In the context of steering product formation in anaerobic digestion systems, the present thesis work elaborates on the potential role of elevated CO2 partial pressures as an environmental driver that may influence end-product selectivity from methane towards compounds from the carboxylic platform. As an emerging field of research, organic acid production via mixed culture fermentation is currently in an exploratory phase and the understanding of basic functional principles driving each of the biochemical conversions of interest is of vital importance.The present investigation forms part of a series of studies conjunctively aimed at elucidating the effect of elevated CO2 partial pressures on glucose fermentation, which consists of a complex network of several metabolic routes. Specifically, this thesis work focuses on the effects of CO2 partial pressure on the degradation of two key metabolites that are central and/or highly relevant to the glucose conversion pathways, namely pyruvate and butyrate.

Interest in biogas has increased due to natural gas shortages and the energy transition. Most digesters in the EU have internal heating for process stability. Unheated reactors could potentially lower costs, but have decreased kinetics due to low operational temperatures, especially in winter. This raises the question how seasonal temperature fluctuations effect biogas production. If this effect could be incorporated in current models such as the Anaerobic Digestion Model no.1 (ADM1), scenarios could be compared to investigate when heating is worthwile. In this thesis, an extension is created to the ADM1 in Python to predict the operational temperature and its effects on the anaerobic digestion process. To lower the stiffness of the ADM1, different pH calculation methods were compared. These methods and the used Python version of the model were validated by digestion of the Benchmark Simulation Model 2 (BSM2) ADM1 influent. It was found that the combination of the Differential Algebraic Equation (DAE) with the Radau numerical solving method produced the smallest errors with the lowest computational burden. A heat-transfer model was included, which calculates bulk liquid temperature as a function of weather data, digester design and some operational factors. A One-At-a-Time (OAT) sensitivity analysis showed the dependence of yearly temperature fluctuations on several design and operational parameters. The calculated operational temperature is coupled to the ADM1 through several temperature inhibition functions for biochemical processes. Furthermore, the dependency of the liquid-gas transfer coefficient was investigated. It was found that this last step in biogas production could potentially become rate-limiting at temperatures lower than 30 °C, but this result was not validated. The coupled heat transfer - ADM1 model is used to study the case of an unheated fixed-dome reactor for the co-digestion of spent wort and Waste Activated Sludge (WAS) of the La Trappe brewery in the Netherlands. The influent substrate concentrations were determined by a biochemical fractionation procedure, and the model was able to simulate methane production curves found in Biochemical Methane Potential (BMP) experiments. Simulations of digestion for the La Trappe case showed seasonality in biogas production, increasing in summer and decreasing in winter. The extent of the effect was found to be dependent on substrate composition, with lipid-rich substrates being affected more than carbohydrate- or protein-rich substrates. Furthermore, the simulation showed seasonality in rate limiting step, with hydrolysis in summer and methanogenesis in winter. This caused the pH to lower with winter, and it was found that higher base dosage is needed in colder winters to prevent acidification. Higher influent temperatures improved process stability, lowered amounts of base needed and increased biogas production. Furthermore, the relation between temperature and volume showed that washout of methanogens and consequent VFA accumulation is dependent on the absolute temperature inhibition function for acetoclastic methanogens. Simulation of the digestion of BSM2 influent showed that the maximum Organic Loading Rate (OLR) increased with temperature till 30 °C, but decreased above this point as free ammonia inhibition starts to limit methanogenic metabolism. For La Trappe, OLR determines the needed base dosage, allowing smaller volumes if more base is added. Lastly, internal heating with the produced biogas was investigated, and optimal heating for biogas production was found. The found results from the model were used to asses the costs and benefits of several design and operational options for the digester at La Trappe. This digester will be used in future research for validation of the model.

Biogas is the well-known product of Anaerobic Digestion (AD), but nowadays, the intermediate products (volatile fatty acids - VFAs) of anaerobic metabolism have gained increasing attention inside the “carboxylate platform”. However, steering and optimizing the process for selective metabolite production is still an unraveled task inside this field since it relies on the manipulation of operational parameters. The objective is to understand the conversion of glucose and glycerol in the mixed culture of anaerobic digestion to unravel possibilities to steer product formation. Glucose and glycerol are the main components in the waste streams of beverage and biodiesel industries. Regarding the degradation pathways in AD, both glucose and glycerol are oxidized to pyruvate by fermentative bacteria to obtain energy and metabolic intermediates under anaerobic conditions through the same intermediate, glyceraldehyde-3-phosphate. Pyruvate, the key branching-point, allows the process to enter different metabolic pathways which lead to the formation of various metabolites. Under the fermentation conditions, redox balance is necessary to be maintained through terminal electron transfer to internally produced compounds. Since glycerol has a higher degree of reduction than glucose (Glucose: 0.33 NADH/C-Glucose; Glycerol: 0.66 NADH/C-Glycerol), the conversion of glycerol into pyruvate generates a double amount of reducing equivalents. On the one hand, this provides the advantage of higher theoretical product yield of reduced compounds. On the other hand, half of the glucose is lost as CO2 during the fermentation, and this reduces the product yield.1 Therefore, we assume that elevated pCO2 could have a more significant detrimental effect on glucose fermentation. In this research, batch experiments at different pCO2 (0.3, 1, 3, 5, 8 bar) were performed, and different types of measurements and analyses were employed to monitor the pCO2 effect on the metabolism. We designed some of the potential pathways of glucose and glycerol conversion under elevated pCO2. The elevated pCO2 converged the product spectrum of both substrates towards propionate production but affected the degradation and production phase of propionate and acetate. Initial pCO2 of 0.3 bar and 1 bar did not cause visible inhibition on the propionate production of both substrates. However, the propionate degradation was kinetically affected under 0.3 and 1 bar initial pCO2. Although propionate was degradable, its degradation phase at 1 bar initial pCO2 was longer than 0.3 bar. On the contrary, when the pCO2 was elevated to 3, 5, and 8 bar, not only the propionate production phase became longer, but also the maximum concentration became lower on both substrates. Moreover, propionate degradation was ceased. The lower propionate production was suspected to be due to the inhibition of NADH production as a consequence of the elevated pCO2 effect. The undegradable propionate might be attributed to unfavored decarboxylation reactions under elevated pCO2. The enrichment approach was applied to examine the adaptability of the microbial consortium under the CO2-exposing environment. Therefore, not only the more predominant metabolic reaction would be favored during the enrichment, but also changes in the community due to CO2 influence were expected. Propionate degradation was achieved with this inoculum at the only tested condition (5 bar initial pCO2). From the community analysis, Smithella was enriched during the enriched and it was suspected to play a significant role in the propionate conversion. Besides, the difference between the substrates on the fermentation has been observed. Due to the higher available reducing power of glycerol than glucose, glycerol was potentially able to generate more propionate. However, the butyrate formation also needs the reducing power to proceed with the reaction, but it was not detected in the glycerol fermentation. Therefore, reducing power distribution from specific substrates with elevated pCO2 might also be affected. Moreover, the substrate was also hypothesized to influence cell viability, where glycerol fermentation increased the cell viability but glucose not. The degradation of acetate and butyrate in the non-enriched inoculum was observed to be kinetically affected by elevated pCO2, with the later becoming undegradable at 8 bar. The reason for this phenomenon needs further investigation.