Harnessing the alkalinity generation during denitrification in the treatment of acidic wastewater from a chemical-manufacturing facility in a wastewater treatment plant (WWTP) results in self-neutralization, offering cost savings and other benefits
Facilities throughout the chemical process industries (CPI) produce large volumes of wastewater, most of which is treated in biological wastewater treatment plants (WWTPs). Since the optimum pH for the biological activity within the WWTP is in the range of 6.5–8.5 [1], it is standard practice that wastewaters that are originally acidic or basic be neutralized to pH of about 7 before they are directed to a biological WWTP. This common practice consumes large volumes of neutralizing chemicals.
An example of this neutralization practice occurs at the Bayer site in Antwerp, Belgium, where a wastewater stream containing nitric acid (HNO3) has been completely neutralized with caustic soda (NaOH) over the past 30 years. This article discusses an alternative approach, wherein the spontaneous pH increase during the involved biological denitrification in the WWTP is harnessed to achieve the neutralization, rather than using NaOH. By harnessing this process for neutralization, the amount of NaOH required for treatment was dramatically reduced, and the approach also proved beneficial in terms of sludge production at the WWTP. This article describes how deep process knowledge and an innovative idea, followed by lab tests and smart virtual testing, can fundamentally change the way a process is run in the CPI.
Denitrification
Nitrate (NO3–) anions can be biologically reduced via nitrite (NO2–) and other intermediates to gaseous nitrogen (N2) in a reaction known as denitrification. Under the conditions of very low or almost zero dissolved oxygen in the wastewater, and sufficient presence of suitable organic carbon compounds (to act as electron donors), certain micro-organisms can remove the NO3– and NO2– from wastewater in a so-called denitrifying, or anoxic, reactor. The optimal pH range for denitrification is 6.5–7.5, and even slightly outside that optimal pH range (for example, pH = 6 or 8), the kinetics already fall to 70% [1].
During the denitrification reaction, the alkalinity (buffering capacity) and pH increase. Equation (1) describes the denitrification with the organic carbon in the water generalized as C10H19O3N [2]:
C10H19O3N + 10NO3– —> 5N2 + 10CO2 + 3H2O + NH3 + 10OH – (1)
This means that for each mole of NO3– converted by the micro-organisms, one mole of OH– is biologically generated. This insight turned out to be a “game-changer.”
Traditional neutralization of wastewater
The wastewater treatment plant at the Bayer Antwerp facility treats a mixture of eight wastewater streams from the production site. All but one of these wastewater streams come together in a buffering pond and make up so-called lagoon wastewater. Besides these streams, the WWTP also treats a separate NO3– -rich wastewater flow. Because the general rule has been to feed neutral wastewater to a biological WWTP, in the past, the HNO 3 production water (originally at pH 1.5) was completely neutralized at the manufacturing plant using caustic soda before the wastewater was sent to the WWTP. The neutralization reaction is shown in Equation (2).
HNO3 + NaOH —> H2O + NaNO3 (2)
In the past, the two flows (lagoon-pond wastewater and NO3– -rich wastewater) were fed separately via a stationary piping system into the anoxic compartment, as depicted at the right-hand side of Figure 1, along with the recycle activated sludge (RAS). As a result of the denitrification shown in Equation (1), the pH in the anoxic compartment typically increased to values of 8.5 to 9, which is beyond the optimal pH range.
FIGURE 1. Shown here is the setup of the denitrification compartment as it used to be (right diagram) and as it is now since the project described here (left diagram). The bunch of velocity streamlines results from CFD modeling to verify the mixing. Before, the NO3 –-rich wastewater was completely neutralized and entered the system via a fixed piping. Since 2024 (left side), the acid wastewater is first mixed with the neutral lagoon wastewater and the resulting acid premix is fed to a Ventoxal unit. In the diagram, RAS is recycled activated sludge and ww is wastewater (CFD simulations by AM-Team)
On top, due to the elevated pH during denitrification, calcium ions (Ca 2+) present in the wastewater precipitated with carbonate (CO32–) anions to form CaCO3 solids, significantly contributing to the sludge generation in the basin.
This is demonstrated in a laboratory set-up (Figure 2). In two batch reactors, the same volumetric mixture was made of RAS, Lagoon wastewater and the NO3– -rich wastewater as it enters into the anoxic compartment of the full-scale WWTP. In the laboratory reactor, wherein the denitrification proceeds as it is (with the pH left uncontrolled) for 180 minutes (data shown in red on Figure 2), the pH increases from about 7.3 to 9. As a result, the dissolved Ca2+ ion concentration is reduced from about 100 ppm to 30 ppm. This is explained by CaCO3 precipitation due to the elevated pH.
FIGURE 2. The pH and concentration of dissolved calcium (in ppm) is shown here as a function of the denitrification time (in min.) in two batch experiments. When the pH is allowed to (spontaneously) increase during the denitrification, the resulting high pH causes precipitation of CaCO3 solids. However, when the pH is controlled in a neutral range by manually adding HCl, calcium ions remain in solution
However, in the case where the pH increase during the denitrification is compensated by manual addition of HCl to control the pH in a small range from 7 to 7.3 (data shown in blue on Figure 2), the Ca2+ ions remain well in solution (the Ca2+ concentration even increases somewhat over the 180-min reaction time. This increase can be attributed to a local dissolution of already present CaCO 3 in the RAS flocs in the areas where the HCl spiking occurs in the mixture). The HCl dosed in the pH-controlled reactor also provided proof-of-concept regarding the mole of OH– generated per mol of NO3–, as described in Equation (1).
In the past, the inorganic fraction of the activated sludge in this WWTP could sometimes reach as much as 80–90% of the sludge, as shown in an earlier Chem. Eng. article [3]. In recent years, this fraction still amounted to 40% of the overall sludge generated, and, hence, contributed significantly to the waste activated sludge (WAS) volume to be handled.
Self-neutralizing wastewater
Since OH – is generated when NO3– is biologically converted into N2 gas in the WWTP, the idea came up to use this biological feature to let the HNO3 wastewater partially self-neutralize. This innovative idea is schematically shown in Figure 3. Instead of dosing all OH- needed to completely neutralize the H+ in the HNO3 wastewater from the manufacturing plant (schematically depicted at the right side of Figure 3), only part of the OH – is dosed at the manufacturing plant, since the rest of the needed OH – is generated by the micro-organisms once the NO3– anions are fed into the biological WWTP (Figure 3, left side).
FIGURE 3. In the past (right side) the HNO3 wastewater was completely neutralized at the plant, resulting in a high pH in the WWTP and CaCO3 sludge solids formation. Since the end of 2024 (left side), the acid stream is only partially neutralized at the plant and the acid stream is further “self-neutralized” thanks to the OH– generation in the biological WWTP
It must be mentioned that more alkalinity is generated by the micro-organisms than is consumed by the incoming un-neutralized H+ in the feed. Obviously, the neutralized NaNO3 part of the feed also results in alkalinity generation. In this way, enough buffering capacity in the compartment is provided to immediately neutralize the fed acid premix (Figure 3).
Now, to turn this idea of self-neutralizing HNO3 wastewater from a paper desk exercise into a real-world working application, the toxicity of a low pH feed on the activated sludge needed to be tackled.
Acidic wastewater in a WWTP
A creative solution was found to cope with the engineering challenge of dosing acidic wastewater into a biological treatment system without hampering its function. Rather than feeding the wastewaters with simple, gravitational piping, like in the past (as depicted at the right side of Figure 1), it was decided to use a Ventoxal unit from Air Liquide, shown in Figure 4, to dynamically feed the wastewaters into the biological WWTP. First, the nitric acid wastewater of pH ~2 (see Figures 1 and 4) is mixed in a static mixer with the neutral lagoon wastewater, yielding a premix at a pH of 2.5–4.5. While this range is somewhat elevated, it remains toxic for the microbes. Next, this premix is sent to the Ventoxal unit, where the reduced diameter at the end of the feed line creates a venturi effect, sucking in an approximately 4–5 times higher volumetric flow of the receiving compartment into the Ventoxal tubes. This way, in a matter of 1–2 seconds, the acid premix is intensively mixed with a large volume of sludge and water from the compartment (still with a vast amount of alkalinity available thanks to the denitrification). This yields a pH of 6.8–7.2 in the mixture just downstream of the Ventoxal unit. This rapid mixing minimizes the contact time of the micro-organisms with the acidic premix.
FIGURE 4. The photo shows the Ventoxal mixing device mounted to the wall of the compartment. The acid premix with a pH between 2.5 and 4.0 is mixed in a very dynamic way with the water and sludge from the compartment so that, after leaving the Ventoxal, the pH of the mixture is already neutral
At this point in the project development, to further substantiate this practical solution with the Ventoxal, the pH of the mixture obtained right after the Ventoxal was simulated at laboratory scale by using a sequential approach of design of experiments (DOE), and ultimately, a Response Surface Model (RSM) – the structured and scientific way to gain knowledge over a process in the CPI, as highlighted in a prior Chem. Eng. article [4].
The graphical representation of the empirical pH model (obtained with the JMP statistical software) is depicted in Figure 5 for a certain fixed set for the volumetric flows of the HNO3 stream, Lagoon stream and water flow from the compartment, providing important information in terms of allowable pH for the HNO3 stream. A pH of at least 6.5 is aimed for the mixture just downstream of the Ventoxal (that is, at the entrance of the wastewater in the anoxic compartment) to remain in the pH range for optimal biological activity. This pH of 6.5 is indicated with the grid in Figure 5. Later on in the project, as discussed below, virtual testing by AM-Team (www.am-team.com) based on theoretical models proved again that the low-pH plume would be marginal at the outlet of the Ventoxal unit.
FIGURE 5. The illustration shows how the pH of the HNO3 wastewater and the pH in the compartment affect the mixture pH right after the Ventoxal unit for a certain fixed set of volumetric flows in the Ventoxal
Low pH and short contact time
During project development, it was important to know upfront that the micro-organisms would not be killed off by contact with the acid premix in the very short residence (contact) time in the Ventoxal unit. In the published literature, long-term effects of lower pH values on sludge can be found, but this information is not entirely satisfying given the extremely short contact time in this context. To obtain more information, a laboratory set-up was constructed. In a first attempt, a simple Jar test mixer (typically encountered in laboratories dealing with wastewater treatment tests) was used at 100 rpm.
At the right side of Figure 6, a beaker is shown filled with compartment wastewater (without the sludge, to allow visual inspection of the neutralization kinetics). It has the same expected (lowered) alkalinity and pH of the receiving anoxic compartment water after the project would have been started.
FIGURE 6. The laboratory setup shows the effect of slow, laminar mixing (right side), and intensive, turbulent mixing (left side) of the acid premix wastewater (colored red by methyl red indicator) into water with the same alkalinity as the receiving compartment mixture in the full-scale WWTP. When the acid premix is poured at the same rate into the beaker equipped with a jet mixer (left), barely red-colored (acid) spots are observed, indicating very fast neutralization, whereas in the case of slow mixing (right), the whole beaker turns red, indicating that the neutralization is slower
The same volumetric ratios were used for the acid premix compared to the (larger) compartment mixture that would be sucked in at the Ventoxal (that is, a ratio of 1 volume premix to 4 volumes of receiving compartment water). The acid premix was poured into the beaker within 1 second. For the purpose of visualization, the acid premix was colored red using methyl red indicator, which turns from red into yellow upon neutralization.
It was observed that the beaker turned completely red for a certain time, as demonstrated at the right side of Figure 6 (that is, the added acid wastewater caused acidic conditions in the beaker, although for a very short time). Next, tests were executed using a volume of receiving sludge and water from the compartment in the beaker and, likewise, acid premix was poured into it within 1 second. Next, the concentration of adenosine triphosphate (ATP), the energy carrier in living cells, was measured in the beaker to serve as an indication of the living micro-organisms left after exposure to different pH conditions.
ATP concentrations were determined by using Luminultra’s bioluminescence method. The tests were executed with different pH values of the (partially) neutralized HNO 3 wastewater used to make-up the premix (with a resulting pH in the range of 4–5 added to the sludge).
The results depicted in Figure 7 show that the pH of the HNO3 wastewater (and, also, the pH of the resulting premix) had a detrimental effect on the sludge, at least when using the slow mixing-in. These laboratory results stress again that, indeed, adding large volumes of acid wastewater into a biological wastewater treatment is certainly not without risk. Proper precautions should be taken (that is, intensified mixing to shorten the contact time between the sludge and the acidic water, like in this project) to ensure the activated-sludge activity would not be hampered by the dosing of the acid water.
FIGURE 7. Shown here is the total ATP concentration as a function of the administered pH of the HNO3 wastewater used to make-up the premix, tested on three different days, when using slow, laminar mixing
Mimicking the intensified mixing of the acid premix with the sludge using a Ventoxal unit was the next step in the laboratory to get a more realistic simulation of the neutralization in the jet.
A completely different mixing device was used in the next stage of this research aimed to better mimic the very turbulent mixing that happens in the Ventoxal. At a rotational speed of 800 rpm (shown at the left side of Figure 6), pouring the acid premix in the receiving compartment water now gave a completely different look. One can see barely red-colored (that is, still acidic) spots in the beaker, indicating overall immediately neutralized water (a totally different view compared to the beaker at the right side of Figure 6).
Next, the same experiments as described above were replicated, pouring acid premix into a mixture of compartment water and sludge, while measuring the ATP in the resulting mixture. This time, the effect of a pH of 2.2 for the HNO3 wastewater was tested, versus a reference of pH 6.2, for the make-up of the premix, on two different days (with two different sludges), as shown in Figure 8.
FIGURE 8. Shown here is the total ATP concentration as function of the administered pH of the HNO3 wastewater used to make-up the premix, tested on two different days, when using intensive turbulent mixing
To evaluate the results on the ATP concentration, an analysis of variance (ANOVA) test was carried out with the sludge (on day 1 and day 2) and the pH (at 2.2 and 6.2) as factors. This ANOVA test indicated that there was no statistically significant effect (p = 0.2) of the HNO3 wastewater pH (used to make up the premix) on the ATP concentration. This indicates that the micro-organisms were not killed by the added acidic premix. Furthermore, after the test, with the observed standard deviation on the ATP measurement of ~100 ng/mL for the 16 tests executed, the probability of correctly rejecting the null hypothesis (H0: there is no effect of the low pH on the ATP) is 80%, if in reality, there would be a true difference of at least 150 ng/mL ATP.
This reduction in ATP would be equivalent to a 5% killing rate of the bacteria, which would be important to know for the project. In conclusion, in case the application of a pH 2.2 water really would result in the death of 5% of the bacteria, we had an 80% chance of effectively showing this (unwanted) effect on the biology with the tests done, but the results showed no statistically significant effect on the sludge.
Virtual testing for upscaling
To upscale the results from the laboratory-scale batch reactors to the full-scale continuous WWTP, a smart combination of kinetic and computational fluid dynamics (CFD) modeling was used by AM-Team [5]. The denitrification experiments that were run in the Bayer laboratory batch reactors were used to calibrate denitrification and pH models. These full-scale models allowed us to predict the new equilibrium pH, while confirming the avoidance of CaCO3 precipitation due to the neutral pH in the WWTP.
The benefit of the modeling approach (in contrast to full-scale pilot testing) is that once the models are available and are validated, the testing can be carried out in a more flexible way, by far. A series of virtual “what-if” tests were executed using different pH values for the HNO3 wastewater, confirming to the project group the feasibility of dosing partially neutralized HNO3 wastewater into the anoxic compartment. CFD modeling (see, for example, Figure 1) indicated that the best placement for the new Ventoxal feeding device was on the wall of the anoxic compartment, as depicted at the left side of Figure 1. For more information, refer to Ref. 5. The modeled process changes were installed in the WWTP.
The move to partial neutralization
After implementation of the project, the pH of the HNO 3 wastewater treated in the biological WWTP had been lowered over different months from neutral pH to finally 2.0. In doing so, the pH in the anoxic compartment went gradually down from the 8.5–9.0 to 6.8–7.2, which is well in line with the outcomes of AM-Team’s virtual testing. As a result, the sludge’s inorganic fraction was reduced from 40% to 10–15%, and the sludge concentration in the basin (so-called mixed-liquor suspended solids; MLSS) was lowered from about 9 g/L to 6 g/L, still attaining the same required active organic sludge concentration of about 5 g/L, as it was in the past.
The downside of the sludge’s reduced inorganic fraction is poorer sludge settleability. This was expected as a logical consequence of the avoidance of the CaCO 3 solids precipitation. In the past, these solids were enmeshed in the activated-sludge flocs, making the flocs literally heavy and improving the overall settling [6]. The worse settling behavior is partially compensated by the lower solids load to the final clarifiers, as a result of the reduced MLSS inventory. Previous experience with the new set-up shows that the sludge settleability is more vulnerable to changes in the organic sludge quality, such as the presence of filaments, for example.
The implementation of the self-neutralizing nitric acid wastewater-treatment process has had a huge effect on raw material usage and filter-cake production. The 40% reduced neutralization of the HNO3 stream has lowered the NaOH usage at the manufacturing plant by hundreds of tons per year. In addition, at the WWTP, since the sludge generation has been reduced by 35%, there was an equally major effect on the waste-sludge-handling operations. Hundreds of tons less of NaOH, FeCl3 and lime are needed per year to dewater the sludge in the filter presses. This results in a significantly reduced amount of produced filter cakes (by many hundreds of tons). Cost savings and sustainability go hand in hand in this innovative optimization process.
Edited by Scott Jenkins
Acknowledgements
Figure 1 appears courtesy of AM-Team. Other images are from Bayer.
References
1. Gray, N.F., “Activated sludge: developments and sustainable solutions”, World Scientific Publishing Europe Ltd., 2023.
2. Khunjar, W., Pitt, P., Bott, C., Chandran, K., Nitrogen, In: Jenkins, D. and Wanner, J., Editors., “Activated sludge — 100 Years and Counting,” IWA Publishing, 2014.
3. Peeters, B. and Vernimmen, L., Challenges of Handling Filamentous and Viscous Wastewater Sludge, Chem. Eng., vol. 123, pp. 52–58, 2016.
4. Peeters, B., Roels, M., Van Aken, S. and Desmarets G., Accelerating Six Sigma Research with the Definitive Screening Design (DSD) technique, Chem. Eng., vol. 126, pp. 54–60, 2019.
5. Muoio, R., Peeters, B. and Audenaert, W., How smart models helped to optimize Bayer’s industrial wastewater, case article, 2025 (https://www.am-team.com/projects/how-smart-models-helped-to-optimize-bayers-industrial-wastewater-treatment).
6. Peeters, B., Dewil, R., Lechat, D., Smets, I.Y., Quantification of the Exchangeable Calcium in Activated Sludge Flocs and its Implication to Sludge Settleability, Separation and Purification Technology, 83, pp. 1–8, 2011.
Authors
Bart Peeters is a senior expert for wastewater treatment at Bayer (Scheldelaan 460, 2040 Antwerp, Belgium; email: bart.peeters@bayer.com) where he has been working since 1998. He first served as a process improvement engineer at a polymer manufacturing plant on site, and since 2004, he has been working at the environmental department of the company. While working at the Bayer (Monsanto legacy) WWTP, he obtained his Ph.D. in engineering from KU Leuven. Prior to that, he received a M.Sc. in chemical engineering degree from KU Leuven, plus a M.Sc. in biochemical engineering technology from university college De Nayer. Throughout his career, he has focused on rigorous R&D studies and the optimization of industrial processes and equipment, leveraging design of experiments and other Six Sigma statistical tools to generate new ideas, insights and breakthroughs. Peeters is a SigmaPro certified Master Black Belt. He has authored 25 publications in scientific journals, technical magazines and international conferences.
Jef Goos is an operations coordinator at Bayer in Antwerp (E-mail: jef.goos@bayer.com). He has been working in the wastewater treatment and utilities department since 1987. He began his career in the maintenance department, where he spearheaded various improvements and optimizations. Subsequently, he worked as an automation engineer. During this period, the department underwent full digitalization. Several projects on the environment, cost control and efficiency improvement were successfully implemented. Later, he became the maintenance and operations coordinator for the wastewater treatment department. Currently, he serves as the operations coordinator for the utilities and wastewater treatment department and remains closely involved in implementing optimization projects and training of new employees. He holds a professional bachelor’s degree in instrumentation and control engineering from KHK Geel.
Tom Rauws is a product supply expert at Bayer in Antwerp (E-mail: tom.rauws@bayer.com), where he has been working since 2013. He obtained a Ph.D. in organic chemistry from the University of Antwerp in 2011, where he also got his bachelor’s and master’s degrees. Tom focuses on the chemistry of the different processes at Bayer to support optimizations mainly in the glyphosate production area, but also across the whole Bayer site in Antwerp. He also leads the global optimization team for glyphosate formulations. Tom is a certified Six Sigma Green Belt and ISO9001 lead auditor. He (co-)authored five publications in scientific journals.
Maiske Noppen is a laboratory analyst at Bayer in Antwerp (E-mail: maiske.noppen@bayer.com), where she has been working since 2005. She obtained her B.S. at Plantijn University College, now AP University College, in Antwerp. Since 2016 she specializes in lab projects concerning wastewater treatment and quality control of the wastewater treatment plant.
Stefan Van Otten is a Process Engineer at Bayer in Antwerp (E-Mail: stefan.vanotten@bayer.com) where he has been working since 2007. Here he works for a polymer manufacturing plant and did process design for several capital projects on capacity increase, quality and safety improvements, and cost and energy savings. He received his M. Sc. in mechanical engineering at the University of Ghent in 2006. Van Otten is a certified Six Sigma Black Belt.
Sverre Goetschalckx is a process improvement engineer at Bayer in Antwerp (E-mail: sverre.goetschalckx@bayer.com), where he has been working since 2021. Here, he works for the Utilities and Wastewater Treatment Plant and focuses mostly on process improvements, energy saving projects, process safety and quality. He is a certified ISO9001 lead auditor. He obtained his M.S.Ch.E. from KU Leuven in 2018 and worked for Worley from 2018 to 2021.
Roberta Muoio is Project Manager at AM-Team (Sint Pieters nieuwstraat 11, 9000 Ghent, Belgium; E-mail: roberta.muoio@am-team.com), a Belgian company specializing in modeling services and digital-twin solutions for the water and wastewater treatment industry. She joined AM-Team in 2020 as advanced simulation engineer and has since progressed to her current role as project manager. Prior to that, she spent two years as a research fellow at the Department of Civil and Environmental Engineering of the University of Florence, where she also obtained her M.Sc. in environmental engineering in 2018. Through her work at AM-Team, Roberta supports the design, optimization and digitalization of water and wastewater treatment plants.
Roel Boussemaere is an international senior expert and domain manager in the Air Liquide Technology Center (ALTEC). In this role, he directly supports Air Liquide’s clients, identifying water treatment challenges and delivering optimal process solutions. Roel brings over 15 years of water treatment experience, including nine years as a gas-injection-in-liquid expert at Air Liquide, preceded by three years as a process engineer at Aquasystems International N.V. and four years at Air Products and Chemicals. His research and industrial innovations primarily focus on sustainable water reuse, membrane technology, industrial wastewater treatment and green technology applications (such as utilizing carbon dioxide and pure oxygen for environmental optimization). He holds a M.Sc. in business engineering: operations management.