Occurrence, removal, and environmental impacts of emerging contaminants detected in water and wastewater in Southern Ontario—Part I: occurrence and removal

A study on the Detroit River watershed (shared between Ontario, Canada and Michigan, USA) indicated the presence of several pharmaceutically active compounds (PhACs) and endocrine disrupting compounds (EDCs) in the effluents from wastewater treatment plants (WWTPs) discharging to the river (Tabe et al. 2009). This encouraged the Ontario Ministry of the Environment and Climate Change (MOECC) to design a comprehensive project to improve understanding of the occurrence, removal, and potential environmental impacts of 220 emerging and legacy contaminants. The study was developed in three stages—Baseline, Phase One, and Phase Two. The Baseline study (5 months) concerned the prevalence of the target substances, the removal efficiency of existing treatment processes, and the toxicity of the WWTP effluents to aquatic organisms. Its outcome helped focus Phase One (13 months) on seasonal variations in occurrence, existing process removal efficiencies, and target substance impacts on aquatic organisms. Phase Two was dedicated to evaluating pilot-scale ozonation in removing or transforming the target substances, and understanding the mechanisms and kinetics of formation of oxidation by-products when selected substances were exposed to ozonation. The results from Phase Two have already been published (Singh et al. 2015; Uslu et al. 2015).

This paper is the first of a pair reporting the Baseline results. It relates to 47 PhACs/EDCs, 17 polybrominated diphenyl ethers (PBDEs), and four alkylphenols and alkylphenol ethoxilates. In this paper (Part I), findings related to the occurrence of these substances at two WWTPs and one water treatment plant (WTP) in Windsor, Ontario, Canada are presented, and the removal efficiencies of the WTP and WWTP processes are discussed. The environmental impacts are presented in Part II.

The presence of trace organic contaminants (TrOCs) in water sources, and their potential negative impacts on the environment and biota, have been the subject of numerous scientific, social, and political debates (Ternes 1998; Heberer 2002; Heberer et al. 2002; Kolpin et al. 2002; Metcalfe et al. 2003; Sedlak et al. 2005; Snyder et al. 2007; Tabe et al. 2009). The development of novel analytical methods and equipment has enabled detection of extremely low TrOC concentrations—e.g., PhACs and EDCs (Yang et al. 2008; Vanderford et al. 2012). WWTP effluents have been recognized as a major point discharge source into surface waters. Their negative impacts on the environment and aquatic life have been reported widely (Jobling et al. 1998; Westergaard et al. 2001; Young et al. 2002; Andreozzi et al. 2003; Richardson 2003; Laville et al. 2004; Sanderson et al. 2004; Jobling et al. 2009; Tetreault et al. 2011; Galus et al. 2013). Some investigations suggest a link between increased loads of these contaminants and various environmental and public health risks.

PhACs are used to prevent or treat disease symptoms, in both livestock and humans. This study focuses on PhACs administered to humans. EDCs are thought to interfere with endocrine system functioning. This involves various mechanisms such as mimicking natural hormones, blocking hormone receiving receptors, and disrupting natural hormone synthesis, metabolism, and excretion (Weyer & Riley 2001). Three major EDC categories comprise estrogenic (which mimic natural estrogen), androgenic (mimicking or blocking testosterone), and thyroidal (affecting thyroid glands directly or indirectly) substances (Snyder et al. 2002). EDCs can be natural or synthetic—e.g., oral contraceptives (Weyer & Riley 2001).

The concentrations of PhACs and EDCs in water sources are much lower than their therapeutic dosages. However, neither their potential short- and long- term impacts, nor their combined (mixture) effects on the environment and public health are well understood (Boxall et al. 2003; Jones et al. 2004; Stackelberg et al. 2004).

Although WTPs and WWTPs are not designed to remove TrOCs, existing processes remove substantial portions, along with other contaminants. However, limited data are available to determine the removal efficiencies of these plants, which jeopardizes optimization efforts.

PBDEs are flame retardants used to reduce fire hazard in plastics, furniture, and electronic items. In the vapor phase, PBDEs delay combustion and inhibit the spread of fire. Depending on the number and location of bromine atoms in their structure, there are 209 possible PBDE compounds. They are of environmental concern due to their health effects including negative impacts on human fertility (Harley et al. 2010).

Several PBDEs have been designated toxic, in Canada, and their manufacture, import, and use were/are being prohibited. (CEPA 2013).

Long-chain alkylphenols and alkylphenol ethoxylates are potential endocrine disruptors and xenoestrogens (Kochukov et al. 2009). Of specific concern are octylphenol (OP), nonylphenol (NP), and nonylphenol ethoxylates (NPEs), which are produced in large volumes. Although their use in detergents is banned they still have applications that lead to widespread release into aquatic environment. According to the US Environmental Protection Agency (USEPA), NP and NPE show low toxicity to humans, but are highly toxic to fish, and aquatic invertebrates and plants (USEPA 2010).

The target substances’ occurrence was studied by monthly 24-hour composite sampling from eight locations in the two WWTPs (WWTP-1 and WWTP-2 hereafter) and one WTP, all in Windsor, Ontario, Canada. Because the WTP is downstream of the discharge from WWTP-1, the sampling locations and times were selected to represent a complete travel path for the target substances from WWTP-1's influent to the finished drinking water leaving the WTP.

The samples for PhACs/EDCs were collected monthly for 4 months from all sampling locations except S1-2, where it was collected weekly for 23 weeks to support the toxicity and biological studies. The PBDE samples were collected monthly for 3 months from all three WWTP-1 locations. The alkylphenol and alkylphenol ethoxylate samples were collected S1-2 monthly for 4 months.

The residence time in each process was taken into account so that the same batch of water was sampled along the sampling path. Samples were poured into pre-cleaned, 1L sampling bottles and shipped immediately to the laboratories in ice-packed coolers. The laboratory of MOECC (LaSB) was responsible for all chemical analyses apart from OP, NP and NPEs, which were analyzed at AXYS Laboratories, Surrey, BC.

Table 1 is a list of the target substances with their limits of detection (LoDs). Several isotope-labelled substances were used as internal standards, and are italicized and marked with an asterisk (*).

The criteria for choosing the target PhACs/EDCs were their abundance in Detroit River water, as obtained through MOECC monitoring projects and reports in the literature (Metcalfe et al. 2003; Hua et al. 2006; Lishman et al. 2006; Tabe et al. 2009; Tabe et al. 2010; Kleywegt et al. 2011), as well as the availability of analytical methods for their detection at ng/L (PPT) levels.

PhACs and EDCs were analyzed using state-of-the art liquid chromatography-tandem mass spectrometry as described in Hao et al. (2008). The maximum acceptable range of recovery was set at 100 ± 20%. All compounds were analyzed using two kinds of standards: standard chemical solutions (of known chemical composition) and isotope labelled standards (readily distinguished by mass spectrometry) to verify specific compound concentrations.

All standard solutions and heptafluorobutyric acid (99%) were obtained from Sigma Aldrich (Oakville, Ontario, Canada). Deuterium-labelled surrogates came from CDN-Isotope (Pointe-Claire, Quebec, Canada), except sulfamethazine (phenyl-¹³C6, 98%) from Cambridge Isotope Laboratories (Andover, MA). Methanol, acetonitrile and sulfuric acid were supplied by Caledon Inc. (Georgetown, Ontario, Canada), ammonium acetate (>99%) by Fluka (Mississauga, Ontario, Canada) and ethylene diamine acetic acid (disodium salt, ACS reagent grade) by BioRad (Mississauga, Ontario, Canada).

C_in is the target pollutant concentration entering the process/plant, (ng/L, μg/L, or mg/L); and,

C_out is the target pollutant concentration in the process/plant effluent, (ng/L, μg/L, or mg/L).

For compounds with undetectable concentrations, a statistical approach using censored data analysis was used, as recommended by Helsel (2012).

A summary of results is shown in Table 2. The 188 analytical results shown for S1-0 and S1-1 refer to the 47 substances sampled monthly for 4 months. The 1,081 results shown for S1-2 relate to the 23 weekly sets of samples taken for toxicology tests and also used for chemical analysis. Removal efficiencies were only calculated for paired samples (those relating to the same batch of water). One set of analytical results, including samples taken from W-1 was lost, leaving just 141 analyses for that location. The median and mean overall concentrations shown in Table 2 refer to the total concentrations of all 47 PhACs and EDCs over the sampling period.

Approximately 70% of the analytes in the WWTP influents were detected at least once over the 4-month sampling period. Some substances detected in the influents were removed by treatment to below their LoDs and were thus reported as undetected in the effluents.

After discharge to the surface water, further reductions in target substance concentrations were noted, arising through natural processes like dilution, biodegradation, photo-degradation, adsorption and sedimentation. As a result, fewer substances were detected at the WTP intake than in WWTP-1's effluent discharge a few kilometers downstream. The detection frequency of substances at the WTP intake was also lower than in WWTP-1's effluent and, for the concentrations of those detected were up three orders of magnitude lower. Consequently, target substance concentrations at W-1 and W-2 were very low and, in many cases, close to their LoDs. This caused inaccuracies in calculating the WTP's removal efficiency, as reported in Table 2.

The median and mean monthly concentrations of the target PhACs/EDCs at both S1-0 and S2-0 were around 200,000 ng/L (0.2 mg/L). The secondary treatments of both WWTPs successfully removed between 95 and 98% of the total concentrations.

Table 3 categorizes the target PPCPs/EDCs by removal efficiency in the WWTPs and WTP. The categories include:

1. Undetected: substances were not detected in any sample.

2. Negative or Zero: substances whose concentrations were higher or similar in the effluent than in the influent.

3. Low: removal efficiencies between 0 and 50%.

4. Moderate: removal efficiencies between 50 and 75%.

5. High: removal efficiencies between 75 and 95%.

6. Excellent: removal efficiencies exceeding 95%.

7. Inconclusive: removal efficiencies with standard errors spread over three or more categories.

Accordingly, the treatment efficiency of WWTP-1 is higher than that of WWTP-2, as the former removed more substances at moderate to excellent efficiency. The two plants, however, performed similarly towards certain substances. For example, both removed acetaminophen, ibuprofen, estriol, and bisphenol A at high/excellent efficiency, clofibric acid, gemfibrozil, and lasalocid A at low efficiency, and had no impact on carbamazepine or diclofenac. The WTP was only efficient in removing ibuprofen and carbamazepine, and to some extent acetaminophen.

The occurrence and removal information of the ten commonest PhACs/EDCs at different sampling locations is shown in Table 4. The highest concentrations in the raw sewage belonged to the three analgesics; acetaminophen, ibuprofen, and naproxen. These, together, formed 96% and 89% of the total concentration of PhACs/EDCs in the influents of WWTP-1 and WWTP-2, respectively. Acetaminophen alone contributed 87% and 72% of the total concentration of PhACs/EDCs in the WWTP-1 and WWTP-2 influents, respectively. Bisphenol A (BPA) was also a major substance in the influent of WWTP-2, contributing 9% to the PhACs/EDCs concentrations. This may demonstrate the impact of the industrial wastewater load entering WWTP-2.

The high general efficiencies of the WWTPs arose from the 99 + % removal of acetaminophen. WWTP-1 also removed 99 + % of ibuprofen and naproxen. The removal efficiencies of the latter two substances at WWTP-2 were 88% and 58%, respectively. This could be due to the difference in the treatment trains of the two plants. The removal efficiencies of the WWTPs for substances other than the three analgesics and BPA were low at 42 ± 7% and 15 ± 7%, respectively. Table 4 also shows the frequency and concentrations of the target substances in transit through the WTP. Only 17-α-estradiol was detected in all 4 sampling events at the WTP intake.

Of the ten substances, four persisted through both wastewater and water treatment processes, with their average concentrations remaining statistically the same from the influent to the effluent of the WWTP-1 and then from the intake to the finished water in WTP. These four included enrofloxacin, chlortetracycline, 17α-estradiol, and meclocycline. The latter had concentrations in the vicinity of its LoD. The performance of the plants in relation to these substances underlined the relative inefficiency of the treatment processes in removing or transforming certain substances at very low concentrations.

Table 5 also shows the Canadian Federal Environmental Quality Guideline (FEQGs) levels for the two WWTP effluents (CEPA 2013). FEQGs have been developed in Canada for certain PBDE congeners in water, fish tissue, sediments, wildlife (including birds’ eggs), to assess the ecological impacts of PBDEs in the environment. Use of FEQGs is voluntary unless prescribed in permits or other regulatory tools.

PBDE-209 (deca-brominated diphenyl ether) alone contributed ∼50% to the total concentration of PBDEs in the influents of both plants, and, together with PBDE-99 and PBDE-47, comprised ∼90% of all PBDEs. The primary treatment at WWTP-1 removed almost half of all influent PBDEs, while the secondary treatment effectively reduced the concentration to ∼7% of the influent. The relatively high removal efficiency of the primary treatment could arise from the hydrophobic nature of these substances and their high-log K_OW, which increased their tendency to adsorb onto and be removed with suspended solids. PBDE-209 showed the highest removal efficiency from the primary effluent at 61%, with others in the vicinity of 40%.

The final concentrations of the target congeners in the plant effluents were in the range of sub- and a few ng/L. PBDE-100 exceeded the FEQGs values in both plants and PBDE-99 at WWTP-1. However, dilution in the Detroit River is expected to bring both concentrations below the guidelines.

Table 6 groups the target congeners according to the WWTPs’ removal efficiencies. They removed approximately three quarters of the congeners at high or excellent efficiencies.

The group included OP, 4-nonylphenol, 4-nonylphenol mono-ethoxylates, and 4-nonylphenol diethoxylates. The two ethoxylates include different isomers. Also, ¹³C₆-4-n-nonylphenol was used as an isotope-labelled surrogate.

Three sets of monthly samples, collected at S1-2, were analyzed for these substances, and the results are reported in Table 7. The most prevalent substance was 4-nonylphenol, contributing approximately 2/3 of the total content of NPEs. Because samples were not collected at S1-0, the plant's removal efficiency for NPEs is unknown.

TrOC occurrences were studied over 4 months at two WWTPs and one WTP in Windsor, Ontario, Canada. The concentration reduction efficiencies of the existing treatment processes in reducing TrOC concentrations were estimated.

Three analgesics – acetaminophen, ibuprofen, and naproxen – were predominant among the 47 PhACs/EDCs studied. Bisphenol A was also detected at quite high concentrations in WWTP-2, which treats mixed industrial and residential wastewaters. Large proportions of these four chemical species were removed in treatment. The removal efficiencies for most other PhACs/EDCs were low.

Thus, although the WWTPs are not designed to remove PhACs/EDCs, a number of them are removed together with other contaminants.

The concentrations of PhACs/EDCs at the WTP intake were low and, sometimes, close to their LoDs. This made calculation of plant removal efficiencies inaccurate.

Three pharmaceuticals – enrofloxacin, chlorotetracycline, and 17-α estradiol – persisted through both the water and wastewater treatment processes. That is, their concentrations at the influent of the WWTP-1 were statistically similar to that at the effluent, and the concentrations at the intake of the WTP were statistically similar to those in the finished drinking water.