Characterizing the free ammonia exposure to the nutrients removal in activated sludge systems †

As a by-product during liquid production, the liquor wastewater exhibits tremendous environmental risks and may cause undesirable e ﬀ ects to the biological systems due to the high concentration of ammonia. This study was carried out to characterize the e ﬀ ects of high concentration free ammonia (FA) exposure in liquor wastewater on the removal of nutrients (NH 4+ – N, NO x (cid:1) – N, PO 43 (cid:1) and organics) in sequencing batch reactors (SBRs). Experimental data demonstrated that the short-term exposure of FA had little e ﬀ ect on the removal of typical pollutants. In the long-term, however, it was found that the removal of NH 4+ – N was determined by the counteraction of negative e ﬀ ects caused by high concentration FA and the contribution brought by initial substrate concentration. Moreover, the high concentration FA had evident inhibitory e ﬀ ects on both the nitri ﬁ cation process and the phosphorus removal. The side-e ﬀ ect of high FA on organics removal was insigni ﬁ cant and the lower removal e ﬃ ciency of apparent COD in high FA reactors was primarily attributed to the accumulation of NO 2 – N, which was commonly neglected in previous studies. Mechanism explorations indicated that the microbes in the SBRs could excrete additional protein-dominated extracellular polymeric substances (EPS) to resist the side-e ﬀ ects caused by high FA. This made the sludge structure more compact, while it blocked the transport of substrates in the process of pollutants removal. Furthermore, FA would also in ﬂ uence the microbial diversity and reduce the abundance of key microorganisms ( i.e. ammonia-oxidizing bacteria, nitrite-oxidizing bacteria and phosphate-accumulating organism) responsible for nutrients removal; thus the corresponding removal e ﬃ ciency subsequently decreased in the presence of high FA.


Introduction
As one of the most featured industry in China, the production of Chinese liquor is extremely high every year. Statistics show that the amount of liquor production was up to 13.6 million tons in 2016, which accounted for about 38% of the world's total production. 1 Consequently, large amounts of liquor wastewater were generated. It has been evaluated that 1 ton liquor production would generate 15 tons of wastewater during the cleaning, fermentation and distillation processes. 2 The wastewater, which contains high concentrations of organics and NH 4 + -N, was difficult to treat, which resulted in great environmental risks. The common disposal of liquor wastewater includes electrodialysis, occulation, ozonation, and biological treatment. 2 In particular, the biological approach was the preferred way considering its environmental friendliness, effectiveness and low costs. 3,4 However, the high concentration of NH 4 + -N, which could also be present in the form of un-ionized ammonia (free ammonia, NH 3 (FA)), would cause a huge shock to the biological systems and inuence their treatment ability and capacity. 5, 6 Chen has demonstrated that the ammonium oxidation ability was greatly inhibited and the removal efficiency of ammonia nitrogen decreased signicantly at FA > 15 mg L À1 . 7 The subsequent nitrite oxidation could also be affected by the high concentration of FA as the process has been reported to be completely inhibited when FA exceeded 16 mg L À1 . 8 In addition, it was found that FA had severe inhibitory effects on the phosphorus uptake during aerobic process and ultimately resulted in a phosphate excess in effluents. 9 Moreover, the removal of organics was also reported to be affected in the presence of FA. The chemical oxygen demand (COD) removal rate and the concentration of dehydrogenase, which were directly related with the removal rate of organics, were both observed to gradually decrease with the increase in FA. 10,11 However, most of the previous studies primarily focused on the effects of FA, on the removal of individual pollutant and its specic removal efficiency. [9][10][11] In general, pollutants co-exist in wastewater and the degradation of one particular pollutant would exhibit competitive or synergistic effects with other pollutants and ultimately inuence their respective removal efficiency. Thus, it was very important and urgent to investigate the inuence of FA exposure to the removal of nutrients and their internal connection in biological systems. Moreover, though previous studies have disclosed that FA had adverse effects on the microbial activity as well as the stability of sludge, which led to the inhibition of pollutants removal, 12 the information on detailed mechanisms for changes in sludge characteristics and shis of microbial community caused by high concentration FA was limited.
In this study, the effects of FA on the simultaneous removal of typical pollutants including NH 4 + -N, NO x À -N, PO 4 3À and organics were rst investigated in the sequencing batch reactors (SBRs). Then, the particular inuenced stages for every specic pollutant by FA and their mutual interactions in the removal process were explored. Furthermore, variations of the structure and characteristics of activated sludge in SBRs with different concentrations of FA and the related mechanisms are discussed in detail. In addition, the shis of microbial communities responsible for the removal of pollutants in SBRs were studied using high-throughput pyrosequencing.

Inoculated sludge and inlet wastewater
The inoculated sludge was taken from the secondary sedimentation tank of wastewater treatment plant in Nanjing, China. Prior to the experiment, the sludge was rinsed with distilled water for 9 h to remove the effects of the original substrate on the test. The initial suspended solids concentration (MLSS) in SBRs was 4000 mg L À1 . According to the survey, the concentration of NH 4 + -N in the liquor wastewater ranges from 80 mg L À1 to 400 mg L À1 in different seasons (Table S1 †); thus the inuent concentrations of NH 4 + -N in four reactors (R1-R4) were set as 50, 100, 200 and 350 mg L À1 , respectively.

Experimental set-up and procedure
The experimental work was performed in four identical labscale SBRs with an effective volume of 2.5 L. The SBRs were operated by alternating anoxic and aerobic reactions with a cycle time of 24 h and consisted of ve phases: 4 h for stirring under anoxic conditions, 16 h for aeration, 1 h for precipitation, 10 min for drainage and the remaining time for idling. The batch tests (on cycle) were initially implemented in SBRs to investigate the short-term effects of FA on the performance of SBRs. Then, the chronic toxic effects of FA were also studied with long-term exposure. The hydraulic retention time and the sludge age were about 48 h and 10 cycles, respectively. To create an anoxic environment, argon was stirred into the inuent for 20 min before entering the water. During aeration stage, the dissolved oxygen concentration was controlled at about 3 mg L À1 by the air blower and the gas owmeter. The room temperature was maintained at 27 AE 1 C.
The nitrite accumulation ratio (NAR) of the reactor was calculated according to the following equation (eqn (2)):

EPS analysis
The method to extract EPS was based on cation exchange using a cation exchange resin (CER) reported by Dubois et al. 13 The anthrone-sulfuric acid method was used to measure the polysaccharides (PS) concentration, 14 and the modied Lowry method was used to measure proteins (PN) concentration. 15 The detailed substances in EPS were explored by 3D uorescence luminescence spectrometer (Hitachi, F7000) at cycle 1 and cycle 27. The excitation wavelength (Ex) and emission wavelength (Em) of EEM spectra were from 200 to 700 nm with the increment of 10 and 2 nm, respectively. The excitation and emission slit were 5 nm and the scanning speed was 12 000 nm min À1 . The data of EEM spectra was analyzed by Origin 8.5.

High-throughput sequencing
For a complete understanding of the microbial community structure of sludge with different FA concentrations, highthroughput sequencing technique was explored when the SBRs were stabilized. The samples were prepared according to the OMEGA Reagent Kit Preparation Guide (E.Z.N.ATM Mag-Bind Soil DNA Kit) and tested by Omega Bio-Tek Company (Agent: Guangzhou ying Biological Engineering Co, Ltd, Nanjing, China). The primers used in high-throughput sequencing analysis were 341F (5 0 -CCTAYGGGRBGCASCAG-3 0 ) and 806R (5 0 -GGACTACNNGGGTATCTAAT-3 0 ). 16

Statistical analysis
All tests in the present study were carried out in triplicate. An analysis of variance (ANOVA, p < 0.05) was used to test the signicance of results.

Effects of FA on the SBR operation
The short-term effects of FA on the performance of SBRs were rst investigated. The obtained results showed that no evident differences on typical pollutants (NH 4 + -N, NO 3 À -N, PO 4 3À and organics) removal were observed among reactors with different FA concentration (data not shown), which excluded the acute effects of FA on the activated sludge system. However, the SBRs are commonly operated continuously in real application and thus the study on the long-term effects of FA was more concerning and attractive. As shown in Fig. 1(a) Unexpectedly, the total reduction of NH 4 + -N was even more in R3 and R4. Previous studies indicated that high concentration of FA would inhibit the activities of ammonia-oxidizing bacteria (AOB) and thus decrease the NH 4 + -N removal. 8 However, it was also reported that the initial high concentration of NH 4 + -N was closely related with its removal rate. 17 In the present study, the contribution of initial high concentration of NH 4 + -N to its removal rate can outcompete the inhibitory effects of FA that resulted in the high removal amounts of NH 4 should be noted that the inhibitory effects of FA could be partly relieved aer cultivation and adaptation as the removal efficiency of NH 4 + -N in the R3 were improved evidently aer 9 cycles, which was consistent with previous literature. 18 As to the nitrication process, high concentrations of NO 2 À -N were accumulated in R3 and R4 reactors. The NAR was about 76% and 83%, for R3 and R4 respectively (as shown in Fig. 1(b)) and the NO 3 À -N concentrations were relatively low. This demonstrated that the nitrite conversion process was severely inhibited in the R3 and R4 reactor. In case of the effects of FA on phosphorus removal, it can be observed from Fig. 1(c) that the phosphorus removal rate was also negatively affected by the FA.
The removal efficiency was about 45% in R1 and 15% in R3 aer operation of 10 cycles. The reason could be attributed to the toxicities of FA to the activities of polyphosphate accumulating organisms (PAO). It has been postulated that FA could inhibit the metabolic process of PAO and thus led to the decrease of phosphorus uptake. 19 The metabolic activity of heterotrophic bacteria has also been reported to be affected by the high FA since the biodegradation rate of organic matters was inhibited. 20 In the present study, the removal of organics in SBRs also appeared to be greatly affected by FA as the COD removal efficiency varied in different reactors and it was relatively low in R3 and R4 (as shown in Fig. 1(d)). However, further exploration has found that the lower COD removal efficiency was primarily attributed to the accumulation of NO 2 À -N. As a reducing substance, 1 mg NO 2 À -N is calculated to be equivalent to 1.14 mg COD. Aer excluding the contribution of NO 2 À -N to COD, the effects of FA on the COD removal in SBRs were actually insignicant ( Fig. 1(d)). The result obtained is quite different from previous studies, which have neglected the contribution of NO 2 À -N to COD and assumed that the degradation of organic matter by heterotrophic bacteria was inhibited by FA. 20 3.2. Mechanisms of the negative effects on pollutants removal by FA It has been discussed above that FA affected the removal of N and P in SBR systems in the long-term run. However, the operation of SBRs included anoxic and aerobic processes. In order to deeply disclose the detailed effects of FA on the longterm performance of SBRs, FA impacts on the particular process in each cycle during SBR operation were further explored.
As shown in Fig. 2, the denitrication rate was very rapid during anoxic period as the concentration of NO 2 À -N and NO 3 À -N in the four reactors all dropped to about 0 within 1 h.
These results indicated that the FA had minor inhibitory effects on the denitrifying bacteria under anoxic condition.
In case of the aerobic stage, the average NH 4 + -N removal rates in R1-R4 were 3.1, 6.2, 8.8 and 8.4 mg (L À1 h À1 ), respectively, indicating that initial high NH 4 + -N concentration offsets the inhibition on AOB by high FA. The generated NO 2 À -N was further converted to NO 3 À -N quickly in R1 and R2 as only minor quantity of NO 2 À -N was observed, while high concentration of NO 3 À -N was accumulated. However, contrary to R1 and R2, the concentration of NO 3 À -N was relatively low in R3, while NO 2 À -N gradually accumulated in the entire stage. Moreover, a signicant lag occurred in R4 as no evident accumulation of NO 2 À -N was observed until 12 h; moreover, the concentration of NO 3 À -N was also very low. The results demonstrated that high FA affected the operation of SBRs by exhibiting negative effects on the oxidation process of NO 2 -N, which resulted in the accumulation of NO 2 À -N in aerobic stage. The reason could be attributed to the increasing biotoxicity of high FA to nitriteoxidizing bacteria (NOB) and thus exhibited inhibitory effects on the nitrication process, which has also been reported by previous studies. 21 In case of the organics removal, it can be observed from Fig. 3(a) that the removal rates of COD during anoxic stage in R3 and R4 were higher than that of R1 and R2, which could be attributed to the utilization of carbon substrates during deni-trication process. In aerobic stage, the concentrations of COD kept decreasing initially because of the consumption of organics by heterotrophic bacteria in all reactors. However, the apparent COD started to increase later in R3 (10 h) and R4 (14 h). This exactly coincided with the time of signicant NO 2 À -N accumulation. Aer subtracting the contribution of NO 2 À -N, The release of phosphate was also rst affected by the FA in the anoxic stage. It can be observed from Fig. 3(b)  -N would inhibit the respiration of PAO under aerobic condition and thus reduced the removal efficiency of phosphate. 9 However, in the present study, the NO 2 À -N accumulation did not appear during this period (Fig. 2). This indicated that the accumulation of NO 2 À -N was not the main reason for the decline of phosphate removal in R3 and R4, while the high concentration of FA would be the key inhibitory factor to the activity of phosphate accumulating microorganisms.

Microstructure properties of sludge characteristics
In order to further explore the mechanisms for the inhibitory effects of FA on typical nutrients removal in SBRs, the changes in sludge morphology and structure were rst observed using a scanning electron microscope aer stable operation. The results showed that the amount of mucus in the sludge ocs increased along with the increase in FA concentration. It can be observed from Fig. 4 that as compared with the loose structure in R1 and R2, the sludge in R3 and R4 was more compact and the oc was thicker. This indicated that large amounts of mucoid substances formed in R3 and R4, where most of the micro-organisms were embedded. The changes in sludge surface structure at high FA concentration resulted in massive losses of the original gully and channel as well as the reduction of mass exchange surface, which would block the exchange and transport of external nutrients and internal metabolites in cells. Therefore, the activities of functional microbes were negatively affected and the removal of typical pollutants was inhibited.

Effects of FA on extracellular polymeric substances (EPS) variations
Furthermore, the mucoid substances were expected to primarily consist of EPS as the microbial cells were reported to be able to produce extra EPS to protect them from adverse conditions, 22,23 such as the toxicity caused by FA in this study. In order to testify this assumption, the overall amounts and components of EPS in SBRs with different FA concentrations were measured. As shown in Fig. 5 and Fig. 6, the main components of EPS in all reactors were PN and PS, while humic acid-like substance was quite low. Moreover, the amount of EPS was clearly more at high FA concentration, particularly in the R3 reactors, in which the EPS increased from 7.2 mg g À1 VSS in cycle 1 to 24.32 mg g À1 VSS in cycle 15.
In addition, it has been found that the increase in EPS was primarily derived from PN (as shown in Fig. 5) as the ratio of PN/ PS in R1-R4 were 1.16, 1.27, 1.98 and 1.76, respectively aer 15 d. It appeared that the microorganisms resisted the adverse environment by excreting more protein, which was consistent with previous studies. 24 More precisely, it also can be observed from Fig. 6 and Table S2 † that aromatic PN-like substances (peak C in Fig. 6) newly emerged aer long-term operation in different reactors. Simultaneously, the high concentrations of FA would inuence microorganisms secreting more tryptophan PN-like (peak A in Fig. 6) and tyrosine PN-like (peak B in Fig. 6) substances, which are consistent with the observations shown in Fig. 5. However, the polysaccharides have been reported to play more important roles in cohesion and adhesion of cells and the granules with lower PN/PS have been considered to have better function. 25 For example, Liu et al. found that the decrease of polysaccharides could negatively inuence the attachment and self-immobilization processes of bacteria in activated sludge. 26 From this perspective, the function of activated sludge systems appeared to be altered due to high FA exposure. Moreover, it is worthy to point out that the largest amount of produced EPS was observed in R3 during stable stage though the concentration of FA was lower than that in R4. This result is consistent with that shown in Fig. 4, in which the sludge was more compact in R3 compared with R4. Thus, it can be concluded that the tolerance of microorganisms under unfavorable conditions by excreting EPS was limited and they would gradually lose such protective mechanism when the concentration of toxic chemicals is beyond the tolerant value. 27 3.5. Shis of microbial community structure in the SBRs with FA Microorganisms play key roles in the removal and biodegradation of typical nutrients in SBRs. Thus, in the current study, high-throughput sequencing was explored to analyze the structure of microbial community and the sample library coverage was more than 90% in all reactors, suggesting that it could well reect the microbial community. The results showed that a total of 39 phyla were detected in SBRs reactors with 16 phyla jointly owned, but the microbial diversity and abundance clearly decreased owing to the existence of FA (as shown in Fig. 7(a) and Table S3 †). For example, the abundances of Bacteroidetes, Verrucomicrobia and Acidobacteria, which played key roles during the nitrication process, [28][29][30] all decreased in the presence of high concentration FA. Their abundance was determined to be 31.5%, 6.8%, and 1.87% in R1, while only 26.7%, 1.83%, and 0.4% was observed in R3, indicating that the high concentration of FA reduced the nitrication process by inhibiting the growth of related bacteria. However, the abundance of Proteobacteria, which has been reported to include large amounts of denitrifying bacteria with the ability of utilizing nitrate, 31,32 increased signicantly from 52.0% in R1 to 63.9% in R3. One of the possible explanations could be that a small amount of nitrate was accumulated in R3 as its deni-trication process is not possibly affected by FA from the perspective of microbial analysis.
The function evolution of microbial community can be demonstrated better based on the genus level. Thus, in the current study, the relative abundance of key functional phylogenetic groups in SBRs at the genus level was also investigated ( Fig. 7(b)). Among these genera, Ferruginibacter, which is known  for its ability of degrading organic substance, 33 showed little difference between R1 and R3 as the abundances were respectively 3.2% and 3.5%. This was consistent with the result of COD removal, which was not affected under different concentrations of FA in SBRs. However, the proportions of Hydrogenophaga, Paracoccus, Nitrosomonas, Comamonas, and Nitrospira, which were all regarded as heterotrophic nitriers responsible for the nitrication, 34-39 decreased in R3 when compared with R1. In particular, the abundance of Nitrospira, as genera of NOB, was reduced by 80% in R3. Thus, the decrease of nitriers in the presence of high concentration FA was the main reason accounting for the reduction of nitrication removal rate and ultimately resulted in the accumulation of NO 2 À -N. Moreover, the abundance of Bacilli, which was proved to be a PAO, 40 shrank by about 83% from R1 to R3. Thus, the phosphorus removal was then negatively affected in the high FA systems.
Overall, the shis of microbial community suggested that the amounts and diversities of microorganisms were obviously affected by different FA concentration, which resulted in the variations of removal efficiency of different pollutants in SBRs.

Conclusions
In conclusion, the inuences of FA on the removal of typical nutrients in the wastewater varied. High concentration of FA would exhibit negative effects on the removal efficiency of NH 4 + -N, but it was also correlated with the initial concentration. The oxidation of NO 2 À -N during nitrication process and the removal of phosphate was clearly inhibited due to the high concentration of FA. However, the degradation of organics was not affected in the presence of high concentration of FA though the apparent concentration of COD varied. This was primarily attributed to the contribution of NO 2 À -N accumulation. Moreover, it was found that the microorganisms resisted the sideeffects of FA by excreting extra protein-dominated EPS, but it can simultaneously inhibit the exchange and transport of substances onto/into the microbes. The diversity and abundance of key microorganisms responsible for the removal of nutrients (e.g. NOB and PAO) were observed to decrease in the high concentration FA reactors and therefore the removal efficiencies of pollutants were negatively affected.

Conflicts of interest
There are no conicts to declare.