Introduction

The release of pharmaceuticals and personal care products (PPCPs) has attracted much public attention (Hollender et al. 2009; Camacho-Muñoz et al. 2010; Sui et al. 2010; Salgado et al. 2012) because of their intrinsic biological activity which may cause adverse effects to aquatic and terrestrial ecosystems (Sorensen et al. 1998; Öllers et al. 2001). These compounds might be excreted by excreta and disposal of unused or expired drugs and eventually reach the wastewater treatment plants (WWTPs) (Escher et al. 2011; Nelson et al. 2011). Ineffective operations of WWTPs make their widespread appearance in the aquatic environment possible (Zhang et al. 2008; Daneshvar et al. 2012), leading to the contamination of surface water, ground water, and eventually, drinking water, where they pose a negative impact to humans (Cleuvers et al. 2004; Schnell et al. 2009). Therefore, knowledge about the occurrence, fate, and transport of PPCPs in WWTPs is of great significance.

The concentrations of PPCPs in WWTP wastewaters were routinely monitored in many countries (Castiglioni et al. 2006; Jones et al. 2007; Santos et al. 2007; Vieno et al. 2007; Jelic et al. 2011; Martín et al. 2012; Salgado et al. 2012; Duan et al. 2013). However, little attention was paid to the daily variation of PPCPs, which was crucial to evaluate their behavior in the WWTPs. Nelson et al. (2011) investigated the diurnal variability of 30 PPCPs in tertiary wastewater effluent, and indicated that trimethoprim, sulfamethoxazole, naproxen, estrone, and triclosan varied greatly during a daily cycle, with relative standard deviations (RSDs) exceeding 100 %. However, some compounds including carbamazepine, primidone, fluoxetine, and triclocarban exhibited little or no variability. Göbel et al. (2005) and Plósz et al. (2010) assessed diurnal variations in the influents using three 8-h flow-proportional composite samples obtained in one sampling day, and suggested that daily antimicrobials and antibiotics loads can correlate with the respective water flows and ammonium loads.

The anaerobic/anoxic/aerobic (A2/O) process is extensively employed in WWTPs throughout the world (Jelic et al. 2011), mostly because it can produce effluents that meet the required quality standards (suitable for disposal or recycling purposes), at reasonable operating and maintenance cost. However, concerning the removal of PPCPs, quite variable removal efficiencies were observed in the previous studies due to chemical and biological properties of individual PPCP as well as wastewater characteristics and operational conditions. For instance, Castiglioni et al. (2006) observed low or no removal for salbutamol, furosemide, and bezafibrate, whereas Kasprzyk-Hordern et al. (2009) noted much higher removal of these compounds (>70 %). Seasonal variation of the removal performance by A2/O was also observed. Castiglioni et al. (2006) reported a removal of 10 % for atenolol during the winter months and much better elimination in summer (55 %) due to different microbial activities. However, these studies simply focused on the PPCPs in raw influent and final effluent, as well as their overall removal efficiencies. In order to comprehensively evaluate the fate, distribution, and mass balance of PPCPs, Nie et al. (2012) and Duan et al. (2013) investigated the occurrence and release of five acidic pharmaceuticals and seven endocrine-disrupting chemicals in different units of municipal WWTPs employing A2/O process, respectively.

Ultraviolet (UV) radiation and UV combined with hydrogen peroxide have been extensively studied as a tertiary treatment for the PPCPs removal from wastewaters. Most of the studies were carried out in lab-scale reactors rather than in full-scale treatment systems. Pereira et al. (2007a, 2007b) monitored the degradation of the selected pharmaceutical active compounds using a UV system in a batch reactor, and proved that the UV system was efficient in degrading some of the targeted compounds (such as ketoprofen and ciprofloxacin) at 100 mJ/cm2. However, when the UV photolysis was implemented in full-scale WWTPs, more factors, including fluence rate distribution, hydraulics, reactor design, and electrical energy, might influence its performances. Recently, an investigation on the removal of 79 PPCPs by UV radiation used for disinfection in a full-scale WWTP in Portugal has been reported and revealed that UV radiation played an important role in reducing the concentrations of some target compounds (Salgado et al. 2012).

Thus, in this study, we conducted a detailed investigation on six PPCPs belonging to different groups (anticonvulsant, stimulant, antibiotics, insect repellent, antipsychotic, and anti-hypertensive) in a full-scale WWTP employing A2/O-UV process in Shanghai, China, to gain an insight into their occurrence and removal along the different units taking into account the raw influent, primary treatment, secondary treatment (anaerobic, anoxic, aerobic tank, respectively), and tertiary treatment process. Special emphasis was laid on the daily fluctuations of PPCPs in the WWTP influent, which was seldom reported in China, to the best of the authors’ knowledge.

Materials and methods

Chemicals and standards

PPCP standards (Table 1) as well as the internal standards (IS), DEET-7D and Phenacetin-13C, were purchased from Dr. Ehrenstorfen (Augsburg, Germany). High-performance liquid chromatography (HPLC)-grade formic acid and methanol were purchased from Fisher Chemical Company (Beijing, China). Sodium hydroxide (NaOH), ammonium acetate and hydrochloric acid (HCl) were of analytical grade and obtained from Merck (Darmstadt, Germany). Ultrapure water was obtained with Nanopure ultrapure water system from Barnstead/Thermolyne (Dubuque, IA). Stock solutions of individual compound were prepared in methanol. Working solutions were prepared by diluting the stock standard solutions and stored at 4 °C in the dark.

Table 1 Therapeutic classes, chemical structures, physicochemical properties and corresponding IS for selected PPCPs
Full size table

Sample collection

The investigated WWTP is located by the Huangpu River, the most important river across Shanghai City. The WWTP, serving 250,000 population equivalents, was designed for a capacity of 50,000 m3/day and fed with mainly domestic wastewater. As shown in Fig. 1, the WWTP employed a physicochemical primary treatment process (screen and primary clarifier) coupled with biological treatment process (anaerobic tank, anoxic tank, oxic tank, and secondary clarifier). After settling in the secondary clarifier, part of the activated sludge was returned to the anaerobic tank (return sludge ratio: 100 %), and the rest of the sludge was conveyed to be dehydrated. The sludge retention time (SRT) was 20 days and hydraulic retention time (HRT) was 13–15 h (the HRTs of anaerobic tank, anoxic tank, and oxic tank were 1.0, 4.0, and 8.3 h, respectively). Finally, effluent of secondary clarifier went through UV disinfection unit (UV3000Plus, Trojan). The UV disinfection unit had two channels. Each was equipped with 120 UV lamps (low pressure, wavelength of 254 nm, 30 W), operating at the UV dosage of 95 mW/cm2, and the total suspended solids was approximately 20 mg/L. The operation parameters of the WWTP during the sampling are presented in Table 2.

Fig. 1
figure 1

Schematic diagram of the treatment processes in the WWTP and sampling sites location (circles and downward arrows)

Full size image
Table 2 Wastewater characterization parameters for influent and effluent of the investigated WWTP during sampling campaign
Full size table

The sampling campaign was conducted in January 2013. Twenty-four-hour composite samples with the sampling interval of 6 h were taken. The sampling sites are displayed in Fig. 1. The wastewater samples were collected in amber glass bottles in duplicate and had been washed with methanol and purified water, immediately placed on ice, transported to the laboratory, and stored at 4 °C until extraction.

Sample preparation, extraction, and analysis

The method for sample preparation and extraction followed a previously developed method (Sui et al. 2009). Water samples were filtered using glass fiber filters (Whatman). Then, filtered water samples (100 mL for influent and primary clarifier and 400 mL for the others) were spiked with 200-μL internal standards and adjusted to pH = 7 with 1 mol/L HCl and NaOH solutions. Following this, solid phase extraction (SPE) was performed using a 12-fold vacuum extraction manifold device (Supelco, USA). The Oasis HLB (Waters) cartridges (200 mg, 6 ml for influent and effluent of primary treatment; 500 mg, 6 ml for others) were conditioned with 5-mL methanol and 3 × 5-mL ultrapure water. Then, samples were loaded onto conditioned SPE cartridges at a flow rate of 3–5 mL/min. After the enrichment step, the cartridges were washed with 5-mL methanol/water (v/v = 1:19) and dried for 30 min under full vacuum. Finally, the analytes were eluted with 5-mL methanol and collected in 10-mL glass tubes. The solvent was evaporated to dryness under gentle N2 flow, and redissolved in 0.4 mL methanol/water (v/v = 1:4).

High-performance liquid chromatography (HPLC, Ultimate3000, Dionex, USA) followed by electrospray ionization and tandem mass spectrometry (ESI-MS/MS, API3200, AB Sciex, USA) was applied to quantify the target PPCPs. The PPCPs were analyzed in positive electrospray ionization (ESI+) and multiple reaction monitoring mode. The analytical method was slightly modified from Chen et al. (2012). Separation was performed on a ZORBAX Eclipse XDB C18 column (3.5 μm, 2.1 × 150 mm), using ultrapure water (solvent A) with 0.1 % formic acid and methanol (solvent B) as the mobile phase. The elution gradient program was as follows: from 15 to 50 % solvent B in 2 min, then to 100 % in 6 min, hold for 8 min, and back to the initial conditions in 14 min. The flow rate was 0.30 mL/min, and the injection volume was set at 10 μL. Quantification was performed following the internal standard calibration approach.

Biochemical oxygen demand (BOD5), chemical oxygen demand (COD), NH3-N, total nitrogen (TN), and total phosphorus (TP) of collected samples were measured using standard methods (APHA et al. 1998).

Method validation

During sampling, field blanks (500-mL ultrapure water in amber glass bottles) were taken to site, exposed to the environment at the sampling sites. For each set of samples analyzed, a procedure blank and a recovery test sample were analyzed identically to wastewater samples. All the procedure and field blanks were below the limit of quantification (LOQ). All the samples were extracted and analyzed in duplicate, and the analytical results were reported as the average of two values.

The relative recovery (RR%) was calculated by spiking a known concentration of the 12 standard target analytes, according to Eq. (1) (Tixier et al., 2003):

$$ RR\left(\%\right)=\frac{C_{ss}-{C}_{os}}{C_s}\times 100 $$
(1)

where C ss and C os are measured concentrations in the spiked and correspondingly unspiked water samples, respectively; C s is the known spiking concentration.

The relative recoveries were proved to be 72–115, 81–128, and 82–122 % in the effluent, influent, and ultrapure water for most compounds, respectively. While for caffeine (CF), the recoveries in the influent samples were not satisfactory due to its high background concentration in the influent. The LOQ, set at S/N ratios ≥10, was 0.2–2.1 ng/L. Detailed information about the LOQ and RR are briefly listed in Table 3.

Table 3 Limit of quantification (LOQ) and relative recovery (RR) for selected PPCPs in the ultrapure water, wastewater effluent, and influent
Full size table

Results and discussion

Concentrations of PPCPs in wastewater

Influent

The PPCP concentrations in the WWTP influent varied from 17 to 11,400 ng/L, as shown in Fig. 2. CF was most abundant in the WWTP influent, which can be attributed to its daily consumption besides pharmaceutical use, such as uptake with beverages and foods (Ogunseitan 1996). Slightly high concentrations of TMP (257 ng/L), sulpiride (SP) (143 ng/L), and metoprolol (MTP) (122 ng/L) were found. The concentration of MTP was consistent with its increased dosage in winter, when the blood pressure is relatively higher than other seasons (Brennan et al. 1982). N, N-diethyl-meta-toluamide (DEET) (66 ng/L) and carbamazepine (CBZ) (17 ng/L) were detected in raw wastewater at low concentrations (Fig. 2). The low concentration of DEET, one of the most efficient insect repellents, was probably because it is normally used less frequently in winter (Fradin et al. 2002).

Fig. 2
figure 2

Concentrations of target PPCPs in the wastewater influents and effluents as well as the removal efficiencies by the treatment process of the WWTP in Shanghai

Full size image

The contamination levels of investigated PPCPs were low in Shanghai, compared to other reported studies. For instance, the concentration of CBZ was at least two times lower than those reported in Spain (70–1,400 ng/L) (Jelic et al. 2011; Martín et al. 2012) and USA (34–350 ng/L) (Yu et al. 2013). TMP, commonly used in combination with sulfonamides as a potentiator (Loftin et al. 2008), was slightly less than those in the influents of WWTPs in Sweden (100–1,300 ng/L) (Lindberg et al. 2005) and Beijing (700 ng/L) (Zhou et al. 2010). The anti-hypertensive MTP in the influent was recorded to be 122 ng/L in Sweden (Bendz et al. 2005), comparable with that in the wastewater influent of Shanghai. The lower contamination level in the WWTP of Shanghai might be ascribed to lower per capita consumption in this area, as defined by Diaz-Cruz et al. (2009) that the concentrations of PPCPs detected in the wastewater reflect their use in the society. Besides, the different types of products in pharmaceutical market in different countries and regions (Miege et al. 2009) can also be an explanation for the lower concentrations of investigated PPCPs.

To assess the daily fluctuation and dynamic behavior of selected PPCPs in the wastewater influent, grab samples were collected every 6 h and analyzed. The concentrations of PPCPs in the grab samples as well as the RSDs between grab and 24-h composite samples are presented in Fig. 3.

Fig. 3
figure 3

Daily fluctuation of the PPCP concentrations (circles) and the relative standard deviation (RSD) (crossed circles) between grab and 24-h composite samples (solid line)

Full size image

In general, the six PPCPs had low variability throughout a day, with the RSDs less than 25 % for most samples, which agreed with that reported in Guangzhou City of China in the dry season (Xu et al. 2006). Also, Sui et al. (2011) observed the RSDs of grab and 2-h composite samples did not exceed 30 % in the influent and 15 % in the effluent in a WWTP of Beijing. It seemed that the concentrations of PPCPs showed no sensitive daily change in many regions of China. One of the explanations for this phenomena might be the complex configuration of sewer systems connected with the WWTPs. The sewer systems receive wastewater in different regions at different times, which lead to a reduction of peak concentrations of some PPCPs. Besides, the configuration of sewer also affects the fraction of deconjugated compounds arriving at WWTPs (Suárez et al. 2008). Furthermore, the WWTPs in these regions usually have relatively large treatment capacity, which might be benefit for minimizing the fluctuation.

However, in the case of TMP and CF, the daily fluctuation still reflected their consumption patterns although it was not significant. The distribution of TMP daily load correlated well with its typical oral prescription of once a day (Göbel et al. 2005), normally in the morning. Therefore, high degree of human excretion is expected in the morning, which can explain the decreasing daily concentration. CF showed a gradually increase increasing daily profile in the daytime and the lowest load in the midnight. It was most likely caused by the consumption pattern and life habit.

Effluent

It is noted that the composition profiles of the target PPCPs in the final effluents were quite similar to the influent samples. CF (733 ng/L) was dominant in the effluent, TMP (186 ng/L), SP (168 ng/L) and MTP (126 ng/L) followed. Also, DEET (40 ng/L) and CBZ (18 ng/L) were detected at low levels. Effluents of WWTPs could be the major sources for PPCPs in the surface waters (Escher et al. 2011; Nelson et al. 2011). Fortunately, concentrations of PPCPs in the WWTP effluent were similar to or lower than those reported before (Gomez et al. 2007; Santos et al. 2007; Vieno et al. 2007; Gulkowska et al. 2008; Camacho-Munoz et al. 2010; Jelic et al. 2011; Martín et al. 2012; Yu et al. 2013). For example, CBZ was found in the wastewater effluents of Spain and USA at the concentrations of 50–150 and 21 ng/L, respectively (Martín et al. 2012; Yu et al. 2013), slightly higher than that in the WWTP effluent of Shanghai. TMP were detected in five WWTPs of Hong Kong and Shenzhen, with the concentrations of 120–230 ng/L in the effluents (Gulkowska et al. 2008), consistent with the finding in our study.

Ecotoxicological risk assessment is evaluated by means of risk quotient (RQ) value, usually expressed as the ratio between the measured environmental concentration and the predicted no-effect concentration of individual compound (Santos et al. 2007). As shown in Fig. 4, the RQ values of investigated PPCPs were all below 1, varied between 0.0038 and 0.0659 in the wastewater effluent. Therefore, environmental risk, with the current use of these compounds, was unlikely.

Fig. 4
figure 4

Risk quotient (RQ) values of some PPCPs in the wastewater effluent

Full size image

Removal efficiency

Primary treatment

Adsorption is the main mechanism in the removal of micropollutants during primary treatment (Suárez et al. 2008). From Fig. 5, the primary clarifier showed poor removal efficiencies for most PPCPs. A reduction of less than 20 % was observed for all the six PPCPs in the primary treatment process, owing to their hydrophilic nature. The compounds studied in this study are of low octanol–water partition coefficient (K ow ) values (<3.0, as shown in Tab. 1), and are expected not to adsorb greatly to the particles, but to dissociate in the aqueous phase (Thomas and Foster. 2005; Salgado et al. 2012).

Fig. 5
figure 5

Removal efficiencies of the target PPCPs in the primary, secondary and tertiary treatment processes

Full size image

Secondary treatment

The removal efficiencies of investigated PPCPs by secondary treatment process ranged from −33 to 87 % (Fig. 5). CF has been proven to be readily biodegradable (>98 %) (Huerta-Fontela et al. 2008; Okuda et al. 2008; Sui et al. 2010; Xue et al. 2010), while only 87 % removal was observed in this study. Likewise, CF was eliminated by 79, 86, and 93 % during the winter months (December, January, and February, respectively) in a Beijing WWTP employing A/A/O process(Sui et al. 2011). The removal efficiency for DEET was 43 %, which was similar to or slightly less than values reported in other studies (Knepper 2004; Costanzo et al. 2007; Sui et al. 2010; Zhou et al. 2010). No significant overall removal rates during secondary treatment were observed for TMP and CBZ. The incomplete removal of these compounds during conventional treatment was reported by several studies (Göbel et al. 2005; Kasprzyk-Hordern et al. 2009; Behera et al. 2011; Jelic et al. 2011;). Göbel et al. (2005) reported that TMP was neither biodegraded nor adsorbed. Pérez et al. (2007), Behera et al. (2011), Jelic et al. (2011), and Martín et al. (2012) reported that CBZ could be hardly removed regardless of secondary treatment process applied. Therefore, CBZ is proposed as an anthropogenic marker (Clara et al. 2004; Nakada et al. 2006). Surprisingly, the concentrations of MTP and SP increased from 116 and 133 ng/L before the secondary treatment to 150 and 177 ng/L after the secondary clarifier, respectively. The case could be explained by cleavage of conjugates (glucuronides, sulfates) of target compounds (Carballa et al. 2004; Galán et al. 2012) and/or analytical deviations.

In general, the removal of most investigated PPCPs by the secondary treatment process was low and even negligible. The SRT and HRT were relatively high (20 days and 13–15 h, respectively) in the investigated WWTP, higher than the critical SRT and HRT proposed by Jones et al. (2007) and Vieno et al. (2007b). Thus, it could be excluded as the main influencing factors. However, as the sampling was conducted during winter (January), the low temperature (9.6 °C in the bioreactor), unfavorable for activated sludge, could be the cause of the inefficient removal during the biological treatment process. Clara et al. (2004) and Castiglioni et al. (2006) also found significantly low removal rates for several pharmaceuticals in winter, for example, the removal rates of atenolol, bezafibrate, and enalapril were 10, 15, and 18 % in winter and 55, 87, and 100 % in summer, respectively, in line with a temperature-dependent increase of microbial activity (average temperatures 9.7 °C in winter and 18.6 °C in summer). Through a lab-scale experiment model, Wick et al. (2009) found that there was a crucial temperature below of which the biodegradation of certain beta blockers and psycho-active drugs was completely inhibited, and suggested that lower temperature led to reduced rate constants. A batch experiment on the removal of TMP under different temperatures by activated sludge collected from the WWTP was conducted, and the preliminary results also confirmed that the temperature had significantly negative effect on its removal from wastewater (as shown in Supporting Information).

The removal of two biodegradable PPCPs (CF and DEET) in each tank (anaerobic, anoxic, and oxic tank) during the secondary treatment process was further studied. As parts of the wastewater were recycled from the secondary clarifier to the anaerobic tank and from the oxic to the anoxic tank, it is improper to calculate the removal efficiency by concentration, due to dilution. Thus, we employ a mass load approach to assess the removal of CF and DEET (Eq.2).

$$ \mathrm{Removal}\left(\%\right)=\frac{C_{in}\cdot \left({Q}_{in}+{Q}_{rec}\right)-{C}_{out}\cdot {Q}_{out}}{C_{in}\cdot \left({Q}_{in}+{Q}_{rec}\right)}\times 100 $$
(2)

where C in , C out , Q in , Q out , and Q rec are the measured compound concentration (nanogram per liter) and the flow mass of the wastewater (cubic meter per day) of inlet, outlet, and recirculation in the investigated tank, respectively.

As shown in Fig. 6, the anaerobic tank seemed to be crucial for CF and DEET removal. A 60–70 % reduction was observed for CF and DEET in the anaerobic tank. Oxic tank followed, with the removal efficiencies of 55 % (for CF) and 3 % (for DEET), respectively. On the contrary, negative removal rates were found for the two PPCPs in anoxic tank.

Fig. 6
figure 6

Removal efficiencies of CF and DEET in each tank of A2/O process. The error bars represent the maximum and minimum removal efficiencies by oxic, anoxic, and anaerobic tank

Full size image

The rapid decrease of PPCP concentrations in the anaerobic tank could ascribe to the sorption or degradation by some types of activated sludge (Xue et al. 2010). For instance, the phosphate accumulating organisms might use PPCPs as carbon sources during the phosphorus release process (Xue et al. 2010). In oxic tank, the available condition allows the development of slow-growing bacteria, such as nitrifiers, and the retention of exocellular enzymes or soluble oxidants (Galán et al. 2012), which might be beneficial for the degradation of PPCPs. The negative removal during anoxic tank was consistent with our previous finding (Sui et al. 2010). We conducted a survey of CF, DEET, BF, and TMP along the A2/O-MBR process in a WWTP of Beijing, and found the amounts of most PPCPs increased after the anoxic tank. We supposed it could be a consequence of (1) deconjugation of conjugated metabolites during the treatment process (Miao et al. 2002); and (2) a change in the adsorption behavior of compounds to particles during treatment process (Lindberg et al. 2005).

Tertiary treatment

As shown in Fig. 5, the tertiary UV treatment seemed to be positively effective in removing investigated PPCPs. Moderate reduction for CF, DEET, MTP, CBZ, and TMP by UV radiation was obtained. It was interesting that the refractory substances CBZ and MTP, which were barely removed by primary and secondary treatment processes, achieved a slightly better removal in the tertiary treatment process (CBZ = 11 % and MTP = 16 %). Salgado et al. (2012) also found that UV radiation had an important role in reducing some PPCPs, such as diclofenac, ibuprofen, clorazepate (25–75 %) and indapamide, enalapril, captopril, atenolol (>75 %). Some PPCPs themselves are readily subjected to photolysis (Pereira et al. 2007a, b), and others might be decomposed by UV treatment in the presence of humic acids existing in the wastewater secondary effluent. Humic acids submitted to UV irradiation are promoted to a transient and excited state, in which they may react with oxygen, forming reactive species as singlet oxygen, or react directly with other organic species, thus promoting their phototransformation (Andreozzi et al. 2003).

Conclusion

The occurrence and removal of six PPCPs (CF, DEET, CBZ, MTP, TMP, and SP) in a municipal WWTP that employed the A2/O-UV process in Shanghai, China were studied. The concentrations ranged from 17 to 11,400 ng/L in the influent and from 18 to 733 ng/L in the effluent, respectively, which presented a lower level compared to other countries. The concentrations of PPCPs in the wastewater influent showed a low variability throughout the day, with the RSDs less than 25 % for most samples. However, for TMP and CF, the slight daily fluctuation still reflected their consumption patterns. The biological treatment process showed inefficient removal for most PPCPs, probably due to the low temperature in winter, and for the two biodegradable PPCPs, CF, and DEET, the anaerobic and oxic tank made contribution to their removal while anoxic tank had negative effect to their elimination. As a tertiary treatment, UV photolysis removed all the investigated PPCPs by 5–38 %, representing a polishing step to compensate for the poor removal performance of PPCPs by biologic treatment process in winter.