POLYETHYLENE PASSIVE SAMPLER METHOD DEVELOPMENT FOR FLUOROTELOMER ALCOHOLS AND OTHER NEUTRAL PERFLUORINATED ALKYL SUBSTANCES

Fluorotelomer alcohols (FTOHs) and other polyand per-fluorinated alkyl substances (PFASs) are common and ubiquitous by-products of various industrial telomerization processes. This class of volatile and semi-volatile compounds has been shown to degrade into a wide variety of perfluorinated carboxylic acids (PFCAs) including perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS), persistent organic pollutants. Recent atmospheric studies have shown fluorotelomer alcohols and their degradation products present in high concentrations spreading out from point sources in North America, Europe, and Asia. This study developed a method for measuring fluorotelomer alcohols through the use of polyethylene (PE) passive samplers coupled to their analysis via gas chromatography-mass spectrometry (GC/MS). Polyethylene-water partitioning coefficients, log KPEW, were determined in a laboratory study and ranged from log KPEW 3.2 to 6.7. Field deployment of PE samplers in aqueous systems was conducted in the outflow of a local waste-water treatment plant. Target FTOH, perfluorosulfonamidoethanol (FOSE), and perfluorosulfonamide (FOSA) compounds were detected in all PE samplers above background concentrations. Maximum accumulated amounts from aqueous exposure ranged from 1860 ng per sheet for 6:2 FTOH down to 3.5 ng per sheet for EtFOSE. Polyethylene-air partitioning coefficients, log KPEA, were estimated using a field deployment of PEs directly compared to active sampling. Atmospheric concentrations of targeted PFASs in Providence (RI, USA) varied daily, with the FTOHs found to be most prevalent (average 10.1 – 14.5 pg/m). Measured concentrations fall within accepted range of literature values for urban environments and indicate the effectiveness of PE passive samplers in detecting FTOHs, FOSAs, and FOSEs in atmospheric deployments. This thesis demonstrated PE samplers are effective in ambient aqueous environments for detecting and quantifying FTOHs, FOSAs, and FOSEs above their blank levels. Additional laboratory experiments are necessary to verify estimated PE-water partitioning constants.


INTRODUCTION
Fluorinated compounds have become ubiquitous across the globe over the past half-century, as increasing industrial production of synthetic organic compounds containing fluorine has led to their introduction into the environment (Key et al. 1997).
These fluorinated compounds are usually composed of a saturated carbon chain with many or all of its hydrogens replaced with fluorine and the terminal carbon ending in various functional groups such as hydroxyls, sulfonates, carboxylic acids, etc. Fluorine is highly electronegative and correspondingly has a high ionization potential and low polarizability (Kissa 1994). When bound to carbon, the large difference in ionization potential results in a strong, highly polarizable bond. The resulting fluorine sheath causes a low-energy surface that is both hydrophilic and lipophilic and consequently repels both oil and water (Lewandowski et al. 2006). As these compounds repel oil and water they reduce the surface tension between the water and another phase surface and are considered surfactants. These surfactant properties have been applied to a wide number of commercial and industrial applications including: paper, textiles, paints, non-stick cookware, polishes, electronics, and water-repellant clothing (Kissa 1994). Production: There are two primary synthetic pathways that are used in the industrial production of these fluorochemicals: the Simons Electro-Chemical Fluorination (ECF) process developed by 3M and the telomerization process used by DuPont®. The ECF process involves the fluorination of long carbon chain feedstocks in a stepwise manner (Abe 1999). Theoretically this leads to the formation of every possible partially fluorinated intermediate compound, and in practice there is a significant amount of byproducts and waste (Sartori & Ignat'ev 1998). These byproducts can be either fully fluorinated or only partially fluorinated with reactive functional groups. The later byproducts of interest include FOSEs, FOSAs, and several other POPs such as PFOS and PFOA (Kissa 1994). The other major industrial synthetic pathway for fluorochemicals involves a controlled polymerization reaction using tetrafluoroethylene (TFE) as the base-polymer unit and an alcohol functional group as the cap. The use of 2-carbon polymer units promotes the formation of linear homologues containing an even-number of carbons (6, 8, 10 etc.) as well as limiting waste byproducts (Schultz et al. 2003).
FTOHs and PFOAs are two of the primary fluorinated intermediates created using this process (Lehmler 2005 (Grandjean et al. 2012;Granum et al. 2013).
In rodents both PFOS and PFOA have been shown to be associated with increased incidences of cancer, decreased body weight, and increased liver weight (Seacat et al. 2003;Kennedy et al. 2010).
Global Distribution: Elevated PFOS concentrations have been found present in wildlife tissue at polar sites in both the Arctic and in the Antarctic (Giesy & Kannan 2001;Houde et al. 2006;Houde et al. 2011). PFOS has a relatively low vapor pressure and high water solubility compared to other POPs (Giesy & Kannan 2002;Krusic et al. 2005) and would not be expected to be transported long distances via the atmosphere. The high concentrations found at both poles indicate long-range transport is occurring and additional processes contribute to long-range transport. A secondary atmospheric source of PFOS and PFOA is possibly from the degradation of FTOHs and other semi-volatile PFASs such as FOSEs and FOSAs. These parent compounds have been identified as precursors for PFOS and PFOA as well as several other perfluorocarboxylic acids (Wallington et al. 2006;Hurley et al. 2004). Hydroxyl radical attack on these precursors is very slow and atmospheric lifetimes range from 10-20 days for FTOHs of varying carbon length and from 20-50 days for selected FOSAs (Stock et al. 2004;Piekarz et al. 2007). The estimated atmospheric residence time for 8:2 FTOH is greater than 50 days (Wania 2007). A 10-50 day lifetime is sufficient to allow for hemispheric transport to the Arctic from primary source regions.
Further evidence of the long-range transport of the precursor compounds exist.
FTOHs and other neutral PFASs have been found all around the world, in both the atmosphere and the surface ocean (Gawor et al. 2014;González-Gaya et al. 2014).
Several cruise transects along the Atlantic have reported elevated atmospheric concentrations in the Arctic, along the coast of Europe, off the coast of Brazil, and Antarctica (Dreyer, Langer, et al. 2009;Xie et al. 2013;Wang et al. 2015). These more remote oceanic areas are far downwind of the elevated concentrations found in populated areas near large-scale industrial production of fluorochemicals, such as Toronto or in Northern Europe, where high concentrations of FTOHs have been observed in the atmosphere as well as in indoor environments (Harner et al. 2006;Jahnke, Ahrens, et al. 2007; Barber et al. 2007;Björklund et al. 2009).
The global distribution of PFASs is rapidly changing as the global fluoropolymer industry continues to grow. For more than a half-century the demand for consumer products using fluoropolymer coatings has been continuously rising; global production has increased at ~5% per year. The global budget for FTOHs produced in 2002 was estimated at 5 x 10 6 kg/yr (Hurley et al. 2004). By 2020, industry experts have predicted that production will approach 4.78 x 10 7 kg/yr, an order of magnitude growth in less than 20 years (Hexa Research 2015). Much of this growth is centered in developing countries such as China, India, and Brazil where there are abundant raw materials, less-rigorous regulations, and low operating costs (Anon 2015). These increases in localized production, particularly in countries such as China and Brazil have already begun altering global distributions. Recent atmospheric and marine studies that reach poleward of 70° S have shown elevated concentrations of fluorinated compounds off the Antarctic coast (Dreyer, Weinberg, et al. 2009;Wei et al. 2007;Del Vento et al. 2012). Measured concentrations are lower than those measured in the Arctic, but nonetheless indicate longrange atmospheric transport is taking place in the Southern Hemisphere. Clearly, the long-range transport of the PFOS and PFOA precursors makes volatile PFASs a global concern and their concentrations are worth monitoring Sampling: The majority of studies that observe FTOHs and PFASs in the environment utilize active sampling methods. These methods typically require a large amount of sample media to be collected (e.g. air, water, sediment, etc.) to quantify the low environmental concentrations that are found. For active sampling of air or water, a large volume of media is pulled through a filter and adsorbent on which the POPs collect over time. The instrumentation associated with active sampling is expensive and time consuming, prohibiting the widespread monitoring of these compounds. In recent years, a variety of passive sampling techniques have been developed in order to measure many POPs in the environment Harner et al. 2006;Lohmann et al. 2012). These techniques include sorbent-impregnated polyurethane discs, solid-phase extraction cartridges, activated carbon felts, and polyethylene sheets (PE). PEs have not yet been applied to monitoring FTOHs and other neutral PFASs, which so far are typically collected using high-volume air samplers (Ahrens et al. 2011;Jahnke, Huber, et al. 2007;Liu et al. 2013).
PE sampling devices are polymer matrices that rely on the accumulation of hydrophobic contaminants through passive diffusive processes. Due to its reliance on diffusion, PE samplers inherently select only for gaseous compounds in the air and freely dissolved compounds in the water (Adams et al. 2007). In comparison to many active and passive methods, PE sheets have advantages in minimal intermolecular and environmental interactions, low cost, and the ease of handling and usage (Lohmann et al. 2012). In addition, the ability to measure both aqueous and atmospheric concentrations at sampling sites provides insight into the transport processes that control a compound's movement through the environment and the quantification of air-water fluxes (Morgan & Lohmann 2008;McDonough et al. 2014;Khairy et al. 2014).

Statement of Problem:
The overarching goal for this research was the validation and field-testing PE as a novel sampling techniques for the monitoring of neutral PFASs. Several of the guiding Instrument Analysis: All samples were analyzed using gas chromatography/mass spectrometry processing. This solution mix was created using individual reference standards. PEs were extracted in individual 60 mL amber vials using ~55 mL of Hexane for 24 hours ( Figure   3). Extracts were concentrated using a Rotovap under a mild nitrogen stream to ~200 µL.
After which, 10 µL of a 100 ng/mL d14 p-terphenyl was added as an injection standard.
PUF/XAD Sandwiches: PUF/XAD sandwiches were prepared using precleaned XAD and PUF materials, a modified Soxhlet extraction thimble, and precleaned aluminum foil. XAD polymeric beads, purchased from Sigma Aldrich, were precleaned using subsequent 24 hours extractions using acetone, dichloromethane, hexane, and methanol respectively. The PUFs used to create the sandwiches were cleaned in a high-temperature pressurized automated solvent extraction system using multiple rinses of hexane and dichloromethane. Soxhlet extraction thimble were modified by removing the bottom portion. Each sandwich was prepared by: 1) A single PUF was wedged in the bottom, 2) 25 grams of XAD was poured in over the first PUF, and 3) a second PUF was placed in the top of the extraction thimble sealing in the XAD ( Figure 2). Assembled PUF/XAD sandwiches were then wrapped in muffled aluminum foil and stored in a refrigerator until deployment.
PUF/XAD sandwiches were extracted in precleaned and muffled Soxhlet extraction apparatuses assembled in series using ~150 mL of hexane. Prior to extraction, each sample was spiked with 25 µL of a 50 ng/mL mass-labeled surrogate reference standard solution. After 24 hours, hexane extracts were concentrated down to ~200 µL in a Rotovap under a mild nitrogen stream. Concentrated extracts were then spiked with 10 µL of a 100 ng/mL d14 p-terphenyl solution as an injection standard and placed in -10 °C freezer until instrument analysis. Aqueous concentrations for each vial were analyzed at the end of the laboratory experiment. PFASs were extracted from the water onto Oasis WAX solid-phase microextraction cartridges (SPME) on a vacuum manifold ( Figure 5). Prior to use, each cartridge was conditioned using 5 mL 0.3% NH4OH in MeOH followed by 5 mL of 0.1M formic acid in water and equilibrated with 5 mL of plasma-grade reagent water. After the sample was passed through the cartridge, the SPME fibers were washed with 5 mL of 20% MeOH in 80% 0.1M formic acid in reagent water followed by 2 mL of 0.3% (v/v) NH4OH in water. Cartridges were dried for 15 minutes and then eluted in clean centrifuge tubes using 3, 15 mL hexane rinses. The 3 hexane rinse extracts were combined and then blown down to ~100 µL under a gentle nitrogen stream. The and 24 (Figures 31 & 32). As both compounds use d7-MeFOSE as the surrogate for their calculations, the large variation in its recovery on these deployment days led to this high variability.
All compounds except for 8:2 FTOH, 10:2 FTOH, and 10:2 FTAcr appeared to reach equilibrium over the course of the 32-day experiment. For these 3 compounds the linear fitting was best when including all data points, indicating that equilibrium was not reached for these compounds. In the absence of other data, we have used these concentrations as the equilibrium concentrations with the understanding that subsequent calculated values represent lower limit concentrations. Lower limit data has value when the alternative is no data.
The water used as the solvent system was then tested in several different manners in order to hopefully ascertain the ending aqueous concentrations. The first method that was employed was a liquid-liquid extraction in the manner of van Leeuwan using a 50:50 hexane:DCM mixture (van Leeuwen & de Boer 2007). The small amount of water used in each sample, ~40 mL, was not conducive to the extraction procedure, leading to low recoveries of the surrogate standards for the test samples. Another attempt was made to determine the ending aqueous concentrations using a solid-phase extraction (SPE) with Oasis WAX cartridges, as described elsewhere (Taniyasu et al. 2005). These extractions were also unsuccessful; higher recoveries were found than in the liquid-liquid extraction attempt, but were too low for accurate calculation of the aqueous concentrations. Due to the difficulties in establishing the ending concentrations, the aqueous concentrations used for the determination of the partitioning constants were simply subtracted from the initial added amount. A major flaw with this method of calculation is that it does not take into account the degradation or loss of any of the compounds. Due to the volatile nature of these compounds it is likely that some losses would occur.