In the experiment, oil wastewater from a WWTP located in Erdos of China was collected and stored at 4°C prior to testing. Analytical reagent grade Ferrous gluconate (Fe≈222 mgL⁻¹), glucose, and other chemical reagents used in the experiment were obtained from Sigma-Aldrich (MO, USA).

The WWTP was located in Erdos of China. Continuous treatment of oil wastewater was performed by separation, dissolved air flotation (DAF), and biodegradation, with DAF No. 1 and No. 2 operating in series. E.1 & 2 and E.S.1 & 2 were the effluent and effluent supernatant of DAF No. 1 and No. 2, respectively. Based on the analysis of in-situ data (see Table 1), the dissolved oxygen in the bioreactor experiment was set to three gradients. Experimental Design for Evaluating the Effects of Different DAF Effluents on Sludge SV (see Table 2)

(1) To obtain Fe concentrations of 0.2, 0.5, 0.8, 1.1, and 2.0 mgL⁻¹, 50 ml of varying concentrations of ferrous gluconate were added to 100 ml of sludge respectively. (2) To obtain oil contents of 68, 145.33, 245.33, 345.33, and 445.33 mgL⁻¹, 50 ml of different concentrations of contaminated oil were added to 100 ml of sludge respectively. (3) Glucose was added to 128, 104, 85, 68, and 55 ml of Milli-Q water, with a chemical oxygen demand (COD) of 68 mgL⁻¹. Activated sludge was placed in a cylinder for 2 h to obtain concentrated sludge (3,880 mgL⁻¹), and subsequently, 100 ml of sludge was mixed with the glucose solution until the concentrations reached 1,700, 1,900, 2,100, 2,300 and 2,500 mgL⁻¹. (4) Eight liters of sludge were removed from the bioreactor. After 2 h of aeration, the DO was maintained at 6–7 mg L⁻¹. One-liter sludge samples were extracted from the bioreactor and placed in three conical flasks. One flask was kept under continuous aeration, another was left standing to maintain the DO at 0.2–1 mgL⁻¹, and the last flask was agitated to maintain the DO at 3–4 mgL⁻¹.

Several experiments were conducted on the changes of SV over time, revealing that there was almost no change after 12 h. Therefore, the cultivation time was set to 12 h (see Supplementary Figure 4).The SV of the sludge was measured after the mixture was shaken at 12–13 rpm for 12 h. For gene detection, 1 ml of sludge was used. Unless otherwise specified, the sludge concentration was always approximately 2,683 mgL⁻¹.

The second-generation gene sequencing method based on 16S rRNA was used for analyzing the microbial community structure. Samples were entrusted to Beijing Kinko Biotechnology Co for testing.

The extraction of bacterial DNA was performed in accordance with the kit instructions (AG Acrey Biotechnology). As shown in Table 3, quantitative real-time polymerase chain reaction (RT-PCR) (Applied Biosystems StepOnePlus Real-Time PCR System, USA) was used for gene PCR amplification of zoogloea and total bacteria.

Zoogloea were classified into Zoogloea ramigera, Zoogloea resiniphila, and Zoogloea caeni. Zoogloea resiniphila.isolate SB3 is a representative bacterial species (An et al., 2016). PepA in the PEP-CTERM protein gene has a high level of transcription and is involved in the formation of Zoogloea resiniphila. Its expression was used to build a floc structure that allowed sludge settling (Gao et al., 2018). Therefore, the pepA gene of Zoogloea resiniphila.isolate SB3 was used in the primer design. Thus, qPCR specific primers were used to detect the expression of pepA in sludge bulking. The sequences of pepA primer are displayed in Table 4.

To date, the 16S rRNA gene sequence, which is found in virtually all bacteria, has been the most commonly used to study bacterial phylogeny and taxonomy (Janda and Abbott, 2007). Therefore, the 16S rRNA gene was used in the primer design. The sequences of 16S rRNA primer are displayed in Table 5. In addition, the NCBI Reference Sequences of the two genes were WBTZ01000035.1 and NZ_CP032515.1:422791-425865, respectively. To confirm the reproducibility of the experimental results, all experiments were performed over three iterations. Unless otherwise stated, all SVs appearing in this article are SV₃₀. In addition, the unit of SV is %.

Due to the predominance of filamentous fungi and their small proporting in activated sludge, they does not exert a significant impact (Lu et al., 2023a). The bacterial community of activated sludge was analyzed using 16S rRNA. As shown in Figure 1A, the main bacteria were Betaproteobatia (15%), Gammaproteobatia (10%), and Bacteroides (7%). Deltaproteobacteria and Epsilon proteobacteria accounted for 5%, while Pedosphaera, Flavobacteria, and Alphaproteobacteria accounted for 3% of the total. Collectively, these bacteria accounted for 51% of the total population. A total of 9 representative strains were selected from each group, with Betaproteobactia including Verminephrobacter aporrectodeae subsp. tuberculatae (34%) and Zoogloea resiniphila TS9 (24%). The matching results between pepA primers and the genes of each representative strain are shown in Figure 1B. Because high-level transcription of the pepA gene was involved in the formation of Zoogloea (see details in Section 2.3), the macrogenomes of strains were compared with the pepA gene in the National Center of Biotechnology Information (NCBI) gene library. Neisseria meningitidis 1076, Rickettsia rickettsii str. Iowa, Desulfovibrio desulfuricans, Helicobacter pylori 10700, Trypanosoma grayi, Flavobacterium columnare, and Bacteroides fragilis lack the pepA gene, with completely mismatched genes. The gene annotation of Escherichia coli str. K-12 substr. MG1655 revealed the presence of a pepA gene fragment that was only 50% matched with the pepA gene of Zoogloea resiniphila TS9. Moreover, the pepA genes of the two bacteria were distantly related. Considering these aspects, this study employed the pepA gene of Zoogloea resinifila TS9 as a reference for primer design, which was specific to the genomes of other strains in the sludge. Furthermore, 16S rRNA primers were used to count the total bacterial number (Qin et al., 2022).

As shown in Figure 2, the concentration of inoculated sludge decreased in the range of 1,700–2,500 mgL⁻¹, which substantially reduced the SV. The measurement of the PCR cycle threshold (Ct) value served to indicate the number of cells, where higher cell concentrations corresponded to a lower number of required cycle thresholds. An increase in pepA Ct value indicates a decrease in pepA gene expression (Spencer et al., 2016; Rehman et al., 2022). The Ct value corresponding to the total number of bacteria increased considerably as sludge concentration decreased, whereas the Ct value corresponding to pepA decreased gradually. The standard curve of Ct value vs. cell numbers is shown in Supplementary Figure 3, based on the assumption that each cell only contains one set of pepA genes. This indicated that the number of zoogloea increased gradually, and the decrease in the number of filamentous bacteria was relatively faster, in accordance with the basic principle of the decrease in the total number of bacteria. This resulted in a reduction in sludge flocs, improved sedimentation performance, and SV recovery. The increase in sludge concentration reduced the amount of substrate available to a single bacterium, whereas filamentous bacteria have the advantage of growing at low substrate concentrations and reproducing in large numbers. Moreover, a rise in sludge causes an increase in flocs and a spike in resistance in the settling process, resulting in a decrease in sedimentation speed and an increase in SV₃₀ (Martins et al., 2003). Thus, reducing the sludge concentration can lower the SV and improve sludge settling performance. All heavy metals are toxic to bacterial life and inhibit microbial processes at medium and high concentrations (Buaisha et al., 2020). The selective discharge of sludge reduces iron enrichment. To improve the settling performance, the sludge concentration should be controlled within the range of 1,700–2,500 mgL⁻¹. However, the SV of normal sludge should be controlled at 15–30%, and the SV can only be adjusted to approximately 40% by regulating the sludge concentration. Because of the demand for bio-concentration for degradation, sludge cannot be discharged indefinitely, hence, alternative strategies must be determined to control the sludge SV.

Iron is essential for life and is pivotal for important metabolic reactions such as the synthesis of hemoglobin, ferrithionein, and other ferritin. A suitable quantity of Fe²⁺ can be employed as a bridge ion to build a network with extracellular polysaccharide polymers (Ren et al., 2018). The extracellular polymer of zoogloea contains a large number of anion groups (Dugan and Pickrum, 1972). These anion groups can combine with Fe to form complexes that can be removed (Dugan, 1987). Fe²⁺ is beneficial for the growth of sludge. For example, Fe is necessary for the synthesis of ferrithione and cytochrome enzymes, which are associated with electron transfer in the respiratory chain and are crucial for microbial respiration. However, excess Fe can inhibit the respiration of sludge (Ma et al., 2015) and potentially cause the destruction of sludge cell structures, flocs, and extracellular polymers (Zhang et al., 2019). According to the analysis shown in Figure 3, the Fe content increased from 0.2 to 2 mgL⁻¹, and sludge SV increased from 31% to 41%. These values were directly proportional to one another.

The Ct value of total bacteria represented by 16S rRNA declined as Fe content rose, whereas the Ct value of zoogloea increased. This suggests that the number of filamentous bacteria gradually increased, while the number of zoogloea gradually decreased, resulting in poor sludge settling performance and increased SV. The excessive addition of heavy metal ions reduces the content of polysaccharides and proteins in sludge bacteria, resulting in a reduction in the extracellular polymeric substances (EPS) content. The abundance, richness, and diversity of microorganisms was reduced because the loss of EPS protection initiates the death and decrease of certain zoogloea (Tsai et al., 2005). Filamentous bacteria proliferated, but the relative abundance of hydrophobic amino acids used in protein synthesis was considerably reduced (Li et al., 2020). Proteins serve as protective layers on the surface. Due to the decrease in hydrophobicin and the weakening of the protective layer, a relatively substantial quantity of Fe may enter filamentous bacteria to promote growth. Sludge bulking is caused by a reduction in zoogloea bacteria and an increase in filamentous bacteria. Therefore, iron enrichment indirectly causes sludge bulking due to the inhibition of zoogloea life activities. Iron is a key component influencing activated sludge SV and its concentration should be maintained below 0.6 mgL⁻¹.

As shown in Figure 4A, the oil content increased from 68 to 445.33 mgL⁻¹, and the sludge SV was stable at approximately 32%. The lowest SV was observed at the oil content of 145.33 mgL⁻¹. The surface volume ratio (A/V) of filamentous bacteria is greater than that of non-filamentous bacteria, offering advantages in terms of growth when the oil content is limited (Martins et al., 2003). Therefore, when the oil content was 68 mgL⁻¹, the Ct value of pepA primer was the highest, and the corresponding number of zoogloea was the lowest. The Ct value of the 16S rRNA primer was the lowest, and the increase of total bacterial population was primarily determined by the growth of filamentous bacteria. This ultimately resulted in poor sludge settling performance. When the oil content increased to 145.33 mgL⁻¹, the Ct value of pepA primer decreased slightly, indicating that the zoogloea grew marginally. The Ct value of the 16S rRNA primer increased rapidly, and the reduction of the total bacterial population was dependent on the slow growth of filamentous bacteria. However, the bacteria fed with high substrate concentration exhibit a high saturation constant K_s and a maximum growth rate μ_max (Chudoba et al., 1973). When the oil content exceeds 145.33 mgL⁻¹, previous experiments have shown that more iron is released into water (Qin et al., 2023). This may be the reason for inducing a decrease in total cell number and a slight increase in the number of zoogloea. The SV was slightly higher than 30%, suggesting a slight viscous bulking. This is because a large number of hydrophilic and highly viscous substances are produced by the proliferation of zoogloea. Microbes absorb a wide range of organic substances but are unable to metabolize them. A large number of highly viscous polysaccharides accumulate outside the cells, causing a rapid increase in cell surface attachments and making it difficult to settle, resulting in viscous bulking (Shi et al., 2022). Thus, the oil acts as a high concentration matrix that may lead to sludge viscous bulking.

As shown in Figure 4B, sludge SV decreased initially and then increased when DO increased from 0.6 to 6.5 mgL⁻¹. The decline in the Ct value of zoogloea was greater than that of total bacteria, indicating that both populations were proliferating, and zoogloea might be the primary driver of total bacteria growth. The number of sludge bacteria peaked at 6.5 mgL⁻¹ DO. Although low DO inhibited the growth of microorganisms, filamentous bacteria easily absorbed DO due to their large surface area and growth advantages (Lau et al., 1984; Martins et al., 2003). Sludge with DO < 1.5 mgL⁻¹ did not flocculate and exhibited poor sedimentation performance. Sludge with DO > 2 mgL⁻¹ had a large particle size and optimal sedimentation performance (Sürücü and Çetin, 1990). Notably, a DO of 3.5 mgL⁻¹ is beneficial for aerobic sludge to update iron in the body (Chen et al., 2018). To offer the sludge microbes a growth advantage, the DO must be enhanced. Therefore, the minimum SV was 40% when the DO was 3.5 mgL⁻¹. However, excess DO indicates strong aeration and flow shear force, which operate on the larger flocs in the sludge to cause its collapse, decrease the sedimentation performance, and increase the SV. In summary, maintaining the oil concentration at 145.33 mgL⁻¹ and the DO at 3.5 mgL⁻¹ was conducive to controlling the sludge SV.

The closer the absolute value of the coefficient is to 1, the stronger the correlation (Table 6). Therefore, SV was strongly correlated with the Fe and sludge concentrations, moderately correlated with DO, and weakly correlated with oil content. The order of correlation with SV was as follows: sludge concentration > Fe amount > DO > oil content.

Adopting the principles of sludge SV management outlined herein, the operation effect in the WWTP is depicted in Figure 5. Figure 5A shows the continuous treatment of wastewater with an MLSS of 13,080 mgL⁻¹ by DAF No. 1 and No. 2. The effluent supernatant was used for sludge culture for 12 h, and its SV ranged between 30 and 35%. The SV of sludge cultured in MilliQ water was 30% with minimal variation. However, the SV of sludge cultured in effluent treated by air floaters No. 1 and No. 2 rose dramatically, reaching 56.33% and 41.33%, respectively. In summary, influent water quality, environmental factors, and operation can all contribute to sludge bulking (Guo et al., 2014). The MLSS of the effluent from the floater impacts the sludge SV. Due to the excessive levels of heavy metal ions, such as iron ions, in oil wastewater, it affects the formation of sludge. Moreover, the SV of effluent sludge from air floaters 1 and 2 increased by 21% and 11%, respectively. The water from floater 1 had a greater capacity to improve SV than that from floater 2. In addition, a low F/M ratio, high or low DO, suspended residence time (SRT), carbon to nitrogen ratio (C/N), pH, sludge concentration, and other factors in sludge culture contribute to bulking (Zaidi et al., 2022). Because floaters 1 and 2 were linked in series, the Fe and COD concentrations in the No. 2 air floater influent were lower than those of the No. 1 floater. However, the No. 2 floater exhibited a more visible control of SV change; hence, we believe that Fe concentration played a major role in sludge bulking.

As shown in Figures 5B, C, the influent sludge, Fe content, and DO were all controlled for 8 consecutive days. The sludge concentration decreased from 4,660 to 1,918 mgL⁻¹, the Fe concentration decreased from 1.35 to 0.34 mgL⁻¹ and the DO decreased from 6.5 to 3.3 mgL⁻¹. However, the oil content increased from 68 to 150 mgL⁻¹. As previously stated, the oil was a high-concentration substrate that had minimal impact on the control of sludge SV. The color of the sludge changed from black to yellow and the odor disappeared. The SV decreased by 5–8% every day, and from 92% to 43% during the 8 days.

The results show that high Fe leads to sludge bulking, and reducing Fe concentration to less than 0.6 mgL⁻¹, selectively discharging sludge to maintain concentration at around 1,700 mgL⁻¹, and controlling oil content and DO at about 145.33 mgL⁻¹ and 3.5 mgL⁻¹, respectively, is the most effective and rapid way to prevent sludge from bulking.