Long-term effect of metal oxide nanoparticles on activated sludge

Nanoparticles (NPs) are incorporated into various domestic and personal care products due to their distinctive characteristics. According to an inventory by Project on Emerging Nanotechnologies (PEN), more than 1,701+ manufacturer-identified, nanotechnology-enabled goods have entered the open market since 2005 (Woodrow 2011). According to PEN, the most frequently found NPs in consumer products are silver (48%), titanium dioxide (10%), carbon (9%), silicon dioxide (5%), zinc oxide (4%) and gold (3%). For example, recent studies identified the release of Ag ion concentration of 1.3 mg/L from nanosilver coated socks while washing (Benn & Westerhoff 2008). Further, Kaegi et al. (2010) reported 145 μg/L Ag⁺ ions in storm water runoff from Ag NPs which are used in outdoor paintings. Due to large scale production and usage, a significant fraction of NPs reaches wastewater treatment plants (WWTPs) and only a small fraction of NPs is discharged into environment (Wang et al. 2012). Currently, anticipated NP concentrations lies in ng/L for natural waters (Gottschalk et al. 2009). However, there are concerns about potential effects of NPs presence in wastewater on performance of WWTPs.

Recently, researchers have reported that NPs such as TiO₂, ZnO, CeO₂, and Ag could possibly affect microbial community, chemical oxygen demand and nitrogen/phosphorus removal from wastewater (Liang et al. 2010; Hou et al. 2012; Wang et al. 2012). Interestingly, Liang et al. (2010) reported the effect of 1 mg/L Ag NPs on activated sludge bacteria (12 h exposure) which resulted in an increase in levels of NH₄⁺ and NO₂^- at the effluent of a modified activated sludge treatment system. Further, studies observed that 0.5 mg/L of Ag NPs had no effect on chemical oxygen demand (COD) removal in sequential batch reaction (SBR) (15-day operation) (Hou et al. 2012). Previous studies have focused on short-term exposure of individual NPs (Ag, TiO₂ Cu, Zn, carbon nanotubes) on mixed culture (activated sludge), however, limited knowledge is available when determining the effects of mixture of NPs (both short- and long-term) on biological reactor performance. Recently, Sundaram & Kumar (2015) showed that short-term exposure (5 h) of mixture of metal oxide NPs to activated sludge can have influence on COD of wastewater and indicated the need for conducting long-term study. Several studies (Zheng et al. 2011; Wang et al. 2012) in the past have investigated the effect of individual NPs on both aerobic and anaerobic microorganisms for a maximum of 120 days, but still much focus had not been laid in determining the effect of mixture of NPs (for both short- and long-term exposure). As wastewater contains more than one type of NP, it is necessary to understand interaction of both individual and mixture of two or more NPs on bacteria in order to quantify their subsequent effects on performance of biological reactors.

The overall objective was to investigate the long-term impacts of two metal oxide NPs (Ag₂O, TiO₂) and their mixture on reduction in COD elimination upon exposure to activated sludge. These NPs were chosen based on the fact that they are produced in greater quantities and are found in a wide range of commercial products (Woodrow 2011). Further, their mobility and persistence leads to interaction with organisms that are potentially harmful in nature (Keller et al. 2010). Experiments were conducted at two different organic concentrations (COD: 250 and 500 mg/L), NP concentrations (1 and 10 mg/L) in a sequential biological reactor (SBR) and parameters such as COD, mixed liquor suspended solids (MLSS), and metal ion concentration were measured. Findings of this study could provide (i) an estimate on reduction in COD due to exposure of NPs (individual and mixture), (ii) fate in wastewater, and (iii) comparison on toxicity (individual versus mixture of NPs).

Two metal oxide NPs, namely silver (I) oxide (Ag₂O) and titanium dioxide (TiO₂) were purchased from Sigma-Aldrich, St Louis, MO, USA (Ag₂O: purity: 99%, color: dark grey; TiO₂: purity: 99.7%, color: white). Respective NP stock suspensions of 1,000 mg/L were prepared by adding 1 g of NPs in 1 L of ultrapure water and suspensions were sonicated for 15 min to prevent NP aggregation (Keller et al. 2010). For a mixture of NPs, equal amounts of respective individual NP solution (Ag₂O and TiO₂) were used to prepare the respective mixture concentration. For example, 1 mg/L of NP mixture solution was prepared using 10 mg/L of Ag₂O and TiO₂ NPs stock solution. According to literature (Gottschalk et al. 2009; Kiser et al. 2009; Zheng et al. 2011), 1 mg/L was considered to be environmentally relevant concentration in wastewater. The reason for choosing higher NP concentration (10 mg/L) is that, due to the large scale production of NPs, their release into wastewater will also increase where studies (USEPA 2009) have reported 6 mg/g of titanium in biosolids of WWTPs in the USA and researchers (Nyberg et al. 2008) in the past have suggested the usage of higher NP concentration to determine its toxicity.

Eight identical biological reactors of 1 L capacity, each in triplicates, were started. Activated sludge (for inoculation) was received from the aeration tank of a municipal WWTP in South Delhi, India (capacity: 14 million L per day; operation: extended aeration process). Collected sludge was aerated for 3 days and then used to inoculate each reactor to achieve desired biomass concentration of approximately 2,500 mg SS/L. Solids were retained in the reactor for a period of 29 days to ensure optimum growth conditions and reactors were run with a 24 h hydraulic retention time. Synthetic wastewater was fed to reactors for 30 min followed by 22 h of aeration, 1 h settling and 30 min for decanting (Gao et al. 2011). This water was prepared in distilled water with glucose and peptone as a prime nutrient source. The chemicals used in preparing synthetic wastewater (per liter) were: NH₄Cl, 125 mg; NaHCO₃, 250 mg; and KH₂PO₄, 25 mg; trace element solution, 0.1 mL. The trace element solution contained (mg/L): MgSO₄·7H₂O, 20; FeCl₃, 15; CuSO₄, 30; MnSO₄·H₂O, 50; CoCl₂·6H₂O, 50; KCl, 18. The reactors were maintained at room temperature (27 ± 1 ̊C) with natural lighting and pH was maintained between 7.5–7.9. No additional UV source of light (through lamp) was provided. While the average outdoor incident luminescence during the test period was found to be 50 klux/h (at noon), 20 klux/h (in the evening) and light intensity inside the room was observed to be 6.7 klux/h. Intensity inside the room was not found to be sufficient to activate TiO₂ NPs and thus activated sludge was exposed only to TiO₂ NPs (without external activation). Air was introduced from the top of reactor through a diffuser using an air pump (0.57 L/min) which was found to be sufficient enough to mix the constituents inside the reactor. The maximum dissolved oxygen concentration during the aeration period reached 6.5–7.5 mg/L, while the average concentration ranged between 4–5 mg/L. Two organic concentrations (input COD), namely 500 and 250 mg/L, were maintained in biological reactors.

Synthetic wastewater spiked with desirable NPs concentration was fed into biological reactors. Before exposing to NPs, all the reactors were operated until a steady state was achieved (for all reactors, COD removal was observed to stabilize within the first 45 days of operation). Two sets of eight biological rectors were started: (a) control reactors (without NPs) (R1 and R2), (b) Ag₂O NPs reactors (1 and 10 mg/L) (R3 and R4), (c) TiO₂ NPs reactors (1 and 10 mg/L) (R5 and R6), and (d) reactors exposed to mixture of both NPs (1 and 10 mg/L) (R7 and R8), for obtaining two independent observations. Reactor performance was monitored by measuring COD in filtered supernatant every alternate day and MLSS for every 15 days. For MLSS measurement, samples were collected at the end of each complete mixing and aeration period (i.e. mixed liquor) prior to the settling period. The schematic representation of the methodology is presented in Figure S1 (available with the online version of this paper).

Z-average size of NPs in synthetic wastewater (i.e. going inside the reactor) was measured using dynamic light scattering (DLS) technique (Particle Sizer Nicomp 380 ZLS). DLS was carried out only before feeding into the reactor and no observations were made in determining the size of NPs in supernatant. Supernatant was analyzed for COD, pH, dissolved oxygen and MLSS as per Standard Methods (APHA et al. 2005). Both exposed and unexposed reactor sludge were analyzed for surface morphologies (transmission electron microscopy, TEM), elemental composition (energy-dispersive X-ray spectroscopy, EDX) and ion content (atomic absorption spectroscopy). TEM was used to determine the size, shape and position of NPs in exposed and unexposed sludge. TEM images were acquired on a Hitachi H8100 TEM instrument at 200 keV. For TEM analysis, samples were dried, suspended and grinded in isopropanol. The prepared solution was sonicated in an ultrasonic bath for 15 min before analysis. EDX analysis was used to confirm the presence of metals in sludge and Ion analysis was carried out for both supernatant and sludge (exposed and unexposed) using atomic absorption spectroscopy. For ion analysis, samples (10 mL) were digested with HNO₃, H₂SO₄ at 180 °C in a hot plate and further diluted to 50 mL before measurement. Total metal ion concentration associated with NPs was measured by atomic absorption spectrometer (AAS-ECIL-4141) and the detection limits for Ag and Ti was 0.1 mg/L.

All results are expressed in mean ± standard deviation. Remaining COD of different samples were statistically compared using student's t-test for a 0.05 level significance. In addition, an analysis of variance (ANOVA) was used to determine the significance of parameters influencing COD removal in exposed reactors and p < 0.05 was considered to be statistically significant. The nature of toxicity of mixture of NPs was determined using the method used by Ince et al. (1999) where the interaction was defined by the following criteria: (1) antagonistic: toxicity of mixture of the NPs is lower than additive toxicity of individual NPs, (2) synergistic: toxicity of mixture of NPs is higher than additive toxicity of individual NPs, and (3) additive: toxicity of mixture of NPs is equal to additive toxicity of individual NPs.

From X-ray diffraction (XRD) analysis, the mineral form of TiO₂ NP was found to be in anatase in nature and crystalline structures of Ag₂O and TiO₂ NPs were found to be cubic tetragonal in nature. From TEM images, the shape of Ag₂O, TiO₂ NPs was found to be spherical and rod in nature respectively (Figure S2, available with the online version of this paper). Brunauer–Emmett–Teller (BET) surface area of dry Ag₂O, TiO₂ NPs were found to be 57.5 m²/g and 74.23 m²/g, respectively. The particle size of dry Ag₂O and TiO₂ NPs determined by TEM were 19.23 ± 7.33 and 22.07 ± 5.06 nm, respectively. The z-average size of NPs going inside the reactor was observed to be 260.9 ± 81.06 nm, 689.17 ± 204.11 and 1,146.87 ± 322.94 nm for Ag₂O, TiO₂ and a mixture of NPs, respectively (average ± one standard deviation). These were found to be significantly different in ultrapure water (p < 0.05) (Figure S3, available with the online version of this paper). The zeta potential of Ag₂O and TiO₂ NPs in ultrapure water were determined to be 7.7 and 5.46 mV which were found to be comparable with available literature data (Romanello & Fidalgo de Cortalezzi 2013).

Figure 1(b) and 1(c) show the variation of dissolved Ag ions in supernatant and sludge after exposure of Ag₂O NPs at 1 and 10 mg/L. Before exposure, Ag ions (in dissolved form) were not detected (below the detection limit of 0.1 mg/L) in both the supernatant and sludge. For both the organic concentrations, Ag in supernatant was found to gradually increase and the maximum dissolved Ag ion content for 500 mg/L of organic concentration was estimated to be 0.26 ± 0.2 mg of Ag/L and 2.7 ± 0.24 mg of Ag/L after the 165th day of exposure for 1 and 10 mg/L NP exposure. It was found to be 0.2 ± 0.12 mg of Ag/L and 2.7 ± 0.24 mg of Ag/L after 180 days for 250 mg/L of organic concentration. When exposed with 10 mg/L of Ag₂O NPs (500 mg/L), the maximum dissolved Ag ions was found to be (3.7 ± 0.51 mg of Ag/g of SS) after 90 days of exposure, while it was found to be (3.34 ± 0.18 mg of Ag/g of SS) after 105 days of exposure for 250 mg/L fed reactors.

Findings of this experiment were compared with that in available literature. As Ag₂O NPs was not studied earlier, results relating to exposure of Ag NPs to aerobic sludge were used for comparison. Our findings at low concentration (1 mg/L) were found to be comparable with the existing literatures. A study by Liang et al. (2010) observed no effect of 0.75 mg/L Ag NPs on COD effluent during a 20-day study (p > 0.05). However, Hou et al. (2012) did not report any effect on COD removal efficiency during a 15-day exposure of mg/L of Ag NPs. However, when exposed to 20 mg/L of Ag NPs for 12 h period peak Ag⁺ concentration of 0.75 mg/L was observed in the mixed liquor. Similarly, Zhang et al. (2014) observed no effect of 0.1 mg/L of Ag NPs on COD removal during a 65-day study. In another study, Yuan et al. (2015) found that 0.1 and 1 mg/L of Ag NPs had no impact on COD (removal: 93 ± 5% and 91 ± 5%) while 5 mg/L Ag NPs was found to have significant impact on COD removal (81 ± 4%) in a 285-day study. From the literature, it is clear that the effect of Ag NPs on COD removal in wastewater treatment depends mainly on NP dosage, duration and the process. It can be understood from the studies (Liang et al. 2010; Hou et al. 2012; Zhang et al. 2014) that short-term exposure of NP to sludge irrespective of NP concentration had no impact on the COD removal, whereas long-term exposure at higher concentration found to have significant effect on COD removal (Yuan et al. 2015). Our study showed about 85.4% removal efficiency of COD during exposure of aerobic sludge to 10 mg/L Ag₂O NPs for 180 days. Overall it appears that Ag₂O NPs up to 1 mg/L concentration do not affect COD removal of aerobic sludge. The differences in removal efficiency with the previously available literature would be due to reactor conditions (duration of exposure, MLSS concentration). In this study, the average MLSS concentration was between 1.6–2.2 g/L and studies in the past (Liang et al. 2010; Hou et al. 2012; Yuan et al. 2015) have used between 2.0–3.0 g/L, while (Zhang et al. 2014) maintained at 5.7 g/L. Thus, it can be understood that, although at different NP concentration and MLSS concentration, Ag NPs had no effect on COD removal efficiency. One of the reasons could be due to their transformations when present in wastewater.

Initially, Kiser et al. (2009) observed around 88% of exposed TiO₂ NPs to be associated with biomass. Further, Wang et al. (2012) reported no effect of TiO₂ NPs (between 0.5 and 2.5 mg/L concentration) on COD removal in a sequential batch reactor during a 30-day study and estimated that around 95% of NPs were removed due to adsorption on to sludge. Additionally, it was determined that 8 mg of Ti/g SS was found in the sludge when exposed to a reactor with 2.5 mg/L TiO₂ NP concentration where the MLSS concentration was found to be 2 g/L. While Ma et al. (2015) observed negligible effects of TiO₂ NPs (concentration range: 0.1 to 20 mg/L) on COD removal effectiveness of laboratory-scale SBRs. Recently, Zhang et al. (2016) found no significant effect on removal performance of SBR reactors after exposure to TiO₂ NPs (10 mg/L) (p > 0.05). The average COD removal ratios of three SBRs were above 89%, thus indicating that the TiO₂ NPs exerted no statistically significant impacts on COD removal during the 21-day operation period of the investigated SBR system. In the present study, a maximum concentration of 14.2 ± 1.2 was found after 180 days of exposure whereas, in the previous study, it was studied only for an exposure period of 27 days. Further, majority of the studies indicated greater retention of TiO₂ NPs (>90%), onto the sludge which is in good agreement with the results of the present study. It can be concluded that sorption to activated sludge is considered as major removal mechanism for TiO₂ NPs.

In order to find the most significant factor influencing the COD removal, all the removal efficiency values for the respective NP type and concentration was used in ANOVA (Table 1). It was found that, for all the three NPs studied, NP type along with COD concentration had no effect on the COD removal (p > 0.05). While COD concentration along with NP type and NP concentration was found to significantly affect the performance of COD removal (p < 0.05). Thus, it can be inferred that all the three parameters chosen in the present study play a significant role in imparting COD removal. When studied for their statistical significance and interactive effects, for both the NP concentrations antagonistic effect was found or, in other words, mixture toxic effect of NPs was found to be smaller than the sum of individual NPs (Table 2). One hypothesis could be that NPs might have interacted with each other along their interaction with activated sludge and suppressed toxic effect of each other NPs on activated sludge. Another hypothesis is that NPs release ions (Xiu et al. 2012) which are toxic to bacteria; it is probable that these ions might have come in contact with ligands (Levard et al. 2012) or been adsorbed, thus reducing the toxicity when present in the mixture. As the aim was limited to study the effect and fate of these NPs in wastewater, much has not been focused on understanding factors influencing toxicity to activated sludge. Future studies are required in order to understand this aspect.

Important findings of this study are presented below.

1. Exposure of TiO₂ was observed to give higher effect on COD reduction than that Ag₂O NPs, irrespective of NP concentration value and input COD concentration.

2. Among all three NPs, mixture of NPs had maximum reduction in COD for both organic concentrations: 1 mg/L had 76.3% after 150 days of exposure (500 mg/L input COD) and 70.2% after 180 days for 10 mg/L exposure (250 mg/L input COD).

3. Mixture of two NPs was found to give higher effect on COD reduction than individual NPs. Following order of NP toxicity was observed (high to low): mixture > TiO₂ > Ag₂O. In contrast, antagonistic effect was observed where combination of individual NPs was found to have greater toxic effect than mixture of NPs.

4. Statistical analysis revealed all the three factors (NP type, NP concentration, organic concentration) play a significant role in imparting COD removal.

5. Irrespective of NPs studied, from the 180th day of exposure of NPs, it was observed that metal ion concentration in sludge was found to be higher than in supernatant.

6. Among individual NPs tested, TiO₂ was found to have larger affinity towards sludge than Ag₂O NPs. When tested among three NPs, dissolved Ti ions concentration was found to be greater than dissolved Ag ion concentration in both sludge and supernatant.

To the best of our knowledge, this study presents information on toxicity of mixture of metal oxide NPs to activated sludge for the first time as per the authors’ knowledge. In addition, it presents information on long-term effects of NPs on biological performance of activated sludge. With the increased dependence of nanotechnology-related goods, NP occurrence in wastewater is inevitable, which makes it extremely important to understand effect on biological performance of activated sludge. This makes findings of the present study important, which can aid WWTP operators in understanding the need, if any, of adjusting operational parameters of biological reactors to avoid any effect of NPs on bacteria. This also indicates the need for monitoring the presence of NPs in wastewater which will help the regulators in arriving at the modification of the operations required in biological processes. Future work must focus on studying the effect of mixture of other NPs on more realistic exposure conditions and identifying/monitoring them in different water types. Further, identifying the nano-sized particles in real complex environmental samples (activated sludge) is also a biggest challenge ahead. Water quality parameters play a major role in deciding the size of NPs when present in water.

This research work was partly supported by Department of Science and Technology (DST), DST Grant Number: DST/TM/WTI/2K11/301; No. SR/FTP/ETA-84/2011) and by Indian Institute of Technology Delhi (IIT Delhi), India. The authors also thank Advanced Instrumentation Research Facility Lab at Jawaharlal Nehru University, New Delhi (India) for their support in characterizing NPs.