Fate of trace organic contaminants in oxic-settling-anoxic (OSA) process applied for biosolids reduction during wastewater treatment

This study investigated the fate of trace organic contaminants (TrOCs) in an oxic-settling-anoxic (OSA) process consisting of a sequencing batch reactor (SBR) with external aerobic/anoxic and anoxic reactors. OSA did not negatively affect TrOC removal of the SBR. Generally, low TrOC removal was observed under anoxic and low substrate conditions, implicating the role of co-metabolism in TrOC biodegradation. Several TrOCs that were recalcitrant in the SBR ( e.g. , benzotriazole) were biodegraded in the external aerobic/anoxic reactor. Some hydrophobic TrOCs ( e.g., triclosan) were desorbed in the anoxic reactor possibly due to loss of sorption sites through volatile solids destruction. In OSA, the sludge was discharged from the aerobic/anoxic reactor which contained lower concentration of TrOCs ( e.g. , triclosan and triclocarban) than that of the control aerobic digester, suggesting that OSA can also help to reduce TrOC concentration in residual biosolids.


Introduction
Wastewater treatment plants (WWTPs) commonly use the conventional activated sludge (CAS) process to treat municipal and industrial wastewater. CAS involves the conversion of soluble organic matter in wastewater into settleable biomass called "sludge" in an aeration tank. The removal of organic matter is affected by the food-tomicroorganism ratio, which is maintained by wasting excess sludge (Tchobanoglus et al., 2003). Previously, sludge was dumped into the ocean but this approach has been banned due to adverse impact on marine life (Foladori et al., 2010). Currently, sludge is treated to reduce volatile solids and water content and then landfilled or incinerated (Semblante et al., 2014;Tchobanoglus et al., 2003). These approaches have high energy requirements and significantly increase the overall cost of sludge management (Semblante et al., 2014). Sludge is also converted into nutrient-rich "biosolids" applied on agricultural land -an approach that facilitates the recovery of carbon, nitrogen, and phosphorous (Foladori et al., 2010;Semblante et al., 2014). However, reducing the volatile solids and pathogen concentration of biosolids to levels that meet regulatory standards can be a technically challenging and expensive exercise involving multiple processes (e.g., anaerobic digestion, reduction when DO is manipulated) (Wei et al., 2003). Others have high process efficiency (e.g., 100% reduction when sludge is destroyed by ozonation) but require high capital investment and operation cost (Foladori et al., 2010).
The oxic-settling-anoxic (OSA) process is an emerging approach to decrease sludge production. OSA involves the insertion of one or more external anoxic reactors in the return activated sludge loop of CAS. Compared to other sludge reduction schemes, OSA has low capital investment and energy requirement. OSA cycles sludge between conditions that are rich (aeration tank) and deficient (external anoxic reactor/s) in oxygen and substrate (Semblante et al., 2014). Consequently, volatile solids are destroyed and converted into inert forms in the external anoxic reactor/s. Using real sewage as feed, recent studies have shown that OSA can reduce the sludge yield (mass of sludge produced/mass of substrate consumed) of sequencing batch reactors (SBRs) by more than This study aims to determine the sorption and biodegradation of TrOCs in OSA operated using real wastewater.
The TrOC concentrations in the effluent and sludge of an OSA system were compared to that of a control system to gain insight on the effects of sludge interchange between different redox regimes on the fate of TrOCs. Furthermore, the fate of TrOCs were determined at different external reactor SRT (SRText), i.e. aerobic/anoxic and anoxic reactors and control aerobic digester. The findings of this study are relevant to the assessment of the TrOC discharge from OSA and in the future development of TrOC mitigation or treatment approaches.

Municipal wastewater
Municipal wastewater was obtained from the beginning of primary sedimentation tank of Wollongong WWTP fortnightly and stored at 4 ºC prior to use to minimise chemical reactions and microbial activity. The basic properties of the municipal wastewater are listed in Supplementary Table S1.

Reactor configuration and operation
The reactors used in this study were previously described (Semblante et al., 2016). Briefly, the OSA system consisted of a sequencing batch reactor, SBROSA (5 L), attached to external aerobic/anoxic (2 L) and anoxic reactors (2 L) ( Figure 1a). Meanwhile, the control system consisted of SBRcontrol (5 L) attached to a single-pass aerobic digester (2 L) (Figure 1b).

[Figure 1]
SBROSA and SBRcontrol were fed with municipal wastewater (Section 2.1). They were operated at 4 cycles/day and HRT of 12 hours. Each cycle comprised of 15 min of filling, 4 hours and 30 min of aeration, 1 hour of settling, and 15 min of decanting. The temperature of both SBRs was maintained at 25 ºC using a water bath.
The SRT of both SBRs (SRTSBR) was maintained at 10 d by regular sludge wastage (W) (Figure 1) throughout the experimental period (Table 1).

[Table 1]
The aerobic/anoxic reactor of the OSA system was intermittently aerated (i.e., aeration was turned on for 8 h and then turned off for 16 h) using an air diffuser placed at the bottom of the reactor (Figure 1a). The DO concentration of the reactor (measured as described in Section 2.3.2) when aeration was turned on and off was 4.6±1.0 mg/L (n=62) and 0.4±0.2 mg/L (n=62), respectively. The anoxic reactor was kept airtight using a silicone-lined cap with inlet and outlet ports. The temperature of both external reactors was maintained at 25 ºC using a water bath.
The aerobic/anoxic reactor was fed with sludge from SBROSA thickened by centrifugation (Beckman Coulter, USA) to 5-10 g/L (q1). Thirty-three percent (33%) of sludge from the aerobic/anoxic reactor was transferred to the anoxic reactor (q2) and 67% was wasted (q3). The total SRT of the external reactors (SRText) was varied from10-40 d (Table 1) by adjusting sludge wastage (q3). The wasted sludge was thickened to 16-24 g/L by centrifugation (Beckman Coulter, USA) for 10 min at 3,267 g. The supernatant was returned to SBROSA, and the pellet was discarded. Sludge from the anoxic reactor was returned to the aerobic/anoxic reactor (q4) and SBROSA (q5).
The aerobic digester (Figure 1b) was continuously aerated using an air diffuser. The DO was 6.2±0.19 mg/L (n=62) and the temperature was maintained at 25 ºC using a water bath. The SRT of the aerobic digester (SRText) was varied from 10-40 d (Table 1) by adjusting sludge wastage (Qout). The aerobic digester was fed with sludge from SBRcontrol (Qin) that has been thickened to 5-10 g/L by centrifugation (Beckman Coulter, USA) for 10 min at 3,267 g. The supernatant produced after thickening was discarded to eliminate the potential impact of return flow on sludge production and/or substrate consumption of SBRcontrol. This facilitated the comparison of two SBRs (SBROSA vs. SBRcontrol) with and without sludge interchange.  The experimental sludge yield (Y) of the SBRs was defined as:

Analytical techniques
where P is the sludge produced in terms of mixed liquor volatile suspended solids (MLVSS) and C is the substrate consumed in terms of total chemical oxygen demand (tCOD). Sludge yield was derived from the slope of the linear regression of the cumulative sludge produced versus the cumulative substrate consumed. The cumulative values were obtained by incrementing the variations in sludge production and substrate consumption in previous sampling intervals.

TrOC concentration in external reactor sludge
To gain further insight in the sorption of and biodegradation of TrOCs, the TrOC concentration of sludge (in ng/L) entering the external reactors i.e., the aerobic/anoxic and anoxic reactors of OSA and the aerobic digester of the control system were estimated as described in Supplementary Table S4.  (Table 1).

Sludge reduction by OSA
Furthermore, regardless of the SRText, SBROSA and SBRcontrol had similar tCOD and ammonia concentration in the effluent (Supplementary Figures S5 and S6). This suggests that OSA did not affect the overall wastewater treatment efficiency of the main aeration tank (SBROSA).

TrOC concentration in the influent
The sampling campaigns at different SRText fell at different seasons (Table 1)

TrOC concentration in SBR sludge
The concentrations of selected TrOCs in the solid phase of sludge are presented in Figure 3b.  Table   S9), and thus dramatic change in the TrOC profile of sludge was not observed (Supplementary Table S7).
Most of the TrOCs that showed a significant difference (>30%) between SBROSA and SBRcontrol sludge (Supplementary Table S10) were non-or partially biodegradable (e.g., TCEP, benzophenone, and others) (Section 3.3.1), which explains why they were detected varying amounts in the sludge. Highly biodegraded compounds such as caffeine, paracetamol, and ibuprofen also showed different concentrations in SBROSA and SBRcontrol sludge because the residual sludge concentration of these compounds was negligible compared to the influent load (1,000-80,000 ng/L). In contrast, highly biodegraded compounds like ketoprofen and naproxen did not have high concentration in sludge and no significant variation between the two SBRs was detected.

Impact of redox regimes in OSA external reactors
The potential impact of additional redox regimes on the fate of TrOCs in OSA was assessed. The aerobic/anoxic reactor received sludge from SBROSA and the anoxic reactor. It had ORP of 120±20 mV (n=34) and 40±20 (n=34) when aeration was turned on and off, respectively. Also, it was deficient in substrate because biodegradable COD has already been consumed in the preceding reactors. Meanwhile, the anoxic reactor was not deficient in substrate and had high methanogenic activity (Wijekoon et al., 2015). In this study, the ORP of the anoxic reactor (-450±30 mV; n= 34) was low but there was no methanogenic activity (indicated by biogas production) due to substrate deficiency. Although a relationship between biodegradation and SRT was observed for the aforementioned compounds, the majority of the load from the incoming sludge was not biodegraded probably because co-metabolic degradation pathways were not activated in the absence of substrate. The residues partitioned in varying concentrations in the aqueous and/or solid phase of anoxic sludge (Supplementary Figure S13).

SBRcontrol vs. aerobic digester: Impact of substrate deficiency
Aerobic digestion involves the treatment of sludge in a completely mixed aerated reactor. The fate of TrOCs in the aerobic digester was investigated to assess TrOC discharge from a conventional sludge treatment unit ( Figure 5). Furthermore, SBRcontrol and aerobic digester were both aerobic reactors, but the former was fed with influent (municipal wastewater) with relatively high concentration of TrOCs and the latter was fed with sludge containing low concentration of readily biodegradable sCOD and reduced concentration of TrOCs (Section 2.2).
Thus comparison of SBROSA and the aerobic digester helps determining the impact of substrate deficiency in TrOC removal (Supplementary Figure S14 and S15).
Generally, with a few exceptions (Section 3.4.1), treatment in SBRcontrol resulted in (i) up to 80% biodegradation of hydrophilic TrOCs especially those with EDG and, (ii) poor biodegradation of hydrophobic TrOCs especially those with EWG. On the contrary, only estrone (a hydrophobic TrOC that was poorly biodegraded in SBRcontrol, Section 3.3) was consistently biodegraded at different SRText in the aerobic digester. Additionally, a few TrOCs (e.g., caffeine, naproxen, and gemfibrozil, discussed in Section 3.5) were highly biodegraded in the aerobic digester at high SRText only (40 d). This demonstrates that the biodegradation of many TrOCs under aerobic condition occurs only when primary substrate is available (Semblante et al., 2015b).

Insights on the TrOC discharge from OSA
TrOC discharge from the particular OSA configuration used in this study was assessed by comparing TrOC concentrations in SBROSA, the aerobic/anoxic reactor (where sludge is discharged from the OSA system, Section 2.2), and the control aerobic digester (where sludge is discharged from the control system, Section 2.2).
The aerobic/anoxic reactor (Figure 4) generally showed lower concentration of many TrOCs in both aqueous and solid phases than SBROSA (Figure 3) given that the majority of the contaminants have already been biodegraded in the main aeration tank.
The aerobic/anoxic reactor also enhanced the biodegradation of estrone, oxybenzone, and benzotriazole (Section 3.5.1). However, non-biodegradable TrOCs (e.g., triclosan and triclocarban) accumulated in the aerobic/anoxic reactor and therefore the solid phase concentration was higher than that of SBROSA (Section 3.5.1). In other words, treatment of sludge in the external reactors enhanced the biodegradation of some TrOCs (e.g., benzotriazole, Figure 4a) but resulted in the accumulation of others (e.g., triclosan, Figure 4b) especially those that are hydrophobic and non-biodegradable in either aerobic or anoxic condition.
Notably, this particular OSA configuration discharges sludge from an aerobic/anoxic reactor rather than an anoxic reactor, which is commonly found in literature (Goel & Noguera, 2006;Semblante et al., 2014). The current study revealed that the aerobic/anoxic treatment results in greater biodegradation of TrOCs than the anoxic treatment (Section 3.5.1). Moreover, the destruction of volatile solids in the anoxic reactor caused desorption of some TrOCs (e.g. paracetamol, sucralose, and bisphenol A) from the solid phase of sludge and consequently increased TrOC concentration in the aqueous phase (Section 3.5.2). This is an indication that the current OSA configuration has potential to have lower TrOC discharge than others involving a single external anoxic reactor.
Generally, the aerobic/anoxic and anoxic reactors of OSA resulted in the biodegradation of a greater number of TrOCs than the aerobic digester. The superior performance of the aerobic/anoxic reactor can be attributed to the variation in redox conditions, which gave rise to nitrifying/denitrifying bacteria that potentially facilitated the biodegradation of some recalcitrant TrOCs (Section 3.5). Furthermore, the concentration of highly sorbing TrOCs (e.g., triclosan and triclocarban) in the aerobic digester (406-10,413 ng/g MLSS) was higher than that of the aerobic/anoxic reactor of OSA (266-8,384 ng/g MLSS). This shows that OSA has potential to yield higher quality biosolids compared to aerobic digestion.

Conclusion
OSA did not affect the effluent TrOC concentration of the SBR. However, the biodegradation of estrone, benzotriazole, and benzophenone was enhanced in the aerobic/anoxic reactor. Generally, aerobic/anoxic condition favoured TrOC biodegradation than anoxic condition. Some TrOCs underwent desorption from sludge due to volatile solids destruction under anoxic condition. The concentration of highly sorbing and recalcitrant TrOCs (e.g., triclosan) in the aerobic/anoxic reactor was lower than that of the control aerobic digester. This suggests that the final sludge residue generated by OSA have potential to have lower TrOC content than that of CAS paired with aerobic digestion.

Acknowledgements
The    The values are the average of two measurements (n=2). The asterisks (*) represent contaminants that were not analysed in a particular sampling campaign. The arrows (→) denote contaminants that were highly biodegraded in the aerobic/anoxic reactor. Only estrone was highly biodegraded in the anoxic reactor.

Figure 5.
Concentration of selected TrOCs the (a) aqueous and (b) solid phase of sludge in the external control aerobic digester when SRTSBR was maintained at 10 d and SRText was varied (10-40 d). The values are the average of two measurements (n=2). The asterisks (*) represent contaminants that were not analysed in a particular sampling campaign. The arrows (→) denote contaminants that were highly biodegraded in the aerobic digester (estrone only).