Recently, a mechanism has been proposed to describe how salts do (not) contribute to bubble coalescence inhibition. This is done by considering how ions (re)distribute themselves along the G/L-interface, expressing this as a Gibbs-Marangoni pressure. The thesis’ goal is to investigate the potential and applicability of this mechanism to predict bubble coalescence inhibition for salts in biotechnological processes, specifically bubble columns. This goal is achieved by recreating the mechanism and evaluating whether the Gibbs- Marangoni pressure is a better predictor than the ionic strength. To make this comparison, experiments are performed to study which ionic strengths/concentrations induce bubble coalescence inhibition. These numbers are then modelled in terms of Gibbs-Marangoni pressures, from where it is assessed which of the two better predicts this inhibition. More- over, additional experimental results from the literature have been tested on the model to support this assessment. The outcome of this thesis is a combination of multiple interesting insights. First, the above mechanism is successfully reproduced into a model. This statement is supported by com- paring the model’s results with previous literature applying this mechanism, showing that results deviate less than 10% from each other. Secondly, from the performed experiments, bubble coalescence inhibition by salts is observable at an average ionic strength of 47.82 mM. Using these results in the model shows that this inhibition happens at an average Gibbs-Marangoni pressure of 1312.91 Pa. How- ever, the inhibiting salt concentrations from the experiments deviate by more than 168% compared to the literature. This difference can be caused by many factors, such as dissimilarities in the experimental setups used. It shows that the critical Gibbs-Marangoni pressure is dependent on these factors and therefore not entirely generalizable for all scenarios. Finally, the model’s results are validated by comparing whether the additionally calculated surface tensions deviate less than 10% from the literature. This showed resulting Gibbs- Marangoni pressures are valid for pure salts at concentrations from 0.02-1.50 M. However, the validity of the model for salt mixtures is questionable. To conclude, the proposed mechanism can describe bubble coalescence inhibition for salts in bubble columns. However, its accuracy is debatable as the model’s results cannot be validated for mixtures. This may require the inclusion of other ion contributions into the model, such as ion polarizability, solvation energy and ion-charge density. Hopefully, this brings us closer to better modelling the effect of salts, and other compounds, on bubble coalescence inhibition.

Anaerobic digestion (AD) is an effective and economical in secondary sludge management. Thermal hydrolysis process (THP) is a reliable pre-treatment method to improve the biodegradability of sludge. Partial nitritation and anammox process (PN/A) is a suitable method used to remove nitrogen in reject water which has a low COD/N ratio. However, some organic compounds produced in THP process, such as melanoidins, have negative effects on AOO and anammox bacteria. Considering that humic substances occupy a large part of the soluble organics in reject water of THP-AD process. Therefore, this study aimed to investigate the acute and long-term toxicity of three types of humic substances (commercial humic acid, melanoidins and reject water) on partial nitritation step of PN/A process. The biodegradability of three types of humic substances under aerobic and anaerobic conditions, and the evolution of humic substances during the biodegradability assays were also analysed. The results showed that all three types of humic substances had higher biodegradability under aerobic condition than that under aerobic condition, and melanoidins had more biodegradable compounds than reject water and humic acid. During the AD process, humic substances acted as electron acceptors, decreasing the cumulative methane production of all three types of humic substances compared to the blank. Also, small MW fractions were transformed to large MW fractions in all three types of humic substances during both aerobic and anaerobic digestion. Moreover, during aerobic and AD, both fulvic acid-like and humic acid-like substances were formed. In addition, the results of acute toxicity tests showed commercial humic acids had a stronger inhibitory effect than reject water and melanoidins on AOO biomass. In addition, a positive correlation between humic substances concentration and AOO activity loss was observed. The long-term operation of PN reactor was failed due to the formation of granular sludge in the reactor and the growth of AOO biomass in the tubing and culture medium.

The removal of iron from groundwater is essential to avoid aesthetic issues of the produced drinking water and to reduce maintenance cost of the system. The most applied iron removal method of oxidation and filtration produces large volumes of aqueous iron sludge of little value and the method is more likely to fail at high iron concentrations. This research investigates the novel concept of removing iron(II) anaerobically from groundwater by precipitation as vivianite (Fe3(PO4)2 ¢ 8 H2O) by dosing phosphate to the water. Natural groundwater was used to investigate the proposed method. A better understanding of the ironphosphate chemistry was obtained in experiments with a synthetic iron solution. To find a possible alternative to the limited resource phosphate, the possibility of removing iron by forming a compact mineral was also tested by dosing sulphate and carbonate. The chemical equilibriums and carried out experiments were evaluated by a geochemical model and the reaction products were analysed by X-ray diffraction. Up to 93% of iron removal by vivianite precipitation was obtained by dosing phosphate to anaerobic groundwater spiked with 100 mg Fe/L. An additional aeration step increased the efficiency to 99.9%, a higher total removal efficiency compared to the conventional aeration-filtration technique. The geochemical model showed that the anaerobic removal stopped when the saturation index (SI) of vivianite drops below 4, which explains why the last 7% were not removed anaerobically. Increasing the pH increases the SI of vivianite and can enhance further removal. Theoretically iron can be removed by vivianite precipitation starting from a concentration of 1 mg/L at a pH of 8.5. A second order kinetics was found for the removal of iron by vivianite precipitation at pH 7 with a rate constant of 2.27 M/s. The corresponding half life of iron is 4 minutes, while the half life of iron oxidation is 16 minutes at the same pH. Vivianite was the only crystalline end product detected and this decreased the sludge volume by a third compared to the sludge currently produced with oxidation and filtration. The total iron removal by sulphate addition only reached 73%, probably caused by the formation of iron(III)- sulphate complexes, and is therefore not a proper alternative to phosphate. The addition of carbonate reached an anaerobic removal of 59% and was increased to 99.9% by aeration. The formed sludge contained of a mixture of several oxidised compounds and the volume was almost 6 times higher compared to the currently produced aqueous iron sludge, which is why carbonate is not considered as an interesting alternative. The possibility of removing iron(II) anaerobically from groundwater by forming a compact mineral is successfully demonstrated. This method can increase the efficiency of drinking water production: higher throughput rates can be reached and a valuable end product with interesting reuse opportunities is created, provided that the phosphate can effectively be recovered from the water. It can decrease the operational costs of groundwater production substantially. The proposed novel method is a promising alternative to the conventional treatment method of oxidation and filtration, especially at plants where large iron sludge volumes are currently produced caused by elevated iron concentrations.

This thesis examines the influence of physical properties of structural EPS on granule morphology. Two types of structural EPS are found to be important. These types of structural EPS are: surface bound EPS and hydrogel forming EPS matrix. VirtualLeaf is used to model granules which contain only binding surface bound EPS. The TST model is used to find the influence of both structural EPS components. In VirtualLeaf compact granules are formed with slight bulging features, similar to granules found in literature. The TST shows that binding surface binding EPS increase the compactness of granules. Furthermore it is shown that porosity of the EPS matrix also determines compactness and maximum size of granules. The strength and measure of the EPS matrix is shown to influence granule stability.