Drinking water production works well, but we can do better.

Anaerobic groundwater is an excellent drinking water source. It presents several advantages over its counterpart, surface water, such as constant quality and temperature, and it is considered to be microbiologically safe. The main contamination sources of anaerobic groundwater are the decomposition of natural organic matter and the dissolution of soil minerals. The first one produces compounds such as ammonia, while the second introduces manganese, iron, and trace metals. Iron, ammonia and manganese must be removed from groundwater to produce drinking water. For this purpose, humans have been using rapid sand filtration - preceded by an aeration step - for over a century.

The purpose of aeration is to strip out undesired gases and to introduce oxygen up to saturation levels. At this oxidation-reduction potential, iron, ammonia and manganese are oxidized by different physical-chemical and biological processes in the subsequent rapid sand filter. As a result, most contaminants precipitate, forming solids that are captured by the filter, and clean water is produced. Although widely used and robust, solid understanding of the intricacies of rapid sand filters is still missing. A high degree of complexity is hidden behind their seemingly simple working principles.

The main reason underneath this extraordinary complexity is the high oxygen load introduced during the aeration step. The saturation of anaerobic groundwater with oxygen onsets a series of simultaneous, interwoven and uncontrolled reactions whose nature and contribution to the overall process remains unknown and generally unpredictable. This process convolution precludes understanding and optimization of rapid sand filtration.

The overarching goal of this thesis is to gain knowledge to advance towards the design of high-flow, resource-efficient sand filters. To do so, we must be able to understand the mechanisms that govern which reactions take place, in which order they occur, and how they affect each other, which will ultimately allow us to predict and control i) microbial community assembly and performance and ii) the interplay between chemical and biological reactions. In the first part of this thesis, we focused on gaining mechanistic understanding of how current sand filters work using laboratory, pilot, and full-scale experiments. In the second one, we used this freshly acquired knowledge to design and test novel systems…
@en

This research explores the potential effect of Subsurface Iron removal (SIR) on the removal of organic micropollutants (OMPs). The presence of OMPs and their transformation products in the water environments globally, have raised concerns due to the potential environmental and human health risks they are posing. Recent technologies often fail to remove OMPs completely or require high energy levels and costs. Previous findings by Vitens water company have indicated that SIR application resulted in the attenuation of certain OMPs. Therefore, the effect of SIR as an alternative, cost-effective removal method of OMPs was investigated.

A continuous flow column experiment was conducted in order to simulate the processes which take place under oxic conditions at SIR for the removal of 5 targeted OMPs (bentazone, metformin, caffeine, carbamazepine and atrazine). Two columns were filled with iron oxides (FeOx) coated sand and manganese oxides (MnOx) coated sand respectively, in order to simulate the two precipitation zones created in the subsurface. The aim of the study was to investigate the removal efficiencies of the selected OMPs in both columns in order to understand if the SIR environment is favourable for the
removal of OMPs.

Results indicated that for most compounds there was removal observed after 65 pore volumes of continuous flow. Metformin was hardly removed (<10% removal rate) in both materials. A higher removal efficiency was observed at the FeOx column for bentazone (65%), caffeine (54%) and carbamazepine (29%), while atrazine was the only compound which had a greater affinity for MnOx, with a removal rate of 87%.

These findings suggest that SIR has a potential on the removal of certain OMPs. It is suggested that during SIR the OMPs existing in the subsurface, pass through the two precipitation zones and undergo removal processes within both zones. However, based on their respective affinities to each of these zones, their removal extents vary accordingly. Ultimately, the water extracted from the undergoing SIR exhibits a reduced concentration of OMPs, which is a result of the combined removal mechanisms operating within both precipitation zones.

In the treatment of industrial wastewater, the biological removal of aniline & nitrogen often occurs simultaneously within the hyper-saline water matrix. This study focuses on pioneering the research
on combined aniline-nitrogen removal in a hyper-saline environment. The experimental approach assessed the cause of nitrite accumulation phenomena in this specific case study as well as the influence
of oxygen concentration and aniline as a carbon source on nitrification. The hypothesis of increasing aniline biodegradability by making it chemically react with nitrite was experimentally tested. In addition, data analysis of the full scale of this study was performed in order to create analytical models to represent its biological processes. It was found that the correlation between oxygen concentrations and nitrification was non-linear and differs from what is observed in the literature due to the significant influence of different boundary conditions. The unavailability of carbon sources during the nitratation process was the cause of nitrite accumulation, which crucially affects nitrification. The chemical reaction between aniline and nitrite (most likely polymerization) resulted in a compound that did not limit nitrification and had more biodegradable potential. Furthermore, an aerobic model was developed but the promising results were most likely caused by the mathematical optimization, which was done in order to bypass the lack of data. Nevertheless, respirometry was used to build an empirical model to detect nitrite accumulation in the effluent of the full scale, which showed promising results. This study can potentially pave the way for the application of Anammox technology in the combined treatment of aniline & nitrogen, which is discussed in the hypothetical design implementations.

Conventional groundwater treatment plants consist of aeration and rapid sand filtration steps, that are merely designed and optimized for iron (Fe), manganese (Mn) and ammonium (NH4+) removal. Understanding the various reduction-oxidation pathways, and interactions of manganese and iron, can play a major role in optimizing the performance of such filters. Interestingly, it is found that under certain conditions, mobilization of dissolved manganese can occur in such filters, which can be critical to the filter operation. Therefore, the main aim of this research is to dive deep into studying the possibilities of manganese reduction pathways occurring at the top layer of the filter media of a groundwater filter. Secondly, the research also focuses on knowing how the removal of manganese is related to the oxidation by MnO2+O2 systems, and also how these systems interact with each other under different pH conditions.
To do so, manganese dioxide (MnO2) coated sand grains were obtained from the second filtration unit of Vitens groundwater treatment plant situated in Holten. Various batch scale experiments were done under aerobic as well as anoxic conditions, in the presence of Mn(II) or Fe(II). Additionally, the influence of pH on manganese removal efficiencies as well as the rates of both manganese and iron oxidation was investigated.
It was found that the dissolved Mn was a reduction product of MnO2-Fe(II) system, where Mn(IV) got reduced to Mn(II), reaching an Fe(II) : Mn(II) molar ratio of 3.65:1 instead of 2:1, as there was a significant difference between the calculated theoretical values and the measured experimental values of both Mn(II) and Fe(II). There was mobilization of Mn(II) which took place from the MnO2 surface, when there was a presence of Fe(II) in the system, which simultaneously got partially oxidized to Fe(III). Also, it was observed that manganese could be removed by MnO2 under anoxic conditions, although under aerobic conditions the removal efficiency was high (93.32% vs 71.83%). Apart from oxidation, there is a possibility of adsorption over MnO2 due to its high sorption capacity towards cations like Mn2+, Mn3+ and Fe2+. This research also showed that a small fraction of Mn(II) reacts with Mn(IV) to form Mn(III) as a reaction product, enhancing the mobilization of Mn(II).

Groundwater is an excellent source for the production of drinking water due to its low concentration of microorganisms and contaminants. The main contaminants present in groundwater are iron, ammonium and manganese, which are sequentially removed in rapid sand filters by a combination of biological and chemical processes. Currently, limited knowledge of the removal mechanisms challenges the design of new drinking water treatment plants (DWTPs). As a result, different treatment schemes are used to treat groundwater with a similar composition. We hypothesize that the difference in plant configuration does not affect the stratification of removal processes. In the present work, the spatial distribution of taxonomic and functional microbiological profiles and physical-chemical removal mechanisms were investigated in two DWTPs with similar incoming water but different treatment schemes. One plant employs a dual bed filter, while the other uses two sequential filters. Concentration profiles and composition of the coating on the sand showed sequential removal of iron and manganese, while ammonium removal was ubiquitous. Activity batch tests revealed section-specific manganese removal mechanisms: adsorption in the first section and oxidation in the second section. The latter is the dominant process in full-scale filters. Manganese oxidation capacity remained constant over the height of the second section. Contrarily, ammonia removal was highly stratified in there. The highest ammonium removal rates were observed at the top section of the filter. Furthermore, ammonium removal rates were higher in the second section compared to the first one. In accordance, metagenomic analysis revealed higher abundances of nitrifying organisms in the second sections. Besides, co-existence of nitrifiers and iron-oxidizers was observed in the top layer, in contrast to the common opinion that iron removal has to be complete before nitrification can start. We (i) conclude that similar distributions in removal mechanisms and genomic profiles were observed in both DWTPs, regardless of plant configuration, (ii) provide the first holistic quantitative analysis of the biological and chemical reactions in full-scale rapid sand filters and, (iii) to our knowledge, prove for the very first time that iron hydroxides on the sand grains adsorb manganese under aerobic conditions, using both adsorption and desorption tests.

Nitrous oxide (N2O) is a potent greenhouse gas, and its stratospheric concentration is already 20% above the pre-industrial level. Over 70% of N2O is produced by microbial processes. Nitrous oxide reductase is the only enzyme able to reduce N2O to innocuous N2. The high abundance of organisms harbouring it and lacking the genetic potential to produce N2O, commonly known as specialist N2O-reducers, has only been recently disclosed in diverse ecosystems. In this work, we aim to understand which mechanisms select for them in open cultures and how N2O affects cellular metabolism. Two continuously-fed stirred-tank membrane reactors (CSTMR) were run at 20 °C, pH 7 and low dilution rate (0.14 d-1) under either acetate- or N2O-limiting conditions. Denitrifying activity could not be completely washed out from the reactors. Yet, a ten orders of magnitude increase in clade II nosZ gene abundance, often associated with specialists, was found under N2O limitation. Moreover, cultivation under N2O excess resulted in 30-50% lower biomass yields and a 25% higher maintenance coefficient as compared to N2O limitation. Lastly, polyhydroxyalkanoates (PHA) consistently accumulated (up to 20 wt%) under both acetate and N2O limitation, and the possible biochemical mechanisms are discussed. We conclude that (i) affinity for N2O was not selective enough to enrich for specialist N2O-reducers under the imposed conditions, and (ii) provide further evidence for the potential cytotoxicity of N2O on the cellular metabolism and, to the best of our knowledge, (iii) report the first evidence of PHA accumulation in N2O-respiring enrichments.