Treatment of Polluted Urban Surface Waters by Sponge Based Aerobic Biolm Reactor: Purication Performances and Resilience

Surface waters are suffering continuous discharging of pollutions, and low DO and black-odorous were easily formed, especially in those slow-owing urban lakes and enclosed small ponds. In-situ treatment by articial aeration or water cycling with a combination of polyurethane sponge as biolm carriers can disentangle this situation without any land occupation. Long-term continuous experiments (187 days) showed that indigenous microorganisms in urban surface waters could form biolms in the novel style of sponge-based aerobic biolm reactors (SABRs). In urban lake waters treatment, the purication performances of SABRs were stable and resilient as the NH 4+ -N and NO 2− -N removals were steady, even facing the abrupt increase of NH 4+ -N and NO 2− -N concentrations in inuent. Enhancing the polyurethane sponge lling ratios in SABRs can reduce DO but did not affect NH 4+ -N removal. SABRs were also able to treat enclosed surface waters or black-odorous waterbodies. Combined SABRs with water cycling, NH 4+ -N removal time was shorter than the time needed by water cycling when biodegradable organic matter was not present. The massive biodegradable organic matter could inhibit nitrication and prolong the purication time. Further results showed that organic matter could be used as carbon sources to eliminate the produced NO 3− -N in SABRs. Therefore, the developed new bioreactor could act as one effective way for treating N-polluted urban surface waters.

Polyurethane sponge is a synthetic polymer hydrophilic biological carrier and an ideal growth medium for bio lm However, there were seldom reports focusing on the resilience of a reactor in treating urban surface waters.
In this study, a novel style of bio lm reactor was developed to purify urban surface waters under various conditions. Polyurethane sponges were used as carriers to enrich the indigenous microorganisms to form bio lms, and the micropolluted urban lakes and enclosed surface waters such as those small ponds in sub-districts were taken as study objects. We prospected that cultured bio lms can bene t contaminant removal in urban surface waters. The objectives of this

Materials And Methods
Aerobic bio lm reactor set-up Cubic-shaped polyurethane sponges (10×10×10 mm) served as bio lm carriersin this study, purchased from AiQin Environmental Technology Co., Ltd., Jiangsu,China. The average speci c surface area of a sponge cube was 91,000m 3 m -2 , and the porosity was 95%.
Sponge-based aerobic bio lm reactors (SABRs) were set in 100L plastic buckets and simulated as treat urban surface waters (Fig.1a).The effective volume of the plastic bucket was 80L.Each SABR consisted of anup-ow zoneand threebio lm zones, all made by acrylics cylinder with 100 mm inner diameter. The up-ow zone was 600 mm in height, and bio lm zones were 280mm in length (consisting of a 100mm cylinder and a 180mm cone) (Fig.1b). Three bio lm zones were connected with the up-ow zoneand arranged 120°between themselves (Fig.1c).The top of the up-ow zone was sealed with a conical acrylic plate to prevent water from owing out of the up-ow zone and ow the surface wateraveragely into the three bio lm zones.Fine iron wires were arranged in the connection area to prevent the carriers from dropping off when introduced into bio lm zones.Small holes were made at the top of the bio lm area to circulate the surface water and prevent the internal carriers from being washed away.This special design considered that in practical application, the reactorswill be placed in the middle of the surface water and reduce the interference of the surrounding people.
Polyurethane spongeswere lledonly in the bio lm zones, and the lling ratiosin bio lm zones were the same in one SABR. Filling ratiosof 20%, 40%, 60% and 80% were adopted andreferred to as FR20, FR40, FR60, and FR80, respectively.The lling ratios in this study werebased on the volume of the bio lm zone, which is different from previous studies as those lling rates were based on the whole reactor (Zhang et al. 2016;Zhao et al. 2019).

Reactor operating conditions
Puri cation performances and resilienceintreatment of urban lake waters The urban lakes are the receiving watersfor urban pollution, such as the e uents from the wastewater treatment plants.
Thus the experiment was fed with continuous ow.Surface waters were collected to laboratory every week from Lake Xuanwu, located in Nanjing City (China).The operation was dividedinto three stages (Table1).In uents were continuously fed to reactors by peristaltic pumps(6.95 mlmin -1 ),and hydraulic retention time (HRT) was 8 daysasthe HRT of Lake Xuanwu was about 7 days.Ammonium chloride (NH 4 Cl) was added to the lake waters(collected from Xuanwu Lake) to prepare the in uent in stage and stage , and simulate the sudden increase ofNH 4 + -Nconcentration in in uent.No sediments were added into the buckets as dredgingis often adopted in urban lakes, such as Lake Xuanwu.After dredging,the organic carbon contents in the surface layer of sediment were low and denitri cation was inhibited ( Puri cation performancesfor treatment of enclosed surface waters Synthetic waters were usedto simulate the enclosed surface waters,which often contain high concentrations of organic matters and NH 4 + -N with black-odorous appearances (such as those small ponds in subdistricts and parks).Duo to no continuous discharge into those enclosed surface waters, batch experiments wereadopted and to examine whether the bio lms cultured byurban lake waters can directlybe used in those stagnant waters. Control and only hydraulic circulation were also set up(same plastic buckets with 80L effective volume).As organic matter may not be present in some enclosed surface waters, two operating conditions were applied in the experiments.
Organic matters present:Acetic acid and glucose (at a ratio of 1 1by chemical oxygen demand) were added into water to simulatethose black-odorous waterbodies (TOC 59.56±3.09mgL -1 )before experiments. Other characters of surface waters wereslightly higher than no organic matters present (Table 1).
Due to low DO and high organic matters in surrounding untreated surface waters, it is bene cial for denitri cationandremoving the produced NO 3 − -N by SABRs. Thus batch experiments were used to removeNO 3 − -N by mixing the e uent from SABR with the untreated surface waters. The mix ratios were 10%,20%, and 50%.Twocontrols were also adopted (e uent of SABR and untreated surface waters).The mixed waters were not purged with N 2 and were open to air during the experiments. Each treatment was repeatedtwo times.

Analytical methods
Samples were taken from the out owand immediately ltered through 0.45μmpore size ltersthen stored in -20℃ until analysis.The concentrations of NH 4 + -N, NO 2 − -N, NO 3 − -N,PO 4 3--P, TN, and TP were determined in accordance with standard methods(SEPA 2002).The DO, water temperature, and pH were in situ measured by a dissolved oxygen analyzer(YSI 5000,USA) and a digital pH meter (PB-10, Sartorius,Germany), respectively.Dissolved organic carbon concentrations(DOC) were measuredbya TOC analyzer (Torch-Teledyne Tekmar).

Statistical analyses
To compare the removal e cienciesofdifferenttreatments, one way-ANOVA followed by Tukey HSD post-hoc test was conducted by using SPSS software with signi cance considered atP< 0.05 (SPSS Statistics version 20.0).

Results And Discussion
Performance of SABRsintreating urban lake waters During stage , the surface of polyurethane sponges in four SABRs changed from white to yellow and then to brown (Fig.  S1), indicating thatbio lmswereformed by the indigenous microorganismsin urban surface waters.Arti cial aeration and water cycling were often used in surface water puri cation, as these methods can improve the DO and change the hydraulic condition (Cao et al. 2020). DO values in four buckets were improved to over 5.00mgL -1 (Fig. 2a), whichbene tted NH 4 + -N removal. With arti cial aeration,DO concentrations in reactors did not show signi cant difference among the different lling ratios. Wherever, DO concentrationsdeclined with the lling ratios when water cycling was applied to change hydraulic conditions (Fig.2a).
Resilience ofSABRswith suddenincrease of NH 4 + -Nconcentrations in in uent of urban lake waters With a slight increase of NH 4 + -N concentrations in in uent in stage , DO in the four SABRs was lower than that in stage and declined with the lling ratios (Fig. 2a).  (Table 2). One exceptionwas that it took 8 days to reduceNO 2 − -Nconcentrationto below 0.01mgL -1 in FR20 becauseNO 2 − -Nconcentrations in the 4 th shock loading weremuch higher than previousshocking loadings( Fig. 3b; Table 2). These resultswere similar tothe nitri cation reactor treating steel wastewater.After the rst shock loading, it took 14 days to recover nitri cation,but only needed1 day after the second and third NH 4 + -N shock loading (Cho et al. 2016).When NH 4 + -N concentrationsin in uent suddenly increased to 15.00mgL -1 in stage , it took 7-8 days to recover thefunction of nitri cation (Table 2). This result further supported that the SABRsfunctionwas robust tovarious disturbances,which could be related to adaptation to uctuating environmental conditions (Berga et al. 2017).In addition, the in uent nitri ermay be related to the resilience of SABRs, as the response of bacterial communities to disturbances can be affected by the dispersal (Yu et al. 2018), and the long HRT may also play an important roleas it can dilute the NH 4 + -N concentration in surface water.

Performances ofSABRsin treatingenclosed surface waters
Condition withoutbiodegradable organic matter Water cycling within SABRs can improve the DO in enclosed surface waters (Fig. 4a). DO in control declined slightlyfrom 6.00mgL -1 to 3.50mgL -1 duringthe initial10days, then decreased to 0.50mgL -1 within 3 days.However,DO values by water cycling were above 9.00mgL -1 during the whole stage, and DO values in four SABRs all declining from 7.90-8.90mgL -1 to 3.50-5.50mgL -1 within 8days, then slowly increased to 10.00mgL -1 .
Although the cultured bio lms had nitri cation ability, it still needed2 days to activate nitri cation to treat surface water containinga high concentration of NH 4 + -N (Fig.4c, d, e).This result consistsof that nitri cationstarted after 46 hours when the initial NH 4 + -N concentrationincreased to 8.10 mg L -1 in TayNinh River water (Le et al. 2019). In other words, the bio lms cultured by urban lake waters can be used in the stagnant waters treatment, and it needed time to activate the microbial activity as the pollutant concentration in enclosed surface water was much higher than urban lake waters,even though the bio lms had surfersuddenly increase of NH 4 + -N concentrations. Thus the limited nitri cationcapacity of bio lms may be the reason for high DO values at the beginning of the experiment (Fig. 4a).
In control, NO 2 − -N concentrationstraightly declined to 0 within 6 days. By water cycling, NO 2 − -Nconcentrationincreasedto 2.29 mgL -1 within 17 days then declined to 0 in 5 days (Fig. 5d).While the changes of NO 2 − -N concentrations in SABRs showeda bimodal pattern, with accumulationsduring initial 6 days and then declined, followed by a second increase and decline tendency, which were different from the results in the conditionwithout organic matters (Fig. 4d). Previous studies showed that low organic loading was favorable fornitri cation(Luo and Meng 2020). But NH 4 + -NandNO 2 − -Nconcentrations in FR20, FR40, and FR60increasedduring initial4 days, which was in accordance withdissimilatory nitrate reductionto ammonium (DNRA) (Carlson et al. 2020;Srinandan et al. 2012).Thus, DNRA was suspected to occur when there were abundant biodegradable organic matters in this study. When organic matterswere used up after 6 days, the subsequent decline ofNO 2 − -Nconcentrationwas due tothe transformation ofNO 2 − -Ninto NO 3 − -N (Fig. 5d). The secondincrease and decline of NO 2 − -N concentrationsin four SABRs were similartothe result of no organic mattercondition (Fig. 4d) and ascribed to the nitri cation.
In control, biodegradable organic matters led to low DO,accompaniedby100% removal of NO 3 − -Nin 5 days.While NO 3 − -Nconcentrationsin SABRs only declined in the initial few days when DO were low (Fig. 5a, e).TN in controlwas removed by 37.56 %, and much higher thanby using water cycling and SABRs (removed by 8.62 %, 11.79 %, 7.78 %, 8.02 %,9.94 %in FR20, FR40, FR60, FR80, and water cycling, respectively) (Fig.S3b). These results revealedthatdenitrifying microbeswere abundant in actual surface water, and denitri cation can be easily achieved when DO was low. It also implied that the addition of nitrate can induce denitri cation in black-odorous waterbodies without any other treatment.
As stated above, denitri cation could occur naturallyby mixing the e uent of SABR with the surrounding-untreated Other water quality parameters Biological conversation of NH 4 + -Nto NO x -N often led to pH decline, but in the continue experiment in this study, pH in e uents was higher than in uent (Fig.2b). It may due to the low ow in uent or high HRT, and buffer effect of surface waters. Moreover, the ne wires used in this study can lead to a pH increase due to Fe corrosion(Di Capua et al. 2019).
While in enclosed surface waters, nitri cationled to the decrease of pH ( Fig. 4b;Fig. S3a),promoted the releaseof Fe 2+ and Fe 3+ from ne wires then precipitates with phosphorus.Thus the removal of PO 4 3--P in the batch experiments was due to the ne wires( Fig. 4f; Fig. S3c).    Puri cationperformances by different treatments (organic matters presented)