Dissolved air flotation (DAF) is a water treatment technology meant for the separation of suspended particles from the water. Since the fifties of the XIX century, it has gained an important role in many sectors of water treatment applications. In general, the flux variability, the small space requirement, and the limited energy consumption excel as some of the most relevant advantages of the DAF system. However, because of its not always excellent removal efficiency, DAF applications are usually considered as pre-treatment technologies to reduce the particles' load from more effective technologies as sand or membrane filtration technologies. Briefly, the amelioration of the suspended solids' separation is of particular interest because it decreases the maintenance cost of the next filtration units and, in some cases, it might substitute them. Coagulation and flocculation processes prove to be of extreme importance for the DAF performance as it is characterized by particles cut-off around 1-10 μm. The purpose of this research was the application of extra-polymeric substances (EPS) as a flocculant to enlarge the floc dimension of an anaerobic digested sludge and improve DAF efficiency. In the literature the flocculation ability of EPS is not completely acknowledged, but there are some succesfully results with microalgae and activated sludge. Specifically, the feasibility of producing EPS directly on-site, is interesting because it makes the treatment plant possibly independent from the flocculant purchase. Moreover, an extra amount of EPS might be sold (1$/kg) or used for the production of other valuable products (e.g. biodiesel production). In collaboration with the Water Lab at the faculty of Civil Engineering in Delft and Royal Haskoning (DHV), it was possible to perform some experiments to understand the flocculation properties of EPS in the dissolved air flotation technologies. More specifically, the effects of the EPS were studied with two experiments: a jar test and a flotation column experiment. The sludge used for the experiment was collected from Harnashpolder facility after being digested for at least 25 days. With the jar test experiments, it was examined the effects of sludge settleability (5 gTSS/L): increasing EPS doses (200-1200 mgEPS/L), increment by 20% the original SVI value, on average. Furthermore, it was performed a small scale flotation experiment with two Alka seltzer pills. After a certain EPS dose (400 mgEPS/L), TSS in the formed foam resulted 1.78 times more concentrated. Finally, it was also analysed the particle size distribution (PSD) variation, but the measurements were characterized by a high standard deviation which reduced the reliability of the results. The scope of these first experiments was the individuation of a certain EPS dose to apply in the flotation column experiment simulating the DAF performance (800 mgEPS/L). The EPS addition did not improve the quality of the effluent (65%), and was slightly inferior to the blank series (69%). However, the obtained foam was almost 1.75 times more concentrated when compared to the blank solution. Towards the end of the experiment, a better removal efficiency was notated within the EPS series and it was correlated to the foam concentration (R2= 0.989). Despite the errors, PSD of the EPS series were characterized by a higher frequency of small particles, between 1-10 μm. In the discussion chapter two explanations are proposed considering the effect of zeta potential variations after the EPS addition. In conclusion, the use of EPS to enhance the solid-liquid separation of the sludge was not successful. However, further research with different doses of EPS and different types of sludge should be studied to assess the potential use of EPS to improve the efficiency of a DAF system.

Purple non-sulphur bacteria (PNSB) are phototrophic bacteria currently under study in the wastewater treatment sector due to their performant nutrient and resource recovery. Until recently, most of researches have focused on closed anaerobic photobioreactors resulting in high selectivity of PNSB and appealing hydrogen, microbial proteins, and carotenoids productivity. Unfortunately, these researches also showed that the implementation of this technology is hampered by its relatively high costs. As is the case for microalgae technology, raceway reactors could possibly overcome this problem thanks to their low investment and maintenance costs. However, a survey of the scientific literature shows that only few researches have investigated the application of PNSB technology with open raceway reactors and that the limited available mechanistic models do not consider the specific conditions which characterize these reactors. Therefore, this study aims to construct a mechanistic mathematical model which includes the mixotrophic metabolism of PNSB and could be used to predict the nutrient removal and recovery and PNSB relative abundance in the raceway reactor. The model was structured mainly considering the photoorganoheterotrophic and respiring chemoorganoheterotrophic growth of PNSB competing with standard (an)aerobic-respiring and fermenting chemoorganoheterophic bacteria under semi-aerobic conditions. The model was tuned with seven batch and sequencing batch reactor (SBR) laboratory experiments. It simulates the reactor performance (COD removal rate 480-780 mgCOD/L/d, yield 0.40– 0.65 mgCODx/mgCODs) and the relative PNSB abundance (10-60%) under different operational conditions (light, dissolved oxygen, surface area) with a relative error around ± 20%. This research proposes an one-at-the-time sensitivity analysis, analysing the impact of those variables (TSS, SRT, COD, light, and biotic competition) which could play an important role in real-case scenarios. From this analysis, it emerged that influent suspended solids (TSS>250 mgTSS/L), hourly variations of the natural light cycle intensity and drops of the available soluble substrate (COD<1000mgCOD/L) could strongly disturb the abundance of PNSB in the system (from 48% to 10%). An extension of the sludge retention time, from 2 to 5 days, was observed to favour the relative abundance of PNSB (from 48% to 60%) and to increase the TSS productivity (from 235 to 400 mgTSS/L/d). The model was intended as a first attempt to simulate the nutrient removal and the PNSB dynamic in a raceway reactor. It has the flexibility to study the impact of the crucial parameters evidenced from the literature review (light, oxygen, carbon source). However, several implementations will be needed in the future, mainly focusing on anoxic chemoheterotrophic growth and hydrolysis.

Purple non-sulphur bacteria (PNSB) are phototrophic bacteria currently under study in the wastewater treatment sector due to their performant nutrient and resource recovery. Until recently, most of researches have focused on closed anaerobic photobioreactors resulting in high selectivity of PNSB and appealing hydrogen, microbial proteins, and carotenoids productivity. Unfortunately, these researches also showed that the implementation of this technology is hampered by its relatively high costs. As is the case for microalgae technology, raceway reactors could possibly overcome this problem thanks to their low investment and maintenance costs. However, a survey of the scientific literature shows that only few researches have investigated the application of PNSB technology with open raceway reactors and that the limited available mechanistic models do not consider the specific conditions which characterize these reactors. Therefore, this study aims to construct a mechanistic mathematical model which includes the mixotrophic metabolism of PNSB and could be used to predict the nutrient removal and recovery and PNSB relative abundance in the raceway reactor. The model was structured mainly considering the photoorganoheterotrophic and respiring chemoorganoheterotrophic growth of PNSB competing with standard (an)aerobic-respiring and fermenting chemoorganoheterophic bacteria under semi-aerobic conditions. The model was tuned with seven batch and sequencing batch reactor (SBR) laboratory experiments. It simulates the reactor performance (COD removal rate 480-780 mgCOD/L/d, yield 0.40– 0.65 mgCODx/mgCODs) and the relative PNSB abundance (10-60%) under different operational conditions (light, dissolved oxygen, surface area) with a relative error around ± 20%. This research proposes an one-at-the-time sensitivity analysis, analysing the impact of those variables (TSS, SRT, COD, light, and biotic competition) which could play an important role in real-case scenarios. From this analysis, it emerged that influent suspended solids (TSS>250 mgTSS/L), hourly variations of the natural light cycle intensity and drops of the available soluble substrate (COD<1000mgCOD/L) could strongly disturb the abundance of PNSB in the system (from 48% to 10%). An extension of the sludge retention time, from 2 to 5 days, was observed to favour the relative abundance of PNSB (from 48% to 60%) and to increase the TSS productivity (from 235 to 400 mgTSS/L/d). The model was intended as a first attempt to simulate the nutrient removal and the PNSB dynamic in a raceway reactor. It has the flexibility to study the impact of the crucial parameters evidenced from the literature review (light, oxygen, carbon source). However, several implementations will be needed in the future, mainly focusing on anoxic chemoheterotrophic growth and hydrolysis.