Selective sludge removal in a segregated aerobic granular biomass system as a strategy to control PAO-GAO competition at high temperatures

An aerobic granular sludge (AGS) reactor was run for 280 days to study the competition between Phosphate and Glycogen Accumulating Organisms (PAOs and GAOs) at high temperatures. Numerous researches have proven that in suspended sludge systems PAOs are outcompeted by GAOs at higher temperatures. In the following study a reactor was operated at 30 °C in which the P-removal efficiency declined from 79% to 32% after 69 days of operation when biomass removal for sludge retention time (SRT) control was established by effluent withdrawal. In a second attempt at 24 °C, efficiency of P-removal remained on average at 71 ± 5% for 76 days. Samples taken from different depths of the sludge bed analysed using Fluorescent in situ hybridization (FISH) microscopy techniques revealed a distinctive microbial community structure: bottom granules contained considerably more Accumulibacter (PAOs) compared to top granules that were dominated by Competibacter (GAOs). In a third phase the SRT was controlled by discharging biomass exclusively from the top of the sludge bed. The application of this method increased the P-removal efficiency up to 100% for 88 days at 30 °C. Granules selected near the bottom of the sludge bed increased in volume, density and overall ash content; resulting in significantly higher settling velocities. With the removal of exclusively bottom biomass in phase four, P-removal efficiency decreased to 36% within 3 weeks. This study shows that biomass segregation in aerobic granular sludge systems offers an extra possibility to influence microbial competition in order to obtain a desired population.     

Optimized aeration strategies for nitrogen and phosphorus removal with aerobic granular sludge

Biological wastewater treatment by aerobic granular sludge biofilms offers the possibility to combine carbon (COD), nitrogen (N) and phosphorus (P) removal in a single reactor. Since denitrification can be affected by suboptimal dissolved oxygen concentrations (DO) and limited availability of COD, different aeration strategies and COD loads were tested to improve N- and P-removal in granular sludge systems. Aeration strategies promoting alternating nitrification and denitrification (AND) were studied to improve reactor efficiencies in comparison with more classical simultaneous nitrification–denitrification (SND) strategies. With nutrient loading rates of 1.6 gCOD L⁻¹ d⁻¹, 0.2 gN L⁻¹ d⁻¹, and 0.08 gP L⁻¹ d⁻¹, and SND aeration strategies, N-removal was limited to 62.3 ± 3.4%. Higher COD loads markedly improved N-removal showing that denitrification was limited by COD. AND strategies were more efficient than SND strategies. Alternating high and low DO phases during the aeration phase increased N-removal to 71.2 ± 5.6% with a COD loading rate of 1.6 gCOD L⁻¹ d⁻¹. Periods of low DO were presumably favorable to denitrifying P-removal saving COD necessary for heterotrophic N-removal. Intermittent aeration with anoxic periods without mixing between the aeration pulses was even more favorable to N-removal, resulting in 78.3 ± 2.9% N-removal with the lowest COD loading rate tested. P-removal was under all tested conditions between 88 and 98%, and was negatively correlated with the concentration of nitrite and nitrate in the effluent (r = −0.74, p < 0.01). With low COD loading rates, important emissions of undesired N₂O gas were observed and a total of 7–9% of N left the reactor as N₂O. However, N₂O emissions significantly decreased with higher COD loads under AND conditions.     

Identification of trigger factors selecting for polyphosphate- and glycogen-accumulating organisms in aerobic granular sludge sequencing batch reactors

Nutrient removal performances of sequencing batch reactors using granular sludge for intensified biological wastewater treatment rely on optimal underlying microbial selection. Trigger factors of bacterial selection and nutrient removal were investigated in these novel biofilm systems with specific emphasis on polyphosphate- (PAO) and glycogen-accumulating organisms (GAO) mainly affiliated with Accumulibacter and Competibacter, respectively. In a first dynamic reactor operated with stepwise changes in concentration and ratio of acetate and propionate (Ac/Pr) under anaerobic feeding and aerobic starvation conditions and without wasting sludge periodically, propionate favorably selected for Accumulibacter (35% relative abundance) and stable production of granular biomass. A Plackett-Burman multifactorial experimental design was then used to screen in eight runs of 50 days at stable sludge retention time of 15 days for the main effects of COD concentration, Ac/Pr ratio, COD/P ratio, pH, temperature, and redox conditions during starvation. At 95% confidence level, pH was mainly triggering direct Accumulibacter selection and nutrient removal. The overall PAO/GAO competition in granular sludge was statistically equally impacted by pH, temperature, and redox factors. High Accumulibacter abundances (30–47%), PAO/GAO ratios (2.8–8.4), and phosphorus removal (80–100%) were selected by slightly alkaline (pH > 7.3) and lower mesophilic (<20 °C) conditions, and under full aeration during fixed 2-h starvation. Nitrogen removal by nitrification and denitrification (84–97%) was positively correlated to pH and temperature. In addition to alkalinity, non-limited organic conditions, 3-carbon propionate substrate, sludge age control, and phase length adaptation under alternating aerobic-anoxic conditions during starvation can lead to efficient nutrient-removing granular sludge biofilm systems.     

Formation of aerobic granules and conversion processes in an aerobic granular sludge reactor at moderate and low temperatures

Temperature changes can influence biological processes considerably. To investigate the effect of temperature changes on the conversion processes and the stability of aerobic granular sludge, an aerobic granular sludge sequencing batch reactor (GSBR) was exposed to short-term and long-term temperature changes. Start-up at 8 degrees C resulted in irregular granules that aggregated as soon as aeration was stopped, which caused severe biomass washout and instable operation. The presence of COD during the aerobic phase is considered to be the major reason for this granule instability. Start-up at 20 degrees C and lowering the temperature to 15 degrees C and 8 degrees C did not have any effect on granule stability and biomass could be easily retained in the system. The temperature dependency of nitrification was lower for aerobic granules than usually found for activated sludge. Due to decreased activity in the outer layers of granules at lower temperatures, the oxygen penetration depth could increase, which resulted in a larger aerobic biomass volume, compensating the decreased activity of individual organisms. Consequently the denitrifying capacity of the granules decreased at reduced temperatures, resulting in an overall poorer nitrogen removal capacity. The overall conclusion that can be drawn from the experiments at low temperatures is that start-up in practice should take place preferentially during warm summer periods, while decreased temperatures during winter periods should not be a problem for granule stability and COD and phosphate removal in a granular sludge system. Nitrogen removal efficiencies should be optimized by changes in reactor operation or cycle time during this season.     

Alternating anoxic feast/aerobic famine condition for improving granular sludge formation in sequencing batch airlift reactor at reduced aeration rate

In this study the influence of a pre-anoxic feast period on granular sludge formation in a sequencing batch airlift reactor is evaluated. Whereas a purely aerobic SBR was operated as a reference (reactor R2), another reactor (R1) was run with a reduced aeration rate and an alternating anoxic-aerobic cycle reinforced by nitrate feeding. The presence of pre-anoxic phase clearly improved the densification of aggregates and allowed granular sludge formation at reduced air flow rate (superficial air velocity (SAV) = 0.63 cm s⁻¹). A low sludge volume index (SVI₃₀ = 45 mL g⁻¹) and a high MLSS concentration (9–10 g L⁻¹) were obtained in the anoxic/aerobic system compared to more conventional results for the aerobic reactor. A granular sludge was observed in the anoxic/aerobic system whilst only flocs were observed in the aerobic reference even when operated at a high aeration rate (SAV = 2.83 cm s⁻¹). Nitrification was maintained efficiently in the anoxic/aerobic system even when organic loading rate (OLR) was increased up to 2.8 kg COD m⁻³ d⁻¹. In the contrary nitrification was unstable in the aerobic system and dropped at high OLR due to competition between autotrophic and heterotrophic growth. The presence of a pre-anoxic period positively affected granulation process via different mechanisms: enhancing heterotrophic growth/storage deeper in the internal anoxic layer of granule, reducing the competition between autotrophic and heterotrophic growth. These processes help to develop dense granular sludge at a moderate aeration rate. This tends to confirm that oxygen transfer is the most limiting factor for granulation at reduced aeration. Hence the use of an alternative electron acceptor (nitrate or nitrite) should be encouraged during feast period for reducing energy demand of the granular sludge process.     

Nitrogen and phosphorus removal from an abattoir wastewater in a SBR with aerobic granular sludge

The formation and performance of granular sludge was studied in an 8 l sequencing batch reactor (SBR) treating an abattoir (slaughterhouse) wastewater. Influent concentrations averaged 1520 mg l⁻¹ volatile suspended solids (VSS), 7685 mg l⁻¹ Chemical oxygen demand (COD), 1057 mg l⁻¹ total kjeldahl nitrogen (TKN), 217 mg l⁻¹ total P. The COD loading was 2.6 kg m⁻³ d⁻¹. The SBR was seeded with flocculating sludge from a SBR with an 1 h settle time, but granules developed within 4 days by reducing the settle time to 2 min. The SBR cycle also had 120 min mixed (anaerobic) fill, 220 min aerated react, and 18 min draw/idle. The granules had a mean diameter of 1.7 mm, a specific gravity of 1.035, a density of 62 g VSS l⁻¹, a zone settling velocity (ZSV) of 51 m h⁻¹, and a sludge volume index (SVI) of 22 ml g⁻¹. Without optimizing process conditions, removal of COD and P were over 98%, and removal of N and VSS were over 97%. Nitrification and denitrification occurred simultaneously during react. The results indicate that conventional SBRs treating wastewaters with flocculating sludge can be converted to granular SBRs by reducing the settle time.     

Characterization of alginate-like exopolysaccharides isolated from aerobic granular sludge in pilot-plant

To understand functional gel-forming exopolysaccharides in aerobic granular sludge, alginate-like exopolysaccharides were specifically extracted from aerobic granular sludge cultivated in a pilot plant treating municipal sewage. The exopolysaccharides were identified by the FAO/WHO alginate identification tests, characterized by biochemical assays, gelation with Ca²⁺, blocks fractionation, spectroscopic analysis as UV-visible, FT-IR and MALDI-TOF MS, and electrophoresis. The yield of extractable alginate-like exopolysaccharides was reached 160 ± 4 mg/g (VSS ratio). They resembled seaweed alginate in UV-visible and MALDI-TOF MS spectra, and distinguished from it in the reactions with acid ferric sulfate, phenol-sulfuric acid and Coomassie brilliant blue G250. Characterized by their high percentage of poly guluronic acid blocks (69.07 ± 8.95%), the isolated exopolysaccharides were capable to form rigid, non-deformable gels in CaCl₂. They were one of the dominant exopolysaccharides in aerobic granular sludge. We suggest that polymers play a significant role in providing aerobic granular sludge a highly hydrophobic, compact, strong and elastic structure.     

The contribution of exopolysaccharides induced struvites accumulation to ammonium adsorption in aerobic granular sludge

Aerobic granular sludge from a lab-scale reactor with simultaneous nitrification/denitrification and enhanced biological phosphorus removal processes exhibited significant amount of ammonium adsorption (1.5 mg NH₄⁺-N/g TSS at an ammonium concentration of 30 mg N/L). Potassium release accompanied ammonium adsorption, indicating an ion exchange process. The existence of potassium magnesium phosphate (K-struvite) as one of potassium sources in the granular sludge was studied by X-ray diffraction analysis (XRD). Artificially prepared K-struvite was indeed shown to adsorb ammonium. Alginate-like exopolysaccharides were isolated and their inducement for struvite formation was investigated as well. Potassium magnesium phosphate proved to be a major factor for ammonium adsorption on the granular sludge. Struvites (potassium/ammonium magnesium phosphate) accumulate in aerobic granular sludge due to inducing of precipitation by alginate-like exopolysaccharides.     

Reducing the startup time of aerobic granular sludge reactors through seeding floccular sludge with crushed aerobic granules

One of the main challenging issues for the aerobic granular sludge technology is the long startup time when dealing with real wastewaters. This study presents a novel strategy to reduce the time required for granulation while ensuring a high level of nutrient removal. This new approach consists of seeding the reactor with a mixture of crushed aerobic granules and floccular sludge. The effectiveness of the strategy was demonstrated using abattoir wastewater, containing nitrogen and phosphorus at approximately 250 mgN/L and 30 mgP/L, respectively. Seven different mixtures of crushed granules and floccular sludge at granular sludge fractions (w/w in dry mass) of 0%, 5%, 10%, 15%, 25%, 30% and 50% were used to start eight granulation processes. The granulation time (defined as the time when the 10th percentile bacterial aggregate size is larger than 200 μm) displayed a strong dependency on the fraction of granular sludge. The shortest granulation time of 18 days was obtained with 50% crushed granules, in comparison with 133 days with 5% crushed granules. Full granulation was not achieved in the two trials without seeding with crushed granules. In contrast to the 100% floccular sludge cases, where a substantial loss of biomass occurred during granulation, the biomass concentration in all other trails did not decrease during granulation. This allowed that good nitrogen removal was maintained in all the reactors during the granulation process. However, enhanced biological phosphorus removal was achieved in only one of the eight trials. This was likely due to the temporary accumulation of nitrite, a strong inhibitor of polyphosphate accumulating organisms.     

Behavior of polymeric substrates in an aerobic granular sludge system

Particulate and slowly biodegradable substrates form an important fraction of industrial wastewater and sewage. To study the influence of suspended solids and colloidal substrate on the morphology and performance of aerobic granular sludge, suspended and soluble starch was used as a model substrate. Degradation was studied using microscopy, micro-electrode measurements, batch experiments and long term laboratory scale reactor operation. Starch was removed by adsorption at the granule surface, followed by hydrolysis and consumption of the hydrolyzed products. Aerobic granules could be maintained on starch as sole influent carbon source, but their structure was filamentous and irregular. It is hypothesized that this is related to the low starch hydrolysis rates, leading to available substrate during the aeration period (extended feast period) and resulting in increased substrate gradients over the granules. The latter induces a less uniform granule development. Starch adsorbed and was consumed at the granule surface instead of being accumulated inside the granules as occurs for soluble substrates. Therefore the simultaneous denitrification efficiencies remained low. Moreover, many protozoa and metazoans were observed in laboratory reactors as well as in pilot- and full-scale Nereda^® reactors, indicating an important role in the removal of suspended solids too.     

2-fluorophenol degradation by aerobic granular sludge in a sequencing batch reactor

Aerobic granular sludge is extremely promising for the treatment of effluents containing toxic compounds, and it can economically compete with conventional activated sludge systems. A laboratory scale granular sequencing batch reactor (SBR) was established and operated during 444 days for the treatment of an aqueous stream containing a toxic compound, 2-fluorophenol (2-FP), in successive phases. Initially during ca. 3 months, the SBR was intermittently fed with 0.22 mM of 2-FP added to an acetate containing medium. No biodegradation of the target compound was observed. Bioaugmentation with a specialized bacterial strain able to degrade 2-FP was subsequently performed. The reactor was thereafter continuously fed with 0.22 and 0.44 mM of 2-FP and with 5.9 mM of acetate (used as co-substrate), for 15 months. Full degradation of the compound was reached with a stoichiometric fluoride release. The 2-FP degrading strain was successfully retained by aerobic granules, as shown through the recovering of the strain from the granular sludge at the end of the experiment. Overall, the granular SBR has shown to be robust, exhibiting a high performance after bioaugmentation with the 2-FP degrading strain. This study corroborates the fact that bioaugmentation is often needed in cases where biodegradation of highly recalcitrant compounds is targeted.     

Nitrogen removal from on-site treated anaerobic effluents using intermittently aerated moving bed biofilm reactors at low temperature

On-site post-treatment of anaerobically pre-treated dairy parlour wastewater (DPWWe; 10 degrees C) and mixture of kitchen waste and black water (BWKWe; 20 degrees C) was studied in moving bed biofilm reactors (MBBR). The focus was on removal of nitrogen and of residual chemical oxygen demand (COD). Moreover, the effect of intermittent aeration and continuous vs. sequencing batch operation was studied. All MBBRs removed 50-60% of nitrogen and 40-70% of total COD (CODt). Complete nitrification was achieved, but denitrification was restricted by lack of carbon. Nitrogen removal was achieved in a single reactor by applying intermittent aeration. Continuous and sequencing batch operation provided similar nitrogen and COD removal, wherefore simpler continuous feeding may be preferred for on-site applications. Combination of pre-treating upflow anaerobic sludge blanket (UASB) -septic tank and MBBR removed over 92% of CODt, 99% of biological oxygen demand (BOD7), and 65-70% of nitrogen.     

Evaluating the solid retention time of bacteria in flocculent and granular sludge

The specific solid retention time for different bacteria within flocculent and granular sludge was determined. Samples were collected from reactor and effluent sludge and the number of a specific bacterial group was evaluated in respect to the total bacterial community with quantitative polymerase chain reaction (qPCR). The ratio of the relative presence of a specific bacterial group in the reactor sludge and wasted sludge was established to observe if preferential wash-out occurred. From the data also the solid retention time for different microbial groups can be estimated. Using this tool, we were able to show that the SRT of populations found on the exterior of granules is slightly lower than the SRT for population in the interior. Archaea were not found in the flocculent system but were present in small amounts within the granular system. It was further observed that protozoa were grazing on the bacterial community within the system indicating that they have the potential to shorten the specific SRT of bacteria.     

Effect of heat recovery from raw wastewater on nitrification and nitrogen removal in activated sludge plants

By recovery of heat from the raw wastewater in the sewer system, the influent temperature of a wastewater treatment plant (WWTP) is reduced. This can have a negative effect on nitrification in the WWTP, since this process strongly depends on temperature. The analysis of the temperature regime in the WWTP of Zurich, Switzerland, revealed that in the cold season, the effluent temperature is about 0.7 degrees C higher than the influent temperature and that nitrification is not affected by a decrease of the influent wastewater temperature lasting for a couple of hours only, but is significantly affected by a longer lasting temperature decrease. Three diagrams were developed with a steady-state model, from which the consequences of a permanent temperature decrease on the nitrification safety factor, aerobic sludge retention time and total nitrogen removal can be evaluated. Using simulations with a dynamic model, calibrated for the Zurich WWTP, a quantitative relationship between the wastewater temperature and the ammonium effluent concentration was established. This relationship can, in combination with measured effluent concentrations of an existing WWTP, be used to predict the increase of the ammonium effluent concentration in this plant resulting from a permanent decrease of the wastewater influent temperature.     

Kinetic modelling of nitrogen and organics removal in vertical and horizontal flow wetlands

This paper provides a comparative evaluation of the kinetic models that were developed to describe the biodegradation of nitrogen and organics removal in wetland systems. Reaction kinetics that were considered in the model development included first order kinetics, Monod and multiple Monod kinetics; these kinetics were combined with continuous-stirred tank reactor (CSTR) or plug flow pattern to produce equations to link inlet and outlet concentrations of each key pollutants across a single wetland. Using three statistical parameters, a critical evaluation of five potential models was made for vertical and horizontal flow wetlands. The results recommended the models that were developed based on Monod models, for predicting the removal of nitrogen and organics in a vertical and horizontal flow wetland system. No clear correlation was observed between influent BOD/COD values and kinetic coefficients of BOD₅ in VF and HF wetlands, illustrating that the removal of biodegradable organics was insensitive to the nature of organic matter. Higher effluent COD/TN values coincided with greater denitrification kinetic coefficients, signifying the dependency of denitrification on the availability of COD in VF wetland systems. In contrast, the trend was opposite in HF wetlands, indicating that availability of NO₃-N was the main limiting step for nitrogen removal. Overall, the results suggested the possible application of the developed alternative predictive models, for understanding the complex biodegradation routes of nitrogen and organics removal in VF and HF wetland systems.     

Estimation of nitrogen removal rate in aqueous phase based on delta15N in microorganisms in solid phase

Microbial nitrification and denitrification are important processes for removing nitrogenous compounds in aqueous systems. Nitrogen removal rate estimation is essential for controlling nitrogen removal processes and modeling the nitrogen cycle in ecosystems. The model described the relationship between ammonium removal rate (aqueous phase) and the nitrogen stable isotope ratio (delta15N) of microorganisms (solid phase) when a coupled nitrification-denitrification process occurs and assimilation and advections are maintained in a steady state. An oxidation ditch in a municipal wastewater treatment plant was evaluated for 3 years using the model. The ammonium removal rate was calculated from the data of delta15N of the activated sludge, it correlated significantly with the observed removal rate. The isotope fractionation factor (epsilon) was determined to be -5.5 per thousand by using a nonlinear method. The model and obtained factor value were applicable for standard activated-sludge processes performed in parallel in the oxidation ditch and a river watershed. The model may help illustrate nitrogen behavior in ecosystems.     

Combined anaerobic and aerobic digestion for increased solids reduction and nitrogen removal

A unique sludge digestion system consisting of anaerobic digestion followed by aerobic digestion and then a recycle step where thickened sludge from the aerobic digester was recirculated back to the anaerobic unit was studied to determine the impact on volatile solids (VS) reduction and nitrogen removal. It was found that the combined anaerobic/aerobic/anaerobic (ANA/AER/ANA) system provided 70% VS reduction compared to 50% for conventional mesophilic anaerobic digestion with a 20 day SRT and 62% for combined anaerobic/aerobic (ANA/AER) digestion with a 15 day anaerobic and a 5 day aerobic SRT. Total Kjeldahl nitrogen (TKN) removal for the ANA/AER/ANA system was 70% for sludge wasted from the aerobic unit and 43.7% when wasted from the anaerobic unit. TKN removal was 64.5% for the ANA/AER system.     

Ammonium adsorption in aerobic granular sludge, activated sludge and anammox granules

The ammonium adsorption properties of aerobic granular sludge, activated sludge and anammox granules have been investigated. During operation of a pilot-scale aerobic granular sludge reactor, a positive relation between the influent ammonium concentration and the ammonium adsorbed was observed. Aerobic granular sludge exhibited much higher adsorption capacity compared to activated sludge and anammox granules. At an equilibrium ammonium concentration of 30 mg N/L, adsorption obtained with activated sludge and anammox granules was around 0.2 mg NH₄-N/g VSS, while aerobic granular sludge from lab- and pilot-scale exhibited an adsorption of 1.7 and 0.9 mg NH₄-N/g VSS, respectively. No difference in the ammonium adsorption was observed in lab-scale reactors operated at different temperatures (20 and 30 °C). In a lab-scale reactor fed with saline wastewater, we observed that the amount of ammonium adsorbed considerably decreased when the salt concentration increased. The results indicate that adsorption or better ion exchange of ammonium should be incorporated into models for nitrification/denitrification, certainly when aerobic granular sludge is used.     

Influence of sand layer depth and percolate impounding regime on nitrogen transformation in vertical-flow constructed wetlands treating faecal sludge

Four laboratory-scale units of vertical-flow constructed wetlands (VFCW) were fed once a week with faecal sludge (FS) at a constant solids loading rate (SLR) of 250 kg TS/(m².year) (equivalent to 260-300 g N/(m².week)) for a period of 12 weeks to study: i) the nitrification and denitrification potential of the sand layer of VFCWs and ii) the effect of percolate impounding regime (permanent or batch-impounding) on nitrogen transformation. The TN content of raw FS was characterised by 65% org-N, 34% NH₄-N and 1% NO_x-N. After FS application and a six-day impounding period, 8-13% TN were recovered in the percolate exhibiting the following composition: 70-80% NH₄-N, 25-30% org-N and <1% NO_x-N. A large fraction of the influent organic N (55%) was filtered in the bed and 24-29% of initial NH₄-N were lost due to nitrification and volatilisation. In permanent impounding systems, 8-11% TN were recovered in the percolate versus 13% in batch-operated beds. N loss was increased with sand layer depth (20-40 cm) under permanent impounding regimes.