Abstract: The present PhD research investigated the possibility of catalytic detoxification of poplar wood based and steam exploded lignocellulosic hydrolyzate with different types of laccase enzymes, with special focus on ethanol and lactic acid products at industrially relevant parameters: high final product concentration, high initial substrate loading and integrated processes. The simultaneous saccharification and fermentation (SSF) process was taken as a base case and five types of laccases were thoroughly investigated on their utilization potential. Phenolic removal from the liquid xylose rich fraction (XRF) was higher with fungal laccases (65-90 %) compared to approximately 30 % removal with bacterial laccase. Moreover, the optimal pH of fungal laccases was close to pH 4.5, the optimum for cellulase, while the bacterial laccase worked at basic pH. Integrating laccase treatment and hydrolysis together showed that fungal laccases have negative impact on final sugar concentration, while bacterial laccase had a strong positive effect. Although bacterial laccase removed less phenol and although its optimal conditions are difficult to integrate with hydrolysis, its enhancing effect on cellulase activity makes it a better candidate for application. The presence of the solid fraction (SF) alters the phenolic concentration evolution significantly, thus screening experiments with the liquid fraction alone do not provide sufficient information for the combined process. Magnetic Cross-Linked Enzyme Aggregates (m-CLEAs) immobilization was assessed for bacterial laccase. m-CLEAs decreased phenolic concentration faster at every pH compared to free bacterial laccase; however, the removal was caused by adsorption rather than by enzyme activity. Although the size of m-CLEAs particles are in the µm range, around 90 % of the initial catalyst mass was recycled from a dense (15 % substrate loading) mixture via magnetic separation. The high recycling rate is promising; m-CLEAs immobilization method can have industrial utilization potential. Minimum sugar revenue (MSR) estimations show that currently hardwood based MSR is 70 % more expensive than corn grain based MSR. About 7-10 fold cellulase activity increase will be needed until MSR will be competitive with corn grain MSR. However, m-CLEAs cellulase can already be competitive if the corn prices are in the higher regime of last year’s prices.

Abstract: This study presents novel sulfolane based non-aqueous CO2 capture solvents, as an alternative solution for capturing CO2 from industrial processes. In order to select the most promising amine system, five different amines were tested by monitoring CO2 absorption and desorption processes using the time-base Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) spectroscopy. During absorption experiments, we observed the formation of Monomethyl Carbonate (MMC) in diisopropylamine (DIPA) and 2-amino-2-methyl-1-propanol (AMP) systems, while carbamate was observed as the main product for the other three amine systems tested. In regeneration experiments, the MMC could be desorbed relatively easily from the amine solution at a mild temperature.

Abstract: Lignosulfonate-based bioactive coatings as soil improvers for lawns were developed using laccase as a biocatalyst. Incorporation of glycerol, xylitol and sorbitol as plasticizers considerably reduced the brittleness of the synthesized coatings of marine carbonate granules while thermal enzyme inactivation at 100 degrees C enabled the production of stable coatings. Heat inactivation produced stable coatings with a molecular weight of 2000 kDa and a viscosity of 4.5 x 10(-3) Pas. The desired plasticity for the spray coating of soil improver granules was achieved by the addition of 2.7% of xylitol. Agriculture beneficial microorganisms (four different Bacillus species) were integrated into the coatings. The stable coatings protected the marine calcium carbonate granules, maintained the viability of the microorganisms and showed no toxic effects on the germination and growth of model plants including corn, wheat, salad, and tomato despite a slight delay in germination. Moreover, the coatings reduced the dust formation of soil improvers by 70%. CO2 emission analysis showed potential for the reduction of up to 3.4 kg CO2-eq. kg(-1) product, making it a viable alternative to fossil-based coatings.

Abstract: Water quality standards (WQS) set the legal definition for safe and desirable water. WQS impose regulatory concentration limits to act as a jurisdiction-specific legislative risk-management tool. Despite its importance in shaping a universal definition of safe, clean water, little information exists with respect to (dis)similarity of chemical WQS worldwide. Therefore, this paper compares chemical WQS for drinking and surface water matrices in eight jurisdictions representing a global geographic distribution: Australia, Brazil, Canada, China, the European Union, the region of Flanders in Belgium, the United States of America, and South Africa. The World Health Organization's list is used as a reference for drinking water standards. Sørensen–Dice indices (SDI) showed little qualitative similarity in the compounds that are regulated in drinking water (median SDI = 40%) and surface water (median SDI = 33%), indicating that the heterogeneity within a matrix is substantial at the level of the standard. Quantitative similarity for matching standards was higher than the qualitative per Kendall correlation (median = 0.73 and 0.58 for drinking water and surface water respectively), yet variance observed within standards remained inexplicably high for organic compounds. Variations in WQS were more pronounced for organic compounds. Most differences cannot be easily explained from a toxicological or risk-based point-of-view. Historical development, ease of measurement, and (toxicological) knowledge gaps on the risk of a vast number of organic compounds are theorized to be the drivers. Therefore, this study argues for a more tailored, risk-based approach in which standards incorporated into water safety plans are dynamically set for compounds that are persistent and could pose a risk for human health and/or aquatic ecosystems. Global variations in WQS should therefore not necessarily be avoided but rather globally harmonized with enough flexibility to ensure a global, up-to-date definition of safe and desirable water everywhere.

Abstract: The shape of leaf laminae exhibits considerable diversity and complexity that reflects adaptations to environmental factors such as ambient light and precipitation as well as phyletic legacy. Many leaves appear to be elliptical which may represent a ‘default’ developmental condition. However, whether their geometry truly conforms to the ellipse equation (EE), i.e., (x/a)2 + (y/b)2 = 1, remains conjectural. One alternative is described by the superellipse equation (SE), a generalized version of EE, i.e., |x/a|n +|y/b|n = 1. To test the efficacy of EE versus SE to describe leaf geometry, the leaf shapes of two Michelia species (i.e., M. cavaleriei var. platypetala, and M. maudiae), were investigated using 60 leaves from each species. Analysis shows that the majority of leaves (118 out of 120) had adjusted root-mean-square errors of < 0.05 for the nonlinear fitting of SE to leaf geometry, i.e., the mean absolute deviation from the polar point to leaf marginal points was smaller than 5% of the radius of a hypothesized circle with its area equaling leaf area. The estimates of n for the two species were ˂ 2, indicating that all sampled leaves conformed to SE and not to EE. This study confirms the existence of SE in leaves, linking this to its potential functional advantages, particularly the possible influence of leaf shape on hydraulic conductance.

Abstract: Exposure to volatile organic compounds (VOCs) remains a major public health concern. Indoor VOC concentrations typically far exceed outdoor levels due to a variety of emission sources and the stringent insulation measures that are imposed today. Many attempts have been made to use photocatalysis for indoor air purification. In an ideal situation, photocatalysis is capable of complete mineralization of VOCs to H2O and CO2, without any byproduct formation. Moreover, the process can take place at standard atmospheric conditions, i.e. ambient temperature and atmospheric pressure. However, successful exploitation is still impeded due to low conversion efficiency, significant pressure loss (and hence a high energy consumption) and byproduct formation. In the first part of this thesis an attempt was made to tackles these problems by designing a novel type of photocatalytic (PCO) reactor. The PCO device consists of a cylindrical vessel filled with TiO2-coated glass tubes and equipped with UV fluorescence lamps. It was investigated in terms of fluid dynamics, coating properties, UV-light distribution and photocatalytic activity. Experimental data was later used to develop and calibrate a Multiphysics model. The model proved to be a useful tool for designing and upscaling the PCO reactor. Consequently, an optimized prototype reactor was constructed and tested according the CEN-EN-16846-1 standard for VOC removal. Although the prototype showed promising results for lab-scale conditions, it struggled with byproduct formation when purifying ppb-level VOCs. In the second part of this thesis, activated carbon adsorption was investigated in order to combine it with photocatalysis. Activated carbon fiber was opted for its fast kinetics, high adsorption capacity and thermo-electrical regeneration. The filter was studied in detail regarding the adsorption of polar and apolar VOCs at indoor air concentration levels and regeneration capabilities. Experimental data was used to develop a Multiphysics model for activated carbon adsorption as well. Consequently, a novel type of ACF filter was developed using the Multiphysics model, which was equipped with electrodes in the tips of the pleats for effective thermal regeneration. In the last part, the combination of both ACF and PCO was studied using a realistic case study. Based on the Multiphysics model, the feasibility of a so-called hybrid air purification device could be investigated. The Multiphysics model shows promising results for this hybrid PCO-ACF system and hence, a demo setup was constructed for future research.

Abstract: Due to the high water- and nutrient-use efficiency, hydroponic cultivation is increasingly vital in progressing to environment-friendly food production. To further alleviate the environmental impacts of synthetic fertilizer production, the use of recovered nutrients should be encouraged in horticulture and agriculture at large. Solid organic fertilizers can largely contribute to this, yet their physical and chemical nature impedes application in hydroponics. This study proposes a bioreactor for mineralization and nitrification followed by a supplementation step for limiting macronutrients to produce nitrate-based solutions from solid fertilizers, here based on a novel microbial fertilizer. Batch tests showed that aerobic conversions at 35 °C could realize a nitrate (NO₃−-N) production efficiency above 90% and a maximum rate of 59 mg N L−1 d−1. In the subsequent bioreactor test, nitrate production efficiencies were lower (44–51%), yet rates were higher (175–212 mg N L−1 d−1). Calcium and magnesium hydroxide were compared to control the bioreactor pH at 6.0 ± 0.2, while also providing macronutrients for plant production. A mass balance estimation to mimic the Hoagland nutrient solution showed that 92.7% of the NO₃−-N in the Ca(OH)₂ scenario could be organically sourced, while this was only 37.4% in the Mg(OH)₂ scenario. Besides, carbon dioxide (CO₂) generated in the bioreactor can be used for greenhouse carbon fertilization to save operational expenditure (OPEX). An estimation of the total OPEX showed that the production of a nutrient solution from solid organic fertilizers can be cost competitive compared to using commercially available liquid inorganic fertilizer solutions.

Abstract: In the biogeochemical flows on Earth, the reactive nitrogen (Nr) level has three times surpassed the safe boundary. The severe transgression of this boundary goes against sustainable planetary development. The modern food production process excessively relies on synthetic Nr fertilizers from the Haber– Bosch process. However, the massive loss of valuable nitrogen resources (i.e., 78-89%) from agriculture has been causing severe nitrogen cascade. Besides, the domestic wastewater in some local areas is discharged without proper treatment, making it a nonnegligible source of Nr pollution for local water bodies. Anthropogenic activities keep pumping out Nr pollution via point-source and non-point-source (NPS) emissions. Compared to the NPS emissions, point sources give visible and identified waste streams. It is vital to intervene the nitrogen cascade from point sources and facilitate humanity back to the safe Nr boundary. The collected and collectible Nr streams from food production, waste management, and recycling secondary raw materials can be used as waste-based fertilizers for agricultural cultivation. Besides the well-investigated recovery of inorganic Nr, organic Nr accounts for a massive Nr proportion on the Earth. Proper handling and treatment make these useful organic fertilizers for soil-based cultivation. However, these organic Nr fertilizers cannot directly apply to fertigation or hydroponic cultivation systems, and further biological conversion via nitrogen mineralization and nitrification to nitrate is essential. Besides the direct Nr cycling, the indirect Nr cycling ‘over the atmosphere’ should also be considered. In this way, the nitrogen cycle can be completed via converting the waste Nr back to nitrogen gas (i.e., Nr removal) and then synthesizing into Nr again. The municipal wastewater treatment plants receive a vast amount of low-strength Nr wastewater (mainly as ammonium) daily. Compared to the conventional nitrification/denitrification process, partial nitritation/anammox (PN/A) is considered a resource- and cost-effective technology for wastewater with a low COD/N ratio. Moreover, the novel autotrophic denitratation/anammox process could be a good Nr removal process for wastewater containing both ammonium and nitrate. This Ph.D. thesis aimed to develop Nr recovery, conversion, and removal bioreactor strategies for different types of waste streams and biomass. Nr recovery was investigated on high-strength Nr waste streams for fertigation or hydroponic applications in Chapters 2 and 3. On the other hand, Nr removal was studied on the medium- to low-strength Nr waste streams in Chapters 4 and 5. In Chapter 2, a novel mineralization and nitrification system was proposed, producing nutrient solutions from solid organic fertilizers for hydroponic systems. Batch tests showed that aerobic incubation at 35°C could realize the NO₃⁻-N production efficiency above 90% from a novel microbial fertilizer. Subsequently, in the stirred tank bioreactor test, NO₃⁻-N production efficiency stabilized in a range of 44-51% under the influent loading rate of 400 mg TN L⁻¹ d⁻¹ at a 5-day HRT. Using Ca(OH)₂ and Mg(OH)₂ as pH control reagents generated the nutrient solutions with different P, Ca, and Mg nutrient levels. After modeling the nutrient balancing process, the proportion of organic-sourced NO₃⁻-N in the Hoagland nutrient solution (HNS) of Ca(OH)₂ scenario was 92.7%, while only 37.4% in the Mg(OH)₂ scenario. Compared to commercial scenarios, the total costs of the organic-sourced HNS can be cost-competitive for hydroponic cultivation. In Chapter 3, the Nr recovery as nitrate (NO₃⁻-N) from diluted human urine (around 670 mg N L⁻¹) was explored in a trickling filter (TF) for the first time. A novel concept of in-situ integrating the TF system into hydroponic systems was proposed as meaningful progress towards sustainable agriculture. The difference between synthetic and real urine in nitrification efficiency was found to be negligible. The full nitrification of alkalinized real urine was realized in the pH-controlled TF by calcium hydroxide (Ca(OH)₂) at around pH 6. The TF could handle different urine collection batches and maintain relatively stable nitrification performance, with NO₃⁻-N production efficiency and rate of 88±3% and 136±4 mg N L⁻¹ d⁻¹, respectively. The optimal HLR to realize this nitrification performance was 2 m³ m⁻² h⁻¹, with energy consumption of 1.8 kWh electricity kg⁻¹ NO₃⁻-N production. Ca(OH)₂, as a cheap base, its triple advantages on urine alkalinization, full nitrification, and macronutrient supplementation were successfully demonstrated in our proposed concept. In Chapter 4, towards more sustainable wastewater treatment, the feasibility of one-stage partial nitritation/anammox (PN/A) was investigated in three parallel packed-bed trickling filters (TFs), with three types of carrier materials of different specific surface areas. Synthetic wastewater containing 100-250 mg NH₄⁺-N L⁻¹ was tested to mimic medium-strength household waste streams after carbon removal. Interestingly, the cheap carrier based on expanded clay achieved similar rates as commercially used plastic carrier materials. The top passive ventilation combined with an optimum hydraulic loading rate of 1.8 m³ m⁻² h⁻¹ could reach approximately 60% total nitrogen (TN) removal at a rate of 300 mg N L⁻¹ d⁻¹. A relatively low NO₃⁻-N production (13%) via PN/A was achieved in TFs. Most of the TN removal took place in the top compartment, where anammox activity was the highest. Energy consumption estimation (0.78 kWh electricity g⁻¹ N removed) suggested that the proposed process could be a suitable low-cost alternative for nitrogen removal. In Chapter 5, coupling sulfur-driven denitratation (SDN) with anammox was proposed to treat the wastewater containing both NO₃⁻-N and NH₄⁺-N, like the secondary effluents of mainstream PN/A processes. To explore the feasibility of sufficient and stable NO₂⁻-N accumulation via SDN in the long term, the effects of pH setpoints, residual NO₃⁻-N level, and biomass-specific NO₃⁻-N loading rate (BSNLR) were investigated. Alternating the pH setpoints between 7.0 and 8.5 could temporarily stimulate the NO₂⁻-N accumulation. Both the residual NO₃⁻-N and BSNLR showed highly positive correlations with the NO₂⁻-N accumulation efficiency. Under the control of pH 8.5, 1.0±0.8 mg NO₃⁻-N L⁻¹ and 150±42 mg NO₃⁻-N g⁻¹ VSS d⁻¹, SDN could produce 6.4±1.0 mg NO₂⁻-N L⁻¹ in the short term. Thiobacillus members may play a crucial role in managing the NO₂⁻-N accumulation, but the reduction of abundance and possible adaptation significantly impaired the efficacy of control strategies in the long run. Overall, novel technologies have been proposed to sustainably convert Nr in waste streams and biomass. The decision for Nr recovery versus removal and synthesis should be based on specific cases with the best environmental, economic, and human-health sustainability. In the future, the Nr management concepts should be further improved to make the nitrogen cycle more sustainable with higher resource use efficiency and less Nr emissions to the environment. Although the thesis is mainly focused on limited types of Nr waste streams, it pointed out the direction of sustainable Nr management and could facilitate the Nr back to the safe boundary in the long run.

Abstract: Energy-autonomous sewage treatment can be achieved if nitrogen (N) removal does not rely on organic carbon (∼chemical oxygen demand, COD), so that a maximum of the COD can be redirected to energy recovery. Shortcut N removal technologies such as partial nitritation/anammox and nitritation/denitritation are therefore essential, enabling carbon- and energy-lean nitrogen removal. In this study, a novel three-reactor pilot design was tested and consisted of a denitrification, an intermittent aeration, and an anammox tank. A vibrating sieve was added for differential sludge retention time (SRT) control. The 13 m3 pilot was operated on pre-treated sewage (A-stage effluent) at 12–24 °C. Selective suppression of unwanted nitrite-oxidizing bacteria over aerobic ammonium-oxidizing bacteria was achieved with strict floccular SRT management combined with innovative aeration control, resulting in a minimal nitrate production ratio of 17 ± 10%. Additionally, anoxic ammonium-oxidizing bacteria (AnAOB) activity could be maintained in the reactor for at least 150 days because of long granular SRT management and the anammox tank. Consequently, the COD/N removal ratio of 2.3 ± 0.7 demonstrated shortcut N removal almost three times lower than the currently applied nitrification/denitrification technology. The effluent total N concentrations of 17 ± 3 mg TN per L (at 21 ± 1 °C) and 17 ± 6 mg TN per L (at 15 ± 1 °C) were however too high for application at the sewage treatment plant Nieuwveer (Breda, The Netherlands). Corresponding N removal efficiencies were 52 ± 12% and 37 ± 21%, respectively. Further development should focus on redirecting more nitrite to AnAOB in the B-stage, exploring effluent-polishing options, or cycling nitrate for increased A-stage denitrification.

Abstract: This thesis provides three potential ways to enhance the nitrogen removal efficiency of mainstream partial nitritation/anammox (PN/A), a key technology to enable energy-positive sewage treatment. In Chapter 1, the typical technologies to promote nitrogen removal efficiency are summarized. In Chapters 2 and 3, the concept ‘winter bioaugmentation with stored summer surplus sludge’ is proposed. Applying that, a cost-effective sludge preservation strategy is required. Preserving PN/A biomass without cooling and redox adjustment proved to be the cost-effective strategy. The reactivation of these stored sludges was also tested in low-temperature systems (15 and 10℃). Respectively 56% and 41% of granules activity compared to pre-storage activity (after Arrhenius-based temperature correction) could be recovered within a month (41% and 32% for flocs activity). In the end, the stored AnAOB bioaugmentation was successfully validated in the lab (20℃). In Chapter 4, a return-sludge nursery concept, applying the sidestream nitritation and blending the resulting effluent with mainstream effluent to achieve an intermediate temperature and nitrogen concentrations, is proposed. That led to a 33 – 36% increase in nitrogen removal efficiency. Arrhenius’ expectations (10 ℃ higher temperature, θ = 1.09) could only explain 49-51% of the activity increase in the nursery reactor, pointing to the role of other factors, e.g., the ~400% elevated electrical conductivity (15-16%), the 56-335% higher effluent nitrogen concentrations (12-14%), and the synergy and unknown factors (20-23%). Thus, the return-sludge biostimulation approach could also enhance nitrogen efficiency in the mainstream. In Chapter 5, the N2O emissions, linked to three typical nitrite-oxidizing bacteria (NOB) suppression strategies (low dissolved oxygen (DO) level, free ammonia (FA), and free nitrous acids (FNA) treatments) were tested in a biofilm system. A low emerged DO level (~0.60 mg O2 L-1) was effective to suppress NOB activity and decrease N2O emissions, but NOB adaptation gradually appeared after 200 days. Further NOB inhibition was successfully achieved by periodical (3 hours per week) FA (~30 mg NH3-N L-1) or FNA (~3 mg HNO2-N L-1) treatments. The FA treatment promoted N2O production, while the FNA treatment had no effect. Thus, PN/A systems should be operated at relatively low DO levels with periodical FNA treatment. In Chapter 6, the major findings proposed and the main conclusions drawn in this thesis are outlined. Beyond that, the possible design of a mainstream PN/A configuration that combined all described three technologies is demonstrated. Overall, the novel insights from this thesis potential to improve nitrogen removal efficiency in the mainstream.

Abstract: The increasing availability of real-time in situ measurements and remote sensing observations have the potential to contribute to the optimization of water resources management. Global challenges such as climate change, intensive agriculture and urbanization put a high pressure on our water resources. Due to recent innovations in measuring both water quantity and quality, river systems can now be monitored in real time at an unprecedented spatial and temporal scale. To interpret the sensor measurements and remote sensing observations additional data for example on: the location of the measurement, upstream and downstream catchment characteristics, horizontal ellipsis are required. In this paper, we present a data management system to support flow-path related functionality for decision making and prediction modelling. Adding meta data sets and facilitating (near) real-time processing of sensor data questions are key concepts for the systems. The potential of the database framework for hydrological applications is demonstrated using different applications for the river system of Flanders. In one, the database framework is used to simulate the daily discharge for each segment within a catchment using a simple data-driven approach. The presented system is useful for numerous applications including pollution tracking, alerting and inter-sensor validation in river systems, or related networks.

Abstract: Bioaugmentation with summer harvested sludge during winter could compensate for bacterial activity loss but requires that sludge activity can be restored after storage. This study assesses the effect of temperature and redox adjustment during the storage over 180 days of partial nitritation/anammox (PN/A) granular resp. floccular sludge from potato processing resp. sludge reject water treatment. Anoxic storage conditions (in the presence of nitrite or nitrate and the absence of oxygen) resulted in a loss of 80-100% of the anammox bacteria (AnAOB) activity capacity at 20 degrees C and 4 degrees C, while anaerobic conditions (without oxygen, nitrite, and nitrate) lost only 45-63%. Storage at 20 degrees C was more cost-effective compared to 4 degrees C, and this was confirmed in the sludge reactivation experiment (20 CC). Furthermore, AnAOB activity correlated negatively with the electrical conductivity level (R-2 > 0.85, p < 0.05), so strong salinity increases should be avoided. No significant differences were found in the activity capacity of aerobic ammonia-oxidizing bacteria (AerAOB) under different storage conditions (p > 0.1). The relative abundance of dominant AnAOB (Candidatus Brocadia) and AerAOB genera (Nitrosomonas) remained constant in both sludges. In conclusion, preserving PN/A biomass without cooling and nitrite or nitrate addition proved to be a cost-effective strategy. (C) 2021 Elsevier B.V. All rights reserved.

Abstract: It is a generally accepted idea that the particle size distribution strongly affects the optical spectra of colloidal plasmonic nanoparticles. It is often quoted as one of the main reasons while explaining the mismatch between the theoretical and experimental optical spectra of such nanoparticles. In this work, these aspects are critically analyzed by means of a bottom up statistical approach that considers variables such as mean, standard deviation and skewness of the nanoparticle size distribution independently from one another. By assuming normal and log-normal distributions of the particle size, the effect of the statistical parameters on the Mie analytical optical spectra of colloidal nanoparticles was studied. The effect of morphology was also studied numerically in order to understand to what extent it can play a role. It is our finding that the particle polydispersity, skewness and surface morphology in fact only weakly impact the optical spectra. While, the selection of suitable optical constants with regard to the crystallinity of the nanoparticles is a far more influential factor for correctly predicting both the plasmon band position and the plasmon bandwidth in theoretical simulations of the optical spectra. It is shown that the mean particle size can be correctly estimated directly from the plasmon band position, as it is the mean that determines the resonance wavelength. The standard deviation can on the other hand be estimated from the intensity distribution data obtained from dynamic light scattering experiments. The results reported herein clear the ambiguity around particle size distribution and optical response of colloidal plasmonic nanoparticles.

Abstract: Purple non-sulfur bacteria (PNSB) form an interesting group of microbes for resource recovery from wastewater. Solid/liquid separation is key for biomass and value-added products recovery, yet insights into PNSB aggregation are thus far limited. This study explored the effects of organic loading rate (OLR), hydraulic retention time (HRT) and water composition on the aggregation of Rhodobacter capsulatus in an anaerobic upflow photobioreactor. Between 2.0 and 14.6 gCOD/(L.d), the optimal OLR for aggregation was 6.1 gCOD/(L.d), resulting in a sedimentation flux of 5.9 kgTSS/(m2.h). With HRT tested between 0.04 and 1.00 d, disaggregation occurred at the relatively long HRT (1 d), possibly due to accumulation of thus far unidentified heat-labile metabolites. Chemical oxygen demand (COD) to nitrogen ratios (6–35 gCOD/gN) and the nitrogen source (ammonium vs. glutamate) also impacted aggregation, highlighting the importance of the type of wastewater and its pre-treatment. These novel insights to improve purple biomass separation pave the way for cost-efficient PNSB applications.