Post-anoxic integrated biofilm activated sludge wastewater treatment process for emerging contaminant removal Open Access

Emerging contaminants (ECs) are a group of chemicals that are present in the environment at very low concentrations but have the potential to cause risks to the environment. Such contaminants include compounds found in pharmaceuticals, personal care products, hormones, artificial sweeteners, etc. The wastewater treatment plants' (WWTPs) effluent is considered one of the largest sources of ECs entering the environment. The activated sludge process, the most widely used wastewater treatment technique, can only partially remove a variety of ECs (Lindberg et al. 2006; Radjenović et al. 2009). The Modified Ludzack Ettinger (MLE) process, a modification of the conventional activated sludge process, is increasingly being adopted to remove biological nutrients such as N and P from wastewater. However, information on the removal of ECs in such advanced biological systems is limited.

The integrated biofilm activated sludge (IBAS) is an advancement for removing biological nitrogen without increasing the plant surface area. The media made of polyvinyl alcohol (PVA) is a highly porous medium with a high surface area (2,500 m² m⁻³) and density (1.025 g cm⁻³), thus providing high biomass concentration and lesser sludge production. Along with the suspended sludge, the development of different microbial communities in the media can promote the removal of organic carbon and biological nitrogen in the presence of oxygen. The IBAS process has also shown promising results in removing phthalic acid esters from wastewater (Gani et al. 2020). Since the development of microbial communities can also facilitate the higher removal of ECs, the IBAS process was integrated with the existing MLE process to check the influence of the modified treatment configuration on the removal of ECs.

The biodegradation of ECs involves complicated chemical processes and is driven by compound structure and agents of biotransformation. At low concentrations, the catabolic biotransformation of compounds predominates over selective degradation by bioenzymes. Under such situations, environmental conditions play a crucial role in the biodegradation of compounds (Fenner et al. 2021).

The hydraulic retention time (HRT), solids retention time (SRT), temperature, and redox conditions influence the removal of ECs in wastewater treatment systems (Koumaki et al. 2021). There has been sufficient research to prove a positive influence of extended SRT (Clara et al. 2005) and HRT (Joss et al. 2005) on the removal of ECs. The extended periods enable the proliferation of a broader consortium of bacteria for the biodegradation of ECs. Higher temperature also enhances the removal efficiency of ECs. At higher temperatures, the growth of the bacteria and the enzyme activity increase, which increases the biodegradation rate and removal efficiency (Sui et al. 2016). However, despite advancements in biological treatment configurations, limited research has systematically explored how varying redox environments influence the removal mechanisms of ECs in integrated systems such as IBAS. Most existing studies focus on overall removal without resolving the stage-specific contributions of biodegradation and sorption under different redox conditions. This lack of mechanistic clarity restricts process optimization efforts for broader EC attenuation. Therefore, this study aims to fill that knowledge gap by evaluating the fate of 20 ECs across anaerobic, anoxic, aerobic, and PVA-based biofilm reactors. Through detailed mass balance, sorption analysis, and redox stage-specific monitoring, we provide novel insights into how compound-specific properties interact with redox conditions to drive EC removal in an IBAS-enhanced MLE configuration.

Spiking of ECs at high concentrations may lead to misinterpretation of their removal (Boethling & Alexander 1979). Hence, the feed (influent) was spiked with the 20 ECs at an environmentally relevant concentration of 1 μg L⁻¹. Spiking was essential to ensure measurable concentrations for the effective tracking of removal efficiency. ECs were also monitored under unspiked conditions to check the influence of spiking on the removal of ECs and the overall performance of the reactor.

The reactor was operated for 7 months. The fate of ECs in the reactor was studied once it reached the steady state, characterized by similar COD removal. The influent and effluent samples were collected on 23 occasions over 4 months to determine the removal efficiency of each contaminant in the dissolved phase. The mass balance and influence of redox conditions were studied by collecting samples every Monday from the influent and each unit process in eight sampling sessions conducted over a period of 2 months.

The samples were collected in carefully rinsed 1 L amber-coloured glass bottles. One sample aliquot was used for conventional parameter analysis, while the second aliquot was used for EC analysis immediately centrifuged and filtered using a 0.45 μm nylon filter paper (Axiva, Delhi) and stored at 4 °C before further processing.

The conventional parameters including chemical oxygen demand (COD), biological oxygen demand (BOD), total suspended solids (TSS), ammonium nitrogen (NH₄-N), nitrate nitrogen (NO₃-N), total nitrogen and ortho-phosphorus (PO₄-P), mixed liquor volatile suspended solid (MLVSS) and mixed liquor suspended solid (MLSS) were measured according to standard methods (APHA 2017). Other parameters, including dissolved oxygen (DO) and oxidation–reduction potential (ORP), were measured from each reactor using a multi-parameter kit (HACH, USA). pH was measured through a pH meter using a glass electrode (Toshniwal, India).

A total of 20 ECs including pharmaceuticals, such as anti-inflammatory drugs – ibuprofen (IBU), diclofenac (DCF), naproxen (NPX), and ketoprofen (KTP); antibiotic – ciprofloxacin (CFX), sulfamethoxazole (SMZ), trimethoprim (TMP), and enrofloxacin (EFX); psychiatric drug – carbamazepine (CBZ); analgesic – acetaminophen (ACMP) and β-blocker- atenolol (ATL); lipid regulator – gemfibrozil (GEM); natural hormone – testosterone (TEST), progesterone (PROG), estrone (E1), 17β estradiol (E2); synthetic hormone – ethinylestradiol (EED); stimulant – caffeine (CAFF); personal care product – triclosan (TCS); industrial product – 2-hydroxybenzothiazole (2-HYD), were selected for the study. Details of their physico-chemical properties are provided in Table S2. The ECs were selected based on the criteria: (a) reported locally in the region, (b) reported across the world (for comparison purposes), and (c) availability of analytical method. The selected ECs represent a wide spectrum of physico-chemical characteristics, including variation in hydrophobicity (log K_ow), acid dissociation constant (pKa), and molecular weight, which are known to affect their behaviour in different redox environments. These properties can influence the tendency of ECs to undergo biodegradation or sorption and are considered in the subsequent analysis of removal trends across the treatment zones.

ECs in all the phases, including dissolved, particulate, and sludge, were studied. Sample clean-up and extraction were performed through solid-phase extraction (SPE) using Oasis HLB 6 cc, 200 mg cartridges (Waters Corporation, Milford, MA, USA) according to a method described in a previous study (Dubey et al. 2022). Briefly, the sample pH was brought down to 2 using HCl (37%). The cartridges were conditioned using 6 mL each of ethyl acetate (Grade-HPLC; SRL, India), acetonitrile (Grade-HPLC; RANKEMTM, India), and methanol (Grade-HPLC; RANKEMTM, India) and equilibrated using 6 mL water at pH 2. About 200 mL of the sample was loaded on the cartridges at a flowrate of 3 mL/min using a vacuum pump (GAST, MI, USA). The cartridges were later washed using 6 mL water:methanol (95:5, pH 2) and dried under vacuum for 20 min. The sample was eluted sequentially using 4 mL of ethyl acetate, acetonitrile, and methanol, each containing 2% ammonia solution. The eluent was evaporated to near dryness under nitrogen using a nitrogen evaporator, reconstituted to 1 mL using a solution of water:acetonitrile (90:10), and spiked with 100 μg/L of internal standard carbutamide-d9. For the particulate and sludge analysis, the sample was centrifuged for 10 min at 8,000 rpm and the moisture content of the collected solids was determined. About 0.5 g (dry weight) of the sample was extracted using a solution of 10 mL acetone and MeOH (1:1) by shaking vigorously for 1 min, followed by ultrasonication in a bath homogenizer for 15 min at 40 °C. The mix was centrifuged at 6,000 rpm for 10 min, and the supernatant was collected. The extraction step was repeated two more times. The final supernatants were mixed, evaporated to 2 mL under nitrogen, and mixed with 200 mL of milli Q (MQ) water at pH 2. The subsequent extraction and clean-up steps are the same for water samples. The extraction and analysis were completed within 48 h of sample collection.

The chromatographic separation was performed using a Sunfire C18 column (5 μm, 4.6 × 250 mm) fitted in a Waters Alliance HPLC coupled to the SQD2 mass detector. The chromatographic and MS parameters are provided in a previous study (Dubey et al. 2022). Briefly, two different combinations of mobile phases (1) water + 0.1% formic acid: methanol + 0.1% formic acid (A:B) and (2) 5 mM ammonium acetate: acetonitrile + 0.1% formic acid (A:B) were used. A 30 min gradient programme was used as follows: 2%B for 3 min, a linear ramp to 70% B for the next 7 min, a hold for 15 min, a linear decline to 2% B for 3 min, and a hold for 2 min for the separation. The other LC and MS parameters, such as injection volume: 10 μL, flow rate: 0.3 mL/min, column temperature: 40 °C, capillary voltage: 3 kV, desolvation temperature: 400 °C, cone gas flow: 50 L/h, and desolvation gas flow: 800 L/h, were kept constant.

The recovery of SPE and process efficiency of the analytical method were determined using the methods described in a previous study (Dubey et al. 2020). An 8-point calibration curve with r² > 0.99 was used to ensure linearity. The intra- and inter-day variability was measured as the standard deviation of five consecutive injections of a sample measured on the same day and over three consecutive days and was found to be 2.1 and 5.4%, respectively. The limit of detection (LOD) and limit of quantification (LOQ) were measured as 3 and 10 times the signal-to-noise (s/n) ratio and were determined in the range from 1 to 7.5 ng/L and 3 to 25 ng/L, respectively, for aqueous samples and 1.5 to 22.5 and 5–75 ng/g, respectively, for the solid (particulates and sludge) samples. ECs detected at a concentration above the LOD but below the LOQ were considered to occur at a concentration of LOQ/2. The recoveries of the SPE process and LOQ values for the 20 ECs in the dissolved and solid phases are provided in Table S3.

Method blanks were injected to check for contamination during sample preparation and analysis as a measure of quality control. A spiked blank was used to check the accuracy of the analytical method. Mobile phase solutions were injected as blanks after every five injections to check for analyte carryover. A 10 μg/L standards mix was injected to check for instrumental drift.

The water quality parameters of the influent and the final effluent are provided in Table S4, and some operating conditions in the individual treatment reactors are provided in Table 1. The IBAS system achieved the required effluent quality, aligning with the regulatory standards throughout the sample collection period.

The DO levels were maintained at 3.48 ± 0.22 mg/L in the PVA reactor and 4.37 ± 0.39 mg/L in the aerobic tank, while ORP values indicated clear redox separation: −136 ± 22 mV (anaerobic), −117.8 ± 28.9 mV (anoxic), and 57.75–67.6 mV in oxic zones. The MLSS concentrations across reactors ranged from 4,530 to 4,801 mg/L, with corresponding MLVSS values between 2,284 and 2,611 mg/L. The food-to-microorganism (F/M) ratio decreased progressively along the treatment chain from 0.39 ± 0.06 kg BOD₅/kg MLVSS/day in the anaerobic reactor to 0.08 ± 0.01 in the PVA reactor and 0.03 ± 0.006 in the aerobic reactor. These parameters confirm the establishment of distinct redox environments and support the evaluation of EC fate across different biological conditions.

The average % removal of ECs (n = 23) in the IBAS configuration is provided in Figure S1. The average removal varied from 11.8% for CBZ to 99.5% for CAFF. As the system was under stable conditions, variation in the % removal of ECs was limited, evident from the low standard deviation. The SRT during the entire period was maintained at around 18.4 days.

Based on the % removed, the ECs were categorized into four groups: excellent: > 80%, good: 50–80%, moderate: 30–50%, and poor: < 30%. Based on this categorization, 9 out of the 20 ECs showed excellent removal. On average, the IBU, KTP, ACMP, ATL, TEST, PROG, E2, CAFF, and TCS were removed by more than 80%. DCF, NPX, CFX, TMP, EFX, GEM, E1, EE2, and 2-HYD were removed from 50 to 80% (9 out of 20 ECs). Though DCF is also considered bio-recalcitrant, the degradation is enhanced with varying redox conditions (anoxic–oxic), while the sorption of DCF to sludge is limited (Zhang & Li 2011). The removal efficiency of EE2 (56.5%) is close to the average removal (68.3%) reported from 282 WWTPs from 29 countries (Tang et al. 2021). In our study, TMP was removed by 52.2%, similar to the degradation reported by Batt et al. (2006). GEM was removed by 69.4%, which is equal to the average removal reported by Gros et al. (2010) from seven WWTPs in Spain. Integration of anoxic and aerobic treatment conditions has been shown to promote the degradation of 2-HYD and can be removed by >50% (Mazioti et al. 2015). 2-HYD is a readily degradable compound with removal efficiency reported from 50 to 70% (Kloepfer et al. 2005).

SMZ was only moderately removed (41.6 ± 5.6%), while CBZ was poorly removed (11.03 ± 3.5%). SMZ is reported with an extensive range of % removal during biological treatment. The % removal is reported as low as 20% (Brown et al. 2006) to 55–74% by other studies (Zhang & Li 2011). Such a variation could be due to the retransformation of metabolites like N4-acetylsulfamethoxazole during the treatment or variation in the sample collection strategy (grab sampling and composite sampling) (Zhang & Li 2011). CBZ is a widely recognized drug reported with negative to <10% removal in biological WWTPs (Zhang et al. 2008). This is because of its recalcitrance towards biological degradation and negligible sorption on sludge (K_d = 1.2 L kg_ss⁻¹) (Ternes et al. 2004).

In line with the results obtained in this study, Tran et al. (2009) studied the role of redox conditions and reported high removal of DCF by heterotrophic bacteria. A higher rate of degradation of NPX and TMP could be due to their chemical structure. NPX can be bio-transformed by o-demethylation of the aromatic methoxy group (Ghattas et al. 2017). While the metabolite (o-desmethylnaproxen) is recalcitrant under anaerobic conditions, it can be degraded under the post-anerobic conditions (Ghattas et al. 2017). In TMP, the substituted pyrimidine ring's structure is degradable under anaerobic conditions (Adrian & Suflita 1994). The natural hormone TEST is reported to be mainly eliminated in the anaerobic tank in an anaerobic–anoxic–oxic configuration (Yu et al. 2019). TCS showed an almost equal degradation rate under PVA tank and anoxic conditions. Overall, the results indicate that aerobic conditions favour most ECs' biodegradation. Some other ECs are rapidly degraded under anaerobic or anoxic conditions, highlighting the required distinctive redox conditions in wastewater treatment schemes to efficiently remove a wide variety of ECs in the incoming wastewater.

Different treatment environments present a broader consortium of bacteria to thrive, which can remove the ECs through different catabolic pathways (Cydzik-Kwiatkowska & Zielińska 2016). In the current study, ECs were removed in different reactors in the order aerobic > anoxic > anaerobic. A similar trend was observed by Joss et al. (2004) with the biodegradation rate (K_bio) as 0.25, 0.125, and 0.0625 L gSS⁻¹ h⁻¹, respectively, under aerobic, anoxic, and anaerobic conditions.

The sorption coefficient, K_d, in the solid phase of the reactors and final sludge were calculated according to Equation (5) based on the ratio of the concentration of ECs in the solid and dissolved phases. The average log K_d values of the solid samples are provided in Table S5. The log K_d values of the final sludge were higher than in the treatment reactors. The differences in the resulting K_d values could be due to differences in the physico-chemical properties of the contaminant and/or the biodegradability of the compound. All the acidic compounds with low pKa values would exist in the ionic phase in the reactors and hence stay in the dissolved form. The log K_d of such ECs, for example non-steroidal anti-inflammatory drugs (NSAIDs), was less than the other ECs. The other compounds with a negative charge but high log K_ow values, like TCS (log K_ow = 4.8), had a high K_d value due to high hydrophobicity. For compounds such as ACMP, CAFF, and IBU, high K_d values in the aerobic reactor than under other redox conditions are due to the high degradability of the ECs, where the removal due to degradation outpaced the sorption of ECs on the solid phase. CBZ showed low K_d values, indicating the occurrence of CBZ predominantly in the dissolved phase. High concentrations of CFX in different reactors and final sludge indicate high sorption of CFX to sludge particles due to electrostatic interactions. The K_d of 2-HYD in the activated sludge was found to be 147 ± 29 L kg⁻¹ in the activated sludge and consistent with the available literature (Mazioti et al. 2015), indicating that sorption contributes negligibly to removing 2-HYD. Similar observations have been reported in related studies (Martín et al. 2012; Dubey et al. 2022).

The removal behaviour of ECs under different redox environments was closely associated with their physico-chemical properties. Compounds with low pKa values (e.g., NSAIDs, such as IBU, NPX, and DCF) predominantly exist in ionized forms at typical wastewater pH, reducing their sorption potential in all reactors, particularly under anaerobic conditions. Conversely, compounds with higher log K_ow values and moderate hydrophobicity, such as TCS (log K_ow = 4.8) and GEM (log K_ow = 4.77), demonstrated higher K_d values, reflecting greater partitioning to the solid phase. Hydrophilic compounds like CAFF and ACMP, despite low log K_ow, were efficiently removed due to their high biodegradability, especially in aerobic zones. These patterns indicate that redox-specific microbial activity and compound-specific chemistry jointly influence removal mechanisms.

The % mass load distribution of ECs is shown in Figure 4(c). All but CBZ and EE2 showed a decrease in the mass load of ECs. The NSAIDs: DCF, NPX, antibiotics, CBZ, and EE2 had a high mass load in the effluent. CFX and TCS were detected with a higher mass load than other ECs in the final sludge. Biodegradation/biotransformation was the dominant removal mechanism of ECs, whereas sorption was the minor removal mechanism, consistent with log K_d trends and the low EC mass detected in sludge. The pharmaceuticals IBU, KTP, ACMP, CAFF and natural hormones, such as TEST and PROG, were highly removed (>80%) via biotransformation/biodegradation. ECs, such as ATL, GEM, E1, E2, and 2-HYD, also showed good removal from 50 to 80%, while the other ECs showed moderate to poor removal. These quantitative observations demonstrate that aerobic and PVA reactors contributed the most to mass removal, especially for readily biodegradable ECs. The reactor-wise load tracking provides a mechanistic basis for evaluating removal performance and complements the concentration-based analysis.

The observed enhancement in EC removal at lower F/M ratios can be attributed to several microbial and biochemical factors. Under substrate-limited conditions (i.e., low F/M), the microbial community tends to shift toward slower-growing organisms capable of degrading more complex and persistent compounds through co-metabolic or stress-induced enzymatic pathways. Low F/M ratios also correspond with extended SRT, which promotes the accumulation of specialized microbial populations and enzymes associated with the degradation of low-concentration ECs (Gallardo-Altamirano et al. 2021). Additionally, reduced substrate availability minimizes competition from fast-growing heterotrophs, allowing for better expression of pollutant-degrading genes in oligotrophic microbial taxa. These factors collectively contribute to the higher EC removal observed in the aerobic reactor, which operated at the lowest F/M ratio in this study.

The study reported the removal of ECs in a lab-scale post-anoxic IBAS process. The treatment configuration was carefully assessed, and the operating conditions impacting the removal of ECs were examined. The total removal efficiency of the 20 ECs was around 87%. The removal in each reactor was observed in the order aerobic > PVA > anoxic > anaerobic. The sorption coefficient, K_d, was the highest for CFX and TCS and varied greatly depending on the chemical properties of the ECs. The % mass load removal for the 20 ECs was around 85%. However, when excluding ACMP and CAFF from the list, the % mass load removal was 56%, highlighting the limited removal of antibiotics, NSAIDs, CBZ, and hormones. Biodegradation/biotransformation was the primary removal mechanism in removing ECs, while sorption was the secondary mechanism. The F/M ratio showed an inverse relation with the ECs removal, with a correlation coefficient of 0.83. However, deeper insights are needed to correlate the mutual functionality of microbial consortia in the presence of various ECs or how the microbes adapt/establish in the presence of ECs.

FAME (Fate and Management of Emerging Contaminants) Project, jointly funded by the Department of Science and Technology, Government of India (DST/TM/INDO-UK/2K17/66(C)), and the UK Natural Environment Research Council (NE/R003548/1) under India-UK Water Quality Programme supported the research work.

M.D. was responsible for conceptualization, visualization, formal analysis, methodology, investigation, validation, data curation, and writing the original draft. P.G. contributed to formal analysis, data curation, and writing the original draft. B.P.V. handled funding acquisition, resources, writing the original draft, writing review and editing, and supervision. A.A.K. managed funding acquisition, project administration, writing review and editing, and supervision.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.