Evides Industriewater (further referred to as ‘EIW’) operates several wastewater treatment plants in the Netherlands. A temporary N2O monitoring campaign at one of its industrial treatment plants exposed highly dynamic N2O emissions with significant peaks. N2O can be formed in the biological treatment process in various ways, depending on local conditions (see box). EIW aims to reduce greenhouse gas emissions, while optimising the treatment efficiency of its plants. As N2O alone accounts for almost 50% of EIWs total carbon footprint, it is considered a priority. The first step in developing mitigation measures was to identify the main causes of N2O formation.
This led to the objectives for this study:
The chemistry of N2O formation
Untreated domestic wastewater contains nitrogen and carbon. Carbon is present in the form of organic matter: faeces, soap and the like. Nitrogen is present mainly as ammonium (NH4+).
At the wastewater treatment plant, bacteria in an aerated (aerobic) zone convert NH4+ into nitrite and nitrate (nitrification) in a biological process. This is followed by denitrification: nitrite and nitrate are converted by other bacteria under oxygen-free conditions (anaerobic or anoxic) into harmless nitrogen gas (N2), which escapes into the air.
While the first step (nitrification) does not require carbon, the available carbon present is still being consumed and escapes as CO2. This is a ‘waste’ of carbon, because carbon is subsequently needed in the denitrification (anaerobic) process as food for the bacteria. External carbon must then be added.
The solution is to allow the water to recirculate. In the anoxic zone, ‘fresh’ bacterial sludge is added to the large stream of nitrate-rich water, as is a small stream of untreated wastewater containing naturally occurring organic matter (inlet in Figure 1a). The organic matter is then effectively used for denitrification, reducing the need for external carbon. Everything then flows back into the aerated zone. This cycle repeats itself many times, with a stream of water being constantly discharged from the aerated zone as ‘clean’ effluent (outlet in Figure 1a). At that point, that stream contains only a very minimal amount of nitrate.
The basic concept of biological treatment therefore lies in the local ratios of dissolved oxygen (DO), NH4+ and carbon. However, when these ratios are disrupted, nitrite and nitrate are not only converted into N2, but also into N2O, a highly potent greenhouse gas. N2O formation fluctuates with the conditions in the reactors. It is therefore important to gain better control over the process in order to minimise N2O emissions.
Approach
The key features of the treatment plant studied are as follows:
Partnering with AM-Team, EIW opted for a special approach: on the spot monitoring combined with modelling to achieve results rapidly. Monitoring showed the emissions level, while the models revealed the root causes of N2O and predicted the impact of mitigation strategies which was impossible to be evaluated on site at the real treatment plant. Two types of models were used: a 3D Computational Fluid Dynamics (CFD) model and a dynamic model. The 3D model mapped where exactly in the reactor N2O was formed, while the dynamic model predicted fluctuations in N2O and treatment efficiency over time [1, 2].
Figure 1. 3D profiles predicted by 3D CFD model for DO [a] and N2O [b]. The DO profile (a) shows areas of O2 concentration even in the aerated zones. Local O2, NH4+ and carbon levels cause three N2O hotspots in different parts of the bioreactor (b).
The models were fed and validated with three types of treatment data (a combination of online and offline data):
‘What-if’-scenarios were then run using the models, thereby testing process changes with the aim of reducing N2O emissions and improving treatment efficiency. Five scenarios were tested in this way and ranked according to impact and practical feasibility.
Findings
3D model: the N2O hotspots in the treatment plant
The 3D profile of dissolved oxygen (DO) (Figure 1a) showed that (undesired) recirculation of DO occurs from the aerated (last) zone of the reactor to the anoxic zone 1. As previously described, the anoxic zone must be free of oxygen (O2). Together with the high NH4+ concentration of the influent, this created a local N2O hotspot at that location (Figure 1b). Next, a lack of carbon source a little more downstream in anoxic zone 1 also created an N2O hotspot.
Finally, we saw that the aerated (aerobic) zones generally had very low O2 concentrations (Figure 1a). It was only in the second part of aerobic zone 2 that O2 levels started to rise to values around 2 mg O2/l. These low O2 concentrations in most of the aerobic volume stimulated local N2O formation in these zones, as shown by the hotspots.
We thus identified three root causes. However, the lack of carbon source, leading to incomplete denitrification, emerged as the dominant cause, with an estimated relative contribution of >90% to total N2O formation.
Dynamic model
Both the dynamic model and the N2O monitoring campaign showed a dynamic emission profile (Figure 2a), with two peaks per day. Calibrated based on available treatment data, the model predicted DO and NH4+ concentrations and ultimately N2O dynamics. The model’s prediction corresponded to a reasonable degree with the monitoring data for N2O (Figure 2a), NH4+ and DO (not shown). Although additional information, particularly on the carbon source in the influent, could further improve the accuracy of the model, it was found to be sufficiently adequate to be applied for virtual testing of mitigation strategies. As this model also showed that the lack of carbon source (leading to incomplete denitrification) was the main cause of the peaks, carbon source levels were the main focus of the mitigation strategies.
Virtual testing of mitigation strategies
The dynamic model was then used to run what-if scenarios. The scenarios differed in terms of amount and timing of carbon dosing. Figure 2a shows that adjusting carbon dosing could reduce (mitigate) N2O emissions very effectively: in a number of scenarios, N2O peaks could be nearly eliminated. Scenario 5 included a dosing strategy proportional to the incoming carbon source that potentially could reduce N2O emissions by 87% and the dosing of external carbon source by 10%. Moreover, the model predicted an improvement in treatment efficiency, with an expected 10–15% reduction in total nitrogen in the effluent (results not shown).
Figure 2. Dynamic model: N2O in the aqueous phase as measured (crosses) and predicted (lines) (a). Optimising carbon dosing is expected to lead to significant reductions in N2O emissions and improve effluent quality, while at the same time saving on the procurement of external carbon source (b).
Conclusion and impact
EIW was able to identify the main causes of N2O emissions during the treatment process. Three causes (hotspots) of N2O were identified: 1) undesired recirculation of O2 from the aerobic to the anoxic zone, 2) very low O2 levels in the aerobic zones and 3) a lack of carbon source in the anoxic zone. The latter was identified as the dominant cause of N2O formation, with an expected share of around 90%.
According to the model scenarios, optimising carbon source dosing will reduce N2O emissions by around 87% and provide additional operational cost savings of 10% on external carbon source. Moreover, the model predicts that this measure will improve effluent quality by 10–15% in terms of total nitrogen removal.
Testing a multitude of scenarios was only possible thanks to modelling.
Since multiple causes of N2O emissions were identified, there is additional potential for optimisation by avoiding low DO levels (in aeration zone 1 and the first half of zone 2) and DO recirculation (at the end of zone 2). This could be achieved by adjusting the aeration control. However, it is clear that adjusting the carbon source dosage would have the highest return on investment if considering 3E optimisation (emissions, effluent, efficiency). The modelling results support the business case for the actual mitigation measures to be taken: at the time of writing, EIW is working on implementing the new carbon dosing strategy. After implementation, EIW will assess the final impact on-site by continued monitoring.
Evides Industriewater operates several WWTPs, including a biological treatment plant for wastewater with municipal characteristics. The company was looking for measures to reduce N2O emissions and improve treatment efficiency. Two N2O emission models were used, calibrated and validated on the basis of available treatment and monitoring data. The first, a CFD model, revealed the root causes and scale of emissions in 3D. The second, a dynamic model, was able to predict and explain emission peaks over time. A lack of carbon source was identified as the dominant cause of N2O emissions. The external carbon dosage is currently being adjusted. The models predict that this will reduce N2O emissions by 87% and save 10% on external carbon source, while reducing total nitrogen in the effluent by 10–15%.