Developing Anammox for mainstream municipal wastewater treatment

Conventional wastewater treatment plants (WWTPs), like activated sludge systems, are energy demanding requiring a large electrical energy supply (e.g. 25 kWh PE-1 year-1) which, especially during peak-load periods, may account for an important quote of the grid installed power of the surrounding area. Only across the EU, there are 16000 WWTPs that consume around 10000 GWh year-1 of electricity. Furthermore, the volume of wastewater treated in WWTPs in the EU is increasing with a rate of around 7% each year. Besides the related financial costs, this energy consumption creates an additional environmental burden. Considering that energy in Europe is mainly produced from the burning of fossil fuels, it has been calculated that energy consumption from WWTPs creates emissions of more than 27 Mtonnes year-1 of CO2 in the EU. Concerns about greenhouse gas emissions on a global level and cost issues on a microeconomic level have become major driving forces towards a more efficient usage of energy in wastewater treatment. In conventional wastewater treatment about 50% of the energy input is consumed in the aeration systems in order to remove organic matter (Chemical Oxygen Demand, COD) while about 25% is consumed in the nitrogen removal process (nitrification/denitrification) (Siegrist et al., 2008). Autotrophic nitrogen removal by anammox bacteria is to date the most efficient and environmentally friendly process for the treatment of ammonium in wastewaters and its application can save up to 60% of the energy input needed for nitrification. Application of anammox to municipal sewage treatment appears as a prerequisite to allow treatment scenarios for wastewater treatment plants with a net energy production (Kartal et al., 2010a). In a treatment scheme where nitrogen is removed via an autotrophic metabolic pathway such as partial nitritation/anammox (PN/anammox),the COD load, which is conventionally oxidized to C02 partly with oxygen and partly with nitrate (N03-) in the denitrification process, can be used to generate energy in the form of methane-rich biogas via the anaerobic digestion process. Whilst the application of anammox related technologies in the side-stream is at present state of the art, the feasibility of this energy-efficient process in main-stream conditions is still under investigation. Lower and variable operating temperatures and ammonium concentrations, together with a demand for high and stable nitrogen removal efficiency, represent the main challenges to overcome for this appealing new frontier of the waste water treatment field. The research described in this thesis aimed at investigating the physiology and kinetic properties of anammox bacteria and their interaction with other microbial communities under municipal wastewater conditions with the ultimate scope of elucidating the boundary conditions for the application of the anammox-based process (PN/anammox) in the treatment of municipal sewage. This fundamental knowledge allow to design and successfully implement at lab- and pilot-scale the completely autotrophic nitrogen removal process for the treatment of municipal sewage. This thesis comprises therefore both fundamental and applied research which main results and achievements are briefly illustrated in this summary. Although anammox related technologies are currently widely applied for nitrogen removal from sewage sludge digester rejection water, many aspects of the anammox process like the kinetic characteristics and the reaction stoichiometry are still under investigation. Parameter values reported in literature are often influenced by mass transfer limitation or by the presence of inactive cells and a significant side population. In Chapter 2 a membrane bioreactor (MBR) based method for growing a highly enriched anammox microbial community is described. The almost pure free­cell suspension of highly active anammox bacteria was used for detailed kinetic and stoichiometric analysis of the anammox process. The yield of biomass production on ammonium uptake was calculated to be 0.071 C-mol N-mol-1, value that was then experimentally confirmed in Chapter 3. The elemental biomass composition was measured as CH1.740o.31No.20So.01Po.01 (22.1 g C-mol-1). From the yield and the elemental biomass composition the macro-chemical reaction equation was identified and validated by long-term reactor operations. The anammox culture described in Chapter 2 exhibited an unreported high biomass specific maximum growth rate of 0.21 d-1 corresponding to a doubling time of 3.3 days at 30°C. Using an experimental methodology based on imposing dynamic process conditions combined with process modeling and parameter estimation, the intrinsic nitrite half saturation constant was identified to be as low as 35 µg-N L-1. This was confirmed to be a stable value in the tested pH range of 6.8-7.5. Using the same system, in Chapter 3 the stoichiometric and kinetic properties of a suspended anammox enrichment culture were investigated at decreasing solid retention times. This procedure enabled the maximum growth rate (µmax) of the anammox enrichment culture to increase to 0.334 d-1, which is four times higher than previously reported in literature and almost 60% higher than observed in Chapter 2. Even though researchers have speculated about the possibility of higher rates before, these speculations were always based on indirect measurements of the kinetic properties. Herewith Chapter 3 reports the first direct experimental evidence for a significant increase in growth rate of an anammox enrichment culture. Since the biomass yield of the enrichment culture established is largely comparable to previous studies, it can be concluded that the increased growth rate results from an equivalent increase in biomass specific electron transfer capacity. Detailed molecular analysis did not reveal either a shift in dominant anammox strain nor major mutations in the dominant strain, suggesting that the actual reasons for the increase in electron transfer capacity is due to small changes in the metabolic machinery. The dominant strain throughout this experiment was closely related to Candidatus Brocadia Sp.40 (99% similarity). In this study anammox bacteria were cultivated applying a novel selection strategy based on the maximization of the electron transfer capacity demonstrating that maximum growth rate is not an intrinsic process property but that it can be increased significantly when the adequate cultivation conditions are imposed. The anammox enrichment became faster through training showing kinetics comparable with other chemolithoautotrophs and it is thereby concluded that anammox can no longer be regarded as intrinsically slow growing microorganism. Nitrite is one of the main substrates of the anammox metabolism, but it is also an inhibitor. Its negative effect on anammox activity has been reported widely during the past decade. Although the adverse effect is clear, conflicting reports exist on the level at which it occurs and its reversible/irreversible nature. In order to elucidate this important aspect, an in-depth study on nitrite inhibition was performed in which the influence of environmental factors was evaluated (Chapter 4). Anammox activity was measured in anammox granules by continuously monitored standardized manometric batch tests extending the interpretation by evaluation of lag times, maximum conversion rates during the tests and substrates/product conversion ratios. The granules, dominated by anammox organisms belonging to the Brocadia type, where sampled from a single-stage anammox full-scale reactor. The observed 50% activity inhibition for nitrite (IC50) was 0.4 g-N L-1. It was shown that biomass relatively quickly (and totally) recovers from high nitrite concentrations. The recovery after exposure indicates that the adverse effect of nitrite is reversible and thus inhibitory rather than toxic in nature. The effect of the presence of ammonium and oxygen during nitrite exposure has also been evaluated. Similarities between exposures at three different pH values suggest that nitrite rather than nitrous acid is the actual inhibiting compound. Overall the results reported in Chapter 4 further underline that the anammox process can be a stable process not prone to temporarily adverse effects of oxygen and nitrite in the reactors. From our experience and previous observations we speculate that cultivation conditions and status of aggregation influence the inhibitory effect of nitrite and that in several cases where high nitrite is reported as a cause of activity loss, it might well be that activity loss has resulted in the accumulation of high nitrite concentrations rather than causing them. The temperature effect on anammox activity is a crucial aspect that needs to be clarified for the successful implementation of anammox related processes at mainstream conditions. Lower operating temperatures in fact, together with lower ammonium concentrations and the demand for high and stable nitrogen removal efficiency, represent the main challenges to overcome for this appealing new frontier of the waste water treatment field. In Chapter 5 is reported the short-term effect of temperature on the maximum biomass specific activity of anaerobic ammonium oxidizing bacteria as evaluated by means of batch tests. The experiments were performed on anammox biomass sampled from two full-scale reactors and two lab-scale reactors, all characterized by different reactor configurations and operating conditions. The results indicate that in the temperature range of 10-30°C the temperature dependency for the anammox conversion cannot be accurately modeled by one single Arrhenius coefficient (i.e. ?) as typically applied for other biological processes. The temperature effect is increasing at lower temperatures, complicating the implementation of a stable mainstream process in winter conditions. Nevertheless, we observed adaptation of anammox bacteria after long term cultivation at 20 and 10°C indicating that also the history of the sludge impacts the temperature effect. Anammox sludge cultivated in an aerated partial nitritation/anammox process and/or in biofilm seemed to be less influenced by a decrease in temperature then anammox sludge grown under non aerated conditions and/or in suspension. The results reported in Chapter 5 indicate that the temperature effect is stronger for anammox than for ammonium oxidizing bacteria (AOB), suggesting that, in order to maintain overall a good nitrogen removal along daily and seasonal temperature fluctuations, process control to balance the activity of both microbial groups needs to be adaptive to changes in relative rates of the two processes. Implications for modeling and process design are finally discussed. In Chapter 6 the application of the single-stage PN/anammox process at conditions relevant for sewage treatment was investigated in a lab-scale gas-lift sequencing batch reactor with granular sludge operated for more than 500 days. The reactor was operated at temperatures between 20 and 10°C and fed with synthetic autotrophic medium with ammonium (60 and 160 mg-N L-1 as only nitrogen compound at an HRT of 0.23-0.3 d. In the presence of ammonium dissolved oxygen was shown to be an effective control parameter, even at higher level than previously assumed (up to 2.5 mg-O2 L-1, for the suppression of the undesired nitratation process catalyzed by nitrite oxidizing bacteria (NOB). This control strategy guaranteed the effective suppression of the nitratation process both at 20 and 15°C, allowing nitrogen removal rates of 0.44 and 0.40 g-NTot L-1 d-1. Unlike previously reported, these high removal rates were obtained together with optimal nitrogen removal efficiencies of 86 and 73%, respectively, fulfilling a decisive prerequisite for the implementation of the PN/anammox process in the main-stream of WWTPs. Anammox bacteria were shown to grow in the system, with estimated growth rate of 0.017 d-1 at 15°C. Operating conditions influencing N20 emissions were also investigated and resulted in the observation of a positive correlation with the nitrite concentration in the bulk whilst no clear correlation could be noticed between N20 emissions and DO concentration or temperature. Unfortunately prolonged operation at 10°C caused a slow but unrestrainable decrease in anammox activity and process efficiency. Nevertheless, since in general these temperatures (winter conditions) do not extend over long time in moderate climates this is not seen as a limitation for the applications of anammox-based technologies in the mainstream of wastewater treatment plants. Chapter 6 represents therefore a proof of concept for the application of the autotrophic nitrogen removal in a single reactor with granular sludge at mainstream conditions. The next logical step in this challenging exploration of anammox bacteria capabilities was to investigate their behavior in the real mainstream of a sewage treatment plant. In Chapter 7 we report the evaluation of the anammox process in a granular sludge fluidized bed lab-scale reactor continuously fed with the actual effluent of the A-stage of the WWTP of Dokhaven, Rotterdam (The Netherlands). In order to exclude the influence of oxygen and the competition for nitrite on anammox growth, the reactor was anoxic and nitrite was dosed continuously to support anammox activity only. The effect of influent COD and related heterotrophic growth by denitrification was instead included in the evaluation. The exclusion of oxygen was also intended in order to better evaluate the effect, if any, of potential toxic compounds in wastewater (e.g. from the influent or the addition of polyelectrolyte and technical grade iron salts in the A-stage). The system was operated for more than ten months at temperatures between 20 and 10°C. Volumetric N-removal rates obtained were comparable or higher than those of conventional N-removal systems, with values higher than 0.4 g-N L-1 d-1 when operated at 10°C. The biomass specific N-removal rate at 10°C was on average 50±7 mg-N g-vss-1 d-1 during last month of operations, almost two times higher than previously reported activities at this temperature. FISH analysis revealed that the dominant anammox species was Candidatus Brocadia Fulgida throughout the experimentation. Evidence for growth of anammox bacteria at main-stream conditions (i.e. anammox biomass increase and nitrate production in absence of oxygen) was demonstrated for the entire temperature range tested (10-20°C). Capability of granulation in the mainstream matrix under operative conditions was also proved since new granules were shown to be actively formed and efficiently retained in the proposed system. COD was also consumed during the process, but heterotrophs could not outcompete anammox bacteria. In Chapter 7 the capability of anammox bacteria to thrive under municipal wastewater conditions (low temperature, low ammonium concentration and presence of COD) was demonstrated for the first time, opening new perspective for the implementation of a more efficient (municipal) wastewater treatment chain. For the application of the autotrophic nitrogen removal process, the first step of partial nitritation performed by ammonium oxidizing bacteria (AOB) has to be also accomplished in order to produce the nitrite used in the anammox process. During partial nitritation the nitratation process performed by nitrite oxidizing bacteria (NOB) has to be suppressed. Even though in Chapter 7 it was demonstrated that anammox itself does not represent a problem, the managing of AOB and NOB activities in order to meet effluent standards might prove more complex for the direct application of the partial nitritation-anammox process on municipal wastewater. With the aim of evaluating the coupling of the anammox and partial nitritation processes at municipal wastewater conditions and the simultaneous suppression of the nitratation process, a pilot-scale experimentation was performed. In Chapter 8 we report the evaluation of the process in a plug-flow granular sludge based pilot-scale reactor (4 m3) continuously fed with the actual effluent of the A-stage of the WWTP of Dokhaven, Rotterdam. The one-stage partial nitritation-anammox system was operated for more than ten months at 19±1 °C. Observed average N-removal and ammonium conversion rates were comparable or higher than those of conventional N-removal systems, with 182±46 and 315±33 mg-N L-1 d-1 respectively. Furthermore, considering the higher biomass concentration obtainable in granular systems and the possibility for further anammox enrichment in the biomass, kinetics much higher than conventional systems appear to be feasible. BOD was also oxidized in the system with average removal efficiency of 90%. The system was shown to efficiently retain granules enriched in anammox bacteria with a small fraction of nitrifiers and heterotrophs located in the outer rim. At the same time, suspended flocs enriched in heterotrophs, and a small fraction of nitrifiers, were preferentially washed-out, allowing the system to withstand occasional COD and solids shock loads. The results reported in Chapter 8 show that the proposed reactor configuration with granular sludge has the potential to be successfully applied for the completely autotrophic nitrogen removal from the mainstream of WWTPs. In summary, the research described in this thesis showed for the first time the feasibility of an innovative technology for the removal of nitrogen from wastewater and posed a solid background to open de facto a new era in which the treatment of wastewater will move from the actual energy depleting to an energy generating process. Tommaso Lotti, september 2015Environmental BiotechnologyApplied Science