory potassium ions was considered a decision criterion. Due to the low cost of HPWS equipment, it was observed that choosing to upgrade biogas to biomethane as opposed to using CHP exhibited higher performance (energy output) and profitability in all process combinations. This was evidenced by the higher ROI and IRR of the AD-HPWS, AD-HPWS-RO/MEE and IEX-AD-HPWS-RO/MEE process options compared to the CHP counterparts. As a result, the choice of biogas utilization was considered an important decision criterion affecting profitability. Because of increased cost savings with upstream reintegration of water and the use of concentrates as fertilizer, the implementation of RO and MEE was observed to increase profitability of all process options including AD-CHP/HPWS and IEX-AD-CHP/HPWS. This was majorly through cost savings from use of RO and MEE concentrates as fertilizer ($250 000/yr) and upstream reintegration of water. This led to the conclusion that the recovery of concentrates from vinasse is an important decision criterion when looking to increase profitability and process sustainability. Overall, based on the techno-economic analyses, the most profitable vinasse treatment process included an anaerobic digester coupled with a high-pressure water scrubbing system for biomethane production and reverse osmosis process for water recovery (ROI: 22.9%, NPV: $540 000). This facilitated both increased energy output from biomethane and cost savings from water reuse. Further research is recommended around the AD modelling aspect to extend functionality to ionic speciation and pH prediction. it is recommended that equipment quotes from suppliers within South Africa be sourced as opposed to costing heuristics in the literature to increase the accuracy of capital and operating expenditure.
  • Thumbnail Image
    Item
    Open Access
    Towards tandem bio-chemo catalytic systems for the activation of alkanes and subsequent oxidation of alcohols to aldehydes using biofabricated Pd and Au catalysts
    (2021) Govender, Mivashya; Kotsiopoulos, Athanasios; Harrison, Susan
    Alkane activation is known to be difficult due to the stable, saturated nature of the C-H bonds. Typical C-H bond activation techniques use harsh conditions, toxic solvents and excessive amounts of energy to produce low-value fuels and solvents. To develop value from the alkane feedstock and promote sustainable chemical manufacture, bio-catalysts are considered for terminal bond activation. This has been successfully demonstrated for the conversion of n-octane to 1-octanol using cytochrome P450 enzymes (Gudiminchi et al., 2012; Julsing et al., 2008; Meissner, 2013; Olaofe, 2013; Pennec et al., 2014). However, due to the low conversion and lowervalue of the alcohol product (Olaofe, 2013), the generation of higher value chemicals is needed to achieve a techno-economically feasible operation. A tandem catalytic process is proposed to valorize alkane activation using a cascade of bio-chemo reactions. Typically, chemical catalysts are synthesized using synthetic supports such as titanium dioxide or activated carbon. However, certain biological supports have a natural affinity for metal ions and have the ability to generate uniform, mono-dispersed nanoparticles, without the use of stabilizers or capping agents. The bacterial cell can be used as a support for chemo-catalysts that are easily accessible and cultivated. In this study, three different strains of Escherichia coli (E. coli) bacteria were considered as catalyst supports; namely E. coli DH5a, E. coli ATCC25922, E. coli BL21DE3. E. coli BL21DE3 was previously used as the host cell for the biocatalytic activation of alkanes. However, due to the known deficiency in hydrogenase for this strain, alternative E. coli strains (E. coli DH5a and E. coli ATCC25922) that could potentially also be used for the expression of this enzyme, were also considered. Reduction on these microbial supports required an electron donor (hydrogen and sodium formate) for manufacture of monometallic Palladium (Pd) and Gold (Au) nanocatalysts. Biosorption studies showed rapid adsorption of Pd and Au ions onto all microbial strains within 1- 5 minutes of cell exposure that was best described by the chemisorption of the metals onto amine, hydroxyl or thiol functional groups. Near-complete adsorption of Pd(II) occurred on all microbial strains, with E. coli ATCC25922 achieving the greatest capacity (ca. 94.4% mol/mol) at 5% (w/w) Pd loading. The adsorption efficiency declined to 27.9% (mol/mol) when the metal loading was increased to 25% (w/w) Pd. Reduction timeframes were dictated by the electron donor, with a hydrogen induced colour change noted within 20 minutes compared to 24 hours using sodium formate. Showing characteristics of ideal nanocatalysts, hydrogen-generated Pd(0) nanoparticles ranging from 3.0-3.5 nm and 3.9-9.3 nm in size were formed across all microbial strains at either 5% (w/w) or 25% (w/w) metal loadings. These nanoparticles were uniform and well-distributed within the cytoplasm. Clustering was most prevalent for the hydrogen-generated 25% (w/w) Pd loaded catalysts on E. coli ATCC25922 and E. coli BL21DE3. Minimal agglomeration and loss of Pd was observed on E. coli DH5a. In comparison to Pd(II), Au(III) was poorly adsorbed on all microbial strains with ca. 45.6% (mol/mol) and 33.7% (mol/mol) adsorbed on E. coli ATCC25922 at the 5% (w/w) and 25% (w/w) loadings. Bioreduction for this metal was only observed with hydrogen as the electron donor on E. coli ATCC25922 with irregularly-shaped Au(0) nanoparticles between 20 nm and 40 nm being formed. Catalyst activity was assessed using the oxidation of benzyl alcohol to benzaldehyde as a control reaction. No notable activity was detected for any of the Au catalysts. The greatest activity was observed by hydrogen-generated Pd catalysts at 25% (w/w) loading with conversions of up to 32.8±2.7% (mol/mol) and selectivity of 94.1±2.8% (mol/mol) across all microbial strains. For 1-octanol, hydrogen-generated 25% (w/w) Pd loaded E. coli ATCC25922 nanocatalysts achieved the highest activity to reach a conversion of 2.4% (mol/mol) with a selectivity of 82.7% (mol/mol) towards octanal. The addition of water limited byproduct poisoning to improve the conversion of 1-octanol to 9.6% (mol/mol) and 99.4% (mol/mol) selectivity to the aldehyde. The successful activity of the biofabricated catalyst on aromatic and aliphatic alcohols shows promise for tandem catalysis to valorize the alcohol product from bio- catalytic activated alkanes. Consequently, this approach can be used to improve the value of noctane via a bio-chemo catalytic cascade reaction where higher selectivity to the aldehyde can be reached.