Technology Selection for Removal of PFAS from Raw Water for Drinking Water Purposes Master’s thesis in the master’s Programme Infrastructure and Environmental Engineering FRIDA HANSSON LISA WU DEPARTMENT OF ARCHITECHTURE AND CIVIL ENGINEERING DIVISION OF GEOLOGY AND GEOTECHNICS CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2024 www.chalmers.se MASTER’S THESIS ACEX30 Technology selection for removal of PFAS from raw water for drinking water purposes Master’s Thesis in the Master’s Programme Infrastructure and Environmental Engineering FRIDA HANSSON LISA WU Department of Architecture and Civil Engineering Division of Geology and Geotechnics CHALMERS UNIVERSITY OF TECHNOLOGY Göteborg, Sweden 2024 I Technology Selection for Removal of PFAS from Raw Water for Drinking Water Purposes Master’s Thesis in the Master’s Programme Infrastructure and Environmental Engineering FRIDA HANSSON LISA WU © FRIDA HANSSON, LISA WU, 2024 Examensarbete ACEX30 Institutionen för arkitektur och samhällsbyggnadsteknik Chalmers tekniska högskola, 2024 Department of Architecture and Civil Engineering Division of Geology and Geotechnics Engineering Geology Chalmers University of Technology SE-412 96 Göteborg Sweden Telephone: + 46 (0)31-772 1000 Cover: Schematic decision support flowchart depicting how the technologies are affected by water chemistry with support from sustainability performance analysis. II Department of Architecture and Civil Engineering Göteborg, Sweden, 2024 I Technology Selection for Removal of PFAS from Raw Water for Drinking Water Purposes Master’s thesis in the Master’s Programme Infrastructure and Environmental Engineering FRIDA HANSSON LISA WU Department of Architecture and Civil Engineering Division of Geology and Geotechnics Engineering Geology Chalmers University of Technology ABSTRACT Per- and polyfluoroalkyl substances (PFAS) are a synthetic group of chemicals that can be harmful for humans and the environment. The use of PFAS has caused the presence of it in drinking water. Two thirds of the groundwater sources in Sweden have found to be contaminated with PFAS. To limit PFAS in drinking water, the European Food Safety Authority, EFSA, has introduced a guideline of 4 ng/L for PFAS 4 and 100 ng/L for PFAS 21 which will be implemented 2026. This study investigates technical and sustainability performance of the technologies Granular Activated Carbon (GAC), Nanofiltration (NF), Ion Exchange (IX) and Foam Fractionation (FF). The study has been performed through a literature study with help from databases such as Web of Science using search strings. The results were presented in a flowchart and a table with results from the sustainability performance analysis. Granular activated carbon and IX alone or in combination with each other or other technologies were to be suggested in the majority of the cases. Nanofiltration showed to have high performance in many aspects, including for short-chained PFAS, with a disadvantage of highly concentrated waste stream that needs further treatment. Foam fractionation needs specific conditions for proper performance but is not suitable for drinking water treatment purposes. Key words: PFAS, nanofiltration, granular activated carbon, foam fractionation, ion exchange, operation and maintenance cost, decision-support, drinking water, water treatment. II Val av reningsteknik för borttagning av PFAS från råvatten för dricksvattenproduktion Examensarbete inom mastersprogrammet infrastruktur och miljöteknik FRIDA HANSSON LISA WU Institutionen för arkitektur och samhällsbyggnadsteknik Avdelningen för geologi och geoteknik Teknisk geologi Chalmers tekniska högskola SAMMANFATTNING Per- och polyfluoralkylämnen (PFAS) är en syntetisk grupp av kemikalier som kan vara skadliga för människor och miljön. Användningen av PFAS har orsakat att det finns i dricksvatten. Två tredjedelar av grundvattenkällorna i Sverige har visat sig vara förorenade med PFAS. För att begränsa PFAS i dricksvatten har Europeiska myndigheten för livsmedelssäkerhet, EFSA, infört en riktlinje på 4 ng/L för PFAS 4 och 100 ng/L för PFAS 21, som kommer att implementeras 2026. Denna studie undersöker teknisk prestanda, såväl som ekologisk och ekonomisk hållbarhet för teknologierna granulärt aktivt kol (GAC), nanofiltrering (NF), jonbytare (IX) och skumfraktionering (FF). Studien har utförts genom en litteraturstudie med hjälp av databaser som Web of Science där söksträngar har använts. Resultaten presenterades i ett flödesschema och en tabell med resultat från hållbarhetsprestandaanalysen. Granulärt aktivt kol och IX, ensamma eller i kombination med varandra eller andra teknologier, föreslogs i majoriteten av fallen. Nanofiltrering visade hög prestanda i många aspekter, inklusive för kortkedjiga PFAS, med en nackdel av en högkoncentrerad avfallsström som behöver ytterligare behandling. Skumfraktionering behöver specifika förhållanden för att fungera korrekt, men är inte lämplig för dricksvattenreningsändamål. Nyckelord: PFAS, nanofiltrering, granulärt aktivt kol, skumfraktionering, jonbytare, underhållskostander, beslutsstöd, dricksvatten, vattenrening. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 III Contents 1 Introduction .......................................................................................................... 1 1.1 Background..............................................................................................................1 1.2 Aim and Objectives .................................................................................................2 1.3 Limitations ...............................................................................................................2 2 Theory ................................................................................................................... 4 2.1 PFAS Compounds ...................................................................................................4 2.2 PFAS in Swedish Water Resources .......................................................................6 2.3 Drinking Water Treatment ....................................................................................8 2.4 Technologies for PFAS Removal ...........................................................................8 2.4.1 Granular Activated Carbon ............................................................................................... 9 2.4.2 Nanofiltration .................................................................................................................. 10 2.4.3 Ion Exchange .................................................................................................................. 11 2.4.4 Foam Fractionation ......................................................................................................... 12 3 Methodology ....................................................................................................... 14 3.1 Collection of Information Regarding Technology Performance ......................16 3.2 Sustainability Analysis ..........................................................................................16 3.3 Compilation of Results for Decision Support .....................................................17 4 Results ................................................................................................................. 19 4.1 Technology Performance Based on Water Chemistry .......................................19 4.1.1 Granular Activated Carbon ............................................................................................. 19 4.1.2 Nanofiltration .................................................................................................................. 21 4.1.3 Ion Exchange .................................................................................................................. 23 4.1.4 Foam Fractionation ......................................................................................................... 25 4.2 Sustainability Performance ..................................................................................26 4.2.1 Granular Activated Carbon ............................................................................................. 27 4.2.2 Nanofiltration .................................................................................................................. 30 4.2.3 Ion Exchange .................................................................................................................. 33 4.2.4 Foam Fractionation ......................................................................................................... 35 4.3 Decision Support for PFAS Removal Technology Selection .............................36 5 Discussion ............................................................................................................ 38 5.1 Technology Performance ......................................................................................38 5.2 Sustainability Performance ..................................................................................40 5.3 Case Application of Results ..................................................................................42 5.3.1 Case A ............................................................................................................................ 42 5.3.2 Case B ............................................................................................................................. 43 5.3.3 Case C ............................................................................................................................. 43 5.4 The Effects of Subjectivity....................................................................................44 6 Conclusions ......................................................................................................... 46 7 Recommendations for Further Work .............................................................. 47 CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 IV References ................................................................................................................... 48 Appendices .................................................................................................................. 55 Appendix A – Water Chemistry and Setup Values of the Studies .................................55 Appendix B – Summary of Motivations for the Scoring of the Sustainability Analysis ............................................................................................ Fel! Bokmärket är inte definierat. Appendix C – Flowcharts for Individual Technologies ..................................................63 CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 V Preface In this report, a literature study was conducted through January 2024 and June 2024. The project has been conducted at the Department of Architecture and Civil Engineering, Engineering Geology, at Chalmers University of Technology, Sweden in collaboration with Norconsult AB. The project has been carried out with Jenny Norrman as supervisor and Andreas Lindhe as examiner from Chalmers University of Technology, and Stephan Köhler as supervisor from Norconsult AB. We would like to thank Stephan Köhler for his expertise, encouragement and enthusiasm towards the work and the field, Jenny Norrman for an exceptional guidance throughout the thesis and together with Andreas Lindhe for the advice we have received. We would also like to thank Emma Johansson from Laholmsbuktens AB and Ludwig Hedberg from Norconsult AB for providing us with much needed information in the field. At last, we would like to thank Chalmers University of Technology and Norconsult AB for making it possible for us to carry out this thesis. Göteborg, June 2024 Frida Hansson & Lisa Wu CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 VI Abbreviations GAC – Granular Activated Carbon NF – Nanofiltration IE – Ion Exchange FF – Foam Fractionation PFAS - Per- and polyfluoroalkyl substances PFCA - perfluoroalkyl carboxylic acids PFSA - perfluoroalkane sulphonic acids PFOS - Perfluorooctanesulphonic acid PFOA - Perfluorooctanoic acid PFHxS - Perfluorohexanesulphonic acid PFNA - Perfluorononanoic acid PFDA - Perfluorodecanoic acid PFUnDA - Perfluoroundecanoic acid PFDoDA - Perfluorododecanoic acid PFTrDA - Perfluorotridecanoic acid 6:2 FTS - 6:2-fluorotelomersulfonic acid PFBS - Perfluorobutanesulphonic acid PFBA - Perfluorobutanoic acid PFPeS - Perfluoropentanesulphonic acid PFPeA - Perfluoropentanoic acid PFHxA - Perfluorohexanoic acid PFHpA - Perfluoroheptanoic acid PFHpS - Perfluoroheptanesulphonic acid PFNS - perfluorononane sulphonic acid PFDS - perfluorodecane sulphonic acid PFUnDS - perfluoroundecane sulphonic acid PFDoDS - perfluorododecane sulphonic acid PFTrDS - perfluorotridecane sulphonic acid TOC - Total organic carbon DOC – Dissolved organic carbon NOM – Natural organic matter DOM – Dissolved organic matter AFFF – Aqueous Film Forming Foam CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 VII CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 1 1 Introduction In this chapter an introduction to the subject will be presented followed by the aims and objectives with this thesis. Lastly, limitations of the study will be discussed. 1.1 Background Per- and polyfluoroalkyl substances, PFAS, are chemicals that have been used since the 1930s and are still commonly used today [1]. PFAS consists of a large and complex group of fluorinated substances that are often used in coatings for products made to resist water, heat, and oil [2]. The substances are synthetic chemicals that can be harmful to both humans and the environment as they accumulate over time. Additionally, PFAS do not break down in the environment and are readily bound to soils and can eventually reach groundwater sources. Due to its persistence in nature, occurrence of bioaccumulation in fish and wildlife is also a possibility. PFAS are PBT substances (Persistent, Bioaccumulative and Toxic). Due to PFAS being used in various products and being a PBT substance, most people have PFAS in their bodies [3]. High levels of PFAS can be found in food as well as drinking water and indoor environments. When it comes to PFAS in drinking water, the water has +most likely originated from groundwater with high contamination levels. Two thirds of the groundwater resources in Sweden have been found to be contaminated with PFAS at various levels, both high and low [4]. The health effects of PFAS are uncertain but studies have shown that PFAS could have a negative impact on human health [5]. Possible risks include decreased fertility in women, increased risks of prostate cancer and kidney cancer, interference with the body’s natural hormones as well as increased risk of obesity and increased cholesterol levels among others [6]. To limit PFAS in drinking water, the European Food Safety Authority, EFSA, has introduced new, stricter guideline values (4 ng/L for PFAS 4 and 100 ng/L PFAS 21) which will be implemented in 2026 [7]. As the guidelines have been introduced there is a need for adapting and improving the drinking water treatment process. Today, there are various technologies available for removing PFAS substances from raw water. These include Granular Activated Carbon (GAC), Ion Exchange (IX), membrane treatment using nanofiltration (NF) and foam fractionation (FF). There are several factors that could influence the choice of the optimal treatment, e.g. the chemical composition of the raw water, technology performance and limitations, energy requirements as well as waste stream management. Granular activated carbon has long been used to eliminate unwanted substances in raw water. This process is easy to implement and well-known. Ion exchange is a treatment that has been used for many years within industrial applications and is based on removing unwanted ions from the water [8]. For GAC and IX, water chemistry affects how long the treatment operates before the material needs replacement. Nanofiltration is a membrane treatment using pressure to separate soluble ions from the water [9]. For NF, the removal of PFAS substances is affected by the size of the membrane pores, which in turn impacts the necessary pressure and thus the energy required. Foam fractionation can also be utilized to separate PFAS where water is subjected to air bubbles that bind PFAS, and thus separated from the water phase [10]. Foam fractionation is very compact, but has a high energy demand. For PFAS substances with lower molecular weight, all the above technologies have a lower performance. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 2 Knowledge regarding the limitations of different technologies is not readily available for water treatment plants and choosing an optimal treatment process can be difficult. At the same time, the water sector has been urged to develop and implement more sustainable treatment processes and many new technologies and developments are emerging. Therefore, there is a need for easily accessed information regarding the performance of a technology and its sustainability. When evaluating potential solutions, it is important to review multiple aspects of sustainability, such as technical, environmental, social, and economic, to gain extensive knowledge regarding the long- term sustainability of a certain treatment process. 1.2 Aim and Objectives The overall aim of this study was to develop a decision support tool for selection of the optimal PFAS-treatment technology for drinking water. The literature search results were limited to two main results: a flowchart suggesting which water treatment process is the most suitable considering the raw water chemistry and the results from the sustainability analysis showing the performance of the four water treatment processes according to the economic and environmental aspects. The flowchart and the sustainability performance analysis together build a decision support for development and building of future as well as retrofitting of existing water treatment plants. The specific objectives were to: • Collect information regarding the performance of the technologies based on water chemistry. • Compare performance and disruptions during drinking water treatment and rank the various processes for removing PFAS from different raw water types. • Compile and present the collected information in a flowchart. • Quantify energy consumption, costs and environmental impact of the PFAS removal technologies. • Collect of information regarding carbon footprints for the treatment technologies currently used for PFAS removal. • Compile and present the sustainability performance of the processes for removing PFAS using a sustainability performance analysis. 1.3 Limitations The project was limited to four water treatment processes: granular active carbon (GAC), ion exchange (IX), nanofiltration (NF) and foam fractionation (FF). Limitations were set to focus on pilot-scale or full-scale studies and laboratory scale experiments that use small-scale setups were excluded. The research did not include results with simulated samples such as spiking the water with excessive PFAS content. During unavailability of information from larger scale experiments, smaller scales were considered. Furthermore, the research focuses on studies with experiments conducted during longer time periods when available. These limitations were set to ensure that the research was as true to reality as possible with sufficient volumes, flows and time to resemble what potential implementations of the technologies in a treatment process would imply. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 3 Limitations of the study included the time frame of the research. During the study, many articles had to be excluded from the research due to the time limit. There was also a limitation of data, where the different technologies rarely had the same water characteristics to create a fair comparison of the technology performances. Another limitation of data was for environmental impact, as no data of could be for nanofiltration and limited data could be found for ion exchange. During contact with external contacts, there had been no calculations on the carbon footprint for the two technologies in question. Another limitation was the inability to have performed the experiments in full-scale for all technologies. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 4 2 Theory This chapter includes theoretical information regarding PFAS compounds, PFAS in Swedish water sources, descriptions of drinking water treatment plants and basic knowledge about the four technologies. 2.1 PFAS Compounds There are thousands of different PFAS compounds [2]. A PFAS compound consists of a fluorinated carbon chain. Where a regular carbon chain would include hydrogen atoms, a fluorinated carbon chain includes fluorine [3]. The compounds can in turn be either perfluorinated compounds or polyfluorinated compounds. Perfluorinated compounds are fully fluorinated carbon chains while polyfluorinated carbon chains are partially fluorinated. The many different PFAS compounds can be divided into two groups, which are polymers or non-polymers [11]. The difference between polymeric and non-polymeric PFAS is that the non-polymeric PFAS are smaller than the polymeric PFAS and therefore more readily adsorbed by humans, animals and in the environment [12]. Polymeric PFAS can therefore be considered safer than non-polymeric PFAS, with some exceptions. Polymeric PFAS can break down and be degraded in the environment or in manufacturing processes which releases non-polymeric PFAS. The polymeric and non-polymeric PFAS are in turn divided into more groups. Polymers include fluoropolymers, side-chain fluorinated polymers and perfluoropolyethers [11]. Non-polymers include perfluoroalkyl acids (PFAA), perfluoroalkane sulphonyl fluorides, fluorotelomers and per- and polyfluoroalkyl ethers. The non-polymeric group PFAA includes, among others, perfluoroalkyl carboxylic acids (PFCA) and perfluoroalkane sulphonic acids (PFSA). A general classification of the PFAS compounds can be seen in Figure 2.1. Figure 2.1: A general classification of PFAS. The compounds of PFCA and PFSA contain functional groups that are bound at the end of the carbon chain [3]. These functional groups can further alter the characteristics of the PFAS compound. The PFCA and PFSA compounds contain a functional group at CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 5 the end of the carbon chain. In the case of PFCA they contain carboxylic acid as a functional group while PFSA contain sulphonic acid as a functional group. These two groups contain the most common PFAS compounds which are PFOA, with a carboxylic acid functional group, and PFOS, with a sulphonic acid functional group [13]. Other than PFOA and PFOS there are plenty PFAS compounds which are of concern in the various divisions of PFAS compounds. The compounds can be classified based on the priority to remove them from the environment [14]. The group PFAS 4 consists of the four most toxic and bioavailable compounds and are the most prioritised to remove from the drinking water. Furthermore, there are groups for PFAS 11 and PFAS 21 which consist of the 11 and 21 most prioritised compounds respectively. The 21 PFAS compounds and their respective groups and properties can be seen in Table 2.1: The PFAS compounds included in PFAS 4, PFAS 11 and PFAS 21 and their respective molecular formula, weight, and functional group [16]. PFAS 4 is also included in PFAS 11 and PFAS 21. PFAS 11 is also included in PFAS The new guideline values that will be implemented in 2026 limits discharge of PFAS 4 to 4 ng/L for and 100 ng/L for PFAS 21 [7]. Further recommendations made by the Swedish Food Agency limits PFAS 11 to 90 ng/L [15]. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 6 Table 2.1: The PFAS compounds included in PFAS 4, PFAS 11 and PFAS 21 and their respective molecular formula, weight, and functional group [16]. PFAS 4 is also included in PFAS 11 and PFAS 21. PFAS 11 is also included in PFAS 21. Compound Molecular formula Molecular weight [g/mol] Functional group Classification PFOS C8HF17O3S 500.13 PFSA PFAS 4 PFAS 11 PFAS 21 PFOA C8HF15O2 414.07 PFCA PFNA C9HF17O2 464.08 PFCA PFHxS C6HF13O3S 400.11 PFSA PFBS C4HF9O3S 300.10 PFSA PFHxA C6HF11O2 314.054 PFCA PFPeA C5HF9O2 264.05 PFCA PFHpA C7HF13O2 364.06 PFCA PFDA C10HF19O2 514.086 PFCA 6:2 FTS C8H5F13O3S 428.16 PFSA PFBA C4HF7O2 214.039 PFCA PFUnDA C11HF21O2 564.09 PFCA PFDoDA C12HF23O2 614.10 PFCA PFTrDA C13HF25O2 664.10 PFCA PFPS C5HF11O3S 350.11 PFSA PFHpS C7HF15O3S 450.12 PFSA PFNS C9HF19O3S 550.14 PFSA PFDS C10HF21O3S 600.15 PFSA PFUnDS C11HF23O3S 650.15 PFSA PFDoDS C12HF25O3S 699.15 PFSA PFTrDS C13HF27O3S 750.17 PFSA 2.2 PFAS in Swedish Water Resources As PFAS are not degraded naturally, and they can persist for a very long time in the environment [2]. The PFAS compounds are readily bound to soils. Long-chain PFAS, tends to remain bound to the soil for longer than short-chain PFAS as shorter chains have a higher water solubility and are therefore more prone to transportation to various groundwater or water bodies. Compound bound to soils can eventually leach out of the soils and reach groundwater. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 7 About 250 groundwater resources in Sweden are at risk of being contaminated with PFAS [4]. When the municipal groundwater resources were investigated in 2016-2017, PFAS was found in two thirds of the groundwater resources [4]. However, the average level of contamination was low, with certain places having higher levels. These high levels of PFAS can mostly be found near sites where firefighting training has occurred. The sites are mostly civil and military airports where they have reported using Aqueous Film Forming Foam (AFFF) containing PFOS as firefighting foam [17]. Furthermore, most sites were reportedly often situated where the ground largely consisted of gravel, and as gravel is a highly permeable material, the foam was likely transported further into the soil and eventually leaching into the groundwater. It has also been shown that there is a correlation between PFAS contamination found in groundwater and landfill sites located near the groundwater resource [4]. There are various municipalities or drinking water producers in Sweden where PFAS has at some point been detected in the raw water source. According to the Swedish Food Agency, 48 municipalities had at some point detected PFAS levels >10 ng/L [1]. Furthermore, 108 municipalities had at some point detected PFAS levels <10 ng/L. In the raw water while 82 municipalities had not detected any PFAS contamination. There are also several municipalities or drinking water producers where the PFAS contamination detected in the drinking water after treatment had at some point been higher than 10 ng/L. 15 municipalities detected PFAS levels >10 ng/L in their drinking water at some point [1]. 59 municipalities had detected PFAS levels <10 ng/L in their drinking water while 80 municipalities had not detected any PFAS in their drinking water. As mentioned earlier, the new guideline values of 4 ng/L for PFAS 4 and 100 ng/L for PFAS 21 limits the allowed PFAS discharge from the drinking water treatment plants and leads to a need for further treatment in many treatment plants. As can be seen in the cases above, there are various instances where removal of PFAS could be needed, depending on what PFAS compounds are present in the waters, as the PFAS levels in both raw waters and drinking waters have been measured to be over the impending 4 ng/L guideline value for PFAS 4. Uppsala is a city where the PFAS levels in the raw water that is used to produce drinking water have been found to be high. The high levels have been caused by firefighting foams used at an airport belonging to the Swedish Armed Forces which was located upstream of the groundwater source used for drinking water in Uppsala [18]. The drinking water in Uppsala is currently being treated to remove PFAS with GAC and reduces PFAS from around 150 ng/L to 5-10 ng/L [19]. Uppsala Vatten has sued the Swedish Armed Forces and has asked that they take responsibility of the costs of removing the PFAS contamination in the drinking water as they were the ones who caused the contamination. The legal case is still ongoing, with the latest occurrence being that the Land and Environment Court of Appeal declared the Swedish Armed Forces responsible for the costs to which the Swedish Armed Forces appealed in May 2024 [20]. Another municipality that has been affected by high concentrations of PFAS is Ronneby and more specifically the village of Kallinge where Swedish Armed Forces has used firefighting foam at an airport similarly to the case in Uppsala [19]. However, the PFAS concentrations found in the drinking water in Kallinge were significantly higher, reaching 10 380 ng/L. This led to over 150 people residing in Ronneby suing the municipality for personal injury as the PFAS-levels in their blood were heightened. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 8 In late 2023, 10 years after the high PFAS contamination was discovered, the supreme court declared that the residents would be paid in damages [21]. Shortly after this happened in 2013, the Swedish Food Agency encouraged all municipalities to check their raw water sources for contamination of PFAS [22]. 2.3 Drinking Water Treatment Treatment of raw water for drinking water purposes can consist of many different treatment processes and combinations of technologies. The treatment process depends on the chemistry of the water to be treated. The conventional treatment process includes coagulation and flocculation, sedimentation, filtration, and disinfection [23]. The purpose of this treatment process is to form larger particles of particulate as well as dissolved matter which will be heavy enough to sediment and separate from the water. This is often followed by filtration of smaller organic and inorganic particles and microorganisms. Lastly, disinfection is used to ensure sufficient inactivation of microorganisms such as bacteria or viruses. In some cases, granular activated carbon is used with the purpose of controlling taste and odour [21]. The treatment process can differ depending on the raw water used as a water resource. Surface water contains different types and levels of contaminants than groundwater, and the different water types will generally require different treatment processes to achieve sufficiently clean water. Surface water typically contains higher levels of microbes, organic carbon and suspended solids which will have been removed through natural infiltration for groundwater. It can be assumed that if a treatment step for removal of PFAS were to be implemented at a drinking water treatment plant, a sufficiently functioning treatment process based on current guideline values is already in place. Adding a treatment step for PFAS removal would be implemented at the end of the treatment process. Lackarebäck’s drinking water treatment plant can be used as an example of a how a typical process train in a drinking water treatment plant without PFAS treatment looks. Lackarebäck’s drinking water treatment plant provides the municipality of Gothenburg with drinking water. At Lackarebäck surface water from Göta Älv is used as raw water source [24]. The treatment steps at Lackarebäck are flocculation with aluminum salts which creates flocs of organic matter, other particles, and microorganisms. This is followed by sedimentation where the flocs sink to the bottom of the tanks and are separated from the water through scraping the tank floor. The next step at Lackarebäck is granular activated carbon, with an approximate EBCT of 15-20 minutes, which removes smaller particles as well as taste and odour [25]. Lastly the water is disinfected by ultrafiltration where microorganisms are removed from the water. This treatment process provides safe drinking water for the population of Gothenburg where the PFAS 4 levels are low and are expected to stay below the limit of 4 ng/L once the limitation is implemented in 2026 [26]. 2.4 Technologies for PFAS Removal The following section presents the background for the four PFAS removal technologies: Granular Activated Carbon (GAC), Nanofiltration (NF), Ion Exchange (IE), and Foam Fractionation (FF). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 9 2.4.1 Granular Activated Carbon Granular activated carbon (GAC) can be made from various materials, such as coconut shell, bituminous coal, petroleum coke, wood, and peat [27]. It has the ability to adsorb organic compounds as well as undesirable tastes and odours among others [28]. This is due to the porosity of the GAC[29]. The pores of the GAC make it possible to have a large surface area, which increases the surface area where the carbon can absorb particles [30]. The pores have the capability to adsorb different molecules as well as PFAS from the water during water treatment. The mechanism of the sorption of PFAS by GAC is mainly caused by the hydrophobic interactions between the PFAS compound and the surface of the activated carbon [31]. Granular activated carbon has the capability of absorbing non-water-soluble organic substances. Granular activated carbon is most commonly designed using a fixed-bed set up [31]. The filters of GAC typically consist of configuration with multiple filters where the first filter removes the bulk of the PFAS and the second filter is used for refining. When reaching breakthrough, the filters are moved so that the second filter becomes the first and a new layer with GAC is added for the refinement. There are several aspects that need to be taken into consideration in the design of a GAC system. These include the pre-treatment, such as sandfilters, where hydrophobic compounds otherwise can directly interfere with the PFAS sorption [31]. Other aspects include the empty bed contact time (EBCT) and bed lifetime. When it comes to waste, GAC can be incinerated or regenerated. The regeneration can happen through multiple ways, where it is mostly done through thermal reactivation [32]. This is done through heating the activated carbon, causing 75-90% of the absorbed content to be volatilized. A steam is thereafter injected and reactivates the carbon through removing the remaining volatiles. The exhaust steam needs to be cleaned from hydrofluoric acid to avoid environmental problems in the surroundings [25]. The regeneration of activated carbon commonly causes a loss of up to 10% that needs to be replaced with new activated carbon to reach the same efficiency. Only certain types of GAC have the capability of being regenerated, but by reactivating GAC, the efficiency is of similar level as a virgin GAC while reducing the carbon footprint that normally is generated during production [31]. A simplified figure of granular activated carbon is presented in Figure 2.2. The necessary contact time (EBCT) for removal of PFAS lies between 10-20 minutes [25]. Sometimes a series of two columns is used where the first (“LEAD”) removes the majority of the PFAS and the latter (“LAG”) assures polishing of PFAS. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 10 Figure 2.2: Simplified figure of granular activated carbon. 2.4.2 Nanofiltration Nanofiltration (NF) is a membrane process driven by pressure [33]. The membranes have the capability to reject multivalent inorganic salts as well as small organic molecules. There are variations of membranes, with dense membranes allowing molecules of the sizes 0.5-1 nm, and less dense typically 1-10 nm [33],[34]. The PFAS rejection can differ depending on membrane characteristics. The membranes of nanofilters are slightly charged when in contact with an aqueous solution. This is due to the dissociation of surface functional groups or adsorption of charge solute [35]. Nanofiltration is considered to have properties in in between ultrafiltration and reverse osmosis [33]. What differs nanofiltration from reverse osmosis, apart from the pore size, is that NF membranes have a low rejection of monovalent ions. Meanwhile the rejection of divalent ions is higher, as well as a higher flux, which is defined by the amount of liquid going through the membrane over time [31], [36]. The process occurring during NF is complex. Within the pores and on the membrane, there are micro-hydrodynamic and interfacial events [36]. The membrane obtains a charge that is dependent on the dissociation of ionizable groups. This can be influenced by specific pH-levels. There is also electrostatic repulsion or attraction that varies dependent on the ion valence, although the concept is not very well understood. Tight nanofiltration membranes separate ions by diffusion gradients while more open membranes also separate larger molecules by size exclusion [25]. What is important to note with nanofiltration is that there is a large waste stream from the separation, with a large amount of water with a high concentration of unwanted particles. This water requires further treatment before going to the wastewater treatment as the level of contaminants are too concentrated [31]. The NF-system typically consists of a feed pump, feeding water to the membrane units, where pre-treatment may be needed to remove particles that may compete with PFAS [31]. Figure 2.3 shows a simplified version of nanofiltration, where permeate water CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 11 enters the nanofiltration process. The technology separates unwanted particles that together with a large amount of water from the treated water that can move forward in the process. Figure 2.3: Simplified figure of nanofiltration. 2.4.3 Ion Exchange Ion exchange is a sorption technology where ions of the same charge are exchanged to remove PFAS from the water [37]. Depending on the type of PFAS targeted, the exchange of molecules typically occurs between either the hydrophobic end of the PFAS molecule (the fluorinated carbon chain), or the hydrophilic end of the PFAS molecule (the functional group), and an ion that is less harmful which is released into the water to adsorb to the PFAS molecule. Due to this, different types of PFAS can behave differently during the Ion Exchange, with the hydrophobic end adsorbing to the resins with hydrophobic and van der Waals interactions, and the functional group absorbs to the resins through electrostatic ionic bonds [38]. Ion exchange consists of solid resin beads that are suspended in the water where the positively charged functional groups in the resin beads bind to the negatively charged PFAS in the water [37]. Ionic bonds are formed between the two different media which leads to sorption of the PFAS onto the resin beads. The reaction that occurs in ion exchange can be written as [39]: 2𝑅𝑁𝑎 + 𝐶𝑎2+ → 𝑅2𝐶𝑎 + 2𝑁𝑎+ What is not shown on the reaction is the presence of 𝑆𝑂3 − which are fixed in the resin. The R represents the resin. To adapt the reaction to PFAS adsorption, 𝑁𝑎+ can be replaced with 𝐶𝑙−, 𝐶𝑎2+ can be replaced with any PFAS compound and 𝑆𝑂3 − can be replaced with 𝑁𝐻4 + as terminal groups in the resin matrix R [25]. The resin contains 𝐶𝑙− which reacts with the PFAS coming in contact with the resin and the resin releasing 2 𝐶𝑙− for each PFAS [39]. The PFAS then binds to the resin and the fixed 𝑁𝐻4 +. After replacing the compounds, the reaction would look like this: 2𝑅𝐶𝑙 + 𝑃𝐹𝐴𝑆2− → 𝑅2𝑃𝐹𝐴𝑆 + 2𝐶𝑙− The resin can be either single-use or it can be regenerated and used multiple times [37]. For removal of PFAS it is most common to use single-use resins as the removal CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 12 efficiency is high and they are easy to use and do not require any further treatment for regeneration. Studies have also shown that the efficiency of regeneration of the resins have achieved mixed results with some challenges to achieve sufficient regeneration [38]. However, in Europe, regeneration of the resin is not allowed for usage in drinking water treatment as certain chemicals are used which are unfit in drinking water production [25]. This appears to make regenerated resins unsuitable and single-use resins the more feasible option [40]. There are studies on how different resin properties affect the PFAS removal, however, they are not readily available or in early stages [38]. But studies comparing different resins tend to show that there can be a difference in how the resins perform for some aspects and that certain types resin work perform better than others [41]. Overall, ion exchange has been shown to be efficient for a broad interval of PFAS concentrations, both low and high [37]. The necessary contact time (EBCT) for removal of PFAS lies between 2.5-4 minutes [25]. As for GAC, two columns can be couples in series where the first (“LEAD”) removes the majority of the PFAS and the latter (“LAG”) assures polishing of PFAS. A simplified figure of how the process of ion exchange works is shown in Figure 2.4. Figure 2.4: Simplified figure of ion exchange. 2.4.4 Foam Fractionation Foam Fractionation is a process that utilizes physical separation of contaminants from the water [37]. Gas and turbulence are introduced into the water which leads to production of bubbles which will rise to the surface and create a foam. The contaminant is adsorbed onto the bubbles which then rises to the surface [42]. The foam is then separated from the surface of the water, removing the contaminants with it. Foam fractionation typically removes dissolved and soluble surface-active compounds [43]. As PFAS compounds are soluble surface-active compounds, foam fractionation could be used to remove PFAS from water. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 13 PFAS are surface-active compounds as they are amphiphilic, with both hydrophilic and hydrophobic ends [43]. This allows the PFAS compounds to adsorb to the bubble’s air- liquid interface as the hydrophobic end repels water and is situated inside the bubble and the hydrophilic end attracts water and is situated outside the bubble. Once the bubbles have risen and created foam, the foam is separated from the water and foamate is created. Foamate is the foam turned into liquid form which occurs after allowing the foam to collapse. It is possible to add co-surfactants into the process which can increase the adsorption of PFAS and in particular short-chain PFAS [37]. However, it is not allowed to add co- surfactants for drinking water production [25]. Previous studies regarding foam fractionation have mainly focused on removing various contaminants from wastewater [43], [44]. Progress has also been made in recent years regarding removal of PFAS from groundwater and leachate using foam fractionation. Foam fractionation generally requires high concentrations of PFAS to be efficiently reducing the PFAS contamination. A simplified figure of how the process of foam fractionation works is shown in Figure 2.5 Figure 2.5: Simplified figure of foam fractionation. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 14 3 Methodology The project has been conducted mainly by collecting information by means of a literature study and receiving data using various tools to generate a general evaluation of the technologies. The tools and methods that have been used are explained below together with a process chart in Figure 3.1. Figure 3.1: Process chart for the working methodology in this study, detailing how the literature research has been conducted and acted as a basis for the results. A literature study has been conducted to collect information and data regarding the four water treatment processes. The literature has been obtained through various databases with Web of Science as the main database and using various search strings (Table 3.1). CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 15 Table 3.1: Search strings used during literature research. Search Strings Used Number of Results Web of Science Number of Relevant Articles Web of Science Articles used “GAC” AND PFAS AND “drinking water” AND (“pilot scale” OR “full scale”) (Field: Abstract) 10 4 3 “GAC” AND cost AND “drinking water” AND (“pilot scale” OR “full scale”) (Field: Abstract) 16 2 1 “Granular activated carbon” AND PFAS AND "drinking water" (“pilot scale” OR “full scale”) 35 14 6 nanofiltration AND PFAS* AND “drinking water” AND (“removal efficiency” OR “pilot scale” OR “full scale”) 9 3 2 “nanofiltration” AND “PFAS” AND (“pilot scale” OR “full scale”) 6 3 2 “nanofiltration” AND cost AND "drinking water" AND (“pilot scale” OR “full scale”) (Field: Abstract) 7 2 1 Ion Exchange AND PFAS AND Pilot Scale AND Drinking Water 16 5 3 (Ion Exchange OR Anion Exchange) AND PFAS AND Drinking Water AND (Pilot Scale OR Full Scale) 26 7 3 (Ion Exchange OR Anion Exchange) AND PFAS AND (Pilot Scale OR Full Scale) AND Cost 11 3 1 Foam Fractionation AND PFAS AND Drinking Water AND (Pilot Scale OR Full Scale) 1 1 1 Foam Fractionation AND PFAS AND (Pilot Scale OR Full Scale) 6 4 3 Additional literature has been provided by the supervisors as well as through snowballing when any references in articles have seemed interesting and relevant. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 16 3.1 Collection of Information Regarding Technology Performance When researching the technology performance, relevant information from the literature, such as water qualities and parameters, as well as removal efficiencies were documented. Through the documented information, a pattern identification has been done for each individual technology. Rough sketches of flowcharts for each technology were made to easily compare the identified patterns of how the PFAS removal was affected by the water chemistry (Fel! Hittar inte referenskälla.). These include noting repeated similarities in results as well as known interconnections between water parameters, results, or other conditions. These were later used to create a flowchart and the parameters which have been identified to inhibit or facilitate the removal efficiency for the technologies are included as steps in the flowchart. Parameters which have been relevant are removal efficiency, presence of organic matter, presence of ions, chain- length of the PFAS compound and functional group of the PFAS compound as well as pH-levels. Furthermore, other parameters such as the concentration of PFAS, contact times and breakthrough of sorption media have been some of the important factors identified which have an impact in the performances of the technologies were discussed but not included in the flowchart as they are not part of the water chemistry or for some reason not suitable in the flowchart. The flowchart was created with the purpose of having a schematic decision support for water treatment plants deciding which water treatment process to implement. The final flowchart along with a sustainability analysis is presented in chapter 4.3. The pros and cons for respective treatment process have been discussed according to environmental, and economic sustainability as well as the technical performances. 3.2 Sustainability Analysis In this project, a sustainability analysis was conducted to get an overview of the overall sustainability of the four water treatment processes. The criteria for the analysis were chosen prior to the collection of information and was chosen according to what was considered relevant and needed for drinking water treatment plants to evaluate the sustainability and feasibility of implementing a treatment technology. Different options for criteria were discussed with supervisors with knowledge in the subject. A tool used for sustainability analysis, in this case WISER, was also used as inspiration for the setup of the criteria. The technical aspect has been discussed in the flowchart and is therefore not included in the sustainability performance analysis, as the purpose of the analysis is to provide additional decision support from other aspects according to sustainability. Criteria were determined and defined with an objective describing each criterion (Table 3.2) and the criteria are implementation cost, maintenance cost, regeneration, CO2 eq. emissions and energy consumption. Other possible criteria that were discussed were for example space requirement and social aspects. These criteria were ultimately excluded due to time restraints and availability of data. Scoring for the degree of which a criteria objective was met was determined to five possible score with a range from - - to ++. When there has been no available information regarding a criterion, the score has been set to N/A and when a criterion not applicable for the technology, the score has been marked with a backslash (\). The scores and definitions can be found in Table 3.3. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 17 Table 3.2: Criteria and objectives for the multi-criteria analysis. Criteria Objective Economic Implementation cost Realistic and reasonable implementation cost for a treatment plant Maintenance cost Cost for operation and maintenance of the treatment plant, e.g. replacement of media, life span Environmental Regeneration Possibility of regeneration and its feasibility CO2 eq. emissions Environmental impact of the life cycle of the technology. Energy Energy consumption Energy consumption due to e.g. aeration and pumping of water Table 3.3: Scoring for the multi-criteria analysis and the degree to which objective is met for each respective score. Score Degree to which objective is met -- Low - Moderately low 0 Neutral + Moderately high ++ High N/A Not available \ Not applicable Information for the analysis was compiled similarly to the flowchart, through a literature study. Information was also obtained from outside sources, such as from Laholmsbuktens VA, and through the Swedish Water and Wastewater Association (Svenskt Vatten). Svenskt Vatten has developed a tool for assessing footprints for different treatment steps. However, this tool is not currently adapted for PFAS treatment, and processes like NF, IX, or FF are not addressed. The tool was further developed by Ludwig Hedberg, which has made it possible to obtain limited information about GAC and NF. 3.3 Compilation of Results for Decision Support The flowchart and the sustainability analysis were combined into one figure which can act as a decision support for drinking water treatment plants choosing what technology to implement. Recommended technologies were suggested in the last row of the flowchart as suggestions of which removal technology or technologies are the most suitable for the water chemistry. The sustainability analysis was included in the figure to take the treatment’s plants priorities into account. Three imaginary examples of drinking water treatment plants, their water chemistry and what sustainability criteria CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 18 they prioritised when implementing a technology for PFAS removal were created to exemplify the decision support. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 19 4 Results The literature findings for technology and sustainability performance are presented in chapter 4.1 and 4.2 and the resulting flowchart along with the sustainability analysis is presented in chapter 4.3. 4.1 Technology Performance Based on Water Chemistry The technology performance for PFAS removal using granular activated carbon, nanofiltration, ion exchange and foam fractionation was assessed based on water chemistry in the incoming water. The findings from the literature search are presented below. Documentations of water chemistry and other relevant parameters can be found in Appendix A. 1 - Appendix A. 15. 4.1.1 Granular Activated Carbon In a study, full-scale water treatment systems in the US that were studied, where the removal efficiency using GAC was tested [45]. In the study, there were a total of 20 utilities with different water treatment processes, where utility #7 (GAC #7) and utility #20 (GAC #20) are included in this report. In the study, GAC #7 was treated by six granular activated carbon contact adsorbers. The time of which the GAC had been in operation was not specified in the article. The empty bed contact time (EBCT) was 10 min, with adsorbers named Norit GAC30, with surface water as raw water source [45]. GAC #20 used Calgon F600, where EBCT was 13 minutes, and the PFAS concentrations had been monitored for 5 years with groundwater as water source. The system consisted of a series set up with a lead and a lag basin. When comparing the removal efficiencies for PFAS belonging to either PFCA or PFSA, the removal efficiency was shown to be higher for the tested PFSA for this case. For GAC #20 higher removal efficiencies were achieved for PFOA and PFOS. GAC #20 had an EBCT of 13 min in each contactor. The removal rates in Table 4.1 are based on the average removal over the course of one year for the lag basin [45]. The concentrations of the lag basin had been monitored for 16 months. Concentrations of PFBA, PFPeA, PFHxA, PFOA, PFBS, PFHxS, and PFOS were monitored for nearly five years on the influent and the lead basin effluent. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 20 Table 4.1: Summarised removal efficiencies with consideration to chain length and functional group [45]. PFAS compound Chain length Removal efficiency GAC #7 [45] Removal efficiency GAC #20 [45] PFSA PFBS C4 – short >96% N/A PFOS C8 – long >89% >95% PFHxS C6 – long >96% >41% PFCA PFBA C4 -short 33% -17% PFPeA C5 - short 74% >22% PFHxA C6 - short 91% >68% PFHpA C7 - long >89% N/A PFOA C8 - long >48% >92% PFNA C9 - long >37% N/A *Treatment train for GAC #7: Riverbank Filtration/Aquifer Recharge and Recovery/Softening/Solids Contact Clarifier/UV Photolysis with Advanced Oxidation (Hydrogen Peroxide)/Granular Filtration/Granular Activated Carbon Filtration [45]. ** Treatment train for GAC #20: Granular Activated Carbon Filtration/Hypochlorous or Hypochlorite [45]. In another study with a pilot-scale, eight GAC adsorbents were tested from different manufacturers [41]. The absorbers achieved a concentration lower than 2 ng/L PFAS initially. The test showed that GAC reached breakthrough at rates between 2 and 5 months for PFOA for an initial concentration of 17.2 ng/L PFOA. Breakthrough happened earlier for non-/sub-bituminous GAC in comparison to GACs from bituminous carbon sources. This results in bituminous GAC having a longer media life for PFOA. Like the statements from the article above, the study suggests that higher removal efficiencies are obtained for PFOS in comparison to PFOA. The initial breakthrough for PFOS was shown to range between 2 and 8.5 months [41]. The initial breakthrough for PFHxS happened between 4 and 8 months. Short-chained PFAS, 70%), meaning that PFAS and TOC have the capability of both being effectively removed in the presence of each other [50]. When it comes to fouling, there are multiple parameters and characteristics that can influence fouling, where the raw water quality characteristics are important. NOM is known to be the primary factor for membrane fouling [51]. The presence of NOM has the capability of altering the characteristics of the surface of the membrane which can influence the electrostatic interactions. NOM can also block the pores of the membrane as well as form complexes with ions leading to a filter cake layer. In another study, pilot-scale field tests with a reject byproduct of 10% and 90% permeate recovery were tested [52]. The NF pilot consisted of NF270 membranes. The groundwater source was monitored for 30 days. The average values of the PFAS compound rejections can be found in Table 4.2 [52]. The study showed a lower rejection of short-chained PFAS during a recovery of >95%, which decreased at a recovery of 97%. The longer-chained PFAS had no significant influence. In a two-stage NF pilot, PFAS was efficiently removed by 98%, with a permeate water containing approximately 77 ng/L [53]. The removal efficiencies are shown in Table 4.2. The table below shows a high removal rate for PFAS compounds belonging to both PFSA and PFCA with varied chain lengths. The removal efficiency of the two-staged NF pilot plant was using full-scale membranes that were monitored for almost 6 months [53]. The two-stage membrane process had 6 membranes in stage 1 and three membranes in stage 2. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 23 Table 4.2: Summarised removal efficiencies with consideration to chain length, functional groups, and molecular weights. [50], [52], [53]. PFAS compound Molecular weight (g/mol) Chain length NF90-400 Removal efficiency* [%] [53] NF270 Removal efficiency** [%] [52] dNF40 Removal efficiency*** [%] [50] PFSA PFPrS 250 C3 - short N/A >85 N/A PFBS 300 C4 – short 96 >90 70 PFPeS 350 C5 - short 97 >95 N/A PFHxS 400 C6 – long 97 >96 >80 FHxSA 400 C6 - long N/A >90 N/A PFHpS 450 C7 - long N/A >96 N/A PFOS 500 C8 – long 99 >98 >90 PFCA PFBA 212 C3/C4 - short N/A N/A 70 PFPeA 264 C5 - short 93 >90 N/A PFHxA 314 C6 - short 97 >95 >80 PFHpA 364 C7 - long 89 >90 >99 PFOA 414 C8 - long 96 >98 >85 PFNA 464 C8/C9 - long N/A N/A >85 *Wellfields, pilot-scale, NF90-400, average for 175 days, 80% recovery **GW, pilot-scale, NF270, average for 30 days *** Surface water, dNF40, 11 rounds (cycles) à 7 days 4.1.3 Ion Exchange Inorganic ions present in the water have been shown to reduce the adsorption of PFAS onto the ion exchange resins [54], [55], [56]. IX has a high selectivity for ions such as sulphate, chloride, nitrate, nitrite among others. That is because these ions compete with the PFAS in adsorption to the IX resins which disturbs the process and reduces the PFAS adsorption and subsequent removal. Additionally, high enough concentrations of ions can replace already adsorbed PFAS and release the PFAS back into the water [25]. In one study, sulphate, bicarbonate, nitrate, and phosphate were increased to 50 mmol/L for 0,5 mmol/L PFHxS [56]. This decreased the adsorption of PFHxS by 10%. Furthermore, increasing the NaCl concentration to 1000 mmol/L decreased PFHxS adsorption by 30%. It can be believed that stronger electrostatic interactions lead to larger reductions of PFAS adsorption and therefore lower removal [54]. Although not as explored, it can also be assumed that the short-chain PFAS will be more affected by the competing ions. The shorter chain will not be as strongly adsorbed onto the resin as long-chain PFAS are which can also lead to short-chain PFAS eventually being re- CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 24 released into the water if long enough time passes [25]. This could potentially lead to the effluent short-chain PFAS concentration being higher than the influent. Like the presence of ions in the water, the presence of organic matter has also been shown to affect the removal of PFAS [54], [55]. The presence of NOM in the water has been shown to reduce PFAS removal due to electrostatic interactions. While the PFAS removal was low in these cases, the removal of DOC was high. What affects the adsorption of organic matter can be the molecular weight of the NOM as well as the carbon content. Furthermore, the resin type could also impact the performance as polyacrylic resins have been shown to be more affected by NOM than polystyrene resins are. When it comes to dissolved organic matter, it has been suggested that it does not matter what the concentration of DOM the water contained once it was above 0 mg/L [48]. The study showed that for concentrations of 2-8 mg DOC/L the removal of PFSA and PFCA decreased by almost 10 percentage points compared to the concentration of 0 mg/L. The difference between 2, 4 and 8 was not significant when the DOC values were changed, and other parameters remained the same. It was therefore a matter of whether DOM was present in the water or not. Presence of DOM had a slight negative influence on PFAS removal. The PFAS removal was reduced by a maximum of 10% with the presence of DOM at any concentration. A study regarding the breakthrough of different PFAS found in groundwater used for drinking water in Anaheim, USA, compared ion exchange and GAC [41]. Breakthrough, in this study, is defined as the effluent concentration divided by the influent concentration occurring instantly, the same day. The EBCT used in the study was 2 minutes and the average concentration for the PFAS compounds was 17.3 ng/L for PFOA, 23.0 ng/L for PFOS, 10.7 ng/L for PFHxS, and 29.4 ng/L for PFBS. The results showed a later breakthrough of the IX resins compared to GAC. The study compared four different resins and their removal of the four PFAS compounds detected in the pilot feedwater (Table 4.3). Table 4.3: The time of breakthrough (months) for four different ion exchange resins and the four detected PFAS types [41]. No breakthrough means that breakthrough has not occurred when the study ended after 26 months. PFOA, PFOS and PFHxS are all included in PFAS 4, PFBS is included in PFAS 21. PFA694E CR2301 LC4 PSR2+ PFOA 3.5 3.5 3.5 7 PFOS No breakthrough No breakthrough No breakthrough No breakthrough PFHxS No breakthrough 16 No breakthrough No breakthrough PFBS 9 9 9 22 As can be seen Table 4.3, for PFOA, three of the resins had a breakthrough at 3.5 months while one resin had a breakthrough at 7 months [41]. All resins had no breakthrough for removal of PFOS and PFHxS when the study ended after 26 months except for CR2301 for PFHxS which had a breakthrough at 16 months. For PFBS, three resins had a breakthrough at 9 months, and one resin had a breakthrough at 22 months. The only PFAS compound of the four with a carboxylic acid functional group is PFOA, the rest are compounds with sulphonate as functional group. What can be seen is that CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 25 PFOA and PFOS, which both have the same chain length of C8, have significantly different times of breakthrough. This suggests that different functional groups have an influence on the removal efficiency. The study suggests that PFOS is more readily adsorbed compared to PFOA [41]. This has been seen in multiple previous studies which implies that the functional group of the PFAS compound has an influence on the adsorption onto the resins, and particularly suggesting that sulphonates is the group with higher adsorption [41], [57]. While the study states that a clear correlation between chain-length and adsorption could not be confirmed, the results imply that the four IX resins tested remove the long-chain as well as the sulphonic acid PFAS compounds better than both the short-chain and the carboxylic acid PFAS compounds [41]. However, a long-chain PFCA could still be more readily removed than a short-chain PFSA. When it comes to the impact of chain- length, a review of multiple studies suggests that many lengths of PFAS compounds, both long-chain and short-chain, can be removed, except for the absolute shortest chains [58]. Many studies show that pH does not significantly affect IX. Varying results have been noted with different results for different PFAS groups and different resin types, but the studies suggest that pH differences do not have a significant impact on the removal efficiency of ion exchange [54]. 4.1.4 Foam Fractionation Due to the scarcity of available research about foam fractionation used for drinking water purposes, information regarding the technology has been collected based on other limitations compared to the previous technologies mentioned. The studies reviewed for foam fractionation have regarded removal of PFAS from various raw water sources such as leachate water and wastewater. Another article, which was the only one found regarding drinking water, discussed the removal of PFAS from concentrate after nanofiltration, however, the concentrate going through foam fractionation becomes retentate and not drinking water. Due to this, the following review of foam fractionation is not necessarily readily applied to drinking water. One of the main aspects that affects the efficiency of foam fractionation is that high concentrations of PFAS are generally required to achieve efficient removal [43], [44], [59]. One study looked at the removal of multiple PFAS compounds in landfill leachate water and found that removal of 90% could be reached when the PFAS concentration was above 50 ng/L for every individual compound [60]. Beyond that, a continuous flow experiment was made with the total PFAS concentration ranging from 3200-25 000 ng/L. A removal of above 90% occurred for the tests with 10 000 ng/L and 25 000 ng/L while the concentrations between 3200-5100 ng/L had removal efficiencies around 75- 85%. Further it can be seen that long-chain PFAS has higher removal rates than short-chain PFAS [43], [44], [61]. One study showed a removal efficiency of 67% for long-chain PFAS while short-chain PFAS had a removal efficiency of 10% [59]. Another study reported a removal efficiency of 81-91% for several long-chain PFAS, such as PFOS and PFOA, and below 30% for short-chain PFCA (chain length below six) and PFSA (chain length below seven) [62]. The lower removal of short-chain PFAS is believed to CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 26 be due to the hydrophobic end having a lower surface activity as the fluorinated carbon chain is shorter [63]. This leads to a weakened adsorption onto the air bubbles. Furthermore, functional groups can also influence removal efficiency of PFAS. It has been shown that sulphonate groups have a lower solubility than carboxylate groups which can increase the hydrophobicity and thus the adsorption of PFAS onto the bubbles [43]. A study made with multiple tests for different PFAS compounds in landfill leachate water, with PFAS concentrations varying between 1100-25 200 ng/L, showed an average removal efficiency of 77% for PFCA and 94% for PFSA [60]. Studies show that a higher concentration of metal cations as well as higher ionic strength in the water increases the removal efficiency [43], [62]. Metal cations that are mono-, di- or tri-valent, such as Na+, K+, Ba2+, Mg2+, Al3+, Fe3+, can be found in groundwater. Studies suggested that the presence of metal cations can increase the removal efficiency and that the efficiency increases further with the charge density of the cation [43], [64]. It has also been observed that higher concentration of ionic salts leads to a lower surface tension of the water allowing faster formation of the foam [64]. It has been suggested that the ionic strength increases removal efficiency due to the solubility of PFAS being reduced as well as the viscosity of the water being increased with increasing ionic strength [42], [43], [60]. However, it was also reported that when the salinity was high, and a co-surfactant (sodium dodecyl sulphate) was used, the removal efficiency decreased as foam formation was decreased. Depending on what is present in the water, pH can influence the removal efficiency [65]. As previously mentioned, metal cations have been shown to increase the removal efficiency. However, it has been observed that for the metal cations to be efficient, a low pH-level would be favourable. One study showed that when metal cations (Fe3+) were present in the water, PFOS and PFOA were removed with a removal efficiency of around 90% when the pH was relatively low at 2.3. Another study, which was done with a pH-level of 7.8, needed an additional Fe3+ amount 20 times higher than the previous study did at a pH-level of 2.3 to reach the same removal efficiency [60]. However, when cationic co-surfactants were present, there was no significant difference and removal remained efficient even at slightly higher pH-levels. Contact time has also been shown to affect the removal efficiency. Longer contact time leads to higher removal. Studies show that most of the PFAS removal occurs early in the process, but a longer time allows for further removal [59]. Increased gas flow rate increases removal as more bubbles are formed and made available for the contaminants to adsorb onto [43]. However, too high flow rate can break the foam. Though not very researched, the removal of long-chain PFAS seems to not be affected by temperature while the removal of short-chain PFAS decreases as the temperature increases [43]. 4.2 Sustainability Performance The sustainability of the technologies was evaluated based on the set criteria and scoring (Table 4.4). The scores were set based on the literature search which has been CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 27 documented in this chapter. An abbreviated summary of the motivation behind the scores can be found in Appendix B – Summary of Motivations for the Scoring of the Sustainability Analysis Appendix B. 1. Table 4.4 Final scoring based on results from the sustainability performance analysis. GAC NF IX FF Implementation cost 0 + + ++ Maintenance cost 0 - + 0 Waste management - -- - ++ Regeneration + \ - \ CO2 eq. emissions - N/A 0 N/A Energy consumption ++ -- ++ -- 4.2.1 Granular Activated Carbon In a study, multiple PFAS were examined in a full-scale drinking water treatment plant [66]. The study performed an economic analysis for the annual regeneration per Filtrasorb 400 filter. The operational costs were also looked at, defined as annual regeneration costs together with the uniform annual costs of the initial purchase cost of a filter of virgin GAC over 10 years with an interest rate of 5% per year. The annual unit operation cost is defined as the annual operations cost over the annual capacity of treated water. A table of the costs can be found in Appendix A. 10. The drinking water treatment plant is for 7 million m3 of water per year. The annual operations costs are at approximately 174 000 000 euros to reach the water goal of 85 ng/L PFAS 11. Another report studied an NF pilot plant used for PFAS removal in drinking water production [67]. The concentrate generated from the NF process was treated with both GAC and ion exchange and all technologies were assessed in the study. The implementation cost of GAC was presented in the supplementary information for the report and was 1142 €/m3 absorbent of Filtrasorb400 [67]. The operational costs to reach the discharge goal of 4 ng/L was 4.41 €/m3 concentrate for Filtrasorb 400 and 19.30 €/m3 concentrate for Norit 1240 W. The operations costs include purchase of virgin GAC, regeneration, transportation to and from the DWTP, pumping energy cost CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 28 to lift the water 6 meters where no operations personal costs or backwashing costs were included. Table 4.5 shows the costs to reach a discharge of 25 ng/L for the concentrate where a 10-year service life is assumed. The costs consider 10% loss of media from replacing the virgin GAC during regeneration. The energy cost for the calculations is 0.095 €/kWh.a highly concentrated water with FPAS levels. In a calculation based on raw water for GAC, the operation costs are at 0.038 €/m3 for reaching the goal of 25 ng/L instead of the 0.80 €/m3 for concentrated water. Table 4.5: Unit costs and operations costs for GAC [67]. Filtrasorb 400 Norit 1240 W Unit costs Unit cost of new GAC or AIX [€/m3 absorbent] 1142 1356 GAC regeneration or AIX incineration cost including transport and GAC handling [€/m3 absorbent] 714 714 Operations costs Annual operations cost new AIX and incineration to meet water goal [€/m3 concentrate] 0.036 3.12 Pumping energy to lift 6 m with pump efficiency 80% [€/m3 concentrate] 0.002 0.002 Total operations costs [€/m3 concentrate] 0.80 3.12 The costs of GAC treatment at one site, according to Laholmbuktens VA AB is presented in Table 4.6. The activated carbon (Brennsorb 1240) is purchased in batches of 2 tonnes when needed and is disposed once exhausted [68]. During the time with the given data, the treatment plant produced 450 000 m3 water [25]. The total cost of the purchase of carbon together with the waste management was approximately 880 000 SEK leading to an operational cost of 2 SEK/m3 when only taking the activated carbon into consideration. CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 29 Table 4.6: Operations costs for activated carbon based on a treatment plant in south of Sweden [68]. Month Purchase of Carbon (SEK) Waste Management (SEK) 2022 January 79 920 - March 79 920 - May - 109 841 June 74 300 20 734 September - - October - 67 240 December 74 300 49 415 2023 March 74 300 - May - 67 453 June 74 300 110 966 October Closed - A life cycle assessment was performed for GAC and ion exchange [69]. Regarding GAC, the study included results for single-use GAC with off-site incineration and reactivated GAC with off-site thermal reactivation from pilot-scale studies. The conditions for the GACs can be found in Appendix A. 11. The capital costs for GAC and operation and maintenance costs for GAC with incineration and GAC with reactivation can be found in Appendix A. 12 and CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 30 Appendix A. 13. The total capital costs for GAC were $78 445 for both single-use and reactivated GAC and the operation and maintenance cost was $17 589 for GAC with incineration and $16 634 for reactivated GAC [69]. Further results from the same study show that for the GAC systems, the production of activated carbon has the largest impact for all the assessed parameters in the study [69]. The assessed parameters include ozone depletion, global warming, smog, acidification, eutrophication, carcinogenics, non-carcinogenics, respiratory effects, ecotoxicity and fossil fuel depletion. For all parameters, the production of GAC stood for over 40% of the overall environmental impact, and for over 70% for majority of the individual parameters [69]. In total, the production of single-use GAC resulted in >80% of the environmental impact of GAC, and the reactivated GAC had similar results. The second largest impact was caused by the incineration of hazardous waste. The values for each of the parameters assessed for environmental impact can be found in Appendix A. 14. It can be seen how the reactivated GAC has lower environmental impact in comparison to single-use GAC for all parameters. For single-use GAC the global warming was 0.4407 kg CO2 eq./m3 treated water and for GAC with thermal reactivation it was 0.0726 kg CO2 eq./m3 treated water [69]. During thermal reactivation approximately 10% of the media goes to losses. There are multiple variants of GAC filters that can be used that might perform differently sustainability wise. Below follows a short comparison between coal, peat, coconut, wood, and reactivated coal GAC. The overall performance of the different filters was looked at in its original condition, with 20% increase of raw materials, 20% decrease of raw materials, 20% increase of electricity demand as well as 20% decreased energy demand [70]. When it comes to global warming potential, kg CO2 eq., the direct emissions caused by coal, peat, coconut, and wood were high whereas reactivated coal was the lowest with the main cause being make-up AC production. Looking at the energy consumption for GAC, it is of similar amount as for IX, consuming <0.05 kWh/m3 [25]. 4.2.2 Nanofiltration A full NF system has been considered to cost 2 to 5 times more than a GAC filter [31]. A cost analysis was performed looking at a pilot-scale NF system fed by a full-scale conventional water treatment plant [71]. The pilot-scale tested both loose nanofiltration (L-NF), with a lower ion rejection capacity, and tight nanofiltration membranes (T-NF). Loose nanofiltration has a molecular weight cut-off (MWCO) of >500 Da and tight nanofiltration has a MWCO <500 Da [72]. The capacity of the pilot-scale NF was 400 000 m3/d, where the investment cost of L-NF was approximately $104 500 000, which included 47 520 membrane modules, construction, and equipment [71]. For T-NF, the cost was approximately $92 500 000 for 28 800 modules along with remineralization. One unit membrane had a price of $1100. The cost analysis for the NF-system was performed considering the investment costs of construction cost, mechanical and electrotechnical equipment cost and membrane cost [71]. The operation and maintenance costs included costs of chemicals used for remineralization for post-treatment as the results from the study for the T-NF would otherwise result in corrosion. Other costs for operation and maintenance included membrane replacement, energy, labour as well as maintenance and repair. In the study, CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 31 electricity consumption for the full-scale NF system was used for calculations to keep margins due to the influence of energy consumption that the size and efficiency of the pumps can cause. The unit cost of electricity was reported as 0.0615 $/kWh in the cost analysis [71]. The chemical costs included costs for chemicals used to prevent membrane clogging and to condition and disinfect the feed water. The lifespan was assumed to be 6.5 years for the average membrane for NF. The final costs from the cost analysis can be found in Table 4.7. For the different membranes, the results showed that the total costs for L-NF and T-NF were just below 29 000 000 $/year [71]. The highest costs per unit could be found in the energy cost of 0.068 $/m3, the membrane replacement cost of 0.054 $/m3 for L-NF and 0.033 $/m3 for T-NF and for the chemical costs that were 0.018 $/m3 for L-NF and 0.038 $/m3 for T-NF. The total unit cost was 0.157 $/m3 for L-NF and 0.156 $/m3 for T-NF. The major operational cost for nanofiltration has been shown to be the cost for energy use [31]. The cost is 0.05-0.10 €/m3 of treated water depending on energy consumption as well as the energy price [31]. According to a study looking at various peer reviewed data on full-scale NF plants, the energy demand is considered to be 0.4 kWh/m3 [73]. For only pumping energy, another study has reported the value of 0.27 kWh/m3 permeate [67]. Meanwhile, in a previously mentioned study, the energy consumption for the pilot NF plant was considered to be 1.1 kWh/m3 [71]. In another pilot-study, where a cost analysis was performed, implementation costs for nanofiltration were reported, which were summarized and are shown in Table 4.8 [67]. The operational costs included energy for feedwater, recirculation, antiscalant dosing pumps, cost of antiscalant, membranes, article filter replacement as well as the cost of the cleaning in place. When it comes to operational and maintenance costs the pumping of the water and the cleaning of the membrane requires maintenance. Table 4.9 shows the summarized costs [67]. There are also other costs that need to be taken into consideration such as prefilters, antiscalants, and chemicals that could be used for example pH adjustments. This would result in a total operation cost of 0.100 €/m3 permeate including the purchase of the membrane [67]. Without the membrane the total cost would be 0.085 €/m3 permeate. The highest costs for nanofiltration are the pumping energy at € 0.026, the antiscalant, to prevent the membrane from scaling or fouling at € 0.036. A summary of all the costs from the two different pilot studies can be found in Table 4.9. CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 32 Table 4.9: Implementation, operation and maintenance costs for nanofiltration [67], [71]. Parameter Description Two-stage, 6 NF90 & 3 NF270 [67] L-NF [71] T-NF [71] Implementation costs NF membrane Cost of membranes per unit € 700 (0.015 €/m3) $ 1100 (0.012 $/m3) $ 1100 (0.012 $/m3) Construction cost - - 0.008 $/m3 0.008 $/m3 Equipment cost - - 0.021 $/m3 0.012 $/m3 Remineralizat ion cost To prevent pipe corrosion (post- treatment) - - 0.003 $/m3 Prefilters Replaced annually 437 €/filter - - Operation and maintenance Pumping energy Feed and recirculation 0.026 €/m3 (0.095 €/kWh) 0.068 $/m3 (including backflush) 0.068 $/m3 (including backflush) Chemical cost All chemicals involved - 0.018 $/m3 0.038 $/m3 Sodium Hydroxide pH adjustment from 7.5 - 7.6 0.002 €/m3 - - Labor cost - - 0.09 $/m3 0.09 $/m3 Maintenance and repair - - 0.09 $/m3 0.09 $/m3 Membrane replacement cost - - 0.054 $/m3 0.033 $/m3 Clean in place Membrane washing acid/chemical s 0.015 €/m3 - - CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 33 Apart from looking at the numbers above, it is important to note that the final cost is case specific [31]. This can be seen in the varying energy use and price. Since nanofiltration is a separation process, there will be retentate water as a rest product/waste. The retentate water contains the rejected particles from the feedwater creating a highly concentrated water containing a large amount of PFAS. The high levels of PFAS in the water must be treated before moving to wastewater treatment. The waste stream produced is of significant volume of up to 20% of the original volume [31]. There are many options for treatment of the waste stream where PFAS will be disposed in the end. In a comparative study, two nanofiltration processes with ultrafiltration as a pre- treatment were tested [74]. The lifespan of the NF membrane was reported to be 10 years. The characteristics of the NF-systems can be found in Appendix A. 15. In the impact assessment, climate change and fossil depletion were discussed. During a nanofiltration installation operationkg CO2 eq. affects climate change and around 0.1 kg oil eq. affects fossil depletion [74]. What was also looked at for the CO2 eq. was the NF manufacturing and the NF installation construction where both were at such low values that they were not visible in the graph. It is important to note that the study does not seem to be based on a full or pilot-scale study. Therefore, the given results from the study might not reflect reality. According to Svenskt Vatten the emission factor of one element of NF membrane is 90.4 kg CO2 eq. for NF from NX filtration and 477 kg CO2 eq. from N64 Pentair [75], [76]. Other information regarding carbon footprint for NF is not available. 4.2.3 Ion Exchange The cost of ion exchange is derived from the cost of a new unit, operation and maintenance costs which includes both energy costs and regular costs for new resin, and cost of disposal of the resin once exhausted. A study examined the costs of ion exchange which was used for the concentrate produced after nanofiltration [67]. The costs varied between the sorbent type and brand used and two types (A600 and PFA694) were examined in the study (Table 4.10). The annual operation cost was for reaching the discharge limit of 4 ng/L from an initial PFAS concentration of 583 ng/L. The PFAS concentration is high as the study uses the concentrate produced after nanofiltration which is generally higher than raw water. Table 4.10: Unit, sorbent, disposal, and annual operation costs for two types of IX [67]. The annual operation costs are for the discharge concentrations of 4 ng/L with an initial influent concentration of 583 ng/L. Component A600 PFA694 New unit cost 5000 €/m3 8500 €/m3 Disposal cost 510 € m3 510 €/m3 Annual operation cost (for a discharge concentration of 4 ng/L) 1.25 €/m3 1.94 €/m3 A life cycle assessment investigated the costs of ion exchange on a basis of a 30-year life cycle and with an assumption of a flow of 6000 L/h, or 144 m3/day. [69]. Single- CHALMERS, Architecture and Civil Engineering, Master’s Thesis ACEX30 34 use ion exchange resin was compared to regenerable ion exchange resin as well as single-use GAC and reactivated GAC. The capital cost was significantly lower for single-use resin compared to the other three. The capital cost for single-use resin was $44 164 while it was $77 571 for regenerable resin and $78 445 for both single-use and reactivated GAC [69]. It was suggested that the shorter EBCT of 2-3 minutes for IX, compared to the EBCT of 10 minutes for GAC, leads to a lower capital cost as IX requires a smaller contactor than GAC does. Additionally, the annual operation cost was also lower than the other three with $11 246 for single-use resin. Regenerable resin had the annual operation cost of $14 746, single-use GAC had the cost of $19 879 while reactivated GAC had the cost of $16 634 [69]. Table 4.11: Various cost components for ion exchange based on the results of a life cycle assessment of a 30-year life cycle [69]. Component Cost Capital Infrastructure cost $44 164 Sorbent cost 15.70 $/kg Annual operation cost $11 246 Cost per m3 of treated water 0.28 $/m3 Media changeout have been suggested to occur 1.62 times per year according to the 30- year life cycle assessment [69]. However, the time between changeout depends on the quality of the water as higher presence of ion or organic matter would decrease lead to decreased time until breakthrough as well as decreased PFAS adsorption. The higher organic matter and the subsequent decrease in adsorption of PFAS would increase the costs as it could lead to more frequent exhaustion of the resins and thereby require more frequent exchange of the resin [77]. The waste that is produced during ion exchange is mostly generated from the exhausted resin. The resin contains the adsorbed PFAS and will therefore need to be taken care of. Commonly, the resin is incinerated at another location which entails transportation of the waste from the water treatment plant. The resin could also be transported to a landfill where the resin is disposed. Regeneration of the resin is possible and available for use. However, it is not suitable for drinking water as the regeneration requires chemicals which are not allowed in drinking water production in the European Union [25]. Furthermore, regeneration requires transportation to another site where the regeneration occurs which leads to increased costs as well as environmental impact [69]. Information regarding CO2 emissions from ion exchange is limited. The CO2 emissions was discussed with a company providing ion exchange products, where the conclusion was that they did not have any knowledge to share regarding the emissions of the products. The topic has been explored in various studies, but the data has been inconclusive. A life cycle assessment was made regarding different technologies for PFAS removal on a time scale of 30 years [69]. Both single-use ion exchange resin and regenerable resin were evaluated and compared to single-use GAC and thermally reactivated GAC. The study evaluated the environmental impact of the technologies for ten different categories which were ozone depletion, global warming, smog, acidification, eutrophication, carcinogenics, non-carcinogenics, respiratory effects, ecotoxicity and fossil fuel depletion. It was shown that the single-use resin had the significantly lowest CHALMERS Architecture and Civil Engineering, Master’s Thesis ACEX30 35 environmental impact on 9 out of 10 categories with the exception being ozone depletion where it had the highest impact [69]. The overall result was that single-use resin had the lowest impact followed by thermally reactivated GAC, regenerable resin and lastly single-use GAC. It was suggested that the low impact of single-use resin is due to the longer bed volumes until breakthrough compared to GAC. The reason for regenerable resins being higher is suggested to be due to the intense usage of chemicals and solvent solutions for regeneration of the resins which also contributes to hazardous wastes. When it comes to the ozone depletion being highest for single-use resin when compared to the other technologies, it is suggested to be due to the polymer synthesis during production of the resins which releases chemicals that are ozone-depleting [69]. As single-use resins require new resins each time they need to be replaced the impact is higher than for regenerable resin as they are reused and do not require replacement as often as single-use resin does. When looking at what causes most of the environmental impact, the study suggests that most of the impact from single-use resin is caused by the resin production followed by the incineration of exhausted resin while the regenerable resin impact is mostly due to incineration. Global warming was presented as kg CO2 eq. released per m3 treated water. For single- use resin the impact was estimated to be 0.03033 kg CO2 eq./m3 [69]. The estimation for regenerable resin, single-use GAC and reactivated GAC was 0.2717 kg CO2 eq./m3, 0.4407 kg CO2 eq./m3 and 0.07264 kg CO2 eq./m3 respectively. The study shows that an estimated 70% of the im