Responses of Nitrogen Removal, Extracellular Polymeric Substances (EPSs), and Physicochemical Properties of Activated Sludge to Different Free Ammonia (FA) Concentrations

Abstract

To investigate the effect of free ammonia (FA) on the nitrogen removal performance, extracellular polymeric substances (EPSs), and physicochemical properties of activated sludge, four laboratory-scale sequencing batch reactors (SBRs) were operated at FA concentrations of 0.5, 5, 10, and 15 mg/L (R_0.5, R₅, R₁₀, and R₁₅, respectively). Results showed that nitrogen removal and the production of EPSs and their components (including polysaccharides, proteins, and nucleic acid) significantly increased with the increased FA concentration from 0.5 to 10 mg/L; however, they decreased with a further increase in FA to 15 mg/L. Moreover, the capillary suction time (CST), specific resistance of filtration (SRF), and sludge volume index (SVI) decreased when FA concentration increased, indicating that better settleability and dewaterability of activated sludge was obtained. Additionally, a path diagram showed that Nitrosomonas was positively correlated, while Denitratisoma was negatively correlated with EPSs and their components. Thauera was positively correlated, while Zoogloea was negatively correlated with the settleability and de-waterability of activated sludge.

1. Introduction

Traditionally, activated sludge was regarded as the core functional body in biological wastewater treatment plants (WWTPs) [1,2,3]. Extracellular polymeric substances (EPSs) that accounted for 80% of the activated sludge matrix are types of metabolism or lysis of microorganisms, which play a significant role in the composition of activated sludge [4,5]. In biological wastewater treatment, ammonia nitrogen, as a kind of main pollutant, widely exists in various wastewaters. Therefore, there must be a certain concentration of free ammonia (FA), because that FA mainly depends on the ammonia concentration, temperature, and pH value of the wastewater [6].

In recent decades, increasing efforts have been devoted to investigating the effects and mechanisms of FA on biological nitrogen processes (nitrification, denitrification, and anammox), biological phosphorus removal, and anaerobic digestion processes [7,8,9,10]. However, during aerobic nitrification, literature related to the effect of FA on EPSs and their components, including proteins (PNs), polysaccharides (PSs), and nucleic acid (DNA), is limited. Zhang et al. [11] reported that the FA pretreatment at 176.5 mg/L could facilitate the break-down of EPSs and kill the living cells. However, the effect of relatively lower FA concentration on the production of EPSs and their components of activated sludge remains unknown.

In the treatment of waste sludge, a few former studies have investigated the correlation between sludge properties and EPS contents under different operating conditions, but no solid conclusion was drawn. For example, Basuvaraj et al. [12] demonstrated that tightly bound EPSs (TB-EPSs), especially the PN content, were positively correlated with physicochemical characteristics of activated sludge in a full-scale system or lab-scale SBRs treating slaughter wastewater. However, Li and Yang. [13] found that loosely bound EPSs (LB-EPSs) were negatively correlated with the physicochemical characteristics of activated sludge in a bioreactor treating synthetic wastewater. Furthermore, no correlation between TB-EPSs and physicochemical characteristics was observed in works of Yang and Li [14] and Ye et al. [15]. Moreover, some other researchers found that rich protein in LB-EPSs could enhance the settleability and dewaterability of activated sludge [15].

Although some works were conducted on the inhibitory effect of FA on the nitrogen variation and EPS contents in wastewater biological treatment, there is still a lack of comprehensive insight into the microbial mechanism of FA on the EPS production and activated sludge properties. Furthermore, the overall relationship among EPSs, microbial species, and FA concentration were not described in detail. In order to fill this knowledge gap, further investigation on the biological response to different FA constraints during aerobic nitrification were imperative in this work. Therefore, the objectives of this study were to (1) elucidate the effect of FA on the EPSs (total EPSs, TB-EPSs, and LB-EPSs) and their components from a microbial perspective; (2) investigate the variations of three kinds of EPSs and their components during nitrification and denitrification; and (3) establish the network correlation among the microorganism, EPS, and settleability and dewaterability of activated sludge.

2. Material and Methods

2.1. Experimental Set-Up and Operational Procedure

Four bench-scale sequencing batch reactors (SBRs) made of Plexiglas with a 5 L effective working volume, 15 cm diameter, and 40 cm effective height were used in the present study. SBRs were operated at four different FA concentrations (0.5, 5, 10, and 15 mg/L). The long-term operation (244 days) was carried out in four SBRs fed with synthetic wastewater. The operational cycles of the four various FA concentration SBRs (R_0.5, R₅, R₁₀, and R₁₅). The filling (5 min), aerobic reaction (240–360 min), anoxic reaction (300–420 min), decanting (25–455 min), and idling period (variable time) involved a cycle of each SBR (

Table 1. Overall operational parameters of the SBRs at four FA concentrations during the entire experimental period.

SBRs	Influent Concentration (mg/L)		Phase Time of the SBR (min)					Operational Parameters
	COD	${\mathbf{NH}}_4^{+}$-N	One Cycle	Filling	Aeration	Anoxic	Decantation	FA (mg/L)	MLSS (mg/L)	Temperature (℃)	pH	DO (mg/L)
R0.5	80	40	620	5	270	300	45	0.5 ± 0.15	3900	20 ± 2.0	7.5 ± 0.2	1.0~2.5
R5	80	90	710	5	300	360	45	5 ± 0.55	4400	25 ± 2.0	8.0 ± 0.2	1.0~2.5
R10	80	130	810	5	360	420	25	10 ± 2.1	4500	30 ± 2.0	8.0 ± 0.2	1.0~2.5
R15	80	55	570	5	240	300	25	15 ± 2.5	4400	35 ± 2.0	8.5 ± 0.2	1.0~2.5

). The aerobic and anoxic reaction durations varied with initial substrate concentrations with the purpose of achieving complete nitrification and denitrification. The different setting time was caused due to various settleability of AS in each SBR. Idle periods of the SBR at four FA levels were also different for maintaining the same total at each cycle time (1 day) for the four SBRs in this present work. In order to reflect the accurate biochemical reaction time, the one cycle time in Table 1 was the sum of filling, aerobic reaction, anoxic reaction, and decanting period, excluded the idle period. It is noted that ethanol, as external carbon source, was added into the reactors at the beginning of denitrification. Table 1 lists the detailed operational conditions of the four SBRs during the entire experimental cycle.

2.2. Inoculated Sludge and Influent Contents

Inoculated activated sludge was collected from a plant mainly treating domestic and brewery wastewater (accounting for approximately 60~70% and 30~40%, respectively). The plant was located at Lanzhou, Gansu Province, China, where the anaerobic-anoxic-oxic process was employed. The initial mixed liquor suspended solids (MLSS) concentration was 3000 mg/L. The inoculated sludge was domesticated for 20 days and fed with synthetic wastewater with the following composition per liter: 115 mg of NH₄Cl, 385 mg of CH₃COONa, 26 mg of KH₂PO₄, and trace element solution. The trace element solution consisted of MgSO₄·7H₂O 5.07 mg/L, MnSO₄·4H₂O 0.31 mg/L, FeSO₄·7H₂O 2.49 mg/L, CuSO₄ 0.25 mg/L, Na₂MoO₄·2H₂O 1.26 mg/L, ZnSO₄·7H₂O 0.44 mg/L, NaCl 0.25 mg/L, CaSO₄·2H₂O 0.43 mg/L, CoCl₂·6H₂O 0.41 mg/L, and EDTA 1.88 mg/L.

2.3. EPS Extraction

The EPS extraction mainly included LB-EPS and TB-EPS extraction. The modified two-step thermal extraction method was adopted to extract LB-EPSs and TB-EPSs in the present study [4,16]

2.3.1. LB-EPS Fraction Extraction

Well-mixed sludge water was centrifuged at 2100× g at 4 °C for 10 min to separate the supernatant from the solids. The supernatant was filtered by a 0.45 μm microporous membrane for analysis, with the collected supernatant regarded as the LB-EPS fraction.

2.3.2. TB-EPS Fraction Extraction

Ringer solution was then added to the residual activated sludge, then the mixture was heated at 80 °C for 60 min in a constant temperature water bath, and subsequently centrifuged again at 12,000× g at 4 °C for 10 min. Finally, the obtained supernatant was filtered again by a 0.45 μm microporous membrane for analysis, with the collected supernatant regarded as the TB-EPS fraction.

2.4. Analytical Methods

For long-term performance detection, the sample was taken from each SBR at the begining and end of nitrification and denitrification, respectively, for every operational cycle. The sample was taken from each SBR every hour to analyze the typical variation of nitrogen, COD, and EPS in a SBR cycle. Each sample was measured parallelly three times and the average was applied.

2.4.1. EPS Quantification

The PN and PS contents of LB-EPSs and TB-EPSs were measured using the Lowry and phenol–sulphuric acid methods, respectively, with bovine serum albumin and glucose used as standards, respectively [17,18]. DNA was determined by using the method of ultraviolet absorption [19]. The sum of PNs, PSs, and DNA fractions in LB-EPSs and TB-EPSs was regarded as the LB-EPS and TB-EPS content, and the sum of LB-EPS and TB-EPS fractions was regarded as the total EPS content.

2.4.2. Conventional Analytical Methods of Wastewater Parameters

COD, NH₄⁺-N, NO₂⁻-N, NO₃⁻-N, and MLSS were analyzed in accordance with the standard methods (APHA, 1998) [20]. Temperature, pH, and dissolved DO values were monitored by using WTW Multi 3420 (WTW Company, Munich, Germany).

The nitrite accumulation ratio (NiAR), and nitrate accumulation ratio (NaAR) were calculated according to the following equation (Equations (1) and (2)) [21]:

$$\mathrm{NiAR}(\%)=\frac{{\mathrm{NO}}_2^{-}-\mathrm{N}}{{\mathrm{NO}}_2^{-}{-\mathrm{N}+\mathrm{NO}}_3^{-}-\mathrm{N}}\times 100\%$$

(1)

$$\mathrm{NaAR}(\%)=\frac{{\mathrm{NO}}_3^{-}-\mathrm{N}}{{\mathrm{NO}}_2^{-}{-\mathrm{N}+\mathrm{NO}}_3^{-}-\mathrm{N}}\times 100\%$$

(2)

The FA concentration was calculated according to the following equation (Equation (4)) which was reported by Anthonisen et al. [6]:

$$\mathrm{FA}\left(\mathrm{mg}/\mathrm{L}\right)=\frac{17}{14}\times \frac{{\mathrm{NH}}_4^{+}{-\mathrm{N}\times 10}^{\mathrm{pH}}}{\exp (\frac{6334}{273+\mathrm{T}}{)+10}^{\mathrm{pH}}}$$

(3)

where NH₄⁺-N is the ammonia concentration (mg/L); NH₄⁺-N_influent and NH₄⁺-N_effluent are the ammonia concentrations (mg/L) in influent and effluent, respectively; NO₂⁻-N is the nitrite concentration (mg/L) at the end of nitrification; NO₃⁻-N is the nitrate concentration (mg/L) at the end of nitrification; T is the temperature (°C); and pH is the pH level.

Based on this equation, FA depends on the ammonia nitrogen concentration, pH value, and temperature. Different FA concentrations could be achieved by adjusting the initial NH₄⁺-N concentration, pH value, and temperature. Various NH₄⁺-N concentrations were achieved by adding a different amount of ammonium chloride (NH₄Cl) (1 mol/L) as the energy and nitrogen source. COD concentration was obtained by adding ethanol (CH₃CH₂OH) as the carbon source. The temperature control system and heating water jacket were used to control the experimental temperature in the range of 20–35 °C. The pH value was controlled at 7.5–8.0 by adding 0.1 M HCl and 0.1 M NaOH. The DO level was kept within the range of 1.0–2.5 mg/L using an air compressor during the aeration period.

2.4.3. Settleability and Dewaterability of Activated Sludge

The settleability of activated sludge was assessed by measuring the sludge volume index (SVI). The dewaterability of activated sludge was assessed by measuring the capillary suction time (CST) and specific resistance of filtration (SRF). SVI was determined by standard methods [20]. CST was detected by a 304B analyzer (Triton, UK) [15]. SRF was calculated according to the methods reported by Li and Yang [13] using the following equation (Equation (4)):

$$\mathrm{SRF}(\mathrm{m}/\mathrm{kg})=\frac{{2\mathrm{PA}}^2\mathrm{b}}{\mathsf{\mu } \mathrm{C}}$$

(4)

where P is the pump pressure (Pa), A is the filter area (m²), μ is the dynamic viscosity of the filtrate (N·s/m²), C is the resistance of the filtrate per unit volume to the medium (kg/m³), b is the slope of the curve between t/V and V, t is filtrate time (s), and V is the filtrate volume (m³).

2.4.4. DNA Extraction, PCR Amplification, Illumina MiSeq Sequencing and Microbial Diversity Analysis

According to the manufacturer’s instructions, DNA was extracted from activated sludge sample (2.5 mL) using the FastDNA^® SPIN extraction kits (MP, Biomedicals, CA, USA), and stored at −20 °C prior to the further analysis. The extracted DNA was quantified using a NanoDrop^® ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and agarose gel electrophoresis.

PCR amplification and Illumina MiSeq sequencing were adopted to analyze the microorganism’s diversity. For 16S rRNA PCR amplification, the V3-V4 region of 16S rRNA was amplified using the primer set 338F (5′-ACTCCTACGGGAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′). The PCR reaction conditions were as follows: preheated at 95 °C for 2 min; 25 cycles of amplification (denatured at 95 °C for 30 s; annealed at 55 °C for 30 s; elongated at 72 °C for 40 s); and, finally, extended at 72 °C for 10 min. Illumina Miseq high-throughput sequencing was conducted at Shanghai Personalbio Technology Co. Ltd. (Shanghai, China).

The MOTHUR software was used for quality control and filtering of the raw sequence. The main steps discarded short sequences less than 50 bp, removed repetitive sequences, and discarded chimeric sequences. The effective sequences were quantified and classified using QIMME software and a 97% sequence similarity threshold was used to classify reads into operational taxonomic units (OTUs).

2.5. Statistical Analysis

The rates of ammonia oxidation, nitrite accumulation, and nitrate production were calculated by using Microsoft Excel 2016. Moreover, one-way analysis of variance (ANOVA) was applied to evaluate the relationship and significant difference among microorganisms, EPSs, and sludge properties. Results were considered statistically significant when p < 0.05 and highly significant when p < 0.01. One-way ANOVA and Pearson’s correlation analysis were conducted by using SPSS software (version 20, SPSS Inc., Chicago, IL, USA). Furthermore, the network analysis was performed by using Cytoscape3.3.0 to construct the network visualization to identify the correlation among FA, EPS, sludge property, and microbial community. Variation partitioning analysis (VPA) was clarified to quantify the contributions of FA, EPS, and sludge property to microbial community using R software 2.15.3.

3. Results and Discussion

3.1. SBR Performance under Different FA Stress

Figure 1a shows the NH₄⁺-N concentration in the influent and effluent under different FA stress during the whole experiment. Influent NH₄⁺-N concentration was in the range of 40–130 mg/L, and the effluent NH₄⁺-N concentration decreased to 0.04–2.90 mg/L. Therefore, the NH₄⁺-N removal efficiency was about 97.6–99.4%, with the average value of 98.7%. Extremely high removal efficiency of NH₄⁺-N (above 99%) clearly indicated that the FA concentration range used in this work did not exert negative effects on the NH₄⁺-N oxidation ability of the SBR. That is to say, FA concentration at 0.5–15 mg/L did not exercise inhibitory effects on the AOB activity in this present work. Xu et al. [21] reported that FA pretreatment (<44.5 mg/L) could improve NH₄⁺-N release, thereby enhancing NH₄⁺-N removal. Moreover, Sun et al. [22] found that FA has an obvious effect on the NH₄⁺-N oxidation rate during nitrification. The NH₄⁺-N oxidation rate was increased from 0.0544 gN/(gVSS⋅d) at 0.5 mg/L to 0.134 gN/(gVSS⋅d) at 10 mg/L. This clearly indicated that FA was favorable to NH₄⁺-N oxidation within suitable FA concentration range.

Figure 1b,c shows the concentration and accumulation rate of NO_x⁻-N during the whole experiment period. The conversion of NH₄⁺-N to NO₃⁻-N was achieved in R_0.5 and R₅, with an average nitrate accumulation rate (NaAR) of 98.5% and 99.2%, respectively, suggesting that ammonia was removed via the nitrate pathway in both R_0.5 and R₅. However, the oxidization of NH₄⁺-N to NO₂⁻-N was obtained in R₁₀ and R₁₅, with a stable average nitrite accumulation rate (NiAR) in 98.4% (from 77 cycles) and 96.7% (from 133 cycles), respectively, demonstrating that ammonia was eliminated via nitrite pathway in these two reactors. Sun et al. [23] reported that the FA concentration at 3.0–33.5 mg/L initiated and maintained a stable nitrite pathway with NiAR of 91.3–95.5% in a SBR treating real municipal landfill leachate. Corresponding, NaAR gradually decreased from 98.2% to 4.7%. Comparing this result with our work, NiAR in Sun et al. [23] was obviously higher than that (0.8% at 0.5 mg/L and 1.5% at 5 mg/L); however, it was slightly slower than that (98.4% at 10 mg/L and 96.7% at 15 mg/L) in our work. For this biological mechanism, a high concentration of FA can promote nitrite accumulation by selectively inhibiting the activity of nitrite oxidation bacteria (NOB) but not ammonia oxidation bacteria (AOB) [6,21]. These results clearly showed that a higher FA concentration of 10 mg/L and 15 mg/L was conducive to establishment of the partial nitrification due to the stronger inhibition of FA on NOB than AOB [3].

Moreover, total nitrogen (TN) removal via simultaneous nitrification and denitrification firstly increased from 7.0 mg/L at 0.5 mg/L, to 31.5 mg/L at 10 mg/L FA, and then decreased to 14.2 mg/L at 15 mg/L, which indicates that appropriate FA concentration (0.5–10 mg/L) made a significant positive contribution to TN removal. It was noted that the NH₄⁺-N concentration in R₁₅ (55 mg/L) was obviously lower than R₅ (90 mg/L) and R₁₀ (130 mg/L). This result is consistent with Wang et al. [24], who found that the TN removal performance significantly improved after FA pretreatment in SBRs which treated synthetic wastewater. They also found that the nitrite pathway was established quickly and maintained stably in the SBR (nitrite accumulation ratio was above 90%) when the FA-treated sludge (at 210 mg/L for 1 day) was returned to the SBR. These results demonstrated that it was feasible to eliminate NOB and enhance nitrogen removal through nitrite pathway via sludge treatment using FA.

3.2. The Effect of FA Concentration on the Production of EPSs and Their Components

To assess the effects of FA on microbial activity, three EPS fractions during the end of nitrification were compared under different FA concentrations (Figure 2a–c). The average content of LB, TB, and EPS increased from 4.68, 36.86, and 41.53 mg/g MLSS to 7.93, 44.09, and 52.02 mg/g MLSS, respectively, with increasing FA concentration from 0.5 to 10 mg/L. The contents of LB and TB increased by 69.44% and 19.61% at 10 mg/L, respectively, compared with 0.5 mg/L, illustrating that FA increased inhibition of outer microorganism activities compared to inner activities. The mechanism for this activity is that microorganisms will produce EPSs which ensure survival in toxic environmental conditions. However, the contents of LB and TB decreased by 31.40% and 22.43%, respectively, at 15 mg/L compared with 10 mg/L, demonstrating that the auto-protection ability of microorganisms is limited. This phenomenon is well explained by Sheng et al. [4], who reported that the enzymes of some bacteria are cell-membrane-bound enzymes which are protected by EPSs. However, microbial cells, especially EPSs, can be disintegrated due to the biocidal impact of FA. A higher FA concentration can cause cell inactivation due to the biocidal impact of FA, which triggers a reduction in the metabolite, thereby reducing EPSs. The result of high-throughput sequencing also supports this phenomenon with OTUs, which reflects microbial communities, decreasing from 2766 at 0.5 mg/L to 1438 at 15 mg/L.

The contents of PN and PS at the end of nitrification also showed an increased tendency when FA increased from 0.5 mg/L to 10 mg/L, to protect the microorganisms from FA bio-toxicity (Figure 2d–f). Specifically, compared to 0.5 mg/L, the PN content showed an increase of 17.33% and 22.82% for LB-EPSs and TB-EPSs, respectively, at 10 mg/L. The PS increased by 176.47% and 15.57% for LB and TB, respectively, compared with 0.5 mg/L. Furthermore, compared to 10 mg/L, the PNs showed a reduction of 7.67% and 16.47% for LB-EPSs and TB-EPSs, respectively, at 15 mg/L. The PS decreased by 54.14% and 29.64% for LB-EPSs and TB-EPSs, respectively, compared with 10 mg/L. The rate of decrease in the PS content in LB-EPSs and TB-EPSs was higher than the PN content, further suggesting that PSs were more sensitive to FA. These results indicate that more EPSs and their components will be secreted by microorganisms under unfavorable environmental conditions. However, this self-defensive ability is limited when the FA concentration is over 10 mg/L.

In general, the content of EPSs and their components increased with increasing FA concentrations, triggered by the denser layers of LB-EPSs and TB-EPSs, which can stop FA to diffuse into the interior of activated sludge and reduce FA toxicity to bacteria [25]. However, at relatively high FA concentrations, microorganisms in the sludge flocs utilize the excess nitrogen to synthesize PN and DNA, thereby accelerating EPS production [15].

3.3. The Typical Profiles of EPS and Their Components under Four FA Treatments

To further assess the variations of EPS and their components during nitrification and denitrification, typical cycles of four SBRs were monitored by measuring the concentrations of these corresponding parameters every 1 h (Figure 3). Under the four FA conditions, EPSs and their components increased during the nitrification process, while they decreased during the denitrification process. Furthermore, during the nitrification process, the variation of EPSs and their components were significantly negatively correlated with the changes of NH₄⁺-N concentration (R = −0.787 to −0.856, p < 0.05); however, they were positively correlated with the changes of NO_x⁻-N concentration, (R = 0.645 to 0.699, p < 0.05). Meanwhile, a significant correlation was positive (R = 0.654 to 0.701, p < 0.05) among EPSs and their components and NO_x⁻-N during the denitrification process. These results clearly suggest that certain correlations were stored intracellularly. This is mainly due to the fact that EPSs were negatively charged and could bind with the positively charged NH₄⁺-N through electrostatic interaction [4]. Hence, more EPSs were produced and used to provide sites for the adsorption of NH₄⁺-N, and then NH₄⁺-N was further biodegraded by nitrifying bacteria, resulting in an increase in NO_x⁻-N and EPSs and their components during nitrification. EPSs and their components subsequently reduced during the denitrification process, because they serve as an energy source and are utilized by denitrifying bacteria for denitrification, leading to a reduction in NO_x⁻-N and EPSs and their components [26,27].

3.4. The Effect of FA on the Dewaterability and Settleability of Activated Sludge

The effect of FA concentration on the dewaterability and settleability separation properties of activated sludge was shown in Figure 4. When the FA concentration gradually increased from 0.5 mg/L to 5 mg/L, the CST and SRF increased from 27 s and 19 × 10⁸ m·kg⁻¹ to 29 s and 22 × 10⁸ m·kg⁻¹, with an increment of 7.4% and 15.8%, respectively. However, at 15 mg/L, the CST and SRF decreased to 20 s and 13 × 10⁸ m·kg⁻¹, respectively, a reduction of 31.0% and 40.9% in comparison with 5 mg/L. Low CST and SRF values are associated with good dewaterability of sludge. Moreover, a better negative linear relationship between SVI and FA concentration was found (R² = 0.912, p < 0.01), although the SVI almost unchanged when the FA concentration increased from 0.5 to 5 mg/L, and then significantly decreased from 97 to 70 mL/g when the FA concentration further increased to 15 mg/L.

On this basis, an appraisal of significant difference for CST, SVI, and SRF under four different FA concentrations was carried out. A significant difference in the CST was observed at 0.5 mg/L compared with other three FA concentrations (ANOVA, p > 0.05, Figure 4a). A significant difference in the CST was found between 0.5~5 mg/L and 10~15 mg/L. There were significant differences in SVI between the two FA concentrations, except between 0.5 and 5 mg/L (ANOVA, p < 0.05, Figure 4b). There were significant differences in SRF between the two FA concentrations, except between 10 and 15 mg/L (ANOVA, p < 0.05, Figure 4c). These results clearly indicate that changes in influent FA concentration have an impact on the performance of sludge–water separation. Liu et al. [28] also reported that FA pretreatment improved the dewaterability of digested sludge by 9.2%. There are two reasons for this phenomenon. The first is that the destruction of EPS by FA can weaken the hydrophilicity of the sludge and reduce the bound water content, thereby improving the dewaterability of the sludge. The second is that FA can reduce the number of negative charges, indicating that the electrostatic repulsion interaction is reduced, and the flocculation of sludge particles is promoted, resulting in better dewatering performance of the sludge.

3.5. Co-Occurrence Network among FA, EPS, Sludge Properties and Microbial Communities

Network analysis structurally and statistically illustrates better patterns of co-occurrence and interrelation with 31 nodes (1 FA parameters, 22 genera, 5 EPS parameters, and 3 activated sludge indexes) and 267 edges (Spearman’s |ρ| ≥ 0.75, p < 0.05) in all the activated samples (Figure 5).

Among these correlations (Figure 5a), for activated sludge properties, it was found that the CST, SVI, and SRF were positively correlated with Zoogloea, Acidovorax and Variovorax (R² varied from 0.592 to 0.920, p < 0.01), while were negatively correlated with Thauera, Pseudomonas, Lewinella, Phaeodactylibacter, and Halomonasana (R² varied from −0.527 to −0.927, p < 0.01) (Table S1). These results were consistent with the former researchers [29,30], who reported that higher distribution of Thauera and lower distribution of Zoogloea were associated with poor dewaterability and settleability of activated sludge properties. However, this result opposed Yang et al. [8] who found that sludge properties were not affected at 16 mg/L of FA in a SBR treating synthetic domestic wastewater. This could be possibly attributed to the fact that the small relative abundance of Zoogloea does not relate to the weak microbial activity of Zoogloea.

For the EPSs and their components, there was only a significant positive correlation between FA concentration and with PNs (R² = 0.643 to 0.715, p < 0.01), suggesting that FA was beneficial for the formation of PN in EPSs. Moreover, EPS, TB-EPSs, LB-EPSs, PN, and PS were significantly positively correlated with genera of Nitrosomonas (R² = 0.839 to 0.981, p < 0.01); however, they were significantly negatively correlated with genera of Denitratisoma (R² = 0.558 to 0.934, p < 0.05). This suggests that Nitrosomonas and Denitratisoma can be used as an indicator microorganisms to better reflect the variation trend of EPSs and their components. Meanwhile, there were significant correlations between PN and various abundant genera in the community. Specially, PN was significantly positively correlated with three genera, including Thauera (R² = 0.752, p < 0.01), Lewinella (R² = 0.814, p < 0.01), and Nitrosomonas (R² = 0.839, p < 0.01), respectively. However, PN was significantly positively correlated with seven genera, including Zoogloea (R² = −0.962, p < 0.01), Nitrospira (R² = −0.873, p < 0.01), Azoarcus (R² = −0.894, p < 0.01), Denitratisoma (R² = −0.588, p < 0.05), Sulfuritalea (R² = −0.930, p < 0.01), Dechloromonas (R² = −0.876, p < 0.01), and Planctomyces (R² = −0.870 p < 0.01). It is well known that PS/PN ratio strongly affects the surface property of activated sludge [31,32]. In this work, the ratio of PS/PN in LB-EPSs firstly increased from 0.51 at 0.5 mg/L to 0.64 at 5 mg/L, and then decreased to 0.60 at 15 mg FA/L. This trend is also consistent with the CST, SVI, and SRF of activated sludge. Nonetheless, the ratio of PS/PN in TB-EPSs gradually decreased from 1.40 at 0.5 mg FA/L to 1.1 at 15 mg FA/L. EPSs are amphipathic molecules and their hydrophobicity and hydrophily are mostly depended upon PN and PS, respectively. In this study, the variation of PN presents significant positive correlation with the CST, SVI, and SRF (R² = 0.577 to 0.709, p < 0.01). Meanwhile, there is no correlation between PS and these parameters. This result suggests that PNs exert significant influence on microbial aggregation related with the performance of sludge flocculation and sedimentation of activated sludge [31,32]. Based on the above analysis, it is inferred that the PS/PN ratio in LB-EPSs (<1) could better reflect the sludge characteristics than TB-EPSs (>1), which was supported by previous works [13,33], in which they reported that TB-EPSs were considered more crucial for sludge dewaterability and settleability compared with LB-EPSs. Furthermore, compared with PS in LB-EPSs, the PN in LB-EPSs played a more important role in affecting the dewaterability and settleability of activated sludge.

Variance partitioning analysis (VPA) was applied to further evaluate the contributions of FA, sludge property (CST, SVI, and SRF), and EPSs (total EPSs, TB-EPSs, and LB-EPSs) to microbial community structure (Figure 5b). This result clearly demonstrated that 97.34% of the variations could be deciphered by these three portions in microbial community. More specifically, sludge property had the greatest contribution of 3.76% to the microbial structure, followed by EPSs (2.48%) and FA (1.44%). The reciprocal action (24.83%) among the sludge property, as well as the EPSs and FA, was more pivotal than three individual components, Nevertheless, there was less effect on the interactions of EPS+FA (38.6%) and EPSs and sludge (25.67%), suggesting that EPSs helped to best explain the discrepancy of the microbial community.

4. Conclusions

This paper presents a comprehensive insight into the response of EPSs and their components, as well as sludge characteristics to different FA concentrations. The following conclusions can be drawn from this work.

(1)

The ammonia oxidation of the Nitrosomonas was affected by FA at a concentration of up to 15 mg/L; however, nitrite oxidation of the Nitrospira was strongly inhibited at an FA concentration of 10–15 mg/L.

(2)

FA at a concentration of lower than 10 mg/L can effectively promote the production of total EPSs, TB-EPSs, LB-EPSs, PNs, and PN. Above this level (<15 mg/L), the production of EPSs and their components are obviously inhibited. These parameters are significantly positively and negatively correlated with Nitrosomonas and Denitratisoma, respectively.

(3)

The settleability and dewaterability of the activated sludge were improved by FA. This characteristics of activated sludge show a significant positive correlation with Thauera and Nitrosomonas, and a significant negative correlation with Zoogloea and Denitratisoma. Furthermore, the PN in LB-EPSs plays an important role in affecting the dewaterability and settleability of activated sludge.