The effect of PO₄^3--P concentration on sludge settleability and nutrients removal performance

The activated sludge process was first invented in the UK, and has been widely used in urban and industrial sewage treatment all over the world (Wanner 2021). However, the frequent occurrence of sludge bulking during its operation will cause the loss of sludge flocs and fluctuation of treatment performance (Shen et al. 2019). In the early stage, activated sludge process was mainly used to remove the organic matter. However, the aerobic environment with biodegradable organics is very easy to stimulate the growth of filamentous bacteria, especially under high OLR condition (no matter flocs or granules). Yang found that sludge volume index (SVI) increased sharply when OLR was increased from 0.45 to 0.75 kg COD·(kg MLSS·d)⁻¹, followed by the disintegration of sludge flocs (Yang et al. 2018). Hamza found that when the OLR was bigger than 1.5 kg COD·(kg MLSS·d)⁻¹, even the sludge settleability of aerobic granular sludge would deteriorate, and serious sludge bulking occurred if the OLR was greater than 2.5 kg COD·(kg MLSS·d)⁻¹) (Hamza et al. 2018).

With the improvement of water environment restoration in the last decade in China, the discharge standard of first-level A (GB18918-2016) and stricter standard (especially for the nutrients to induce eutrophication) have been widely implemented in most wastewater treatment plants (WWTPs). In order to enhance nutrients removal, the OLR of the treatment process such as AAO has been decreased to 0.05–0.15 kg COD·(kg MLSS·d)⁻¹. Compared with biological nitrogen removal, the biological phosphorus removal is not only difficult to increase removal performance, but also has a closer relationship with sludge settleability. Firstly, phosphorus content could influence the density of activated sludge, and its slight change will significantly affect the settleability of activated sludge whose density is close to water (Schuler & Jang 2007). Secondly, filamentous bacteria are more competitive to acquire nutrients in phosphorus-deficient condition, which affects the activity of PAO. Thirdly, phosphorus concentration could affect the metabolic activity, floc structure, flocculation characteristics of activated sludge, etc. At present, due to the organic pollution of groundwater and the dilution and scour of rainwater, the influent COD and -P of WWTPs in various regions are mostly in the range of 200–350 mg·L⁻¹ and 4–15 mg·L⁻¹, respectively (Heinonen et al. 2013; Carrey et al. 2020). Under such long-term changes of OLR and C/P ratio, some studies have found that the reaction duration (anaerobic and aerobic) can be adjusted to control sludge settleability and enhance biological phosphorus removal. Zhang reduced the cycle duration of SBR, the sludge settling ability was efficiently improved and the system showed excellent phosphorus removal ability (Zhang et al. 2016). Mandel suggested that aeration phase duration of 60 min was the most suitable for maintaining -P removal through his experiment (Mandel et al. 2019).

Based on the above introductions, the effects of OLR and influent -P concentration on filamentous sludge bulking, nutrients removal, and organics utilization were investigated through altering reaction duration in this study. It also provides theoretical and practical guidance for WWTPs to control sludge settleability.

The volume of inflow synthetic wastewater was 3 L per cycle and its composition was shown in Table 2. The inoculated sludge was obtained from the anaerobic tank of a lab scale anaerobic/aerobic (AO) reactor in our laboratory. The inoculated sludge was in filamentous bulking state with the sludge volume index (SVI) and MLSS of 470 mL·g⁻¹ and 2,000 mg·L⁻¹, respectively. It was mixed with purified water with a ratio of 1:1, followed by 2.0 h of sedimentation. Then the supernatant with the same volume of the purified water was discharged. This procedure was repeated three times to wash the inoculated sludge clean. Then the sedimentary sludge was distributed into three SBRs evenly.

Table 2

The composition of synthetic wastewater

Water quality indicators .	COD .	-N .	-P .	Alkalinity .	Microelements .
Items	CH3COONa·3H2O	NH4Cl	KH2PO4	NaHCO3	CaCl2·2H2O	MgSO4	Nutrient solution
Dosage (mg)	663.80	166.90	18.80–75.20	375	40	80	0.3(mL)
Concentration (mg·L−1)	326.90	43.7	4.3–17.2	400	–	–	–

Water quality indicators .	COD .	-N .	-P .	Alkalinity .	Microelements .
Items	CH3COONa·3H2O	NH4Cl	KH2PO4	NaHCO3	CaCl2·2H2O	MgSO4	Nutrient solution
Dosage (mg)	663.80	166.90	18.80–75.20	375	40	80	0.3(mL)
Concentration (mg·L−1)	326.90	43.7	4.3–17.2	400	–	–	–

The composition of nutrient solution was 1.5 g of FeCl₃·6H₂O, 0.15 g of H₃BO₃, 0.03 g of CuSO₄·5H₂O, 0.18 g of KI, 0.12 g of MnCl₂·4H₂O, 0.06 g of Na₂MoO₄·2H₂O, 0.12 g of ZnSO₄·7H₂O, 0.15 g of CoCl₂·6H₂O and 10 g of EDTA per liter.

View Large

Samples for analysis were filtered with 0.45 mm filter to separate suspended solids from the liquid. MLSS, MLVSS (mixed liquid volatile suspended solids), SV (sludge settling velocity), SVI, CODcr, -N, -N, -N and -P were measured according to standard methods (APHA 2005). Total inorganic nitrogen (TIN) was regarded as the sum of -N, -N, -N (He et al. 2016). The morphology of microbe was observed by optical microscope (OLYMPUSBX51). The method proposed by Ed was used to determine the abundance index of filamentous bacteria FI (Eikelboom 2000), the method proposed by Zhou was used to determine the total phosphorus of activated sludge (Zhou 2005). DO, pH and ORP were measured by WTW Multi 3401 DO tester. Specific oxygen utilization rate (SOUR) was measured by breathing method. Pyrosequencing analysis was carried out to investigate microbial species and abundance by the Bio-engineering (Shanghai) Co., Ltd. The samples were taken from the inoculated sludge and the sludge at the last cycle of each SBR (named as N0, N1, N2 and N3).

When the influent -P were 4.3 mg·L⁻¹ and 8.6 mg·L⁻¹, increasing the OLR through reducing the aerobic duration improved the sludge settleability significantly. When the aerobic duration decreased from 2.0 h to 1.5 h (stage 1 and stage 2), the average variation rates of SVI in SBR1 and SBR2 increased from 1.79 mL·(g·cycle)⁻¹ and 1.74 mL·(g·cycle)⁻¹ to 2.28 mL·(g·cycle)⁻¹ and 2.12 mL·(g·cycle)⁻¹, respectively. The reason was most filamentous bacteria grown under aerobic environment in the presence of COD (De Kreuk et al. 2006). When the influent -P was 17.2 mg·L⁻¹, the average variation rate of SVI decreased from 1.92 mL·(g·cycle)⁻¹ to 1.26 mL·(g·cycle)⁻¹. The reason was that the PAOs absorbed most short chain fatty acids (SCFAs) during anaerobic period, and only limited COD was left in the aerobic period (Yang et al. 2017). In addition, the shearing force during aeration period decreased with the shorten of the aerobic duration, which was not conducive to form the dense sludge flocs (Sousa et al. 2018).

Increasing the OLR through decreasing the anaerobic duration had limited effect to improve sludge settleability. When the anaerobic time decreased from 1.0 h to 0.5 h (stage 2 and stage 3), the average variation rates of SVI decreased from 2.28 mL·(g·cycle)⁻¹, 2.12 mL·(g·cycle)⁻¹ and 1.26 mL·(g·cycle)⁻¹ to 0.63 mL·(g·cycle)⁻¹, 0.64 mL·(g·cycle)⁻¹ and 0.52 mL·(g·cycle)⁻¹, respectively. This was because the PAOs could uptake the SCFAs to convert poly-β-hydroxybutyrate (PHB) during anaerobic period (Hansaem & Yun 2014). When the anaerobic duration was decreased, it was difficult for PAOs to uptake SCFAs sufficiently, and the residual COD would be used by filamentous bacteria in following aerobic period (Pronk et al. 2015).

During the second stage, the OLR increased from 0.26 to 0.33 kg COD·(kg MLSS·d)⁻¹ with the aerobic duration decreased from 2.0 h to 1.5 h (shown in Figure 3). The sludge activities were all improved, and the average SOUR increased from 2.95, 2.93 and 2.71 (mg O₂/g MLSS·h)⁻¹ (first stage) to 4.81, 5.72 and 4.56 (mg O₂/g MLSS·h)⁻¹ (second stage). Obvious -N accumulation was observed in the first stage, and the short cut nitrification changed to complete nitrification in the second stage, which would increase the amount of COD consumed by denitrification (Rocher et al. 2015). In theory, to reduce 1 g of -N and -N to N₂, 8.60 g and 5.14 g COD would be consumed (Peng et al. 2010). Taking the typical cycle for example, the amount of COD consumed for denitrification in SBR3 increased from 200.91 mg (first stage) to 496.63 mg (second stage). While there was only 980.7 mg COD in the influent wastewater, so the phosphorus releasing process would be inhibited (Xie et al. 2018). Since the influent -P concentration of SBR1 and SBR2 were little, the above effect was limited.

During the third stage, the OLR increased to 0.40 kg COD·(kg MLSS·d)⁻¹ with the anaerobic duration decreased from 1.0 h to 0.5 h, and the average SOUR rose up to 7.58 mg O₂·(g MLSS·h)⁻¹, 5.80 mg O₂·(g MLSS·h)⁻¹ and 4.66 mg O₂·(g MLSS·h)⁻¹, respectively. The -N accumulation occurred in the first two stages almost disappeared, and the competition for organics between denitrification and phosphorus release was more serious (Zhang et al. 2020). Consequently, the phosphorus removal performance all deteriorated, especially for SBR3. The average phosphorus contents of sludge in each SBR were 31.56 mg·g⁻¹, 46.84 mg·g⁻¹ and 52.46 mg·g⁻¹ in the first stage, and they increased to 39.86 mg·g⁻¹, 54.41 mg·g⁻¹ and 56.44 mg·g⁻¹ in the second stage, respectively. It was strange although the phosphorus removal of SBR3 deteriorated, the phosphorus content of sludge increased, the reason was the crystals formed by -P and cations would adhere to the sludge flocs (Angela et al. 2011). This verified the explanation for the variations of sludge settleability in section 3.1. During the third stage, the phosphorus contents in the sludge of each SBR decreased to 36.75 mg·g⁻¹, 54.38 mg·g⁻¹ and 53.32 mg·g⁻¹, respectively, which were consistent with the phosphorus removal performance.

During the second stage, as the sludge activity increased with the increase of OLR, the SOUR of ammonia oxidizing bacteria (AOB) and nitrite oxidizing bacteria (NOB) both increased significantly (Caluwé et al. 2017; Dong et al. 2017). From the end of first stage to the initial of second stage, the average SOUR_(AOB) increased from 1.20, 1.13 and 0.85 mg O₂·(g MLSS·h)⁻¹ to 2.03, 2.40 and 1.30 mg O₂·(g MLSS·h)⁻¹, respectively. The average SOUR_(NOB) increased from 0.77, 0.89 and 0.71 mg O₂·(g MLSS·h)⁻¹ to 1.25, 1.30 and 0.98 mg O₂·(g MLSS·h)⁻¹, respectively. The effluent TIN of each reactor decreased from the first stage to the second stage due to the enhancement of microbial assimilation. It was found that SBR1 has been completely nitrated at the end of the first stage, and the effluent -N was almost 0 (shown in Figure 6(a)). As a result of the overall decreasing trend of TIN, the effluent -N of SBR1 has a downward trend at the beginning of the second stage. At the same time, the -N accumulation that still existed in SBR2 and SBR3 during the second stage were also weakened.

During the third stage, the average effluent TIN of each SBR decreased from 6.99 mg·L⁻¹, 6.95 mg·L⁻¹ and 7.08 mg·L⁻¹ to 6.65 mg·L⁻¹, 6.73 mg·L⁻¹ and 6.34 mg·L⁻¹, respectively. As the assimilation effect was enhanced with the increase of OLR, the average MLSS of SBR1 and SBR2 increased from 2,325 mg·L⁻¹ and 2,520 mg·L⁻¹ to 2,402 mg·L⁻¹ and 2,587 mg·L⁻¹. But the average MLSS of SBR3 dropped from 2,677 mg·L⁻¹ to 2,582 mg·L⁻¹. It could be seen that assimilation was the main reason for the decrease of effluent TIN except for SBR3 (Phatak et al. 2016). This was because the sufficient nitrification in SBR3 produced the most -N, and the initial -N concentration in the following cycle was high (5.7–5.9 mg·L⁻¹). This was benefit to enriching DPAO (Sousa et al. 2018).

Both denitrification and phosphorus release require organic matter as electron donors (Li et al. 2019). Considering the assimilation, 8.67 g COD and 5.20 g COD are required to completely denitrify 1 g -N and 1 g -N into N₂ (Peng et al. 2010). In the process of phosphorus release, PAO absorbed sodium acetate to release -P. This study used acetate as the organics to calculate the COD. To release 1 g of -P during anaerobic period, 1.2 g COD would be used.

According to the above calculations, it was found the sum of COD used for denitrification and phosphorus release sometimes exceeded the influent COD, this was because DPAO which could use -N as electron acceptor to absorb -P was enriched (one carbon two usage) (Hansaem & Yun 2014). The percentages of COD used by DPAO all increased during the second stage, and decreased during the third stage. Increasing the influent -P was beneficial to enriching DPAO (Wang et al. 2014), the percentages of COD used by DPAO in SBR3 was significantly higher than those of SBR1 and SBR2. This indicated that DPAO would be accumulated under the condition of high -P concentration and sufficient nitrification, this was also the reason why the effluent TIN decreased in SBR3.

In the level of genus, the main PAO in the inoculated sludge was Candidatus Accumulibacter with a relative abundance of 0.223%, and it decreased to 0.217%, 0.162% and 0.136% at the end of the experiment, respectively (shown in Figure 8). In the second and third stages, more COD was used to denitrify the -N during the anaerobic period, this was the main reason for the decrease of the abundance of PAO (Rubio-Rincon et al. 2017). The DPAOs in the inoculated sludge mainly consisted of Thauera and Flavobacterium with a total relative abundance of 0.96%, and it increased to 1.21%, 3.42% and 2.13% at the end of the experiment, suggesting that increasing influent -P was beneficial to enriching DPAO (Wang et al. 2018). Although the relative abundance of DPAO of SBR3 is lower than that of SBR2, it can be seen from Figure 5 that the ratio of COD used by DPAO of SBR3 was the biggest, which might be because the activity of Thauera and Flavobacterium were bigger under higher -P condition (Fu et al. 2019).

There were four kinds of denitrifying bacteria (DNB) in the inoculated sludge (Lysobacter, Diaphorobacter, Roseomonas and Legionellaceae), among which Lysobacter accounted for the highest proportion and was the dominant species with a relative abundance of 0.397%, and it increased to 0.404%, 0.422% and 0.414% at the end of experiment, respectively. -N accumulation occurred in the first two stages, while the COD consumed by -N denitrification was only 60% of that by -N denitrification. As the -N accumulation in SBR1 was less than the others, the COD used for denitrification was the highest (Sahinkaya et al. 2011). Therefore it could be seen from Figure 7 that the percentage of COD used for denitrification in the second stage of SBR1 reached 51.05%.

AOB of the inoculated sludge consisted of Nitrosococcus and Nitrosomonas with a total relative abundance of 0.49%, and it increased to 0.53%, 0.53% and 0.51% at the end of experiment, respectively. Nitrospira was the only NOB bacterium observed; its relative abundance in the inoculated sludge and the end of experiment were 0.39%, 0.67%, 0.53% and 0.52%, respectively. It could be found that the average SOUR and the relative abundance of AOB and NOB increased with the increase of OLR (Li et al. 2020). Although the relative abundance of both AOB and NOB were bigger than those of the inoculated sludge, the NOB was inhibited with the increase of influent -P concentration, and this was the reason why -N accumulation was more obvious in SBR2 and SBR3 (Wang et al. 2016).

The effects of OLR and influent -P on sludge settleability, nitrogen and phosphorus removal were studied in three lab-scale SBR reactors using synthetic domestic wastewater. We found that when the influent COD was constant, increasing influent -P concentration could not absolutely enhance the sludge settleability. When the influent -P concentration were 4.3 mg·L⁻¹ and 8.6 mg·L⁻¹, increasing the OLR by reducing the aerobic duration was beneficial to improve sludge settleability, and the short-cut nitrification would change to complete nitrification. Meanwhile, the denitrification would compete with phosphorus removal for carbon source, which would deteriorate the phosphorus removal performance. We also observed that DPAO were enriched under the high -P concentration and sufficient nitrification, thus saving carbon source.

Not applicable.

Not applicable.

Not applicable.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Zhaoxu Peng: Methodology, Formal analysis, Writing–review, Resources. Ju Wang: Methodology, Data formal analysis, Writing–original draft. Ningqi Niu: Formal analysis, Writing–review. Ao Liu: Methodology, Writing–review. Yongqing Niu: Part of the experiment. Jing Qin: Part of the experiment. Minghui Liu: Part of the experiment. Ying Li: Part of the experiment.

This research was supported by the National Natural Science Fund (42107427), and the Science and Technology Foundation of Henan Province (222102320426).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.