The impact of the phosphorus limitation on denitrification efficiency and specific sludge production in industrial wastewater treatment

CSTR

completely stirred tank reactor

SS

suspended solid

VSS

volatile suspended solids

SRB

sulphate reducing bacteria

ORP

oxidation-reduction potential

COD

chemical oxygen demand

Nitratax

probe measuring oxidated nitrogen concentration

LDO

probe measuring soluble oxygen concentration

P

phosphorus

N-NO_x

sum of oxidated forms of nitrogen (N-⁠, N-⁠)

Phosphorus is a crucial nutrient for denitrifying microorganisms, with its abundance positively influencing their activities. Heterotrophic denitrifying bacteria are among the most commonly used denitrifying bacteria which can convert nitrate and nitrite to gaseous nitrogen or nitrogen oxides, utilizing organics as electron donors (Feng et al. 2020). Multiple bacterial strains with diverse functions and synergistic effects are integrated to establish a more reliable micro-ecosystem. This integration is achieved through the functional complementation of composite bacterial strains, leading to the effective removal of various forms of nitrogen from wastewater. To prevent excessive phosphate in the effluent, and to retain denitrification in the reactor, phosphate supply should be maintained at least of 0.3–0.6 mg/L (Feng et al. 2023).

Denitrification was incomplete without the sufficient supply of phosphorus. Phosphorus limitation influenced the structural and functional diversity of bacteria (Samaddar et al. 2019). Phosphorus limitation reduces microbial nitrogen use efficiency by increasing extracellular enzyme investment (Sun et al. 2023). The composition of the mixed-denitrification culture adapts to changes in conditions, resulting in adjustments to the representation of individual species, including specific phosphorus requirements (Xu et al. 2015; Fan et al. 2018). Microorganisms may replace phosphorus with other elements in a P-limited environment, only to a certain extent. A genome of any microbial cell, carrying all the information necessary for an organism also can respond to changes in its microenvironment. Organisms can respond to environmental changes by changing their enzymatic composition and use more enzymes such as, iron, copper, and molybdenum (Moura & Moura 2001; de Vet et al. 2012).

The extremely low concentration of ortho-phosphorus (⁠-P), which is the dissolved reactive component of total phosphorus, is a primary indicator of rate-limiting conditions in denitrification reactors. The threshold below which -P becomes the rate-limiting nutrient was found: S_P/S_N = 0.0086 g P/g N and S_P/S_M = 0.0013 g P/g COD (Boltz et al. 2012).

Heterotrophic denitrification was more effective and more cost-effective than autotrophic denitrification, but phosphorus limitation inhibited heterotrophic denitrifying bacteria more greatly than autotrophic denitrifiers. The denitrification performance was periodically impacted by P limitation, particularly at low water temperatures as found in the denitrification of polluted surface water (Wang et al. 2018). The difference in efficiency is significant and can be applied to denitrifying culture in suspended and biofilm. This finding confirms the suitability of choosing heterotrophic denitrification for the treatment of water with nitrates without organics.

Industrial wastewater with a high content of nitrates and sulphates represents a great load for the conventional technology of wastewater treatment plants (De Lucas et al. 2005; Feng et al. 2023). It is possible to apply denitrification technology for the treatment of industrial wastewater but with regard to extreme composition requiring special conditions, such as wastewater from the uranium industry (Biradar et al. 2008), agro-food industry, or electronic wastewater (Song et al. 2024). Many chemical productions use directly concentrated nitric and sulphuric acids in a mixture, especially for the production of nitro-substances and the resulting wastewaters contain nitrates and sulphates. There has been a concern that sulphate could be reduced by sulphate-reducing bacteria (SRB) to a toxic and strongly smelling hydrogen sulphide in a presence of a high concentration of sulphates and readily degradable substrates. However, several studies have shown that denitrifiers can inhibit SRB, which can reduce SRB activity and even its abundance, leading to a reduction in the concentration of sulphides generated (Zhao et al. 2013). Sulphate reduction can occur only after complete nitrate removal if an organic substrate is still present and oxidation–reduction potential (ORP) in the system sufficiently decreases (Stanek et al. 2012). The system ORP could reflect the effect of denitrification inhibiting sulphate reduction, when ORP was from −50 to –150 mV, denitrification predominated (Zhang et al. 2008).

The objective of this study was to explore the potential for optimizing the existing full-scale denitrification technology to industrial wastewater from the production of nitrocellulose. Wastewater contains only residual nitric and sulphuric acid, the concentration of nitrate nitrogen in raw wastewater ranges from 100 to 400 mg/L, sulphate sulphur from 330 to 650 mg/L, with a pH of around 1.5 and a temperature of 15–35 °C. The content of phosphorus and biodegradable organic substances is negligible; therefore, organic substrate and phosphoric acid need to be added for successful denitrification.

The main aim of experiments was to find out the limited concentration of phosphorus, which will still ensure the required denitrification efficiency and by reducing the phosphorus dose could reduce the specific production of denitrifying biomass.

The volume of actual wastewater produced is approximately 2 million m³/year. The treatment of wastewater consists of filtration of fibres of nitrocellulose, neutralization, biological denitrification, and chemical post-treatment. From the filtration station, wastewater was pumped to the pre-neutralization reactor of volume 100 m³ stirred by submersible mixers. Pre-neutralization was carried out using sodium hydroxide. Currently, to reduce operating costs, lime is used as the primary neutralizing agent and is dosed directly into the denitrification reactor of volume 3,000 m³ through the static mixer, where methanol and phosphoric acid are also added. A possibility of 50% sodium hydroxide additional dosage serves as a backup in the case of a sharp drop in pH in the denitrification reactor due to the influence of extremely acidic wastewater. Denitrifying sludge in the denitrification reactor is mixed with two submersible mixers. Mixed liquor flows from a denitrification reactor to a degassing tank. Waste gas bubbles, containing nitrogen and carbon dioxide, are removed at this stage to avoid the deterioration of the sludge sedimentation in the following settling tank. For the final treatment of effluent from the settling tank, chemical coagulation was implemented, with ferric sulphate and a polymer flocculant was added. The excess sludge from the settler and the chemical sludge from coagulation are combined, processed together, and dewatered using a screw press. The dewatered sludge is then disposed of by composting. The costs associated with the transport and composting of the dewatered sludge constitute a significant portion of the operating expenses.

The pilot model consists of the radial stainless-steel denitrification reactor with a volume of 40 L operated as CSTR, stirred by a Heidolph laboratory vertical stirrer. The reactor is followed by a stainless-steel cylindrical degassing tank stirred by the laboratory vertical mixer removing bubbles of waste gas. The sludge sedimentation in the subsequent separation stage was carried out in a vertical square settler made of plexiglass with 18 L. The wastewater is pumped into the pilot model by a Prominent Gama membrane pump, with the inflow pumping set to 3 L/h. The return sludge is also pumped by a Prominent Gama membrane pump. The outflow water is accumulated in the stainless-steel cylinder for sample collection. In the reactor, three online probes are installed: Nitratax (Hach), which measures the concentration of nitrates and nitrites using the UV absorbance principle, an LDO probe (Hach) for measuring dissolved oxygen concentration, and a pH probe (Gryf HB).

The aim of experiments was to establish conditions that closely resemble real-world scenarios in full-scale wastewater treatment plants, to ensure that the results are applicable and can potentially improve the efficiency and economics of the denitrification process. The pilot model has the same set-up as the full-scale plant and is operated using real wastewater, with the necessary amount sourced directly from the full-scale wastewater filtration station. The pilot denitrification reactor was inoculated with the denitrifying sludge obtained from the full-scale denitrification reactor.

The dosing of substrate and neutralizing agents was automatically controlled by the Gryf XBC system. Lime milk with a concentration of 5% served as the main neutralizing agent and was dosed to the denitrification reactor to maintain pH in the range of 6.6–6.8. As a backup neutralizing agent, a 10% solution of sodium hydroxide was prepared and its dosage was activated in the case of a drop in pH in the denitrification reactor below 6.5. Diluted methanol with a concentration of COD 44 g/L was used as a substrate and its dosing was regulated by a Nitratax probe to maintain the oxidized nitrogen concentration around 10 mg/L. When the concentration of N-NO_x exceeded 10.5 mg/L, methanol dosing was turned on and when it dropped below 9.5, it was turned off. If the immediate nitrogen loading rate was too high or the denitrification ability of the biomass was impaired, the oxidized nitrogen concentration in the reactor could increase uncontrollably and lead to substrate overdose. To mitigate this risk, the inflow pump was stopped when the oxidized nitrogen concentration in the reactor exceeded 13.5 mg/L and was restarted only when the nitrogen concentration fell below 9.5. The nitrogen loading rate was dependent on the concentration of nitrate in the wastewater and the operating time of the inflow pump as the pump flow rate was constant. The average nitrogen loading rate was 0.39 g L⁻¹ d⁻¹ with a standard deviation of 0.1 g L⁻¹ d⁻¹ and the sludge loading rate was 0.15 g g⁻¹ d⁻¹ with a standard deviation of 0.05 g g⁻¹ d⁻¹.

The excess sludge was removed by taking the exact volume of sludge mixture from the denitrification reactor in which suspended solids (SS) were analyzed. The volume of the removed sludge was adjusted so that the concentration of SS in the reactor was approximately 3 g/L.

Influent samples were taken daily for the determination of N-⁠, N-_, and acidity. Effluent samples were analyzed 3 times a week for parameters P-⁠, N-⁠, N-⁠, SS, and COD. Sludge samples from the denitrification reactor for the determination of SS and VSS were taken three times a week. Analyses were performed according to U.S. EPA methods.

Phosphoric acid was dosed into the denitrification reactor once a day, the presented phosphorus concentrations were initial in the dosing cycle. During experiments its dose was gradually reduced until the ceasing of denitrification, then the dosage was resumed, and again gradually increased as shown in Table 1.

The daily consumption of substrate and neutralizing agent, concentrations of the oxidized form of nitrogen, concentrations of consumed phosphorus, the sedimentation properties of the sludge, and the amount of excess sludge were monitored.

When decreasing the phosphorus dose, it was important to identify the point of this nutrient deficit. An indicative parameter of denitrification activity of biomass was the concentration of oxidized forms of nitrogen, which remained untreated even with the organic substrate. However, with pronounced fluctuations in nitrate concentration in the influent real wastewater, this may not be readily detectable.

When denitrification activity was high, nitrate-containing wastewater was continuously pumped into the reactor. However, as the denitrification rate decreased, the sludge could not remove all the supplied nitrates efficiently. This continued until the oxidized nitrogen concentration in the reactor exceeded 13.5 mg/L, at which point the inflow pump was automatically turned off. Then the system waited until the oxidized nitrogen concentration dropped below 9.5 mg/L, at which time the inflow pump was restarted. The volume of wastewater pumped to the reactor could be used as an indicator to estimate the denitrification capability of the biomass.

The most substantial decrease in wastewater inflow, which was controlled according to the oxidized nitrogen concentration in the reactor, occurred at the end of period 5, when phosphoric acid dosing was completely stopped. Upon starting the influent pump, the nitrate concentration in the reactor immediately exceeded the limit, prompting the pump to shut down. In the 190th day of the experiment, the reactor could not process even a relatively small nitrogen load. Concurrently, the organic substrate was no longer consumed causing the concentration to rise to 1,000 mg/L COD.

In experiment 6, phosphoric acid dosing was restored to the previous dose of 0.1 mL/d which represented a phosphorus concentration of 0.24 mg/L. The system recovered quickly, with nitrate concentration returning to the regulated value within 14 h. The decrease in COD was slower and nitrate concentrations returned to normal values within a week. The lowest dose of phosphoric acid (0.1 mL) representing the concentration of P 0.24 mg/L was applied to the denitrification reactor during experiments 4 and 6. The efficiency of denitrification was sufficient to fulfil the requirements of effluent quality. The value of phosphorus concentration for successful denitrification in these cases is a little lower, but close to 0.3 mg/L as recommended by Feng et al. (2023) as the lowest limit in the range of 0.3–0.6 mg/L P.

The green line denotes the start of substrate dosing and the red the end of dosing. Immediately after the start of dosing, the concentration of oxygen drops to a minimum and is followed by a decrease in the concentration of oxidized nitrogen. When the concentration of N-NO_x 9.5 mg/L is reached, the substrate dosing is switched off. After some time, the concentration of oxygen begins to slowly rise due to the diffusion of atmospheric oxygen through the open surface of the reactor until the next dose of substrate is added.

Experiments were also evaluated for the effect of the phosphorus dose to specific substrate consumption per N-NO_x removed. Table 2 presents data of the specific consumption of substrate per removed nitrate nitrogen and specific amount of P to removed nitrogen and COD. The removed nitrate nitrogen and consumed COD are calculated as the difference between daily mass inflow and outflow.

Many bacteria are able to accumulate organic substrates during a shortage of phosphorus (Song et al. 2024). This is often linked to their ability to survive in adverse conditions. In the absence of phosphorus, bacteria can limit growth and increase the accumulation of storage substances. Denitrifying bacteria of the genus Pseudomonas, Paracoccus, Bacillus and Achromobacter have a high need for phosphorus, which is crucial for their growth and metabolic functions. In the absence of phosphorus, they are able to accumulate organic substrates, which allows them to survive in adverse conditions and use the accumulated reserves as soon as conditions improve. The concentration of extracellular polymeric substances (EPSs) was related to the denitrification performance (Tang et al. 2024).

However, after slight restoration of P dosing from zero to 0.03 g/day in the experiment 6, the ratio of COD/N reverted to the similar values of experiment 2 but with P dose only 0.13 g/d. After restoring the denitrification capacity of organisms, also the storage compounds were utilized. Specific COD consumption further decreased when the specific dose of phosphorus was increased. From results of experiments can be concluded that specific substrate demand depends on availability of phosphorus and increases upon lack of phosphorus in consequences with the species composition of bacterial biomass. Bai et al. in their experiments found out that 15.8% of nitrate nitrogen were removed by denitrifying phosphorus accumulating organisms. 16S rRNA gene analysis and stoichiometric ratios indicated the system was rich in phosphorus accumulating organisms Dechloromonas and Ca Accumulibacter (Bai et al. 2022)

The ratio of available phosphorus to organic substrate as COD is in the literature reported as 0.0013 g/g P/COD (Boltz et al. 2012). In our case 0.0006 g/g was apparently enough, but it due to the length of the experiments supported changing the composition of the denitrification culture to species with lower specific requirements for the amount of available phosphorus, as stated by (Wang et al. 2018; Feng et al. 2023).

The further objective of experiments was to determine the effect of the phosphorus dosage on the specific sludge production, expressed as the ratio of the amount of produced organic dry matter to the amount of removed COD or removed N-NO_x. During experiments, the amount of sludge production, the amount of sludge leaving the effluent, the concentration of SS in the influent, and the organic part of sludge were accurately measured. Furthermore, the sludge production was corrected for the difference between the initial and final dry matter concentration in the reactor for each individual step of the experiment. The obtained data are presented in Table 3.

The calculations of specific sludge production are represented in Figure 7.

The reduction in the phosphoric acid dose to 0.1 mL/d resulted in a decline in sludge production. After stopping dosing, there was a further significant decrease in biomass production, associated with the denitrification collapse. After the resumption of phosphorus dosing, biomass growth restarted. However, at a level of specific production observed at the beginning of the test with a higher dose of phosphorus. Given the duration of the experiment the reason could be a higher representation of organisms tolerant to low phosphorus concentrations, which initiated deferred reproduction after the restoration of phosphorus supply and utilization of stock substances (Samaddar et al. 2019). Available phosphorus influenced the composition of the mixed-denitrification culture that are able to adapt to changes in conditions, resulting in the different representation of individual species with specific phosphorus requirements (Xu et al. 2015; Fan et al. 2018).

From the measured data, we attempted to describe the development of phosphorus concentration in the organic fraction of sludge using daily balances of incoming and outgoing phosphorus. This involved the amount of phosphorus leaving in the outflow in the form of dissolved phosphate phosphorus for each day and also the amount of phosphorus leaving the denitrification reactor with excess sludge and in an undissolved form in the outflow. Based on the sludge amount in the reactor, it was possible to calculate the concentration of phosphorus in the organic dry matter of the sludge.

The initial concentration of phosphorus in the sludge was determined as the average value of the ratio of removed phosphorus to the production of organic sludge dry matter during the first test period when the dose of phosphoric acid was 1 mL/d. Considering the concentration of phosphorus in the effluent during this period, it was evident that there was an adequate surplus of phosphorus. Therefore, the newly formed biomass should contain the same amount of phosphorus as all existing biomass in the reactor.

At the beginning of the experiments, during the high phosphorus dose of 0.27 g/d, the concentration of phosphorus in the organic dry matter of the sludge ranged between 5 and 6%. After reducing the phosphorus dose to half, the concentration in the sludge did not decrease immediately, but phosphorus stopped leaving into the effluent. Practically, all of it was utilized by the biomass. After further decreasing the dose of phosphoric acid to 0.25 mL/d, there was a rapid decrease in the average concentration of phosphorus in the biomass, accompanied by an increase in sludge production. After further reducing the dose, the concentration of phosphorus in the biomass stabilized somewhere below 2%, accompanied by a decrease in biomass production.

Upon stopping the dosing, the phosphorus concentration in the biomass quickly dropped to less than 1%. The minimum value of 0.6% was reached just before the collapse of the denitrification capacity of the sludge. Sludge production was also minimal during this period.

After restoring the dosage of phosphoric acid with a dose of 0.1 mL/d, the concentration of phosphorus in the biomass remained at around 0.8%. These conditions, however, might not be as uncomfortable anymore, as the sludge quickly restored its denitrification capacity and reproduction. After a further increase in dosage, dissolved phosphorus was even found in the effluent, indicating that the sludge is unable to utilize such an amount of phosphorus as at the beginning of the experiment. Nevertheless, its specific production and denitrification ability remained practically the same. This would confirm a higher proportion of organisms with lower phosphorus requirements.

A continuous denitrification pilot model was operated using the real wastewater influent and methanol as the denitrification substrate. Experiments were conducted with specific phosphorus dosages varying from 0 to 16.8 mg P/g N-NOx.

The minimum phosphorus concentration of 0.24 mg/L, corresponding to a specific dose of 1.8 mg P/g of removed oxidized nitrogen, was sufficient for successful denitrification efficiency. However, this was observed after 70 days of operation, allowing adequate time for the denitrifying culture to adapt to low phosphorus concentrations in the reactor.

The identified minimum safe dose of phosphorus was approximately 1 mg P per g of introduced COD, which is lower than the values reported in the literature. Under long-term low phosphorus conditions, selected organisms were satisfied with a phosphorus content of around 1% in their organic dry matter, compared to the initial 5%.

The results indicated that potential operational savings could be achieved by reducing phosphoric acid demand although a limited phosphorus supply did not decrease specific sludge production.

R.S. contributed to methodology, data curation, formal analysis, investigating, visualization, writing original draft.

J.Z. contributed to conceptualization, supervision, review & editing.

The study was supported by Kralovehradecka provozni, a.s. Hradec Kralove, Czech Republic and Synthesia Pardubice, Czech Republic.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.