ARTICLE pubs.acs.org/est
Diurnal Variability of Pharmaceutical, Personal Care Product, Estrogen and Alkylphenol Concentrations in Effluent from a Tertiary Wastewater Treatment Facility Eric D. Nelson,* Huy Do, Roger S. Lewis, and Steve A. Carr San Jose Creek Water Quality Laboratory, Sanitation Districts of Los Angeles County, 1965 Workman Mill Rd., Whittier California 90601, United States
bS Supporting Information ABSTRACT: Hourly samples of tertiary wastewater effluent were analyzed for 30 pharmaceuticals, personal care products, estrogenic steroids, and alkylphenols in order to better understand the rate at which these compounds enter the environment. Several distinct patterns of daily cycling were observed, and were characterized as three separate categories. The concentrations of compounds such as trimethoprim, sulfamethoxazole, naproxen, estrone, and triclosan varied greatly during a daily cycle, with relative standard deviations exceeding 100% of their daily mean. Less extreme daily cycles were seen for other compounds such as azithromycin, atenolol, tert-octylphenol, iopromide and gemfibrozil. Peak concentrations for most compounds occurred in the early evening (5-8 pm). However, some compounds including carbamazepine, primidone, fluoxetine, and triclocarban exhibited little or no variability.
’ INTRODUCTION The 1999-2000 United States Geological Survey reconnaissance of rivers and streams throughout the United States highlighted the prevalence of endocrine disrupting compounds (EDCs) and pharmaceuticals and personal care products (PPCPs) in the environment.1 Increasingly, researchers are focusing attention on such organic chemical contaminants that enter surface waters via their discharge from wastewater treatment facilities or agricultural runoff. Concern over three particular groups of compounds (human and artificial hormones, PPCPs, and metabolites of surfactants) stems both from their ability to impact fish reproduction at trace concentrations2 and the increasing reliance on using treated wastewater for groundwater recharge or indirect production of potable water.3,4 Several researchers have shown that wastewater effluent has the potential to disrupt the endocrine function of aquatic life in receiving waters.5-8 Natural and artificial human hormones may be a factor in faunal endocrine problems,9 but some metabolites of industrial/household chemicals may also be responsible, for example, 4-nonyl and 4-octylphenols.8,10 Additionally, the presence of PPCPs has been confirmed in many surface waters,1,11-14 though it is still not clear what impact these concentrations have on aquatic fauna. The increasing high usage of both over-the-counter and prescription drugs along with reports of bacterial antibiotic resistance15 amplify the need for evaluating PPCP's environmental loads coming from treatment systems. These concerns over EDCs and PPCPs in effluent are most pressing in the arid southwestern United States where surface waters may be dominated by a wastewater component and where large metropolitan areas have an increasing demand for potable water that is in short supply.16 This demand is partially satisfied by the use of highly treated effluents for groundwater recharge. Recognizing the links between wastewater and drinking water r 2010 American Chemical Society
has led to more scrutiny of sources of these contaminants of emerging concern.17,18 PPCPs coming from effluent-impacted surface water, as well as from leaking septic systems, are increasingly found in groundwater,19-24 with carbamazepine, primidone, sulfamethoxazole, meprobamate, phenytoin, tris(2-chloroethyl) phosphate (TCEP), and several iodinated contrast media being detected most frequently. To better understand and control chemicals of emerging concern such as PPCPs and EDCs in wastewater, groundwater and the environment, it is critical to know more about their presence and variability in treated wastewater. The current understanding of any potential diurnal fluctuations of emerging contaminants in wastewater effluent is limited. To date there has been no study of temporal behavior of these compounds in daily plant effluent with a time resolution that is higher than 8 h.25,26 Martinovic et al. noted the changing estrogenic equivalence of wastewater effluent in 2-h composite samples taken between 10 a.m. and 8 p.m. (5 samples/day), though they did not analyze for hormones or other chemicals.27 The purpose of the work presented here is to increase our understanding of PPCPs and EDCs in wastewater effluent, and thus the impact on receiving waters, through the use of high frequency sampling coupled with analysis for a wide range of contaminants of emerging concern. The suite of compounds selected covers a range of chemical/usage classes including antibiotics, β-blockers, nonsteroidal anti-inflammatory drugs (NSAID), anticonvulsants, antidepressants, diuretics, lipid lowering drugs, flame retardants, iodinated X-ray contrast agents, Received: July 21, 2010 Accepted: December 10, 2010 Revised: November 10, 2010 Published: December 28, 2010 1228
dx.doi.org/10.1021/es102452f | Environ. Sci. Technol. 2011, 45, 1228–1234
Environmental Science & Technology
ARTICLE
Figure 1. Simplified schematic of treatment plant train, and location of sample collection.
antimicrobial compounds, human and synthetic hormones, industrial chemicals/metabolites, and caffeine.
’ MATERIALS AND METHODS Reagents and Chemicals. All PPCPs, steroid and alkylphenol (AP) standards were purchased from Sigma-Aldrich (Milwaukee, WI), except atorvastatin, anhydro-erythromycin A (Toronto Research Chemicals: Toronto, Canada), iopromide and iohexol (U.S. Pharmacopeia: Rockville, MD). Isotope-labeled internal standards were purchased from C/D/N Isotopes (Pointe-Claire, Canada), Cambridge Isotope Laboratories (Andover, MA), SigmaAldrich and Toronto Research Chemicals. Desmethoxy iopromide was kindly provided by Bayer Pharma (Berlin, Germany). See the Supporting Information for details of labeled internal standards. HPLC grade formic acid, ammonium acetate, and ammonium formate were obtained from Sigma-Aldrich. Gas chromatographic grade solvents were purchased from EMD (Gibbstown, NJ), and Fisher Scientific (Pittsburgh, PA). Reagent water was prepared using an in-house system designed by US Filter (Warrendale, PA). The dechlorinating agent sodium thiosulfate was purchased from Fisher Scientific. Analytical standards were prepared in methanol. Anhydroerythromycin A (erythromycin[-H2O]) stock was prepared in acetonitrile to reduce the likelihood of its conversion to erythromycin enol ether or erythromycin A. Plant Description. Sampling was conducted at a large water reclamation facility in the Los Angeles, CA metropolitan area. This facility is a step-feed nitrification/denitrification, tertiary wastewater treatment plant employing both free chlorine and chloramines as disinfection agents. The filtration media is anthracite coal sitting on top of sand and gravel. Figure 1 shows a schematic of the treatment process, and the location where samples were collected. The plant has a design capacity of 62.5 million gallons/day, mostly from its approximately 650 000 residential customers. Detention time of sewage prior to entering this plant is relatively short (0.5-2 h). Hydraulic residence time (HRT) is 8-12 h generally broken down as follows: 1.5 h primary treatment, 3 h secondary (activated sludge) digestion, 1.5 h secondary settling, 0.25 h filtration, and 1.5 h chlorine contact time. The plant does not receive any waste from known pharmaceutical manufacturers, but does receive wastewater from nine hospitals within the service area. Sampling. Three separate sampling events were conducted: July 8-9, 2008 (Tuesday-Wednesday); May 18-19, 2009 (MondayTuesday); and October 19-20, 2009 (Monday-Tuesday). A composite graph of influent/effluent flows for the days sampled is depicted in Figure 2. Samples were collected every hour in 24 separate, acid washed glass bottles using an
Figure 2. Composite daily plant flow (combined of three events). Less consistent effluent flow is due to periodic back flushing of 3° filter beds.
automated ISCO sampler (Lincoln, NE). Sodium thiosulfate (∼1 g) was added to individual collection bottles prior to sampling. Samples were transferred to precleaned, amber glass bottles, and stored at 4 °C until extraction. Each onehour sample (225-350 mL, depending on the sampler) was kept in a separate container, then extracted and analyzed individually. During the second sampling event, two onehour samples were combined for extraction of steroids and alkylphenols (a total of 12 extracts). Additionally, in the second and third sampling event, a flow-weighted, composite sample from the same 24 h period was collected to compare to individual, hourly samples. All samples were extracted within three days of collection, within the established method holding times (greater than 4 days for all compounds; see the Supporting Information). Additionally, total organic carbon (TOC) was measured in the hourly samples from the third event using EPA SW846 Method 9060. Extraction and Analysis. Methods were validated using USEPA guidelines,28 and validation statistics are provided in the Supporting Information. Reporting limits (RL) for steroids were determined empirically by adding standard to pre-extracted matrix samples. All calibration curves were linear fit, weighted 1/x and had r g 0.995, except triclocarban, which used a quadratic curve due to responses outside of the instrument's linear range (See Supporting Information). During the course of the study, more labeled internal standards were introduced, and the analytical method became more precise and robust with each sampling event. Only 17 PPCPs were quantified in samples from July 2008, but the analyte list was expanded to 30 compounds for the later events. Likewise, the list of isotope-labeled internal standards increased with each sampling. 1229
dx.doi.org/10.1021/es102452f |Environ. Sci. Technol. 2011, 45, 1228–1234
Environmental Science & Technology SPE Extraction. Samples were extracted on polymeric, solid phase extraction cartridges (SPE) using an Autotrace automated SPE extractor (Caliper Life Sciences: Hopkinton, MA). PPCP and EDC compounds were split into two separate extraction procedures. Procedures are summarized here, with details available in the Supporting Information. Each day's extraction contained a method blank and a laboratory control standard (LCS). PPCPs were extracted from 200 mL of sample spiked with isotope-labeled internal standard. These samples were loaded onto Oasis HLB (Waters: Milford, MA) 200 mg SPE cartridges at a rate of 10 mL/min, rinsed, dried, and then eluted with methanol followed by methanol/dichloromethane (70/30). Extracts were reduced to 1.0 mL final volume, transferred into autosampler vials and stored at 4 °C until analysis. EDCs, which include steroidal hormones and APs, were extracted using 500 mL of sample spiked with isotope-labeled internal standard. Samples were loaded onto 500 mg SPE cartridges at a rate of 12 mL/min. Oasis HLB or Strata X (Phenomenex: Torrence, CA) cartridges were used interchangeably, both providing acceptable results. Samples were rinsed with methanol/water, dried, and then eluted with methanol/dichloromethane (50/50). Extracts were reduced to 1.0 mL final volume, transferred into autosampler vials and stored at 4 °C until analysis. Analysis. Four separate analyses were performed on the extracts. PPCP extracts were analyzed in positive electrospray ionization (ESIþ) mode and then negative (ESI-) mode. EDC extracts were analyzed in ESIþ mode for steroidal hormones and ESI- for APs. All analyses were performed on a Shimazdu HPLC system (Columbia, MD: 10ADvp pumps and SIL-HTc autosample unit) coupled with an Applied Biosystems API5000 triple quadrupole mass spectrometer (MS/MS; Foster City, CA). HPLC, and MS parameters may be found in the Supporting Information. PPCPþ. Eight point calibration curves ranged from 10 to 2000 ng/L. For some compounds that regularly exceeded this range, wider calibration ranges were used: 30 to 6000 ng/L for atenolol and 100 to 20 000 ng/L for iohexol. The HPLC separation used for ESI þ is similar to that described by Gros, et al.14 Chromatographic separation was performed using a Dionex (Sunnyvale, CA) Acclaim Polar Advantage II C18 column (150 mm 2.1 mm, 3 μm particle diameter) at a flow rate of 0.35 mL/minute (2 μL injection). An HPLC gradient using a pH 4 buffered aqueous mixture of formic acid/ammonium formate and a 1:1 mixture of acetonitrile/methanol with 0.1% formic acid started with organic solvent at 3%, and increased to 95% at 10 min, then held for 6 min. PPCP-. Eight point calibration curves ranged from 10 to 1000 ng/L. For some compounds that regularly exceeded this range, wider calibration ranges were used: 20 to 2000 for gemfibrozil and 30 to 3000 ng/L for iopromide. Chromatographic separation was performed using an Agilent (San Jose, CA) Zorbax RX-C18 column (150 mm 2.1 mm, 5 μm particle diameter) at a flow rate of 0.4 mL/minute (3 μL injection). An HPLC gradient using 40 mg/L aqueous ammonium acetate and methanol started with organic solvent at 3%, increased to 25% at 1 min, and then increased to 99% by 6 min and held for 5 min. Steroid. Five point calibration curves ranged from 2 to 50 ng/L. Chromatographic separation was performed using a Thermo (Waltham, MA) Aquasil C18 column (50 mm 2.1 mm, 3 μm particle diameter) at a flow rate of 0.4 mL/minute (10 μL injection). An HPLC gradient using 0.1% aqueous formic acid and methanol
ARTICLE
with 0.1% formic acid started with organic solvent at 30%, was held for 0.4 min and then increased to 40% at 2 min, then 50% at 3, and 90% at 8 min, then held for 1.4 min. Alkylphenol. Seven point calibration curves ranged from 5 to 200 ng/L for 4-tert-octylphenol (OP) and 25 to 1000 ng/L for technical 4-nonylphenol (NP). Chromatographic separation was performed using the same Thermo Aquasil column as steroid analysis, at a flow rate of 0.4 mL/minute (5 μL injection). An HPLC gradient using 40 mg/L aqueous ammonium acetate and methanol started with organic solvent at 65%, then increased to 95% at 2.2 min, and was held for 4 min. Reporting limits were 10 ng/L for most PPCP compounds, 2 ng/L for steroids, 5 ng/L for OP and 25 ng/L for NP (Table 1). Carryover below reporting limits was noticed for azithromycin, triclocarban and carbamazepine. There was some residual system/environmental contamination for triclocarban, triclosan, BPA, and nonylphenol, which accounted for their 25 ng/L reporting limit.
’ RESULTS AND DISCUSSION Extraction and Analysis QA/QC. Extractions were found to be effective and the analytical methods were reliable. All extracted (method) blanks had concentrations of target analytes
