Development and Calibration of a Passive ... - ACS Publications

Apr 9, 2012 - Office of Environment and Heritage, P.O. Box 29, Lidcombe NSW 1825, Australia. ⊥. NIOZ Royal Netherlands Institute for Sea Research, P...
1 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Development and Calibration of a Passive Sampler for Perfluorinated Alkyl Carboxylates and Sulfonates in Water Sarit L. Kaserzon,*,† Karen Kennedy,† Darryl W. Hawker,‡ Jack Thompson,† Steve Carter,§ Anthony C. Roach,∥ Kees Booij,⊥ and Jochen F. Mueller† †

The University of Queensland, The National Research Centre for Environmental Toxicology (Entox), 39 Kessels Road, Coopers Plains QLD 4108, Australia ‡ Griffith University, School of Environment, Nathan QLD 4111, Australia § Queensland Health Forensic and Scientific Services, Coopers Plains QLD 4108, Australia ∥ Office of Environment and Heritage, P.O. Box 29, Lidcombe NSW 1825, Australia ⊥ NIOZ Royal Netherlands Institute for Sea Research, P.O. Box 59, 1790 AB Texel, The Netherlands S Supporting Information *

ABSTRACT: Perfluorinated chemicals (PFCs) are emerging environmental contaminants with a global distribution. Due to their moderate water solubility, the majority of the environmental burden is assumed to be in the water phase. This work describes the application of the first passive sampler for the quantitative assessment of concentrations of perfluorinated alkylcarboxylates (PFCAs) and sulfonates (PFSAs) in water. The sampler is based on a modified Polar Organic Chemical Integrative Sampler (POCIS) with a weak anion exchange sorbent as a receiving phase. Sampling rates were between 0.16 and 0.37 L d−1, and the duration of the kinetic sampling stage was between 2.2 and 13 d. A field deployment in the most urbanized estuary in Australia (Sydney Harbour) showed trace level concentrations from passive samplers (0.1−12 ng L−1), in good agreement with parallel grab sampling (0.2−16 ng L−1). A separate field comparison of the modified POCIS with standard POCIS suggests the latter may have application for PFC sampling, but with a more limited range of analytes than the modified POCIS which contains a sorbent with a mixed mode of action.



in the effluent.13,14 This makes WWTP outfalls potential point sources in the environment.15,16 This tendency to pass through water treatment also extends to drinking water, and numerous studies have found low level (ng L−1) concentrations in municipal drinking waters.17,18 In some areas contaminated water supplies have provided a significant contribution to daily intakes and raised serum concentrations in residents affected.8,19,20 Passive sampling provides a low-cost and time-integrative sampling approach that has already proven useful for a broad range of environmental contaminants.21−27 However, application of passive sampling devices has been most effective when applied to neutral compounds. Challenges remain for sampling ionized chemicals as most passive sampling phases are incompatible with these species. To our knowledge no quantitative method has been described for a passive sampler for PFCAs and PFSAs in water. Alvarez et al.28 reported the detection of PFOS and

INTRODUCTION Poly- and perfluoroalkyl compounds (PFCs) have been manufactured on an industrial scale for over 50 years.1,2 The perfluoroalkylcarboxylates (PFCAs) and sulfonates (PFSAs) have surfactant properties, are anionic at the pH of environmental waters, and resistant to both chemical and metabolic degradation.3−5 The persistence of these compounds in the environment, their consistent detection in monitoring studies,6,7 and possible impact on organisms8 have prompted growing concern. In 2010 perfluorooctanesulfonate (PFOS) and its salts were included under the Stockholm Convention as substances whose production and use should be restricted9 and in 2012 the European Commission proposed an Environmental Quality Standard for inland surface waters of 650 pg L−1 for PFOS.10 Perfluoroctanoate (PFOA) has also been subject to increased regulation with a number of steps taken to limit its emission to the environment.11 The moderate water solubility of these compounds means they are often found in aquatic environments, and transport via waterways appears to be a major distribution pathway both locally and globally.5,6,12 Multiple studies have shown that PFCAs and PFSAs are not effectively removed in wastewater treatment plants (WWTPs) and in some instances are enriched © 2012 American Chemical Society

Received: Revised: Accepted: Published: 4985

September 19, 2011 April 2, 2012 April 9, 2012 April 9, 2012 dx.doi.org/10.1021/es300593a | Environ. Sci. Technol. 2012, 46, 4985−4993

Environmental Science & Technology

Article

Figure 1. PFC concentrations in POCIS (left) and water (right) vs time for PFHxA, PFOA, and PFDA during the calibration study.

monitoring of aquatic persistent organic pollutants. POCIS are passive samplers developed for relatively polar (log Kow < 4.0) compounds comprising a sorbent enclosed between two polyethersulfone membranes.23,29,30 Grabic et al.31 reported the accumulation of PFOS, PFOA and perfluorononanoate

PFOA in polar organic chemical integrated samplers (POCIS) extracts as part of a monitoring program of three wastewater treatment plant effluents and other natural waters in the United States. This suggests that POCIS may be suitable passive samplers for PFCs including PFCAs and PFSAs for global 4986

dx.doi.org/10.1021/es300593a | Environ. Sci. Technol. 2012, 46, 4985−4993

Environmental Science & Technology

Article

limiting membranes (Pall Supor 0.45 μm pore size, 47 mm diameter, 140 μm thickness). The total exposed surface area of these POCIS when assembled was 16 cm2, less than a standard commercially available POCIS (41 cm2) and the surface areato-sorbent mass ratio was 27 cm2 g−1. A higher sorbent mass (600 vs 200 mg) and a different sorbent type (compared with the standard POCIS configuration) were chosen to increase the sorption capacity of the sampler. A larger pore size (0.45 vs 0.1 μm) was chosen to enhance the sampling rate although increased biofouling may become a consideration. However no fouling was observed on deployed samplers from this study. Prior to deployment, each assembled POCIS was conditioned in a 100-mL beaker using 20 mL of 0.1% (v/v) ammonia solution in methanol followed by 20 mL of methanol and 40 mL of water (10 min for each solvent). This conditioning, based on sorbent manufacturer recommendations, prepares the sorbent to interact efficiently with the sample matrix. Individual POCIS were sealed in solvent-rinsed aluminum foil and stored at 4 °C prior to and after deployment in both the calibration and field studies. PES membranes were cleaned before POCIS construction in 200 mL of methanol for 20 min followed by 400 mL of water for 10 min. Calibration of Passive Samplers. POCIS were calibrated for PFCs using a static renewal experimental design in a 1400-L water tank described in detail previously.24 A methanol solution (0.75 L) containing seven PFCs (PFHxA, PFHpA, PFOA, PFNA, PFDA, PFBS, PFOS) was spiked into potable water and mixed well to achieve nominal concentrations of between 250 and 350 ng L−1. POCIS were secured in parallel within stainless steel deployment cages, using steel spacers, and the cages were mounted on the rotating arms (set to 11 rpm) of the tank. POCIS were exposed for 1 to 26 days in a staggered consecutive deployment design (Figure 1). The water in the tank was exchanged, using the same dosing regime described above, after every 3−7 days across this period (static renewal) to ensure minimal depletion of the PFCs. Grab samples (500 mL) were collected daily and a portion (50 mL) was analyzed for PFCs. Two samples (before and after exchange) were collected on days when the water was exchanged. The tank remained covered with a stainless steel lid throughout the course of the study to minimize volatilization. The water flow over the samplers was estimated to be 0.23 ± 0.04 m s−1 using codeployed passive flow monitors (PFMs) as described by O’Brien et al.33 Temperature was recorded every 20 min using a submersible data recorder (Thermochron ibutton, Dallas, TX, USA) and averaged 27 °C. The pH of the water was measured daily using pH strips and averaged 6.5. All mean exposure parameters (temperature, pH, salinity, flow velocity) in the calibration study are summarized in Table S1. Sydney Harbour Field Study. Replicate POCIS were deployed in Homebush Bay, Sydney Harbour (Australia) in October 2010 for 2, 4, and 7 days. Grab samples (1 L) were collected on initial deployment (t = 0) and on each day that POCIS were retrieved. Average exposure conditions (temperature, salinity, and flow velocity) during the field study are summarized (Table S1) along with a representative pH for the deployment site based on previous monitoring.34 Flow velocity was estimated using PFMs.35 Retrieved POCIS and grab water samples were stored at 4 °C and transported on ice. Comparison with Standard POCIS. Standard POCIS of the pharmaceutical configuration (containing 200 mg of Oasis HLB sorbent, enclosed between two polyethersulfone Pall

(PFNA) (22−152 ng), in a POCIS with Oasis HLB sorbent as the receiving phase, when deployed in rivers in the Czech Republic. However a lack of calibration data and characterization of sampling kinetics means that these data cannot be used quantitatively. PFCAs and PFSAs exist in water of neutral pH as relatively nonvolatile anions. The weak anion exchange material Oasis WAX has been shown to be a suitable solid phase extraction (SPE) sorbent for anionic perfluorinated species in water32 due to modification of Oasis HLB sorbent with piperazine groups that offer a weak anionic mechanism of retention. A comparison between Oasis HLB and Oasis WAX for extraction of PFCAs and PFSAs from water has revealed a generally similar performance, but with the latter superior for relatively short-chained species.32 Therefore it was hypothesized that a similar weak anion exchange sorbent would provide a suitable sequestration phase for a range of PFCs within a POCIS. The aim of this work was to calibrate a modified POCIS as a passive sampler for PFCAs and PFSAs in water by incorporating a weak anion exchange sorbent (Strata XAW) in the POCIS. Using a suite of selected compounds from these families, uptake kinetics and sampling rates were investigated in the laboratory. To explore the sampler’s applicability under field conditions, an initial deployment in the most urbanized Australian estuary (Sydney Harbour) with parallel grab sampling was conducted. The field performance of the modified POCIS was also compared with that of a standard POCIS.



MATERIALS AND METHODS Chemicals, Materials, and Reagents. PFCAs investigated in this work were perfluoropentanoate (PFPeA), perfluorohexanoate (PFHxA), perfluoroheptanoate (PFHpA), perfluorooctanoate (PFOA), perfluorononanoate (PFNA), perfluorodecanoate (PFDA), perfluoroundecanoate (PFUnDA), and perfluorododecanoate (PFDoDA). The PFSAs w ere perfluorobutanesulfonate (PFBS), perfluorohexanesulfonate (PFHxS), and perfluorooctanesulfonate (PFOS). The acid forms of native PFCAs were purchased from Wako Pure Chemical Industries (Osaka, Japan). The acid forms of PFSAs were purchased from AccuStandard Inc., (New Haven, CT, USA). A 13C-labeled PFCA and PFSA solution mixture (MPFAC MXA, Wellington Laboratories, Guelph, Ontario, Canada) was used as a quantification and recovery internal standard in all samples (Supporting Information (SI) Table S2). A 13C-labeled instrument performance internal standard perfluoro[13C8]octanoic acid (M8PFOA) was purchased from Wellington Laboratories (Guelph, Ontario, Canada). Stock solutions of each native PFC were prepared at a concentration of 1 mg mL−1 in methanol. Working solutions and spiking mixtures were prepared by dilution of the stock in methanol. All standards and solutions were stored at 4 °C. HPLC grade methanol and ammonia solution (32% w/w) were purchased from Merck, Germany. Ultra pure water was used in sampler construction, sample cleanup, and chemical analysis (HI-PURE water system, Permutit, Australia). All laboratory equipment and glassware was solvent-rinsed with acetone followed by methanol and dried prior to use. Modified POCIS Samplers. POCIS passive samplers were constructed as described previously,23 but using a specific configuration different from the commercially available POCIS. This modified POCIS contained 600 mg of Strata XAW (Phenomenex, Sydney, Australia) as the receiving phase, enclosed between two polyethersulfone (PES) diffusion4987

dx.doi.org/10.1021/es300593a | Environ. Sci. Technol. 2012, 46, 4985−4993

Environmental Science & Technology

Article

Supor 100 microporous (0.1 μm) membranes; with a surface area of 45.8 cm2) were deployed in triplicate alongside modified POCIS (triplicate) for 14 days (during June−July 2011) at the outflow of a Wastewater Treatment Plant in Brno − Modřice (Czech Republic). The study was conducted as part of an interlaboratory comparison study of passive samplers for emerging pollutants, organized by the Network of reference laboratories and related organizations for monitoring and biomonitoring of emerging environmental pollutants (NORMAN).36 Two field blanks (one for each POCIS configuration) and a laboratory blank were used for QA/QC measures. Water temperature, water velocity, and pH averaged 18 ± 0.6 °C, 0.1 ± 0.02 m s−1, and 7.7 ± 0.1, respectively. Retrieved POCIS were transported on ice and stored at 4 °C until extraction and analysis. Extraction of POCIS. Each POCIS was disassembled and the sorbent was transferred while moist, using a stainless steel spatula, into a precleaned empty 6-mL SPE cartridge with a 20μm glass fiber frit. PES membranes were removed for extraction from POCIS deployed in the calibration study for 1, 7, 13, and 26 d and placed in separate cartridges without rinsing to avoid loss of PFCs. Data from the membranes were used to estimate membrane−water sorption coefficients (Kmw) (SI Figure S3 and Table S7). Quantification and recovery internal standard (MPFAC MXA) was spiked (50 μL; 0.08 ng μL−1) onto the sorbent and membrane samples. After 1 h, samples were eluted on an SPE manifold under vacuum with 6 mL 0.1% (v/v) ammonia solution in methanol followed by 6 mL of methanol. The eluate was reduced under a gentle stream of nitrogen to about 0.5 mL, and made up to a final volume of 1 mL (1:1 methanol/water) for LC/MS/MS analysis. The instrument performance standard (M8PFOA) (10 μL; 0.4 ng μL−1) was added just prior to analysis. Extraction of Grab Samples. Grab samples (50 mL for laboratory calibration; 1 L in Sydney Harbour) were extracted using 6 mL, 150 mg Strata XAW anion exchange SPE cartridges (Phenomenex, Sydney Australia). Cartridges were conditioned prior to sample loading with 4 mL of 0.1% (v/v) ammonia solution in methanol followed by 4 mL of methanol and 4 mL of water. Water samples were spiked (50 μL; 0.08 ng μL−1) with quantification and recovery internal standard (MPFAC MXA) prior to loading onto the cartridges. The cartridges were left to dry under vacuum for 1 h after loading before being eluted as described above for the POCIS sorbent. Analysis of PFCs. Samples were analyzed by LC/MS/MS following Thompson et al.18 Briefly, a HPLC (Shimadzu Corp., Kyoto, Japan) coupled to a QTrap4000 triple quadrupole mass spectrometer (AB Sciex, Concord, Ontario, Canada) was used. The target PFCs were separated by gradient elution on the HPLC. The mass spectrometer was operated in negative electrospray ionization mode using scheduled multiple reaction monitoring (MRM). Target PFCs were identified and confirmed by retention time and by comparing MRM transition intensity ratios between the sample and an appropriate concentration standard from the same run. Mass labeled PFCs (i.e., quantification and recovery internal standard, MPFAC MXA) were used to quantify the corresponding native compounds in the sample and adjust for method recovery based on isotope dilution. PFCs without a corresponding mass labeled standard (i.e., PFHpA and PFBS) were quantified using the closest standard in terms of molar mass and functional group (i.e., PFBS was quantified using the PFHxS internal standard). M8PFOA was used to check

instrument performance and calculate MPFAC MXA recoveries. The limit of quantification for the suite of PFCs employed, determined by signal-to-noise ratio, was 0.5−0.9 ng L−1 (with an instrumental detection limit of 0.1−0.2 ng L−1). QA/QC. Field and fabrication control (i.e., nonexposed) POCIS were employed alongside the exposed POCIS during the preparation, deployment, retrieval, extraction, and analysis of these samplers during both the laboratory calibration (n = 2) and field study (n = 2), respectively. In addition, blank water samples (n = 2) were transported to and from the field study site and extracted alongside the grab water samples. Two POCIS were deployed for two consecutive 13-day periods to assess the reproducibility of uptake in the laboratory calibration study. Differences in accumulation of PFCs in these POCIS were