Susanne Lackner

Full-scale partial nitritation/anammox experiences: An application survey

Partial nitritation/anammox (PN/A) has been one of the most innovative developments in biological wastewater treatment in recent years. With its discovery in the 1990s a completely new way of ammonium removal from wastewater became available. Over the past decade many technologies have been developed and studied for their applicability to the PN/A concept and several have made it into full-scale. With the perspective of reaching 100 full-scale installations in operation worldwide by 2014 this work presents a summary of PN/A technologies that have been successfully developed, implemented and optimized for high-strength ammonium wastewaters with low C:N ratios and elevated temperatures. The data revealed that more than 50% of all PN/A installations are sequencing batch reactors, 88% of all plants being operated as single-stage systems, and 75% for sidestream treatment of municipal wastewater. Additionally an in-depth survey of 14 full-scale installations was conducted to evaluate practical experiences and report on operational control and troubleshooting. Incoming solids, aeration control and nitrate built up were revealed as the main operational difficulties. The information provided gives a unique/new perspective throughout all the major technologies and discusses the remaining obstacles.

Low temperature partial nitritation/anammox in a moving bed biofilm reactor treating low strength wastewater.

Municipal wastewater collected in areas with moderate climate is subjected to a gradual temperature decrease from around 20 °C in summer to about 10 °C in winter. A lab-scale moving bed biofilm reactor (MBBR) with carrier material (K3 from AnoxKaldnes) was used to test the tolerance of the overall partial nitritation/anammox process to this temperature gradient. A synthetic influent, containing only ammonium and no organic carbon was used to minimize denitrification effects. After stable reactor operation at 20 °C, the temperature was slowly reduced by 2 °C per month and afterward held constant at 10 °C. Along the temperature decrease, the ammonium conversion dropped from an average of 40 gN m(-3) d(-1) (0.2 gN kgTSS h(-1)) at 20 °C to about 15 gN m(-3) d(-1) (0.07 gN kg TSS h(-1)) at 10 °C, while the effluent concentration was kept <8 mg NH4-N l(-1) during the whole operation. This also resulted in doubling of the hydraulic retention time over the temperature ramp. The MBBR with its biofilm on 10 mm thick carriers proved to sufficiently sustain enough biomass to allow anammox activity even at 10 °C. Even though there was a minor nitrite-build up when the temperature dropped below 12.5 °C, reactor performance recovered as the temperature decrease continued. Microbial community analysis by 16S rRNA amplicon analysis revealed a relatively stable community composition over the entire experimental period.

Sequential Aeration of Membrane-Aerated Biofilm Reactors for High-Rate Autotrophic Nitrogen Removal: Experimental Demonstration

One-stage autotrophic nitrogen (N) removal, requiring the simultaneous activity of aerobic and anaerobic ammonium oxidizing bacteria (AOB and AnAOB), can be obtained in spatially redox-stratified biofilms. However, previous experience with Membrane-Aerated Biofilm Reactors (MABRs) has revealed a difficulty in reducing the abundance and activity of nitrite oxidizing bacteria (NOB), which drastically lowers process efficiency. Here we show how sequential aeration is an effective strategy to attain autotrophic N removal in MABRs: Two separate MABRs, which displayed limited or no N removal under continuous aeration, could remove more than 5.5 g N/m(2)/day (at loads up to 8 g N/m(2)/day) by controlled variation of sequential aeration regimes. Daily averaged ratios of the surficial loads of O(2) (oxygen) to NH(4)(+) (ammonium) (L(O(2))/L(NH(4))) were close to 1.73 at this optimum. Real-time quantitative PCR based on 16S rRNA gene confirmed that sequential aeration, even at elevated average O(2) loads, stimulated the abundance of AnAOB and AOB and prevented the increase in NOB. Nitrous oxide (N(2)O) emissions were 100-fold lower compared to other anaerobic ammonium oxidation (Anammox)-nitritation systems. Hence, by applying periodic aeration to MABRs, one-stage autotrophic N removal biofilm reactors can be easily obtained, displaying very competitive removal rates, and negligible N(2)O emissions.

Response of Different Nitrospira Species To Anoxic Periods Depends on Operational DO

The exploitation of a lag phase in nitrate production after anoxic periods is a promising approach to suppress nitrite oxidizing bacteria, which is crucial for implementation of the combined partial nitritation-anammox process. An in-depth study of the actual lag phase in nitrate production after short anoxic periods was performed with varied temperatures and air flow rates. In monitored batch experiments, biomass from four different full-scale partial nitritation-anammox plants was subjected to anoxic periods of 5-60 min. Ammonium and the nitrite that was produced were present to reproduce reactor conditions and enable ammonium and nitrite oxidation at the same time. The lag phase observed in nitrite oxidation exceeded the lag phase in ammonium oxidation after anoxic periods of more than 15-20 min. Lower temperatures slowed down the conversion rates but did not affect the lag phases. The operational oxygen concentration in the originating full scale plants strongly affected the length of the lag phase, which could be attributed to different species of Nitrospira spp. detected by DGGE and sequencing analysis.

Inoculum effects on community composition and nitritation performance of autotrophic nitrifying biofilm reactors with counter-diffusion geometry.

The link between nitritation success in a membrane-aerated biofilm reactor (MABR) and the composition of the initial ammonia- and nitrite-oxidizing bacterial (AOB and NOB) population was investigated. Four identically operated flat-sheet type MABRs were initiated with two different inocula: from an autotrophic nitrifying bioreactor (Inoculum A) or from a municipal wastewater treatment plant (Inoculum B). Higher nitritation efficiencies (NO(2)(-)-N/NH(4)(+)-N) were obtained in the Inoculum B- (55.2-56.4%) versus the Inoculum A- (20.2-22.1%) initiated reactors. The biofilms had similar oxygen penetration depths (100-150 µm), but the AOB profiles [based on 16S rRNA gene targeted real-time quantitative PCR (qPCR)] revealed different peak densities at or distant from the membrane surface in the Inoculum B- versus A-initiated reactors, respectively. Quantitative fluorescence in situ hybridization (FISH) revealed that the predominant AOB in the Inoculum A- and B-initiated reactors were Nitrosospira spp. (48.9-61.2%) versus halophilic and halotolerant Nitrosomonas spp. (54.8-63.7%), respectively. The latter biofilm displayed a higher specific AOB activity than the former biofilm (1.65 fmol cell(-1) h(-1) versus 0.79 fmol cell(-1) h(-1) ). These observations suggest that the AOB and NOB population compositions of the inoculum may determine dominant AOB in the MABR biofilm, which in turn affects the degree of attainable nitritation in an MABR.

Modeling of Biofilm Systems: A Review

The modeling of biochemical processes in biofilms is more complex compared to those in suspended biomass due to the existence of substrate gradients. The diffusion and reaction of substrates within the biofilms were simulated in 1D models in the 1970s. The quality of these simulation results was later improved by consideration of mass transfer at the bulk/biofilm interface and detachment of biomass from the surface. Furthermore, modeling of species distribution along the axis perpendicular to the substratum helped to simulate productivity and long-term behavior in multispecies biofilms. Multidimensional models that were able to give a realistic prediction of the surface structure of biofilms were published in the 1990s. The 2D or 3D representation of the distribution of the species in a matrix of extracellular polymeric substances (EPS) helped predict the behavior of multispecies biofilm systems. The influence of shear forces on such 2D or 3D biofilm structures was included by solving the Navier-Stokes equation for the liquid phase above the biofilm. More recently, the interaction between the fluid and biofilm structures was addressed more seriously by no longer considering the biofilm structures as being rigid. The latter approach opened a new door, enabling one to describe biofilms as viscoelastic systems that are not only able to grow and produce but also be deformed or even dislodged if external forces are applied.

Evaluating operation strategies and process stability of a single stage nitritation–anammox SBR by use of the oxidation–reduction potential (ORP)

A single stage nitritation-anammox SBR was operated for 300 days to investigate the impact of cycle operation strategies on process performance and the oxidation-reduction potential (ORP) as process monitoring parameter. Different combinations of feeding (interval, continuous, one-time) and aeration (interval, continuous) strategies were tested revealing that interval feeding and interval aeration was the most suitable case in terms of process performance (ammonium removal, nitrate production and pH stability) and use of the ORP value as indicator parameter. Further investigations into the use of the ORP value showed clear correlations of the ORP slope with the air flow rate and the maximum ORP peak with the ammonium loading under varying operation conditions. Depletion of the main substrates (ammonium and oxygen) was also detectable fastest following the ORP value proofing its worth for process control.

Nitritation performance and biofilm development of co- and counter-diffusion biofilm reactors: Modeling and experimental comparison

A comparative study was conducted on the start-up performance and biofilm development in two different biofilm reactors with aim of obtaining partial nitritation. The reactors were both operated under oxygen limited conditions, but differed in geometry. While substrates (O2, NH3) co-diffused in one geometry, they counter-diffused in the other. Mathematical simulations of these two geometries were implemented in two 1-D multispecies biofilm models using the AQUASIM software. Sensitivity analysis results showed that the oxygen mass transfer coefficient (Ki) and maximum specific growth rate of ammonia-oxidizing (AOB) and nitrite-oxidizing bacteria (NOB) were the determinant parameters in nitrogen conversion simulations. The modeling simulations demonstrated that Ki had stronger effects on nitrogen conversion at lower (0-10 m d(-1)) than at the higher values (>10 m d(-1)). The experimental results showed that the counter-diffusion biofilms developed faster and attained a larger maximum biofilm thickness than the co-diffusion biofilms. Under oxygen limited condition (DO<0.1 mg L(-1)) and high pH (8.0-8.3), nitrite accumulation was triggered more significantly in co-diffusion than counter-diffusion biofilms by increasing the applied ammonia loading from 0.21 to 0.78 g NH4+-NL(-1) d(-1). The co- and counter-diffusion biofilms displayed very different spatial structures and population distributions after 120 days of operation. AOB were dominant throughout the biofilm depth in co-diffusion biofilms, while the counter-diffusion biofilms presented a stratified structure with an abundance of AOB and NOB at the base and putative heterotrophs at the surface of the biofilm, respectively.

Enhancing the formation and shear resistance of nitrifying biofilms on membranes by surface modification

Polypropylene (PP) membranes and polyethylene (PE) surfaces were modified to enhance formation and shear resistance of nitrifying biofilms for wastewater treatment applications. A combination of plasma polymerization and wet chemistry was employed to ultimately introduce poly(ethyleneglycol) (PEG) chains with two different functional groups (-PEG-NH(2) and -PEG-CH(3)). Biofilm growth experiments using a mixed nitrifying bacterial culture revealed that the specific combination of PEG chains with amino groups resulted in most biofilm formation on both PP and PE samples. Detachment experiments showed similar trends: biofilms on -PEG-NH(2) modified surfaces were much stronger compared to the other modifications and the unmodified reference surfaces. Electrostatic interactions between the protonated amino group and negatively charged bacteria as well as PEG chain density which can affect the surface structure might be possible explanations of the superiority of the -PEG-NH(2) modification. The success of the-PEG-NH(2) modification was independent of the original surface and might, therefore, be used in wastewater treatment bioreactors to improve reactor performance by making biofilm formation more stable and predictable.

Comparing the performance and operation stability of an SBR and MBBR for single-stage nitritation-anammox treating wastewater with high organic load

Single stage nitritation-anammox reactors have gained increasing attention for their application in municipal and industrial wastewater treatment. The most commonly used system in municipal reject water treatment is at present the sequencing batch reactor (SBR), the moving-bed biofilm reactor (MBBR) is the second most common. However, little is known about their applicability to industrial wastewaters with high C/N ratios. This study presents a comparative approach to evaluate the performance of these two systems by changing the influent from reject water (C:N ratio 1:1) stepwise to an industrial wastewater (C:N ratio 3:1). An intentionally induced temperature drop that led to nitrite accumulation was also tested. The results showed that the MBBR (1.9 kg-N m−3 d−1) was superior to the SBR (0.5 kg-N m−3 d−1) with at maximum up to four times higher volumetric nitrogen removal rates. Both systems accumulated nitrite (>100 mg-N L−1) during the temperature drop from 30°C to as low as 18°C (MBBR) and 20°C (SBR), which subsequently resulted in almost complete loss in the removal capacities. However, the previous removal rates could be re-established in both systems within approximately 40 days. In comparison, the MBBR showed the more stable and higher performance even though higher nitrite concentrations (up to 500 mg-N L−1) were encountered. Overall, MBBR operation and handling was also easier and the system was more robust to disturbances compared to the SBR.