IN DEGREE PROJECT ENVIRONMENTAL ENGINEERING,SECOND CYCLE, 30 CREDITS

, STOCKHOLM SWEDEN 2018

Electrodialytic Remediation of PFAS-Contaminated Soil

GEORGIOS NIARCHOS

KTH ROYAL INSTITUTE OF TECHNOLOGYSCHOOL OF ARCHITECTURE AND THE BUILT ENVIRONMENT

Electrodialytic Remediation of PFAS-Contaminated Soil Georgios Niarchos

Supervisor Jon Petter Gustafsson

Examiner Jon Petter Gustafsson

Supervisors at Swedish University of Agricultural Sciences Mattias Sörengård Lutz Ahrens

Degree Project in Environmental Engineering and Sustainable Infrastructure

KTH Royal Institute of Technology

School of Architecture and Built Environment

Department of Sustainable Development, Environmental Science and Engineering

SE-100 44 Stockholm, Sweden

TRITA-ABE-MBT-18480

i

Abstract Per- and polyfluoroalkyl substances (PFASs) are a group of anthropogenic aliphatic compounds, widely known for their environmental persistence and toxicity to living beings. While they are ubiquitous in the environment, interest has been focused on contaminated soil, which can act as a primary recipient and source of groundwater contamination. Electrokinetic technology is based on the movement of ions under the effect of an electric field. This could be a promising remediation solution, since PFASs are usually present in their anionic form. The contaminants can then be concentrated towards the anode, thus reducing a plume’s volume and possibly extracting the substances from soil. The preliminary aim of the present study was to evaluate the potential of using electrodialysis for the remediation of PFAS-contaminated soil for the first time. Experiments were run with natural contaminated soil samples, originating from a fire-fighting training site at Arlanda Airport, and at Kallinge, Sweden, as well as in artificially spikedsoil. Electrodes were placed in electrolyte-filled chambers and separated by the soil with ion-exchange membranes for pH-control. In total, five experiments were conducted. Two different setups were tested, a typical 3-compartment EKR cell and a 2-compartment setup, to allow for pH increase and facilitate PFAS desorption. Two different current densities were tested; 0.19 mA cm-2 and 0.38 mA cm-2. After twenty-one days, soil was cut in ten parts lengthwise and triplicate samples were analysed for PFAS concentrations, with HPLC-MS/MS. Sixteen out of the twenty-six screened PFASs were detected above MDL in the natural soil samples. The majority of the detected PFASs showed a positive trend of electromigration towards the anode, under both current densities, with only longer-chained compounds (c>8) being immobile. This can be attributed to the stronger sorption potential of long-chained PFAS molecules, as has been reported in previous sorption studies. Mass balance distribution for a high current density (0.38 mA cm-2) experiment revealed that 73.2% of Σ26PFAS was concentrated towards the anode, with 59% at the soil closer to the anode, 5.7% at the anion exchange membrane and 8.5% at the anolyte. It also showed higher mobility for short-chained molecules (c≤6), as they were the only compounds to be extracted from soil and be concentrated in the anolyte. Higher current densities were not directly correlated with higher electromigration rates, as to the lack of mass balance data for the low current density experiments. Regardless, electrodialysis could be a viable option for PFAS soil remediation and further research to encourage the understanding of the migration mechanism, as well as combination with other treatment methods is encouraged.

Keywords Per- and polyfluoroalkyl substances, Electrokinetic, Ground, Groundwater, Treatment, Electrochemistry, Environment

ii

Sammanfattning Per- och polyfluoralkylsubstanser (PFAS) är en grupp av antropogena alifatiska föreningar, allmänt kända för sin miljöpåverkan och toxicitet för levande varelser. Medan de är allestädes närvarande i miljön har intresset varit inriktat på förorenad mark, som kan fungera som primär mottagare och källa till grundvattenförorening. Elektrokinetisk teknik är baserad på jonernas rörelse under effekten av ett elektriskt fält. Detta kan vara en lovande lösningsmedel, eftersom PFAS är vanligtvis närvarande i sin anjoniska form. Föroreningarna kan sedan koncentreras mot anoden, vilket reducerar en plums volym och eventuellt extraherar ämnena från jorden. Det preliminära målet med den föreliggande studien var att utvärdera potentialen att använda elektrodialys för sanering av PFAS-förorenad jord för första gången. Experimenten kördes med naturliga förorenade jordprover, härrörande från en brandbekämpningsplats vid Arlanda flygplats, och i Kallinge, Sverige, samt i konstgjort spikedsol. Elektroder placerades i elektrolytfyllda kamrar och separerades av jorden med jonbytesmembran för pH-kontroll. Totalt genomfördes fem experiment. Två olika inställningar testades, en typisk 3-facks EKR-cell och en 2-facksinställning, vilket möjliggör pH-ökning och underlättar PFAS-desorption. Två olika strömtätheter testades; 0,19 mA cm-2 och 0,38 mA cm-2. Efter tjugo dagar skärs jorden i tio delar i längdriktningen och trippelprover analyserades för PFAS-koncentrationer, med HPLC-MS / MS. Sexton av de tjugosex screenade PFAS: erna detekterades över MDL i de naturliga markproverna. Majoriteten av de upptäckta PFAS-värdena visade en positiv trend av elektromigration mot anoden under båda strömtätheten, varvid endast längre kedjiga föreningar (c> 8) var immobila. Detta kan hänföras till den starkare sorptionspotentialen hos långkedjiga PFAS-molekyler, vilket har rapporterats i tidigare sorptionsstudier. Massbalansfördelning för ett experiment med hög strömtäthet (0,38 mA cm-2) visade att 73,2% av Σ26PFAS koncentrerades mot anoden, med 59% vid jorden närmare anoden, 5,7% vid anjonbytarmembranet och 8,5% vid anolyten. Det visade också högre rörlighet för kortkedjiga molekyler (c≤6), eftersom de var de enda föreningarna som skulle extraheras från jord och koncentreras i anolyten. Högre strömtätheter var inte direkt korrelerade med högre elektromigrationshastigheter, avseende bristen på massbalansdata för experimenten med låg strömtäthet. Oavsett elektrodialys kan det vara ett lönsamt alternativ för PFAS-markrening och ytterligare forskning för att uppmuntra förståelsen för migrationsmekanismen, liksom kombinationen med andra behandlingsmetoder främjas.

Nyckelord Per- och polyfluoralkylsubstanser, Electrokinetisk, Grundvatten, Förorening, Elektrokemi, Miljö

iii

Acknowledgements

This dissertation marks the completion of my master’s studies in Environmental Engineering and Sustainable Infrastructure at KTH Royal Institute of Technology, in Stockholm, Sweden. The project’s research part was conducted at the Department of Aquatic Sciences and Assessment at the Swedish University of Agricultural Sciences (SLU), in Uppsala, and partly at the Department of Civil Engineering, at the Technical University of Denmark (DTU), in Copenhagen.

At this point I would like to express my gratitude to all the people who were directly and indirectly involved in the realisation of the project. First and foremost, I want to thank my examiner and supervisor Jon Petter Gustafsson for giving me the spark of motivation and providing me the necessary means to conduct research on this subject. I want to acknowledge the contribution of SLU and specifically my main supervisor Mattias Sörengård for his on-point guidance, support and resourceful attitude. A big thank you goes to my co-supervisor Lutz Ahrens for showing trust in me and sharing his immense knowledge, which was essential for the completion of the project. I also want to acknowledge the people at DTU, who provided me with the necessary tools to conduct my research and made my short stay in Denmark enjoyable. Also, thank you Clara for helping me with the translation.

Lastly, I want to thank everyone who stood by me during the whole course of my studies. My family, partner, friends and colleagues have all been my source of inspiration and supported me in many ways along this path.

iv

Table of Contents Introduction ................................................................................................................................................... 1

Per- and Polyfluoroalkyl Substances ........................................................................................................ 1 Structure and Classification ................................................................................................................. 1 Physicochemical Properties .................................................................................................................. 2 Fate and Transport ............................................................................................................................... 3 Toxicity and Exposure Pathways ......................................................................................................... 4 Regulations and Guidelines .................................................................................................................. 4

Electrodialytic Remediation ...................................................................................................................... 5 Principles ............................................................................................................................................... 6 Efficiency Parameters ........................................................................................................................... 7

Hypothesis and Research Questions ........................................................................................................ 8 Materials and Methods .................................................................................................................................. 8

Soil Sampling ............................................................................................................................................. 8 Soil Characterisation ................................................................................................................................. 9 Electrodialytic Remediation ...................................................................................................................... 9

Soil Preparation .................................................................................................................................... 9 Experimental Setup ............................................................................................................................ 10 Electric conductivity & pH ................................................................................................................. 10

PFAS Analysis .......................................................................................................................................... 12 Extraction ............................................................................................................................................ 12 Quantification ..................................................................................................................................... 12 Quality Control and Assurance .......................................................................................................... 13

Results .......................................................................................................................................................... 13 Soil Characteristics .................................................................................................................................. 13 Operating Parameters .............................................................................................................................. 14

Electroosmotic flow ............................................................................................................................ 14 Conductivity ........................................................................................................................................ 15 pH ........................................................................................................................................................ 16

PFAS Migration ........................................................................................................................................ 17 Natural soils ........................................................................................................................................ 17 Spiked Soil ...........................................................................................................................................20 Overall Distribution ............................................................................................................................20 Non-ionic PFASs ................................................................................................................................. 21

Discussion ..................................................................................................................................................... 22 References .................................................................................................................................................... 24 Appendix A ................................................................................................................................................... 29 Appendix B ................................................................................................................................................... 31 Appendix C ................................................................................................................................................... 33

v

Abbreviations

AFFF Aqueous film forming foam

AEM Anion exchange membrane

CEM Cation exchange membrane

DC Direct current

ECF Electrochemical fluorination process

EDL Electrochemical double layer

EOF Electroosmotic flow

foc Organic carbon fraction

FOSA Perfluorooctane sulfonamides

FOSAA Perfluoroalkane sulfonamidoacetic acids

FTSA Fluorotelomer sulfonic acid

IS Internal standard

Kd Sediment-water partition coefficient

KOC Organic-carbon partition coefficient

LOI Loss on ignition

MDL Method detection limit

PFAA Perfluoroalkyl acid

PFAS Per- and polyfluoroalkyl substance

PFBA Perfluorobutyric acid

PFBS Perfluorobutane sulfonic acid

PFCA Perfluorinated carboxylic acid

PFDA Pefluorodecanoic acid

PFDoDA Perfluorododecanoic acid

PFHpA Perfluoroheptanoic acid

PFHxA Perfluorohexanoic acid

PFHxDA Perfluorohexadecanoic acid

PFHxS Perfluorohexane sulfonic acid

PFNA Perfluorononanoic acid

PFOA Perfluorooctanoic acid

PFOcDA Perfluorooctadecanoic acid

PFOS Perfluorooctane sulfonic acid

PFPeA Perfluoropentanoic acid

PFSA Perfluorinated sulfonic acid

vi

PFTeDA Perfluorotetradecanoic acid

PFTriDA Perfluorotridecanoic acid

PFUnDA Perfluoroundecanoic acid

UHPLC-MS/MS Ultra-high-performance liquid chromatography tandem mass spectrometry

WHC Water holding capacity

1

Introduction

Per- and Polyfluoroalkyl Substances Per and polyfluoroalkyl substances, denoted by the acronym PFASs, refer to a group of man-made emerging contaminants that have been produced since the early 1950s (Giesy & Kannan, 2002). They are aliphatic substances, containing one or more carbon atoms, with hydrogen atoms on their alkyl chain being replaced by fluorine atoms (Buck et al., 2011). The C-F bond of PFASs is characterized by extreme chemical stability (Smart, 1994). Furthermore, they consist of a hydrophobic carbon tail, and a hydrophilic ionic head, and they exhibit excellent surfactant properties (Yasuda et al., 1995).

Due to these unique physicochemical properties, PFASs have been applied in a wide variety of industrial activities and consumer goods. These include surface protection agents (textile coatings, anti-grease cookware), paper and cardboard packaging, pesticides and herbicides, paints, and personal care products (Krafft & Riess, 2015). One other major use of PFASs is in fire-fighting facilities, such as military bases and civil airports, where they are included in aqueous film forming foams (AFFFs) to extinguish hydrocarbon-fuel fires (Baduel et al., 2017).

The widespread use of PFASs has led to their dispersal in the environment, thus affecting wildlife and humans with potential toxic effects (Kennedy et al., 2004; Cui et al., 2009; Butenhoff et al., 2012). The extent of the environmental problem was accentuated at the start of the second millennium. Giesy and Kannan (2001) were the first to document the occurrence of PFASs in wildlife, with findings that verified the contamination on a global extent, even in pristine ecosystems. Around the same time, prevalence of the substances in the blood of non-industrially exposed humans was detected by Hansen et al. (2001), yet again denoting the high level of PFAS contamination.

Since then, studies such as the above have increased the attention of the public and the scientific community and gave rise to new research that nowadays primarily aims to identify the potential toxicity of the substances, as well as to find ways to isolate them and remove them from exposure pathways (Kucharzyk et al., 2017). Towards this direction, the present study intends to examine a novel methodology of PFAS extraction from polluted soil, using electrodialytic remediation technology. The project´s aim is to investigate the viability of the technology for PFAS remediation.

Structure and Classification

PFASs were first synthesised with the use of the electrochemical fluorination process (ECF), which uses electricity to replace hydrogen with fluorine bonds (H-C H-F) (3M, 2017). Their structure comprises of a hydrophobic/oleophobic carbon tail and a hydrophilic charged head. Their head is their functional group and is typically a sulfonic or carboxylic acid. PFASs can include different amounts of hydrogen, oxygen and nitrogen atoms. Their added advantage compared to the conventional hydrocarbon surfactants is that the covalent C-F bonds provide chemical stability against temperature fluctuations as well as extreme pH conditions (AMAP, 2004).

Figure 1. Chemical structure of PFOA. Image source: NIEHS, 2018

2

Figure 2. Classification and chemical formulas of the PFASs that were screened and quantified in the current project. The nomenclature is based on terminology suggested by Buck et al. (2011). L- and B- refer to the linear and branched

forms of the same compound.

The term perfluorinated is used to describe molecules, in which all hydrogen atoms are replaced by fluorine atoms, while in polyfluorinated substances at least one but not all hydrogen atoms are replaced by fluorine (Buck et al., 2011). Some polyfluorinated substances, so-called precursors, are capable of transformations to more stable compounds of the perfluoroalkyl group. PFASs contain the general moiety C𝑛𝑛F2𝑛𝑛+1− − R, where R represents the charged functional group and n refers to the number of carbon atoms in the alkyl chain (AMAP, 2004). According to the functional group and chain length, they can be divided into groups and sub-classes, as can be seen in Figure 1. Regarding their classification, the chain length of PFASs is of utmost importance, since substances with more carbon atoms exhibit higher persistence and are more bioaccumulative and ubiquitous in the environment (Wang et al., 2013). Consequently, PFASs of longer chain, especially PFOS (n=8) and PFOA (n=7) have gained the most attention and have gradually been phased out and replaced by shorter chain PFASs (3M, 2000). PFASs are thus commonly referred to as “long” and “short” chain in literature (Buck et al., 2011). To avoid subjectivity, hereafter, the OECD (2012) definition is used, which dictates that “long chain” PFASs refer to perfluoroalkyl carboxylic acids (PFCAs) with n≥8, perfluoroalkyl sulfonates (PFSAs) with n≥6, and all other PFASs with n≥7 (Buck et al., 2011).

Physicochemical Properties

The widespread use of PFASs is largely attributed to their unique physicochemical properties. One key property is their repellency of water and oil, which is due to their hydrophilic functional group head and lipophilic carbon tail. Another important characteristic of PFASs is their chemical stability. When fluorine, the most electronegative element of the periodic table, is paired with carbon, they form the strongest bond possible in organic chemistry (O'Hagan, 2008). This results in chemical stability against fluctuations of temperature and oxidation state, as well as extreme pH conditions. In addition, due to their low vapor pressures, volatilization from ground or water is hindered, especially in longer chain compounds (Rayne & Forest, 2009).

3

Table 1. Typical physicochemical properties of PFOS and PFOA. Source: (USEPA, 2014) Property PFOS PFOA

Molecular weight (g mol-1) 538 414

Water solubility at 25oC (mg L-1) 550-570 9.5*103 Melting point (oC) >400 45-54

Boiling point (oC) Not measurable 188-192

Vapor pressure at 20oC (mm Hg) 2.48 * 10-6 0.017 a

Organic-carbon partition coefficient (logKoc)

2.57 (Value estimated on anion and not the salt) 2.06

Half-Life Atmospheric: 114 days Water: > 41 years (at 25oC)

Atmospheric: 90 days b

Water: > 92 years (at 25oC) a Extrapolation from measurement

b The atmospheric half-life value identified for PFOA estimated based on available data determined from short study periods

The solvation properties of PFASs, i.e. their partitioning and solubility, govern their transportation patterns through environmental media. According to Ahrens et al. (2011), one significant parameter is the sediment-water partition coefficient Kd, which depends both on the chemical properties and on the sediment characteristics. Low Kd values mean that the substances occur predominantly in the dissolved phase and are more likely to disperse from soil to aquatic environments. Based on their findings, particularly PFOS and PFOA tend to desorb from sediments and be transported in the aqueous phase. The solubility of the substances in water decreases with increasing chain length, while it is generally higher for ionizable PFASs (PFCAs & PFSAs). The water solubility is a highly temperature-dependent parameter, which increases sharply with temperature (Prevedouros et al., 2006).

The adsorption of PFASs to soils and sediments depends on their structure. Two principal mechanisms affect the process: the interaction of the hydrophobic carbon tail with the organic carbon fraction of the soil, and the electrostatic interactions of the polar head with the charged clay particles (Kucharzyk et al., 2017; Higgins & Luthy, 2006). Key soil parameters that affect this sorption behaviour are the pH value and the organic matter content. More specifically, adsorption is enhanced under acidic conditions, as is the case with most anionic surfactants, while desorption is inversely proportional to the organic carbon fraction (foc) (Du et al., 2014; Ahrens et al., 2011). This is due to the fact that increased pH creates more negatively charged surfaces, which results in electrostatic repulsion of PFASs. It is important to note that inorganic materials present in soil, such as clay, sand, and iron oxides, can also be effective PFAS sorbents (Higgins & Luthy, 2006). The ionic behaviour of PFASs is dependent on the pH of the environmental matrix and the properties of the specific substance (Buck et al., 2011). Due to their low pKa values, most PFASs are mainly present as anions in natural conditions, except for FOSA, which has a pKa value of 6.56 (Ahrens et al., 2012).

Fate and Transport

Due to the recalcitrance of PFASs, produced amounts almost entirely end up in the environment. Environmental pollution originates from point or non-point sources. Significant point sources are landfills, wastewater treatment plants, and areas where fire-fighting activities occur, such as civil airports, military bases and fire disaster sites (Ahrens, 2011; Ahrens et al., 2015; Dauchy et al., 2017). Non-point sources are more ambiguous, since they are not directly related to human activities and are harder to identify. However, atmospheric wet or dry deposition is the main recorded non-point source. Dreyer et al. (2010) found thirty-four PFASs in twenty precipitation samples in northern Germany in concentrations of 1.6-48.6 ng L-1 for ΣPFAS, while Scott et al. (2006) have detected 0.6-89 ng L-1 of PFOA in north America. Another important non-point source is surface runoff from contaminated sites (Ahrens, 2011). To distinguish between direct and indirect sources, Simcik and Dorweiler (2005) suggested that the ratio of PFHpA:PFOA could be used as an indicator of the source of PFASs, with

4

values ≤0.9 indicative of urban sources and values markedly >0.9 representative of atmospheric deposition as a source.

Although sources of PFASs are anthropogenic, their distribution in the environment is ubiquitous. Numerous studies have detected concentrations even in pristine ecosystems and remote areas such as the arctic (Giesy & Kannan, 2002; Kwok et al., 2013; Åkerblom et al., 2017). This phenomenon is due to long-range transportation and deposition of the substances, which commonly happens by means of oceanic currents and sea spray, or atmospheric degradation of precursors (Ahrens, 2011). According to Prevedouros et al. (2006), oceanic transport contributes the most to the long-range transport of PFASs, with 2-12 tons year-1 of PFOA being transferred by oceanic means to the Arctic. In addition, neutral precursor compounds (e.g. fluorotelomers and perfluoroalkyl sulfonamides), which are volatile, can also end up in the atmosphere and precipitate, even in remote regions.

Generally, the main transportation paths of PFASs in the environment are aquatic and atmospheric. Riverine pathways are key, as they are usually located in populated watersheds with pollution sources and at the same time they provide an input to the marine environment (Mclachlan et al., 2007). Ground polluted areas can be sources of groundwater infiltration, which is of primary concern especially in Sweden, where 50% of the country´s drinking water originates from groundwater (Banzhaf et al., 2017).

It is important to mention that physicochemical properties affect the environmental fate of PFASs. Neutral PFASs, such as fluorotelomer alcohols (FTOHS), normally have higher vapor pressure and lower water solubility compared to ionizable PFASs and are therefore more prone to long-range atmospheric transport. The partitioning behaviour also influences the fate of the substances. For instance, long chain PFASs tend to bind more strongly on solid particles and are therefore more likely to be stored in soil and sediment. (Ahrens, 2011)

One important sink of PFASs in the environment is the ocean. However, soils and sediments also constitute significant sinks, because of their capability to store polluted surface water, as well as their potential to contaminate groundwater aquifers (Ahrens et al., 2010). Therefore, remediation of polluted ground near hotspots could prevent the diffusion of the substances in the environment.

Toxicity and Exposure Pathways

PFASs are potential toxicants when in contact with living beings. Based on numerous epidemiological studies, adverse health effects associated with exposure to them include hepatotoxicity (Gleason et al., 2015), thyroid and sex hormone disruption (Olsen & Zobel, 2007), carcinogenesis (Alexander & Olsen, 2007; Butenhoff et al., 2012; Chang et al., 2014), and immunosuppression (Dalsager et al., 2016; Kielsen et al., 2016). Toxicity of the substances is typically expressed after long-term exposure. Long-chain PFASs have estimated serum half-lives that range from 2-9 years (PFOA 2-4, PFOS 5-6, PFHxS 8-9), which leads to their bioaccumulation (ATDSR, 2017). As proteinophilic substances, they tend to accumulate in the blood, the kidneys, and the liver (Martin et al., 2003). The bioaccumulation characteristics of PFASs are relevant with the food web. Specifically, accumulation and biomagnification are noticed more in air-breathing animals than water-respiring organisms, due to the high aqueous solubility and low volatility of PFASs (Table 1).

Routes of human exposure to PFASs are dietary intake, inhalation of indoor air and dust, and dermal contact with soil or other polluted surfaces (Jian et al., 2017). The importance of each exposure pathway is region-specific; still, trends can be noticed. As reviewed by Noorlander et al. (2011) and Denys et al. (2014) a significant contributor to exposure is consumption of fish and crustaceans.

Regulations and Guidelines

Concerns regarding the environmental persistence and toxicity of PFASs have led to a set of regulatory actions on a global, national, and regional level, as well as voluntary steps towards reducing production and use of PFASs. In 2002, the chief global manufacturer of PFASs, 3M, voluntarily announced the production cessation of two major PFAS compound groups, PFOS and PFOA (US EPA, 2000; Buck et al., 2011). European Union with regulation No 757/2010, has enlisted PFOS and its precursors under the Stockholm Convention, Annex B, on persistent organic pollutants (POPs), restricting their use in the 181 participant countries (Vierke et al., 2012; UNEP, 2009; European Comission, 2010). In the case of

5

Sweden, the country falls under the European chemicals’ regulation, “REACH”, which since 2009 dictates the restriction of production and use of PFOS, in chemical products and goods (KEMI, 2009).

According to the Swedish environmental protection agency, different generic guideline values are set, depending on the land use. The two types of land classification are “sensitive land use” (KM) and “less sensitive land use” (MKM) (Table 2). This classification governs the level of protection that is required. For example, a residential area would normally be under KM protection, while industrial zones would be under MKM. According to this classification different guidelines are set (Table 3).

Table 2. Classification of areas for various risk objects, according to the Swedish EPA. (Naturvårdsverket, 2009)

Risk object Sensitive land use (KM)

Less sensitive land use (MKM)

People staying within area Full-time residents Part-time visitors Land environment in the area

Protection of the soil’s ecological function

Limited protection of the soil’s ecological function

Groundwater Groundwater within and adjacent to the area

Groundwater in distance ≥200m downstream

Surface Water Protection of the surface water Protection of the surface water

Table 3. Target values for PFOS in land for different exposure paths and protection objects (mg kg-1). (Swedish Geotechnical Institute, 2015)

Guideline value Sensitive land use (KM)

Less sensitive land use

(MKM) Health risk-based target value 0.031 11

Soil intake 1.9 17 Skin contact 6.8 34 Inhalation of dust 2 100 21 000 Inhalation of vapours 3 600 36 000 Intake of groundwater 0.033 - Intake of plants 0.6 -

Protection of soil environment 0.003 0.3 Protection of groundwater 0.0066 0.021 Protection of surface water 0.027 0.027 Preliminary guideline value 0.003 0.020

Despite the mentioned initiatives, regulations, and guidelines, the production and distribution of these compounds has not yet declined fittingly on a global level. A study by the Swedish Chemicals Agency (KEMI) (2009) revealed that more than 3000 PFASs were at the time on the global market. What is more, the chemical footprint of PFAS pollution is quite large, as detailed in the previous chapters. These are strong indicators that PFASs are still emerging pollutants and that solutions are called-for, both for the cessation of PFAS production, as well as their removal from environmental inventories.

Electrodialytic Remediation The concept of applying electric current in soil dates back to 1809, when F.F. Reuss discovered that the application of direct current (DC) in porous media results in transportation of water (Reuss, 1809). Since then, various aspects of the electrokinetic technology, including soil consolidation, desalination, and dewatering, have been studied and applied (Hansen et al., 2015). In the last three decades, focus has been placed on the technology’s potential to remediate contaminated soil, using electric field as a cleaning agent (Hansen et al., 2015). While it has been extensively studied for the removal of heavy metals from soil (Virkutyte et al., 2002), until now its efficiency has not been tested for the removal of PFASs. Here, electrodialytic remediation (EDR) is examined, which is an enhanced version of electrokinetic remediation with the addition of ion-exchange membranes (Ottosen et al., 1997). So far, popular technologies, including air-stripping, thermal treatment, soil vapor extraction, and hydroxyl-based chemical oxidation have proven ineffective for PFAS removal (Kucharzyk et al., 2017) and EDR could provide an alternative in-situ solution.

6

Principles

EDR is a concentration remediation technique, and its intention is to confine contaminants in a small area. Under the effect of an electric field in a soil column, the pore groundwater or an externally added fluid act as conductive mediums. By applying direct current in the order of mA cm-2, electrons flow in one direction only, and physical, chemical and hydrological changes occur in the soil, leading to various transportation phenomena. (Acar & Alshawabkeh, 1990).

Figure 3 Main transportation processes in a soil matrix under electrokinetic remediation. On the anionic soil surface, cations are attracted forming the electric double layer (EDL).

The main transportation processes that take place in the soil matrix are: • Electromigration: the mobilisation of ionic particles towards the electrode of the opposite charge.

Cations and anions will move towards the cathode and anode respectively. This is the main transportation mechanism that PFASs present in their anionic form are expected to move through. Since most PFASs are negatively charged, they will theoretically electromigrate towards the anode. Electromigration velocity (vem) [cm s-1] is directly proportional to the electrical field (E) and can be calculated by the equation (1) (Baraud et al., 1997).

𝑣𝑣𝑒𝑒𝑒𝑒 = 𝑧𝑧|𝑧𝑧|∗ 𝑢𝑢 ∗ ∇(−E) (1)

where u is the ionic electric mobility [cm2 s-1V-1], z is the ionic electric charge [C], ∇(−𝐸𝐸) is the electrical potential gradient [V cm-1]

• Electroosmosis: the mobilisation of ions due to the electric field exerts a viscous drag on water particles, which results in hydraulic flow. The direction of the flow is dependent on the net electromigration pattern. Typically, soil pores have a negatively charged surface, which serves as a cation-accumulation area, the so-called electrical double-layer (Figure 2). There will therefore be a net flow of cations towards the cathode, and accordingly the electroosmotic flow (EOF) will follow this direction. Uncharged particles are mainly expected to transport through this mechanism. Electroosmosis (Qeos) [cm3 s-1] can be quantified by the equation (2). (Baraud et al., 1997)

𝑄𝑄𝑒𝑒𝑒𝑒𝑒𝑒 =𝐾𝐾𝑒𝑒𝑆𝑆∇(−𝐸𝐸) (2)

where Ke is the electroosmotic permeability coefficient, S is the cross-sectional area [cm2], ∇(−𝐸𝐸) is the electrical potential gradient [V cm-1]

• Electrophoresis: the mobilisation of charged soil particles, which normally happens towards the anode in the negatively charged soil. Due to the tortuosity of the soil fabric, particles usually do not travel long distances. However, this mechanism becomes relevant when micelles are formed, for example with the addition of surfactants or in the treatment of slurries (Hansen et al., 2015; Acar & Alshawabkeh, 1990; Baraud et al., 1997).

7

Upon the successful migration of the contaminants to the respective electrode, extraction methods can be further used to remove them. These include for instance electroplating, pumping, precipitation, complexation, etc according to the characteristics of the specific contaminant (Virkutyte et al., 2002).

Efficiency Parameters

Various authors have summarized the parameters that affect the efficiency of the extraction process. Arguably, one of the most important control parameters is the soil and pore-water pH. Hydrogen ion concentrations are inconstant during the remediation, since electrolysis of water takes place at the electrodes according to reactions (3) and (4) (Virkutyte et al., 2002).

Anode (+): 2H2O − 4e− → O2 ↑ +H+, Eo = −1.229 (3)

Cathode (-): 2H2O + 2e− → H2 ↑ +2OH−, Eo = −0.828 (4)

Essentially, every anode is an acidic source and every cathode is an alkali source. This leads to the formation of an acid and basic front respectively, that advance towards the opposite electrodes through electromigration, diffusion and advection. Since soils act as cation exchangers (Gustafsson et al., 2007), the movement of the acid front can be naturally retarded inside the soil matrix. This buffering capacity of the soil can be favourable for PFAS removal; in the case of PFAS remediation, high pH values favour the desorption from soil (Du et al., 2014) and thus prevalence of the basic front is preferred, as it could result in higher electromigration rates. However, alkaline conditions could also be undesirable, since they can cause precipitation of metals as insoluble metal hydroxides (Figure 3) and reduce the resistivity of the soil (Hansen et al., 1999). Precipitation of metals can also cause the formation of sediments that can clog soil pores and reduce dispersion over time. In the case of reduced resistivity, the enhancement of the method with electrolytes is proposed, as explained further in the following chapter.

Another significant efficiency parameter is the applied current density. Previous similar studies have used a current density on the range of 0.1 to few tens of mA cm-2 (Hansen et al., 1999; Ottosen et al., 1997; Alshawabkeh et al., 1999). Contaminant removal rates are proportional to the current density that is applied during the process, as regards heavy metals (Ottosen et al., 1997). Nevertheless, the optimum current density is relevant with the soil type, as well as the type of pollutant that is to be removed. So far, no data exist for PFASs. It is worth mentioning that high current densities (≈20 mA cm-2) have been reported to cause degradation of the substances of up to 96.7% ± 3 (Niu et al., 2012). The process of degradation involves the transformation of the substances to radicals and subsequently results in the production of CO2 and HF. However, the current study aims primarily to identify the transportation rather than degradation phenomena.

Figure 4. pC-pH diagram of ferrihydrite, showing its solubility in the water matrix. Oversaturation exists in the shaded area, where precipitation of the metal occurs as Fe(OH)3. As a result, basic conditions enhance precipitation of metals such as iron, reducing the conductivity of the soil. Image source: (Brezonik & Arnold, 2011)

8

One unique asset of the technology is its ability to decontaminate low-permeability soils, where other treatments techniques usually fail. While EDR is applicable for both coarse and fine-grained soils, its efficiency has proven higher for the latter. As reviewed by Virkutyte et al. (2002), electrokinetic remediation has demonstrated heavy metal removal efficiency of 85-95% for low-permeability soils, while for porous soils, such as peat, the efficiency drops at 65%. A similar trend is noticed for the organic contaminants phenol and acetate, with >94% removal rates.

To improve the efficiency of the technology, several enhancement techniques have been investigated in former studies. As described above, one of the most important parameters that affect the efficiency of the process is the pH. For this reason, EDR introduces the use of ion-exchange membranes, differentiating it from conventional electrokinetic treatment. The way the membranes work is by attracting counter-ions, i.e. ions with the opposite charge than the surface of the membranes, while they reject co-ions. Therefore, depending on the setup, ion-exchange membranes can be used to control the acid and base fronts that are formed in the electrodes, through a less invasive and costly way than the addition of pH regulating chemicals. Using an anion exchange membrane (AEM) on the interface between the anode and the soil, transportation of H+ is hindered, which is desirable both for PFAS extraction, as well as from an environmental point. Correspondingly, cation exchange membranes (CEM) are used at the cathode side, to control the movement of hydroxides, thus forming three compartments: anolyte, soil, and catholyte. In the present study, a variant of this setup will also be tested; the two-compartment EDR setup modified from Pedersen et al. (2015). During the two-compartment setup the cation exchange membrane will be removed, thus allowing free movement of OH- in the soil column, with the intention of increasing PFAS desorption potential.

Hypothesis and Research Questions The overall aim of the study is to assess the potential of electrodialysis for the remediation of PFAS-contaminated soil. The main hypothesis that will be tested is thus formulated as: “Application of an electric field in soil will result in the migration of ionizable PFASs towards the anode”.

Based on concerns arising from previous studies, the methodological development described in the following chapter aimed to provide answers to the following research questions:

• Do PFASs migrate under the effect of an electric field? • If so, what is the applicability of the method depending on different background conditions

(different soils, electricity, etc)? • Which process dominates the transportation pattern of various PFASs (i.e. electromigration, EOF,

electrophoresis)? • Is there a correlation between removal efficiency and PFASs’ structural properties, (i.e. functional

group, chain length, mass)? • What is the optimum experimental setup that provides the highest removal efficiency (Two- versus

three-compartment, see chapter 2)? • What is the role of pH in the remediation process?

Materials and Methods

Soil Sampling

Remediation experiments were conducted for natural soil samples, as well as for artificially contaminated soil (spiked). The natural soil samples that were used for the remediation experiments were provided by SLU and they were collected from known PFAS-contaminated locations. First, soil that originated from a fire training facility at the Arlanda airport, located near Stockholm and Uppsala (17°55´E 59°39´N) was tested. At this facility, AFFF containing PFASs was being used from 1980s’ until 2011, when it got replaced by PFAS-free AFFF (Gobelius et al., 2017). Also, soil from the area of Brantafors, Kallinge (55°43'24.39"N 13°28'40.88"E) was examined, which has also been contaminated by a nearby air-force site more than 25 years ago (Ronneby Kommun, 2018). After collection and transportation, samples were stored at 8 oC. (Ahrens et al., 2010; Du et al., 2014; Kirchmann et al., 2005)

9

Soil Characterisation

Soil characteristics were measured to gain supplementary information on key parameters that could affect the remediation process. Variables that were studied related to hydraulic and chemical properties of the soil, including carbon content, moisture content, pH, porosity, concentration of metals, and particle size. The carbon content was calculated by the loss-on-ignition (LOI) method. Soil was first dried and then placed in an oven at 550 oC for 24 hours, and the organic carbon fraction was estimated as foc = LOI/1.72 (Kupryianchyk et al., 2016). The content of heavy metals was analysed by acid digestion Danish standard (DS259) at DTU (see Appendix C) and measurement with a inductively coupled plasma-optical emission spectrometer (ICP-OES, iCAP 7200 Thermo Scientific). Soil texture was determined for the soils by the Swedish Geotechnical Institue and results can be seen in Appendix C. The gravimetric soil water content of the samples was calculated according to Formula 5. Dry mass was estimated after freeze-drying of the samples for one week.

𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌𝐌 𝐜𝐜𝐌𝐌𝐜𝐜𝐌𝐌𝐌𝐌𝐜𝐜𝐌𝐌 = �𝐌𝐌𝐰𝐰𝐌𝐌𝐌𝐌 − 𝐌𝐌𝐝𝐝𝐌𝐌𝐝𝐝�/𝐌𝐌𝐰𝐰𝐌𝐌𝐌𝐌 ∗ 𝟏𝟏𝟏𝟏𝟏𝟏%

(5)

Electrodialytic Remediation

Electrodialytic remediation experiments were conducted in a bench scale design at the Department of Aquatic Sciences and Assessment, Swedish University of Agricultural Sciences (SLU), in Uppsala, Sweden, and at the Technical University of Denmark (DTU), in Copenhagen, Denmark. In total, five experiments were conducted, with variations in the experimental setup, operational parameters, and soil type, as can be seen in Table 4. Experiment I was used as a screening test to check if there was any migration of PFASs during EDR and therefore there were less data available. Since the soil resistivity was alternating during each experiment, the voltage was automatically adjusted by the power supply (PS 200B Elektro-Automatik GmbH) following Ohm’s law (𝑉𝑉 = 𝐼𝐼 ∗ 𝑅𝑅) to keep constant current. Current and voltage were monitored daily with a multimeter.

Table 4. Overview of EDR experiments.

Experiment Setup Soil Origin Current [mA]

Current Density

[mA cm-2] Laboratory

I 3-compartment Arlanda 10 0.19 DTU II 3-compartment Arlanda 20 0.39 DTU III 2-compartment Arlanda 20 0.39 DTU IV 2-compartment Kallinge 10 0.19 SLU V 2-compartment Spiked 10 0.19 SLU

Soil Preparation

All soil samples that were used in the electrokinetic experiments were pre-processed with a mortar and thoroughly mixed, to ensure a homogenous soil contamination. The soil was passed through a 2 mm sieve to remove stones, larger aggregates and debris. Subsequently, it was homogenized in an end-to-end shaking machine for at least 24 h. Before being placed in the remediation cell, each soil sample was saturated with purified MilliQ-water, approximately to the soil’s water holding capacity (WHC), to simulate saturated aquifer conditions.

Since not all PFASs of interest were found in the natural soil samples, in the last electrokinetic experiment an artificially spiked soil was tested. The soil was prepared with a particle composition resembling the OECD’s guidelines (2009). In addition, having a known starting concentration, would help to avoid biased results. Previous experiments by Schedin (2013) on PFAS soil adsorption showed that aging positively influences adsorption strength, hence the spiked samples that were used have formerly been aged for more than twelve months. Consequently, aged spiked soils with 15 compounds (see Appendix A. table 2) were homogenised with natural, non-contaminated soil to form a final sample aiming at a PFAS concentration 0.06 mg kg-1 (60 ppb) and the following composition:

10

- 75% Sandy soil with particle size ≤2 mm (non-contaminated), from Högåsa, as 0.35 – 0.45 cm in Kirchmann et al., (2005)

- 15% Clay soil with ≥30% kaolinite content from Vreta Kloster (non-contaminated), as 0.35 – 0.45 cm in Kirchmann et al., (2005)

- 5% Clay soil with ≥30% kaolinite content from Vreta Kloster (spiked with 0.6 mg -1 kg PFAS mix) - 5% Humified peat from Econova AB, Sweden (spiked with 0.6 mg kg-1 PFAS mix)

Experimental Setup

The electrodialytic remediation was conducted in cylindrical Plexiglass cells (d= 8.1 cm, L = 10 cm), under full control of variables. Each experiment lasted for twenty-one days, according to typical duration times that were used in previous experiments at DTU (Ottosen et al., 1997). Electrodialytic remediation differs from conventional electrokinetic remediation setups in that it is pH-regulated with the use of ion exchange membranes. Lab scale trials were mainly categorized into two-compartment (2C) and three-compartment (3C) experiments, where the compartments refer to the chambers that are formed, due to the addition of ion-exchange membranes according to the experimental setup. In both setups, the soil was placed in a chamber of 10 cm length and 8 cm diameter. The difference is that in a typical 3C setup (Figure 5), the electrodes were both placed in the electrolyte compartment, while in a 2C setup, the cathode was placed directly into the soil and the catholyte is absent. With the 3C setup the effect of pH alteration due to the water electrolysis reactions was hindered, while the 2C setup was expected to cause an increase in the pH of the soil, with the intention of enhancing the partitioning of PFASs to the aqueous phase, since PFAS adsorption is known to be pH dependant (Du et al., 2014). In addition, different current densities were tested (0.19 and 0.39 mA cm-2), as higher current density is generally correlated with faster movement of ions according to previous studies (Hansen et al., 1999).

The soil was carefully packed in the middle chamber so that it contained no large pockets of air, as this could later lead to sedimentation of the soil and loss of resistivity, R. A thin layer of silicon grease was applied between each chamber connection, so that leaking was avoided. Electrode rods were inserted in the electrolyte chamber, or directly in the soil for the 2-compartment setup. Finally, the cell was held together using four threaded rods with screws, that were tightened at each of the cell’s corners. The equipment and reagents that were used in the experiments can be found in the Appendix C.

An electrolyte solution was used to enhance the conductivity of the electrolyte medium. Sodium nitrate (NaNO3) 0.01 M solution with a pH of 2 was prepared by adding 8.5 g of NaNO3 in 10 L of distilled water and 20 mL of 1:1 nitric acid (HNO3). The electrolyte solution was added in volumetric cylinders, where peristaltic pumps were connected for constant recirculation (≈15.7 mL min-1). Changes in water volume were indicative of the effect of EOF, or electrolysis of H2O, and were therefore monitored. In addition, pH changes of the electrolyte were monitored daily, since an increase of the anolyte´s pH could indicate the migration of the basic front formed at the cathode towards the anode and vice-versa. Upon termination of the experiment, the cell was taken apart and the soil, ion exchange membranes and electrolyte solution were analysed, to measure their PFAS concentrations, as described hereinafter.

Electric conductivity & pH

Electric conductivity measures the ability of a medium to pass electric current. The conductivity was measured directly in samples from the electrolytes during the experiment, and for the soil it was measured for each soil slice after the end of the experiments, by preparing a 1:10 soil:water suspension shaken in an end-over-end machine for 7 days, at 200 rpm. The conductivity of the solution was measured with the use of an EC meter (HI 2211 pH/ORP meter by Hanna Instruments). For the same samples, pH measurements were conducted with a pH meter (611 pH meter by Metrohm). Both instruments were calibrated prior to measurements.

11

Figure 5. Schematic representation of experimental setup for 2-compartment electrodialytic remediation.

Figure 6. Schematic representation of experimental setup for 3-compartment electrodialytic remediation.

12

PFAS Analysis

The specific concentrations of individual PFASs in the soil samples were measured by using liquid chromatography coupled to tandem mass spectroscopy (UHPLC-MS/MS; Quantiva TSQ; Thermo Fisher), after extracting the PFASs from the soil and membranes, using liquid-solid extraction when methanol was the eluent. Analysis was conducted both for all remediated soil sample slices, as well as the reference samples and membranes, all in triplicates. The materials and equipment used for PFAS analysis are included in the Appendix (A).

Extraction

After completion of the electrodialytic remediation, the soil columns were cut into approximately 10 slices of equal thickness (1 cm). The soil samples were then prepared for analysis, first by freeze-drying for 7 days, to remove any moisture content while avoiding PFAS losses and cross-contamination due to evaporation. Weighing was done before and after the freeze-drying, to estimate the moisture content (Formula 5). After drying, the soil samples were homogenised by crushing with mortar and pestle, and then using end-over-end shaking for 24 h. Finally, the samples were taken for PFAS analysis, liquid-solid extraction, as described elsewhere (Ahrens et al., 2010).

In brief, the main principle of this method is that the analyte which is bound to the solid matrix gets dispersed and partitioned to the liquid phase of the soil solvent solution. In this case, PFAS is the target compound bound to the soil or membrane, and the separation techniques involve shaking, centrifugation and filtration, in combination with pH adjustment, to enhance dissociation of PFAS from the solids.

For each sample, approximately 5 g of soil sample were weighed and put into a methanol rinsed 50 mL PP-tube. Then, NaOH was added and left to soak for 30 minutes, to enhance desorption of PFAS from the soil. Subsequently, 20 mL of methanol were added along with 100 μL of internal standard (IS). The sample was shaken in a wrist-action shaker for 60 minutes at 200 rpm, followed by centrifugation at 3500 rpm for 15 minutes, to separate the solvent from the solid phase. The supernatant aliquot was decanted into a new methanol rinsed 50 mL PP-tube. The extraction process was repeated with 10 mL of methanol and no addition of internal standards, with shorter shaking time (30 minutes).

The samples were concentrated down to a final volume of 0.5 mL, using a nitrogen blow-down evaporator. The method relies on the use of the inert gas N2, which is gently blown on the surface of the liquid with the use of needles (Figure 1 in Appendix B). The equilibrium between the liquid and vapor phases is thus altered, and there is faster vaporisation of the solvent. The tubes were rinsed with methanol three times during the evaporation process, to minimize analyte losses due to adsorption on the tubes’ walls. Finally, the samples were filtrated through a recycled cellulose syringe filter (0.45 μm, Sartorius), into 2 mL auto injector vials. After filtration, samples were first diluted to 1:1 (v/v) with NaOH for pH neutralisation and further to 1:1 (v/v) with milli-Q water for HPLC/MS-MS analysis. 10 μL aliquots were then stored in special glass vials and directly injected in HPLC/MS-MS.

The PFAS analysis for the membranes followed the same principle as for the soil/sediment extraction, however with doubled shaking times (120 and 60 minutes). After their use, all excess soil was brushed off, membranes were freeze-dried and cut into small pieces before extraction. Membranes were analysed for experiments II and III. Finally, electrolyte samples were analysed for PFAS concentrations before and after treatment to investigate the potential of isolating the compounds in the electrolytic compartment, for experiments II, III, and IV. To measure the PFASs in the electrolyte, direct injection was used. Specifically, 500 μL of each electrolyte sample (n=2) were diluted to 1:1 (v/v) with methanol in an Eppendorf tube, vortexed for 10 min and centrifuged at 15000 rpm. The samples were filtered through a recycled cellulose syringe filter (0.45 μm, Sartorius). In total, 159 soil extractions, 9 membrane extractions, and 14 electrolyte samples were conducted and measured.

Quantification

Quantification of PFASs was conducted for all the soil and water extracts using high performance liquid chromatography, coupled with tandem mass spectrometry (UHPLC-MS/MS; Quantiva TSQ; Thermo Fisher). The principle of the method is based on the fact that the different compounds that exist in the extracts get separated through the chromatographic column at different rates. Thereafter, target

13

compounds are ionised with electrospray and fragmented with the use of collision gas and finally go through a coupled mass spectrometer, where a detector returns the final signal intensity. The consistent use of internal standards in the samples and in the 9-point calibration curve containing all individual PFASs (0.01 – 100 ng mL-1) allowed for accurate quantification. The analytical instrument that was used for the analysis of PFAS in the samples, was TSQ Quantiva (ThermoFisher). The interpretation of peaks and mass quantification was performed with the TraceFinder software (Agilent Technologies) and further analysis of the data was done using Microsoft Office 2016 Excel.

Quality Control and Assurance

To avoid cross-contamination, all glassware equipment was carefully cleaned prior to use. The cleaning process involved rinsing with ethanol and water, dishwashing, burning at 400 oC, and lastly rinsing three times with methanol. The polypropylene tubes were also rinsed with methanol three times before use. Internal standardization method (ISM) was used to monitor and improve the precision of quantitative analysis. Internal standards are compounds of similar structure with the PFAS compounds, added in the samples at the start of the extraction process with a known concentration. They act as tracers of analyte losses during the sample preparation, extraction, sample storage, and inlet. Also, with the use of IS the recovery percentages of the extraction process could be calculated, according to the area of the peaks that were detected during quantification. Recoveries were calculated at 104%±34% for individual compounds in the solid phase and at 84%±22% in the aqueous phase.

Recovery = areaIS ∗ areasamples−1 ∗ DF, where in this case the dilution factor DF = 8

To ensure the quality of the results, as well as the repeatability of the method, quality assurance testing was conducted. Negative blank tests (n=3) were conducted for each experiment (Table 4). The negative blanks had the same extraction process, but without the addition of soil sample, just spiking with IS.

The method detection limits (MDL) were calculated for each individual compound included in the study. The MDLs were defined as the mean concentration of negative blanks plus three times the standard deviation. In case of absence of the substances in the blanks, the lowest acceptable point in the calibration curve is used as the MDL. Concentrations found under the MDL were not considered. A minimum signal-to-noise ratio of 5 was a prerequisite for detection of the compounds.

MDL = 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚(𝑏𝑏𝑏𝑏𝑏𝑏𝑛𝑛𝑏𝑏𝑒𝑒) + 3 ∗ 𝜎𝜎(𝑏𝑏𝑏𝑏𝑏𝑏𝑛𝑛𝑏𝑏𝑒𝑒)

Results

Soil Characteristics Results from the soil characterisation are provided in Table 1 and Appendix C Table 4. For the Kallinge soil, it was noticed that it had a much higher water holding capacity compared to the other samples (Arlanda and Artificial soil), which can be related to its particle size distribution and pore size. In addition, Kallinge soil had a higher carbon content, which could indicate a higher adsorption capacity for the specific sample. Further background data on soil characteristics, namely texture and concentrations of main elements, can be found in the Appendix C.

Table 5. Water and carbon content of soil samples.

Experiment Soil origin WHC (%)

Carbon content

(%) σ I Arlanda 23 1.92 0.16 II Arlanda 22 1.92 0.16 III Arlanda 22.99 1.92 0.16 IV Kallinge 34.57 4.48 0.33 V Artificial 20.67 3.55 0

14

Operating Parameters

Electroosmotic flow

By relating the profile of the water content in the soil column before and after the treatment we can gain information on the water flow during the experiments. Water moved from the anode to the cathode in all experiments, as was expected due to the EOF’s effect, Figure 7. The water profile showed greater variability for experiment IV, while it remained relatively constant for experiments II and V. The soil used for experiment IV had higher WHC, which means that it was more difficult to reach a steady state of water content, especially under the effect of EOF. EOF is also dependent on the net charge of ions in the soil, with higher ionic concentrations resulting in higher EOF rates. Experiments II and III also had higher current density, which can cause faster flow of water from the electrolyte compartment, thus compensating losses due to water electrolysis.

Figure 7. Final to initial water content ratio (Wc/Wc0) in the soil for experiments II-IV.

In general, the soil was expected to be drier towards the end of the experiments compared to at the start because of the electrolysis of water (reactions 1 and 2). A decrease in water volume is noticed for most experiments, however, for experiment V the opposite phenomenon was noticed, with an approximately 10% increase in water content throughout the whole soil sample. The independent variable in experiment V is the soil type, which can explain this discrepancy. According to the soil texture analysis (see Appendix C), the artificial soil used in experiment V had a high amount of low porosity material. Therefore, it is possible that the WHC of the soil was not reached during preparation, or possible presence of air pockets during packing allowed oversaturated conditions. Nevertheless, for experiment V it is obvious that the water level reached a steady state after completion of the experiment. For experiments IV and V, the volume loss of electrolyte was monitored. Cumulative volume losses reached up to 100 mL in 21 days, or 4.76 mL d-1, which could either be attributed to EOF or electrolysis of water (Figure 8).

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

0.00 2.00 4.00 6.00 8.00 10.00

Wc/

Wc 0

Distance from anode (cm)

Initial

II

III

IV

V

15

Figure 8. Electrolyte volume drop in the anolyte chamber during experiments IV and V

Conductivity

The electrical conductivity is proportional to the number of ions present in the soil and their mobility. As can be seen in figure 9, conductivity is generally noticed to be higher near the electrodes, at 0 and 10 cm. This phenomenon could have three explanations; i) ions are transported and concentrated at these areas through electromigration, thus creating an ion-depletion zone in the middle chamber ii) ions are generated at the electrodes through electrolysis and other reactions iii) the soil around the electrodes is heated through the Joule effect, resulting in higher solubility of ions.

Figure 9. Distribution of soil conductivity after remediation for experiments II to V.

A difference in the results between the 2- and 3-compartment setups can also be noticed. For 2-compartment experiments (III, IV, V), there was a higher conductivity towards the cathode, while for the 3-compartment (II) experiments it was much higher at the anode. Since the 3-compartment setup is the optimum for the remediation of metals, this discrepancy could be explained by the fact that in a 3-compartment setup metals migrate towards the anode. Further investigations on the concentration of metals after remediation could help validate this hypothesis.

0

20

40

60

80

100

0 5 10 15

V0lu

me

(mL)

Time (days)

Experiment IV Experiment V

0

100

200

300

400

500

600

700

0 2 4 6 8 10

Cond

uctiv

ity (m

S cm

-1)

Distance from anode + (cm)

II III IV V

16

Figure 10. Voltage variation during the remediation experiments IV, V.

For experiment IV the resistivity was too high from the start of the experiment and the power source reached its maximum voltage capacity (≈82 Volt) to achieve constant current density at 19 mA cm-2. However, the voltage dropped after the middle of the experiment (=12 days). The dryness of the soil near the anode is also related to the increased resistivity of the soil (see Appendix B). Generally, a lower voltage is noticed in 3-compartment experiments, which could be related to the pH distribution, as high 2-compartment setups resulted in pH increase, which in turn could result in precipitation of metals and thus loss of conductivity.

pH

The pH profile in the soil varied largely according to the experimental setup. In figure 11, it can be noticed that for the 3-compartment experiments (I, II) pH is kept relatively constant at a pH range of 2.7 to 4.5. At this range, PFCA, PFSA and FTSA are predominantly in their ionised forms and therefore susceptible to electromigration. As was intended, the 2C setup resulted in alkalinisation of the soil during remediation and in the creation of the basic front at the cathode part, with pH values up to 10. At high pH it is expected to have higher electrostatic repulsion between soil particles and anionic PFASs, leading to increased mobility (Higgins & Luthy, 2006) and thus increased remediation efficiency. However, as mentioned above, high pH values could be the reason for the increase in resistivity in the soil in the 2-compartment experiments, hence reducing the system’s power efficiency. Therefore, there could be a trade-off between the efficiency of PFAS electromigration and the current efficiency.

0

20

40

60

80

100

120

140

0 1 2 3 4 7 8 9 10 11 12 15 16 17 18 19

Volta

ge (V

)

Time (days)

I II III IV V

17

Figure 11. pH distribution for experiments I to IV after application of EDR

PFAS Migration

Natural soils

Thirteen out of the twenty-six screened PFASs were detected in the Arlanda soil samples (Appendix C. Table 1). Since EDR is a concentration technique, some PFASs that could not be found in the reference samples were detected after the remediation, concentrated at the anode area and especially on the anion exchange membranes. More specifically, 16 compounds were detected in total on the membrane of the 2-compartment experiment III (non-detectable in reference soil but on AEM at anode: PFBA, PFNA, EtFOSAA, 10:2 FTSA) and 15 compounds on the membrane of 3-compartment experiment II (non-detectable in reference soil but on AEM at anode: PFBA, PFNA, 10:2 FTSA). These results are indicators that some PFASs may exist at low, undetectable concentrations at the area of Arlanda airport based on the applied method detection limits. Soil from Kallinge had higher PFAS contamination. In these soil samples, 19 PFASs were found, with much higher PFCA concentrations than in the Arlanda samples (Appendix C. Table 1).

2

4

6

8

10

0 2 4 6 8 10

pH

Distance from anode (cm)

I II III IV V

18

Figure 12. Distribution of Σ26PFAS at the cathode and the anode part of the soil after remediation for experiments I to V.

PFOS was detected above the Swedish guideline values for the protection of the soil environment (≤ 3 μg kg-1 soil). Similar PFOS concentrations were measured in the areas of Arlanda and Kallinge, 9.66 and 9.37 μg kg-1 soil respectively. However, PFOA contamination was 56.83% higher in Kallinge. As can be seen in the following figures, PFASs followed a trend of electromigration towards the anode. While the trend is noticed for all ionized PFASs, the extent of electromigration varies for each experiment. For Σ26PFAS, accumulation rates at the part of the soil that is closer to the anode range from 59% to 74% of the total soil concentration.

Highest accumulation rates towards the anode soil part were noticed for the experiment III, which was a 2-compartpent setup with the highest current density. By comparing the experiments I and II, which test the same soil and setup but with different current densities, we can see that there is an increase of 9% with a doubled current density. It is important to note that figure 12 depicts the migration profile for compounds detected in the soil compartment. However, this does not present a complete picture of the distribution of PFASs, as they could have migrated towards other parts of the setup, such as the membranes and electrolytes. For this reason, the results of mass balance distribution experiments are presented later in this chapter.

59%

68%

74%

65%

63%

41%

32%

26%

35%

37%

I

II

III

IV

V

% Σ26PFAS

Anode Part Cathode Part

19

Figure 13. Post-remediation distribution of Σ26PFAS (A), PFOS (B) and PFOA (C) concentrations in the soil for

experiments I to IV (natural soil samples). Each data point represents an average of triplicates. Error bars refer to standard deviation (σ).

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10

Soil

conc

entr

atio

n (μ

g kg

-1so

il)

(A)

I

II

II

IV

0

5

10

15

20

25

30

0 1 2 3 4 5 6 7 8 9 10

Soil

conc

entr

atio

n (μ

g kg

-1so

il)

(B)

I

II

III

IV

0

0.5

1

1.5

2

2.5

3

0 1 2 3 4 5 6 7 8 9 10

Soil

conc

entr

atio

n (μ

g kg

-1so

il)

Distance from cathode (-) (cm)

(C)

I

II

III

IV

20

Spiked Soil

The spiked soil experiment (V) served the purpose of investigating PFASs that were not detectable in all the natural soil samples and to compare the electromigration rates with a known initial concentration. In total, 15 compounds were spiked in the soil, as can be seen in the Figure 13. The electromigration trend is noticed independently of the functional group, but the accumulation rates are much lower for longer chain PFASs, such as C10 PFUnDA (56%) and C11 PFDoDA (53%). There was noticed a discrepancy between spiked initial concentrations and final concentrations detected in soil and therefore mass balance studies were required to assess the overall distribution of the compounds.

Figure 14. Distribution and concentrations of PFASs for experiment V).

Overall Distribution

It is possible that accumulation of PFASs occurs not only at the soil chamber, but also at the other components of the experimental setup, i.e. the ion exchange membranes and the electrolyte chamber. To examine this possibility, experiments II and III were used to gain the distribution of the substances across these different components. From the distribution profile in Figure 14 we can see that in experiment II, 11 of the 16 detected substances accumulated towards the anode part. As discussed earlier, certain compounds were only detectable at the anode part, such as 10:2 FTSA. In experiment II, 73.24% of Σ26PFASs were concentrated towards the components of the anode part. Specifically, 58.99% of Σ26PFASs accumulated at the anode’s soil (half of the total soil), 5.7% at the AEM, 8,54% at the anolyte, 26.72% at the cathode’s soil and 0.05% at the CEM. For experiment III the accumulation rate towards the anode at the anode’s soil, AEM, and anolyte for Σ26PFASs is 90.88%. Specifically, 57.67% of the total amount of PFASs was concentrated at the 5cm of soil towards the anode, 17.71% at the AEM and 15.50% at the electrolyte, while merely 9.12% of the Σ26PFASs was left at the part of the soil column that is closer to the cathode. Low electromigration rates are noticed mainly for longer chain PFASs (C≥8). No contaminants were noticed in the catholyte, and low concentrations were found in the cation exchange membrane for experiment II. In the 2-compartment setup of experiment III, distribution of the substances across the different components shows an almost complete accumulation towards the anode (+). In addition, nine of the 17 detected substances are only found on the AEM. PFASs that were found in the anolyte, were in higher concentrations for PFBA (C3), PFPeA (C4), PFHxA (C5), PFHpA (C6), and PFHxS (C6) (Figure 13 and 14). Therefore, a trend is noticed; short-chain PFASs (C≤6) can pass through the membrane and be accumulated in the anolyte, thus being extracted from the soil matrix.

232

191 198 194179

155136

11395

82

0

50

100

150

200

250

0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5

Soil

conc

entr

atio

n (μ

g kg

-1so

il)

Distance from anode (cm)

8:2 FTSA

6:2 FTSA

FOSA

PFOS

PFHxS

PFBS

PFDoDA

PFUnDA

PFDA

PFNA

PFOA

PFHpA

PFHxA

PFPeA

21

Figure 15. Total distribution of PFASs across different components of the experiments II and III.

Non-ionic PFASs

PFASs that are usually not present in their ionised form typically belong to the group of sulfonamides, i.e. FOSAs, FOSAAs and their ethyl and methyl derivatives. FOSA, which was detected in all of the soils, behaved differently according to the setup. For FOSA, the pKa value is 6.56, which means that the specific substance can become ionised at basic environments (pH> 6.56). In the following graphs it evident that a trend of FOSA migration is noticed more for the 2-compartment experiments (III, IV, V), and especially for experiment V. This phenomenon validates the effect of pH on the ionisation and electromigration of FOSAs, for which 2-compartment setups are proposed for remediation.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

PFBA

PFPe

A

PFHx

A

PFHp

A

PFO

A

PFN

A

PFDA

PFU

nDA

PFBS

PFHx

S

PFO

S

FOSA

EtFO

SA

FOSA

A

6:2

FTSA

10:2

FTS

A

(II)Electrolyte (+) Membrane (+) Soil (+) Soil (-) Membrane (-) Electrolyte (-)

0%10%20%30%40%50%60%70%80%90%

100%

PFBA

PFPe

A

PFHx

A

PFHp

A

PFO

A

PFN

A

PFDA

PFU

nDA

PFBS

PFHx

S

PFO

S

FOSA

EtFO

SA

FOSA

A

EtFO

SAA

6:2

FTSA

10:2

FTS

A

(III)Electrolyte (+) Membrane (+) Soil (+) Soil (-)

22

Figure 16. Distribution of FOSA in experiments I to V.

Discussion As is evident by the results of the experiments, the main hypothesis of the project was validated; ionic PFASs migrated under the effect of low density electricity towards the anode. The dominant transportation mechanism for anionic PFASs was therefore electromigration and not electrophoresis, or EOF. The efficiency of the transportation mechanism was higher in the 2-compartment setup. In addition, the 2-compartment setup succeeded in the formation and prevalence of the basic over the acidic front, which is desirable both for the desorption of PFASs and ionisation of FOSA, and to prevent acidification of soils and groundwater, which is an undesirable effect of the conventional electrokinetic system. However, the movement of the basic front in the soil could be influenced by the groundwater’s movement pattern in up-scaled conditions, and thus further research is needed to understand this influence in saturated and unsaturated soils. If hydraulic or electroosmotic flow is high, it could affect the pH profile in the soil and subsequently influence the dissociation of PFASs. In any case, there was a positive correlation between the increase of the soil pH during the 2-compartment setup and the efficiency of the technology. However, there is a trade-off between the efficiency of the technology and the increase of resistance in the soil, as was evident by the results. For one experiment (IV) there was high resistivity during the remediation, which could increase the required voltage and, consequently, the required costs for applying the technology. As was also apparent by the results, intrusion of the electrolyte into the soil compartment can result in increased electric conductivity. Consequently,

y = 0.021x + 0.3743R² = 0.5855

0.00

0.20

0.40

0.60

0.80

0 5 10

Soil

conc

entr

ratio

n (μ

g kg

-1) FOSA I

y = 0.0357x + 0.2656R² = 0.0649

0.00

0.50

1.00

1.50

2.00

0 5 10

FOSA II

y = 0.01x + 0.2853R² = 0.3386

0.00

0.10

0.20

0.30

0.40

0 5 10

Soil

conc

entr

ratio

n (μ

g kg

-1) FOSA III

y = 0.0067x + 0.1671R² = 0.5251

0.000.050.100.150.200.250.30

0 5 10

FOSA IV

y = 1.7174x + 14.252R² = 0.8105

0

10

20

30

40

0 5 10Soil

cond

entr

atio

n (μ

g kg

-1)

Distance from cathode (cm)

FOSA V

23

potential increases in voltage could be avoided by preconditioning the soil with an electrolyte with high salt content. Concerning the applied current density, while higher migration rates were also expected to correlate with higher current densities, results presented in the study are not enough to support such claim. Further studies are encouraged, biased towards current density to investigate this.

One added benefit of the EDR setup proved to be the concentration of the contaminants on the AEMs. This could greatly decrease the costs of a wider remediation strategy, that could aim to concentrate PFAS plumes at the membranes and reduce the volume to be further treated. Extraction from the soil and concentration of PFASs in the electrolyte was the study’s original aim, but was finally achievable mainly for short chain PFASs, which is primarily attributed to their faster migration rates. Further experimental work is required to determine if this potential exists for longer-chained PFASs, for example by applying higher current density or longer remediation time.

Since the migration of PFASs was validated on lab-scale with this study, future work should focus on ways to further examine the underlying mechanisms that would allow up-scaling of the method. One important parameter that was not tested in this project was the kinetics of the process. To test how fast the migration of PFASs takes place, time should be used as an independent variable, while keeping the other variables constant. This could be done by conducting identical experiments in different time intervals, or by sampling the soil or contained water during the experiment. Transport modelling of the process could also help predict the kinetic behaviour of PFASs and facilitate the simulation of field conditions. Future work could also include different setups. One possibility is to investigate the effect of gravity on EOF by using a horizontal setup instead of the vertical one that was used in this study. Setups with horizonal positioning of the electrodes could theoretically help to stop the effect of EOF due to gravity and even induce hydraulic flow towards the anode, thus implementing a flushing effect in addition to electromigration. Another possible improvement of the method could be the enhancement of the soil’s conductivity by preconditioning with electrolytes or surfactants. Finally, it would be interesting to test the potential of the method to not only transport substances in soil, but also to degrade them by increasing the current density. Recent studies have shown that such a potential is possible through oxidation and degradation to shorter-chain molecules (Schaefer et al., 2017; Trautmann et al. 2015). This could result in a combined effect of transformation to shorter, mobile compounds and their subsequent migration out of the soil matrix

24

References 3M, 2000. Phase-out plan for POSF-based products. St Paul, US Environmental Protection Agency.

3M, 2017. 3M and fluorochemicals. [Online] Available at: https://www.3m.com/3M/en_US/sustainability-us/policies-reports/3m-and-fluorochemicals/ [Accessed 12 2017].

Acar, Y. B. & Alshawabkeh, A. N., 1990. Principles of Electrodialytic Remediation. Environmental Science and Technology, 27(13), pp. 2638-2647.

Ahrens, L., 2011. Polyfluoroalkyl compounds in the aquatic environment: a review of their occurrence and fate. J. Environ. Monit., Volume 13, pp. 20-31.

Ahrens, L. et al., 2012. Improved Characterization of Gas−Particle Partitioning for Per- and Polyfluoroalkyl Substances in the Atmosphere Using Annular Diffusion Denuder Samplers. Environ. Sci. Technol., Issue 46, p. 7199−7206.

Ahrens, L. et al., 2015. Stockholm Arlanda Airport as a source of per- and polyfluoroalkyl substances to water, sediment and fish. Chemosphere, Volume 129, pp. 33-38.

Ahrens, L. et al., 2010. Distribution of polyfluoroalkyl compounds in water, suspended particulate matter and sediment from Tokyo Bay, Japan. Chemosphere, Volume 79, pp. 266-272.

Ahrens, L. et al., 2011. Partitioning of perfluorooctanoate (PFOA), perfluorooctane sulfonate (PFOS) and perfluorooctane sulfonamide (PFOSA) between water and sediment. Chemosphere, pp. 731-737.

Alexander, B. H. & Olsen, G. W., 2007. Bladder Cancer in Perfluorooctanesulfonyl Fluoride Manufacturing Workers. Annals of pidemiology, 17(6), pp. 471-478.

Alshawabkeh, A. N., Yeung, A. T. & Bricka, M. R., 1999. Practical Aspects of In-situ Electrokinetic Extraction. Journal of Environmental Engineering, Volume 27, pp. 27-35.

AMAP, 2004. AMAP Assessment 2002: Persistent Organic Pollutants in the Arctic., Oslo, Norway: Arctic Monitoring and Assessment Programme (AMAP).

ATDSR, 2017. An Overview of Perfluoroalkyl and Polyfluoroalkyl Substances and Interim Guidance for Clinicians Responding to Patient Exposure Concerns, s.l.: Agency for Toxic Substances and Disease Registry.

Baduel, C. et al., 2017. Discovery of novel per- and polyfluoroalkyl substances (PFASs) at a fire fighting training ground and preliminary investigation of their fate and mobility. Chemosphere, 10, pp. 1030-1038.

Banzhaf, S. et al., 2017. A review of contamination of surface-, ground-, and drinking water in Sweden by perfluoroalkyl and polyfluoroalkyl substances (PFASs). Ambio, Volume 3, pp. 335-346.

Baraud, F., Tellier, S. & Astruc, M., 1997. Ion velocity in soil solution during electrokinetic remediation. Journal of Hazardous Materials, Volume 56, pp. 315-332.

Brezonik, P. L. & Arnold, W. A., 2011. Water Chemistry: An Introduction to the Chemistry of Natural and Engineered Aquatic Systems. New York: Oxford University Press.

Buck, R. C. et al., 2011. Perfluoroalkyl and Polyfluoroalkyl Substances in the Environment: Terminology, Classification, and Origins. Integrated Environmental Assessment and Management, 7(4), pp. 513-541.

Butenhoff, J. et al., 2012. Toxicological evaluation of ammonium perfluorobutyrate in rats: twenty-eight-day and ninety-day oral gavage studies. Reprod Toxicol., 33(4), pp. 513-30.

25

Butenhoff, J. L., Chang, S.-C., Olsen, G. W. & Thomford, P. J., 2012. Chronic dietary toxicity and carcinogenicity study with potassium perfluorooctanesulfonate in Sprague Dawley rats. Toxicology, pp. 1-15.

Chang, E. et al., 2014. A critical review of perfluorooctanoate and perfluorooctanesulfonate exposure and cancer risk in humans. Crit Rev Toxicol, 44(1), pp. 1-81.

Cui, L. et al., 2009. Studies on the toxicological effects of PFOA and PFOS on rats using histological observation and chemical analysis. Archives of Environmental Contamination and Toxicology, pp. 338-349.

Dalsager, L. et al., 2016. Association between prenatal exposure to perfluorinated compounds and symptoms of infections at age 1–4 years among 359 children in the Odense Child Cohort. Environment International, Volume 96, pp. 58-64.

Dauchy, X. et al., 2017. Per- and polyfluoroalkyl substances in firefighting foam concentrates and water samples collected near sites impacted by the use of these foams. Chemosphere, Volume 183, pp. 53-61.

Denys, S. et al., 2014. Is the fresh water fish consumption a significant determinant of the internal exposure to perfluoroalkylated substances (PFAS)?. Toxicology Letters, 231(2), pp. 233-238.

Dreyer, A., Matthias, V., Weinberg, I. & Ebinghaus, R., 2010. Wet deposition of poly- and perfluorinated compounds in Northern Germany. Environmental Pollution, 158(5), pp. 1221-1227.

Du, Z. et al., 2014. Adsorption behavior and mechanism of perfluorinated compounds on various adsorbents - A review. Journal of Hazardous Materials, pp. 443-453.

European Comission, 2010. Persistent Organic Pollutants (POPs) - Amendment of the POP regulation. [Online] Available at: http://ec.europa.eu/environment/chemicals/international_conventions/index_en.htm [Accessed 11 2017].

Giesy, J. P. & Kannan, K., 2002. Perfluorochemical surfactants in the environment, s.l.: Environmental Science and Technology.

Gleason, J. A., Post, G. B. & Fagliano, J. A., 2015. Associations of perfluorinated chemical serum concentrations and biomarkers of liver function and uric acid in the US population (NHANES), 2007–2010. Environmental Research, pp. 8-14.

Gobelius, L., Lewis, J. & Ahrens, L., 2017. Plant Uptake of Per- and Polyfluoroalkyl Substances at a Contaminated Fire Training Facility to Evaluate the Phytoremediation Potential of Various Plant Species. Environ. Sci. Technol., Issue 51, pp. 12602-12610.

Gustafsson, J. P., Jacks, G., Simonsson, M. & Nilsson, I., 2007. Soil and Water Chemistry, Stockholm: KTH Department of Land and Water Resources Engineering.

Hansen, H. K. et al., 1999. Electrodialytic Remediation of Soil Polluted With Heavy Metals Key Parameters for Optimization of the Process. Institution of Chemical Engineers, Volume 77.

Hansen, H. K., Ottosen, L. M. & Ribeiro, A. B., 2015. Electrokinetic Soil Remediation: An Overview. In: A. B. Ribeiro, E. P. Mateus & N. Couto, eds. Electrokinetics Across Disciplines and Continents. s.l.:Springer, pp. 3-18.

Hansen, K., Clemen, L., Ellefson, M. & Johnson, H., 2001. Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environmental Science & Technology, pp. 766-770.

Higgins, C. P. & Luthy, R. G., 2006. Sorption of perfluorinated surfactants on sediments. Environmental Science & Technology, pp. 7251-7256.

Jakobsson, K. et al., 2013. Exponering för perfluorerade ämnen (PFAS) i dricksvatten i Ronneby kommun, Region Skåne: s.n.

26

Jian, J.-M.et al., 2017. Global distribution of perfluorochemicals (PFCs) in potential human exposure source–A review. Environment International, Volume 108, pp. 51-62.

KEMI, 2009. Högfluorerade ämnen i kläder, skor och kemiska produkter, Sundyberg: Kemikalieinspektionen .

Kennedy, G. et al., 2004. The toxicology of perfluorooctanoate. Critical Reviews in Toxicology, pp. 351-384.

Kielsen, K. et al., 2016. Vaccination Efficacy and Environmental Pollution. In: C. Esser, ed. Environmental Influences on the Immune System. s.l.:Springer, pp. 181-204.

Kirchmann, H., Snäll, S., Eriksson, J. & Mattsson, L., 2005. Properties and classification of soils of the Swedish long-term fertility experiments: V. Sites at Vreta Kloster and Högåsa. Acta Agriculturae Scandinavica, Section B — Soil & Plant Science, 55(2), pp. 98-110.

Kirkelund, G. M., Ottosen, L. M. & Villumsen, A., 2009. Electrodialytic remediation of harbour sediment in suspension—Evaluation of effects induced by changes in stirring velocity and current density on heavy metal removal and pH. Journal of Hazardous Materials, 169(1-3), pp. 685-690.

Krafft, M. P. & Riess, J. G., 2015. Per- and polyfluorinated substances (PFASs): Environmental challenges. Current Opinion in Colloid & Interface Science, pp. 192-212.

Kucharzyk, K. H. et al., 2017. Novel treatment technologies for PFAS compounds: A critical review. Journal of Environmental Management, 15 12, pp. 757-764.

Kupryianchyk, D., Hale, S.E., Breedveld, G.D., Cornelissen, G., 2016. Treatment of sites contaminated with perfluorinated compounds using biochar amendment. Chemosphere, Biochars multifunctional role as a novel technology in the agricultural, environmental, and industrial sectors 142, 35–40. https://doi.org/10.1016/j.chemosphere.2015.04.085

Kwok, K. Y. et al., 2013. Transport of Perfluoroalkyl substances (PFAS) from an arctic glacier to downstream locations: Implications for sources. Science of the Total Environmnt, 447(1), pp. 46-55.

Martin, J. W., Mabury, S. A., Solomon, K. R. & Muir, D. C., 2003. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ Toxicol Chem., Volume 1, pp. 196-204.

Mclachlan, M. S., Holmstrom, K. E., Reth, M. & Berger, U., 2007. Riverine Discharge of Perfluorinated Carboxylates from the European Continent. Environmental Science & Technology, Volume 41, pp. 7260-7265.

Naturvårdsverket, 2009. Riktvärden för förorenad mark: Modellbeskrivning och vägledning, Bromma: Naturvårdsverket.

Niu, J. et al., 2012. Electrochemical Mineralization of Perfluorocarboxylic Acids (PFCAs) by Ce-Doped Modified Porous Nanocrystalline PbO2 Film Electrode. Environmental Science and Technology, Volume 46, pp. 10191-10198.

Noorlander, C. W. et al., 2011. Levels of Perfluorinated Compounds in Food and Dietary Intake of PFOS and PFOA in The Netherlands. Journal of Agricultural and Food Chemistry, Volume 59, p. 7496–7505.

O'Hagan, D., 2008. Understanding organofluorine chemistry. An introduction to the C–F bond. Chemical Society Reviews, pp. 308-319.

Olsen, G. W. & Zobel, L. R., 2007. Assessment of lipid, hepatic, and thyroid parameters with serum perfluorooctanoate (PFOA) concentrations in fluorochemical production workers. International Archives of Occupational and Environmental Health, 81(2), pp. 231-246.

Organisation for Economic Co-operation and Development, 2009. OECD guidelines for the testing of chemicals, s.l.: s.n.

27

Organisation for Economic Co-operation and Development, 2012. Portal on Per and Poly Fluorinated Chemicals. [Online] Available at: http://www.oecd.org/chemicalsafety/portal-perfluorinated-chemicals/ [Accessed 2017].

Ottosen, L. M., Hanseen, H. K., Laursen, S. & Villumsen, A., 1997. Electrodialytic Remediation of Soil Polluted with Copper from Wood Preservation Industry. Environmental Science and Technology, Volume 31, pp. 1711-1715.

Ottosen, L. M., Hansen, H. K. & Hansen, C. B., 2000. Water splitting at ion-exchange membranes and potential differences in soil during electrodialytic soil remediation. Journal of Applied Electrochemistry, Volume 30, p. 1199±1207.

Pedersen, K. B., Jensen, P. E., Ottosen, L. M. & Lejon, T., 2015. An optimised method for electrodialytic removal of heavy metals from harbour sediments. Electrochimica Acta, Volume 173, pp. 432-439.

Prevedouros, K., Cousins, I. T., Buck, R. C. & Korzeniowski, S. H., 2006. Sources, Fate and Transport of Perfluorocarboxylates. Environmental Science & Technology, 40(1).

Rayne, S. & Forest, K., 2009. Perfluoroalkyl sulfonic and carboxylic acids: A critical review of physicochemical properties, levels and patterns in waters and wastewaters, and treatment methods. Journal of Environmental Science and Health, Part A, 44(12), pp. 1145-1199.

Reuss, F. F., 1809. Notice sur un nouvel effet de l'électricité galvanique. s.l.:s.n.

Ronneby Kommun, 2018. Frågor och svar om PFAS. [Online] Available at: https://www.ronneby.se/bygga-bo--miljo/vatten-och-avlopp/pfas-information/fragor-och-svar-om-pfas.html [Accessed 2018].

Schaefer, C.E., Andaya, C., Burant, A., Condee, C.W., Urtiaga, A., Strathmann, T.J., Higgins, C.P., 2017. Electrochemical treatment of perfluorooctanoic acid and perfluorooctane sulfonate: Insights into mechanisms and application to groundwater treatment. Chemical Engineering Journal 317, 424–432. https://doi.org/10.1016/j.cej.2017.02.107

Schedin, E., 2013. Effect of organic carbon, black carbon, calcium ions and aging on the sorption of per- and polyfluoroalkylated substances (PFASs) to soil., Uppsala: SLU.

Scott, B. F., Spncer, C., Mabury, S. A. & Muir, D. C., 2006. Poly and Perfluorinated Carboxylates in North American Precipitation. Environ. Sci. Technol.2006, 40(23), pp. 7167-7174.

Simcik, M. F. & Dorweiler, K. J., 2005. Ratio of Perfluorochemical Concentrations as a Tracer of Atmospheric Deposition to Surface Waters. Environmental Science & Technology, Volume 39, pp. 8678-8683.

Smart, B. E., 1994. Characteristics of C-F systems. In: Organofluorine chemistry: Principles and commercial applications. New York: s.n., pp. 57-88.

Swedish Geotechnical Institute, 2015. Preliminära riktvärden för högfluorerade ämnen (PFAS) i mark och grundvatten, Linköping: Statens geotekniska institut.

Trautmann, A.M., Schell, H., Schmidt, K.R., Mangold, K.-M., Tiehm, A., 2015. Electrochemical degradation of perfluoroalkyl and polyfluoroalkyl substances (PFASs) in groundwater. Water Sci. Technol. 71, 1569–1575. https://doi.org/10.2166/wst.2015.143

UNEP, 2009. Governments unite to step-up reduction on global DDT reliance and add nine new chemicals under international treaty. [Online] Available at: http://chm.pops.int/Convention/Pressrelease/COP4Geneva9May2009/tabid/542/languages/en-US/Default.aspx [Accessed 11 2017].

28

US EPA, 2000. EPA and 3M anounce phase out of PFOS. [Online] Available at: https://yosemite.epa.gov/opa/admpress.nsf/0/33aa946e6cb11f35852568e1005246b4 [Accessed 11 2017].

USEPA, 2014. Emerging Contaminants – Perfluorooctane Sulfonate (PFOS) and Perfluorooctanoic Acid (PFOA), s.l.: USEPA.

Wang, Z., Cousins, I. T., Scheringer, M. & Hungerbühler, K., 2013. Fluorinated alternatives to long-chain perfluoroalkyl carboxylic acids (PFCAs), perfluoroalkane sulfonic acids (PFSAs) and their potential precursors. Environment International, pp. 242-248.

Vierke, L. et al., 2012. Perfluorooctanoic acid (PFOA) — main concerns and regulatory developments in Europe from an environmental point of view. Environmental Sciences Europe, 24(12).

Virkutyte, J., Sillanpää, M. & Latostenmaa, P., 2002. Electrokinetic soil remediation critical overview. The Science of the Total Environment, Volume 289, pp. 97-121.

Xenia, T., Granby, K. & Christensen, J. H., 2011. Tools to discover anionic and nonionic polyfluorinated alkyl surfactants by liquid chromatography electrospray ionisation mass spectrometry. Journal of Chromatography A, pp. 7094-7104.

Yasuda, H., Okuno, T., Sawa, Y. & Yasuda, T., 1995. Surface Configuration Change Observed for Teflon-PFA on Immersion in Water. Langmuir, pp. 3255-3260.

Åkerblom, S. et al., 2017. Variation and accumulation patterns of poly- and perfluoroalkyl substances (PFAS) in European perch (Perca fluviatilis) across a gradient of pristine Swedish lakes. Science of The Total Environment, 599-600(1), pp. 1685-1692

29

Appendix A Materials and Equipment

Appendix A. Table 1. Lab equipment used. Equipment Manufacturer Electrodialytic experiments - Platinum-coated titanium rod electrodes - - Power supply kit E3612A - Peristaltic pump - Neoprene tubing - EDR cells (plastic) - Anion exchange membranes (204 SZRA B02249C) - Cation exchange membranes (CR67HUY N12116B) - pH meter, pH 1000H - Multimeter UT33D - Silicon grease - Silicon plugs - Polypropylene (PP) tubing

DTU, department of civil engineering, Denmark Hewlett Packard MasterFlex® MasterFlex® DTU, department of civil engineering, Denmark Ionics MKIII France Ionics MKIII France VWR International, USA UNI-Trend Technology, China - - -

Soil extractions - 50 mL PP centrifuge tubes - 15 mL PP tubes - 1.5 ml PP microcentrifuge tubes - Syringes 5 mL

- Syringe filters (0.45 μm) - Wrist-action shaker - End-over-end shaker - Nitrogen Evaporator N-EVAP 112

Corning Corning Scantec Nordic Norm-Ject®, Henke Sass Wolf, Germany VWR International, VWR® syringe filters, 0.45μm pore size, Nylon membrane, USA Gerhardt Heidolph Organimation Associates, USA

General equipment - Ultrasonic bath Branson - Parafilm Bemis, Parafilm M®, USA PFAS Analysis - UHPLC-MS/MS Quantiva TSQ - Chromatography column Acquity UPLC BEH-C18 Waters,

100 mm × 2.1 i.d., 1.7 µm particle size - 2 mL glass vials - Vial caps

Thermo Fisher Scientific, US Waters Corporation, Manchester, UK Agilent Technologies, Poland Agilent Technologies, USA

30

Appendix A. Table 2. Compounds used for spiked soil. Compound Abbreviation Purity Manufacturer Potassium heptadecafluorooctanesulfonate

PFOS-K 98% Sigma-Aldrich Sweden AB

Perfluorooctanoic acid PFOA 96% Sigma-Aldrich Sweden AB

Perfluoropentanoic acid PFPeA 97% Sigma-Aldrich Sweden AB

Perfluorohexanoic acid PFHxA 97% Sigma-Aldrich Sweden AB

Perfluoroheptanoic acid PFHpA 99% Sigma-Aldrich Sweden AB

Perfluorononanoic acid PFNA 97% Sigma-Aldrich Sweden AB

Perfluorodecanoic acid PFDA 98% Sigma-Aldrich Sweden AB

Perfluoroundecanoic acid PFUnDA 95% Sigma-Aldrich Sweden AB

Perfluorododecanoic acid PFDoDA 95% Sigma-Aldrich Sweden AB

Perfluorooctadecanoic acid PFOcDA 97% Alfa Aesar GmbH & Co.KG

Potassium nonafluoro-1-butanesulfonate

PFBS-K 98% Sigma-Aldrich Sweden AB

Potassium tridecafluorohexane-1-sulfonate

PFHxS-K 98% Sigma-Aldrich Sweden AB

1H,1H,2H,2H-Perfluorodecanesulfonic acid

8:2 FTSA ≤100% Apollo Scientific Ltd

1H,1H,2H,2H-Tridecafluorooctane-1- sulfonic acid

6:2 FTSA 98% Apollo Scientific Ltd

Perfluorooctanesulfonamide FOSA ≤100% Sigma-Aldrich Sweden AB

Appendix A. Table 3. Internal standard mixture contents. Coumpound Analyte description [13C4]PFBA PFBA [13C2]PFHxA PFPeA,PFBS, PFHxA [13C4]PFOA PFOA, PFHpA [13C5]PFNA PFNA [13C2]PFDA PFDA [13C2]PFUnDA PFUnDA [13C2]PFDoDA PFDoDA, PFTeDA, PFHxDA, PFOcDA [18O2]PFHxS PFHxS, 6:2 FTSA [13C4]PFOS PFOS, 8:2 FTSA [13C8]FOSA FOSA d5-N-EtFOSA N-EtFOSA

31

Appendix B

Pictures of experiments

Appendix B. Picture 1. N2 blow down evaporation of extracts

Appendix B. Picture 2. EDR cell for experiment IV. Sedimentation of the soil and intrusion of the electrolyte is visible in

the top of the cell.

32

Appendix B. Picture 3. Dry soil noticed at the end of the experiment IV, towards the anode.

33

Appendix C

PFAS concentrations & Soil characteristics

Appendix C. Table 1. PFAS concentrations for soil samples before remediation. Concentrations are expressed in μg kg-1 of soil. Soil Origin PFBA PFPeA PFHxA PFHpA PFOA PFNA PFDA PFUnDA PFDoDA PFTriDA PFTeDA PFHxDA PFOcDA ΣPFCAs

Arlanda SD

<1.31 0.03

0.73 0.36

0.45 0.00

0.33 0.00

0.49 0.03

<0.52 0.00

<0.10 0.01

<0.05 0.00

<0.10 0.00

<0.05 0.00

<0.45 0.00

<0.50 0.01

1.09 0.05

3.08

Kallinge SD 1.96

0.90 7.93 0.18

4.57 0.24

2.25 0.18

1.13 0.26

1.19 0.36

0.21 0.06

0.15 0.03

0.17 0.07

<0.05 0.04

<0.45 0.31

<0.5 0.35

1.00 0.58

20.5

Spiked soil SD

<1.31 0.00

18.33 1.12

16.39 1.15

20.12 1.12

20.69 1.39

19.16 1.79

21.86 2.74

17.77 0.80

27.85 4.41

<0.05 0.00

<0.45 0.00

<0.50 0.00

<1.00 0.05

163.48

MDL 1.31 0.60 0.39 0.05 0.05 0.52 0.10 0.05 0.10 0.05 0.45 0.50 1.00

Soil Origin PFBS PFHxS PFOS PFDS ΣPFSAs FOSA MeFOSA EtFOSA ΣFOSAs MeFOSE EtFOSE ΣFOSEs

Arlanda 0.12 1.14 9.66 <0.05 10.92 0.22 <0.05 0.07 0.29 <0.5 <0.05 0.00

SD 0.01 0.18 0.87 0.00 0.00 0.00 0.00 0.00 0.00

Kallinge 0.28 1.32 9.37 0.02 10.98 0.13 <0.05 0.08 0.21 0.23 <0.05 0.23

SD 0.04 0.03 0.69 0.02 0.00 0.00 0.02 0.01 0.00

Spiked soil 23.73 19.36 24.98 <0.05 68.07 24.11 <0.05 <0.17 24.11 <0.5 <0.05 0.00

SD 1.27 4.83 4.03 0.00 4.13 0.00 0.00 0.00 0.00

MDL 0.16 0.05 0.10 0.05 0.24 0.05 0.17 0.50 0.05

34

Soil Origin FOSAA MeFOSAA EtFOSAA ΣFOSAAs 6:2 FTSA 8:2 FTSA 10:2 FTSA ΣFTSAs

Arlanda <0.65 <0.05 <0.01 0.00 0.15 0.20 <0.05 0.35

SD 0.00 0.00 0.00 0.04 0.28 0.00

Kallinge <0.65 <0.05 <0.01 0.00 0.20 2.91 <0.05 3.11

SD 0.00 0.00 0.00 0.02 0.42 0.00

Spiked soil <0.65 <0.05 <0.01 0.00 42.46 39.57 <0.05 82.03

SD 0.00 0.00 0.00 9.20 4.25 0.00

MDL 0.65 0.05 0.01 0.27 0.50 0.05

Appendix C. Table 2. Method detection limits.

Compound PFBA PFPeA PFHxA PFHpA PFOA PFNA PFDA PFUnDA PFDoDA PFTriDA PFTeDA PFHxDA PFOcDA MDL 1.31 0.60 0.39 0.05 0.05 0.52 0.10 0.05 0.10 0.05 0.45 0.50 1.00

Compound PFBS PFHxS B-PFHxS PFOS L-PFOS B-PFOS PFDS FOSA B-FOSA MeFOSA EtFOSA

MDL 0.16 0.05 0.05 0.10 0.10 0.10 0.05 0.24 5.00 0.05 0.17

Compound MeFOSE EtFOSE FOSAA MeFOSAA EtFOSAA 6:2 FTSA 8:2 FTSA 10:2 FTSA MDL 0.5 0.05 0.65 0.05 0.01 0.27 0.50 0.05

Appendix C. Table 3. Recovery rates for analysed compounds. Compound PFBA PFHxA PFOA PFNA PFDA PFUnDA PFDoDA PFHxS Recovery rate (%)

89,92 88,18 88,30 87,64 89,74 95 99,77 84,47

Compound PFOS FOSA MeFOSA EtFOSA MeFOSE EtFOSE MeFOSAA EtFOSAA Recovery rate (%)

88,73 120,69 187,61 221,43 107,17 120,16 109,08 113,34

35

Appendix C. Table 4. Concentration of metals for soil sampled in Arlanda, corresponding to experiments I,II,III. Metal Al 308.215 As 197.198 Ba 493.408 Ca 317.933 Cd 228.802 Cr 206.550 Cu 324.754 Fe 259.940 K 769.897

Concentration [mg/L] 254.88 0.11 0.90 45.67 0.01 0.48 0.19 383.18 53.06

σ 7.69 0.04 0.05 1.19 0.01 0.02 0.03 18.56 2.02 Metal Mg 280.270 Mn 294.921 Na 589.592 Ni 231.604 P 214.914 Pb 283.305 S 182.562 Zn 334.502

Concentration [mg/L] 91.62 6.79 3.53 0.28 3.89 1.01 0.21 0.82

σ 2.35 2.62 0.81 0.03 0.79 0.09 0.22 0.07

Appendix C. Table 5. Texture of the studied soils. Classification Arlanda soil Kallinge soil Artificial soil

Clay (d≤0.002 mm) 59% 5.7% 43% Silt (0.002mm<d<0.06mm) 34% 43% 28%

Sand (0.06mm≤d≤2mm) 7% 51.3% 29%

TRITA TRITA-ABE-MBT-18480

www.kth.se