CECHA EUROPEAN CHEMICALS AGENCY Annex to the ANNEX XV RESTRICTION REPORT PROPOSAL FOR A RESTRICTION SUBSTANCE NAME(S): Per- and polyfluoroalkyl substances (PFASs) IUPAC NAME(S): n.a. EC NUMBER(S): n.a. CAS NUMBER(S): n.a. CONTACT DETAILS OF THE DOSSIER SUBMITTERS: BAuA Federal Institute for Occupational Safety and Health Division 5 - Federal Office for Chemicals Friedrich- Henkel-Weg 1-25 D-44149 Dortmund, Germany Bureau REACH, National Institute for Public Health and the Environment (RIVM) Antonie van Leeuwenhoeklaan 9 3721 MA Bilthoven, The Netherlands Swedish Chemicals Agency (KEMI) PO Box 2, SE-172 13 Sundbyberg, Sweden Norwegian Environment Agency P.O. Box 5672 Torgarden N-7485 Trondheim, Norway The Danish Environmental Protection Agency Tolderlundsvej 5 5000 Odense C, Denmark VERSION NUMBER: 2 DATE: 22.03.2023 P.O. Box 400, FI-00121 Helsinki, Finland I Tel. I Fax +358 9 68618210 I echa.europa.eu ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Version history Version Changes Date 1 Original version for pre-publication February 2023 2 B.9.10.2. – correction of emission numbers for fluorinated gases March 2023 B.9.18.2.5. – minor text addition TABLE OF CONTENTS Annex B Information on hazard and risk ........................................................1 B.1. Identity of the substance(s) and physical and chemical properties ....................................... 1 B.1.1. Name and other identifiers of the substance(s) ........................................................... 1 B.1.2. Physicochemical properties ....................................................................................... 2 B.1.3. Justification for grouping .......................................................................................... 3 B.2. Manufacture and uses (summary) .................................................................................... 7 B.3. C lassification and labelling ............................................................................................... 8 B.3.1. C lassification and labelling in Annex VI of Regulation (EC ) No 1272/2008 (C LP Regulation) ....................................................................................................................................... 8 B.3.2. C lassification and labelling in classification and labelling inventory/ Industry’s self classification(s) and labelling 1............................................................................................11 B.4. Environmental fate properties .........................................................................................17 B.4.1. Degradation...........................................................................................................17 B.4.2. Environmental distribution .......................................................................................70 B.4.3. Persistence compensating low bioaccumulation potential for mobile substances .......... 130 B.4.4. Accumulation in plants .......................................................................................... 132 B.4.5. Removal from the environment, decontamination and purification ............................. 136 B.5. Human health hazard assessment ................................................................................. 141 B.5.1. Toxicokinetics/ADME (absorption, metabolism, distribution and elimination) ............... 149 B.5.2. Evidence from experimental animal data ................................................................. 156 B.5.3. Evidence from epidemiological data ........................................................................ 170 B.5.4. Human data on oligomeric/polymeric PFASs ............................................................ 188 B.5.5. C ombined toxicity................................................................................................. 189 B.5.6. Derivation of DNEL(s)/DMEL(s) .............................................................................. 191 B.6. Human health hazard assessment of physicochemical properties ...................................... 192 B.7. Environmental hazard assessment................................................................................. 193 i ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.1. Ecotoxicity ........................................................................................................... 193 B.7.2. Effects on wildlife ................................................................................................. 204 B.7.3. Atmospheric compartment – global warming potential .............................................. 208 B.7.4. Microbiological activity in sewage treatment systems ............................................... 212 B.7.5. Endocrine activity and endocrine disruption ............................................................. 213 B.7.6. Hazard and occurrence of fluoropoly mers................................................................ 219 B.8. PBT and vPvB assessment ............................................................................................ 223 B.9. Exposure assessment................................................................................................... 224 B.9.1. Introduction to hazard and risk .............................................................................. 226 B.9.2. PFASs manufacturing ............................................................................................ 227 B.9.3. Textiles, upholstery, leather, apparel and carpets .................................................... 232 B.9.4. Food contact materials and packaging .................................................................... 241 B.9.5. Metal plating and manufacturing of metal products .................................................. 253 B.9.6. C onsumer mixtures .............................................................................................. 256 B.9.7. C osmetics............................................................................................................ 257 B.9.8. Ski wax ............................................................................................................... 260 B.9.9. Applications of fluorinated gases ............................................................................ 265 B.9.10. Medical devices .................................................................................................. 272 B.9.11. Transport........................................................................................................... 275 B.9.12. Electronics and semiconductors ............................................................................ 277 B.9.13. Energy .............................................................................................................. 280 B.9.14. C onstruction products ......................................................................................... 283 B.9.15. Lubricants .......................................................................................................... 288 B.9.16. Petroleum and mining ......................................................................................... 292 B.9.17. Active substances in Plant Protection Products (PPP), Biocidal Products (BP) and Medicinal Products (MP) ................................................................................................................ 298 B.9.18. Waste................................................................................................................ 299 B.9.19. Other sources (for example natural sources, unintentional releases) ........................ 320 B.9.20. Overall environmental exposure assessment.......................................................... 321 B.9.21. Human exposure ................................................................................................ 324 B.9.22. C ombined human exposure assessment ................................................................ 330 Appendices to Annex B............................................................................... 345 ii ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.1.2. Physicochemical Properties .......................................................................... 345 Appendix B.4.1.2. Key arrowhead subgroups......................................................................... 389 Appendix B.4.1.3.1. Degradation of PFC A precursors.............................................................. 392 Appendix B.4.2.1.3. Mobility in water ................................................................................... 398 Appendix B.4.2.3. Soil ........................................................................................................ 419 Appendix B.4.2.6.1. EOF/AOF .............................................................................................. 422 Appendix B.4.2.7. Monitoring of specific PFASs in environmental samples ................................ 429 Appendix B.4.2.9. Bioaccumulation ...................................................................................... 442 Appendix B.4.4. Accumulation in plants ................................................................................ 444 Appendix B.5.1.1.2. Distribution .......................................................................................... 460 Appendix B.7.5. Endocrine activity and endocrine disruption ................................................... 503 Appendix B.9.3. Textiles, upholstery, leather, apparel and carpets .......................................... 527 Appendix B.9.4. Food contact materials and packaging........................................................... 534 Appendix B.9.8. Ski wax ..................................................................................................... 536 Appendix B.9.11. Transport................................................................................................. 539 Appendix B.9.12. Electronics and semiconductors .................................................................. 540 Appendix B.9.13. Energy..................................................................................................... 542 Appendix B.9.14. C onstruction products ............................................................................... 544 Appendix B.9.15. Lubricants ................................................................................................ 545 Appendix B.9.16. Petroleum and mining ............................................................................... 546 Appendix B.9.21. Human exposure ...................................................................................... 551 References ............................................................................................... 571 TABLES Table B.1. PFAS applications and researched uses. ...................................................7 Table B.2. 43 PFASs with harmonised classification ...................................................8 Table B.3. PFASs (subdivided into PFAS categories) ................................................ 11 Table B.4. Details on available classification information for the endpoints of concern ...... 15 Table B.5. List of substances selected for QSAR modelling of degradation. .................... 18 Table B.6. Properties of compared substances. ...................................................... 67 Table B.7. Log KOC for PFAAs............................................................................. 74 Table B.8. Log KOC for PFCs .............................................................................. 76 Table B.9. Log KOC for perfluoroalkylamines and perfluoroethers................................. 77 Table B.10. Estimated characteristic travel distances of fluorotelomeric alcohols........... 110 Table B.11. Estimated characteristic travel distances of selected PFCAs...................... 110 Table B.12. Overview on bioconcentration factors (BCFs) and bioaccumulation factors ... 121 iii ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.13. Non-exhaustive list of PFASs............................................................ 141 Table B.14. Ecotoxicological threshold values for different PFASs .............................. 197 Table B.15. GWP-values (GWP-100) collected from the IPCC fourth Assessment Report.. 208 Table B.16. GWP-values (GWP-100) for selected fluranes. ...................................... 209 Table B.17. Overview over EA/ED of different PFASs. ............................................ 214 Table B.18. Methodology used to estimate emissions per sector. .............................. 225 Table B.19. Average emission factors to environmental compartments....................... 228 Table B.20 Estimated annual EEA emissions ........................................................ 228 Table B.21. Comparison of past, current, and expected future emissions .................... 230 Table B.22. Assigning major use applications to release groupings. ........................... 234 Table B.23. Market splits for situations where more than one release grouping exists. ... 235 Table B.24. Estimated annual EEA emissions ....................................................... 236 Table B.25. Assumed TULAC usage rates for PFASs 1990 – 2020.............................. 239 Table B.26. Emissions of PFASs and methods used in the manufacturing stage. ........... 243 Table B.27. Quantities of PTFE in Industrial Cookware re-coating.............................. 247 Table B.28. Emissions of PFASs and methods used. Service life stage ........................ 249 Table B.29. Estimated annual EEA emissions in the FCM and packaging sector ............. 250 Table B.30. Estimated annual EEA emission in the metal plating subsector. ................. 254 Table B.31. Estimated annual EEA emissions in the manufacturing of metal products..... 254 Table B.32. Estimated annual EEA emissions in the consumer mixtures sector ............. 256 Table B.33. Total emissions of PFAS from use of cosmetic products/year .................... 258 Table B.34. Estimates for total emissions for the different cosmetic product categories .. 258 Table B.35. Estimated annual EEA emission in the ski wax sector ............................. 262 Table B.36. Estimates of total emissions of fluorinated gases................................... 268 Table B.37. Estimated total emissions of HFCs and PFCs ........................................ 270 Table B.38. Estimated annual EEA emissions in the medical devices sector ................. 273 Table B.39. Estimated annual EEA emissions in transportation products and articles ...... 276 Table B.40. Estimated annual EEA emissions in electronics and semiconductor sectors ... 278 Table B.41. Estimated annual EEA emissions in the energy sector............................. 281 Table B.42. Estimated annual EEA emissions in the construction sector ...................... 285 Table B.43. Estimated annual EEA emissions in the lubricants sector ......................... 289 Table B.44. Content of monomeric PFASs in fluoropolymers "high" scenario. ............... 293 Table B.45. Estimated annual EEA emissions in the petroleum and mining industry ....... 294 Table B.46. Concentrations of monomeric PFAS in fluoropolymer (high scenario) .......... 296 Table B.47. Concentrations of non-polymeric PFASs in fluoropolymer (low scenario) ...... 296 Table B.48. Groups of PFASs used in the various sectors ........................................ 302 Table B.49. Calculated total PFAS annual loads .................................................... 304 Table B.50. End products of the thermal degradation of reported PFAS compounds. ...... 306 Table B.51. Calculated mean and median concentrations in incineration bottom ash ...... 307 Table B.52. Amount of bottom ash (rounded) ...................................................... 307 Table B.53. Calculated total PFAS amounts in incinerator bottom and fly ash. .............. 308 Table B.54. Calculated mean and median concentrations in the influent/effluent water... 310 Table B.55. Reported influent, effluent and sludge data from Eurostat for 2016. ........... 310 Table B.56. Calculated total yearly PFAS load for selected EU-Member States .............. 310 Table B.57. PFAS emissions (rounded) in EEA for three main waste treatment methods . 311 Table B.58. ECHA ERCs for the waste stage (ECHA, 2012a)..................................... 315 Table B.59. Waste stage emissions based on ECHA ERCs........................................ 316 Table B.60. Summary of EEA yearly emissions of PFASs per use sector ...................... 322 Table B.61. Biomonitoring studies of extractable organofluorine (EOF)....................... 331 Table B.62. Median levels (ng/mL) of restricted (or proposed for restriction) PFAAs....... 333 Table B.63. Detection frequencies of PFAAs ........................................................ 335 Table B.64. Findings of 6:2 Cl-PFESA, ADONA, HFPO-DA and PFECHS in serum ............ 339 Table B.65. Serum levels (ng/mL) of PFAEs ........................................................ 342 Table B.66. Basic substance information and physical chemical properties of PFCAs ...... 345 Table B.67. Basic substance information and physical chemical properties of PFSAs ...... 353 Table B.68. Basic substance information and physical chemical properties of PFCs ........ 366 Table B.69. Basic substance information and physical chemical properties of PFAEs....... 372 Table B.70. Basic substance information and physical chemical properties of PFPAs....... 376 iv ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.71. Perfluoroalkylamines. Basic substance information ................................ 382 Table B.72. Basic substance information and physical chemical properties of FTOHs ...... 385 Table B.73. Presence (YES) or absence (NO) of fragments for different BIOWIN models . 389 Table B.74. Maximum number of instances of each fragment in the training set ........... 391 Table B.75. Summary of formed PFCAs during degradation of n:2 FTOHs ................... 392 Table B.76. Summary of formed PFCAs during degradation of n:2 fluorotelomer........... 394 Table B.77. Concentrations of PFASs in ground water described by detection frequency . 398 Table B.78. Concentrations of PFASs in surface water described by detection frequency . 399 Table B.79. Concentrations of PFASs in drinking water described by detection frequency 411 Table B.80. Concentrations of PFASs in precipitation described by detection frequency... 416 Table B.81. Concentrations of PFASs in sediment and soil ....................................... 419 Table B.82. Studies of EOF, AOF and/or TF in abiotic environmental samples ............... 422 Table B.83. Studies of EOF and/or TF in biota...................................................... 427 Table B.84. Concentrations of PFASs in air described by detection frequency ............... 429 Table B.85. Concentrations of PFASs in WWTP influent and effluent........................... 430 Table B.86. Concentrations of PFASs in WWTP sludge............................................ 431 Table B.87. Concentrations of PFASs in biota described by detection frequency ............ 432 Table B.88. Concentrations of PFASs in plants described by detection frequency........... 441 Table B.89. Summary of bioaccumulation studies in fish according to OECD 305 .......... 442 Table B.90. Uptake of PFASs in plants. .............................................................. 444 Table B.91. Serum (S), plasma (P) or whole blood (WB) concentrations (ng/mL) of PFAAs ............................................................................................................... 460 Table B.92. Serum (S), plasma (P) or whole blood (WB) concentrations (ng/mL) of PFAEs ............................................................................................................... 486 Table B.93. Breast milk concentrations (ng/mL) of PFASs ....................................... 493 Table B.94. Concentrations of PFAAs in urine, follicular fluid, semen.......................... 495 Table B.95. Concentrations of precursors, PFAEs and PFECHS in urine ....................... 500 Table B.96. Studies investigating the EA / ED of PFAS. .......................................... 503 Table B.97. TULAC – emission estimates (ERC assumptions). .................................. 527 Table B.98. Assumptions for backward facing usage rates....................................... 530 Table B.99. Default ERFs for ERCs 10a and 11a. .................................................. 535 Table B.100. Emission factors for “Other uses” during service life. ............................ 535 Table B.101. Summary of assumptions and factors applied to (emission) data. ............ 536 Table B.102. Annual emissions to air, water and soil ............................................. 539 Table B.103. Selected Environmental Release Categories (ERC). .............................. 540 Table B.104. Calculated one compartment based on default release factors (%) per ERC.540 Table B.105. Emission factors (%) to the environment........................................... 540 Table B.106. Selected Environmental Release Categories (ERC). .............................. 542 Table B.107. Calculated one compartment based on default release factors (%) per ERC.542 Table B.108. Emission factors (%) to the environment........................................... 543 Table B.109. ERC for the total release to the environment. ..................................... 544 Table B.110. ERC for the total release to the environment. ..................................... 545 Table B.111. Concentration of PFAS in indoor air (pg/m3) ....................................... 551 Table B.112. Concentrations (ng/g) of PFASs in indoor dust .................................... 555 FIGURES Figure B.1. Extract of Figure 1 of Wang et al. (2021c) ...............................................2 Figure B.2. Relative frequency of hazard classes regarding health effects ..................... 14 Figure B.3. Self-classifications on hazardous to the aquatic environment and ozone layer. 16 Figure B.4. Predicted biodegradability of the analysed PFAS categories and molecules ..... 20 Figure B.5. Example structure: perfluorohexane. .................................................... 21 Figure B.6. Example structure: chloropentafluoroethane........................................... 21 Figure B.7. Example structure: 1,1,2,2-tetrafluoro-1,2-bis(trifluoromethoxy) ethane. ...... 22 Figure B.8. Example structure: Perfluorobutanoic acid (PFBA). ................................... 23 Figure B.9. Example structure: Perfluorobutane sulfonic acid (PFBS). .......................... 25 Figure B.10. Example structure: Perfluorohexyl phosphonic acid................................. 26 Figure B.11. Example structure: Perfluamine (=perfluoro(tripropyl)amine).................... 27 v ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.12. Examples of substances with a combination of structural elements. ............ 29 Figure B.13. Examples of fluoropolymers.............................................................. 30 Figure B.14. Example of an n:2 fluorotelomer alcohol: 4:2 FTOH. ............................... 32 Figure B.15. Proposed 6:2 FTOH biotransformation pathways in aerobic sediment system. 33 Figure B.16. Aerobic degradation pathways of 8:2 FTOH in soil and activated sludge ....... 34 Figure B.17. Example of an n:2 fluorotelomer iodide: 4:2 FTI. ................................... 35 Figure B.18. Example of an ester of an n:2 FTOH ................................................... 35 Figure B.19. Example of an n:2 monoPAP/diPAP..................................................... 36 Figure B.20. Example of an n:2 fluorotelomer urethane monomer .............................. 36 Figure B.21. Example of an n:2 fluorotelomer sulfonic acid ....................................... 36 Figure B.22. Example of an n:2 fluorotelomer thioether amido sulfonate ...................... 37 Figure B.23. Example of an n:2 fluorotelomer silane ............................................... 37 Figure B.24. Example of an n:2 fluorotelomer olefin ................................................ 38 Figure B.25. Proposed atmospheric degradation pathway for n:2 fluorotelomer olefins ..... 38 Figure B.26. Example of an n:2 fluorotelomer-based side-chain fluorinated polymer ........ 38 Figure B.27. Example of an amide of a perfluoroalkyl carboxylic acid........................... 39 Figure B.28. Example of a n:1 fluorotelomer alcohol ............................................... 39 Figure B.29. Example of a perfluoroalkyl alcohol .................................................... 40 Figure B.30. Example of a perfluoroalkyl iodide...................................................... 40 Figure B.31. Example of a perfluorinated olefin ...................................................... 40 Figure B.32. Mechanism for the atmospheric oxidation of perfluorobut-2-ene ................ 41 Figure B.33. Example of a side-chain fluorinated aromatic: (Heptafluoropropyl)benzene. .. 41 Figure B.34. Structural formula of flurtamone. ....................................................... 42 Figure B.35 Structural formula of saflufenacil. ....................................................... 42 Figure B.36 Structural formula of fluazinam. ......................................................... 43 Figure B.37 Structural formulas of fluometuron (a); trifloxystrobin (b); cyflumetofen (c). . 43 Figure B.38. Pharmaceutical active substances ...................................................... 44 Figure B.39. Degradation pathways for fluoxetine based on the predicted intermediates ... 46 Figure B.40. Photolysis degradation pathways of TFM .............................................. 47 Figure B.41. Degradation routes of some key intermediates from fluorinated gases. ........ 49 Figure B.42. Degradation scheme of a selection of PFSA precursors ............................ 53 Figure B.43. Oxidation processes from sulphides/thiols to the corresponding PFSAs......... 54 Figure B.44. Generic structures of PFSA precursors. ................................................ 55 Figure B.45. Degradation pathway of 6:6 perfluoroalkyl phosphinic acid (6:6 PFPiA). ....... 55 Figure B.46. Inherently unstable key fluorinated compounds. .................................... 57 Figure B.47. Degradation pathway of trifluoromethanol. ........................................... 58 Figure B.48. Degradation pathway of trifluoromethylamine ....................................... 58 Figure B.49. Postulated hydrolysis via difluoromethanediol ....................................... 58 Figure B.50. Degradation pathways of 10-(trifluoromethoxy)decane-1-sulfonate ............ 60 Figure B.51. Degradation pathways of TFMHxOH .................................................... 61 Figure B.52. Degradation pathways of TFMPrOH..................................................... 62 Figure B.53. CYP-mediated degradation of OSI-930. ............................................... 62 Figure B.54. Degradation pathways of 4-(14C-trifluoromethoxy)benzoic acid.................. 63 Figure B.55. Representative examples of trifluoromethylamino-derivatives ................... 64 Figure B.56. Biodegradation pathways of 2,2-difluoro-1,3-benzodioxole (DFBD) ............. 65 Figure B.57. Fludioxonil. .................................................................................. 65 Figure B.58. Implications of high persistence......................................................... 68 Figure B.59. EOF (ng F/g) detected in different matrices .......................................... 88 Figure B.60. Time trends (percent per year) in suspended particulate matter .............. 101 Figure B.61. Five‐year moving average deposition fluxes of TFA, PFPrA, and PFBA ........ 102 Figure B.62. Time trends of trifluoroacetic acid (TFA) in leaves (µg/g dry weight) ......... 103 Figure B.63. Time trend of HFC-134a (+5.19 ppt per year) in air.............................. 104 Figure B.64. Measured BCFs and BAFs ............................................................... 125 Figure B.65. Reproduction of model for biota concentration development.................... 130 Figure B.66. Affected areas in contamination cases. .............................................. 136 Figure B.67. Estimated annual emissions of PFCA ................................................. 230 Figure B.68. Indicative PFAS tonnage flows (t/y) for paper & board packaging ............. 246 Figure B.69. Indicative PFAS (Fluoropolymers) tonnage flows (t/y) ........................... 248 vi ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.70. Loss of PFAS coating during use. ..................................................... 251 Figure B.71. Concentration PFAS found in rinsate of packaging containers (ppb)........... 252 Figure B.72. Material flow diagram for PFASs (all species) in ski-waxes ...................... 261 Figure B.73. Emissions of fluorinated gases from stocks 2018. Source EEA (2022). ....... 266 Figure B.74. Quantity of projected demand and emissions of sum HFCs/HCFCs ............ 267 Figure B.75. EU imports of fluorinated-gases within air-conditioning equipment............ 269 Figure B.76. Overview waste stage emissions. ..................................................... 300 Figure B.77. Schematic overview of PFAS emissions with waste treatment facilities ....... 301 Figure B.78. PFAS mass imbalance landfill. Source: SANBORN (2019)........................ 305 Figure B.79. PFAS exposure pathways to humans from Oliaei et al. (2013). ................ 325 Figure B.80. Default worst-case release factors.................................................... 529 Figure B.81. Source flow diagram for estimated emissions from tracer products. .......... 547 Figure B.82. Source flow diagram for estimated emissions from anti-foaming agents ..... 548 Figure B.83. Source flow diagram for estimated emissions (low scenario) from FP......... 549 Figure B.84. Source flow diagram for estimates emissions (high scenario) from FP........ 550 vii ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Annex B Information on hazard and risk B.1. Identity of the substance(s) and physical and chemical properties Please, refer to section 1.1.1. in the main report for the main definitions. In the following subsections only additional information is provided. B.1.1. Name and other identifiers of the substance(s) Per- and polyfluoroalkyl substances (PFASs) are a class of synthetic compounds that have attracted much public attention since the late 1990s, when the hazards and ubiquitous occurrence in the environment of two PFASs, perfluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS), started to be reported and recognized. Early communications used many different terminologies for what nowadays are called PFASs (e.g. per- and polyfluorinated chemicals, perfluorinated organics, perfluorochemical surfactants, highly fluorinated compounds). It is noted that the definition of PFASs historically encompasses both per- and polyfluoroalkyl substances, however the polyfluoroalkyl substances belong to the scope of the PFASs only when containing also at least one perfluorinated alkyl moiety (one fully fluorinated methyl or methylene group) and hence can also be called perfluoroalkyl substances. Polyfluoroalkyl substances which only contain partially fluorinated carbon atoms are not within the scope of the restriction proposal. OECD (2021) provides example molecular structures of included and excluded substances (see also Figure B.1 below). 1 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.1. Extract of Figure 1 of Wang et al. (2021c) illustrating examples of substances, which belong or do not belong to the scope of the PFASs . (Remark: The substance identified by CAS No. 24210-45-5 is a PFAS under the new OECD definition but out of scope for this restriction proposal by derogation.) The new OECD definition also mentioned in the Figure B.1 above and given in section 1.1.1. in the main report reads as: “PFASs are defined as fluorinated substances that contain at least one fully fluorinated methyl or methylene carbon atom (without any H/Cl/Br/I atom attached to it), i.e. with a few noted exceptions, any chemical with at least a perfluorinated methyl group (–CF 3) or a perfluorinated methylene group (–CF 2–) is a PFAS.” (OECD, 2021). B.1.2. Physicochemical properties Certain trends in the physicochemical properties within smaller, limited and structurally similar subgroups can be found. For some PFAS subgroups (e.g. Perfluoroalkylcarboxylic acids (PFCAs), Perfluoroalkane sulfonic acids (PFSAs), Perfluoroalkanes, Perfluoroalkylphosphonic acids (PFPAs) and Perfluoroalkylamines) the physicochemical properties of several corresponding molecules have been compiled (see Appendix B.1.2.). Within the individual subgroups (e.g. the PFCAs or the PFSAs), an increase in the log KOW and the boiling point can be observed with increasing chain length (equivalent to an increasing number of F atoms). In addition, a decrease in water solubility and vapour pressure of the substances can be observed with increasing chain length within the 2 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) homologous series of the subgroups. However, due to the wide scope of the proposed restriction entry and the large number of substances covered by this restriction proposal, general conclusions on the physicochemical properties of all PFASs are complicated or even impossible. See Table B.66 to Table B.72 in Appendix B.1.2. for the basic substance information and physical chemical properties of PFCAs, PFSAs, Perfluoroalkanes, Haloperfluoroalkanes and Perfluoroalkylethers, PFPAs and Perfluoroalkylamines. B.1.3. Justification for grouping Generally, due to the perfluoroalkyl moieties, PFASs are either very persistent themselves or degrade to form (over a short or long timescale) terminal degradation products which still contain one or several perfluoroalkyl moieties (rendering them very persistent). Hence, all members of the PFAS group share a common hazard and risk (described in sections 1.1.4. and 1.1.6. of the main report). For analogy, according to ECHA guidance R.11 (ECHA, 2017b), if transformation/degradation products with PBT/vPvB properties are generated, the substances themselves must be regarded as PBT/vPvB substances and should be treated like PBT/vPvB substances with regard to emission estimation and exposure control. If there are specific PFASs for which sufficient evidence is provided that the perfluorinated moiety is fully degraded at a rate which indicates them to be not persistent, resulting in a substance/substances which is/are not a PFAS, then those substances/groups should be considered excluded from the scope. The terminal PFAS degradation products are often referred to as arrowhead substances, while the parent substances degrading to the arrowheads are referred to as precursors (e.g. 6:2 FTOH is a precursor of PFHxA). The term related substance(s) is used interchangeably with the term precursor. Over sufficient time horizons all precursor substances will contribute to environmental stocks of their corresponding arrowhead substances (see section 1.1.4. of the main report for further details). PFASs have so far been subjected to regulatory risk management on a subgroup-bysubgroup basis (see section B.4.1). For the following PFASs the Committee for risk assessment (RAC) and the Committee for Socio-economic analysis (SEAC) adopted the suggested restrictions as appropriate on reducing the risk to human health and/or the environment due to the PBT/vPvB properties of the terminal PFAS degradation products: PFOA, its salts and PFOA-related substances (ECHA, 2015b); PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA including their salts and precursors (ECHA, 2018a); PFHxS including its salts and related substances (ECHA, 2019b); PFHxA, its salts and related substances (ECHA, 2019c). All of the above substances were also identified as SVHC (ECHA, 2012b; ECHA, 2012c; ECHA, 2012d; ECHA, 2012e; ECHA, 2013; ECHA, 2015b; ECHA, 2016b; ECHA, 2017c). Additionally, PFBS and HFPO-DA have been identified as SVHC (ECHA, 2019d; ECHA, 2019e). Due to the high number of PFAS subgroups (see section 1.1.1. of the main report) on the global market, it would take a significant amount of time to submit and process restriction proposals on all PFASs on a subgroup-by-subgroup basis, whereas the environmental stock of the very persistent PFASs would simultaneously continue to increase. Ban of single PFAS substances or subgroups may also lead to the substitution by other PFASs as the number of substances in this group is very high, so-called regrettable substitution. For some applications, production volumes may be low for specific PFASs (or even zero currently). Also, novel unregulated PFASs could theoretically be developed for these or other uses in the future. The precise identities of the PFASs currently used are oftentimes largely unknown due to manufacturer confidentiality. It is noted that the overall PFAS volume 3 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) across all uses is assumed to be significant (EC, 2020b). Consequently, the European Commission’s Chemicals Strategy for Sustainability (CSS) reiterates the concern for the class of PFAS substances and suggests a group approach under relevant regulations in order to address PFASs. They state their aim as phasing out persistent substances such as per- and polyfluoroalkyl substances (PFASs), unless their use is proven essential for society (EC, 2020a). Based on the above considerations, managing all PFASs together as a group is a clear benefit to environment and humans. A class-based approach has been chosen for the current restriction proposal in order to prevent the possibility for regrettable substitution. This dossier has put some weight on the link between the physicochemical properties of PFASs (persistence) and their environmental and toxicological effects. This is in line with the findings from the examination of strategies for grouping of PFASs by Cousins et al. (2020a), although these authors went one step further and recommended to regulate PFASs solely on the basis of persistence (“the P-sufficient approach”). The selected grouping approach in this dossier is based on the persistence of PFASs as its main concern. However, there are support ing properties triggering additional concerns in combination with persistence that add to the overall assessment. Those are other environmental or toxicological concerns like bioaccumulation, aqueous mobility, long-range transport, effects on humans or the environment, and high global warming potential. Cousins et al. (2022) compiled information about PFAS in rainwater and compared the concentrations found with various regulatory limit values. They found that, for PFOA and PFOS, levels in rainwater often greatly exceed Lifetime Drinking Water Health Advisory levels from the US Environmental Protection Agency (EPA) and concluded that for these substances the planetary boundary for chemical pollution has been exceeded. In a followup paper, the authors discussed the property of persistence in relation to the study on PFAS in rainwater (Scheringer et al., 2022). The authors argue that there is no need to consider chemical partitioning for persistent substances such as PFAS as the substances anyway cover virtually the entire partitioning space. T he continuous release of PFAS will lead to their accumulation somewhere in the environment until some known or unknown effect threshold is exceeded. Again, regulation of chemicals on the basis of persistence alone is the recommended approach. It is noted that the first example of regulation of PFASs as a chemical class according to the P-sufficient approach has been introduced in California. Here a regulation of PFASs as a class is in place for certain consumer products under the California Saf er Consumer Products Program (Balan et al., 2021). In a review paper, Cousins et al. (2016) looked at the precautionary principle and chemicals management in relation to PFAA contamination of groundwater. The authors argue that all PFASs entering groundwater, irrespective of their perfluoroalkyl chain length and bioaccumulation potential, will result in poorly reversible exposures and risks, as well as further clean-up costs for society. In order to protect groundwater resources for future generations, the authors call for a precautionary approach and prevention of use and release of highly persistent and mobile chemicals such as PFASs. For most of the investigated PFASs at least one of the mentioned additional hazardous properties applies. For the majority of PFASs, data is still lacking, but the current restriction dossier justifies that the probability for harmful effects for the less-studied PFASs, in addition to the intrinsic persistence, is sufficient for a preventive approach and a class -based restriction. A preventive approach of not using highly persistent synthetic organic subs tances is more protective and also overall less costly for society, both in terms of fewer tests and reduction in externalized societal costs including the expected costs of health care, loss of biodiversity, loss of ecosystem services, loss of property value and remediation (Cousins et al., 2020b). 4 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The proposed scope definition is in line with previous PFAS restriction proposals with an analogue approach to, e.g. PFOA, PFHxS and PFHxA with a scope definition based on a molecular structure formula. The inclusion of perfluorinated alkyl moieties as short as one perfluorinated carbon atom is in agreement with the definition of PFASs according to the UNEP/OECD Global PFC Group. Trifluoromethyl fragments are also linked via degradation to trifluoroacetic acid (TFA) which has been demonstrated to be a persistent substance with harmful properties in a comparable way as other PFASs with longer fluorinated alkyl chains. Support for the outlined justification for grouping may be found in the scient ific literature, and in particular in the key papers of Cousins et al. (2020a); Kwiatkowski et al. (2020); Wang et al. (2017b). In summary, the grouping is based on structural similarity (common perfluoroalkyl moieties) that triggers equivalent hazards and risks among the substances covered, primarily related to the very persistent property of the substances. However, the grouping is also justified by the desire to avoid regrettable substitution and prevention of future exposures of those PFASs which are not currently in use. B.1.3.1. Naturally occurring organofluorine substances In a review article on fluorine-containing natural products from 1999, O'Hagan and Harper explain that although ca. 3000 secondary metabolites containing the halogens chlorine, bromine and iodine have been reported, only 13 secondary metabolites containing fluorine have been discovered. This is in contrast to fluorine being the most abundant halogen in the earth's crust (O’Hagan and Harper, 1999). The majority of the natural fluorinecontaining substances are fatty acids with a single fluorine atom at the end of the carbon chain. None of the reported substances were per- or polyfluorinated. On several of the major continents, plants have been found that biosynthesise the highly toxic monofluoroacetate, presumably for the purpose of defence. In a study of the concentrations of trifluoroacetate (TFA) in ocean waters, Frank et al. (2002) concluded that TFA in oceans have mostly a natural origin, while in the atmosphere, precipitate, freshwaters and needles of conifers, TFA most likely stems from anthropogenic sources. The authors stated that the total amount of trifluoroacetate (TFA) present in the global environment greatly exceeds what may be expected to be contributed from various industrial sources. Scott et al. (2005) further investigated whether TFA concentrations in the marine environment could have natural sources by determining a series of depth profiles of TFA in the Arctic, North and South Atlantic, and Pacific Oceans. They concluded that underwater vents could contribute to the TFA concentrations in the oceans. It was indicated that the heterogeneous distribution of TFA can only be partially explained by recent anthropogenic sources, while the total inventory of TFA in the oceans cannot be explained entirely by human activities. TFA in freshwaters is thought to have solely anthropogenic sources, while TFA found in oceans may be of both natural and anthropogenic origin (Fleet D et al., 2017). Solomon et al. (2016) investigated the sources, fates, toxicity, and risks of TFA and its salts. They concluded that TFA is both produced naturally and synthetically and that it could be a breakdown product of more than 1 million chemicals, including refrigerants, pharmaceuticals, pesticides and polymers. The study indicates an annual input of 6 T from natural sources, while the total cumulative contribution of TFA to the environment from HFCs/HFOs (refrigerants) from 1990 to 2050 was estimated at more than 20 million tonnes. However, in a classical risk assessment the authors concluded that it is not likely that the environmental concentrations of TFA will exceed the effect levels as TFA has only a low to moderate toxicity. The evidence for natural TFA was carefully examined by Joudan et al. (2021), and they came to a different conclusion with regards to the origin of TFA in the oceans. Their methods included: (i) critical evaluation of measurements of TFA in pre-industrial samples; 5 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) (ii) examination of the likelihood of TFA formation by hypothesized mechanisms; (iii) exploration of other potential TFA sources to the deep ocean; and (iv) examination of global budgets of TFA. The authors concluded that the presence of TFA in t he deep ocean and lack of closed TFA budget is not sufficient evidence that TFA occurs naturally, especially without a reasonable mechanism of formation. Zhai et al. (2015) measured the concentrations of TFA in urban landscape waters, tap water and snow in Beijing, China. A comparison between 2002- and 2012-values demonstrated a 17-fold increase from 23–98 ng/L to 345–828 ng/L in urban landscape waters. In the same period an increase of TFA from not detected to 155 ng/L occurred in tap water. TFA in precipitation was measured by Freeling et al. (2020) in samples collected in Germany over one year. The article points to anthropogenic sources, and in particular formation of TFA in the atmosphere by photodegradation of certain hydrofluorocarbons (HFCs), hydrochlorofluorocarbons (HCFCs), and unsaturated hydrofluorocarbons (hydrofluoroolefins, HFOs) as sources of atmospheric TFA. Their findings indicate a considerable increase in the atmospheric deposition of TFA in Germany over the last two decades. Substitution of HFCs and HFOs by halogen-free gases was suggested as an effective measure to reduce the TFA load in precipitation, as HFCs and HFOs are considered as rising sources. In summary, the number of naturally occurring organic fluorine-containing substances is low compared to other halogenated substances. TFA has been indicated to have natural sources in oceans (underwater vents) but evidence for the existence of natural trifluoroacetic acid is considered insufficient. Oceans are considered the final environmental sink of the substance. However, natural sources for TFA in the oceans have been questioned. TFA in the atmosphere, precipitate, freshwaters and needles of conifers most likely stems from anthropogenic sources. Concentrations of TFA in urban waters and tap water have been increasing over the last decades. 6 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.2. Manufacture and uses (summary) PFASs consist of substances with very variable characteristics (gaseous, liquids and solids) and thus are applied in a wide variety of applications as can be seen in Table B.1. Especially PFAS polymers have a very broad application in hundreds if not thousands of sub-applications. In Table B.1 detailed researched PFAS applications (including PFAS manufacturing and waste stage) as well as less researched PFAS applications are presented. A detailed description can be found in Annex A. Table B.1. PFAS applications and researched uses. PFAS applications PFAS manufacturing Textile, upholstery, leather apparel and carpets Food contact materials and packaging Metal plating and manufacturing of metal products C onsumer mixtures C osmetics Ski wax Medical devices Transport C onstruction products Lubricants Electronics and semiconductors Petroleum and mining Applications of fluorinated gases Energy Laboratory equipment & filtration Medicinal products Plant protection products and biocides Plastics (other than packaging) and rubber/elastomer production (including flame retardants) C hemical industry Pyrotechnics Personal care products other than cosmetics Fracking (currently hardly applicable in EEA) Immersion cooling (currently not a use in the EEA) Defence industry Printing inks Other niche appications Uses (yet) unknown C ement industry     Professional cleaning and polishing Green uses are researched in detail Blue uses are researched in general Orange uses not researched in detail Purple use: Separate restriction proposal Waste stage PFAS applications Firefighting foam 7 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.3. Classification and labelling B.3.1. Classification and labelling in Annex VI of Regulation (EC) No 1272/2008 (CLP Regulation) Harmonised classifications on human health hazards: In a screening of Annex VI of CLP for PFASs with harmonised classifications, human health endpoints carcinogenicity (Carc.), mutagenicity (Muta.), reproductive toxicity (Repr.) including effects on or via lactation (Lact.), and specific target organ toxicity following repeated exposure (STOT RE) were considered of most concern following long -term exposure. In total, 41 PFASs were identified at the time of screening (Q4 2020) as having such a classification for one or more of these five endpoints. Since the time of screening, the Committee for Risk Assessment (RAC) has evaluated classification proposals for additional PFASs. PFHpA is listed as Repr. 1B and STOT RE since February 2022 and 6:2 FTOH is included in Table B.2 indicated as “CLH proposal agreed”, because so far (June 2022) it has not been officially inserted in Annex VI of CLP. Only 6:2 FTOH was considered in the assessment for Table B.3 and Table B.4 and Figure B.2, below (since it was included already in ECHA dataset from 2020). Thus, 43 PFASs in sum are listed in Table B.2, having a classification for one or more of the above mentioned five endpoints. Please note that most of these substances have additional harmonised classifications for other endpoints (human health, environment and/or physicochemical properties) as well; these are however not listed in Table B.2, which is limited to Carc., Muta., Repr., Lact. and STOT RE classifications. Of further note, a number of the harmonised classifications were based on read-across and not on actual data on the substance. The list contains a number of PFASs known to be used as active substances in plant protection products and biocides that are further known to be TFA precursors. Table B.2. 43 PFASs with harmonised classification for carcinogenicity (Carc.), mutagenicity (Muta.), reproductive toxicity (Repr.), effects on or via lactation (Lact.) and/or specific target organ toxicity following repeated exposure (STOT RE). PFASs known as active substances in plant protection products (PPP) and biocidal products (BP) are listed in the second part of the table. EC Number CAS number Substance name/abbr. 206-397-9 335-67-1 PFOA 206-400-3 335-76-2 PFDA 206-801-3 375-95-1 PFNA 217-179-8 1763-23-1 PFOS 220-527-1 2795-39-3 PFOS-Potassium salt 223-320-4 3825-26-1 APFO (PFOA Ammonium salt) 249-415-0 29081-56-9 PFOS Ammonium salt Harmonised classification for Carc./Muta./Repr./Lact . and/or STOT RE C arc. 2; H351 / Repr. 1B; H360D / Lact.; H362 / STOT RE 1; H372 [liver] C arc. 2; H351 / Repr. 1B; H360Df / Lact.; H362 C arc. 2; H351 / Repr. 1B; H360Df / Lact.; H362 / STOT RE 1; H372 [liver; thymus; spleen] C arc. 2; H351 / Repr. 1B; H360D / Lact.; H362 / STOT RE 1; H372 C arc. 2; H351 / Repr. 1B; H360D / Lact.; H362 / STOT RE 1; H372 C arc. 2; H351 / Repr. 1B; H360D / Lact.; H362 / STOT RE 1; H372 [liver] C arc. 2; H351 / Repr. 1B; 8 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) EC Number CAS number Substance name/abbr. Harmonised classification for Carc./Muta./Repr./Lact . and/or STOT RE H360D / Lact.; H362 / STOT RE 1; H372 C arc. 2; H351 / Repr. 1B; H360D / Lact.; H362 / STOT RE 1; H372 249-644-6 29457-72-5 PFOS Lithium salt 274-460-8 70225-14-8 PFOS Diethanolamine C arc. 2; H351 / Repr. 1B; H360D / Lact.; H362 / STOT RE 1; H372 402-190-4 113674-956 STOT RE 2; H373 / STOT RE 1; H372 406-740-4 1939-27-1 4-(2-chloro-4trifluoromethyl)phenoxy-2fluoroaniline hydrochloride 3'-trifluoromethylisobutyranilide STOT RE 2; H373 407-810-7 - (PFAS mixture) STOT RE 2; H373 413-640-4 - (PFAS mixture) STOT RE 2; H373 415-300-0 90076-65-6 lithium bis(trifluoromethylsulfonyl)imide STOT RE 2; H373 415-500-8 145963-844 2-amino-4-dimethylamino-6trifluoroethoxy-1,3,5-triazine STOT RE 2; H373 421-080-7 161462-357 STOT RE 2; H373 422-500-1 - 1-cyclopropyl-3-(2-methylthio-4trifluoromethylphenyl)-1,3propanedione (PFAS reaction mass) 423-180-6 - (PFAS mixture) STOT RE 2; H373 424-520-6 65753-47-1 2-chloro-3-trifluoromethylpyridine STOT RE 1; H372 427-880-2 90357-53-2 STOT RE 2; H373 429-560-8 4274-38-8 432-190-1 182176-529 214353-170 N-[4-cyano-3trifluoromethylphenyl]methacrylamid e 2-amino-4(trifluoromethyl)benzenethiol hydrochloride (PFAS reaction mass) 433-580-2 STOT RE 2; H373 STOT RE 2; H373 STOT RE 2; H373 1-(2-amino-5-chlorophenyl)-2,2,2trifluoro-1,1-ethanediol, hydrochloride [containing ≥0.1% 4chloroaniline (EC No 203-401-0)] C arc. 1B; H350 C arc. 2; H351 / Repr. 1B; H360Df / Lact.; H362 / STOT RE 1; H372 [liver; thymus; spleen] C arc. 2; H351 / Repr. 1B; H360Df / Lact.; H362 C arc. 2; H351 / Repr. 1B; H360Df / Lact.; H362 C arc. 2; H351 / Repr. 1B; H360Df / Lact.; H362 / STOT RE 1; H372 [liver; thymus; spleen] - 21049-39-8 PFNA Sodium salt 221-470-5 3108-42-7 PFDA Ammonium salt - 3830-45-3 PFDA Sodium salt - 4149-60-4 PFNA Ammonium salt PFASs with harmonised classification agreed after Q4 2020: 9 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) EC Number CAS number Substance name/abbr. Harmonised classification for Carc./Muta./Repr./Lact . and/or STOT RE 206-798-9 375-85-9 PFHpA 211-477-1 647-42-7 6:2 FTOH Repr. 1B, H360D / STOT RE1 [liver] C LH proposal agreed (STOT RE 2; H373 [teeth, bone] PFASs known as active substances in plant protection products a./o. biocidal products (approved or not approved) 216-428-8 1582-09-8 Trifluralin (containing <0.5 ppm C arc. 2; H351 NPDA)* 274-125-6 69806-50-4 Fluazifop-butyl* Repr. 1B; H360D 405-090-9 67485-29-4 Hydramethylnon* STOT RE 1; H372 417-680-3 101463-698 Flufenoxuron** Lact.; H362 421-960-0 90035-08-8 Flocoumafen** Repr. 1B; H360D / STOT RE 1; H372 [blood] 424-610-5 120068-373 126535-157 141112-290 142459-583 335104-842 Fipronil** STOT RE 1; H372 Trisulfuron-methyl* C arc. 2; H351 Isoxaflutole* Repr. 2; H361d Flufenacet* STOT RE 2; H373 Tembotrione* 614-708-8 68694-11-1 Triflumizole* 616-669-2 79241-46-6 Fluazifop-P-butyl* Repr. 2; H361d / STOT RE 2; H373 [eyes; kidneys; liver] Repr. 1B; H360D / STOT RE 2; H373 [liver] Repr. 2; H361d 616-712-5 79622-59-6 Fluazinam* Repr. 2; H361d 617-373-6 82657-04-3 Bifenthrin*/** C arc. 2; H351 / STOT RE 1; H372 [nervous system] 603-146-9 604-222-4 604-290-5 608-879-8 * = PPP; ** = BP 10 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.3.2. Classification and labelling in classification and labelling inventory/ Industry’s self-classification(s) and labelling 1 Self-classifications on human health hazards: Of the approximately 6 800 PFASs listed in the ECHA database, around 6 600 PFASs have a (self-)classification indicated in the registrations or notifications for at least one environmental, human health and/or physicochemical endpoint at the time of screening (Q4 2020). Among these, 316 substances (additional to the 41 (PFHpA and 6:2 FTOH not included) already identified in B.3.1) have a self-classification for one or more of the five human health endpoints considered of most concern following long term exposure of humans to PFASs (Carc., Muta., Repr., Lact., STOT RE). Table B.3 presents the total of 357 PFASs classified for Carc., Muta., Repr., Lact. and/or STOT RE (mostly self -classifications), subdivided at screening into PFAS categories and subsequently re -arranged into ‘arrowheads’ and ‘possible PFAA precursors’, as per Figure 1 in section 1.1.1. of the main report. Around 200 PFASs from ECHA’s database have no classification indicated for ecotoxicological, toxicological and/or physicochemical properties in either the registrations or notifications. This might be because of absence of data and not because of data showing that classification is not required. Table B.3. PFASs (subdivided into PFAS categories) with harmonised or self-classification (ECHA database status: Q4 2020) for carcinogenicity (Carc. ), mutagenicity (Muta.), reproductive toxicity (Repr.), effects on or via lactation (Lact.) and/or specific target organ toxicity following repeated exposure (STOT RE), re -arranged into ‘arrowheads’ and ‘possible PFAA precursors’. PFASs can be classified f or multiple endpoints of concern; thus, the sum of percentages of total PFASs classifications can exceed 100%. EC-no. ECHA categories Counts Possible PFAA precursors Carc. Muta. Repr. STOT RE Lact. 13 C nF2n* 3 5 9 9 0 5 C omplex** 3 0 1 1 0 5 Fluorotelomer alcohol 1 0 2 4 0 2 Fluorotelomer epoxides 2 2 0 0 0 1 Fluorotelomer methacrylates (other) 0 0 0 1 0 4 Hydrofluoroethers 2 0 1 2 0 2 Hydrofluoroolefins 0 2 0 1 0 4 n:1 Fluorotelomer alcohols (FTOHs) 1 0 2 2 0 27 n:1 Fluorotelomer-based nonpolymers 4 2 15 13 1 1 n:1 FT (meth)acrylate 1 0 0 0 0 1 n:2 Fluorotelomer acrylates 0 0 0 1 0 1 n:2 Fluorotelomer alcohols 0 0 1 1 0 11 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) EC-no. Counts 1 ECHA categories Carc. Muta. Repr. Lact. 1 STOT RE 0 n:2 Fluorotelomer ethoxylates 0 1 1 n:2 Fluorotelomer methacrylates 0 0 0 1 0 4 n:2 Fluorotelomer olefins 0 0 1 4 0 1 n:2 Fluorotelomer phosphate esters 0 0 0 1 0 2 n:2 Fluorotelomer silanes 0 0 1 2 0 6 n:2 Fluorotelomer sulfonic acids 0 0 0 6 0 1 n:2 Fluorotelomer sulfonyl based compounds 0 0 0 1 0 5 n:2 Fluorotelomer-based nonpolymers 0 0 0 5 0 1 n:2 Fluorotelomer-thiol derivatives 0 0 1 1 0 4 Other carbonyl-based non-polymers 3 0 0 2 0 20 Other fluorotelomer-based nonpolymers 2 0 16 9 8 9 Other per- and polyfluoroalkyl ether based substances 0 0 2 7 1 4 Pther sulfonyl-based non-polymers 0 0 0 4 0 2 Perfluoroalkenes 1 0 0 2 0 5 Perfluoroalkyl carbonyl amides 3 1 3 1 0 2 Perfluoroalkyl carbonyl halides 0 0 0 2 0 2 Perfluoroalkyl carboxylic acids (PFC A) esters 0 1 1 0 1 1 Perfluoroalkyl epoxides 0 0 0 1 0 4 Perfluoroalkyl iodides 0 0 2 3 0 2 Perfluoroalkyl ketones 0 0 2 1 0 5 Perfluoroalkane sulfonamides 0 0 2 4 0 2 Perfluoroalkane sulfonamidoethanols 0 0 1 2 0 1 Per- and polyfluoroether carboxylic acids (PFEC As) halides 0 0 0 1 0 0 12 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) EC-no. Counts 1 ECHA categories Carc. Muta. Repr. Lact. 1 STOT RE 1 Perfluoroalkane sulfonyl halides 0 0 7 Perfluoroalkane sulfonic acids (PFSA) esters 2 2 2 2 2 154 Side-chain fluorinated aromatics 25 14 81 89 14 2 Perfluoroalkyl phosphate esters (PAPs) 0 0 0 2 0 315 Possible PFAA precursors 53 30 148 189 27 88% Of total PFAS (n=357) 17% 10% 46% 60% 9% 0 Arrowheads (PFAAs) 4 No EC HA category, OEC D category: perfluoroalkyl carboxylic acids (PFC As) + salts 4 0 4 2 4 5 Per- and polyfluoroether carboxylic acids (PFEC As) 0 0 1 4 0 8 Perfluoroalkyl carboxylic acids (PFC As) + salts 5 0 7 4 4 10 Perfluoroalkane sulfonic acids (PFSAs) + salts 8 0 7 8 6 27 Arrowheads (PFAAs) 17 0 19 18 14 8% Of total PFAS (n=357) 63% 0% 70% 67% 52% Not assignable to a category (currently) 10 No EC HA category, no OEC D category, entry in EC HA list 3 0 3 7 0 5 Other PFAS 0 2 2 3 0 15 Not assignable to a category 3 2 5 10 0 4% Of total PFAS (n=357) 20% 13% 33% 67% 0% * The PFAS category ‘C nF2n’ refers to PFASs containing a -C F2- moiety. These substances fulfil the PFAS definition in section 1.1.1. of the main report. ** The PFAS category ‘complex’ refers to metal complexes. The 357 substances fall into 43 PFAS categories in total. The listed PFASs are most often classified for STOT RE (217), followed by Repr. (172), Carc. (73), Lact. (41), and Muta. (32) (Table B.3 and Figure B.2). 13 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Lact.; 11% Carc.; 20% Muta.; 9% STOT RE; 61% Repr.; 48% Figure B.2. Relative frequency of hazard classes regarding health effects of 357 PFASs on the five endpoints of concern (Carc. = Carcinogenicity, Muta. = Mutagenicity, Repr. = Reproductive toxicity, Lact. = effects on or via lactation, STOT RE = specific target organ toxicity following repeated exposure). The fact that several thousands of PFASs do not bear (self-)classifications for the endpoints of most concern (Carc., Muta., Repr., Lact. and/or STOT RE) does not mean that these PFASs do not have these properties, but most likely that study data are lacking for the majority of them to base classification on. Hence, the fact that the current list of PFASs with (self-)classifications already includes so many different PFAS categories suggests that PFASs currently without (self-)classifications may exhibit one or more similar properties of concern. From Table B.3, it can be seen that the majority (88%) of the 357 substances classified for Carc., Muta., Repr., Lact. and/or STOT RE are possible PFAA precursors (such as fluorotelomers, perfluoroalkyl carbonyl amides, hydrofluoroethers, etc.) including potential TFA precursors. Moreover, these PFASs are not only PFAA precursors, because the groups are very diverse and degradation pathways are complex/uncertain. Repeated exposure to the classified PFASs affected various organs, such as liver, kidney, thymus, endocrine system, immune system, nervous system, respiratory system, spleen, blood, heart and cardiovascular system, brain, bone marrow, skin, lymph nodes, testicles, uterus, and gastrointestinal tract. Target effects of reproductive and developmental toxicity are reported as adverse effects on e.g. fertility, pup survival, offspring viability and on the foetal skeleton. Carcinogenicity and mutagenicity in categories 1 and 2 are reported as well as effects on or via lactation. Further details are listed in Table B.4. 14 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.4. Details on available classification information for the endpoints of concern for classified PFASs in two main PFAS classes, and for not assignable PFASs. PFASs can be classified for multiple endpoints of concern; thus, the sum of percentages of total PFASs classifications can exceed 100%.Further details are listed in Table B.3. Number (percentage) PFAS class of 357 Carc. Muta. Repr. Lact. STOT RE STOT RE affected organs: classified PFASs Possible PFAA precursors incl. TFA precursors (incl. telomers, epoxides, 315 (88%) halides, fluorinated gases, olefins, esters, sidechain aromatics, etc.) 15% 8% 41% 8% 53% Damage to multiple tissues and organs, for example: liver, hepatobiliary system, kidneys, adrenal glands, lower uri-nary tract, blood, nervous system, heart, respiratory sys-tem, reproductive organs, thyroid, immune system, thy-mus, spleen, endocrine system, bones, eyes, gastrointesti-nal tract, teeth. Arrowheads (incl. carboxylic, sulfonic acids and salts) 27 (8%) 63% 0% 70% 52% 67% Damage to organs; liver, kidneys, blood, lung, central nervous system, cardiovascular system. Not assignable PFASs 15 (4%) 20% 13% 33% 0% 67% Liver, kidneys, adrenals, renal system, ovary, testes, gastrointestinal tract, nervous system, hematopoietic system, immune system, respiratory system. C arc. = carcinogenicity, Muta. = mutagenicity, Repr. = reproductive toxicity, Lact. = effects on or via lactation, STOT RE = specific target organ toxicity – repeated exposure 15 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Self-classifications on environmental hazards: 1 129 PFASs of the 6 800 PFASs listed in the ECHA database on registration and C&L notification have a self-classification on hazardous to the aquatic environment (Aquatic Acute and/or Aquatic Chronic) or hazardous to the ozone layer. Further 4 substances have selfclassifications on hazardous to the aquatic environment (Aquatic Chronic 2 and 3) as well as to the ozone layer. The PFASs are most often classified for Aquatic Chronic 4 (444) followed by Aquatic Acute 1 + Aquatic Chronic 1 (322), Aquatic Chronic 2 (135) and Aquatic Chronic 3 (110). Further self-classifications are shown in Figure B.3. For information on the evaluation of the database: If for one substance different self-classifications on long-term aquatic hazard (different categories on Aquatic Chronic) were listed, the classification with the most entries were chosen. If the number was the same, the more stringent was selected. Aquatic Acute 1 + Aquatic Chronic 1 322 Aquatic Acute 1 + Aquatic Chronic 2 13 Aquatic Acute 1 + Aquatic Chronic 3 3 solely Aquatic Acute 1 53 solely Aquatic Chronic 1 30 solely Aquatic Chronic 2 135 solely Aquatic Chronic 3 110 solely Aquatic Chronic 4 444 Hazardous to the Ozone Layer (Category 1) 23 0 50 100 150 200 250 300 350 400 450 500 Figure B.3. Self-classifications on hazardous to the aquatic envir onment and ozone layer. 16 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4. Environmental fate properties B.4.1. Degradation B.4.1.1. Degradation - general There is a broad diversity in the chemical and physical properties of PFASs. However, despite this high diversity, members of the PFAS group have in common that the y contain perfluoroalkyl moieties that are extremely resistant to environmental and metabolic degradation. There are a few exceptions to this general rule, see section B.4.1.4. The resistance to degradation of the perfluoroalkyl moiety is primarily due to the high electronegativity and low polarisability of fluorine, which results in the strongest covalent bond known in organic chemistry: the C-F bond (Kissa, 2001). The C-F bond is resistant to acids, bases, oxidation and reduction, and even high temperatures. Multiple C–F bonds on the same geminal carbon lead to additional strengthening of the C–F bond. The strong electron- withdrawing effect of the fluorine atoms in perfluoroalkyl moieties also strengthens the skeletal bonds in the carbon chain (Cousins et al., 2020b). It is not expected that the length of the perfluoroalkyl chain has any major impact on the inherent stability of PF ASs. As illustrated in Figure 1 in section 1.1.1. in the main report, many PFASs have non-fluorinated moieties attached to the perfluorinated moiety. During degradation processes non-fluorinated moieties of molecules are transformed and oxidative processes often lead to a gradual conversion of non-fluorinated carbon atoms to CO2, which is emitted to the atmosphere while the degrading substance structure is gradually getting smaller. In the end, most of the nonfluorinated parts are fully degraded while the perfluoroalkyl part remains, often attached to a functional group at its highest oxidation state (e.g. carboxylic acid). For additional details, see section B.4.1.3. Cousins et al. (2020a) considered that PFASs in general are, or ultimately transform into, stable substances, often PFAAs. Wang et al. (2017b) explained that perfluoroalkyl (CnF 2n+1-) and perfluoroether (CnF 2n+1−O−CmF 2m+1−) moieties are very persistent under natural conditions. Even though some PFASs may partially degrade in abiotic and biotic degradation processes, they will all ultimately transform into highly stable end products, which are usually the very persistent perfluoroalkyl or perfluoroalkyl(poly)ether acids (here collectively termed “PFAAs”), for example, PFCAs, PFSAs, PFECAs, and PFESAs. In laboratory photochemical experiments it has been possible to achieve degradation of PFCAs (for C1 to C4 PFCAs) in reaction with OH radicals. However, this transformation was found to be a minor pathway in the environment , while the main atmospheric removal mechanism is via wet and dry deposition which occurs in a timescale of the order of 10 days (Hurley et al., 2004a). Liou et al. (2010) investigated the biodegradability of PFOA and found the substance to be microbiologically inert and environmentally persistent. If PFAAs degrade, they do it so slowly that it is not observable and their half-lives could be in the order of decades, centuries or even greater (Parsons et al., 2008). Parsons et al. (2008) reviewed the biodegradation of perfluorinated compounds. The authors pointed out that the most theoretically plausible degradation pathway for PFASs is via reductive defluorination, which could occur under anaerobic conditions. The same authors reported for PFOS that no biodegradation was observed under aerobic conditions, while there were some observations of degradation of PFOS under anaerobic conditions though no metabolites were measured in these studies. In principle, it cannot be ruled out that some degradation of other PFASs under anaerobic conditions can occur (e.g. in hypoxic groundwater, marine water or sediments), or will occur in the future if bacteria adapt to utilise the energy present in the PFAS substrates. Indications for such bacterial behaviour were found when a PFOA-degrading bacterial strain was isolated from soil near a PFAS production plant (Yi et al., 2016). The PFOA-degradation has been demonstrated at lab conditions with a low degradation efficiency only, and have yet to be observed at significant rates in the 17 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) environment (Arp and Slinde, 2018). B.4.1.2. Key arrowhead subgroups In the present section lack of degradation of selected PFAS subgroups is considered in more detail. The subgroups have been selected on the basis of their expected stability in the environment, and they are representing arrowhead substances or arrowhead subgroups, i.e. final degradation products that do not undergo any further degradation in the environment. The arrowhead approach may be defined as a risk management approach when a representative PFAS is managed together with its salts and precursors. The approach has been recognized by scientists (e.g. (Cousins et al., 2020a)) and represents the dominant current approach to grouping PFASs for risk assessment and risk management globally. It is noted that the information presented below covers PFASs where the perfluoroalkyl moieties constitute the largest part of the substance. However, there are many PFASs in which the perfluoroalkyl moiety contributes to a relatively small part of the substance compared to the non-fluorinated parts. For these substances it can generally be expected that the primary degradation will target the non-fluorinated part, see section B.4.1.3. B.4.1.2.1. QSAR modelling of degradation The persistence of selected subgroups was investigated with QSAR modelling of abiotic/biotic degradation of three representative members of the different subgroups (apart from haloperfluoroalkanes). The QSAR models used in this study were selected according to their capacity and competence of predicting abiotic and biotic degradation of the selected PFASs. For the purpose of this study, capacity means that the substance or similar substances are part of the training set of the model, while competence refers to the applicability domain (AD) for the endpoint to be predicted. For additional details on the training sets for the QSAR models and their reliability, see Appendix B.4.1.2. A complete list of substances selected for QSAR modelling is found in Table B.5. Table B.5. List of substances selected for QSAR modelling of degradation. PFAS category PFAS name CAS number Perfluoroalkyl carboxylic acids (PFC As) Perfluoroalkane sulfonic acids (PFSAs) Perfluoroalkyl phosphonic acids (PFPAs) Perfluoroalkanes PFOA 335-67-1 PFHxA 307-24-4 PFBA 375-22-4 PFOS 1763-23-1 PFHxS 355-46-4 PFBS 375-73-5 Perfluorooctyl phosphonic acid (PFOPA) Perfluorohexyl phosphonic acid (PFHxPA) Perfluorobutyl phosphonic acid (PFBPA) Perfluorohexane 40143-78-0 Perfluorooctane Perfluoroalkylamines 40143-76-8 52299-24-8 355-42-0 307-34-6 Perfluorodecaline (perflunafene) 306-94-5 Perfluamine 338-83-0 Perfluoromethyldiethylamine 758-48-5 Perfluorotrihexylamine 432-08-6 18 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS category PFAS name CAS number Perfluoroalkyl ethers (PFAEs) Perfluorodiethylether, C F3C F2-OC F2C F3 358-21-4 C F3-O-C F2-C F2-O-CF3 378-11-0 2,2,3,3,4,4,5-heptafluorotetrahydro5-(nonafluorobutyl)furan 335-36-4 B.4.1.2.2. Abiotic degradation For abiotic degradation, the modelling suites used were AOPWIN and HYDROWIN from EPI Suite, which predict the atmospheric and water degradation, respectively; OPERA, which predicts the hydroxylation rate; and VEGA, which predicts persistence in air, water, sediment, and soil in seven (7) different models. Abiotic degradation predictions were of low reliability for all analysed PFASs in water, sediment and soil. The abiotic degradation models have a very low coverage of perfluorinated compounds in their training sets, and overall, the models were found unsuitable to reliably predict photodegradation of PFASs. Hence, the Dossier Submitters do not recommend the use of any of the models investigated to estimate abiotic degradation of PFASs, as the currently available versions of the QSARs are not adapted to accurately model perfluoroalkyl compounds. B.4.1.2.3. Biotic degradation For biodegradation, the modelling suites used were BIOWIN v4.11 of EpiSuite, which incorporates seven (7) different models to predict different endpoints related to biodegradability; OPERA, which predicts biodegradation and ready biodegradability; and VEGA, which predicts ready biodegradability. BIOWIN1 (linear probability model) and BIOWIN2 (non-linear probability model) are intended to convey a general indication of biodegradability under aerobic conditions. BIOWIN3 (expert survey ultimate biodegradation model) and BIOWIN4 (expert survey primary biodegradation model) rate the ultimate and primary biodegradation of each compound on a semiquantitative scale of 1 (longer than months) to 5 (hours). Primary biodegradation is the transformation of a parent compound to an initial metabolite. Ultimate biodegradation is the transformation of a parent compound to carbon dioxide and water, mineral oxides of any other elements present in the test compound, and new cell material. BIOWIN5 (MITI linear model) and BIOWIN6 (MITI non-linear model) are predictive models for assessing a compound’s biodegradability in the Japanese MITI ready biodegradation test (OECD 301C). The critical biodegradation evaluations (results of the MITI tests) are either "readily degradable" (value of 1) or "not readily degradable” (value of 0). 0 to 1 is the full probability range. BIOWIN7 (anaerobic biodegradation model) estimates the probability of fast biodegradation under methanogenic anaerobic conditions; specifically, under the conditions of the "serum bottle" anaerobic biodegradation screening test. This endpoint is assumed to be predictive of degradation in a typical anaerobic digestor. The screening criteria for persistence in the environment are BIOWIN2 <0.5 or BIOWIN6 <0.5 and BIOWIN3 <2.25 as described in ECHA’s Guidance on Information Requirements and Chemical Safety Assessment – Endpoint specific guidance (ECHA, 2017b). Apart from the fragment [-CF 3], which is included in the training sets of BIOWIN 1 to 4, BIOWIN is generic when it considers “C bonded to atoms other than H”, not being specific to C-F bonds. In addition, the fragment [-F] is included only in the training sets of BIOWIN 5 and 6 (MITI models). These limitations hamper BIOWIN’s biodegradability prediction reliability, and the outputs should be interpreted individually and with caution. 19 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) BIOWIN may be used as a supporting tool for aerobic biodegradability predictions, if results are interpreted individually, and all limitations stated. In addition, due to the apparent lack of [-F] and [-CF3] fragments in the training set of BIOWIN 7, the prediction of PFASs anaerobic biodegradation using BIOWIN should also be interpreted with great care. OPERA and VEGA returned a low reliability in the predictions of PFAS biodegradability for all subclasses. The results from these models are therefore not used in the assessment and only outputs from BIOWIN (for which the applicability domain is not explicit in the outputs) will be further discussed. For BIOWIN, there is no universally accepted definition of the applicability domain, and therefore different parameters should be considered to evaluate the prediction reliability. Having these limitations in mind, the main finding of the QSAR modelling study is that all BIOWIN biotic models predict a (very) slow degradation of PFASs. QSAR modelling results of biotic degradation for the individual subgroups are summarized below. For all investigated substances, the estimated values are within the BIOWIN criteria indicating potentially persistent substances as described in ECHA’s Endpoint specific guidance (ECHA, 2017b). The predicted biodegradability of individual PFASs is found in Figure B.4. Figure B.4. Predicted biodegradability of the analysed PFAS categories and molecules (numbers on the left are CAS numbers). The results from the BIOWIN 4 predictions of primary biodegradation show unexpected trends which are not in line with the knowledge from experimental studies for these substances. Hence, these results are regarded as indication that the model has limited suitability for these kinds of compounds. 20 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Perfluoroalkanes Figure B.5. Example structure: perfluorohexane. Ravishankara et al. (1993) investigated the atmospheric lifetimes of long-lived halogenated species, including CF 4, C2F 6, c-C4F 8, (CF 3)2c-C4F 6, C5F 12 and C6F 14. The possible atmospheric loss processes of these gases were assessed by determining the rate coefficients for the reactions of these gases with O( 1D), H, and OH and the absorption cross sections at 121.6 nm in the laboratory and using these data as input to a two-dimensional atmospheric model. The lifetimes of all the studied perfluoroalkane compounds were found to be more than 2000 years. These findings were confirmed by Say et al. (2021) who looked at the global trends and European emissions of tetrafluoromethane (CF 4), hexafluoroethane (C2F 6) and octafluoropropane (C3F 8). The fully fluorinated hydrocarbons were described to be potent greenhouse gases with lifetimes in the order of thousands to tens of thousands of years (50 000 years for CF 4). Thermal decomposition of perfluoroalkanes starts above 800 °C (compounds with tertiary carbon atoms above 600 °C) with the formation of saturated and unsaturated decomposition products and some carbon (Siegemund et al., 2012). From the group of perfluoroalkanes, perfluorohexane, perfluorooctane and perfluorodecaline were investigated with QSAR biodegradation models. The three substances were predicted to biodegrade slowly (BIOWIN 1, 2, 5, 6 and 7). Ultimate biodegradation (BIOWIN 3) predicted the three substances to be recalcitrant. Primary biodegradation (BIOWIN 4) estimated perfluorooctane and perfluorodecaline to be recalcitrant, while half -life of perfluorohexane was estimated to be months. All perfluoroalkanes were predicted as not readily biodegradable. Haloperfluoroalkanes Figure B.6. Example structure: chloropentafluoroethane. Fully halogenated compounds (Cl and Br in addition to F) with a high fluorine content have excellent thermal stability and are non-flammable. Chlorofluoroalkanes are characterized by high chemical and thermal stabilities, which increase with their fluorine content. At high temperature, thermal cleavage of the C-Br bond of bromoperfluoroalkanes into radicals may occur. The chemical stability of bromofluoroalkanes is slightly lower than that of the corresponding chlorofluoroalkanes. However, as with the chlorofluoroalkanes, stability 21 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) increases with the fluorine:bromine ratio. In contrast to chloro- and bromofluoroalkanes, iodofluoroalkanes readily undergo chemical reactions, reacting preferentially by homolytic cleavage of the C-I bond (Siegemund et al., 2012). However, when exposed to ultraviolet light in the upper atmosphere, chloro- and bromofluoroalkanes may suffer cleavage with release of chlorine or bromine radicals that can go on to destroy ozone in catalytic cycles. The atmospheric lifetime of the compound chloropentafluoroethane was reported to be 1 700 years (Forster, 2007). This PFAS subgroup has not been investigated so far with QSAR modelling. Perfluoroalkylethers (PFAEs) Figure B.7. Example structure: 1,1,2,2-tetrafluoro-1,2-bis(trifluoromethoxy) ethane. The thermal stability of perfluoroalkylethers (PFAEs) was studied by Helmick (1990) in relation to the potential application of the substances as high temperature engine lubricants. A range of PFAEs was studied and found to have decomposition temperatures in the interval 301389 °C which demonstrates the high thermal stability of the substances. Based on the investigated fluids, the authors concluded that t he stability of the PFAEs was not affected by intrinsic factors such as carbon chain length, branching, or cumulated -CF 2- groups. Hori et al. (2009) investigated the oxygen-induced mineralization of perfluoroalkylether sulfonates in subcritical water. They pointed out that ether linkages originally were inserted into the perfluoroalkyl chains so that the molecules should contain only short perfluoroalkyl fragments. In the first place these molecules were expected to decompose more easily than other PFASs because of the presence of the ether linkages, but no one has co nfirmed that they do in fact decompose more easily. Indeed, the authors observed that perfluoroalkylether sulfonates decomposed only at 350 °C in the presence of oxygen gas in supercritical water, while below 300 °C no reaction was observed. Under environmentally relevant conditions perfluoroalkylether chains are similarly resistant to abiotic (photolysis, reactions with OH radicals, and hydrolysis) and biotic degradation as the perfluoroalkyl chains (Wang et al., 2015c). Three ether substances were investigated with QSAR modelling for biodegradation: perfluorodiethylether, 1,1,2,2-tetrafluoro-1,2-bis(trifluoromethoxy)ethane and 2,2,3,3,4,4,5-heptafluorotetrahydro-5-(nonafluorobutyl)furan. All were predicted not to biodegrade fast (BIOWIN 1, 2, 5, 6 and 7). Ultimate biodegradation (BIOWIN 3) categorized all substances as recalcitrant, while primary biodegradation (BIOWIN 4) was weeks -months for perfluorodiethylether and 1,1,2,2-tetrafluoro-1,2-bis(trifluoromethoxy)ethane, and of months for 2,2,3,3,4,4,5-heptafluorotetrahydro-5-(nonafluorobutyl)furan. All three compounds were predicted as not readily biodegradable. The QSAR models include a negative fragment contribution of the aliphatic ether bond on the degradation potential. This indicates that the ether bond in PFECAs and PFESAs is not expected to decrease the environmental persistence as compared to PFCAs or PFSAs. The ether substance 2,3,3,3-tetrafluoro-2-(heptafluoropropoxy)propionic acid (HFPODA/GenX), its salts and its acyl halides were recognized as very persistant by the Member State Committee and the substances identified as substances of very high concern, among others on the basis of an equivalent level of concern having probable serious effects to the environment which give rise to an equivalent level of concern to those of PBT/vPvB substances (ECHA, 2019d). 22 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Perfluoroalkylcarboxylic acids (PFCAs) Figure B.8. Example structure: Perfluorobutanoic acid (PFBA). The sources, fate and transport of PFCAs were reviewed by Prevedouros et al. (2006). The PFCAs were evaluated as highly water-soluble and persistent with a high potential for longrange aquatic transport to the Arctic. PFCAs were shown to not undergo degradation in the environment. The global historical industry-wide emissions of total PFCAs were estimated to be 3 200-7 300 t. It was assumed that the majority (∼80%) of this was released to the environment from fluoropolymer manufacture and use. Qu et al. (2016) looked at the photochemical decomposition of the environmentally persistent PFCA-class. It was emphasized that the class of PFCAs are chemically inert due to the strong electronegativity of fluorine and very strong C-F bond, making them resistant to normal environmental degradation. In the study, the photodegradation of a series of PFCAs (C 2-C12) in water by a medium-pressure mercury lamp was experimentally and theoretically examined. The PFCAs were mainly decomposed into shorter carbon chain length PFCAs in a stepwise manner, with the accumulation of TFA and fluoride ions as the end products. These findings could enhance the general understanding of the photodegradation of PFCAs, although the conditions investigated are not directly environmentally relevant. Taniyasu et al. (2013) studied the environmental photolysis of PFASs in natural environment at high altitudes in Mt. Mauna Kea (Hawaii, USA; 4 200 m) and Mt. Tateyama (Toyama, Japan; 2 500 m). They observed decomposition of long-chain PFCAs (and PFSAs) with successive dealkylation and formation of short-chain compounds such as PFBA (and PFBS), typically with 20-30% decomposition after 106 days. However, these observations were disputed by Wang et al. (2015b), who argued that the perfluoroalkyl carboxylic and sulfonic acids are too stable to undergo atmospheric photolysis and asked for information on whether adsorption of longchain substances on the surface of the vials was considered in the experiments. Among the perfluoroalkylcarboxylic acids, PFOA, PFHxA and PFBA were investigated for biodegradability with QSAR modelling. BIOWIN 1 and 2 predicted all of them not to biodegrade fast. BIOWIN 3 (ultimate biodegradation) predicted that PFOA and PFHxA are recalcitrant, while for PFBA a half-life of months was predicted. For primary biodegradation, BIOWIN 4 predicted semi-quantitative half-lives as PFOA (weeks-months) > PFHxA (weeks) > PFBA (days-weeks). For PFOA and PFHxA BIOWIN 5 and 6 predicted that these compounds do not biodegrade fast. For PFBA the linear model of BIOWIN 5 predicted fast biodegradability, while the non-linear model of BIOWIN 6 predicted that the substance does not biodegrade fast. The overall assessment of biodegradability by BIOWIN for all three perfluoroalkylcarboxylic acids was that they are not readily biodegradable. BIOWIN 1-4 were reliable to predict all fragments except [-F]. Considering that [-F] contributes positively to biodegradation in BIOWIN 5 (linear MITI model), these results should be interpreted with caution, particularly for PFBA, which was predicted to biodegrade fast by BIOWIN 5, which is an unexpected result that does not match well with the range of observations of PFBA in environmental samples. See Appendix B.4.1.2. for further details of inclusion of fragments in training sets. 23 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Perfluorooctanoic acid (PFOA) as well as six long-chained perfluorinated carboxylic acids (C9C14 PFCAs) have been identified as substances of very high concern (SVHC) fulfilling the P and vP criteria according to REACH Annex XIII (ECHA, 2012b; ECHA, 2012c; ECHA, 2012d; ECHA, 2012e; ECHA, 2013; ECHA, 2015b; ECHA, 2016b). Furthermore, a REACH restriction on C9-C14 PFCAs including their salts and precursors has recently been adopted (ECHA, 2018a) due to their P and vP properties. A restriction on PFHxA its salts and related substances has been proposed based on its high persistence exceeding by far the P and vP criteria (ECHA, 2019c) and RAC has agreed to the extreme persistence as a key concern. In 2019, PFOA its salts and PFOA-related substances were listed in the Stockholm Convention, and the restriction is included in the EU POPs regulation since 2020. Just recently a proposal has been submitted for long-chain perfluorocarboxylic acids, their salts and related compounds 1 under the Stockholm Convention. Trifluoroacetic acid (TFA) The environmental fate of TFA, together with trichloro-, dichloro-, and monochloroacetic acids, was investigated using field aquatic microcosms and laboratory sediment –water systems (Ellis et al., 2001). TFA was extremely persistent and showed no degradation during a one-year field study. Biodegradation of mono-, di- and trifluoroacetate by microbial cultures with different origins was investigated by Alexandrino et al. (2018). Microbial inocula samples collected from a site with a long history of industrial contamination and activated sludge obtained from a municipal wastewater treatment plant were used in the study. Defluorination was obtained in the cultures fed with monofluoroacetate, while difluoroacetate and TFA were recalcitrant in all tested conditions. The authors pointed out that the persistence and accumulation of these substances in the environment is a relevant issue that may lead to disturbance in ecosystems. TFA and its sources, pathways, and consequences for drinking water were assessed by Scheurer et al. (2017). It was pointed out that there are contradictory results in the scientific literature with regards to microbial degradation of TFA. Some studies have observed TFA to be persistent, while some other studies have reported microbial degradation of TFA; Visscher et al. (1994) reported the rapid (in the order of one week) microbial degradation of TFA in sediments under oxic and anoxic conditions, with formation of fluoroform. Kim et al. (2000) performed a long-term (90 weeks) study to assess biodegradation of TFA in an engineered anaerobic reactor. TFA was found to be co-metabolically degradable, and the authors indicated that anaerobic degradation is a potential removal pathway for TFA in freshwater sediments and may limit their accumulation in the environment. In their own study of degradation of TFA in a WWTP, Scheurer et al. (2017) observed no decrease of TFA concentrations over 28 days during biological wastewater treatment in accordance with a modified OECD guideline 302 B Zahn-Wellens test. TFA is registered in the 1 000 – 10 000 t/y tonnage band in the ECHA database. In the registration dossier, the registrants have concluded, based on experimental evidence, that TFA fulfils both vP and P criteria, and TFA was found to be not readily nor inherently biodegradable, as well as not biodegradable under aerobic conditions. However, it was also stated by the registrant that a 'not assignable study' shows that co-metabolic degradation in anaerobic conditions can happen. TFA was not investigated specifically in QSAR modelling of biodegradation. 1 Long-chain PFC As and their salts are a homologous series of substances with the molecular formula of C nF2n+1C O2H (where 8 ≤ n ≤ 20). Related compounds are viewed as any substance that is a precursor and may degrade or transform to long-chain PFC As, where the perfluorinated alkyl moiety has the formula C nF2n+1 (where 8 ≤ n ≤ 20) and is directly bonded to any chemical moiety other than a fluorine, chlorine or bromine atom. 24 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Perfluoroalkane sulfonic acids (PFSAs) Figure B.9. Example structure: Perfluorobutane sulfonic acid (PFBS). The Global PFC Group refers to PFSAs as very persistent in the environment, while their potential precursors are transformed into PFSAs abiotically or biotically (OECD/UNEP, 2013). Due to the high resistance to heat and chemical agents, the perfluoroalkyl substances have been frequently used in products with high versatility, strength, resilience and durability. However, the high persistence allows for a wide distribution in the environment, and many PFSAs have been detected globally in the environment. Perfluoroalkane sulfonic acids (PFSAs) are very stable, with a very high thermal and chemical stability. Anhydrous PFSAs are stable at 400 °C in the absence of air, but they may form hydrogen fluoride at this temperature when moisture is present. The sulfur atoms in PFSAs are at their maximum oxidation state, and cannot be oxidised further (Arp and Slinde, 2018). Defluorination of fluorinated sulfonates by a Pseudomonas strain was investigated by Key et al. (1998). Trifluoromethane sulfonate, PFOS and some related not fully fluorinated substances were subjected to biodegradation by Pseudomonas under aerobic, sulfur-limiting conditions. Growth and defluorination were observed for the compounds containing hydrogen on the carbon chain, while it is reported that trifluoromethane sulfonate and PFOS were not degraded. Saez et al. (2008) studied the degradation of PFASs, including the sulfonic acids PFBS and PFOS, in closed bottle tests with municipal sewage sludge. Bacterial communities from sewage sludge were exposed to a mixture of PFASs under aerobic and anaerobic conditions. Individual PFAS concentrations were determined after solid phase extraction. The experiments were based on the OECD guideline 301D (closed bottle test) with slight modifications. It was found that the PFASs tested in these experiments are non-biodegradable under the conditions used. A few studies have reported the degradation of PFOS by isolated bacterial strains under special laboratory conditions or by a specific enzyme when incubated with a mediator substance in laboratory conditions. A summary of the studies may be found in the SVHC Support Document for PFBS (ECHA, 2019e). These results show that bacteria may adapt to utilise the energy present in the PFAS substrates. However, such transformations have not been observed at environmentally relevant conditions. PFOS, PFHxS and PFBS were investigated for biodegradation potential in QSAR modelling. All three substances were predicted not to biodegrade fast (BIOWIN 1, 2, 5, 6 and 7). BIOWIN 3 estimated the compounds to be recalcitrant, while the primary biodegradation model (BIOWIN 4) estimation was PFOS (months) > PFHxS (weeks-months) > PFBS (weeks). All three PFASs were predicted as not readily biodegradable. It should be noted that the sulfonic acid structure fragment is not included in the MITI models of BIOWIN 5 and 6. Hence, the results from these models should be given little weight in the assessment. In 2009, PFOS and its derivatives were included in the Stockholm Convention on Persistent 25 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Organic Pollutants to eliminate their use before being restricted in the EU under the POPs Regulation. PFBS and its salts have been included in the REACH Candidate List meeting the criteria under REACH Article 57(f), due to its very high persistence (ECHA, 2019e). PFHxS fulfils the criteria for being “very persistent” and has been adopted as SVHC by the Member State Committee in 2017 (ECHA, 2017c). PFHxS, its salts and PFHxS-related compounds were also recently listed on Annex A of the Stockholm Convention without any exemptions for use. Perfluoroalkyl phosphonic acids (PFPAs) Figure B.10. Example structure: Perfluorohexyl phosphonic acid. PFPAs are not expected to undergo hydrolysis under environmentally relevant conditions. They are resistant to basic hydrolysis (Emeleus and Smith, 1959) and stable in water at elevated temperatures up to 180 °C (Mahmood and Shreeve, 1986), similarly as PFCAs and PFSAs. Biotransformation of PFPAs has not been observed. In a metabolism study, rats dosed with C8 PFPA did not produce 1H-perfluorooctane, which has been observed for the similar substances perfluoroalkyl phosphinic acids at a lower oxidation stage (Joudan et al., 2017). A microbial degradation study conducted according to OECD Test Guideline (TG) 309 examined biodegradation of C6, C8 or C10 PFPA (Llorca-Casamayor, 2012). Water used in the experiment was wastewater effluent taken from Beuerbach WWTP (Hesse, Germany). The samples were distributed in amber glass bottles and spiked with PFPA mixture. The bottles were stirred 24 h/d in an orbital digester at 100 rpm and the pH was controlled. Dark conditions were used in order to minimize the algae growth. Aerobic conditions were maintained by aeration 30 min/d. The samples were compared with non-spiked blank samples, as well as with spiked samples treated with NaN3 to stop all biological activity. Samples were regularly collected from the flasks and analyzed by LC-MS for quantification of the PFPA substances. The experiment showed that no degradation had occurred for PFHxPA and PFOPA over 30 days. For PFDPA the results were inconclusive due to practical problems and formation of a biofilm on the walls of the experimental flasks. Wang et al. (2016b) reviewed the environmental properties of e.g. perfluoroalkyl phosphonic acids. Existing evidence demonstrated high resistance of these substances to heat, oxidants, bases and aerobic degradation in surface waters. The authors concluded that the data suggested a high or very high persistence of PFPAs in the environment and biota, and a high long-range transport potential. PFOPA, PFHxPA and PFBPA were investigated with the QSAR models in BIOWIN v4.11 of EpiSuite for biodegradability. All substances were predicted not to biodegrade fast (BIOWIN 1, 2, 5, 6 and 7). Ultimate biodegradation (BIOWIN 3) categorized all substances as recalcitrant, and primary biodegradation predicted half-life of PFOPA (months) > PFHxPA (months) > PFBPA (weeks-months). All three PFASs were predicted as not readily biodegradable. However, all the BIOWIN models lack coefficients for phosphonate (C-P bond), 26 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) which reduces the strength of the BIOWIN modelling results for this subclass. Perfluoroalkylamines Figure B.11. Example structure: Perfluamine (=perfluoro(tripropyl)amine). Siegemund et al. (2012) examined the properties of several fluoroorganic compounds. The perfluorinated tertiary amines were found to be chemically inert and thermally stable. The substances are deprived of the usual basic character and reactivity of amines due to the electron-withdrawing nature of the perfluoroalkyl substituents. Tertiary perfluoroalkylamines do not form salts or complexes with strong acids and are not attacked by most oxidizing or reducing agents. Laboratory experiments were performed by Bernard et al. (2020) in order to assess the atmospheric lifetimes of perfluoroalkylamines N(C 2F 5)3, N(C3F 7)3, and N(C4F 9)3. The O(1D) reaction and UV photolysis loss processes evaluated in this work were used in 2-D atmospheric model simulations to evaluate the global total atmospheric lifetimes. The atmospheric lifetime was found to be more than 3 000 years for all three substances. Among the perfluoroalkylamines, perfluamine, perfluoromethyldiethylamine and perfluorotrihexylamine were selected for QSAR modelling of biodegradation. All substances were predicted not to biodegrade fast (BIOWIN 1, 2, 5, 6 and 7). Ultimate biodegradation (BIOWIN 3) predicted the three substances to be recalcitrant. Primary biodegradation (BIOWIN 4) estimated perfluamine and perfluorotrihexylamine to be recalcitrant, while the half-life of perfluoromethyldiethylamine was estimated to be months. All perfluoroalkylamines were predicted as not readily biodegradable. For the compound perfluorotrihexylamine, the number of instances of the fragments [C with 4 single bond and no H] and [-F] exceeds too much the number in the training set. Therefore, the results of the BIOWIN modelling for this substance, should be interpreted with care. B.4.1.2.4. The effects of chain length, branching and cyclic structure elements Within each of the above PFAS subclasses the substances differ only in the number of perfluorinated carbon atoms in the carbon chain(s), i.e. the chain length. There is no evidence in the literature that the length of the perfluorinated carbon chain has a significant influence on the degradability/stability of these substances. Hence, all members of the same PFAS subclass are to be considered equally persistent. Neither ultrashort -chain (C1-C3) nor ultralong-chain PFAAs will biodegrade under environmentally relevant conditions, and PFAAs are regarded as highly stable substances in which several precursors ultimately degrade into. The stability of organic fluorine compounds has been described in detail by Siegemund et al. (2012): “When all valences of a carbon chain are satisfied by fluorine, the zig-zag-shaped carbon skeleton is twisted out of its plane in the form of a helix. This situation allows the electronegative fluorine substituents to envelop the carbon skeleton completely and shield it from chemical (especially nucleophilic) attack. Several other properties of the C -F bond 27 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) contribute to the fact that highly fluorinated alkanes are the most stable organic compounds. These include low polarizability and high bond energies, which increase with increasing substitution by fluorine”. It is not expected that branching of the perfluoroalkyl chain will significantly affect the high persistence of the corresponding unbranched PFAS substances as long as the alkyl chain is fully fluorinated. Likewise, cyclic perfluoroalkyl structures are expected to be as persistent as linear or branched PFASs, potentially with the exception of very small ring structures with high ring strain (i.e. 3- or 4-membered rings). However, perfluorocyclobutane (PFC-318) has an atmospheric lifetime of 3 200 years, which demonstrates the high persistence even of 4-membered ring structures (Forster, 2007). The cyclic PFAS substance ammonium difluoro[[2,2,4,5tetrafluoro-5-(trifluoromethoxy)-1,3-dioxolan-4-yl]oxy]acetate (CAS no 1190931-27-1) was shown to be not readily biodegradable (5% DOC removal) in a screening test for ready biodegradability (OECD 301A), see registration dossier2. Monitoring programs have detected the presence of perfluoroethylcyclohexane sulfonate (PFECHS, "cyclic PFOS") in water, sediments and biota from different places in the world, as well as in the Devon Ice Cap, see section B.4.2.7.4 on monitoring data. In the same section it is explained that perfluoropropylcyclopentanesulfonate (PFPCPeS) was found in environmental samples downstream of Beijing airport, which again demonstrates the high environmental stability of the substances. B.4.1.2.5. Structural elements in combination As described above, selected PFAS subclasses, or structural elements (different moieties), have been investigated and shown to be persistent and highly stable in the environment. The remarkable stability arises from the high strength of the C-F bond, in combination with structural elements which are not transformed under environmental conditions and which do not inflict sufficient reactivity to neighbouring C-F units. Hence, the PFAS subgroups described represent final degradation products that do not undergo any further degradation in the environment and are designated arrowhead substances or arrowhead subgroups. Perfluoroalkyl acids with an acid functional group at its highest oxidation state, i.e. carboxylic, sulfonic and phosphonic acids represent structural endpoints in an oxidative environment. Degradation studies and monitoring data show that these substances are extremely persistent and do not undergo biotic or abiotic degradation in the environment. It should be noted that all substances in the above assessment are perfluorinated substances with fully fluorinated carbon chains in combination with selected functional groups. If other functional groups or C-H bonds are included in the substance structure, further assessment of the stability may need to be conducted. Any substance with a combination of the above-mentioned structure elements is also expected to be persistent. There is no reason to expect that t hese structure elements in combination will induce considerably higher reactivity in a perfluorinated substance as compared to substances containing these elements separately. Examples of substances that contain a combination of several of the mentioned st ructure elements include HFPO-DA (GenX), ADONA, F53B, other perfluoroalkyl ether sulfonic acids and perfluoro-N-methylmorpholine (PMM), which can be found in Figure B.12. 2 http://echa.europa.eu/registration-dossier/-/registered-dossier/5712, date of access: 2022-10-21. 28 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.12. Examples of substances with a combination of structural elements. In the SVHC identification of HFPO-DA (GenX) it was concluded, based on available experimental evidence and QSAR information, that the substance exceeds the P and vPcriteria of REACH by far (ECHA, 2019d). Gordon (2011) evaluated toxicological aspects of ADONA and indicated that ADONA is a non-reactive, stable and not readily biodegradable substance that decomposes only at 125-175 °C. Perfluoroalkylether sulfonic acids, including C2F 5OC2F 4SO3H, were shown to not undergo any reaction even in supercritical water with oxygen gas for 6 h at temperatures up to 300 °C (Hori et al., 2009). Wang et al. (2013b) looked at the environmental occurrence of F -53B in China and assessed its toxicity and persistence. Ready biodegradability of F-53B was measured in a Closed Bottle Test (CBT) according to OECD Guideline 301D. In addition, the stability of F -53B under various advanced oxidation process (AOP) conditions was assessed. Although the compound showed a slow degradation throughout the test period in the CBT, it did not meet the OECD criteria to satisfy ready biodegradation. Under all AOP test conditions, the degradation of F -53B was very low. The authors concluded that F-53B is not readily biodegradable, and their data suggested that F-53B is as persistent as PFOS. This is supported by the ubiquitous presence of F -53B in the environment in China, US, UK, Sweden, the Netherlands and Korea, while China is the only known location for emissions of the substance. This is a strong indication of high persistence of the substance. However, indications of slow de-chlorination of F-53B were found in an increased molar ratio of the H-analogue as compared to the manufactured mixture (Pan et al., 2018). Perfluoro-N-methylmorpholine (PMM), which contains both an ether structure unit and a tertiary amine in combination, is considered as very persistent by the registrant in the ECHA database (tonnage band: 100-1 000 t/y). The substance was assessed as very persistent by UBA (2019) on the basis of QSAR and biodegradation screening tests, and it has been added to the ChemSec SIN list 3 as a very persistent and very mobile substance. Fluoropolymers and side-chain fluorinated polymers 3 https://sinlist.chemsec.org/, date of access: 2022-09-29. 29 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Many of the mentioned structural elements may be combined multiple times in polymeric chains. Again, the combination of persistent structural elements is expected to result in a persistent overall structure, where the number of persistent parts is high. Examples of fluoropolymers constructed from the mentioned persistent structure elements include e.g. polytetrafluoroethylene (PTFE), perfluoroalkoxy polymer (PFA), etc (see Figure B.13). Figure B.13. Examples of fluoropolymers. Lohmann et al. (2020) investigated whether fluoropolymers may be regarded as polymers of low concern for human and environmental health. With regards to persistence, they pointed out that fluoropolymers are very persistent under environmental conditions, which, in the same way as for other polymers, can lead to a wide array of concerns, particularly with respect to disposal of fluoropolymer-containing wastes and products. Current concern over microplastics present in the oceans is also related to fluoropolymers. Henry et al. (2018) reviewed regulatory criteria of fluoropolymers in relation to the polymers' physical and chemical properties. It was indicated that fluoropolymers in general have high thermal, chemical, photochemical, hydrolytic, oxidative, and biological stability and that they are persistent. The authors looked in more detail at the four fluoropolymers PTFE (polytetrafluoroethylene), ETFE (ethylene tetrafluoroethylene), FEP (fluorinated ethylene proplene) and PFA (perfluoroalkoxy polymer) and concluded that these are stable under hydrolytic, light, oxidative, biodegrading (aerobic and anaerobic) and thermal conditions. Side-chain fluorinated polymers (SFPs) are different from the fluoropolymers in that they usually contain a non-fluorinated backbone with per- or polyfluorinated alkyl side-chains attached to the backbone via a linker (Buck et al., 2011). These linkers are often labile and may be cleaved under environmental conditions with liberation of well-known PFASs. SFPs are similar to from non-polymeric PFAS precursors and are expected to follow the same reactivity pattern. See next section for examples on degradation of SFPs to PFAAs. Further discussion on the degradation of SFPs is provided, e.g. in the restriction proposal of PFHxA, its salts and related substances (ECHA, 2019c). 30 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.1.3. Degradation of precursors into the corresponding PFAS arrowheads There are many PFASs which contain degradable non-perfluorinated moieties. These precursor substances are not completely persistent themselves but will ultimately degrade to very persistent PFAS arrowheads (PFAAs), through reactions such as atmospheric oxidation, metabolic transformations, and hydrolysis. It is expected that the degradation will primarily target the non-fluorinated parts. During the degradation process the non-fluorinated moieties are transformed and oxidative processes often lead to the gradual conversion of nonfluorinated carbon atoms into oxidized molecules such as CO 2 while the degrading substance structure is gradually getting smaller. In the end most of the non-fluorinated parts are usually lost, while the perfluoroalkyl part is remaining (although defluorination of the carbon atom next to the non-fluorinated part can occur), attached to a functional group at its highest oxidation state (e.g. carboxylic acid). Such functional groups often carry a negative charge which leads to an increased polarity for the degradation products. Hence, the fluorinated degradation products of PFASs may be assumed to have elevated mobility with water compared to precursor substances. In the coming subsections, relevant available information on degradation of t he following precursors is summarised: PFCA precursors:    n:2 Fluorotelomer alcohols (FTOHs) Other n:2 fluorotelomer derivatives o n:2 Fluorotelomer iodides (FTIs) o Esters of n:2 fluorotelomer alcohols o n:2 Polyfluoroalkyl phosphoric acid mono-/diesters (monoPAPs/diPAPs) o n:2 Fluorotelomer urethane (monomers) o n:2 Fluorotelomer sulfonic acids (FTSAs) o n:2 Fluorotelomer thioether amido sulfonates (FTTAoSs) o n:2 Fluorotelomer silanes o n:2 Fluorotelomer olefins (FTOs) o n:2 Fluorotelomer-based side-chain fluorinated polymers Other PFCA precursor o Perfluoroalkyl carboxylic acid halides o Amides of perfluoroalkyl carboxylic acids o n:1 Fluorotelomer alcohols o Perfluoroalkyl alcohols o Perfluoroalkyl iodides (PFAIs) o Perfluorinated olefins (with the formula CnF2n+1-CF=CF-CmF2m+1) o Side-chain fluorinated aromatics Fluorinated gases: o o o o Hydrofluorocarbons (HFCs) and hydrofluoroolefins (HFOs) Hydrofluoroethers (HFEs) Perfluoroalkyl ketones Perfluoroalkyl nitrile compounds PFSA precursors: o o o o o o o Perfluoroalkane sulfonyl halides Perfluoroalkane sulfonic acid esters Perfluoroalkane sulfonamides (FASAs) Perfluoroalkane sulfide derivatives Perfluoroalkane sulfoxide derivatives Perfluoroalkane sulfinic acid derivatives Side-chain fluorinated polymers based on sulfonic acid derivative 31 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFPA precursors: o Perfluoroalkyl phosphinic acids (PFPiAs) These subsections include degradation data of the various PFAS subgroups in different compartments, such as air, soil, and water and involves different degradation mechanisms and pathways. Hence, the actual fate for a specific PFAS in the environment depends both on available degradation pathways for that PFAS subgroup and the physicochemical properties of the specific substance, like volatility and solubility that determines the partitioning to different compartments. The presence of microorganisms in those compartments is an additional factor that influences the degradation. B.4.1.3.1. Degradation of PFCA precursors The degradation pathways of several PFCA precursors into the corresponding PFCAs are extensively described in the Background documents to the Opinion on the Annex XV dossier proposing restrictions on PFOA (ECHA, 2018b), C9-C14 PFCAs (ECHA, 2018a), and PFHxA (ECHA, 2019c). Hence, this section is to a large extent based on the degradation information in these documents which has been discussed and accepted by RAC for the purpose of these restrictions. However, the subsection “Other PFCA precursors” are mainly PFAS subgroups that were not included in the background documents to previous restrictions. The following PFAS subgroups are expected to degrade into PFCAs:  n:2 Fluorotelomer alcohols (FTOHs) Figure B.14. Example of an n:2 fluorotelomer alcohol: 4:2 FTOH. Based on the available data it can be expected that n:2 FTOHs will be degraded and transformed into Cx-PFCAs (with x= n-2, n-1, n, n+1; see Appendix B.4.1.3.1. Table B.75). This means that up to three -CF2- groups can be defluorinated and mineralized to CO2 and HF until the ultimate persistent PFCA is formed. The degradation pathways of n:2 FTOHs exemplified by the degradation of 6:2 FTOH: 6:2 FTOH degrades to the corresponding PFCAs under various conditions (see Appendix B.4.1.3.1. Table B.75). The degradation pathways of 6:2 FTOH in an aerobic river sediment system proposed by Zhao et al. (2013a) are illustrated in Figure B.15 and these pathways are typical for 6:2 FTOHs. In this specific study, after 100 days, 22.4 mol% 5:3 FTCA (5:3 fluorotelomer carboxylic acid; also referred to as 5:3 acid), 10.4 mol% C5-PFCA (PFPeA), 8.4 mol% C6-PFCA (PFHxA), and 1.5 mol% C4-PFCA (PFBA) were detected. Major intermediates during biotransformation of 6:2 FTOH were 6:2 FTCA (6:2 fluorotelomer carboxylic acid; also referred to as 6:2 acid), 6:2 FTUCA (6:2 fluorotelomer unsaturated carboxylic acid), 5:2 ketone, and 5:2 sFTOH (5:2 secondary fluorotelomer alcohol). The recovery of 6:2 FTOH and quantifiable transformation products ranged 71-88 mol% of initially applied 6:2 FTOH. The lower mass balance can be explained by formation of bound residues. Another study investigated the biotransformation of the intermediate degradation product 5:3 FTCA in activated sludge (Wang et al., 2012). After 90 days 5:3 FTCA biotransformation yielded 14.2 mol% 4:3 FTCA, 5.9 mol% C5-PFCA (PFPeA) and 0.8 mol% C4-PFCA (PFBA). 32 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) These results implicate that 5:3 FTCA should not be regarded as an arrowhead, but rather a relatively stable intermediate that will ultimately degrade to the corresponding PFCAs. Still, the degradation rate of 5:3 FTCA is highly dependent on the specific environmental conditions. E.g. Liu et al. (2010c) incubated 5:3 FTCA in aerobic soil and after 60 days only 2% 4:3 FTCA was observed. According to the authors, this indicates that 5:3 FTCA is relatively resistant to biodegradation in soil due to its strong tendency to become irreversibly adsorbed to soil. Figure B.15. Proposed 6:2 FTOH biotransformation pathways in aerobic sediment system (based on Zhao et al. (2013a)). In one biodegradation study with 6:2 FTOH in an aerobic microbial culture (Sun et al., 2020) C2-PFCA (trifluoroacetic acid) (2.3 mol%) was formed along with other degradation products, meaning that in this specific case up to five -CF 2- groups were defluorinated for a minor fraction of the starting material. However, once the PFCA, for example C4-PFCA (PFBA) is formed, that specific PFCA is persistent and will not show any further defluorination of -CF2groups. The photooxidation of 6:2 FTOH was investigated at the surface of TiO2, SiO2, Fe2O3, Mauritanian sand, and Icelandic volcanic ash (Styler et al., 2013). At all surfaces the photooxidation resulted in the production of surface-sorbed PFCAs (PFHpA, PFHxA and PFPeA). These results provide evidence that the heterogeneous photooxidation of FTOHs at metal-rich atmospheric surface may provide a significant loss mechanism for FTOHs and also act as a source of aerosol-phase PFCAs close to source regions. The long-range transport of these aerosols is a possible source of PFCAs to remote areas. The degradation pathways of n:2 FTOHs exemplified by the degradation of 8:2 FTOH: 8:2 FTOH degrades to the corresponding PFCAs under various conditions (see Appendix B.4.1.3.1. Table B.75). 8:2 FTOH metabolism universally shows the formation of PFOA and, to a smaller fraction, PFNA and lower-chain-length PFCAs (Butt et al., 2014). 33 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The proposed degradation pathways of 8:2 FTOH in soil and activated sludge are illustrated in Figure B.16. These pathways are very similar to those proposed for 6:2 FTOH in Figure B.15. The percentages of the degradation products refer to several studies (Dinglasan et al., 2004; Wang et al., 2005a; Wang et al., 2009; Wang et al., 2005b). 8:2 FTOH metabolism universally shows the formation of PFOA and, to a smaller fraction, PFNA and lower-chain-length PFCAs – mainly PFHPA and PFHxA (Butt et al., 2014). 7:3 FTCA (7:3 fluorotelomer carboxylic acid; also referred to as 7:3 acid) is usually also a major degradation product but should not be regarded as an arrowhead, but rather a relatively stable intermediate that will ultimately degrade to the corresponding PFCAs (Butt et al., 2010; Li et al., 2018b). Figure B.16. Aerobic degradation pathways of 8:2 FTOH in soil and activated sludge (figure based on Liu and Mejia Avendano (2013)). Stable and semi-stable compounds are shown inside dashed boxes. 2H-PFOA (2H-C8-PFCA) has been proposed, but it has not been successfully validated. In one biodegradation study with 8:2 FTOH in an anaerobic activated sludge (Li et al., 2018b) perfluoropentanoic acid (1.2%) and perfluorobutanoic acid (1.9%) were formed along with other degradation products, meaning that in this specific case up to five -CF2- groups were defluorinated. However, once the PFCA, for example C6-PFCA (PFHxA) is formed, that specific PFCA is persistent and will not show any further defluorination of -CF2- groups. Atmospheric degradation was studied in a smog chamber (Ellis et al., 2004). Experiments were performed in 750 Torr of air at 296 K. Reaction mixtures were subject to 0.5 to 15 min 34 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) UV radiation leading to a consumption of FTOH in the range of 66 to >98%. It was shown that 8:2 FTOH is oxidized, initiated by Cl atoms which represent OH radicals, and forms PFNA, PFOA (1.5% C mass balance of 8:2 FTOH) and PFCAs containing a carbon chain of less than eight carbon atoms. The formation of PFOA is expected to be greater because intermediate transformation products were still observed (e.g. 26% 8:2 FTCA, 6% 8:2 FTAL (8:2 fluorotelomer aldehyde)). The authors stress that the formation of PFOA is small but significant and postulate that FTOH degradation is likely an important source of PFOA and other PFCAs in remote areas. It can be assumed that the degradation mechanisms for n:2 FTOHs are independent from the chain length (see also section B.4.1.3 concerning the effects of chain length, branching and cyclic structure elements on persistence). A limited number of available degradation studies on n:2 FTOHs and the intermediate products are summarized in Appendix B.4.1.3.1. Table B.75.  Other n:2 fluorotelomer derivatives o n:2 Fluorotelomer iodides (FTIs) Figure B.17. Example of an n:2 fluorotelomer iodide: 4:2 FTI. Based on the available data it can be expected that n:2 FTIs will be degraded and transformed into Cx-PFCAs (with x= n-1, n, n+1; see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). n:2 FTIs follow a similar degradation pattern as the n:2 fluorotelomer alcohols forming the corresponding PFCAs, generally via an initial hydrolysis step forming the corresponding n:2 FTOH. o Esters of n:2 fluorotelomer alcohols Figure B.18. Example of an ester of an n:2 FTOH: 4:2 fluorotelomer methacrylate (4:2 FTMA). Based on the available data it can be expected that esters of n:2 fluorotelomer alcohols (e.g. n:2 fluorotelomer (meth)acrylates (FT(M)As), n:2 fluorotelomer stearate monoesters, and n:2 fluorotelomer citrate trimesters) will be degraded and transformed into Cx-PFCAs (with x= n-2, n-1, n; see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). Esters of n:2 fluorotelomer alcohols follow the same degradation pattern as the n:2 fluorotelomer alcohols forming the corresponding PFCAs, generally via an initial hydrolysis step forming the corresponding n:2 FTOH. 35 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) o n:2 Polyfluoroalkyl phosphoric acid mono-/diesters (monoPAPs/diPAPs) Figure B.19. Example of an n:2 monoPAP/diPAP : 4:2 monoPAP. Based on the available data it can be expected that n:2 monoPAPs and n:2 diPAPs will be degraded and transformed into Cx-PFCAs (with x= n-2, n-1, n, n+1; see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). n:2 monoPAPs and n:2 diPAPs follow the same degradation pattern as the n:2 fluorotelomer alcohols forming the corresponding PFCAs, generally via an initial hydrolysis step forming the corresponding n:2 FTOH. o n:2 Fluorotelomer urethane (monomers) Figure B.20. Example of an n:2 fluorotelomer urethane monomer: hexamethylene-1,6-di-(4:2 fluorotelomer urethane). Based on the available data it can be expected that n:2 fluorotelomer urethane (monomers) will be degraded and transformed into Cx-PFCAs (with x= n-2, n-1, n; see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). n:2 Fluorotelomer urethane monomers follow a similar degradation pattern as the n:2 fluorotelomer alcohols forming the corresponding PFCAs (generally via an initial hydrolysis step forming the corresponding n:2 FTOH). o n:2 Fluorotelomer sulfonic acids (FTSAs) Figure B.21. Example of an n:2 fluorotelomer sulfonic acid: 4:2 fluorotelomer sulfonic acid. Based on the available data it can be expected that n:2 FTSAs will be degraded under aerobic conditions and transformed into Cx-PFCAs (with x= n-2, n-1, n, n+1; see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). n:2 Fluorotelomer sulfonic acids follow a similar degradation pattern as the n:2 fluorotelomer alcohols forming the corresponding PFCAs, generally via an initial de-sulfonation step forming the corresponding n:2 FTOH or the corresponding n:2 FTAL (fluorotelomer aldehyde) directly, depending on the conditions. 36 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) o n:2 Fluorotelomer thioether amido sulfonates (FTTAoSs) [belonging to the PFAS subgroup n:2 fluorotelomer-thiol derivatives] Figure B.22. Example of an n:2 fluorotelomer thioether amido sulfonate : 4:2 FTTAoS (FTTAoSs are used in aqueous film-forming foam (AFF) formulations). Based on the available data it can be expected that FTTAoS (and similar substances belonging to the PFAS subgroup n:2 fluorotelomer-thiol derivatives) will be degraded under aerobic conditions and transformed into PFCAs (see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). n:2 FTTAoSs follow a similar degradation pattern as the n:2 fluorotelomer alcohols forming the corresponding PFCAs, generally via initial oxidation of the thioethergroup into a sulfinyl group followed by an oxidation/C-S-cleavage step to form a n:2 fluorotelomer sulfonic acid, followed by a de-sulfonation step primarily forming the corresponding n:2 FTOH or the corresponding n:2 FTAL (fluorotelomer aldehyde) directly, depending on the conditions. o n:2 Fluorotelomer silanes Figure B.23. Example of an n:2 fluorotelomer silane : triethoxy(3,3,4,4,5,5,6,6,7,7,8,8,8tridecafluorooctyl)silane. Based on the available data it can be expected that n:2 fluorotelomer silanes will be degraded and transformed into corresponding PFCAs in the atmosphere (see references and % PFCAs in Appendix B.4.1.3.1. Table B.76). Nielsen (2014) proposed a photo-oxidation-mediated mechanism in which the corresponding n:2 fluorotelomer carboxylic acids (n:2 FTCAs) are initially formed by oxidation/Si-C cleavage, followed by further degradation into the corresponding PFCAs (in analogy with degradation of FTOHs). In a study by Zhu et al. (2019) 8:2 polyfluoroalkyl trimethoxysilane (8:2 PTrMeOSi) was degraded in a hydroxyl radical-based total oxidizable precursor assay, yielding perfluoroheptanoic acid (PFHpA, 49 ± 11%) (-2 CF2), perfluorooctanoic acid (PFOA, 14 ± 3%) (-1 CF2), perfluorohexanoic acid (PFHxA, 12 ± 3%) (-3 CF2), perfluorononanoic acid (PFNA, 2 ± 0.2%) (-0 CF2) and the other shorter-chain analogues with dec reasing molar yields. 37 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) o n:2 Fluorotelomer olefins (FTOs) Figure B.24. Example of an n:2 fluorotelomer olefin: 4:2 FTO. Based on the available data it can be expected that n:2 fluorotelomer olefins will be degraded and transformed into corresponding PFCAs in the atmosphere via a photo-oxidation-mediated mechanism (see Figure B.25 and references in Appendix B.4.1.3.1. Table B.76). Figure B.25. Proposed atmospheric degradation pathway for n:2 fluorotelomer olefins into PFCAs from Nielsen (2014). o n:2 Fluorotelomer-based side-chain fluorinated polymers Figure B.26. Example of an n:2 fluorotelomer -based side-chain fluorinated polymer: 4:2 fluorotelomer acrylate polymer. Based on the available data it can be expected that n:2 fluorotelomer-based side-chain fluorinated polymers will degrade in the same way as the small-molecule precursors into the corresponding PFCAs (Russell et al., 2008; Russell et al., 2010; Washington et al., 2009) (Rankin et al., 2014; Washington and Jenkins, 2015), generally via an initial hydrolysis step forming the corresponding FTOH. It can be assumed that the degradation mechanisms of the n:2 fluorotelomer derivatives described above are independent from the chain length (see also section B.4.1.2 concerning the effects of chain length, branching and cyclic st ructure elements on persistence). A limited 38 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) number of available degradation studies on fluorotelomer derivatives are summarized in Appendix B.4.1.3.1. Table B.76.  Other PFCA precursors (majority of these PFAS subgroups were not included in the background documents to previous PFCA restrictions) o Perfluoroalkyl carboxylic acid halides Based on the available data it can be expected that perfluoroalkyl carboxylic acid halides will undergo hydrolysis with formation of the corresponding PFCAs. In the manufacturing of PFCAs, perfluoroalkyl carboxylic acid fluorides, CnF 2n+1C(O)F are hydrolysed in the last synthesis step to yield the corresponding PFCA C nF 2n+1C(O)OH (Buck et al., 2011). Perfluoroalkyl carboxylic acid fluorides, as well as other perfluoroalkyl carboxy lic acid halides, are expected to undergo hydrolysis also under environmental conditions (Young and Mabury, 2010). o Amides of perfluoroalkyl carboxylic acids Figure B.27. Example of an amide of a perfluoroalkyl carboxylic acid : N-ethyl-perfluorobutyramide. Based on the available data it can be expected that amides of perfluoroalkyl carboxylic acids (CnF 2n+1C(O)NRR’) can be abiotic ally degraded (primarily in the atmosphere via a photooxidation-mediated mechanism) and transformed into corresponding PFCAs (mainly CnF 2n+1C(O)OH; (Jackson and Mabury, 2013; Jackson et al., 2013)). o n:1 Fluorotelomer alcohols Figure B.28. Example of a n:1 fluorotelomer alcohol: 5:1 FTOH. Based on the available data it can be expected that in the atmosphere n:1 fluorotelomer alcohols (n:1 FTOHs) can undergo OH-radical-mediated oxidation to form perfluoroaldehydes F(CF 2)nC(O)H, which can be further oxidized to form the corresponding PFCAs (Hurley et al., 2006; Hurley et al., 2004b; Wang et al., 2014c). In a study by Hurley et al. (2004b), n:1 FTOHs F(CF 2)nCH2OH (n = 1-4) were reacted with Cl radicals in a smog chamber (UV irradiation of F(CF 2)nCH2OH/Cl2 in air; the Cl radicals represent OH radicals). The reaction was followed by FTIR analysis. In all cases, the perfluoroaldehyde, F(CF 2)nC(O)H, was the sole primary product. F(CF 2)nCOOH, C(O)F2, CF 3OH, and CF 3O3CF 3 were observed as secondary products. According to the authors, reaction of F(CF2)nCH2OH (n = 1-4) is initiated by the abstraction of hydrogen, followed by reaction with oxygen, leading to formation of the perfluoroaldehyde, F(CF 2)nC(O)H. In a separate study by Hurley et al. (2006), it was suggested that perfluoroaldehydes F(CF 2)nC(O)H can be further oxidized in the atmosphere via initial formation of perfluoroacyl peroxy radicals F(CF 2)nC(O)O2 that can react further with HO2 radicals, forming the 39 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) corresponding PFCAs. The perfluoroaldehydes can also degrade via another route, forming F(CF 2)n radicals and CO, that can react further via a chain shortening mechanism. The authors argued that the relative importance of these two reaction pat hways in the atmosphere requires detailed knowledge of the temperature- and pressure-dependence of respective pathway. o Perfluoroalkyl alcohols Figure B.29. Example of a perfluoroalkyl alcohol: perfluorohexanol. Based on the available data it can be expected that in the atmosphere perfluoroalkyl alcohols (CnF 2n+1OH) can undergo heterogeneous elimination of HF to give the acyl fluoride s Cn-1F 2n1C(O)F, which can hydrolyze to give the corresponding PFCAs C n-1F 2n-1C(O)OH (Ellis et al., 2004). o Perfluoroalkyl iodides (PFAIs) Figure B.30. Example of a perfluoroalkyl iodide : perfluoropentyl iodide. Based on the available data it can be expected that perfluoroalkyl iodides can be abiotically degraded and transformed into PFCAs. Perfluoroalkyl iodides are known to readily undergo chemical reactions under certain laboratory conditions, such as gas phase photolysis, reacting preferentially by homolytic cleavage of the C–I bond (Nielsen, 2014; Siegemund et al., 2012). Based on this intrinsic property, perfluoroalkyl iodides can be expected to generate the radical CnF2n+1 under certain environmental conditions, e.g. in the atmosphere, via a photooxidation-mediated mechanism. CnF2n+1· is a potential source for PFCAs. o Perfluorinated olefins (with the formula CnF2n+1-CF=CF-CmF2m+1) Figure B.31. Example of a perfluorinated olefin: perfluoropent-2-ene. Based on the available data it can be expected that in the atmosphere perfluorinated olefins (CnF 2n+1-CF=CF-CmF2m+1) can undergo OH-radical-mediated degradation forming the perfluoroacyl fluorides CnF 2n+1C(O)F + CmF 2m+1C(O)F that can subsequently hydrolyse to the corresponding PFCAs CnF 2n+1C(O)OH + CmF 2m+1C(O)OH ((Young et al., 2009; Young and Mabury, 2010); see Figure B.32). 40 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.32. Mechanism for the atmospheric oxidation of perfluorobut-2-ene (Young et al. 2009). Young et al. (2009) investigated the expected atmospheric fate of two perfluorobutenes, CF 3CF=CFCF 3 and CF 3CF 2CF=CF 2, using smog chamber techniques. Rate constants for reaction with chlorine atoms and hydroxyl radicals were measured with relative rate techniques. The atmospheric lifet imes of CF 3CF=CFCF3 and CF 3CF 2CF=CF 2 are determined by reaction with OH radicals and are approximately 24 and 6 days, respectively. The chlorine atom- and OH radical-initiated oxidation of CF 3CF=CFCF3 in 700 Torr of air gives CF 3C(O)F in a molar yield indistinguishable from 200%, while the oxidation of CF 3CF 2CF=CF 2 gives CF 3CF 2C(O)F and COF 2 in molar yields indistinguishable from 100%. The atmospheric fate of CF 3C(O)F and CF 3CF 2C(O)F is hydrolysis to give perfluoroalkyl carboxylic acids (PFCAs), CF 3C(O)OH and CF 3CF 2C(O)OH. In a review article on atmospheric perfluorinated acid precursors, Young and Mabury (2010) argued that because perfluorinated chain length has not been shown to affect reaction mechanism in experiments with alcohols and acids, it is likely that any alkene, in which a C– F bond and a perfluorinated alkane chain appear on one side of the double bond, would be expected to follow this pathway, producing perfluoroacyl fluorides, and subsequently, PFCAs, in 100% molar yield. Fluorinated alkenes form PFCAs in 100 or 200% yield, under typical atmospheric conditions, in the presence or absence of NOx. o Side-chain fluorinated aromatics Figure B.33. Example of a side-chain fluorinated aromatic: (Heptafluoropropyl)benzene. Based on the available data it can be expected that side-chain fluorinated aromatics (CnF 2n+1Ar) can be degraded and transformed into the corresponding PFCAs (CnF 2n+1C(O)OH). The PFAS subgroup side-chain fluorinated aromatics is a very diverse group defined as “aromatics that have one or more aliphatic fully fluorinated saturated carbon moiety on the side chain(s) attached to the aromatic ring(s)” (OECD, 2021). The aromatic ring can be a phenyl group or any heteroaromatic group, with or without additional substituents. The fluorinated side-chain can have different carbon chain lengths and branching. The trifluoromethyl group is the most widely applied fluorinated side-chain. There are many pesticides that contain a trifluoromethyl-substituted aromatic ring. The 41 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) corresponding PFCA, Trifluoroacetic acid (TFA), has been identified as a significant degradation produc t in numerous studies conducted as part of the EU evaluation of plant protecting active substances (with a variety of structural features), including: Flurtamone (Figure B.34): Metabolism of flurtamone in primary crops was investigated in the cereals/grass (wheat, barley) and in oilseeds/pulses (sunflower, peanuts) crop groups, using 14 C-flurtamone (EFSA, 2016). TFA metabolite was identified as the most abundant compound of the total residues in wheat grain (86–93% total radioactive residue - TRR), in wheat forage (44% TRR) and in wheat straw (49% TRR), while 3-(trifluoromethyl)benzoic acid metabolite was predominantly identified in sunflower seed (19% TRR). The metabolism of flurtamone in primary crops proceeds mainly by hydroxylation, respectively, of the phenyl and trifluoromethylphenyl rings, followed by conjugation with malonic acid and glucose, Ndemethylation, oxidative cleavage of the trifluoromethylphenyl moiety leading to TFA metabolite, and oxidative ring opening of the furanone moiety with subsequent cleavage and degradation of the carbon chain. Figure B.34. Structural formula of flurtamone. Saflufenacil (Figure B.35): Metabolism of saflufenacil in primary crops was investigated in maize, soybean and tomatoes, using 14C- saflufenacil (EFSA, 2014). In maize, TFA was the predominant constituent (30.5% to 88% TRR, in grain it accounted for 0.004 mg/kg). Since the potentially corresponding 14C-phenyl-labelled metabolites as counter parts of TFA were not detected at adequate quantities, the occurrence of TFA was explained by the uptake of this metabolite or a respective precursor molecule from the soil. In soya beans after preemergence application to the soil surface, TFA was the major compound identified (65.4%TRR (beans) to 85.2%TRR (forage)). In tomatoes TFA was also found being the predominant constituent (48.6% TRR (in fruit) to 82.2%TRR (in tomato plant)). The occurrence of TFA was explained by the uptake of this metabolite or of a respective precursor molecule from the soil. Figure B.35 Structural formula of saflufenacil. 42 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Fluazinam (Figure B.36): Fluazinam is proposed to be used on crops that can be grown in rotation with other crops. In the confined rotational crop studies (EFSA, 2017b), TFA was the only relevant compound in rotational crops (lettuces, barley grains, carrots). Fluazinam or any of its primary metabolites were not found. Figure B.36 Structural formula of fluazinam. Fluometuron, trifloxystrobin and cyflumetofen (Figure B.37) are additional examples of plant protecting active substances where TFA has been identified as a significant metabolite in their respective EU evaluation (EFSA, 2017a; EFSA, 2019b; EFSA, 2021). a) b) c) Figure B.37 Structural formulas of fluometuron (a); trifloxystrobin (b); cyflumetofen (c). In a study by Scheurer et al. (2017) the potential TFA formation in wastewater treatment plants (WWTP) was investigated for a number CF 3-containing compounds, including five trifluoromethyl substituted aromatics: the pharmaceutical active substances fluoxetine and sitagliptine and the plant protecting active substances flufenacet, flurtamone and fluopyram (see Figure B.38). 43 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) a) d) b) c) e) Figure B.38. Pharmaceutical active substances: fluoxetine (a) and sitagliptine (b) and the plant protecting active substances flufenacet (c), flurtamone (d) and fluopyram (e) were examined for their potential degradation to TFA in wastewater treatment plants during activated sludge treatment or upon ozonation (Scheurer et al., 2017). The TFA-evolution-potential of the test compounds by chemical oxidation were examined. A concentration of 100 μg/L of the test compound was applied in demineralized water and two different ozone dosages (0.4–0.5 mg/L and 4–5 mg/L) were used. Samples were taken after contact times between 5 min and 60 min. Fluoxetine and flurtamone were rapidly degraded by ozone and could not be detected after 5 min contact time at both ozone concentrations applied. Approx. 40% TFA had been formed on a molar base from fluoxetine and flurtamone after 60 min. The respective precursor compounds were completely oxidized after 5 min contact time but a steady increase of TFA over the course of the test indicated that intermediates are formed, which are further oxidized to TFA. A comparatively fast but incomplete oxidation after 60 min was also observed for flufenacet and fluopyram. TFA yields were 19% and 32%, respectively. Sitagliptine was completely degraded after 60 min contact time in the batches with 4 mg/L, but the TFA yield was lower (4%). The biological degradation of the test compounds was investigated by conducting a modified OECD guideline 302 B Zahn-Wellens test. In the test, sewage sludge directly taken from the activated sludge basin of the local WWTP was used as inoculum and were spiked with an aqueous solution of the test compound to obtain a final concentration of 1 mg/L. Samples were collected at least once a week and at days 27 and 28 according to the guideline. The primary degradation of the compounds and the formation of TFA was followed by LC/MS/MS. At the end of the test (28 d) removal of 67% fluoxetine, 56% flufenacet, 51% flurtamone, 25% fluopyram and 20% sitagliptine was observed. A steady increase of TFA was observed and after 28 d the following TFA concentrations were measured: 1.4 μg/L (fluoxetine), 7.4 μg/L (flufenacet), 1.4 μg/L (flurtamone), 1.2 μg/L (fluopyram), 0.3 μg/L (sitagliptine). These TFA concentrations correspond to up to 5% molar transformation of the degraded parent compounds. 44 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) In a mechanistic study by Khan and Murphy (2021), the microbial degradation pathways of the pharmaceutical active substance fluoxetine by common environmental bacteria were investigated by 19F NMR and GC–MS analyses. After fluoxetine had been incubated with bacteria, TFMP was shown to accumulate, and it is proposed that the ether bond in fluoxetine is initially hydrolysed yielding 4-(trifluoromethyl)phenol (TFMP) and 3-(methylamino)-1phenylpropan-1-ol. The latter degraded further while TFMP remained in the culture supernatant. In a subsequent experiment, when TFMP was incubated with bacteria separately, it was degraded further and TFA was ultimately formed. In addition to TFA, 19F NMR signals from the meta-cleavage products were detected as well as for fluoride ions. The formation of fluoride ions was explained by a competing photolytic degradation of the meta-cleavage products, resulting in defluorination. The extent of this defluorination pathway was facilitated by the exposure of light. Based on these experimental observations and predicted intermediates from the EAWAG Biocatalysis/Biodegradation Database, the overall degradation pathways for fluoxetine were proposed by the authors (Figure B.39). 45 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.39. Degradation pathway s for fluoxetine based on the predicted intermediates from the EAWAG Biocatalysis/Biodegradation Database and the experimental observations. Degradation products that were observed with 19F NMR and/or GC–MS analyses are shown inside dashed boxes. In a study by Tisler et al. (2019), the degradation of fluoxetine by photolysis (solar simulator at 295-3 000 nm) during a 28-h experimental period in phosphate buffered water was monitored by LC-MS/MS analysis. A complex mixture of degradation products was identified including intermediate degradation products retaining the CF 3-group such as 4(trifluoromethyl)phenol (TFMP), but also defluorinated degradation products such as hydroxylated benzoic acid. TFA occurrence during direct photolysis was in average 0.3%. During indirect photolysis 1.5% of degraded fluoxetine was transformed to TFA. According to the authors, since TFA can be considered as a mineralization dead-end product, it will presumably occur in low amounts only, if photodegradation is incomplete and the concentrations of intermediates are still increasing. Hence, the level of TFA can be expected 46 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) to increase after extended photolysis (>28 hours). In a mechanistic study by Ellis and Mabury (2000), photolysis degradation experiments of 3trifluoromethyl-4-nitrophenol (TFM) were carried out at 365 nm in buffered deionized water (pH 7 and pH 9) and analysed by 19F NMR and HPLC–UV. The half-life of TFM at was found to be 22 h at pH 9 yielding 5.1% TFA, and 91.7 h at pH 7 yielding 17.8% TFA. In addition to TFA, the formation of fluoride ions was also observed and explained by competing degradation pathways that seemed to be facilitated by a higher pH (see proposed degradation pathways in Figure B.40). This type of defluorination degradation pathway (via the deprotonated TFM) has been reported previously for orto- and para trifluoromethyl phenol (as the only degradation pathway under abiotic degradation conditions; (Sakai and Santi, 1973). This was also verified in separate experiments with orto- and para trifluoromethyl phenol within this study. Figure B.40. Photolysis degradation pathways of TFM proposed by Ellis and Mabury (2000). Conclusions on degradation of PFCA precursors In conclusion, all PFCA precursor share the same basic structural features: a perfluorinated aliphatic part (F(CF2)n-) (linear/branched/cyclic ) attached to a degradable moiety, including for example -CH2-R, -aromatic group, -C(O)NRR’. These substances can be degraded to PFCAs by abiotic and/or biotic processes in the environment. However, there may be a large variation in the degradation rates, pathways and to what extent the corresponding PFCAs are formed depending on the specific environmental conditions. For those substances where no specific degradation studies are available, degradation pathways can in many cases be assumed based on the chemical similarity with related substances, irrespective of chain length or branching of the perfluoroalkyl moieties. Still, physicochemical properties like volatility and solubility of a specific substance will influence the partitioning to different compartments in the environment and thus affect the environmental fate. The above conclusions are to a large extent based on the reasoning in the Background documents to the Opinion on the Annex XV dossier proposing restrictions on PFOA (ECHA, 2018b), C9-C14 PFCAs (ECHA, 2018a), and PFHxA (ECHA, 2019c) which has been discussed and accepted by RAC for the purpose of these restrictions. 47 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.1.3.2. Degradation of fluorinated gases A part of the PFAS universe are fluorinated gases which end up in the atmosphere after releases and therefore degrade under different conditions, as compared to PFAS that mainly partition to water and soil. Following release into the environment, fluorinated gases reside in the atmosphere where they are oxidized into a variety of degradation products. Some degrade easily in the atmosphere, while others are more stable and require much longer times. Some degradation routes lead to formation of TFA as a single product in high yields which precipitates with rain and snow. In some cases, intermediate degradation products form, which may degrade further via several pathways. For example certain fluorinated gases have more complex degradation mechanisms which partly lead to complete degradation and formation of degradation products like CO2 and HF, while TFA formation is a minor degradation product. o Hydrofluorocarbons (HFCs) and hydrofluoroolefins (HFOs) When evaluating the degradation of fluorinated gases like HFCs and HFOs, there are some key intermediates which are formed from several different starting gases. These include trifluoroacetaldehyde (CF 3CHO), trifluoroacetyl fluoride (CF 3COF) and trifluoromethanol (CF 3OH), see Figure B.41 below. For example, fluorinated gases containing one or more C-H bonds are susceptible to attack by OH radicals in the lower atmosphere (Wallington et al., 1994). These radical processes lead to carbonyl compound intermediates, e.g. trifluoroacetaldehyde (CF 3CHO) or trifluoroacetyl fluoride (CF 3COF). It is known that the atmospheric decomposition of e.g. HFO-1234ze (CHF=CH-CF 3) yields trifluoroacetaldehyde (CF 3CHO) with 100% molar yield (Hansen et al., 2021; Nilsson et al., 2009; Qing et al., 2018). 48 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) 100% Trifluoroacetyl fluoride (CF3COF), Atm. lifetime short Trifluoroacetic acid (TFA), Precipitates CO2, HF Trifluoroacetaldehyde (CF3CHO) Atm. lifetime ca. 4 days Trifluoroacetic acid (TFA), Precipitates Three different degradation pathways: a) the OH-initiated abstraction reaction, b) hydrolysis or c) photolysis. TFA formation indicated at up to 10%. HF CO2, HF Trifluoromethanol (CF3OH) Atm. lifetime ca. 5 years GWP unknown, Toxic Figure B.41. Degradation routes of some key intermediates from fluorinated gases. 49 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The sequence of gas-phase reactions that follow from an initial attack of OH radicals on the parent halocarbon are sufficiently rapid that heterogeneous and aqueous processes play no role. In contrast, the lifetimes of the carbonyl products are relatively long (weeks) and hydrolysis in water droplets may be relevant for the removal of halogenated halogen compounds (Wallington et al., 1994). For trifluoroacetaldehyde (CF 3CHO) reaction with OH radicals is important, while trifluoroacetyl fluoride (CF 3COF) is removed almost entirely into water droplets. Although acid fluorides are almost insoluble in water, they hydrolyze quickly with formation of HF and the corresponding carboxylic acid which are very water soluble. Hence, hydrolysis removes trifluoroacetyl fluoride (CF 3COF) from the gas phase irreversibly as TFA. For trifluoroacetaldehyde (CF 3CHO) further gas-phase oxidation processes are important. The atmospheric degradation of trifluoroacetaldehyde (CF 3CHO) can occur via three competing reactions: a) the OH-initiated abstraction reaction, b) hydrolysis or c) photolysis ((UBA, 2021c), page 106, and references therein). TFA may be the outcome of some of these processes and subprocesses (e.g. path b, hydrolysis, Figure B.41), with CO2 and HF indicated as the final end products in the other processes. How important the three different degradation processes are relative to each other is unclear, while up to 10% formation of TFA from trifluoroacetaldehyde (CF 3CHO) has been estimated by UBA (2021c), page 109. Buszek and Francisco (2009) looked at the gas-phase decomposition of trifluoromethanol (CF 3OH) with water. They pointed out that it is known that trifluoromethanol quickly degrades into carbonyl fluoride (CF 2O) and HF at room temperature, while the photolytic lifetime of the substance in the atmosphere below 40 km is very long, see section B.4.1.4. Hence, trifluoromethanol in the atmosphere is acting as a sink for hydrofluorocarbons (HFCs) and hydrofluoroethers (HFEs). However, the authors identified a catalytic mechanism with water and OH radical to be relevant for the decomposition of trifluoromethanol and formation of carbonyl fluoride (CF 2O) and HF. Sulbaek Andersen et al. (2018) investigated the atmospheric degradation of HCFO-1233zd(E), E-CF 3CH=CHCl in a 3-dimensional global atmospheric chemistry and transport model. Atmospheric degradation of E-CF 3CH=CHCl is initiated by reaction with OH radicals, which leads to several chemical oxidation products. The atmospheric lifetime was estimated to ca. 36 days, and GWP at <5. The degradation pathways were shown to go via CF 3CHO as a key intermediate, which over time degrades further to HF and CO 2 or TFA. In this model TFA formation was indicated at approximately 2%. As HFO-1234ze also degrades via the intermediate CF 3CHO, a similar yield of TFA is expected, while HFO-1336mzz(Z) degrades with formation of 2 molecules of CF 3CHO and would therefore give approximately 4% TFA. In a comprehensive overview of the current knowledge on HFO degradation to TFA and the consequences for human health and the environment following from increasing TFAconcentrations in the environment, the group ATMOsphere pointed at the strong evidence for TFA levels increasing in the environment as a result of increasing HFO use (ATMOsphere, 2022). They strongly underlined the urgent need for policymakers to take action, as it is impossible to remove TFA from the environment at a later stage. In conference presentations and in a preprint publication Hansen et al. (2021) looked further into the atmospheric photodissociation of trifluoroacetaldehyde (CF 3CHO) as a degradation intermediate from HFO-1234ze. They found indications that although photolysis of trifluoroacetaldehyde (CF 3CHO) with formation of trifluoromethyl (CF 3) and formyl (CHO) radicals, which is further transformed into CO 2 and HF, is the dominating decomposition pathway (79%), up to 11% of the trifluoroacetaldehyde (CF 3CHO) in the atmosphere could decompose with formation of CO and fluoroform (CHF 3, HFC-23). Fluoroform has a GWP = 12 690, while its parent HFOs may have GWPs of less than 1 (e.g. HFO-1234ze). However, the authors point at uncertainties in the study and call for experiments to investigate these considerable findings further. The atmospheric lifetime of fluoroform (HFC-23) is ca. 228 years (Stanley et al., 2020). 50 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) In a recent study of the tropospheric photolysis of CF 3CHO, Sulbaek Andersen and Nielsen (2022) came to a different conclusion when they, in a chamber study, used broadband actinic radiation and FTIR spectroscopy for detection of the photolysis products. No formation of CF 3H (HFC-23), was observed under any of the experimental conditions and an estimated upper limit for the yield of HFC-23 of 0.3% was established. The atmospheric chemistry of short -chain haloolefins (e.g. HFO-1234ze) was investigated by Wallington et al. (2015). They concluded that haloolefins containing the CF 3CF= group leads to TFA as a persistent degradation product, while haloolefins containing the CF 3CH= group were reported to degrade with formation of CF 3CHO as the primary key intermediate. This general rule will identify among others the substances HCFO-1224yd(Z) (CF 3CF=CHCl), HFO1234yf (CF 3CF=CH2) and HFO-1216 (CF 3CF=CF 2) as TFA-precursors. The report by UBA on persistent degradation products of halogenated refrigerants and blowing agents (UBA, 2021c) investigates the degradation of e.g. hydrofluorocarbons (HFCs). For HFC-134a (CF3CFH2) degradation to 7-20% TFA is reported, while for HFC-227ea (CF3CHFCF3) 100% yield of TFA is stated. Trifluoroacetyl fluoride (CF 3COF), which hydrolyses to TFA in droplets, is reported as the key intermediate also for HFC-236ea (CHF 2CHFCF3). For HFC-4310mee (CF 3CF 2CHFCHFCF3) degradation to TFA, as well as PFPrA, is indicated in high yields. For HFC-125 (pentafluoroethane, CF3CHF2), Young and Mabury (2010) observed that TFA was formed as a degradation product under conditions of limited NOx. o Hydrofluoroethers (HFEs) Tsai (2005a) looked into the degradation of hydrofluoroethers (HFEs) as these are being used as third generation replacements to chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs) and perfluorocarbons (PFCs). With regards to atmospheric degradation, the author uses HFE-7100, C4F 9OCH3, as a typical example and explains that OH radical initiated hydrogen abstraction is common. Following oxidative reactions then lead to the corresponding formate ester C4F 9OC(O)H. Such substances are rather unreactive towards further radical processes but may undergo hydrolysis in droplets. Nohara et al. (2001) found that for the formate esters such, as C4F 9OC(O)H, the formate group is cleaved off with formation of the corresponding alcohol C4F 9OH, which suffers loss of F-atoms from the carbon atom attached the hydroxyl group and formation of the carboxylic acid C 3F 7CO2H (PFBA) with one perfluorinated C-atom less in the fluoroalkyl chain, i.e.: CnF 2n+1OCH3 --> Cn-1F 2n-1CO2H. o Perfluoroalkyl ketones According to a study by Taniguchi et al. (2003) the ketone substance CF 3CF 2C(O)CF(CF3)2 (FK-5-1-12) in the atmosphere suffers photolytic cleavage which results in CF 3C(O)F and COF 2. As indicated above, CF 3C(O)F will be incorporated into rain/cloud/seawater where it will undergo hydrolysis to give TFA, while COF 2 will be converted to CO2 and HF. The half-life of trifluoromethanol (CF 3OH) with respect to decomposition into COF 2 was found to be 4-5 h in this study. Ren et al. (2019) found that photolysis of FK-5-1-12 was the dominant loss pathway in the troposphere, with the substance having an atmospheric lifetime of 3-11 days. TFA and PFPrA were identified as degradation products. According to information from the manufacturer 3M (3M, 2017), the fluoroketone (CF 3)2CFC(O)CF 3 (C5-FK) will quickly hydrolyze in contact with water with TFA as t he primary hydrolysis product, and 3M recommends against uses of the compound where substantial amounts are intentionally released. o Perfluoroalkyl nitrile compounds The atmospheric chemistry of the nitrile coumpound (CF 3)2CFCN, (C4-FN) was studied by Sulbaek Andersen et al. (2017) in FTIR/smog chamber experiments and ab initio quantum calculations. They estimated the atmospheric lifetime of C4-FN at approximately 22 years and GWP at 1 490. The sole atmospheric degradation products were found to be NO, COF 2, and CF 3C(O)F. The latter is known to hydrolyze in droplets with formation of TFA. The yield of 51 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) formation of CF 3C(O)F from C4-FN was indicated at 100%. Conclusions on degradation of fluorinated gases In conclusion, fluorinated gases have a complex atmospheric chemistry often based on radical oxidation processes via trifluoroacetaldehyde (CF3CHO) or trifluoroacetyl fluoride (CF 3COF) as intermediates. The latter of these key intermediates is further hydrolyzed in high yield to trifluoroacetic acid (TFA) in water droplets which precipitates with rain and snow, while trifluoroacetaldehyde may degrade to TFA in 2 to 10% yield in one of three degradation pathways. Other key degradation products from fluorinated gases include longer chain substances like PFBA which may be formed e.g. from the hydrofluoroether HFE-7100 (C4F 9OCH3). B.4.1.3.3. Degradation of PFSA precursors The degradation pathways of several PFSA precursors into the corresponding PFSAs (mainly C4, C6 and C8 perfluoroalkane sulfonic acids) are described in the Annex XV restriction report for the restriction proposal on perfluorohexane sulfonic acid (PFHxS, (ECHA, 2019b)). Hence, this section is based on the information in the PFHxS Annex XV restriction report. It can be assumed that the degradation mechanisms for PFSAs of other chain lengths are the same as for C4, C6 and C8 PFSA precursors. In a literature study carried out by the University of Oslo (Nielsen, 2017), the formation of PFBS and PFHxS through abiotic degradation of precursors was investigated. PFBS/PFHxSrelated substances were found to include PFBS/PFHxS sulfonic acid halides, sulfonic esters (alkyl, olefinic and aryl) and sulfonamides, side-chain fluorinated polymers containing the PFBS/PFHxS moiety, as well as subclasses of PFBS/PFHxS-related substances like sulfones and sulfinic acids. Abiotic degradation of the identified precursors to PFSAs may proceed either via reaction with water or via oxidative radical processes in the atmosphere. However, in the radical processes, for the sulfonyl group may also be cleaved off in a different degradation pathway with formation of perfluoroalkyl radicals that may suffer sequential CF 2-loss and formation of shorter chain-length PFCAs (this pathway has been reported for perfluoroalkane sulfonamides, (D'Eon J et al., 2006; Martin et al., 2006); see Figure B.42). To what extent the precursors will end up as PFSAs or PFCAs may vary with the environmental conditions and is difficult to predict. The rate of degradation may vary for the different precursors, and in some cases the process may take years. Little information about the rate of degradation of such substances has been published. 52 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFHxS sulfonic halides, X = F, Cl, Br hydrolysis or radical processes PFHxS PFHxS sulfonic ester, R = alkyl, olefin, aryl radical processes ... PFHxS sulfonamides, R, R' = H, alkyl, olefin, aryl PFCAs Figure B.42. Degradation scheme of a selection of PFSA precursors, exemplified by PFHxS precursors. A review article on the atmospheric oxidation of organic sulfur-containing substances shows that dimethyl sulphide is oxidized in radical initiated oxidation processes in the atmosphere via dimethylsulfoxide and methane sulfinic acid to methane sulfonic acid as the end product (Barnes et al., 2006). Oxidation of the relevant sulfinic acids to PFBS and PFHxS is also described in a study of potential precursors to PFBS and PFHxS (Nielsen, 2017). The findings suggest sulphides, thiols and intermediate oxidation products as precursors to PFSAs, as shown in Figure B.43. 53 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.43. Oxidation processes from sulphides/thiols to the corresponding PFSA s. A review of the microbial degradation of polyfluoroalkyl chemicals in the environment points out that perfluoroalkane sulfonamido derivatives may undergo aerobic biodegradation, via the relatively stable intermediate sulfonamides, to the corresponding PFSAs as the final degradation products (Liu and Mejia Avendano, 2013). Liu et al. (2019a) investigated the biotransformation of perfluoroalkane sulfonamide compounds in aerobic soil and looked specifically at differences between the linear and branched isomers in the transformation of PFOS-precursors to PFOS. However, as there are several degradation pathways for the different precursors, there was no clear overall trend that differentiates between the linear and the branched precursors. In biological systems it has been demonstrated that perfluoroalkane sulfonamides like NEtFOSA are precursors to PFOS in fish (Tomy et al., 2004) and N-EtFOSA was biotransformed by earthworms to PFOS after in vivo and in vitro exposure (Zhao et al., 2018b). Further in vitro depletion of PFOS precursors (N-EtFOSA and perfluorooctane sulfonamide (FOSA)) was confirmed in a liver microsomal assay approach in polar bear, ringed seal and laboratory rat (Letcher et al., 2014). Perfluoroalkane sulfonamido alcohols like N-EtFOSE are degraded to PFOS in activated sludge (Rhoads et al., 2008) and levels of PFSA-precursors in sludge from WWTP exceeded those of PFSAs itself (Eriksson et al., 2017a). Conclusions on degradation of PFSA precursors In conclusion, all molecules that contain a C nF 2n+1SO2-, CnF 2n+1SO- or CnF 2n+1S- moiety (Figure B.44) can form the corresponding PFSAs (C nF 2n+1SO3H) through abiotic and/or biotic degradation in the environment. However, concerning perfluoroalkane sulfonamides, the sulfonyl group may also be cleaved off in a different degradation pathway in the atmosphere with formation of perfluoroalkyl radicals that may suffer sequential CF 2-loss and formation of shorter chain-length PFCAs. For those substances where no degradation studies are available it can be assumed that based on the chemical similarity, irrespective of chain length or branching of the perfluorinated moieties, they will most likely be degraded in a similar way (see also section B.4.1.2 concerning the effects of chain length, branching and cyclic structure elements on persistence). It was also concluded in the Annex XV restriction report for PFHxS that side-chain fluorinated polymers containing .e.g. perfluoroalkane sulfonamide-based sidechains, will degrade in the same way as the corresponding small-molecule precursors. 54 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.44. Generic structures of PFSA precursors.R = any chemical group, n = 0 or higher. PFAS subgroups that are considered as PFSA precursors includes, e.g.: o o o o o o o Perfluoroalkane sulfonyl halides Perfluoroalkane sulfonic acid esters Perfluoroalkane sulfonamides (FASAs) Perfluoroalkane sulfide derivatives Perfluoroalkane sulfoxide derivatives Perfluoroalkane sulfinic acid derivatives Side-chain fluorinated polymers based on sulfonic acid derivatives (mainly sulfonamides) B.4.1.3.4. Degradation of PFPA precursors o Perfluoroalkyl phosphinic acids (PFPiAs) Figure B.45. Degradation pathway of 6:6 perfluoroalkyl phosphinic acid (6:6 PFPiA). In a review by Wang et al. (2016b), available information on degradation of perfluoroalkyl phosphinic acids (PFPiAs) were evaluated. PFPiAs were found to degrade to perfluoroalkyl phosphonic acids (PFPAs) and 1H-perfluoroalkanes CnF 2n+1H under various laboratory conditions. The environmental relevance of this degradation remains however somewhat unclear. Biodegradation of PFPiAs into PFPAs were found in some in vivo studies, while no degradation of PFPiAa was observed in a 28-day OECD 301-F test on ready biodegradability. 1H-Perfluoroalkanes can also potentially be oxidized to form corresponding PFCAs (e.g. via reaction with OH radicals in the atmosphere (Wang et al., 2014c; Young and Mabury, 2010). Conclusions on degradation of PFPA precursors Based on the available data it can be expected that perfluoroalkyl phosphinic acids can be abiotically or biotically degraded and transformed int o the corresponding perfluoroalkyl phosphonic acids (PFPAs) and 1H-perfluoroalkanes CnF 2n+1H. 1H-Perfluoroalkanes could potentially degrade further to form the corresponding PFCAs. B.4.1.3.5. Degradation of other precursors Structural elements in combination: Above, the degradation patterns of different structural elements in per- and polyfluorinated substances have been investigated for representative PFAS substances. Most often perfluoroalkyl chains remain mainly intact, while degradation processes take place else where 55 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) in the precursor molecules, especially in non-fluorinated moieties. In many cases degradation stops when a perfluoroalkyl group attached to a functional group at its highest oxidation step has been reached, i.e. PFAAs. However, in some cases molecule s suffer loss of fluorine from the carbon atom next to the non-fluorinated part of the substance. Persistent final degradation products, i.e. arrowheads, have been investigated in detail in section B.4.1.2. The degradation processes of molecules are dictated by their chemical structure and the conditions that prevail where the substances are found. Hence, it is important to take into consideration if a substance partitions to air, soil or water. However, under identical conditions one specific type of functional group often behaves in the same way with similar neighbouring groups in a molecule. The length of a perfluoroalkyl chain, branching or the presence of cyclic structures is not expected to affect the reactivity and degradation of a functional group considerably. The same applies for the different types of polymers within the PFAS scope, which are in general assumed to follow the same degradation pattern for eac h specific functional group. The degradability of a substance can often be assessed by looking at one reactive structure element at the time, when these elements are separated by non-reactive moieties, like in many PFASs. Based on the understanding of the reactivity of structural elements in per- and polyfluorinated substances, one can assess expected degradation routes of similar compounds for which experimental studies of degradation have not been published. In combination with the knowledge summarised in section B.4.1.2 on persistent structural elements, one can estimate the degradation patterns, and in many cases the final degradation products, of a large part of the PFAS universe. B.4.1.4. Fully degradable PFASs Generally, PFASs are either very persistent themselves or will ultimately degrade to very persistent degradation products which are still PFASs (arrowheads). There are , however, a few specific PFAS subgroups with certain structural elements in combination that are fully degradable and cannot form persistent PFAS arrowheads (such as perfluoroalkyl acids). For the vast majority of PFASs regulatory standard tests for persistence are missing, and the assessment of full degradability is based on the general scientific literature and expert judgement. In this section, relevant available degradation data on substances belonging to these PFAS subgroups are summarised: fully degradable trifluoromethoxy-derivatives, fully degradable trifluoromethylamino-derivatives, fully degradable difluoromethanedioxyderivatives. First a brief chemical overview of degradation pathways is given, followed by experimental observations. B.4.1.4.1. Trifluoromethanol, trifluoromethylamine and difluoromethanediol In the first instance, the three parent compounds, trifluoromethanol, trifluoromethylamine and difluoromethanediol, are all inherently unstable compounds that spontaneously decompose, as shown in Figure B.46. 56 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Spontaneous decomposition above -20 C Trifluoromethanol, CF 3OH Spontaneous decomposition above -20 C Trifluoromethylamine, CF 3NH2 Never isolated due to instability, postulated as a transient intermediate in decompositions Difluoromethanediol, CF 2(OH)2 Figure B.46. Inherently unstable key fluorinated compounds. Trifluoromethanol (CF 3-OH) is a chemically unstable gas at room temperature and decomposes spontaneously (slow decomposition was found already at –20 °C). It degrades via the initial elimination of hydrogen fluoride (Christe et al., 2007; Redwood and Willis, 1965; Seppelt, 1977; Taniguchi et al., 2003) to form carbonyl difluoride that, in the presence of water, subsequently hydrolyses to yield carbon dioxide and hydrogen fluoride ((Francisco, 1993; Taniguchi et al., 2003; Uchimaru et al., 2004); see Figure B.47). 57 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.47. Degradation pathway of trifluoromethanol. Trifluoromethylamine (CF 3NH2) is a chemically unstable gas at room temperature and decomposes spontaneously at t emperatures above -20 °C via HF elimination (see Figure B.48). Figure B.48. Degradation pathway of trifluoromethylamine proposed by Klöter and Seppelt (1979). Difluoromethanediol (CF 2(OH)2) has never been isolated as such due to its low stability, but it has been postulated as a transient intermediate in the hydrolysis of carbonyl difluoride by Francisco (1993). It is considered a chemically unstable substance and decomposes spontaneously. Hence, it leads to full mineralization to HF and CO2 (see Figure B.49). Figure B.49. Postulated hydrolysis via difluoromethanediol by Francisco (1993). As the above three key perfluorinated compounds are chemically unstable, all compounds that lead to these substances during degradation, and no other perfluorinated degradation products, will degrade under any relevant environmental conditions. The perfluorinated part of these substances will eventually fully mineralize. Below, different substances that degrade via the three key substances with full mineralization are discussed.  Fully degradable trifluoromethoxy-derivatives A larger molecule containing a trifluoromethoxy-group attached to a degradable moiety (CF3X, where X = -OR and where R = methyl (-CH3), methylene (-CH2-), an aromatic group or a carbonyl group (-C(O)-)) is expected to degrade with formation of trifluoromethanol. As this substance is inherently unstable, it will quickly mineralize under environmental conditions as described above. Hence, it cannot be persistent in itself nor degrade to a persistent PFAS 58 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) arrowhead.  Fully degradable trifluoromethylamino-derivatives A larger molecule with a trifluoromethylamino-group attached to a degradable moiety (CF3X, where X = -NRR’ and where R and R’ = hydrogen (-H), methyl (-CH3), methylene (-CH2-), an aromatic group or a carbonyl group (-C(O)-)) is expected to degrade with formation of trifluoromethylamine. This substance is inherently unstable, as described above, and will mineralize under environmental conditions. Hence, it cannot be persistent in itself nor degrade to a persistent PFAS arrowhead.  Fully degradable difluoromethanedioxy-derivatives A larger molecule containing a difluoromethanedioxy-group attached to a degradable moiety (X-CF 2-X’, where X = -OR and X’ = -OR’’ and where R and R’’ = methyl (-CH3), methylene (CH2-), an aromatic group or a carbonyl group (-C(O)-)) is expected to degrade to difluoromethanediol. This substance is inherently unstable, as described above, and will quickly mineralize under environmental conditions. Hence, it cannot be persistent in itself nor degrade to a persistent PFAS arrowhead. In the following text examples of experimental observations of the degradation of compounds in the groups trifluoromethoxy-, trifluoromethylamino- and difluoromethanedioxy-derivatives are presented. Although the experiments have generally not been run under standard persistence test conditions, the results are regarded as supporting evidence for the above assessment based on general chemical principles.  Experimental observations - fully degradable trifluoromethoxy-derivatives In a study by Peschka et al. (2008), 10-(trifluoromethoxy)decane-1-sulfonate was exposed to a fixed bed bioreactor (FBBR) that was run with surface water from lakes and rivers and the biodegradation pathways were investigated by LC–MS, TOF-MS and ion chromatography (IC) analyses. Two degradation pathways were proposed by the authors (Figure B.50). 59 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.50. Degradation pathways of 10-(trifluoromethoxy)decane-1-sulfonate proposed by Peschka et al. (2008). The main degradation pathway (I) proceeds via a desulfonation step, oxidation and chain shortening via β-oxidation. Trifluoromethanol is subsequently formed and degrades immediately to inorganic fluoride and carbon dioxide. The minor degradation pathway (II) proceeds via oxidation of the alkyl chain forming the proposed oxo-10(trifluoromethoxy)decane-1-sulfonate, that was slowly degraded further, with assumed mineralization into fluoride, sulfate and carbon dioxide as final degradation products. The degradation via pathway I was nearly complete after 28 days, while a slow release of fluoride was continuing to the end of the experiment (87 days) which was presumably ascribed to a slow degradation of the oxo-10-(trifluoromethoxy)decane-1-sulfonate (pathway II). The overall mineralization of 10-(trifluoromethoxy)decane-1-sulfonate was ca 90% complete after 87 days based on the measured fluoride ion concentration. These results represent a water test that indicate biotic degradation of the substance in the environment. Biodegradation of two trifluoromethoxy-substituted alcohols, 6-(trifluoromethoxy)-hexan-1ol (TFMHxOH) and 3-(trifluoromethoxy)-propan-1-ol (TFMPrOH) were examined by Frömel and Knepper (2015). Experiments were carried out under aerobic conditions with effluent water from a municipal wastewater treatment plant and the biodegradation pathways were investigated by LC–MS and ion chromatography (IC) analyses. TFMHxOH was almost completely mineralized after 37 days as judged from the fluoride release. 6-Trifluoromethoxy hexanoic acid (TFMHxA) and trifluoromethyl carbonate (TFMC) where identified as transient transformation products (proposed degradation pathways in Figure B.51). 60 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.51. Degradation pathways of TF MHxOH proposed by Frömel and Knepper (2015). TFMPrOH on the other hand only mineralized to a 15% extent as judged from the fluoride release. TFMC was also detected during biotransformation of TFMPrOH, however, the major transformation product was 3-trifluoromethoxy-propanoic acid (TFMPrA), which did not show any further degradation under these conditions within a 47 days period (proposed degradation pathways in Figure B.52). The authors pointed out that the long-term fate of the remaining major transformation product TFMPrA is questionable and cannot be assessed without further laborious studies. However, based on the chemical structure of TFMPrA (non-perfluorinated methylene groups), it should not be regarded as an arrowhead, but rather a relatively stable intermediate that is likely to mineralize under environmentally relevant conditions over time. EAWAG-BBD pathway prediction4 of TFMPrA reveals plausible degradation pathways for 4 EAWAG-BBD Pathway Prediction System predicts plausible pathways for microbial degradation of chemical compounds. Predictions use biotransformation rules, based on reactions found in the EAWAG-BBD database or in the scientific literature. http://eawag-bbd.ethz.ch/predict/, date of access: 2022-09-29. 61 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) microbial degradation into trifluoromethanol. The results indicate biotic degradation of the substance in the environment. Figure B.52. Degradation pathways of TFMPrOH proposed by Frömel and Knepper (2015). In a study by Dihel et al. (2009), Sprague Dawley rats were given a single oral dose of the preclinical candidate OSI-930. LC-MS/MS analysis of the rat plasma revealed a substantial metabolite peak that was assigned to a glucuronide conjugate of OSI-930 (dTFM + Glucuronide, see Figure B.53). Based on these findings and further experiments of OSI-930 in vitro (in rat and human hepatic microsomes), the authors concluded that this biotransformation is CYP-mediated and they proposed an ipso-substitution mechanism for the displacement of the trifluoromethoxy group from the phenyl moiety (producing trifluoromethanol), followed by a rapid glucuronidation. CYP3A4 was found as the major facilitator of this pathway with a contribution from CYP2D6. These results show that trifluoromethoxy substituted aromatic substances may be metabolized with full mineralization of the perfluoroalkyl part by mammals in the environment. Although this study applies only to certain specific conditions, the results are taken as indications that biotic degradation may take place in the environment. Figure B.53. CYP-mediated degradation of OSI-930. One stakeholder has initiated a biodegradation study on the 14C radiolabelled 4-(14Ctrifluoromethoxy)benzoic acid in Q2 2022 (CropLife Europe (2022), see Figure B.54). A summary of the interim results (data collected at day 14) is presented below. This aerobic 62 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) soil degradation study is conducted according to OECD guideline 307 (OECD, 2002). According to the study protocol, soil samples were treated with the radiolabelled test item at 0.67 mg/kg (dry weight equivalent) and incubated under controlled and dark conditions at 20 °C. A closed incubation system with continuous aeration was used with an attached trapping system for the determination of volatile compounds. The incubations were performed in three different soils with 8 samplings over a planned period of 28 days. At each sampling date, evolved and trapped volatiles (e.g. 14CO2 in NaOH trapping solution) were quantified by LSC (liquid scintillation counting). For each soil and sampling time, two replicates were taken and extracted with acetonitrile/water mixtures and acetone, and the extracts were analysed by LSC and radio-HPLC in order to determine the amount of test substance remaining in soil and the number and amount of potential metabolites. The non-extractable residues were quantified by combustion in an Oxidizer and subsequent LSC measurement. All quantitative measurements of 14C are expressed as % of the total applied radioactivity (% TAR). Interim results after 14 days are available for two of the soils and show that the parent compound was almost completely degraded (<2% TAR of parent recovered in extracts). For both soils, the major degradation product was 14CO2 accounting for 57–58% TAR, meaning that complete defluorination/mineralization of the O-CF 3-group has occurred to this extent after 14-day incubation. No other volatile 14C-containing metabolite could be detected. Nonextractable residues (NER) were observed for both soils which reached maximum levels of 46% TAR at day 6 and 51% TAR at day 3, respectively, before they decreased again to levels of 32 and 36% TAR at day 14. According to the authors of the report, this indicates that part of the parent test item was bound to the humic matrix in soil. They argue further that the proceeding defluorination and mineralization (increasing 14CO2-levels) also show that this does not stop the further degradation of trifluoromethoxy-group, but just slows it down to a certain extent. Figure B.54. Degradation pathways of 4-(14 C-trifluoromethoxy)benzoic acid in a soil degradation study conducted according to OECD guideline 307. The above assessment explains that substances that degrade via trifluoromethanol (and no other perfluoroalkyl fragments) are not to be considered as persistent PFASs, as there are no possible stable PFAS that may be formed from the indicated substances. Furthermore, examples of degradation of specific substances have been included to show that there is experimental data to support this consideration.  Experimental observations - fully degradable trifluoromethylamino-derivatives Schiesser et al. (2020) investigated the hydrolytic stability of twelve trifluoromethylaminoderivatives in aqueous media (2 mM water/dimethyl sulfoxide 4:1 solution at room temperature; degradation monitored by LC-MS analysis). All trifluoromethylamino-derivatives were found hydrolytically unstable as they were hydrolysed to a significant degree already after 24 hours under these mild conditions (100% completion for ten of the derivatives and 75% and 10% completion respectively for two of the derivatives), yielding the corresponding carbamoyl fluorides (RR’NC(O)F), which are no longer PFAS, by elimination of 2 HF (see Figure B.55). 63 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.55. Representative examples of trifluoromethylamino-derivatives that were found hydrolytically unstable (Schiesser et al., 2020). The above assessment explains that this type of trifluoromethylamino-derivatives are not to be considered as persistent PFAS as there are no possible stable PFAS that may be formed from the indicated substances. Furthermore, examples of degradation of spec ific substances have been included to show that there is experimental evidence to support this consideration.  Experimental observations - fully degradable difluoromethanedioxy-derivatives In a mechanistic study by Bygd et al. (2021), the microbial degradation of model compound 2,2-difluoro-1,3-benzodioxole (DFBD) by Pseudomonas strains was investigated by 19F NMR, 1 H NMR, and GC–MS analyses. It was found that for bacterial strains expressing toluene dioxygenase (e.g. Pseudomonas putida F1), a rapid fluoride release was observed that declined within hours. The authors proposed a degradation pathway with the initial formation of 4,5-dihyro-dihydroxy-DFBD (4,5-DD-DFBD) and subsequent reactions producing fluoride ions (Figure B.56). This major path is highlighted by darker arrows. The three intermediates between 4,5-DD-DFBD and fluoride plus pyrogallol are expected to have a very short lifetime and represent a proposed mechanism for fluoride release. The minor products (thin arrows) were found stable under these conditions and within the timeframe for this study. These results only demonstrate degradation under certain specific conditions. However, the results are taken as indications that biotic degradation of such substances may take place in the environment. The Pseudomonas genome database and other databases revealed hundreds of bacteria with enzymes sharing high amino acid sequence identity to toluene dioxygenase from P. putida F1, which according to the authors suggest that the mechanism revealed here may apply to the defluorination of DFBD-containing compounds in the environment . 64 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.56. Biodegradation pathways of 2,2-difluoro-1,3-benzodioxole (DFBD) proposed by Bygd et al. (2021). In a study by Alexandrino et al. (2020), the biodegradation of fludioxonil by microbial consortia enriched from estuarine sediment and agricultural soil was investigated by pontentiometry and HPLC analyses (see Figure B.57). After an enrichment period of 6 months, four microbial consortia were able to completely remove and defluorinate fludioxonil under co-metabolic conditions in concentrations up to 10 mg/L, in a maximum of 21 days. Results suggest that defluorination is not a primary catabolic step, as fludioxonil removal was always faster than its defluorination. Three of the four enriched consortia had similar biodegradation performances in the absence of sodium acetate as a co-substrate. According to the authors, estimated half-life values were significantly lower than those report ed in literature, highlighting the unique metabolic performance of the obtained consortia. Analysis of their microbial composition revealed that they integrate several bacterial species belonging to the Proteobacteria phylum, with the most common genera being Pseudomonas, Ochrobactrum and Comamonas. The environmental relevance of the degradation in these consortia was not further elaborated by the authors. Figure B.57. Fludioxonil. 65 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The above assessment explains that substances that degrade via difluoromethanediol (and no other perfluoroalkyl fragments) are not to be considered as persistent PFAS as there are no possible stable PFAS that may be formed from the indicated substances. Furthermore, examples of degradation of specific substances have been included to show that there is experimental data to support this consideration. B.4.1.4.2. Conclusions on fully degradable PFASs In conclusion, it can be expected that PFASs from the subgroups described above can fully mineralize in the environment with basis in the understanding of the reactivity of key structural elements. Mineralization pathways have been identified and persistent arrowheads, such as perfluoroalkyl acids, cannot be formed along the expected degradation pathways of these PFASs. B.4.1.5. Persistence of PFASs under regulatory and scientific scrutiny The European Environment Agency stresses that the major concern of PFASs is due to their persistence, and that PFASs either are, or degrade to, persistent chemicals, with many of them accumulating in humans and animals, and all of them ultimately accumulating in the environment (EEA, 2020). The Global PFC Group points out that PFAAs are very persistent in the environment, whereas their potential precursors are transformed in the environment abiotically or biotically into PFCAs and/or PFSAs. PFAAs and their potential precursors are ubiquitous in the environment, even in remote regions. Several PFASs have lately been recognised as very persistent, potentially bioaccumulative and toxic (OECD/UNEP, 2013). According to the Californian toxic substances' authorities, all PFASs or their degradation, reaction, or metabolism products, are environmentally persistent. And for this reason, PFASs as a c lass are regulated in certain consumer products in California (Balan et al., 2021). It is emphasised by these authorities that persistence of a chemical in the environment promotes sustained exposure and contributes to accumulation in the environment. Because persistence is an inherent property of a chemical in the environment that results in increased exposure to the chemical and consequently potential for health risks, it can appropriately be identified as a hazard trait. In the Helsingør Statement on PFASs a group of scientists pointed out that the current knowledge demonstrates that the perfluorinated parts of any PFAS are recalcitrant and will form terminal transformation products, including PFCAs and PFSAs, which are persistent in the environment (Scheringer et al., 2014). Extensive and increasing use and emissions of fluorinated alternatives will lead to increasing levels of PFCAs, PFSAs and other stable perfluorinated degradation products in the environment, biota and humans. In the follow-up Madrid statement, it was warned that PFASs are very persistent man-made substances found everywhere. PFASs contain perfluorinated chains that only degrade very slowly, if at all, under environmental conditions (Blum et al., 2015). The high persistence of PFASs allows for a wide distribution in the environment, and many PFASs have been detected globally in the environment. A large group of scientists has reached a consensus that PFASs are the most environmentally persistent substances among organic chemicals and support a broad scope in restricting the use of PFAS in society (Cousins et al., 2020b). PFAS has been given the nickname “forever chemicals” in the popular press. Cousins et al. (2019) investigated the consequences of persistence for organic substances and provided case studies for three different classes of very persistent substances: chlorofluorocarbons, polychlorinated biphenyls, and PFASs. They argue that high persistence has important implications for the behaviour of chemicals in the environment. Persistent chemicals are distributed widely, often globally, and reach (much) higher concentrations than 66 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) short-lived chemicals emitted at the same rate. Looking more at detail at the Cousins et al. (2019) paper the Dossier Submitters find the following: The implications of high persistence for the levels and time trends of chemicals in the environment were modelled using a simple multimedia environmental fate model. The model was a so-called unit-world model with three compartments: the global troposphere (height 6 000 m, volume 3.06 x 1018 m3), the global surface ocean water (depth 100 m, volume 3.62 x 1016 m3), and the global surface soil (depth 0.1 m, volume 1.48 x 10 13 m3). In each compartment, a first-order degradation process takes place. In addition, there are three nondegradative losses: diffusion to the stratosphere, settling to deep ocean water, and burial in deep soil. Substances C and D were assigned the following properties (see Table B.6): Table B.6. Properties of compared substances. Log KAW Log KOW Substance C -1 8 t1/2 (d) 2 Substance D 2 000 -1 8 A half-life of 2 000 days or 5.5 years is long, but not excessively high. PFASs, for example most PFAAs, can have much longer half-lives. If PFAAs degrade, they do it so slowly that it is not observable and their half-lives could be on the order of decades, centuries or even greater. Log KOW for the two substances compared in the study was 8, which is typical for more lipophilic substances. In a first emission scenario, the same constant emission rate to air (100 mol h -1) was assumed for each c hemical and the concentrations in air, water and soil calculated at steady-state and in a dynamic scenario where the initial concentrations in all media are equal to zero. The model showed that an increase in the degradation half -life by a factor of 1 000 (from 2 days to 2 000 days) leads to an increase in the time to steady-state by a factor of 600–880 (from 20 days to 33–48.5 years). Similarly, the increase in the total inventory of chemical in the model system is only around a factor of 550 to 600 because of the increasing effect of the non-degradation losses. The long-lived chemical (chemical D) shows a marked overshoot with increasing concentrations for more than 4 years after the emission peak in year 10 (concentration peak in year 14.5, see Figure B.58 below). Moreover, the decreasing concentrations form a long tail that extends for many years after the stop of the emissions in year 20. An important finding from the model results is that the KOW is of less importance and does not modify the general implications of high persistence. 67 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.58. Implications of high persistence. Concentrations of chemicals C (panel (i)) and D (panel (ii)) as function of time in the scenario with dynamic emissions. For both chemicals, emissions start in year 0, increase by 10 mol h -1 every year, peak in year 10 at a value of 100 mol h-1, then decrease by 10 mol h -1 every year, and end in y ear 20. Note the much higher levels of chemical D compared to chemical C. In the case that unexpected effects are caused by a short -lived chemical, it is possible to rapidly cease environmental contamination by restricting or banning its use, which then also means that no additional effects will be caused by that chemical. In contrast, in the case of very persistent chemicals, it is not possible to cease environmental contamination within a reasonable time frame by simply restricting or banning their use. Environmental contamination by very persistent chemicals – and the effects related to this contamination – will continue for years to decades. This poor reversibility of contamination is because very persistent chemicals are, by definition, difficult to degrade. In summary, the main concerns with very persistent chemicals according to Cousins et al. (2019) are: (1) The continuous release of very persistent chemicals will lead to widespread, long-lasting, and increasing contamination. (2) Increasing concentrations will result in increasing probabilities that known and unknown effects occur, be it by a single chemical and/or in a mixture with other substances. (3) Once adverse effects are identified, it will be technically challenging, energy intensive, and thus costly, to reverse the chemical contamination and therefore the effects. These measures are limited to contamination hotspots, whereas, for most of the environment, no remediation or clean-up will be possible. It is argued that high persistence should be given particular emphasis in chemicals assessment and management and that very persistent chemicals should be regulated on the basis of their persistence alone (P-sufficient approach). Monitoring results underline the above listed main concerns of persistent chemicals PFASs' ubiquitously presence in the environment confirms their widespread occurrence. PFASs which have been monitored are present in soils, ground- and freshwater as well as drinking water across the world (see B.4.2). PFASs are even found in remote areas without anthropogenic activity. This also demonstrates their long-range transport potential (see B.4.2.8). Though their transport potential is also determined by other properties their persistence is crucial. 68 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The poor reversibility of contamination within reasonable time due to persistence can be demonstrated by time trends (see B.4.2.7.9). Though industry has voluntary phased out PFOS and it is also regulated globally as a POP environmental monitoring data do not show a clear downward trend. Though concentrations of marine predators in Europe level off, increasing PFOS (and PFCAs) concentrations were reported for biota living in contaminated regions with slow water exchange such as the Balt ic Sea which indicates a dilution in the environment rather than a degradation (see B.4.2.7.9). 69 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.2. Environmental distribution B.4.2.1. Adsorption/desorption/mobility in water B.4.2.1.1. Factors influencing adsorption potential of PFASs Chain length Adsorption happens dominantly via hydrophobic attraction. An increase in sorption with increasing chain length at least for longer chain PFASs (C5- C15) has been observed across the main PFAA subclasses (PFCAs, PFSAs, PFPAs, PFPiAs), in sediments (Higgins and Luthy, 2006), sludge (Arvaniti et al., 2014; Zhou et al., 2010a) and soil (Lee and Mabury, 2017). Elmoznino et al. (2018) demonstrated that an increase in log KOC correlates to the alkyl chain length for PFCAs (C7-C10) and PFSAs (C6 and C8). Baduel et al. (2017) demonstrated a predictable pattern for the effect of alkyl-chain length on mobility in soil for PFSAs (C1-C11), PFCAs (C 4-C10), FASAs (C3-C8) and ketone PFSAs (C6-C11) where the vertical distribution is a function of the alkyl chain length with higher mobility for shorter chain lengths. In sewage sludge, Zhang et al. (2013c) also recorded an increased sorption with increasing chain length for PFCAs(C3-C12) and PFSAs. Milinovic et al. (2015) reported that among three studied PFASs (PFOS, PFOA and PFBS), PFOS was the most strongly adsorbed by s ix different soils. The authors attributed the strong interaction of PFOS with soil particles to hydrophobic interaction, as indicated by a strong correlation between the log KOW values of the three PFASs, the functional hydrophilic group, i.e. sulfonic vs. carboxylic acid, and the log KOC values of the soils. Campos Pereira et al. (2018) showed that the PFCA and PFSA sorption was further found to increase with increasing perfluorocarbon chain length with 0.60 log KOC units per CF2 moiety for C3-C10 PFCAs and 0.83 log KOC units per CF2 moiety for C4, C6, and C8 PFSAs. Shorter-chained PFASs such as PFBA, PFPeA and PFBS were weakly sorbed (less than 10% on average), while longer-chained PFASs (PFOS, FOSA, C9-C11 and C13 PFCA) sorbed strongly (on average, 99-100%). In general, the shorter the chain length the more important the polar-polar interaction becomes (Zhao et al., 2012). For C2-C4 PFAAs, the adsorption on sludge increases with decrease in chain length (Zhang et al., 2013c). The interpretation of this phenomenon given by the authors is that the hydrophobicity of short -chain PFCAs and PFSAs decreases with decreasing chain length so that electrostatic interaction was dominant for sludge -water interactions for C2-C4 PFAAs in contrast to hydrophobic interactions, which dominate the sorption for longer chain PFAAs. Therefore, electrostatic interactions are also an importa nt factor though adsorption/desorption to soils is commonly normalised to the organic carbon fraction (i.e. KOC value) assuming hydrophobic interaction mainly governing adsorption/desorption. However, the clay fraction is for instance also considered a relevant sorption phase for organic cations such as the PFAAs (Droge and Goss, 2013). Functional groups Elmoznino et al. (2018) observed that PFSAs would partition more strongly to effluent-derived suspended particulate matter than PFCAs with the same number of perfluorinated carbons. The authors attribute this to differences in sorption, as log KOC values are one and two units lower for PFHxS and PFBS, respectively, than for PFOS. Also Campos Pereira et al. (2018) found that PFSAs sorbed more strongly than PFCAs. Lee and Mabury (2017) concluded that PFPAs are more sorptive than PFCAs at equal chain length by comparing the Kd values calculated for PFPiAs and PFPAs via aqueous loss method and direct soil analysis in a soil-sorption experiment with those reported in literature for other PFAAs. The differences between the sorption of PFPAs and PFSAs of equal perfluorocarbon chain length were not consistent between the direct soil analysis and aqueous loss method. Although there is no data available to compare the Kd of PFPiAS with other PFAAs of the same chain length, based on the data from Lee and Mabury (2017) it is expected that PFPiAs are at 70 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) least as sorptive as other PFAAs. Differences in sorption between sediment soil and sludge and the water phase have been reported for PFCAs, PFSAs and PFPAs with the same chain length. The partition coefficients between sediments and the overlaying water phase by direct analysis have been reported to be higher for PFOS than PFOA (Ahrens, 2011). In soil, Campos Pereira et al. (2018) demonstrated that the sorption of PFSAs was stronger than of PFCAs and the sorption increased with increasing perfluorocarbon chain length. An increase of 0.60 and 0.83 log KOC units per CF 2 moiety for PFCAs and PFSAs, respectively, was observed. Higher partition coefficient of PFSAs than PFCAs has also been reported in sorption experiments via the aqueous loss method in sediments (Higgins and Luthy, 2006) and activated sludge (Zhou et al., 2010b). Cyclic structure and ether groups Based on modelled data (COSMOlogic), cyclic PFAAs (C5-C7) can be (highly) adsorbed by soil (LogKOC >3.5), with increasing sorption with increasing number of perfluorinated carbons. The presence of ether groups in the carbon chain does not alter the electron density in the carbon chain. Thus, short-chain PFECAs and PFESAs are expected to behave similar as PFCAs and PFSAs, with shorter chain substances having low adsorption potential, which is expected to increase with increasing chain length. The high mobility of HFPO-DA (5 Carbons) is described in the Annex XV dossier on the proposal for identification of HFPO-DA as a substance of very high concern (ECHA, 2019d). Role of the sorbent Soils consist of organic matter, minerals and pore spaces filled with air and water (Bradry, 2010; Hellsing et al., 2016). Sand, silt and clay all provide minerals and surface area for the sorption of PFASs. Sand, silt and clay differ in their particle size, and smaller clay particles have colloidal properties carrying positive and/or negative charges. Influence of ions and pH versus fraction of organic components PFAS sorption is influenced by the soil pH and soil solution ionic strength. Campos Pereira et al. (2018) investigated the effect of solution pH and concentrations of Al3+ , Ca2+ and Na+ on the sorption of PFCAs and PFSAs in soils. The KOC showed a negative relationship to both pH (Δlog KOC /ΔpH = −0.32 ± 0.11 log units) and the soil organic matter (SOM) bulk net negative charge (Δlog KOC = −1.41 ± 0.40 per log unit molc g−1). The effects of cation treatment and SOM bulk net charge were evident for many PFASs with low to moderate sorption (C5– C8 PFCAs and C6 PFSA). However, for the most strongly sorbing and most long-chained PFASs (C9–C11 and C13 PFCAs, C8 PFSA and perfluorooctane sulfonamide (FOSA)), smaller effects of cations were seen, and instead sorption was more strongly related to the pH value. Hellsing et al. (2016) found that a negatively charged silica surface was not able to adsorb anionic PFASs such as PFHxA, PFOA, PFOS, and PFNA. On the contrary, positively charged alumina surface adsorbed significant amounts of these compounds, indicating that an electrostatic mechanism might come into partial effect for adsorbing PFCAs (C8-C11), PFSAs (PFOS and PFDS) and N-EtFOSAA, N-MeFOSAA on electrically charged soil components (Higgins and Luthy, 2006). One study investigated sorption to the phyllosilicate clay minerals illite, kaolinite, and bentonite (Droge and Goss, 2013). The authors point out that clays and minerals can have widely differing available surface areas for sorption and cation-exchange capacity values. It therefore would be challenging to include a generic parameter to account for clay sorption. It has been proposed that an increase in the fraction of organic components in the soil leads to more pronounced hydrophobic interaction as shown for PFOs, PFOS and 8:2 FTOH and PFBS (Brusseau, 2018; Milinovic et al., 2015). Influence of humic acid and formation of complexes Humic acid or other dissolved organic matter might form complexes with PFASs in the soil solution and inhibit sorption of those chemicals on to soil components such as clay minerals 71 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) and particulate organic matter. In river water, PFAAs like PFHxS, PFHxA, PFBS and PFOA were shown to be co-transported with dissolved organic carbon (DOC). These PFCAs and PFSAs carry a negative charge in natural waters, and the correlation was observed only for the shorter chained (≤C7 for PFCAs and ≤C6 for PFSAs) and more hydrophilic substances. The authors suggest that one possible explanation for the observed phenomenon could be that these PFASs readily bind to positively charged ions that are complex bound to DOC (e.g. Ca 2+) with negatively charged head groups, while the longer chained PFAAs (≥C7 for PFCAs and ≥C6 for PFSAs) rather partition to even more hydrophobic phases in the water, such as the organic carbon fraction of suspended particulate matter (Nguyen et al., 2019). Influence of proteins Protein binding as discussed in B.4.2.9 can also be an important factor influencing the adsorption for instance to sewage sludge. The presence of protein may lead to an increased adsorption of PFCASs (C2-C15) to sludge (Zhang et al., 2013a). For PFSAs, unlike PFCAs, carbohydrates were found to lead to an increase of adsorption. B.4.2.1.2. Adsorption and desorption of arrowhead PFASs As pointed out in section B.4.1., most PFASs degrade to their corresponding arrowhead PFASs in the environment. These arrowhead PFASs are therefore looked at with regard to adsorption potential. Physicochemical properties of PFASs are characterised in section B.1.2. pKa values With regard to electrostatic interactions it is important to differentiate between neutral and charged PFASs. PFCAs, PFSAs and PFPAs have low pKa values and are therefore almost completely dissociated at environmental relevant pH-values and therefore have a negatively charged headgroup. In contrast, perfluoroalkylamines have very high pKa values and thus will exist predominantly in the neutral form. Perfluoroalkylamines are therefore volatile depending on theier molecular mass and water solubility (B.4.2.4).Van der Waals interactions will however play a role for both groups, PFAAs and amines, dependent on the length of the hydrophic fluorinated carbon chain. Because soils are generally anionically charged, anionically charged PFAAs become more mobile whereas cationic charged bases such as perfluoroalkylamines become less mobile as they can be retained by cation exchange processes. However, for some soil types, such as those with metal oxides, which can have a large anionic exchange capacity, this general rule of thumb may not apply. In sum, an assessment merely on the KOC value may underestimate mobility of negatively charged PFAAs (UBA, 2019). KOC values According to the physicochemical data for PFCAs, PFSAs and PFPAs there is a trend of increasing KOC values with increasing chain length. The increasing adsorption potential pattern from PFCAs over PFPAs to PFSAs, as reported in the previous section, however, is not that clearly reflected in the KOC values. Perfluoroalkanes, which lack a functional group, have higher KOC values than the PFAAs of the same chain length. It is thus expected that PFASs without a functional group will be more adsorptive. It should, however, be noted that up to a chain length of 4 carbons perfluoroalkanes have boiling points below 0 °C and their Henry’s Law constants indicate that they are volatile (see B.4.2.4 on volatilisation). It is thus more likely that these short-chained perfluoroalkanes evaporate into the air when released to the environment. The same applies to the shorter chained perfluoroethers without any other functional groups. The two C4 perfluoroethers (1,1,1,2,2-Pentafluoro-2-(pentafluoroethoxy) ethane and 1,1,2,2Tetrafluoro-1,2-bis(trifluoromethoxy)ethane) have boiling points below 2.5 °C and 13 °C and their Henry’s Law constants indicate that they are volatile (see B.4.2.4 on volatilisation). 72 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) HFPO-DA, a branched C5 PFECA with an additional carboxylic acid functional group is described in the Annex XV dossier on the proposal for identification of HFPO-DA as a substance of very high concern (ECHA, 2019d). In the dossier log KOC values for HFPO-DA are 2.48 and 1.92 based on molecular connectivity indices and on estimated log KOW, respectively. Also, ADONA, a diether with five perfluorinated carbons and a carboxylic acid functional group can be considered mobile on the basis of its estimated log KOC <1.3. With respect to the carbon chain length, it should be noted that precursor substances which have a non-fluorinated moiety are expected to degrade to arrowhead PFASs with less carbons (see section B.4.1). Hence, those precursors are expected to form more arrowhead PFASs once they have been released to the environment. 73 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.7. Log KOC for PFAAs (sediment organic carbon-normalised distribution coefficient) shown in dependence of the carbon chain length. 10 11 1 Carbon 2 Carbons 3 Carbons 4 Carbons 6 Carbons 8 Carbons 9 Carbons Carbons Carbons Perfluoroalkylcarboxylic TFA PFBA PFHxA PFOA PFNA PFDA PFUnDA acids 2.76 2.39 3.3 (Higgins 2.06 (Higgins and (Higgins and and Luthy, (Higgins and 1.767 Luthy, Luthy, 2006) Luthy, (Predicted 2006) 2006) 1.63 – 2.35 4.8 (Ahrens 2006) using US 2.4 (Ahrens 3.6 (Ahrens (Sepulvado 0.437 et al., 1.09 EPA EPIet al., et al., et al., 2011) (Predicted 2010b; (Ahrens et Suite 2010b; 2010b; log K OC (sediment organic using US Ahrens et al., 2010b; (PC KOC WIN Ahrens et Ahrens et carbon-normalised distribution EPA EPIal., 2010c) Ahrens et v1.66)) al., 2010c) al., 2010c) coefficient) Suite al., 2010c) (PC KOC WIN v1.66)) C 1-PFSA Perfluoroalkane sulfonic acids C 2-PFSA C 3-PFSA 2.345 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 3.675 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 74 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) log K OC (sediment organic carbon-normalised distribution coefficient) Perfluoroalkyl phosphonic acids log K OC (sediment organic carbon-normalised distribution coefficient) 1 Carbon 2 Carbons 3 Carbons 0.352 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 1.016 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 1.681 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) PFMPA PFEPA 0.654 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 1.318 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 4 Carbons 6 Carbons PFBPA PFHxPA n.a. 3.977 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 8 Carbons 9 Carbons 10 Carbons 11 Carbons 75 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.8. Log KOC for PFCs shown in dependence of the carbon chain length (sediment organic carbon-normalised distribution coefficient). Perfluoroalkanes log K OC (sediment organic carbon-normalised distribution coefficient) 1 carbon 2 carbons 3 carbons 4 carbons 5 carbons 6 carbons 8 carbons C -PFC C 2-PFC C 3-PFC C 4-PFC C 5-PFC C 6-PFC C 8-PFC 1.687 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 2.352 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 3.016 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 3.681 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) n.a. 5.010 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 6.339 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 76 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.9. Log KOC for perfluoroalkylamines and perfluoroethers (sediment organic carbon-normalised distribution coefficient). Perfluoroalkylamines Acronym molecular formula log K OC (sediment organic carbonnormalised distribution coefficient) PFMAm PFEAm PFPrAm PFBAm PFHxAm C 3F9N; C 6F15N; C 9F21N; C 12F27N; C 18F39N; [(C F3)3N] [(C 2F5)3N] [(C 3F7)3N] [(C 4F9)3N] [(C 6F13)3N] 3.104 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 5.098 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) n.a. n.a. n.a. 4.433 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) C 2F6O C 4F10O C 4F10O 2 C 6F14O 3 C 2F4O C 3F6O 1.330 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 2.660 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 4.706 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.946 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 0.932 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 1.596 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) C 5F13N Perfluoroethers molecular formula log K OC (sediment organic carbonnormalised distribution coefficient) 77 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Conclusion on sorption Several studies have confirmed a relationship between sorption and perfluorocarbon chain length. Although generally there is a linear relationship between the carbon chain length and log KOC value, it should also be noted that for the substances with shorter carbon chain length, the polar-polar interaction, determined by a functional group such as carboxylic acid or amine, gains importance. Most of the studies investigating distribution focus sed on PFCAs and PFSAs but also PFPAs, PFPiAs, PFECAs and PFESAs, additionally cyclic PFAAs (C5-C7) have been investigated to some extent. PFPAs are more sorptive than PFCAs at equal chain length. Sorption differences of PFPAs and PFSAs did not follow a sy stematic pattern. PFAASs have both hydrophobic fluoroalkyl chains and hydrophilic ionizable functional group and thus show complex behaviours in the environment in terms of their sorption and desorption processes (Ahrens, 2011; Kannan, 2011). Neutral PFASs are more likely to adsorb to organic matter or partition into the gas phase. It is difficult to predict the sorption of the PFASs universe from a single sorbent bulk property. The properties of the sorbent needs to be considered as well. B.4.2.1.3. Mobility in water Specific criteria have been proposed for identifying mobile or very mobile substances and are currently under consideration for including into legislation. For example, the German Environment Agency (UBA) has proposed the following: M is indicated by water solubility ≥0.15 mg/L and log Koc ≤4.0 or log Kow ≤4.0, and vM by water solubility ≥0.15 mg/L and log Koc ≤3.0 or log Kow ≤3.0 (UBA, 2017a). In the continuing discussion water solubility has been c onsidered not to be a suitable property to set a threshold for the assessment of mobility. The principal reasons are difficulties when assessing ionic and ionizable substances, in which water solubility is dependent on counter ions (Rüdel et al., 2020; UBA, 2019). For the purpose of this restriction proposal no specific cut off values are proposed to be used but the comparisons below are for background information. With respect to the suggested mobility criteria described above and on the basis of their KOC , all PFCAs up to a chain length of 11 carbons can be considered either vM (≤ C10_PFCA, PFDA) or M (PFUnDA). Likewise, all PFSAs up to 6 carbons can either be considered vM (≤ C4-PFSA, PFBS) or M (PFHxS) based on their KOC values. The same can be concluded for the PFPAs. Considering the perfluoroalkylamines, only 1,1,1-Trifluoro-N,Nbis(trifluoromethyl)methanamine (PFMAm) fulfils the M criterion based on the Koc value. With the exception of C4F 10O2 all perfluoroethers fulfil either the vM or M criterion based on their KOC values. It is noted that perfluorinated olefins, and alkenes in general are expected to degrade to PFCAs (see section B.4.1.3), which are mobile. Perfluoroalkanes up to four carbons would be considered mobile. However, as mentioned above, it should be noted that up to chain length of 4 carbons perfluoroalkanes have boiling points below 0° C. It is thus more likely that these short chained perfluoroalkanes evaporate into the air when released to the environment. The same applies to the short-chain perfluoroethers without any other functional groups. (see also B.4.2.1.2) Three PFASs have been accepted as being mobile so far: - - PFBS has been identified as a substance of very high concern based on its equivalent level of concern as compared to PBT/vPvB substances: very high persistence, high mobility in water and soil, high potential for long-range transport, and difficulty of remediation and water purification as well as moderate bioaccumulation in humans (ECHA, 2019e) HFPO-DA has been identified as a substance of very high concern based on its equivalent level of concern due to its mobility and persistence (ECHA, 2019d); RAC has agreed that PFHxA, its salts and related substances possess properties, in particular very high persistence combined with mobility, that can be considered to constitute an intrinsic hazard (ECHA, 2021a). 78 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Many PFASs are precursors of PFAAs, which have been demonstrated above to be either mobile or very mobile. Hence a large part of PFASs can be considered as mobile in water, either by themselves or as result of their degradation into PFAAs. However, no or insufficient data are available on physico-chemical properties and fate of many PFASs (especially precursors). Therefore, uncertainties remain regarding mobility of several PFASs in water. Several PFASs have been detected in fresh and ocean water as well as ground and drinking water indicating their mobility. Though routine target analyses mainly focus on PFCAs and PFSAs and some of the precursors of these PFAAs, other PFASs can also be present in these compartments, which is further supported by the large fraction of unknown organofluorine compounds in aquatic samples (see B.4.2.6.1). For example, less studied PFASs, such as HFPO-DA, HFPO-TA, ADONA, 6:2 Cl-PFESA, 8:2 Cl-PFESA, 6:2 H-PFESA and 6:2 FTSA, have been ubiquitously detected in worldwide surface waters (Wang et al., 2019c). In addition, studies on the aquatic environment published between 2009 and 2017 have (tentatively) identified 455 new PFASs (nine fully and 446 partially fluorinated compounds), although for most of these compounds there are no quantitative monitoring data due to the lack of reference standards (Xiao, 2017) (see also B.4.2.6.3). The review of review papers by Sims et al. (2022) on surface water monitoring data note that >C9 PFCAs and PFSAs were generally detected at lower concentrations compared to shorter-chained PFCAs and PFSAs in both surface waters and groundwater (see B.4.2.7.2). Physico-chemical properties and environmental partitioning dynamics clearly affect fate in the environment as discussed in B.4.2.1.1 and B.4.2.1.2. Shorter-chained PFAAs are more likely to be distributed to drinking water than longer-chained PFAAs based on their properties. However, also for PFASs considered not as mobile due to their adsorptive properties it cannot be excluded that they, due to the high persistence, enter drinking water. In this context it has to be also noted that ultra-short chained PFAAs (C1C3) are often lacking in monitoring programs as they are analytically challenging (Björnsdotter et al., 2020). However, available data indicate that TFA generally occurs in considerably higher concentrations than other PFASs in aquatic samples. Other ultra short PFAAs may also significantly contribute to the total amount of PFASs in water samples (see B.4.2.7.3 and Table B.77 to Table B.80). Mobility as a concern Mobility is a contributing factor for 1. 2. 3. 4. Potential for long range transport via water (see section B.4.2.8) Potential for drinking water contamination (see also next subsection) Uptake in plants and crops (see section B.4.4) Making very persistent substances available for increasing internal concentrations in biota along the increase of the environmental exposures (see section B.4.2.9, “Persistence compensating low bioaccumulation potential” and section 1.1.4. of the main report “High potential for ubiquitous, increasing and irreversible exposure of the environment and humans”). For substances mobile in water phase, there are no local or intermittent sinks for the pollution stock, and therefore mobile substances have a high potential for continuously increasing environmental concentrations and exposure of wildlife and humans. Oceanwater is important as a sink and for transport of these compounds. The occurrence of high concentrations of PFASs in coastal waters could possibly be problematic because the substances will be bioavailable and can accumulate in the marine food chain (Cai et al., 2012). Furthermore, it is difficult in practice to manage exposures due to high mobility and exposures at different locations and times. Mobile PFASs may end up in drinking water, posing a potential risk to human health. Apart from soil as a long-term source of 79 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) groundwater contamination, statistical analysis showed a correlation between the major contaminants in the river and those in the groundwater, indicating the potential linkage of PFASs in the groundwater to the surface water ((Li et al., 2020a), see also B.4.2). Reemtsma et al. (2016) concluded that persistent and mobile organic compounds may be of concern for water quality because they are persistent in the environment and are not removed from water by sorption processes due to their high polarity and excellent water solubility. A problem which has been underestimated due to an analytical gap in the past. Raw water which is used for drinking water is obtained either from groundwater, bank filtration or surface waters. On average about 50% of the water for drinking water production is taken from groundwater, whereas the amount from surface water is about 36% (EC, 2016). With regard to groundwater and bank filtration, adsorption and desorption in soil and sediment are thus crucial parameters for drinking water quality. Due to their persistence, the residence time of PFASs in groundwater is at least the residence time of groundwater (>40 days up to an order of millennia (Małoszewski and Zuber, 1982; McGuire et al., 2005)) because transport away from the site in water is the only removal mechanism. As a consequence, PFAS-contaminated groundwater can act as a long-lasting source, leading to poorly reversible exposure (Cousins et al., 2016). Human exposure via drinking water is further discussed in B.9.21.2. Cases of contaminated sites provided below in this section and in Annex E illustrate the long-lasting problems and hardly reversible contaminations with groundwater and drinking water contamination. However, not only incidences cause elevated levels of PFAS in drinking water. A three-year monitoring campaign (2010–2013) investigated the occurrence, sources and fate of nine PFCAs and three PFSAs, in the most industrialized region of Italy (Castiglioni et al., 2015), also discussed in section B.4.2.3 (soil). Samples were collected in influents and effluents of wastewater treatment plants (WWTPs), in the main rivers flowing through the basin, and in raw groundwater and finished drinking water. The mass balance of the emissions in the River Lambro basin showed continuously increasing contamination from north to south and differences in the composition of homologues in the west and east sides of the basin. Ground and drinking water were contaminated in industrial areas, but these substances were removed well in Milan. Contamination from industrial sources was prevalent over urban sources, contributing to 90% of the loads measured at the closure of the basin. The River Lambro was confirmed as one of the main sources of contamination in the Po River. Moreover, an epidemiological investigation was carried out on the population living in an area contaminated by PFASs in drinking water and possibly affecting the food chain. The ongoing investigation shows an important set of a priori evidenced adverse health effects (Mastrantonio et al., 2018) (see B.5). Several studies have demonstrated PFAS contamination of water bodies due to the use of aqueous film forming fire-fighting foams (AFFF) at e.g. airports. For example, contamination away from its point source (which was likely a military airport in the north of the city) was demonstrated in Uppsala, Sweden. The PFAA levels in groundwater measured in 2012–2014 decreased downstream from the point source, although high ΣPFAA levels (>100 ng/L) were still found several kilometres from the point source in the Uppsala aquifer (Gyllenhammar et al., 2015). Several studies have also shown that fluorochemical production facilities are potential point-sources of several PFASs, including perfluoroalkyl ethers, to surrounding environmental compartment including water bodies (see B.4.2.7 and B.4.2.3). WWTPs are also an important source for contamination. For example, the effluent of WWTP in Halle was found to contain four times higher levels of PFAAs than river water and was dominated by PFBS with 32 times higher concentration than the riverine level. In this case annual flux of PFAAs from River Saale was estimated to be 164 ± 23 kg/y (Shafique et al., 2017). 80 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Drinking water contamination Following summarized monitoring data on detection of PFASs in drinking water underline that drinking water contamination is a concern already not limited to a few locations but ubiquitously. Further monitoring data for specific PFAS groups are also summarized in B.4.2.6, B.4.2.7, and Appendix B.4.2.1.3. Mobility in water Table B.79. Jian et al. (2017) reviewed the global distribution of PFASs in drinking water from Europe in France, Spain, Norway, Netherlands, Faroe Islands and Germany and Thailand, Korea, Japan, Ghana and the US (Boiteux et al., 2017; Domingo, 2012; Eriksson et al., 2013; Eschauzier et al., 2013; Haug et al., 2010a; Llorca et al., 2012a; Schwanz et al., 2016). The reviewed monitoring programs of targeted analysis mainly focussed on PFAAs. In this review, PFAS levels varied in the order of well water > tap water > bottled water > drinking water > raw water. The highest PFAS contamination in well water indicated that point sources could be the main cause. For results diverging from this order, the authors assume that greater PFAS concentrations in tap and drinking water than in raw water indicate a role of drinking water treatment 5 processes in PFAS contamination. As discussed in section B.4.5, shorter-chain PFASs are generally less efficiently removed during water treatment process compared to the longer-chain PFASs. For example, Eschauzier et al. (2012) demonstrated that granular activated carbon (GAC) could efficiently reduce longer-chain PFAAs, but had little effect on shorter-chain PFAAs such as PFBA and PFBS. It was proposed that the degradation of precursor compounds and long-chain chemicals could lead to PFAS increase at the WWTP (Eriksson et al., 2013). Thus, purification processes greatly influence PFAS levels in post-treatment water, while the carbon chain lengths of PFASs in untreated water also have significant influence on the purification efficiency (Schwanz et al., 2016). Castiglioni et al. (2015) analysed drinking water collected in the whole Lambro basin in Milan, Italy. This study confirmed PFAAs as ubiquitous contaminants in the aqueous environment and identified the main sources and the environmental fate of these substances in the most industrialized and heavily inhabited area in Italy. All the substances (C4-C12 PFCA, C4, C6 and C8 PFSA) were detected in drinking water at frequencies higher than 60% in northern Milan (an industrialized area), around 30–50% in the metropolitan area of Milan (except long-chain PFCA, which was not detected) and were never detected in southern Milan (agricultural area). In comparison to the above-mentioned studies, the levels of longer-chained PFASs such as PFOA, PFOS and PFHxS in the Milan area were the highest (up to 1 886 ng/L for PFOA, 150 ng/L for PFOS and 141 ng/L for PFHxS in drinking/tap water according to data given in the IPCHEM database). Same observations are made for shorter-chained PFASs - the highest concentrations have been measured in Italian drinking/tap water: up to 556 ng/L (PFBA), 347 ng/L (PFBS), 267 ng/L (PFPeA), 240 ng/L (PFHxA) and 100 ng/L (PFHpA) according to data given in the IPCHEM database6. In comparison PFBA has been detected up to 27 ng/L maximum and 10 ng/L average in tap water in Spain; PFBS up to 36 ng/L maximum and 8.3 ng/L average in Spain, PFHpA up to 16 ng/L maximum and 8.1 ng/L average in Spain and 1.2 ng/L average concentration in France , PFHxA up to 11 ng/L maximum and 4.7 ng/L average in Spain and 3.7 ng/L average and France, and PFPeA up to 17 ng/L maximum and 3.8 ng/L average in Spain and 2.7 ng/L average in France (Kaboré et al., 2018; Llorca et al., 2012a). Other PFASs have been also detected in tap water but the concentrations have been mostly below 1 ng/L. Schwanz et al. (2016) investigated bottled drinking water (n=38) from France, Spain and Brazil. The profile of compounds in bottled water was significantly different for the samples from the three studied markets. For example, PFOS was only detected in France (present in 26% of the samples), PFBS was detected in France and Brazil but not in Spain, whereby, Spain was the only country where PFHxA was quantified in 20% of the samples. The median 5 https://www.sciencedirect.com/topics/earth-and-planetary-sciences/water-purification, date of access: 2022-09-29. 6 https://ipchem.jrc.ec.europa.eu/, date of access: 2022-09-29. 81 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) concentration of the total PFASs in bottled water was as follows: Brazil > France > Spain with 15.0, 14.9 and 11.3 ng/L, respectively. However, the highest level of contamination in a single sample was found in a bottled water sample that was commercialised in France, with 116 ng/L as the sum of all individual PFASs. In the present work, in bottled water the most frequently detected compound was PFHpA, which was detected up to 47% in the samples from France and 40% in Spain. While French and Brazilian samples showed contamination by various PFAS, above the limit of detection, in Spanish samples only PFHxA, PFHpA, PFOA and PFNA were found. However, the authors hypothesise that the source of contamination is the migration from the dopants used in the polymers (in general, PET) in the fabrication of water bottles and in the ink from labels can be the source of part of PFASs in water though residual environmental contamination can be a source. PFASs have also been detected in bottled drinking water from Ireland (Harrad et al., 2019) as well as Germany (Llorca et al., 2012a). The detection frequency of bottled water from Ireland (n=21) was 87% for PFOA, 29% for PFOS, 29% for PFBS, 19% for MetFOSA and PFNA and 42% for MetFOSE. The detection frequency was 100% (n=2) for PFHpA in bottled water from Germany. The concentration in these samples was 6.6 and 12 ng/L, and one of them contained also PFOS at 1.0 ng/L. Likewise it was hypothesized that the origin of PFHpA is the plastic of the bottles, although there is no available data to confirm it. A recent global survey found a widespread distribution of short -chain PFAAs in drinking water (Kaboré et al., 2018), with PFBA being detected in 58% of bottled water and 92% of tap water, PFPeA detected in 32% of bottled water and 68% of tap water, PFHxA in 50% of bottled water and 64% of tap water, PFHpA in 42% of bottled water and 90% of tap water, PFBS in 47% of bottled water and 88% of tap water. All the above-mentioned studies have focussed on analysing PFCAs and PFSAs in targeted analysis. However, the ban on PFOA has led to the production and use of alternative fluorinated compounds, such as HFPO-DA which now has been identified as a substance of similar concern due to its persistence and mobility. HFPO-DA and other perfluoroalkyl ethers have been detected in drinking water especially in the vicinity of fluoropolymer production plants (see B.4.2.7.6), but also further away from the point -sources which illustrates the problem of mobile PFASs. Taken together, the concentrations and relative contributions of different conventionally analysed PFAAs in drinking water differ between studies, but the individual levels of the most abundant PFAAs, i.e. PFOA, PFOS, PFHpA, PFBA, PFPeA, PFBS, PFNA are generally within the same order of magnitude. Furthermore, studies of ultrashort PFAAs show that TFA (shortest PFCA) is present in drinking water at roughly 2 orders of magnitude higher concentrations compared to the above mentioned PFAAs and other ultrashort PFAAs have been found in largely comparable levels to the conventionally measured PFAAs (see B.4.2.7). Besides the few routinely analysed PFASs, a large fraction of the total organofluorine in drinking water remains unidentified (see B.4.2.6). Apart from target analyses, studies of total organic fluorine as well as screening studies show that other, but not yet reliably identified and quantifiable PFASs are present in drinking water (see B.4.2.6). In a study by Kaboré et al. (2018), 104 suspect-target PFASs were screened in drinking water samples from Canada and other countries (Burkina Faso, Chile, Ivory Coast, France, Japan, Mexico, Norway, and the USA). The study is the first to observe perfluoroalkane sulfonate (PFECHS) and C4–C6 perfluoroalkane sulfonamides (FBSA, FHxSA) in drinking water. In addition, Guardian et al. (2021) was the first study to show the presence of 2-(Nmethylperfluorooctanesulfonamido) acetic acid (N-MeFOSAA) in drinking water and 3 novel PFASs (C5H5OF 8, C6H4O2F 6, and C9H2O2F 16) were detected using suspect screening in source water in Philippines and Thailand. 82 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Conclusion on mobility of PFASs Mobility is of concern for several reasons as these properties contribute to the LRTP, uptake in plants and crop and the potential to contaminate drinking water. There are no local or intermittent sinks for the pollution stock, and therefore mobile substances have a high potential for continuously increasing environmental concentrations and exposure of wildlife and humans. As discussed above many of the arrowhead PFASs can be considered as mobile or very mobile based on their KOC values with respect to the suggested mobility criteria. Several PFASs have been detected in fresh and ocean water as well as ground and drinking water indicating their mobility. Monitoring data on detection of PFAS in drinking water underline that drinking water contamination is a concern already not limited to a few locations but ubiquitously. B.4.2.2. Sewage sludge Sewage sludge is a by-product of wastewater treatment in municipal and industrial sewage treatment plants and has been identified as relevant anthropogenic source for the release of PFASs into the environment (Bossi et al., 2008; Kallenborn et al., 2004). The recycling of sewage sludge as amendment for agricultural soils is a regular practice in the EU (Hudcová et al., 2019; Kacprzak et al., 2017) and is widely considered a potent release pathway of PFASs from sewage sludge (Aro et al., 2021b; Bossi et al., 2008; Navarro et al., 2016; Schultz et al., 2006; Semerád et al., 2020). PFAS contamination of agricultural soil has been directly linked to sewage sludge amendments in the US (Washington et al., 2010) and China (Wen et al., 2014). EU regulations do not require testing of sewage sludge for PFAS contamination before application (Hudcová et al., 2019). Therefore, data is limited. Bossi et al. (2008) and Semerád et al. (2020) hypothesized that soil amendments with PFAS contaminated sewage sludge can lead to the contamination of agricultural soils and leaching into ground water (see also B.4.2.3). The uptake of PFASs from sewage sludge contaminated soil into plants has been demonstrated for PFAAs (Lee et al., 2014; Wen et al., 2014), similar to the uptake into soil organisms (Navarro et al., 2016). The migration of pollutants from soil to plants or soil organisms could facilitate a probable entry pathway into the food chain (Navarro et al., 2016). B.4.2.3. Soil A recent review demonstrated that PFASs are present in soils across the globe, and indicated that soil is a significant reservoir for PFASs (Brusseau et al., 2020). A major aspect of concern is the long-term migration potential to surface water, groundwater, and the atmosphere as well as potential human exposure via the soil-groundwater-drinking water-path or the soil-groundwater-nutrition (plant or animal) path. Brusseau et al. (2020) compiled data on soil concentrations of PFASs for “uncontaminated” sites (to derive background concentrations of PFASs in soil), contaminated sites as well as secondary source sites. Contaminated sites include PFAS manufacturing sites, fire training sites and other AFFF-associated locations at airports and military installations, and an airplane crash site. The secondary-source sites include sites that are adjacent to PFAScontaminated primary-source sites, or sites for which PFAS-contaminated media were used for different purposes e.g. for irrigation or as fertilizer. Examples of such cases are described below. The potential importance of soil as a global reservoir for PFASs was first quantified by Strynar et al. (2012) who measured the concentrations of 13 PFAAs (PFTrDA, PFTeDA, PFDoA, PFUdA, PFDA, PFNA, PFOA, PFHpA, PFHxA, PFDS, PFOS, PFHS, PFBS) in samples of surface soil collected from 60 locations in 6 countries. Strynar et al. estimated global soil loadings of 1 860 and >7 000 t of PFOA and PFOS, respectively. Rankin et al. (2016) reported concentrations of 32 PFASs (PFCAs and PFSAs) in surface soil samples collected 83 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) from 62 locations across all continents. Quantifiable levels of more than one PFAS were present in all samples tested, including soils collected from remote locations. Washington et al. (2019) used data from Rankin et al. (2016) to calculate global soil loadings for eight PFAAs (PFCAs with chain lengths C6 – C12 and PFOS). The combined estimated load for all eight PFAAs ranged from 1 500 to 9 000 t, with mean estimates of approximately 1 000 t for both PFOA and PFOS. These results indicate that soil has the potential to be a substantial reservoir for PFASs. Another study reported a meta-analysis of PFASs soil-to-groundwater concentration ratios for samples collected from 324 AFFF source-zone sites across 56 military installations distributed throughout the U.S. (Hunter Anderson et al., 2019). The results demonstrated that soil is a significant reservoir for PFASs at these contaminated sites. Transport modeling conducted at individual contaminated sites also indicates that soils and the vadose zone serve as a significant long-term source of PFASs (Weber et al., 2017). In this context it is worth noting that point -source contamination may be challenging to identify unless locations of use or production has been identified or is suspected in an area. Results from groundwater monitoring near point sources indicate that PFAS concentrations in groundwater and aquifers are highly variable even within a few kilometres, at least based on the currently available information, which inherently is influenced by site-specific geological factors. For example, levels of PFOA around an industrial park in China were on average around 15 000 ng/L, however, when compared to observations only 1-4 km away, mean PFOA dropped to 23 ng/L, and then further decreased to 2.6 ng/L when sampled 4-10 km away from the facility (Liu and Liu, 2016). A similar gradient was seen by Filipovic et al. in Stockholm, Sweden for groundwater contamination by AFFF at a former air force base. Concentrations of PFOS in some wells reached 42 200 ng/L while others within a few kilometers were as low as 7 ng/L (Filipovic et al., 2015). Another point of interest is the relative ranges of soil versus groundwater concentrations reported for PFASs (Hunter Anderson et al., 2019) in order to estimate the potential for groundwater contamination via soil-leaching. In this context soil monitoring studies investigating depth profiles of PFASs can reveal how PFASs with different properties distribute in soil and subsequently their potential for leaching into groundwater. The basic assumption is, that the differing physico-chemical properties of different PFASs lead to differences in their distribution within the soil column (.e.g. (Baduel et al., 2017; Buck et al., 2011; Casson and Chiang, 2018; Dauchy et al., 2019; Sepulvado et al., 2011; Washington et al., 2010)). This leads to different PFAS concentrations depending on the sampling depth. Brusseau et al. (2020) - based on findings from the above-mentioned studies - described a trend according to which the greatest concentration of longer-chain PFASs ≥C7 (referring to Buck et al. (2011)) are mostly found closer towards the surface (within 1 m, according to Baduel et al. (2017)). Accordingly, maximum concentrations for short-chain PFASs are measured at greater depths (>2 m according to Baduel et al. (2017)). Brusseau et al. (2020) summarized the observations from different studies as follows: “The majority of depth-profile data sets show high concentrations present at shallow depths and exponential decreases at greater depths”. Hunter Anderson et al. (2019) described positive c soil/cgroundwater ratios for the vast majority (87%) of data they took into account, reflecting greater soil than groundwat er concentrations. In this context soil concentrations ~100-times greater than groundwater were observed. This may pose a long-term source for groundwater contamination due to leaching. The following examples for PFAS contamination in soil are provided to support the link between the mobile and persistent properties of PFASs and the type of contamination. Further monitoring data from contaminated areas are presented in Appendix B.4.2.3. Table B.81. 84 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.2.3.1. Examples from Europe Cases of land application of industrial-waste derived amendment are for instance described in the study by Wilhelm et al. (2008). Industrial waste with high concentrations of PFASs was manufactured into a soil improver by a recycling company and spread by farmers on agricultural land of the rural area Sauerland, Germany. This led to substantial environmental pollution and had an impact on drinking water. A second site with a similar cause of contamination is located in Baden-Wuerttemberg, Germany. In the surroundings of Rastatt, 480 hectares of former arable land are contaminated with short-chain PFASs. The pollution was detected in 2013 and has probably been caused by the longstanding application of compost mixed with sludge from paper production, contaminated with various precursors. Over time, shorter-chained PFASs (e.g. PFBS, PFBA, PFHxA) and precursors (e.g. 4:2 FTOH, 6:2 FTOH) in the soil wash out into the groundwater. Two groundwater wells for drinking water production had to be closed. Until now, no practicable solution for removing the short -chain PFASs from the soil or groundwater has been found. Furthermore, there are still high concentrations of PFASs (Brendel et al., 2018). In Italy, on January 2014, drinking water contamination in an area of the Veneto Region was detected mainly due to the drain of fluorinated chemicals by a manufacturing company operating since 1964. More details on the drinking water contamination are provided in B.4.2.1.3 Drinking water works in the municipality of Kallinge, Sweden were immediately closed down after contamination of the groundwater was discovered near a Swedish Air Force and civil aviation base where PFAS-containing AFFFs have been used (Jakobsson et al., 2014). Field investigations were carried out in the vicinity of four sites where AFFFs are or were intensively used (two airports, a training center for firefighters and an oil storage depot after a large explosion). In case of the incident of a fire at an oil storage depot 28 years had passed until the field investigation. PFAS profiles were influenced by parameters such as route of PFAS transport after use (runoff, seepage, direct discharge), time elapsed since the cessation of firefighting activities, and firefighting foam composition. The authors conclude that the PFAS concentrations found around the investigated sites are the highest recorded in France (Dauchy et al., 2017). A study investigating the impact of two fluoropolymer manufacturing facilities on downstream contamination of a river and drinking water resources clearly demonstrated the impact on the water quality of the drinking water resources (Bach et al., 2016a). Landfills have also been shown to contaminate adjacent areas. For example, in a survey of leachates from 117 Swedish landfills, all samples contained detectable levels PFAAs. The most abundant PFAAs were PFBA, PFBS, PFPeA, PFHxA and PFOS with mean concentrations between 427 and 1 242 ng/L (maximum concentrations of 22 800 – 74 600 ng/L). PFAAs were also found in nearby ground water at mean concentrations ranging between 67 to 129 ng/L and in surface water at generally lower mean concentrations ranging between 15 and 30 ng/L (Miljösamverkan Sverige, 2022). Another Swedish study has shown that ultrashort (C1-C3) PFAAs also contribute significantly to the total PFAA content in leachate (Björnsdotter et al., 2019). It should be noted that most of these cases were discovered by chance. Other cases may be yet undetected. Consequently, the use of PFAS-contaminated media such as biosolids and irrigation water can result in soil contamination, subsequent distribution to other media, and ultimately the potential for human exposure at locations far from the original PFAS source (Lindstrom et al., 2011; Liu et al., 2017). 85 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.2.3.2. Examples outside Europe In June 2000, 22 000 L of fire-retardant foam containing perfluorinated surfactants was accidentally released at L. B. Pearson International Airport, Toronto, ON, and subsequently entered into Etobicoke Creek, a tributary to Lake Ontario (Moody et al., 2002). A few years later, the same creek was subject to contamination with approx. 48 000 L of AFFF which were applied to a burning aircraft, which had overrun the runway (Oakes et al., 2010). This incidences may be seen as an involuntary long-term field study on the distribution of PFASs. For instance, even a decade after the first spill, sediment PFOS concentrations are were still elevated in Spring Creek Pond which received the foam discharge (Awad et al., 2011). The decades-long disposal of manufacturing waste containing PFASs in landfills in Minnesota resulted in contamination of groundwater serving as the drinking water supply for the eastern Twin Cities metropolitan region. In 2004, local and state agencies in Minnesota were alerted to the presence of PFAAss in the drinking water supplies of several eastern Twin Cities suburbs. A study conducted in 2010, six years after the discovery, measured PFASs in garden produce due to past/ongoing water contamination (Scher et al., 2018). B.4.2.3.3. Conclusion on soil distribution PFASs are present ubiquitously in soils, not only at contaminated sites but they are also found in “uncontaminated” sites far from point sources. At the most contaminated sites, PFAS concentrations in soil range up to ppm levels. Many cases of soil contamination with PFASs are known inside and outside of Europe and the assumption is that many contaminated sites are yet to be discovered. Soil can thus be seen as a global reservoir for PFAS contamination and human exposure via drinking water or food. PFASs in agricultural soils are a potential point of entry into the food web via crops or animals. Additionally, PFASs are retained at high concentrations in the vadose zone where they eventually can reach groundwater which is a potential source for drinking water. B.4.2.4. Volatilisation The threshold for volatile substances is Henrys Law Constant >250 Pa*m³/mol, according to REACH Guidance R.16 (ECHA, 2016a). For PFASs which are below this threshold aqueous compartments are more relevant regarding their distribution compared to the atmosphere. Neutral PFASs can have a relatively high vapour pressure whereas dissociated, charged PFASs have a negligible vapour pressure, are soluble in water, and have a very low air−water partition coefficient (Barton et al., 2007; Kaiser et al., 2005). Vierke and coworkers concluded in their work on PFCAs that the extent of volatilization of PFCAs in the environment will depend on the water pH and their pKa. Knowledge of the pKas of PFCAs is therefore vital for understanding their environmental transport and fate (Vierke et al., 2013). PFAAs may exist as neutral PFASs with higher volatility and lower water solubility or ionic PFAS with lower volatility and higher water solubility. Considering their low pKa value s, it can be assumed that the PFAAs almost completely dissociate at environmentally relevant pH values and are therefore charged, have a low vapour pressure and higher water solubility than their neutral forms and their volatility can be regarded as negligible. In their review on distribution modelling the authors consider volatilization not a major concern for most PFASs with functional groups which dissociate such as PFAAs (Sima and Jaffé, 2021). Furthermore, it can be expected that larger molecules with a high molecular weight are non-volatile. The largest molecules among the PFASs, e.g, side-chain fluorinated polymers, gradually degrade into the PFAAs, and are therefore expected to have negligible volatilisation (see also section B.4.1.3). Same applies to other PFASs which contain a large 86 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) non-fluorinated aromatic moiety. Neutral PFASs such as the perfluorinated olefins, perfluoroethers, perfluoroalky lamines and halofluoroalkanes are volatile depending on their molecular size and water solubility (see section B.1.2). As discussed in the section on long-range transport (B.4.2.8) precursors such as fluorotelomer alcohols (FTOHs) are very volatile due to their high vapour pressures and non-ionic status (Chen et al., 2020a). Similarly, uncharged PFASs like perfluoroalkane sulphonamides (FASAs), perfluoroalkane sulfonamidoethanols (FASEs) are less watersoluble and more volatile. For many of these PFASs, degradation to a less volatile and more water-soluble arrowhead PFAS applies as well, depending on the chain length of the resulting arrowhead (see section B.4.1.3). B.4.2.5. Distribution modelling Not assessed. B.4.2.6. Monitoring of unknown PFASs in environmental compartments Targeted PFAS analysis commonly includes up to 40 individual PFASs. However, given the much larger number of PFASs in total, methods that measure unspecific organic fluorine are used to estimate the total amount of PFASs in a sample. In this section, environmental monitoring data for PFASs of unknown identity are summarised, including analyses by extractable organic fluorine (EOF), adsorbable organic fluorine (AOF), the total oxidizable precursor assay (TOPA) as well as non-target and suspect screening. For a description of the analytical methods and their applicability see Annex E.4. B.4.2.6.1. EOF/AOF Adsorbable organofluorine (AOF) and extractable organofluorine (EOF) analysis provide two different sum measures of the organic fluorine in the sample, including known and unknown organofluorine compounds (including non-PFAS compounds) with different physico-chemical properties. The covered organofluorine compounds may vary depending on which of the two methods is used. Fluorine mass balance analysis of EOF/AOF and the sum of identified PFASs via targeted analysis allows the estimation of the fraction of unidentified organic fluorine (UOF) in a sample. Available measurements of EOF/AOF in the environment cover a wide range of environmental media, including surface water, drinking water, ground water, wastewater, sediment, sludge, soil, dust, and wildlife. For a detailed list of studies, including main results, see Appendix B.4.2.6.1. Table B.82 and Table B.83. For human matrices, see section B.9.22. Below follows a summary of the main result s from selected environmental studies. EOF/AOF and fluorine mass balance analyses in abiotic environmental matrices show varying, but significant, fractions of unidentified PFASs (40-99.9%). The varying fractions could be due to sample location (e.g. a contaminated area), but also to e.g. sample characteristics, study design, choice of analytical method (e.g. extraction method, detection limit) and the number of PFASs included in the targeted analysis. While in most studies, a few to a few dozen analytes were measured in the targeted analysis, a recent study (Aro et al., 2021b) on effluent and sludge water in the Nordic environment monitored 73 PFASs in addition to EOF. Surface- and groundwater overall displayed a large fraction of UOF. In samples without known point-sources, the UOF fraction was generally >90%, whereas contaminated water samples normally contained a lower but highly variable fraction of UOF. One study reported in European surface waters without known contamination, EOF levels in the range <30-849 ng F/L in Finland, Sweden, Iceland, Denmark, Greenland, Norway and the Faroe Islands, and an average UOF of 92% (Kärrman et al., 2019). The EOF concentrations in 87 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) this study were similar to two studies of German rivers reporting EOF levels (without targeted analysis) of 40–60 ng F/L in Moselle, 50–300 ng F/L in Rhine (Metzger et al., 2019), and 50–550 ng F/L in river Spree (Gehrenkemper et al., 2021). The highest overall European EOF concentrations (800–4 000 ng F/L) were detected in contaminated river and lake surface waters in Norway, where UOF was on average 99% (Aro et al., 2021b). Outside Europe, the highest level of EOF in surface water (440 000 ng F/L) was reported from the Cape Fear River in the US, downstream a fluoropolymer manufacturing plant (Han et al., 2021a). The EOF was largely composed of perfluoro-2-methoxyaacetic acid, used in fluoropolymer manufacturing, demonstrating that UOF can be low in areas contaminated with specific PFASs. EOF/AOF and fluorine mass balance analyses in biota show, similar to the abiotic samples, varying fractions of UOF (range 10-99%). Kärrman et al. (2019) analysed EOF as well as 78 targeted PFASs in samples of biota (n=51) from the Nordic countries. The results showed average EOF levels ranging from 221–707 ng F/g (Figure B.59). Despite the relatively large number of targeted PFASs analysed, significant and varying fractions of UOF were present in the samples, ranging from 32% UOF in bird eggs to 82% in reindeer. In a study on marine mammals from the northern Hemisphere, the EOF could in most of the samples be explained by the sum of target PFASs, and in particular the bioaccumulative long-chain PFAAs (Spaan et al., 2020). However, for samples taken at the US east coast, 30-75% of the EOF remained unidentified, which indicates sources of unidentified organofluorine. In another study (Schultes et al., 2020b), samples of killer whales from east Greenland showed that although most of the EOF in liver, blood, kidney, lungs and ovaries could be explained by targeted and known PFASs, almost none of the EOF in blubber, which displayed the highest concentration of EOF, was attributed to PFASs in the targeted analysis. Figure B.59. EOF (ng F/g) detected in different matrices and comparison between target PFASs and EOF, expressed as mass balance (% of known PFASs) (Kärrman et al., 2019). In the different studies, PFASs detected via targeted analysis in biota are predominantly the long-chain PFSAs (especially PFOS) and PFCAs, most likely due to their bioaccumulating properties. Conversely, ultrashort-chain PFAAs (C1−C3 PFAAs), that are often not included in the target analyses, may contribute significantly to the fluorine mass balance in abiotic and aqueous samples, due to their high water solubility and low adsorption. PFAAs detected in environmental samples may to a varying degree originate from degradation of known or unknown precursor compounds. These precursors are also likely to yield several intermediate products before degrading into their respective stable arrowhead PFAAs. Since only fractions of the original precursors degrade fully into PFAAs, the unknown intermediates are likely significant contributors to the fluorine mass balance and UOF. In the study by Kärrman et al. (2019), there was an indication of decreasing UOF from abiotic samples to biotic samples (Figure B.59), indicating possible metabolism of precursors to arrowhead PFAAs. In addition, by increasing the oxidation in a sample using the total 88 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) oxidizable precursor assay (TOPA), as discussed in section B.4.2.6.2, more arrowheads are being formed, decreasing the UOF. In conclusion, studies on EOF/AOF combined with targeted PFAS analyses show that varying but significant fractions of organofluorine in environmental samples are unknown and not captured by conventional monitoring using only targeted PFAS analysis. There are indications that higher trophic levels have lower fractions of UOF, possibly due to metabolism of precursors into the stable PFAAs which are often included in the targeted analyses and among which many tend to biomagnify within food webs. B.4.2.6.2. TOPA The total oxidizable precursor assay (TOPA), which estimates the amount of unknown oxidizable PFCA precursors in a sample, has been applied to abiotic samples such as water, soil and suspended particular matter as well as to biotic samples. Overall, studies show that unknown PFASs constitute varying fractions of the total PFAS content in environmental samples, ranging from small to very large. A number of relevant studies are provided below. Joerss et al. (2020a) analysed PFASs in river water from Germany and China. Samples were collected close to seven suspected point sources and analysed for 29 individual PFASs, with and without pre-treatment with TOPA. Upon oxidative conversion by TOPA, a marked increase in C4-C7 PFCAs was observed in the samples from Germany (59 ± 19%), with an 88 ± 30% increase of PFBA (C4 PFCA), demonstrating the presence of unknown precursor substances to unregulated PFASs. The increase in C4-C7 PFCAs was less marked in the samples from China (15 ± 10%). In one study, 83 freshwater, marine and soil samples from the German environmental specimen bank were analyzed for 32 PFASs using target analysis, before and after TOPA treatment (Gockener et al., 2021). After TOPA, only C4-C14 PFCAs were analysed based on the assumption that all precursors are oxidized to PFCAs. The result showed that the concentrations of primarily C4-C10 PFCAs increased in most samples, by factors of up to 90, demonstrating that substantial fractions of precursor substances were present in the samples which would not have been detected without the use of TOPA. A follow -up study (Gockener et al., 2022) on 100 samples of suspended particular matter from the German environmental specimen bank, collected between 2005-2019, where 41 PFASs were analysed, showed that ΣPFASs were 1.3 – 145 times higher after direct oxidizable precursos assay (dTOP) as compared to the targeted analysis, demonstrating large amounts of unknown precursors in the samples. Simonnet-Laprade et al. (2019) examined the trophic transfer of PFASs in the Orge River near Paris, France, with the aim to investigate the contribution of precursors to the biomagnification of PFAAs. 30 PFASs were analyzed in samples that were subjec t to treatment with TOPA in water (n=1), sediment (n=1) and biota (n=15), including biofilm, invertebrates, and fish. The result showed that substantial proportions of extractable unknown precursors were present in all samples (15–80% of the sum of PFASs) and that a systematic increase in the sum of PFCAs was observed (+15–424%) following TOPA. The largest relative increases were found at the base of the trophic web, in biofilm (424%), sediment (319%), leaf litter (298%), and macrophytes (196%) whereas lowe r relative increases were found in fish (15–22%), pointing to biotransformation of unknown prePFAAs in the trophic web to known PFAAs. In a feeding experiment with laying hens and feed grown on a PFAS-contaminated site, 25 PFASs were analyzed in the feed and eggs with/without TOPA to assess the transfer of PFASs into eggs (Gockener et al., 2020a). The results showed that application of TOPA demonstrated the presence of significant fractions of precursor substances in the feed and in the eggs. In the feed, the levels of C4-C8 PFCAs increased after TOPA, ranging from 232% (PFPeA) to 786% (PFOA). Similarly, in the egg yolk PFOA increased by 647%, 89 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) whereas C5-C7 PFCAs increased by more than 1 000% from levels below LOQ. Altogether this demonstrated the presence of unknown precursor substances in the feed as well as in the eggs. Janda et al. (2019) analysed the impact of TOPA treatment on concentrations of PFCAs including ultrashort C2-C3 substances. From a soil core of the PFAS-contaminated Upper Rhine Valley area in Germany, samples at different depths were analysed for C2–C14 PFCAs as well as C4–C8 and C10 PFSAs, with and without TOPA treatment. After the TOPA, concentrations of C2–C14 PFCAs increased considerably in the upper soil layers (0-30 cm), with total molar PFAA-concentrations increasing by factors of 1.6 to 5.0. More than 40 cm below the surface, precursors of TFA, PFPrA and PFBA (i.e. C2-C4 PFCA) accounted for >50% of the measured PFASs, showing their importance in mass balances based on TOPA. Chen et al. (2019a) analysed 39 precipitation samples from 28 cities in mainland China for 22 PFASs, including the ultrashort TFA and PFPrA (C2-C3 PFCAs), with and without TOPA treatment. The result showed that unknown precursors accounted for 6-56% of the total molar concentrations of PFASs, demonstrating an underestimation of PFAS mass load from precipitation if only monitoring data on known PFASs were analysed. TFA and PFPrA constituted large fractions (22-91%) of the ΣPFASs after TOPA treatment. Similarly, in leachates from municipal solid waste facilities in China, Wang et al. (2020b) showed from the use of TOPA that unknown C2-C12 PFAA-precursors constituted 0.35 – 68 molar% of the leachates and that unknown precursors to C2-C3 PFCAs represented major fractions. In conclusion, studies of abiotic and biotic samples utilizing TOPA demonstrate that considerable fractions of PFASs in the samples may be comprised of unknown oxidizable PFASs that are not detected in routine target analyses, including precursors to C2-C3 PFCAs that are rarely analysed. Thus, environmental samples commonly contain P FASs with unknown identity and regulatory status and the total PFAS mass balances may be underestimated without treatment with TOPA. B.4.2.6.3. Non-target and suspect screening of PFAS Non-target and suspect screening are qualitative or semi-quantitative analytical methods that have been used to identify unknown PFASs. In suspect screening analysis (SSA), the accurate mass of molecular features obtained from HRMS are compared to databases with known PFASs (Koch et al., 2020), e.g. the Norman Network PFAS Suspect List, the OECD's New Comprehensive Global Database for PFASs 7 and the USEPA CompTox Chemistry Dashboard. Further confirmation of the structure will then be compared with available mass spectra from literature and databases, such as the MassBank8. Non-target screening (NTS) applies different data filtration/data mining techniques to identify compounds without a priori knowledge about which substances that may be present in a sample. Non-target and suspect screening have been applied to environmental samples, including water, sediment, sludge, air, soil, airborne particulate matter and biota. For screening of human matrices, see section B.9.21. The paragraphs below summarise the results from a selection of relevant papers. In a review by Liu et al. (2019d), on NTS studies published until October 2018, it was shown that NTS methods using chromatography coupled to high-resolution mass spectrometry (HRMS) had been used to (tentatively) identify more than 750 PFASs belonging to more than 130 classes with small (e.g. C 2F 4) to very large (e.g. C44F 79) PFAS moieties in selected environmental samples, biofluids or commercial products. Different PFCAs, PFSAs, fluorotelomers and per/polyfluoroalkyl sulfonamides together composed 82% of all discovered PFAS subclasses and 74% of all PFAS analytes. This estimate of total PFAS discoveries may still underestimate the total number of PFASs molecules observed 7 http://www.oecd.org/chemicalsafety/portal-perfluorinated-chemicals/, date of access: 2022-09-29. 8 www.massbank.eu, date of access: 2022-09-29. 90 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) because of unresolved structural isomers and a possibility of the occurrence of coelution (i.e. inadequate chromatographic separation of similar substances). In 2019, Ng et al. (2022) investigated a large number of PFASs by target and suspect screening in 95 samples of river water, wastewater, ground water, sediment and biota from the Danube River Basin (passing through 11 European countries). For the targeted PFAS screening, 56 commercially available reference standards were used. The suspect screening covered 4 777 PFASs retrieved from t he NORMAN Substance Database 9. In total, 82 PFASs were detected in the samples, of which 10 were detected by targeted screening and 72 via the suspect screening. This demonstrated that targeted PFAS analyses with reference standards only covered a small fraction of all PFASs in the studied environmental compartments of the Danube River Basin and indicates a potentially large -scale migration of PFASs in Europe. Spaan et al. (2020) performed a fluorine mass balance on 25 liver tissues from 11 different species in the northern hemisphere, sampled between the years 2000 to 2017, using a combination of targeted PFAS analysis, EOF and total fluorine determination, and suspect screening. In 7/25 samples the EOF was significantly greater than the sum of the targeted analytes, with 30−75% of the EOF unidentified. Of the identified 63 PFASs, the suspect screening provided 37 PFASs that were not part of the targeted analysis, demonstrating the value of a multiplatform approach for determining PFAS mass balance in environmental samples. Barrett et al. (2021) employed non-target and suspect screening to examine temporal trends of 22 legacy PFASs as well as unregulated PFASs in livers of 98 beluga whales in Canada collected from 2000 to 2017. In total, 54 PFASs were identified via integration of non-target (34 identified PFASs) and suspect screening (39 identified PFASs), demonstrating the advantage of using the two complementary screening methods. 21 of the 54 PFASs were confirmed via traditional targeted analyses using reference standards. In conclusion, non-target and suspect screening have identified hundreds of different PFASs in various environmental samples. These substances would go undetected if only targeted analyses with available reference standards were performed. Consequently, the targeted analyses underestimate the number of PFASs as well as the total PFAS mass balance, however the suspect and non-target screening can only provide qualitative data of substances, and not quantitative, giving limited value to mass balance calculations. B.4.2.7. Monitoring of specific PFASs in environmental samples In this section monitoring data are summarised for specific PFAS groups, with focus on PFASs that are not covered by existing or proposed restrictions under REACH and/or the POPs Regulation ((EU) No 2019/1021). The data presented here is not a complete review of studies. Instead, key studies have been selected to exemplify environmental occurrence of PFASs, especially of those not routinely included in monitoring studies. Reported concentrations of PFASs in these studies are presented in Appendix B.4.2.1.3. Mobility in water, Appendix B.4.2.3. and Appendix B.4.2.7. Environmental matrices discussed here include biota as well as abiotic matrices, such as water (ground-, raw-, surface- and drinking water rain, snow, ice), suspended particulate matter, air, soil and sediment as well as wastewater influent, effluent and sludge. Concentrations of PFASs in humans, food, indoor air and dust are discussed B.9.21 and B.9.22. The current section of targeted analyses does not cover monitoring of precursors. Instead, occurrence of precursors in the environment is discussed in section B.4.1.3.5 and B.4.2.6.2. 9 https://www.norman-network.com/nds/susdat/, date of access: 2022-09-29. 91 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.2.7.1. ≥C4 PFCAs and ≥C4 PFSAs covered by existing or proposed restrictions PFASs covered by existing or proposed restrictions under REACH and/or the POPs Regulation include PFOA, PFOS, PFHxS, PFHxA and C9-C14 PFCAs, as well as their salts and related compounds. These PFAAs (especially PFOS and PFOA) have been thoroughly monitored and thus there is a wealth of data on their occurrence in the environment. Monitoring data for these compounds are partly presented in sections B.4.2.1-B.4.2.4 to substantiate the environmental fate properties of PFASs. In brief, PFOS and PFOA are generally the PFAAs detected at highest levels in soil, sediment and sludge (see B.4.2.3), whereas PFOS and C9-C14 PFCAs are dominating in biota (see B.4.2.9). The higher concentrations of the longer chain PFAAs in these matrices, in comparison to the shorter chain PFAAs, are partly due to differences in sorption and bioaccumulation, which is further discussed under B.4.2.1 and B.4.2.9. PFOA, PFOS, PFHxS and PFHxA are among the most abundant PFASs in aquatic compartments and can generally be found in the majority of analysed ground-, surfaceand tap water samples from various countries, which has been summarised by e.g. Sims et al. (2022). For further details, see B.4.2.1.3. In summary, despite the phase-out of PFOS and PFOA they are still abundantly detected worldwide, including Europe and remote regions, which illustrates that once released environmental contamination with PFASs is poorly reversable. Furthermore, the presence of PFAA precursors (including slowly degradable side-chain fluorinated polymers) will remain a long-lasting source of PFAAs in the environment. B.4.2.7.2. ≥C4 PFCAs and ≥C4 PFSAs not covered by existing or proposed restrictions C4-C7 PFCAs and PFSAs that are not included in existing or proposed restrictions include PFBA, PFPeA, PFHpA, PFBS, PFPeS and PFHpS. These compounds are found in surface water from the global oceans, rivers, and lakes (Muir and Miaz, 2021; Sims et al., 2022), groundwater, tap water and raw water (Kaboré et al., 2018; Llorca et al., 2012a; Sims et al., 2022), soil and sediment (Brusseau et al., 2020; Janda et al., 2019), the atmosphere (Wong et al., 2021), and in snow and ice of remote regions (Kwok et al., 2013; MacInnis et al., 2017). Furthermore, these compounds are found in influent, effluent and sludge of WWTPs (Lenka et al., 2021) and in landfill leachates (Hamid et al., 2018). This is further exemplified below. PFBA and PFBS are the dominating ≥C4 PFAAs in European and Arctic surface waters (Muir and Miaz, 2021; Pan et al., 2018; Zhao et al., 2015). In general, the levels of these compounds are in the same order of magnitude as those of PFOA and PFOS (Sims et al., 2022). In a recent review of the occurrence of PFAAs in global surface and ground water, PFBA, PFPeA, PFHpA, PFBS and PFPeS were each detected in 35-79% of the samples (Sims et al., 2022). Furthermore, ≥C9 PFCAs and PFSAs were present in lower concentrations compared to the shorter-chain homologues in both surface waters and groundwater. In a monitoring study of surface water from the lake Mälaren (Sweden), Thames River (UK) and Rhine River (Germany) the combined concentration of C4-C7 PFCAs and PFBS represented approximately half of the total PFAS content (Pan et al., 2018). In another example from the North Sea, PFBS and PFBA accounted for approximately 50% of the sum of PFASs (Zhao et al., 2015). Similarly, PFBA, PFPeA, PFHxA, PFHpA and PFBS accounted for approximately 60% of the sum of PFASs in coastal waters of the German Bight and Baltic Sea (Joerss et al., 2019). Poor correlation between PFBS/PFBA and other PFAAs in aquatic environments suggests that these compounds have other sources than longer chain PFAAs, which could reflect a different use pattern and/or environmental behaviour (Joerss et al., 2019; Zhao et al., 2015). For example, in a study investigating source-specific patterns of PFASs in river 92 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) samples up- and downstream potential industrial point -sources, PFBS represented 66-70% of the total PFAS in rivers downstream pharmaceutical and pesticide intermediate manufacturing sites, which is probably due to the use of PFBS in these industries (Joerss et al., 2020a). A survey of drinking water from several countries showed that PFBA, PFPeA, PFHpA and PFBS were each detected in 68-92% of the tap water and in 32-58% of the bottled water samples. The sum of these compounds constituted approximately half of the total PFAS concentration in the drinking water (Kaboré et al., 2018). The ubiquity of these short chain PFAAs in drinking water support that they are highly mobile and well distributed in the freshwater environment, even far away from point sources and that they are poorly removed by current drinking water treatment technologies (see B.4.2.1). In the atmosphere at locations far away from point -sources, PFBA has been shown to be the dominant ≥C4 PFAA (Fredricsson et al., 2021; Wong et al., 2018). In contrast to other studied PFAAs, the levels of PFBA in these remote areas were comparable to reported levels in urban environments, which indicates that PFBA is uniformly distributed in the global atmosphere (Wong et al., 2018). This could be explained by PFBA being more susceptible to long-range transport or/and that PFBA is being formed from fluorinated gases, such as HFE-7100, in the atmosphere (see B.4.1.3.2 and B.4.2.8). In the air directly above landfills, incineration plants, transfer stations and WWTPs, reported levels of PFASs (dominated by FTOHs) are considerably higher (up to 30 times) compared to reference sites (Ahrens et al., 2011; Tian et al., 2018; Wang et al., 2020b). PFBA was generally the dominant ≥C4 PFAA in the air above these sites, although the levels of PFOA or PFOS occasionally reached or even exceeded the PFBA concentration. In biota, the concentrations of shorter chained PFAAs are often below the reporting limits or, when detected, considerably lower compared to longer chained PFAAs. This is probably a consequence of their shorter biological half-life and lower potential to partition into protein and lipids (see B.4.2.9). For example, PFBS has been found in livers from arctic polar bears and killer whales, however in concentrations five orders of magnitude lower than those of PFOS (Gebbink et al., 2016). In soil, the shorter chain PFAAs are generally found at lower concentrations than longer chain PFAAs. However, whereas the longer PFAAs strongly dominate at the shallow depths, the shorter PFAAs become increasingly more abundant at the deeper depths. This is due to their higher mobility in soil compared to the longer PFAAs, which increases the risk for these substances to reach the water table over time (Brusseau et al., 2020). In the meantime, the soil is a significant reservoir for these and other PFASs. PFASs in wastewater influents and effluents have been reviewed by Lenka et al. (2021). For PFCAs, the sorption is expected to increase with carbon chain length and P FCAs with less than 10 carbon are therefore expected to be present in treated wastewater, whereas longer chain PFCAs and PFOS are expected to sorb on sewage sludge (Arvaniti et al., 2014). In general, the majority of studies demonstrate poor removal of most PFASs, which is demonstrated by the concentrations of PFASs generally being higher in the effluents after wastewater treatment, which can result in considerable releases into the environment (Lenka et al., 2021) (see also B.4.5). In a study of Nordic WWTP effluent samples, C4-C7 PFAAs contributed with 43% to the ΣPFAS, whereas the longer chained PFCAs and PFOS together accounted for 26% (Aro et al., 2021b). C4-C6 PFAAs were also the dominant class in WWTP influents and effluents in the U.S., contributing with 55 ± 19% of the ΣPFAS (Tavasoli et al., 2021). In contrast, longer chained PFAAs are relatively more abundant than shorter chained PFAAs in WWTP sludge, even though shorter chained PFAAs are also detected in these matrices (Aro et al., 2021b; Tavasoli et al., 2021). In addition to WWTPs, landfills constitute another point -source of PFASs. A review of PFASs in landfill leachate from several continents concluded that C4-C7 PFAAs were more abundant than their longer-chain (>C8) homologues in leachate (Hamid et al., 2018). 93 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The text above regards what is often referred to as short-chain PFCAs and PFSAs. However, there are also longer chained PFAAs that are not covered by existing or proposed restrictions and are sometimes included in target analysis of environmental samples. Examples of these are ≥C9 PFSAs, such as PFNS (C9) and PFDS (C10). In general, these compounds are of relatively low importance in aquatic samples. In conclusion, ≥C4 PFCAs and PFSAs that are not subject to existing or proposed restrictions constitute a large part of the total PFAA levels in European aquatic environments, including drinking water. Furthermore, PFBA is the dominating ≥C4 PFAA in the atmosphere. However, these compounds are found in lower levels in top-layers of soil, sludge and biota compared to the longer chain PFCAs and PFOS. B.4.2.7.3. C1-C3 PFCAs and PFSAs C2-C3 PFCAs (i.e. TFA, PFPrA) and C1-C3 PFSAs (i.e. TFMS, PFEtS and PFPrS), often referred to as ultrashort PFAAs, have not usually been included in monitoring studies, which could partly be explained by analytical limitations (Björnsdotter et al., 2020). Among these compounds, most monitoring and source information is available for TFA. Studies reporting detectable levels of TFA in deep ocean waters and in pre-industrial glacier ice and firn samples have indicated that there can be natural sources of TFA (Frank et al., 2002; Scott et al., 2005; von Sydow et al., 2000). However, whether non-anthropogenic sources of TFA exist is still debated (Joudan et al., 2021; Nielsen et al., 2001). In any case, natural sources will only apply for TFA in the oceans, while TFA in fresh water and in the terrestrial environment originates from anthropogenic sources. It is known that TFA is formed through atmospheric oxidation of some hydrofluorocarbons (HFCs), hydrochlorofluorocarbons (HCFCs) and hydrofluoroolefins (HFOs). TFA can also be formed during fluoropolymer thermolysis and through oxidation of precursor compounds, such as other PFASs as well as pharmaceuticals or pesticides containing a carbon-bound trifluoromethyl moiety (Cui et al., 2019; Freeling et al., 2020). Industries that manufacture TFA or use TFA in the production of e.g. plant protection agents are also potential point sources (Scheurer et al., 2017). Information about the sources of the other ultrashort PFAAs is limited, but PFPrA could potentially arise from impurities and byproducts in either TFA or PFBAchemistry,whereas PFEts and PFPrS have been found in AFFFs and can be present as a residual formulation and/or by-product from products manufactured by electrochemical fluorination (Björnsdotter et al., 2020). Two Swedish studies have investigated the presence of C1-C3 PFAAs in waters close to potential point sources, such as airports, landfills and waste management facilities (Björnsdotter et al., 2019; Ericson Jogsten and Yeung, 2017). In the study by Björnsdotter, concentrations of ΣC1−C3 ranged up to 84 000 ng/L and represented more than 30% of the ΣPFAS in 20 out of 32 analysed samples (Björnsdotter et al., 2019). The Ericson study, which only analysed PFEtS and PFPrS, found that all samples contained at least one of these compounds and that PFPrS contributed with 0.4 to 17% to the ΣPFASs in these samples. In addition, leachate from seven municipal solid waste disposal facilities in Tianjin, China showed that TFA contributed with 52-68% to total PFAS concentration in leachate from incineration plants and transfer stations, whereas PFPrA was dominant (4983% of ΣPFAS) in leachate from landfills (Wang et al., 2020a). C1-C3 PFAAs have also been found in aquatic matrices without known historic PFAS contamination. For example, in two Swedish drinking water samples without known contamination PFPrA, TFMS and PFPrS were detected in levels up to 28, 7.8 and 0.18 ng/L, respectively, which resulted in a combined contribution of 43% and 84% to the ΣPFAS (Björnsdotter et al., 2019). In a study of US bottled drinking water, PFPrA was detected in 22 out of 101 samples and accounted for 42% of the total PFAS content on a mass basis, whereas PFPrS contributed with less than 1% (Chow et al., 2021). In a multinational study of drinking water, PFPrS and/or PFEtS were detected in samples from Burkina Faso, Ivory Coast and China but not from Norway, Chile, Canada and USA (Kaboré et al., 2018). The drinking water studies above did not include TFA in their analyses. However, in a recent 94 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) study of direct or indirect source water for drinking water production, TFA was found at a median concentration of 900 ng/L, representing more than 90% of the total PFAS content (Neuwald et al., 2022). In addition, TFMS, PFPrA, and PFPrS were detected in more than 80% of these samples. In another study of Danish ground water, TFA was detected in 89% of the 247 analysed samples. The TFA levels were mostly below 1 000 ng/L but concentrations up to 2 400 ng/L were detected (Miljøstyrelsen, 2020). TFA mainly enters the terrestrial environment via precipitation, where it has been reported to be the most abundant known PFAS. For example, in a nation-wide German study including 1 187 wet deposition samples taken over a year (2018-2019), the precipitationweighted average of TFA was 335 ng/L (Freeling et al., 2020). Higher levels were observed in the summer, which was suggested to result from higher concentrations of photochemically generated oxidants (e.g. hydroxyl radicals) in the troposphere, which enhances transformation of TFA precursors. In a study of precipitation from mainland China, TFA was found in concentrations of 9-1 800 ng/L, which was 1-2 orders of magnitude higher than those of any other individual PFAA. The sum of TFA and PFPrA accounted for 22-91% of the total PFASs in these samples (Chen et al., 2019a). Another Chinese study reported average TFA levels of 643±169 and 282±68 ng/L in Beijing urban landscape water and snow, respectively (Zhai et al., 2015). C2-C3 PFAAs have also been detected in WWTP effluents. In a Nordic study, PFEtS, PFPrS and PFPrA were found in all WWTP effluent (dissolved + particle phase) samples and accounted for 6% of the total PFAS concentration in the samples from Finland, Sweden, Denmark, Norway, and Faroe Islands (Kärrman et al., 2019). In Greenland and Norway they were the most abundant class of PFAAs with a mean contribution of 39%. In another study of fluorine mass balance analysis of Nordic WWTP effluent samples, C2-C3 PFAAs was the group with the largest contribution (on average 35%) to the identified extractable organofluorine (Aro et al., 2021b). In 144 air samples collected during a year (2012-13) in Beijing, China, the annual mean concentration of TFA was 1 580±558 pg/m3 (Wu et al., 2014). In addition, TFA was the dominant PFAA in the air above Chinese landfills and constituted 80% of the ΣPFCA, whereas the levels of PFPrA were in the same range as PFBA (i.e. the most abundant ≥4C PFAA). TFA was also the most abundant PFAA in plant leaves and dry deposition the at these sites (Tian et al., 2018). In a study of the Nordic environment, C2-C3 PFAAs were not detected in any of the biota samples including bird eggs, fish and terrestrial and marine mammals, which probably reflects their relatively low bioaccumulation potential (Kärrman et al., 2019). In conclusion, TFA has been detected at high levels in aquatic environments (including drinking water) and precipitation worldwide, whereas there is less information about other C1-C3 PFAAs. The few available studies reveal that C1-C3 PFAAs are being released into the environment and constitute a considerable fraction of the total content of known PFASs in aquatic matrices, including drinking water. The presence of these compounds in drinking water also reflects the difficulties to remove them with common water treatment methods (see B.4.5). B.4.2.7.4. Cyclic PFAAs Perfluoro-4-ethylcyclohexanesulfonate (PFECHS), which is a cyclic C8-PFSA, is the most monitored cyclic PFAA. It has been used as an erosion inhibitor in aircraft hydraulic fluids (De Silva et al., 2011) and could also be present as an impurity in POSF -based products, such as AFFFs (MacInnis et al., 2017). PFECHS has been detected, together with other PFASs, in water, sediment and biota in the vicinity of airports in China, Canada and the Nordic countries, probably as a consequence of its use in aircraft hydraulic fluids (de Solla et al., 2012; Kärrman et al., 2019; Lescord 95 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) et al., 2015; Wang et al., 2016a). Furthermore, PFECHS has been found in Chinese coastal water and sediment, for which the primary source was suggested to be erosion inhibitor factories (Liu et al., 2019d). PFECHS has also been found in Nordic WWTP effluent samples, which indicates that consumer products, other than aircraft hydraulic fluids, may be source of PFECHS (Kärrman et al., 2019). PFECHS has been detected in all surface water samples from the North American Great Lakes (De Silva et al., 2011) and in 86% of Baltic Sea coastal water samples, but in none of the water samples from the German Bight (Joerss et al., 2019). In the Baltic Sea, levels of PFECHS were in the same magnitude as those of PFOS which indicates that a direct source (e.g. hydraulic fluids) is more likely than PFECHS being an impurity in POSF -based products (Joerss et al., 2019). PFECHS has also been detected in tap water samples from 2 source locations from the Great Lakes but not in other analysed samples from other countries included in a study by Kaboré et al. (2018). PFECHS, together with other PFASs, has also been detected on the Devon Ice Cap (1 800 m above sea level) (MacInnis et al., 2017). The presence of PFASs in the ice cap is suggested by the authors to be attributable to atmospheric deposition due to the high altitude, which source could be the direct emission due to leakage of the compound from aircraft during usage. In biota, PFECHS has been detected in polar bears in the Canadian Arctic (Letcher et al., 2018) and in marine mammals from the Denmark and Greenland (Kärrman et al., 2019), however in concentrations several orders of magnitude lower compared to PFOS. Furthermore, PFECHS has been detected in herring gull eggs and in top predator fish in the Great Lakes of Canada (De Silva et al., 2011; Letcher et al., 2015), but not in bird eggs in a study of the Nordic environment (Kärrman et al., 2019). The Dossier Submitters have found few monitoring studies on other cyclic PFAAs. Downstream of Beijing international airport, perfluoropropylcyclopentanesulfonate (PFPCPeS) was found in surface water, sediment and Crucian carp but at lower detection frequencies and levels than those of PFECHS (Wang et al., 2016a). None of these compounds were found in the aquatic reference samples. In another study of fish from the Great Lakes of Canada, perfluoromethylcyclohexane sulfonate (PFMeCHS), was detected in the dissolved phase but not above the detection limits in fish tissue (De Silva et al., 2011). In conclusion, PFECHS has been detected in aquatic environments, especially in connection to aviation activities. PFECHS has also been found in low levels in biota and ice of remote areas. Monitoring of other cyclic PFAAs is scarce. B.4.2.7.5. Perfluoroalkyl phosphonic acids (PFPAs) Perfluoroalkyl phosphonic acids (PFPAs) have been included only in a limited number of studies, where they generally have been detected at lower levels and detection frequencies than frequently analysed PFCAs and PFSAs (Xiao, 2017). For example, PFPAs have been detected in surface water in the Netherlands (Esparza et al., 2011) and Canada (D'eon et al., 2009), in Spanish and German tap water and in German WWTP effluent water (Llorca et al., 2012a). The lower concentrations and detection frequencies could be explained by a lower global historical use and releases of these substances compared to PFCAs and PFSAs (Wang et al., 2016b). B.4.2.7.6. Perfluoroalkyl ethers (PFAEs) The most studied PFAEs to date are HFPO-DA, ADONA and Cl-PFESAs. Other PFAEs have rarely been measured with targeted (semi-)quantitative analyses in environmental samples. Further insight on the occurrence of additional PFAEs can be gained by applying non-target or suspect screening methods which are further discussed in B.4.2.6.3. 96 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) HFPO-DA has been found in surface waterbodies, including the marine environment in Europe (Heydebreck et al., 2015; Pan et al., 2018), in the US (Sun et al., 2016) and China (Pan et al., 2017). For example, HFPO-DA was detected in all surface water samples from the lake Mälaren (Sweden), Rhine River (Germany) and Thames River (UK), which showed occurrence in European waters not limited to fluorochemical industry zones (Pan et al., 2018). Furthermore, the HFPO trimer acid (HFPO-TrA) was detected in 73% of these samples but in levels generally lower than those of HFPO-DA. In contrast, in the highly contaminated Xiaoqing River Estuary, China, HFPO-TrA was found at higher detection frequencies and concentrations in surface water, invertebrates and fish compared to HFPODA (Li et al., 2021e). In a study by Joerss et al. (2019), the levels of HFPO-DA in the German Bight coastal water were three times higher than those of PFOA and contributed with 27% to the ΣPFAS. On the contrary, the same study reported negligible levels of HFPO-DA in the Baltic Sea, probably reflecting the lower water exchange and fewer pointsources. HFPO-DA has also been detected in Arctic seawater (albeit in considerably lower levels than in European coastal waters) and snow (Joerss et al., 2020b; Xie et al., 2020b). The occurrence of HFPO-DA has also been studied in the vicinity of fluorochemical production plants in Europe. For example, HFPO-DA was detected in all samples downstream from a fluorochemical production plant in Dordrecht, the Netherlands, with the highest level (812 ng/L) being 13 times higher than the sum of PFAAs. In contrast, the levels were below the reporting limit at the control sites and in two out of three samples taken upstream from the plant. HFPO-DA was also found in tap water within 50 km from the fluoropolymer manufacturing plant (Brandsma et al., 2019; Gebbink et al., 2017). In the same area, HFPO-DA was detected in all grass and leave samples collected within 3 km from the plant (Brandsma et al., 2019), but not in a reference area 85 km away. This indicates that locally grown foods may contain HFPO-DA, which was confirmed by Mengelers et al. (2018) who showed that HFPO-DA could be detected in 40% of 74 vegetable samples within 4 km from the plant. Both in leaves/grass and river water, there was a declining concentration gradient in relation to the distance from the plant, which further strengthens that the manufacturing plant was the source of HFPO-DA to the adjacent environment. Another example of the impact of point-sources on environmental HFPO-DA levels was investigated by Joerss et al. (2020a), who found detectable levels in all 29 German river samples up- and downstream potential point -sources. Levels were particularly high downstream from a German fluoropolymer manufacturing site, where HFPO-DA constituted approximately 50% of the total PFASs in the samples. In a study by Pan et al. (2018), ADONA was found only in the German Rhine River but not in other surface waterbodies in Europe, US and China. Furthermore, in another study by Joerss et al. (2020a) ADONA was detected in 73% of the analysed German river samples up- and downstream potential point -sources, . The levels where markedly high (up to 2 500 ng/L) downstream from a German fluoropolymer manufacturing site where ADONA constituted approximately 40% of the total PFASs. In a study of the Nordic environment, ADONA was not detected in either water or biological samples (egg birds, fish and marine and terrestrial mammals) in areas without known contamination (Kärrman et al., 2019). China is the only known emission source of 6:2 Cl-PFESA, where its ubiquitously present in the environment (Pan et al., 2018). In European surface waters, 6:2 Cl-PFESA has either been below the reporting limit (Gebbink et al., 2017; Joerss et al., 2020a) or detected at low levels (Pan et al., 2018). 6:2 Cl-PFESA has also been detected in Greenland marine mammals (Gebbink et al., 2016). These data strongly suggest that 6:2 Cl-PFESA can undergo long-range transport to the Arctic and has the potential for global distribution and to a certain degree to bioaccumulate. Several PFECAs have been detected in the Cape Fear River in the US downstream from a fluorochemical manufacturer plant (Strynar et al., 2015; Sun et al., 2016) where a sulfonated tetrafluoroethylene based fluoropolymer-copolymer was produced. Due to the 97 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) lack of authentic standards, the concentration of these PFECAs could not be quantified. Based on the comparison of chromatographic peak areas Sun et al. (2016) determined that the dominant PFECA was perfluoro-2-methoxyacetic acid (PFMOAA), with a peak area ~100 times that of HFPO-DA, followed by perfluoro-3,5-dioxahexanoic acid (PFO2HxA; peak area ~25 times that of HFPO-DA) and perfluoro-3,5,7-trioxaoctanoic acid (PFO3OA; peak area ~5 times that of HFPO-DA). Hopkins et al. (2018) estimated the added concentration of PFAEs in the Cape Fear River to be 40 000 ng/L with contributions from PFMOAA >> PFO2HxA > PFO3OA ~ HFPO-DA > Nafion byproduct 2 > PFO4DA. When the manufacturer began to capture the process wastewater, the sum concentration of targeted PFAEs in the river dropped sharply to an estimated concentration of approximately 4 200 ng/L. Furthermore, analysis of water samples taken at the water treatment plant located 80 km downstream from the industry showed similar levels of PFAEs in the ingoing raw water and the treated drinking water, which demonstrates the difficulties to remove these compounds from drinking water. Studies of increased blood levels of these PFAEs in the population consuming the contaminated drinking water are further discussed in B.9.21.2. Song et al. (2018c) identified the presence of several PFAEs in the Xiaoqing River, which receives water discharge from one of the major fluoropolymer manufacturing facilities in China. They identified a total of 42 PFASs, including 3 HFPO oligomers (dimer acid, DA; trimer acid, TA and tetramer acid, TeA) and numerous tentatively detected isomers of C9−C14 PFECAs. The water concentrations of HFPO-TrA and HFPO-DA were approximately 1 and 2 orders of magnitude lower than the concentrations of PFOA. Besides the HFPO oligomers, monoether PFECAs and polyether PFECAs were the most abundant emerging PFASs, whereas the levels of polyhydrogen-substituted PFCAs, monohydrogen-substituted PFCAs and monochlorine-substituted perfluoroalkyl carboxylic acids were significantly lower. In conclusion, studies clearly demonstrate significant releases of well-known as well as less studied PFAEs from fluorochemical industries resulting in high levels of these compounds in the adjacent environments, including drinking water. Furthermore, there are evidence of PFAEs also being found in water environments without known point -sources. B.4.2.7.7. Fluorinated Gases Atmospheric levels of fluorinated gases are routinely monitored for example in the Advanced Global Atmospheric Gases Experiment (AGAGE, 2022). Time trends from these monitoring programs are presented B.4.2.7.9. B.4.2.7.8. Polymeric PFASs Polymeric PFASs (including fluoropolymers, perfluoropolyethers and side -chain fluorinated polymers (SCFP)) can generally not be quantified by standard analytical methods used for low-molecular weight non-polymeric PFASs as reference standards are not available and the methods are unsuitable (see E.4.). However, there are methods that can be used to determine the type of polymer, the molecular weight and the layer thickness of the polymer (NCM, 2022). Furthermore, some total fluorine methods can include fluorine originating from polymeric PFASs (although polymeric PFASs will be lost if extraction methods are applied, in e.g. EOF analysis). Furthermore, TOPA can to a certain degree be applied to side-chain fluorinated polymers to cleave off and release PFAAs, which can subsequently be measured by conventional analytical methods (NCM, 2022). Using mass spectrometry, two side-chain fluorinated polymers have been quantitatively analysed. The active components of the common fabric protector Scotchguard™, one perfluorooctane sulfonamide-based urethan side-chain polymer used until 2002 and perfluorobutane sulfonate-based urethan side-chain polymer used after 2002, have been detected in several matrices including aquatic sediment and agricultural soil (Chu and Letcher, 2017) as well as biosolids from WWTPs in the US (Letcher et al., 2020) and in 98 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) WWTP sludge and landfill leachates in Sweden (Fredriksson F et al., 2020). At all sites, the levels of these side-chain fluorinated polymers greatly exceeded the sum of all other individual non-polymeric PFASs (22-83 analysed by targeted analysis in the different studies). side-chain fluorinated polymers in the environment is the detection of the fluorinated siloxanes D3F and D4F (which are degradation products of and impurities in silicone side-chain fluorinated polymers) close to a fluorinated methylsiloxane manufacturing plant in China. The concentrations of D3F and D4F in surface water near the plant ranged between 3.3-291 ng/L and 7.0-168 ng/L, respectively, and in sediment between 12-5 478 ng/g and 17-6 277 ng/g, respectively (Zhi et al., 2018). D3F and D4F have also been detected in sediment, sewage sludge and WWTP effluent in Europe (McLachlan et al., 2014; van Bavel B. et al., 2016). According to a review of Lohmann et al. (2020), incomplete polymerization during the synthesis of fluoropolymers may result in residual monomers, oligomers and smaller “polymers” with up to about 100 monomer units, which may be released to air upon heating during manufacturing and processing and to water through wastewater streams. In one example, a series of polyfluorinated carboxylic acids differing by one 1,1-difluoroethene (CF 2CH2) unit, which were likely used as building blocks for production of PVDF, were found close to a manufacturing facility in the US (Lohmann et al., 2020). In another example, PFAS oligomers were detected in stack emission samples at a fluorochemical production site. Fluoropolymer microparticles have been found in the environment and biota, even in remote areas, which is further discussed in B.7.5. In addition to the release of polymers to the environment, several examples of the release of other PFASs (such as HFPO-DA and PFOA) from fluoropolymer production facilities are given in the sections above. B.4.2.7.9. Environmental time trends Temporal trend analysis of PFASs in the environment is challenging due to the statistical power of individual studies often being too low to identify significant trends. Furthermore, comparison of different studies is hampered by the use of different study designs including geographical regions, time intervals, sampled matrices and analytical as well as statistical methods. For example, recent changes in the environment may be masked in longer time series (sometimes several decades), unless the time trends before and after identified change-points (i.e. a timepoint when the trend changes direction) are assessed separately. It is important to mention that phased-out PFASs that may show declining trends in Europe are not disappearing on a global scale due to their potential for long-range transport and persistency in various compartments, which is further discussed in B.4.2.8. Time trends for PFASs covered by existing or proposed restrictions Global environmental trends for PFASs have been reviewed by Land et al. (2018). Overall, they concluded that while there are clear declining trends for PFOS and PFOA in humans (see B.9.22.5), the trends in biota are mixed and often not significant (which can be influenced by biota being more heterogenous). For example, monitoring of biota from North America, Europe and the Arctic revealed no clear trend for PFSAs (i.e. PFOS, its precursor FOSA and PFHxS) or PFOA, whereas longer chained PFCAs (PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA and PFTeDA) showed increasing or insignificant temporal trends. Another review of time series of Arctic biota, which included the 1980s/1990s, generally found increasing PFOS/PFSA trends but these seem to be levelling off from the mid-2000s, which becomes apparent when trend analysis only considers more recent years (Muir et al., 2019). Similar results were obtained for PFCAs, but the decline of certain PFCAs in biota seems to be delayed and weaker compared to the trends for PFOS. In individual studies, increasing PFOS (and PFCA) concentrations have been reported for biota living in contaminated 99 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) regions with slow water exchange, such as the Baltic Sea (Faxneld et al., 2016; Huber et al., 2012; Roos et al., 2013; Schultes et al., 2020b; Sun et al., 2019). In summary, the results indicate that the phase-out of PFOA and PFOS did not result in clear declining trends in biota on a global scale yet, which is likely attributable to their high persistence. Regarding time trends of PFASs in abiotic matrices, the review by Land et al. (2018) found declining trends for PFOS and PFOA in surface waters, whereas most studies on sediment cores reported increasing concentrations. The authors concluded that the PFOS and PFOA restrictions resulted in decreasing concentrations especially in surface waters with high water exchange, whereas remote areas might have a delayed response to regulatory measures. A recent review by Muir and Miaz (2021) compiled data on global temporal trends of PFASs in ocean and coastal waters. Significantly declining trends were observed for ΣC7-C12 PFCAs in recent years in the North Sea and Baltic Sea (2015-2017) and in the North American Great Lakes (2004-2017). For the Mediterranean Sea, PFOA and PFOS showed significantly declining trends in 2012-2018. Furthermore, mass discharge estimates indicated continued emissions of C7-C12 PFCAs to the Chinese Coasts in recent years, whereas the European riverine discharges have been reduced. In Germany, levels of commonly targeted PFASs in suspended particulate matter (SPM) from the main river systems showed declining trends between 2005-2019, whereas the decline was less pronounced after applying the direct total oxidizable precursors (dTOP) assay. This indicates that declining trends of legacy PFASs may be overestimated when precursors are not taken into account (Gockener et al., 2022). Analyses of European wastewater sludge have shown decreasing levels of PFOA and PFOS over time (Alder and van der Voet, 2015; Ulrich et al., 2016). For example, in a study of 4 981 sludge samples collected between 2008-2013 from 1 165 German WWTPs, the estimated load of PFAA in the sewage sludge decreased by 90% over the study period (Ulrich et al., 2016). In conclusion, in contrast to clear decreasing trends of PFOA and PFOS in humans, the trends of these compounds in biota are conflicting. In addition, there is indication of potentially increasing levels of longer chain PFCAs in biota from contaminated regions with slow water exchange. Regarding time trends in aquatic environments, the levels of PFOS and PFOA seem to be decreasing in European and North American coastal, sea and river waters. Time trends for PFASs not covered by existing or proposed restrictions While there are numerous time trend studies for the most studied PFAAs, such as PFOA and PFOS, there is very limited time trend information available for the majority of PFASs, including precursors, PFAEs, PFPAs and shorter chain PFCAs and PFSAs. In the review of temporal trends by Land et al. (2018), no firm conclusions could be drawn for non-restricted PFAAs in biota as these compounds are rarely analysed in biota and often are close to or below the detection limits. For PFBA (C4 PFCA) and PFHpA (C7 PFCA), there were mostly insignificant trends. For PFDS (C10 PFSA), there was a tendency towards increasing trends in mammals, decreasing trends in fish and mixed trends for birds. Some studies have indicated a shift from longer chain to shorter chain PFAAs over time. For example, a recent study of temporal trends in Canadian beluga whale livers showed that unregulated C4-C7 PFASs, H-PFCAs, and odd-chain FTCAs increased between 2000 and 2017, whereas longer chain PFASs significantly decreased over time (Barrett et al., 2021). In another example, time trends of PFAAs in suspended particulate matter (SPM) 100 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) from German main river systems were examined with direct TOP analysis (dTOP) 10, which accounts for precursors (Gockener et al., 2022). There were no significant trends for C4C7 PFCAs between 2005-2019 but declining trends for C8-C18 PFCAs and PFSAs (Figure B.60). In a study by Pickard et al. (2020) PFBA increased in ice cores from the Canadian Arctic between approximately 1990 and 2015 (Figure B.60). In concordance, Kirchgeorg et al. (2013) reported that PFBA was the most abundant ≥C4 PFAA in a firn core sample from the European Alps and that the levels significantly increased between the years 1997 and 2007. At the same time, decreasing trends were seen for PFOA and PFHpA. Figure B.60. Time trends (percent per year) in suspended particulate matter (2005-2019) sampled in the main German rivers and important tributaries (Danube, Elbe, Mulde, Rhine, Saale, Saar) presented by Gockener et al. (2022). Samples were analysed by direct Total Oxidizable Precursor (dTOP) Assay. Trend analysis took variations in river system and total organic carbon content into account. n.s = non-significant trend. In the arctic atmosphere, the concentration of PFBA was reported to be 1 to 2 orders of magnitude higher than that of other individual PFAAs and an increasing trend of PFBA with a doubling time of 2.5 y was reported at the Alert station in Canada between 2006 and 2014 (Wong et al., 2018). Increasing concentrations were also reported for PFOA, PFOS, and FTOHs at the Alert station with doubling times of 3.7, 2.9 and 5-7 years, respectively. In contrast, decreasing or non-changing levels of PFOA and PFOS, with half-times of 1.97.5 and 11-67 years, respectively, were reported from two stations in Northern Norway (PFBA was not monitored at these sites) (Wong et al., 2018). There are few time trend studies of C1-C3 PFAAs. Pickard et al. (2020) studied atmospheric deposition of TFA and PFPrA using two Arctic ice cores. They concluded that the deposition fluxes for these compounds have increased since approximately 1990. The mean fluxes of TFA were almost one order of magnitude higher in the post -2000 samples, compared to the pre-2000 samples (Figure B.61). The authors concluded that the trend is in consistency with the production rate of fluorinated gases (HCFCs and HFCs) that degrade to TFA, which suggests that these are the major sources to TFA in arctic ice cores. 10 The dTOP assay is a modification of the TOP assay in which the sample is directly and fully oxidized in a more concentrated oxidation solution compared to the original method where extracts are oxidized. 101 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.61. Five‐year moving average deposition fluxes of TFA, PFPrA, and PFBA from Devon Ice Cap (solid lines) and Mt. Oxford icefield (dashed lines) ice cores (taken from Pickard et al. (2020)). In studies of TFA in German precipitation, the precipitation-weighted average concentrations in 2018-2019 (Freeling et al., 2020) were 3 to 4 times higher compared to 1995-1996 (Klein, 1997). These results indicate a substantial increase of wet -deposited TFA during the past two decades in the German environment. An even stronger increase was observed for urban landscape waters in Beijing, China, where TFA concentration showed up to 17-fold increase between 2002-2012 (Zhai et al., 2015). An increase in TFA concentrations has not only been observed in abiotic matrices but also in leaves of European beech in Germany between 1989-2020 (Figure B.62) and in leaves of black poplar (UBA, 2021b). A positive trend was also indicated in coniferous shoots from the European spruce and in shoots from pines (Freeling et al., 2022). Based on these results, the authors concluded that TFA concentrations continue to increase in Germany, which may result in an increased exposure to the wider environment including other biota. 102 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.62. Time trends of trifluoroacetic acid (TFA) in leaves (µg/g dry weight) of the European beech from four locations in Germany (taken from Freeling et al. (2022)). The Zeppelin station in Svalbard represents a background site for global climate gas monitoring and is used to study t he transport of polluted air episodes. At this location, increasing trends for the majority of the 13 investigated fluorinated gases in air were reported (NILU, 2021). The results demonstrate that four gases, namely HFC-125, HFC134a, HFC-152a, and HCFC-141b, showed increasing trends from 2001-2020 ranging between 0.4–5.19 ppt/y (see Figure B.63 for HFC-134a). The other nine gases were analysed between 2010-2020 and eight of these (HFC-32, HFC-43-10mee, HFC-143a, HFC227ea, HFC-236fa, HFC-245fa, HFC365mfc, and HFC-23) showed increasing trends between 0.01-2.31 ppt/y. The only fluorinated gas showing a declining trend was HCFC124 with -0.056 ppt/y between 2010-2020 (NILU, 2021). Similar results were reported globally under the Advanced Global Atmospheric Gases Experiment (AGAGE), which includes monitoring data from coastal or mountain sites in Norway (the Zeppelin station mentioned above), Italy, Ireland, Switzerland, American Samoa, Barbados, the US, and Australia (AGAGE, 2022). Taken together, the results available for fluorinated gases show overall increasing concentrations in air. 103 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.63. Time trend of HFC-134a (+5.19 ppt per year) in air sampled at the Zeppelin station in Svalbard between 2001-2020 in ppt (NILU, 2021). In conclusion, time trend studies clearly demonstrate steadily increasing levels of fluorinated gases increasingly used after the implementation of the Montreal Protocol in 1989. Concurrently, a clear increase of TFA in air, precipitation, ice cores and plants has been observed, which is likely a result of the growing abundance of TFA-yielding gases. In addition, analyses of ice/firn cores show increasing levels of PFBA and PFPrA over time. Regarding non-restricted PFAAs in biota, no firm conclusions can be drawn due to few existing time trend studies combined with low concentrations and detections frequencies. However, limited environmental evidence from individual studies of different environmental matrices suggests a shift from longer chain to shorter chain PFAAs in the environment. However, more evidence is needed to draw clear conclusions. B.4.2.7.10. Conclusions for environmental monitoring Studies on EOF/AOF and mass balance analysis show that varying but significant fractions of organofluorine in environmental samples are unknown and are therefore not captured by monitoring using only targeted PFAS analysis. There are indications that higher trophic levels display a lower fraction of unknown organofluorine, possibly due to metabolism of precursors into the stable PFAAs which are often included in the targeted analyses. Studies of abiotic and biotic samples utilizing TOPA demonstrate that considerable fractions of PFASs in the samples may be comprised of unknown oxidizable PFASs that are not detected in routine target analyses, including precursors to C1-C3 PFCAs that are rarely analyzed. Thus, environmental samples commonly contain PFASs with unknown identity and regulatory status and the total PFAS mass balances may be underestimated without treatment with TOPA. Non-target and suspect screening methods have been applied with the aim to identify the compounds constituting the unknown organofluorine fraction. These screening methods have (tentatively) identified hundreds of different PFASs in various environmental samples. These substances would go undetected if only targeted analyses with available reference standards were performed. However, the suspect and non-target screening can only 104 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) provide qualitative and semi-quantitative data of substances, and not quantitative, giving limited value to mass balance calculations. Although the information above clearly demonstrates that targeted analyses of individual PFASs do not provide the full picture of PFASs contamination, the vast majority of studies have applied such methods. While most of these studies have analysed a limited number of compounds, limited data are available on the occurrence and/or concentration of other PFASs, such as PFAEs and C1-C3 PFAAs. PFASs are ubiquitously found in European environments and biota. Furthermore, numerous examples of drinking water highly contaminated from different types of point-sources have been reported and many more cases are likely to go undetected. The sum of shorter chain PFAAs have been shown to often account for a major part of the total known PFAS content in water samples, including drinking water. In the light of the high persistence of these non-restricted compounds, their high mobility, low adsorption to organic carbon and the difficulty to remove them from wat er (section B.4.2.2), the concentrations of these compounds will increase if emissions of these compounds and/or their precursors to the environment continue. Targeted monitoring studies of PFASs in environmental matrices show that PFOS and PFOA, which are restricted, still are the dominating PFASs in soil, sediment, sludge and biota, and among the most abundant PFAAs in aquatic environments. Thus, despite the phase-out of PFOS and PFOA, they are still detected in high levels worldwide, which illustrates that contamination of PFASs is poorly reversable and underpins the need to ban also other PFASs to avoid similar problems in the future. Furthermore, the presence of precursors, such as side-chain fluorinated polymers, will remain a long-lasting source of PFAAs in the environment even after a phase-out of production. The bans of PFOS and PFOA have resulted in a transition to other PFASs, such as shorter chain PFAAs and PFAEs. For example, HFPO-DA is widely detected in the European environment, whereas 6:2 Cl-PFESA is found in high levels in China, but currently not in Europe. Besides these most studied PFAEs, studies have clearly showed the presence of other, sometimes even more abundant PFAEs in the vicinity of fluorochemical industries. However, little is known about the environmental levels of these and other PFAEs. Furthermore, very little is known about the levels of polymeric PFASs in the environment. Whereas time trends of PFOA and PFOS are decreasing in humans, the trends in biota are inconsistent. Regarding time trends in aquatic environments, the levels of PFOS and PFOA seem to be decreasing in European and North American coastal, sea and river waters. However, it has to be noted that phased-out PFASs that may show declining trends locally are not disappearing on a global scale due to their potential for long-range transport and persistence in various compartments. For most PFASs, including PFAEs and shorter chain PFAAs, there is limited or no temporal trend data. The clearest upgoing time trend is observed for the fluorinated gases that have been increasingly used after the implementation of the Montreal Protocol. A simultaneous increase of TFA in air, precipitation and plants is likely a result of the increase of TFA-yielding gases. In addition, analyses of ice/firn cores show increasing atmospheric deposition of TFA, PFPrA and PFBA over time. B.4.2.8. Long-range transport potential The potential for environmental long-range transport (LRTP) is one major concern for persistent pollutants. By long range transport a shift of potential risks occurs off of the point of emission and often time-delayed. According to the OECD definition, long-range transport (LRT) refers to the transport of substances within the moving mass to locations distant from its sources (mainly for a distance greater than 100 kilometres). LRTP is indicated, if these substances are measured 105 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) in distant locations in concentrations that are of “potential concern”. The moving mass could be air, water or particles, as discussed below. B.4.2.8.1. Transport pathways As outlined below contamination is not geographically limited, so PFASs are found ubiquitously in the environment. Because of non-degradability, the movement of their carriers leads to global drift of PFASs over long distances from the point of release. Depending on their specific physiochemical properties PFASs distribute between the respective compartments. Three main hypotheses are currently proposed f or the global transport of PFASs. - - Non-charged, volatile precursor compounds could undergo long-range atmospheric transport and be degraded to persistent arrowhead PFASs being deposited via wet or dry atmospheric deposition in atmosphere and reaching remot e areas (Ellis et al., 2004; Martin et al., 2006; Schenker et al., 2008; Wong et al., 2018). ionic and water soluble PFASs could be transported directly by river waters into estuaries and coastal waters. additionally, PFASs can be transported by particles to which they are adsorbed or absorbed, such as dust, sediments, or through matrices in which they are included as additive, e.g. in polymers. Moreover, a long-range transport of PFASs may occur by biota e.g. migratory birds. Further, due to complex interactions between a substance and the compartments and a broad variety of environmental conditions the transport into remote areas is not limited to a single pathway. PFASs distribute to different compartments depending on their physicochemical properties, as well as the environmental conditions (e.g. temperature or pH). By changing environmental conditions, a substance-shift between the compartments may occur. Most existing studies focus on legacy PFASs but emerging and novel per- and polyfluoroalkyl substances currently come into research focus (e.g. Gomis et al. (2015)). This will narrow the gap of knowledge regarding other PFASs. Air and water Generally, charged short-chain PFASs have a higher LRTP in aquatic environments (Muir et al., 2019). The physicochemical characteristics also influence the type of long-range transport in the aqueous environment, e.g. sea spray, microlayer, surface water, deep ocean water (Ahrens et al., 2010a; Ahrens et al., 2010c). Early modelling studies indicated that PFAAs are more likely to be transported via oceanic currents than by the atmosphere (Armitage et al., 2006; Wania, 2007) but a recent study by Yeung et al. (2017) suggested that atmospheric input accounts for the majority of measured PFOA in the Arctic Ocean. Global transport by marine ocean currents was indicated as the major pathway of PFASs delivery to non-emission regions by both monitoring and modelling results (Stemmler and Lammel, 2010; Yeung et al., 2017) though the single processes are not yet fully understood. PFASs are globally distributed in the marine environment (Yamashita et al., 2011). The movement of PFASs, from coastal areas that are influenced by urban emissions, to subArctic and Arctic Ocean waters, was illustrated by Ahrens et al. (2010c), who found C6–C10 PFCAs averaging about 700 pg/L in coastal seawater of southern Norway and at detection limits (∼10 pg/L) in the open Norwegian Sea. As well industrial coastal areas as river based large industrial metropolitan areas and atmospheric deposition are considered as main sources of PFASs in sea water. Ocean currents and related dilution effects have a crucial 106 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) influence on global PFAS distribution in marine ecosystems (Ahrens et al., 2010a). Above that, ocean waters are important sinks for PFASs (Lohmann et al., 2013; Yeung et al., 2017) The transport of PFASs with water also could be time delayed. PFAS-loaded river water often runs not directly into oceans, but into more or less enclosed estuaries and adjacent seas, like into the North Sea, the Baltic Sea or the Mediterranean. Despite the North Sea having a wide-open connection to the Atlantic Ocean, the exchange between both seas is partially limited due to the morphology of the North Sea. The water circulates in the shallow North Sea continental shelf and the formation of thermo- and haloclines further hampers the water exchange. It is assumed that the North Sea water is totally exchanged within one or two years (Gyory J., 2013). The trapping effect of persistent pollutants is much higher in areas with lower exchange rates with the oceans due to deep basins and narrow links or due to natural barriers to the ocean. So, the water in the Baltic Sea is exchanged every 25-35 years (Kraatz, 2004). The residence time of water in the Mediterranean of approximately 100 years, making the Mediterranean ecosystem especially sensitive to the increase of the pollution stock of highly persistent substances like PFASs (Millot, 1989). Depending on water exchange rates from the adjacent seas with the oceans the translocation of PFASs sulfonyl groupParticularly volatile precursors, such as FTOHs, can undergo long-range atmospheric transport (Ellis et al., 2004). The detection of FTOHs for instance in the Arctic and Antarctic air agreed with the model prediction and conclusion, which supported the hypothesis of atmospheric transport toward remote regions (Bengtson Nash et al., 2010; Dreyer et al., 2009; Paul et al., 2009). Finally, sea spray aerosols (SSA) could be an important source of PFASs to the atmosphere and, over certain areas where sea spray deposition is important, a significant source to terrestrial environments, too (Johansson et al., 2019). SSA formation and their subsequent atmospheric transport and deposition have been suggested to play a prominent role in the occurrence of ionizable PFASs in the maritime Antarctica and other remote regions. However, field studies on SSA's role as vector of transport of PFAS are lacking. The effective enrichment of certain PFASs, such as PFAAs and possibly other PFASs in sea spray aerosols was recently demonstrated in laboratory studies, suggesting that SSA is a potential source of PFAAs to the atmosphere (Sha et al., 2022). The first field work by Casas et al. (2020) assessed the simultaneous occurrence of PFASs simultaneously at South Bay (Livingston Island, Antarctica) in seawater (SW), the sea-surface microlayer (SML) and SSA. Average PFASs concentrations were 313 pg L−1, 447 pg L−1, and 0.67 pg m−3 in SW, the SML and SSA, respectively. The enrichment factors of PFASs in the SML and SSA ranged between 1.2 and 5, and between 522 and 4 690, respec tively. This amplification of concentrations in the SML is consistent with the surfactant properties of ionic PFASs, while the large enrichment of PFASs in atmospheric SSA may be facilitated by the large surface area of SSA and the sorption of PFASs to aerosol organic matter. The measured large amplification of concentrations in marine aerosols supports the role of SSA as a relevant vector for long-range atmospheric transport mainly of charged PFASs. The transport via SSA may impact large areas of inland Europe and other continents in addition to coastal areas. Thus, SSA may currently be an important source PFASs to the atmosphere and, over certain areas, to terrestrial environments triggering also long-range transport. Particles and plastic debris Several PFASs may adsorb to particles. These particles such as dust or sediments may be easily drifted by air and water, and could be deposited far away from the point of release. Longer chain PFASs, in particular those with a sulfonyl group like PFSAs, Et FOSAA and FOSA are preferentially distributed in biota or the abiotic environment such as sediments, which could act as a sink for PFASs (Muir et al., 2019). However, considering the persistent nature of PFASs or their degradation products it is important to note, that sinks can become sources again. 107 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Dust can be transported vertically and horizontally. The main anthropogenic sources for particulate organic matter are the transport and the industrial sector. About 3.1 Mt/a of these particles are emitted in Europe annually (Koolen and Rothenberg, 2019). Global annual dust emissions are currently estimated to range between 1 000 and 3 000 Mt/a, whereas, beside anthropogenic sources, major dust source regions include the Sahara, the Arabian and Asian deserts (Tegen and Schepanski, 2009). Anthropogenic as well as natural particles could be loaded with PFASs. Dust deposition in remote areas occurs through both dry deposition and wet deposition associated with cloud and precipitation processes. As such, deposition involves a complex set of physical processes. Global dust deposition rates are strongly interlinked with the origin of the particles and with the meteorological and seasonal conditions (Knippertz and Todd, 2012). Rivers carry enormous amounts of sediments into coastal areas. For example, the annual transfer of sand, gravel and cobbles from the hinterland towards the Rhine delta was estimated by Frings et al. (2014) with 0.66 Mt/y. Not only natural sediments are transported by rivers and marine currents. The long-range transport of plastic debris and microplastics in the marine environment has been extensively documented (Eriksen et al., 2014; Howell et al., 2012; Maximenko et al., 2012; Obbard, 2018; Van Sebille et al., 2020). Plastics enter the oceans in massive amounts every year (4.8 to 12.7 Mt) (Jambeck et al., 2015) and accumulate in the oceans as plastic gyres (Eriksen et al., 2014) or in sediments. The transport of PFASs often does not take the direct and fastest routes from the point of release to remote areas. Depending on particles size, particles could remain for long times in the atmosphere and a deposition may occur even many years later. Adsorptive PFASs undergo long-range marine transport via plastic debris to a vast extent (Rani et al., 2017; Tanaka et al., 2020a; Tanaka et al., 2020b) . A large share of microparticles as well as larger particles found in the environment already consists of plastic debris. Non-fluorinated as well as fluorinated plastics absorb large amounts of nonpolymeric PFASs. Consequently, both polymeric PFASs and non-polymeric PFASs like additives can be transported as and with plastic debris. Larger plastic particles become suspended microplastics over time by mechanical crushing and by chemical transformation processes, which much easier could be moved to remote areas. Due to its density PE and PP float in the ocean surface and are easily transported by surface ocean currents and winds, whereas PVC tends to deposit near sources. An estimated amount of about 35 000 t of microplastics are floating in the world’s oceans (Cózar et al., 2014; Eriksen et al., 2014). However, according to Koelmans et al. (2017), this represents less than 1% of the floating ac cumulated plastic discharge. The remaining share is settled below the surface, at deep seafloor and in coastal sediments. The floating particles are transported with the ocean currents and a large part is trapped in the five subtropical gyres for about 40 years. It is estimated, that plastic debris are transported across the global oceans for more than 70 years (Wu et al., 2021). The Arctic Ocean appears to be a dead-end for plastic debris due to the poleward transport from sub-polar North Atlantic Ocean. The Arctic Ocean seafloor (e.g. Barents Seas) is thus an important sink of marine plastics (Cózar et al., 2014). It can be concluded, that PFASs associated with plastic debris are transported over long distances for many decades. As mentioned in chapter B.7.6, plastic debris (macro- as well as microparticles) is ingested by migratory species e.g. seabirds (Tanaka et al., 2020b). Partially, these organisms cover long distances on their extended migrations. So, migratory organisms have to be considered as important vector for PFASs and for longrange transport. Moreover, substances are subject to complex exchange processes between the different compartments. Thus, sediments may serve as temporary sinks for a certain time for instance for substances which are adsorbed to particles. 108 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.4.2.8.2. LRTP based on physicochemical data PFASs can be expected to be more volatile the higher their air water coefficient logKAW is. Volatile PFASs can undergo long-range atmospheric transport. It is highly likely that, e.g. fluorinated gases, short -chain fluorotelomer alcohols and perfluorinated olefins are transported this way due to their volatility. Of these the PFASs which gradually degrade into ionic PFCAs (see sections B.4.1.3 and B.4.2.1 “Mobility in water”) may change their route of long-range transport from air to water, depending on the chain length of the resulting PFCA. In general, uncharged PFASs like perfluoroalkane sulphonamides (FASAs), perfluoroalkane sulfonamidoethanols (FASEs) and fluorotelomer alcohols (FTOHs) are less water-soluble and more volatile than ionic PFASs. Once released in the environment, these PFASs can be (bio)degraded in the atmosphere or in other compartments under aerobic conditions to PFCAs and PFSAs (Ellis and Mabury, 2003; Martin et al., 2006; Rhoads et al., 2008; Schenker et al., 2008) (see also section B.4.1.3). PFCAs and PFSAs are PFAS subgroups that exist in anionic form in water. Therefore, these substances are highly water soluble. They could be transported by river waters and by ocean currents to remote regions (Armitage et al., 2009; McMurdo et al., 2008; Prevedouros et al., 2008). See also sections B.1.2 and B.4.2.1 for data on the properties relevant for LRTP. Substances with higher log KOC -values (>3.5) like long-chain PFASs and cyclic PFAAs can be (highly) adsorbed by particles. For ionic and ionizable substances, the water solubility and/or the sorption potential is dependent on counter ions (section B.4.2.1). So, cationic charged PFAS like perfluoroalkylamines also may adsorb to particles because of mainly anionically charged soil particles. Depending on natural circumstances, such substances could be moved either by water or adsorbed with particles. Not only natural particles serve as PFAS acceptor. As described above, particles of plastic debris are an important vector for highly adsorptive PFASs. Because of their small size, especially microplastics (<1 mm) have a large ratio of surface area to volume. That promotes adsorption of chemical contaminants to their surface. Microplastic particles therefore have a very high capacity to facilitate the transport of PFASs. Using substance physicochemical property data, the potential for long-range transport can be estimated. Different models use different matrices as basis for calculation. The OECD Tool (LRTP-Tool; ©OECD, 2009), which is used for the LRTP estimates, is a generic multimedia box model that yields estimates of numerical indicators for LRTP like the Characteristic Travel Distance (CTD [km]) for screening purposes. CTD is defined as the point in space at which the concentration as a function of place has decreased to 1/e ( about 37%) of the initial value. The CTD is applied for water and for air (CTDwater, CTDair) (Bennett et al., 1998). High CTD values were calculated for fluorotelomer alcohols. For the calculation an atmospheric lifetime of about 20 days was used from Ellis et al. (2003). For degradation half-life in water a value of 93 h from Gauthier and Mabury (2005) for 8:2 FTOH was used for all FTOHs 11. The respective log KAW values based on Arp et al. (2006) (Episuite calculation) and the log KOW values based on Arp’s COSMOTHERM estimations were used as input parameter to estimate the LRTP with the OECD-tool for fluorotelomer alcohols. The models standard setting was retained for the calculation. Based on their log KOW and the log KAW, FTOHs are mainly emitted to air. It is noted that with increasing chain length the CTD is decreasing (see Table B.10). 11 It is noted that this value is only used here for the purpose of LRTP calculation whereas it is not provided here for the purpose of the degradation assessment. 109 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.10. Estimated characteristic travel distances of fluorotelomeric alcohols and the respective input parameter for the OECD tool. [unit] C AS-Nr. 4:2 FTOH 6:2 FTOH 8:2 FTOH 10:2 FTOH 2043-47-2 647-42-7 678-39-7 865-86-1 C TD air km 9 405 6 727 3 816 2 839 C TD water km 11 9 10 10 molecular weight g x mol-1 264.09 364.10 464.12 564.13 logKaw -1.35 -2.39 -3.25 -4.23 logKow 3.21 4.44 5.66 6.91 h 480 480 480 480 h 107 93 93 93 h 72 72 72 72 degradation half-life air degradation half-life water half-life soil Because FTOHs are forming corresponding PFCAs (see section B.4.1.3), PFCAs may be released into the environment secondarily from a release of FTOHs within a very long distance. For PFCAs the respective calculated logKAW values (Episuite) and the log KOW values calculated with COSMOTHERM based on Arp et al. (2006) were used as input parameters for the OECD-tool to estimate the LRTP for perfluoroalkyl carboxylic acids. The standard half-life’s for non-degradable substances of 106 h for all compartments, provided in the OECD tool, were used for the calculation. Table B.11. Estimated characteristic travel distances of selected PFCAs and the respective input parameter for the OECD tool. [unit] PFPeA PFHxA PFHpA PFOA PFNA PFDA C AS-Nr. 2706-90-3 307-24-4 375-85-9 335-67-1 375-95-1 335-76-2 C TD air km 398 785 667 771 852 725 1 010 029 1 042 881 1 045 656 C TD water km 25 917 18 187 11 993 6 598 4 005 2 249 molecular weight log Kaw gx mol-1 264.05 314.05 364.06 414.07 464.08 514.09 -3.04 -2.66 -2.37 -2.03 -1.79 -1.52 log Kow 3.43 3.26 3.82 4.30 4.84 5.30 half-life air h 1 000 000 1 000 000 1 000 000 1 000 000 1 000 000 1 000 000 h 1 000 000 1 000 000 1 000 000 1 000 000 1 000 000 1 000 000 h 1 000 000 1 000 000 1 000 000 1 000 000 1 000 000 1 000 000 half-life water half-life soil CTD estimations show, that PFCAs theoretically could be transported by air multiple times across the globe, due to their non-degradability. Because, PFCAs partition to air occurs only to a very limited extent, transport via water appears more relevant. Sho rter chain PFCAs like PEPeA are distributed with water more than 20 000 km. Assuming a multidirectional substance distribution, that means a distribution over the complete earth 110 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) surface. With increasing chain length, the CTD in water decreases, however, long-chain PFCAs are still transported for several thousand kilometres (see Table B.11). As described in chapter B.4.2.1, PFASs with high log KOC and log KOW strongly adsorb to particles. Highly adsorptive PFASs are PFOS, PFOA and PFBS, PFOS. Especially cyclic PFAAs (C5-C7, and greater) strongly adsorb to soil and particles. For these substances a particle mediated long-range transport is highly likely. It is not common to provide CTD values for particles. But looking at data provided before, particle associated PFASs, especially those which are adsorbed to plastic debris, may transported over the whole earth surface for many decades, too. B.4.2.8.3. LRTP evaluation based on monitoring data Generally, field data are accepted as evidence for the long-range transport of a chemical if (1) measured levels are available in locations distant from the sources of its release; (2) monitoring data show that long-range environmental transport of the chemical may have occurred via air, water, or migratory species. However, the mere detection of a chemical in a remote region cannot necessarily be understood as evidence of long-range transport, as the potential influence of local sources has to be considered as well (Scheringer, 2009), since both long-range transport as well as local pollution may contribute to the presence of PFASs. Various PFASs are already ubiquitously detectable in remote areas like in arctic -, antarcticor glacier firns, at open sea or even in the higher atmosphere (see B.4.2.7, Appendix B.4.2.3. and Appendix B.4.2.7.). All these data confirm the long-range transport. While little information on PFASs in Antarctica is available, new data show a wider range of PFASs in the Arctic since 2009. Several novel PFASs were detected in biota, water and air from Nordic and Arctic countries showing the wide distribution and potential for long-range transport of precursors and PFASs. Precursor -PFASs contribute to the total PFASs detected in all environmental compartments of remote areas and were frequently detected in many matrices. Data for tracking PFASs along the way from its point of release to remote areas are rare. Following, the results of such a study are discussed in regard to LRTP of PFASs. In their comprehensive studies Möller et al. (2010) investigated the distribution and sources of 40 PFASs in the river Rhine watershed in the Rhine-Waal-Scheldt-Estuary and at open North Sea. PFOS, PFOA, PFBS and PFBA usually were the main detected substances. In the North Sea, about 175 kilometres offshore, an average summarised PFAS concentration of 0.35 ng/L was provided in this publication. The measured concentration of PFASs in the North Sea seems to be low, however, if linked to the volume of the North Sea of 54 000 cubic kilometres (areal of the North Sea 575 000 square kilometres with an average depth of 94 m) it results in 2 000 t PFAS. Many large rivers, carrying high loads of pollutants, drain into the North Sea. The river basins are densely populated and large industries occur in this area. Going from open sea via river estuaries upstream the rivers, an increasing PFAS concentration could be measured. Along the Dutch coast, in the RhineWaal-Scheldt-Estuary, an average PFAS concentration of about 12 ng/L could be detected. The average PFAS concentration at the mouth of the rivers Rhine and Waal, in the Haringvliet, is 121 ng/L. Going further upstream the rivers, a high amount of PFASs is released into the Nederrijn and in the river Waal (average concentration of summarised PFAS: 260 ng/L). Large amounts of PFASs are also drained into the North Sea by the river Scheldt. At the Scheldt rivers’ mouth (Western Scheldt) an average PFASs concentration of 95 ng/L was measured. The highest mean concentration of PFASs was measured in the river Scheldt with 498 ng/L. In the river Ruhr (into which the river Moehne is drained) at the inflow into the Rhine an average PFASs concentration of 47 ng/L was detected. Downstream Leverkusen the mean concentration of summarised PFASs raised to 181 ng/L instantaneously, compared to the average summarised PFASs concentration of about 21 ng/L in the Rhine upstream Leverkusen. 111 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The respective increasing PFASs concentration in the rivers Rhine, Waal and Scheldt was caused by direct industrial emissions or indirectly via wastewater treatment plants or by the inflow of contaminated water from the several tributaries. Another source of PFASs contamination in river water may result from application of contaminated sludge to fields. The PFASs are eroded from the soil by rain. The substances are transported via river effluent over long distances into the Rhine-Waal-Scheldt-Estruary. So, for example, the linear distances from the point of emission are >200 km, from Leverkusen to a monitoring point in the North Sea or >300 km from contaminated soil in the river Moehnes drainage area to the same monitoring point at the open sea. On the one hand, the substances are diluted within the North Sea due to the large amount of marine water, but also trapped in the North Sea for a while. The studies from Möller et al. (2010) not only show an increased PFAS concentration in the vicinity of heavy industrialised and dense populated areas but also indicate that composition of measured PFASs concentrations at different sampling points along the River Rhine watershed and in the North Sea is very distinct. Whereas in the North Sea and in the Lake Constance a high share of PFOS was detected, the share of PFHxS and PFBA is much higher in the river sections. Due to the restriction of PFOA and PFOS the industries substituted these substances with shorter chain PFASs already some years ago. So, the high share of PFOS in the Lake Constance and in the North Sea mainly could be related to previous environmental releases. In contrast to this, the measured concentration of shortchain PFASs in the river sections represent current PFAS releases. This is an indication that enclosed water bodies may serve as temporary sinks for PFASs, too. Due to trapping effects PFASs are transported time delayed e.g. from large lakes into the rivers or from est uaries and adjacent seas into the oceans. Conclusion: Many PFASs have potential for long-range transport mainly due to their high persistence. Mobility in water and volatility contribute to the LRTP as well. Some precursors, such as FTOH, are themselves long-range transported. The same can be expected for other volatile PFASs. Precursor PFASs degrade over time to PFAAs which can be expected to be long-range transported, based on their physicochemical properties and high persistence. This was demonstrated with model calculation for selected PFCAs. The LRT pathways are different, depending on the PFASs, and may change when a precursor degrades to the corresponding arrowhead. Volatile PFAS such as fluorinated gases and uncharged PFASs like perfluoroalkane sulphonamides (FASAs), perfluoroalkane sulfonamidoethanols (FASEs) and partially fluorotelomer alcohols (FTOHs) are mainly transported via air. The long-range transport via water is the predominant pathway for anionic PFASs like PFCAs and PFSAs). An important vector to remote areas is plastic debris for adsorptive PFASs. Further, all compartments (e.g. air, water and sediments) may serve as a temporary sink for PFASs. And, the combination of different sinks and considering the persistent nature of PFASs or their degradation products, sinks can become sources for environmental releases again over very long periods of time. However, for the majority of PFASs no data on transport pathways or point sources are available and thus substantial uncertainties on the concern of the long-range transport potential remain. B.4.2.9. Bioaccumulation For this dossier, a review of recent peer-reviewed articles and scientific reports was carried out. The recent data illustrate the specific mechanisms of bioconcentration and biomagnification and list in particular results of non-regulated or assessed PFASs. In the following, data from modelling, laboratory and field studies as well as from monitoring campaigns are presented. B.4.2.9.1. Toxicokinetics of PFASs in animals The overall body burden and target site concentration of a chemical and its metabolites is governed by its toxicokinetics (i.e., processes of absorption, distribution, metabolism, and excretion, ADME). Yet, available toxicokinetic data primarily focus on PFAAs (De Silva et 112 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) al., 2021; Kudo, 2015; Kudo and Kawashima, 2003). In a recent preliminary study targeted as well as non-targeted and suspect screening of striped dolphins of the Mediterranean Sea stranded along Tuscany coasts revealed a broader spectrum of PFAS uptake (Roscioli C et al., 2022). Non-target analysis allowed identifying FOSA as well as n:3 PFCAs. FOSA was the most abundant followed by FHxSA and FBSA in all the tissues. n:3 FTCA were mainly detected in high levels in liver and only in traces in some blood samples. This suggests that they are generated from precursors such as FTOHs and FTS present at high concentration in liver cells, but they are not distributed to different tissues through the bloodstream (Roscioli C et al., 2022). Overall, there are more robust data regarding the ADME of PFASs in humans and rodent model species than in other species. Furthermore, PFOA and PFOS are the most frequently studied PFAS. As discussed in the chapter on toxicokinetic processes (see section B.5.1), studies with mammalian species show that PFCAs and many other PFASs are readily absorbed and distributed especially among protein-rich tissues like liver, serum, and kidney (Kudo and Kawashima, 2003). Due to the high sorption capacities (e.g. (Armitage et al., 2012; Luebker et al., 2002), the toxicokinetic behaviours of many PFASs (uptake, translocation, bioaccumulation, biotransformation, elimination, etc.) differ considerably from the common hydrophobic and persistent organic pollutants (Ng and Hungerbühler, 2013). Indeed, PFAAs were found to serum albumin, α globulins, and fatty acid-binding proteins For instance, PFOA and PFOS preferentially distribute to the liver in most species; PFBA and PFHxS appear to preferentially distribute to the serum and, to a lesser extent, to the liver in animals (reviewed by Ebert et al. (2020)). The enterohepatic circulation of PFAAs likely contributes to their extended elimination half lives in humans. It was demonstrated that PFBS, PFHxS, and PFOS were transported into hepatocytes both in a sodium-dependent and a sodium-independent manner by Na+ /taurocholate co-transporting polypeptide (NTCP). PFBS, PFHxS, PFOS and PFCAs with 7-10 carbons are substrates of organic anion transporting polypeptides (OATPs). Chinese Hamster Ovary and Human Embryonic Kidney cells were used to demonstrate that human OATPs can transport PFBS, PFHxS, PFOS and the 2 PFCAs (C8 and C9). In addition, it was shown that rats possess different OATPs able to transport all 3 sulfonates. This study suggests that besides NTCP and the human apical sodium-dependent bile salt transporter, OATPs are also capable of contributing to the enterohepatic circulation and extended human serum elimination half-lives of PFBS and other PFAAs (Zhao et al., 2016). As outlined in chapter B.5.1 on toxicokinetics, in mammals the major route of excretion for PFASs is renal elimination and to a smaller extent biliary and faecal (ATSDR, 2021; Consoer et al., 2014; EFSA, 2020; Kudo and Kawashima, 2003). Likewise, elimination of PFOA in rainbow trout occurred primarily via the renal route, which is consistent with numerous studies also in mammals suggesting that fish possess membrane transporters that facilitate the movement of PFOA from plasma to urine (Consoer et al., 2014). In both, humans and animals, PFASs are transferred to the foetus via the placenta and to the offspring via breast milk (e.g. DeWitt (2015)). PFAAs do not readily cross the mature blood-brain barrier. This is supported by findings from Harada et al. (2007) in which PFOA and PFOS cerebral spinal fluid concentrations in adult humans were more than 500-fold lower than serum concentrations. However, high levels of PFSAs and PFCAs were found in the brain of wild mammals and birds, e.g. in polar bears or gulls (Leranth et al., 2008; Verreault et al., 2005). PFASs are transferred to offspring, milk and eggs in many taxa, including livestock species (see review by Death et al. (2021). For instance, Sharpe et al. (2010) showed that when zebrafish underwent a reproductive cycle in the presence of PFOS, approximately 10% of the adult PFOS body burden was transferred to the developing embryos, resulting in a higher total PFOS concentration in eggs (116±13.3 µg/g) than in the parent fish (72.1±7.6 µg/g). Gronnestad et al. (2017) demonstrated in hooded seals (Cystophora cristata) how 8 PFASs were transferred from mother to offspring via maternal transfer via both milk and the placenta, of which placental transfer is the dominant pathway reaching high levels in pub plasma. 113 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Also in birds, maternal transfer is a major exposure route (Gockener et al., 2020b; Jouanneau et al., 2021; Kowalczyk et al., 2020) though little is known about the extent of the transfer of the different PFASs to the eggs, especially for alternative fluorinated compounds. A recent study by Jouanneau et al. (2021) showed that contamination of females and their eggs were dominated by linPFOS, then PFUnA or PFTriA. They measured PFASs, including HFPO-DA, ADONA, and F-53B, in the plasma of prelaying black-legged kittiwake females breeding in Svalbard. There was a linear association between females and eggs for most of the tested PFASs and maternal transfer ratios in females and eggs suggest that the transfer is increasing with PFAAs carbon chain length, therefore the longest chain PFCAs were preferentially transferred to the eggs. PFCAs and PFSAs are not metabolised in animals (Kudo, 2015). Studies on PFOA as well as PFSAs such as PFOS (C8-PFSA) and PFDS (C10-PFSA) in rats have shown that they are excreted untransformed without forming any metabolites or conjugates. Thus, PFCAs are believed to represent metabolically inert and stable end-stage products. However, certain precursors have been shown to transform, to various extents in rodents, e.g. into their perfluorinated carboxylate “backbone structures”, such as 8:2 FTOH that is metabolised into e.g. PFOA and PFNA (C9-PFCA) (Henderson and Smith, 2007). Neutral volatile atmospheric precursors such as FTOH and FASA can biotransform in humans and wildlife, contributing to overall exposures of the arrowhead PFAAs (including PFOS and PFOA) (De Silva et al., 2021). Based on the available literature on elimination half-lives in animal experiments, it can be assumed that several replacement substances have similar bioaccumulative properties as the arrowhead PFASs (Brantsaeter et al., 2013). Still, data providing evidence for bioaccumulative potential of less-well studied PFASs is rare. Rice et al. (2021) reported that HFPO-DA and ADONA may have bioaccumulative potential, but no clear conclusion could be drawn based on available data in rats. Polychlorotrifluoroethylene (PCTFE) oligomers accumulated preferentially in the liver during long-term oral exposure of rats (Kinkead et al., 1991). 6:2 FTSA, but not 6:2 FTCA, was detected at high levels in serum and liver following repeated-dose exposure of mice (Sheng et al., 2017), whereas 6:2 ClPFESA accumulated in serum and gut (Pan et al., 2019b) and liver of mice (Pan et al., 2021). Similarly, sodium ρ-perfluorous nonenoxybenzene sulfonate (OBS) bioaccumulated in mouse liver, but also in the gut (colon, ileum). At the highest dose, OBS was also detected in serum, kidney and faeces (Wang et al., 2019a). It is challenging to demonstrate bioaccumulation of PFAAs in humans, because of the lack of constant exposure over several years. However, findings from several studies provide indirect evidence. For example, a study on predictors of selected PFASs in pregnant women found that reported time since the most recent pregnancy was positively associated with measured blood concentrations of PFOA, PFNA and PFOS, thus indicating bioa ccumulation during the time between pregnancies (Brantsaeter et al., 2013). In a study on levels of 6:2 Cl-PFESA and 8:2 Cl-PFESA in metal plating workers, lower median values were observed in individuals which had been employed less than one year (18.6 ng/mL) as compared to those with more than one year of employment (1 347 ng/mL) (Shi et al., 2016), which indirectly indicates bioaccumulation. Conclusion: Many PFASs are readily absorbed and distributed especially among proteinrich tissues (especially liver, serum, kidney) and, thus, the toxicokinetic behaviours of PFASs differs considerably from the traditional hydrophobic chemicals. Many PFASs are transferred to the fetus via the placenta and via eggs, and to the offspring via breast milk. B.4.2.9.2. Characteristics influencing behaviour bioaccumulation and toxicokinetic Protein binding Some PFASs, mainly PFAAs, have been discussed as being proteinophilic rather than 114 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) lipophilic substances questioning the usual bioaccumulation assessment scheme. As discussed in the PFHxA background document, the relationship between structure (e.g. chain length) and affinity to proteins is complex and thus still a matter of research. In mammals, serum albumin, fatty acid binding proteins (FABPs) and organic anion transporters (OATs) have been identified as important to the tissue distribution, speciesspecific accumulation, and species- and gender-specific elimination rates of PFCAs and PFSAs (Zhang et al., 2013d). Several biological matrices with high sorption capacities and the corresponding distribution coefficients or binding affinities have been reported for some PFAAs. These include serum albumin as a transport protein in blood, phospholipids as the major component of cellular membranes, alpha globulins, and liver fatty acid binding proteins (FABPs) that belong to the intracellular lipid-binding protein superfamily (Allendorf et al., 2019; Allendorf et al., 2021; Armitage et al., 2012; Bischel et al., 2011; Droge, 2019; Luebker et al., 2002; Weaver et al., 2009). Unlike the accumulation in adipose tissue, binding to proteins and accumulation in specific organs has a higher potential to cause adverse effects, since organ toxic effects may arise (see B.5.1). Several PFCAs and PFSAs bind to albumin and other proteins (EFSA, 2020). Protein binding contributes to the distribution of PFASs to blood rich organs and liver. PFOA and PFOS have a high affinity for human uptake transport proteins such as organic anion transporters OAT1, OAT3 and urate transporter URAT1 (EFSA, 2020). The data for PFECAs is more limited. A recent study showed that HFPO-DA binds to human liver fatty acid-binding protein (Sheng et al., 2018a). There are no studies available investigating direct binding between HFPO-DA and albumin, and therefore it currently remains unknown whether HFPO-DA interacts with albumin directly or not. However, highest concentrations of HFPODA in rodents are found in the liver and the blood, indicating that HFPO-DA follows the same pattern of protein binding as other PFASs (Chen and Guo, 2009; Sheng et al., 2016). In addition to this, no data is available on OAT efficacy for HFPO-DA. It is therefore not known what effect HFPO-DA has on the functioning of the OATs and if resorption of HFPODA in the lumen of the kidney will occur in humans or not. Differences in protein affinity may have implications for renal reabsorption of PFASs by the organic anion transporter in which higher affinity may imply higher reabsorption (Lu et al., 2021). Lu et al. (2021) estimated binding affinities of various PFASs on OAT4 which indicated that the affinity increased with carbon chain length. PFASs can also bind to hLFABP and it is reported that binding affinity increases in the order of 6:2 fluorotelomer carboxylic acid (FTCA) < 6:2 fluorotelomer sulfonic acid (FTSA) < HFPO-DA < PFOA (Sheng et al., 2018a). Higher binding affinity of (6:2 Cl-PFESA) and HFPO-DA compared to PFOA was reported (Rice et al., 2021; Sheng et al., 2018a). Bischel et al. (2011), investigated with equilibrium dialysis the binding of PFCAs (C2–C12) and PFSAs (C4, C6, and C8) to bovine serum albumin (BSA). An increase in the protein water distribution (KPW) with increasing chain length was observed for PFCAs with four to six perfluorinated carbons. Log KPW values for C4 to C12 PFAAs range from 3.3 to 4.3. Affinity for BSA increases with PFAA hydrophobicity but decreases from the C8 to C12 PFCAs, likely due to steric hindrances associated with longer and more rigid perfluoroalkyl chains. With the exception of PFDoDA over 90% of all PFAAs were bound to BSA (Bischel et al., 2011). Allendorf (2021) analysed a consistent set of distribution coefficients for a series of PFAAs and 4 other PFASs (HFPO-DA, DONA, 9Cl-PF3ONS and PFECHS) to physiologically relevant matrices including albumin, membrane lipids, structural proteins, and storage lipids. The results of the physiologically based distribution calculations showed that albumin with the highest partitioning coefficients as well as membrane lipids, and structural proteins are of major relevance in estimating the accumulation of PFAAs in different organs. 115 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Log KPW values for C7 to C11 PFCAs range from 4.6 to 4.86. For PFSAs with four to eight carbons (C4, C6, and C8) Log KPW values are 3.34, 4.94 and 4.81, respectively. For HFPODA, DONA, 9Cl-PF3ONS and PFECHS Log KPW values are 3.19, 4.06, 5.14 and 4.68, respectively. The albumin/water partition coefficients for the alternatives are in the same range as for classical PFAAs. Structural modifications such as the introduction of ether groups into the chain do not reduce sorption to albumin, whereas the chlorine atom in 9ClPF3ONS seems to even increase the sorption to albumin (Allendorf et al., 2019). This study concludes that the introduced ether groups do not considerably alter the distribution properties compared to PFCAs. Apart from serum albumin the binding to other proteins may have an impact. In the study by Zhang et al. (2013d) the binding strength to a fatty acid binding protein, a highabundance protein in liver, was found to be dependent on the length of the fluorocarbon tail and the polar headgroup. According to the authors, this dependence can be rationalized by the binding mode inside the protein’s ligand-binding cavity, as revealed by molecular docking analysis. The authors conclude that based on their calculation, the potential disruption of the uptake and transport of fatty acids cannot be ignored. Affinity to serum proteins increases with decreasing chain length (Bischel et al., 2011). This might, besides placental transfer, also affect other aspects of accumulation, such as passage of the blood-cerebral barrier. Notably, a previous comparative study demonstrated occurrence of PFASs in human brain, using autopsies of various organs that had been sampled from 20 subjects (Perez et al., 2013). In this study, the concentrations of 21 PFASs (C4-C18 PFCAs, C4, C6 and C8 PFSAs, and perfluorohexyl ethanoic acid (FHEA), perfluorooctyl ethanoic acid FOEA, and perfluorodecyl ethanoic acid FDEA; perfluorooctanesulfonamide (PFOSA)) were analysed in 99 samples of autopsy tissues (brain, liver, lung, bone, and kidney) from subjects who had been living in Tarragona (Catalonia, Spain). Note that PFBA levels in lung and kidney tissues are likely overestimated by Perez et al. (2013) as discussed under B.5.1.2 samples showed detectable values of at least two of the investigated compounds. Although PFASs accumulation followed different trends depending on the specific tissue, some similarities were observed between liver and brain, on one hand, and between kidney and lung, on the other hand. In liver, PFHxA, PFOS and FHEA were the most prevalent compounds, with median concentrations of 68.3, 41.9 and 16.7 ng/g, respectively. PFOS was present in 90% of the samples, while PFOA could be quantified in 45% of the samples (median: 4.0 ng/g). In brain, PFHxA was the main compound, being detected in all the samples at concentrations ranging from 10.1 to 486 ng/g. The contributions of PFNA (median: 13.5 ng/g) and PFDA (median: 12.4 ng/g) were also relatively important in brain samples. In contrast, PFOS was only quantified in 20% of the samples (median: 1.9 ng/g), whereas PFOA was not detected in any of them. In general terms, lung was the tissue showing the highest accumulation of PFASs. In contrast, detection of PFHxS and other PFASs was much lower in the investigated brain samples (Ao et al., 2019). Again, further studies are required to clarify whether this effect is related to neurological or neurobehavioural health risks. Protein binding is assumed as one of the main mechanisms explaining facilitated tissue distribution. This hypothesis is supported by the findings of Numata et al. (2014). Four PFSAs and 3 PFCAs were quantifiable in feed, plasma, edible tissues, and urine of pigs. Unexcreted PFAAs accumulated in plasma (up to 51%), fat, and muscle tissues (collectively, meat 40–49%), liver (under 7%), and kidney (under 2%) for most substances. An exception was PFOS, with lower affinity for plasma (23%) and higher for liver (35%) in the body of pigs. Nevertheless, the potential to bind to BSA may not fully explain the toxicokinetics. Transporter proteins as described above may also have an impact on toxicokinetics. Apart from this, a study has shown that PFAAs bind to peroxisome proliferator-activated receptors. This plays a role in lipid metabolism, induces conformational changes of this receptor and may thus change the function of the protein (Zhang et al., 2014). 116 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Conclusion: Protein binding is assumed as one of the main mechanisms explaining facilitated tissue distribution. The relationship between structure (e.g. chain length) and affinity to proteins is complex and thus still a matter of research. Several biological matrices with high sorption capacities and the corresponding distribution coefficients or binding affinities have been reported for some different groups of PFAS. These include serum albumin as a transport protein in blood, phospholipids as the major component of cellular membranes, alpha globulins, and liver fatty ac id binding proteins (FABPs) that belong to the intracellular lipid-binding protein superfamily. Chain length and chemical structure Depending on chain length and functional groups, PFASs vary in their ability to bioaccumulate. This has been mainly investigated for PFCAs and PFSAs. As discussed in the PFHxA background document, the affinity of PFAAs to proteins is chain-length dependent and increases up to a certain number of perfluorinated carbons depending on the protein (Ng and Hungerbühler, 2014; Zhang et al., 2013a). Shi et al. (2018b) investigated how the differential tissue distribution and bioaccumulation behaviour of 25 PFASs in crucian carp from two field sites impacted by point sources can provide information about the processes governing uptake, distribution and elimination of PFASs. Transformation of concentration data into relative body burden demonstrated that blood, gonads, and muscle together accounted for >90% of the amount of PFASs in the organism. Functional group was a relatively more important predictor of internal distribution than chain-length for PFASs. PFCAs and PFSAs with six or more perfluorinated carbons displayed a positive trend in BAFs with increasing chain-length and distinct differences for PFASs with different functional groups (sulfonic acid > sulfonamide > carboxylic acid). Branched isomers of PFOS and PFOA had consistently lower BAFs compared to the linear isomer while the inclusion of an ether bond and/or terminal chlorine leads to a higher BAF for C8 Cl-PFESA compared to n-PFOS. This may be one reason why whole-body bioaccumulation factors (BAFs) for short -chain PFASs deviated from the positive relationship with hydrophobicity observed for longer-chain PFASs. Overall, the study suggests that BAF patterns were most consistent with protein-binding mechanisms, although partitioning to phospholipids may contribute to the accumulation of long-chain PFASs in specific tissues (Shi et al., 2018b). Ahrens and Bundschuh (2014) published a review paper on the behaviour and impacts of different PFASs in aquatic systems, including bioaccumulation in various taxa. They showed that the average sum of PFOS and PFOA concentrations were typically in the ranges of 0.110 µg/kg ww for invertebrates, 1-100 µg/kg ww for fish and reptiles, 1-500 µg/kg ww for birds and 5-10 000 µg/kg ww for mammals. PFOS concentrations were typically up to 3 orders of magnitude higher compared with PFOA. The lower bioaccumulation potential of PFOA in comparison to PFOS was believed to be driven by both a shorter C-chain and differences in the functional group (e.g. carboxylic acids vs. sulfonates). Branched isomers were measured and were generally more readily excreted than linear isomers, which lead linear isomers to appear as more bioaccumulative than the comparable branched isomers in the addressed aquatic taxa. De Silva et al. (2009) conducted dietary exposure studies in rainbow trout, administering PFOA isomers (manufactured with Electrochemical Fluorination (ECF)), linear PFNA, and branched PFNA for 36 d. Throughout exposure and depuration phases, blood and tissue sampling ensued. The accumulation ratio revealed similar accumulation propensity o f nPFOA and two minor branched PFOA isomers; however, the majority of branched isomers (n=5) had lower accumulation ratio values than n-PFOA. Sharpe et al. (2010) investigated the bioaccumulation of branched and linear PFOS isomers in rainbow trout and zebrafish. They found the branched PFOS isomers to bioaccumulate significantly less than the linear PFOS isomers, which may explain the relative lower concentration of branched PFOS isomers in some aquatic species in the field. The study 117 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) remained unclear about whether this observed difference was due to differences in the uptake phase or in the elimination phase. A different study on fish found linear PFOS isomers to accumulate more readily than the branched PFOS isomers, possibly being the result of a more efficient elimination of the branched PFOS isomers (number of branched isomers: n=5, (Hassell et al., 2020)). Conclusion: It is evident that protein binding of various PFASs efficiently distribute them into different organs and tissues, allowing for passage across brain and placental barriers, yielding in maternal- offspring transfer also via the milk. Due to surfactant properties of some PFASs and the broad range of chemical structures, the affinity of different PFAAs for proteins varies widely, suggesting binding site-specific interactions, while the toxicokinetic behaviour is complex. B.4.2.9.3. Modelling partitioning and bioaccumulation behaviour Mechanistic bioaccumulation models developed for neutral lipophilic contaminants such as polychlorinated biphenyls and organochlorine pesticides have been widely used by academics, risk-assessment professionals, and regulatory authorities (Arnot and Gobas, 2004). Strong relationships between empirically derived bioaccumulation metrics (BCF, BAF, TMF) and distribution ratios for protein–water (DPW) and membrane–water (DMW) of individual PFAAs have been demonstrated for some ecosystems (Anderson et al., 2019; Byrne et al., 2017). DPW and DMW are key parameters that may be useful for predicting PFAS bioaccumulation potential. However, distinction between proteins is warranted, since different protein types can exhibit different sorptive capacities (Henneberger et al., 2016). Accordingly, simple equilibrium partitioning‐based models may require utilizing a series of distribution coefficients for different proteins (e.g. transporter protein–water distribution ratios and structural protein–water distribution ratios). This approach was e.g. used to assess tissue‐specific bioaccumulation of PFAAs and other ionic compounds in laboratory exposed fish (Chen et al., 2016; Chen et al., 2017a). Chen et al. (2017a) showed a positive linear relationship between log BCF ss values and physicochemical properties such as octanol–water distribution coefficients (log Dow ), membrane–water distribution coefficients (log Dmw ), albumin–water distribution coefficients (log DBSAw ), and muscle protein–water distribution coefficients (log Dmpw ), indicating the importance of lipid–, phospholipid–, and protein–water partitioning. A chemical activity–based approach to ecological risk assessment bridges some gaps between traditional empirical modelling efforts and mechanistic models (Gobas et al., 2018a; Gobas et al., 2018b; Gobas et al., 2020b). This approach was used to assess bioaccumulation and exposure risks of several PFASs in wildlife at AFFF -impacted sites (Gobas et al., 2020a). The chemical activities of PFOS and other PFAAs indicated that these compounds tend to primarily biomagnify in food webs composed of air-breathing wildlife (birds, mammals, terrestrial reptiles) compared to those comprising only aquatic organisms. An advantage of this approach that is particularly relevant to PFAS is that it can be used effectively for both neutral and ionic substances, including anionic, cationic, and zwitterionic compounds (De Silva et al., 2021). Beyond simple partitioning-based models for substance screening, more sophisticated approaches may be required for higher-resolution modelling. For example, ionic compounds binding to intra- and extracellular protein (serum albumin, L-FABP), as well as membrane-associated organic anion transporters, may act to provide both enhanced sorption capacity and advective transport across biological membranes (Ng and Hungerbühler, 2013). This affects uptake and elimination rates as well as tissue distribution and helps explain the long elimination half-lives of PFAA in organisms. Wang et al. (2011c) conducted studies employing the COSMOtherm model to estimate physicochemical properties for 130 individual PFASs, namely perfluoroalkyl acids (including branched isomers for C4–C8 perfluoroalkylcarboxylic acids), their precursors and some 118 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) important intermediates. The estimated physico-chemical properties are interpreted using structure-property relationships and rationalised with insight into molecular interactions. Within a homologous series of linear PFASs with the same functional group, both air–water and octanol–water partition coefficients increase with increasing perfluorinated chain length, likely due to increasing molecular volume. For PFASs with the same perfluorinated chain length but different functional groups, the ability of the functional group to form hydrogen bonds strongly influences the chemicals’ partitioning behaviour). The partitioning behaviour of all theoretically possible branched isomers can vary considerably; however, the predominant isopropyl and monomethyl branched isomers in technical mixtures have similar properties as their linear counterparts (differences below 0.5 log units). Even with the large number of studies available, physiologically based toxicokinetic models for predicting the bioaccumulation of PFASs are, however, still in developmental phases (Armitage et al., 2017) and thus highly uncertain. Two fish bioaccumulation models have been developed that account for some of the physico-chemical characteristics of PFASs. One is the BIOconcentration model for Ionogenic Organic Compounds (BIONIC) (Armitage et al., 2013), and the other is a proteinpartitioning bioaccumulation model from Ng and Hungerbühler (2013); Ng and Hungerbühler (2014). Due to data availability, these models were built on training sets limited to PFCAs with carbon chain lengths greater than 7. Both models focused on predictions in freshwater fish. Given the lack of protein-partitioning values for fish, the protein-partitioning component of the bioaccumulation model (i.e., Ng and Hungerbühler (2013) used rat and human protein-partitioning values. The BIONIC model (Armitage et al., 2013) considers phospholipids, rather than proteins, as the primary repository for PFASs, but the model does recognize the ionization potential of these substances. However, because PFCAs and PFSAs bind primarily to fatty acid–binding proteins and lipoproteins/albumin, and then are sequestered in protein-rich tissues, these proteins are important to consider. However, the protein-partitioning model underestimated the bioaccumulation of PFHxS (6 carbons) and generally underestimated whole -body bioaccumulation (Ankley et al., 2021). Thus, it is possible that t he active clearance and reabsorption processes described in the protein-partitioning model do not operate in the same way or to the same extent in fish as in rats or humans (Ankley et al., 2021). None of these models offer predictions for short -chain (less than 7 carbon) PFCAs and PFSAs, but available empirical data from Martin et al. (2003a) and Martin et al. (2003b), used in both models, showed that short -chain PFCAs (i.e., PFHxA, PFHpA) did not bioconcentrate in any rainbow trout tissue. Physiologically based toxicokinetic (PBTK) models incorporating absorption, distribution, metabolism, and excretion metrics have been developed to assess the toxicokinetic of PFOS and PFOA in various animal models, including fish and mammals (Andersen et al., 2008; Consoer et al., 2014; Khazaee and Ng, 2018), but are still in the developmental phase. Conclusion: Several modelling studies indicated that simple equilibrium partitioning‐based models may require utilizing a series of distribution coefficients for different proteins. Positive linear relationships BCF values and physicochemical properties such as octanol– water distribution coefficients, membrane–water distribution coefficients, albumin–water distribution coefficients, and muscle prot ein–water distribution coefficients. Even with the large number of studies available, physiologically based toxicokinetic models for predicting the bioaccumulation of PFASs are, however, still in developmental phases and thus highly uncertain. B.4.2.9.4. Laboratory and field studies on bioaccumulation A recent review article by Burkhard (2021) evaluated literature for bioconcentration factors (BCFs) and bioaccumulation factors (BAFs) for PFASs in freshwater species from 22 taxonomic classes. The assembled data were evaluated for quality, and for gaps and limitations in bioaccumulation information for the PFAS universe of chemicals. For BCFs, measurements have been focused on perfluorohexane sulfonic acid (PFHxS), PFOS, 119 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) perfluorooctanoic acid (PFOA), PFNA, and PFDA. Accordingly, in general, carbonyl and sulfonyl PFAS classes are relatively data rich, whereas phosphate, fluorotelomer, and ether PFAS classes are data limited for fish and non-existent for most other taxonomic classes. Laboratory studies are limited to species with standardized testing protocols. Taxonomic classes with the most measurements were, in descending order, Teleostei (fish), Bivalvia, and Malacostraca. The numbers of PFAS chemicals with available measured BCFs (and BAFs) are summarized in Table B.12 by structure category, and these counts cover all tissue types (BCFs and BAFs for whole body, muscle, fillet, liver, and other organs). The BCFs reported were measured in the laboratory using standardized protocols with aquatic organisms and a water-borne exposure procedure. Steady state was demonstrated for all tests. Tests are typically run with a 28-d uptake phase followed by a depuration phase running from 14 to 28 d depending on the chemical. As commonly noted in the literature, BCFs and BAFs with 8 or more carbons increase uniformly with increasing number of carbons in the alkyl chain (Babut et al., 2017; Kwadijk et al., 2010). For the other taxonomic groups, a similarly large increase in BCFs and BAFs generally occurs between PFOA and PFNA. BCF values of different PFASs groups followed the order phosphinic acids > PFCAs > PFSAs > others. Among the 43 PFASs compounds for which BCF and BAF studies are available in different aquatic species 62% (27 compounds) have a median ± SD BCF and/or BAF values above the REACH threshold for B (log BCF >3.3). Table B.12 and Figure B.64 show large inter- and intra-species variability and that differences between individual studies and the medians across all studies exist for BCFs, as also confirmed by Wassenaar et al. (2020). These may have several reasons: there is an inherent variability in BCF and BAF measurements, which is commonly observed in BCF and BAF measurements for a chemical (Arnot and Gobas, 2006; Wassenaar et al., 2020). Furthermore, bioaccumulation of PFAS chemicals appears to have a slight dependency on concentration of the chemical (Chen et al., 2016; Dai et al., 2013; Hoke et al., 2015; Inoue et al., 2012), though other reporting demonstrated significant correlation of concentrations for PFOA (RIVM, 2017) and different studies have different exposure concentrations. Laboratory measurements suggest that BCFs decline with increasing exposure concentration (Chen et al., 2016; Dai et al., 2013; Hoke et al., 2015; Inoue et al., 2012). For example, Inoue et al. (2012) reported BCFs of 720 and 1300 with aqueous concentrations of 16 and 1.88 µg/L, respectively, for PFOS with common carp, a 1.8-fold increase in BCF with an 8.5-fold decrease in exposure concentration. As reviewed by Burkhard (2021), within a study, most often, lower concentrations provide higher measured BCFs, for example, studies with PFOS in 3 amphibian species (Abercrombie et al., 2019), PFASs with common carp (Inoue et al., 2012) , and with zebrafish (Chen et al., 2016), and with Daphnia magna (Dai et al., 2013). Cause(s) of the decrease in BCFs with higher concentrations are unknown. However, PFAS residues are known to be controlled by a combination of passive diffusion and active transport processes (Ankley et al., 2021) and are generally not concentration dependent, which suggests there might be some type of capacity limitation in the active transport processes. 120 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.12. Overview on bioconcentration factors (BCFs) and bioaccumulation factors (BAFs; L/kg wet wt) for fish for whole-body, muscle/fillet, and liver tissues (median ± standard deviation, n) provided by Burkhard (2021). Bold numbers = above the B criterion of REACH Annex XIII. SS = steady state; OECD = Organisation for Economic Co -operation and Development. Chemical Tissue B metric CAS no. Whole body Log BCF SS and Log BCF kinetic Muscle/ fillet Log BCF SS and Log BCF kinetic Liver Log BCF SS Log BCF kinetic Whole body Log BAF Muscle/fi llet Log BAF Liver Log BAF 4504862-2 4516747-3 9261252-7 12088 5-29-2 4528551-6 7200768-2 7382936-4 1.18 ± 0.08 (2) –0.05 ± — (1) 0.98 ± 0.3 (3) 1.26 ± — (1) 1.38 ± 0.61 (14) 2.78 ± 0.51 (6) 3.79 ± 0.48 (3) –0.22 ± 1.15 (5) –0.64 ± 1.21 (3) 0.40 ± 1.01 (5) 0.51 ± 1.24 (3) 0.82 ± 1.18 (7) 2.80 ± 0.40 (4) 3.81 ± 0.54 (4) 1.80 ± 1.23 (5) 1.15 ± 1.53 (3) 1.73 ± 0.99 (5) 1.78 ± 1.20 (3) 1.93 ± 1.00 (14) 3.79 ± 0.24 (4) 2.98 ± 0.93 (8) 0.47 ± 0.96 (40) 0.15 ± 1.46 (18) 0.09 ± 1.34 (19) –0.16 ± 1.27 (32) 0.90 ± 1.14 (105) 2.07 ± 0.76 (88) 3.06 ± 0.49 (72) PFUnDA 19685 9-54-8 3.57 ± 0.31 (5) 3.97 ± 0.88 (4) 3.41 ± 0.74 (8) PFDoDA 17197 8-95-3 3.64 ± 0.60 (8) 4.12 ± 0.83 (4) 4.46 ± 1.16 (4) 2.16 ± 1.68 (6) 1.74 ± 2.45 (5) 1.40 ± 1.51 (11) 1.80 ± 1.24 (10) 2.16 ± 0.87 (48) 2.80 ± 1.15 (41) 3.45 ± 0.62 (43) 3.47 ± 1.01 (21) 2.18 ± — (1) 0.37 ± 1.11 (3) 1.48 ± 1.43 (4) 2.79 ± 1.62 (6) 0.92 ± 0.99 (6) 1.97 ± 1.05 (48) 2.84 ± 0.73 (20) 3.72 ± 0.65 (30) 4.34 ± 0.72 (28) 4.32 ± 1.52 (17) PFTrDA 86237 4-87-6 4.34 ± 0.46 (2) 4.51 ± 0.85 (2) 5.22 ± 1.08 (2) — 4.66 ± 0.16 (3) 5.43 ± — (1) 36597 1-87-5 PFHxDA 6790519-5 PFOcDA 1651711-6 Sulfonyl compounds: 4.40 ± 0.56 (4) 3.68 ± 0.01 (2) 2.57 ± 0.09 (2) 4.96 ± 0.62 (2) — 4.74 ± 1.09 (3) — — — 4.38 ± — (1) — 5.08 ± — (1) — — — — — — 1.06 ± 0.49 (7) 2.07 ± 0.25 (6) — 0.09 ± 1.15 (4) 1.34 ± 0.19 (2) — 1.74 ± 1.06 (5) 2.41 ± 0.4 (4) — 2.00 ± 1.13 (5) 2.30 ± 0.74 (25) — 3.01 ± 0.66 (21) 3.27 ± 0.96 (7) 3.17 ± 0.88 (18) 3.52 ± 0.78 (81) 1.35 ± 0.84 (16) 1.28 ± 0.86 (56) 2.20 ± — (1) 3.09 ± 0.60 (155) 1.18 ± 0.34 (5) 2.18 ± 0.58 (17) 3.20 ± 0.10 (3) 3.74 ± 0.84 (61) C arbonyl compounds: C arboxylic acids: PFBA PFPeA PFHxA PFHpA PFOA PFNA PFDA PFTeDA 3.89 ± 0.77 (54) 4.50 ± 1.57 (28) Sulfonic acids: PFBS PFHxS PFHpS PFOS 37573-5 35546-4 37592-8 176323-1 121 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Chemical Tissue B metric CAS no. PFDS 33577-3 — Liver Log BCF SS Log BCF kinetic Whole body Log BAF Muscle/fi llet Log BAF Liver Log BAF 1.30 ± — (1) — — — 3.21 ± — (1) — — 4.20 ± 0.15 (3) — — — 3.70 ± 0.53 (12) 2.95 ± 0.94 (24) 4.00 ± 0.2 (11) — — 4.10 ± 0.16 (2) — — 2991— — — 3.50 ± 50-6 0.06 (2) Perfluoroalkane sulfonyl amido ethanols, phosphate esters (SAmPAPs): SAmPAP 29651.42 ± — — — — 52-8 (1) — 3.45 ± 0.21 (2) — — PFEC HS Sulfonamides: FOSA 75491-6 Whole body Log BCF SS and Log BCF kinetic — Muscle/ fillet Log BCF SS and Log BCF kinetic — — — Sulfonamidoacetic acids: MeFOSAA 235531-9 — EtFOSAA Perfluoroalkyl phosphate compounds: Phosphonic acids: PFHxPA 4014376-8 1.46 ± 0.36 (2) 1.44 ± 0.14 (2) 2.83 ± 0.35 (2) — — — PFOPA 4014378-0 1.92 ± 0.13 (2) 1.74 ± 0.12 (2) 2.66 ± 0.18 (2) — — — 5229926-0 2.21 ± 0.13 (2) 1.87 ± 0.04 (2) 3.36 ± 0.67 (2) — — — Phosphinic acids: C 6/C 6 40143PFPiA 77-9 C 6/C 8 PFPiA — 5.12 ± 0.71 (2) 7.36 ± 1.24 (2) 5.51 ± 0.16 (2) 7.03 ± 0.88 (2) 5.50 ± 0.69 (2) 7.35 ± 0.93 (2) — — — — — — C 8/C 8 PFPiA 4014379-1 8.30 ± 0.89 (2) 7.44 ± 1.43 (2) 7.36 ± 1.65 (2) — — — C 6/C 10 PFPiA — 8.52 ± 1.01 (2) 7.78 ± 1.43 (2) 7.43 ± 1.36 (2) — — — C 8/C 10 PFPiA — 5.79 ± 0.07 (2) 2.71 ± 3.80 (2) 1.85 ± 2.28 (2) — — — C 6/C 12 PFPiA — 6.30 ± 0.98 (2) 5.35 ± 1.56 (2) 4.89 ± 2.73 (2) — — — — — — PFDPA Fluorotelomer-related compounds: Fluorotelomer alcohols: 8:2 FTOH 6782.50 ± 39-7 0.42 (2) — — n:2 Fluorotelomer alcohol, phosphate esters (PAPs): 6:2 diPAP 8:2 diPAP 5767795-9 — — 3.50 ± — (1) — — — 67841-1 — — 2.39 ± — (1) — — — — 4.14 ± — (1) 1.49 ± — (1) — Fluorotelomer sulfonate: 4:2 FTSA 75712 4-72-4 — — 122 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Chemical Tissue B metric CAS no. 6:2 FTSA Whole body Log BCF SS and Log BCF kinetic 1.54 ± 0.62 (3) Muscle/ fillet Log BCF SS and Log BCF kinetic Liver Log BCF SS Log BCF kinetic 2761997-2 — — 8:2 FTSA 3910834-4 — — — Per- and polyfluoroalkyl ether-based compounds: Whole body Log BAF Muscle/fi llet Log BAF Liver Log BAF — 4.86 ± 0.41 (2) — 3.94 ± 0.56 (4) — — PFESAs, salts, and esters–monoethers: 4:2 C lPFESA — — — F-53B (6:2 73606- 2.85 ± C l-PFESA) 19-6 1.16 (6) — — 3.43 ± 0.20 (4) — 4.33 ± 0.19 (5) — 3.85 ± 0.42 (6) 3.27 ± — (1) 4.76 ± 0.4 (6) 8:2 C lPFESA — — — 4.69 ± — (1) 5.64 ± — (1) PFEC As, salts and esters–diethers: HFPO-DA 1325213-6 — — HFPO-TA 1325214-7 — — — — — — 0.61 ± — (1) 1.05 ± — (1) 0.50 ± — (1) 1.75 ± — (1) 8332989-9 0.40 ± — (1) A synthesis of 513 laboratory-based and 931 field-based measurements indicates that long-chain PFCAs with a 12 to 14 carbon-chain length generally exhibit the highest bioaccumulation potential, with whole-body BCF values ranging between 18 000 and 40 000 L/kg (Gobas et al., 2020b). Laboratory-based whole-body BCFs of PFCA with 8 to 11 carbon-chain lengths are generally lower (BCF range 4.0–4 900 L/kg). As described in the monitoring section B.4.2.4, field studies show that air-breathing organisms are more likely to bioaccumulate PFAAs compared to gill breathing organisms since they cannot eliminate PFAAs via ventilation. Accordingly, certain PFASs like C8-C10 PFCAs and C6 PFSA are more likely to bioaccumulate in air-breathing organisms, including humans, as compared to gill breathing organisms and that trophic magnification occurs in certain food webs in the environment where air breathing organisms are top-predators in the food chains (e.g. (De Silva et al., 2021)). Conclusion: For BCFs, measurements have been focused on PFHxS, PFOS, PFOA, PFNA, and PFDA. However, also for other members of these substance classes BCFs have been reported in several studies. Accordingly, in general, carbonyl and sulfonyl PFAS classes are relatively data rich, whereas phosphate, fluorotelomer, and ether and all other PFAS subgroups are data limited for fish and non-existent for most other taxonomic classes. Generally, measured BCF values show extremely large inter and intra species v ariability for the same compounds indicating large uncertainties. Among the 43 PFASs for which BCF and BAF studies are available in different aquatic species, 62% (27 compounds) have a median ± SD BCF and/or BAF values above the REACH threshold for B (log BCF >3.3). BCF values of PFAS groups follow the order phosphinic acids > sulfonyl acids > carbonyl acids > others. In general, BCFs decline with increasing exposure concentration, while BCFs/BAFs of PFASs with 8 or more carbons increase uniformly with increasing number of carbons in the alkyl chain, with highest bioaccumulation potential of compounds with 12 to 14 carbon-chain length. However, aqueous testing underestimates the potential of bioaccumulation of PFASs, since air-breathing organisms are more likely to bioaccumulate PFASs compared to gill breathing organisms. Thus, established assessment methods of bioaccumulation based on bioconcentration testing in aquatic organisms do not suffice to 123 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) assess the bioaccumulation behaviour of PFASs. B.4.2.9.5. Bioaccumulation and trophic magnification in wildlife Analyses of biota at different trophic levels of the ecosystem show clearly that many PFASs that do not accumulate in aqueous species, bioaccumulate and biomagnify to high concentration levels in air breathers including top predator animals like polar bears, whales and seals (for example De Silva et al. (2021) and Chen et al. (2021b)) . In top-predators dietary uptake routes indeed are important in PFASs toxicokinetics, especially for uptake in mammalian and top predator species. The recent review by De Silva et al. (2021) demonstrated that elevated exposures of wildlife to PFAS represent a concern for their health directly and for human populations that consume wildlife (Fair et al., 2019; Guillette et al., 2020). PFASs were detected in endangered species like green turtles and polar bears (Eggers Pedersen et al., 2015; Wood et al., 2021). In 2001, the first report on the global occurrence of PFOS in wildlife was released, illustrating widespread presence in biological tissues even in remote regions such as the Arctic (Giesy and Kannan, 2001). Concentrations of PFOS and other PFAA have been detected in invertebrates, fish, amphibians, reptiles, birds, and mammals worldwide (Ahrens et al., 2011; Houde et al., 2011; Muir et al., 2019). Several comprehensive reviews (Ahrens et al., 2011; Houde et al., 2011; Muir et al., 2019) have summarized data from available biomonitoring studies. The highest PFAS concentrations in wildlife tend to be associated with proximity to contaminated sites. For example, one of the highest reported fish PFOS concentrations (maximum 9 349 ng/g dry wt in whole fish tissue) was from an AFFF -impacted site downstream from Barksdale Air Force Base in Louisiana (Lanza et al., 2017). Many biomonitoring studies have identified elevated exposures to legacy and emerging PFASs as the result of industrial activities (Death et al., 2021; Groffen et al., 2019; Guillette et al., 2020; Liu et al., 2017). As shown by Burkhard (2021), for species commonly consumed by humans (e.g.e.g. fish, clams, mussels, oysters, and scallops, lobsters, crabs, shrimp, and prawns and winkles), there are BCF and BAF measurements for the PFCAs (n‐PFHxA through n-PFUnDA), PFSAs (n‐PFBS, n‐PFHxS, and n‐PFOS) and FOSA (except for Gastropoda). For other chemicals, BCF and BAF data are limited, and measurements for chemicals beyond the PFASs just listed are needed. Comparison of laboratory BCFs with field BAFs revealed that 60% (26 of 43 comparisons) of the BAFs are greater than their corresponding BCFs, and similar proportions exist for BAFs based on whole body, muscle, and liver (Burkhard, 2021), Table B.12, Figure B.64). The BAFs include all exposure routes and, as suggested by the modelling efforts of Larson et al. (2018), sedimentary sources of the PFAS can cause BAFs to be greater than BCFs. Trophic magnification factors (which are food web average biomagnification factors) are slightly >1 for some PFAS compounds (Fang et al., 2014; Loi et al., 2011; Martin et al., 2004; Munoz et al., 2017). These field data suggest that BAFs for some PFASs should be larger than their BCFs for aquatic species because of biomagnification processes within the food web. 124 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.64. Measured BCFs and BAFs in log units from different aquatic species (provided in Burkhard (2021). Avian and marine mammalian food webs exhibit the highest reported trophic magnification factors (TMFs) for PFAAs (Kelly et al., 2009). In aquatic piscivorous food webs, TMFs tend to be much lower. For example, TMFs of PFOS in the Lake Ontario aquatic piscivorous food webs including air breathing top predators, e.g. dolphins range between 1.9 and 5.9 (Houde et al., 2008). In particular, PFOS and several other PFAS of concern, which are likewise moderately hydrophobic and poorly metabolizable substances, may not biomagnify extensively in aquatic food webs because of efficient respiratory elimination to water via gills (De Silva et al., 2021; Kelly et al., 2009). Conversely, these substances can biomagnify to a high degree in food webs containing air-breathing animals because elimination of these substances via lung–air exchange is negligible. The contribution of PFAA precursors to field-based measurements of BAFs represents a major gap in understanding of PFAS bioaccumulation. For example, one study noted higher than expected accumulation of PFCA with 5 and 6 carbons in marine plankton from the north-western Atlantic and posited that this reflects the accumulation of degraded precursor compounds (Zhang et al., 2019d). Another study that included liver tissues from marine mammals from the same region found a large fraction (30–75%) of unidentified organofluorine (Spaan et al., 2020). Occurrence and accumulation of PFAS in apex species 125 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFOS is the dominant PFAS in apex predators, with an average proportion of approximately 60%, and the proportion of PFOS in many cases (>25%) is above 80%. Meanwhile, longchain PFCAs (C9-C14) are major PFASs in apex predators as well (Chen et al., 2021c). In particular, PFOS/PFOSA as well as PFCAs of C8/9-C11/12 chain length have shown to biomagnify in terrestrial (e.g. Müller et al., 2011; Zhao et al., 2013c) and aquatic food webs (Kelly et al., 2009; Xu et al., 2014). As a consequence, species of high trophic position (e.g. apex predators) accumulate high concentrations of certain PFASs in their tissues (Chen et al., 2021c). Among European predators, especially Eurasian otters (Lutra lutra) have shown to have comparably high PFOS levels compared to other predators (Androulakakis et al., 2022; Badry et al., 2022). The fact that PFOS is still the most dominant PFAS in biota was suggested to be related to its persistence and the continued use of PFOS precursors like fluorotelomer alcohols and polyfluoroalkyl phosphate (Houde et al., 2011). Apex predators of aquatic food webs have shown to accumulate high PFOS levels in proximity of potential point pollution sources (e.g. Badry et al., 2022) . More detailed information on occurence and concentrations of regulated and non-regulated PFASs are provided in section B.4.2.6 and B.4.2.7. Conclusion: As outlined in this section, analyses of biota at different trophic levels of ecosystems clearly show that many PFASs bioaccumulate and biomagnify to high concentration levels in top predator animals. Comparison of laboratory BCFs with field BAFs revealed that in most cases the BAFs are greater than their corresponding BCFs likely due to multiple exposure pathways in wildlife. Avian and marine mammalian food webs exhibit the highest reported TMFs for PFAA while in aquatic piscivorous food webs TMFs tend to be much lower. In particular, PFOS, which are likewise moderately hydrophobic and poorly metabolizable substances, may not biomagnify extensively in aquatic food webs because of efficient respiratory elimination to water via gills but can biomagnify to a high degree in food webs containing air-breathing animals because elimination of these substances via lung–air exchange is negligible. Especially PFOS and long-chain PFASs are frequently detected in protein-rich tissues of almost all wildlife species, whereas novel analytical techniques demonstrate the presence of emerging PFASs such as 7:3 FTCA or 6:2 ClPFESA. B.4.2.9.6. Substance-specific bioaccumulation data Bioaccumulation of long-chain (C8-C14) PFCAs and (C6-C8) PFSAs     C11-C14 PFCAs have been assessed to fulfil the vB-criterion of REACH Annex XIII. C8-C10-PFCA, as well as their salts meet the B-criterion (vB not assessed). C6-PFSA has been assessed to meet the vB criterion of REACH Annex XIII. PFOS and its salts have been assessed to meet the POP criterion for bioaccumulation due to its potential to bioaccumulate and biomagnify in mammals and piscivorous birds Details of the assessment can be found in the supporting documentation of the listing in the Candidate List (ECHA, 2012b; ECHA, 2012c; ECHA, 2012d; ECHA, 2012e; ECHA, 2013; ECHA, 2015b; ECHA, 2016b; ECHA, 2017c) 12. Bioaccumulation of short-chain (C4-C6) PFCAs and PFSAs, and C7 PFCA For the shorter-chain PFCAs and PFSAs no bioconcentration in aquatic organisms due to uptake from the aqueous phase by diffusion via the gills is expected. Due to their high water-solubility SC-PFAAs are, unlike C8-C10-PFCAs and C6-PFSA, expected to be quickly excreted via gill permeation (Martin et al., 2003a; Martin et al., 2003b). A study was conducted on both bioconcentration and biomagnification in rainbow trout (Martin et al., 2003a; Martin et al., 2003b). Both studies investigated a homologous series of PFCAs and 12 http://chm.pops.int/TheC onvention/ThePOPs/TheNewPOPs/tabid/2511/Default.aspx , date of access: 2022-12-22. 126 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFSAs. and perfluoroalkyl chain lengths shorter than eight and six carbons, respectively, could not be detected in most tissues and were considered to have insignificant bioconcentration factors (BCF). SC-PFCAs and SC-PFSAs are to some extent bioaccumulative in air breathing organisms, as far as this has been possible to assess. Elimination half -lives of short-chained-PFAAs, which have been recently used for long-chained PFAAs as a metric to estimate bioaccumulation potential in air-breathing organisms, are shorter in comparison with longchained PFAAs. Depending on the species, half-lives range from a couple of hours to several days in mammals (Chengelis et al., 2009a; Gannon et al., 2016; Numata et al., 2014) and up to over a month-almost a year in humans (Bolan et al., 2021), see toxicokinetic section B.5.1.1.4). Bioaccumulation of cyclic PFAAs De Silva et al. (2011) estimated the Bioaccumulation factors (BAFs whole-body) for two fish species (lake trout and Walleye), of several PFAAs, including PFECHS based on the ratio of fish to water concentrations measured in samples from the Great Lakes. The mean log BAF (whole body homogenate, wet weight) values for fish corresponded to 2.8 for PFECHS, 2.1 for PFOA, and 4.5 for PFOS. BAFs are calculated with the assumption that the concentration of the pollutant observed in the fish is the result of exposure to the same pollutant in the water and diet. As such, if the pollutant in the fish is the result of biotransformation of a precursor, then the resulting BAF may be an overestimate. It is not probable that the fishes could have been exposed to precursors to PFECHS, as no precursors to this substance are known. A trend in tissue/blood ratios (liver > kidney > bladder > muscle) was observed for PFECHS, PFPCPeS, br-PFOS, lin-PFOS, and F-53B suggesting that these compounds share similar mechanisms for uptake and distribution in the body (Wang et al., 2016a). Overall, the trend of mean Log BAF whole-body F-53B (4.6) ≈ lin-PFOS (4.6) > br-PFOS (3.8) > PFECHS (2.7) > PFPCPeS (1.9) appeared to follow the hydrophobicity pattern, with lowest BAF for the less hydrophobic cyclic PFAAs. Isomer-specific differences in the tissue/blood distribution ratios and BAF whole-body for PFECHS and PFPCPeS indicate that ring structure and position of the sulfonic acid group affect the bioaccumulation potential. No studies of the biomagnification or trophic magnification of PFECHS or other cyclic PFAAs have been identified. Bioaccumulation of PFECAs and PFESAs Few laboratory studies are available on the bioaccumulation potential of PFECAs in fish (see Appendix B.4.2.9.). Additional data on field studies for HFPO-DA are presented in section B.4.2.7.6 and in the Annex XV dossier on the proposal for identification of HFPODA as a substance of very high concern (ECHA, 2019d). According to the SVHC support document for HFPO-DA (ECHA, 2019d), the BCFs for HFPODA are below 2 000. Based on the structural similarities with PFOA, it can be expected that bioaccumulation factors are higher at low environmental concentrations. Although bioaccumulation of HFPO-DA is still low at environmental concentrations, fish consumption could be a relevant exposure route for humans as it is for PFOA (ECHA, 2019d). Chlorinated polyfluoroalkyl ether sulfonic acid (Cl-PFESA) is regarded as a Chinese perfluorooctane sulfonate (PFOS) alternative with a commercial name of F -53B, in which the main component is 6:2 Cl-PFESA with its two homologues (8:2 and 10:2 Cl-PFESA) as impurities (Jin et al., 2020c). Shi et al. (2015) recently reported on the first detection of F-53B in biological samples and determined the tissue distribution and whole -body bioaccumulation factors (BAF whole body) in crucian carp (Carassius carassius). Tissue/blood ratios showed that distribution of F-53B primarily occurs to the kidney, gonad, liver, and heart. Median Log BAF whole body values for F-53B exceeded regulatory bioaccumulation criterion and were significantly higher than those of PFOS in the same data sets. On the basis of its apparent omnipresence and strong bioaccumulation propensity, it was 127 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) hypothesized that F-53B could explain a significant fraction of previously unidentified organofluorine in biological samples from China. However, no laboratory study on the bioaccumulation of F-53B is available. The available data indicate that bioconcentration factors and bioaccumulation factors (BAF) are low for shorter-chain PFECAs. Similarly, as for LC-PFCAs, it can be expected that PFECAs with more than the seven perfluorinated carbons would show a higher potential for bioconcentration. However, there are no bioconcentration studies available for longer chain PFECAs. The only PFESA that has been commercialised is 6:2 Cl-PFESA, the major component of F53B, a substitute for PFOA used for decades in China in the chrome plating industry see Annex A chapter 3.5. In a Chinese field bioaccumulation study 6:2 Cl-PFESA (substitute of PFOS) was detected in the tissues (kidney, gill, muscle, brain, heart, gladder, gonad, liver) and blood of all sampled Crucian carp (n=43) from two different locations (Shi et al., 2015). The median Log BAF whole body (L/kg ww) values for 6:2 Cl-PFESA ranged from 3.80 to 5.23. Wang et al. (2016a) conducted a study on the bioaccumulation factors and tissue distribution of and other PFAAs in Crucian carp (Carassius carassius) downstream of the Beijing Airport in China. The median concentration for F53-B was highest in blood (11.42 ng/g ww) followed by kidney (10.02 ng/g ww), liver (8.73 ng/g ww) and bladder (8.54 ng/g ww) and lowest in muscle (1.75 ng/g ww). The mean concentration in the surface water for the sampling sites ranged from 90% removal efficiency) removing PFAAs (≥C4 PFSAs and ≥C4 PFCAs from water) in bench- (Appleman et al., 2013) and full-scale studies (Appleman et al., 2014; Thompson et al., 2011). High efficiency of removal could also be expected for PFECAs and PFESAs (Hopkins et al., 2018). The disposal of concentrate, which will contain elevated concentrations of PFASs, will need to be addressed. To summarise, the properties of especially the most stable PFASs resulting from the degradation of other PFASs in the environment, are such that render water treatment very difficult hence increasing the technical demands (and costs) of the treatment of water obtained for drinking water, process water and household water uses. The increasing number of findings of PFASs in surface waters, groundwaters and drinking water (see section B.4.2.1.3) demonstrates the need for purification of drinking water. EurEau (2022) has also assessed the purification methods for water suppliers and concludes that PFASs should be managed at their source due to the challenges in the water supply. B.4.5.8. Wastewater treatment Several studies showed that conventional wastewater treatment has a limited efficiency in removing both, short-chain and long-chain PFCAs and PFSAs from aqueous waste streams. Shorter-chained and longer-chained PFCAs and PFSAs accumulate in sludge and are released to receiving waters via WWTP effluents (Arvaniti and Stasinakis, 2015; Bossi et al., 2008; Eriksson et al., 2017b). Shorter-chained and longer-chained PFCAs are generally found in higher c oncentrations in the effluent water than the influent water (Bossi et al., 2008; Eriksson et al., 2017b); Sinclair and Kannan (2006), which indicates that they are hard to remove from water during wastewater treatment process and that precursor PFASs can degrade into PFCAs and PFSAs during the wastewater treatment. Several PFCAs precursors (FTSAs, FTCA/FTUCA, diPAP) were found in lower concentration in the effluent water than the influent water of three WWTP in Sweden (Eriksson et al., 2017b), which, together with the calculated increase in the daily discharge of PFCAs (effluent and sludge) compared with the daily incoming amount in the influent water indicates that PFCA precursors can potentially degraded to PFCAs during wastewater treatment process. More than 75% of the PFASs detected in sludge from 3 WWT P in Sweden were precursor compounds and intermediates to PFAAs (FTSAs, FTCAs, FTUCAs, diPAP, monoPAP; (Eriksson et al., 2017b). The formation of PFAAs from precursors in WWTP is dependent on process temperature and treatment type, with higher rates of formation in biological WWTP (during the acerbic biological step) at longer hydraulic retention times and higher temperatures (Guerra et al., 2014). Care should be taken when comparing concentration of PFASs measured in the waste streams between WWTPs, due to the effect of the treatment process and the specific 139 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) sampling strategies (e.g. sampling at high or low flow periods and seasonal effects (Guerra et al., 2014; Sinclair and Kannan, 2006). PFPAs with short perfluoroalkyl chains (≤C6) are expected to combine high water solubility and low sorption to organic matter, therefore there will be hardly removed from water in WWTP. However, sorption is expected increase with increasing chain length PFAAs (Wang et al., 2016b); see section 1.1.1.4. of the main report), with larger fraction of the shorterchained PFPAs partitioning into the sludge. The adsorption of shorter-chained PFAAs in sludge has been observed to increase with increasing chain length (increase in distribution coefficient between influent water and sludge; (Eriksson et al., 2017b) which can be explained by increasing hydrophobicity of the molecules (Zhang et al., 2013a). However, for shorter-chained PFAAs, the electrostatic interactions between the anionic functional group and the sludge are expected to play a more important role than hydrophobic interactions (Zhang et al., 2013a) . Higher solid-liquid distribution coefficients have been observed in WWTPs for PFOS compared to PFOA (Eriksson et al., 2017b; Guerra et al., 2014) and laboratory incubation (Zhang et al., 2013a), which can be explained due to the stronger hydrophobic properties of the sulfonic analogues. The lower distribution coefficient of PFHxS compared to PFHxA in sludge from a WWTP observed by Eriksson et al. (2017b) could be due to the higher formation of PFHxA from precursors in the sludge. Municipal WWTPs are not able to effectively remove shorter-chained or longer-chained PFAAs and the discharge of municipal sewage water is a significant source of PFAAs to the aquatic environment (Becker et al., 2008; Filipovic et al., 2015; Loos et al., 2013). In addition, the disposal of the sludge from industrial and municipal WWTPs can also be a significant source of PFAAs to the terrestrial environment (Eriksson et al., 2017b; GomezCanela et al., 2012; Washington et al., 2010). B.4.5.9. Remediation of contaminated sites Leaching of PFASs to groundwater from contaminated soils can be reduced by sorbent amendment using activated carbon (94 and 99.9% PFOS reduction), compost soil (29 and 34% PFOS reduction) and montmorillonite (28 and 40% PFOS reduction) (Hale et al., 2017). However, an assessment of the suitability of the method for field application with higher PFAS concentrations and different PFASs is still required. 140 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5. Human health hazard assessment The majority of available data on human health effects address the toxicity of PFAAs (mainly PFCAs and PFSAs; in particular PFOA and PFOS), while less or no data are available for other PFAS groups. For the vast majority of PFASs (estimated >99%), no data on repeated-dose toxicity, carcinogenicity, or reproductive toxicity is available. Existing recent literature reviews and assessments were explored (ATSDR, 2021; EFSA, 2020; EPA-US, 2021b; Fenton et al., 2021; NTP, 2016a; PFAS-TOX-DATABASE, 2021; Rice et al., 2021; Zeng et al., 2021) to identify the main types of effects of PFASs. Besides literature reviews and assessment reports, also registration data from the ECHA dissemination site and IUCLID were screened and available information used for the assessment. In addition, information from literature searches was included (Embase, Medline, Web of Science, PubMed, Scifinder). The literature searches were performed in 2021 to identify published data with relevance to human health that has not been covered by recent reviews. Search terms were selected based on EFSA (2020). Table B.13 lists the PFASs for which data have been reported in one or more of these reviews, assessments, and studies. Table B.13. Non-exhaustive list of PFASs in assessed literature and study reports and examples for polymeric PFASs. PFAS group No. of C PFAS PFCAs Perfluoroalkyl carboxylic acids / carboxylates C 2 Trifluoroacetic acid TFA Abbreviation C4 Perfluorobutanoic acid, Perfluorobutyric acid PFBA C5 Perfluoropentanoic acid PFPeA C6 Perfluorohexanoic acid PFHxA C7 Perfluoroheptanoic acid PFHpA C8 Perfluorooctanoic acid PFOA APFO C9 Perfluorononanoic acid PFNA CAS No EC No 76-05-1 2923-184 (Na + salt) 200-9293 220-8796 375-22-4 206-786104953 86-0 (NH4+ salt) 2706-90- 220-3003 7 307-24-4 206-1962923-26- 6 4 (Na + 220-881salt) 7 2161547-4 244-479(NH4+ 6 salt) 375-85-9 206-798201099 59-5 243-518(Na + salt) 4 335-67-1 206-3973825-26- 9 1 223-320+ (NH4 salt) 4 375-95-1 4149-604 (NH4+ salt) 206-8013 - 141 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS group PFSAs PFSIAs No. of C PFAS C 10 Perfluorodecanoic acid Abbreviation PFDA CAS No 335-76-2 3108-427 (NH4+ salt) EC No 206-4003 221-4705 C 11 Perfluoroundecanoic acid PFUnDA C 12 Perfluorododecanoic acid PFDoDA 2058-948 307-55-1 C 13 Perfluorotridecanoic acid PFTrDA C 14 Perfluorotetradecanoic acid PFTeDA C 16 Perfluorohexadecanoic acid PFHxDA C 18 Perfluorooctadecanoic acid PFODA 218-1654 206-2032 276-7452 206-8034 267-6381 240-5825 6790519-5 1651711-6 Perfluoroalkane sulfonic acids / sulfonates C1 Trifluoromethanesulfonic acid TFMS 1493-136 2926-274 (K + salt) 216-0875 608-3344 C4 Perfluorobutane sulfonic acid PFBS C5 Perfluoropentane sulfonic acid PFPeS C6 Perfluorohexane sulfonic acid PFHxS 375-73-5 2942049-3 (K + salt) 2706-914 355-46-4 3871-996 (K + salt) 206-7931 249-6163 220-3012 206-5871 223-3932 C7 Perfluoroheptane sulfonic acid PFHpS 375-92-8 206-8008 C8 Perfluorooctane sulfonic acid PFOS 217-1798 220-5271 C9 Perfluorononane sulfonic acid PFNS C 10 Perfluorodecane sulfonic acid PFDS 1763-231 2795-393 (K + salt) 6825912-1 335-77-3 PFOSI 647-29-0 - C 6/C 8 Perfluoroalkyl phosphinic acids C 6/C 8 PFPiA C 8/C 8 Perfluoroalkyl phosphinic acids C 8/C 8 PFPiA 61080034-5 4014379-1 1240600- - 206-4019 Perfluoroalkane sulfinic acids C8 PFPiAs 7262994-8 376-06-7 Perfluorooctane sulfinic acid Perfluoroalkyl phosphinic acids C 6/C 12 Perfluoroalkyl phosphinic acids C 6/C 12 PFPiA 142 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS group PFECAs No. of C PFAS Abbreviation CAS No 41-2 EC No C 8/C 10 Perfluoroalkyl phosphinic acids C 8/C 10 PFPiA 50077681-8 - Perfluoroalkyl ether carboxylic acids Hexafluoropropylene oxide dimer acid HFPO-DA, GenX 1325213-6 6203780-3 (NH4+ salt) DONA ADONA 95844544-8 (NH4+ salt) 236-2368 - 2,3,3,3-tetrafluoro-2-[1,1,2,3,3,3hexafluoro-2(heptafluoropropoxy)propoxy]propionic acid Hexafluoropropylene oxide tetramer acid Ammonium difluoro[1,1,2,2tetrafluoro-2(pentafluoroethoxy)ethoxy]acetate HFPO-TA 1325214-7 236-2373 HFPO-tetramer acid EEA-NH4 6529416-8 90802052-0 (NH4+ salt) - Potassium 2-(3-trifluoromethoxy1,1,2,2,3,3-hexafluoropropoxy)2,3,3,3-tetrafluoropropionate mv31 K+ 49680564-2 (K + salt) - Ammonium difluoro{[2,2,4,5tetrafluoro-5-(trifluoromethoxy)-1,3dioxolan-4-yl]oxy}acetate F-DIOX - Perfluoro-2-methoxypropanoic acid PFMOPrA Perfluoro-4-methoxybutanioc acid PFMOBA 119093127-1 (NH4+ salt) 1314029-9 86309089-5 17476706-7 3949292-7 Dodecafluoro-3H-4,8-dioxanonanoic acid PFESAs 480-3104 - - 2,2,3,3,4,4-Hexafluoro-4(nonafluorobutoxy)butanoic acid 2,2,4,4,6,6,8,8,10,10,12,12,14,14,14Pentadecafluoro-3,5,7,9,11,13hexaoxatetradecanoic acid Perfluoroalkyl ether sulfonic acids 4:4 C 8 PFEC A 1,1,2,2,3,3,4,4-Octafluoro-4(1,1,2,2,3,3,4,4,4-nonafluorobutoxy)1-butanesulfonic acid 4:2 chlorinated polyfluorinated ether sulfonic acid (2-(4-C hloro-1,1,2,2,3,3,4,4octafluorobutoxy)-1,1,2,2tetrafluoroethanesulfonic acid) 4:4 C 8 PFESA 1237850- 23-5 4:2 C l-PFESA 73772896-0 - 6:2 chlorinated polyfluorinated ether sulfonic acid 6:2 C l-PFESA 75642658-1 - C 8 Po-PFEC A - 143 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS group PASFs No. of C PFAS (2-[(6-C hloro-1,1,2,2,3,3,4,4,5,5,6,6dodecafluorohexyl)oxy]-1,1,2,2tetrafluoroethanesulfonic acid) 8:2 chlorinated polyfluorinated ether sulfonic acid (2-[(8-C hloro1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8hexadecafluorooctyl)oxy]-1,1,2,2tetrafluoroethanesulfonic acid) Mixture of 6:2 C l-PFESA and 8:2 C lPFESA Perfluoroethylcyclohexane sulfonate CAS No EC No 8:2 C l-PFESA; PFESA-BP2, Nafion Byproduct 2 76305192-9 - F-53B 7360619-6 646-833; 335-24-0 (K + salt); 6758442-3 (isomer to K + salt); 113530834-7 (isomer to K + salt) PFEC HS Perfluoroalkane sulfonyl fluorides Perfluorooctanesulfonyl fluoride n:2 FTOHs Abbreviation POSF 307-35-7 206-2006 6:2 Fluorotelomer alcohol 6:2 FTOH 647-42-7 8:2 Fluorotelomer alcohol 8:2 FTOH 678-39-7 211-4771 211-6480 6:2 Fluorotelomer sulfonic acid Ammonium Perfluorohexylethylsulfonate 6:2 FTSA 2761997-2 5958739-2 (NH4+ salt) 248-5806 8:2 Fluorotelomer sulfonic acid 8:2 FTSA 3910834-4 254-2958 - 8:2 diPAP 5767803-2 678-41-1 Perfluorooctane sulfonamide PFOSA, FOSA 754-91-6 N-ethyl perfluorooctane sulfonamide EtFOSA 4151-50- 212-0460 223-980- n:2 Fluorotelomer alcohols n:2 FTSAs n:2 Fluorotelomer sulfonic acids n:2 PAPs n:2 Polyfluoroalkyl phosphoric acid esters (PAPs) 8:2 Fluorotelomer phosphate monoester 8:2 Fluorotelomer phosphate diester FASAs 8:2 monoPAP 211-6496 Perfluoroalkane sulfonamides 144 ANNEX XV RESTRICTION REPORT - Per- and polyfluoroalkyl substances (PFASs) PFAS group No. of C PFAS N-Methyl perfluorooctane sulfonamidoacetic acid Abbreviation CAS No 2 EC No 3 (N)MeFOSAA, Me-PFOSAAcOH 2355-319 - 2991-506 1691-992 221-0611 216-8874 FC-807 3038198-7 250-1665 Hostinert 216 - 403-0505 N-Ethyl perfluorooctane (N)EtFOSAA, Etsulfonamidoacetic acid PFOSA-AcOH N-ethyl EtFOSE perfluoroocta nesulfona midoetha nol SAmPAPs Perfluoroalkane sulfonamidoethanol phosphate esters PFE alkanes Ammonium bis[2-[N-ethyl (heptadecafluorooctane) sulfonylamino]ethyl]phosphate Perfluoroether alkanes A 3:1 mixture of perfluoro(5,8,9,12tetramethyl-4,7,10,13tetraoxahexadecane) and perfluoro(5,6,9,12-tetramethyl4,7,10,13-tetraoxahexadecane) PFE alkenes Perfluoroether alkenes 1,1,2-Trifluoro-2(trifluoromethoxy)ethene 1,1,1,2,2,3,3-heptafluoro-3[(trifluorovinyl)oxy]propane perfluoropropylvinylether 1,1,2-Trifluoro-2(pentafluoroethoxy)ethene 1,1,2,2,3,3-hexafluoro-1trifluoromethoxy-3trifluorovinyloxypropane 1[difluoro(trifluoronnethoxy)nnethoxy]1,2,2-trifluoroethylene Hexafluoropropylene Other PFEAS (3:1mixture of 404-7105 and 404-7304) PMVE 1187-935 1623-058 214-7037 216-6002 1049343-3 4057309-9 234-0187 442-3909 Move3 70087487-9 - HFO-1216 116-15-4 204-1274 FC-3284; PF5052 PF-310 382-28-5 5949372-0 206-8411 407-4008 11228177-3 407-7606 61-2 473-3907 PPVE PEVE Mv31 Other ether-based PFASs 2,2,3,3,5,5,6,6-octafluoro-4(trifluoromethyl)morpholine 1-[3-[4-((heptadecafluorononyl)oxy)benzamido]propyl]-N,N,Ntrimethylammonium iodide 1-[2-(2,4-Dichlorophenyl)-3-(1,1,2,2tetrafluoroethoxy)propyl]-1H-1,2,4triazole reaction mass of 2,2,3,3,5,5,6,6octafluoro-4-(1,1,1,2,3,3,3heptafluoropropan-2-yl)morpholine Tetraconazole FC- 145 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS group HFCs No. of C PFAS Abbreviation and 2,2,3,3,5,5,6,6-octafluoro-4(heptafluoropropyl)morpholine 1-[3,5-dichloro-2-fluoro-4Noviflumuron (1,1,2,3,3,3hexafluoropropoxy)phenyl]-3-(2,6difluorobenzoyl)urea 2,3,3,4,4-pentafluoro-2,5bis(1,1,1,2,3,3,3-heptafluoropropan-2yl)-5-methoxytetrahydrofuran EC No 12145102-3 - 95720918-6 - 206-5578 212-3770 207-0792 425-3201 207-0766 420-6408; - Hydrofluorocarbons C2 HFEs CAS No Pentafluoroethane HFC -125 354-33-6 1,1,1,2-Tetrafluoroethane 811-97-2 1,1,1,2,3,3,3-heptafluoropropane HFC -134a; Norflurane HFC -227ea 1,1,1,3,3,3-hexafluoropropane HFC -236fa 690-39-1 1,1,1,2,3,3-hexafluoropropane HFC -236ea 431-63-0 reaction mass of (R, R)1,1,1,2,2,3,4,5,5,5-decafluoropentane and (S, S)-1,1,1,2,2,3,4,5,5,5decafluoropentane 1,1,1,2,2,3,3,4,4,5,5,6,6tridecafluorohexane 1,1,1-trifluoroethane HFC -4310mee HFC -5213 14234707-7; 13849542-8 355-37-3 HFC -143a 420-46-2 1,1,1,3,3-pentafluoropropane HFC -245fa 460-73-1 1,1,1,3,3-pentafluorobutane HFC -365mfc 406-58-6 1,1,1,2,2,3,3,4,4,5,5,6,6tridecafluorooctane 1,1,2,2,3,3,4-heptafluorocyclopentane HFC -7613; C 6ethane HFC PA; ZEORORA 8079317-5 1529077-4 1,1,1,2,2,3,3,4,4-Nonafluoro-4methoxybutane - 2[difluoro(methoxy)methyl]1,1,1,2,3,3,3-heptafluoropropane (1:1) 1-ethoxy-1,1,2,2,3,3,4,4,4nonafluorobutane; reaction mass of: 1-ethoxy-1,1,2,3,3,3-hexafluoro-2(trifluoromethyl)propane 1,1,1,2,2,3,4,5,5,5-Decafluoro-3methoxy-4-(trifluoromethyl)pentane HFE-7100 - 422-2702 HFE-7200 - 425-3400 HFE-7300 13218292-4 459-5205 3-Ethoxyperfluoro(2-methylhexane) HFE-7500 29773093-9 435-7901 3-(Difluoromethoxy)-1,1,2,2tetrafluoropropane HFE-356pcf3 3504299-0 - 431-89-0 206-5819 206-9965 419-1706 430-2501 430-7101 Hydrofluoroethers 146 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS group No. of C PFAS HFOs Hydrofluoroolefins Abbreviation Sevoflurane CAS No 2852386-6 EC No - 2,3,3,3-Tetrafluoro-1-propene HFO-1234yf 754-12-1 468-7107 3,3,3-trifluoropropene HFO-1243zf 677-21-4 (1E)-1,3,3,3-Tetrafluoro-1-propene HFO-1234ze(E) (2Z)-1,1,1,4,4,4-Hexafluoro-2-butene 1645-836 692-49-9 HFO1336mzz(Z) HFO667111336mzz(E) 86-2 HFO-1132a; 75-38-7 Vinylidenfluoride 211-6370 471-4800 - (2E)-1,1,1,4,4,4-Hexafluoro-2-butene 1,1-difluoroethene 200-8677 TFA PRECURSOR 2,2,2-Trifluoroethanol - 75-89-8 200-9136 Polychlorotrifluoroethylene PC TFE - Polytetrafluoroethylene PTFE Polyvinylidene fluoride PVDF Perfluoropolyether PFPE fluorotelomer-based acrylate polymer - 9002-839 9002-840 2493779-9 6999167-9 - fluorotelomer-based urethane polymer - - - Oligomeric and polymeric PFASs - i.p. = in particular From these reviews, assessments, and experimental data (peer-reviewed publications as well as study reports from industry), it can be inferred that exposure to PFASs can result in various health effects. The strength of evidence is not the same for all eff ects and all PFASs, given that not all endpoints and all PFASs have been studied extensively. EFSA (2018) and EFSA (2020) reviewed the epidemiological evidence for association between PFAS exposure and adverse effects in humans. Most of the available information covers PFOS and PFOA, but information was also available for other PFASs, in particular for PFCAs and PFSAs. Based on human data, EFSA c oncluded that there is sufficient evidence for an association between serum levels of at least PFOS and PFOA and a reduction in antibody response after vaccination, increased serum cholesterol, increased serum alanine transferase (ALT) and reduced birth weight. There was also some evidence to suggest associations with increased propensity of infections. For other outcomes investigated, there was insufficient evidence to conclude such associations. EFSA selected effects on the immune system as the critical effect in both experimental animals and humans, and derived the TWI for four PFASs based on a human study (EFSA, 2020). For PFOS, PFOA, PFNA, PFDA, and PFHpA and their salts, sufficient data for harmonised classifications for carcinogenicity (Carc. 2, except PFHpA), reproductive toxicity (Repr. 1B), effects on or via lactation (Lact., except PFHpA) and STOT RE 1 (except for PFDA) are available, see Table B.2. A harmonised classification has recently also been agreed on by RAC for 6:2 FTOH (STOT RE 2), which awaits official publication (ECHA, 2021b). 147 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Despite a lack of harmonised classification, available studies indicate similar concerns for some other PFASs. Of note, HFPO-DA, POSF, 6:2 FTSA and 8:2 FTSA have selfclassifications for STOT RE, and POSF as well for reproductive toxicity . Of further note, and supporting evidence for similarities in the toxicity profile, is that several other PFAAs and PFAA precursors have self-classifications for Carc., Repr., Lact. and STOT RE (see Table B.2 and Table B.3). A summary of the human health concerns from the available information is presented in the sections below. The Dossier Submitters applied a qualitative approach for the description of human health concerns with focus on endpoints considered most relevant for long-term exposure: repeated-dose toxicity (with targets most consistently affected by PFASs in experimental animals: liver, kidney, thyroid, immune system, and serum lipids), carcinogenicity, and toxicity to reproduction. The endpoints acute toxicity, irritation, corrosiveness, and sensitisation are not considered relevant for the human health risk assessment of this restriction proposal for PFASs. Mutagenic effects are reported only for a minority of PFASs; and hence, this endpoint is also not considered further for the PFAS hazard assessment. An update of epidemiological dat a not included in the EFSA 2020 scientific opinion is presented in B.5.3. The following sections only include statements for substances which were screened to date. 148 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5.1. Toxicokinetics/ADME (absorption, metabolism, distribution and elimination) B.5.1.1. Toxicokinetics/ADME of non-polymeric PFASs The current state of knowledge on toxicokinetics of numerous and structurally diverse PFASs relies on studies with considerable heterogeneity. Existing data have recently been thoroughly reviewed for PFCAs and PFSAs, and some precursors (ATSDR, 2021; EFSA, 2020; Fenton et al., 2021; Pizzurro et al., 2019), and for PFECAs (Rice et al., 2021) and is summarised below. Also, updated literature searches for these compounds as well as for PFASs not included in the scopes of these reviews have been performed and additional ADME data has been identified. Inter-species differences in toxicokinetics, in particular tissue distribution and elimination, should be taken into account when extrapolating animal data to human health (Pizzurro et al., 2019). B.5.1.1.1. Absorption EFSA (2020) and ATSDR (2021) have concluded that based on both experimental animal studies and human studies a range of PFASs and in particular PFCAs and PFSAs are readily absorbed upon oral exposure. PFASs are also absorbed via inhalation and dermal contact, but no quantitative estimates are currently available (ATSDR, 2021). Experimental animals Data on absorption after oral exposure in rats, mice and monkeys are available for several PFASs, e.g. PFCAs (PFBA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA and PFTeDA), for some PFSAs (PFBS, PFHxS, PFOS), for two PFECAs (HFPO-DA, ADONA). Fractional absorption ranges from >50% for PFBS to >95% for PFHxS, PFOA, PFBA, PFNA, PFDA, PFUnDA, and PFDoDA (ATSDR, 2021). PFPAs and PFPiAs are also absorbed into the bloodstream of rats, even high molecular weight compounds such as C6/C12 PFPiA and C8/C10 PFPiA (1002 amu) (D'Eon and Mabury, 2010; Joudan et al., 2017). Absorption halflives were 2.7 ± 0.5 h for C6/C8 PFPiA, 2.1 ± 1.2 h for C8 PFPA, and 1.3 ± 0.4 h for C8/C8 PFPiA (Joudan et al., 2017). The fluorotelomer alcohol 8:2 FTOH is rapidly absorbed at rates of 27-57% in rats (EFSA, 2020). Also for Noviflumuron (a complex ether-based PFAS) high absorption rates of 74-93% of the orally administered dose have been reported in rats (TERC, 2003). PFASs can also be absorbed via inhalation (evidence for PFOA in rats) or dermal contact (evidence for PFBA, PFOA, and PFOS in rabbits and rodents), but no quantitative estimates on the fractional absorption in animals were identified (ATSDR, 2021; Weatherly et al., 2021). Humans EFSA (2020) concluded that based on observations of elevated levels of PFCAs and PFSAs in humans exposed to contaminated water the gastrointestinal absorption of these PFASs occurs to a significant degree. For example, adults exposed to specific PFASs via contaminated drinking water in Italy and Sweden had elevated serum concentrations of PFOA, PFOS and PFHxA when compared to the general population (Ingelido et al., 2018; Shi et al., 2016). Furthermore, elevated concentrations of 6:2 Cl-PFESA and 8:2 Cl-PFESA were observed in high fish consumers from China, indicating high oral absorption also of these two replacement PFASs (Shi et al., 2016), demonstrating similarities to already restricted PFASs such as PFOA and PFOS. Data on blood concentrations of a range of PFASs are described in detail in section B.9.21. Absorption by inhalation and dermal contact is also based on indirect evidence. For example, in adults occupationally exposed to PFASs during fluorochemical production (primarily exposure via inhalation), elevated blood concentrations were observed, with 149 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) median PFOA, PFOS and PFHxS concentrations in the range 500 – 7 000 ng/mL (Fromme et al., 2009; Olsen, 2015). Also, studies on professional ski waxers reported elevated blood concentrations of a range of PFCAs (Freberg et al., 2010; Nilsson et al., 2013b; Nilsson et al., 2010b), probably because of inhalation of very high levels of PFASs measured in air in waxing cabins (Nilsson et al., 2010a). Fluorinated gases are absorbed via inhalation at low rates. For HFCs, HFC-134a, HFC-143a, HFC-152a, HFC-145fa (Ernstgård et al., 2009; Ernstgård et al., 2012a; Gunnare et al., 2006; Gunnare et al., 2007) absorption rates from these studies have been summarised to be <4% (Ernstgård et al., 2012b). B.5.1.1.2. Distribution As concluded by EFSA (2020) and ATSDR (2021), and based on data from both animal and human studies, PFCAs and PFSAs are distributed widely in the body and the highest concentrations have been observed in the protein-rich tissues liver, kidneys, and blood. Similar distribution patterns have been reported for some other PFASs, while different patterns have also been shown for certain PFASs. PFAS protein binding in different animal species is detailed in B.4.2.9. PFCAs and PFSAs bind in particular to serum albumin and some intracellular proteins including liver fatty acid binding protein (L-FABP) with implications for blood and tissue distribution. Experimental animals For ADONA (Rice et al., 2021), PFPAs and PFPiAs (D'Eon and Mabury, 2010; Joudan et al., 2017), wide distribution to protein-rich tissues have been observed, similar to the distribution of PFCAs and PFSAs (ATSDR, 2021; EFSA, 2020). 8:2 FTOH distributes rapidly to blood and tissues in rats with highest levels of 8:2 FTOH in fat, liver, thyroid and adrenals (EFSA, 2020; Fasano et al., 2006). Noviflumuron (a complex ether-based PFAS) is distributed to organs in the order of fat >> adrenal = skin > ovaries followed by spleen and liver (TERC, 2004). For Noviflumuron, there was only limited transplacental transfer in rats (TERC, 2005). 6:2 FTSA, but not 6:2 FTCA, was detected at high levels in serum and liver following repeated-dose exposure (Sheng et al., 2017), whereas 6:2 Cl-PFESA has been shown to distribute to serum, gut (Pan et al., 2019b) and liver (Pan et al., 2021). Similarly, sodium ρ-perfluorous nonenoxybenzene sulfonate (OBS) distribute to mouse liver, but also to the gut (colon, ileum). At the highest dose, OBS was also detected in serum, kidney and faeces (Wang et al., 2019a). In addition to organ distribution, PFCAs and PFSAs can be distributed to breast milk and cross the placenta and can thus be transferred from the dams to the foetus, resulting in maternal excretion of these compounds (see B.5.1.1.4) and exposure of the foetus/infant (ATSDR, 2021; Bartels et al., 2020; EFSA, 2020). Humans In humans, very few studies on organ distribution have been published, likely because of limited sample availability. Levels of PFASs measured in human organs and other human samples relevant for distribution are presented in section B.9.22 on human biomonitoring and Appendix B.5.1.1.2.. In three studies of PFASs in human liver tissues, PFOS was the most abundant PFAS at mean levels ranging from 14 – 27 ng/g (Karrman et al., 2010; Maestri et al., 2006; Olsen et al., 2003). PFBA detected at high levels in lung autopsy tissue in Spain by Perez et al. (2013) (median value of 807 ng/g lung tissue). Abraham et al. (2021) questioned PFBA accumulation in lung tissue as they found low or non-detectable levels in lung autopsy 150 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) samples from France. This is more in line with what is expected based on low bioaccumulation potential and currently estimated short biological half -life of PFBA (Chang et al., 2008). A study by Wang et al. (2018a) demonstrated that several PFASs can cross the cerebrospinal fluid barrier. Detectable levels were observed in more than half of the cerebrospinal fluid samples for PFOA, PFNA, PFDA, PFHxS, PFOS and 6:2 Cl-PFESA. These compounds were also the most prominent compounds in the serum samples. This study supports a previous study (Fujii et al., 2015) that show the presence of PFASs in cerebrospinal fluid, but at significantly lower concentrations than in serum presumably due to the blood brain barrier in adults. PFASs can cross the blood-follicle barrier and the resulting concentration in follicular fluid indicates exposure to maturing oocytes that grow in the ovarian follicle. The median bloodfollicle transfer efficiencies (ratio of follicular fluid:serum) of PF AAs measured in four different studies ranged from 0.47 (PFDoDA) to 1.04 (PFHxS) (Hallberg et al., 2021; Heffernan et al., 2018; Kang et al., 2020; McCoy et al., 2017). The follicle transfer efficiencies for Cl-PFESAs have been reported to be 0.73, 0.75 and 0.91 for 4:2 Cl-PFESA, 6:2 Cl-PFESA and 8:2 Cl-PFESA, respectively, demonstrating similar properties for replacement PFASs as for already restricted PFASs such as PFOS and PFOA (Kang et al., 2020). Other investigated PFASs, e.g. 4:4 C8 PFESA, 4:4 C8 PFECA, C8 Polyether PFECA as well as 6:2 Cl-PFESA and PFECHS also had blood-follicle transfer efficiencies in the range from 0.72 to 0.94 (Hallberg et al., 2021; Kang et al., 2020). Strong correlations between PFASs in serum/plasma and follicular fluid indicate that serum/plasma is a good proxy for determining the oocyte exposure levels. PFCAs and PFSAs as well as replacement PFASs such as 6:2 Cl-PFESA and 8:2 Cl-PFESA, can readily cross the placenta and thus be transferred to the foetus (see section B.9.21 on human biomonitoring). A recent study on human embryonic and foetal organs found that concentrations in embryo/foetal tissue were lower than maternal serum and similar to placenta levels. The concentration of the sum of five PFASs (PFOS, PFOA, PFNA, PFDA and PFUnDA) was highest in lung and liver tissue and lowest in CNS samples (Mamsen et al., 2019). B.5.1.1.3. Metabolism Several PFASs have been shown to be excreted untransformed, i.e. without forming any metabolites or conjugates while others are readily metabolised, often to arrowhead PFASs and will thus contribute to exposure to these compounds (ATSDR, 2021; EFSA, 2020). Experimental animals No metabolism has been observed or is expected for PFCAs and PFSAs (ATSDR, 2021; EFSA, 2020). Similarly, no metabolism has been reported for PFPAs (Joudan et al., 2017) and 6:2 Cl-PFESA (Wang et al., 2013b), as well as for PFECAs HFPO-DA, ADONA, and EEA (Rice et al., 2021). These substances are thus considered metabolically inert and stable end-stage products. In contrast, studies on experimental animals have demonstrated that other PFAS groups such as FTOHs, PAPs and FASAs are transformed to the arrowhead groups PFCAs and PFSAs (EFSA, 2020). 6:2 FTOH is also shown to be metabolised in rats, the primary stable metabolite being 5:3 fluorotelomer carboxylic acid (5:3 FTCA), but also PFCAs, such as PFBA, PFPeA, PFHxA, and PFHpA (Kabadi et al., 2018; Kabadi et al., 2020; Ruan et al., 2014). In addition, fluoride appears to be released during the metabolism (Mukerji et al., 2015). 8:2 FTOH metabolises to glucuronide and glutathione conjugates of the parent compound, oxidised and reduced intermediates and PFOA, PFNA, PFHpA and PFHxA (EFSA, 2020). In rats, PFPiAs are extensively metabolised by cleavage of one C–P bond yielding the corresponding PFPA and 1H-perfluoroalkanes (Joudan et al., 2017). In a mouse study with exposure to aqueous film-forming foam (AFFF) PFAS mixture of C6 151 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) and C7 perfluoroalkane sulfonates (PFSAs) were enriched in mouse serum, suggesting in vivo transformation of sulfonamide precursors. Some substituted C8 PFSAs [ketoperfluorooctane sulfonate (PFOS), hydrogen-PFOS, and unsaturated PFOS] appeared to be more bioaccumulative than linear PFOS or were formed in vivo from unidentified precursors. Sulfonimide dimers were detected in serum that may have been a minor component of the AFFF or formed via metabolism (McDonough et al., 2020a). Some metabolism was reported for Noviflumuron in rats (>5% of administered dose): urinary metabolites included a glucuronide, sulphate conjugates of hexafluoroalkoxyfluorodichloroaniline and a mercapturic acid conjugate of parent compound, but 84-100% was parent test material (TERC, 2003). Exposure to HFO-1234yf led to low extent (<0.1% of dose received) of biotransformation in rabbits (Schuster et al., 2010). Urinary metabolites included N-acetyl-S-(3,3,3-trifluoro2-hydroxypropanyl)-l-cysteine (predominant metabolite), S-(3,3,3-Trifluoro-2hydroxypropanyl)mercaptolactic acid, 3,3,3-trifluoro-1,2-dihydroxypropane, 3,3,3trifluoro-2-propanol, and inorganic fluoride (Schuster et al., 2010). Humans Similar to experimental animals, no metabolism has been observed for PFCAs and PFSAs (EFSA, 2020), and this is likely also the case for other PFASs that are not metabolised in experimental animals. It is shown that humans are able to transform precursors to PFCAs and PFSAs, similar to experimental animals (EFSA, 2020). For example, 8:2 FTOH is transformed to FTCAs and FTUCAs and further to PFOA and PFNA (EFSA, 2020; Nilsson et al., 2013a). For HFCs, no urinary metabolites were detected after exposure to HFC-134a (Gunnare et al., 2006), HFC-143a (Gunnare et al., 2007), HFC152a (Ernstgård et al., 2012a), or HFC245fa (Ernstgård et al., 2009), summarized by Ernstgård et al. (2012b). B.5.1.1.4. Excretion As summarised by EFSA (2020) and ATSDR (2021) a range of PFASs, and in particular PFCAs and PFSASs, are eliminated both through urine and via faeces. In addition, excretion through transplacental transfer, menstruation and breastfeeding has been demonstrated. In humans, long elimination half-lives have been observed for several PFASs, while other PFASs are excreted faster and do thus have shorter elimination half-lives. Experimental animals Urine is the main excretion route for PFCAs with less than ten carbon atoms (C <10) (EFSA, 2020), while for PFCAs with ten or more carbon atoms, biliary excretion and subsequent excretion via faeces is the main elimination route (EFSA, 2020). Elimination of PFCAs in rats is faster in females than in males (EFSA, 2020). PFSAs (only data available for C ≤ 8) are primarily eliminated in urine, and to a lesser extent in faeces (EFSA, 2020). In rats, 8:2 FTOH is mainly eliminated via faeces (70%), to a lesser extent through biliary excretion (20-45%), and by less than 4% via urine (EFSA, 2020). For Noviflumuron (complex etherbased PFAS), 63% of the dose was eliminated via faeces (cumulatively) in rats and 5390% of the was recovered in the faeces within 72 h (TERC, 2002; TERC, 2004). Urine is considered a minor excretion route for Noviflumuron, as only 0.7-5.0% was recovered in the urine independent of the dose (TERC, 2002) and the total cumulative urinary elimination was about 12% of the dose (TERC, 2004). Breast milk also represents a significant route of excretion for Noviflumuron in rats (TERC, 2005). For fluorinated gases, metabolites of HFO-1234yf were mostly excreted in urine within 12 h after the end of exposure in rabbits (t 1/2 of ca. 9.5 h) (Schuster et al., 2010). 152 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Serum half-lives of several PFCAs and PFSAs in experimental animals have been summarised by EFSA (2020), ATSDR (2021) and Fenton et al. (2021). In rats, serum halflives from less than an hour or few hours for PFHxA and PFBS up to more than 100 days for PFDA and one of the isomers of PFOS (1m-PFOS) have been reported. In mice, halflives between 3 h (PFBA, females) and 228 days (PFNA, males) have been reported (EFSA, 2020). In male rats, the blood half-life was 0.95 h for C8 PFPA and 2.8 h for C8/C8 PFPiA (Joudan et al., 2017). A recent study on CD-1 mice reported serum half-lives of PFBS of 5.8 (M) and 4.5 (F) h, respectively (Lau et al., 2020). Large sex differences in elimination half-lives have been observed in rats, while the differences are smaller in mice (ATSDR, 2021; EFSA, 2020; Fenton et al., 2021). The 6:2 FTOH metabolite, 5:3 FTCA is estimated to have plasma half-lives of 64 (m) – 67 (f) days in rats, and under conditions of repeated oral 6:2 FTOH exposure it would take approximately one year for the 5:3 FTCA to reach steady state (Kabadi et al., 2020). The half-life of 8:2 FTOH was 7 – 9 days in rats (EFSA, 2020; Fasano et al., 2006). Humans Detailed data on PFAS concentrations in urine, breast milk, placenta and cord blood have been summarised in section B.9.21 and Appendix B.5.1.1.2. Table B.91 to Table B.95. In humans, short chain PFCAs, are mostly excreted in urine, whereas PFNA and longer chain PFASs are preferentially eliminated through the bile (EFSA, 2020). 6:2 Cl-PFESA and 8:2 Cl-PFESA have also been detected in urine, demonstrating urinary excretion (Shi et al., 2016). 6:2 Cl-PFESA was not detected in urine from USA (Calafat et al., 2019; Kato et al., 2018), likely due to low human exposure to these compounds. Breast milk also represents a significant route of excretion for PFCAs, and PFSAs (ATSDR, 2021; EFSA, 2020). Replacement PFASs such as 6:2 Cl-PFESA and 8:2 Cl-PFESA have also been detected in relatively high concentrations in breast milk in China, indicating similarities with restricted PFASs such as PFOS and PFOA (Awad et al., 2020; Jin et al., 2020a). In contrast, 6:2 Cl-PFESA and 8:2 Cl-PFESA were not detected in breast milk from Sweden and Czech Republic (Awad et al., 2020; Cerna et al., 2020), likely because of low human exposure to these compounds. Other routes of elimination are transplacental transfer and blood loss during menstruation as summarised for PFCAs and PFSAs by EFSA (2020) and ATSDR (2021). Median ratios between foetal:maternal blood are in the range of 0.36-0.74 for PFHxS, PFOS, PFOA, PFNA and 6:2 Cl-PFESA (Chen et al., 2017b; EFSA, 2020). The transfer rates probably depend on the structure of the compound, where longer fluoroalkyl chain length and a terminal sulfonate group are associated with a lower foetal:maternal blood ratio (EFSA, 2020). Transplacental transfer, menstruation and breastfeeding may at least partly explain gender differences in levels of PFASs between men and women (EFSA, 2020). For 6:2 Cl-PFESA and 8:2 Cl-PFESA transplacental transfer has also been demonstrated, and both compounds were detected both in placenta samples as well as in cord blood (Cai et al., 2020; Chen et al., 2017b; Gao et al., 2019; Lu et al., 2021). Again, this shows similarities between these replacement PFASs and the restricted PFASs such as PFOS and PFOA. Total elimination half-lives of PFCAs and PFSAs in humans range from a few days (PFBA) to several years (for example PFOS, PFOA and PFHxS) (ATSDR, 2021; EFSA, 2020; Fenton et al., 2021). In a study by Li et al. (2018c), marked sex differences in elimination halflives were observed for PFHxS and PFOS, with more rapid elimination in women compared to men, while the difference between sexes was only marginal for PFOA (Li et al., 2018c). The long human elimination half-lives of long-chain PFASs are mainly attributed to active renal and intestinal reabsorption via organic anion transporters (OATs) (Ducatman et al., 2021; EFSA, 2020; Fenton et al., 2021). PFOA and PFOS have been shown to have a high affinity for human transport proteins such as OAT1, OAT3 and urate transporter URAT1 (EFSA, 2020), contributing to their long half-lives in humans. 153 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Differences in renal reabsorption is considered a main reason for variability in the half -lives of PFASs between species. Serum albumin binding is also important for the long half -life of many PFASs as it limits renal excretion due to less free PFASs in the circulation (Fenton et al., 2021). In a study including individuals with high exposure to 6:2 Cl-PFESA, a median elimination half-life of 15.3 years was reported, again indicating similarities between these replacement PFASs and the restricted PFASs such as PFOS and PFOA (Shi et al., 2016). It may be suspected that PFPA and PFPiA have longer half-lives in humans than in rats, though predictions are complicated by the metabolism of PFPiAs (Joudan et al., 2017). B.5.1.2. Toxicokinetics/ADME of polymeric PFASs ADME properties of oligomeric and polymeric PFASs can vary widely. The chemical composition of the principal monomers as well as the identification of the oligomeric/polymeric substance on the basis of general chemical descriptors, such as relevant names and numbers (such as CAS- and/or EC-no., chemical name or trade names) alone are not sufficient for human health assessment. Oligomers/polymers can vary in terms of molecular weight distribution, physical state, and possible inclusion of comonomers and others, but can carry the same name and/or CAS number. Also, additives and non-intentionally added substances (NIAS) can play a relevant role in the final oligomeric/polymeric product. For further details see ECETOC (2019) and the following box. From ECETOC (2019): “Structural and morphological descriptors and/or physical, and chemical properties Depending on the type of polymer under investigation, relevant key parameters may be structural and/or morphological descriptors as well as physico -chemical and screening-level fate properties (no order of properties is inferred): ­ ­ ­ Structural descriptors include e.g. chemical formula, degree of substitution, tacticity, Mw, Mw distribution (polydispersity), number average molecular weight (NAMW), and RFG(s); Morphological descriptors include e.g. physical state at ambient temperature and pressure (solid, liquid), shape (e.g. spherical, fibre, tubular), physical form (e.g. amorphous, crystalline); Physico-chemical properties include e.g. water solubility, n-octanol/water partition coefficient (log Pow), acid dissociation constant (pKa), net charge (under conditions that are relevant for ecological and human health hazard assessment), vapour pressure, viscosity / melt -flow index / glass transition temperature, density, degradability.” Only a few studies with toxicokinetic/ADME information are available for this diverse group of oligomeric/polymeric PFASs. No studies are available on toxicokinetics of polymeric PFASs. All available studies on toxicokinetics/ADME studied oligomeric PCTFE (oils and/or pure oligomers). Under the REACH regulation, oligomers are not defined and do not fall under the polymer definition according to REACH (Article 5(3)). Oligomers are composed of a small number of linked monomer units and thus differ from non-polymeric PFASs. Most publications on PCTFE oligomer ADME and toxicity provide insufficient information on the identification of the substance or study details and results investigated. Some evidence for absorption of PCTFE oligomers was reported for experimental animals (rodents and rhesus monkey) into blood, liver, and kidney, via repeated exposure to vapour via inhalation (Kinkead et al., 1989; Kinkead et al., 1987), and after oral administration via the GI tract (DelRaso et al., 1991; Jones et al., 1991; Kinkead et al., 154 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) 1987). Absorption via the skin cannot be excluded, because small insignificant increases of plasma fluoride concentrations after dermal absorption of PCTFE oligomers were shown in rodent urine and plasma (Kinkead et al., 1987). There is evidence of distribution of PCTFE oligomers, after repeated inhalation into various tissues, such as kidney, liver, lung, testes, brain and fat (Kinkead et al., 1989; Vinegar et al., 1992). After oral administration, concentrations of pure PCTFE tetramers in the liver of rats were significantly higher compared to PCTFE halocarbon oil and to pure PCTFE trimers (DelRaso et al., 1991). Furthermore, composition of the ratio percentage of trimer to tetramer present in PCTFE HC 3.1 oil-C6:C8 (55:45) was found to be altered when measured in the liver (32:68) (DelRaso et al., 1991). Data on metabolism are scarce. PCTFE oligomers were reported to be metabolised to carboxylic acid derivatives in rats based on physiologically based modelling and measurements of decreasing trimer and tetramer concentrations in blood and liver and release of fluoride (Vinegar et al., 1992). PCTFE oligomers and/or unbound fluoride ions were excreted in urine after exposure to PCTFE oligomers (halocarbon oils of different compositions) via inhalation, dermal contact and oral uptake (Kinkead et al., 1989; Kinkead et al., 1987; Vinegar et al., 1992). Because of the lack of data, it is not possible to conclude on ADME characteristics of oligomeric or polymeric PFASs. However, considering the few available toxicity data (B.5.2.1) ADME characteristics that allow interactions with biota cannot be excluded for oligomeric/polymeric PFASs. 155 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5.2. Evidence from experimental animal data B.5.2.1. Repeated dose toxicity (experimental animal data) With respect to the large array of potential health effects and the vast number of different PFASs, this section focuses on the most prominent advers e effects that occur most consistently across different groups of PFASs. These are effects on liver, kidney, thyroid, immune system, and serum lipids. Toxicity to reproduction is covered in a separate section (B.5.2.2). For oligomeric/polymeric PFASs, only few experimental studies are available. Predominantly oligomeric PCTFE oils were studied, some of which show effects dependent on chain length and mixture-ratio of different chain lengths (DelRaso et al., 1991; Kinkead et al., 1991). However, differences of oligomer/polymer properties and lack of data do not allow an in-depth assessment of the human health hazards. Thus, the RDT effects in this section give only an exemplary insight into possible health impact of oligomeric/polymeric PFASs. For side-chain fluorinated polymers (SCFPs) it is known that the PFAS moieties on the side chains can separate from the carbon backbone of at least some commercial SCFPs over time, and may further degrade to PFCAs and PFSAs in the environment and biota (OECD, 2022). Since most PFCAs and PFSAs were shown to cause adverse effects in humans and animals, it can be concluded that SCFPs contribute to adverse health effects, e.g. in o rgans such as liver, kidney, thyroid, immunological tissues after repeated exposure. No further toxicity data are available for SCFPs and will not be addressed in the following section on specific effects. B.5.2.1.1. Liver effects in experimental animals Various reports and reviews demonstrate that the liver is a sensitive and one of the most consistently affected targets of PFAS toxicity in experimental animals (ATSDR, 2021; EFSA, 2020; Fenton et al., 2021). A recent review (Fenton et al., 2021) summarised liver effects for humans and animal studies and concluded that the liver is a primary target organ of PFASs, in particular PFCAs with C ≥ 8 and PFSAs with C ≥ 6. Effects on the liver include specific P450 (CYP) pathway induction, significantly increased liver weight, hepatic steatosis, apoptosis, hepatocellular adenomas and carcinomas, and disrupted fatty acid metabolism that can be peroxisome proliferator–activated receptor alpha (PPARα)– dependent or –independent and present across species (Cui et al., 2009; Filgo et al., 2015; Huang et al., 2013; Hui et al., 2017; Li et al., 2017; Maestri et al., 2006; NTP, 2020; Perez et al., 2013; Wan et al., 2012; Xu et al., 2016; Xu et al., 2020b; Yao et al., 2016; Zhang et al., 2016b). This section summarises the evidence from animal experiments. Epidemiological evidence for liver toxicity was discussed by EFSA (2020) and an update to EFSA (2020) is provided in section B.5.3.1.3. Non-polymeric PFASs Among all studied non-polymeric PFASs, the most consistent effect in the available animal studies is a dose-dependent increase in liver weight, mostly accompanied with hepatocellular hypertrophy and increases in liver enzymes (ALT, AST, ALP) indicative for liver cell dysfunction, for example in rodent models repeatedly exposed to     PFCAs, PFSAs (numerous studies reviewed by (ATSDR, 2021; EFSA, 2020; NTP, 2019a; NTP, 2019b), TFA (BayerCropScience, 2007; Blake, 1970; Just et al., 1989), Perfluamine (Triskelion, 2019b), PFECAs o HFPO-DA (e.g. (Rushing et al., 2017; WIL, 2008a; WIL, 2008b; WIL, 2009)), o ADONA (e.g. (Gordon, 2011; NOTOX, 2007a)), 156 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs)        o Mv31 K+ salt (Aventis Pharma Deutschland GmbH, 2002), o F-DIOX (RTC, 2013), PFESAs o 6:2 Cl-PFESA (Zhang et al., 2018a), o PFESA-BP2 (Lang et al., 2020), PFE alkenes o PMVE (WIL, 2016), o Mv31 (NOTOX, 2007b), complex ether-based PFASs, e.g. o Tetraconazole (Huntingdon, N/A-b; Huntingdon, N/A-d), o Noviflumuron (Dow AgroSciences, 2002a), Fluorotelomers o 8:2 FTOH ((Ladics et al., 2008; Wang et al., 2019b), reviewed by EFSA (2020)), o EtFOSE ((Xie et al., 2009), reviewed by EFSA (2020)), o 6:2 FTSA (6: 2 fluorotelomer sulfonic acid (Sheng et al., 2017), HFCs o HFC-134a (ICI, 1979; Zeneca, 1993), o HFC-245fa (Frauenhofer Institut für Toxikologie und experim. Medizin, 2005), o HFC-365mfc (TNO, N/A), o HFC-4310mee (DuPont Haskell, 2007b; Haskell, 1996b), o HFC-52-13 (Hita, 1994), HFC-76-13 (Hita, 2007a), o HFCPA (Huntingdon, 1998a; Huntingdon, 1998c; Huntingdon, 1998d), HFOs o HFO-1234yf (Tveit et al., 2013), o HFO-1234ze(E) (TNO, 2006), o HFO-1336mzz(Z) (DuPont Haskell, 2014), o HFO-1132a (Litton Bionetics, 1984), HFEs o HFE-7100 (Huntingdon, 1995; Huntingdon, 1996a; Huntingdon, 1996b), o HFE-7200 (Huntingdon, 1997), o HFE-7300 (Charles River, 2019b; Hita, 2004), o HFE-7500 (3M, 1998), o HFE-356pcf3 (DuPont Haskell, 2009; DuPont Haskell, 2011). Moreover, there are indications for hepatocellular necrosis from experimental rodent models after repeated exposure (28 or 90 days) for       most PFCAs and PFSAs, except PFBS and PFHxDA (ATSDR, 2021; EFSA, 2020), several PFECAs, e.g. o HFPO-DA (Blake et al., 2020; Caverly Rae et al., 2015; DuPont Haskell, 2008a; Haskell, 2010; MPI Research Inc., 2013; Sheng et al., 2018b; Wang et al., 2017a; WIL, 2008a; WIL, 2010a; WIL, 2011a), o HFPO-TA (Sheng et al., 2018b), o EEA-NH4 (Hita, 2006), o F-DIOX (RTC, 2010; RTC, 2011; RTC, 2012), PFESAs, e.g. o 6:2 Cl-PFESA (Zhang et al., 2018a), complex ether-based PFASs, .e.g. o Tetraconazole (Huntingdon, N/A-c), o Noviflumuron (Dow AgroSciences, 2002b), Fluorotelomers, e.g. o 6:2 FTOH (Hita, 2007b; Mukerji et al., 2015; Rice et al., 2020; Serex et al., 2014; WIL, 2005), o 8:2 FTOH (Ladics et al., 2008), and o 6:2 FTSA (Sheng et al., 2017), and HFE-7300 (Hita, 2004). 157 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Examples of PFASs that did not induce hepatotoxicity in animal studies are several perfluoroether-alkenes and -alkanes as well as some complex perfluoroether alkylic substances. Within fluorinated gases, hydrofluoroethers generally show typical increased liver weight and hepatocellular hypertrophy, but most hydrofluorocarbons and hydrofluoroolefins only increased liver weights, mainly without histopathological changes, e.g. HFC-134a (ICI, 1979), HFC-52-13 (Hita, 1994), HFO-1234yf (Tveit et al., 2013). For PFASs of different groups (PFCAs, PFSAs, PFECAs), it has been demonstrated that potency differences for liver effects are largely determined by kinetics (serum half -lives) (Gomis et al., 2018). The mechanisms underlying PFAS-induced hepatotoxicity has been extensively studied. As reviewed by EFSA (2020), ATSDR (2021), and Fenton et al. (2021), several PFASs lead to a transcriptional activation of mouse and human PPARα‐related genes in liver in adultexposed models. Additionally, activation of other nuclear receptors such as PPARγ, constitutive androstane rec eptor (CAR), and pregnane X-receptor (PXR) has also been reported. These nuclear receptors regulate lipid and glucose metabolism and transport as well as inflammation and some nuclear receptors, including PPARα, are considered more responsive in tissues of rodents than in humans (Fenton et al., 2021; Rosen et al., 2017; Wolf et al., 2012). Predictions of the binding mode of PFOS, PFOA and PFHxS to PPARα, PPARγ and ERα are available and show that hydrophobic interactions play a key role in the binding of these PFAAs (Jiang et al., 2015). However, ATSDR (2021) concluded that hepatic effects of PFCAs and PFSAs in rodents likely result from a combination of PPARα-dependent and independent changes, which is in line with a CLH RAC opinion for PFOA (ECHA, 2011), demonstrating that human relevance of such effects should be expected. Oligomeric/polymeric PFASs Predominantly oligomeric PFASs were studied within repeated dosed inhalation, and oral toxicity studies. For the oligomeric PCTFE (CAS-no. 9002-83-9), liver related effects were reported in rats exposed to various oligomer ratios: - - Pure PCTFE trimers (C6) and pure PCTFE tetramers (C8) (DelRaso et al., 1991; Kinkead et al., 1991) PCTFE Halocarbon (HC) oils (base oils) of different (trade) names: o S-27 oil (C8 : C10 unknown ratio) (Kinkead et al., 1990a), o Safetol®3.1/ MLO-87-124 (70% C6 : 30% C8) (Kinkead et al., 1989; Kinkead et al., 1990b), o PCTFE 3.1 (55% C6 : 45% C8 and 95% C6 : 5% C8) (DelRaso et al., 1991), o PCTFE 3.1 (C6 : C8, unknown ratio) ((Jones et al., 1991) o PCTFE 3.1 (C6 : C8, unknown ratio with additives) (Mattie et al., 1993) o PCTFE HC (70% C6 : 30% C8 with additives) (Vinegar et al., 1992) o PCTFE HC (55% C6 : 45% C8 without additives) (Vinegar et al., 1992) pure PCTFE trimers (C6) and pure PCTFE tetramers (C8) (DelRaso et al., 1991; Kinkead et al., 1991). On the basis of repeated dose oral toxicity studies with oligomeric PCTFE (DelRaso et al., 1991; Kinkead et al., 1991; Kinkead et al., 1989; Kinkead et al., 1990a; Kinkead et al., 1990b; Mattie et al., 1993), common hepatotoxic findings in rats were increased relative liver weights as well as hepatocellular cytomegaly. Effects observed in oral studies were similar to those reported for two repeated inhalation studies (Kinkead et al., 1990b; Vinegar et al., 1992). After oral administration (Mattie et al., 1993) and after inhalation (Kinkead et al., 1989) of oligomeric PCTFE oil, increases in liver enzymes (ALT, AST) indicative for liver cell dysfunction were observed in rats. Additional liver effects in rats associated with oligomeric PCTFE, differing between studies of pure tri- or tetramers and of various oil compositions, were e.g. altered hepatocellular architecture, elevated serum liver-associated enzymes, hepatocytic eosinophilic granular cytoplasm, and loss of hepatocytic cytoplasmic basophilia (DelRaso et al., 1991; Kinkead 158 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) et al., 1991). Kinkead et al. (1990a) concluded on the basis of further observations that the liver was probably the primary target organ of oligomeric PCTFE. For PCTFE oligomers in rats, repeated orally administered tetramers were observed to be more hepatotoxic than trimers (Kinkead et al., 1991; Vinegar et al., 1992). For polymeric PTFE, the only available information on repeated-dose toxicity was an older study from Zapp (1962). In this publication it was stated that a laboratory study with rats, fed with a diet of 25% finely ground PTFE resins (polymer composition and further key parameters unknown) for 90 days, showed no signs of toxic effects and no pathological changes detectable by gross or microscopic examination of the tissues. Also in a book by Sheftel (2000) that summarises data of indirect food additives and polymers, PTFE was reported to cause no effect on rats after ten months oral administration. However, insufficiently reported study details in both sources (e.g. polymer properties, study design) weaken the power of the available information. For one perfluoropolyether (PFPE) surfactant, Johnston et al. (1996b) communicated concerns in a letter to the editor, indicating that - based on unpublished data - the oligomeric PFPE surfactant cannot be assumed biologically inert, because it caused an increase in liver weight in rats. The authors state that the used PFPE surfactant was similar to the ammonium carboxylate PFPE surfactant used by Johnston et al. (1996a) (CF 3O(CF 2CF(CF3)O)3CF2-COO-NH4+ ), with an average molecular weight of 740 g/mol. For further details on the PFPE substance, Johnston et al. (1996a) cited Chittofrati et al. (1989), who described the PFPE surfactant with the formula Rf-CF 2-COO-NH4+ , with Rf described as CF 3-[(O-CF 2-CF(CF3))n-(O-CF2)m]-O- and Rf having n/m = 3 ÷ 4 monomer units with a molecular weight of 710 g/mol. There is indication for significantly increased hepatic peroxisomal ß-oxidation after oral administration of oligomeric PCTFE oils to rats (DelRaso et al., 1991; Kinkead et al., 1991; Kinkead et al., 1990a; Mattie et al., 1993). In monkeys the increase in peroxisomal ßoxidation was not significant (Jones et al., 1991). This effect was not observed in rats following inhalation of PFPE fluorinated oil (KrytoxTM) (Kelly et al., 1993). In summary, there are indications that oligomeric PFASs can cause adverse liver effects, which is generally in line with the typical effects observed for non-oligomeric/nonpolymeric PFASs. Clarity on liver effects of the highly diverse group of oligomeric/polymeric PFASs cannot be given on the basis of available data. B.5.2.1.2. Serum lipids in experimental animals Regarding non-polymeric PFASs, most animal studies addressing serum cholesterol have been conducted at much higher PFAS exposure levels than human studies (EFSA, 2020; Fragki et al., 2021). In general, studies with experimental animals mainly measured serum total cholesterol, which is rather reduced in rodents than increased as observed in epidemiological studies. For instance, total serum cholesterol was reduced after repeated doses for 28 or 90 days in male rodents by PFSAs: PFBS, PFHxS, and PFOS (NTP, 2019b); PFCAs: PFBA (Butenhoff et al., 2012a), PFHxA (Chengelis et al., 2009b), PFOA (e.g. (Loveless et al., 2006)), PFNA, PFDA (NTP, 2019a), PFDoDA (Kato et al., 2015), PFODA (Hirata-Koizumi et al., 2012). For PFECAs, HFPO-DA (Conley et al., 2021; Conley et al., 2019; DuPont Haskell, 2008b; Sheng et al., 2018b; WIL, 2009), EEA-NH4 (Hita, 2006), and F-DIOX (RTC, 2013) reduced total cholesterol. However, some studies report increases in serum cholesterol in mice exposed to human relevant PFOA-levels in combination with high fat or westernised diet (Rebholz et al., 2016; Schlezinger et al., 2020), suggesting that both exposure concentration, diet and sex may influence the effect of PFASs on lipid metabolism. Increases of total cholesterol were reported for HFPO-DA (Blake et al., 2020) and F-DIOX (RTC, 2011). An increase in total serum cholesterol was also reported for 6:2 Cl-PFESA (Zhang et al., 159 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) 2018a), Tetraconazole (Huntingdon, N/A-b; Huntingdon, N/A-f) and Noviflumoron (Dow AgroSciences, 2002a; Dow AgroSciences, 2002b; Dow AgroSciences, 2005b) . Two hydrofluoroethers (fluorinated gases) induced a decrease in total serum cholesterol in rats after repeated exposure to HFE-7100 for 28 days (Huntingdon, 1995) or HFE-7500 for 5 days (3M, 1998). Animal studies show a clear association of PFASs with changes in lipid metabolism. However, because of significant species differences affecting both lipid metabolism and PFAS toxicokinetics (Fragki et al., 2021) and often large differences in exposure levels, the animal experiments have not been able to elucidate the causality of the PFAS – serum cholesterol association demonstrated in epidemiological studies. Serum triglycerides are mostly reduced in experimental animals (rodents) after repeated exposure to  PFCAs: o PFHxA (Klaunig et al., 2015), o PFOA (DeWitt et al., 2009; Loveless et al., 2006; Qazi et al., 2010; Wu et al., 2018; Xie et al., 2003), o PFNA (NTP, 2019a; Wang et al., 2015c), o PFDA (NTP, 2019a), o PFTeDA (Hirata-Koizumi et al., 2015), o PFODA (Hirata-Koizumi et al., 2012);  PFSAs: o PFBS (Bijland et al., 2011), o PFOS (Lai et al., 2018); PFECAs: o HFPO-DA (Blake et al., 2020; Conley et al., 2021; Conley et al., 2019; DuPont Haskell, 2008b), o HFPO-TA (Sheng et al., 2018b), o mv31 K+ -salt (Aventis Pharma Deutschland GmbH, 2002), o F-DIOX (RTC, 2013); other ether PFASs: o PEVE (Haskell, 1997), and some HFCs/HFOs: o HFC-236fa (Haskell, 1996a), o HFO-1336mzz(Z) (DuPont Haskell, 2010a).    However, similar to cholesterol effects, also considerable increases in serum triglycerides are reported in some rodent studies for         PFHpA (Anonymous, 2017), PFOA (Loveless et al., 2006; Minata et al., 2010), PFDoDA (Zhang et al., 2008), PFOS (Huck et al., 2018; Su et al., 2019), 6:2 Cl-PFESA (Zhang et al., 2018a), HFPO-DA (Conley et al., 2021), F-DIOX (RTC, 2011), or HFCPA (Huntingdon, 1998c; Huntingdon, 1998d). 160 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Regarding oligomeric/polymeric PFASs, serum triglycerides were increased after repeated oral exposure to oligomeric PCTFE oil (unknown ratio) in rhesus monkeys (Jones et al., 1991). Serum bile acids are consistently increased in experimental animals after repeated exposure to some non-polymeric PFASs, e.g. PFCAs (NTP, 2019a) or PFSAs (NTP, 2019b). EFSA (2020) concluded that such an increase in serum bile acids indicates cholestasis. Increased serum bilirubin as another indicator for cholestasis was also evident for some PFCAs (NTP, 2019a) and PFSAs (NTP, 2019b). B.5.2.1.3. Kidney effects in experimental animals Non-polymeric PFASs In experimental animal models (mainly in rats, sometimes mice), kidney weights relative to body mass were increased for some PFAAs, PFECAs, perfluoroether alkenes, complex other ether-based PFASs, and a variety of fluorinated gases (HFCs including HFOs, and HFEs), underlining the variety of chemical structures of PFASs that can induce changes on kidney physiology. Epidemiological evidence for effects on the kidney are summarised by EFSA (2020) and updated in B.5.3.1.4. Among substances that induced increases in relative kidney weights in rodents are         PFCAs: o PFHxA (Chengelis et al., 2009b; Loveless et al., 2009; NTP, 2019a), o PFOA (Butenhoff et al., 2004; Cui et al., 2009; Griffith and Long, 1980; Loveless et al., 2006; Loveless et al., 2008; NTP, 2019a), o PFNA (NTP, 2019a), o PFDA (Frawley et al., 2018; NTP, 2019a), o PFDoDA (Kato et al., 2015), PFSAs: o PFBS (NTP, 2019b); PFECAs: o HFPO-DA (Blake et al., 2020; Caverly Rae et al., 2015; DuPont Haskell, 2008b; MPI Research Inc., 2013; WIL, 2008b; WIL, 2009; WIL, 2010a; WIL, 2010b), o EEA-NH4 (Hita, 2006; WIL, 2011b), o F-DIOX (RTC, 2011; RTC, 2013); PFE alkenes: o PMVE (DuPont Haskell, 2007a; WIL, 2016), o PEVE (Haskell, 1997); other ether-based PFAS: o Tetraconazole (Huntingdon, N/A-b; Huntingdon, N/A-d; Huntingdon, N/A-e; Huntingdon, N/A-f), o Noviflumuron (Dow AgroSciences, 2002a; Dow AgroSciences, 2002b; Dow AgroSciences, 2005a; Dow AgroSciences, 2005b); FTOHs: o 6:2 FTOH (Hita, 2007b; Kirkpatrick, 2005; Mukerji et al., 2015; Rice et al., 2020; Serex et al., 2014; WIL, 2005) HFCs: o HFC-134a (ICI, 1979), o HFC-236fa (Haskell, 1996a), o HFC-365mfc (TNO, N/A), o HFC-76-13 (Hita, 2007a), o HFCPA (ZEORORA) (Huntingdon, 1998a); HFEs: o HFE-7100 (Huntingdon, 1996b), o HFE, 7200 (Mitsubishi, 1996), 161 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) o o HFE-7300 (Hita, 2004), and HFE-365pcf3 (DuPont Haskell, 2009; DuPont Haskell, 2011). For several of these substances, increased kidney weights were accompanied by histopathological changes in the kidneys. For example, minor microscopic findings, tubular epithelial hypertrophy or degeneration were found for          PFCA o PFHxA (Klaunig et al., 2015; WIL, 2005); PFSA o PFBS (Lieder et al., 2009); PFECAs: o HFPO-DA (Caverly Rae et al., 2015; Haskell, 2010; MPI Research Inc., 2013; WIL, 2010a; WIL, 2010b), o ADONA (Charles River, 2007a), o EEA-NH4 (Hita, 2006), o F-DIOX (Toxicology Centre S.p.A., 2011); PFE alkenes: o PMVE (DuPont Haskell, 2007a), o PPVE (Charles River, 2017), o PEVE (Haskell, 1997); other ether-based PFASs: o Tetraconazole (Huntingdon, N/A-e) o Noviflumuron (Dow AgroSciences, 2005b); fluorotelomer alcohols o 6:2 FTOH (Kirkpatrick, 2005; Rice et al., 2020; Serex et al., 2014; WIL, 2005) HFCs: o HFC-245fa (Frauenhofer Institut für Toxikologie und experim. Medizin, 2005), o HFC-76-13 (Hita, 2007a), HFOs: o HFO-1216 (Haskell, 1989), HFEs: o HFE-7300 (Hita, 2004), o HFE-356pcf3 (DuPont Haskell, 2011), o Sevoflurane (Gonsowski et al., 1994a; Gonsowski et al., 1994b). For some substances from different PFAS categories, necrotic effects were reported (PFHxA, HFPO-DA, ADONA, PEVE, HFE-7300, Sevoflurane, HFO-1216). For some other PFASs, including PFOA, PFOS, PFHxS, PFDA, PFUnA, PFBA, and PFDoDA, animal studies did not indicate impaired renal function or morphological damage (ATSDR, 2021). Oligomeric/polymeric PFASs Kidney effects of oligomeric/polymeric PFASs were studied in repeated dose inhalation and oral toxicity studies. For the PCTFE oligomers, kidney related effects, such as increase of relative kidney weights, were reported in rats exposed to various compositions of the fluoropolymer: - - PCTFE Halocarbon (HC) oils (base oils) of different trade names: o 27-S oil with C8-C10 chain length (Kinkead et al., 1990a), o Safetol®3.1/ MLO-87-124 (70% C6 : 30% C8) (Kinkead et al., 1989; Kinkead et al., 1990b), o PCTFE 3.1 (C6 : C8, unknown ratio) ((Jones et al., 1991) o PCTFE 3.1 (C6 : C8, unknown rat io with additives) (Mattie et al., 1993) pure PCTFE tetramers (DelRaso et al., 1991; Kinkead et al., 1991). Significant increases in blood urea nitrogen (BUN) in rats were reported either after 162 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) inhalation of PCTFE oligomers (3.1 oil) (Kinkead et al., 1989; Kinkead et al., 1990b) or after oral administration of PCTFE pure trimers or PCTFE pure tetramers (Kinkead et al., 1991). Comparable effects were reported for PCTFE oils, such as 3.1 oil or HC 27-S oil (Kinkead et al., 1990a; Kinkead et al., 1990b). Moreover, Jones et al. (1991) observed significantly increased BUN in all treated rhesus monkeys (15-day, oral) exposed to 725 mg/kg bw/d PCTFE oligomers (HC 3.1 oil). For perfluoropolyethers (PFPEs), one study reported histologic alterations, but no further effects on kidney: In a limit test with rats exposed to 1 000 mg/kg bw/d of polymeric PFPE Fomblin HC/25 product (Mn = 3 200, content of low molecular weight <1 000 less than 0.1%), compared to controls, three of five males (but not the females) showed significantly increased (localised) basophilia (Malinverno et al., 1996). The lower sensitivity of female rats compared to male rats is also known from the non-polymeric PFAS PFOA related to lower serum levels of PFOA in female rats (EFSA, 2020). In summary, there are indications that low molecular weight oligomeric/polymeric PFASs (oligomeric PCTFE and polymeric PFPE Fomblin HC/25) can cause kidney effects, which is generally in line with the effects observed for non-polymeric PFASs. Clarity on kidney effects of oligomeric/polymeric PFASs cannot be given on the basis of available data. B.5.2.1.4. Thyroid effects in experimental animals Non-polymeric PFASs In experimental animal models, several PFASs can increase thyroid weight, induce follicular hypertrophy and decrease serum T3 and T4, but only some PFASs induce all of these effects combined in rats (PFBA, PFOA, PFNA, PFDA, PFHxDA, PFHxS, (e.g., Butenhoff et al., 2012a; Hirata-Koizumi et al., 2015; MPI Research Inc., 2013; NTP, 2019a; NTP, 2019b; Ramhoj et al., 2018; Ramhoj et al., 2020)). For several other PFASs (especially PFAAs, including some PFEASs), an indication for thyroid effects is available. Thyroid hormones were reduced without changes in thyroid weight or histopathological changes in rats (and one mouse study for PFHpA) after repeated exposure to:      PFCA: o PFHxA (NTP, 2019a), o PFHpA (Anonymous, 2017); PFSA: o PFBS (Feng et al., 2017; NTP, 2019b), o PFOS (NTP, 2019b); PFECA: o HFPO-DA (Conley et al., 2021; Conley et al., 2019); PFESA: o 6:2 Cl-PFESA (Hong et al., 2020); HFEs: o HFE-7300 (Charles River, 2019b). In contrast, for TFA and HFCs or HFOs there are almost no indications for thyroid effects, which may partially be explained by the limited available data. The US EPA used perturbation of thyroid hormone levels as the critical effect to derive a subchronic and chronic reference dose for PFBS. It was acknowledged that there is uncertainty with regard to the potential for adverse developmental effects because of the lack of studies investigating neurodevelopmental effects. Nevertheless, taking all available data together, the US EPA concluded that the evidence in animals for thyroid effects supports a hazard (EPA-US, 2021a). 163 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) With respect to the mode of action, PFASs can interfere with thyroid metabolism on several levels in thyroid hormone biogenesis, distribution and receptor binding (EFSA, 2020; Köhrle, 2008). On the level of biogenesis, PFASs can potentially disturb sodium iodide symporters, haemoprotein thyroperoxidase, iodinases, or deiodinases (EFSA, 2020). On the level of distribution, there is evidence for competitive binding of PFAS to thyroid hormone binding proteins (transtyhretin and thyroxine-binding globulin), but at lower affinities than thyroid hormones (Behnisch et al., 2021; Berg et al., 2015; EFSA, 2020; Fenton et al., 2021; Ren et al., 2016; Weiss et al., 2009; Zhang et al., 2016a). Binding of PFCAs and PFSAs to thyroid receptors was shown for PFBA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA, PFBS, PFHxS and PFOS, but not for FTOH (8:2) (Ren et al., 2015). Coperchini et al. (2021) concluded in their review that available evidence, mainly provided by in vitro studies and in animal models, supports a thyroiddisrupting effect of the exposure to both old- and new-generation PFASs, while stating that epidemiological data provided contrasting results. Regarding oligomeric/polymeric PFASs, no studies observing thyroid parameters are known. B.5.2.1.5. Immune effects in experimental animals Strong evidence demonstrates that exposure to several non-polymeric PFASs modifies the immune response, with inhibition/suppression of the immune response as most consistent immunotoxic effect relevant for human health (DeWitt et al., 2019; Fenton et al., 2021; NTP, 2016b; Zeng et al., 2021). In the most recent EFSA opinion on PFASs (EFSA, 2020), effects on the immune system were considered the most critical effects and the risk assessment was based on reduction in vaccine antibody response in children allowing to derive a tolerable weekly intake (TWI) of 4.4 ng/kg bw per week for the sum of four PFASs (PFOS, PFHxS, PFNA, and PFOA). Recent epidemiological literature not included in EFSA (2020) is summarised in section B.5.3.1.1. In support of the epidemiological evidence, immunotoxic effects have been observed in animal studies for a variety of different PFASs. The following effects have been reported for PFASs across many PFAS subgroups, such as PFAAs, PFEASs, and some fluorinated gases. Reduction of lymphoid organ weights was observed for, e.g. PFHxA (Loveless et al., 2009), PFOA, PFNA, PFDA (NTP, 2019a), F-DIOX (RTC, 2011), CAS No. 524709-77-1 (NonClinical Saftey, 2017). Changes in lymphocyte counts or proliferation were observed for, e.g.:           TFA (BayerCropScience, 2014), PFOA and PFOS (Vetvicka and Vetvickova, 2013), PFDoDA (Kato et al., 2015), HFPO-DA (WIL, 2008a), F-DIOX (RTC, 2013), PMVE (DuPont Haskell, 2007a), Tetraconazole (Huntingdon, N/A-b), HFC-236fa (Haskell, 1996a), HFC-365mfc (Huntingdon, N/A-a), HFO-1216 (Triskelion, 2019a). Whereas most immunotoxic effects are reported in response to higher PFAS doses, reductions in T-cell-dependent antibody responses (TDAR) are observed also at non-toxic doses of PFOA and PFOS in mice (DeWitt et al., 2009; DeWitt et al., 2008; DeWitt et al., 2016; Dong et al., 2011; Dong et al., 2009; EFSA, 2020; Peden-Adams et al., 2008; Zheng et al., 2009). The TDAR is a measure of func tional immune response and has only been performed for a few PFASs. Serum antibody levels were reduced after exposure to PFOA (e.g., DeWitt et al., 2009; DeWitt et al., 2008; DeWitt et al., 2016), PFOS (e.g., Dong et al., 2009; Peden-Adams et al., 2008; Suo et al., 2017; Zheng et al., 2009; Zheng et al., 2011), and histological alterations in immune organs were observed after exposure to, e.g. 164 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFHxA (WIL, 2005), PFOS (Wang et al., 2011a), F-DIOX (RTC, 2011), PMVE (WIL, 2016). The most common immunotoxic observations in animal studies are effects on spleen weight followed by thymus weight (reduction in most cases). A recent study in mice exposed to three different perfluoroalkylether carboxylic acids (PFECAs) demonstrated that perfluoro2-methoxypropanoic acid (PFMOPrA) induced changes in splenic cellularity and perfluoro4-methoxybutanioc acid (PFMOBA) decreased numbers of B and natural killer (NK) cells (Woodlief et al., 2021). Significant changes in NK cell cytotoxicity or T cell-dependent antibody responses were not observed in this study (Woodlief et al., 2021). While the immunotoxic effects of some PFASs are well established, the mechanisms leading to the effects are still unclear (ATSDR, 2021; Beans, 2021; EFSA, 2020). For oligomeric/polymeric PFASs, immunotoxic effects were shown in terms of decreased absolute spleen weight of rats after inhalation of aerosolised oligomeric PCTFE oil (Kinkead et al., 1989). After oral administration of oligomeric PCTFE oil, absolute thymus weight was significantly increased in rats (Kinkead et al., 1990a). As detailed in section B.5.1.1, PFAS surfactants may increase the intestinal permeability, as demonstrated in vitro for PFOS, which may contribute to effects on the human immune system (Groh et al., 2017). B.5.2.1.6. Other effects (experimental animal data) Non-polymeric PFASs For a specific group of compounds, the n:2 fluorotelomer alcohols, another adverse effect finding has been consistently reported in the performed animal studies. Adverse effects on teeth and bone, consistent with fluoride toxicosis, have been seen with increasing doses of 6:2 FTOH (Hita, 2007b; Kirkpatrick, 2005; Mukerji et al., 2015; Rice et al., 2020; Serex et al., 2014; WIL, 2005). Likewise, Serex et al. (2014) report that similar effects have been seen for 8:2 FTOH exposed animals. The observed eff ects include: discolouration of the incisors, irregular ameloblast alignment and pigmentation, missing/broken/misaligned incisors, degeneration and atrophy of ameloblastic epithelium, accentuation of the normal laminar pattern of dentin and an increase in incomplete decalcification of enamel. Also incomplete decalcification of nasal bones and long bones (tibia and femur) has also been observed. The animal groups that show adverse effects on teeth and bone have been shown to have statistically significant increases in fluoride concentrations in serum and urine (Rice et al., 2020; Serex et al., 2014). Oligomeric/Polymeric PFASs In an older study, fluorocarbon vapour animal experiments with polymers of PTFE, PCTFE and PCTFE including plasticiser of unknown compositions (continuous inhalation exposure of mice to concentrations greater than 10 ppm; 1 h/d, 28 days and 1 h/d, 2 weeks) resulted in death (Hagemeyer and Stubblebine, 1954). Daily exposures appeared to have a cumulative effect, particularly in the region of 10 ppm. Autopsies showed inflammation and destruction of the tissues, oedema of the lungs, and enlargement of the heart. However, insufficiently reported study details (e.g. no information on control group, chemical/polymer characteristics, study design, etc.) weaken the power of the available information. Moreover, significant loss of body weight below the initial weight (>10%, manifested at day 7) during a 14-day repeated oral exposure of adult rats with pure PCTFE tetramers as well as with oligomeric PCTFE oil (3.1 oil-C6:C8, composed of 55% trimer and 45% tetramer oligomers) was observed for example by DelRaso et al. (1991). Furthermore, compared to controls, significantly decreased body weights and body weight gains of adult rats exposed to PCTFE oligomers/polymers were reported for either the oral route (pure trimers, pure tetramers, oligomeric PCTFE oils) (DelRaso et al., 1991; Hagemeyer and 165 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Stubblebine, 1954; Kinkead et al., 1991; Kinkead et al., 1989; Kinkead et al., 1990a; Kinkead et al., 1990b), or for the inhalational route of an oligomeric PCTFE oil (70% trimer: 30% tetramer) (Kinkead et al., 1989; Kinkead et al., 1990b). Also for polymeric PFPE surfactant, Johnston et al. (1996b) reported reduced body weights of rats, based on unpublished data. Affected body weights in adult rodents are also observed for nonpolymeric PFASs, such as PFHxA, PFNA, PFOS (EFSA, 2020). Furthermore, oral oligomeric PCTFE oil administration to rats led to significant changes in total protein, glucose and in serum albumin (Kinkead et al., 1990a; Mattie et al., 1993) as well as in erythroid parameters (Kinkead et al., 1989; Mattie et al., 1993), and in signs of neurotoxicity accompanied by increased brain weight (Kinkead et al., 1990a), were reported. Oral, dermal and inhalation repeated dose toxicity studies with experimental animals were summarised in an evaluating report regarding fluoropolymers as cosmetic ingredients, focussing on PTFE (Johnson and Zhu, 2018), but also on PCTFE in earlier versions (Johnson, 2018a; Johnson, 2018b). KrytoxTM (PFPE base oil) inhalation resulted in mild effects (increased lung weight) at the highest tested concentration (1 000 mg/m³), but no liver effects (such as liver weight, liver enzyme, peroxisomal ß-oxidation or histopathologic examination) in rats (DelRaso, 1996; Kelly et al., 1993). Some additional data on other effects of oligomeric/polymeric PFASs on humans are addressed under B.5.4. B.5.2.2. Toxicity to reproduction (experimental animal data) In a recent review, Fenton and colleagues concluded on reproductive toxicity that exposure to PFASs has adverse effects on conception, pregnancy, and foetal development. The underlying birth weight data are largely supportive, while the literature on growth and obesity is inconclusive (Fenton et al., 2021). B.5.2.2.1. Developmental effects in experimental animals Adverse effects on reproduction in experimental animal models, such as total litter loss and perinatal/postnatal mortality, have been observed for a variety of non-polymeric PFASs with different chemical structures. (Total) Litter loss was observed in experimental animal models after exposure to    TFA (Covance Laboratories, 2020a), as well as after exposure to PFCAs of various chain lengths: o PFBA (Das et al., 2008), o PFOA (Abbott et al., 2007), o PFNA (Das et al., 2015; Singh and Singh, 2019a; Wolf et al., 2010), o PFDA (Harris and Birnbaum, 1989), o PFODA (Hirata-Koizumi et al., 2012) PFECA o F-DIOX (RTC, 2011). Neonatal/Postnatal mortality was observed in experimental animal models for exposures to  some PFCAs and PFSAs: o PFHxA (Iwai and Hoberman, 2014), o PFHpA (Anonymous, 2017), o PFOA (Abbott et al., 2007; Lau et al., 2006; Song et al., 2018a; Yahia et al., 2010), 166 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFNA (Das et al., 2015; Wolf et al., 2010), PFODA (Hirata-Koizumi et al., 2012), and PFOS (Lau et al., 2003; Luebker et al., 2005a; Luebker et al., 2005b; Yahia et al., 2008). Furthermore, PFECAs: o HFPO-DA (Conley et al., 2021), o ADONA (Charles River, 2007b), o EEA-NH4 (WIL, 2011b). Complex ether-based PFAS o Noviflumuron also resulted in postnatal mortality (Dow AgroSciences, 2004). FTOHs o 6:2 FTOH has also been shown to increase pup mortality (Kirkpatrick, 2005; O'Connor et al., 2014; WIL, 2005). o o o    Reduction of offspring body weight and/or reduced body weight gain is also one of the more consistent developmental effects throughout different PFAS groups in experimental animal models. It can be a secondary effect of mat ernal toxicity (or influenced by it), caused by prenatal exposure, or by exposure through lactation (also see B.5.1.1.2 and B.5.1.1.4). Reduced offspring body weight (gain) was detected for all of the abovementioned substances (except PFBA) and in addition for     PFUnDA (Takahashi et al., 2014), PFTeDA (Hirata-Koizumi et al., 2015), PFBS (Feng et al., 2017), and Perfluamine/PTPA (Charles River, 2019a). A variety of fluorinated gases (HFCs and HFOs) also caused reduced pup body weight (gain):    HFC-245fa (Frauenhofer Institut für Toxikologie und experim. Medizin, 2005), HFC-4310mee (Haskell, 1994), HFCPA (ZEORORA) (Huntingdon, 1998b),    HFO-1336mzz(Z) (DuPont Haskell, 2010b; DuPont Haskell, 2014; WIL, 2014), HFO-1336mzz(E) (Triskelion, 2016), HFO-1216 (Triskelion, 2019a). Decreased pup body weight was also seen after 6:2 FTOH and 8:2 FTOH exposure (Kirkpatrick, 2005; Mukerji et al., 2015; Mylchreest et al., 2005; O'Connor et al., 2014; WIL, 2005). Other effects indicating toxicity to reproduction were only evident for single substances or single groups of PFASs. Those effects include the impaired development of mammary glands, which is to date only investigated for PFOA exposure (Macon et al., 2011; Tucker et al., 2015; White et al., 2011). Delayed ossification during development occurred in several PFAS groups (various PFAAs, HFPO-DA, several fluorinated gases, 6:2 FTOH as well as 8:2 FTOH) in experimental animal models (e.g., Lau et al., 2006; Mylchreest et al., 2005; O'Connor et al., 2014; Tveit et al., 2013; WIL, 2010a). Delayed ossification is often a secondary effect due to reduced pup weight or maternal toxicity. Still, there is emerging evidence that bone mineralisation may be affected by PFASs also in human children (see B.5.3.4.1) (Cluett et al., 2019; Fenton et al., 2021). Some PFASs caused developmental malformations in offspring of experimental animal models: TFA (Covance Laboratories, 2020b), PFOA (Lau et al., 2006), and HFC-125. However, for the latter the effect was only found in few pups at higher concentrations ((Huntingdon, 1992)). 167 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) For oligomeric/polymeric PFASs, no studies observing developmental toxicity are known. B.5.2.2.2. Fertility effects in experimental animals Reduced weight of reproductive organs might further affect fertility and is a consistent effect across various non-polymeric PFASs and various PFAS classes:        PFCAs o TFA (Covance Laboratories, 2020b), o PFNA (NTP, 2019a), o PFDA (NTP, 2019a), o PFDoDA (Chen et al., 2019b; Kato et al., 2015), PFSAs o PFBS (Feng et al., 2017), o PFOS (Lee et al., 2015), PFECAs o HFPO-DA (Conley et al., 2019; Haskell, 2010; WIL, 2008a; WIL, 2010a; WIL, 2010b), o EEA-NH4 (WIL, 2011b), o F-DIOX (RTC, 2011), o 6:2 Cl-PFESA (Zhou et al., 2018), Other ether-based PFASs o Move3 (Triskelion, 2017), o Noviflumuron (Dow AgroSciences, 2002a), HFCs o HFC-134a (ICI, 1979), o HFCPA (ZEORORA) (Huntingdon, 1998b), FTOHs o 6:2 FTOH (Mukerji et al., 2015; O'Connor et al., 2014), and the TFA-precursor 2,2,2-Trifluoroethanol (Wilkenfeld RM, 1981). Several different PFASs from different PFAS groups impaired sperm quality in experimental animal models after exposure to PFNA (Singh and Singh, 2019a; Singh and Singh, 2019b), PFDA (NTP, 2019a), and PFDoDA (Kato et al., 2015), as well as PFOS (Zhang et al., 2019b). Moreover, sperm quality was diminished in animal experiments by 6:2 Cl-PFESA (Zhou et al., 2018) as well as the fluorinated gas HFC-245fa (Frauenhofer Institut für Toxikologie und experim. Medizin, 2005) and the TFA precursor 2,2,2-Trifluoroethanol (Wilkenfeld RM, 1981). Furthermore, a reduction of sex hormones (estradiol and/or testosterone) was measured in parental animals after exposures to different PFAAs: PFNA (Feng et al., 2010; Feng et al., 2009; NTP, 2019a; Singh and Singh, 2019a), PFDA (NTP, 2019a), PFDoDA (Shi et al., 2009; Shi et al., 2007), and PFBS (Cao et al., 2020), PFOS (Zhang et al., 2020c). Impairment of oestrus cyclicity was observed in rodents after exposure to PFDoDA (Kato et al., 2015), PFBS (Feng et al., 2017)(also in pup oestrus cycle), PFOS (Du et al., 2019; NTP, 2019b), and HFPO-DA (WIL, 2008a) and 6:2 FTOH (Mukerji et al., 2015). Reduced fertility indices were observed in experimental animal models for Noviflumuron (Dow AgroSciences, 2004), for the fluorinated gas HFC-4310mee (DuPont Haskell, 2007b), and the TFA precursor 2,2,2-Trifluoroethanol (Wilkenfeld RM, 1981). Infertility was associated with PFOA and PFHxS in humans (Fei et al., 2009; Fenton et al., 2021). For oligomeric/polymeric PFASs, no studies observing fertility effects are known. B.5.2.2.3. Effects on or via lactation in experimental animals Reduced pup weight gain during the lactation period was observed after exposures to TFA 168 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) (Covance Laboratories, 2020a), PFHxA (Iwai and Hoberman, 2014), F-DIOX (RTC, 2012) and Noviflumuron (Dow AgroSciences, 2004). However, reduced pup weight during lactation can have multiple reasons and only cross-fostering experiments with lactation exposure excluding prenatal exposure can clearly indicate lactational effects. For example, PFOA induced impairment of mammary gland development only via lactation without prenatal exposure (EFSA, 2020; White et al., 2009). For Noviflumuron, data from a crossfostering study “indicate that the exposure of the test material to foetuses through the placenta is limited and the majority of exposure oc curs postnatally through nursing.” (TERC, 2005) (also see B.5.1.2). B.5.2.3. Carcinogenicity (experimental animal data) Several non-polymeric PFASs have recently been classified as possibly or probably carcinogenic to humans. The International Agency for Research on Cancer (IARC) class ified PFOA as possibly carcinogenic (Group 2B; (IARC, 2017)) based on positive associations observed for cancers of the testis and kidney as well as increased incidences of testicular Leydig c ell adenoma, hepatocellular adenoma, and pancreatic acinar cell adenoma in animal studies. The US EPA found that there is suggestive evidence that PFOA (EPA-US, 2016b), PFOS (EPA-US, 2016a), and HFPO-DA (EPA-US, 2018) may cause cancer. The EFSA Contam Panel concluded in 2018 that there is insufficient support for carcinogenicity of PFOS and PFOA in humans from epidemiological studies, but that both compounds induced tumours in rats (EFSA, 2018). According to the CLP Regulation in the EU, 17 PFASs currently exhibit a harmonised classification as carcinogenic (Carc. 2 or Carc. 1B; e.g. PFOA and its ammonium salt, PFNA and its sodium and ammonium salts, PFDA and its sodium and ammonium salts, PFOS and its ammonium, lithium and potassium salts, trifluralin). Additionally, amongst the PFASs registered in the EU, 82 PFASs are selfclassified by registrants as Carc. 1A/B or Carc. 2. For PFOA (3M, 1983; Biegel et al., 2001; Kamendulis et al., 2022; NTP, 2019a), PFOS (Butenhoff et al., 2012b), HFPO-DA (EPA-US, 2018; MPI Research Inc., 2013), Tetraconazole (Huntingdon, N/A-e), Noviflumuron (Dow AgroSciences, 2005b), and HFC134a (Collins et al., 1995) there is evidence for carcinogenic effects from animal studies (e.g. Leydig cell adenoma, hepatocellular adenoma and carcinoma, pancreatic acinar c ell adenoma and adenocarcinomas, follicular adenomas and carcinomas of the thyroid, renal tubule adenoma or carcinoma). For the observed carcinogenic effects, the available information is not sufficient to rule out human relevance of the underlying mode of action (ATSDR, 2021). Regarding polymeric PTFE, several studies found carcinogenic effects (e.g. malignant sarcomas/fibrosarcomas) after subcutaneous implantation of PTFE implants in rodents (Bryson and Bischoff, 1969; IARC, 1999; Oppenheimer et al., 1955; Russell et al., 1959; Tomatis, 1963). Comparable number of tumours were seen in experimental animals with similarly sized implants of glass (IARC, 1999; Tomatis, 1963) and non-fluorinated polyethylene (Oppenheimer et al., 1955). Solid-state carcinogenesis (IARC, 1999) might therefore be an explanation for the carcinogenic effects of PTFE implants. For the vast majority of PFASs, long-term toxicity or carcinogenicity studies as well as epidemiological studies informative on potential carcinogenic effects are not available and thus, human relevance of carcinogenicity of most PFASs is unclear. 169 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5.3. Evidence from epidemiological data In 2021, ATSDR published Toxicological Profile for Perfluoroalkyls (ATSDR, 2021) including literature published before September 2018. In 2018 and 2020, EFSA published two opinions on PFASs in food including literature searched covering the time period between January 2008 and July 2019. In the EFSA opinion from 2020, 27 PFASs 13 were included but only PFOS, PFHxS, PFOA and PFNA were included in the risk characterization (EFSA, 2018; EFSA, 2020). For the present restriction dossier, a new literature search (Medline) was performed in August 2021 to identify papers published on PFASs not included in the EFSA opinions, as well as papers published after July 2019 for the 27 PFASs covered by EFSA. This section focuses particularly on evaluating whether new studies strengthen or weaken the conclusions made by EFSA (2020) and if data on additional PFASs follow the same patterns. A narrative review of the new information is provided below (sections B.5.3.1 to B.5.3.5). Several studies have measured PFASs that were not included in the final statistical analyses due to low detection frequencies (e.g. detected in less than 70% of the study population). This does not rule out possible associations between these PFASs and health outcomes. For the less studied PFASs, the evidence for exposure-health association is limited due to the small number of studies. B.5.3.1. Epidemiological evidence for effects on organ function, energy metabolism and the immune system B.5.3.1.1. Immune outcomes Since effects of PFASs on the immune system were considered the most critical for the risk assessment by EFSA (2020), the literature search for the immune outcomes in humans was extended to cover the period until 25 th January 2022. The outcomes have been categorised into vaccination responses, common infectious diseases, asthma- and allergyrelated outcomes and autoimmune diseases. Vaccination response EFSA concluded in 2020 that PFOS and PFOA are associated with reduced antibody responses to vaccination. Given the lower concentrations of other PFASs, the evidence for associations for these were weaker (EFSA, 2020). Further, ATSDR (2021) stated that PFOA, PFOS, PFHxS and PFDA are inversely related to vaccination responses, and reported also some limited evidence for similar associations for PFNA, PFUnDa and PFDoDA. The new literature search identified three studies published between July 2019 and January 25th 2022. No studies on additional PFASs than the 27 included in EFSA (2020) were identified in the new literature search. In a study based on a subset of data from a randomised controlled trial of early measles vaccination conducted in Guinea-Bissau (Timmermann et al., 2020), children’s (n = 237) PFOS and PFDA levels at inclusion (4-7 months of age) were associated with lower measles antibody concentrations at 9-months among the children who had received a measles vaccine. Furthermore, there was a trend towards reduced pre -vaccination measles antibody levels with elevated PFAS levels (PFHxS, PFOA, PFNA, PFUnDA) indicated by 13 The PFASs included in EFSA (2020): PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA, PFPeDA, PFHxDA, PFODA, PFBS, PFHxS, PFHpS, PFOS, PFDS, PFOSI, 8:2 FTOH, 8:2 monoPAP, 8:2 diPAP, FOSA, EtFOSA, EtFOSE, FC -807. 170 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) sensitivity analysis where the most influential PFAS observations were removed. In another prospective study from the Faroe Islands (Shih et al., 2021), the association between several PFASs (PFOS, PFOA, PFHxS, PFNA, PFDA) at birth, 7, 14, 22 and 28 years of age and serum antibody levels in adults vaccinated at age 28 against hepatitis type A and B, diphteria and tetanus (n=172-399) were evaluated. Inverse, but not statistically significant, trends were observed between PFOA at ages 14 and 28 years and hepatitis type A antibody levels, as well as PFOA at 22 and 28 years and hepatitis type B antibody levels. No inverse associations of PFAS exposure were found with diphtheria and tetanus antibody concentrations. In a cross-sectional study in vaccinated Greenlandic children (7 — 12 years of age), serum levels of PFHxS and PFOS was associated with lower diphtheria antibody levels in children with known vaccination record (n = 169). No similar associations were observed for tetanus antibody levels (Timmermann et al., 2022). No associations were observed for PFHpS, PFOA, PFNA, PFDA and PFUnDA. All three studies reported inverse trends between levels of PFASs and vaccine antibody levels, hence supporting the conclusions of ATSDR and EFSA. Common infectious diseases EFSA concluded that there is some evidence to suggest that exposures to PFASs are associated with increased propensity of infections but that more studies with objective measures of infections (not self-reports) are needed (EFSA, 2020). Twelve studies were identified in the new literature search. Of these studies, f ive examined the effect of prenatal PFAS exposure and health outcome in children, two examined childhood exposure and outcome in children/adolescents (prospective design), and five were cross-sectional studies in adults. With the exception of Zeng et al. t hat included 6:2 Cl-PFESA and 8:2 Cl-PFESA (Zeng et al., 2020), no data on additional PFASs than the ones included in EFSA (2020) were identified. With regard to upper respiratory tract infections (URTI) (Dalsager et al., 2021; Kvalem et al., 2020; Wang et al., 2022) and gastric infections (Dalsager et al., 2021; Timmermann et al., 2020; Wang et al., 2022), there are inconsistent findings. When it comes to lower respiratory tract infections (LRTI), however, four of six prospective studies reported positive associations between either prenatal or childhood exposure to PFASs and LRTI (PFOS, PFHxS, PFHpS, PFOA, PFNA, PFDA, PFHpA and/or PFDoDA) (Ait Bamai et al., 2020; Dalsager et al., 2021; Impinen et al., 2019; Kvalem et al., 2020). A positive association was also reported for PFBS and respiratory tract infections, including both URTI and LRTI (Huang et al., 2020). Furthermore, two of the three prospective studies examining groups of infectious diseases as the outcome reported positive associations between either prenatal or childhood exposure to PFASs and infections (Timmermann et al., 2020): Fever, diarrhea, coughing or vomiting; (Dalsager et al., 2021): URTI, LRTI, GI or other infections). In the cross-sectional study on groups of outcomes, Bulka et al. Bulka et al. (2021) reported on a positive association between PFAS exposure (PFOS, PFHxS, PFOA, PFNA) and pathogen burden (sum of seropositivity: cytomegalov irus, Epstein Barr virus, hepatitis C and E, herpes simplex 1 and 2, HIV, T. gondii, and Toxocara spp). The association between PFOS, PFOA, PFBA, 6:2 Cl-PFESA, 8:2 Cl-PFESA and the antibody levels of hepatitis B virus (HBsAb) were investigated in a cross-sectional study on Chinese adults. Lower serum HBsAb levels were observed for an increase in linear PFOS, 6:2 ClPFESA, 8:2 Cl-PFESA and PFBA. The association seemed to be stronger for 6:2 Cl-PFESA than for PFOS. Furthermore, cross-sectional studies in adults on COVID-19 infections reported positive associations with several PFASs (Grandjean et al., 2020; Ji et al., 2021; Nielsen and Joud, 2021). Overall, the new studies strengthen the evidence for an association between PFASs and 171 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) common infectious diseases, and particularly with respect to lower respiratory tract infections. Asthma and allergies In a prospective study no associations between prenatal exposure to PFASs and doctordiagnosed asthma were observed, whereas in another prospective study an association between PFNA and self-reported asthma (not doctor-diagnosed) was reported (Beck et al., 2019; Zeng et al., 2019b). Two cross-sectional studies were identified, one reported an increased risk of asthma in young children with exposure to PFOS and PFOA (JacksonBrowne et al., 2020), whereas no statistically significant associations were found in a study in adolescents (Gaylord et al., 2019). There is little evidence for an effect of tested PFASs on lung function in humans. Three prospective studies have investigated the effect of prenatal PFAS exposures on lung function in childhood (4-12 years of age) (Agier et al., 2019; Kung et al., 2021; ManzanoSalgado et al., 2019), whereas one study has investigated the effect of exposure at age 10 years and lung function at the age of 16 years (Kvalem et al., 2020). No statistically significant findings were reported except an inverse association between prenatal exposure to PFOA and forced vital capacity (FVC) at age 4 years (Manzano-Salgado et al., 2019). With regard to cross-sectional studies in children and/or young adults, no statistically significant associations have been observed (Agier et al., 2019; Gaylord et al., 2019; Impinen et al., 2018; Kung et al., 2021; Kvalem et al., 2020). Regarding atopic dermatitis in relation to PFAS exposure, no statistically significant associations were observed in a cross-sectional study including children (Kvalem et al., 2020). In studies with a prospective design, however, prenatal (Ait Bamai et al., 2020; Impinen et al., 2019; Lowe et al., 2019; Manzano-Salgado et al., 2019) or childhood (Kvalem et al., 2020) exposure to some PFASs (PFHxS, PFOS, PFHxA, PFOA, PFUnDA, PFTrDA, PFDoDA and/or N-MeFOSAA) have been associated with a reduced risk of atopic dermatitis, while one study observed a positive association with prenatal PFOA exposure (Wen et al., 2019). Most of the associations with atopic dermatitis are seen in relation to exposure to PFCAs. One study on total IgE levels were identified. Using piecewise linear regression, inverse relations were observed between PFASs (PFOS, PFOA, PFNA, PFDA) and total IgE below the turning point and positive relations above the turning point (Zeng et al. 2019). As concluded by EFSA, there are no or inconsistent associations between prenatal and postnatal PFAS exposure and asthma in children and young adults. Furthermore, there is little evidence for an effect of PFAS exposure on lung function. With regard to allergyrelated outcomes, there is increasing evidence for an inverse association between PFASs and atopic dermatitis, whereas no conclusion can be drawn with regard to the effect of PFAS exposures on rhinitis, urticaria, serum IgE or skin prick tests due to few studies and inconsistent findings. No data on additional PFASs than the 27 included in EFSA (2020) were identified. Autoimmunity According to EFSA (2018) and ATSDR (2021), there is some suggestive association between serum PFOA and an increased risk of ulcerative colitis, but not for other autoimmune diseases. No studies on autoimmune diseases were retrieved from the new literature search. B.5.3.1.2. Thyroid function and disease EFSA (2018) and EFSA (2020) concluded that the available evidence was insufficient to suggest that exposures to PFASs are associated with thyroid disease or changes in thyroid hormones. The vast majority of the reviewed studies were cross-sectional. 172 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The updated literature search identified 14 studies. Several of the recent studies investigated PFAS mixture effects. Six publications used data from birth cohort studies and examined prospective associations between PFASs measured in maternal blood sampled in pregnancy and thyroid hormones measured in neonatal cord-blood. Of these, five also reported cross-sectional associations with maternal thyroid function and/or disease (Itoh et al., 2019; Lebeaux et al., 2020; Liang et al., 2020; Preston et al., 2020b; Preston et al., 2018; Xiao et al., 2020). Two studies reported cross-sectional associations where PFASs and thyroid function markers were measured in neonatal cord-blood (Guo et al., 2021) or in child blood at age 2, 4 and 6 years (Kim et al., 2020a). Six studies reported associations between PFASs and thyroid hormones in adult populations, one was prospective (Blake et al., 2018), one a case-control study (Li et al., 2021d), and four were cross-sectional (Aimuzi et al., 2020; Gallo et al., 2022; Inoue et al., 2019; Sarzo et al., 2021). Overall, the results from prospective studies indicated positive associations bet ween prenatal PFAS mixtures and single PFASs (elaborated below) with T3 and fT3, inverse associations with T4 and/or fT4, and null associations with TSH. However, it is noted that the results were not consistent between the studies regarding which of the P FASs that contributed to the thyroid hormone associations, sex-differences, and/or maternal thyroid antibody status. For example, in Preston et al. (2020b) higher concentrations of PFAS mixtures (BKMR and WQS) showed that PFHxS and N-MeFOSAA contributed most to lower T4, and associations were predominantly found in male neonates, whereas Liang et al. (2020) showed that PFNA, PFDA and PFUnDA contributed most to the PFAS mixture (BKMR) association with higher T3, while Itoh et al. (2019) and Lebeaux et al. (2020) found that PFOA, PFOS and PFHxS contributed most to lower cord fT4 but was only found in mothers with higher TPO antibody levels. For the studies in adults, some associations between individual PFASs and PFAS mixtures and thyroid function were indicat ed, while the two cross-sectional studies in neonates showed inconsistent associations. Overall, the new studies strengthen the concern that PFAS exposures might be associated with thyroid disease or changes in thyroid hormones. B.5.3.1.3. Liver effect and metabolic outcomes Liver enzymes and liver disease EFSA (2020) concluded that “epidemiological studies provide evidence for an association between exposure to PFASs and increased serum levels of the liver enzyme alanine transferase (ALT). The magnitude of the associations was small, however, and few studies found associations with ALT outside the reference range. There were no associations with liver disease.” Two cross-sectional (Attanasio, 2019; Jin et al., 2020d) and two prospective studies (Mora et al., 2018; Stratakis et al., 2020) were identified in the new literature search. A crosssectional analysis of NHANES data (2013-2016) reported sex-differences for the associations between PFOA and PFNA and the liver enzymes ALT and aspartate aminotransferase (AST) in adolescents aged 12 – 19 years, with positive associations in females and inverse associations in males (Attanasio, 2019). Also, PFHxS was inversely associated with AST in males, but not in females. PFOA, PFNA, PFOS and PFHxS were positively associated with bilirubin levels in both male and female adolescents (Attanasio, 2019). A prospective study in the US project Viva cohort examined associations between prenatal PFASs exposure and ALT levels in children (mean age 7.7 years). This study also reported sex differences and showed that in girls, prenatal exposures to N-MeFOSAA and N-EtFOSAA were associated with lower ALT. The other PFASs showed similar tendency (Mora et al., 2018). These two studies suggest that prenatal and early life exposures to PFASs may 173 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) show sex and developmental stage dependencies. A study by Jin et al. (2020d) in children and adolescents with physician-diagnosed nonalcoholic fatty liver disease (NAFLD), reported associations between alterations in key amino-acids and lipids pathways underlying NAFLD pathophysiology and exposure to PFOA, PFOS, PFHxS and a mixture of PFAS. In a subcohort from the Human Early-Life Exposome (HELIX) project, Stratakis et al. (2020) evaluated whether prenatal exposure to PFASs in 1 105 mothers and their children was associated with established serum biomarkers of liver injury and alterations in se rum metabolome in children. Higher exposure to PFAS mixture (PFOS, PFOA, PFNA, PFHxS, and PFUnDA) during pregnancy was associated with higher liver enzyme levels of ALT, AST, and gamma-glutamyltransferase (GGT) in child serum. They also found significant perturbations in amino acid and glycerophospholipid metabolism associated with prenatal PFAS exposure. Furthermore, a systematic review and meta-analysis of 24 human epidemiological studies and 85 rodent studies provide new evidence for a causal associatio n between PFASs exposure and liver disease (Costello et al., 2022). Meta-analyses of the human studies revealed that exposure to PFOA, PFOS, and PFNA were associated with higher ALT levels. PFOA exposure was also associated with higher aspartate aminotransferase and gammaglutamyl transferase levels in humans. In rodents, PFAS exposures consistently resulted in higher ALT levels and steatosis. Thus, the results for associations between PFASs and markers of liver function in observational human studies strengthen the evidence for PFASs hepatotoxicity from rodent studies. The results from the new studies are in line with the findings in EFSA (2020) and strengthen the evidence for an association between exposure to PFASs and increased levels of the liver enzymes, mainly alanine transferase (ALT). Blood lipids In 2018, EFSA concluded that human epidemiological studies provided strong support for associations between exposure to PFOS and PFOA and increased serum levels of total cholesterol (TC) and LDL cholesterol (LDL-C). Evidence for a similar association for PFNA were reported by EFSA in 2020. However, there is uncertainty connected to the causality of this association because of potential confounding caused by common mechanisms of reabsorption of bile acids and PFASs in the small intestine (enterohepatic circulation) (EFSA 2020). Reabsorbed bile acids are negatively correlated with serum cholesterol, meaning that individuals with low enterohepatic circulation (e.g. by natural individual variability) can have low levels of PFASs and low levels of cholesterol, which leads to a significant but possibly not causal correlation (EFSA, 2020). In the new literature search, one prospective birth cohort study and 15 cross -sectional studies were identified of which one had both a cross-sectional and prospective design (Liu et al., 2020b). Three of the studies (Cong et al., 2021; Han et al., 2021a; Yu et al., 2021c) included 6:2 Cl-PFESA and 8:2 Cl-PFESA that were not covered by EFSA. A study in the Odense Child cohort examined PFASs in maternal pregnancy serum (PFOS, PFOA, PFHxS, PFNA and PFDA) and repeated lipid measurements in infancy (up to 18 months). The results showed that PFNA and PFDA were positively associated a higher TC (Jensen et al., 2020b). Liu et al. (2020b) performed a 2-year follow-up study in adults investigating the effect of plasma PFASs (PFOS, PFOA, PFHxS, PFNA, PFDA) at baseline with lipoprotein and apolipoprotein subspecies both at baseline and at 2 years. At baseline, elevated plasma PFOA concentrations were significantly associated with higher apoB and apoC -III concentrations. PFASs were not associated with TC or triglycerides (TG). PFAS levels were 174 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) primarily associated with lipids or apolipoprotein concentrations in intermediate -to-low density lipoprotein (IDL+LDL) and HDL-C that contain apoC-III. Similar patterns of associations were observed at the two-year follow-up. Several cross-sectional studies examined associations between PFASs and blood lipids in adolescents or adults. Positive association for PFASs (PFOA, PFOS, PFHxS, PFNA, PFDA, PFUnDA and/or 6:2 Cl-PFESA) with TC and/or LDL-C were reported in six studies (Averina et al., 2021; Canova et al., 2020; Cong et al., 2021; Han et al., 2021a; Lin et al., 2020a; Lin et al., 2020c). One study reported a positive association between PFOA and small dense LDL particles (Liou and Kaptoge, 2020). Two studies reported no associations between the PFASs and TC or LDL-C (Chen et al., 2020b; Menzel et al., 2021). The majority of the studies on high-density lipoprotein cholesterol (HDL-C) and TG observed null associations with PFASs. However, positive and inverse associations were also reported (Canova et al., 2020; Chen et al., 2020b; Cong et al., 2021; Han et al., 2021a; Lin et al., 2020a; Lin et al., 2020c; Ye et al., 2021; Yu et al., 2021c; Zare Jeddi et al., 2021). This is in line with the previously described variability in these associations as described by EFSA (EFSA, 2018; EFSA, 2020). Two studies examined PFAS lipid associations in pregnant women (Dalla Zuanna et al., 2021; Yang et al., 2021). In a cross-sectional study by Dalla Zuanna et al. (2021) plasma levels of TC, HDL-C and LDL-C were reported to increase throughout the trimesters. PFOA was inversely associated with TC, whereas both PFOA and PFHxS were inversely associated with LDL-C. All the three PFASs were positively associated with HDL-C. In a cohort of pregnant women in China, Yang et al. Yang et al. (2021) reported positive associations between PFHxS level and TC, LDL and LDL-C, as well as between PFUnDA level and HDL-C. Inverse associations were reported for PFDA and LDL-C, and for PFOA and the LDL-C/HDLC ratio. No associations were reported for PFOS, PFNA, PFHpS and the sum of PFASs and the blood lipids. Of note, serum cholesterol is normally elevated during pregnancy. In a cross-sectional study in Taiwanese children using lipidomics, 13 PFASs were analysed (Lee et al., 2021). PFTrDA and PFDA exposures were associated with serum lipid profiles, while PFOS exposure showed a borderline significant association. When examining the associations of exposure to each of these PFASs with each lipid, several lipids were identified associated with the levels of the PFASs. In the cross-sectional Fit Futures study, the effect of serum PFASs in adolescents on dyslipidemia (defined as: TC ≥ 5.17 mmol/L and/or LDL-C ≥ 3.36 mmol/L and/or apolipoprotein B ≥1.10 g/L) was examined. The highest vs. lowest quartiles of total PFAS (ΣPFAS), PFNA and PFDA concentrations were positively associated with the risk of dyslipidemia (Averina et al., 2021). In line with the findings in EFSA (2020), the review of new studies supports the association in non-pregnant adults between PFASs and total cholesterol, LDL cholesterol and inconsistency in the associations with HDL cholesterol and triglycerides. More data on PFASs other than PFOS and PFOA indicate similar associations also for PFHxS, PFNA, PFDA and PFUnDA, 6:2 Cl-PFESA and 8:2 Cl-PFESA. Diabetes In the EFSA Opinion on PFOS and PFOA (EFSA, 2018), 15 studies were reviewed on the associations between PFOS and/or PFOA and glucose homoeostasis or diagnosis of diabetes. Overall, the results did not indicate adverse effects on glucose homoeostasis or increased risk of diabetes. In EFSA (2020), eight studies were reviewed on associations between PFASs other than PFOS and PFOA, and diabetes (or metabolic syndrome), and the results were not consistent. EFSA (2020) thus concluded on insufficient evidence for associations between PFASs and diabetes. 175 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The new literature search identified 11 studies published after July 2019 reporting on associations between PFASs and risk of diabetes, and one study published before 2019 addressing PFASs not discussed by EFSA in 2020. Of the four prospective studies on gestational diabetes, three studies observed a statistically significant association with PFAS exposure (Rahman et al., 2019; Xu et al., 2020a; Yu et al., 2021a). The PFASs that were positively associated with gestational diabetes were PFOS, PFNA and PFHpA in Yu et al. (2021a) and PFNA, PFOA, PFHpA, PFDoDA in Rahman et al. (2019), and PFBS and PFDoA in Xu et al. (2020a). In the fourth study by Zhang et al. (2015), no associations were observed between N-ETFOSAA, N-MeFOSAA, PFDA, and PFOSA and gestational diabetes. Two cross-sectional studies on gestational diabetes reported positive associations with PFAS exposure (Preston et al., 2020a; Ren et al., 2020). Preston et al. (2020a) reported a positive association between PFOS and MeFOSAA and biological markers linked to gestational diabetes. Ren et al. (2020) reported positive associations between PFOS, PFNA, PFDA, PFUnDA and PFDoA with glucose levels in pregnant women, linked to the risk of gestational diabetes. Regarding diabetes type 2, a prospective adult cohort study which included six PFASs (PFOS, PFOA, PFHxS, N-EtFOSAA, N-MeFOSAA), found that PFOA was significantly associated with an increased risk of diabetes and that PFOS and N-EtFOSAA were associated with microvascular disease (Cardenas et al., 2019). The four case-control studies on diabetes type 2 in adults showed inconsistent findings with positive, no or inverse associations between PFASs and diabetes type 2 (Charles et al., 2020; Duan et al., 2021; Han et al., 2021a; Schillemans et al., 2021). Positive associations between PFAS exposure and diabetes were reported in two crosssectional studies (Duan et al., 2020; Zeeshan et al., 2021). Zeeshan et al. reported positive associations between branched PFOS, linear PFOS, branched PFHxS, linear PFHxS, PFOA, PFHpS, PFNA, PFDA, PFUnDA, PFTrDA and total PFAS and diabetes type 2. Duan et al. reported that PFHxA, PFOA, PFNA, PFHxS, ∑PFCAs C2-C3 (sum ultrashort-chain perfluoroalkyl carboxylic acids, i.e. 2-3 carbon atoms), ∑PFCAs C4-C7 (sum short chain PFCAs i.e., 4-7 carbon atoms), ∑PFCAs C8-C12 (sum long chain PFCAs, i.e., 8-12 carbon atoms), ∑PFCAs (sum all PFCAs), and ∑PFASs (sum all PFASs) were positively associated with glycemic biomarkers linked to increased risk of diabetes (Duan et al., 2020). In line with the findings in EFSA (2020), the review of new studies shows inconsistent results with positive, no- or inverse associations between PFASs and diabetes type 2. However, there are new studies suggesting a positive association between several PFASs and gestational diabetes. Obesity EFSA (2020) concluded that “there is insufficient evidence for associations between PFASs and obesity”. The literature review in 2021, identified 11 studies published after July 2019 reporting associations between PFASs and obesity, and one prospective cohort with repeated measures of both PFASs and outcome (Blake et al., 2018) reporting PFASs (Me-PFOSA and Et-PFOSA) not included in EFSA 2020. Five of these were studies in adults (Blake et al., 2018; Chen et al., 2020b; Ding et al., 2021; Mitro et al., 2020a; Mitro et al., 2020b) and seven in children and adolescents. Of the studies in children, those with PFASs assessed during pregnancy (Jensen et al., 2020a; Martinsson et al., 2020; Starling et al., 2019) are described in section B.5.3.2.2. In adults, positive associations between PFAS and obesity were reported for PFOS (Chen 176 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) et al., 2020b; Ding et al., 2021; Mitro et al., 2020b), PFOA (Chen et al., 2020b; Ding et al., 2021; Mitro et al., 2020a; Mitro et al., 2020b)), PFHxS (Averina et al., 2021; Chen et al., 2020b), PFHpS (Averina et al., 2021), N-EtFOSAA (Ding et al., 2021), and N-MeFOSAA (Ding et al., 2021). In the study by Blake et al. (2018) none of the investigated PFASs (PFOA, PFOS, PFOSA, PFNA, PFHxS, PFDeA, Me-PFOSA and Et-PFOSA) were associated with obesity. In children and/or adolescents, inverse associations between PFAS and obesity were reported for PFOS (Thomsen et al., 2021) and PFOA (Pinney et al., 2019), PFNA (Domazet et al., 2020), PFHxS (Domazet et al., 2020), and PFDA (Domazet et al., 2020; Thomsen et al., 2021). The review of new studies suggests differences in associations between PFASs and obesity in adults and children/adolescents. Whereas the majority of the new studies report positive associations between PFAS exposures and obesity in adults, inverse associations were reported for children/adolescents. Overall, some associations between PFASs and obesity are indicated, but the evidence is still insufficient with regard to a link between PFASs and obesity. Metabolic syndrome EFSA (2020) concluded that “there is insufficient evidence for associations between exposure to PFASs and metabolic syndrome” (EFSA, 2020). Three cross-sectional studies in adults were identified in the new literature search. Ye et al. (2021) measured a total of 20 PFASs. Positive associations were found for total PFOS, branched PFOS, linear PFOA and perfluoro-6-methylpheptanoic acid (6 m-PFOA), PFNA and PFDA, while null association was observed for linear isomers of PFOS, PFHxS, PFDoA, PFTrDA and PFUnDA. Yu et al. (2021c) quantified 32 PFASs in serum from Chinese adults. They found significantly positive associations for 6:2 Cl-PFESA, 8:2 Cl-PFESA and PFOS with metabolic syndrome. In addition, they found statistically significant positive associations between the PFAS mixture and metabolic syndrome when holding all PFAS at the 55th percentile or above compared to the median concentrations. There were no associations between PFOA, PFBA, PFHxS, PFHpA, PFNA, PFDA, PFUnDA, PFDoDA or PFTrDA with metabolic syndrome. Zare Jeddi et al. (2021) examined the associations between serum PFAS levels and the prevalence of metabolic syndrome among highly exposed young adult (ages 20 – 39) residents in the Veneto Region (North-Eastern Italy). PFOA, PFHxS, and PFNA were not associated with the risk of metabolic syndrome, whereas PFOS was associated with a lower risk of metabolic syndrome. One of the new studies included 6:2 Cl-PFESA and 8:2 Cl-PFESA, in which positive associations with metabolic syndrome were reported. However, due to the few additional studies and inconsistent findings, these studies do not strengthen the evidence for an association between PFASs and metabolic syndrome. Of note, different criteria were used for the identification of metabolic syndrome. B.5.3.1.4. Kidney function and uric acid EFSA (2020) summarised that there was insufficient evidence for associations between exposure to PFASs and kidney function and furthermore that “although several studies show association between PFASs and glomerular filtration rate, there is insufficient evidence to conclude that the associations are causal, due to plausible confounding and reverse causality”. The literature search conducted in 2021 identified seven studies published after July 2019, and one study (Blake et al., 2018) including PFASs (Me-PFOSA and Et-PFOSA) not evaluated by EFSA (2020). 177 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Three cross-sectional studies reported positive associations between several PFASs (PFOS, PFOA, PFNA, PFHxS) and two markers of reduced kidney function; increased serum uric acid/hyperuricemia (Scinicariello et al., 2020; Zeng et al., 2019a) and reduced estimated glomerular filtration rate (eGRF) (Moon, 2021). However, in the cross-sectional study by Arrebola et al. (2019), no associations were reported between PFASs (PFHxS, PFOA, PFOS, PFNA, PFDA and MeFOSAA) and serum uric acid levels in the study population. Associations with markers of reduced kidney function were reported in a randomised control trial and a prospective study. In the randomised controlled trial (Blake et al., 2018), that included lifestyle interventions, it was found that plasma PFAS concentrations (PFOS and PFOA) during follow-up were inversely associated with eGFR. According to the prospective cohort study with repeated measures by Lin et al. (2021), serum PFNA, PFHxS, and PFDA were inversely associated with eGRF. In contrasts, the PFOS precursor Me-PFOSA was associated with an increase in eGFR. Two cross-sectional studies reported positive association with urge urinary incontinence (Cui et al., 2021) and risk of kidney stones (Blake et al., 2018). The review of new studies strengthens the support for associations between PFASs and outcomes related to reduced kidney function. However, as pointed out by EFSA (2020) and ATSDR (2021) confounding and reverse causation is plausible and thus a causal association between PFAS exposures and kidney function is uncertain. B.5.3.1.5. Cardiovascular disease and mortality EFSA (2018) concluded on no clear association between PFOS/PFOA and cardiovascular disease and in 2020 EFSA concluded that there was not sufficient evidence for other PFASs. The new literature search identified eight studies examining associations between PFAS and cardiovascular outcomes published after July 2019, and one study from before July 2019 including other PFASs. A cross-sectional study with 10 859 participants from NHANES observed positive associations were between N-EtFOSAA and ∑12 PFASs and total cardiovascular disease, and between N-MeFOSAA and congestive heart failure (Huang et al., 2018). In a prospective study, prenatal exposure to PFOA was positively associated with cardiometabolic risk score in children at age 12 (Li et al., 2021b). In another prospective study, PFOS and N-EtFOSAA were positively associated with risk of any microvascular disease (Cardenas et al., 2019). Studies on blood pressure and/or hypertension observed positive associations with several PFASs; PFOS (Averina et al., 2021; Liao et al., 2020; Pitter et al., 2020b), PFOA (Averina et al., 2021; Liao et al., 2020; Mitro et al., 2020a; Pitter et al., 2020b), PFNA (Liao et al., 2020; Pitter et al., 2020b), PFHxS (Averina et al., 2021; Pitter et al., 2020b), and Σ18 PFAS (Averina et al., 2021). The results from the new literature search strengthen the evidence for an association between PFASs and increased risk of cardiovascular diseases. B.5.3.2. Reproductive outcomes (epidemiological evidence) B.5.3.2.1. Fertility, pregnancy and sex hormones outcomes EFSA (2018) and EFSA (2020) concluded that “the available evidence is insufficient to suggest that pre- or postnatal exposures to PFASs are associated with effects on male fertility”. Concerning female fertility, EFSA (2018) and EFSA (2020) concluded that “there is insufficient evidence that exposure to PFOS or PFOA is related to menstrual cycle length, 178 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) menopause or puberty development” and that “there are no consistent associations between pre- or postnatal exposures to PFASs and female reproduction outcomes or puberty”. Furthermore, EFSA (2020) concluded “that there is insufficient evidence that exposure to PFOS/PFOA or other PFASs may adversely affect fecundity”. Male and female fertility The updated search identified seven studies on fertility or fertility -related parameters in adults or adolescents. A cross-sectional study with couples undergoing in vitro fertilization (IVF) treatment suggested that both male and female exposure to several PFASs (PFOS, PFHxS, PFNA, PFDA, PFOSA, and PFHpA) were negatively associated with one of more intermediate IVF outcomes, i.e. the numbers of retrieved oocytes, mature oocytes, 2 PN zygotes and goodquality embryos (Ma et al., 2021). Furthermore, male exposure to PFNA and PFDA concentrations were associated with reduced sperm concentration and PFOSA was associated with reduced sperm count. None of the maternal or paternal PFAS concentrations were associated with IVF clinical outcomes. Three cross-sectional studies on male fertility conducted in China were identified. The first (Luo et al., 2021) assessed associations of 24 PFASs, including seven branched PFOS isomers, two branched PFOA isomers, 8:2 Cl PFESA and 6:2 Cl-PFESA with male reproductive hormones. Hormones measured in serum included total testosterone (TT), oestradiol (E2), follicular stimulating hormone (FSH), luteinizing hormone (LH) and insulin like factor 3 (INSL3), and sex hormone binding globulin (SHBG). The results showed that in the young men PFAS mixture was significantly associated with reduced levels of E2, which was accompanied by inverse association with E2/TT ratio; and PFUnDA was identified as the main contributor. However, the overall associations between PFAS mixture and other reproductive hormones were generally non-significant (Luo et al., 2021). The second study (Pan et al., 2019a) investigated associations between PFAS measured in matched semen and serum samples with markers of semen quality. The results showed an adverse association between serum 6:2 Cl-PFESA and semen quality (Pan et al., 2019a). The third study (Cui et al., 2020) used the same study population as Pan et al. (2019a) and reported inverse associations between plasma concentrations of PFOA, PFNA, PFOS, and 6:2 ClPFESA) with total testosterone and indicated inverse associations also for other reproductive hormones (Cui et al., 2020). In females, two prospective studies examined associations between PFASs and menopause/hormone levels during the menopausal transition, and both used data from the US Study of Women’s Health Across the Nation (SWAN). In the first paper, 1 120 women with 17-years follow up (5 466 person-years), all PFAS concentrations (linear PFOS, sum of branched PFOS isomers, linear PFOA, and PFNA) were classified into four PFASs clusters (low, low–medium, medium–high, and high) using k-means clustering to account for PFASs mixtures. Compared with the low exposure cluster, women in the high cluster had significantly increased risk, i.e. equivalent to 2 years earlier median time to natural menopause (Ding et al., 2020). In the second paper, associations between serum concentrations of 11 PFASs at baseline (1999-2000) and longitudinal serum concentrations of follicle-stimulating hormone (FSH), oestradiol, testosterone, and sex hormone-binding globulin (SHBG) at baseline and through 2015-2016 were examined in 1 371 women. PFOA and PFOS were positively associated with FSH, while PFNA and PFOA were inversely associated with oestradiol (Harlow et al., 2021). The results from both papers thus indicate that PFASs (linear PFOS, sum of branched PFOS isomers, linear PFOA, PFNA, PFHxS) are associated with earlier time of menopause and linear-PFOA, linear-PFOS, sum of branched PFOS isomers, total PFOS and N-EtFOSAA are positively associated with FSH, and PFNA and PFOA are associated with lower oestradiol-2 concentrations. The magnitude of the increase in TSH and decrease in oestradiol-2 was estimated to correspond to 2.5 and 4 years of aging (earlier age at menopause), respectively. 179 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) A prospective study in the UK ALSPAC cohort observed no associations between a range of prenatal PFAS exposures, including PFOS, PFOA, PFHxS, and PFNA, and Anti-Müllerian hormone, a marker of ovarian reserve in female offspring at 15 years of age (Donley et al., 2019). A nested case control study within the Ronneby cohort in Sweden examined high/low exposure to PFASs through contaminated drinking water and polycystic ovarian syndrome (PCOS), uterine leiomyoma, and endometriosis. Ronneby is an ecological cohort where one third of all inhabitants in Ronneby municipality were exposed to high PFASs (mainly PFHxS and PFOS) through contaminated drinking water between 1985 and 2013. Exposure to high levels of PFASs in drinking water was associated with increased risk of PCOS and possibly uterine leiomyoma, but not with endometriosis (Hammarstrand et al., 2021). The review of new studies provides limited suggestive evidence for an association between PFASs and adverse effects on male and female fertility. Sex hormones and related outcomes in infants In the updated literature search, two prospective investigations of PFASs and anogenital distance (AGD), one cross-sectional study on concentrations of oestrogens in cord blood, and one prospective study of serum hormones and hormone precursors in infants were identified. In a study from the Faroe Islands maternal exposure to PFOA, PFOS, PFNA and PFDA, but not PFHxS, were significantly associated with longer AGD in male infants (n = 232) (Christensen et al., 2021). A study within the Canadian MIREC cohort did not find any clear evidence that maternal plasma concentrations of PFOS, PFOA or PFHxS were associated with infant AGD in either male (n = 205) or female (n = 196) infants (Arbuckle et al., 2020). In infants with a mean age of 3.9 months in the Odense Birth cohort, there were no associations between five maternal PFASs (PFOS, PFOA, PFHxS, PFNA, PFDA) individually or combined with any of five sex hormones measured in boys (Jensen et al., 2020b). In girls, maternal PFDA concentration was associated with reduced dehydroepiandrosterone (DHEA). No associations were found for other PFASs or for the five PFASs combined. A study from China (Liu et al., 2021a) reported signific ant cross-sectional associations between six PFASs (PFOA, PFDA, PFNA, PFOS, PFHxS, 6:2 Cl-PFESA) and three oestrogen concentrations (oestrone (E1), oestradiol (E2), oestriol (E3)) in newborn cord-blood. Results from analyses stratified by infant sex showed that in boys five PFASs (PFDA, PFNA, PFOS, PFHxS and 6:2 Cl-PFESA) were positively associated with oestrone, four PFAS (PFDA, PFNA, PFOS and 6:2 Cl-PFESA) were associated with oestradiol, and PFOA and PFDA were positively associated with oestriol (Liu et al., 2021a). The results in girls were less clear. Of the six PFASs, only PFNA was positively associated with oestrone, PFOA, PFDA and PFNA were positively associated with oestradiol, and no PFASs were associated with oestriol. The review of new studies provide limited suggestive evidence for an association between PFASs and adverse effects on sex hormones and related outcomes in infants. Timing of puberty One new prospective study on pubertal timing was identified in the literature search. In the US Project Viva cohort, mid-childhood plasma PFASs were significantly associated with markers of delayed puberty (less pubertal development in early adolescence and older age at peak height velocity) in girls, but not in boys. In girls, the associations were significant for PFOA, PFOS, and PFDA, not significant for PFHxS and N-MeFOSAA, and borderline significant for an overall PFAS mixture (Carwile et al., 2021). The new study showed a consistent association between exposure to PFASs and delayed timing of puberty in girls 180 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) but no association with puberty in boys. Miscarriage and preeclampsia EFSA (2020) concluded that “there is limited evidence to suggest any association between PFASs and miscarriage”. Furthermore, EFSA concluded on insufficient evidence to suggest that PFASs are associated with pregnancy induc ed hypertension or preeclampsia (EFSA, 2018; EFSA, 2020). The new search identified two studies that examined associations between PFAS e xposure and the risk of pregnancy loss. A prospective study in China in women undergoing in vitro fertilisation-embryo transfer (IVF-ET) measured twelve PFASs measured in serum samples prior to IVF-ET treatment. There were no clear associations between the PFASs and early pregnancy loss (Wang et al., 2021b). A nested case control study within the Danish National Birth Cohort examined associations between plasma concentrations of seven PFASs in early pregnancy (mean gestational week 8) and miscarriage in gestational weeks 12-22. The results showed that PFOA, PFHpS and the PFAS mixture (weighted quantile sum, WQS) were associated with increased risk of miscarriage, particularly in parous women. The authors stated that larger studies among nulliparous women are needed to diminish the potential confounding by reproductive history (Liew et al., 2020). In a case-control study in Sweden, serum samples collected early in pregnancy were obtained for 296 women diagnosed with preeclampsia (cases) and 580 healthy controls and analysed for PFOA, PFOS, PFNA and PFHxS. In the adjusted analyses, there were no associations between PFASs and preeclampsia (Rylander et al., 2020). The review of new studies are in line with EFSA (2020) that the evidence for an association between PFASs and miscarriage and preeclampsia is still limited. Preterm delivery EFSA (2020) concluded that “the evidence for an association between PFASs and preterm delivery is limited”. In the Chinese Guangzhou Birth Cohort, both maternal serum 6:2 ClPFESA and PFOS were associated with significantly higher risk of preterm delivery (Chu et al., 2020). In contrast, in a nested case-control study in China there were no associations between maternal PFASs concentrations and spontaneous preterm delivery. However, PFOS and 6:2 Cl-PFESA were both positively associated with one of three markers of inflammation measured in the same blood samples as PFAS (Liu et al., 2020c). The review of new studies are in line with EFSA (2020) that the evidence for an association between PFASs and preterm delivery is still limited. B.5.3.2.2. Developmental outcomes EFSA (2018) and EFSA (2020) concluded that there “may well be a causal association between PFOS and PFOA and birth weight” but that for PFASs, which were found in lower concentrations than PFOS and PFOA, the reported associations were inconsistent. Birth weight Seven new birth cohort studies published after July 2019 have continued to report associations for maternal PFOS and PFOA with reduced birth weight, birth length and/or foetal growth restriction (Chen et al., 2021a; Chu et al., 2020; Kashino et al., 2020; Souza et al., 2020; Wikstrom et al., 2020; Xiao et al., 2020). A study from the Swedish Ronneby cohort published after the search has also been reviewed (Engström et al., 2022). This is an ecological cohort where one third of all inhabitants in Ronneby municipality were exposed to high PFASs (mainly PFHxS and PFOS) through contaminated drinking water between 1985 and 2013. Engström et al. (2022) compared infants with mothers exposed 181 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) to PFASs contaminated drinking water (high exposure) and infants born to mothers not exposed to contaminated drinking water (low exposure) to infants born to mother in another county (referents). High exposure to PFASs was associated with infant birthweight in a sex-specific manner and was inversely associated with birthweight in boys, and positively associated with birthweight in girls. Other studies have also reported inverse associations with birth weight for PFNA, PFDA, PFOSA, PFHxS and PFUnDA, but the results are not consistent across all studies. In the Chinese Guangzhou Birth Cohort, a study in 372 mother-child dyads showed that the PFOS alternatives 6:2 Cl-PFESA and 8:2 Cl-PFESA in maternal serum were inversely associated with birthweight with clinically relevant effect sizes (Chu et al., 2020). Overall, the new literature on associations between PFASs and birth weight support the conclusion from EFSA (2020) that there may well be a causal association between PFOS and PFOA and birth weight and strengthen the concern that other PFASs exert similar effects. Birth defects With regard to birth defects, EFSA (2018) reviewed several studies from the C8 cohort in the US, conducted in residents exposed to PFAS contaminated drinking water. There were no consistent findings between prenatal PFAS exposures and birth defects. The new literature search identified only one new study, which is a nested case-control study conducted in a cohort of 11 578 new-borns in Guangzhou, China (Ou et al., 2021). Prenatal maternal serum PFASs and cord-blood PFASs were compared between 158 cases (newborn with congenital heart defects) and 158 healthy controls. The results showed that gestational exposure to most PFASs, especially linear PFOS, 6 m-PFOS, PFDA, and PFDoA, were associated with increased risks for congenital heart defects. Neurodevelopmental outcomes EFSA (2020) concluded that “there is insufficient evidence to suggest that prenatal exposure to PFASs may adversely affect neurobehavioral development.” With regard to neurodevelopmental outcomes including emotional behaviour problems, attention deficit/hyperactivity (ADHD), autism spectrum disorder (ASD) and cognition, the updated search identified eleven new studies, nine prospective (Harris et al., 2021; Jedynak et al., 2021; Oh et al., 2021a; Oh et al., 2021b; Skogheim et al., 2020; Vuong et al., 2021a; Vuong et al., 2019; Vuong et al., 2021b; Yu et al., 2021b) and two case-control studies (Skogheim et al., 2021). A prospective study in US Project Viva, reported that prenatal PFOS, but not other PFASs, was associated with more child behaviour problems measured by parent and teacher assessment of children’s emotional behaviour problems at age 6 – 10 years (Harris et al., 2021). This study also reported cross-sectional associations for PFAS measured in the children, which showed that PFOA, PFOS, PFHxS, and PFDA were associated with parent rated behaviour problems, but not teacher rated problems, while PFNA was associated with both parent and teacher rated behaviour problems. In a prospective study including 708 mother-child pairs from five European cohorts, there were no significant associations for five PFASs (PFHxS, PFOS, PFOA, PFNA, PFUnDA) measured in maternal blood with child emotional behaviour problems assessed between three and seven years of age (Jedynak et al., 2021). Two prospective studies used data from the MARBELS cohort in the US. This cohort recruited mothers who previously had a child diagnosed with ASD and therefore had higher risk of having a child with delays or deficits in areas of development or behaviour. One study examined associations between prenatal PFAS exposures and children’s cognitive 182 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) development assessed at 6, 12, 24, and 36 months of age (Oh et al., 2021b). This study included 140 mother-child pair with neurocognitive data at all time points. Of six PFASs detected in more than 60% of the samples, PFOA was consistently associated with lower neurodevelopmental scores at 24 months, 36 months and in the repeated measure trajectory (Oh et al., 2021b). The other study examined associations between prenatal exposure to individual and combined PFASs and clinically confirmed ASD in children at 3 years of age and inc luded 173 mother-child pairs. The results showed that PFOA, PFNA, and the PFAS mixture (principal component) were associated with increased ASD risk (Oh et al., 2021a). Two studies used data from The Norwegian Mother, F ather and Child Cohort study (MoBa). PFASs were measured in blood around gestation week 18. The first study included 944 mother-child pairs and investigated the prospective associations between prenatal exposure to PFASs and symptoms of ADHD and cognitive functioning in a subgroup of children aged three and a half years (Skogheim et al., 2020). The children underwent extensive clinical assessment. The results showed no associations with ADHD symptoms, language skills or estimated IQ. However, PFOS, PFOA, PFHxS, PFHpS were negatively associated with nonverbal working memory, and PFNA, PFDA, PFUnDA were positively associated with verbal working memory in boys (Skogheim et al., 2020). The second study is a nested case-control study which included 821 ADHD cases, 400 ASD cases, and 980 controls. Diagnoses were obtained by linkage to the Norwegian Patient Registry. PFASs included PFOA, PFNA, PFDA, PFUnDA, PFHxS, PFHpS and PFOS. The results showed that for some PFASs (PFDA, PFOS, PFUnDA), as well as the mixture of all PFASs, there were inverse associations with ASD and/or ADHD, while PFOA was associated with a non-linear increased odd of ASD and ADHD in children (Skogheim et al., 2021). A study using data from the Shanghai Birth cohort examined the association between prenatal PFAS exposure measured in early pregnancy and foetal brain-derived neurotrophic factor (BDNF, an important factor in neurodevelopment) level in the umbilical cord blood (Yu et al., 2021b). Quantile-based g-computation was applied to explore the joint and independent effects of PFAS on BDNF level. The results showed that PFHxS exposure was significantly associated with an increased BDNF level. There was no association between the PFAS mixture and BDNF. Shin et al. used data from CHARGE (CHildhood Autism Risk from Genetics and Environment), a population-based case-control study comprising index children and their families, with a goal to identify causes and contributing factors for autism (Shin et al., 2020). The study included 450 mothers with blood sampled at the time of study enrolment (when the child was 2 – 5 years old) and 453 children who completed the study with a final diagnosis of ASD or typical development. The study lacked prenatal samples because of its case control design and therefore, both model-based reconstructed estimates during pregnancy and concurrent (post -diagnosis) measurements of nine PFAS concentrations were used as surrogate of in utero exposure. The analyses showed that modelled prenatal maternal PFHxS and PFOS exposure, but not other PFASs (PFOA, PFNA, PFDA, PFUA, PFDoDA, N-MeFOSAA and N-EtFOSAA), were associated with increased odds of child diagnosis of ASD. Three prospective studies used data from the US HOME cohort. One study investigated prenatal and childhood PFAS and cognitive abilities at eight years and whether sex modifies these associations (Vuong et al., 2019). PFOS, PFOA, PFHxS and PFNA were quantified in maternal serum during gestation and in child serum at thee and eight years. There were no adverse prospective associations between prenatal and childhood PFAS and cognitive function at eight years. Another st udy investigated prenatal PFAS and neurobehavior through eight years of age (Vuong et al., 2021a). PFOA, PFOS, PFHxS and PFNA were quantified in maternal serum during gest ation. Prenatal PFOS and PFNA were consistently associated with measures related to hyperactive-impulsive type ADHD. PFHxS was associated with increased problems for both externalizing and internalizing behaviours, while PFOA was not associated with any outcome (Vuong et al., 2021a). The third study 183 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) (Vuong et al., 2021b) investigated childhood concentrations of PFOS, PFOA, PFHxS and PFNA measured at three and eight years and neurobehavioral domains in children at eight years. The results showed that childhood serum concentrations of PFAS at ages three and six years were not associated with externalizing or internalizing behaviours in children at age eight years. However, there was suggestive evidence that child sex modified the associations between childhood PFAS and neurobehavior, where higher concentrations of PFOA and PFNA were associated with more externalizing problems, internalizing problems, and behaviour symptoms in boys compared to girls. Furthermore, this study confirmed that the previously reported associations between prenatal PFAS concentrations and more externalizing and internalizing problems in children reported by Vuong et al. (2021a) remain even after taking childhood concentrations into account (Vuong et al., 2021b). Overall, the new studies revealed some indication of adverse associations between prenatal PFAS exposures and neurodevelopment, and several studies reported sex differences. However, as for the studies reviewed by EFSA (2018) and EFSA (2020), the new studies reported inconsistent results with positive, no or inverse associations between PFASs and the outcomes. The most consistent associations were observed for prenatal PFOS and PFHxS exposure and child neurodevelopment. Other developmental outcomes EFSA (2020) concluded that “there is insufficient evidence to suggest that exposure to PFASs may adversely affect overweight”. For other developmental outcomes including growth in infancy and childhood, overweight, and metabolic risk factors, the updated search identified one cross-sectional (Spratlen et al., 2020), one case-control study (Martinsson et al., 2020) and six prospective studies (Blomberg et al., 2021; Carwile et al., 2021; Jensen et al., 2020a; Mora et al., 2018; Shih et al., 2022; Starling et al., 2019; Tian et al., 2021). In a longitudinal study, prenatal PFASs were associated with sex-specific and chemical specific differences in infant weight and adiposity up to 5 months of age (Starling et al., 2019). A study in the Odense Child cohort examined PFASs in maternal pregnancy serum (PFOS, PFOA, PFHxS, PFNA and PFDA) and repeated markers of adiposity and lipid metabolism in infancy (up to 18 months). The results showed that PFNA and PFDA were positively associated with several markers of adiposity and higher total cholesterol (Jensen et al., 2020a). In a case-control study by Martinsson et al. (2020), none of the PFASs investigated (PFOS, PFOA, PFHxS and PFNA) were associated with obesity at age 4 years. A study within the Faroe Islands birth cohort reported consistent associations between PFASs concentrations in serum collected at age 18 months, 5 years, and 9 years with serum adipokine hormones measured in the serum at 9 years (Shih et al., 2022). PFOS, PFOA, PFHxS, PFNA and PFDA at 18 months, 5 years and 9 years were associated with leptin, leptin receptor, and resistin at age 9 years, and the results were also evident for a PFAS mixture. Three prospective studies reported that prenatal PFAS exposure were associated with disrupted lipid metabolism in neonates/children, but the results were inconsistent between studies and between different PFASs (Blomberg et al., 2021; Mora et al., 2018; Tian et al., 2021). Tian et al. (2021) included Bayesian Kernel Machine Regression (BKMR) models, which showed that the mixture of eight PFASs (PFHxS, PFOS, PFOA, PFNA, PFDA, PFUdA, PFDoA, PFTrDA) were inversely associated with all lipid concentrations, i.e., total cholesterol, triglycerides, HDL-C, and LDL-C. PFASs were among the pollutants suspected to have been released during the collapse of the World Trade Center on 9/11/2001. A cross-sectional study in pregnant women exposed to the World Trade Center disaster reported that several cord-blood PFASs were associated with higher cord-blood lipid levels, including evidence of a strong linear trend between triglycerides and both PFOA and PFHxS (Spratlen et al., 2020). 184 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Overall, the new studies indicate associations between multiple prenatal PFASs and other developmental outcomes, including growth in infancy and childhood, overweight, and metabolic risk factors in children. However, the evidence is still limited. B.5.3.3. Carcinogenicity (epidemiological evidence) A number of studies on cancer incidence or cancer mortality and occupational or environmental exposure reviewed by EFSA (2018) provided insufficient support for carcinogenicity of PFOS and PFOA in humans. The studies reviewed by EFSA in 2020 did not change the previous conclusion for PFOS and PFOA. For other PFASs, limited information was identified (EFSA, 2020). Six studies were identified in the new literature search, of which five included breast cancer (Cohn et al., 2020; Itoh et al., 2021; Mancini et al., 2020; Tsai et al., 2020), one also included ovarian, uterine and prostate cancer (Omoike et al., 2021), and one study included renal cell carcinoma (Shearer et al., 2021). A study published after the main search was also included because it was highly relevant for this narrative review (Li et al., 2022a). Cohn et al. (2020) used prospective data from a US cohort and examined nested casecontrol associations between in utero PFAS concentrations and risk of breast cancer over 54 years follow-up. The results showed that high maternal N-EtFOSAA, a precursor of PFOS, in combination with high maternal total cholesterol predicted a 3.6-fold increased risk of breast cancer in daughters. Conversely, maternal PFOS was associated with decreased daughters’ breast cancer risk. No associations were reported for PFHxS and PFOA and breast cancer. A prospective nested case control within the French E3N cohort (Mancini et al., 2020) reported no association between PFOS and PFOA and overall risk of breast cancer. However, the results showed positive linear associations between PFOS and the risk of receptor-positive (ER+ and PR+ tumours). There was no association between PFOA and receptor-positive breast cancer risk. In Taiwan, Tsai et al. (2020) enrolled 120 cases with breast cancer and 119 controls and found no significant associations between PFASs and breast cancer (Tsai et al., 2020). After stratifying for age, a positive association for PFOS were observed for the <50 years age group. After stratifying for the oestrogen receptor status and age group, PFHxS and PFOS concentrations were associated with higher risk of ER positive tumours in the <50 years age group, while PFNA and PFDA exposure was associated with lower risk of ER positive tumours in the <50 years age group. In a Japanese case-control study, 20 included PFAS congeners (PFHxS, PFHpS, linear and five branched PFOS, linear and branched PFOA, linear and branched PFNA, linear and branched PFDA, linear and branched PFUnDA, linear and branched PFDoDA, linear and branched PFTrDA) were all significantly inversely associated with risk of breast cancer (Itoh et al., 2021). In a cross-sectional study within the US NHANES, Omoike et al. (2021) reported that PFOA, PFOS, PFHxS, and PFDA were associated with ovarian cancer (OR range: 1.01-1.03) and PFOA, PFOS, PFNA. PFHxS and PFDA were associated with breast cancer (OR range: 1.011.09). PFDA and PFNA were inversely associated with prostate cancer, and PFDA, PFNA and PFHxS were inversely associated with uterine cancer. Shearer et al. (2021) used prospective population-based data in the US for nested case control study (324/324) examining the associations between seven PFASs (PFOA, PFOS, PFHxS, PFNA, N-EtFOSAA, N-MeFOSAA and PFDA) and risk of renal cell carcinoma (Shearer et al., 2021). The results showed a statistically significant positive exposure-response association between prediagnostic serum PFOA concentrations and subsequent risk of renal 185 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) cell carcinoma. Cancer incidence in a Swedish cohort with high exposure to perfluoroalkyl substances in drinking water caused by use of firefighting foams at a nearby airport was reported (Li et al., 2022a). This prospective ecological study in the Swedish Ronneby cohort included more than 60 000 individuals who ever lived in the Ronneby municipality during the period of PFAS contamination (in particular PFHxS and PFOS) of drinking water between 1985 and 2013. The study population and study period was fully covered by the Swedish Cancer Register. The results showed that high PFASs exposure was not associated with an overall higher risk of cancer, but that the highly exposed group had moderately increased risk of kidney cancer (HR 1.27; 95% CI 0.85, 1.89) and bladder cancer (HR 1.32; 95% CI 1.01– 1.72). The results for kidney cancer are in accordance with previous findings for PFASs exposure in the C8 study. In summary, due to inconsistent findings, the new studies do not provide clear support for an association between PFASs and breast, ovarian and prostate cancer. However, two new prospective studies strengthen the evidence that PFOA and multiple PFASs exposure are associated with renal cell carcinoma and kidney cancer. B.5.3.4. Other outcomes (epidemiological evidence) B.5.3.4.1. Bone mineral density There were two studies related to bone mineral density in EFSA (2018) and EFSA (2020), and EFSA concluded that there was insufficient evidence for associations between exposure to PFASs and low bone mineral density (BMD) or osteoporosis. The updated search identified one new study which examined associations between PFASs and BMD (Hu et al., 2019). In this randomised intervention study with four energy restricted diets assessing long-term weight loss in healthy overweight or obese individuals, cross-sectional and prospective associations between five PFASs (PFOS, PFOA, PFHxS, PFNA and PFDA) and BMD at six bone sites at baseline and after two years in participants aged 30-70 years were examined. A BMI related genetic risk score was calculated on the basis of 97 single nucleotide polymorphisms (SNPs). In cross-sectional analyses, inverse associations were observed between baseline PFOS and PFOA concentrations and BMD at all bone sites, and inverse associations were also indicated for the other PFASs. In prospective analyses, baseline concentrations of PFOS, PFNA and PFDA were associated with a faster decline in BMD at some bone sites, and PFOA concentration was associated with a faster decline in BMD at one bone site. These associations were independent of weight changes during the trial (Hu et al., 2019). PFASs-related declines in BMD were influenced by genetic predisposition to obesity. More studies are needed to determine potential associations between PFAS exposures and accelerated bone loss. B.5.3.5. General conclusion from new epidemiological data Immune outcomes were chosen as the critical effect for risk assessment of PFASs in the EFSA 2020 Opinion (EFSA, 2020) and inverse relation between PFASs and vaccination response was also concluded by ATSDR (2021). The new studies identified in the literature search strengthen the evidence for an association between exposure to several PFASs, in particular PFOS, PFOA, PFHxS, PFNA, PFUnDA, and PFDA, and reduced vaccine responses in children. The adversity of the immunosuppressive effects of PFASs is f urther supported by increasing evidence for an association also with increased propensity of lower respiratory tract infections (LRTI) and possibly also other infectious diseases. The evidence for associations with asthma- and allergy-related outcomes (hypersensitivity) remain weak, but there is increasing evidence for an inverse association between PFASs and atopic dermatitis. 186 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Liver toxicity and metabolic disruption are also outcomes of concern in relation to PFAS exposures. The evidence is still considered strong for a positive association between PFASs and plasma levels of the liver enzyme ALT, which is a marker of liver toxicity and fatty liver diseases. Furthermore, the new publications support the association between PFOS and PFOA and increased total and LDL-cholesterol in adults. Similar associations are suggested also for PFHxS, PFNA, PFDA and PFUnDA, 6:2 Cl-PFESA and 8:2 Cl-PFESA. However, the potential contribution of confounding due to similar intestinal reuptake mechanisms for PFASs and bile acids is still not clear. In the EFSA 2020 opinion, no clear association between PFOS or PFOA and cardiovascular disease was found (EFSA, 2020). However, recent studies strengthen the evidence for a positive association. The review of new studies provides limited suggestive evidence for associations between PFASs and adverse effects on reproduction and development, including male and female fertility, miscarriage and preeclampsia, delayed timing of puberty, adverse effects on sex hormones and related outcomes in infants. The evidence for an association between PFASs and preterm delivery is still limited, however, the new literature supports the conclusion from EFSA (2020) that there may well be a causal association between PFOS and PFOA and birth weight and strengthen the concern that other PFASs exert similar effects. Overall, the new studies revealed some adverse associations between prenatal PFAS exposures and neurodevelopment, and several studies reported sex differences. However, as for the studies reviewed by EFSA (2018) and EFSA (2020), the new studies reported inconsistent results. Regarding cancer risk, two new prospective studies strengthen the evidence that PFOA and multiple PFAS exposures are associated with renal cell carcinoma and kidney cancer. However, the new studies do not provide clear support for an association between PFASs and breast, ovarian and prostate cancer. With respect to other organ toxicities, the recent publications strengthen the concern that PFAS exposures might be associated with thyroid disease or changes in thyroid hormones, whereas potential associations with accelerated bone loss are still unclear. New studies strengthen the support for positive associations between PFASs and outcomes related to reduced kidney function, but confounding and reverse causation is plausible and thus a causal association between PFAS exposures and kidney function is uncertain. 187 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5.4. Human data on oligomeric/polymeric PFASs Inhalation of aerosolized waterproofing agents, sealants and ski waxes containing polymeric PFASs as well as inhalation of pyrolysis products of fluoropolymers (e.g. during manufacture or use) is reported to cause respiratory illness, such as acute chemical pneumonitis, reactive airway dysfunction syndrome, pulmonary oedema, occasionally accompanied by non-specific systemic symptoms, such as fever, chills, malaise, arthralgias, and nausea (influenza-like syndrome, polymer fume fever) in humans (Greenberg and Vearrier, 2015; Hays and Spiller, 2014). These effects are of unclear etiology but demonstrate a toxicological relevance of polymeric PFASs and their degradation products in acute inhalation exposure scenarios. For example, reduced lung function after occupational exposure of professional ski waxers, can be explained by particle exposure (Dahlqvist et al., 1992) or by exposure to degradation products from ski waxes, e.g. PFCAs (Freberg et al., 2010; Freberg et al., 2016; Nilsson et al., 2013b). Case studies of accidents with acute excessive inhalation exposure to f luoropolymers report severe symptoms of pulmonary oedema and polymer fume fever, e.g. after occupational inhalation of fumes from heated PTFE in a plastic factory (Lee et al., 1997) or after smoking ski wax-contaminated cigarettes (Strøm and Alexandersen, 1990). Occupational long-term exposure (for 28 years) to PTFE spraying can cause granulomatous lung lesions such as pneumoconiosis (Lee et al., 2018a). The authors conclude, such lesions appear to be caused not by the degradation products of PTFE from high temperatures but by spraying the particles of PTFE. Scanning electron microscopic (SEM) features of the lesion revealing fluorine elements show multiple round to oval granular material measuring 2 – 6 μm. Further results in comparison with standard PTFE and PTFE spray solution as used in the factory showed the presence of PTFE in the lung tissue. The particles found in the personal samples measured 1 – 22 µm by SEM, but particles smaller than 1 μm were also found. Choi et al. (2014) also report three cases of small airway-centered granulomatous lesions in workers employed at facilities that apply coatings to pans and other utensils. The workers were repeatedly exposed (inhalational PTFE exposure was between 7 and 20 years) to PTFE particles that were probably generated by the drying process when PTFE coatings are dried in a convection oven at high temperatures (380-420 °C). PTFE is self-classified as STOT RE 1 (respiratory system/inhalation) by one notifier14. Moreover, when polymeric PFASs, such as PTFE and PCTFE are heated to very high temperatures above thermal stability, they may emit toxic decomposition products, such as tetrafluoroethylene (TFE), which itself is not in the scope of the proposed restriction (Huber et al., 2009). TFE is self-classified as STOT RE 2 (kidney) by one notifier and has a harmonised classification as Carc. 1B15 and was reported to be probably carcinogenic to humans (IARC, 2014). 14 https://echa.europa.eu/de/information-on-chemicals/cl-inventory-database/- /discli/details/70383, date of access: 2022-06-02. 15 https://echa.europa.eu/de/substance-information/-/substanceinfo/100.003.752, date of access: 2022-06-02. 188 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5.5. Combined toxicity Due to their persistence, mobility, bioaccumulation potential and thus tendency for longrange transport, PFASs are ubiquitously dispersed in the environment. Therefore, many different PFASs are co-occurring in the environment, drinking water, and food, resulting in a combined exposure to multiple PFASs that, based on the available data on some groups of PFASs, show similarity of effects. Accordingly, an assessment of hazards and risks taking into account such combined exposure would reflect exposure conditions more realistically than single compound assessments. A precise modelling of combined effects of all PFASs in the scope of this restriction proposal is realistically not achievable because of lack of data on toxicokinetics, toxicodynamics, slope of dose response curves as well as limited knowledge of the mode -of action (Borgert et al., 2004). For most PFASs, no data on effects are available. However, the lack of data should not preclude to consider the risks and hazards from combined exposure to different PFASs because of the following: First, it has been demonstrated in multiple studies that concentration addition may give a realistic worst-case estimation of combined toxicities for risk assessment procedures even if similarity of components is unknown (Backhaus et al., 2000; Martin et al., 2021). Therefore, Backhaus and Faust (2012) suggested to apply concentration addition as a precautious first tier, irrespective of the modes/mechanisms of action of the mixture components. Dose-addition has also been adopted as the default assessment approach in EFSA’s “Guidance on harmonised methodologies for human health, animal health and ecological risk assessment of combined exposure to multiple chemicals” (EFSA, 2019a). In recent studies, it has been demonstrated that cytotoxicity results of mixtures of PFASs were approximately additive (Hoover et al., 2019) or even more than additive (Ojo et al., 2021). Given similar effect patterns for many PFASs in animal studies, such as effects on liver, thyroid hormone system, and immune system, additive effects may be considered as realistic worst-case estimation. Second, a common mode of act ion is not a prerequisite for grouping chemicals for a combined exposure assessment and thus evidence for the same target organ is considered sufficient (EFSA, 2019a). It has been demonstrated in vitro that very diverse chemical classes with different molecular mechanisms can act in a concentration additive manner if triggering a similar adverse outcome (Stalter et al., 2020). Third, even low or insignificantly toxic concentrations of the individual components can add up to observable combined effects (Altenburger et al., 2013). This something-from“nothing” phenomenon has been first observed for endocrine disrupting compounds (Silva et al., 2002) but applies also to other mechanisms of action (Versieren et al., 2016). Combined exposure to multiple PFASs is documented in an increasing number of epidemiological studies, many of which demonstrate an association between increased exposure to PFAS mixtures and increased incidences of various health outcomes, such as immunotoxicity, metabolic effects, developmental toxicity, thyroid hormone effects, and others (see B.5.3). As conc luded by Rosato et al. (2022), specific guidelines and tools for the assessment of mixture observational studies are warranted. In vivo studies on effects of PFAS mixtures are scarce. Recently Marques et al. (2021) demonstrated cumulative mixture effects of PFOA, PFOS, and PFHxS on metabolic endpoints in mice. Combined effects of various PFASs have been suggested before for the endpoint immunotoxicity (McDonough et al., 2020b). For liver effects, a risk assessment of mixtures of 16 PFASs has been applied by Bil et al. (2021) by use of a relative potency approach (RIVM et al., 2018). In this study, the authors derived relative potency factors (RPFs) for 14 PFAAs and 2 PFAA precursors taking into account potency for liver effects relative to PFOA. Based on the RPFs and the combined 189 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS exposure in two separate exposure scenarios (drinking water ingestion and fish consumption), total PFOA equivalents did not exceed the drinking water limit for PFOA in The Netherlands of 87.5 ng/L, whereas the total PFOA equivalents resulting from fish consumption did exceed the fish c onsumption limit for PFOA in the Netherlands of 1.5 ng/g wet weight. This example demonstrates the high relevance of combined exposure assessments of PFASs. It can be expected that more limit values are exceeded in other exposure scenarios and when other endpoints are considered, in particular when more PFASs are taken into account. In an earlier study, a cumulative health risk assessment of 17 PFASs has been applied by use of the Hazard Index (HI) approach (Borg et al., 2013). The risk characterization showed a concern for hepat otoxicity and reproductive toxicity in a subpopulation eating PFOS-contaminated fish, illustrating that high local exposure may be of concern. For the occupationally exposed (e.g. ski waxers) there was concern for hepatotoxicity by PFOA and all congeners in combination as well as for reproductive toxicity by all congeners in combination. The immense number of PFASs in addition to the fact that appropriate toxicological data are not available for the vast majority of them, renders approaches for combined ris k assessments unattainable for all the PFASs within the scope of this restriction proposal. Additionally, a large fraction of total organic fluorine from various exposure scenarios cannot be accounted for by commonly analysed PFASs indicating the co -occurrence of unidentified PFASs (Borg and Ivarsson, 2017). In conclusion, it is emphasized that combined exposure to different PFASs affecting the same target organs may result in combined additive effects, rendering exceedance of effect thresholds or limit values more likely than single compound assessments. 190 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.5.6. Derivation of DNEL(s)/DMEL(s) Derivation of DNELs/DMELs is not considered relevant for this dossier since PFASs should be treated as non-threshold substances for the purposes of risk assessment. Any release of PFASs to the environment (see B.9) can be regarded as an unacceptable risk to the environment and human health (similar to PBT/vPvB substances under the REACH regulation). Since the non-threshold approach requires minimisat ion of exposures/releases, there is no need to compare in quantitative way exposure levels with effect thresholds (i.e. DNEL or DMEL values)(Wilkenfeld RM, 1981). Moreover, given the large number of substances covered, which are to a large extent still insufficiently researched in terms of human toxicity, it is considered impossible to establish a safe level for individual PFASs, let alone for the group in total. There is increasing evidence that effects of well-studied PFASs occur at lower levels than previously anticipated (e.g., EFSA, 2020), and that exposure to these few PFASs already exceeds existing limit values. Hence, any additional exposure to other PFASs will result in more severe effects and/or a larger part of the population exceeding the limit values. Exposure therefore needs to be minimised, as per non-threshold approach. 191 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.6. Human health hazard assessment of physicochemical properties Not relevant for this proposal. 192 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7. Environmental hazard assessment B.7.1. Ecotoxicity B.7.1.1. Notes on the procedure PFASs represent one of the most rapidly increasing study fields in ecotoxicology, contributing with several hundred new scientific publications in each of the recent years. To get an overview over the scientific literature, investigating ecotoxicological effects of PFASs, an initial literature search was performed by using the search term PFAS* and effect* in isi web of knowledge 16 in early 2021. The initial search was restricted to review papers. Further refinement of the search results was performed by screening titles and abstract. The initial literature search resulted in a list of 76 papers. After screening the titles for relevance, 20 papers remained. Two of these were only available in Chinese and were excluded. The remaining 18 review papers were downloaded in full text and their relevance for the present review work was assessed. Not all of the 18 review papers were considered useful and those are consequently not mentioned in this chapter. The collection of studies from the initial search was complemented non-systematically by other studies that came up during the process of writing. The PFASTox Database 17 was used to get an overview over the ecotoxicological threshold values which are being reported by scientific literature. Functional details, as well as information on quality assurance for the data compiled in that database are provided by Pelch et al. (2019). All entries for in vitro and in vivo studies except those of the effect category “endocrine disruption” (those are already included in section B.7.5) were downloaded as a study list generated from the database. The full study list was filtered to exclude entries with a focus on human health (e.g. studies with human cell lines, rats or mice). The effects of PFASs on human health are assessed in section B.5. The abstract entries of all references in the filtered study list were screened for threshold values which were then collected in a separate table (see Table B.14). In a review paper from Ankley et al. (2021), Ankley and colleagues summarized the information on ecological effects of PFASs from the US EPAs ECOTOX Knowledgebase 18 according to environmental compartments as well as different species with environmental relevance. The main findings from that compilation will be presented below together with complementary data from other studies not mentioned in the review from Ankley and colleagues. B.7.1.2. Aquatic invertebrates According to Ankley et al. (2021), crustaceans are often the most sensitive aquatic invertebrates to be affected by PFASs. Furthermore, they observed, that “toxicity in the same PFAS class tends to increase with increased fluorocarbon chain length”. Other authors observed a similar pattern in their investigation of the effects of PFBA, PFHxA and PFOA on Daphnia magna (Barmentlo et al., 2015). Ankley et al. (2021) found, that toxicity values (e.g. ECX) from chronic exposure were mostly within the same order of magnitude as values from acute exposure for the same species. Effects on the development of invertebrates occur at lower levels concentrations than effects on growth and reproduction. Studies with saltwater organisms are rare, compared to studies with freshwater organisms. Ankley et al. (2021) concluded that “marine invertebrates tend to show a higher sensitivity to PFOA and PFOS” compared to freshwater species. However, they highlighted that the data at hand is limited, and that the conclusion is subject to a high degree of uncertainty. 16 www.isiwebofknowledge.com, date of access: 2022-12-01. https://pfastoxdatabase.org/, date of access: 2021-10-07. 18 https://www.epa.gov/chemical-research/ecotoxicology-database, date of access: 2022-12-01. 17 193 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) In contrast, Mhadhbi et al. (2012) reported in their study on the effects of PFOA and PFOS on marine species from different trophic levels that effects occurred in marine species at higher concentrations compared to freshwater species suggesting that marine species are less sensitive. B.7.1.3. Terrestrial invertebrates Ankley et al. (2021) observed that ecotoxicological studies on the effects of PFASs with terrestrial invertebrates are scarce, compared to the number of studies with aquatic invertebrates. Thus, firm conclusions regarding the differences of the ecotoxicological effects of different PFAS groups in terrestrial invertebrates are not possible. The same applies for interspecies differences. However, Ankley and colleagues concluded that developmental toxicity in terrestrial invertebrates increases with carbon chain length within the same PFAS group. Additionally, they observed a trend regarding the toxicity of different PFAS groups which they described as perfluoroalkane sulfonamides (FASAs) > Perfluoroalkane sulfonic acids (PFSAs) > Perfluoroalkylcarboxylic acids (PFCAs) > fluorotelomer alcohols (FTOHs). Ankley and colleagues suggested a few different mechanisms through which PFASs exert their toxic effects in invertebrates (terrestrial as well as aquatic): oxidative stress which affects the antioxidative defense systems; genotoxic effects such as DNA strand breaks, chromosomal breaks or apoptosis; neurotoxic effects expressed e.g. via altered brain morphology; metabolic effects e.g. fat accumulation due to alternations of different regulating pathways (e.g. interaction with peroxisome proliferator-activated receptors (PPARs)); effects on the immune system via adversely affecting immune -related cell viability. B.7.1.4. Fish Most of the available literature (90%), that was evaluated by Ankley and colleagues regarding the effects of PFASs on fish, focuses on PFCAs and PFSAs. Similar to aquatic invertebrates, most studies were performed with freshwater species (95% of the evaluated studies) while studies investigating the effects of PFAS on marine fish are rare (5%). Regarding acute toxicity, Ankley and colleagues observed a lower acute toxicity of PFASs compared to invertebrates. Based on the order of magnitude of ecotoxicological threshold values in Cyprinids, C4 PFAAs appear to be less acutely and chronically toxic than PFAAs with chain lengths > C6. Addit ionally, Ankley and colleagues concluded that “within the same chain length (C8), sulfonates are typically more toxic than carboxylates”. Similar results, regarding the relation of toxicity and chain length as well as a higher toxicity for sulfonic PFAAs were also reported in an earlier study by Ulhaq et al. (2013). They investigated the effects of 7 PFAAs (TFA, PFBA, PFOA, PFNA, PFDA, PFBS, PFOS) on zebrafish. Rericha et al. (2021) investigated behavioral effects of 58 different PFASs on early life stage zebrafish as means to assess their developmental toxicity. They observed that at least 2 substances each from the groups of PFCAs, PFSAs, phosphate ethers and phosphinic acids “induced larval behavior effects”. Regarding chain lengths of the investigated compounds they reported, that “PFCAs and PFSAs associated with abnormal larval behavior had 3−17 and 3−8 continuously fluorinated carbons (CFCs), respectively”. PFASs with unsaturated C-F chains did not cause behavioral effects in this study. Mortality was also not observed within the tests performed by Rericha and colleagues. PFOA and PFNA were the only two of the 58 PFASs tested, that caused morphological effects. In relation to chronic effects, Ankley and colleagues concluded, that NOEC or LOEC values for effects on reproduction were lower than values for effects on growth (based on data for PFOA and PFOS from studies with Cyprinids only). Regarding the mechanisms of toxicity, it was concluded, that “different PFAS have been shown to elicit oxidative stress and apoptosis in fish both in vitro and in vivo”. Other than that, Ankley and colleagues mentioned that different PFASs act toxic via an activation of nuclear receptors involved in lipid metabolism e.g. PPARs which is associated with an increased liver lipid content and 194 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) steatosis (fatty liver) in mammals, but hepatic effects from PFOS exposure have also been observed in zebrafish. B.7.1.5. Amphibians Acc ording to Ankley et al. (2021), studies investigating effects of PFASs on amphibians are only available for 9 different PFASs (PFOA, PFNA, PFDA, PFUnA, PFOS, PFHxS, 6:2 FTS, 10:2 FTCA, 10:2 FTUCA) and the majority of the studies investigated the effects of PFOS and PFOA. Similar to other species, toxicity of PFASs towards amphibians is influenced by the carbon chain length, as well as the functional group of the individual PFAS. However, the data at hand was not sufficient to derive trends. Ankley and colleagues describe that effects from PFASs on growth and development in early live stages have been reported for several species. B.7.1.6. Birds From the limited data available (acute and chronic toxicity on birds has only been investigated for PFOS, PFBS, PFHxS and PFOA) Ankley and colleagues concluded that results from acute toxicity studies with birds are comparable to those from studies with rodents: PFASs with C8 carbon chains are more toxic, than short -chain PFASs, and sulfonates are more toxic than carboxylates with the same chain length. Results from different studies with PFOS indicate that bird species show different sensitivities to the same substance, but results are within the same order of magnitude. In two recent studies, Dennis and colleagues investigated the liver and eggs of northern bobwhite quail (Colinus virginianus) individuals that were chronically exposed to PFASs (PFHxA, PFOS, or PFHxS or a mixture of PFHxA and PFOS or PFHxS and PFOS) (Dennis et al., 2022; Dennis et al., 2021). From the residue values from the most sensitive subgroup, they derived chronic toxicity values and found them to be lower than similar values reported for birds from earlier studies. The reported values are in the <50 ng/g ww range. Additionally, Dennis and colleagues observed that “PFOS and PFHxS were more bioaccumulative than PFHxA in avian tissues, but PFHxA was more toxic to reproducing birds than either PFOS or a PFOS:PFHxS mixture”. B.7.1.7. Reptiles Ankley et al. (2021) compiled 5 studies that investigated the effects of PFASs (namely PFOS and PFHxS) in reptiles. Effects from those studies encompass decreased growth of juveniles and decreased egg viability. B.7.1.8. Other species Laboratory experiments with other species (i.e. species not commonly used for ecotoxicological tests) examined the effects of PFASs on different endpoints, including immunological, neurological, and histopathological endpoints. Ankley et al. (2021) collated studies with polar bears, sled dogs, marine mammals, which in total investigated the effects of 9 PFCAs, 3 PFSAs, 3 FTOHs, and 3 “novel” PFASs (HFPO-DA, ADONA and 6:2 ClPFESA (major component of F-53B)). On the basis of this data, Ankley and colleagues concluded that “PFOS followed by PFOA has the greatest amount of toxicological data”. Regarding the toxicity of different PFAS groups, the data from other species shows similar patterns as the data from e.g. fish: PFCAs with chain lengths ≤ C6 appear to be less toxic than PFCAs with chain lengths from C8 – C12 (while PFTrDA (C14-PFCA) again was considered less toxic). Within the group of sulfonic acids, Ankley and collea gues also observed a trend in toxicity according to chain length: PFOS (C8; most toxic) < PFHxS (C6) < PFBS (C4; least toxic). Data for 6:2 FTOH and 8:2 FTOH suggests, that these compounds exert “less toxicity than other C6 PFAS compounds” Interspecies differences, e.g. regarding sensitivity, are difficult to evaluate in this context 195 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) due to limited data. Ankley and colleagues made an attempt considering only data for PFOS and concluded “Based on body weight/body weight gain effects, there were no discernible differences between species exposed to PFOS during gestation. Even when reproductive effects such as pup viability and/or weight gain are considered, PFOS effect levels between species were typically 10‐fold or less based on dose”. A review by Krafft and Riess (2015) observed that PFAAs with shorter chain lengths exert a lower toxicity, which concurs with observations from the studies mentioned previously. Additionally, they mentioned that “PFAS exposure has been consistently associated with lipid and carbohydrate metabolism disorders”, based on studies with exposed humans, which is in line with the observations from Ankley et al. (2021) regarding metabolism disorders in mammals. Accordingly, Krafft and Riess (2015) stated, that, concerning the toxicity of PFAAs, “the liver has been identified as a specific target organ.” but confined that these findings are “highly dependent on PFAS, dose, species, strain and gender”. Furthermore, Krafft and colleagues described effects of PFASs on the immune system as well as developmental effects e.g. a reduced antibody response and IgM antibody production in mice exposed to PFOS or PFOA, or developmental toxicity observed in adult rodents exposed in utero to PFOS, PFHxS or PFOA. Similar effects have been reported in the review from Ankley and colleagues for various species (see above). Several of the studies, compiled by Ankley and colleagues, measured tissue concentrations of PFASs in different wild living animals that were expose d to PFASs (and other contaminants) in the field. Subsequently, these studies tried to link the measured tissue concentrations to adverse effects. The findings of Ankley and colleagues regarding the effects of PFASs in wild living animals, as well as complementary results from other studies, are discussed in more detail in chapter B.7.2. B.7.1.9. Plants As reviewed by Li et al. (2022b) occurrence of PFASs rarely lead to obvious phenotypic/physiological damages in plants, but markedly perturb some biological activities at biochemical and molecular scales. According to Li et al. (2022b) PFAS exposure induces the over-generated reactive oxygen species and further damages plant cell structure and organelle functions. A number of biochemical activities in plant cells are perturbed, such as photosynthesis, gene expression, protein synthesis, carbon and nitrogen metabolisms. B.7.1.10. PFASTox Database The filtered study list from the PFASTox database (see “Notes on the procedure” at the beginning of this chapter) comprised 167 entries. Only 23 entries reported threshold values in the abstract. These are listed in Table B.14. Several studies reported threshold values for more than one PFAS thus the table contains more than 23 entries. 196 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.14. Ecotoxicological threshold values for different PFASs from scientific literature recorded in the PFASTox database. Additional information (Endpoint, species/cell line) are included to allow for a supe rficial classification of the threshold values. Substance Study type Effect value Effect Endpoint Species/Cell line Reference LOEC Danio rerio (Annunziato et al., 2019) LOEC Danio rerio (Annunziato et al., 2019) PFHxS in vivo 2 µM morphometric effects in the larvae, specifically increased length and yolk sac area 6:2 FTOH* in vivo 2 µM Behavioral endpoints: distance traveled & mean velocity PFHxA in vivo 6:2 C l-PFESA in ovo 6:2 C l-PFESA in ovo 0.2 µM 150 ng/g (egg w) 1 500 ng/g (egg w) in ovo 38 000 ng/g (egg w) PFHxS reduction in the overall length and yolk sac size (however not observed at higher doses) LOEC lower heart rate LOEC enlarged liver (8%) LOEC Danio rerio C hicken (Gallus gallus domesticus) C hicken (Gallus gallus domesticus) decreased tarsus length and embryo mass LOEC C hicken (Gallus gallus domesticus) (C assone et al., 2012) mortality LC 50 Nematode (Caenorhabditis elegans) (C hen et al., 2018a) mortality LC 50 Nematode (Caenorhabditis elegans) (Annunziato et al., 2019) (Briels et al., 2018) (Briels et al., 2018) PFOS in vivo PFBS in vivo 1.4 µM (95% C I, Range 1.1-1.6 µM) 794 µM (95% C I, Range 6241 009 µM) PFDA in vitro 7.8 µM oocyte viability LC 50 Oocyte cells from pigs PFDA PFDoDA, PFNA, PFOA PFOS, PFOA, PFBS in vitro 3.8 µM maturation IM50 in vitro 107-647 µM cytotoxicity EC 50 in vitro 57 – 80 µM IC 50 PFDoDA in vivo 1.2 mg/L aromatase inhibition level of dopamine was upregulated significantly Oocyte cells from pigs human placental choriocarcinoma cell line JEG-3 human placental choriocarcinoma cell line JEG-3 (C hen et al., 2018a) (Domínguez et al., 2019) (Domínguez et al., 2019) (Gorrochategui et al., 2014) (Gorrochategui et al., 2014) LOEC Danio rerio (Guo et al., 2018) 197 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Study type Effect value Effect Endpoint Species/Cell line PFNA in vitro 10 µg/L LOEC Oocyte cells from cows PFNA PFBS, PFHxS, PFNA, PFHpA in vitro LOEC Oocyte cells from cows increased mortality LOEC Eisenia fetida PFNA, PFHpA in vivo 0.1 µg/L 100 000 µg/kg dw 100 000 µg/kg dw negative effect on blastocyst formation lipid droplet distribution significantly altered significant weight loss LOEC Eisenia fetida (Hallberg et al., 2019) (Hallberg et al., 2019) (Karnjanapiboonwo ng et al., 2018) (Karnjanapiboonwo ng et al., 2018) genotoxicity (DNA strand breaks and fragmentation, chromosomal breaks and apoptosis) EC 50 Marine mussel (Perna viridis) (Liu et al., 2014) genotoxicity (DNA strand breaks and fragmentation, chromosomal breaks and apoptosis) EC 50 Marine mussel (Perna viridis) (Liu et al., 2014) EC 50 Marine mussel (Perna viridis) (Liu et al., 2014) EC 50 Marine mussel (Perna viridis) LOEC (only one concentratio n tested) Chironomus riparius (Liu et al., 2014) (Marziali et al., 2019) in vivo PFNA in vivo PFDA in vivo 33 µg/L (95% C I, Range 29 – 37 µg/L) 594 µg/L (95% C I, Range 3411 063 µg/L) 195 µg/L (95% C I, Range 144 – 265 µg/L) 78 µg/L (95% C I, Range 7384 µg/L) PFOS, PFOA, PFBS in vivo 10 µg/L PFOS PFOA in vivo in vivo genotoxicity (DNA strand breaks and fragmentation, chromosomal breaks and apoptosis) genotoxicity (DNA strand breaks and fragmentation, chromosomal breaks and apoptosis) reduced growth PFOS in vivo 29.8+/4.1 µM PFOA in vivo 424.1+/124.0 µM behavior: prolonged backward swimming (indicating modified cellular cadmium conductance) behavior: prolonged backward swimming (indicating modified cellular cadmium conductance) FOSAPrTMA in vivo 19.1+/- behavior: shortened backward swimming Reference EC 50 Paramecium caudatum (Matsubara et al., 2006) EC 50 Paramecium caudatum (Matsubara et al., 2006) EC 50 Paramecium caudatum (Matsubara et al., 198 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Study type Effect value Effect 17.3 µM PFNA in vivo 98.7+/20.1 µM PFDA in vivo 60.4+/10.1 µM (indicating modified cellular cadmium conductance) behavior: prolonged backward swimming (indicating modified cellular cadmium conductance) behavior: prolonged backward swimming (indicating modified cellular cadmium conductance) 6:2 C l-PFESA in vivo 6 mg/L 6:2 C l-PFESA in vivo 5 µg/L PFNA in vitro 4.8 µM PFDA in vitro TFA in vivo PFBA in vivo PFOA in vivo PFNA in vivo PFDA in vivo Endpoint Species/Cell line Reference 2006) EC 50 Paramecium caudatum (Matsubara et al., 2006) EC 50 Paramecium caudatum (Matsubara et al., 2006) decreased survival LOEC Danio rerio (Shi et al., 2017) LOEC 7.1 µM 700 mg/L (95% C I, Range 4601 000 mg/L) decreased liver triglyceride levels inhibition of p-glycoprotein (p-gp) cellular efflux transporter inhibition of p-glycoprotein (p-gp) cellular efflux transporter sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal curvature) IC 50 Danio rerio gill cells from marine mussel (Mytilus californianus) gill cells from marine mussel (Mytilus californianus) (Shi et al., 2019a) (Stevenson et al., 2006) (Stevenson et al., 2006) EC 50 Danio rerio (Ulhaq et al., 2013) 2 200 mg/L (95% C I, Range 1 2002 200 mg/L) 350 mg/L (95% C I, Range 290430 mg/L) 16 mg/L (95% C I, Range 7.7450 mg/L) 5.0 mg/L (95% C I, Range 3.8- sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal curvature) sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal curvature) sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal curvature) sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal EC 50 Danio rerio (Ulhaq et al., 2013) EC 50 Danio rerio (Ulhaq et al., 2013) EC 50 Danio rerio (Ulhaq et al., 2013) EC 50 Danio rerio (Ulhaq et al., 2013) IC 50 199 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Study type Effect value Effect 6.6 mg/L) curvature) 450 mg/L (95% C I, Range 350600 mg/L) 1.5 mg/L (95% C I, Range 1.1.1.9 mg/L) sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal curvature) sublethal endpoints (edema, malformations, non-hatched eggs, lack of circulation, reduced pigmentation, spinal curvature) Endpoint Species/Cell line Reference EC 50 Danio rerio (Ulhaq et al., 2013) EC 50 Danio rerio (Ulhaq et al., 2013) PFBS in vivo PFOS in vivo TFA, PFBA in vivo >3 000 mg/L mortality LC 50 Danio rerio (Ulhaq et al., 2013) PFOA in vivo 430 mg/L (95% C I, Range 290710 mg/L) mortality LC 50 Danio rerio (Ulhaq et al., 2013) PFNA in vivo >10 mg/L mortality LC 50 Danio rerio (Ulhaq et al., 2013) PFDA in vivo PFBS in vivo 8.4 mg/L (95% C I, Range 5.315 mg/L) 1 500 mg/L (95% C I, Range 1 1001 900 mg/L) PFOS in vivo <10 mg/L mortality LC 50 Danio rerio 6:2 C l-PFESA in vivo 15.5 mg/L mortality LC 50 Danio rerio TFA in vivo 70 mg/L mortality LC 50 Brachionus calyciflorus PFPrA in vivo 80 mg/L mortality LC 50 Brachionus calyciflorus (Ulhaq et al., 2013) (Wang et al., 2013b) (Wang et al., 2014a) (Wang et al., 2014a) PFBA in vivo 110 mg/L mortality LC 50 Brachionus calyciflorus (Wang et al., 2014a) mortality LC 50 Danio rerio (Ulhaq et al., 2013) mortality LC 50 Danio rerio (Ulhaq et al., 2013) 200 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Study type Effect value Effect Endpoint Species/Cell line PFPeA in vivo 130 mg/L mortality LC 50 Brachionus calyciflorus PFHxA in vivo Reference (Wang et al., 2014a) (Wang et al., 2014a) 140 mg/L mortality LC 50 Brachionus calyciflorus 1.2 mg/L, upregulated gene expression levels of 1.2 mg/L and thyrotropin-releasing hormone (trh), 6 mg/L corticotrophin-releasing hormone (crh) (Zhang et al., PFDoDA in vivo respectively and iodothyronine deiodinases (dio2) LOEC Danio rerio 2018b) 1.2 mg/L and downregulated gene expression levels of 6 mg/L thyroglobulin (tg) and thyroid receptor (Zhang et al., PFDoDA in vivo respectively (trbeta) LOEC Danio rerio 2018b) *6:2 FTOH has recently been classified via decision of EC HAs Risk Assessment C ommittee (RAC ) as a substance that is very toxi c to aquatic life with long lasting effects (Aquatic C hronic 1; H410, M=1) (EC HA News: https://echa.europa.eu/documents/10162/2082415/news_annex_rac_seac_dec_2021_en.pdf/92b14f83-580d-323a-486a32fade778505?t=1639041622096; RAC protocoll: https://echa.europa.eu/documents/10162/17090/rac59_final+minutes_en.pdf/2f350729-0880-57da9a1e-b812d24df808?t=1639990293786) 201 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Ecotoxicological threshold values reported in the PFASTox Database differ over several orders of magnitude (ng-scale to mg-scale). A comparison, or averaging of the different values is not feasible due to the large differences between the different studies (e.g. test-design, sensitivity of species, reporting of results). The key message of this compilation-exercise is, that some of the concentration values at which certain PFASs cause ecotoxicological effects (mostly organ specific effects) are within the same order of magnitude at which PFASs have been detected in the environment (see chapter B.4.2). The significance of this comparison is of course very limited as it ignores e.g. inter-species differences regarding the sensitivity to PFAS-mediated effects as well as differences between laboratory conditions (under which the above values were derived) and real-world conditions. Still, it underlines the need to minimize emissions of PFASs to the environment. This is without prejudice to the assumption that PFASs are regarded as nonthreshold substances from a regulatory perspective, due to their very high persistence. Additionally, this compilation highlights the large knowledge gap which exists around the study of environmental effects of PFASs. The PFASTox Database, at the time of access, contains information for 29 out of potentially >10 000 individual PFASs. This shows, that the investigation of potential adverse effects of PFASs in the environment has not even begun to comprehensively assess this large class of substances despite severe efforts that have been put into this area of research over the past decades. B.7.1.11. Conclusions Despite the growing amount of studies investigating the ecotoxicity of PFASs, the available data on adverse effects of PFASs in the environment is limited to a small number of substances. Moreover, most studies invest igate the aquatic toxicity of PFASs, leaving a huge gap of knowledge regarding the toxicity towards terrestrial organisms. Additionally, conventional ecotoxicological tests may not be suitable to detect long term effects from exposure to PFASs. PFASs can remain in the environment for long time periods (decades-centuries or even longer) due to their high persistence but ecotoxicological test systems usually cover only time spans of a few days – weeks. Yet, for a small subset of PFASs, there is information av ailable that suggests that these substances cause adverse effects in the environment. Some of the PFASs (especially PFOA and PFOS) have been investigated thoroughly and suggestions for possible mechanisms of action have been made. Based mostly on information for PFOS and PFOA, there is “ample basis to suspect that at least a subset of PFASs can be considered persistent, bioaccumulative, and/or toxic” as Ankley and colleagues phrased it in their review paper on the ecological risks of PFASs (Ankley et al., 2021). PFASs may also cause adverse effects, that are relevant for whole populations if they affect endpoints such as reproduction or survival of offspring. Two trends regarding the toxicity of PFASs could be derived from the currently data available: 1. Toxicity in the same PFAS class tends to increase with increased C-chain length. 2. PFSAs are usually more toxic than PFCAs with the same chain length (C8) (Ankley et al., 2021). However, interactions of substances with the environment are complex and depend on various factors. For example, the environmental behaviour and fate of PFASs depend both on the inherent physicochemical properties for eac h respective PFAS and their degradation products, the physico-chemical conditions of the abiotic environmental compartments that act as recipient systems (e.g. organic carbon content of sediments, or temperature, salinity, concentration of oxidants in seawater), and the physiological status and conditions of the recipient organisms which may take up and accumulate the given PFAS. Considering that PFASs are very persistent and mobile, organisms living in different environmental compartments are continuously exposed to PFASs. Without reliable data that is suitable to detect also long-term effects of PFASs (inter-generational effects after intergenerational exposure), it is not possible to demonstrate safe use of PFASs. This warrants for a restriction of the use of PFASs to minimize emissions to the environment. At the time notable 202 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) effects from PFASs exposure occur in the environment it will be difficult, if not impossible, to remove the contamination. Thus, there is a threat of irreversible damage. 203 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.2. Effects on wildlife B.7.2.1. Notes on the procedure The basis for the following paragraphs is the chapter “Field-based effects studies in wildlife” in the review paper on ecological risks of PFASs from Ankley et al. (2021). The main findings from this chapter were summarized according to different animal groups (avian species, mammals, reptiles, fish and invertebrates). These findings were complemented by results from other studies that were compiled in the PFASTox database, see chapter B.7.1 for more details. The study list from that database was filtered for studies that investigated effects on wildlife and are not included in the review by Ankley and colleagues. B.7.2.2. PFASs effects on wildlife Several studies investigated tissue concentrations of PFASs in wild living animals that were exposed to PFASs (and other contaminants) in the field. Subsequently, these studies assessed the link of the measured tissue concentrations to adverse effects o r endpoints related to impairments (e.g. oxidative stress) – in distinction to the studies compiled in chapter B.7.1 where “wildlife” species were subject to ecotoxicological tests under laboratory conditions. Ankley et al. (2021) stated in their review on ecological risks of PFASs that these types of studies are of special interest due to the fact that “field studies can provide a critical perspective on whether actual exposures in free‐living organisms can reach levels that cause observable adverse effects”. However, they highlighted also limitations regarding the toxicity data from field studies, as it is “possible that other unidentified factors contribute to observed effects, and that any effects observed are influenced by the choice of endpoint” (Ankley et al., 2021). Thus, establishing links between contaminant exposure and health outcome is a difficult task (Rodríguez-Estival and Mateo, 2019). The studies evaluated for this report reported various endpoints associated with PFAS exposure in the field. These endpoints comprise changes in brain chemistry (e.g. brain hormone levels), oxidative stress (e.g. induction of antioxidant enzymes), reproduction (e.g. eggshell thinning), changes in metabolism (e.g. PPARα and cyp4a), reduced biomass, endocrine activity (e.g. changes in hormone levels), or changes in immunologic parameters (e.g. a decreased antibody response). Ankley et al. (2021) based their review on PFAS effects on wildlife on 39 studies that assessed avian species (19), mammals (11), reptiles (3), fish (2) and invertebrates (2). The PFASTox database yielded 8 complementing studies, 5 on avian species, 2 on mammals and 1 on fish. B.7.2.3. Avian species The endpoints from field studies with avian species, compiled by Ankley et al. (2021), comprise measures of oxidative stress and reproduction. They focused their evaluation on reproduction endpoints, presumably because this is a highly relevant endpoint to assess adverse effects on the population or ecosystem level. Two studies reported decreased hatching succe ss for eggs of tree swallows (Tachycineta bicolor) from different regions in Minnesota. The effect was linked to PFOS concentrations in eggs of 150 ng/g ww and upwards. However, in a more recent study, where tree swallow eggs were sampled near an Air Force base in Michigan (coordinated by the same author), no impact on hatching was observed while concentration measurements in eggs resulted in concentrations >600 ng/g ww. According to Ankley et al. (2021), the authors of that study considered the effects from the earlier studies to be linked to the presence of cocontaminants. Two different studies, reviewed by Ankley et al. (2021), reported negative associations between PFAS contamination and reproductive endpoints. One study reported reduced eggshell thickness and hatching success for individuals of a population of Great tits (Parus major) that was nesting in a PFAS-contaminated environment (PFASs were also measured in the eggs) and the other linked concentrations of plasma PFDoA with a reduced hatching success in Arctic black-legged kittiwakes (Riss tridactyla). However, Ankley and colleagues highlighted, that for these studies “mechanism­based quantitative linkages between exposure 204 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) and effect have not yet been established” (Ankley et al., 2021). Additionally, Blévin et al. (2018) investigated the effects of environmental contaminants (organochlorines, PFASs and mercury) on the incubation temperature (relevant for breeding success) of wild breeding Arctic black-legged kittiwakes and found that PFAS body burdens were not related to the incubation temperature measured at the brood patches of the birds. Costantini et al. (2019) observed protein oxidative damage as well as other oxidative status markers in male individuals of the Arctic black-legged kittiwakes with higher body burdens of long-chain PFCAs (C11-C13). Furthermore, they observed, that the “non-enzymatic antioxidant capacity (including antioxidants of protein origin) was significantly lower in those birds having higher plasma concentration of linear perfluorooctanesulfonic acid (PFOSlin)”. However, activity of glutathione peroxidase in erythrocytes could not be correlated to PFAS body burdens. The authors state, that more experimental work is needed “to demonstrate whether PFASs cause toxic effects on free-living vertebrates through increased oxidative stress”. Miljeteig et al. (2012) did not find associations between PFAS levels in ivory gull (Pagophila eburnea) individuals and three endpoints (eggshell thickness, retinol (vitamin A) and alphatocopherol (vitamin E)) when they investigated effects of environmental contamination on the Arctic seabird. Several PFASs (PFOS, PFNA, PFOSA, PFHxS and PFOA) were detected in the livers of wild common cormorants (Phalacrocorax carbo) from Lake Biwa, Japan (Nakayama et al., 2008). Subsequently they investigated the potential adverse effects of environmental contaminants through gene expression profiling. Nakayama et al. (2008) report that their results “suggest the induction of antioxidant enzymes in response to oxidative stress caused by PFCs and the suppression of molecular chaperones, leading to reduction in protein stability”. In addition, they aimed to identify possible combination effects of PFASs with dioxins and related compounds (DRCs) via multiple regression analysis and concluded that “the regression models suggested the potential of PFCs to enhance toxicities of DRCs”. In a study to assess the correlation between thyroid hormone concentrations and plasma concentrations of halogenated organic contaminants in young individuals of two se abird species (Rissa tridactyla and Fulmarus glacialis), “Positive associations between total thyroxin (TT4) and PFCs (PFHpS, PFOS, PFNA)” were found (Nøst et al., 2012). The authors further qualified that “Although correlations do not implicate causal relationships per se, the correlations are of concern as disruption of TH homeostasis may cause developmental effects in young birds”. B.7.2.4. Mammalian wildlife Field-effect studies with mammals, included in the review from Ankley et al. (2021), “consistently show an association between PFAS exposure and alterations in biomarkers of exposure and effect.”. Ankley et al. (2021) indicated one study with polar bears, where changes in enzymes relevant for the neurotransmitter-system were significantly correlated with brain concentrations of PFOS and PFCAs. In a follow up study (Pedersen et al., 2016) investigated changes in the brain steroid concentrations of polar bears. Pedersen et al. (2016) report that ΣPFSAs (PFBS, PFHxS, PFOS and PFDS) as well as ΣPFCAs (PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA and PFPeDA) could be linked with increased testosterone concentrations in specific brain regions. However, that authors could not verify the exact mechanistic connection and suggested a “interference with de novo steroid synthesis” or a “disruption of peripheral steroidogenic tissues mainly in gonads and feedback mechanisms” as possible mechanistic explanations. Kurtz et al. (2019), reviewed in Ankley et al. (2021), reported that ΣPFCAs (PFOA, PFNA, PFDA, PFUnA, PFDoA, PFTriA, PFTA) as well as ΣPFAAs (Ankley et al., 2021) ΣPFCAs + PFOS and PFOSA) positively correlated with PPARα mRNA and cytochrome P4504a (cyp4a) protein expression in the kidney of stranded marine mammals (cataceans). Three studies mentioned in the review of Ankley and colleagues investigated links between environmental PFASs exposure and 205 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) immunologic parameters in marine mammals. Scotter et al. (2019) did not observe a relation between antibody prevalence and exposure to contaminants in their study with Atlantic walruses (Odobenus rosmarus rosmarus). Fair et al. (2013) detected significant positive associations between PFAS exposure and immunological parameters in bottlenose dolphins (Tursiops truncates; absolute numbers of CD2+ T cells, CD4+ helper T cells, CD19+ immature B cells, CD21+ mature B cells, CD2/CD21 ratio, MHCII+ cells, B cell proliferation, serum IgG1, granulocytic, and monocytic phagocytosis). Kannan et al. (2006) found in their study that individuals of southern sea otters (Enhydra lutris nereis), that died of infectous disesases, had significantly higher concentrations of PFOS and PFOA compared to individuals that died of nonifectious causes. Persson and Magnusson (2015) investigated liver concentrations of PFAAs (PFOA, PFNA, PFDA, PFunA, PFDoA, PFTrDA, PFBS, PFHxS, PFOS) in male individuals of wild American mink ( Neovison vison) sampled in Sweden. The authors subsequently examined associations between contaminant-body-burdens and effects on the reproductive system and reported a negative association between anogenital distance - a potential marker for endocrine disruptor exposure (Foster, 2006; Mammadov et al., 2018) - and concentrations of PFOS, PFDA, PFUnDA as well as ΣPFAAs. Soloff et al. (2017) evaluated PFAA concentrations in bottlenose dolphins (Tursiops truncatus) highly exposed via their environment and assessed the effects on the cellular immune system in an ex vivo flow cytometry-based assay. The authors report that “Baseline PFOS concentrations were associated with significantly increased CD4(+) and CD8(+) T cell proliferation”. Together with further data from in vitro exposure Soloff and colleagues concluded that their results “suggest that PFOS directly dysregulates the dolphin cellular immune system and has implications for health hazards”. B.7.2.5. Reptiles With regard to reptiles, Ankley et al. (2021) reviewed three studies that linked environmental PFAS exposure with reduced biomass (PFOS and PFNA for male individuals of Malaclemys terrapin and PFHxA for both sexes), adverse effects on uroporphyrin concentrations (PFOS; observed in Caretta caretta) as well as immunosuppression (PFOS; observed in four different turtle species). B.7.2.6. Fish Field-effect studies focusing on fish, that were discussed in the review by Ankley et al. (2021) reported ambiguous results regarding the effects of environmental PFAS exposure. A study from Bangma et al. (2018) with wild c aught striped mullet (Mugil cephalus), described a positive connection between PFAS-liver-concentration and the number of eggs, which the authors assumed to be due to a higher food consumption. No other effects on reproduction endpoints were observed in this study. Oakes et al. (2010) observed liver effects in fishes (species not specified) shortly after AFFF contamination of the creek they inhabited, but the effects did not persist over time. A study from Guillette et al. (2020) with striped bass (Morone saxatilis) related serum concentrations of 23 different PFASs with endpoints relevant for the liver and immunesystem. In addition, a study from Houde et al. (2013) quantified twelve PFASs (PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA, PFPeDa, PFHxS, PFOS, PFDS, PFECHS) in muscle, liver and plasma of northern pike (Esox lucius) environmentally exposed to contaminants via municipal wastewater effluents in St. Lawrence River, Canada. Next to significantly up regulated stress response genes (MT, GST, SOD and CYP1A1), the authors observed significant relationships between VTG gene expression in liver, VTG-plasma-activity and plasma concentrations of PFTrDA, PFTeDA and PFDS. 206 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.2.7. Invertebrates Ankley et al. (2021) included two field-effect studies with invertebrates in their review. A study from Rusconi et al. (2015) reported that more sensitive taxa of invertebrates were found upstream of a wastewater influent from a fluorotelomer plant while downstream of the discharge point, these species were not present. Statistical analysis, however, did not result in significant differences between the two loc ations. Differences in the genetic diversity between the two locations were furthermore quantified for one species (Hydropsyche modesta) but could not be associated to the contamination. Bakke et al. (2010) assessed training facilities for firefighting (former as well as active) in order to evaluate if these sites are affected by the discharge of PFAS-containing firefighting foams. Sampling of earthworms (Eisenia fetida) to assess the bioaccumulation of PFASs was part of the assessment. Adverse effects were not recorded during the assessment, however the authors of the study concluded: “Based upon predictions of no effect concentrations (PNEC) for 6:2 FTS, PFOS and PFOA in soils, soil organisms living within about 100 meters from these four sites may be at risk”. However, the authors state with reservations that this assumption is based solely on PNEC values for only one species and that more information for a robust assessment is warranted (Amundsen et al., 2008). B.7.2.8. Conclusion A recent review paper from Ankley et al. (2021), as well as complementary studies, compiled in the PFASTox database, provide some evidence, that environmental exposure to PFASs is correlated to effects observed wildlife animals of different animal groups (avian species, mammals, reptiles, fish and invertebrates). Observable effects comprise various endpoints e.g. changes in brain chemistry or metabolism, oxidative stress, reproduction, reduced bioma ss, endocrine disruption, or changes in immunologic parameters. Due to the limitations of the studies, a clear link between PFASs measurements in the environment, or PFAS-body-burdens in the animals and the observed effects can rarely be established. Other factors such as cocontaminants, environmental conditions, age, or sex might have contributed to the effects. Complementary laboratory studies that can plausibly link effects in these species to PFAS exposure would be needed but are in most cases not available. Yet, the available studies provide evidence, that PFASs can cause adverse effects on wildlife species at currently relevant concentrations. Following a precautionary approach, this warrants for measures to minimize the emissions of PFASs into the environment. As emissions will continue, the pollution stock will increase over time due to the high persistence of PFASs (or their final degradation products) and adverse effects on population- or ecosystem-level will become more likely. At the time notable effects from PFASs exposure occur in the environment it will be difficult, if not impossible, to remove the contamination. Thus, there is a threat of irreversible damage. 207 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.3. Atmospheric compartment – global warming potential Some PFASs are persistent and volatile and will partition to the atmosphere where they will stay for a very long time. These substances may have a considerable global warming potential which contributes to the greenhouse effect and global warming. In fact, some of the strongest greenhouse gases known are PFASs. One of the most relevant subclasses of PFASs that contribute to global warming are the fluorinated gases, e.g. hydrofluorocarbons (HFCs) and hydrofluoroethers (HFEs). These are highvolume substances used as refrigerants, blowing agents, and solvents etc. with considerable emissions. Emitted substances evaporate and reside in the atmosphere (Oltersdorf et al., 2021). The demand for fluorinated gases is currently increasing. In an expert insight paper by Sovacool et al. (2021) it is pointed out that fluorinated gases have been termed “supergreenhouse gases” given their severe and powerful impact on the climate. They are the most potent greenhouse gases known to modern science, with global warming potentials far greater than carbon dioxide, some up to almost 24 000 times more so. At the same time, they are also the fastest growing class of greenhouse gas emissions around the world, especially in developing countries. The global warming potential (GWP) of a substance depends, inter alia, on its lifetime in the atmosphere. Short-lived substances may therefore often have a lower GWP compared to long-lived substances if they otherwise are comparable in their contribution to radiative forcing. The Global Warming Potential over 100 years (GWP-100) is used for countries' greenhouse gas emission inventories reported to the UNFCCC. In this report the Dossier Submitters simply refer to GWP for GWP-100, unless otherwise stated. GWP is a relative measurement which measures the global warming potential of a substance relative to that of carbon dioxide (CO2) with GWP = 1 by definition. Radiative forcing is a term that is closely related to GWP, but unlike GWP, it does not have a time horizon. Radiative forcing describes how strongly the radiation balance of the atmosphere is influenced if the concentration of a given gas, chemical, or substance increases. Changes in the Earth's overall radiative forcing can be caused by changes in the concentration of greenhouse gases in the atmosphere, thus leading to global warming. Generally, the more fluorine atoms in a compound, the greater its GWP and radiative forcing (Sovacool et al., 2021). Atmospheric lifetimes and global warming potential for fluorinated gases may be found in IPCC Assessment Reports. Emissions, up to 2020, are to be reported using the GWP -values from the IPCC fourth Assessment Report (Forster, 2007). Updated values from the IPCC fifth Assessment Report will be used for reporting under the Paris Agreement. A few examples are listed in Table B.15 below for comparison. Table B.15. GWP-values (GWP-100) collected from the IPCC fourth Assessment Report. Gas Chemical formula Lifetime (y) GWP C arbon dioxide C O2 Methane C H4 12 25 Nitrous oxide N2O 114 298 HFC -23 C HF3 270 14 800 1 HFC -32 C H2F2 4.90 675 HFC -236fa C F3C H2C F3 240 9 810 PFC -14 C F4 50 000 7 390 PFC -5-1-14 C F3C F2C F2CF2CF2CF3 3 200 9 300 HFE-125 C HF2OC F3 136 14 900 HFE-143a C H3OC F3 4.30 756 HFE-7100 C F3C F2C F2CF2OCH3 3.8 297 208 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) According to ECOS (2021) the latest generation of fluorinated gases (e.g. HFOs) were developed to replace earlier generations of gases that had high global warming potential . Although the HFOs often have low GWP, they may lead to formation of TFA during degradation in the atmosphere. The authors called for replacement of fluorinated gases with truly sustainable and futureproof alternatives, such as natural refrigerants. Dudita and Kauffeld (2021) summarized the environmental impacts of HFO refrigerants and their alternatives, including their degradation products. The authors concluded that alternative refrigerants should be halogen-free, avoiding thus atmospheric TFA and HF formation. The study also looked into the refrigerants' indirect contribution to global warming through their manufacturing process and concluded that natural refrigerants (ammonia, CO 2 and hydrocarbons) have considerably lower CO 2 equivalent emissions during manufacture. Cousins et al. (2020b) elaborate on the consequences of the high persistence of PFASs and points out that some PFASs, or their breakdown products, may hav e environmental effects in addition to the concerns usually considered under REACH, like e.g. high climate impact (e.g. in the case of perfluoroalkanes and perfluoro-tert-amines). In some cases, such PFASs are not covered by the F-gas regulation and its measures. One example is perfluorotributylamine (N(C 4F 9)3) which was studied by Tsai (2017). It was found that the substance has a very low solubility in water and relatively high vaporization from the water bodies, suggesting that perfluorotributylamine will sink into the atmosphere. The substance was reported to have an atmospheric lifetime of 500 years and a GWP = 7 100. Bernard et al. (2020) investigated the perfluorinated trialkylamines further and found the fully fluorinated triethyl and tripropyl analogues to have GWPs of 9 900 and 8 700, respectively. The fluranes, e.g. sevoflurane, isoflurane, desflurane and enflurane, are perfluorinated alkyl ether substances with a considerable atmospheric lifetime and GWP, see Table B.16 with data collected from Hodnebrog et al. (2020). These substances are known for their anaesthetic effects. Two of them are listed in Annex II (reporting obligations only), of the F-gas regulation, while the other two are not included in the F-gas regulation. Table B.16. GWP-values (GWP-100) for selected fluranes. Gas Chemical formula Lifetime (y) GWP Sevoflurane, HFE‐347mmz1 Isoflurane, HC FE‐235da2 (C F3)2C HOC H2F C HF2OC HC lC F3 1.9 3.5 205 565 Desflurane, HFE‐236ea2 C HF2OC HFC F3 14.1 2 720 Enflurane, HC FE‐235ca2 C HF2OC F2C HFC l 4.4 686 The high climatic effect of fluoroform (HFC-23) is well known and measures to reduce the emissions of the substance has previously been introduced. Stanley et al. (2020) calculated that these measures should have seen global emissions drop by 87% between 2014 and 2017. Instead, atmospheric observations show that emissions of fluoroform (HFC-23) have increased and in 2018 were higher than at any point in history. The authors speculated that the magnitude of the discrepancy between reported emissions reductions and emissions inferred from t he atmospheric data could result from developing countries have been unsuccessful in meeting their reported emissions reductions, or that there may be substantial unreported production of HCFC22 at unknown locations which has been regarded as the main source of fluoroform (HFC-23) emissions. The discrepancy between the inventory-based emissions estimates of fluoroform (HFC-23) and the emissions based on atmospheric measurements is roughly equivalent to the total green-house gas emissions of Spain in 2017 (Stanley et al., 2020). A literature survey on the emissions from incineration of fluoropolymer materials performed by the Norwegian Institute for Air Research (NILU) on behalf of the Norwegian Environment Agency disclosed information indicating that incineration of fluoropolymers leads to the formation of substances like CF 4 (PFC-14), CHF 3 (HFC-23), C2F 6 (PFC-116), tetrafluoroethene (TFE) and 209 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) hexafluoropropene (HFP), some of which are potent greenhouse gases (Huber et al., 2009). The kind of compounds formed is strongly dependent on the incineration conditions like temperature, moisture, oxygen content, use of catalysts etc. The report concludes that incineration of fluoropolymer containing products has a great potential to contribute considerably to the total greenhouse gas emissions of Norway, but due to the lack of sound data on the fate of fluoropolymers in Norway as well as of the chemical reactions in the different types of waste incineration plants, no exact amounts can be given. Stoiber et al. (2020) looked further into the disposal of products and materials containing PFASs and concluded that incineration of PFAS wastes can release toxic air pollutants and greenhouse gases, which may represent a cyclical problem as disposal of PFAS-containing wastes creates repeated cycles of contamination. Volatile PFASs may also be emitted into the air from landfills and wastewater treatment plants. In the Chemicals Strategy for Sustainability the European Commission points out that in a safe and sustainable-by-design approach to chemicals, overall sustainability should be ensured by minimising the environmental footprint of chemicals in particular on climate change, resource use, ecosystems and biodiversity from a lifecycle perspective (EEB, 2020). B.7.3.1. Legislation Two legislative acts have already been adopted to control emissions from fluorinated greenhouse gases, including hydrofluorocarbons (HFCs), in the European Union: the F -gas Regulation and the MAC Directive. B.7.3.1.1. F-gas Regulation The current F-gas Regulation (Regulation (EU) No 517/2014), which applies since 1 January 2015, replaces the original F-gas Regulation adopted in 2006 (EC, 2014). The F-gas regulation has the following ambitions:  Limiting the total amount of the most potent fluorinated greenhouse gases (called Fgases in the regulation) that can be produced and imported into the EU from 2015 onwards and phasing them down in steps to one-fifth of the level of 2014 in 2030. This will be the main driver of the move towards more climate-friendly technologies;  Banning the use of F-gases (or F-gases with a GWP above a certain threshold) in many new types of equipment where less harmful alternatives are widely available, such as fridges in homes or supermarkets, air conditioning and foams and aerosols;  Preventing emissions of F-gases from existing equipment by requiring leak checks, proper servicing and recovery of the gases at the end of the equipment's life . Annex I of the regulation lists 27 specific fluorinated greenhouse gases (F -gases) for which the above regulations apply and for which the intentional release into the atmosphere shall be prohibited where the release is not technically necessary for the intended use. Annex II lists 43 additional fluorinated greenhouse gases (F-gases) that are subject to reporting obligations. The basis for the F-gas regulation is the GWP of the substances in scope and their contribution to global warming, while other concerns are not taken into account, e.g. atmospheric degradation to TFA which precipitates and causes exposure to humans and the environment. A discussion of degradation of some relevant fluorinated gases may be found in section B.4.1.3.2. The German Environment Agency recently published a comprehensive investigation of degradation of fluorinated gases: "Persistent degradation products of halogenated refrigerants and blowing agents in the environment: type, environmental concentrations, and fate with particular regard to new halogenated substitutes with low global warming potential" (UBA, 2021c). 210 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) The F-gas regulation is currently under revision and a proposal for a revised regulation was published in April 2022 (EC, 2022). B.7.3.1.2. MAC Directive The Mobile Air-Conditioning (MAC) Directive (Directive 2006/40/EC) covers the use of fluorinated gases with a GWP of more than 150 in passenger cars (vehicle category M1) and light commercial vehicles (category N1, class 1) (EC, 2006). The directive is enforced over three phases, and from 1 January 2017, the use of fluorinated greenhouse gases with a GWP higher than 150 in all new vehicles put on the EU market has been totally banned. introduced from 2011, and in all new cars and vans produced from 2017. The traditionally used refrigerant in MAC systems, HFC-134a (CH2FCF3), has a GWP of 1 430 and has been phased out for use in air condition equipment in new cars in the EU. The Directive does not specify any particular refrigerant or system, leaving the technical choice on the car manufacturers. The MAC Directive is limited to the use of fluorinated gases in air-conditioning systems in cars and vans, but not in buses, trains, ships etc. Air condition equipment is only one of several applications of fluorinated gases. 211 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.4. Microbiological activity in sewage treatment systems Not specifically assessed. See section B.4.5 on the challenges to remove PFASs in STPs. 212 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.5. Endocrine activity and endocrine disruption B.7.5.1. Notes of procedure To obtain relevant literature that addresses the endocrine activity (EA) or endocrine disruption (ED) of PFASs in the environment an initial literature research for review papers was performed, using Scopus 19 and PubMed20. A first screening of the results showed that there were many overlaps between the results from the two different databases but the research in Scopus produced more relevant results. Thus, only references yielded from Scopus were screened thoroughly. In total 754 titles were screened. Twenty papers were selected for further reading after title screening. The results reported in the (mostly review) papers were collected in tabular form. In some cases, they were supplemented with more details from the primary references. Furthermore, the literature research was complemented by references compiled in the PFASTox database21. The database was filtered for in-vitro and animal studies on the endocrine system. PFASTox Database entries were extracted in tabular form and the most important results reported in abstracts were added to the tabular compilation of results. In case of ambiguities or insufficient description of results or methodologies the references were read in more detail. Lastly, it is important to note, that the EA or ED of PFAS which are already restricted (e.g. PFOS, PFOA, C9-C14 PFCAs) were not recorded in the tabular compilation unless for reasons of comparison or to describe trends. B.7.5.2. Overview of results The studies that were assessed in the course of this research reported EA/ED for 32 individual PFASs, and different mixtures of these substances. Considering the group of PFASs contains >10 000 individual substances this is a rather small dataset. From the present data no trends regarding chain length or functional groups that promote EA/ED could be derived. Thus, it is not possible to focus the assessment on a certain subset of PFASs. Most likely not all relevant studies describing EA or ED of PFASs in the environment were collected. Still, this shows that the overall data on endocrine effects of PFASs in the environment is scarce. Four studies described adverse effects of different PFASs, evoked by an e ndocrine mode of action, which are considered relevant on population level (i.e. having the potential to negatively affect a whole population of animals e.g. through reduction of fecundity or fertility of individuals). In particular, 8:2 FTOH, PFBS, 6:2 Cl-PFESA and a PFAS mixture of PFOA, PFOS, PFBS and PFBA were found to cause adverse effects that can be considered relevant on the population level. In total 69 cases of EA/ED of 32 PFASs were recorded (in silico 1, in vitro 51, in vivo 17; incl. the 4 cases with effects relevant on population level). 11 cases with inconclusive results ( in vitro 9, in vivo 2) and 12 cases of no EA/ED after exposure to PFASs were reported (in vitro 8, in vivo 4). A tabular overview of the EA/ED of each PFAS together with the respective study type and reference can be found in Table B.17. A bias in the search results towards studies reporting EA/ED of PFASs has to be assumed due to the fact that the research was designed to discover studies describing such effects. Furthermore, the publication of negative results unfortunately is still not common practice in the scientific community which further strengthens the bias. Still the results of this research provide the insight, that some PFAS - with heterogenous structures and functional groups (e.g. carboxylic acids, sulfonic acids, telomer alcohols, ether, ester, sulfonamides, or cyclic PFAS) - show EA/ED in in silico, in vitro and in vivo tests and cause adverse effects through disruption of the hormone system – in some cases with the potential to negatively affect whole populations. Additionally, there is a very large group of PFASs for which no information regarding their EA/ED is available (as this research indicates). Some of these compounds, e.g. those with structural similarities to already known PFASs with properties causing EA/ED, might have the potential to cause adverse effects through interaction with the hormone system of organisms in the environment. To this 19 www.scopus.com, date of access: 2022-09-29. https://pubmed.ncbi.nlm.nih.gov/, date of access: 2022-09-29. 21 https://pfastoxdatabase.org/, date of access: 2022-09-29. 20 213 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) moment it is however not possible to say with certainty which structural elements of PFASs are responsible for their EA/ED properties. Table B.17. Overview over EA/ED of different PFASs . Substance Study Type Endocrine activity/ (C AS-, EC -No.) Endorcrine disruption 4:2 FTOH (2043-47-2, 218-0509) Reference In vitro + (Rosenmai et al., 2016) In vitro + (Liu et al., 2007) In vitro 0 (Weiss et al., 2009) In vitro In vitro In vitro + + + (Rosenmai et al., 2016) (Ishibashi et al., 2007) (Liu et al., 2007) In vitro In vitro In vitro + + + In vitro 0 (Liu et al., 2009) (Maras et al., 2006) (Benninghoff et al., 2011) (Weiss et al., 2009) In vitro In vitro In vitro + + + (Rosenmai et al., 2016) (Ishibashi et al., 2007) (Maras et al., 2006) In vitro + (Benninghoff et al., 2011) In vivo + (Liu et al., 2010a) In vitro + (Benninghoff et al., 2011) In vitro + (Rosenmai et al., 2016) In vitro + (Rosenmai et al., 2013) In vivo - (Rosenmai et al., 2016) 10:2 diPAP (1895-26-7, 217-5855) In vitro - (Rosenmai et al., 2016) 8:2 triPAP (N.A., N.A.) In vitro + (Rosenmai et al., 2016) In vitro In vitro In vitro 0 + + (Weiss et al., 2009) (Rosenmai et al., 2016) (Behr et al., 2018) In vitro 0 (Vongphachan et al., 2011) In vitro In vitro In vivo + + (C roce et al., 2019) (Ishibashi et al., 2007) (Godfrey et al., 2017) In vivo + (Godfrey et al., 2019) In vitro + (Rosenmai et al., 2016) In vitro + (Rosenmai et al., 2018) In vitro - (Wielogórska et al., 2015) In vitro + (Weiss et al., 2009) In vitro + (Rosenmai et al., 2016) 6:2 FTOH (647-42-7, 211-477-1) 8:2 FTOH (678-39-7, 211-648-0) 8:2 FTOAcr (27905-45-9, 248-7227) 8:2 monoPAP (5767803-2, N.A.) 8:2 diPAP (678-41-1, 211-649-6) PFBA (375-22-4, 206-786-3) PFPeA (2706-90-3, 220-3007) PFHxA (307-24-4, 206-196-6) 214 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance (C AS-, EC -No.) PFHpA (375-85-9, 206-798-9) Study Type In vitro Endocrine activity/ Endorcrine disruption + In vitro + In vitro + In vitro - In vitro + PFHxS (355-46-4, 206-587-1) PFHpS (375-85-9, 206-798-9) PFDS (335-77-3. 206-401-9) HFPO-DA (13252-13-6, 236-2368) HFPO-TA (2641-34-1, 220-1413) (Benninghoff et al., 2011) (Vongphachan et al., 2011) (Vongphachan et al., 2011) (Wielogórska et al., 2015) (Ishibashi et al., 2007) In vivo - (C assone et al., 2012) In vitro + (Weiss et al., 2009) In vitro In vitro + + In vitro + In vitro + (Rosenmai et al., 2016) (Benninghoff et al., 2011) (Vongphachan et al., 2011) (Ishibashi et al., 2007) In vitro - (Wielogórska et al., 2015) In vitro + + In vitro + (Weiss et al., 2009) (Vongphachan et al., 2011) (Behr et al., 2018) In vitro In vitro In vivo + - (C roce et al., 2019) (Ishibashi et al., 2011) (Newsted et al., 2008) In vivo In vivo In vivo + + + (C hen et al., 2018b) (Sant et al., 2019) (Lou et al., 2013) In vitro In vitro In vitro + + + In vitro - (Weiss et al., 2009) (Watkins et al., 2015) (Vongphachan et al., 2011) (Behr et al., 2018) In vitro In vitro + - In vivo In vivo + In vitro 0 In vivo + In vitro + In vivo 0 In vitro + (Vongphachan et al., 2011) (Nøst et al., 2012) (Benninghoff et al., 2011) (Benninghoff et al., 2011) (Li et al., 2019) In vitro In vivo + + (C operchini et al., 2020) (C onley et al., 2019) In vitro + (Li et al., 2019) In vitro PFBS (375-73-5, 206-793-1) Reference (Ishibashi et al., 2011) (Wielogórska et al., 2015) (Ramhoj et al., 2020) (C assone et al., 2012) 215 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance (C AS-, EC -No.) 6:6 PFPiA (70609-44-8, N.A.) 6:8 PFPiA (610800-34-5, N.A.) Study Type Endocrine activity/ Endorcrine disruption Reference In vivo + (Liu et al., 2019b) In vivo + (Liu et al., 2019b) 8:8 PFPiA (500776-69-2, N.A.) In vivo + (Liu et al., 2019b) In vivo + (Kim et al., 2020b) In vitro + (Wågbø et al., 2012) PFOSA (754-91-6, 212-046-0) PFEC HS (646-83-3, N.A.) 6:2 FTUA (N.A., N.A.) N-MeFOSE (24448-09-7, 246-2621) N-EtFOSE (1691-99-2, 216-8874) FOSA (754-91-6, 212-046-0) N-MeFOSA (31506-32-8, 250-6658) N-EtFOSA (4151-50-2, 223-9803) F-53B (MIxture of 6:2 C l-PFESA and 8:2 C lPFESA) (73606-19-6, N.A.) 6:2 C l-PFESA (73606-19-6, N.A.) In vivo + (Houde et al., 2016) In vitro + (Weiss et al., 2009) In vitro 0 (Weiss et al., 2009) In vitro 0 (Weiss et al., 2009) In vitro In vitro + + (Weiss et al., 2009) (Rosenmai et al., 2018) In vitro 0 (Weiss et al., 2009) In vitro 0 (Weiss et al., 2009) In silico + (Deng et al., 2018) In vitro + (Deng et al., 2018) In vivo + (Deng et al., 2018) In vitro + (Li et al., 2018a) In vivo In vivo In vitro + 0 + (Shi et al., 2018a) (Zhou et al., 2018) 8:2 C l-PFESA (N.A., N.A.) PFAS mixture (PFOA, PFOS, PFBS, PFNA) In vivo + (N.A., N.A.) + = Activity/effects related to PFAS exposure - = No activity/effects related to PFAS exposure 0 = inconclusive results XX = Effects with presumed relevance on population level (Li et al., 2018a) (Lee et al., 2017) B.7.5.3. Summary of main results from studies A short summary and reference for each study which is mentioned in the following paragraphs can be found in Appendix B.7.5.. B.7.5.3.1. In silico Two in silico studies show that different PFASs have the theoretical ability to interact with 216 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) receptors in the endocrine system of different species, or to bind to hormone transporting proteins and thus have the potential to interfere with hormone homeostasis. B.7.5.3.2. In vitro Different reporter gene assays show that PFASs can interact with hormone receptors from different cell systems incl. estrogen receptors (ER), androgen receptors (AR) or peroxisome proliferatoractivated receptors (PPAR). Via these interactions PFASs can evoke (anti) estrogenic or (anti) androgenic activity and influence steroidogenesis-. Examples for effects are an alternation of the receptor activity or an induction of VTG production or cell proliferation. Besides PPAR and relevant receptors from the hypothalamic –pituitary–gonadal (HPG) axis, PFASs have been shown to interact with the hypothalamic –pituitary–thyorid (HPT) axis too. In vitro binding to thyroxine transport proteins transthyretin (TTR) and thyroxine-binding-glubolin (TBG) has been reported in different studies (see also B.5.2.1.4). Furthermore, some PFASs can alter the expression of thyroid-hormone responsive genes. Some in vitro studies made an attempt at describing the relation between EA/ED of some PFASs and their structure (i.e. chain length or functional group). Binding to the human PPARγ ligand binding domain was reported to increase with increasing chain length of PFCAs until a chain length of 11 carbon atoms. Similarly, PPARα activity in HepG2 cells was observed to increase with chain length up until a chain length of C8. Interestingly, an in vitro study with Baikal seal (Pusa sibirica) PPARα (BS PPARα) made contradictory observations and reported that PFCAs with a chain length > C7 had a negative correlation between chain length and induction potency towards BS PPARα. A test with HEK 293 cells found the activity of PFOA and substitutes towards the PPARγ to increase in the order HFPO-DA < PFOA < HFPO-TA. A radioligand-binding assay testing for binding capacity to human TTR reported a maximum binding potency for PFCAs with a chain length of C8. The authors of this study suggested that the binding potency is directly linked to the number of fluorinated alkyl groups in the carbon-chain. Regarding the influence of functional groups on EA/ED, one in vitro study described the binding affinity towards PPAR to be stronger for PFSAs, compared to their PFCA homologues. Another study however, reported that they found the transactivation potencies of PFCAs to be stronger than the ones of PFSAs with similar chain length. TTR binding potency was observed to be stronger for PFSAs than for PFCAs (at least for C4-C8 PFASs). Especially with regard to the interaction of PFAS with PPAR there is evidence suggesting a relationship between the chain length as well as the head group and the binding affinity of the compounds. Similar indications exist from one study investigating TTR binding potencies of different PFAS. But the results from different studies contradict each other and overall, there is not enough information to identify clear trends. Statements like “PFAS above/below a certain chain length or PFAS with/without a certain headgroup have a stronger/weaker endocrine activity” cannot be made on the basis of the data at hand. B.7.5.3.3. In vivo There is evidence for changes in the gene expression for genes regulated by hormones in fish and crustaceans after in vivo exposure to certain PFASs (e.g. vtg, a common indicator for an (anti)estrogenic MoA). Other in vivo studies report changes in the act ivity of enzymes that catalyze the biosynthesis of hormones in fish, or birds. Adverse effects observed in in vivo studies after exposure to certain PFASs include morphological changes in endocrine and other organs in fish, e.g. follicle cell degeneration and atrophy, or changes in swim bladder size (an indicator for disruption of HPT-axis as recently summarized by Dang et al. (2021), a decreased fecundity in fish, a reduced number of eggs spawned and other adverse effects on the F1 generation of exposed fish (e.g. changes in sex ratio, reduced spermiation) or reduced pipping success (pip = first crack in eggshell) and decreased tarsus length and embryo mass in birds exposed in ovo. All of the aforementioned adverse effects have the potential to negatively affect whole populations. Additionally, several in vivo studies observed changes in hormone levels in different species after exposure to certain PFASs. Two studies with fish reported that changes in hormone levels were transferred to the F1 generation and adverse effects related to hormone level changes were 217 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) observed in F1 generation even though the F1 generation was not exposed. Suc h crossgenerational effects in combination with the high persistence of PFASs (or their degradation products) can have severe impacts on whole wildlife populations and in the long run also disrupt networks of ecosystems. A positive relationship between changes in hormone levels and exposure to some PFASs was already observed in wildlife birds. Correlations do not necessarily show a causal relationship, but it still raises a concern about effects of certain PFASs on the endocrine system in situ (i.e. in the natural environment). B.7.5.3.4. Conclusion In summary, the in silico, in vitro and in vivo indications of interactions of some PFASs with the endocrine system of environmental species, adverse effects (some occurring crossgenerational), and first observations of possible influences of PFAS body-burden on hormone levels in wildlife raise concerns about the presence of PFASs in the environment. The environmental presence of PFASs and their concentrations in wildlife will increase under continued use due to thehigh persistence of these substances or their final degradation products, increasing the probability for adverse and irreversible effects. At the time notable effects from PFASs exposure occur in the environment it will be difficult, if not impossible, to remove the contamination. Thus, there is a threat of irreversible damage. 218 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.7.6. Hazard and occurrence of fluoropolymers Fluoropolymers are fluorinated polymers characterised by at most times fully -fluorinated carbon backbone (Buck et al., 2011). One example is polytetrafluoroethylene (PTFE). Other fluoropolymers are fluorinated ethylene propylene (FEP), perfluoroalkoxy alkanes (PFA), ethylene tetrafluoroethylene (ETFE), other tetrafluoroethylene-copolymers, polyvinylidene fluoride (PVDF), and fluoroelastomers (Gardiner, 2015). Fluoropolymers are indirectly of concern because ­ ­ During their production and use monomers, oligomers smaller polymers and by -products are emitted into the environment During waste incineration at end of life, non-polymeric PFAS may be formed and emitted PFCAs (including trifluoroacetic acid, TFA) and other fluorinated compounds are formed when fluoropolymers are incinerated. Today no safe option for end-of-life of fluoropolymers and articles containing fluoropolymers is known, recycling is not sufficiently possible (Drobny, 2008; Schlipf and Schwalm, 2014). Fluoropolymers are produced with different production processes, leading to different products (e.g. granulate or powder) and are incorporated in different articles (e.g. cookware, membranes for clothing) (Lohmann et al., 2020). The same type of fluoropolymer (e.g. PTFE) can be produced by different production processes, using different building blocks or different manufacturing conditions (including different surfactants) in order to obtain specific properties suitable for different applications. From an environmental point of view the PFAS –based processing aids used in the production of fluoropolymers are of concern. They can be emitted to the environment during production of fluoropolymers, during the production of article containing fluoropolymers and during use and disposal of those articles. Processing aids are for example per- and polyfluoroalkylether carboxylic acids (PFECAs) like the ammonium salt of HFPO-DA (already identified as SVHC), cC604, is the ammonium salt of [perfluoro-{acetic acid}], 2-[(5methoxy-1)], ammonium 4,8-dioxa-3H-perfluorononanoate (CAS 958445-44-8, ADONA) (Lohmann et al., 2020). Certain types/forms of fluoropolymers can be produced without processing aid (emulsion polymerization is done with PFAS-based processing aids, suspension polymerization does not need PFAS-based processing aids). Further information can be found in Annex A section A.2.1.4.2. Fluoropolymers themselves can pose an environmental hazard. Like for other polymers, fluoropolymer microplastics can be formed during their use phase or end-of-life phase. It is therefore important not only to look at the use phase but the whole life cycle of fluoropolymers. Monitoring data show that fluoropolymer microplastics are present in the environment. (see B.4.2.7.8). Lohmann et al. (2020) state that fluoropolymers are extremely persistent under environmental conditions. The authors refer to Dams and Hintzer (2017). The Dossier Submitters are currently not aware of any study investigating the persistence of fluoropolymers under environmental conditions. Even though fluoropolymers are a large group of polymers with different properties, there are no indications that fluoropolymers will degrade in particular as one of their key properties are their thermal and weather stability for the uses as described (Dams and Hintzer, 2017). Therefore, persistence can be concluded. B.7.6.1. Environmental hazard of fluoropolymers Henry et al. (2018) argued that fluoropolymers are not toxic, based on a data set that was restricted only to a few fluoropolymer types typically >100 000 Da (Lohmann et al., 2020). Lohmann et al. have doubt concerning this assessment, because concentrations of leachable components were very low (1 ppm for PTFE fine particles) in the data set from Henry et al. 219 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) (2018), whereby leachable components are mainly processing aids. Other studies reported concentrations of 1–10 ppm in PTFE fine powder and much higher in PTFE aqueous dispersion (Wang et al., 2014c). Similar levels of PFAAs (0.3–24 ppm) were found in personal care articles that contained PTFE fine particles (assuming the cosmetics contained 1% PTFE, the range of leachables is 0.3–24 ppm; if the total organofluorine measurements represented PTFE fine powder, then the range of PFAA-leachables is 15 – 1 000 ppm) (Lohmann et al., 2020). one study analysed twenty targeted PFASs in fluoropolymer raw material and product samples made in China. The total concentration of PFASs except PFOA in the fluoropolymer product samples ranged from below the detection limit (0.1 ng/g; 0.0001 ppm) to 50.3 ng/g (0.0503 ppm). The total concentration of PFASs (except PFOA) in the fluoropolymer raw material samples ranged from < MDL (0.1 ng/g) to 8.3 × 105 ng/g (83 ppm). HFPO-TrA, HFPO-TeA and HFPO-DA, hexafluoropropylene oxide (HFPO, C3F6O) oligomers, were typical alternatives to PFOA in fluoropolymer manufacturing (Meng et al., 2021). The following passages are similarly also addressed in the Microplastic restriction proposal (ECHA, 2019a) and a more detailed assessment can be found there. In this context RAC has agreed that although there are uncertainties in the understanding of the hazard and risk of microplastics, there is sufficient evidence to conclude that that they constitute an intrinsic hazard because of their long-term persistence in the environment in combination with their particulate form and potential to cause adverse effects (ECHA, 2020). According to the hazard assessment included in the Microplastic restriction proposal many studies in earlier years focused on ingestion of microplastics and their occurrence in the gut, rather than exploring adverse effects on organisms. Nevertheless, ingestion in laboratory studies has since been linked to a diverse range of sub-lethal endpoints, including reduced food intake, false satiation and reduced energy reserves, as well as mortality and sub-lethal ‘apical effects’, such as effects on growth rates or reproduction (summarized by Besseling et al. (2018)). Furthermore, in the microplastics dossier the following effects have been collected from publications: altered survival, feeding, growth, reproduction, moulting, behaviour, photosynthesis, oxidative stress, enzyme activity, gene expression, and nutrient cycling, as well as malformation, and inflammation.; all of which were considered by the authors to be relevant to population or community-level effects. Generally, it has to be noted that investigated effects were observed more in acute than chronic studies. Discussion about microplastics often addressed whether microplastics facilitate the uptake of organic pollutants such as POPs or metals. According to the microplastics Dossier current scientific consensus on this issue would suggest that ingestion of microplastics does not significantly enhance bioaccumulation of POPs relevant to other types of particulates present in the environment. Furthermore, fluoropolymer microplastics as vectors for organic pollutants may be even less relevant due to their minor adsorptive properties. The bioaccumulation potential for polymers in general is poorly understood so far. Based on theoretical conclusions, nanoplastics as (bio)degradation products of microplastics may have a higher bioaccumulation potential and with that more likely evoke adverse effects. Cell membrane penetration cannot be excluded. Several studies have investigated adverse effects of microplastics in general. No negative effects on population level have been demonstrated so far but there is an emerging understanding of the potential effects of microplastics. More details on uptake and distribution are given in B.5.1.2. B.7.6.2. Findings of fluoropolymers in the environment There is a wealth of monitoring studies which investigated the occurrence of microplastics in the environment and biota. There are however to a lesser extent monitoring data which distinguish between different types of polymers. PTFE microparticles have been found in fish and sediment (from remote Arctic Ocean). Bergmann et al. (2017) investigated microplastics in sediment samples from the Arctic. Samples were collected at 2 340 – 5 570 m water depth. Sampling took place in 2015. Analysis of microplastics took place with attenuated total reflection Fourier-transformed infrared (ATR-FTIR) spectroscopy 220 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) for particles >500 µm. For the small size fraction (<500 µm) organic material was removed before analysis with Fourier-transform infrared microscopy (µFTIR). Results from spectroscopy were compared to library material, automated in the case of small particles. 4 356 particles/kg sediment (overall mean number) have been found. Eighteen different types of polymers were identified, among them PTFE. In the larger size fraction all analysed particles were PTFE. For the smaller size fraction it was not possible to analyse for PTFE because “PTFE cannot be detected within the spectral region of 3 600−1 250 cm−1 available for the particles <500 μm in μFTIR” (Bergmann et al., 2017). The occurrence of PTFE in all sediment samples is explained by the high density (2.10−2.30 g cm−3), which exceeds the density of seawater. A density dependent distribution pattern between water and sediment were also observed in a study of small-scale Japanese rivers which investigated the abundance and distribution and characterised microplastics (Kabir et al., 2022). Polyvinyl chloride (PVC), polyethylene (PE), and polypropylene (PP) were the major polymers in this study. The polymers—polyvinyl chloride, polymethylmethacrylate, polyurethane, fluorinated ethylene propylene, and polybutylene in sediments were distinct from those detected in surface water, as were the predominance of large-size (1–5 mm) and fragment-shape microplastics. Fluorinated ethylene propylene (FEP) is a copolymer manufactured from hexafluoropropylene and tetrafluoroethylene. The authors conclude that it is theoretically obvious that MP particles of high-density (>1.0 g/cm3) polymers are prone to be settled easily in the freshwater environments and low-density particles float on the surface water or in the water column. Kabir and coworkers also lists other studies which investigated microplastics qualitatively. Though there is a density dependent distribution pattern of polymers in the environment, in all listed studies the dominating polymers were not fluorinated. For instance, the polymers PE, PP, and PS made up >75% of all microplastics identified in the sediments in the rivers Rhine and Main in Germany (Klein et al., 2015). It should be noted that PE, PP, and PS, cover 55.7% of the European plastic demand, which is a reason for their large abundance (Plastics Europe, 2013). Most common polymer types found in 106 fish from 22 species inhabiting three sites of the Han River, South Korea, were polypropylene (PP) (≥40%) and polyethylene (PE) (≥23%), followed by polytetrafluoroethylene (PTFE) (≥16%) at all sampling locations (Park et al., 2022). Omnivorous and insectivorous fish contained more Microplastics than carnivorous and herbivorous fish. In addition, studies investigating fish habitats showed that pelagic fish contained slightly higher levels of microplastics than demersal fish. Interestingly, it is essential to emphasize the detection of PTFE in fish because no PTFE was observed in water. This result was due to the high density of PTFE, and it was challenging to detect heavy plastic in water because the water sampling was conducted from the upper part of the river using a Manta net. The number of microplastics is corrected for the fish weight, the average number of Microplastics was 16.26 ± 12.51 Microplastics 100 g−1 in the demersal fish; 17.69 ± 12.84 microplastics 100 g−1 in the pelagic fish. In fact, the study found that the most common polymers in the fish were relatively light PP and PE. In contrast, sediment can be a sink for microplastics, especially high-density microplastics (Woodall et al., 2014). Therefore, demersal and benthic fishes may have been more exposed to high-density microplastics than pelagic species. Capillo et al. (2020) analysed microplastics also in demersal fish from the Southern Tyrrhenian area (Central Mediterranean). Sampling took place in 2017. Selected particles were analysed with micro-Raman or ATR-FTIR spectroscopy. Five polymers where identified, among them PTFE. The major contaminant that affected examined individuals was represented by polytetrafluoroethylene (PTFE, 75%) in Mullus barbatus barbatus which is widely regarded as a bio-indicator species for its benthic behaviour, habitat and feeding modalities and its reduced mobility. B.7.6.3. Conclusion In conclusion, fluoropolymers are a group of materials with different properties, and for some of 221 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) the fluoropolymers persistence is well known. Other types of fluoropolymers would be expected to be persistent in the environment based on their chemical composition, see section B.4.1.2. Environmental exposure does take place as shown by monitoring data. Also, remote regions are already exposed, which confirms the persistent nature of the materials. The bioaccumulation potential is poorly understood so far. Based on theoretical conclusions nanoplastics formed due to weathering of microplastics may have a higher bioaccumulation potential and with that more likely evoke adverse effects. Cell membrane penetration cannot be excluded. Several studies have investigated several adverse effects of microplastics in general. No negative effects on population level have been demonstrated so far but there is an emerging understanding of the potential effects of microplastics. 222 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.8. PBT and vPvB assessment See section 1.1.4. of the main report. No further assessment carried out. 223 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9. Exposure assessment In the exposure assessment both the environmental exposure and the human exposure is assessed. In section B.7 the environmental exposure is assesed. In section B.5 the human exposure is assessed. Environmental emssions are described as service life emissions. Tonnage put on the market will lead to service life emissions, presented as [t]. (Next year, if for instance the same PFAS tonnage is put on the market, a similar service life emission is applicable). The emission value is generated e.g. by multiplying yearly tonnage put on the market by the default release factors (for instance an ECHA Environmental Release Category (ERC)). This is not an emission rate and is the emission over the whole service life. Preferably industry specific emission factors were used and applied but this information was not always available. In case no industry specific information was available, ECHA generic ERC were applied. There are however no dedicated ERC for PFASs. Because of the lack of better alternatives, ERC for organic substances were applied. Industry specific emission factors were used for PFASs manufacturing and the sectors food contact material and packaging (partly), metal plating and manufacturing of me tal products, cosmetics, ski wax and HVACR. For medical devices consultant expertise was used. For the other sectors, no industry specific information was available. For these sectors, default parameters for environmental release rates according to the REACH methodology described in the ECHA Guidance on information requirements and Chemical Safety Assessment, Chapter R.16 were used (ECHA, 2016a), see also Table B.18. The environmental release category (ERC) describes the broad conditions of use from the environmental perspective. In applying ECHA ERC, emissions from the lifestage (which can be several years) is allocated to the year of production. Waste stage emissions are also allocated to the same year, and the integrated emission over 20 years is taken for the emission from waste, see Guidance ECHA (2012a), p 46. Uncertainties and limitation due to the use of ERC  Correctness and completeness of the volumes of PFASs used for e mission estimates are not fully known Uncertainties and possible inaccuracies in the selection and application of ERC  The wide range of substances. (If emission data from industry is lacking, ERCs have to be applied. The ERC for organic substances is used for emission estimation of a very broad group of vP PFAS substances with very different substance characteristics)  The wide range of applications in the sectors  The identification of relevant emissions during different life cycle stages of the substances  Adding ERF for three compartments air, water, soil  The emissions of polymeric PFASs in the use stage might be an ove restimation due to lack of industry data and due to lack of specific ERCs for polymers. For high molecular, non-powdery PFASs the use of ERC may lead to an overestimation of emissions, whereas for micro-powder PFASs, the use of ERC may lead to an underestimation of emissions. For the environmental impact assessment, the total release to the environment (onecompartment model) was used, based on the summation of the individual Emission Release Factors (ERF) to air, water and soil. The reasoning for assuming the environment as one compartment is that the overarching concern with PFASs in scope of this restriction proposal is persistence, which is independent of the compartment the PFASs are emitted to. Specific negative impacts may occur in connection to point sources due to high local concentrations. 224 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) However, these are not included in the overall impact assessment. The ERF is derived from Chapter R.16 of the REACH guidance. In case the addition of the separate ERF to water, soil, and air exceeds 100%, a total emission to the environment of 100% is assumed. The chosen approach therefore comes with considerable uncertainties. See Annex F (Uncertainties) for more details. In Table B.18 the methodology used to estimate emissions per sector is presented. Table B.18. Methodology used to estimate emissions per sector. Sector REACH methodology Other methodology Manufacturing No Yes TULAC Yes Food contact material and packaging Metal plating Yes, partly No Yes, partly (Industry data for paper and board manufacturing) Yes C onsumer mixtures Yes No C osmetics No Yes Ski wax Yes Yes HVAC R No Yes Medical devices No Yes Transport Yes Electronics and semiconductors Yes Energy Yes C onstruction Yes Lubricants Yes Petroleum and mining Yes Waste Yes Adapted ERC for landfill and literature research 225 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.1. Introduction to hazard and risk PFASs are non-threshold substances due to their properties that are similar to PBT/vPvB-. Emissions are a proxy for risk. Minimiszing emissions is key to reduce risks in the long term. Therefore, a proper estimation of emissions is of importance. It is important to realise that because of the persistence properties of PFASs in the long term, the volumes put on the market will lead to environmental emissions (in or outside EEA). In section B.9, data on emission of PFASs are provided and emission calculation/estimation will be described. The purpose is to provide: a. Estimates of the total amount of PFASs released to the environment in the EEA in the base year (i.e., the expected year of entering into force (EIF) of the restriction). b. Project total emissions for different timelines (20, 30 and 45 years after entry into force of the restriction). Emissions are estimated for all investigated (use) sectors (see Table B.18), starting with emissions of PFASs during manufacturing and ending with emissions of PFASs in the waste stage. Due to the combination of high persistence and broad use in articles across various sectors with wide variation in service-life duration and type of use PFASs emissions may last , many years following initial placing on the maket , including the end-of-life stage. For many uses, significant volumes of PFASs enter the waste stage, due to a combination of the persistence of PFASs and low emissions during service-life of polymer containing articles. Waste emissions can for instance occur in waste transfer stations, landfills, waste-water treatment plants and waste incinerators. Additionally, also during recycling of materials, PFAS emissions are possible. This is described in more detail in B.9.18. Since sector specific emission factors often were not available to the Dossier Submitters, for most sectors, default Environmental Release Categories (ERC) from the ECHA guidance (ECHA, 2016a) were used to estimate emissions. OECD SPERCs, which are more sector specific compared to ERCs, have not been considered useful as not for all researched uses SPERCs are available and the status of SPERCs for sectors where they are available is often unclear. The quality of SPERCs is diverse (some are reviewed/evaluated, others not) and for several SPERCs, substance specific physico-chemical information is needed. In Table B.18 the methodology to estimate emissions is provided for each sector. Yearly emissions in the EEA are considerably lower than yearly estimates of EEA PFAS production and use volumes. Some uses lead to direct emissions (like use of impregnation sprays) within the year of use of a product, whereas other uses lead to prolonged emissions stretching over several years (i.e., the use of coatings containing PFASs). The difference between estimates of PFAS volumes in production and use (Annex A) and emissions (part of this Annex B) is discussed in more detail in Annex F (Uncertainty). Under real life conditions, emissions of PFASs do not only occur during current use and during the time periods addressed in the impact assessment. Several sectors have a long history of PFASs use, and PFASs may have been emitting for decades. Addtionally, there is a significant technical stock of PFASs in articles (cars, fashion, etc.). As a consequence, PFASs have been acc umulating in the environment for decades, leading to an environmental stock. Since there is insufficient data on the volume of PFASs used and emitted by sectors in the past, as well as on PFASs properties and fate of the individual PFASs, it is not possible to estimate the volume of this environmental stock therefore, the estimates of volumes and emission for the the base year of the assessment (2025, being the year in which the restriction is expected to enter into force). Implications will be further disc ussed in Annex E.2., and in section 2.4. of the main report. In the impact assessment (Annex E.2.) the emissions to air, water and soil are considered together as emissions to the environment. Therefore a total environmental emission is also presented in the Annex. 226 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.2. PFASs manufacturing B.9.2.1. Introduction Initially, most PFASs were produced without consideration of the impact (Goldenman et al., 2019). This caused serious incidents in the past years at multiple PFASs manufacturing sites in Europe:     3M, Zwijndrecht, Belgium Chemours, Dordrecht, the Netherlands Solvay Miteni, Veneto, Italy Arkema/Daikin, Pierre-Bénite, France. A description of PFASs manufacturing activities in the EU/EEA, the type of PFASs produced and the location of the different sites is presented in Annex A section A.2.1. The production, import and export of PFASs used in the EEA has been evaluated. The methodology for estimating emissions consisted of two main elements:   Review of relevant literature. Stakeholder surveys and interviews. Information on emissions of PFASs to water, soil and air during production and processing activities was asked from stakeholders. Even though information was provided by some of the major producers and processors of PFASs present in the EEA, not all emissions were captured. For instance emission to soil was not presented although the situation at 3M Zwijndrecht suggest that PFAS soil emission for PFAS manufacturing cannot be neglected. In order to extrapolate, two values are needed. Firstly, the total volume of PFASs produced and processed per year. Secondly, an emission factor to predict the emissions to air and water of the aforementioned activities. (Emissions to soil were not applicable according to industry). The emission factor represents the percentage of PFASs which is released to the environment when a certain amount of PFASs is being manufactured or processed. The emission factors were derived from the information received from stakeholders on both volume and emissions. The following equation was used (Equation 1): Equation 1. Calculation of emission factors 𝐸𝐹𝑐𝑜𝑚𝑝 = 𝐸 𝑐𝑜𝑚𝑝 / 𝑚 × 100, Where: EF emission factor of the produced/ processed PFASs [%] comp receiving environmental compartment i.e.: water or air [-] E Tonnes of PFASs emitted per year [t/y] m Tonnes of PFASs produced/ processed per year [t/y] After deriving the emission factors to air and water for each individual company, an average emission factor to air and water was calculated. These average values were then applied to the 227 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) volume of produced/processed PFASs, which gave the results of the estimated PFASs annually released to the environment. The derived emission factors are shown in Table B.19. Table B.19. Average emission factors to environmental compartments based on stakeholder information. PFAA and precursors + other PFASs Number of datapoints Average emission factors [%] PFAA Fluorinated gases Polymeric PFASs Emission factor to water Emission factor to air Emission factor to water Emission factor to air Emission factor to water Emission factor to air 5 5 6 6 7 6 0.04 0.06 0.00 2.06 0.01 0.02 B.9.2.2. Emissions Some companies provided their annual production volume as a range. In this case, the average values were used for emission calculations, e.g. in case a manufacturer reported a production volume of 100 – 1 000 t/y, the average of 550 t was used. For some manufacturing sites (3M and Chemours), the information was cross checked with permit and enforcement data to get feeling with realistic numbers. Information on emissions resulting from PFASs manufacturing from literature was not used because it includes diff erent year and types of PFASs. The investigation learned that there are about 20 PFASs production facilities in Europe (see Annex A section A.2.1.). Table B.20 presents the estimated annual EEA emissions, split into PFAA and PFAA precursors (+ other PFASs), fluorinated gases, and polymeric PFASs to air and water. Table B.20 Estimated annual EEA emissions of PFAA and PFAA precursors (+ other PFASs), fluorinated gases, polymeric PFASs and total PFASs from PFASs manufacturing . PFAA and PFAA fluorinated gases Polymeric PFASs Total PFASs (t/y) precursors (+ other (t/y) (t/y) PFASs) (t/y) low high low high low high low high Volume 53 902 118 051 15 000 176 548 49 000 101 763 117 902 396 362 Total emission 54 118 309 3 637 15 30 378 3 785 Emission to air 32 71 309 3 637 10 20 351 3 728 Emission to water 22 47 - - 5 10 27 57 Emission of fluorinated gases to air by Chemours Dordrecht, The Netherlands, is approximately 70 t/y. However, efforts are made to drastically reduce emissions (blauw, 2021). Emissions of CF4 to air by 3M Zwijndrecht, Belgium, is about 33 t/y (based on permit (p.51) (Provincie Antwerpen, 2018; Provincie Antwerpen, 2020). The total emission of fluorinated gases was 222 t 228 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) according to 3M22. Especially HFC-23 (not a PFAS) emissions were overseen. So in fluoropolymer production emissions of F-gases/fluorinated gases are of relevance. Fluorinated gases, used to manufacture fluoropolymers and fluoroelastomers, have not been restricted by the provisions of the Montreal Protocol, because chemicals used as feedstock in the manufacture of other chemicals are excluded from the Montreal Protocol. Emissions of fluorinated gases used for fluoropolymer production seem to be significant. The emissions of PFAA and PFAA precursors (+ other PFASs) and polymeric PFASs to the environment are far lower than emissions of fluorinated gases (Rodriguez et al., 2021) (see Table B.20). In the past (2013), Chemours Dordrecht in the Netherlands, the main production site in Europe for fluoropolymers such as Teflon™ and Viton™, was allowed to emit 7 t Hexafluoropropylene oxide dimer acid (HFPO-DA) per year. Currently, these numbers are lower23. It is unknown if all permits for PFASs production sites in EEA have been updated to new, stricter, levels. Neither is known if enforcement takes place. Recently, potential illegal emissions of PFAA and PFAA precursors to water were reported 24. The total emissions from PFAS manufacturing are estimated between 400 and 4 000 t/y (rounded numbers). The numbers are, order of magnitude, in line with site specific permit data (keeping in mind about 20 production facilities in EEA). It is important to note that indirect emissions to air, water and soil, from waste are are not included in the total emissions. These emissions could be significant , see section 0. Emissions of PFASs from waste from Chemours in The Netherlands for instance was 13 times higher than via water. Total waste volume at this site is about 1 800 t/y (ILT, 2018; Tweede Kamer, 2019). In the majority of cases, fluorinated polymer production requires not only fluorine-containing feedstock chemicals but also fluorinated production aids for polymers such as dispersion aids and polymerisation initiators. Production of such fluorinated production aids can also lead to emissions poly- and perfluorinated by-products, both highly volatile and water soluble (Hopkins et al., 2018; Lohmann et al., 2020). Past emissions: According to literature, production facilities of PFASs were responsible for >75% of total global perfluorocarboxylates (PFCA) emissions between 1950 and 2002 (Prevedouros et al., 2006). Based on literature, emissions in the recent past likely have increased as can be seen in Figure B.6725. This very likely is the result of increased production as is described in Appendix A. 22 https://www.apache.be/nl/2021/06/15/3m-stootte-ongemerkt-gigantische-hoeveelheden-zwaar- broeikasgas-uit, date of access: 2022-12-14. 23 https://cms.dordrecht.nl/Inwoners/Overzicht_Inwoners/Dossier_C hemours_en_DuPont/Nieuws/Nieuwe_v ergunningen_C hemours_gepubliceerd, date of access: 2022-12-14. 24 https://www.taylordailypress.net/chemical-company-3m-illegally-discharges-toxic-substance-fbsa-insche/, date of access: 2022-12-14. 25 https://www.unep.org/explore-topics/chemicals-waste/what-we-do/policy-and-governance/globalchemicals-outlook, date of access: 2022-12-14. 229 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.67. Estimated annual emissions of PFCA from PFOA production sites (left) and fluoropolymer sites (right) in the United States, Western Europe and Japan (blue), as well as in China, Russia, Poland and India (orange) (OECD/UNEP, 2015). According to the OECD/UNEP (2015), in 2015 there was an increase in the total global annual C4-C14 PFCA emissions between 1951 and 2002. This period was followed by a considerable decrease (approximately 40% reduction in emissions). Emissions continued to decrease in the US, Western Europe and Japan, while in the rest of the world an increase was observed. Although PFCA are only a fraction of the PFASs universe, these findings are in line with what the EU producers reported. A total of 27 stakeholders provided information on past, current, and future emissions of PFASs. A summary is provided in Table B.21. Stakeholders think both past and future emissions are lower than current emissions. Table B.21. Comparison of past, current, and expected future emissions according to stakeholders. Past emissions (compared to current status) Future emissions (compared to current status) Higher No change Lower Higher No change Lower 1 4 11 0 4 12 Notes: The values indicate the number of sites Stakeholders think that, despite an increase in production, emissions of PFASs will remain the same or decrease because of the implementation of best available technologies. B.9.2.3. Human exposure It has been known for a long time that exposure to PFASs is of relevance and harmful, especially for higher exposed populations like factory workers and people living close to PFAS manufacturing sites. Greater plasma contamination levels have been detected in workers in the PFAS manufacturing Industry. Recent studies demonstrate that the human population living close to a PFAS factory in 230 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Zwijndrecht, Belgium is highly exposed to PFASs (ewg, N/T)26. The water in the river Westerschelde contains levels of PFASs that are directly or indirectly linked t o PFAS manufacturing. The Fishing industry is affected, and eating of fish from the Westerschelde is discouraged (RIVM, 2022). In the Netherlands, citizens nearby a PFAS manufacturing facility are advised not to eat vegetables from gardens nearby the PFAS manufacturing plant (Générations Futures, 2022; RIVM, 2021). Since summer 2022, there are PFAS exposure issues (elevated concentrations in environmen, breast milk, etc.) nearby a PFAS manufacturing plants in Lyon (France) 27. Distance to a PFAS manufacturing facility is linked to the potential exposure level, as can be seen from Gebbink and van Leeuwen (2020). Exposure to PFASs via air decreases with distance from the manufacturing facility as indicated by D'Ambro et al. (2021). B.9.2.4. PFAS processing After PFAS manufacturing and before article production often processing of PFAS is required: drying, sintering, compounding, etc. There is no emission data from processors of PFASs, but there is evidence that the drying step (sintering) of fluoropolymers has led to substantial emissions of processing aids to air at sites that produce PTFE (West Virginia (US) and The Netherlands) (Lohmann et al., 2020). https://vu.nl/en/news/2021/3m-pfas-plant-in-antwerpen-closed-down, date of access: 2022-12-14. 27 https://www.euronews.com/next/2022/08/08/forever-chemicals-how-toxic-are-the-levels-of-pfasfound-in-french-tap-water-and-breast-mi, date of access: 2022-12-16. 26 231 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.3. Textiles, upholstery, leather, apparel and carpets B.9.3.1. Introduction The use of PFASs in Textile, Upholstery, Leather, Apparel and Carpets (TULAC) is described in Annex A. Total emissions of PFASs to the environment have been assessed. For this assessment a basic source-flow model has been developed to make use of the data from the market analysis and substance identification. One key caveat here is that on a more general level a very large number of substances have been identified as being in use or potentially in use (currently around 120 unique substances). Additionally, many of these unique substances appear in mixtures as combinations of substances, and furthermore in some mixtures it may be the case that specific substances are present as an impurity, rather than an intentional use. (Note that the data from stakeholders under the CfE does not always specify which substances are impurities and which are intentional uses). The quality of market data also varies significantly from substance to substance. Therefore, the approach was not to estimate on a substance-by-substance basis, but by grouping. A key benefit of using grouping approaches is that impacts of varying specific data are lessened. The trade-off of this approach is that the estimates provided will have a higher uncertainty attached to them overall. However, this approach can still provide useful data to estimate the orders of magnitude for emissions when comparing PFAS groups and different sectors. Therefore, the emission characterisation was developed to answer three key questions:  What is the magnitude of the estimated emissions by different PFAS groups for uses covering textiles? (assumed to cover the major TULAC categories)  What are the key life-cycle stages for emissions?  What are the key distributions amongst the environmental compartments (air, land, water) for emissions? The estimates provided will help answer these questions and act as a guide as to the key groups and points of release (both in terms of life-cycle stage and receiving environment). The key caveat being that the estimates included in this section should be treated as indicative orders of magnitude and not definitive estimates to the nearest tonne. One final additional caveat to include here, is that the European Commission conducted a study in 2019 – 2020 (Wood, 2020b) covering ‘the use of PFAS and fluorine-free alternatives in textiles, upholstery, carpets, leather and apparel’. The Commission study also included an estimate of emissions for 15 non-polymeric PFASs (mostly longer chain ≥C8) as a life-cycle approach. It is not the intention of the current body of work to repeat the work already completed and presented within the Commission study. However, to maintain continuity the current approach does where possible aim to align the approach with the European Commission work, to allow some further comparisons to be made. The low estimate calculates total PFAS usage per year in the EEA (including non-polymeric PFAS and fluoropolymers) as ~41 000 t/y. The high estimate, using the same approach, calculates total usage as ~143 000 t/y. As a means of comparison and reflection, the study commissioned by the European Commission in 2020 (Wood, 2020b) provides estimates of 45 000 – 80 000 t of PFAS used in TULAC (which also includes both polymeric and non-polymeric PFAS). 232 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.3.2. Emissions In general, major life-cycle stages are broadly disaggregated into three stages:    The TULAC production stage (including manufacture of components and mixtures, and textile treating). The use stage. The waste stage. Not described here, see 0. For assumptions used, see Appendix B.9.3. Textiles, upholstery, leather, apparel and carpets. B.9.3.2.1. Production stage Based on the CfE, stakeholder interviews and literature review, the production stage covers three discrete sets of activities which have been managed separately for emission estimation:  Manufacture of non-woven membranes (an example). This production process covers the manufacture of membranes and mesh, which can be carried out as cold or hot processes using physical manipulation of fluoropolymers (e.g. extrusion). Based on tonnage, PTFE is the dominant fluoropolymer in use. However, note the possible blending of fluoropolymers with fluoroelastomers (like FKM and FFKM) to attain the desired physical properties. The membranes produced by this process can be used as the final product (e.g. technical textiles such as filters) or as a component in other textiles (e.g. as an interliner).  Manufacture of mixtures (formulations) for textile treatment. This production process covers the manufacture of commercial mixtures for textile treating. In this case the finalised mixture is a product in it’s own right and the emissions associated with treating of textiles have been included as a separate activity. The production of these mixtures will use all forms of non-polymeric PFASs, fluoropolymers and other polymeric PFASs. Furthermore, the working concentrations within mixtures is likely to vary both by final application and by PFAS used. The process of blending and mixing liquid mixtures is carried out in process vessels in various but suitable processing conditions.  Manufacture of components using processing aids. As a sub-activity to the above the CfE and discussions with industry stakeholders identified t he use of nonpolymeric PFASs as a processing aid in the manufacture of textiles. This could include the manufacture of fluoropolymer membranes (above) and non-fluorine based components for use in textile applications. The distinction made here is that the use of processing aids does not intentionally leave any fluoropolymers in the finished article. Whereas the first bullet point above used fluoropolymer as the raw material in the finished article.  Textile treating. This activity covers the follow-on step from the manufacture of mixtures and includes the use of commercial mixtures to treat textiles as a finishing step for production of finished textile articles. This process is likely to vary on a siteby-site basis, but it is assumed that mixtures are applied by a general set of approaches covering dipping, spraying, brushing, and/or rolling. The activity may take place within a wider set of steps for textile finishing, or a separate activity applied to imported/pre-manufactured textiles. B.9.3.2.2. Use stage TULAC c overs a very wide set of applications with the specific emissions varying on an application-by-application basis. Broadly the current approach identified the following major application types:  Home textiles. 233 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs)      Consumer apparel (including clothing). Professional apparel (including PPE). Technical textiles. Medical textiles 28. Other textiles. Furthermore, the study completed by the European Commission in 2020, aimed to provide a holistic approach to how emissions may occur across all these application types, bas ed on setting (indoor/outdoor use) and frequency of cleaning (including laundry)/wetting (Wood, 2020b). The Commission study also made use of the ECHA R.16 Environmental exposure assessment guidance, including ERC default emission factors to guide estimates (ECHA, 2016a). To align continuity, the same approach has been adopted within the current study. Three overall release groups were identified:  Release Group 1. (Indoors) treated textile or leather articles that are subject to frequent cleaning/washing and so have higher potential for release. Examples could include all types of clothing, PPE where visibility is an issue (i.e., High-Vis jackets), and medical textiles.  Release Group 2. (Indoors) treated textile or leather articles that are subject to infrequent cleaning/washing. Examples could include carpets, rugs, and curtains.  Release Group 3. (Outdoors) Other treated textile or leather articles, not included in group 1 or 2, that are used in situations where a low level of release could occur. Examples could include footwear, outdoor clothing, outdoor technical textiles, other textiles used outdoors. Based on the three groups detailed above the major application types have been assigned as described in Table B.22. Table B.22. Assigning major use applications to release groupings . Major application type Home textiles C onsumer apparel Professional apparel Technical textiles Medical textiles Other Sub-uses C arpets and rugs C urtains Release grouping Group 2 Group 2 Upholstery Indoor and outdoor wear Group 2 Group 1 Sports wear Footwear Professional sportswear Footwear PPE Outdoor technical textiles High performance membranes Medical textiles All other uses Group 1 Group 3 Group 1 Group 3 Group 1 Group 3 Group 2 Group 1 Group 2 or 3 Note that where several of the major application types in Table B.22 cover more than one release group, it has been necessary to apply a split to the market data to apportion the quantity of PFAS used within the application type and its connected release grouping. Table B.23 provides details of how these splits have been applied. 28 Note that medical devices can also include textiles in some cases. Medical devices are covered separately in B.9.10. To maintain clarity, medical textiles used in this section refers to any use of textiles in a medical setting, excluding use within or on the patient (i.e., implantable textiles li ke gauzes or applications used upon the body like bandages are included under medical devices). As examples of the textiles included within the current definition this could extend to articles such as mattress protectors upon hospital beds, curtains/drapes around the bed, and gowns / PPE used by medical professionals. 234 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.23. Market splits for situations where more than one release grouping exists . Assigning major use applications to release groupings. Justification Release Group Release Group Major Release Group 3 2 applica1 outdoor – indoor – tion type indoor – assumed low infrequent frequent release as % wetting as % wetting as % All uses under this application are covered by Group 2 only. This application includes clothing more generally. Based on the C fE, outdoor clothing such as ‘raincoats’ and hiking gear may be a more dominant use than items like sportswear. This group is likely to be highly diverse. In lieu of data assume equal splits. Very diverse application type. In lieu of data assume equal splits. Home textiles 0 100 0 C onsumer apparel 40 0 60 Professiona l apparel 50 0 50 Technical textiles 33.3 33.3 33.3 Medical textiles 100 0 0 This application appears under group 1 only. Other textiles 0 50 50 In lieu of data split tonnages equally. Table B.97 and Figure B.80 in Appendix B.9.3. provide a summary of all factors applied within the model and key assumptions which guide the flow of PFASs through various stages of products life-cycles. ECHA ERCs were used to estimate emissions bec ause more specific emissions factors could not be found. The overview of the estimated EEA yearly emissions for 2020 is shown in Table B.24. Most of the emissions go to air since PFAA precursors are volatile to a large extent, whereas the water soluble PFAAs occur at a lower degree in processing mixtures. For non-polymeric PFASs the primary release point to environment (key life -cycle stages for emission) is either during the treatment of textiles with PFAS-based mixtures, or from the use stage for indoors textiles or leather articles subject that are subject to frequent cleaning/washing and so have higher potential for release. Overall, the combined production of commercial mixtures and treating of textiles as a finishing step may be more emissive to the environment than the production of non-woven membranes. Based on the existing quantity data of PFASs, the "low estimate" appears to be the more credible estimate of quantities of PFASs in the EU/EEA 29. 29 C onclusive remarks on the provided quantity data of various PFASs reported in the study on TULAC carried out within the work on this dossier (see section G.2.3) were that based on existing quantity data of PFASs, "other sources of information" and comparative assumptions and calculations made by the consultant for "low estimate" and "high estimate", then the "low estimate" appears to be the more credible estimate of quantities of PFAS in the EU / EEA. 235 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.24. Estimated annual EEA emissions of PFAA and PFAA precursors (+ other PFASs), fluorinated gases, polymeric PFASs and total PFASs in 2020 (baseline) in the TULAC sector. C2- C3 PFA S substances (t/y) PFA A ≥C4 (t/y) Sidechain fluorinated polymers (t/y) Total PFA As and PFA A precursors (t/y) Fluoro polymers (t/y) PFPE (t/y) low high low high low high low high low high low Total volume 2 150 10 300 3 512 8 612 2 430 14 236 8 092 33 148 32 305 107 861 Total emission 508 2 434 605 1 491 945 6 370 2 058 10 295 7 914 24 076 Total polymeric PFA Ss (t/y) Total PFA Ss (t/y) high low high low high 786 1 683 33 091 109 544 41 183 142 692 412 884 8 326 24 960 10 384 35 255 236 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.3.2.3. Past emissions The previous section provided a set of emission estimates for PFASs used in TULAC for the year 2020 using a source-flow approach to provide a baseline. This section further provides additional estimates to map trends covering the backward-looking time-series (1990-2020). As part of the forward projections, assumptions around ongoing REACH restrictions have been implemented to best knowledge using a conservative approach as follows:  C9-C14 non-polymeric PFASs that include PFAA precursors for the formation of side chain fluorinated polymers. In August 2021, a group restriction was included in Annex XVII, REACH on perfluorinated carboxyl acids (C9-C14 PFCAs) and those substances that may degrade to them. This means that they are restricted from 25 February 2023. C9-C14 PFCAs and their related substances in the following certain textiles shall not, from 4 July 2023, be used in, or placed on the market in: o textiles for oil- and water-repellency for the protection of workers from dangerous liquids that comprise risks to their health and safety; o the manufacture of polytetrafluoroethylene (PTFE) and polyvinylidene fluoride (PVDF) for the production of: o high performance, corrosion resistant gas filter membranes, water filter membranes and membranes for medical textiles; o industrial waste heat exchanger equipment; o industrial sealants c apable of preventing leakage of volatile organic compounds and PM 2,5 particulates. The projections in the August 2021 C9-C14 PFCA restriction report (which was finished before the publication of the restriction) assume a more conservative estimate with implementation by early 2022 at the latest and close of the transition window by the end of 2024, after which all use will cease. The market data from the CfE only, identifies two tonnes of C9-C14 PFCAs in TULAC in use for the 2020 baseline, and therefore it is possible that significant steps to transition or reduce impurities have already been taken. As a general point regarding the development of emissions projections it is important to recognise that prediction of future usage rates and technologies are v ery challenging and cannot consider unpredicted sudden world events (with Covid-19 being a good example). Furthermore, the more forward in time the projection is the greater the uncertainty in future trends or events. Therefore, any estimate of usage (and associated emissions) as far forward as 2050 should be treated with a great deal of care and used only as indicative of possible usage rates and their associated emissions. In the current context to help guide both the backward-looking estimates and the projections some common rules have had to be established, which are detailed below:  Firstly, it is assumed that activity equals emissions. The source-flow model provides data as tonnages of PFASs (disaggregated between the different substance groups) for different application types (e.g. home textiles). It is assumed that where there is demand for PFAS in these applications that the emissions will be generated proportionate to the baseline. For example, if the demand for home textiles grows strongly in the next 30 years, it is assumed that the emissions of PFAS from that sector will also grow strongly and proportionate to demand. Therefore, growth rates in different application types can be used to provide proxies for emission trends.  Secondly, based on the baseline emissions for those substance groupings and lifecycle stage combinations where there was no activity/emission it is assumed that this has always and will always be the case. For example, the manufacture of nonwoven membranes utilises only fluoropolymers and side-chain fluorinated polymers 237 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) and no non-polymeric PFASs, and therefore it is assumed this has always and will always be the case.  Where market trend data has not been identified a uniform growth rate of 2% per year is applied. The CfE and many stakeholder interviews demonstrated that the physical properties of PFASs are highly desirable within specific applications, and that therefore, as the economy of the EEA has grown since 1990 (and will continue to do so, see OECD (2015)), it can be assumed there will be increased demand for PFAS from 1990 (first year of the estimates) onwards.  Changes in technology and processes have not been considered within the projections. It is possible that even under a business-as-usual scenario that the science will continue to evolve and new types of PFASs brought into commercial use. No effort has been made to consider these elements for the forward projections.The unpredictability of such developments is high. (In the automobile sector for instance the electrification has a disruptive effect). Approach for backward looking time-series (1990-2020) – Assumptions and caveats For the back-ward looking trends, Wang et al. (2014b) provides some useful steer based on a global level set of emission estimates for PFCAs. The authors comment that the usage and emission rates for PFCAs broadly go in three movements, which commences from 1951 – 2002, during which time the commercialisation of PFAS was completed and exponential growth in their use took hold 30. In the second movement from 2003 – 2015, mounting concerns around the longer chain PFASs, particularly PFOS saw a shift in production and use. Longer chain C8 PFSAs such as PFOS were phased out by the major global producers of PFAS with a diversification towards shorter chain (C4 and C6) PFASs. This changes the emission profile, with emissions of PFOS in particular falling sharply from recognised point sources (particularly manufacture). In the third and final movement from 2016 – 2030 with the global level ban on PFOA and multiple regulatory restrictions covering not only the PFOA chemistry (C8 PFCA) and PFOS chemistry (C8 PFSA), but also the PFHxS chemistry (C6 PFSA) and longer chain C9-C14 PFCAs, there is a further diversification and move away from longer chain PFAS into other types of PFASs. Table B.98 in the appendix provides some further details of the assumptions used to help map the back-ward looking (1990 – 2020) trends in usage of PFASs in specific application sectors. These growth trends in different application sectors drive the overall use of PFASs. Therefore, the average growth rate across all six major application sectors (home textiles, consumer apparel, professional apparel, technical textiles, medical textiles, and other textiles) has been derived and applied to the baseline market data from 2020 to map total usage rates from 1990 – 2020. The same averaged growth rate has also been applied to the overall emissions from the 2020 baseline to derive annual total emission estimates from 1990-2020. As a caveat, it is recognised that based on the three release groupings from the previous section B.9.3.2, i.e., o o o Group 1 – Indoor use – frequent cleaning/wetting; Group 2 – Indoor use – infrequent cleaning/wetting; and Group 3 – Outdoor – low release, that specific application types may be more emissive than others. Applying an averaged growth rate in the current approach will miss any nuance in the changing balance of PFAS-based substance used across the different application types. However, given the uncertainty in the overall emission estimates and variances in market data (and not least the assumptions in applied growth rates), these changes are likely to be within the overall uncertainty. Deriving and 30 As a side note, “The Introduction to Fluoropolymers“ (2nd Edition), authored by Ebnesajjad (2021), and published by Elsevier, comment that the first fluoropolymer, PTFE, was commercialised in 1947. 238 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) applying an averaged growth rate based on underlying changes at the application type level provides a transparent and simplified approach to provide emission estimates that can be used to understand what the future picture may look like. The usage rates for PFASs are provided in Table B.25 while the results of the overall emission estimates as a time-series are provided in the sub-section that follows after the discussion on the 2020-2050 assumptions. Table B.25. Assumed TULAC usage rates for PFASs 1990 – 2020 (using low estimate value as baseline). All values in tonnes. PFASs groups Non-polymeric C2-C3 substances Non-polymeric C4 Substances Non-polymeric C5 substances Non-polymeric C6 substances Non-polymeric C9-C14 substances Non-polymeric other substances Fluoropolymers and other polymeric PFAS excl side chain fluorinated polymers Side-chain fluorinated polymers Overall total use 1990 1995 2000 2005 2010 2015 2020 1 173 1 297 1 435 1 588 1 757 1 943 2 150 8.8 9.7 10.7 12.4 14.7 16.5 17.0 0.50 0.55 0.61 0.72 0.88 1.00 1.00 1 506 1 666 1 843 2 170 2 661 3 104 3 399 1.1 1.2 1.3 1.5 1.6 1.8 2.0 50 56 62 68 76 84 93 15 452 17 094 18 911 20 922 23 145 25 606 33 091 1 483 1 615 1 760 1 921 2 100 2 297 2 430 19 675 21 740 24 025 26 683 29 756 33 053 41 183 B.9.3.3. Human exposure Exposure to PFASs in textiles may occur directly via dermal uptake and, particularly in children, via hand-to-mouth and object-to-mouth exposure (e.g. when sucking on clothes). However, these exposure routes have been estimated to be insignificant compared to other routes (EPADK, 2015a). During the use stage, PFASs can be released from textiles through a number of mechanisms, such as evaporation of volatile PFAS residuals from the fabric, loss of particles and fiber fragments by abrasion, breaking of the carbon-oxygen bond of side-chain fluorinated polymers present in the fabric, and wash-out of water-soluble residuals, such as PFCAs (Xiong and Haddad, 2021). Evaporated PFASs and PFASs in abraded particles and textile fibers can end up in indoor dust and air, which are known exposure routes to PFASs in humans (see B.9.21.4). In certain indoor environments with large quantities of textiles, such as furnishing, carpets and clothes, the levels of PFASs in air and dust may be elevated. For example, high levels of FTOHs have been reported in shops for outdoor wear and equipment (Langer et al., 2010; Schlummer et al., 2013). For the staff working in these shops, the exposure to PFOA from the indoor air (based on 5% conversion of FTOH to PFOA) was estimated to be of the same magnitude as the normal dietary intake of PFOA. In a study of a Chinese texile manufacturing plant, the levels of FTOHs were reported to be 2-3 orders of magnitude higher than the levels measured in the shops mentioned above, indicating a potential of high exposure to textile workers (Heydebreck et al., 2016). 239 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.3.4. Summary PFAS emissions occur during the production and use stage during the life cycle of TULAC products. These emissions were calculated using either default ERC information specified in the REACH methodology or with more specific emission data where available. Total emission was estimated between 10 000 and 35 000 t/y (rounded numbers). 240 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.4. Food contact materials and packaging B.9.4.1. Introduction Uses of PFASs in food (and feed) contact materials (FCM) and packaging are described in Annex A. In the following section the emissions for three main food contact and packaging uses, listed below, are described under the manufacturing and service life stage.  Packaging    Food (and feed) contact packaging Generic packaging Other packaging: Coating of beverage cans, ink and lacquer residues in labels and car wrapping.  Consumer and industrial cookware  Industrial food and feed production equipment Waste stage emissions are covered elsewhere (see B.9.18). Understanding of waste management (recycling, landfill, incineration or composting) and the fate of PFASs sent to waste management is very important for understanding the potential for full life cycle emissions. This applies especially to this uses as (food) packaging material (paper as well as plastics) are recyclables with Extended Producer Responsibility regimes and are stimulated to be recycled. B.9.4.2. Manufacturing stage emissions Manufacturing stage emissions are described below and summarized in Table B.26. B.9.4.2.1. A - Packaging manufacturing A1) Food (and feed) contact packaging manufacturing For packaging, emissions of PFASs are far more relevant in the manufacturing stage than in the use stage as the production process is a ‘wet process’ (i.e. for paper & board packaging, large quantities of cellulose pulp in water are moved through various stages of processing to get to the final product). Limited specific emission data have been identified that enable a comprehensive estimation of the emissions of PFASs from the manufacturing of articles. However, some data were identified that allowed an estimation of emissions especially for the manufacture of Paper & Board (P&B) packaging. Information was based on US EPA data which was based on industry data . Data on PFASs paper & board, was available in the US Inventory of Effective Food Contact Substance (FCS) Notifications (FDA-US, 2021). Specifically, some FCS notifications have an associated environmental assessment which may include estimates of PFAS releases from the paper-making process, i.e. product manufacture. These estimates have been used to derive emission estimates across the European Economic Area (EEA). According to the REACH methodology the total emissions in the manufacturing phase to air, water and soil compartments is between 943 – 6 881 t/y. The calculations based on FCS results in emissions between 113.1 – 825.7 t/y. Since the FCS data are actual data from the paper and board manufacturing they are considered to be more accurate than the REACH data. The paper and board manufacturing emissions are based on USA FCS Notifications industry FDA data and are considered to be worst case scenario emissions (FDA-US, 2021). See appendix for more information on release factors and assumptions. 241 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) A2) Generic plastic packaging manufacturing For plastic packaging, emissions, although largely unknown, also mainly occur in the manufacturing stage. PFASs in generic (thermoplastic) packaging have no function in the final article. It has been assumed that the proportion of non-paper and board packaging such as plastics containing PFASs is low to negligible, and that the plastics that do contain PFASs have it in low concentrations (BfR, 2020). Despite extensive stakeholder consultation and literature search, data collected on PFASs emission in generic plastic packaging were limited. Nevertheless, with the available data, emissions of PFASs from generic packaging manufacturing have been addressed in a qualitative manner. For generic plastic packaging PFAS processing aids are a possible source of emission during manufacturing. An additional source could also be external lubrication of processing equipment where sometimes PFASs are also used. Due to the absence of industry data it was not possible to provide an estimate of PFAS emissions from plastic packaging manufacturing. However as PFAS processing aid residues are estimated to be present in non neglectable tonnages (300 – 600 t, see Annex A.3.4.2.) and the production process is a ‘wet’ process where external (PFAS containing) lubrication is of relev ance, emissions likely cannot be neglected. A3) f-HDPE containers manufacturing Based on US EPA data, it is seen that fluorinated high-density polyethylene (f-HDPE) containers and similar plastics (i.e., fluorinated polyolefins) can result in PFAS contamination of the product (being e.g. food)31. Fluorination of plastic containers creates a barrier on the plastic’s surface and increases packaging strength. According to the EDF investigation, the types of containers currently undergoing the fluorination process range from “packaged food and consumer products to larger containers used by retailers such as restaurants to even larger drums used by manufacturers to store and transport fluids.” There is however no data on PFAS emissions from fluorination of HDPE containers. Potentially, emissions of fluorinated gases could occur but no data was available. Information what was available is mentioned below. During the fluorination process, HDPE containers are subjected to fluorine elemental gas under elevated temperatures. The anticipated chemical reaction results in formation of partially fluorinated long chain polymers and possibly fully fluorinated short chain polymers. Similarly, when a mixture of fluorine and oxygen is used during fluorination, oxy -fluorinated polymers are formed. If short chain polyethylene compounds (either impurities/by -products in the manufacturing of HDPE containers or from the degradation of HDPE during fluorination) are present during fluorination, further chemical reaction could occur and transform the oxyfluorinated polymers into perfluorinated carboxylic acids. US EPA - EPA’s Analytical Chemistry Branch PFAS Testing - Rinses from Selected Fluorinated and Non-Fluorinated HDPE Containers. Hundreds of million containers, articles, and plastic bottles are fluorinated each year 31. A4) Other uses: Coating of beverage cans, ink and lacquer residues in labels and car wrapping: Industry specific data on PFASs used in beverage can coating manufacturing, ink and lacquer manufacturing, and car wrapping production were scarce. Emissions for manufacturing of PFAS coated cans, ink and lacquer residues and car wrapping were not available. During the extensive stakeholder consultation, data returns on this issue were very low and manufacturing stage https://www.foodpackagingforum.org/news/plastic-container-fluorine-treatments-createpfas, date of access: 2022-12-14. 31 242 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) emissions could not be calculated. Therefore ECHA ERC 10a (outdoor use) was applied. B.9.4.2.2. B - Consumer and industrial cookware manfacturing: The information was insufficient to allow a quantitative estimation of the emissions during the manufacture of consumer and industrial cookware in which fluoropolymers are used. Due to the absence of detailed manufacturing process information, even a default calculation using the REACH methodology was not enough to derive a quantitative estimate. Emissions from manufacturing consumer and industrial cookware containing PFASs were estimated on a qualitative basis due to the absence of sufficient data for a quantitative analysis. Qualitative evidence was provided by stakeholders (CfE and 2 nd stakeholder consultation) and was therefore used to estimate t he emissions during normal operating conditions of manufacturing installations. Accidental emissions were not accounted for. Fluoropolymer emissions from re-coating operations of cookware (mainly industrial cookware) could be significant because of the quantities of coatings removed, but data has not been received on the fate of this fluoropolymer. PFAS emissions from re -coating industrial cookware were estimated from data received during the stakeholder consultation. The wider question of emissions from disposal is described in B.9.18. B.9.4.2.3. C - Industrial food and feed manufacturing: For manufacturing of industrial food and feed equipment the production process is a ‘dry ’ process and emissions during production are usually low. Analysis of the emissions from the manufacturing of PFAS (mainly fluoropolymers) for use in industrial food processing suggest emissions are lower than in ‘wet processes’ like the manufacturing of paper & board packaging. Although there are also indications that in the production of PFAS coated materials (PFAS) emissions cannot be neglected. See for instance the remarks from US EPA permit information (Super-Temp Wire & Cable Inc., 2014). In PTFE paste extrusion lines, used for instance to coat cables and tubes used in a.o. electric equipment for food and feed production/preparation, emissions of particulate matter (PM), volatile organic compounds (VOCs), hazardous air contaminants (HACs), hazardous air pollutants (HAPs) and toxic thermal decomposition products are of relevance (Super-Temp Wire & Cable Inc., 2014). Paste extrusion lines, printing lines, melt extrusion lines and fuse lines of fluoropolymers all have the potential to generate toxic thermal decomposition products. The carbonyl fluoride produced hydrolyses to hydrogen fluoride. Emissions of these toxic gases are not quantifiable as they are highly variable due to variations in oven t emperature and line speed. However, permit data suggests that decomposition products can be minimised by maintaining oven temperatures below 540 °C. Table B.26. Emissions of PFASs and methods used in the manufacturing stage. Sub-use in PFAS PFAS Type Estimation Methodology Other Details Manufacturing emission *** Scenario stage estimate (t/y) A1) Food & Feed Packaging: Paper & Board 113 – 826 Telomers and polymers with side-chain telomers Reasonable worst-case C alculation based on FDA data. The FDA approach is preferred as it accounts for data from manufacturing Based on data for large installations in the USA: A wet process with significant emissions 243 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Sub-use in Manufacturing stage PFAS emission estimate (t/y) PFAS Type *** Estimation Scenario Methodology Other Details sites A2) Generic plastic packaging No data PPAs A3) f-HDPE containers No data PFAA (precursors) A4) Other uses No data – likely low Fluoropolymers B) Consumer Cookware Very low Fluoropolymers including fluoropolymer PPA Normal operating conditions Estimation from stakeholder data Qualitative assessment C) Industrial food and feed production equipment Very low Fluoropolymers and PPAs Normal operating conditions Estimation from stakeholder data Qualitative assessment C2) Re-coating Industrial Cookware 0.4 – 1.2**** Fluoropolymers and PPAs Realisticcase C alculation from stakeholder data C alculation is for PTFE only TOTAL 113-827     No data High tonnage (1 640 – 3 280 t) but no specific emission data for thermo plastic production. ERC s applied Important topic in USA No data Mainly caused by paper & board manufacturing * Standard REAC H methodology (EC HA, 2016a). ** Based on 2% loss of fluoropolymer-coating during use. May be an under-estimate as can be seen in Figure B.70. *** Type of PFAS from which emissions originate using expert judgement. † May be an over estimate. **** Emission to soil and water. Based on 90% disposal via waste. 244 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.4.2.4. Service life emissions A - Packaging service life Service life emissions of food contact material and (food/feed) packaging are of lesser relevance than manufacturing emissions. No industry specific emission factors were available. Estimates presented here are derived using the first level (tier 1) methodology. This relies upon having tonnage data for the substance(s) of interest at the relevant point in the life cycle of that substance. Of the various ERCs defined the most appropriate were chosen, that being ERCs 10a and 11a. A description and justification behind the chosen ERCs are described in Appendix B.9.4. Food contact materials and packaging. A1) Food (and feed) contact packaging service life For emissions during the service-life of paper & board used in food and feed contact, the outdoor emissions of PFASs predicted from the REACH methodology are recommended to be used as a worst case scenario. In real world conditions it is unlikely food and feed packaging will be left outdoors and exposed to weathering during its service-life for any length of time. However for takeaway food, outdoor emissions could occur for example during transportation. In addition, paper & board packaging used for food and feed contact that has been discarded i.e. litter, will be exposed to weathering and therefore these emissions may be realised in part. A precautionary approach is recommended for instance to use the higher outdoor figures which could be seen as a worst-case scenario. In Figure B.68 a tonnage and emission flow chart for paper & board packaging (the main packaging use where PFASs are applied) is plotted. Service lifes emissions as well as manufacturing stage emissions are covered. Values in the blue boxes represent tonnage flows of PFAS/year while values in the red boxes represent emissions of PFAS/year. As can be seen below in the red boxes the manufacturing emissions are by far the most relevant. More details on emission are presented in the Appendix. A distinction between indoor and outdoor emissions is made. 245 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.68. Indicative PFAS tonnage flows (t/y) for paper & board packaging in service life and manufacturing stage. Values in blue boxes are tonnage flows of PFAS/year. Values in red boxes are emissions of PFAS/year. Left branch: detected (Table A.22. in Annex A), right branch: intentionally added (Table A.23. in Annex A): Giving similar results. EU-27 & NO & UK. A2) Generic plastic packaging service life Due to limited research on PFASs in plastic packing, data on service life emissions for PFASs in generic packaging were unavailable. A precautionary approach is used for the emission estimate i.e. to use the higher outdoor figures which could be seen as a worst-case scenario. Therefore for service life the ‘standard REACH methodology’ ,ERC 10a, has been used. A3) Other uses: Coating of beverage cans, ink and lacquer residues in labels and car wrapping Data on PFASs used in lacquers & ink, beverage can coating and car wrapping were scarce. During the extensive stakeholder consultation, data returns on this issue were very low. Nevertheless, on basis of available tonnage data (see Annex A, section A.3.4.2.), emissions of PFASs have been addressed in a qualitative manner. Service-life emissions as a result from these uses were calculated using the ERCs mentioned in Appendix B.9.4. Table B.99. A4) f-HDPE PFAS emission into the environment from fluorinated plastic like f-HDPE containers is likely to be low. However, PFAS migration from fluorination of HDPE containers into stored products might be a serious concern (EPA-US, 2021c). B - Consumer and industrial cookware service life B1) Consumer cookware No publications have been identified that indicate the quantity of fluoropolymer lost during cookware use by consumers, although at the end of 2022 an article was published on micro plastics release from consumers cookware (Luo et al., 2022). However, numerous examples have been found online on consumer websites of pans that have clearly lost a significant part (2% or even 10% or more) of the fluoropolymer-coating during use, revealing the bare metal as shown in Figure B.70. An illustrative upper bound can be calculated assuming a 2% loss of 246 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) material per year for the consumer cookware sector, equivalent to an emission of 70 t/y (3 500 * 0,02). The position of manufacturers is that consumers should purchase new kitchenware when the non-stick finish no longer performs as well as desired. This will likely correspond to only a small loss of material to consumers or the environment. No data sources have been identified that indicate the quantity of fluoropolymer lost during use by consumers. B2) Industrial cookware Given the criticality of the coating for industrial food production it seems unlikely that damage similar to that seen in the consumer cookware market would be tolerated. Emissions during service life (re-coating excluded) likely is very low. In Figure B.69 a tonnage and emission flow chart for consumer cookware and industrial food & feed production equipment is plotted in which the service life as well as manufacturing stage emissions are covered. Emissions in the use stage are expected to be low unless equipment is improperly used, for example scratched or overheated (Luo et al., 2022; Schlummer et al., 2015). Most polymeric PFASs in these articles enter the end-of-life stage where additional emissions might occur for instance from landfills or incineration. During the preparation of the dossier, industrial re-coating of cookware has been seen as a potential emission source 32. This use is not describe in Annex A. Commercial bakeries use nonstick cookware and the coating is regularly replaced. In the case of non-stick baking pans after 12-24 months the old PTFE coating is removed and a new coating is applied. Comprehensive data on the quantities of fluoropolymers removed from industrial cookware for recycling purposes (which is then re-coated) have not been identified. However, for Sweden and Germany, information was available as shown in Table B.27 and has been extrapolated to the whole of EEA. Recoating is also done for cheese forms and likely other food contact items as well. Given the uncertainties with the tonnage and emission estimates, the values should be regarded as indicative only. Table B.27. Quantities of PTFE in Industrial Cookware re -coating according to data. Parameter Quantity Source Pans re-coated in Sweden each year Pans re-coated in Germany each year Total surface area of the Swedish + German yearly recoated pans 20 000* 100 000* (RIVM, 2020) Stakeholder information 12 000 m 2 (0.1 m2/pan) Stakeholder information Density of PTFE 2.16 g/cm 3 = 2.16 t/m3 (WSH, 2021) Average coating thickness industrial cookware 35 – 100 µm (PTFE C oatings, 2021) and stakeholder information Population of Sweden + Germany 10.38+83.16= 95.54 million (Statista, 2020) Population of EEA (without UK) 454 million (Eurostat, 2020) Notes: *Assumed to be number of units stripped and re -coated per year. 32 Re-coaters advertised on Google for instance as seen here www.itn­coatings.com or https://hoffmann­ germany.de/pages/wiederbeschichtungs_service, date of access: for both: 2022­12­14. 247 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Each year in Sweden and Germany an estimated 0.42 – 1.2 m3 PTFE coating is replaced. Therefore 0.9 – 2.6 t of PTFE are removed from industrial bakeware annually in the two countries. Extrapolating to EEA via population figures leads to: 454/95.54*0.9=4.3 t up to 454/95.54*2.6=12.4 t removed volume fluoropolymer, which has to be split into emissions and waste. In case 10% of the re-coated tonnes are emitted, as realistic case, this leads to an emission of 0.4–1.24 t/y. The other 90% is discarded as waste (where also emissions might occur). According to one stakeholder, the old coating is partially burnt out at 420 °C in a special circulating air oven. The resulting processed air is post -combusted at >1 000 °C. Closed-circuit gas scrubbing water is disposed of as hazardous waste. Sandblasting waste is said to be disposed of as hazardous waste. Other sources mentioned less rigorous destruction. Finally, not all hazardous waste is disposed as hazardous waste. Therefore, the recoating activity remains a potential source of emissions. C - Industrial food and feed equipment service life Industrial food and feed equipment consist of conveyor belts, O-rings, etc. used in factories which produce food or feed. Emissions of PFASs during the service-life of industrial food and feed equipment when used in the food industry are based solely on indoor use and represent a reasonable worst-case scenario, based on ERC 10a (indoor use). Emissions from the use of fluoropolymers in industrial food processing and the pharmaceutical sector 33 in the EU indicates a low emission of 0.2 t/y as can be seen in the red box on the right side of Figure B.69. As mentioned in Annex A, volumes for industrial food processing cannot be disa ggregated from the pharmaceutical sector. Figure B.69. Indicative PFAS (Fluoropolymers) tonnage flows (t/y) for both consumer cookware and industrial food products in 2015 (EU-28) in service life as well as manufacturing stage. Tonnages and emission are represented by the blue box and red box respectively. Emission of fluoropolymers into the environment are represented by the red boxes shown in Figure B.69. Values in blue boxes are tonnage flows of PFASs from manufacturing through the 33 The separate proportions going to food and pharmaceuticals has not been able to be disaggregated. 248 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) service-life of products to waste/recycling. Polymeric PFAS emissions for consumer as well as industrial use are expected to be low during (proper) use. For consumer as well as industrial cookware overheating of fluoropolymer coatings is not recommended due to the release/emission of fumes that may have detrimental health effects. Modern ovens, both for consumers and professionals, however, could easily exceed 300 °C posing a risk to fluoropolymer coating emission. For service life, the ‘standard REACH methodology’/ERCs has been employed. Estimates presented here are derived using the first level (tier 1) methodology. In Table B.28 the PFAS emissions from all distinguished sub-uses are summarised for the service-life stage. Table B.28. Emissions of PFASs and methods used. Service life stage (proper use assumed). PFAS emission estimate (t/y) 3 – 20 (based on two different calculation methods) PFAS Type *** Estimation Scenario Methodology Other Details Telomers and polymers with side-chain telomers Worst-case ERC REAC H* Based on outdoor emissions, ERC 10a A2) Generic plastic packaging Lacquers and ink 5 – 11 PPAs (residues) Worst case ERC REAC H 2 Non polymeric Based on outdoor emissions, ERC 10a A3) Coated beverage cans 16 Fluoropolymers (PTFE wax) Estimation ERC REAC H Based on outdoor emissions, ERC 10a A4) f-HDPE No data, likely fluorinated gas 13 Fluoropolymers - - Actual topic in the USA Fluoropolymers Worst case estimation ERC REAC H Based on outdoor emissions, ERC 10a B1) Consumer cookware 70** Fluoropolymers Reasonable worst-case Own calculation assuming 2% loss per year, although sometimes 10% was seen (see Figure B.70) Severe loss of polymer coating possible in case of low quality product and/or improper use B2) Industrial cookware Very low Fluoropolymers Reasonable worst case No data Based on industry quality standards Sub-use A1) Paper & Board FCM Packaging A5) Other uses (Car wrapping) 249 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Sub-use C) Industrial food and feed production equipment TOTAL service life, excl. waste stage (rounded numbers) PFAS emission estimate (t/y) 0.2 PFAS Type *** Estimation Scenario Methodology Other Details Fluoropolymers Worst-case REAC H Based on indoor emissions, ERC 11a 109 – 132 PFAS used in paper & board packaging are surfactants (telomers and polymers with telomeric side -chains/ side chain fluorinated polymers; PFAS used in consumer and industrial cookware are fluoropolymers.    Standard REAC H methodology (EC HA, 2016a). ** Based on 2% loss of fluoropolymer-coating during use. May be an under-estimate as can be seen in Figure B.70 *** Type of PFAS from which emissions originate using expert judgement. The emissions for food and feed contact material and packaging are further summarized in Table B.29. The main source of emission is attributed to the manufacturing of paper & board packaging. Table B.29. Estimated annual EEA emissions in the FCM and packaging sector of PFAA and PFAA precursors, polymeric PFASs and total PFASs in 2020 (baseline) based on Table B.26 and Table B.28. Total polymeric PFASs Total PFAS Total PFAA and PFAA (t/y) (t/y) precursors (t/y) low midpoint high low midpoint high low midpoint high Total volume 3 267 6 305 9 342 15 330 17 880 20 430 18 597 24 185 29 772 Manufacturing stage emissions Service life stage emissions 113 470 826 0 0,5 1 113* 470 827* 10 21 32 99 99 99 109 121 132 Total emission 123 491 858 99 100 100 222 591 959 *: Based on USA FC S Notifications industry FDA data. C annot be fully disaggregated but mainly telomers and sidechain fluoropolymers. B.9.4.3. Human exposure The high price of fluoropolymers is mentioned by producers as a driver for minimising use and exposure. According to stakeholders, PFAS exposure to humans is very limited as (local) strict regulations and approval schemes are in place, although not always strictly enforced. On the other hand, available literature data suggest that diet is the major human exposure pathway for some PFASs (especially non polymeric PFASs), but that there is large variability in exposure levels across populations (De Silva et al., 2021). A possible source of fluoropolymer exposure 250 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) can be losses of fluoropolymer material from wear and tear of consumer cookware (Luo et al., 2022). These losses can be very high, particularly for low quality goods, see Figure B.70 shown below. According to a BfR FAQ of 2018 (BfR, 2018), PTFE minute particles, released from scratched coatings of coated cookware, ovenware and frying pans can be ingested. This is considered to not pose a risk to human health, since PTFE particles are excreted from the body unchanged. However, this applies only as long as the recommended conditions of use of fluoropolymers to protect human health are applied (e.g. no overheating). In the use stage PFAS migration to food from paper and board contributes significantly to tolerable weekly intake (Lerch et al., 2022). Additionally, it is stated that toxic vapours from fluorinated compounds and particles develop at temperatures above 360 °C which means an overheating of coated cooking, baking and frying utensils, especially when empty (BfR, 2018). Figure B.70. Loss of PFAS coating during use. FCM from paper and board (e.g. fast food packaging) can be a source of human exposure towards non-polymeric PFASs. However, with the exception of very specific exposure scenarios like very frequent consumption of microwave popcorn, the impact of FCM as a source of human exposure towards PFAS seems to be low compared to other sources such as food. EFSA (2020) concluded in their risk assessment of PFASs in food that exposure to residual PFASs in PTFE cookware was low in comparison to other sources (see also section B.9.21.1). Moreover, fluorinated HDPE containers could lead to PFAS exposure of consumers when food is stored in these containers (Rand and Mabury, 2011) (Figure B.71). Fluorination of HDPE packaging is however commonly used in the US to treat polyethylene and polypropylene containers each year and if (illegally?) applied in Europe could be a source of human exposure towards non polymer PFAS 31. 251 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.71. Concentration PFAS found in rinsate of packaging containers (ppb). Source: EPA-US (2021c). Fluorination of plastic is commonly used to treat hundreds of millions of polyethylene and polypropylene containers each year and could be a serious source of human exposure (especially if food is packed in such containers). B.9.4.4. Summary Emissions from the food contact and material sector were calculated for the manufacturing and service life stage using either default ERC information specified in the REACH methodology or with more specific emission data where available. Manufacturing stage emissions are of higher importance than service life emissions. Total emission of PFASs calculated for the manufacturing stage and service life stage together, amounted to 222 to 959 t/y. This very likely is an underestimation as for instance in the manufacturing stage quite some emission values are lacking. Human exposure to PFASs in food contact material may include lost material from consumer cookware from wear and tear, food packaging and storage of food in f -HDPE containers. 252 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.5. Metal plating and manufacturing of metal products B.9.5.1. Introduction The use of PFASs in this sector is described in Annex A.3.5. B.9.5.2. Emissions A distinction between emissions during metal plating and the manufacture of metal products is made. B.9.5.2.1. Metal plating Emissions of PFASs during metal plating processes originate e.g. from the rinsing steps between the electrolytes and from replacement of used solutions (UBA, 2017b). The galvanic bath is used for a couple of years. It is renewed e.g. when the emerging sludge may hamper the galvanic process. The galvanic solution without further treatment must be considered as process solution and becomes liquid waste. It is forbidden to dump the solution directly into a WWTP, not even if the solution is diluted. It has to be treated in a physical-chemical processing plant. In this plant metals are reclaimed and other substances are removed. PFAS can be removed via ion exchanger from the galvanic solution. But this technology is not state of the art yet and therefore it is not mandatory. Currently, the Reference Document on Best Available Techniques for the Surface Treatment of Metals and Plastics (STM-BREF) is revised. It is likely that this treatment step would become obligatory. According to the German central association for surface technology (Zentralverband Oberflächentechnik e.V., ZVO) in Germany 20% of the applied surfactant is lost by the plating processes itself over time, e.g. by trickling from treated devices (EPA-DK, 2011). For the lifetime of the surfactant in the galvanic bath no data are available. The Dossier Submitters assume that this share can be generalized across Europe. Release of PFASs from metal/chrome plating processes into the environment is possible via air, wastewater and also via waste (e.g. metal hydroxide sludge, ion exchange resins). As already mentioned, ~20% of the applied surfactants are emitted into the environment. The following emission pathways for the fluorosurfactants were quantified: 50-85% via wastewater, 0.1-24% via waste (e.g. metal hydroxide sludge) and <0.1% via exhaust air (Hauser H., 2020). Waste related emissions are included B.9.18. Short-chain PFASs have a low adsorption potential and are difficult to remove during water treatment processes and will be released in receiving waters. PFASs with higher adsorption potential will be bound to the metal hydroxide sludge and removed via active carbon or ion exchangers (UBA, 2017b; UBA, 2022). Based on information from the PFHxA dossier a current annual release of 6 t (central estimate, range 0.5-11.4 t/y) into the environment was estimated for 6:2 FTS in Europe during chrome plating. This central estimation was confirmed in the 2nd stakeholder consultation during preparation for this restriction dossier: A survey of suppliers of 6:2 FTS for surface treatment identified an emission of about 5 – 8 t/y. Furthermore, one stakeholder noted an emission by wastewater of 25-50% of the applied 6:2 FTS during plastic etching. Another stakeholder described a carryover of <1% in the functional chrome plating process. A further source of PFAS-emissions is de-chroming plating of defective batches and product carriers. Due to adsorption and desorption processes on plastic coated plating parts, PFASs can get to other partial streams of the electroplating plant and emitted via wastewater. Used process solutions from degreasing baths could also release PFASs. The degreasing solutions and their rinse waters are normally not specifically treated in order to remove PFASs (UBA, 2022). Due to the lack of information, a quantification of the PFAS release by dechroming is not possible. Chromate solutions containing mist suppressing agents have a limited usage lifetime and have to be changed regularly. The used solutions are treated as chemical waste, where chromium is 253 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) isolated. The rest is disposed of with the risk of long-term leaching of PFAS to the groundwater or emission of PFAS-contaminated wastewater (NEA, 2017). It is not possible to derive sufficiently representative numbers from the literature to calculate use quantities and emissions for the industry as a whole or for individual metal plating baths. Based on information from a stakeholder consultation the final article does not contain PFAS. In Table B.30 the overview of estimated PFAS emission is given. Table B.30. Estimated annual EEA emission in the metal plating subsector . Metal plating 6:2 FTS (PFHxA related substances) Total PFASs (t/y) (t/y) Low High Low High Total volume 2 57 2 57 Total emission 0.5 11.4 0.5 11.4 B.9.5.2.2. Manufacture of metal products As stated in Annex A it is estimated that approximately 960 t/y of fluoropolymers are used in the production of metal products. Information on emission of PFASs during production and use of metal products is not available. Shares that are used for consumer uses like the production of cookware and bakeware and for industrial applications/coatings respectively are unknown. Information on emission of PFASs during production and use of metal products is not available (see Table B.31). Table B.31. Estimated annual EEA emissions in the manufacturing of metal products. Manufacture of metal products Total PFASs (t/y) Total volume 960 Total emission No data B.9.5.3. Human exposure PFASs are used in metal plating processes as mist suppressant to decrease aerosol emissions (e.g. chromium) and reduce worker exposure. In the manufacturing of metal products PFAS are applied for various purposes (coating, corrosion inhibitor, cleaning, electrical insulation, etc. see Annex A.3.5.). Stakeholders indicate that exposure levels of workers are low. Exposure to workers is supressed with protective clothing, protective goggles, local exhaust, masks, etc. Exposure to consumers is low to negligible. E.g. PFAS are not present in the chromium-plated article as PFAS has no function in the final product. B.9.5.4. Summary Emissions from metal plating and manufacturing of metal products are estimated to be between 0.5 and 11.4 t/y for 2020. This emission is solely linked to mist supressants in the metal plating subuse. (Emissions from manufacture of metal products is unclear, so total emission for the sector is an underestimation). As stated in Annex A.3.5., it is estimated that approximately 960 t/y of fluoropolymers are used in the manufacturing of metal products. Information on the concentration of PFASs per use, annual production tonnages, import volumes for metal plating/production of metal products is hardly available. The same applies for information on annual emissions/release and future 254 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) emissions. 255 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.6. Consumer mixtures B.9.6.1. Introduction The use of PFASs in consumer mixtures is described in Annex A. B.9.6.2. Emissions In the CfE, one company estimated that up to 100% of PFASs used in cleaning products could potentially be emitted to the environment. This is in line with the respective ECHA standard exposure scenarios (ERC 8A and ERC 8D) for widespread use of non-reactive processing aids which are not included into or onto articles (ECHA, 2016a). So, based on data provided before, it must be assumed that 20 t PFASs/y is released into the environment by consumer mixtures in the EEA. For guitars strings alone, the yearly EEA tonnage is estimated at 3 t PFASs/y. The emission is estimated to be 3 t PFASs/y as well. These estimations are listed in Table B.32. Table B.32. Estimated annual EEA emissions in the consumer mixtures sector of total PFASs in 2020 (baseline). Total PFASs (t/y) Total volume 21-30 C leaning agents, polishes, waxes emissions 20 C oated guitar strings emissions ** 3* Total emissions 23 *: Either fluoropolymers or a side-chain fluorinated polymer **: The emission value is the same as the total volume. The reasoning is that, according to stakeholders, the coating is over time released to the hands, which after washing release to the water compartment. B.9.6.3. Human exposure Apart from direct exposure to PFAS due to application of cleaning agents, polishes, and waxes, indirect exposure via ingestion of house dust takes place for consumers. In general, a variety of PFAS can be found in dust in European households and offices, indicating that consumers are exposed to PFAS on a daily basis (Bohlin-Nizzetto et al., 2015; de la Torre et al., 2019; Eriksson and Kärrman, 2015; Harrad et al., 2019; Karásková et al., 2016; Padilla -Sánchez and Haug, 2016; Poothong et al., 2020; Weiss et al., 2021; Winkens et al., 2018) . Cleaning agents, polishes, and waxes are one source of PFAS in dust. Other uses of PFASs, such as uses in textiles and in c onstruction materials or articles containing PFASs, also contribute to the presence of PFASs in house dust. Further information on PFAS in house dust is provided in section B.9.21.4. PFAS concentrations can vary by several orders of magnitude, showcasing the effect of variable furnishing and consumer behaviour. B.9.6.4. Summary Limited information is available on the emission of PFASs from consumer mixtures. Up to 100% may be emitted, the emission estimate is therefore the same as the use estimate, which is 23 t/y. PFASs from consumer mixtures may also gather in house dust, which leads to human exposure. 256 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.7. Cosmetics B.9.7.1. Introduction PFASs in cosmetic products is a research area which came into focus during the last decade. Three different databases were consulted to get an overview of the identity and frequency of occurrence of PFASs in cosmetic products. About 170 unique PFAS ingredients potentially occurring in cosmetic products were identified within the cosmetic ingredient database (CosIng). Of these, 42 were present in products within three European cosmetic databases, among which polytetrafluoroethylene (PTFE; a PFAS polymer) and C9-15 fluoroalcohol phosphate were most frequent. Based on measured PFAS concentrations, the share of products containing PFASs, sales data from Cosmetics Europe, as well as other parameters and assumptions, the total emission of PFASs from cosmetic products after use was estimated for the European Economic Area (EEA). B.9.7.2. Emissions Generally, four parameters have to be considered calculating a chemical’s emission, in this case for PFASs (EPFAS in kg/y) from products (see Equation):     the concentration of a chemical in the products (C PFAS in µg PFAS/g product), the total amount, or tonnage of the products sold per year (A products in t/y), the share of products containing the chemical (f PFAS products), and the fraction of the chemical released from the product into a certain compartment (f release) (e.g. wastewater or solid waste etc.). 𝐸𝑃𝐹𝐴𝑆 [ 𝑘𝑔 µ𝑔 𝑡 ] = 𝐶𝑃𝐹𝐴𝑆 [ ] × 𝐴𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑠 [ ] × 𝑓𝑃𝐹𝐴𝑆 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑠 × 𝑓𝑟𝑒𝑙𝑒𝑎𝑠𝑒 × 10−3 𝑦𝑒𝑎𝑟 𝑔 𝑦𝑒𝑎𝑟 The factor 10-3 in the equation is a conversion factor from g/y to kg/y. The f release part can be neglected (i.e. set equal to one) in order to calculate the total emission or total content of PFASs in the products that can be emitted into the environment as a worst-case scenario. All emission estimates are annual values for the EEA (not including Lichtenstein and Iceland), or correspondingly the EU27 and Norway. The total emissions are assumed to occur after product use only and to exclusively end up in wastewater or to solid waste. No other emissions were considered. The emissions to solid waste in all worst -case scenarios are equal to zero, as the entire emissions are assumed to go into wastewater. The total emissions in those worst -case scenarios are therefore equal to the emissions into wastewater. The concentrations of PFASs in the cosmetics products were derived using different analytical methods due to the complexities to characterize and quantify the large diversity of PFASs occurring in these products. These methods included measuring total fluorine (TF), the extractable organic fluorine (EOF) and individual PFASs (targeted analysis). The EEA emission estimates from cosmetic products were based on Total Fluorine (TF) and Extractable organic fluorine (EOF) concentrations. These were deemed as a more realistic concentration representation of emissions from listed ingredients. The emissions based on TF represents any kind of PFASs (low and high molecular weight PFASs, including polymers, non-polar and polar, as well as ionisable and non-ionisable PFASs), but can also represent inorganic fluorine if present in the product. To avoid the potential for overestimating the PFAS content due to presence of inorganic fluorine, EOF was used to capture a wide range of organofluorines and removes inorganic fluorine. The EOF-based emission calculations are the best estimate for non-polymeric and polar (i.e., soluble in methanol) PFASs that are present in the cosmetic products. Targeted PFAS analysis, on the 257 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) other hand, is highly specific and sensitive, but does not capture PFASs not specifically included in the method. In the tables below fluorine content was converted to PFAS using a conversion factor of 1.7. Emissions are presented in Table B.33 below. Table B.33. Total emissions of PFAS from use of cosmetic products/year , best case and worst scenario. A conversion factor of 1.7 (midpoint of 1.4-2) is used to make the tonnages and emissions based on PFAS content enabling comparison with other uses. Emission (to wastewater*) Total PFASs (kg/y) Low High Volume 28 64 000 Total emission use phase (worst case) 15 64 000 *Emissions to solid waste is described in KEMI (2021b). As Table B.34 demonstrates, skin care is the dominant cosmetic category in terms of use and emissions. For more details about t he different categories, see KEMI (2021b). Table B.34. Estimates for total emissions for the different cosmetic product categories in a best, average- and worst-case scenario. Product Emissions (kg PFASs/y) category TOTAL TOTAL TOTAL best case average case worst case Skin C are 14 13 940 49 300 Toiletries 2 950 2 550 Hair C are 3 1 700 4 590 Perfumes and Fragrances Decorative C osmetics Total emission (rounded) 0 0 0 10 2 040 6 970 29 (part to solid waste) 18 600 (part to solid waste) 64 000 The Skin Care product category contributed the most to the sum (PFCA) emission estimates, followed by Hair Care and Decorative cosmetics. More information on analytical challenges and emissions estimates can be found in the full report (KEMI, 2021b). Emission estimates and other collected data in this study show, while subject to several uncertainties, that cosmetic products contribute to the occurrence of PFAS in the environment, both via wastewater and via solid waste. There may be environmental risks associated with PFAS in cosmetic products (Ahrens and Bundschuh, 2014). PFASs may be released to the environment from cosmetics during product manufacturing use and disposal. PFAS-containing cosmetics that are removed from the skin or body will enter wastewater streams or be directed to solid waste. It is unclear how these substances behave during wastewater treatment. However, wastewater treatment plants (WWTPs) have been identified as a significant source of PFASs releasing either to receiving water (via effluent outfalls), air (via off gassing from settling tanks), and fields (via application of sludge as fertilizer) (Ahrens et al., 2011; Lee et al., 2014; Yeung, 2017). B.9.7.3. Human exposure Food, dust, air, and drinking water are regarded as the major exposure pathways for known PFASs in the general population (Trudel et al., 2008; Vestergren et al., 2008). Dermal exposure to PFASs has so far been considered negligible, most likely due to a lack of data on occurrence 258 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) and dermal permeation of PFAS from cosmetic products (Winkens et al., 2017b). Given the large surface area of the skin (~22 m2; Godin and Touitou (2007)) and the liberal application of some cosmetic products, dermal uptake may represent an important route of PFAS exposure particularly in children. Further investigation is needed. B.9.7.4. Summary Total emission of PFASs estimated in the EEA from cosmetic product use was based on parameters including measured PFAS concentrations (i.e TF, EOF, targeted PFAS analysis), the share of products containing PFASs, and sales data from Cosmetics Europe. Emissions were assumed to occur after product use only and to exclusively occur to wastewater and solid waste. As worst case scenario with emissions going 100% to waste water, the calculated total emiss ions of PFASs from use of cosmetic products is 64 000 kg PFAS/y and equal to the tonnage put on the market. Skin care products are the largest contributor to emissions. Despite limited data, it is thought that dermal exposure to PFASs may represent a relevant uptake route. However further investigation is needed. 259 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.8. Ski wax B.9.8.1. Introduction PFASs commonly used in the production of gliders and other ski wax products can be released into the environment, as the wax containing PFASs will be worn off through abrasion against the snow, causing diffuse spreading of PFASs through the snow, water and sediment. During application, wax can be brushed off the ski, ending up on the ground. This may eventually end up on the snow if not correctly disposed of. Direct human exposure to PFAS can occur when applying ski wax treatments, as the application often includes heating, melting, brushing and sanding of mixtures containing PFASs close to the airways, meaning users can be exposed to high concentrations of PFASs (Nilsson et al., 2010b). B.9.8.2. Emissions For this assessment, a basic source-flow model was developed. The development of this sourceflow approach began with a consideration of the key life-cycle stages and the kinds of emissions that may occur at each life-cycle stage. ERCs were only applied for the manufacturing stage. For other parts of the flow, an analysis for each step in the life-cycle was performed with a basis in what is known from practical experience and stakeholder information. Four basic life-cycle stages are considered based on the likelihood for emissions to occur, or for material to flow through into the next life cycle stage (Figure B.72):     Formulation of the ski wax: This includes consideration of the PFAS substances used within the ski wax. Note, that it was assumed that the life-cycle begins at this stage rather than the manufacture of the PFASs themselves. Storage: For other sectors beyond ski-waxes, storage can potentially be an important life-cycle stage for potential emissions. Therefore, for completeness ‘storage’ has been included within the current source-flow approach. However, during storage of ski waxes it is unlikely that leaks or spillages are to occur, which would directly contribute to environmental emissions. Therefore, emissions from this stage are assumed to be zero. Use: Active use of ski waxes is likely the most important life-cycle stage. To provide a higher level of disaggregation, this stage is split into two sub-components: o Emissions to the environment during the application of ski-wax to skis/snowboards etc. o Further emissions to the environment associated with the continued service life (i.e., during skiing). Waste: The waste cycle includes three key pathways, landfill of wastes and/or end of life product (i.e., any unused final quantities of wax discarded), thermal destruction (incineration), and wastewater treatment plants. In Figure B.72 shown below, the PFAS quantities (material flow) and PFAS emissions are presented for the different life cycle stages. 260 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) All material not lost directly to the environment will enter the waste stage through three main pathways: Waste generated during manufacturing, excess wax discarded during application, and unused ski wax in service life (i.e. because of date expiration). Figure B.72. Material flow diagram for PFASs (all species) in ski-waxes for the European Union. Active use of ski waxes is likely the most important life-cycle stage. To provide a higher level of disaggregation this stage is split into two sub-components based on a basic source-flow model that has been developed to make use of the data from the market analysis and substance identification:   Emissions may occur during the application of ski-wax to skis/snowboards. This is potentially the single biggest point of release in the life-cycle. Based on consultation with industry, the efficiency of applying ski-wax to skis/snowboards for instanceis relatively poor. The average of the values provided suggests that 80% of ski-wax applied is lost during the application process. This can occur both indoors and outdoors over snow. Further feedback suggests a significant proportion of lost material may be recaptured and consigned to waste (e.g. via vacuuming). Therefore, it is assumed that 80% is initially lost, with 40% (half) recaptured and consigned to waste and 40% (half) truly lost to the environment, with an equal split between water and land. The remaining 20% is retained on the skis/snowboard. The application life-cycle stage is estimated to emit 624 kg of PFASs to the environment annually (equivalent to 38% of the total quantity of PFASs used in ski-waxes per year). Secondly, after application there may be further losses during service life (i.e., skiing). It is assumed that, worst case, 100% of the wax applied is lost during this stage through 261 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs)  erosion of the wax with further emissions to air, water, and soil during active skiing. This amounts to 280 kg lost directly to the environment per year. In lieu of better data it is assumed 10% (approximatley 30 kg) of the wax purchased is not used and enters the waste cycle (this is where the ski-wax bars are worn particularly thin, or liquid products are residue amounts in containers). All material not lost directly to the environment during formulation or use will enter the waste stage through three main pathways (i.e., wastes generated during formulation; excess wax discarded during application; and unused ski wax which has expired). Figure B.72 and Table B.35 provide the results of the emission estimates, which highlights the main receiving environments to be equally shared between surface water and soil. This is based on the use of the ERC emission factors. In practice it could be expected that the main receiving environment is snow itself, which in due course (i.e., spring melts) will lead to contamination of surface waters and soil. Figure B.72 also highlights that of the 1 640 kg of PFASs (within skiwax) used per year a total of 945 kg is lost to the environment (58% of the total used), primarily during application, and 728 kg/y enter the waste stage (equivalent to 44% of total PFASs used). Table B.35. Estimated annual EEA emission in the ski wax sector of total PFASs in 2020 (baseline). Compartment Total PFASs (t/y) Low High Total volume 1.6 Emission to soil 0.452 Emission to surface water 0.452 Emission to air 0.041 Total emission 0.945 It should be noted that ski waxes account for only a small share of total PFAS use. As a consequence, ski waxes are most likely not very significant in terms of total environmental PFAS exposure. However, environmental exposure does occur when waxes are released during skiing onto the track where people are skiing. Stakeholder consultation highlighted contrasting opinions with respect to the amount of ski wax released to the environment during skiing, ranging from 1-5% (depending on the contact area between the ski and the snow and on the s now conditions) to 100% of the wax. One interviewee claimed, without providing evidence, that given the very small amounts of PFAS in the waxes it should be considered that the amounts of PFAS found in snow could come from other sources like air, the ski c lothes, lubricants used in lifts and vehicles, or fire-fighting foams 34. However, a scientific study examining the presence of PFASs in soil, earthworms (Eisenia fetida), and bank voles (Myodes glareolus) from a skiing area in Trondheim, Norway, found long chain PFCAs at high concentrations, compared to the reference area with no skiing activities which was dominated instead by short-chained PFCAs (Gronnestad et al., 2019). on an outdoor recreation area with significant cross-country ski activity, found evidence of high levels of both long and short chain PFASs in the snow at the starting line of a ski race, with the longer-chain analytes (C10-C14) predominating. Importantly, the 14 PFASs detected in the snow matched what has been found in ski wax. PFAS levels diminished at around 4 km from the starting line, but soil and groundwater were also found to be contaminated with PFASs. The authors conclude that ski wax use, from which fluorocarbons abrade onto snow during a ski race, was the main source of PFAS contamination at that specific sampling site. The soil was contaminated by the same four dominant PFASs that were present in the snow samples, although at a lower concentration. Groundwater contained only short chain PFAS at low concentrations (Carlson and Tupper, 2020). 34 Interview with FIS. 262 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.8.3. Human exposure Direct human exposure to PFASs can occur when applying ski wax treatments, as the application often includes heating, melting, brushing and sanding of mixtures containing PFAS close to the airways, meaning users can be exposed to high concentrations of PFASs. Prior to application, the sole of the skis is usually cleaned with a liquid non-fluorinated base cleaner and a cloth. To provide context about the potential magnitude of human exposure to PFASs from ski waxes, the magnitude of the use should be considered. About 21 t of PFAS-based waxes are produced per year in the EEA35 (see Annex A section A.3.8.2.). It is expected that a similar quantity of PFAS-based ski waxes is currently used in the EEA, because they are mostly produced and used in the EU (interviewees suggested the EU accounts for some 60% of the global production and 70% of the global market 36). PFAS-based ski wax may contain up to 100% PFAS, although the concentrations depend on the formulation. However as noted previously, the reported concentration data only looks at a subset of PFAS substances for which analytical methods and standards are available and does not quantify total PFAS content or allow differentiation between those substances present in ski wax (including as impurities) or those actively used/added. The following input was received from stakeholders regarding the potential for human exposure to PFASs from ski waxes: No specific information was provided regarding PFAS exposure during production of ski waxes, although producers suggested that the main potential for exposure is during application of the wax to the ski. Prior to application of ski wax to the ski, the sole of the skis is usually cleaned with a liquid non-fluorinated base cleaner and a cloth. The high-end ski wax is then placed on the sole of the ski as a powder, melted with an iron and distributed evenly on the ski sole. Upon cooling the wax becomes solid again and much of the wax is removed from the ski base by scraping, and brushing, leaving a thin layer of wax on the ski sole. This has led to several concerns: ­ ­ ­ ­ ­ The range between melting point and boiling point of the compound is very narrow, so fumes can be released even when the boiling point is not reached. As a result PFAS has been found in the blood of people applying ski wax. It should be noted however that emissions of fumes from the application of fluorine-free alternatives may also have health concerns. A proportion of the wax applied will fall to the floor. Scraping and brushing can lead to formation of dust that could be inhaled. Professional ski technicians use protection equipment to shield themselves from potential exposure, including gas masks, fume hoods, gloves and protective clothing. However, this is not as common for non-professionals. Often, especially with non-professionals, the contained wax is disposed of in general waste or even in the snow/outside, but some ski wax producers and EEA countries (e.g. Norway) have recommendations in place for waste wax to be disposed of by waste handling companies (e.g. through incineration). Another formulation of PFAS-containing ski wax is as a liquid with the fluorinated ingredients as a suspension. Applications of these will result in much lower exposure as they do not require melting with an iron or the same level of scraping and brushing. The waste generated is also considerably less. However, the performanc e of liquid wax during skiing does not match the best powder waxes. Interviews with ski wax manufacturers seemed to suggest that at least professionals applying ski waxes are expected to use the high-end powder wax, but there may be others (notably athletes at lower or amateur level competitions) that use the same. There have been studies 35 36 Interview with Rodewax. Interviews with Rodewax and Swix. 263 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) which document a direct correlation between years exposed as a ski waxer and concentration of several different perfluoroalkyl carboxylate (PFCA) compounds in blood, with one study showing that Swedish wax technicians' median blood level of PFOA is 112 ng/mL compared to 2.5 ng/mL in the general population (Freberg et al., 2010; Nilsson et al., 2010b). Recently, one of the main producers of ski wax, Swix, had a campaign where anyone (at least in Norway) could deliver fluorine-based ski wax from any manufacturer to Swix for safe waste handling. The customer was even awarded a €50 certificate for shopping in the Swix online store37. Starting from the ski wax producers, the potential number of workers through the supply chain and users of PFAS-based ski waxes is estimated in the following.  Producers of ski waxes: According to interviews with some of the main ski wax producers, around 100-200 people are employed by at least 20-25 ski wax producers (many of which are small companies) in the EEA 38. This includes the production of both PFAS-based and fluorine-free ski waxes. Considering that most producers offer both, it is not possible to distinguish workers relating to only PFAS-based ski waxes.  Wholesale, distribution and retail: No information on the number of workers in wholesale, distribution and retail of ski waxes was available. However, these workers are very unlikely to be directly exposed to PFAS-based ski waxes, because the waxes are packed during these supply chain stages. Hence their numbers of employees are less important in the context of this proposed restriction.  Application of waxes to skis:  o The total number of people involved in the waxing of skis with PFAS-based ski waxes is highly uncertain, because at lower or amateur level skiing competitions skiers will likely manage their equipment themselves (or their parents will, in the case of junior competitors). In some cases sports shops offer to wax skis as a service, but it is assumed that a minor part of skis are prepared in this way. o For higher level competitions, the number of professionals involved in applying PFAS based ski waxes can be estimated to be around a few hundred based on interviews with ski wax manufacturers . End-users (skiers): Exposure of both workers and skiiers is negliglible during the usestage. However, from each skier a small amount of wax is released to the environment during use. PFAS-based ski waxes are mostly used during competitions. Competition skiers are then estimated to be in the order of 7 million at all levels in the EEA (tourist skiers hardly use PFAS wax). However, such estimate is based on uncertain assumptions so it should be considered as a rough indication of the potential order of magnitude only. B.9.8.4. Summary PFAS emissions from ski wax were calculated for the manufacturing and use stages (during application and emission during service life). Total emissions were estimated to be 0.945 t/y for 2020. Around 58% of the the total quantity of PFASs used in the manufacture of ski waxes are emitted into the environment, while 44% goes to waste. Direct human exposure to PFAS can occur when applying ski wax treatments. A direct correlation between years exposed as a ski waxer and concentration of several different perfluoroalkyl carboxylate (PFCA) compounds in blood have been reported in literature. It is not clear how many people are actually involved in the waxing of skis with PFAS-based ski waxes. For higher level competitions this is estimated to be a few hundreds, while a higher number of active amateurs are usually waxing their own skis, with varying level of protective equipment. 37 38 https://www.swixsport.com/no/kampanjer/fluorfri/, date of access: 2022-12-14. Interviews with Rodewax and Swix. 264 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.9. Applications of fluorinated gases B.9.9.1. Introduction The use of fluorinated gases in the HVACR sector is described in Annex A. B.9.9.2. Emissions Data on emissions of fluorinated gases from different applications have primarily been collected from the GHG Inventory data, estimated on the basis of UN methodology (EEA, 2022). It is to be noted that the GHG Inventory data include HFCs and PFCs, while HFOs, HFEs, fluoroketones etc. are not considered. The same methodology was used for use volume estimates and additional details on the information sources may be found in Section A.3.9.2. Based on market volumes, HFOs are increasing and currently constitute 24% of the market of fluorinated gases, the other subgroups are expected to be of limited volumes. Emissions can be assumed to follow a similar trend. Estimates for HFO emissions have been derived by combining market volumes from the F-gas report and implied emission factors (IEF) derived from the GHG Inventory based on data for market volumes and emissions for HFCs in similar applications. Stock emissions are to be understood as emissions from products and equipment currently in service life. Appendix IV shows the estimated total emissions of fluorinated gases from the different subapplications. The estimates are based on the GHG Inventory data and are limited to the substances reported there (HFCs and PFCs, which constitute the major part of the fluorinated gas uses and emissions). However, they do not include volumes of e.g. HFOs which is an emerging fluorinated gas subgroup. All emissions are to air. Emissions are disaggregated at the subapplication level and not on the substance level. Appendix V compares the activity data and emission estimates for fluorinated gases in 2018. Emissions from manufacturing, as well as emissions from the use stage, are included. For comparative purposes, manufacturing and stock emissions as a proportion of the total emissions have been calculated and presented. These figures are calculated as follows: Tonnes of fluorinated gas emitted (single application)/Total tonnes fluorinated gas emitted (all applications) x 100. From the GHG Inventory data it is apparent that the emissions during the manufacturing of products and equipment is generally low, between 0 – 3% of the fluorinated gases used for all activities other than foam blowing. By contrast emissions from stocks are significantly higher as would generally be expected and are in the range 0 – 13% for all activities other than aerosol use. Commercial and industrial refrigeration, mobile and stationary air conditioning accounting for 82% of the total emissions from stocks. Figure B.73 below shows the emissions of HFCs/PFCs from the different applications with basis in the GHG Inventory data. 265 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.73. Emissions of fluorinated gases from stocks 2018. Source EEA (2022). For open cell foams, foam blowing agents are almost entirely released at initial use and therefore it is expected that after the initial use there would be very low emissions from stoc ks since there are very little if any left in the stocks. With regards to closed cell foam stakeholder discussions in this assessment indicate generally fluorinated gases remain locked up in the foam until decommissioning and losses from stocks would be of the order of 1% per year. In total, the emissions of fluorinated gases in 2018 (HFOs excluded) from the different uses were as follows: Manufacturing of products and equipment: 1 696 t/y Stocks (i.e., service-life): 38 806 t/y For comparison, Öko-Recherche developed an estimate on behalf of the German Environment Agency for the total fluorinated gas emissions (HFCs and HFOs) at ca. 60 000 t/y for 2020 (UBA, 2021c). The estimate was based on modelling with the AnaFgas model which projects the demand and emissions of fluorinated greenhouse gases in the EU in various scenarios. The model is based on market data and estimates of the annual sales of equipment containing these substances and the number of substances needed to manufacture and/or maintain appliances within the EU. Emissions from stationary air conditioning was estimated at 12 193 t/y in 2020 by UBA, which is higher than 7 458 t/y for stationary air conditioning and heat pumps that is estimated in the present assessment based on the GHG Invent ory data (HFCs only, 2018). Industry stakeholders have underlined the importance of hydrofluoroolefins (HFOs) and fluoroketone (FK) alternatives during the development of this dossier. Primarily this is because in several applications they can substitute t he function provided by other fluorinated gases alone or in blends, whilst at the same time having significantly lower global warming potential (GWP). The GHG Inventory data contains emission estimates, but does not include HFOs and FKs. On the other hand, the F-gas report includes HFOs (no data on FKs), but emissions are not estimated. In the market section it was estimated that the HFOs/HCFOs currently constitutes 24% of the market of fluorinated gases. The majority of HFOs/HCFOs currently being used commercially as a single substance (rather than a blend) are in mobile air conditioning (MAC) systems for passenger cars and in light goods vehicles and commercial refrigeration, commercial air conditioning, heat pumps and process cooling. Figure B.74 below shows the projected demand and emissions of fluorinated gases (HFCs and 266 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) HFOs) for various applications modelled with the AnaFgas model and as presented by ÖkoRecherche (UBA, 2021c). Figure B.74. Quantity of projected demand and emissions of sum HFCs/HCFCs and HFOs/HCFOs in Europe (EU-28) in metric kilotonnes (kt) in the years 2000 to 2050 in 10-year increments by sector, from UBA (2021c). B.9.9.2.1. Estimated emissions from mobile air conditioning (MAC) Emissions from mobile air conditioning have been looked at in more detail as this is a sector with large uses and emissions. A key change in the sector is the transition from HFC-134a as the main refrigerant to HFO-1234yf which is accompanied by a large reduction in climate impact as the GWP is much lower for the latter (GWP 1 430 vs. 4). As HFOs are not reported in the GHG Inventory, emissions from MAC are largely underestimated if HFOs (i.e. HFO-1234yf) are not taken into consideration. The Dossier Submitters estimated overall emissions of fluorinated gases from MAC with basis in GHG Inventory data uses the annual emissions of HFCs and PFCs from MAC technical stocks as a starting point: 11 648 t, see Table B.36. EC Directive 2006/40/EC on MAC required all new passenger cars (and light goods vehicles) from 1 January 2017 to be filled with a refrigerant with a global warming potential (GWP) no more than 150. As a result, almost 100% of new cars after this date use HFO-1234yf refrigerant instead of HFC-134a (EFCTC, 2022). However, the requirements in the MAC Directive were introduced gradually, so the change from HFC to HFOs had started even before 2017. Data for 2017 – 2018 demonstrates that approximately 30 million new passenger cars are introduced in EU-28 each year, which equates to approximately 11% of the total number of 267 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) passenger cars of 271.5 million39. In 2018, due to the MAC Directive, one year of vehicles with HFO-refrigerants had been introduced on the market (= 11%), in addition to the fraction that had already entered the market due to the gradual phase-in of the MAC Directive requirements, assumed at 9%. Hence, a total of 11%+9%=20% of the vehicles in 2018 are estimated to contain HFO refrigerants. While 20% of vehicles in the EU in 2018 used HFO-refrigerants, the remaining 80% used HFCrefrigerants, corresponding to emissions of 11 648 t according to the GHG Inventory and Table B.36 below. Assuming a similar emission rate, HFO emissions in 2018 would amount to 2 912 t. In total, fluorinated gas emissions from MAC in 2018 would then be 11 648 t (HFCs)+2 912 t (HFOs)=14 560 t (fluorinated gases). The 30 million new passenger cars (= to 11% of the total number) put on the market each year, all with HFO-refrigerants, would then correspond to 1 602 t of HFO emissions with basis in 2018 numbers, see Table B.36. Table B.36. Estimates of total emissions of fluorinated gases from MAC (2018), in tonnes per year. Fluorinated gases HFCs HFOs total (t/y) (t/y) (t/y) Stocks total volume Stocks total emission New use total volume New use total emission 493 173 11 648 2 912 14 560 30 671 0 1 602 1 602 The fraction of cars with HFO-refrigerant in the total technical stock of fluorinated gases in MAC will gradually increase with time until HFC refrigerants have been phased out in essentially all new vehicles. Figure B.75 is copied from the EEA F-gas report in 2021 and provides information on the amounts of HFCs and HFOs imported in mobile air condition equipment since 2014. The figure below confirms that HFOs were gradually phased in before 2017. However, as explained previously, it is likely that the import of HFOs in vehicles is under-reported due to the high reporting threshold. 39 https://www.acea.be/statistics/tag/category/key-figures, date of access: 2022-05-19. 268 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.75. EU imports of fluorinated-gases within air-conditioning equipment for vehicles EU-27 + UK, from EEA (2021). An alternative approach to estimate emissions of fluorinated gases from MAC starts with the number of vehicles in the EU/EEA (243 million passenger cars, 34 million trucks and vans and 690.000 buses)39 and the amount of fluorinated gases used in AC in each vehicle category (0.6 kg in cars, 1 kg in trucks/vans and 6 kg in buses). Furthermore, a leakage rate of 1.67-5%40 is assumed. A total fluorinated gas emission from MAC could then be estimated at ca. 3 073 t/y – 9 200 t/y. This estimate is somewhat lower compared to the estimate derived from the GHG Inventory. Noting that in the GHG Inventory, an emission factor of 9. 94 has been used, the difference is not dramatic. For comparison, the estimate developed by Öko-Recherche for total emissions of fluorinated gases from MAC for 2020 was 29 466 t based on the AnaFgas model (UBA, 2021c). A 15% overall increase in emissions was estimated from 2020 to 2050. The estimate of 14 560 t total emissions of fluorinated gases from MAC with a basis in GHG Inventory data is taken as the main estimate of fluorinated gases from mobile air conditioning in this assessment. Two alternative approaches to the same estimate were explored for comparison, one giving a lower and the other giving a higher emission volume from this sector. The estimated total emissions of fluorinated gases from the different sub-applications are presented inTable B.37. 40 1.67% is based on stakeholder information. 5% represents ERC 9b Widespread use of functional fluid (outdoor). 269 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.37. Estimated total emissions of HFCs and PFCs from the different sub-applications (in t/y). Total fluorinated gases [t/y] % HVACR emission from total all fluorinated gas applications (separate % for manufactured products and stocks) High Total, all applications Manufactured Products 1 696 100 Stocks 38 806 100 C ommercial refrigeration Manufactured products 121 7 Stocks 9 426 24 Domestic refrigeration Manufactured products 1 0 Stocks 16 0 Industrial refrigeration Manufactured products 77 5 Stocks 3 603 9 Transport refrigeration Manufactured products 29 2 Stocks 1 312 3 Mobile air conditioning Manufactured products 78 5 Stocks 11 648 30 Stationary air conditioning Manufactured products 47 3 Stocks 7 411 19 Foam Blowing Agent (C losed cell) Manufactured Products 1 272 75 Stocks 2 914 8 Foam Blowing Agent (Open cell) Manufactured Products 45 3 Stocks 1 029 3 Fire suppressing Agents Manufactured Products 1 0 Stocks 702 2 Propellants (nonMDI) Manufactured Products 2 0 Stocks 699 2 Manufactured Products No data No data Solvents C over gas Other Stocks 11 0 Manufactured Products 23 1 Stocks No data No data Manufactured Products 0 0 Stocks 35 0 B.9.9.3. Human exposure Worker exposures to fluorinated gases in plants manufacturing equipment or putting the gases into a product should be small, given the need to use closed systems (mentioned by many respondents to the CfE). Outside of the manufacturing plant, however, there may be greater exposure of workers, for example during service and at sites reclaiming refrigeration equipment at the end of its service life and in case of accidental emissions. 270 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.9.4. Summary Emissions of HFCs and PFCs from HVACR and other applications of fluo rinated gases are estimated to be 1 696 t/y for manufactured products and 38 806 t/y for products and equipment in service. Emissions from the sector were estimated using GHG Inventory data and data from the F-gas report (i.e. data reported under the climate convention and F-gas regulation). HFOs represent an emerging subclass of fluorinated refrigerants that have increased from 6% to 24% of total volumes of fluorinated gases from 2016 to 2019. In mobile air conditioning (MAC) HFOs represent a considerable fraction of the refrigerants, and estimates based solely on HFC/PFC data will be an underestimate. The overall emissions of fluorinated gases in MAC was estimated at 14 560 t/y for 2018 (11 648 t/y HFC+2 912 t/y HFO), and the HFO/HFC ratio is gradually increasing as HFO-1234yf is introduced as a substitute for HFCs with higher GWP. Worker exposures to fluorinated gases in manufacturing sites is expected to be limited (e.g. due to closed systems), however for workers outside the manufacturing plant exposure is expected to be higher (e.g. during service and at sites reclaiming refrigeration equipment). 271 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.10. Medical devices B.9.10.1. Introduction Medical devices are regulated under EU Regulation 2017/745 and are, generally speaking, instruments, materials, reagents or other articles intended for diagnosis, prevention, monitoring or treatment of diseases in human beings (see for the full definition the EU Regulation). The use of PFASs in medical devices is described in Annex A. B.9.10.2. Emission Production of medical devices requires a high degree of cleanliness, purity, chemical stability and thermal resistance. PFAS are therefore already used in medical device production processes. Emissions from these processes are not taken into account here as information is lacking. In Table B.38, an overview of PFAS volumes and estimated emissions is given for the medical sector. Fluoropolymers in medical applications generally have a low emission profile during use. The highest emissions are from the MDI propellant. Their arrowheads are mainly TFA and/or other PFAAs. 1,1-difluoroethane (HFC-152a) is an exemption: this substance is not in the chemical scope of the proposed PFASs restriction. 272 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.38. Estimated annual EEA emissions in the medical devices sector of PFAA and PFAA precursors, fluorinated gases, polymeric PFASs and total PFASs in 2020 (baseline). Total volume for all medical use categories PFAAs and PFAA precursors (t/y) Fluorinated gases (t/y) Fluoro polymers (t/y) Total PFASs (t/y) low high low high low high low high 1 279 3 495 20 160 46 000 3 233 12 032 24 672 61 527 MDI: 144 MDI: 5 400 Other: 1 000 Other: 2 000 TOTAL: 1 144 TOTAL: 7 400 Emission 128 350 7 870 32 120 3 932 Median: 5 901 273 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Emission factors were based on consultant expertise, not on ERCs for this use:    90% for (MDI) propellants and 5% for non propellant use of fluorinated gases; 10% for non-polymer PFAS in industrial processes (heat transfer agents, specialty cleaners, and end-products) and 1% for PFAS polymers (by wear and tear). This is lower thant ERC 11a but for wear and tear indoors in a medical setting 1% seems more realistic. Regarding MDIs, a distinction should be made between MDI used by patients and the production stage. The emission factor of propellants by MDI use is high (70-100%) whereas the emission factor of propellants during production is much lower, probably <10%. Production phase emissions were not taken into account. An uncertainty factor of at least 2 is applicable. Not all MDIs are (completely) used before they are disposed of. According to a stakeholder, about 30% of fluorinated gases remain in the MDI after proper use. A stakeholder mentioned that unused propellant is captured and sent back to the original supplier of the propellant. The amount recycled at one EU site ranges from 30 to 40 t and the amount recycled at the other EU site ranges from 100 to 150 t. Another stakeholder mentioned that separate collection of MDIs was highly unsuccessful. If the return, recovery and/or destruction rates increase, the specific emission value might be reduced in future. Polymeric PFASs emission in the use stage are likely very low, therefore these mostly reach the waste stage, which is described in section B.9.18. B.9.10.3. Human exposure In general, human exposure to emissions coming from medical devices is considered to be low. Some remarks made by stakeholders regarding exposure:     It is unlikely that polymeric PFASs will be released from the devices during their lifetime because PFAS polymers tear and wear is limited during (human) lifetime. During normal use PTFE articles will not pose harm to human health and are not expected to deteriorate during their use. It is only during the waste stage where PFAS could be released to the environment. Regarding applications of PFAS fluids in medical devices, including solvent cleaning, carrier solvent and heat transfer, these products can be emitted to the air during use. When used in industrial applications like lubricant and coating deposition and precision cleaning, workers are typically exposed to solvent vapor during their shifts, which is why determining acceptable exposure limits is a critical factor in selecting process solvents. For MDI use the exposure to fluorinated gases is high. B.9.10.4. Summary PFASs emission from medical devices was calculated using emission factors; emission from the use phase of medical devices only is estimated from 4 000 to 8 000 t/y (rounded numbers). The emission profile of medical devices is generally low, most notable is emission of fluorinated gases during the use stage of MDIs. MDIs causes significant human exposure to PFASs. Emissions from the manufacturing of medical devices have not been taken into account as information is lacking. 274 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.11. Transport B.9.11.1. Introduction The use of PFASs in the transport sector is described in Annex A.3.11. B.9.11.2. Emissions Information that can be used to estimate emissions in the transport sector was only available for fluorinated gases and polymeric PFASs. The emission of fluorinated gases is described in B.9.9.2. B.9.11.2.1. Fluoropolymers (e.g. applied in applications) coatings and finishes or sealing Since information was only available on volumes used in the automotive sector (Annex A.3.11.) no estimates were made for the other sectors. To estimate emissions of polymeric PFASs, it is assumed, that they can be released from transportation vehicles through wear processes, e.g. abrasion or erosion. The stronger the various effecting factors that are described below are, the higher the wear rates. It is likely that indoors (e.g. in the passenger compartment) wear rates are lower than those outdoor and in parts that are subject to harsh conditions (e.g. in the engine bay area) wear rates are higher. However, because of a lack of data it is not possible to make a precise distinction between indoors and outdoors nor between different conditions. Therefore, it is assumed that emissions of FPs in automotive applications occur indoors and outdoors, eac h with a share of 50%. Emissions were estimated based on ERC and the volume of polymeric PFASs in road transportation vehicles. In Annex A.3.11.2.4., the lower estimate for polymeric PFASs in cars is 0.35 kg (Améduri, 2020) and the higher estimate is 0.8 kg (stakeholder information). These estimates, together with information on the number of road transportation vehicles registered in the EU in 2019 (277 759 682 sum of all personal vehicles, trucks and busses) (ACEA, 2020), were used to estimate the total amount of polymeric PFASs emitted from road transportation vehicles. The average service life of road transportation vehicles is assumed to be 11.95 years (ACEA, 2021). Details on calculations of emissions are included in Appendix B.9.11. Total emission estimates are provided in Table B.39. B.9.11.2.2. HVACR systems in transportation vehicles and HVACR-systems for transport refrigeration Estimations on the emissions of fluorinated gases from HVACR systems in transportation vehicles and transport refrigeration are presented in section B.9.9.2. Stakeholders did not provide information on emissions of fluorinated gases from from HVACRsystems in ships trains and aircrafts. 275 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.39. Estimated annual EEA emissions in transportation products and articles of PFAA and PFAA precursors (+ other PFASs), fluorinated gases, polymeric PFASs and total PFASs in 2020 (baseline) (t/y). Total fluorinated Total polymeric PFASs Total PFASs (t/y) gases (t/y) (t/y) Total volume 13 232 Total emission Road transportation vehicles HVAC R-systems in transportation See section B.9.9.2 Low High 6 410 14 652 19 642 27 884 269 609 269 + amount reported under section B.9.9.2 609 + amount reported under section B.9.9.2 269 609 269 609 See section B.9.9.2 See section B.9.9.2 See section B.9.9.2 Low High It is important to note that the emissions presented in Table B.39 are very likely a significant underestimation as tonnages (and therefore also emissions) from PFAS use in ships, trains and aircrafts are missing due to a lack of information. B.9.11.3. Human exposure Polymeric PFASs which make up the greatest share of PFASs used in the transportation sector are considered to be stable up to 300 °C. Therefore, exposure of workers or consumers is assumed to be low during the service life of transportation vehicles. During production, or end of life, other conditions can apply that may lead to an exposure. B.9.11.4. Summary Emissions from the transporation sector were calculated using ERC. Total emissions of polymeric PFASs amount to between 269-609 t/y. Most emissions likely occur in the production stage or during end of life. Only qualitative information is available on exposure of workers and consumers. Stakeholders believe exposure is limited because of stability of PFASs up to 300 °C. 276 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.12. Electronics and semiconductors B.9.12.1. Introduction The use of PFASs in the Electronics and Semiconductors industries is described in Annex A. 3.12. B.9.12.2. Emissions Emissions were calculated for the different life-cycle stages. Because of uncertainties in the estimated volumes and many unspecified actual and patented uses, an emission calculation per substance is not possible. Instead, an approach was used with default emission factors corresponding to different emissions scenarios described by ECHA. Default release factors from the REACH environmental release categories (ERC) were used in the emission characterisation. Estimates of product service-life emission to the environment were derived using the latest version of the standard REACH methodology (ECHA, 2016a). Emissions were calculated for combined values of soil, water and air. The following life cycle stages are considered in the assessment:   Production The manufacturers included in this section are mainly companies that specialise in modifying or combining raw materials to create new products. Use In this stage all PFASs that are produced in the production stage are assumed to enter the use stage, except for the solvents (non-ionic non-polymeric PFASs). There are regulations for the collection and handling of e-Waste e.g. the WEEE directive and part of the used electronic devices is incinerated or exported to low-income countries. Waste stage emission are included in section 0. For emission c alculations, a distinction is made between polymeric PFASs and non-polymeric PFASs. Polymeric PFASs include fluoropolymers and perfluoropolyethers, and non-polymeric PFASs include side-chain fluorinated polymers and ionic and non-ionic PFASs. This is done because the various PFASs groups differ in their uses and emission pathways. According to stakeholders, the following uses are most common: 1. Fluoropolymers are generally used in or on articles and form a solid matrix. 2. Perfluoropolyethers are usually applied as oils or waxes, meant to treat surfaces, but they can also contain one or more functional groups which make them able to bond to substrates or be useful as crosslinkers. 3. Side-chain fluorinated polymers are generally used in or on articles and form a solid matrix. 4. Non-polymeric PFASs are smaller molecules and are mostly used as processing aids, solvents, cleaning liquids, surfactants, coatings and heat transfer fluids. Environmental release categories (ERC), that describe the use of PFASs most closely including release factors were assigned to the production and use stages. A detailed description on emission calculations using volumes in Annex A.3.12., ERC, release factors and calculation factors, as well as a justification for the choices made, is included in Appendix B.9.12. The emissions are summarised in Table B.40. The emissions in Table B.40 are rounded values as it is estimates based on stakeholder information. Therefore, values may slightly differ when calculating emissions from the use stage and production stage compared to the total emissions. 277 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.40. Estimated annual EEA emissions in electronics and semiconductor sectors of PFAA and PFAA precursors, fluorinated gases, polymeric PFASs and total PFASs in 2020 (baseline ). C2- C3 PFA S substances (t/y) PFA A ≥C4 (t/y) Side-chain fluorinated polymers (t/y) PFA As and PFA A precursors Fluorinated gases (t/y) Fluoro polymers (t/y) PFPE (t/y) Polymeric PFA Ss (t/y) Total PFA Ss (t/y) low high low high low high low high low high low high low high low high low high Total volum e 159 221 671 1 315 11 13 841 1 549 140 140 1 551 4 063 9 552 1 560 4 615 2 541 6 304 Production stage 8 11 335 658 5 7 348 676 7 7 5 12 4 276 9 288 364 971 0 0 0.3 1 0.005 0.007 0.3 1 0 0 2 4 0.004 0.3 2 4 2 5 8 11 335 659 5 7 348 677 7 7 7 16 4 276 11 292 366 976 Use stage Total emission 278 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.12.3. Human exposure Stakeholders indicate that exposure levels of workers are low. Instructions on safe handling and minimisation of release are included in the technical data sheet and safety data sheets (SDS). Industrially, frequent handling of parts may cause skin exposure. According to industry, the use of filters, personal protective equipment, clean rooms and closed systems, as well as training, leakage tests, and frequent servicing, contribute to minimising exposure of workers. Some stakeholders stated that their applications of PTFE cause no emissions, but rather serve to protect humans and environment from dangerous media. An example is the protection against acids in the semiconductor industry. PFAS-based solvents are used to clean electronic components. This is commonly done in special sealed equipment where solvents can be recovered and reused. According to a stakeholder, exposure during the application of solvent is expected to be low, as the safety margin for PFASs solvens is higher than in non PFASs solvents. The stakeholder stated that it is important to establish acceptable exposure limits and set a standard for personal protection. Stakeholders stated that in parts used in electronics dev ices, like gaskets, tubes and membranes, the fluoropolymers and elastomers are utilised in a bound state (and stable at up to 300 °C), and therefore human exposure is hardly possible. According to industry, microelectromechanical systems device manufacturing is highly automated and takes place in a closed system in a clean room environment. It is designed in such a way that there is minimal potential for human contact. For some of the processes exhaust hoods are used, and equipment in the manufacturing process has internal cold trap filters that remove more than 99% from the exhaust stream. The semiconductor industry has, according to industry, implemented stringent risk management measures and safety practices to prevent the release of chemicals during all stages of the manufacturing process. According to stakeholders there is no release to the workplace environment during normal production due to closed system manufacturing equipment, which is installed in a cleanroom environment. Automated chemical delivery systems create a barrier between workers and the manufacturing process and protect against chemical and physical hazards. The stakeholders also stated that emission of gaseous PFASs in the exhaust gas of the manufacturing equipment can be almost completely eliminated by incineration. A lot of electronics are produced outside EEA. Certain regulations are in place to manage and minimise emissions/exposure, although the Industrial Emission Directive seems to be not applicalbe for this Industry. A stakeholder commented that relevant regulations include the General Product Safety Directive 2001/95/EC and the F -Gas Regulation (EU) No 517/2014. Additionally, standards such as ISO 9001, ISO 14001, ISO 45001 are used to minimize release and ensure safety. Lastly, it is important to notice, that multiple stakeholders stated that there is no release from multiple uses, but release may differ based on the use and knowledge on the material. A stakeholder mentioned that their supplier stated that no protective measures need to be taken when coated metal products and solid components made of fluoropolymers or fluoroelastomers are used as intended. B.9.12.4. Summary Emissions from the electronics and semiconductor sectors in the production and use stages were calculated using ERC. Total emissions of non-polymeric PFASs (including side-chain polymers) amount to between 300 and 700 t/y (rounded numbers), of polymeric PFASs to between 10 and 300 t/y (rounded numbers) and of fluorinated gases to 7 t/y. Almost all emissions occur in the production stage. Only qualitative information is available on exposure of workers and consumers. Stakeholders generally believe exposure is limited because of the use of closed systems, filters, PPE and training of workers. 279 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.13. Energy B.9.13.1. Introduction The use of PFASs in the energy industry is described in Annex A.3.13. B.9.13.2. Emissions Emissions were calculated for different stages in the life cycle of PFASs. Because uncertainties in the estimated volumes and many unspecified actual and patented uses, an accurate emission calculation per substance is not possible. Instead, an approach was used with default emission factors corresponding to different emissions scenarios described by REACH (ECHA, 2016a). Emissions are estimated based on combined values for soil, water and air. Each stage in the life cycle of a product has a different likelihood of emissions. The following life-stages are considered in the assessment:   Production The industry that produces PFASs from raw materials is included in section B.9.2. The manufacturers included in this section are mainly companies that specialise in modifying or combining raw materials to create new products. Use (service life) Most uses of PFASs in the energy sector are industrial or professional There are Regulations for the collection and handling of e-waste e.g. the WEEE directive and the Battery directive. Waste stage emission are included in section B.9.18. According to responses from stakeholders, the following applications are most common: 1. Fluoropolymers are generally used in or on articles and form a solid matrix. 2. Perfluoropolyethers are usually applied as oils or waxes meant to treat surfaces but they can also contain one or more functional groups which enable them to bond to substrates or be useful as crosslinkers 3. Side-chain fluorinated polymers are generally used in or on articles and form a solid matrix 4. The non-polymeric PFASs are smaller molecules and are used for applications such as cleaning agents for precision cleaning and insulation gas in switchgear in the energy sector. Environmental release categories (ERC), that describe the use of PFASs most closely including release factors were assigned to the production and use stages. A detailed description on emission calculations using volumes in Annex A.3.13., ERC, release factors and calculation factors, as well as a justification for the choices made, is included in Appendix B.9.13.. The emissions are summarised in Table B.41. The emissions in Table B.41 are rounded values as it are estimates based on stakeholder information. Therefore, values may slightly differ when calculating emissions from the use stage and production stage compared with the total emissions. 280 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.41. Estimated annual EEA emissions in the energy sector of PFAA and PFAA precursors, fluorinated gases, polymeric PFASs and total PFASs in 2020 (baseline). Side-chain PFAAs and C2- C3 PFAS Fluoropolyme PFPE Polymeric Total PFASs PFAA ≥C4 fluorinated PFAA substances rs (t/y) (t/y) PFASs (t/y) (t/y) (t/y) polymers precursors (t/y) (t/y) (t/y) low high low high low high low high low high low high low high low high Total volume 233 233 20 20 40 41 293 294 2 590 2 917 2 3 2 592 2 920 2 884 3 214 Production stage 12 12 10 10 20 21 42 42 8 9 1 2 9 10 50 53 Use stage 0.00 0.00 0.01 0.01 0.02 0.02 0.03 0.03 2 3 0 0 3 3 3 3 Total emission 12 12 10 10 20 21 42 42 10 12 1 2 12 13 53 56 281 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.13.3. Human exposure Stakeholders indicated that exposure levels of workers in the energy sector are low. Instructions on safe handling and minimisation of release are included in the technical data sheet and safety data sheets (SDS). Industrially, frequent handling of parts may cause skin exposure. In multiple cases, fluoropolymers and elastomers are utilised in a bound state (and stable at up to 300 °C) and do therefore barely contribute to human exposure. The use of filters, personal protective equipment, clean rooms and closed systems, as well as training, leakage tests , and frequent servicing, contribute to minimising exposure of workers. Stakeholders stated that Proton Exchange Membrane (PEM) fuel cells and PEM electrolyser are usually operated at low temperature of 40-90 oC. Studies into these fuel cells are conducted at 200-400 oC and prove that e.g. Nafion ionomers will not degrade at operating temperatures. Additionally, stakeholders mentioned that they have no knowledge on human or environmental exposure during operation of gas diffusion layers in fuel cells or electrolysers. According to stakeholders, human exposure to PFASs in batteries can only occur during the manufacturing and recycling stages. No release of or exposure to PFASs is expected during use. One stakeholder commented that due to the presence of (other) hazardous substances (metal salts and oxides used for electrodes) in batteries, safety measures are in place, both for the environment as well as for workers. Since waste batteries are recycled in accordance with the Battery Directive 2006/66/EC PFASs are, according to stakeholders, not released into the environment. For medium voltage switchgear, a stakeholder mentioned that PFAS insulation gases are always used in closed loop systems. Leakage rates are small (<0.1% annually, <5% over the full device lifetime; according to the standard IEC 62271-200) because of the use of leak-tight compartments (Dansk Standard (DS), 2021). The stakeholder further stated that human exposure is at 0.002 x the exposure limits when assuming a leakage rate of 1% per year (IEC 62271-4 ed.2) (Dansk Standard (DS), 2022). In high-voltage equipment, gas-insulated switchgear needs to be extremely gas-tight. In DIN EN IEC 62271-203 (and linked standards), requirements for emission reducing measures and designs for Switchgear technologies is described. According to stakeholders, the use of such standards, will ensure minimal human exposure to polymeric PFASs or lubricants containing PFASs, under normal conditions. B.9.13.4. Summary Emissions from the energy sector in the production and use stages were calculated using ERC. Total emissions of non-polymeric PFASs (including side-chain polymers) amount 42 t/y (rounded number) and total emissions of polymeric PFASs amount to 13 t/y (rounded number). Almost all emissions occur in the production stage. Only qualitative information is available on exposure of workers and consumers. St akeholders generally believe exposure is limited because of the use of closed systems, filters, PPE and training of workers. 282 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.14. Construction products B.9.14.1. Introduction The use of PFASs in building materials/construction products is described in Annex A.3.14. For emission estimations, the PFASs used in construction products are divided into polymeric PFASs and non-polymeric PFASs The polymeric PFASs that are included in the emissions estimate are:      PTFE (polytetrafluoroethylene) ETFE (ethylene tetrafluoroethylene) PVDF (polyvinylidene fluoride) Other fluoropolymer substances (including the fluoroelastomers FKM, FFKM and THV) Side chain fluorinated polymers The first three substances (fluoropolymers) account for 97% of the reported total usage of polymeric PFASs in the construction sector while the latter two covers 3% of the usage. It should be noted that for the emission estimation, side -chain fluorinated polymers are considered to be PFAA precursors. B.9.14.2. Emissions The PFASs and uses identified and included in Annex A.3.14. broadly cover one of the following three major categories: 1. Articles. In many cases the articles produced (such as architectural membranes) are 100% fluoropolymer manufactured through extrusion as a hot or cold process. Extruded fluoropolymer can also be manipulated or woven to form meshes, tapes, or other thin film technologies for use across a very wide range of applications. 2. Processing aids. Non-polymeric PFASs are used as processing aid in the manufacturing of articles. PFASs are used as a processing aid (PA) in the production of these products but are not retained (intentionally) within the final products. Polymer processing aids (PPAs), used to produce thermoplastics, thermosetting plastics and elastomers, are included in the resin and therefore becomes a part of the final article. 3. Construction mixtures. This category covers blended liquid mixtures which are either 100% PFASs or more commonly where PFASs is added to tailor the physical properties of the mixture or the final article to which it is added. A good example here are paints, where PFASs can be added as a surfactant to aid spreading and levelling within specialised applications. Four basic life-cycle stages are considered in which emissions may occur, or from which PFASs flow through into the next life cycle stage: 1. Formulation (articles and commercial construction mixtures only). This lifecycle stage covers the first step in the production of articles and commercial construction mixtures. Two different types are considered:  Type 1 which covers the extrusion and/or manipulation of fluoropolymers to form articles like meshes, membranes, tapes, thin films, or other solid forms for use in construction  Type 2 which covers the blending/mixing and formulation of commercial construction mixtures, and PFASs are used to achieve desired technical functions and physical properties. 2. Processing aids. This stage covers the use of PFASs based processing aids in the manufacture of articles. The processing aid is expected to be expended and is not included (intentionally) in the final article. Stakeholders state that capture systems are 283 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) in place to retain and re-use the processing aid as many times as possible and will after the final use be thermally destructed. Those applications in which PPA are used and PFASs are intended to still be present in the final article, are covered in the formulation step under point 1. 3. Application (commercial construction mixtures only). This stage covers the application of commercial construction mixtures. These applications can take place both indoors and outdoors. The mechanism of application for commercial construction mixtures can vary widely (e.g. by spray, roller, cloth, brush, ‘dipping’, etc.). 4. Use stage (articles and commercial construction mixtures only). This stage covers the use service life. Again, these uses can be split between outdoor and indoor applications. Primary emission for outdoor applications is likely to be through a combination of weathering and abrasion depending on the specific application. This makes applying emission factors at a high-level (i.e., all indoor articles) challenging. For articles, this may include both use by professionals and DIY products used by consumers. It also means that use itself can be both within public buildings and domestic properties, which also affects the potential rates of emission, pathways, and exposure. No further efforts have been made to disaggregate between articles used in public and private buildings. Emissions in the waste stage is included in section B.9.18. Environmental release categories (ERC), that describe the use of PFASs most closely including release factors were assigned to each of the stages described above. A detailed description on emission calculations using volumes in Annex A.3.14., ERC and release factors, as well as a justification for the choices made, is included in Appendix B.9.14.. Estimated PFASs emissions from construction products, based on the source-flow model using the year 2020 as a baseline, are provided in Table B.107. In the table emissions from each lifecycle stage is shown for both construction articles and construction mixtures. The source-flow model used for the emission estimations is based on a grouping approach rather than a substance-by-substance. The trade-off of using such an approach is that the estimates provided will have a higher uncertainty. However, this approach can still provide useful data to estimate the orders of magnitude for emissions when comparing PFAS groups, life -cycle stags, sub-uses and different sectors. As can be seen in Table B.42, emissions are dominated by construction mixtures which is dominated by the life-cycle stage ‘Application’. Furthermore, it can be seen that the use stage outdoors contributes more than indoor and that fluoropolymers contribute more than nonpolymeric PFASs. 284 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.42. Estimated annual EEA emissions in the construction sector of PFAA and PFAA precursors, polymeric PFASs and total PFASs in 2020 (baseline). Side-chain PFAAs and PFAA Fluoropolymer Total PFASs Emission PFAA C2PFAA ≥C4 fluorinated precursors a s (t/y) (t/y) estimates C3 (t/y) (t/y) (t/y) polymers (t/y) low high low high low high low high low high low high Total volume Articles Formulation 2 405 4 254 10 320 5 241 12 725 4 11 4 0.04 11 0.1 8 20 12 0.04 30 0.1 6 15 6 15 13 31 19 47 0.1 0.2 0.1 0.2 0.2 0.5 0.3 1 11 26 11 26 21 52 32 78 Formulation Application outdoor 5 44 11 106 0.3 3 1 10 5 47 12 116 77 827 193 2 003 82 874 205 2 119 Application - indoor 22 53 2 5 24 58 411 997 435 1 055 Use stage - outdoor 1 3 0.1 0.3 2 4 27 65 28 69 Use stage - indoor 0.02 0.1 0.002 0.005 0.02 0.1 0.4 1 0.4 1 Total 72 174 5 16 77 190 1 343 3 259 1 420 3 449 Processing aids b 0.04 0.1 Use stage - outdoor Use stage - indoor Total Mixtures 987 0.04 0.1 Total articles and mixtures 1 451 3 527 The volume of non-polymeric PFASs and the split between processing aids (C 2-C 3 PFAAs) and PFAAs ≥ 4 is estimated based on Glüge et al. (2020). b Processing aids is here placed under C 2-C 3 PFAAs, however, they can also be PFAS-based solvents/functional fluids a 285 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.14.3. Human exposure The construction sector employs about 18 million people in the EU 41. 16 companies provided data on the number of workers employed in the production and handling of well over 100 different PFAS-based construction products. In total, approximately 12 000 people work in these companies (rounded to closest 1 000). Considering the uncertainty of the estimate of the overall market for PFASs in construction products and given that many of the stakeholders who provided their number of workers, did not specify the related tonnage of PFASs in construction products, extrapolation to t he whole market was not possible. Given the diverse uses of PFASs in construction products, it can be assumed that the share of construction workers that handle PFAS -based products, at least occasionally, is significant. Industrial exposure during production of articles could occur through inhalation of vapours or of dust generated during mechanical manipulation of materials. Stakeholders mentioned that many production plants make use of abatement equipment as well as personal protection equipment to reduce exposure to both vapours and dust. Some stakeholders stated that only trained professionals are handling installations and decommissioning of articles containing PFASs. Stakeholders mentioned that the processes to produce processing aids are largely auto mated and take place in closed systems to capture and re-use PFASs as far as possible. There may be some exposure, particularly during maintenance, but overall exposure is expected to be low. It should be noted that there are several examples of environmental releases of PFASs from factories that use processing aids for the production of fluoropolymers (Lohmann et al., 2020). The addition of PFASs to mixtures during blending and mixing takes place at higher temperatures during which vapours are generated. Stakeholders highlighted that fluoropolymer such as PTFE, ETFE and PVDF are highly stable and have a low vapour pressure. It should be noted that also non-polymeric PFASs and side-chain fluorinated polymers are used in paints and coatings. The highest worker exposure will likely occur during the application stage of paints and coatings (and potentially adhesives). PFAS-based mixtures can be added as coatings/impregnation by spraying, rolling, or brushing onto finished articles, this may occur both indoors and outdoors. B.9.14.3.1. Consumers use For Do-It-Yourself (DIY) sealant and adhesive products, tapes, DIY durable water repellent, impregnation and aftermarket floor protection and paint exposure is possible. (Though, it is not clear to what extent paint used by consumers contain PFASs). However, there is no information on the number of consumers using these products. It is possible that PFAS-based mixtures are occasionally used in DIY adhesives and sealants, most likely in small quantities. This suggests that potential exposure is limited. Products for DIY durable water repellent impregnation of stone, glass, tiles and aftermarket floor protection can contain between 0.5% and 2% PFASs (usually side chain fluorinated polymers). If applied as an aerosol spray, high exposure can occur, and cases of acute intoxication have been observed (both consumers and professional workers) (ECHA, 2017a). Fluoropolymers (such as PTFE, ETFE, and PVDF) are largely stable from degradation, while nonpolymeric PFASs may be bound within the matrix of the article. The greater potential for release and exposure may occur where articles are abraded during use (i.e., moving parts) or where maintenance involves cutting, which would generate dust. Exposure to materials in buildings in which PFAS-based construction products are used is 41 https://ec.europa.eu/growth/sectors/construction_en, date of access: 2022-12-14. 286 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) possible. However, if undisturbed and not subject to direct weathering or washing, the potential release of PFASs is likely low. B.9.14.4. Summary Emissions from construction articles, processing aids and mixtures may occur during formulation (articles and commercial construction mixtures only) in the manufacturing of articles (of processing aids), in the application stage, (commercial construction mixtures only) and in the use stage (articles and commercial construction mixtures only). Emissions were calculated using ERC. Total yearly EEA emissions of non-polymeric PFASs (including side-chain fluorinated polymers) amount to between 100 and 200 t/y (rounded number) and total yearly EEA emissions of polymeric PFASs amount to between 1 300 and 3 300 t/y (rounded number). Almost all emissions occur in the production stage. Only qualitative information is available on exposure of workers and consumers. Stakeholders generally believe exposure is limited because of the use of closed systems, abatement systems, personal protection equipment and handling only by trained employees. High exposure of consumers may take place using DIY products containing PFASs especially when applied as aerosol sprays. 287 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.15. Lubricants B.9.15.1. Introduction The use of PFASs in the lubricants industry is described in Annex A.3.15. B.9.15.2. Emissions Three basic life-cycle stages are considered in whic h emissions may occur, or from which PFASs flow through into the next life cycle stage: 1. Formulation of lubricants (including manufacture of sealed articles). This stage covers the formulation of lubricant products, and manufacture of articles containing lubricants. Lubricants include low viscosity lubricants (oils), greases or other mixtures ( e.g. solid/dry-film lubricant which is a solvent containing a solid additive such as micropowder PTFE). Stakeholders highlighted that 95-97% of all PFAS-based lubricants are used within sealed articles, where lubricants are added once (at point of manufacture). For the remaining 3-5% a total loss is assumed during use and there is a need for relubrication at later stages of use. The formulation stage will cover the emissions associated with formulating lubricants (blending/mixing) and losses during filling/injection of lubricants into the manufacture of sealed articles. 2. Use emissions from sealed articles. This life-cycle stage covers the release of PFASs from sealed articles during service life (e.g. leaks, faulty equipment, accidental releases). The applications in this life-cycle stage have been further disaggregated into:  Outdoor applications (automotive, aviation hydraulic systems and landing gears, industrial machinery used outdoor, diving equipment etc).  Indoor applications (ovens, belts and chains, gears, industrial machinery used indoors, automotive uses (e.g. inside a car). It is assumed that any PFASs released outdoors end up directly in the environment, while PFASs released indoors more likely end up in wastewater systems. 3. Use emissions from open applications. This life-cycle stage covers the smaller set of applications identified previously (3-5% of total use). The applications from this life-cycle stage have been further disaggregated into:   Outdoor’ applications (e.g. use of specialist dry-film lubrication for bike chains etc). Here it is assumed all PFASs are released into the environment. Indoor applications (mould release agents, re-lubrication of belts and chains etc). Besides the life-cycle stages mentioned above, emission of PFAS-based solvents for cleaning before lubrication is also included. Environmental release categories (ERC), that describe the use of PFASs most closely including release factors were assigned to each of the stages described above. A detailed description on emission calculations using volumes in Annex A.3.15., ERC and release factors, as well as a justification for the choices made, is included in Appendix B.9.15.. Estimated PFASs emissions (t/y) from lubricants, based on the source-flow model using the year 2020 as a baseline, are provided in Table B.43. In the table, emissions from each life-cycle stage are shown. Emissions in the waste stage is included in section 0. 288 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.43. Estimated annual EEA emissions in the lubricants sector of PFAA and PFAA precursors, polymeric PFASs and total PFASs in 2020 (baseline). Emission estimates Nonpolymeric Fluorinated PFASs a gases (other (solvents)b additives than PTFE) Fluoropoly mers (micropowder PTFE, solid additive) PFPEs (base oil) Total fluoropoly mers and PFPEs Total PFASs low low high low high low high low high low Total volume 70 150 1 10 800 1 200 300 800 1 100 2 000 1 171 2 160 Formulation stage 1 2 0.02 0.2 19 29 7 19 26 48 27 50 Use stage - Sealed equipment 1 Outdoor application 2 0.03 0.3 22 33 8 22 30 54 31 57 3 0.03 0.3 25 38 10 25 35 64 36 66 0 0.01 0.1 11 16 4 11 15 27 15 27 0.02 0.2 13 19 5 13 17 32 17 32 Use stage - Sealed equipment -Indoor 1 application Use stage - open 0 outdoor application high high Use stage - open indoor application 0 0 Use of cleaning agents outdoor 18 38 18 38 Use of cleaning agents indoor 8 17 8 17 152 287 Total Assumed to be PFAAs ≥ 4 b The use of PFAS-based solvents (functional fluids/fluorinated liquids) for industrial cleaning is described in Annex A.3.9. on Fluorinated gases. a The source-flow model used for the emission estimations is based on a grouping approach rather than a substance-by-substance. The trade-off of using such an approach is that the estimates provided will have a higher uncertainty. However, this approach can still provide useful data to estimate the orders of magnitude for emissions when comparing PFAS groups , life-cycle stags, sub-uses and different sectors. B.9.15.3. Human exposure B.9.15.3.1. Lubricant formulation/production It is estimated by various stakeholders that fewer than 1 000 workers are involved in the formulation of PFAS containing lubricants in the EU/EEA. It is st ated by stakeholders that worker exposure is very low due to strict controls, closed systems or encapsulation and low volatility of PTFE and PFPE. No exposure data were provided or identified. The same applies for formulation of lubricants containing PFAS-based solvents where some inhalation exposure cannot be excluded. B.9.15.3.2. Immediate downstream users The immediate downstream users are those typically incorporating the lubricants containing PFASs into parts/articles such as bearings. As indicated in section A.3.15, there are many different types of parts and equipment and an estimate of the number of workers exposed to PFASs from handling lubricants are not possible to make. Stakeholders estimate there are 10 000 289 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) – 100 000 workers. They stated that many processes have been or are being automated (estimated for applications covering 30 – 70% of the lubricant volume), which reduces worker exposure. Further, it is stated that many applications take place in clean rooms or under conditions that are similar to clean rooms. It is mentioned that the sector has a strong focus on occupational health and safety compliance. In open applications (3-5% of the lubricants containing PFASs) there might be a higher potential for worker exposure than for the above-mentioned more automated processes. B.9.15.3.3. Equipment end-users Stakeholders believed that workers using equipment with lubricants containing PFASs are generally not exposed to PFASs, or the potential for exposure is very limited, due to the low potential for release. B.9.15.3.4. Cleaning PFAS-based solvents are used to clean mass-manufactured parts (such as electronic components). This is commonly done in specially sealed equipment in which solvents can be recovered and reused. For these types of applications, stakeholders argue that exposure is expected to be low. Cleaning of articles that were previously lubricated with PFAS-based lubricants, using PFASbased solvents (cleaning of parts/articles prior to adding lubricants, cleaning of production machinery or during maintenance of e.g. wind power installations) may take place during the entire lubricant life cycle. Worker exposure during these type of cleaning activities is possible and is likely to be relevant because the solvents are volatile and sometimes applied as aerosols. In addition, it is assumed that these activities might be less strictly controlled as they are only carried out occasionally and not always in industrial settings. It is therefore assumed that this could lead to inhalation as well as dermal exposures. No exposure data on this type of exposure were identified or received from stakeholders. B.9.15.3.5. Note on professional uses Most of the above-described worker exposure situations can be seen as industrial uses, but some of the re-lubrication ('open'/'total loss' applications) and some cleaning uses might be seen as professional uses. Further, some of the uses described below for consumers might also be relevant for some professionals (e.g. bike repair). B.9.15.3.6. Consumer use of lubricants Consumer exposure is expected by industry stakeholders to be very limited as consumers generally do not use this type of lubricants because of the expense. One exemption from this are speciality products for bicycle chains. Consumer PTFE lubricants for bicycle lubrication is available on the web at a cost of DKK 70 (equalling around €9.4) for 400 mL. Other multiple purpose PTFE lubricants, which can be acquired via websites have also been identified. Some of these latter lubricants can be bought by consumers but might be intended for professional users. While no evidence of widespread use of PFAS-based lubricants by consumers has been found, the price quoted above does not support the claim that PFAS-based lubricant is too expensive for consumers, considering more expensive fluorine-free lubricants are also being sold. In any case, all industry stakeholders interviewed do not consider fluorinated lubricants for consumers use as critical or essential. 290 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.15.3.7. Consumer use of lubricated equipment Similarly, as for worker end-users, industry stakeholders consider consumer exposure to PFASbased lubricants in consumer articles (cars, household appliances, power tools etc.) to be very limited, due to the low potential for releases in these lifetime lubrication applications. B.9.15.4. Summary Emissions from lubricants may occur during formulation, use emissions of sealed articles, and use emissions from open applications. Emissions were calculated using ERC. Total emissions of non-polymeric PFASs amount to between 0.1 and 1 t/y (rounded numbers) and total emissions of polymeric PFASs amount to between 110 and 230 t/y (rounded number). Emission of fluorinated gases abount to between 30 and 60 t/y (rounded numbers). Only qualitative information is available on exposure of workers and consumers. Stakeholders generally believe exposure is limited because of the use of closed systems, abatement systems, personal protection equipment and handling only by trained employees. 291 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.16. Petroleum and mining B.9.16.1. Introduction The use of PFASs in the petroleum and mining sector is described in Annex A.3.16. B.9.16.2. Emission B.9.16.2.1. Emissions from tracers Limited specific data have been obtained in the assessment on the precise release patterns of PFAS-based tracers (e.g. in which environmental compartment release will occur), or exactly how these substances will behave when released to the environment. It should be noted that the volume of emissions for these products estimated during formulation using the ECHA (2016a) guidance ERCs are broadly in line with an estimate provided by a supplier. B.9.16.2.2. Emissions from anti-foaming agents Very limited specific data have been obtained in the assessment on the precise release patterns of PFAS-based anti-foaming agents e.g. in which environmental compartments release will occur, or exactly how these substances will behave when released into the environment. B.9.16.2.3. Emissions from fluoropolymers An estimate has been made for the total sales and use of fluoropolymers in Europe for the 2020 baseline year for petroleum and mining sector (3 500-7 500 t), however it has not been possible to differentiate between different types of fluoropolymers within this estimate. Therefore, the emissions estimate conducted takes a general approach in t erms of fluoropolymers used in the petroleum and mining sector rather than attempting to disaggregate for different polymer types. The estimate of emissions from this part of the life cycle only considers the ‘new’ usage of fluoropolymers in the baseline year. As noted by Lohmann et al. (2020), it is important to distinguish between fluoropolymer substances, fluoropolymer products, and fluoropolymers in finished articles, as there are many different processes of making fluoropolymer products (Lohmann et al., 2020). While it could be expected that fluoropolymer itself could be released to an extent at various stages of the product lifecycle (e.g. through abrasion of tubes, linings and joint sealants during use, release from end-of-life processes such as landfill, incineration, etc.), the availability of data to provide quantitative estimates of releases from these routes is currently lacking so this aspect has not been considered in the present study. While it has been considered that fluoropolymers are distinct from other non-polymeric PFASs and should be separated from them for hazard assessment or regulatory purposes (Henry et al., 2018), the possible releases of ‘leachable’ non-polymeric PFASs (e.g. processing aids, synthesis by-products and oligomers) from the use of fluoropolymer in specific downstream uses need to be considered (Lohmann et al., 2020). The level of residual monomeric PFASs in the fluoropolymer matrix is highly uncertain and limited specific data were available in this assessment. The ECHA ERCs for the formulation and use stages were applied to the total PFAS content estimated for the total fluoropolymer. Broad ‘high’ and ‘low’ [residual concentration] scenarios were used to produce an estimate and wide range of total quantity of monomeric PFASs present in the fluoropolymer used in Europe were presented between these two scenarios, to illustrate this uncertainty. The ‘high’ scenario was used by the ECHA (2015a) Annex XV restriction report on PFOA to derive estimates for the proportion of total fluoropolymer used containing different concentrations of 292 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) monomeric PFASs. The assumptions made and PFAS concentrations used in deriving an estimate for the residual PFASs content for the ‘high’ scenario are presented in Table B.44. Table B.44. Content of monomeric PFASs in fluoropolymers "high" scenario. Proportion of FP PFAS PFAS ECHA (2014) concentration concentration PFOA Annex XV market applicable (2020) range (ppm) used (ppm) restriction scenario (%) 0 20-50 1 000-2 000 35 1 500 Worst case/Refined Worst case/Refined Refined 67 17 17 The ‘low’ scenario makes the broad assumption of 1ppm concentration for PFASs (unspecified) in all fluoropolymers. Using this approach, an estimate for the residual quantity of monomeric PFASs in the fluoropolymer is 4-8 kg (low scenario) and 900-1 900 kg (high scenario). These estimates only consider the emissions from ‘new’ fluoropolymer being used in this sector each year. It has not been possible to estimate the tonnage of ‘in-situ’ fluoropolymer in this sector, which could act as a source of residual monomeric PFASs to the environment. B.9.16.2.4. Emissions from non-polymer PFASs An overview of the total emissions (baseline 2020) of non-polymeric -PFAS to the environment from the petroleum and mining industry in the EEA is provided in Table B.45. Emissions from waste are further described in section B.9.18. 293 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.45. Estimated annual EEA emissions in the petroleum and mining industry of PFAA and PFAA precursors, polymeric PFASsand total PFASs in 2020 (baseline). Emissions Emissions Total Total Group PFAS Quantity Quantity Emissions Emissions to to marine to land Emissions quantity compound of product of PFAS to air (kg) freshwater water (kg) (kg) (kg) entering used (kg) (kg) (of which waste (kg) transferred to waste) (kg) 1 000 25-70 20-25 0-110 0-5 165-185 20-145[1] Water and Fluoro-alkane 1 000 (20) gas tracers tracers + other PFAScontaining tracers Drilling/ Production chemicals Fluorosiloxane antifoaming agents Monomeric PFAS (not specified) 170 000 3 400 – 8 500 85-635 70-210 (15-40) 20-760 0-45 3 500 000 4-8 1-2 <1 <1 <1 Fluoro– (<1) polymers 7 500 500 (all) Low scenario Monomeric 3 500 000 900-1 900 270-580 3-6 20-40 20-45 FluoroPFAS (not – (2-4) (1-2) polymers specified) 7 500 500 (all) High scenario [1] ~20 kg from the formulation stage (freshwater and soil); 0-125 kg from the use stage (recaptured at surface) 170–1 650 70-360 [2] 1-3 <1 (formulation) 1-3 (end of life) 310-670 3-6 (formulation) 3 310-670 (end of life)3 [2] 70-170 kg from the formulation stage (water (before STP) and soil); 0-190 kg from the use stage (recaptured at surface) [3] C alculated using the total calculated residual monomeric PFAS content of the fluoropolymer and the total emissions across all use stages 294 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) For tracer and anti-foam products, the levels of emissions estimated in Baseline year: 2020 Notes on ranges of data: [1] Scenario 3; [2] Scenario 1; [3] Scenario 1 and 2; For a full description of each scenario, see NEA (2021). * C alculated using Environmental Release C ategory (ERC ) emission scenario no.2 for ‘formulation into a mixture’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 2.5% w.w to air; 2% w.w to wastewater and 0.01% w.w to soil. Figure B.81 and Baseline year: 2020 Notes on ranges of data: [1] Scenario 3; [2] Scenario 1; [3] Scenario 2 and 3; For a full description of each scenario see NEA (2021). * C alculated using Environmental Release C ategory (ERC ) emission scenario no.2 for ‘formulation into a mixture’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 2.5% w.w to air; 2% w.w to wastewater and 0.01% w.w to soil. Figure B.82 are expected to be directly proportional to volumes of production and use of these products. Therefore, the pattern of emission levels (relative to the 2020 baseline) from 1990 to 2050 period, is expected to follow the same trend as presented in the baseline scenario for expected tonnage of sales and use for these products as presented in A.3.16.2. B.9.16.2.5. Historic emissions Backward looking (1990 – 2020) For tracer and anti-foam products, the levels of emissions estimated are expected to be directly proportional to volumes of production and use of these products, therefore the pattern of emission levels (relative to the 2020 baseline) from 1990 to 2050 period, is expected to follow the same trend as presented in the baseline scenario for expected tonnage of sales and use for these products (see Annex A.3.16.2.). In the case of fluoropolymers the emission levels of monomeric PFAS from fluoropolymer will be dependent on the overall residual levels present in the fluoropolymer material after manufacture. A general assumption has been made (based on the discussion in Lohmann et al. (2020), and informed by consultation with manufacturers and suppliers in the assessment) that a higher concentration of monomeric PFAS in fluoropolymer can be expected in the fluoropolymer used in previous years (e.g. pre-2020) compared to the 2020 baseline year. For purposes of the ‘high’ scenario presented in this assessment, the quantities of monomeric PFASs in the 1990-2020 period have been based on the scenarios outlined in the ECHA (2015a) PFOA Annex XV restriction report. In this assessment, the 2020 baseline, has been based on the ‘refined scenario’. It is assumed that the ‘worst case’ scenario presented in the ECHA (2015a) PFOA Annex XV restriction report applies in the period 1990-2000, with an assumed transition between the ‘refined’ and ‘worst case’ scenarios applying in 2010. An overview of the assumed proportions of fluoropolymer containing different concentrations of residual PFASs for the period 1990 to 2020 (low scenario) is provided in Table B.46 below. 295 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.46. Concentrations of monomeric PFAS in fluoropolymer (high scenario), assumed for the period 1990-2020. PFAS concentrati on range (ppm) PFAS concentra tion used (ppm) ECHA (2014) PFOA Annex XV restrictio n scenario Proportio n of market (1990) (%) Proportio n of market (2000) (%) Proportio n of market (2010) (%) Proportio n of market (2020) (%) 0 0 Worst case / Refined 67 67 67 67 20-50 35 Worst case / Refined 17 17 17 17 1 000-2 000 1500 Refined 8 17 1 00050 000 25 500 Worst C ase 17 17 8 For purposes of the ‘low’ scenario presented in this assessment, the quantities of monomeric PFASs in the 1990-2020 are based on simple general assumption (based on the discussion in Lohmann et al. (2020) that all fluoropolymer pre-2000 will have a residual PFASs concentration of 10 ppm, and all fluoropolymer from 2020 will have a residual PFAS concentration of 1 ppm, with an assumed transition between these scenarios applying in 2010. An overview of the assumed concentrations and proportions of PFASs in the fluoropolymer for the period 1990 to 2020 (low scenario) is provided in Table B.47 below. Table B.47. Concentrations of non-polymeric PFASs in fluoropolymer (low scenario), assumed for the period 1990-2020. PFAS Proportion Proportion Proportion Proportion concentrati of market of market of market of market on range (1990) (2000) (2010) (2020) (ppm) (%) (%) (%) (%) 1 50 100 10 100 100 50 For fluoropolymers, the emissions of monomeric -PFASs will be dependent on the residual content of PFASs in the polymer matrix. The emission time series has been derived, based on scenarios of estimated PFAS concentrations on fluoropolymer in past and future years. In general, it is expected that pre-2000 fluoropolymer will have much higher concentrations than the 2020 baseline and concentrations would decline after the 2020 baseline. This explains that the projected emission of PFAS from use of fluoropolyme rs (high scenario) has been higher before 2000 than the projected emission of PFASs from antifoaming agents. However, projected emission of PFAS from use of fluoropolymers have decreased below the level of the antifoaming agents by 2020. B.9.16.3. Human exposure a) It is not expected that direct exposure to consumers or the general public would occur. It is expected that direct exposure will only occur potentially for occupational workers across the supply chain. This could include: Workers involved in formulation of PFASs/fluoropolymers in products. 296 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) b) Workers handing PFASs/fluoropolymer products on site. c) Workers involved in handling wastes containing PFASs on site. Exposure to fluoropolymers and non-polymeric PFASs can occur at various points along the supply chain, but this is expected to be minimal due to the use of adequate PPE, closed equipment and ventilation. Both the landfilling and incineration of fluoropolymer could potentially lead to the release of PFASs to the environment (Lohmann et al., 2020). It has not been possible to derive an estimate for the total number of workers impacted by the use of PFASs and/or fluoropolymers in these sectors, due to a lack of data and the relatively large number of steps and complexity in the supply chain. B.9.16.4. Summary Total emissions to the environment from tracer products amount is between 165–185 kg, total emissions from anti-foaming agents amount is between 170–1 650 kg, and total emissions from fluoropolymers amount between 310–670 kg for the high scenario and between 1–3 kg for the low scenario. Emissions from the petroleum and mining sector were calculated using ERC. However, for the use phase, the default values in the ECHA (2016a) Guidance have been adjusted to take into account the estimated distribution of use between onshore and offshore installations and the assumed impact this has on the emission pattern of chemicals at these facilities. This is a 90:10 split between offshore: onshore installations, assuming a 100% release to marine waters at offshore installations, and a 50% release to soil, 50% release to fresh water at onshore installations. Direct exposure to consumers or the general public is not expected t o occur. It is expected that direct exposure will only occur potentially for occupational workers across the supply chain. Unfortunately, it has not been possible to derive an estimate for the total number of workers impacted by the use of PFASs and/or fluoropolymers in these sectors, due to a lack of data and the relatively large number of steps and complexity in the supply chain. 297 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.17. Active substances in Plant Protection Products (PPP), Biocidal Products (BP) and Medicinal Products (MP) As the uses of active substances in Plant Protection Products (PPP), Biocidal Products (BP) and Medicinal Products (MP) are derogated from the restriction proposal without a timelimit, this sector has not been studied in detail. Specific information on emissions from these uses can therefore not be presented in this section. 298 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.18. Waste B.9.18.1. Introduction The waste stage is part of the life cycle of a substance as is, in a mixture or in an article. Therefore, its distribution via the supply chain including service-life of articles and waste stage as an emission source with its associated risks is taken into account in this Annex XV proposal. Understanding of waste management (recycling, landfill, incineration or composting) and the fate of PFASs in the waste stage, is important for understanding the potential for full life cycle emissions. The waste stage, following use of substances, mixtures and articles, is included in the rules and guidelines for chemical safety assesment. This means also for the waste stage of substances registrants have the obligation to proof that risks are adequatelly controlled for human health and the environment (ECHA, 2008; ECHA, 2010). PFAS emissions heavily depend on the physical state of the substance (gaseous, liquid or solid), its application (open or closed, process application, its presence as a constituent of another substance, or its presence in a mixture or article) and may arise during all the life cycle steps: production, application, use and end-of-life/waste stage. This is reflected in Figure A.1 in Annex A. Waste stage emissions may be of relevance for all uses where PFASs are still present in the products or articles at the end-of-life stage. So, the waste stage emissions would be typically relevant in case emissions during use represent only a fraction of the total PFASs content in the product or article. In this section the end of life stage and emissions in this phase are described. For a number of PFAS uses, a significant percentage and tonnage of PFASs enter the waste stage. Waste operations (from collecting, bulking, sorting to composting/landfilling/incineration/ recycling) could contribute significantly to the environmental releases of PFASs, and thus a qualitative or quantitative assessment is needed. In case of a (partial) restriction, for impact assessment purposes it is helpful to know if waste stage emissions are (partly) prevented. In the impact assessment (Annex E.2.) the waste stage emissions are not taken into account in the different use sectors, mainly because of the high (emis sion) uncertainty related to the routes of disposal and the final waste treatment method applied. In Figure B.76 the emission routes from waste collection to final waste treatment are presented. 299 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.76. Overview waste stage emissions. Some waste treatment processes are interlinked, e.g. landfilling and waste water treatment. Emissions from waste treatment may affect different resources important to society (farmland, food, drinking water, etc.). A number of relations between PFAS emissions from waste (treatment) and affected resources is shown in Figure B.77. 300 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.77. Schematic overview of PFAS emissions with waste treatment facilities as important source(s). This overview is meant as an illustrative example that is non exhaustive. Adopted from: Helmer et al. (2022). B.9.18.2. Overview of waste stage emissions The Dossier Submitters consider waste stage emissions for PFAS uses in the following sectors: TULAC, food contact material and packaging, construction products, transportation, medical applications, HVACR, electronics and semiconductor, and energy. These sectors are given a high priority forr the assessment, based on anticipated waste volumes containing PFASs. In Table B.48 the groups of PFASs used in the various sectors, their presence in open or closed applications and high waste stage emission relevance is provided. The relevant uses in which high emissions are expected, are highlighted in green. The uses that are not highlighted are expected to contribute to emissions from waste to a lesser extent. For fluorinated gases much remains unclear regarding potential emissions in the waste stage. A high tonnage enters the waste stage (see Table B.59) but it is unclear to the Dossier Submitters if destruction and recovery of fluorinated gases fully prevent emissions of fluorinated substances in these waste treatment processes as information is rare on this topic. 301 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.48. Groups of PFASs used in the various sectors , presence in open or closed applications and high waste stage emission relevance (light green). Recyclin g x Articles (high relevance in waste stage and recycling if ticked) x x x x x x x x x x PFAS use Polymeri c PFASs TULAC x Food contact materials and packaging Metal plating and manufacturing of metal products Fluorinate d gases C onsumer mixtures PFAA and PFAA precurs ors Open applicatio n (emission mainly in use stage) C osmetics x Ski wax x HVAC R x Medical applications x Transportation x Electronics and semiconductor x Energy x x x x x x x x x x x x x x x C onstruction products x x x Lubricants x x Petroleum & Mining x x x x B.9.18.2.1. Main shortcomings in linking waste stage emissions to use categories Linking PFAS use categories (the uses presented in the table above) to potential waste stage emissions are associated with a number of uncertainties. There is no clear connection between PFAS use categories and EWC waste categories (Eurostat, 2010), and it is not always clear in which waste treatment operations (e.g. landfill, incineration, recycling) the PFASs containing articles are processed. Besides this, the effectiveness of waste treatment is not always clear and many uncertainties remain (e.g. EPA-US (2020)). Furthermore, it is not fully clear where the main emissions during the waste stage occur. Not all emission routes are clear and very often only some (specific) PFASs are measured. In a recent study on leachate from landfills, 88% ± 4% of the extractable organofluorine substances were unknown (Bjorklund et al., 2021). Ideally, PFAS emission factors per main disposal or recovery operation should be derived. This information is currently not available as for deriving a reliable emission factor, information on full PFAS input as well as full PFAS output with regard to the waste operation 302 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) is needed. This should be done for each waste category and treatment method. This often is not possible as i.e. PFAS output measurement (in leachate, flue gas, treated wastewater, etc.) are reported in literature but information on total PF AS input (i.e. to landfill or to incinerate) is generally not available or scarce. B.9.18.2.2. Methods applied Two methods were used to get an overview of waste stage emissions where possible: Literature study and using ECHA ERCs. A literature study was conducted on yearly EEA PFAS waste stage emissions. The following main waste treatment options were analysed to try to get emission estimates (with all limitations):    Landfilling Incineration Waste water treatment PFAS emissions related to these main waste treatment methods are described with a brief overview of literature and the limitations encountered. For these three waste treatment methods a brief overview of conclusions from former PFAS restriction dossiers is included as well. The three waste treatment methods that are covered are potentially important emission sources. Additional emissions can occur from sewage sludge applications and from waste transfer stations for instance. In section B.9.18.2.10 the PFAS waste stage emissions for the relevant PFAS uses of Table B.59 are described. B.9.18.2.3. Landfilling Landfilling is a common waste treatment method for numerous waste streams. The intention of the European Commission is to lower the amount of waste being landfilled 42. Landfilling in general does not fully destruct PFAS in waste streams. The individual contribution of certain PFAS uses to PFAS landfill emissions is hard to determine. It can be assumed that waste streams containing high PFAS concentrations lik some industrial waste will likely contribute more to PFAS emissions from the waste stage than waste streams with low PFAS concentrations (ITRC, 2021; Lin et al., 2022; Solo-Gabriele et al., 2020). Based on the data published from 10 countries (>250 landfills), C4-C7 PFCA are predominant in leachate, while FOTH is dominant in landfill air (Zhang et al., 2023). The relevance of landfills for PFAS emissions has been shown in multiple studies. It was found that a waste water treatment plant (WWTP) receiving leachate from a landfill can have up to three-time higher PFAS concentration in the influent than a WWTP that does not receive leachate (Masoner et al., 2020). This indicates the risk of PFAS emissions from landfills. Recently, this was also seen in the Netherlands where landfill leachate appeared to be a major PFAS source in a study on PFAS in WWTPs (STOWA, 2021b). PFAS emissions to air are less well researched but could be of importance, especially for telomeric alcohols (Lin et al., 2022). It has also been shown that the climate and the age of waste deposited on the landfill have an effect on the PFAS concentration in the leachate (Lang et al., 2017). It is assumed, that over time 100% of all containing PFASs or their degradation products will eventually end 42 https://environment.ec.europa.eu/topics/waste -and-recycling/landfill-waste_en, date of access: 2022-12-15. 303 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) up in the leachate or in the ground. Additionally, intentional and unintentional fires on landfills may release PFASs to air. Based on an extensive literature review performed by the Dossier Submitters, an annual EEA emission of PFASs from landfill of 1-6 t/y was estimated. Combining the concentration of PFASs in leachate (around 7 460 ng/L) with the leachate generation factor from Brennan et al. (2016)(0.2 – 1 m³ leachate generated per tonne of waste landfilled per year) and the total amount of landfilled waste in 2018 (838 861 071 t), the PFAS load generated in the EU from the waste landfilled per year can be calculated (Table B.49). Table B.49. Calculated total PFAS annual loads based on the available Eurostat data (landfills). PFAS Annual load PFAS load mean PFAS load median PFAS load median min (t/y) max (t/y) min (t/y) max (t/y) 1.14 5.72 0.597 2.98 Research in Sweden showed that there likely are significant unknown PFASs present in leachate as “PFAS 11 analysis” could only explain 5-40% of Total Organic Fluorine (TOF) analysis (Avfall, 2021). Therefore total PFAS load could well be 2-20 times higher than targeted measurements of PFASs in samples. In addition, PFASs continue to be released from landfills for many years, even after closure of a landfill (Propp et al., 2021). Landfill mass balance For landfill emissions there are numerous studies on PFAS leachate concentrations, however information on a mass balance is scarce. In the USA, in the PFAS Waste source testing report, an attempt was made to make a PFAS mass balance for landfilling by measuring PFASs in certain (not all) waste streams entering the landfill and PFAS leachate measurements (SANBORN, 2019). To help evaluate what fractions of the PFAS compounds potentially leach from waste, versus what is sequestered (i.e., what is held on to) in the landfill a mass balance was derived. Information on PFAS load entering the landfill via waste, estima ted via waste sampling, and PFAS load leaving the landfill, estimated with leachate sampling, was collected and analysed. A large PFAS mass imbalance was seen, see also Figure B.78. It was noticed that:   A significant portion of PFASs is sequestered in the landfill for a long time. Longer chain PFASs are sequestered to a higher degree than short chain PFASs. PFAS emissions to air from the landfill were not considered but could be of relevance for volatile PFASs. 304 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.78. PFAS mass imbalance landfill. Source: SANBORN (2019). Because of sequestration and delay between PFASs entering the landfill and PFASs leaving the landfill via leachate (or air), a PFAS restriction will on the short term not lead to lower PFASs leachate concentrations. There will be a strong delay before lower PFASs loads entering the landfill will result in lower PFASs leachate concentrations. Even closed landfills continue to emit PFASs via leachate (and possibly via air)(Propp et al., 2021). Landfilling in previous restriction dossiers In the Annex XV dossiers for both PFOA and PFHxA, landfill has been mentioned as important sinks for these subst ances in specific product groups. The PFOA Annex XV report indicated that landfills may pose a potential source of PFASs in the environment through volatilization and through leaching to soil and groundwater (Bossi et al., 2008; Busch et al., 2010). The report further indicated that although untreated waste landfilling is not permitted since 2005, old landfills may still pose a problem and that the treatment of leachate is applied to most, but not to all landfills. A German case study indicated that more than 20% of active and non-active landfills are not treated at all. Further uncertainty was recognized concerning the disposal of the sludge and filters applied in the treated landfills. The dossier recognized that emissions might be higher than expected due to degradation and indicated a potential lag in time due to persistency. No quantitative data on PFOA mass flows in landfills were provided. The PFHxA annex XV report reported between 55 and 60% landfilling based on Eurostat data for 2018, but indicated considerable differences between the European member states. The report also indicated that no specific disposal data and recycling rates for paper and cardboard and for textiles were provided, which was considered essential for a proper evaluation. Based on recent Eurostat data for emission calculation it was assumed that 38% of waste is landfilled. The DS is aware that this is a heavily simplified approach as landfill rates vary strongly among waste streams as well as among EEA countries and time. 305 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.18.2.4. Incineration Waste incineration is a major waste treatment method in the EEA and its share in waste treatment methods has grown in recent years. In most European Member States, incineration is applied43. The amount of waste incinerated in EU member st ates increased by 105% between 1995 and 201844. Currently a total of about 100 million tonnes of waste are incinerated in the EEA per year (Bjorklund et al., 2021). The main waste categories that are incinerated are industrial hazardous waste, healthcare waste and municipal waste. For municipal waste, 29% (2018) was incinerated with energy recovery 45. PFASs are more difficult to break down than other halogenated organic chemicals such as PCB, due to PFAS fluorine’s electronegativity and the chemical stability of fluorinated compounds (EPA-US, 2020). Waste incineration can be divided into: a. Municipal waste incineration with and without energy recovery. (There are about 500 active municipal waste incinerators in the EEA). b. Hazardous waste incineration (including co-incineration in cement kilns). Hazardous waste incinerators have higher operating temperatures than municipal waste incinerators. There are over 100 installations in Europe. Waste incineration is a high potential source of PFAS emissions as volumes of waste incinerated are increasingly large. The Dossier Submitters are of the opinion that incineration conditions (e.g. temperature) are not likely adequate to fully break down PFASs into non-hazardous fractions like HF, CO2 and H2O. Consequently, not properly controlled incineration may lead to air emissions and a solid ash fractions containing PFASs. Considering the broad application of PFASs the Dossier Submitters assume that most PFASs will end up in non-hazardous waste streams handled by municipal waste incinerators. Specific waste streams that are classified as hazardous (such as infectious medical waste and certain industrial waste) will partly end up in dedicated hazardous waste incinerators. Limited PFAS releases from incineration plants are indicated in literature, and they depend on the type of material and type of incineration. Studies which take into account practical incineration conditions are scarce. Previous Annex XV dossiers concerning PFASs also indicate that there are still large uncertainties on the releases, as a full fluorine mass balance is lacking for incinerators. In Table B.50 the degradation products from the incineration of PFASs found in literature are listed (García et al., 2007; Geertinger et al., 2019; Huber et al., 2009). Table B.50. End products of the thermal degradation of reported PFAS compounds. PFASs End products Fluoropolymers PFOS C O 2, C F4, C 2F6, C HF3, C 3F6, C C lF3, C 4F8, C 2C l3F3, HF, trifluoroacetic acid and other perfluorinated gases C F4, C 2F6, C HF3, C 2H2F2, HF PTFE C O 2, C O, C F4, C 2F6, C 3F6, C 2F4, and other fluorinated compounds Based on waste incineration Best available techniques REFerence documents (BREF) municipal solid waste incineration plants are required to operate at 850 °C with a residence time of two seconds. These conditions may not be suitable for the destruction of PFAS43 https://www.cewep.eu/waste-to-energy-plants-in-europe-in-2019/, date of access: 2022-12-14. 44 https://ec.europa.eu/eurostat/statistics- explained/index.php?title=File:Municipal_waste_landfilled,_incinerated,_recycled_and_composted, _EU,_1995-2020.png, date of access: 2022-12-16. 45 https://www.cewep.eu/municipal-waste-treatment-2017/, date of access: 2022-12-14. 306 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) containing waste. Incomplete destruction of PFASs during incineration can result in the formation of smaller PFASs or mixed halogenated organic byproducts. For complete destruction, without side products, temperatures above 1 400 °C are needed which can be reached in cement kilns (Winchell et al., 2021). It is not clear if PFAS containing waste is fully mineralized to CO2, H2O and HF in real life operational conditions in other incinerators than cement kilns. Waste incinerators cannot process a large intake of fluor containing waste as the HF that is formed can destroy critical parts of the incinerator. Research states that temperatures need to be between 1 600 and 2 000 °C for full destruction. Hazardous waste incinerators (not many capacity available however in Europe) and cement kilns may therefore be better suited for (nearly) full PFAS destruction than municipal waste incinerators (Bolan et al., 2021). Wang et al. (2015a) showed that the addition of calcium hydroxide can catalyse the defluorination process. At temperatures of 900 °C, this method showed high transformation rates, indicated by the formation of calcium fluoride. Flue gas, bottom ash and fly ash Although PFASs can rarely be found in the flue gas and no quantitative data is available, some public ations reported the occurrence of PFASs in the fly and bottom ash of representative waste incinerators, indicating that the PFASs are not fully destroyed. . Additionally, one publication also analysed PFASs in fly ash (Sandblom, 2014), however no fly ash data was available for Europe. Abis et al. (2020) reported that 3% of the incinerated material becomes fly ash and BiPRO (2005) stated that 2.25% of the incinerated material becomes fly ash. Ultimately, the 3% from Abis et al. (2020) was chosen to reflect a worst case scenario. The data generated for fly ash should however be analysed with care, as the data is based on only one publication and data set. From this data, an emission value can be calculated to compare the PFAS occurrence in bottom ash to the landfill and WWTP data (Table B.51, Table B.52, Table B.53). Three publications from Sweden and the Netherlands were identified resulting in the following mean and median values (Rijkswaterstaat, 2020; RWS, 2020; Sandblom, 2014; Wohlin, 2020): Table B.51. Calculated mean and median concentrations in incineration bottom ash and fly ash for the PFAS groups (rounded). Data taken from three studies. n.A.: not analysed. Substance Mean (bottom ash) Median (bottom Mean (fly ash) (pg/g) ash) (pg/g) (pg/g) ∑PFC A 811 43 8 981 ∑PFSA 312 0.00 1 805 ∑PFPA n.A. n.A. n.A ∑Precursors 578 0.00 n.A Total PFAS 1 701 43 10 786 Table B.52. Amount of bottom ash (rounded) from Blasenbauer et al. (2020) and incinerated waste in 2018 in Europe as well as the amount of fly ash generated from that waste. A generation factor of 3% for the fly ash was used from Abis et al. (2020). Amount of bottom ash [t] Amount of incinerated Amount of generated fly waste [t] 2018 [t] ash [t] 15 323 000 144 077 000 4 322 000 307 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) With this data the amount of PFAS in bottom and fly ash can be calculated (Table B.53). Table B.53. Calculated total PFAS amounts in incinerator bottom and fly ash. Mean PFAS amount Median PFAS amount Mean PFAS amount (bottom ash) [kg] (bottom ash) [kg] (fly ash) [kg] 25.5 1.3 46.6 While most bottom and fly ash is landfilled in Europe it can also be used in pavement or highway foundation (Blasenbauer et al., 2020; ECHA, 2012a). Fly ash is often used in cement, concrete and gypsum as well as restoration and filling material in open cast mines, quarries and pits. Most publications conclude that very low amounts of sampled PFASs can be found in the flue gas and that most sampled PFASs are destroyed >99%. Incineration studies monitor a limited number of (PFAS) compounds, leaving the question of unmonitored PFAS and Particles of Incomplete Combustion (PIC) unanswered (Stoiber et al., 2020). Most of these sources are based on laboratory studies and do not necessarily represent circumstances in reality. Neither are full fluorine mass balances provided to make clear where fluorine ends up. Excessive operating temperatures could lead to sludge ash fusion and clinker forming. Very high operating temperatures will be prevented to extend refractory life and increase the time between costly maintenance shutdowns. As recently stated in literature a major weakness in studies is that they do not mimic fullscale operational variations in thermal treatment systems. These variations are significantly different from those in lab scale systems. The lack of closed mass balances (>90% F accounted for) for nearly all experiments strongly implies that fluorinated products may escape thermal and post-treatment processes, elude detection, and be released to the environment . US EPA recently stated that very little information is published on PFAS destruction. Fluorine chemistry is sufficiently different than chlorine chemistry. Chlorine chemistry cannot be extrapolated (Linak and Lee, N/T). US EPA also indicated that temperatures in incinerators are very heterogenous. Waste is heterogenous (composition, humidity, caloric value, etc.) and incineration conditions are heterogenous as well. Residence time, temperature, waste humidity and the presence of catalytic surfaces, often metals, are crucial variables. HF, just as HCL, is affecting the waste incinerator as it can be very aggressive for equipment. Waste incinerator operators and permit requirement will steer toward low HF emissions. (Sorbent injection is often used by operators to prevent damage caused by HF ). Collection of highly volatile PFASs is not a standard requirement for incinerators and limited information is available on PFASs in the scrubber systems of incinerators which could lead to a higher amount of PFASs in scrubber water. Winchell et al. (2021) mentioned that little is known about the fate of PFASs through incineration. PFASs in flue gas, ash, or water streams used for incinerator pollution control may be undet ectable. Water used in the air pollution control systems contained PFASs and may mask the levels emitted by the furnace. The authors also noted that the inorganic fluoride content of the wet scrubber discharge water was over 10 000 times that of the influent flow plus the measured PFASs fed to the first fluidized bed furnace. The impact from energy consumption associated with operating any incineration PFASs destruction technology is negligible compared to the reduction in greenhouse gas emissions due to the destruction of fluorocarbons (UNEP, 2018). 308 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Incineration in previous dossiers Incineration has been mentioned as important waste treatment option in both the PFOA and PFHxA Annex XV report. For the PFOA restriction the dossier concluded that ‘Although incineration might destroy PFOA, a final conclusion cannot be made since insufficient information is available on the behavior of PFOA and PFOA-related substances during the incineration process (see chapter B.4.4.4 in the corresponding dossier).’ The PFHxA Annex XV report also discusses incineration as a disposal route but potential releases from incineration are mentioned only qualitatively in the PFHxA Annex XV report. The current restriction proposal on PFAS in AFFF remarks the following: Page 66. One of the measures to achieve minimized emissions is the safe disposal of PFAS -containing waste. The exposure assessment assumes incineration as disposal method to estimate the emissions to the environment from disposal. However, it is noted that the nature and quantities of emissions of PFASs or other fluorinated substances resulting from these disposal processes are not well known and further research should be carried out in real industrial conditions to ascertain their efficiency. Also, the impact on the emissions of greenhouse gases has not been calculated. B.9.18.2.5. Waste water treatment Waste water treatment from both domestic and industrial sources is known as an important source of PFAS release and has been discussed in previous restriction dossiers. Possible primary sources of PFASs are industry sectors producing or using PFASs (e.g. fluoropolymer production industry, the paper – and textile industry, the photo industry, electroplating c ompanies) as well as landfills. Domestic sources such as households can be secondary PFAS sources (i.e. via toilet paper and personal care products discharge). Generally, PFASs are difficult to remove in Waste Water Treatment Plants (WWTPs) and often higher concentrations are observed in the effluent than in the influent. These increases in concentration in the effluent compared to the influent can be as high as 35 times and are attributed to the conversions of (unknown) PFAS precursors to stable PFAS end-products (STOWA, 2021a). WWTP sludge can serve as a sink for PFASs and thus lead to PFASs emissions and pollution to agricultural soil in case the sludge is used as a fertilizer, which is common practice in some EU countries. Most authors concluded that long-chain PFAS tend to accumulate more in the solid phase than in the liquid phase due to higher intermolecular interactions between the long PFAS chain and the solids (Coggan et al., 2019; Glimstedt, 2016). Some publications also measured PFASs in the air above the water and around the waste water treatment plants. However, these emissions are assumed to be up to a factor 10 lower compared to the PFAS emissions from the effluent (Ahrens et al., 2011). The following mean and median concentrations were calculated (Table B.54, Table B.55, Table B.56) 309 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.54. Calculated mean and median concentrations in the influent/effluent water as well as sludge(WWTP) for the PFAS groups. Data taken from 7 studies. Substance ∑PFC A Mean influent [ng/L] 618 Median influent [ng/L] 150 Mean effluent [ng/L] 388 Median effluent [ng/L] 102 Mean sludge [ng/g] 51 Median sludge [ng/g] 2 ∑PFSA 59 12 162 21 16 2 ∑PFPA 4 0.00 0.90 0.00 0.00 0.00 ∑Precursors 116 8 17 2 47 35 Total PFAS 797 170 567.9 125 114 39 Table B.55. Reported influent, effluent and sludge data from Eurostat for 2016 46 . Influent Effluent Sludge production [million m³] [million m³] [t in d.s.] 14 299 16 944 6 577 000 Table B.56. Calculated total yearly PFAS load for selected EU-Member States based on the available Eurostat data (WWTP). All values in [kg]. Mean PFAS Median Median Mean PFAS Median Mean PFAS load sludge PFAS load PFAS load load PFAS load load sludge effluent effluent influent influent 11 374 5 657 9 614 3 097 748 406 Recent, sampling based, research in the Netherlands estimated a yearly PFAS load in effluent of WWTPs of 65 – 180 kg PFAS/y for the whole of the Netherlands, while 15– 45 kg PFAS/y leaves the treatment plants via sewage sludge (STOWA, 2021a). Given the order of magnitude, this is in line with the calculations above. The effluent from the WWTP flows directly into European surface waters and as such the contained PFASs are directly emitted into the freshwater or marine aquatic environment. It is assumed that approximately 2/3 of the total load of PFASs contained in sewage s ludge is not destroyed as it is either applied on land, composted or disposed of in a landfill. Only roughly a 1/3 of the sewage sludge is incinerated. Waste water treatment in previous restriction dossiers In the PFOA Annex XV report municipal wastewater treatment plants (WWTPs) were identified as important sources of PFOA and other PFASs, but the report also indicated that it was almost impossible to trace back the origin of PFOA and precursor emissions from municipal sewage treatment plants and to estimate the share in overall emissions from different industry sectors and consumer households. The report identified some important sources (textile- and photo industry, landfills and electroplating). The report further indicated that it is difficult to remove PFOA, that PFOA may be formed in the WWTP processes and that air emissions of PFAS may occur. PFOA concentrations in the effluent are often higher than in the influent. It was further indicated that PFOA may bind to sewage sludge and that it can be a pot ential PFOA source in the environment 46 https://ec.europa.eu/eurostat/databrowser/view/TEN00030/default/table?lang=en&category=env. env_wat.env_nwat, date of access: 2022-12-19. 310 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) through the application of sludge for soil fertilization, which is still common in many European countries. The PFHxA Annex XV dossier indicates that it is very difficult to remove PFHxA from the wastewater and thus WWTPs act as a source of PFHxA through their effluents and through the generation of sludge. Modelling results showed that 81% of the incoming PFHxA will be passed to the effluent, 12% to sludge and 8% to air. PFHxA is found by various studies in wastewater effluent and sludge, which confirms the model distribution. The aerial emissions may be related to volatile substances, such as 6:2 FTOH, which may be formed by the degradation of from side-chain fluorinated polymers. The behaviour of PFHxA in treatment plants is triggered by its low adsorption potential (see B.4.2.1) and the persistence. The PFHxS Annex XV dossier describes various industries that may be important sources for the WWTPs such as the chrome plating and pulp and paper mills. The dossier indicates that ‘Emissions might be severely underestimated because potentially large direct emissions to the environment are possible from pulp and paper mill sludge that is produced during the treatment of wastewater derived from a paper mill.’ B.9.18.2.6. Conclusion on PFAS emissions from landfilling, incineration and WWTPs It has been shown that, keeping in mind all the uncertainties mentioned in B.9.18.2.1, WWTP via its effluent contributes the most to PFAS load to the environment with landfilling and incinerations contributing less. However, insufficient data was available for flue gas. Hence, an underestimation for waste incineration is likely. In the Table B.57 below the mean and median PFAS loads are presented. Table B.57. PFAS emissions (rounded) in EEA for three main waste treatment methods based on literature research. Waste treatment method Mean PFAS load (t/y) Median PFAS load (t/y) Landfilling Leachate (water) 1.1-5.7 0.6-3.0 Air No data No data Flue gas No data No data Bottom ash 0.026 0.001 Fly ash 0.047 no data Effluent (water) 9.6 3.1 Sludge 0.75 0.4 Air No data No data Incineration Waste water treatment B.9.18.2.7. Sewage sludge Sewage sludge is generated in WWTP by separating the undissolved particles from the water, which is done in lagoons or basins. As WWTP receives waters from urban and industrial sources they can contain PFASs originating from the production and use stage of the PFAS products (e.g. cosmetics, washing of textiles, etc.). Throughout the wastewater treatment, the PFASs also end up in the sludge, especially the longer-chain PFASs as the intermolecular interaction between the PFAS molecule and the surface of the particles is higher the longer the chain is (Glimstedt, 2016). After the treatment in the WWTP, sludge 311 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) from urban wastewater is recovered (agricultural use or compost and other uses) or disposed of (incinerated or disposed of otherwise). Within the EEA sewage sludge from urban wastewater treatment is disposed of or recovered (land application) in roughly equal proportions 46. Most sewage sludge is generated in Western Europe. In Southern and Northern Europe more sewage sludge tends to be recovered, while more is disposed of in Western and Eastern Europe. Bianchini et al. (2016) reported, that due to the Landfill Directive (1999/31/EC) amounts of sewage sludge disposed of in landfills will rapidly decrease in upcoming years as Member States reduce the amount of biodegradable waste sent to landfills by 2016. Generally, the total sludge quantities fluctuate slightly but remain between 5,78 and 6,53 million tonnes dry substance (d.s.) since 2011. The PFAS load in sludge is estimated at 0.40 t in the EEA per year. Treatment methods vary among the EU Member States and also develop differently. In Central Europe, it is to be expected that the agricultural use of sewage sludge will be further reduced or restricted by regulations. In The Netherlands and some Austrian federal states, land application is already prohibited. In Central Europe, it is to be expected that the agricultural use of sewage sludge will be further reduced or restricted by regulations. In Eastern Europe, an increasingly larger percentage of households connected to treatment plants can be expected. There, agricultural use of sludge is still considered the preferred disposal method. The use of alternatives such as co-incineration are expected, but the market implementation might be influenced by public concerns or possible administrative obstacles due to national legislation47. The reuse of biosolids as soil improver/fertilizer in arable crops represented the most used disposal/recovery option in some European countries. This has led to restrictions in t he use of biosolids with Directive 86/278/EEC, however, an evaluation of the directive in 2014 has found shortcomings also with regards to contaminations such as PFASs. These are currently not regulated. Currently Directive 86/278/EEC is under evaluation and PFASs are likely to be an important attention point 48. Examples in EEA (Rastatt, Germany) and the USA illustrate the need to be careful with land application of sludge and biosolids (Johnson, 2022; Treat, 2021). France, Finland, Germany, Ireland, Italy and Spain currently have the highest share of biosolids recycled to land. Most countries in the EU have prohibited the use of untreated sludge on land, while some Member States (France, Ireland and the UK) permit the use of untreated sludge (Collivignarelli et al., 2019). B.9.18.2.8. Waste transfer stations Waste transfer station are generally used to sort, crush and bulk waste in order to enable an efficient waste packing for further transport. During this process liquids are usually decanted into a separate container (JRC, 2019). During the sorting and crushing process liquids may be spilled from the waste and can end up in the leachate form the waste transfer station. It has been shown in literature, that the amount of PFASs in the leachate from waste transfer stations can reach levels of up to 46 µg/L (Wang et al., 2020a). This study was performed in China and may not be representative of the European waste transfer stations in Europe, however it indicates the risk arising from these transfer stations (RWS, 2020). According to the European Waste Management Association (FEAD) there are 2 400 recycling and sorting centers in Europe (FEAD, 2021). It is unclear how many of these are 47 https://eu-recycling.com/Archive/22374, date of access: 2022-12-19. https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12328-Sewage-sludgeuse-in-farming-evaluation_en, date of access: 2022-12-19. 48 312 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) waste transfer stations. The actual number of waste transfer stations in Europe is unknown. Waste transfer stations might be an important source of PFAS contaminated surface runoff water as was indicated by RWS (2020). Leachates from waste stored at waste incinerators maybe an important source of emissions also (Liu et al., 2021b). B.9.18.2.9. Recycling The aim of the restriction is to end all emissions of PFAS substances mentioned in the entry. As the dossier has demonstrated a concern resulting from PFAS exposure, there is a concern for recycled materials containing these substances as well. Therefore this restriction proposal also includes PFASs released during the recycling process and PFASs in recycled material and articles made from recycled materials. The recycling market is a worldwide market. Materials can be exported outside EEA to be recycled/treated. Next to that, recycling is extending the lifetime of an article but in the end landfilling, incineration, composting or dumping/littering remain. Recycling may appear for all types of PFASs that are applied in articles, but is e xpected to be high in textiles and paper, construction products, End-of-Live vehicles and electronics. Especially for lacquers and ink, PFASs likely stays either in the recycled paper and/or is released to water in recycling facilities. For plastics, athough much less studies are availabe, the same risk could be applicable as both paper and plastics are major recycling flows. From moulded fibre, elevated levels of PFASs have been reported (see Straková et al. (2021)). If such fibres are composted, all PFAS content will be released in the environment, leading to soil, water, and crop contamination 49.For paper and board for instance it is observed that increased recycling will lead to spread of PFASs throughout all the paper and board market, as cited for instance by (Lowe et al., 2021). The Single Use Plastic Directive could lead to more moulded fibre food packaging being used, most of which contains PFASs for water and grease resistance. This could (further) contaminate paper and board recycling streams. Eurostat (2019) has provided 2019 recycling percentages for plastic packaging and Ewaste with 42% and 41% respectively. Recycling of textiles is considered low by the European Commission50. Fluoropolymers are very broadly applied but recycling of fluoropolymers is hardly possible as they are often coated on other materials. It has for instance to be noted that, as also mentioned by stakeholders, the fluoropolymer-content of metal cook- and bakeware is <1% weight/weight and is only removable via sand blasting or burning of the non-stick coating. And this applies to very many other articles: Recycling of fluoropolymers is hardly possible/cost effective. Currently, lithium-ion batteries (LIB) as well as solar panels are niche waste streams. This will quickly change as both categories experience tremendous growth 51. PFAS emissions from improper waste treatment and accidental fires of these applications may become a potential risk. There are for instance a lot of fires at recyclers caused by Li-ion batteries. There will be (PFAS) emissions (Larsson et al., 2017). For solar panels the fluoropolymer coatings in front and back sheets currently can hardly be recycled and end up in shredder fractions which can be applied in (road) foundations or 49 https://www.fidra.org.uk/news/pfas-in-compostable-packaging/, date of access: 2022-12-14. 50 The EC indicate that textile recycling is low: Strategy for textiles (europa.eu) https://www.rystadenergy.com/news/reduce -reuse-solar-pv-recycling-market-to-be-worth-2-7billion-by-2030, date of access: 2022-12-14. 51 313 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) incineration (Späth et al., 2022). Recollection of fluorinated gases from air conditioners and cooling equipment is crucial to prevent emissions in the waste stage. A crucial way to prevent the release of environmentally persistent fluorinated gases into the atmosphere is through collection and immediate destruction of the gases. Destruction methods for fluorinated gases require extreme conditions due to their chemical and physical inertness (Sheldon and Crimmin, 2022). In case recollection of fluorinated gases and/or destruction is not properly conducted, gas emission from waste treatment of cooling equipment can le ad to serious PFAS emissions in the waste stage. Recycling materials enter the production phase again. This is for instance an important factor in paper production where very often new and recycled material are mixed to produce paper. ECHA (2016a) R.16 guidance page 36) states that ‘Where the process leading to the recovery of the substance is the same as the substance manufacturing (like e.g. for some metals), this last step in the waste operation (i.e. when recov ery follows other waste treatments) may be already covered in the assessment of the manufacturing process. Other operations regularly carried out in the context of recycling/recovery (e.g. dismantling processes, milling and separation processes) may need a particular assessment.’. For the release to water and air the default release factors in Table R.1810: Defaults for the paper scenario in the R.18 Guidance may be used. Values provided in Table R18-30 for shredding may be used as well. For other recycling streams, such defaults are not provided. Recycled products, although its lifecycle is extended, will finally be incinerated, landfilled or littered. Recycling in previous restriction dossiers Recycling was only mentioned qualitatively in the PFOA restriction dossier, which indicated that reuse and recycling should not be prohibited to facilitate the sustainable management of resources. Recycling of paper and cardboard was mentioned in the dossier, but could not be quantified because limited information was available. Recycling was not mentioned for the other product groups. Except for transferring PFOA to next generation products, emissions to air and water were also expected due to landfilling and incineration of waste. According to the PFHxA restriction dossier there might be a recycling issue with regard to paper, textiles, semiconductors and building materials. Although textiles and paper and cardboard were considered very relevant sources of PFHxA in the PFHxA restriction dossier, releases from recycling of these materials (i.e. shredding, washing and pulping of wastepaper and cardboard prior to re-use) were not considered in the assessment as no data on PFHxA emissions from the process of wastepaper recycling were available to the Dossier Submitters. B.9.18.2.10. Estimation of waste stage emissions based on ERC Because of all uncertainty regarding waste stage emission as stated in literature, also emission calculation based on ECHA ERCs was applied. To get an indication of the amounts of PFASs that are released from the waste stage ERCs for waste treatment were applied to the volumes of PFASs that enter the waste stage. It was assumed that 38% of waste containing PFASs is landfilled, 38% is recycled and 24% is incinerated. For paper uses, ECHA has different ERC (with higher emission percentages) The volume of PFASs entering the waste stage based on Eurostat 2018 numbers. The 314 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) website in the footnote 52 is a starting point for the estimations. ECHA is assuming a residence time of 20 years for waste in landfills. For landfill emissions in Table B.58 ECHA R.18 Landfill ERC were used. This likely leads to an underestimation of life time PFAS landfill emissions as landfills have a longer lifetime than the default 20 year period. From Figure B.78 above it can be seen that an ERC of around 10% seems more likely. Taking into account that PFAS leachate emissions will continue for many years after closure of a landfill, the 10% could still an underestimation of the PFAS landfill life-time emission fac tor (Propp et al., 2021)53. As explained in Figure B.78, and due to te persistence of PFASs, the fact that the applied ERC is for organic substances and not very persistent PFASs, and the fact that leachate will continue to emit many years after landfill closure, a worst case emission factor of 10% is applied as well in Table B.58. It should be noted that after 20 years (the residence time in landfills asumed via the ERC), PFASs will still be present in the landfill and for decades if not centuries. Part of the PFASs will be sequesterd in the landfill but how much is unclear. Table B.58. ECHA ERCs for the waste stage (ECHA, 2012a). Air Water Soil Incineration 0.0001 0.0001 0 Landfill 0.0005 0.032 0.0016 Recycling 0.025 0.0025 0 Paper 0.15 0.9014 0.00144 WWTP 0.001 0.09 0 Sludge Remark For paper recycling there are separate (much higher numbers) 0.91 The PFAS load entering the waste stage is a starting point for the estimations and applying a generic division among the different waste treatment techniques (landfill, recycling and incineration) based on Eurostat 2018 numbers 52. Subsequently the standard ERC release factors from the ECHA guidance (ECHA, 2012a) for estimating the releases for landfilling, incineration and recycling are applied. ECHA is assuming a residence time of 20 years for waste in landfills. 52 https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Waste_statistics, date of access: 2022-12-23. 53 https://www.wastetodaymagazine.com/article /minnesota-pfas-groundwater-contaminationlandfills/, date of access: 2022-12-19. 315 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.59. Waste stage emissions based on ECHA ERCs. Air, water and soil emissions presented as environmental emissions. PFAS load (t) entering solid waste stage midpoint unless stated others TOTAL ENVIRONMENTAL COMPARTMENT Lubricants 1 447 34 71 TULAC 69 118 1 621 3 407 FCM and Packaging 4 897 115 241 6 429 151 317 15 240 357 751 Consumer mixtures 0 0 0 Construction products 6 495 152 320 Cosmetics 0 0 0 Metal plating and manufacturing of metal products 954 22 47 Production 36 000* 540* 540 Ski wax 1 0.946 n.a. Transportation (stock) 12 850 301 633 Transportation (new on market) 6 141 144 303 Petroleum & Mining 1 0 0 Medical applications HVACR and 5 901 138 291 19 724 Unknown unknown Use Paper and board packaging (midpoint) C onsumer bakeware & professional bakeware Other uses (packaging, can's etc) TOTAL ENVIRONMENTAL COMPARTMENT life time emissions ERC 10% landfill (water) life time emissions 316 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Use PFAS load (t) entering solid waste stage midpoint unless stated others TOTAL ENVIRONMENTAL COMPARTMENT TOTAL ENVIRONMENTAL COMPARTMENT life time emissions ERC 10% landfill (water) 3 752 88 185 2 995 70 148 3 736 7 255 life time emissions other fluorinated gases applications Electronics and semiconductor Energy TOTAL * Extrapolated based on C hemours, the Netherlands data. Total waste stage emissions based on ERCs are (far) higher than the literature estimates and range between 4 000 and 7 000 t/y. However the emissions of fluorinated gases is not taken into account as there was no data available. B.9.18.2.11. Summary PFASs are used in thousands of applications. All these applications sooner or later reach an end of life stage: The waste stage. In final waste treatment, landfilling and incineration are the most important waste treatment methods. Recycling can extend the lifetime but in the end, for almost all substances, mixtures or articles, only emission in the use phase or during landfilling and incineration apply. Landfilling of PFAS containing material leads to very slow release of PFASs. This can continue for decades or even longer after closure of the landfill. Landfills have more or less a symbiotic relationship with Waste Water Treatment Plants (WWTP) as landfills send leachate to WWTP and WWTP send sludge to landfills. Waste water treatment plants (WWTP) are not capable of removing PFAS: In many cases even higher PFAS concentrations are measured in effluent compared to influent. Spreading of (WWTP) sludge is a high risk for PFAS spreading. In waste incineration many remains unclear. Due to the fluorine electronegativity PFAS are even harder to breakdown via high temperature than other halogenated substances like PCB’s. Incomplete destruction of PFAS compounds can result in the formation of smaller PFAS and/or products of incomplete combustion. The latter compounds have not been researched and could be a potential concern. No study has been found with a full fluorine mass balance. Scrubwater of incinerator flue gas cleaning systems is less well researched but could be a significant source of PFAS. In recycling, PFAS cannot be removed. It will be present in the recycled articles such as paper or plastics. Based on literature research the yearly emissions from landfilling, incineration and waste water treatment are 1-6 t/y. There are serious limitations in literature f.i. incomplete PFAS analysis and lack of closed fluorine mass balances. Therefore additional emission calculations have been also performed. ERC calculations: 317 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) No specific emission factors for waste treatment methods have been found. In some cases like landfilling and incineration (some) PFAS output measurements are available, though often incomplete. But PFAS input to a landfill or incinerator is hardly ever clear. Therefore for additional emission estimates the Dossier Submitters had to rely on generic ECHA ERCs for organic substances. Especially for food contact and packaging waste stage emissions seem to be significant. (Because ECHA has different ERCs for paper packaging). TULAC waste stage emissions are also relatively high due to the high tonnage entering the waste stage. Waste stage emissions in PFAS manufacturing could also be relevant as was shown for Chemours in the Netherlands (and 3M Zwijndrecht and Miteni in Italy). These are however limited examples whereas there are about 20 PFAS manufacturers in EEA. In general waste stage emissions, using ERCs are not to be neglected and for landfilling will continue for a very long period. The risk that PFAS emissions in waste stage are (severly) underestimated is prominent because of the many uncertainties in PFAS waste treatment emissions (incomplet e analysis, lack of mass balances, no clear picture on emission of fluorinated gases in HVACR treatment, etc.) and the application of an ERC for organic substances for a very persistent group of substances. Based on ECHA ERCs the lifetime waste stage emissions (20 year) coming from yearly PFAS loads entering the waste stage, are between 4 000 and 7 000 t (rounded numbers). The fate of fluorinated gases, the PFAS group with the highest emission in manufacturing/ use phase, is unclear in waste stage. Fluorinated gases are recovered or destroyed but for the Dossier Submitters it remains unclear if and to what extent PFAS emissions (emission of fluorinated gases) are to be expected. Because of societal stock and because PFAS use has increased over the last decades, the waste stage will remain an important source of PFAS emissions for many years to come, even in case of a swift full PFAS ban. B.9.18.2.12. Human exposure As a lot of waste streams contain (some) PFASs. Waste treatment can be a source of human and worker exposure. In waste collection until final waste treatment the most risk full phase regarding PFAS worker exposure likely is the waste pre-treatment and sorting phase where dust exposure is often prominent. In waste sorting higher levels of PFAS were found in waste recycling plant workers serum and urine compared with other job assignments, such as managers, suggesting that sorters may be directly exposed to PFAS. And in dust samples collected from e-waste areas in China PFAS c oncentration was 15 times higher than in clothing shops (Peng et al., 2022). Acc ording to another study, waste workers are chronically exposed to PFAS released during the recycling/repurposing of materials at e-waste processing sites. Although the study was performed in China, exposure at EU (e-waste) recycling facilities likely is non-neglectable as well (Garg et al., 2020). Also in recycling/processing of construction materials including contruction waste, (PFAS) dust exposure is a risk. As construction waste is a high volume, often dusty, waste stream, this could be of relevance for worker exposure (Chamberlain et al., 2022; Fernández et al., 2021). Exposure during waste treatment can result in PFAS intake through inhalation, ingestion, and dermal routes (Tansel, 2022). Not only waste sorters and waste recycling workers are exposed to PFAS, also people and communities around waste treatment sites are (disproportionately) affected. Studies from the U.S. and Europe reported that waste facilities such as incinerators and landfills are more frequently located in disadvantaged communities, causing pollution and health inequalities in addition to economic and social injustices (Stoiber et al., 2020) .Risks to 318 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) occupational and public safety from waste are greater in low-income and middle-income countries. Export of PFAS containing waste (i.e. WEEE) to countries outside EEA could lead to serious (worker) exposure issues in the waste receiving countries, especially in case informal waste treatment methods like open burning etc. are applied (Cook and Velis, 2020). 319 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.19. Other sources (for example releases) natural sources, unintentional Besides the manufacturing of PFASs, some PFASs may also be formed as a by -product of other manufacturing processes. A significant emission source of PFCs (specifically tetrafluoromethane and hexafluoroethane) is the aluminium and rare earth metal smelting industry (Kim et al., 2021)54. In the primary production of aluminium, anode effects during the Hall-Heroult process result in the production of tetrafluoromethane and hexafluoroethane as by-products, which is generally emitted to air. Reduction in the emissions of these PFCs cannot be regulated through the EU REACH regulation, as it involves industrial emissions of by-products. Other regulatory actions are being applied to reduce these sources of PFCs to the environment, as these substances are also very longlived greenhouse gases with a high global warming potential, giving these subtances the potential of becoming a near-permanent legacy of the industrial era (Victor et al., 1997). Another assumed source of a specific PFAS – trifluoroacetic acid (TFA) – that cannot be regulated, are natural processes under extreme conditions – like undersea hydrothermal vents (Solomon et al., 2016; UNEP, 2015). However, there is discussion on the relevance of natural sources of TFA in the environment, especially when it comes to the non-marine environments (Joudan et al., 2021; UNEP, 2015). 54 https://www.epa.gov/f-gas-partnership-programs/aluminum-industry, date of access: 2022-12-19. 320 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.20. Overall environmental exposure assessment Assuming a one-compartment model, all emissions to the environment can be considered as direct exposure to the environment. Although the type of exposure may differ, e.g. direct exposure by emissions of fluorinated gases and slow release from fluoropolymers over a long period of time, eventually the environment is exposed to all PFASs that are emitted. In Table B.60 total yearly EEA emissions to the environment are provided for the researched PFAS uses. Emissions from Plant Protection Products (PPP), Biocidal Products (BP) and Medicinal Products (MP) and emissions in the waste stage are not included mainly because of the high (emission) uncertainty related to the routes of disposal and the final waste treatment method applied. 321 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.60. Summary of EEA yearly emissions of PFASs per use sector (waste stage and active ingredients from PPPs, MPs and BPs not included). A pplication PFA As and PFA A precursors (t/y) Fluorinated gases (t/y) Polymeric PFA Ss (t/y) Total PFA Ss (t/y) low mid high low mid high low mid high low mid high Manufacture 54 86 118 309 1 973 3 637 15 23 30 378 2 082 3 785 TULAC b 2 058 6 177 10 295 8 326 16 643 24 960 10 384 22 820 35 255 123 491 858 99 100 100 222 591 959 0.5 6 11.4 1 6 11 C onsumer mixtures 23 23 23 C osmetics 0.015 32 64 1 1 1 38 806 1 696 38 806 1 696 38 806 1 696 Food contact m ate rials and pack aging Me tal plating and m anufacture of metal productsc Sk i wax Applications of fluorinated gasesd,e Me dical devices Transport 128 239 350 38 806 1 696 38 806 1 696 38 806 1 696 3 772 5 586 7 400 d 32 76 120 3 932 5 901 7 870 269 35 439 58 609 80 269 35 439 58 609 80 11 152 292 366 671 976 Ele ctronics and se m iconductors 348 513 677 Ene rgy se ctor 42 42 42 12 13 13 53 55 56 C onstruction products 88 152 216 1 364 2 338 3 311 1 451 2 489 3 527 Lubricants 0.11 0.6 1.1 123 174 225 152 220 288 Pe troleum and m iningc 0.3 0 1 2 TOTA L 2 842 7 707 12 571 42 923 46 418 49 912 10 251 19 958 29 660 56 038 74 137 92 232 Total 2 842 7 707 12 571 5 813 9 308 12 802 10 017 19 577 29 131 18 694 36 646 54 593 f g 7 29 7 46 7 62 2.3 a: In some cases a basis for providing a range is lack ing. There the available e stimate is applied throughout; b: TULAC = Textile, upholstery, leather, apparel and carpets; c: No data available for e mission of polymeric PFASs; d: For the se sectors the e missions re late to stock (pre sented in italics). For re ference only, the emissions from tonnage brought new to m arket in 2020 are also given; e : Includes emissions of fluorinated gases in transport sector. f: Total based on emissions from best available data (stock if available, new to m arket if stock is not available); g: For re ference only, also the total emissions from tonnage brought new to m arket in 2020 are presented. 322 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFASs (including the group Other PFASs) used for the impact assessment (see Annex E.2.). These emissions may be an underestimation of environmental emissions and exposure, as not all PFAS uses were researched, emissions from technical stock (PFASs in products already in use) are not taken into account, except for HVACR (as stock information for this sector was available). Waste stage emissions are not taken into account in Emissions from technical stock may have a considerable contribution to the yearly emissions, especially for products that have a long life cycle. Besides PFASs in technical stock, the environmental monitoring data (see B.4.2.6 and B.4.2.7) show that there already is an environmental stock of PFASs. This environmental pollution of PFAS may already exceed planetary boundaries (Cousins et al., 2022), so all additional emissions contribute further to the environmental stock and associated risks of PFAS. There is a delay in time and spa ce when and where the emissions could be demonstrated in the environment. Based on the available information from the different sectors it is clear that polymeric PFASs and fluorinated gases have the highest emissions. For fluorinated gases, this can also be explained by the fact that technical stock emissions are used for HVACR. 323 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.21. Human exposure In each of the PFAS use sections described above, human exposure is briefly described for each use sector. In the text below the general picture for human exposure is described in more detail. Human exposure to PFASs has been extensively studied through biomonitoring studies and intake estimations via different exposure routes. The wealth of studies unambiguously demonstrates that humans world-wide are exposed to a wide range of PFASs with the highest levels reported for workers and community residents living close to contaminated sites (hot spots). Exposure to PFAAs occurs either via direct intake of arrowhead PFAAs, which are not further metabolised in the body, or indirectly via intake of precursor compounds that can be metabolised to arrowhead PFAAs and thus contribute to the internal levels of these compounds. The ubiquitous presence of PFASs in environmental media (see B.4.2.6 and B.4.2.7) and widespread use in many types of consumer products lead to an array of potential exposure sources (Figure B.79). For the general population (i.e. individuals who are not occupationally exposed or living in contaminated hot spots), exposure routes include ingestion of food and water, intake of indoor dust, inhalation of air and contact with consumer products (De Silva et al., 2021; Haug et al., 2011a; Poothong et al., 2020; Vestergren et al., 2012; Vestergren and Cousins, 2009). The dominating exposure route varies greatly for different PFAAs, reflecting their physico-chemical properties and use patterns. For hydrophobic and bioaccumulative long-chain PFAAs, dietary intake (especially of fish and meat) is typically the most important exposure route, whereas for highly water soluble short-chain PFAAs, drinking water and other food categories, such as vegetables, tend to be the dominating exposure routes (EFSA, 2020; Vestergren et al., 2012). For precursor compounds, exposure to consumer products (e.g. impregnation products) via the indoor environment (air and dust) is probably the major exposure route (Vestergren et al., 2008). Regarding other non-polymeric PFASs, such as PFAEs, the relative contribution from different exposure routes in the general population has not been described. For occupationally exposed individuals, who may have a higher expo sure, the primary routes to PFAA exposure are inhalation, ingestion of dust and dermal uptake at the workplace (De Silva et al., 2021; Fu et al., 2015). The bioavailability and thus the potential for human exposure to fluoropolymers has been an issue for discussion and is further described in B.5.1.2. In summary, it has been proposed that absorption of fluoropolymers in humans is obstructed due to their large sizes (Henry et al., 2018). On the contrary, it has been argued that the production, processing, use, and end-of-life treatment of fluoropolymers lead to emissions of bioavailable compounds (e.g. monomers, oligomers, decomposition and combustion products, PFAA/PFEA polymerization aids, additives, unintentional PFAS by-products, impurities, and particles, etc.), which may be relevant for human exposure (Lohmann et al., 2020). Regarding side-chain fluorinated polymers, potential degradation to more bioavailable PFASs may add to the exposure to these compounds in humans. 324 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.79. PFAS exposure pathways to humans from Oliaei et al. (2013). Children are exposed to PFASs prenatally via placental transfer and postnatally via breast milk, as demonstrated by the presence of PFASs in umbilical cord blood, placenta, breast milk and in the blood of breastfed children. There is a known association between the length of the breastfeeding period and higher levels of PFASs in the child, and conversely, lower levels in the nursing mother, illustrating the importance of breast milk as an exposure route for the infant (Brantsaeter et al., 2013; Haug et al., 2011a; Haug et al., 2010b; Koponen et al., 2018). Results of several studies have highlighted notable differences among exposure to PFASs across different life stages (e.g. breastfed infants versus adults) and genders. In adults, PFAS blood levels are often higher in men compared to women of the same population, which can partly be explained by physiological factors, such as menstruation, pregnancy, breastfeeding and gender-dependent differences in urinary elimination, leading to a faster excretion in women (Brantsaeter et al., 2013; Gomis et al., 2017; Wong et al., 2014). However, the lower blood levels in women may also reflect different expos ure patterns between genders. 325 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.21.1. Dietary exposure In recent years, the environmental persistence of PFASs and their biomagnification in food webs have generated a notable interest and a large number of publications in this regard. For the already regulated PFASs, of chain lengths ranging between eight to fourteen carbons, food (primarily fish, egg, meat) is generally the major exposure route (EFSA, 2020). For shorter chain PFASs, the available data set is less robust, however the data indicate that these substances are present in food, such as vegetables, but that the dietary intake may be less important for the total exposure (EFSA, 2020; Vestergren et al., 2012). The most extensive overview of PFASs in food in Europe was presented in the EFSA risk assessment of PFASs in food (EFSA, 2020). The dataset contained >97 000 food samples collected between 2000 – 2016 in 16 European countries. Quantitative data were available for 17 PFASs, of which eight are not restricted and in the scope of this restriction proposal - PFBA, PFPeA, PFHpA, PFPeDA, PFHxDA, PFDoDA, PFBS, and PFHpS. A cut -off value of 1 µg/kg was applied based on Commission recommendation 2010/161/EC (EC, 2010). The result showed that PFBA, PFPeA, PFHpA, PFBS, and PFHpS were detected in food samples, however close to or below the cut -off value. Of the >97 000 samples, the levels in approximately 24 500 samples were below the cut-off value 1 µg/kg. In the total dietary exposure assessment, PFBA was estimated to contribute with 16% to the total PFAS exposure via diet. PFPeDA, PFHxDA and PFDoDA were not detected in any samples. A number of scientific studies has also analysed PFASs in food in other European countries not covered by the EFSA report. The results confirmed the findings by EFSA, i.e., that the levels of unregulated PFAAs were in the low µg/kg range or lower (data not shown). Contaminated sites may result in highly contaminated foodstuffs, which is illustrated by several studies. For example, high levels of PFOS, up to 230 µg/kg, was found in veal meat close to a firefighting training site in Denmark where the animals had been consuming contaminated grass (DVFA, 2021). This subsequently led to highly elevated blood levels of PFOS in a population consuming the veal meat. Another example is a recent study of individuals living within a 3 km radius from the 3M facility in Zwijndrecht, Belgium, which showed that consumption of locally produced food (especially eggs) was associated with higher blood levels of e.g. PFOS (VITO and PIH, 2021). Furthermore, HFPO-DA (which was not covered by the overview by EFSA) has been detected in the Netherlands at low µg/kg levels in vegetables in the vicinity of a fluoropolymer production plant and close to a company processing products from the fluoropolymer production plant (reviewed in Gebbink and van Leeuwen (2020)). In addition to populations living in areas with higher levels of PFASs in food, certain dietary habits may lead to high exposure to PFASs. A potentially highly exposed group is fishermen, with a diet high in fish, where elevated serum levels of PFASs (including unregulated PFASs) have been shown (Hu et al., 2018; Shi et al., 2016; Zhou et al., 2014). In addition to exposure to PFASs via foodstuff, there is also a possibility for additional exposure via the use of PFASs in food contact materials. Transfer of PFCAs/PFSAs, PAPs and fluorotelomers from food contact materials to the food have been shown (Elizalde et al., 2018; Gebbink et al., 2013; Lerch et al., 2022; Zabaleta et al., 2020) with higher transfer at increased temperatures (Elizalde et al., 2018) and where the tolerable daily intake (TDI) for PFAS in food set by (EFSA, 2020) was exceeded in children (Lerch et al., 2022). In the risk assessment of PFAS in food, EFSA concluded that although PTFE cookware may contain residual PFOA in the low µg/kg range and that food packaging may contain PFASs because of their grease-resistant properties, and that the use of this type of material is likely to contribute to human exposure to PFASs, the contribution is small compared with other sources of exposure, see also section B.9.4.3. In the literature search for this section, no studies on TF, EOF or TOPA in food were found. 326 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.21.2. Exposure via drinking water PFASs are often found in European drinking water (Domingo and Nadal, 2019; Kaboré et al., 2018; Lindfeldt et al., 2021). For example, a recent Swedish national survey reported that approximately half of drinking water samples from 154 Swedish municipal water works contained detectable levels of one or several analysed PFASs (Lindfeldt et al., 2021). Thus, drinking water can be a relevant exposure route for the general population, especially for short-chained PFASs (EFSA, 2020; Vestergren et al., 2012). Monitoring of PFASs in drinking water is further discussed in B.4.2.1.3, B.4.2.6 and B.4.2.7. Drinking water is a particularly important exposure route in areas near point sources of PFASs. For example, elevated levels of PFASs, including non-regulated compounds, have been found in drinking water in areas near fluorochemical manufacturing facilities, firefighting training sites, landfills and agricultural land treated with industrial-waste derived soil improvers in Europe, the US and China (Frisbee et al., 2009; Gebbink and van Leeuwen, 2020; Glynn et al., 2015; Gyllenhammar et al., 2015; Ingelido et al., 2018; Kotlarz et al., 2020; Scher et al., 2018; Wilhelm et al., 2008; Xu et al., 2020c) . As a consequence of the high levels in drinking water, populations residing these contaminated sites often display elevated PFAS levels in blood. For example, Glynn et al. (2020) showed that blood levels in children can be significantly elevated already at relatively low levels of PFAAs in the drinking water (<10 ng/L of single PFAAs). B.9.21.3. Exposure via outdoor air Outdoor air is not considered to be an important exposure route to PFASs in the general population (Egeghy and Lorber, 2011). However, for populations living in areas close to atmospheric point sources, exposure via outdoor air may be of relevance (De Silva et al., 2021). For example, elevated internal levels of PFOA were detected in a population living near the Dupont/Chemours facility in Dordrecht, the Netherlands. The drinking water did not contain elevated levels of PFOA and thus, the air was deemed the most likely exposure route. B.9.21.4. Exposure via consumer products and the indoor environment PFASs are present in a wide range of consumer products in the indoor environment. From these products, non-volatile PFASs can be released by abrasion, whereas volatile and semivolatile PFASs can be released by volatilisation (Winkens et al., 2018). Released PFAS partition to dust and air, which constitute potential exposure routes to individuals residing the indoor environment. Consumers may also be exposed to PFASs directly from products via e.g. inhalation, dermal uptake or hand-to-mouth exposure. Impregnation products for floors, carpets and furniture have been hypothesised to be especially important for elevated indoor levels of PFASs (Beesoon et al., 2012; Knobeloch et al., 2012; Kubwabo et al., 2005; Zhou et al., 2022). The hypothesis that consumer products are the origin to PFASs in indoor environments is supported by the composition of PFASs in dust and air, which shows that precursors used in consumer products dominate the concentration profiles (see Appendix B.9.21. Table B.111 and Table B.112). In indoor dust, n:2 monoPAPs/diPAPs (precursors to PFCAs) are found at the highest levels (De Silva et al., 2012; Eriksson and Kärrman, 2015; Winkens et al., 2018) . Other precursors (e.g. FOSAs, FOSEs and FTOHs) and arrowhead PFAAs (mainly PFOS and PFOA) are found at varying but largely comparable levels. Among the PFAEs, 6:2 Cl-PFESA have been detected in indoor dust from China (Xu et al., 2021a; Zhang et al., 2020a), but only at a low detection frequency (2%) in Finnish dust samples (Winkens et al., 2018). ADONA and HFPO-DA were not detected in any dust sample in these studies (Winkens et al., 2018; Zhang et al., 2020a). Significant correlations between PFAA levels in household dust and in resident’s serum, which could indicate that dust can be a relevant exposure route, have been reported but not consistently found in all studies (Haug et al., 2011a; Kim et al., 327 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) 2019a; Poothong et al., 2020; Wu et al., 2015). Although the relative contribution from dust to the total exposure is generally regarded to be low (but highly variable ) in adults, this exposure route can be important for small children, due to their higher absolute intake of dust (EPA-US, 2011). In a recent review, the median contribution from dust was estimated to account for 25%, 13%, 7% and 3% of the total internal levels of PFHxS, PFOA, PFNA and PFOS, respectively (DeLuca et al., 2022). Indoor air generally contains higher concentrations of PFASs compared to outdoor air (Yao et al., 2018). FTOHs dominate the concentration profile in indoor air, followed by other precursors, such as FOSAs, FOSEs and FTAs. Partitioning of these compounds to air is expected due to their low water solubility and high vapor pressure (Jian et al., 2017; Winkens et al., 2018). Biotransformation of inhaled precursors add to the indirect exposure to arrowhead PFASs. Levels of ionic PFASs, such as PFOS and PFOA, in indoor air are generally low and inhalation of indoor air is not regarded to be a large direct exposure route to these substances. In a recent review, indoor air generally contributed with 2-4% of individual PFAA serum levels (DeLuca et al., 2022). Regarding exposure to PFASs from the use of consumer products, the very limited information found (Vestergren et al., 2008; Washburn et al., 2005), related to the already regulated substances PFOS and PFOA, estimate the contribution to systemic exposure from consumer products to be low. Washburn et al. (2005) estimated that the use of consumer products containing PFOA would not lead to detectable levels of PFOA in serum (<0.5 ng/mL). However, acute inhalation toxicity has been reported from use of impregnation sprays, indicating that high short-term exposure can occur (summarised in ECHA (2017a)). Little is known about the exposure to PFASs from cosmetics or personal care products, but it has been suggested to be of relevance, in particular for children (Winkens et al., 2017b). The very limited available data on exposure to fluorinated gases in consumer products suggests that the consumer exposure is very low (EPA-DK, 2015b). For more information on consumer exposure related to different use sectors, see the following sections: B.9.3.3, B.9.4.3, B.9.6.3, B.9.7.3, B.9.8.3, B.9.10.3, B.9.11.3, B.9.12.3, B.9.13.3, B.9.14.3, B.9.15.3, B.9.16.3. B.9.21.5. Occupational exposure For industrial and professional workers, the primary exposure routes to PFASs are inhalation, ingestion of dust and dermal uptake (De Silva et al., 2021; Fu et al., 2015) during PFASs production, product manufacturing as well as professional use and disposal of PFAS-containing products (Lohmann et al., 2020). Workers at several fluorochemical and fluoropolymer production facilities have been shown to have elevated levels of PFAAs and related precursors in their blood (Gao et al., 2018; Gebbink and van Leeuwen, 2020; Girardi and Merler, 2019; Olsen and Zobel, 2007). For example, at the Rimar-Miteni factory in the Italian Veneto region, where PFOA had been produced since 1968, workers displayed heavily elevated levels of PFOA in serum (geometric mean: 4 048 ng/mL; range 19–91 900 ng/mL) when sampled in the period 2000–2013 (Girardi and Merler, 2019). Workers at facilities manufacturing PFAScontaining products, such as textiles, may also be highly exposed to (precursors to) PFAAs (Heydebreck et al., 2016). In addition, high serum levels of Cl-PFESAs have been reported in Chinese metal plating workers, due to the use of mist suppressant F -53B (containing ClPFESAs) in this particular industry (Shi et al., 2016). Groups of professional workers that may be exposed to PFASs are for example firefighters (Rotander et al., 2015b), ski waxers using old methods (see B.9.8.3)(Freberg et al., 2010; Nilsson et al., 2010a), waste recycling workers (Peng et al., 2022) and workers who 328 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) frequently use impregnation sprays (summarised in ECHA (2017a)). For case studies of acute and long-term occupational inhalation exposures and health effects, see B.5.4. Occupational exposures to fluorinated gases in plant manufacturing equipment or when putting the gases into products should be small, given the need to use closed systems (mentioned by many respondents to the CfE). Outside of the manufacturing plant, however, there may be greater exposure of workers, for example during service and at sites reclaiming refrigeration equipment at the end of its service life in case of accidental releases (EPA-DK, 2015b). According to a review by Tsai (2005b) exposure to HFCs could occur via inhalation during accidental spills or leakages from the refrigeration system, recycling system for electronic appliances, degreasing process for precision cleaning, gas delivery pipeline of the semiconductor manufacturing or medical delivery systems for the treatment of asthma. Short-term moderate exposure to HFCs has been reported during refrigeration repair work (Gjølstad et al., 2003). In addition, several studies have demonstrated that health care workers and veterinarians may be exposed to fluorinated anesthetics (e.g. isoflurane, sevoflurane) via inhalation during surgical operations (Saber and Hougaard, 2009; Scapellato et al., 2014). For more information on worker exposure related to different use sectors, see B.9.5.3, B.9.8.3, B.9.10.3, B.9.11.3, B.9.12.3, B.9.13.3, B.9.14.3, B.9.15.3, B.9.16.3. 329 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.22. Combined human exposure assessment Blood serum and plasma are the most common matrices for PFAS analyses in humans. To a lesser extent, PFASs have been measured in other human matrices, such as whole blood, breast milk, placenta, urine, hair and nails, etc. For most of the studied PFASs, the levels in matched plasma and serum are comparable, whereas the levels in whole blood are generally half of the corresponding serum or plasma levels (Ehresman et al., 2007; Jin et al., 2020b; Poothong et al., 2017). In contrast to other studied PFASs, PFHxA and PFOSA are detected in much higher levels in whole blood compared to serum or plasma (Poothong et al., 2017). Consequently, whole blood is deemed to be the most suitable blood matrix for analysis of PFHxA and PFOSA. For most PFASs, the ratios between serum, plasma and whole blood are not known and should therefore not be presumed. There are several reviews and reports presenting compilations of PFAS levels in human matrices, mostly focusing on PFAAs in blood (e.g. ECHA (2018a); EFSA (2020)). Several of the most frequently analysed PFAAs and their respective related substances are covered by the Stockholm Convention on POPs or by existing or proposed REACH restrictions (PFOS, PFOA, PFHxA, PFHxS, C9-C14 PFCAs), whereas other non-restricted PFASs are often omitted from the compilations. Therefore, the Dossier Submitters have done a nonexhaustive compilation of non-restricted PFASs measured in different human matrices. This compilation also includes studies of total extractable organofluorine and non-target and suspect screening studies. Summaries of these studies are tabulated further down in this chapter and in Appendix B.5.1.1.2. Table B.91 and Table B.95. B.9.22.1. Unknown organic fluorine in humans This restriction proposal covers known PFASs, but also PFASs of unknown identity. Thus, unspecific methods, such as extractable organic fluorine (EOF) analysis, are needed to estimate the total exposure to organofluorine compounds. Part of these unknown compounds will be currently non-restricted PFASs, but a fraction of these unknown compounds could be precursors to already restricted PFASs. In addition, EOF may contain organofluorine compounds originating from medicinal products, biocidal products or plant protection products. More information about the analytical techniques is given in E.7. Total organic fluorine measurements in environmental matrices are summarised in B.4.2.6.1. The Dossier Submitters literature search found eight studies of EOF in human matrices ( Table B.61). For the EOF analysis, all except one study used the combustion ion chromatography (CIC) method, which has not been standardized to date. However, in a Swedish interlaboratory comparison of water and sludge samples, the EOF-CIC method showed good accuracy and reasonable interlaboratory precision (KEMI, 2021a). Based on fluorine mass balance analysis 55, the proportion of unidentified organic fluorine (UOF) to the total EOF varied substantially between the studies, ranging from 0 up to 97% ( Table B.61). The differences in the proportion of UOF between the studies can be explained by variations in the exposure of the studied populations but also by the study design, including the choice of reported summary statistics, matrix, extraction method, detection limit and the number of PFASs included in the target analysis (i.e. the known organofluorine fraction). Thus, comparisons between different studies should be made with caution. Furthermore, the reported EOF levels are often close to their respective LOD/LOQ, which entails uncertainties in the quantification of EOF and estimated UOF. 55 C omparison of the total EOF concentration to the amount of fluorine from the identified PFASs (analysed with targeted methods) in the sample. 330 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.61. Biomonitoring studies of extractable organofluorine (EOF). DF=detection frequency, CIC=combustion ion chromatography, CNAA=cyclic neutron activation analysis, WB=whole blood, S=serum, P=plasma, N PFAS=the number of individual PFASs constituting the known fraction. Country, Matrix N DF Mean EOF % UOF Method LOD/ Reference year (n) PFAS EOF (ng F/mL or LOQ ng F/g) Sweden, 2018 Sweden, 2014 2015 Austria, 2017-19 WB 63 (147) WB Exposed 63 (20) WB Reference 63 (9) Maternal S 61 (21 pools) 88% 9.0a 54 C IC 95% 234a 36 C IC 100% 24.8 84 C IC 3 (mean) (Aro et al., 2021c) 7.1-107 (Aro et al., 2022) 7.1 29% 3.8b 24 C IC 2.7 (Kaiser et al., 2021b) C ord S (11 pools) 61 64% 1.6b 9 C IC 1.1 Placenta (13 pools) 61 46% 1.3b 51 C IC 0.92 C hina, 2009 S (60) 11 100% 210 87 C NAA 20 (Liu et al., 2020a) Sweden, 1996-2017 S (57 pools) 61 33% 41b 60 C IC 25 (Miaz et al., 2020) USA, 2001 P (4) 22 4 of 4 18-59 (range) 0-15 C IC 6 (Miyake et al., 2007b) Japan, 2003 WB (3) 22 1 of 3 <6-8.9 (range) 16 C IC 6 Japan, 2004 WB,S Worker 22 (2) 2 of 2 C IC 30 C hina, 2004 WB (47) 52 >70% 424-505 0-3 (WB, range) 1020-1070 0 (S, range) ND-94 14-69 (range) C IC 4 Germany (Muster), 1982-2009 P (80) 52 100% 24 0-48 C IC 4 Germany (Halle), 1995-2009 C hina, 2004 P (42) 52 100% 16 0-43 C IC 4 WB (30) 13 N.R. <6-43c 15-66d C IC 6 a) b) c) d) (Yeung and Mabury, 2016) (Yeung et al., 2008) Values LOD 331 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Aro et al. (2021c) performed a fluorine mass balance analysis of 147 whole blood samples from residents in five Swedish municipalities. The average EOF concentration was 9.0 ng F/mL (range 0.59-57 ng F/mL) and the UOF accounted for 54% of the total EOF. In a second study, the same research group analysed 20 whole blood samples from re sidents in Ronneby (Sweden) who had been exposed to drinking water contaminated with firefighting foam (Aro et al., 2022). The exposure had ceased less than one year prior to the sampling. The exposed population had an average EOF concentration of 234 ng F/mL (range <116-592 ng F/mL), which was more than twenty times higher than in the general population. In addition, the average level of UOF was higher in the exposed population (79 ng F/mL) compared to the general population (6.2 ng F/mL). Although the levels of UOF was higher in the exposed population, this population had a lower fraction (36%) of UOF in blood compared to the general population (54%). In another Swedish study, fluorine mass balance analysis was applied in serum samples from first-time mothers, collected in yearly sampling campaigns between 1996 and 2017 (Miaz et al., 2020). Some of these women had been exposed t o PFAS contaminated drinking water until the year 2012. Only 33% of the samples had detectable EOF levels, which could be due to a relatively high detection limit (25 ng F/mL). Over the entire period, detectable levels of EOF ranged from 26 to 74 ng F/mL and UOF accounted for 25-89% (mean 60%) of the total EOF. The concentration of EOF did not significantly decrease over time, whereas the sum of identified PFASs declined significantly, probably reflecting the known declining trend for some legacy PFAAs (see section B.9.22.5). Consequently, the proportion of UOF to the EOF increased by approximately 4% per year. However, the time trend should be interpreted with caution as it was based on only 19 samples. Yeung and Mabury (2016) measured EOF in 122 plasma samples collected in the German cities Halle and Münster between the years 1982 and 2009. The fraction of UOF accounted for 0-48% of the EOF. The levels of identified PFASs decreased significantly over time, whereas there was no significant trend in EOF concentrations. After the year 2000, the amount and proportion of UOF increased in the Münster samples, whereas no trend was observed in the Halle samples. Kaiser et al. (2021a) measured EOF in maternal serum, placental tissue, and cord serum from Austrian women. On average, 42% of the samples showed detectable EOF levels. The mean EOF concentration was 3.8 ng F/mL in maternal serum, 1.3 ng F/g in placenta and 1.6 ng F/mL in cord serum, which is relatively low compared to other studies. The fraction of UOF was highest in placenta (51%), followed by maternal serum (24%) and cord serum (9%). In addition to studies in the European population, fluorine mass balance analysis has been applied to human blood samples from China, Japan and USA and in nails and hair from China (Liu et al., 2020a; Miyake et al., 2007b; Yeung and Mabury, 2016; Yeung et al., 2008). The levels of EOF in these studies are generally comparable to the levels reported for Europeans except for a Chinese study that reported significantly higher levels (Liu et al., 2020a). However, the latter study analysed EOF with the CNAA method, for which the comparability to the CIC method is uncertain. Another method to analyse unknown PFASs is the total oxidizable precursor assay (TOPA), which estimates the amount of unknown precursors in a sample. This method has been successfully applied to e.g. water, soil and biota (see B.4.2.6.2). However, to the best of the Dossier Submitters knowledge, TOPA has not been applied to human matrices to date. In conclusion, studies of European populations demonstrate that a considerable fraction of the extractable organofluorine detected in human samples is not explained by the PFASs that are routinely analysed in target analyses. Consequently, humans are exposed to a considerable amount of PFASs with unknown identity and regulatory status. Furthermore, limited data sets suggest that the proportion of unknown PFASs in relation to known PFASs 332 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) in humans increases over time. B.9.22.2. Targeted analyses of PFAAs in humans The EFSA risk assessment of PFASs in food (EFSA, 2020) included a comprehensive compilation of European biomonitoring studies conducted between the years 2007 and 2016. In these studies, the most frequently analysed and most abundant PFAAs in serum or plasma from adults were PFOS, PFOA, PFHxS and PFNA, contributing with 64%, 16%, 5.6% and 5.1% to the sum of PFASs, respectively. The median serum or plasma levels of these and other PFASs that are already covered by (proposed) restrictions are presented in Table B.62 and can serve as reference for comparison to the levels of non-restricted PFAAs discussed below (Table B.63) and in Appendix B.5.1.1.2. Table B.91 to Table B.92. The levels of precursors (i.e. FOSA, 8:2 monoPAP, 8:2 diPAP, EtFOSA, EtFOSAA and PFOSI) were found in low levels compared to the PFAAs in Table B.62. Table B.62. Median levels (ng/mL) of restricted (or proposed for restriction) PFAAs in plasma or serum of European adults (EFSA, 2020). PFHxA PFHxA PFNA PFDA PFUnDA PFDoDA PFTrDA PFHxS PFOS PFOA (WBb ) Median of 0.03 medians 0.62 1.9 0.61 0.30 0.28 0.05 - 0.67 7.7 Range of medians 0.620.62 0.764.9 0.302.6 0.071.3 0.062.5 0.030.54 - 0.20152 1.727 0.010.045 Number of 2 1 32 37 33 22 7 0 37 32 studies a For studies with median concentration lower than LOQ, the LOQ was used for calculations. a) Only studies where median is > LOQ are included in this number. b) WB = whole blood. Within the recently finalized European Human Biomonitoring Initiative (HBM4EU), PFASs were measured in serum or plasma of 1957 teenagers from 9 European countries. Overall, 14.3% of the teenagers exceeded the guidance value of 6.9 µg/L for the sum of PFOA, PFOS, PFNA and PFHxS based on the EFSA tolerable weekly intake (EFSA, 2020; VITO, 2022). The reported detection frequencies of the non-restricted PFSAs and PFCAs vary substantially between studies (Table B.63). In addition to the possibility of varying exposure patterns between European subpopulations, the variability of detection frequencies can depend on the magnitude of the reporting limits, which complicates the comparability of the studies. Blood levels of non-restricted PFAAs in the European population are generally lower than the levels of e.g. PFOS, PFOA, PFHxS and PFNA. PFHpS is generally the most frequently detected and most abundant non-restricted PFAA in European blood samples, with median levels in the range of 0.04-0.5 ng/mL (Table B.63). Detection frequencies of PFBA, PFPrA, PFBA, PFPeA, PFHpA, PFHxDA, PFEtS, PFPrS, PFBS, PFPeS PFNS, and PFDS are varying but generally moderate to low, whereas the levels of PFOcDA, PFUnDS and PFDoDS were below the reporting limits in all studies included in the Dossier Submitters compilation. A PFAA that has rarely been included in target analysis is TFA. In Swedish studies of whole blood, TFA was found in 62% of the samples from the general population (Aro et al., 2021c) and in 1 out of 20 samples from consumers of contaminated drinking water (Aro et al., 2020b). In these two studies, the levels of TFA in blood were not reported due to low recovery. In a Chinese study, TFA was detected in 97% of the analysed serum samples with a median level of 8.6 ng/mL, which accounted for 17% of the sum of the 21 PFASs analysed in the populatio (Duan et al., 2020). PFPAs and PFPiAs have been analysed in blood matrices in a few European studies. In two Norwegian studies, PFHxPA was detected in 52-100% of plasma, serum and whole 333 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) blood samples from adults (Poothong et al., 2017; Thépaut et al., 2021). On the contrary, PFHxPA was not detected in Austrian or Swedish studies (Aro et al., 2021c; Aro et al., 2020b; Kaiser et al., 2021a). The remaining analysed PFPAs (PFOPA and PFDPA) and PFPiAs (C6/C6 PFPiA) were generally not found in these studies (Aro et al., 2021c; Aro et al., 2020b; Kaiser et al., 2021a; Poothong et al., 2017; Thépaut et al., 2021). Fetuses are exposed to PFASs in their mothers’ blood via placental transfer but measured levels in cord blood are generally lower than in matched maternal blood (EFSA, 2020). In human breast milk, PFOA and PFOS are the most abundant PFAAs (Awad et al., 2020; Cariou et al., 2015; Karrman et al., 2007; Lee et al., 2018b). Detectable levels of other long-chain PFCAs, PFHxS and PFHxA are also reported in some studies, whereas the levels of other PFAAs generally are very low or below the reporting limits. Breast milk concentrations are considerably lower than serum concentrations, with milk:serum ratios in the range of 0.01-0.12 for different PFAAs (PFOS, PFOA, PFHxS, PFNA, PFDA, PFUnDA and FOSA) (Cariou et al., 2015; EFSA, 2020; Karrman et al., 2007; Liu et al., 2011). Although the levels in breast milk are low compared to blood, breast milk is the major exposure source to PFASs in nursing children (Haug et al., 2011a) and it is therefore important to keep these levels as low as possible. In addition to blood matrices, urine has been analysed in several studies, which show generally lower levels of PFAAs in urine compared to serum (Kato et al., 2018; Wang et al., 2018c; Xu et al., 2020c). The urine:serum ratios are generally highest for short -chain PFAAs, which illustrates their faster renal clearance due to their relatively higher watersolubility and shorter biological half-lives (EFSA, 2020). There are some studies reporting higher detection rates and/or levels of PFPeA, PFBA and PFHpA in urine compared to matched serum (Kato et al., 2018; Kim et al., 2019b; Kim et al., 2014; Zhou et al., 2014). These studies indicate that urine may be a relevant matrix for detecting exposure to certain short-chained PFASs that may otherwise be overlooked. PFAA levels in hair and nails have been analysed in several studies, however the relevance of these non-invasive methods for assessing the internal exposure remains unclear, for example due to the risk of contamination from the surrounding environment (Kim et al., 2019b; Kim and Oh, 2017; Liu et al., 2020a; Martin et al., 2019; Piva et al., 2021; Wang et al., 2018b). Concentrations of PFASs in internal organs and other matrices (liver, follicular fluid, cerebrospinal fluid, semen, etc) are tabulated in Appendix B.5.1.1.2. Table B.94 and Table B.95 and discussed in section B.5.1.1.2. 334 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.63. Detection frequencies of PFAAs (not covered by existing or proposed restrictions) in human blood samples from the general and highly exposed populations in Europe. S=serum, P=plasma, WB=whole blood. For details and full references, see Appendix B.5.1.1.2. Table B.91. PFHxD PFOcD PFUnD PFDoD Country Samples Year N TFA PFPrA PFBA PFPeA PFHpA PFEtS PFPrS PFBS PFPeS PFHpS PFNS PFDS A A S S GENERAL EUROPEAN POPULATIONS 15-19 years, S Faroe I France 201011 2010Maternal, S 13 13 years, S 2011 Adult, S 2011 Sweden Adult, S Faroe I 5 years, S Norway Belgium Adult, S Adult, S Norway Adult, S Sweden 2013Maternal, S 17 15 Pool 11-12 years, S 2014 200 201416 744 3-17 years, 2014Germany P 17 1 109 UK Belgium Norway France Sweden France Adult, S 201114 2012 201215 2014 201314 940 75% 3% 0% 98% 14% 0% 100 0% 0% 0% 1% 50% 51 277 0% 0% 0% 0% 96% 0.4% 0% 0% 100% 7% 0% 12% 0% 18% 0% 92% 33% 45% 37% 3% 0% 13% 51% 0% 13% 0% 0% 66% 0% 79% 0% 0% 2% 1% 579 51 63% 4% 0% 158 205 61 13% 1% 0% 3% 0% 0% 1% 0% 0% 96% 100% 53% 0% 0% 77% 0% 0% Adult, S Adult, S 2015 2015 59 242 0% 49% 0% 17% 21% 7% 0% 0% C zech R. Adult, S 2015 201516 2016 2016 300 0% 0% 45% 20% 2% 250 51 158 17% 40% 66% 73% 18% Italy Adult, S Sweden Adult, S Germany S 0% 0% 335 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Country Samples Belgium 14-15 years, S Sweden Adult, S Germany Year 201718 201719 200919 N TFA PFPrA PFBA 410 110 100 Italy Adult, P 6-11 years, S 2016 61 Norway Adult, S 201617 142 Sweden Adolescents 2016,S 17 1 098 PFPeA PFHpA 0% 0% 0% 1% 0% 4% 5% 0% 0% 0% 6% 0% 22% 3% 8% 0% 50% 0% 19% 59% 68% 100% 0.5% 0.4% 0% 3% 0% 2% 67% 1% 9% 0% 49% 1% 90% 27% 44% 4% 17% 28% 101 0% 13% Sweden S Adult, S 201516 Italy 5% 0% 90% 2.8% 9.5% 0% 0% 76% 49% 100% 99% 0% 0% 60% 0% 0% 0% 0% 5% 0% 100% 100% 901 257 26% 0% Germany Adult, P 20 2% 0% 33% 4% Sweden 0% 4% 0% 0% 13 0% 0% 10 90 164 Adult, S Adult, WB 43% 0% Faroe I Adult, S 2006 Germany C hild, P 2006 Germany Maternal, P 2006 Norway PFUnD PFDoD S S 1% Germany Adult, S 2006 200809 2014 201415 PFDS PFPeS PFHpS PFNS 0% 11% 2017 72 2017- 21 Austria Maternal, S 19 Pool C hild&Adult Sweden 2018 148 62% 22% WB 2018C zech R. Adult, S 19 395 2019C zech R. Adult, S 20 242 HIGHLY EXPOSED POPULATIONS IN EUROPE PFHxD PFOcD PFEtS PFPrS PFBS A A 5% 0% 15% 0% 5% 100% 100% 100% 0% 60% 0% 0% 0% 56% 7% 26% 51% 31% 336 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Country Samples Year C hild&Adult, 2017S 19 C hild&Adult, Germany S 2018 Sweden Adult, S 2018 Italy Belgium C hild&Adult 2021 S N 18 122 TFA PFPrA PFBA 0 PFPeA PFHpA 0.1% 906 26 796 5% 0% PFHxD PFOcD PFEtS PFPrS PFBS A A 1% 3% 100% 16% 81% 11% 3% PFPeS PFHpS PFNS 100% PFDS PFUnD PFDoD S S 100% 30% 337 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) B.9.22.3. Targeted analyses of PFAEs and cyclic PFASs in humans According to the Dossier Submitters literature search, the most studied PFAEs to date are 6:2 Cl-PFESA, ADONA, and HFPO-DA (Table B.64). In addition, the cyclic PFECHS is sometimes included in target analyses of human matrices. 6:2 Cl-PFESA has been analysed in populations from Sweden, Germany and Austria. These studies all indicate a low exposure, with detection frequencies ranging from 0 to 12% (Aro et al., 2021c; Aro et al., 2020b; Awad et al., 2020; Gockener et al., 2020a; Gyllenhammar et al., 2020; Kaiser et al., 2021a; Miaz et al., 2020; Nystrom et al., 2022) . In contrast to European populations, several biomonitoring studies show a wide exposure to Cl-PFESAs in the Chinese population. In these studies, 6:2 Cl-PFESA has been found in close to all analysed blood samples from Chinese adults at median levels ranging from 1.5 to 8.6 ng/mL (Chen et al., 2017b; Duan et al., 2020; Jin et al., 2020b; Kang et al., 2020; Li et al., 2020c; Yu et al., 2021c). At these levels, 6:2 Cl-PFESA is generally the third most abundant PFAS in blood, following PFOA and PFOS. In addition to blood matrices, Cl-PFESAs have been detected in placenta, breast milk, follicular fluid, semen, urine, hair and cerebrospinal fluid from the Chinese population (Awad et al., 2020; Chen et al., 2017b; Kang et al., 2020; Lu et al., 2021; Pan et al., 2019a; Wang et al., 2018a; Wang et al., 2018b). Metal plating workers and high consumers of freshwater fish have been identified as highly exposed groups, showing median serum levels of 51 ng/mL and 94 ng/mL, respectively (Shi et al., 2016). Although the levels of Cl-PFESAs are currently low in Europe, the available data from China demonstrate that these compounds can reach high internal levels in several human matrices. ADONA has been measured in European studies from Sweden, Germany, Austria and the Czech Republic. These studies report non-detectable or low (≤0.1 ng/mL) levels of ADONA in serum, whole blood, placenta, and breast milk (Aro et al., 2020a; Awad et al., 2020; Cerna et al., 2020; Fromme et al., 2018; Gockener et al., 2020b; Gyllenhammar et al., 2020; Kaiser et al., 2021a; Livsmedelsverket Naturvårdsverket, 2020; Menzel et al., 2021; Miaz et al., 2020; Wang et al., 2018a). In a German study of adult populations living close to or 80 km away from a production plant where ADONA has been used in the production of fluoropolymers, ADONA was found in 9% and 16% of the plasma samples, respective ly, but in none of the samples from a reference population (Fromme et al., 2017). Regarding nonEuropean populations, ADONA has not been detected in blood or urine form the US (Calafat et al., 2019; Kato et al., 2018) and in low or non-detectable levels in blood serum and breast milk from China (Awad et al., 2020; Duan et al., 2020). HFPO-DA has not been detected in any whole blood or serum sample from Sweden or Germany (Aro et al., 2021c; Aro et al., 2020a; Gockener et al., 2020b; Gyllenhammar et al., 2020; Miaz et al., 2020) or in breast milk from the Czech Republic (Cerna et al., 2020). In accordance with these European studies, HFPO-DA has not been detected in blood or urine from North American populations without known elevated exposure to PFASs (Kato et al., 2018; Kotlarz et al., 2020; Mottaleb et al., 2020), with exception of a detection frequency of 1.2% in urinary samples (N=2 682) analysed within the US NHANES program (Calafat et al., 2019). PFECHS has been found in Swedish, Austrian and German studies with detection frequencies of 0-84% and at low median levels ranging from 0.002 to 0.02 ng/mL (Aro et al., 2021c; Aro et al., 2020a; Gockener et al., 2020b; Kaiser et al., 2021a). 338 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.64. Findings of 6:2 Cl-PFESA, ADONA, HFPO-DA and PFECHS in serum presented with detection frequencies (%), median concentrations (bold) and ranges (ng/m L). A detailed summary is found in Appendix B.5.1.1.2. Table B.92. Year, country N 6:2 Cl-PFESA ADONA HFPO-DA 110 FC P (Fromme et al., 2017) 16% (<0.2-0.7) 2014, 200 2016 Ref Germany 0% (<0.2) 2014 Sweden, WB 0% 20 (<0.01) C DW 0% (<0.007/<0.02) 0% (<0.022) 19962017 Sweden 30 P 0% 0% 10% 2016 158 Germany Ref 0% (<0.25) 2018 906 Germany C DW 0.3% (<0.25-1.0) 2009100 2019 Germany 20162017 Sweden 1098 20172019 Sweden 110 20172019, Austria 21 P 20172019, Austria, 11 P 15% (Aro et al., (<0.01-0.06) 2020b) (Miaz et al., 2019) (Fromme et al., 2018) 0% (<0.25) 2017 72 Germany 148 Reference 9% (<0.2-14) 2009, 86 2015 FC P Germany 2018 Sweden, WB PFECHS (Menzel et al., 2021) 12% 16% 0% (<0.01-0.02) (<0.007/<0.020.04) (<0.022) 80% 0.02 (Aro et al., (<0.01-0.11) 2021c) 0% (<0.25) 0% (<0.25) 0% (<0.25) 0% (<0.25) 5.4% (<0.005/<0.06-1.2) 0.1% (<0.02/<3.6-0.07) 1% (<0.08-0.2) 0% (<0.08) 0% 52% 0.006 (<0.0006/<0.010.074) 84% 0.003 (C9) vary between studies (Gockener et al., 2020b; Land et al., 2018; Miaz et al., 2020). Levels of non-restricted PFCAs and PFSAs are often close to the reporting limits and the data is therefore less robust for time trend analyses. Furthermore, differences in e.g. study period, location and reporting limits make the studies difficult to reliably compare. Thus, no firm conclusions on time trends can be drawn for these compounds. To the best of the Dossier Submitters knowledge, there are no European time trend studies of other non-restricted PFASs, except for one study of PFECHS, which showed declining levels in Swedish mothers between the years 2000 and 2017. This is in line with the 3M phase -out of PFECHS in the year 2002, concurrent with other perfluorooctyl-based substances (Miaz et al., 2020). Time trends of PFASs in the environment are discussed under B.4.2.7.9. B.9.22.6. Non-target and suspect screening of PFASs in humans Non-target and suspect screening analyses have been used to tentatively identify previously unanalysed PFASs in Australian firefighters (Rotander et al., 2015a), Chinese women (Kang et al., 2020; Li et al., 2020c; Wang et al., 2021a), Swedish pregnant women (Miaz et al., 2020) as well as American populations of fishermen (Baygi et al., 2021), firefighters (Grashow et al., 2020), pregnant women (Gerona et al., 2018), mothers and newborns (Wang et al., 2021a) and consumers of contaminat ed drinking water (McDonough et al., 2021b). Thus, available screening studies have mostly been applied in non-European populations and several of them have focused on highly exposed individuals. Non-target or suspect screening studies of the general European population are scarce. The confidence level of the identification of each compound is indicated by the Schymanski scale, where 1 is the highest and 5 is the lowest level of confidence (Schymanski et al., 2014). Non-regulated PFASs identified with confidence levels 2 or 3 include PFESAs, PFECAs, ketonePFSAs, unsaturated perfluorinated alcohols and unsaturated PFSAs. These results indicate that humans are exposed to non-regulated PFASs that have not been reliably identified or quantified to date. The reported semi-quantification of the tentatively identified PFASs have inherent uncertainties. Thus, it is not possible to estimate the levels of these compounds in relation to the unknown fraction of EOF in blood. However, these studies do not consistently indicate that there are any new dominating PFASs that could close the gap of between the fraction of known PFASs and the total EOF in human blood samples. In addition to human samples, non-target and suspect screening methods have been used to identify PFASs in various environmental compartments, which is described in chapter B.4.2.6.3. B.9.22.7. Summary of combined human exposure assessment Fluorine mass balance analyses of human blood show substantial variation in the proportion of unknown organofluorine to the total concentration of EOF reported in different studies (097%). Thus, humans are exposed to a highly variable but considerable amount of PFASs for which the identity and consequently the regulatory status are unknown. Attempts to 343 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) characterise the unknown PFAS fraction have been performed with non-target and suspect screening methods. These studies suggest that for example PFESAs, PFECAs, ketone-PFSAs, unsaturated perfluorinated alcohols and unsaturated PFSAs may partly constitute the unknown fraction. However, the gap between the known and the total EOF concentration remains unexplained. Human biomonitoring studies of known PFASs unambiguously demonstrate world -wide exposure to a wide range of PFASs, with especially high exposure levels in populations living in areas close to PFAS point sources as well as in occupationally exposed individuals. In the general European population, PFOS, PFOA, PFHxS and PFNA are the most studied and most abundant PFAAs. In a considerable part of the population, the serum levels of these four PFAAs exceed the tolerably weekly intake. The high detection frequencies for PFASs that have been phased out show that the historic exposure takes a long time to reverse. Non-regulated PFAAs are found at lower levels and generally at moderate to low detection frequencies. However, it is evident that the European population is exposed to non-regulated PFAAs. Furthermore, the few studies of TFA in human matrices indicate that part of the European population have detectable levels of TFA in blood. Blood levels of PFAEs (ADONA, HFPO-DA) are low in the general European population. However, in the US and China, several PFAEs have been detected in blood from populations living close to fluorochemical manufacturing facilities. These studies indicate that a potential increased use of PFAEs as substitutes for legacy PFASs could lead to increased human exposure also in Europe. In addition, studies from China show a widespread human exposure to 6:2 Cl-PFESA, which is generally the third most abundant PFAS in blood of the Chinese population. European biomonitoring studies indicate low current exposure to Cl-PFESAs in the general population. However, the available Chinese data demonstrate that these compounds can reach high internal levels in several human matrices. 344 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendices to Annex B Appendix B.1.2. Physicochemical Properties Table B.66. Basic substance information and physical chemical properties of PFCAs (Perfluoroalkylcarboxylic acids). C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA C10C11C12 abbrevi C2 -PFCA PFCA PFCA PFCA a-tion acrony TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA PFDA PFUnDA PFDoDA m IUPAC trifluoroa butanoic pentanoi hexanoic heptanoic octanoic nonanoic decanoic undecano dodecano name cetic acid acid, c acid, acid, acid, acid, acid, acid, ic acid, ic acid, heptafluo nonafluo undecaflu tridecaflu pentadec heptadec nonadeca henicosa- tricosaflu roroorooroa-fluoroa-fluoro-fluorofluorooromolecul C F3C F3(C F2)2 C F3(C F2)3 C F3(C F2)4 C F3(C F2)5 C F3(C F2)6 C F3(C F2)7 C F3(C F2)8 C F3(C F2)9 C F3(C F2)1 0-C OOH -C OOH -C OOH -C OOH -C OOH -C OOH -C OOH -C OOH -C OOH ar C OOH formula C AS number 76-05-1 375-22-4 physico-chemical data molecul 114.02 214.04 ar weight g/mol partitio ning coeffici ent log K OW 0.79 ± 0.48 at 25 °C (calculate d with QSAR; REAC H registrati on data (2021- 3.39 ± 0.60 at 25 °C (calculate d using Advanced C hemistr y Developm ent (AC D/Lab s) C13 PFCA PFTrDA C14 PFCA PFTeDA tridecanoi c acid, pentacos a-fluoroC F3(C F2)1 1-C OOH tetradeca noic acid, heptacos a-fluoroC F3(C F2)1 2-C OOH 2706-903 307-24-4 375-85-9 335-67-1 375-95-1 335-76-2 2058-948 307-55-1 7262994-8 376-06-7 264.05 314.05 364.06 414.07 464.08 514.08 564.09 614.10 664.11 714.11 3.43 (calc. using C OSMOt herm (Arp et al., 2006) 3.40 (Predicte d using US EPA 4.06 (calc., C OSMOth erm (temp. not specified) (Wang et al., 2011c) 4.13 (exp. 3.82 (calc. using C OSMOth erm (Arp et al., 2006) 5.33 (Predicte d using US EPA EPI-Suite 5.30 (calc., C OSMOth erm (temp. not specified) (Wang et al., 2011c) 4.30 (calc. 5.9 (calc., C OSMOth erm, (Wang et al., 2011c)) 7.27 (Predicte d using US EPA EPI-Suite 6.5 (calc., C OSMOth erm, (Wang et al., 2011c)) 7.667 (exp. value, MSDS LabNetwo 7.2 (calc., C OSMOth erm, (Wang et al., 2011c)) 8.548 (exp. value, MSDS LabNetwo 7.8 (calc., C OSMOth erm, (Wang et al., 2011c)) 9.429 (exp. value, MSDS LabNetwo 8.25 (calc., C OSMOth erm, (Wang et al., 2011c)) 8.90 (calc., C OSMOth erm, (Wang et al., 2011c)) 11.191 (exp. value, MSDS LabNetwo 345 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m log K OA 56 C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA 05-31))56 0.50 (Predicte d using US EPA EPI-Suite (KOWWI N v1.67)) Software V11.02) 2.43 (Predicte d using US EPA EPI-Suite (KOWWIN v1.67)) EPI-Suite (KOWWI N v1.67)) (KOWWI N v1.67)) using C OSMOth erm (Arp et al., 2006) 5.843 at 25 °C (Estimate from Log Kow [0,50 (KowWin estimate) ] and log 4.743 at 25 °C (Estimate from Log Kow [2.43 (KowWin estimate) ] and log 4.992 at 25 °C (Estimat e from Log Kow [3.40 (KowWin estimate )] and value, MSDS LabNetw ork) [Royal Society of C hemistr y, C hemSpi der database ]57 3.26 (calc. using C OSMOth erm (Arp et al., 2006) 6.63 (calc., C OSMOth erm (Wang et al., 2011c) (KOWWI N v1.67)) 4.84 (calc. using C OSMOth erm (Arp et al., 2006) rk) [Royal Society of C hemistr y, C hemSpi der database ]57 5.30 (calc. using C OSMOth erm (Arp et al., 2006) rk) [Royal Society of C hemistr y, C hemSpi der database ]57 rk) [Royal Society of C hemistr y, C hemSpi der database ]57 5.480 at 25 °C (Estimate from Log Kow [5.33 (KowWin estimate) ] and log 7.23 (calc., C OSMOth erm, (Wang et al., 2011c)) 7.50 (calc., C OSMOth erm, (Wang et al., 2011c)) 7.77 (calc., C OSMOth erm, (Wang et al., 2011c)) 8.08 (calc., C OSMOth erm, (Wang et al., 2011c)) 8.36 (calc., C OSMOth erm, (Wang et al., 2011c)) C13 PFCA PFTrDA C14 PFCA PFTeDA rk) [Royal Society of C hemistr y, C hemSpi der database ] 57 8.63 (calc., C OSMOth erm, (Wang et al., 2011c)) 8.87 (calc., C OSMOth erm, (Wang et al., 2011c)) https://echa.europa.eu/information-on-chemicals/registered-substances, date of access: 2022-09-30. 57 http://www.chemspider.com/, date of access: 2022-09-30. 346 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA PFNA C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA TFA PFBA PFPeA PFHxA PFHpA PFOA Kaw [-5.343 (exp. database )]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate) ) Kaw [-2.313 (HenryWi n estimate) ]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate) ) log Kaw [-1.592 (HenryWi n estimate )]; Predicted using US EPA EPISuite (KOAWI N v1.10 estimate )) log K AW -5.343 at 25 °C (exp. database US EPA) [Royal Society of C hemistr y, C hemSpi der database ]57 -2.313 at 25 °C (Predicte d using US EPA EPI-Suite (HenryWi n v3.10 estimate) ) -3.04 (calc. using EPIsuite (Arp et al., 2006)) -1.592 at 25 °C (Predicte d using US EPA EPI-Suite (HenryWi n v3.10 estimate )) -2.66 (calc., (EC HA, 2016a)) -2.66 (calc. using EPIsuite (Arp et al., 2006)) -2,37 (calc. using EPIsuite (Arp et al., 2006)) -0.150 at 25 °C (Predicte d using US EPA EPI-Suite (HenryWi n v3.10 estimate) ) -1.93 (calc., C OSMOth erm, (Wang et al., 2011c)) -2.03 (calc. using EPIsuite (Arp et al., 2006)) -1.58 (calc., C OSMOth erm, (Wang et al., 2011c)) -1.79 (calc. using EPIsuite (Arp et al., 2006)) -1.27 (calc., C OSMOth erm, (Wang et al., 2011c)) -1.52 (calc. using EPIsuite (Arp et al., 2006)) -0.92 (calc., C OSMOth erm, (Wang et al., 2011c)) -0.58 (calc., C OSMOth erm, (Wang et al., 2011c)) -0.38 (calc., C OSMOth erm, (Wang et al., 2011c)) 0.03 (calc., C OSMOth erm, (Wang et al., 2011c)) dissocia pKa 0.32– pKa -0.16 pKa 0.5 <1.6 <1.6 <1.6 pKa pKa pKa Kaw [0.150 (HenryWi n estimate) ]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate) ) 347 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m tion constan t partitio n coeffici ents log Kd (sedime C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA 0.05±0.1 0 (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) 0.42 (exp. value, potentio metric titration of aq. sol.; (C abala et al., 2017) 0.40±0.1 0 (calculat ed using Advance d C hemistr y Develop ment (AC D/La bs) Software V11.02) (Zhao et al., 2014) 0.47±0.1 0 (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) (calculate d from exp. values, (Vierke, 2014)) 1.3 (LópezFontán et al., 2005) (calculate d from exp. values, (Vierke, 2014)) 0.82 (calc., C OSMOth erm, (Wang et al., 2011c)) 2.58 (exp. value, measure ment of the PFC As solubility change with pH; (C abala et al., 2017) (calculate d from exp. values, (Vierke, 2014)) 2.58 (Moroi et al., 2001) 2.61 (exp. value, measure ment of the PFC As solubility change with pH; (C abala et al., 2017) (calculate d from exp. values, (Vierke, 2014)) 0.52±0.1 0 (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) 0.52±0.1 0 (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) 0.52±0.1 0 (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) n.a. n.a. n.a. 1.4 – 3.1 (Li et al., 2011) n.a. 0.04 (Ahrens et al., 2010b)* 0.6 (Ahrens et al., 2010b)* 1.8 (Ahrens et al., 2010b)* 3.0 (Ahrens et al., 2010b)* n.a. n.a. n.a. 3.13 (exp. value, measure ment of the PFC As solubility change with pH; (C abala et al., 2017) 348 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m nt and overlap ping dissolve d phase) log K OC (sedime nt organic carbonnormali sed distribu tion coeffici ent) water solubilit y C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA PFNA C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA TFA PFBA PFPeA PFHxA PFHpA PFOA 0.437 (Predicte d using US EPA EPI-Suite (PC KOCW IN v1.66)) 1.767 (Predicte d using US EPA EPI-Suite (PC KOC W IN v1.66)) 2.431 (Predicte d using US EPA EPI-Suite (PC KOC WIN v1.66)) 1.63 – 2.35 (Sepulva do et al., 2011) 3.761 (Predicte d using US EPA EPI-Suite (PC KOC WIN v1.66)) 2.06 (Higgins and Luthy, 2006) 1.09 (Ahrens et al., 2010b)* 2.39 (Higgins and Luthy, 2006) 2.4 (Ahrens et al., 2010b)* 2.76 (Higgins and Luthy, 2006) 3.6 (Ahrens et al., 2010b)* 3.3 (Higgins and Luthy, 2006) 4.8 (Ahrens et al., 2010b)* n.a. n.a. n.a. miscible with water (>10 g/cm³)(e xp. result; REAC H registrati on data (202105-31))56 1 000 g/L at 20 °C , fully 0.7657 g/L at 25 °C (Estimate from Log Kow [2.43 (KowWin est)]; Predicted using US EPA EPISuite (WSKOW v1.41)) 61.11 mg/L at 25 °C (Estimat e from Log Kow [3.40 (KowWin est)]; Predicted using US EPA EPISuite (WSKOW v1.41)) 15.7 g/L (25 °C ) (Zhao et al., 2014) 0.3527 mg/L at 25 °C (Estimate from Log Kow [5.33 (KowWin est)]; Predicted using US EPA EPISuite (WSKOW v1.41)) 9.5 g/L (25 °C ) 4.14 g/L (22 °C ) (EC HA, 2013) practicall y insoluble in water (exp. result; MSDS Alfa Aesar) [Royal Society of C hemistr y, C hemSpi der 5.14 g/L at 25 °C (EC HA, 2016b) 2.9⋅10g/L pH 1 at 25 °C 2.2⋅104 g/L pH 2 at 25 °C 2.0⋅103 g/L pH 3 at 25 °C 0.014 g/L pH 4 at 25 °C 7.3⋅10g/L; pH 1 at 25 °C 5.5⋅105 g/L; pH 2 at 25 °C 5.1⋅104 g/L; pH 3 at 25 °C 3.5⋅103 g/L; pH 4 at 1.9⋅106 g/L; pH 1 at 25 °C 1.4⋅105 g/L; pH 2 at 25 °C 1.3⋅104 g/L; pH 3 at 25 °C 9.3⋅104 g/L; pH 4 at 1.2⋅10g/L; pH 1 at 25 °C 9.0⋅104 g/L; pH 2 at 25 °C 8.5⋅103 g/L; pH 3 at 25 °C 0.056 g/L; pH 4 at 25 °C 4 5 6 349 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA database ]57 1.882⋅106 g/L (Estimate from Log Kow [7.27 (KowWin est)]; Predicted using US EPA EPISuite (WSKOW v1.41)) miscible (exp. data, SRC PhysProp database, REAC H registrati on data (202105-31))56 vapour pressur e 12.4 kPa at 20 °C (interpola ted from exp. results; REAC H registrati on data (202105-31))56 15.5 kPa at 25 °C C10PFCA PFDA 1.3 kPa at 25 °C (calculate d using Advanced C hemistr y Developm ent (AC D/Lab s) Software V11.02) 1.06 kPa at 25 °C (calculat ed using Advance d C hemistr y Develop ment (AC D/La bs) Software 1.98 mm Hg at 25 °C ; equals to 263.93 Pa US EPA; Estimatio n Program Interface (EPI) Suite. 71.9 Pa at 25 °C (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) 4.2 Pa (25 °C ) extrapola ted from measured data 2.3 Pa (20 ° C ) extrapola ted from measured data 128 Pa 22.8 Pa at 25 °C (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) 3.1 to 99.97 kP a (129.6 to 218.9 C ) (calculate d) (EC HA, 2016b) C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA 0.14 g/L; pH 5 at 25 °C 0.16 g/L; pH 6-10 at 25 °C (calculate d) (EC HA, 2012b) 0.034 g/L pH 5 at 25 °C 0.039 g/L pH 6 at 25 °C 0.040 g/L pH 7 at 25 °C 0.041 g/L pH 8-10 at 25 °C (calculate d) (EC HA, 2012e) 25 °C 8.6⋅103 g/L; pH 5 at 25 °C 0.0100 g/L; pH 6-10 at 25 °C (calculate d) (EC HA, 2012d) 25 °C 2.2⋅103 g/L; pH 5 at 25 °C 2.6⋅103 g/L; pH 6-10 at 25 °C (calculate d) (EC HA, 2012c) 0.6 to 99.97 kP a (112 to 237.7 °C) (calculate d) (EC HA, 2012b) 1.25 Pa at 25 °C (calculate d) (EC HA, 2012e) 0.48 Pa at 25 °C (calculate d) (EC HA, 2012d) 0.18 Pa at 25 °C (calculate d) (EC HA, 2012c) 350 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m boiling point C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA V11.02) Ver. 4.11. Nov, 2012. Available from, as of Jan 11, 2015 140 °C (exp. value, Alfa Aesar) [Royal Society of 157 °C (Savu, 2000) (Predicte d using US EPA EPI-Suite (Mean VP of Antoine & Grain methods, MPBPWIN v1.42)) 12.8 kPa at 25 °C (calculate d using Advanced C hemistr y Develop ment (AC D/Lab s) Software V11.02) 71.78 °C (extrapol ated, exp. result, ebulliome ter; REAC H 120.0 °C (C abala et al., 2017) C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA 218 °C measured (EC HA, 2016b) 238.4 °C (calculate d) (EC HA, 2012b) 249 °C (EC HA, 2012e) 260.7 °C (calculate d) (EC HA, 2012d) 270 °C (EC HA, 2012c) (59.3 °C ) measured (EC HA, 2013) -63.5 °C (exp. result; SRC [Syracus e Research C orporati 189.0 °C (C abala et al., 2017) 218 °C (EC HA, 2015b) 351 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m Henrys Law constan t C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA 9.08E002 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond 4.77E001 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite registrati on data (202105-31))56 73 °C (Handboo k data: C RC ; REAC H registrati on data (202105-31))56 72.4 °C (Handboo k data: Merck index; REAC H registrati on data (202105-31))56 4.31E006 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond C hemistr y, C hemSpi der database ]57 1.19E004 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond 6.26E004 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA n.a. n.a. n.a. n.a. n.a. on of Syracuse , New York (US)]) 3.29E003 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond 1.73E002 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond 352 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbrevi a-tion acrony m C2 -PFCA C4 -PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA TFA PFBA PFPeA PFHxA PFHpA PFOA PFNA Method, HENRYWI N v3.10)) Method, HENRYWI N v3.10)) Method, HENRYW IN v3.10)) Method, HENRYWI N v3.10)) Method, HENRYWI N v3.10)) Method, HENRYWI N v3.10)) (Bond Method, HENRYWI N v3.10)) Hcp=8.9× 101 [mol/(m 3 Pa)] (Sander, 2015) Hcp= 8.2×10−2 [mol/(m 3 Pa)] (Sander, 2015) n.a. Hcp=4.4× 10−1 [mol/(m 3 Pa)] (Sander, 2015) Hcp=5.0× 10−2 [mol/(m 3 Pa)] (Sander, 2015) Hcp=4.9× 10−2 [mol/(m 3 Pa)] (Sander, 2015) Hcp=4.3× 10−2 [mol/(m 3 Pa)] (Sander, 2015) C10PFCA PFDA C11PFCA PFUnDA C12 PFCA PFDoDA C13 PFCA PFTrDA C14 PFCA PFTeDA Hcp=2.5× 10−2 [mol/(m 3 Pa)] (Sander, 2015) Hcp=1.3× 10−2 [mol/(m 3 Pa)] (Sander, 2015) Hcp=6.4× 10−3 [mol/(m 3 Pa)] (Sander, 2015) n.a. Hcp=1.6× 10−3 [mol/(m 3 Pa)] (Sander, 2015) Table B.67. Basic substance information and physical chemical properties of PFSAs (Perfluoroalkane sulfonic acids). abb CC2 C3 C4 C6 C8-PFSA C10-PFSA C12-PFSA C13 -PFSA rev PFSA PFSA PFSA PFSA PFSA iatio n acr TFMS, PFBS PFHxS PFOS PFDS PFDoDS PFTrDS ony TFSA, m HOTf or TfOH 1,1,2,2,3, 1,1,2,2,3,3,4, 1,1,2,2,3,3,4,4,5, 1,1,2,2,3,3,4,4,5,5, 1,1,2,2 1,1,2,2, IUP Trifluor Pentaflu 1,1,2,2 4,5,5,6,6,7,7, 5,6,6,7,7,8,8,9,9, 6,6,7,7,8,8,9,9,10,1 ,3,3,4, 3,3,4,4, 3,4,4,5,5, AC ometha oroetha ,3,3,30,11,11,12,12,13,13 4,45,5,6,6, 6,6,7,7,8, 8,8,9,9,10,10, 10,10,11,11,12,1 na nesulfo nesulfo Heptafl 102,12,13uoro-1- Nonafl 68,8me nic acid nic acid Tridecaf Heptahenicosafluoro Pentacosafluorodo Heptacosafluorotride propan uo-rodecane-1cane-1-sulfonic acid esulfoni 1luoro-1- decafluoro decane-1sulfonic acid sulphonic acid c acid butane hexane- -1- C14 -PFSA PFTeDS 1,1,2,2,3,3,4,4,5,5,6, 6,7,7,8,8,9,9,10,10,1 1,11,12,12,13,13,14,1 4,14Nonacosafluorotetrade cane-1-sulfonic acid 353 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS sulfoni c acid Perfluo robuta nesulfo nic acid sulfonic acid octanesulf onic acid Perfluor ohexan esulfoni c acid Perfluoroo ctanesulfo nic acid Perfluorodeca nesulfonic acid Perfluorododecane sulfonic acid Perfluorotridecanesu lfonic acid Perfluorotetradecanes ulfonic acid oth er na me s Triflic acid Perfluor oethane sulfonic acid Perfluor opropa nesulfo nic acid mol ecu lar for mul a C F3SO 3H C F3(C F2 )-SO 3H C F3(C F2 )2SO 3H C F3(C F 2)3SO 3H C F3(C F2 )5-SO 3H C F3(C F2)7SO 3H C F3(C F2)9SO 3H C F3(C F2)11-SO 3H C F3(C F2)12-SO 3H C F3(C F2)13-SO 3H 1493354-88CA 13-6 1 S nu mb er EC 216nu 087-5 mb er physico-chemical data 42341-6 37573-5 355-464 1763-23-1 335-77-3 79780-39-5 791563-89-8 1379460-39-5 - 206793-1 206587-1 217-179-8 206-401-9 279-259-9 - - mol ecu 250.1 300.1 400.1 500.1 600.2 700.2 750.2 800.2 150.1 200.1 354 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m lar wei ght g/ mol par titi oni ng coe ffici ent log K OW CPFSA C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH <0.3 at 25 °C and pH 1, (exp. result, HPLC method , OEC D 117; REAC H registra tion data (20210521))56 -0.49 (QSAR estimat ion (KOWW 0.48 (Predict ed using US EPA EPISuite (KOWW IN v1.67 estimat e)) 1.45 (Predict ed using US EPA EPISuite (KOWW IN v1.67 estimat e)) C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS -0.34 at 23 °C and pH 1.7 (exp. result, shake flask metho d EU Method A.8; REAC H registr ation data (20210526)) 2.41 (Predic 5.17 (calc., C OSMO therm, (Wang et al., 2011c)) 4.57 (exp. value, MSDS LabNet work) [Royal Society of C hemist ry, C hemS pider databas e]57 4.512±0.8 62 at 25 °C (calculate d using Advanced C hemistry Developm ent (AC D/Labs ) Software V11.02) 4.49 (Predicted using US EPA EPISuite; HSDB, National Library of Medicine 5.972±0.891 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 7.432±0.916 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 8.161±0.927 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 8.891±0.939 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 355 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA C3 PFSA C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA TFMS, TFSA, HOTf or TfOH PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS IN); REAC H registra tion data (20210521))56 ted using US EPA EPISuite (KOW WIN v1.67 estimat e)) 2.808 (exp. value, MSDS LabNet work) [Royal Society of C hemis try, C hemS pider databa se]57 4.34 (Predict ed using US EPA EPISuite (KOWW IN v1.67 estimat e)) (US))58 58 https://pubchem.ncbi.nlm.nih.gov/, date of access: 2022-09-30. 356 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA log K OA 4.902 at 25 °C (Estima te from Log Kow [0.49 (KowWi n estimat e)] and log Kaw [-5.392 (Henry Win estimat e)]; Predict ed using US EPA EPISuite (KOAW IN v1.10 C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH 5.152 at 25 °C (Estima te from Log Kow [0.48 (KowWi n estimat e)] and log Kaw [-4.672 (Henry Win estimat e)]; Predicte d using US EPA EPISuite (KOAWI N v1.10 estimat e)) 5.401 at 25 °C (Estima te from Log Kow [1.45 (KowWi n estimat e)] and log Kaw [-3.951 (Henry Win estimat e)]; Predict ed using US EPA EPISuite (KOAW IN v1.10 C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS 5.640 at 25 °C (Estim ate from Log Kow [2.41 (KowW in estimat e)] and log Kaw [-3.23 0 (Henry Win estimat e)]; Predict ed using US EPA EPISuite 7.55 (calc., C OSMO therm, (Wang et al., 2011c)) 6.130 at 25 °C (Estima te from Log Kow [4.34 (KowWi n estimat e)] and log Kaw [-1.790 (Henry Win estimat e)]; Predicte d using n.a. n.a. n.a. n.a. n.a. 357 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS estimat e)) (KOAW IN v1.10 estimat e)) -3.951 at 25 °C (Predict ed using US EPA EPISuite (Henry Win v3.10 estimat e)) -3.230 at 25 °C (Predic ted using US EPA EPISuite (Henry Win v3.10 estimat e)) US EPA EPISuite (KOAWI N v1.10 estimat e)) -2.38 (calc., C OSMO therm, (Wang et al., 2011c)) -1.790 at 25 °C (Predict ed using US EPA EPISuite (Henry Win v3.10 estimat n.a. n.a. n.a. n.a. n.a. TFMS, TFSA, HOTf or TfOH estimat e)) log K AW C3 PFSA -5.392 at 25 °C (Predict ed using US EPA EPISuite (Henry Win v3.10 estimat e)) -4.672 at 25 °C (Predict ed using US EPA EPISuite (Henry Win v3.10 estimat e)) 358 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA dis soc iati on con sta nt pKa<0 (REAC H registra tion data (20210525))56 pKa -3. 86±0.5 0 (calcula ted using Advanc ed C hemist ry Develop ment (AC D/L abs) Softwar e V11.02) pKa -3. 63±0.5 0 (calcula ted using Advanc ed C hemis try Develo pment (AC D/L abs) Softwar e V11.02 ) pKa -3. 57±0.5 0 (calcul ated using Advanc ed C hemis try Develo pment (AC D/L abs) Softwa re V11.02 ) par titi n.a. n.a. n.a. n.a. C3 PFSA TFMS, TFSA, HOTf or TfOH C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS pKa -3.27 ±0.50 (calculate d using Advanced C hemistry Developm ent (AC D/Labs ) Software V11.02) pKa -3.26±0. 50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -3.26±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -3.26±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -3.26±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) n.a. n.a. n.a. n.a. n.a. e)) -3.45 (calc., C OSMO therm, (Wang et al., 2011c)) pKa -3. 34±0.5 0 (calcula ted using Advanc ed C hemist ry Develop ment (AC D/L abs) Softwar e V11.02) n.a. 359 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS 2.345 (Predic ted using US EPA EPISuite 3.675 (Predict ed using US EPA EPISuite n.a. n.a. n.a. n.a. n.a. on coe ffici ent s log Kd (se dim ent and ove rla ppi ng dis sol ved pha se) log K OC (se dim ent org ani 0.352 (Predict ed using US EPA EPISuite 1.016 (Predict ed using US EPA EPISuite 1.681 (Predict ed using US EPA EPISuite 360 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA c car bon nor mal ise d dist rib uti on coe ffici ent ) wat er sol ubil ity (PC KOC WIN v1.66 estimat e)) (PC KOC WIN v1.66 estimat e)) ≥1 604 g/L at 20 °C (exp. result, flask method , OEC D 105; REAC H registra 17.04 g/L at 25 °C (Estima te from Log Kow [0.48 (KowWi n est)]; Predicte C3 PFSA C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS (PC KOC WIN v1.66 estimat e)) (PC KO C WIN v1.66 estimat e)) (PC KOC WIN v1.66 estimat e)) 1.378 g/L at 25 °C (Estima te from Log Kow [1.45 (KowWi n est)]; Predict ≥1 000 g/L at 20 °C (exp. result, flask metho d, EU Method A.6; REAC H 2.3 g/L (calc., C OSMO therm, (Wang et al., 2011c)) 7.5 g/L in unbuffere d water (pH 1.82) at 25 °C (calculate d using Advanced C hemistry Developm ent 0.42 g/L in unbuffered water (pH 1.82) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software 0.026 g/L in unbuffered water (pH 4.43) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 6.7 x 10-3 g/L in unbuffered water (pH 5.05) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.8 x 10-3 g/L in unbuffered water (pH 5.65) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) TFMS, TFSA, HOTf or TfOH 361 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH tion data (20210525))56 1.975⋅1 05 g/L at 25 °C (Estima te from Log Kow [0.49 (KowWi n est)]; Predict ed using US EPA EPISuite (WSKO W v1.41)) d using US EPA EPISuite (WSKO W v1.41)) ed using US EPA EPISuite (WSKO W v1.41)) C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS (AC D/Labs ) Software V11.02) V11.02) registr ation data (20210526))56 999 g/L in unbuff ered water (pH 0.52) at 25 °C (calcul ated using Advanc ed C hemis try Develo pment (AC D/L abs) Softwa 362 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m vap our pre ssu re CPFSA C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH 2.4 hPa (at 20 °C ), 3.2 hPa (at 25 °C ), 12.9 hPa (at 50 °C ) (exp. result, OEC D 104 (Vapou r Pressur e C urve); REAC H registra tion data (202105- 36.3 Pa at 25 °C (Predict ed using US EPA EPISuite (Mean VP of Antoine & Grain method s, MPBPWI N v1.42)) 10.01 Pa at 25 °C (Predict ed using US EPA EPISuite (Modifi ed Grain method , MPBPW IN v1.42)) C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS 58.9 Pa (calc., (Wang et al., 2011c)) 0.267 Pa at 25 °C (Predicted using US EPA EPISuite (Antoine method); HSDB, National Library of Medicine (US))58 n.a. n.a. n.a. n.a. re V11.02 ) 7 Pa at 20 °C (exp. result; OEC D 104 (Vapou r Pressur e C urve) ; REAC H registr ation data (20210526))56 6.9 Pa at 25 °C (Predic ted 363 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m boil ing poi nt CPFSA C2 PFSA C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA TFMS, TFSA, HOTf or TfOH PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS 25))56 using US EPA EPISuite (Modifi ed Grain metho d, MPBPW IN v1.42)) 198 °C at 1013 hPa (exp. result; EU Method A.2; REAC H registr ation data (202105- 238239 °C (exp. result; HSDB, National Library of Medicin e (US))58 249 °C (exp. result; HSDB, National Library of Medicine (US))58 n.a. n.a. n.a. n.a. 161162 °C (exp. result; MSDS Alfa Aesar) [Royal Society of C hemis try, C hemS pider databas 178 °C (exp. databas e US EPA) [Royal Society of C hemist ry, C hemS pider databas e]57 C3 PFSA 196 °C (exp. databas e US EPA) [Royal Society of C hemis try, C hemS pider databas e]57 364 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m He nry s La w con sta nt CPFSA C2 PFSA C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA TFMS, TFSA, HOTf or TfOH PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS e]57 162 °C (handb ook data; REAC H registra tion data (20210525))56 9.92⋅10 -8 atm3 m /mol e at 25 °C (Predict ed using US EPA EPISuite (Bond Method , HENRY 26))56 3.97⋅10atmm3/mol e at 25 °C (Predict ed using US EPA EPISuite (Bond Method, HENRY WIN n.a. n.a. n.a. n.a. n.a. 5.21⋅107 atmm3/mol e at 25 °C (Predict ed using US EPA EPISuite (Bond Method, HENRY WIN C3 PFSA 2.74⋅10 -6 atm3 m /mol e at 25 °C (Predict ed using US EPA EPISuite (Bond Method , HENRY 1.44⋅1 05 atmm3/mol e at 25 °C (Predic ted using US EPA EPISuite (Bond Method , 4 365 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abb rev iatio n acr ony m CPFSA C2 PFSA C3 PFSA TFMS, TFSA, HOTf or TfOH WIN v3.10)) v3.10)) WIN v3.10)) - - - C4 PFSA C6 PFSA C8-PFSA C10-PFSA C12-PFSA C13 -PFSA C14 -PFSA PFBS PFHxS PFOS PFDS PFDoDS PFTrDS PFTeDS HENRY WIN v3.10)) Hcp=2. 0 [mol/( m3Pa)] (Sande r, 2015) v3.10)) Hcp=9.0×1 0−4 [mol/(m 3P a)] (Sander, 2015) - - - - Hcp=5.1 ×10−1 [mol/( m3Pa)] (Sander , 2015) Table B.68. Basic substance information and physical chemical properties of PFCs (Perfluoroalkanes). abbreviati C-PFC C2-PFC C3-PFC C4-PFC C6-PFC C8-PFC on IUPAC Tetrafluorometh Hexafluoroeth Octafluoropro Decafluorobut Tetradecafluo Octadecafluo name ane ane -pane ane ro-hexane ro-octane other C arbon Perfluoroethan Perfluoroprop Perfluorobutan Perfluorohexa Perfluoroocta names tetrafluoride e ane e ne ne molecular C F4 C 2F6 C 3F8 C 4F10 C 6F14 C 8F18 formula C10 -PFC C12-PFC Docosafluorodec ane Perfluorodecane Hexacosafluorodode cane Perfluorododecane C 10F22 C 12F26 C AS number 75-73-0 76-16-4 76-19-7 355-25-9 355-42-0 307-34-6 307-45-9 307-59-5 EC number 200-896-5 200-939-8 200-941-9 206-580-3 206-585-0 206-199-2 - 206-204-8 physico-chemical data 366 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviati on molecular weight g/mol C-PFC C2-PFC C3-PFC C4-PFC C6-PFC C8-PFC C10 -PFC C12-PFC 88.01 138.01 188.02 238.03 338.04 438.06 538.07 638.09 partitionin g coefficient log K OW 1.18 (HSDB, National Library of Medicine (US); REAC H registration data (2021-05-31))56 1.19 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 2.00 (exp. database US EPA, Hansch, C et al. (1995)) [Royal Society of C hemistry, C hemSpider database]57 2.15 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 2.15 (estimated with QSAR; REAC H registration data (202105-31))56 2.8 at 25 °C (calculated; REAC H registration data (202105-31))56 3.12 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 4.09 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 4.822 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database]57 6.02 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) ≥ 4.5 (OEC D 107 (Shake Flask Method), estimation method (solubility ratio); REAC H registration data (202106-08))56 6.584 (exp. value, MSDS LabNetwork) 7.95 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) ca. 6.2 at 25 °C , pH 7 (estimated using US EPA EPI-Suite; REAC H registration data (202106-09))56 8.346 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database]57 8.011±0.865 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 9.470±0.893 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 11.87 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database]57 log K OA -1.143 at 25 °C (Estimate from Log Kow [1.18 (exp. database)] and log Kaw [2.323 (exp. -0.919 at 25 °C (Estimate from Log Kow [2.00 (exp. database)] and log Kaw -0.010 at 25 °C (Estimate from Log Kow [3.12 (KowWin estimate)] -0.346 at 25 °C (Estimate from Log Kow [4.09 (KowWin estimate)] 0.144 at 25 °C (Estimate from Log Kow [6.02 (KowWin estimate)] 0.633 at 25 °C (estimate from Log Kow [7.95 (KowWin estimate)] n.a. n.a. 367 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviati on log K AW dissociatio n constant partition coefficient s log K d (sediment and overlappin C-PFC C2-PFC C3-PFC C4-PFC C6-PFC C8-PFC database)]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate)) -0.950 (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 -0.95 (HSDB, National Library of Medicine (US))58 2.323 (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 [2.919 (exp. database)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) and log Kaw [3.130 (exp. database)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) and log Kaw [4.436 (HenryWin estim.)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) and log Kaw [5.876 (HenryWin estim.)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) and log Kaw [7.317 (HenryWin estim.)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 2.919 (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 3.130 (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 no dissociable groups no dissociable groups no dissociable groups 4.436 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) no dissociable groups n.a. n.a. n.a. n.a. 5.876 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) no dissociable groups n.a. 7.317 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) no dissociable groups n.a. C10 -PFC C12-PFC n.a. n.a. no dissociable groups no dissociable groups n.a. n.a. 368 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviati on g dissolved phase) C-PFC C2-PFC C3-PFC C4-PFC C6-PFC C8-PFC C10 -PFC C12-PFC log K OC (sediment organic carbonnormalise d distributio n coefficient ) 1.687 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 2.352 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 3.016 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 3.681 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 5.010 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 6.339 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) n.a. n.a. water solubility 18.8-20 mg/L (exp. result; REAC H registration data (2021-05-31))56 4.1 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.52 g/L at 25 °C (exp. result, OEC D 105, column elution method; REAC H registration data (202105-31))56 0.36 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software 5.7 mg/L at 20 °C (calculated value; REAC H registration data (202105-31))56 5.7 mg/L at 15 °C (HSDB, National Library of Medicine (US))58 1.612 mg/L at 25 °C (Estimate from Log Kow [4.09 (KowWin est)]; Predicted using US EPA EPI-Suite (WSKOW v1.41)) 1.4 mg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development ≤0.1 mg/L at 20 °C , pH 7.1-7.2 (exp. result, extrapolated, slow-stirring flask method; REAC H registration data (202106-09))56 0.0096 mg/L at 25 °C (Estimate from Log Kow [6.02 (KowWin est)]; Predicted using US EPA EPI-Suite ca. 10 µg/L at 20 °C , pH 7 (exp. result; WoE, REAC H registration data (202106-09)) 56 0.052 µg/L at 25 °C (Estimate from Log Kow [7.95 (KowWin est)]; Predicted using US EPA EPI-Suite (WSKOW v1.41)) 0.048 µg/L 0.00031 µg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.00031 µg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 369 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviati on C-PFC C2-PFC C3-PFC V11.02) C4-PFC C6-PFC C8-PFC (AC D/Labs) Software V11.02) (WSKOW v1.41)) 0.0081 mg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) C10 -PFC C12-PFC vapour pressure 2.3 ⋅107 Pa at 25 °C (exp. result; REAC H registration data (2021-05-31))56 1.13 ⋅107 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 2902.1 kPa at 18 °C (exp. result, static cell method; REAC H registration data (202105-31))56 3.2 ⋅106 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 883.9 kPa at 25 °C (exp. result; HSDB, National Library of Medicine (US))58 767 kPa at 20 °C (exp. value, handbook data; REAC H registration data (202105-31))56 268.0 kPa at 25 °C (exp. result; HSDB, National Library of Medicine (US))58 258.6 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) ca. 26.5 kPa at 25 °C (calculated with QSAR; REAC H registration data (202106-09))56 30.4 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) ca. 3 kPa at 25 °C (exp. result; WoE, REAC H registration data (202106-09))56 5.17 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 599.95 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 92.39 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) boiling point -127.9 °C (handbook -78.1 °C (exp. result; HSDB, -37 °C (exp. value, -2.1 °C (handbook 58.45 °C (exp. result, 105.9 °C (exp. result; 150 °C (exp. result; source: 178 °C (exp. result; source: Haszeldine, 370 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviati on Henrys Law constant C-PFC C2-PFC C3-PFC C4-PFC C6-PFC C8-PFC C10 -PFC C12-PFC data; C RC Handbook of C hemistry and Physics. 95th Edition) 5.15 atmm3/mole (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 4.59 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) National Library of Medicine (US))58 handbook data; REAC H registration data (202105-31))56 33.0 atmm3/mole (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 24.5 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Group Method, HENRYWIN v3.10)) data; C RC Handbook of C hemistry and Physics. 95th Edition) 245 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Group Method, HENRYWIN v3.10)) EU Method A.2; REAC H registration data (202106-09))56 2.45⋅104 atm-m3/mole at 25 °C (Predicted using US EPA EPI-Suite (Group Method, HENRYWIN v3.10)) HSDB, National Library of Medicine (US))58 2.45⋅106 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Group Method, HENRYWIN v3.10)) Benning, Anthony F.; US2490764, 1949; Scifinder [C AS]) n.a. R. N.; Journal of the C hemical Society, (1950); Scifinder [C AS]) Hcp=1.2×10−7 [mol/(m 3Pa)] (Sander, 2015) Hcp=1.5×10−8 [mol/(m 3Pa)] (Sander, 2015) Hcp=5.4×10− - - - Hcp=2.1×10−6 [mol/(m 3Pa)] (Sander, 2015) 20.3 atmm3/mole (exp. database US EPA) [Royal Society of C hemistry, C hemSpider database]57 24.1 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) Hcp=6.5×10−7 [mol/(m 3Pa)] (Sander, 2015) n.a. 10 [mol/(m 3Pa)] (Sander, 2015) 371 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.69. Basic substance information and physical chemical properties of PFAEs (Haloperfluoroalkanes and Perfluoroalkylethers; acyclic and cyclic). abbreviatio Cryofluoran Fluobrene PFME PFEE Perfluorogly Perfluorodigl Tetrafluor HFPO n/acronym e me yme ooxirane 2,2,3,32,2,3-Trifluoro-31,1,2,21,1,2,2IUPAC 1,21,2Trifluoro(trifluorom 1,1,1,2,2Tetrafluoro (trifluoromethyl)o Tetrafluoro-1name Dichlorotetraf Dibromotetraf ethoxy)methane Pentafluoro-2- Tetrafluoro1,2[1,1,2,2oxirane xirane luoroethane luoroethane (pentafluoroet bis(trifluoromet tetrafluoro-2hoxy)ethane hoxy)ethane (trifluorometh oxy)ethoxy]2(trifluorometh oxy)ethane Perfluorodieth yl ether; Perfluoroethyl ether C 4F10O Perfluoroglyme Perfluorodigly me Tetrafluoro oxirane C 2Br2F4 Perfluorodimethyl ether; Perfluoromethyl ether C 2F6O C 4F10O 2 C 6F14O 3 C 2F4O Trifluoro(trifluoro methyl)oxirane; Hexafluoro-1,2epoxypropane C 3F6O 76-14-2 124-73-2 1479-49-8 358-21-4 378-11-0 40891-99-4 694-17-7 428-59-1 200-937-7 204-711-9 - - - - 211-767-8 207-050-4 other names C ryofluorane Fluobrene molecular formula structural formula C 2C l2F4 C AS number EC number physico-chemical data 372 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviatio n/acronym molecular weight g/mol Cryofluoran e 170.92 partitioning coefficient log K OW Fluobrene PFME PFEE Perfluorogly me 270.03 Perfluorodigl yme 386.04 Tetrafluor ooxirane 116.01 HFPO 259.82 154.01 254.03 2.82 (exp. database US EPA; source: Hansch,C et al. (1995)) 2.96 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 2.00 (Predicted using US EPA EPISuite (KOWWIN v1.67 estimate)) 3.93 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 6.120±0.807 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 5.55 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 1.10 (Predicted using US EPA EPISuite (KOWWIN v1.67 estimate)) 1.72 (Predicted using US EPA EPISuite (KOWWIN v1.67 estimate)) log K OA 0.761 at 25 °C (Estimate from Log Kow [2.82 (exp. database)] and log Kaw [2.059 (exp. database)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 2.139 at 25 °C (Estimate from Log Kow [2.96 (KowWin estimate)] and log Kaw [0.821 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 1.660 at 25 °C (Estimate from Log Kow [2.00 (KowWin estimate)] and log Kaw [0.340 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 2.148 at 25 °C (Estimate from Log Kow [3.93 (KowWin estimate)] and log Kaw [1.782 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) n.a. 6.437 at 25 °C (Estimate from Log Kow [5.55 (KowWin estimate)] and log Kaw [-0.887 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 2.113 at 25 °C (Estimate from Log Kow [1.72 (KowWin estimate)] and log Kaw [-0.393 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 0.340 at 25 °C (Predicted using US 1.782 at 25 °C (Predicted n.a. -0.887 at 25 °C 2.214 at 25 °C (Estimate from Log Kow [1.10 (KowWin estimate)] and log Kaw [-1.114 (HenryWin estimate)]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate)) -1.114 at 25 °C log K AW 2.059 (exp. database US 0.821 at 25 °C 166.02 -0.393 at 25 °C (Predicted using 373 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviatio n/acronym Cryofluoran e EPA) Fluobrene PFME PFEE (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) EPA EPI-Suite (HenryWin v3.10 estimate)) using US EPA EPI-Suite (HenryWin v3.10 estimate)) dissociation constant no dissociable groups no dissociable groups n.a. n.a. partition coefficients log K d (sediment and overlapping dissolved phase) log K OC (sediment organic carbonnormalised distribution coefficient) n.a. n.a. n.a. 2.352 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 2.352 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 130 mg/L at 25 °C (exp. result; HSDB; source: Riddick et al. (1985)) 3 mg/L at 25 °C (exp. result; HSDB; source: Horvath et al. (1999)) water solubility Perfluorogly me Perfluorodigl yme (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) Tetrafluor ooxirane (Predicted using US EPA EPISuite (HenryWin v3.10 estimate)) HFPO n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1.330 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 2.660 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 1.946 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 0.932 (Predicted using US EPA EPISuite (PC KOC WI N v1.66 estimate)) 1.596 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 7.4 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) 0.23 g/L unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry 4.706 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.17 g/L unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry 5 mg/L unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry 110 g/L unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry 1.2 g/L unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) US EPA EPI-Suite (HenryWin v3.10 estimate)) 374 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviatio n/acronym Cryofluoran e Fluobrene PFME PFEE Software V11.02) 1.431 g/L at 25 °C (Estimate from Log Kow [2.00 (KowWin est)]; Predicted using US EPA EPI-Suite (WSKOW v1.41)) Development (AC D/Labs) Software V11.02) 9.882 mg/L at 25 °C (Estimate from Log Kow [3.93 (KowWin est)]; Predicted using US EPA EPI-Suite (WSKOW v1.41)) 127.86 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) vapour pressure 268.51 kPa at 25 °C (exp. result; HSDB; source: Riddick et al. (1985)) 43.33 kPa at 25 °C (exp. result; HSDB; source: Daubert et al. (1989)) 3.60⋅106 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) boiling point 3.5 °C (handbook data; C RC Handbook of C hemistry and Physics. 91th Edition) 47.35 °C (handbook data; C RC Handbook of C hemistry and Physics. 86th Edition) -59 °C (exp. result; source: Simons, J. H.; US2500388, 1950; Scifinder (C AS)) 2.5 °C (exp. result; source: Dresdner, R. D.; Journal of Organic C hemistry, (1959), 24, Perfluorogly me Development (AC D/Labs) Software V11.02) Perfluorodigl yme Development (AC D/Labs) Software V11.02) Tetrafluor ooxirane Developme nt (AC D/Labs) Software V11.02) 110.92 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 4.27 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 13 °C (exp. result; source: Simons, J. H.; US2500388, 1950; Scifinder (C AS)) 66 °C (exp. result; Modena, S.; Journal of Fluorine C hemistry, (1988), 40(2- 3.57⋅106 Pa at 25 °C (calculated using Advanced C hemistry Developme nt (AC D/Labs) Software V11.02) -63.5 °C (exp. result; SRC [Syracuse Research C orporatio n of HFPO Software V11.02) 946.59 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) -27.4 °C (exp. result; SRC [Syracuse Research C orporation of Syracuse, New York (US)]) 375 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) abbreviatio n/acronym Cryofluoran e Fluobrene PFME PFEE Perfluorogly me Perfluorodigl yme 3), 349-57; Scifinder (C AS)) Tetrafluor ooxirane Syracuse, New York (US)]) HFPO 0.00188 atmm3/mole at 25 °C (Predicted using US EPA EPISuite (Bond Method, HENRYWIN v3.10)) 0.0099 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 698-700; Scifinder (C AS)) Henrys Law constant 1.51 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 0.162 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 0.0535 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 1.48 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) n.a. 0.00317 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) Hcp=9.0×10−6 [mol/(m 3Pa)] (Sander, 2015) Hcp=2.7×10−7 [mol/(m 3Pa)] (Sander, 2015) - - - - Hcp=8.8×10−6 [mol/(m 3Pa)] (Sander, 2015) Table B.70. Basic substance information and physical chemical properties of PFPAs (Perfluoroalkylphosphonic acids). acron PFMPA PFEPA PFBPA PFPPA PFHxPA PFOPA PFDPA ym IUPAC name other name s (Trifluoro methyl)phosphon ic acid Trifluoro methylphosphic acid; Perfluoro (Pentafluo roethyl)phosphoni c acid Pentafluor oethylphosphoni c acid; Perfluoroe PFDoPA (Nonafluorobut yl)phosphonic acid; (Undecafluoropen ty)lphosphonic acid (Tridecafluorohex yl)phosphonic acid Heptadecafluoroo ctylphosphonic acid (Henicosafluorode cyl)phosphonic acid (Pentacosafluorodod ecyl)phosphonic acid Perfluorobutylphosphonic acid Perfluoropentylphosphonic acid Perfluorohexanep hosphonic acid Perfluorooctylphosphonic acid Perfluorodecylpho sphonic acid Perfluorododecylpho sphonic acid 376 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acron ym PFMPA PFEPA methyl phosponic acid thyl phosponic acid molec C H2F3O 3P C 2H2F5O 3P ular formul a C AS 374-09-4 103305numb 01-7 er EC numb er physico-chemical data molec 149.99 200.00 ular weigh t g/mol PFBPA PFPPA PFHxPA PFOPA PFDPA PFDoPA C 4H2F9O 3P C 5H2F11O 3P C 6H2F13O 3P C 8H2F17O 3P C 10H2F21O 3P C 12H2F25O 3P 52299-24-8 2109769-70-0 40143-76-8 40143-78-0 52299-26-0 63225-55-8 - - - - - - 300.02 350.02 400.03 500.05 600.06 700.08 partiti oning coeffic ient log K OW -0.28 (Predicte d using US EPA EPI-Suite (KOWWI N v1.67 estimate) ) 0.68 (Predicted using US EPA EPISuite (KOWWIN v1.67 estimate)) 4.093±0.674 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 4.659±0.696 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 5.389±0.741 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 4.55 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 6.849±0.834 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 8.308±0.917 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 9.768±0.993 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) log 8.126 at 8.366 at n.a. n.a. 9.353 at 25 °C n.a. n.a. n.a. 377 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acron ym K OA log K AW dissoc iation PFMPA PFEPA PFBPA PFPPA 25 °C (Estimate from Log Kow [0.28 (KowWin estimate) ] and log Kaw [-8.406 (HenryWi n estimate) ]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate) ) 25 °C (Estimate from Log Kow [0.68 (KowWin estimate)] and log Kaw [-7.686 (HenryWin estimate)] ; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate)) -8.406 at 25 °C (Predicte d using US EPA EPI-Suite (HenryWi n v3.10 estimate) ) pKa 0.37±0.1 -7.686 at 25 °C (Predicted using US EPA EPISuite (HenryWin v3.10 estimate)) n.a. n.a. pKa 0.64±0.1 pKa 0.64±0.10 (calculated pKa 0.72±0.10 (calculated using PFHxPA PFOPA PFDPA PFDoPA -4.803 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) n.a. n.a. n.a. pKa 0.74±0.10 (calculated using pKa 0.78±0.10 (calculated using pKa 0.78±0.10 (calculated using pKa 0.78±0.10 (calculated using (Estimate from Log Kow [4.55 (KowWin estimate)] and log Kaw [-4.803 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 378 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acron ym consta nt partiti on coeffic ients log K d (sedi ment and overla pping dissol ved phase ) log K OC (sedi ment organi c carbo n- PFMPA PFEPA PFBPA PFPPA PFHxPA PFOPA PFDPA PFDoPA 0 (calculate d using Advanced C hemistr y Developm ent (AC D/Lab s) Software V11.02) n.a. 0 (calculate d using Advanced C hemistry Developm ent (AC D/Lab s) Software V11.02) using Advanced C hemistry Development (AC D/Labs) Software V11.02) Advanced C hemistry Development (AC D/Labs) Software V11.02) Advanced C hemistry Development (AC D/Labs) Software V11.02) Advanced C hemistry Development (AC D/Labs) Software V11.02) Advanced C hemistry Development (AC D/Labs) Software V11.02) Advanced C hemistry Development (AC D/Labs) Software V11.02) n.a. n.a. n.a. n.a. n.a. n.a. n.a. 0.654 (Predicte d using US EPA EPI-Suite (PC KOC W IN v1.66 estimate) 1.318 (Predicted using US EPA EPISuite (PC KOCWI N v1.66 estimate)) n.a. n.a. 3.977 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) n.a. n.a. n.a. 379 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acron ym norma lised distrib ution coeffic ient) PFMPA PFEPA PFBPA PFPPA PFHxPA PFOPA PFDPA PFDoPA ) water solubil ity 85 g/L in unbuffere d water (pH 0.49) at 25 °C (calculate d using Advanced C hemistr y Developm ent (AC D/Lab s) Software V11.02) 22 g/L in unbuffere d water (pH 1.10) at 25 °C (calculate d using Advanced C hemistry Developm ent (AC D/Lab s) Software V11.02) 3.6 g/L in unbuffered water (pH 1.96) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.7 g/L in unbuffered water (pH 2.33) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.8 g/L in unbuffered water (pH 2.71) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.19 g/L in unbuffered water (pH 3.44) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.038 g/L in unbuffered water (pH 4.18) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 5.7 mg/L in unbuffered water (pH 4.98) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) vapou r press ure 10.15 Pa at 25 °C (calculate d using Advanced C hemistr y Developm ent (AC D/Lab 105.59 Pa at 25 °C (calculate d using Advanced C hemistry Developm ent (AC D/Lab s) 40.80 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 18.67 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 8.12 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.40 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.227 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 35.60 mPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 380 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acron ym PFMPA PFEPA PFBPA PFPPA PFHxPA PFOPA PFDPA PFDoPA s) Software V11.02) Software V11.02) boilin g point 210.5± 45.0 °C (calculate d using Advanced C hemistr y Developm ent (AC D/Lab s) Software V11.02) 426.42 °C at 101325 Pa (exp. result, extrapolat ed; MSDS Oakwood) 186.0±50.0 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 199.9±50.0 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 214.3±50.0 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 243.6±50.0 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 272.4±50.0 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 300.2±52.0 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) Henry s Law consta nt 9.60⋅10-11 atmm3/mole at 25 °C (Predicte d using US EPA EPI-Suite (Bond Method, HENRYWI N v3.10)) - 5.04⋅10-10 atmm3/mole at 25 °C (Predicted using US EPA EPISuite (Bond Method, HENRYWI N v3.10)) - n.a. n.a. 3.85⋅10-7 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) n.a. n.a. n.a. - - - - - - 381 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.71. Perfluoroalkylamines. Basic substance information and physical chemical properties. acronym IUPAC name PFMAm 1,1,1-Trifluoro-N,Nbis(trifluoromethyl)met hanamine PFEAm 1,1,2,2,2Pentafluoro-N,Nbis(pentafluoroethyl) ethanamine PFPrAm 1,1,2,2,3,3,3Heptafluoro-N,Nbis(1,1,2,2,3,3,3heptafluoropropyl)1-propanamine PFBAm 1,1,2,2,3,3,4,4,4Nonafluoro-N,Nbis(nonafluorobuty l)-1-butanamine PFHxAm 1,1,2,2,3,3,4,4,5,5, 6,6,6Tridecafluoro-N,Nbis(tridecafluorohex yl)-1-hexanamine other names Tris(trifluoromethyl)ami ne; Perfluorotrimethylamine C 3F9N; [(C F3)3N] Heneicosafluorotripr opylamine; Perfluorotripropylam ine; Perfluamine C 9F21N; [(C 3F7)3N] Tris(perfluorobutyl)amine; Perfluorotributyla mine C 12F27N; [(C 4F9)3N] Perfluorotrihexylam ine molecular formula Pentadecafluorotriet hylamine; Perfluorotriethylamin e C 6F15N; [(C 2F5)3N] C 18F39N; [(C 6F13)3N] C AS number 432-03-1 EC number physico-chemical data 359-70-6 206-632-5 338-83-0 206-420-2 311-89-7 206-223-1 432-08-6 - 758-48-5 - molecular weight g/mol 221.02 371.05 521.07 671.09 971.14 321.04 partitioning coefficient log K OW 1.29 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 4.18 (Predicted using US EPA EPISuite (KOWWIN v1.67 estimate)) 6.462 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database]57 ≥5.3-≤6.1 (readacross: log Kow of 5.3 is for PFHp (perfluoroheptanes); log Kow of 6.1 is for PTBA (perfluorotributylami nes); REAC H registration data (2021-06-14)) 9.105 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database]57 15.109±0.941 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 11.748 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database]57 19.103±0.998 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 3.22 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 1,1,2,2,2Pentafluoro-N(pentafluoroethyl )-N(trifluoromethyl)e thanamine Perfluoromethyldi ethyl-amine; Perfluorodiethylm ethylamine C 5F13N 382 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acronym log K OA PFMAm 1.428 at 25 °C (Estimate from Log Kow [1.29 (KowWin estimate)] and log Kaw [-0.138 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) PFEAm 2.155 at 25 °C (Estimate from Log Kow [4.18 (KowWin estimate)] and log Kaw [2.025 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) PFPrAm n.a. PFBAm n.a. PFHxAm n.a. log K AW -0.138 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) 2.025 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) n.a. n.a. n.a. 1.304 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) dissociation constant pKa -28.74±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -27.46±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -27.02±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -26.84±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -26.31±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) pKa -28.57±0.50 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) partition coefficients log K d (sediment and overlapping dissolved phase) n.a. n.a. n.a. n.a. n.a. n.a. 1.916 at 25 °C (Estimate from Log Kow [3.22 (KowWin estimate)] and log Kaw [1.304 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 383 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acronym log K OC (sediment organic carbonnormalised distribution coefficient) PFMAm 3.104 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) PFEAm 5.098 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) PFPrAm n.a. PFBAm n.a. PFHxAm n.a. water solubility 0.21 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.2 mg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.381 ug/L (exp. result, mean value, EPA OPPTS 830.7840 (Water Solubility), flask method; REAC H registration data (2021-06-14)) insoluble (exp. result; HSDB, National Library of Medicine (US)) 0.081 ug/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 5.2⋅10-11 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 5.5 mg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) vapour pressure 394.6 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 17.3 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.516 kPa at 20 °C (exp. result, ASTM E1719-97; REAC H registration data (2021-06-14)) 73.33 Pa (exp. result; HSDB, National Library of Medicine (US)) boiling point -10.5 °C (exp. result; source: Young, John A.; Journal of the American C hemical Society, (1958), 80, 1889-92; Scifinder (C AS)) 72 °C (exp. result; source: Felling, Kyle W.; Journal of Fluorine C hemistry, (2003), 123(2), 233-236; Scifinder (C AS)) 132 °C (exp. result, ASTM E-1719-97 and ASTM D112094; REAC H registration data (2021-06-14)) 178 °C (handbook data; C RC Handbook of C hemistry and Physics. 83rd Edition) 0.45 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 258 °C (exp. result; source: Kauck, Edward A.; GB666733, 1952https://scifind ern.cas.org/navigate/ 46.93 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 45 °C at 978.59 hPa (exp. result; source: Kauck, Edward A.; GB666733, 1952https://scifi nder- 4.433 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 384 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acronym Henrys Law constant PFMAm PFEAm PFPrAm PFBAm PFHxAm ?appId=c45de92fc9dd-48de-ac3f512e39d817e2&cle arSearch=true&res ultType=reference& resultView=DETAIL &state=searchDetai l.reference&suppres sNavigation=true& uiC ontext=369&uiS ubC ontext=607&uri ForDetails=docume nt/pt/document/18 419826; Scifinder (C AS)) 0.0178 atm-m3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 2.59 atm-m3/mole at 25 °C (Predicted using US EPA EPISuite (Bond Method, HENRYWIN v3.10)) n.a. n.a. n.a. - - - Hcp= 1.8×10−10 [mol/(m 3Pa)] (Sander, 2015) - n.cas.org/navigat e/?appId=c45de9 2f-c9dd-48deac3f512e39d817e2&c learSearch=true& resultType=refer ence&resultView =DETAIL&state= searchDetail.refer ence&suppressNa vigation=true&ui C ontext=369&uiS ubC ontext=607& uriForDetails=doc ument/pt/docum ent/18419826; Scifinder (C AS)) 0.493 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) - Table B.72. Basic substance information and physical chemical properties of FTOHs (Fluorotelomer alcohols, C4 to C10). acronym 4:2 FTOH 6:2 FTOH 8:2 FTOH 10:2 FTOH IUPAC 3,3,4,4,5,5,6,6,63,3,4,4,5,5,6,6,7,7,8,8,8 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,1 3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,1 name Nonafluoro-1-hexanol -Tridecafluoro-1-octanol 0-Heptadecafluoro-1-decanol 2-Henicosafluoro-1-dodecanol 385 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acronym other names molecular formula C AS number EC number 4:2 FTOH 2(Perfluorobutyl)ethan ol C 6H5F9O 6:2 FTOH 2(Perfluorohexyl)ethanol 8:2 FTOH 2-(Perfluorooctyl)ethanol 10:2 FTOH 2-(Perfluorodecyl)ethanol C 8H5F13O C 10H5F17O C 12H5F21O 2043-47-2 647-42-7 678-39-7 865-86-1 218-050-9 211-477-1 211-648-0 212-748-7 364.1 464.1 564.1 6.995±0.855 at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 9.213 (exp. value, MSDS LabNetwork) [Royal Society of C hemistry, C hemSpider database] 57 n.a. physico-chemical data molecular 264.1 weight g/mol partitionin g coefficient log K OW 3.66 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 5.60 (Predicted using US EPA EPI-Suite (KOWWIN v1.67 estimate)) 7.53 (Predicted using US EPA EPISuite (KOWWIN v1.67 estimate)) log K OA 4.314 at 25 °C (Estimate from Log Kow [3.66 (KowWin estimate)] and log Kaw [-0.654 (HenryWin estimate)]; Predicted using US EPA EPISuite (KOAWIN v1.10 estimate)) 4.812 at 25 °C (Estimate from Log Kow [5.60 (KowWin estimate)] and log Kaw [0.788 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) 5.301 at 25 °C (Estimate from Log Kow [7.53 (KowWin estimate)] and log Kaw [2.229 (HenryWin estimate)]; Predicted using US EPA EPI-Suite (KOAWIN v1.10 estimate)) log K AW -0.138 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) pKa 14.16±0.10 (calculated using -0.138 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) -0.138 at 25 °C (Predicted using US EPA EPI-Suite (HenryWin v3.10 estimate)) n.a. pKa 14.26±0.10 (calculated using pKa 14.28±0.10 (calculated using Advanced C hemistry Development pKa 14.28±0.10 (calculated using Advanced C hemistry Development (AC D/Labs) Software dissociatio n constant 386 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acronym 4:2 FTOH Advanced C hemistry Development (AC D/Labs) Software V11.02) 6:2 FTOH Advanced C hemistry Development (AC D/Labs) Software V11.02) 8:2 FTOH (AC D/Labs) Software V11.02) 10:2 FTOH V11.02) partition coefficients log K d (sediment and overlappin g dissolved phase) log K OC (sediment organic carbonnormalised distribution coefficient) n.a. n.a. n.a. n.a. 2.726 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 4.055 (Predicted using US EPA EPI-Suite (PC KOC WIN v1.66 estimate)) 5.384 (Predicted using US EPA EPISuite (PC KOC WIN v1.66 estimate)) 5.2 (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.1 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 1.72 kPa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 140-143 °C (exp. value, Sigma Aldrich, 0.058 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 50.9 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 88-95 °C at 28 mmHg (exp. value, Sigma 0.003 g/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 0.17 mg/L in unbuffered water (pH 7) at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 22.7 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 2.77 Pa at 25 °C (calculated using Advanced C hemistry Development (AC D/Labs) Software V11.02) 112-114 °C at 10 mmHg (exp. value, Alfa Aesar, MSDS) [Royal 145 °C at 10 mmHg (exp. value, Alfa Aesar, MSDS) [Royal Society of C hemistry, C hemSpider water solubility vapour pressure boiling point 387 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) acronym 4:2 FTOH MSDS) [Royal Society of C hemistry, C hemSpider database]57 6:2 FTOH Aldrich, MSDS) [Royal Society of C hemistry, C hemSpider database] 57 8:2 FTOH Society of C hemistry, C hemSpider database]57 10:2 FTOH database]57 Henrys Law constant 5.42E-003 atmm3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 1.50E-001 atm-m3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) 4.14 atm-m3/mole at 25 °C (Predicted using US EPA EPI-Suite (Bond Method, HENRYWIN v3.10)) n.a. Hcp= 6.6×10−3 [mol/(m 3Pa)] (Sander, 2015) Hcp= 1.7×10−4 [mol/(m 3Pa)] (Sander, 2015) Hcp= 2.0×10−4 [mol/(m 3Pa)] (Sander, 2015) Hcp= 1.3×10−4 [mol/(m 3Pa)] (Sander, 2015) 388 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.4.1.2. Key arrowhead subgroups On the reliability and training sets of QSAR models The below assessment of the training sets of the QSAR models relates to the modelling of PFAS persistence described in section B.4.1.2. Biodegradability estimates are likely to be less accurate for compounds outside the molecular weight (MW) range of the training set compounds. Therefore, the substance’s MW should be in the range of the training set substance’s MW. The training set encompasses substances with MW ranging from 31–698 (BIOWIN 1 and 2), MW 53-698 (BIOWIN 3 and 4), MW 30– 1215 (BIOWIN 5 and 6), and MW 46 – 885 (BIOWIN 7). In the present study, the analysed PFAS’ MW ranged from 214 to 971. In this sense, all predictions may be considered reliable except for the perfluorotrihexylamine (MW = 971.143), which was outside the training set MW for most BIOWIN models. Furthermore, it is important that fragment coefficients have been developed for all functional group(s) or other structural features of the substances, as they might be relevant for biodegradation. Table B.73 shows all fragments present in each of the 18 substances analysed herein, followed by the presence or absence of this fragment in the training set, for different BIOWIN models. One major issue in predicting biodegradability of PFAS is that PFAS contain fluorine [-F] fragments, whose coefficients were not developed either for the linear/non-linear models (BIOWIN 1 and 2), nor for the primary/ultimate biodegradation models (BIOWIN 3 and 4), compromising the robustness and reliability of the results. On the other hand, MITI models (BIOWIN 5 and 6) lack a fragment coefficient for trifluoromethyl groups [ -CF3]. Therefore, the influence of the fluorine atoms and bonds on PFAS aerobic biodegradation will be underestimated, as no BIOWIN model is complete to predict all fluorine fragments in the molecule. For anaerobic biodegradation, fragment coefficients values for the fragments [-F] and [-CF3] returned as zero for all 18 studied PFAS. If a chemical for estimation contains unique or unusual substructures not included in a model’s fragment library, these structural features should not be considered in the prediction process. It should be noted, however, that a model can still have value even if there is a “missing fragment” deemed important. Considering that these fragments are not listed in the training set of BIOWIN 7 (anaerobic), it is likely that coefficients were not developed despite displayed in the output tables. As for fragments that categorize the PFAS substances into different groups, aliphatic acid [C(=O)-OH], sulfonic acid, tertiary amine and aliphatic ether [C-O-C] have fragment coefficients developed for BIOWIN models, while the fragment phosphonic acid is not included in the training sets. Table B.73. Presence (YES) or absence (NO) of fragments for different BIOWIN models present in each modelled PFAS in the training set. Group and Fragment BW 1, 2 BW 3, 4 BW 5, 6 BW 7 substance Carboxylic C arbon with 4 single bonds & no YES YES YES YES acids hydrogens Aliphatic acid [-C (=O)-OH] YES YES YES YES PFOA PFHxA Trifluoromethyl group [-C F3] YES YES NO UNL PFBA Fluorine [-F] NO NO YES UNL Sulfonic acids PFOS PFHxS PFBS C arbon with 4 single bonds & no hydrogens YES YES YES YES Sulfonic acid/salt -> aliphatic attach YES YES NO YES Trifluoromethyl group [-C F3] YES YES NO UNL 389 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Group and substance Phosphonic acids PFOPA PFHxPA PFBPA Perfluoroalka nes Perfluorohexan e Perfluorooctan e Perfluorodecali ne* Fragment BW 1, 2 BW 3, 4 BW 5, 6 BW 7 Fluorine [-F] C arbon with 4 single bonds & no hydrogens Phosphonic acid -> aliphatic attach NO YES NO YES YES YES UNL YES NO NO NO NO Trifluoromethyl group [-C F3] Fluorine [-F] C arbon with 4 single bonds & no hydrogens Trifluoromethyl group [-C F3] YES NO YES YES NO YES NO YES YES UNL UNL YES YES YES NO UNL Fluorine [-F] NO NO YES UNL Perfluoroalky C arbon with 4 single bonds & no YES YES YES hydrogens lamines Perfluamine Tertiary amine YES YES YES Perfluorometh Trifluoromethyl group [-C F3] YES YES NO yldiethylamine Fluorine [-F] NO NO YES Perfluorotrihex ylamine C arbon with 4 single bonds & no Ethers YES YES YES Perfluorodiethy hydrogens lether Aliphatic ether [C -O-C ] YES YES YES C F3-O-C F2Trifluoromethyl group [-C F3] YES YES NO C F2-O-C F3 Fluorine [-F] NO NO YES 2,2,3,3,4,4,5heptafluorotetr ahydro-5(nonafluorobut yl)furan BW – BIOWIN model; *Perfluorodecaline does not have [-C F3] groups; UNL - unlikely YES YES UNL UNL YES YES UNL UNL There are maximum instances of fragments in the substances used in the training set. Therefore, it is recommended that the substance to be predicted has no more than the maximum fragment instances used in the training set. Table B.74 shows the maximum instances of each fragment in the training set library (grey cells), as well as the number of each fragment in the modelled substances. As it can be seen from Table B.73 and Table B.74, no PFAS could be accurately predicted by all BIOWIN models. Similarly, different applicability domains were found to different PFAS groups, mainly due to the presence and instances of characteristic fragments. 390 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.74. Maximum number of instances of each fragment in the training set (grey) versus the number of instances of each fragment in the modelled PFAS (colours). The colours indicate which set of BIOWIN models cover the instances of the fragments in the traini ng set. Fragment C arbon with 4 single bonds & no hydrogens Trifluoromethyl group [-C F3] Fluorine [-F] Aliphatic acid [-C (=O)-OH] Sulfonic acid / salt -> aliphatic attach Phosphonic acid - aliphatic attach Tertiary amine Maximum instances of the fragment in the training set Instances of the fragment in the modelled PFAS Carboxylic Sulfonic Phosphoni Perfluoro acid acid c acid alkane BW 1,2 BW 3,4 BW 5,6 BW 7 A B C D E F G H I J K 2 3 12 3 6 4 2 7 5 3 7 5 3 4 1 1 0 0 1 1 1 1 1 1 1 1 1 0 4 0 1 27 3 0 3 15 1 11 1 7 1 17 0 13 0 9 0 17 0 13 0 1 2 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 4 2 1 1 0 0 0 0 0 0 Perfluoro alkylamine Ether L M N O P Q R 6 10 6 2 15 2 2 7 2 2 0 3 3 3 2 2 1 9 0 14 0 18 0 18 0 21 0 13 0 39 0 10 0 10 0 16 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 Aliphatic ether [C -O-C ] 2 5 5 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 1 A. PFOA; B. PFHxA; C . PFBA; D. PFOS; E. PFHxS; F. PFBS; G. PFPOA; H. PFHxPA, I. PFBPA; J. Perfluorohexane; K. Perfluorooctane ; L. Perfluorodecaline; M. Perfluamine; N. Perfluoromethyldiethylamine; O. Perfluorotrihexylamine; P. Perfluorodiethylether; Q. 1,1,2,2 -Tetrafluoro-1,2-bis(trifluoromethoxy)ethane; R. 2,2,3,3,4,4,5-heptafluorotetrahydro-5-(nonafluorobutyl)furan Within training set of all models Within training set of models 3-7 Within training sets of models 1-4 Within training set of MITI models (5 and 6) Outside training sets 391 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.4.1.3.1. Degradation of PFCA precursors Table B.75. Summary of formed PFCAs during degradation of n:2 FTOHs and the intermediate products (5:3 acid, fluorotelomer carboxylic acid (FTCA) and fluorotelomer unsaturated carboxylic acid (FTUCA)). Substance Compartment Study C4-PFCA C5-PFCA C6-PFCA C7-PFCA C8-PFCA C9-PFCA C10-PFCA Reference duration [%] [%] [%] [%] [%] [%] [%] 6:2 FTOH Atmosphere + + + + (Ellis et al., 2004) Atmosphere + + + (Styler et al., 2013) Soil (flow 84 d 0.8 4.2 4.5 (Liu et al., through) 2010b) Soil (closed 180 d 1.8 30 8.1 (Liu et al., system) 2010c) Mixed bacterial 90 d <0.5 <0.5 5 (Liu et al., culture 2010c) WWTP-activated 60 d 4.4 mol% 11 mol% (Zhao et al., sludge 2013b) Aerobic river 100 d 1.5 mol% 10.4 8.4 mol% (Zhao et al., sediment system mol% 2013a) Anaerobic digester sludge 5:3 acid (5:3 FTC A) 8:2 FTOH Anaerobic sediment WWTP-activated sludge Atmosphere Aqueous photolysis – H2O 2 solution Aqueous photolysis – synthetic field water 90 d 176 d 100 d - 90 d 0.1 - 0.2 mol% 0.4 mol% 0.6 mol% 0.8 mol% 5.9 mol% 0.1 0.24 - 0.32 (Zhang et al., 2013b) 1.5 1.6 10 h 40 + 140-146 h 1-8 + (Zhang et al., 2016c) (Wang et al., 2012) (Ellis et al., 2004) (Gauthier and Mabury, 2005) 392 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Compartment Aqueous photolysis –Lake Ontario mixed microbial system (sediment and groundwater mixed bacterial culture activated sludge Study duration C4-PFCA [%] C5-PFCA [%] C6-PFCA [%] 81 d C9-PFCA [%] + (but below LOQ) - 3 - (Dinglasan et al., 2004) 6 - 2.1 - (Wang et al., 2005a) (Wang et al., 2005b) (Wang et al., 2009) Soil 197 d 1-4 Anaerobic digester sludge Anaerobic activated sludge Sediment-water system 181 d - - 10-40 (average 25) 0.3 mol% 5.4 8.9 17 8:2 FTUC A Sediment-water system 35 d 10:2 FTOH soil 30 d 10:2 FTC A 10:2 FTUC A 1 C8-PFCA [%] 3 Not evaluated Not evaluated - 8:2 FTC A 90 d C7-PFCA [%] 28 d 150 d 1.9 1.2 50 d C10-PFCA [%] Reference (Zhang et al., 2013b) (Li et al., 2018b) (Myers and Mabury, 2010) 21 mol% (water) 9.3 mol% (sed.) 27 mol% (water) 9 mol% (sed.) 5.1 mol% <1 mol% <1 mol% (Myers and Mabury, 2010) 4.3 mol% 59.7 mol% Soil-earthworm 8.7 mol% 7.3 mol% 74.9 mol% (Zhao and Zhu, 2017) Soil-wheat 8.9 mol% 5.9 mol% 77.8 mol% Soil-earthwormwheat 9.9 mol% 6.0 mol% 74.8 mol% 1.1 mol% (water) 11 mol% (sed.) 6 mol% (water) Sediment-water system Sediment-water system 12 mol% (water, at day 22) 50 d 35 d 0.37 mol% 1.9 mol% (sed.) (Myers and Mabury, 2010) (Myers and Mabury, 2010) 393 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Compartment Study duration C4-PFCA [%] C5-PFCA [%] C6-PFCA [%] C7-PFCA [%] (sed.) C8-PFCA [%] C9-PFCA [%] 1.7 mol% (sed.) C10-PFCA [%] 22 mol% (sed.) Reference [+] detected, but not quantified; [-] not detected; [ ] not evaluated Table B.76. Summary of formed PFCAs during degradation of n:2 fluorotelomer derivatives. Substance Compartment Study C4-PFCA C5-PFCA C6-PFCA C7-PFCA duration [%] [%] [%] [%] n:2 Fluorotelomer iodides (FTIs) 6:2 FTI Soil 4:2 FTI Atmosphere FTI Hydrolysis (modelling) 91 d - 20 mol% + + 3.8 mol% C8-PFCA [%] C9-PFCA [%] 16 mol% C10-PFCA [%] Reference (Ruan et al., 2010) (Young et al., 2008; Young and Mabury, 2010) (Nielsen, 2014; Rayne and Forest, 2010) C orresponding FTOHs and PFC As Esters of n:2 fluorotelomer alcohols 8:2 fluorotelomer stearate monoester 8:2 fluorotelomer stearate monoester 8:2 Fluorotelomer citrate triester n:2FT(M)A (n=2-12) Agricultural soil Forest soil 80 d 94 d 0.16 mol% 0.2 mol% 0.38 mol% 0.9 mol% 1.7 mol% 4 mol% 218 d 0.2 mol% 0.8 mol% 4 mol% Forest soil 8:2 FTA Soil 105 d <0.4 mol% 1.3 mol% 8 mol% (Royer et al., 2015) 8:2 FTMA Soil 105 d <0.4 mol% 3.4 mol% 10.3 mol% (Royer et al., 2015) Hydrolysis (modelling) C orresponding FTOHs and PFC As 0.009 mol% (Dasu et al., 2013) (Dasu et al., 2013) (Dasu et al., 2013) (Nielsen, 2014; Rayne and Forest, 2010) 394 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Compartment 4:2 FTA Atmosphere Study duration 10 d C4-PFCA C5-PFCA C6-PFCA [%] [%] [%] C orresponding PFC As (1-10 mol%) C7-PFCA [%] C8-PFCA [%] C9-PFCA [%] C10-PFCA [%] Reference (Butt et al., 2009) n:2 Polyfluoroalkyl phosphoric acid mono-/diesters (monoPAPs/diPAPs) n:2 diPAPs (n = 4, 6, 8, 10) Rats 6:2 monoPAP Wastewater and sewage sludge Wastewater and sewage sludge Wastewater and sewage sludge Soil and plant 6:2 diPAP n:2 monoPAPs (n = 4, 6, 8, 10) 6:2 diPAP 6:2 diPAP 6:2 diPAP 8:2 diPAP 8:2 diPAP 8:2 monoPAP and diPAP Activated sludge Soil Soil compost amended soil 2.4 compost amended substrate in presence of crops (carrot) in presence of crops (lettuce) Hydrolysis C orresponding FTOHs and PFC As (D'Eon J and Mabury, 2011) 92 d 0.7 mol% 2.1 mol% 8.4 mol% 92 d 1.5 mol% 6.2 mol% 7.3 mol% (Lee et al., 2010) 92 d C orresponding FTOHs (1-2% after 92 days) and PFC As 5.5 months 30 d + + + + 2 mol% 0.04 mol% 112 d 112 d 0.73 0.47 mol% 6.4 108 d 3 months + + 6 0.34 0.25 2.1 + + 10% + + 62% + + + 1 month + 14 d 8:2 FTOH (Lee et al., 2014) (Lewis et al., 2016) (Liu and Liu, 2016) (Bizkarguena ga et al., 2016) + (D'Eon and Mabury, 2007; 395 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Compartment Study duration C4-PFCA [%] C5-PFCA [%] C6-PFCA [%] C7-PFCA [%] C8-PFCA [%] C9-PFCA [%] C10-PFCA [%] Reference Nielsen, 2014; Rayne and Forest, 2010) 8:2 monoPAP and diPAP Rats 15 d - + n:2 Fluorotelomer urethane (monomers) toluene-2,4-di-(8:2 Agricultural 180 d fluorotelomer soil urethane) (FTU) Forest soil + + (from residual 8:2 FTOH) 117 d 0.07 mol% 0.11 mol% 0.84 mol% Forest soil 180 d hexamethylene-1,6di(8:2 fluorotelomer urethane) (HMU) n:2 Fluorotelomer sulfonic acids (FTSAs) 0.06 mol% 0.14 mol% 0.94 mol% - (D'Eon and Mabury, 2007) (Dasu and Lee, 2016) 6:2 FTSA WWTPactivated sludge 90 d 0.14 1.5 1.1 - (Wang et al., 2011b) 6:2 FTSA Aerobic sediment 90 d - 21 mol% 20 mol% 0.55 mol% (Zhang et al., 2016c) Anaerobic sediment 100 d - - - - n:2 Fluorotelomer thioether amido sulfonates (FTTAoSs) [belonging to the PFAS subgroup n:2 fluorotelomer -thiol derivatives] n:2 FTTAoS Soil amended 60 d + + + + (Harding(n=4,6,8) with an AFFF Marjanovic solution et al., 2015) n:2 Fluorotelomer olefins (FTOs) n:2 Fluorotelomer Atmosphere olefins C orresponding FTOHs and PFC As (Nielsen, 2014; Sulbaek Andersen et 396 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Compartment Study duration C4-PFCA [%] C5-PFCA [%] C6-PFCA [%] C7-PFCA [%] C8-PFCA [%] C9-PFCA [%] C10-PFCA [%] Reference al., 2005; Young and Mabury, 2010) n:2 Fluorotelomer silanes n:2 Fluorotelomer Atmosphere silanes C orresponding PFC As (Nielsen, 2014; Zhu et al., 2019) [+] detected, but not quantified; [-] not detected; [ ] not evaluated 397 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.4.2.1.3. Mobility in water Table B.77. Concentrations of PFASs in ground water described by detection frequency (% or x/xx), median value (ng/L) in bold and range in brackets. Grey cells indicate sampling sites near known PFAS point sources. NR=not reported. Location Pollution N TFA PFEtS PFPrA PFBA PFPeA PFHxA PFHpA PFOA Reference Denmark 247 89% (Miljøstyrelsen, (<50-2 2020) 400) 11 European Partly 2 000-7 46% 29% 25% 24% (WFD C IS, countries polluted areas 000 2020) (Janda et al., Germany 42 ~800 2019) (NR-7 500) Germany, Polluted 5 (~800(NRBadensampling 2 200) 56) Württemberg sites Sweden (Ground/surface water) Sweden Suspected 23 contamination 100% (0.075 700) 100% 100% (4.8(4.824 000) 65 000) Near landfills 52-196 100% 100% 100% (Ericson Jogsten (3.4(2.6(1.4and Yeung, 90 000) 29 000) 26 000) 2017) 61% 60% 66% 56% 71% (Miljösamverkan 129 80 77 23 54 Sverige, 2022) (2). Long-chain PFAAs increasingly accumulated in the roots with increasing carbon chain (shoot:root ratio <1) (Krippner et al., 2014) Root concentration factors (RC F) for PFSAs increased with increasing carbon chain length RC F for PFC As decrease with increasing carbon chain length from C 4 to C 6, then increase between C 6 and C 11 and are quite similar for C 11 to C 14 (U-shaped relationship with chain length) Foliage to root concentration factor (FRC F) for PFAAs decreased exponentially with increasing carbon chain length (FRC F >1 for C 4 and C 5 PFC As) Translocation of PFAAs from root to foliage can by described by transpiration stream concentration factor (TSC F) (assumption: substance is not degraded in the plant, elimination from the plant is negligible, substance is only taken up through roots). Except for C 4 PFC A (TSC F ~ 0.8) the TSC F values were less than 1, which means that transfer from nutrient solution to leaves was inhibited. No simple relationship between the TSC F and carbon chain length of PFAAs could be observed. RC Fs increased with increasing chain length and were highest for long-chain PFC As (≥ C 11) in all three plant species. RC Fs (Felizeter et al., 2012)  PFC As (C 4-C 14) PFSAs (C 4, C 6, C 8)     PFC As (C 4-C 14) PFSAs (C 4, C 6, C 8)  (Felizeter et al., 2014) 444 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method system (PFAA-spiked nutrient solution with nominal concentration of 10 ng/L to 10 µg/L of each spiked PFAA) Plant species (plant parts) lycopersicum var. Moneymaker), cabbage (Brassica oleracea convar. capitata var. alba) and zucchini (Cucurbita pepo var. Black Beauty) (root, stem, leaf, twig, and edible parts) substances Results       semi-static mesocosm study (exposure phase 14 days, depuration period 14 days, individual PFAA aquatic macrophytes: Echinodorus horemanii (submerged species) and Eichhornia PFC As (C 5-C 14) PFSAs (C 4, C 6, C 8)     were a factor of about 2-3 higher for PFSAs than for PFC As (with same carbon chain) Higher stem concentration factors were observed for C 8-C 11 PFAAs (sharp decrease of stem concentration factors ≥ C 11 PFC As) Leaf concentration factor were in all plant species >1 (except long-chain PFC As). C ompared with the concentration factors of the other aboveground parts of the plants, the leaves show the highest concentration factors for most of the PFA As. Edible part concentration factor (cabbage head, tomato and zucchini fruit) were highest for the short-chain PFC As and decrease with increasing chain length. No concentrations were detected above LOQ for C 12-C 14 PFC As. EC F was >1 for C 4C 6 PFC As in cabbage and tomato and for C 5 PFC A in zucchini All other PFAAs than C 12-C 14 PFC As were present in the above ground plant parts. They are taken up with the transpiration stream and accumulate primarily in the leaves. Distribution within the plants were similar between plant species and among PFAAs. Transpiration stream concentration factor was <1 for all compounds in all plant species, which means transfer from the nutrient solution to the vegetative parts of the plants was inhibited. All edible part /leaf transfer factor factors were <1, which indicates that leafy crops with open leaves (spinach or some lettuce) accumulate higher amounts in the edible part than fruit-bearing crops. Leafy crops pose a higher risk for human exposure. BC Fs increased with increasing carbon chain length Higher BC Fs for PFSAs compared to PFC As with identical chain length Leaf: BC F = 16.6 – 835 (E. horemanii), BC F = 21.2-349 (E. Crassipes) (higher leaf BC F in submerged species) Root: BC F = 3.5 – 1620 (E. horemanii), BC F = 3.1-2 590 (E. Crassipes) Reference (Pi et al., 2017) 445 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method concentration = 20 µg/L) Plant species (plant parts) crassipes (freefloating species) substances Results    greenhouse, water augmented with varying concentrations of PFAAs (nominal 0.2-40 µg/L). lettuce (Lactuca sativa) and strawberry (Fragaria ananassa) (lettuce leaf and strawberry root, shoot, fruit) PFC As (C 4-C 9) PFSAs (C 4, C 6, C 8)        Reference Whole-plant: BC F = 13.7 – 910 (E. horemanii), BC F = 18.8 – 1280 (E. Crassipes) Translocation factors (TF): After exposure phase only C 4 PFSA, C 5 PFC A and C 6-PFC A allocate more in leaf compared to roots (TF >1) in both species. At the end of depuration phase additionally C 6 PFSA, C 7 PFC A and C 8 PFC A were observed to be more allocated in leaf than in roots. During exposure phase TFs were generally higher for the submerged species compared to the free-floating specie (except for C 5 and C 6 PFC A). During depuration phase TFs were higher in the free-floating species (except for C 13 and C 14 PFC A). A sigmoidal relationship between BC Fs and chain length as well as membrane-water distribution coefficient, protein-water distribution coefficient and organic carbon-water partition coefficient was observed. PFAA plant concentrations increased linearly with the aqueous concentration of PFAAs PFC As bioaccumulated to a greater degree than PFSAs C hain length dependency trends in lettuce shoot and strawberry fruit: decreasing concentrations with increasing chain length In strawberry fruit concentrations for C 7-C 9 PFC As and C 6, C8 PFSAs were below LOQ (suggestion that other specific transport mechanisms exist for long-chain PFAAs) Strawberry: fruit-soil concentration factor (at 10 µg/L water concentration): C 4 PFC A = 203, C 5 PFC A 243, C 6 PFC A 34.5 (overall average decrease of fruit-soil concentration factor per C F2 group was ~0.3 log units) Lettuce: The BAFs for PFC As decreased ~0.4 -0.7 log units per additional C F2 group. Dependent on organic carbon content of the soil the BAFs for PFC As ranged from 0.938 to 3390 and for PFSA from 0.759 to 316 (at 10 µg/L water concentration) Bioaccumulation potential depends on analyte functional group and chain length, concentration in the reclaimed water, (Blaine et al., 2014b) 446 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) substances Results greenhouse study, designed to assess PFAA phytoremediation (PFAA were added weekly to irrigation water (nominal 1 mg/L of each compound), 14 to 18-week establishment period, sand was used as growth media) herbaceus plant species: amaranth (Amaranthus tricolor), mustard (Brassica juncea), bemudagrass (Cynodon dactylon), horsetail (Esquisetum hyemale), sunflower (Helianthus annuus), tall fescue (Schedonorus arundinaceus), red fescue (Festuca rubra) and C rimson clover (Trifolium incarnatum) woody plant species: river birch (Betula nigra), green ash (Fraxinus pennsylvanica), sweetgun PFC As (C 5, C 6, C 8) PFSAs (C 4, C 6, C 8, C 10)  Reference and organic carbon content of the soil    Highest BC F was observed for C 5 PFC A (18.0 – 174.6 in herbaceous plant species, 0.3 – 156.2 in woody plant species) and lowest for C 8 PFSA (0.7 – 5.5 in herbaceous plant species, 0.0 – 16.4 in woody plant species) BC Fs for PFC As were higher than for PFSAs BC Fs decreased with increasing chain length (except for C 4 PFSA) BC F of the best performing tree species (based on foliage concentrations, Salix nigra) were lower than BC F of the best performing herbaceous species (Festuca rubra) but were generally in the same overall range of the herbaceous plants. 447 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) (Liquidambar styraciflua), tulip poplar (Liriodendron tulipfera), sycamore (Platanus occidentalis), loblollv pine (Pinus taeda), black willow (Salix nigra) substances Results Reference hydroponic model plant system (PFAA concentrations nominal 2 µg/L) wall cress (Arabidopsis thaliana) (root, shoot) PFC As (C 4-C 10) PFSAs (C 4, C 6, C 8)  (Muller et al., 2016) climate chamber - hydroponic system (nutrient solution spiked with PFAAs at 0.5 mg/L and 1 mg/L; sampling after 1, 2, 6, 13 and 20 days) climate chamber hydroponic system grass (Bromus diandrus) PFC As (C 4, C 8, C 10) PFSAs (C 4, C 6, C 8) Rapid saturation of root concentration occurred for all PFAAs (except C 4 PFC A). Shoot concentrations increased continuously. RC Fs and SC Fs of PFC As followed U-shaped trend with increasing chain length (highest for C 4, C 9 and C 10 PFC As). For PFSAs the RC Fs and SC Fs increased with increasing chainlength. The transfer factor from nutrient solution to the aerial part of plant increased with exposure time for both concentration levels, except from day 13 to day 20 at the higher concentration level where no increasing transfer factor was observed. For PFC As the transfer factor decreased with increasing chain length. No statistical differences in transfer factor values were observed for PFSAs. After 20 days transfer factors were highest for C 4-PFC A, but all PFAAs were greater than 1 (2.036 – 5.65) Time-dependent uptake: All PFAAs were efficiently absorbed by wheat roots. C 2 and C 3 PFC As were rapidly taken up within the 80-hour exposure period and no steady state was observed. The uptake rates of the C 4-C 8 PFAAs slowed during (Zhang et al., 2019c)     wheat (Triticum acstivnm L.) (root, shoot) PFC As (C 2-C 4, C 6, C 8) PFSA (C 8)  (GarciaValcarcel et al., 2014) 448 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) substances (1. timedependent: nutrient solution spiked with each PFAA at 0.1 mg/L; sampling after 2, 4, 8, 6, 32 and 80 hours; 2. concentrationdependent: nutrient solution spiked with PFAA at 0, 0.1, 0.5, 0.7, 1.0, 1.5, 2.0, 2.5 mg/L; sampling after 4 hours;3. effect of metabolic, aquaporin and anion-channel inhibitors) greenhouse study, hydroponic system exposed with three different solutions: effluents from two WWTPs (C 4C 8 PFC As and C 4, C 8 PFSAs Results     lettuce (Lactuca sativa L.) and spinach (Spinacia oleracea L.) (root, shoot) PFC As (C 4-C 14) PFSAs (C 4, C 6, C 8)   the late exposure period. The C 8 PFAAs reached a steady state at the end of exposure period. Root concentrations of PFC As decreased with increasing chain length, achieving a minimum at C 6 PFC A, and then increased again with the chain length. The concentrations in shoots were lower than in roots, except C 4 PFC A. The final shoot concentration decreased with increasing chain length. The translocation factors (concentration ratio shoot:root) were all less than 1 and declined with increasing chain length except for C 4 PFC A (translocation factor = 1.1). The relatively low translocation factor for C 2 and C 3 PFC A could be a result due to their high concentrations in roots and the short exposure time (80 hours). Therefore, an increase is expected as the bioaccumulation proceeds for a longer time. C oncentration-dependent uptake: The uptake of PFAAs was nonlinear and followed Michaelis-Menden model well, indicating that the uptake is a carrier-mediated process. The absorption of different types of PFAAs (PFC As and PFSAs) and different chain lengths may follow different pathways in wheats. The uptake of PFAAs by wheat is mainly an energy dependent active process, whereas for C 2 and C 3 PFC As, anion channels and aquaporins also participate in the uptake process. The root concentration factor showed similar pattern under different treatments. The root concentration factor values ranged from 5.4 (C 4 PFSA) to 2400 (C 11 PFC A) for lettuce and from 1.7 (C 4 PFSA) to 1500 (C 12 PFC A) for spinach, respectively. The root concentration factor was generally higher in lettuce than spinach. The values for PFC As decreased from C 4 to C 6 and from C 11 to C 14, while it increased in the range between C 6 and C 11. For PFSAs the root concentration factor increased with increasing chain length. The leaf concentration factor showed similar pattern under Reference (Dal Ferro et al., 2021) 449 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) substances detected in effluent) and a PFAAs-spiked drinking water solution (nominal concentration of 500 ng/L for each PFAA); exposure time: 45 (lettuce) and 55 (spinach) days Results  climate chamber - hydroponic system (nutrient solution spiked with 6:2 FTSA at 1.100 nmol/mL; exposure time: 12 days) pumpkin (Cucurbita maxima L.) (root, shoot) hydroponic greenhouse study (nutrient solution spiked with FOSA at 1.856 nmol/mL; exposure time: 12 days) soybean (Glycine max L. Merrill) and pumpkin (Cucurbita maxima L.) (root, shoot) 6:2 fluorotelomer sulfonic acid (6:2 FTSA) PFC As (C 2-C 8) PFSAs (C 4, C 6, C 8)    Perfluorooctane sulfonamide (FOSA) + degradation products (C 8 PFSA is also present as impurity (~8 mol%) in the FOSA standard)   different treatments. The highest leaf concentration factor was observed for C 4 PFC A (median values between 47 and 440) followed by C 5 PFC A (49-64) and C 6 PFC A (9.8-22). The translocation factor from root to shoot decreased with increasing carbon chain length (except for C 5 PFC A in spinach which has a higher TF than C 4 PFC A). For lettuce, the translocation factor was >1 only for C 4 PFC A and C 5 PFC A for the treatment with WWTP effluents. For all other PFAAs and treatments the translocation factor was less than 1. For spinach, translocation factors >1 were observed for C 4 – C 6 PFC As and C 4 PFSA in all treatments and additionally C 7 PFC A in treatment with PFAA-spiked solution. C omparing PFC As and PFSAs with the same carbon chain length, show that PFC As had a greater affinity to shoots accumulation than PFSAs. 6:2 FTSA and all tested PFAAs had a root concentration factor >1. The root concentration factor of 6:2 FTSA was 2.6-24.2fold as high as those of PFAAs of the same or much shorter carbon chain length, demonstrating much higher bioaccumulative ability of 6:2 FTSA in pumpkin roots. C 2-C 7 PFC As were found as metabolites in roots and shoots. C 7 PFC A and C 4 PFC A were the major metabolite in roots, while C 4 PFC A was the major metabolite in shoots. Neither the PFAAs nor 6:2 FTSA had higher concentrations in shoots than in roots, leading to translocation factors lower than 1. The translocation factor decreased with increasing perfluorinated carbon chain length (and increasing log Kow). The translocation factor of FOSA from roots to shoots was 1.5-fold higher for pumpkin (0.09) than for soybean (0.06). The higher root lipid content of pumpkin might be an explanation. C 4, C 6 and C 8 PFSA were found as metabolites in roots and shoots of pumpkin and soybean. The concentration in roots were higher than in shoots for all metabolites. The concentration increased with increasing chain length. Reference (Zhao et al., 2019) (Zhao et al., 2018a) 450 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method growth chamber (nutrient solution spiked with C 8 PFSA or F53-B concentrations at 0.05 and 0.1 µg/mL; exposure time: 7 days) Plant species (plant parts) wheat (Triticum aestivum L.) (root, shoot) substances Results Reference C 8 PFSA, commercial F53-B (6:2 C l-PFESA as major component and 8:2 C l-PFESA as impurity)  Root concentrations of C 8 PFSA and C l-PFESAs were an order of magnitude higher than those in shoots. There was no significant difference in BC Fshoot values between C 8 PFSA and 6:2 C l-PFESAs at both tested levels (BC Fshoot = 2.614 – 4.182). The BC Fshoot values for 8:2 C l-PFESA were about 70-76% lower than for C 8 PFSA. On the contrary, the BC Froot values for 8:2 C l-PFESA were 1.4-1.9-fold higher than for C 8 PFSA (BC Froot 8:2 C l-PFESAs = 194.7 and 266.7). The translocation factor values from root to shoot were similar for C 8 PFSA and 6:2 C l-PFESA (0.023-0.029) and quite lower for 8:2 C l-PFESA (0.004 and 0.005). 6:2 C l-PFESAs had a similar accumulation pattern as C 8 PFSA, whereas 8:2 C l-PFESA was predominantly restricted to the roots, which might be attributed to their hydrophobicity and carbon chain length. (Lin et al., 2020b) No correlation between the root concentration factor and carbon chain length of PFAAs Transfer potential from roots to straws and further to the grains was higher for short-chain PFAAs than for long-chain PFAAs Transfer factor from roots to straws: higher for PFC As than for PFSAs (with same carbon chain) Transfer factor from straws to grains: higher for PFSAs than for PFC As (with same carbon chain) PFC A concentrations in grain increased logarithmically with increasing PFC A concentrations in soils while PFSA concentrations in grain were correlated linearly with PFSA concentrations in soils, indicating that PFC As and PFSAs may have different transport pathways from soil to grain. Straw: PFAA concentrations decreased significantly with increasing chain length 4-fold increase in spiking (1.00 mg of the individual substances/kg soil versus 0.25 mg/kg soil) leads to a 4-fold (Wen et al., 2014)    Soil studies field study (biosolid amended soils; biosolids applied = 4.5, 9, 18, 36 t/(ha∙y)) wheat (Triticum aestivum L.) (root, straw, husk, grain) PFC As (C 4-C 11, C 14) PFSAs (C 4, C 6, C 8)      pot experiment (soil was spiked with an aqueous solution of 0.25 maize (Zea mays) (straw, kernel) PFC As (C 4-C 10)  PFSAs (C 4, C 6, C 8)  (Krippner et al., 2015) 451 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) substances mg individual PFAA/kg soil and 1.00 mg individual PFAA/kg soil). After 128 days straw and kernels were harvested. Results      field study (grass from biosolidapplied fields) greenhouse study with biosolidsamended soil (industrially impacted soil (PFAA contaminated biosolids), grass (tall fescue, barley, Bermuda grass, Kentucky bluegrass) PFC As (C 6-C 14) PFSAs (C 4, C 6, C 8) n:2 FTOH (n= 6, 8, 10, 12, 14), n:2 sFTOH (n= 7, 9, 11, 13), 8:2 FTA lettuce (Lactuca sativa) and tomato (Lycopersicon lycopersicum) (lettuce leaf and tomato fruit) PFC As (C 5-C 10) PFSAs (C 4, C 6, C 7, C 8, C 10)      Reference higher concentration of PFAAs with chain lengths ≥C 6 in the straw of maize plants. 4-fold increase in spiking concentration leads only to a 2-fold increase of C 4 PFC A, C 5 PFC A and C 4 PFSA in the straw. Kernel: only C 4-C 7 PFC As and C 4 PFSA were detected (highest concentration for C 5 PFC A) C oncentrations of PFAAs detected in kernels were lower than those measured in the straw PFC As are always found in higher concentrations than PFSAs Highest soil-to-plant transfer for both straw and kernels was determined for short-chain PFAAs. The PFAA transfer from soil to straw decreased with increasing chain length. Transfer factors in straw was >1 for C 4-C 7 PFC As and C 4 PFSA. Transfer factors in kernels were all below 1. PFAAs accumulate to a greater degree in vegetative plant parts and only to a small degree in generative organs. The shortest PFC A had the highest grass /soil accumulation factor (GSAF) and the accumulation potentials decreased with increasing carbon chain length (C 6 to C 9 mean decrease of 32-fold, C 9 to C 14 mean decrease 2-fold) The accumulation potential is higher for C 8 PFC A than for C 8 PFSA (GSAF for other PFSAs were not provided) FTOHs, sFTOH were quantifiable in only a few plant samples and only at very low concentrations compared to PFC As (e.g. 8:2 FTOH 10-fold lower concentrations than C 8 PFC A) (Yoo et al., 2011) Greenhouse study: The BAFs for PFAAs in greenhouse lettuce decreased approximately 0.3 log units per additional C F2 group (C 4 PFSA excluded from regression calculation) BAFs in lettuce in industrially impacted soil higher than in municipal soil. BAFs (lettuce, municipal soil) = 0.19 – 28.4, BAFs (lettuce, industrially impacted soil) = 0.52 – 56.8 (C 10 PFDS < LOQ); greatest accumulation was seen for C 4 PFC A No apparent linear trend for PFAA log BAFs in tomato. BAF (Blaine et al., 2013) 452 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) substances Results municipal soil)  limited-scale field study (fertilization via biosolids with different agronomic rate for nitrogen (0.5x, 1x, 2x, 4x))      greenhouse study with biosolidsamended soil (industrially impacted biosolids and municipal biosolids) radish (Raphanus sativus), celery (Apium graveolens var. dulce), tomato (Lycopersicon lycopersicum), and sugar snap pea (Pisum sativum var. macrocarpon) PFC As (C 4-C 10) PFSAs (C 4, C 6, C 8)     Reference decreases approximately 0.5 – 0.9 log units (C 4 PFC A excluded from regression calculation) BAFs (tomato, industrially impacted soil) = 0.1 – 17.1 (C 9, C 10 PFC As and C 7, C 8, C 10 PFSAs < LOQ); greatest accumulation was seen for C 5 PFC A Bioaccumulation of PFAAs from biosolid-amended soils depend strongly on PFAA concentrations, soil properties, the type of crop, and analyte transpiration stream concentration factor (TSC F) lettuce: except for C 4 PFC A in municipal soil (TSC F ~ 1.25) the TSC F values were less than 1 for all analytes and both biosolidsamended soils Limited-scale field study: BAFs decreased with increasing chain length (only calculated for 4x plot) Lettuce: BAFs could only be calculated for C 4 PFC A (40.0), C 5 PFC A (16.3), C 4 PFSA (2.02), C 6 PFSA (1.51), C 8 PFSA (0.1); concentrations for all other PFAAs were below LOQ Tomato: BAFs could only be calculated for C 4 PFC A (18.2), C 5 PFC A (14.9) and C 6 PFC A (6.84); concentrations for all other PFAAs were below LOQ Root-soil concentration factors (RC F) for tomato and pea were independent of PFC A chain length, while radish and celery RC Fs showed a slight decrease with increasing chain length (chain length trends were not calculated for PFSAs as only three analytes were studied). PFC As: Shoot-soil concentration factors (SC F) for all crops showed a decrease with increasing chain length (0.11 to 0.36 log decrease per C F2 group). PFC As: Fruit-soil concentration factors (FC F) decreased with increasing chain length (0.54−0.58 log decrease per C F2 group). Fruit crops were found to accumulate fewer long-chain PFAAs than shoot or root crops presumably due to an increasing (Blaine et al., 2014a) 453 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Outdoor lysimeter (soil spiked with a mixture of PFAAs – nominal concentration of each PFAA: 0.1 mg/kg dw, 1 mg/kg dw, 5 mg/kg dw, 10 mg/kg dw) climate chamber, two biosolid amended soils (anaerobically digested thermal drying sludge Plant species (plant parts) (radish and celery: root and shoot; tomato and pea: root shoot and fruit) Radish (Rapahnus sativus), lettuce (Lactuca sativa), pea (Pisum sativum) and maize (Zea mays) (radish: roots, bulb, foliage; lettuce: roots, foliage; pea: roots, stem, twigs, leaves, pods, peas; maize: roots, stem, leaves, hull leaves, cobs, kernels) substances spinach (Spinacia oleracea), tomato (Solanum lycopersicum PFC As (C 4-C 16) PFSAs (C 4, C 6, C 8, C 10) FOSA, N-MeFOSA, N-EtFOSA Results  PFC As (C 4-C 14) PFSAs (C 4, C 8)       Reference number of biological barriers as the contaminant is transported throughout the plant (root to shoot to fruit). Edible parts: radish root (RC F >1 for C 4, C 6, C 9 PFC A and C 4, C 6 PFSA), celery shoot (SC F >1 for C 4-C 7 PFC A and all three PFSAs), tomato fruit and pea fruit (FC F >1 for C 4-C 6 PFC As) Phytotoxic effects of PFAAs: radish and lettuce plants grown in the highest exposure level soil were smaller at the time of harvest compared to those growing in lower exposure levels. Pea and maize plants showed no visible effects of phytotoxicity. Edible part/soil concentration factors ranged over seven orders of magnitude and decreased strongly with increasing PFAA chain length, by a factor of 10 for each additional C F2 group for pea. Root retention factors increased by a factor 1.7 for each C F2 group. Fruit/leaf concentration factors decreased by a factor 2.5 for each C F2 group. Independent of the plant species the highest concentrations were found in leaves and roots of the plants and the lowest in the fruits. Therefore, leafy and root vegetables pose the highest risk for dietary exposure followed by fruit-bearing crops. (Felizeter et al., 2021) Tomato: C 8 PFSA (75%) and C 7–C 10 PFC As (54–96%) preferentially remained in roots and C 4–C 6 PFC As tended to be translocated to above-ground tissues (leaf: 31–56%, fruit: 32–48%). Transfer factor values for PFASs in fruit were >1 for C 4 (30.87 and 69.82), C 5 (31.22) and C 6 PFC As (3.64). Transfer factors between tomato plant parts were determined (Navarro et al., 2017) 454 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method and anaerobically digested municipal solid waste compost); exposure time: 28 days (spinach) and 6 months (tomato) Plant species (plant parts) L.), (tomato: root, stem, leaf, fruit) substances Results   climate chamber, soils fortified by addition of technical mixture of C 8 PFSA (~50 mg/kg soil); exposure time: 28 days maize (Zea mays) (root, leaf) greenhouse, water-soil-plant system, derived from a surface flow constructed wetland; surface water enriched with PFAA at two different concentrations (low PFAA concentration 0.18–4.22 µg/L; high PFAA concentration 64-4 300 µg/L) pot experiments bulrush (Typha angustifolia) (root, shoot) technical mixture of C 8 PFSA   PFC As (C 4-C 8) PFSAs (C 4, C 6, C 8)     wheat PFC As (C 4, C 6, C 8)  Reference to evaluate the translocation and distribution within the plant: leaf concentration factor for C 5–C 8 PFC A >1 (2.69 – 6.92), edible part concentration factor only for C 5-PFC A >1 (1.12) Spinach: only C 5, C 8 PFC A and C 8 PFSA were detected (transfer factor 1.08 – 4.47) C 6 PFSA, C 10 PFSA, C 13-C 16 PFC A, FOSA, N-MeFOSA and NEtFOSA were not detected in any sample. High levels of C 4 and C 6 PFSA were also detected in roots and leaves as the commercial mixture of C 8 PFSA also contains ~ 1.5% of C 4 and C 6 PFSAs. C 8 PFSA is highly presented in roots (89%), whereas C 4 PFSA (88%) and C 6 PFSA (82%) were preferentially found in leaves. The transfer factor values in roots were highest for C 8 PFSA (8.82) followed by C 4 PFSA (5.00) and C 6 PFSA (2.62). Transfer factor values in leaves were: 9.39 (C 6 PFSA), 4.00 (C 4 PFSA) and 0.80 (C 8 PFSA). Longer-chain PFAAs showed higher root uptake potential compared to shorter-chain PFAAs. PFSAs exhibited higher concentrations in the roots compared to PFC As. BAFroot/water = 8.1 – 45.7 at low concentrations and 0.7 – 15.6 at higher concentrations; BAFshoots/water = 7.3 – 26.6 at low concentrations and 1.4 – 3.5 at higher concentrations. BAFplant/water decreased with increasing PFAA initial concentrations. A positive correlation between BAFplant/soil and chain length was observed. Translocation factor from roots to shoots decreased with increasing chain length (TF >1 for C 4-C 7 PFC As and C 4 PFSA). Higher translocation factors were observed for PFC As compared to PFSAs with similar chain length. Both the root and shoot accumulated PFAAs from soil. The (Navarro et al., 2017) (Zhang et al., 2020b) (Lan et al., 455 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method in greenhouse with wheat-soil system (PFAA concentrations: 200 and 2 000 µg/kg soil) greenhouse microcosm experiment (duration 5.5 months; four different treatments: 1. Soil without biosolids; 2. WWTP biosolidsamended soil (16 g biosolids/kg soil); 3. WWTPand paper fibre biosolidsamended soil Plant species (plant parts) (Triticum aestivum L.) (root, shoot) substances Results Reference PFSAs (C 4, C 6, C 8) 2018) alfalfa plants (Medicago truncatula) pumpkin (Cucurbita maxima) (root, stalk, flower, leaf, fruit) DiPAPs PFC As (C 4-C 14) accumulation of PFAAs was enhanced at the higher spiking levels of PFAAs and increased with decreasing chain length. PFC As showed higher accumulation in wheat than PFSAs of the same chain lengths, which corresponds to their differences of log Kow values.  Root: BAF values were all greater than 1. BAF values for PFC As decreased with increasing chain length, while the BAF were at similar level for PFSAs  Shoot: BAF values were greater than 1 for C 4 and C 6 PFAAs.  Transfer factor from root to shoot: no significant difference was observed between the two spiking levels. For C 4 and C 6 PFAAs the transfer factors were greater than 1 (C 4 PFC A: 11.1 and 11.9; C 4 PFSA: 4.51 and 1.86; C 6 PFC A: 2.6 and 1.9; C 6 PFSA: 1.04 and 1.28) indicating their higher accumulation potential in shoot. The limited transfer potential of long-chain PFAAs from root to shoot may be due to their low solubility and greater interaction with biological macromolecules (e.g. protein and lipid) in root. Greenhouse study:  DiPAPs (6:2 diPAP, 6:2/8:2 diPAP, 8:2 diPAP, 8:2/10:2 diPAP, 10:2 diPAP, 10:2/12:2 diPAP) as well as PFC As (C 4-C 13) were detected in control soil and biosolid-amended soils prior application.  The majority of the 6:2 diPAP resided in the soil (99%), with minor uptake observed in the plants (1%)  FTOHs were not analysed and the microcosms were open to the atmosphere of the greenhouse.  The following transformation products were observed in the soil-plant microcosm: 6:2 FTC A, 6:2 FTUC A, 5:3 FTC A, C 4-C 6 PFC As  The plant-soil accumulation factor for PFC As decreased with increasing carbon chain length Field study:  several PFC As were present in the field soil collected prior to 456 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method (mixture 1:4); 4. 6:2 diPAP-spiked WWTP biosolidsamended soil) and field experiment (compost and paper fibre biosolids (1:4) were applied to two farmfields) pot experiment with two different compost amended soils (soil 2.4 and substrate, 8:2 diPAP concentration nominal 500 ng/g) Plant species (plant parts) substances Results   lettuce (Lactuca sativa) and carrot (Daucus carota ssp sativus) 8:2 diPAP + degradation products    (carrot: peel, core, leaf; lettuce: leaf, heart) (growth periods: carrot ~ 3 months, lettuce ~1 month)  climate chamber with plant-soil longhair sedge (Carex PFAS-ether: 2,2,3,3-Tetrafluoro-3-  Reference the application. No diPAPs were detected. DiPAPs can migrate to lower soil depths. 2 months after application 6:2/8:2 diPAP, 8:2 diPAP, 8:2/10:2 diPAP were observed at a soil depth of 5-10 and 6:2 diPAP even at a depth of 10-15 inches Uptake of diPAPs and PFC As in the root, stalk, flower, leaf and fruit of pumpkins were observed after 3.5 months In the presence of crops different degradation products were detected in the soil compared to the experiment in absence of crop. Degradation products in the absence of crop: 8:2 monoPAP, 8:2 saturated and unsaturated fluorotelomer carboxylate (8:2 FTC A and 8:2 FTUC A), 7:3 FTC A, C 6-C 8 PFC As C arrot experiments: degradation products: C 4-C 9 PFC As, 7:3 FTC A; Lower 8:2 diPAP concentrations were observed in the carrot parts compared to the soil. This could mean that the substance retained to the soil or that metabolization occurs. The highest concentrations of PFC As were found in leaves. BC F and translocation of PFC As increased with increasing solubility. Independent of type of soil BC F values decreased with increasing chain length. Lettuce experiments: only C 8 PFC A was detected as degradation product in the heart of the lettuce. No translocation through the plant was observed as in the leaves concentrations of 8:2 diPAP and C 8 PFC A were below method detection limits (MDL for 8:2 diPAP = 3 ng/g; MDL for C 8 PFC A not mentioned). (Bizkarguenaga et al., 2016) All PFAS-ether were taken up by plant roots, translocated to shoots, and accumulated in plant tissues. Exposure 457 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method system (solutions of PFAS-ether at 500 or 2 000 ng/L of each PFAS-ether; sampling at day 52 and 80) pot experiment (soil was spiked with a stock solution of target PFASs to achieve nominal concentrations of 200 ng/g and 500 ng/g; exposure time: 60 days) Plant species (plant parts) comosa) (root, shoot) wheat (Triticum aestivum L.), maize (Zea mays L.), soybean (Glycine max L. Merrill), and pumpkin (Cucurbita maxima L.) (Wheat: root, substances (trifluoromethoxy)propionic acid (PFMOPrA), perfluoro(4-methoxybutanoic) acid (PFMOBA), 2,3,3,3-tetrafluoro-2(heptafluoropropoxy)propionic acid (HFPO-DA), ammonium 2,2,3 trifluoro-3(1,1,2,2,3,3-hexafluoro-3trifluoromethoxypropoxy), propionate (ADONA), Potassium 2-(6-chloro1,1,2,2,3,3,4,4,5,5,6,6dodecafluorohexyloxy)1,1,2,2tetrafluoroethane sulfonate (F-53B) 6:2 fluorotelomer sulfonate (6:2 FTS), 6:2 chlorinated polyfluoroalkyl ether sulfonates (6:2 C l-PFESA, trade name F-53B) and perfluorophosphinates (C 6/C 6 and C 8/C 8 PFPiAs) Results      Reference concentration and time positively affected the plant uptake of the PFAS-ether. The uptake in shoots was significantly higher for the other PFAS-ether than for F-53B. F-53B, which has the longest carbon chain among the PFAS-ether in this study and a sulfonic functional group, was largely accumulated in roots with very limited upwards translocation. Plants exposed to HFPO-DA (500 ng/L, 52 days) had the highest translocation factor. Only F-53B had a translocation factor less than 1. The relatively higher hydrophobicity and lipophilicity of F-53B (log Kow =7.03) could cause greater interactions between the substance and the biological macromolecules in plant roots (e.g. proteins, lipids). This could lead to decreasing penetration of F-53B through the C asparian strip. The concentration of PFAS-ether in water-soluble fraction increased with decreasing carbon chain length and logKow values and had a positive linear relationship with PFAS -ether mass in whole plants and plant shoots. PFAS-ether that had higher concentration in water-soluble fraction (e.g. PFMOPrA and PFMOBA) may have higher leachability to surrounding environments. Aging process could facilitate PFAS -ether to become non-extractable, hence reducing their mobility in soil and bioavailability to plants. All four PFASs were detected in all compartments of all four plant species in both spiked treatments. The concentration of the four PFASs in the plant roots were significantly higher than those in shoots and leaves. The mean concentrations in the plant roots and aboveground tissues decreased in order of 6:2 FTS > F53-B > C 6/C 6 PFPiA > C 8/C 8 PFPiA with increasing hydrophobicity (increasing log Kow). The root soil concentration factor as well as the translocation factor from root to shoot were inversely proportional to their log Kow for all four plant species. All translocation factor values were below 1. (Zhou et al., 2020) 458 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Method Plant species (plant parts) shoot; maize, soybean, pumpkin: root, shoot, leaf) substances Results Reference Pot experiment in outdoor transparent shelter (irrigation with PFAS contaminated river water from Lee's C reek (AFFF contamination, <1 µg/L PFAS) and PFAS-free water) Tomato (Solanum lycopersicum), lettuce (Lactuca sativa) and beet (Beta vulgaris ssp. vulgaris) (tomato: root, stem, foliage, flower, foliage; lettuce and beet: root, foliage) PFC As (C 4-C 12) PFSAs (C 4, C 6, C 8, C 10), PFOSA (perfluorooctanesulfonamide)  (McDonough et al., 2021a)    Short-chain PFASs: tomato plants (exception of roots) showed significant increases of short-chain PFAS concentrations with river treatment (highest concentration in flowers). This effect could not be observed in lettuce or beet. Long-chain PFASs: C oncentration of long-chain PFASs only increased in tomato foliage and lettuce roots and foliage. Tomato: In small fruits higher concentrations of short-chain PFAS but lower absolute PFAS mass were observed. Whereas in larger fruits lower concentration of short-chain PFAS but higher absolute PFAS mass was detected. Biomagnification of short-chain PFASs in flowers could have implications for pollinators depending on where the contaminants reside (e.g. in pollen or nectar, exposure to insects could be meaningful). 459 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.5.1.1.2. Distribution Table B.91 to Table B.95 summarise human biomonitoring studies of PFASs. The tables include precursors and arrowhead PFASs that are not covered by existing or proposed restrictions. Thus, data for PFOA, PFOS, PFHxA, PFHxS and C9-C14 PFCAs and related precursors are not included in these tables. Table B.91. Serum (S), plasma (P) or whole blood (WB) concentrations (ng/mL) of PFAAs (PFSAs, PFCAs and PFPAs) that are not covered by existing or proposed restrictions described by detection frequency (% or x/xx), median (bold) and range (within brackets). No te that the levels of whole blood are not comparable to plasma/serum. Grey cells indicate sampling sites near known PFAS point sources. NR=not reported, w pp=weeks postpartum; P=pool; cont drw=contaminated drinking water, FPP=fluoropolymer production plant, FCM=fluorochemical manufacturing facility, ref=reference area. Year Age (gender) Country, area Sam ple N PFBA PFPeA PFHpA 1977 40-50 (M) 2006 S Norway 24 P 0% 54% 63% 0.076 0.081 (<0.05- (<0.050.14) 0.63) 1982 20-29 (F,M) 2009 S Germany, Muster, Halle (<0.1) 420PFDS 122PFHp S 420PFHp A 100% (0.019-2.2) PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As 67% 83% 0% (Haug 0.07 et al., 0.20 4 (<0.05 2009) (<0.05 (<0. -0.47) ) 050.18 ) 100% (0.0542.0) 46% Munster (<0.00 50.084) 69% Halle (<0.00 50.097) (Yeung and Mabury , 2016; Yeung et al., 2013b; Yeung et al., 2013c) 460 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age (gender) Country, area Sam ple N 1982 20-29 (F,M) 2010 P 258PFCAs11% 2% 158PFSAs (<0.5-1.4) (<0.514) 53% (<0.5-2.3) 0% (<0. 5) 1996 Primipara, 4 w Sweden, pp Uppsala 2010 S 36 P (5-25 /P) 100% (0.0560.14) 100% (0.011 -0.26) 1997 20-21 (F,M) 2004 P Germany, Arns-berg (cont. drw) 30 0% (<1.0) 72% (<0. 0130.10 ) 0% (<0. 1) 1997 20-42 Primipara, 4 w pp 2012 S Sweden, Uppsala 27 P 0% (10/P) (<0.3) 0% (<0.1) 70% (<0. 0090.05 8) 81% (<0.00 50.091) 2000 2008 P Norway, National survey 3174/ 4295 11% (<0.050.077P95) 2001 70 (F,M) 2004 Sweden, Uppsala 1006 0% 74% 0.05 (<0.02) (F) 78% 0.06 Germany, Munster PFBA PFPeA PFHpA 100% (0.0140.10) PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As 45% (<0.52.3) 0% (<0.5) (Wilhel m et al., 2009) 90% 0.15 (<0.05 0.34P95 ) 14% (F) 11% (Schrot erKerman i et al., 2013) (Glynn et al., 2012) (Gebbin k et al., 2015) (HBM4E U, 2022) 0% (<0.03 ) (Saliho vic et al., 461 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age (gender) Country, area Sam ple P 2002 Newborn (F,M) 2005 C ord P 2002 14-15 (F,M) 2005 S 2002 50-65 (F,M) 2005 S 2003 Adults S (F,M) N PFBA PFPeA PFHpA (M) Belgium, 8 locations Norway, High fish consumption 8P (48- 0% 162/P) (<0.1) PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As (M) 0% (<0. 02) 8P(149 0% -186/P) (<0.1) 0% (<0. 02) 8P 0% (182- (<0.1) 197/P) 0% (<0. 02) 175 37% 99.5% 0.42 (<0.03 5-2.5) (<0.0350.49) 2003 pregnant 2008 S Spain 1216 2005 Pregnant P Denmark, Aarhus 88 2005 1-100 USA (OH & WV) 69 030 2015) (Roose ns et al., 2010) 0% 65% 0.08 (<0.051.95) 4.9% 38% 100% 0.25 (0.030.7) (Haug et al., 2010b) (Manza noSalgado et al., 2016) (Bach et al., 2015) (Frisbe 462 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age (gender) Country, area Sam ple N PFBA (F,M) 2006 S 2006 47-80 S (M) Cont. drw Faroe Island Whalers 10 2006 Adults (F,M) WB Spain, C atalonia 48 2006 C hildren P (F,M) Germany, Siegen (ref) 80 0% (<1) 153 0% (<1) Mothers PFPeA PFHpA (<0.5NR) (<0.5-NR) 60% 0.39 (<0.030.49) 0% (<0.78) Adults (M) Germany, Brilon (ref) 103 0% (<1) C hildren (F,M) Germany, Arnsberg 90 0% (<1) PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As e et al., 2009) 100% 0.65 (0.161.1) 0% (<0. 73) 5% (<0. 10.1P9 5) 0.7 % (<0. 1NR) 3% (<0. 1NR) 60% 0.07 (<0.03 -0.09) 0% (<0.58 ) (Hu et al., 2018) (Ericso n et al., 2007) (Holzer et al., 2008) 33% (<0. 10.2P9 463 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age (gender) Country, area Sam ple N PFBA PFPeA PFHpA PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As 5) Mothers Cont. drw 164 0% (<1) 4% (<0. 1NR) 101 0% (<1) 13% (<0. 10.2P9 5 ) Spain, Barcelona 30 20% (80% 0.6 (10/P ) (0.0380.057) 2009 25-42 S (F) 0% 100% 0.083 ⱡ 30% (<0.00 (0.0252) 0.16) (<0. 010.05 9) 0% (<0. 78) 0% (<0. 5) Belgium, Antwerp 20 0% (<1.3) 0% (<1.4) 2009 20-29 Germany, (F,M) Munster 2019 P 2009 32.3 Pregnant Japan 2010 S 100 0% (<0.25) 5% 4% (<0.25- (<0.250.31) 1.1) 0% (<0.5) 399 0% (<0.25) 6% 0% (<0.19) (<0.15) 0% 0% 1% (<0.13) (<0.14) (<0. 080.11 ) 2009 12-79 1524 C anada 0% (<0.5) 0% 6% (<0.25 -0.42) 3% (<0.14 -0.46) PFH xPA 100% : 0% 0.051 (<0. ⱡ 009) (0.041 PFO PA: 0.082) 0% (<0. 005) PFD PA: 0% (<0. 07) 0% (Petro (<0.57 et al., ) 2014) 0% (Gocke (<0.25 ner et ) al., 2020b) 0% (Nakay (<0.13 ama et ) al., 2020) (Health 467 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age (gender) Country, area Sam ple (F,M) 2011 P 2010 15-19 2011 S 2010 18-80 (F) 2011 S 2010 Newborns (F,M) 2012 C ord S N PFBA PFPeA PFHpA PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As (<0. 4) 98% 0% 14% 0% (<0.01 0.15 (<0.02 (<0.01 -1.2) ) -7.6) Norway, Troms arctic 940 75% 0.11 (<0.07-1.5) Sweden 270 4.4% (<0.1-NR) Slovakia, National 322 6% (<0.010.019P95) 13% 20% (<0.01- (<0.010.04P95) 0.09P95) 0.3 % (<0. 01NR) 2010 20-46 Pregnant 2013 S C ord S France 100 0% (<0.38) 0% 0% (<0.24) (<0.24) 1% (<0. 2NR) 1% (<0. 2NR) 50% 0.18 ⱡ 2010 18-44 Pregnant 2013 Greenland 0% 87% 0.19 (0.06- 0% (<0.38) 209 0% 0% (<0.24) (<0.24) 0% 16% (12 S (F,M) Belgium, Zwijndrecht <3km from 3M facility N 796 PFBA PFPeA PFHpA (<0.066.3) (<0.09- (<0.03-2.0) 0.60) 4.5% 0% (<0.2-NR) (<0.2) 11% (<0.20.53P95) PFHxD PFOcD PFB PFPeS PFHpS PFNS PFDS Oth Refere A A S er nce PFC As/ PFS As/ PFP As (<0. 051.6) 2.6 % (<0. 2NR) al., 2020b) 30% (<0.21.0P95) (VITO and PIH, 2021) ⱡ Arithmetic mean, ¥ Geometric mean, § Mean per area. 485 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.92. Serum (S), plasma (P) or whole blood (WB) concentrations (ng/mL) of PFAEs , cyclic PFAS and precursors that are not covered by existing or proposed restrictions, described by detection frequency (% or x/xx), median (bold) and range (within brackets) . Note that the levels of whole blood are not comparable to plasma/serum. Grey cells indicate sampling sites near known PFAS point sources. NR=not reported, P95=95 th percentile, w pp=weeks postpartum; P=pool; cont. drw =contaminated drinking water, FPP= fluoropolymer production plant, FCM=fluorochemical manufacturing facility, ref=reference area. Other PFASs Reference 4:2 3:3 PFECH C6/C Country N 6:2 Cl8:2 ClADON HFP Age Year FTSA FTC S 6 PFESA PFESA A O(gen , area Sam /FTS PFPiA A DA der) ple 1982 2009 S 1996 2017 S 2029 (F,M) 1997 2012 S 2042 Primi para, 4w pp 2029 (F,M) 2009 2019 P 2009, 2015 P 2014 , 2016 Primi para, 4 w. pp Germany , Munster, Halle Sweden, Uppsala Sweden, Uppsala Germany , Munster 32 0 57 P (525 /P) 27 P (1 0/ P) 5% (<0.010.017) 10 0 0% (<0.25) 0% (<0.09) 0% (<0.25) 0% (<0.00 4) 0% (<0.25 ) Adults Germany 86 (F,M) Near FPP 9% (<0.214.4) Germany 110 80 km from FPP Adult Germany 20 0 ,Munich s (F,M) Ref. 16% (<0.20.7) 0% (<0.2) 0% (<0. 72) 0% (<0. 25) NR% (0.060.28) 0% (<0.25 ) 0% (<0.0 01) 0% (<0.5) 4:2/4:2 diPAP: 40% (<0.0007-0.095) (Yeung et al., 2013c) 3:3 FTA (FPrPA): 0% (<0.34) (Miaz et al., 2020) 4:2 monoPAP: 0% (<0.3) 4:2/4:2 diPAP: 0% (<0.001) (Gebbink et al., 2015) 7H-PFHpA: 0% (<0.5%) (Gockener et al., 2020b) (Fromme et al., 2017) 486 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Sam ple P Age (gen der) 2014 WB 20-42 (F,M) 2016 2017 S 2017 P 2017 2019 S 2017 2019 S 8:2 ClPFESA ADON A HFP ODA 2018 N 6:2 ClPFESA PFECH S Sweden,Ro 20 0/20 0/20 nneby (<0.01) (<0.01) Cont. drw Sweden, 10 5.4% 0.6% 107 cities 98 (<0.005 (<0.06/ 21 /<0.06<0.03(F,M) 1.2) 0.03) Germany 72 30, Berlin 57 (F,M) 0/20 (<0.02) 0/20 3/20 (<0.02 (<0.012) 0.06) 2145 Primi para Preg nant (at deliv ery) C6/C 6 PFPiA 3:3 FTC A 4:2 FTSA /FTS Other PFASs 0% 0/20 0/20 FBSA: 0% (<0.0 (<0.005 (<0.015/<0.03) 5) ) MeFSBA: 0% (<0.21) 0% 0.1% (<0. (<0.0 1/< 612) 0.56) Reference group 0.1% (<0.02 /<3.60.07) 0% (<0.25 ) 11 0 1% (<0.080.2) 0% (<0.14) 0% (<0.08 ) Austria, 3 areas (samples with highest 54 in 21 P 0% (<0.000 2/ <0.004) 0% (<0.000 4/ <0.006) PFASlevels out of 136 samples) 27 in 11 P 0% 0% (<0.000 2/0.004 ) (<0.000 4/0.006 ) 1179 (F,M) Sweden, 5 areas 14 8 12% 0% 11/21 0.006 (<0.00 06/ <0.010.074) 8/11 0.008 (<0.00 06/ <0.010.027) 16% 0% 80% 0.02 ⱡ (<0.010.02) (<0.01) (<0.02 -0.04) (<0. 022) (<0.01 -0.11) 7-85 Germany, 906 0.3% (Aro et al., 2022) (Nystrom et al., 2022) (Menzel et al., 2021) Sweden, Uppsala C ord S 2018 WB Country , area 0% 0% (<0.0 8) 3:3 FTA (FPrPA): 0% (<0.29) (Gyllenhammar et al., 2020) 0% (<0.0 002/ <0.00 4) FBSA: 0% (<0.001/<0.02) MeFBSA: 0% (<0.008/<0.12) (Kaiser et al., 2021a) 0% (<0.0 002/ <0.00 4) FBSA: 0% (<0.001/<0.02) MeFBSA: 0% (<0.008/<0.12) FBSA: 2% (<0.015/<0.030.05) MeFBSA: 0% (<0.21) 16/19 0.003 ( 10 /P) 20 6:2 ClPFESA 8:2 ClPFESA ADON A HFP ODA PFECH S C6/C 6 PFPiA 3:3 FTC A 4:2 FTSA /FTS Other PFASs (<0.251.0) 0% (<0.25 ) Reference 2018) 0% (<0.06 ) 0% (<0.1 0) 0% (<0.06 ) 0% (<0.1 0) (Toms et al., 2019) 50% (<0.0 010.05) 10% (<0.0 050.018) 4:2 diPAP: 12% (<0.008-0.022) 90% (<0.0 010.05) 0% (<0.0 05) 4:2 diPAP: 0% (<0.008) Nafion by-prod.2: 0% (<0.1) PFO4DA: 0% (<0.1) PFO5DoA: 0% (<0.1) (Lee and Mabury, 2011) (Kotlarz et al., 2020) 488 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Age Country N Year (gen , area Sam ple der) 2017- 6-86 USA, 344 2018 (F,M) Wilmington S (NC ) Cont. drw 6:2 ClPFESA 8:2 ClPFESA ADON A 10 0 NR year S 2016 S 1780 (F,M) NR USA, Kentuck y USA 50 0% (<0.1) 0% (<0.1) 2018 S 14-79 (F,M) USA, 30 Bladen county, Area near FCM. Bottled water last 0% (<0.1) 0% (<0.1) HFP ODA 0% (<2) 0% (<0. 05) 0% (<0. 1) 0% (<0.1) PFECH S C6/C 6 PFPiA 3:3 FTC A 4:2 FTSA /FTS Other PFASs Reference Nafion by-prod.1: 0.3% (<0.1-0.4) Nafion by-prod.2: 99% 2.7 (<0.1-8.4P95) PEPA: 9.6% (<0.1/0.5-2) PMPA: 0.3% (<0.4/1.5-1.5) PFO2HxA: 0.3% (<0.5/5.8-1.9) PFO3OA: 28% (<0.1/1.3-3.7) PFO4DA: 99% 2.5 (<0.1-12.8P95) PFO5DoA: 88% 0.3 (<0.1-1P95) NVHOS: 15% (<0.1/0.2-4.6) 0% (<0.0 5) (Mottaleb et al., 2020) (Kato et al., 2018) (NC DHHS, 2018) 489 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Sam ple Age (gen der) Country , area N 6:2 ClPFESA 8:2 ClPFESA ADON A HFP ODA 0% (<0.1) 0% (<0.1) PFECH S C6/C 6 PFPiA 3:3 FTC A 4:2 FTSA /FTS Other PFASs Reference 4-14 months. 2018 S 19-93 (F,M) NR year S USA, El 213 Paso C ounty. Cont. drw 2012-15 C hina, 19 Shandong Metal plating workers C hina, 45 Hubei fresh water fish consumer s C hina, 8 Wuhan, Hubei (ref) 0% (<0.1) NR S 1-91 (F,M) C hina, 7 cities 15 16 2015 2016 S 27.1 ±2.8 Preg nant C hina, Wuhan 32 C ord S 32 100% 51.5 (2.05040) 100% 1.6 (0.3790.7) (Barton et al., 2020) 10:2 Cl-PFESA: 0% (<0.045) 100% NR 93.7 1.6 (2.0-357) (<0.043.4) 10:2 Cl-PFESA: 0% (<0.045) 100% 4.8 (1.95.9) 100% 0.08 (0.040.10) 10:2 Cl-PFESA: 0% (<0.045) 98% 2.2 (<0.04500) 100% 1.5 (0.234.4) 13% (<0.0567.5) 10:2 Cl-PFESA: 0.3% (<0.07-3.3) 100% 0.6 81% 0.01 84% 0.01 (<0.004 -0.21) (Shi et al., 2016) (Jin et al., 2020b) (C hen et al., 2017b) 490 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Sam ple 2015 2016 S Age (gen der) 2441 Preg nant Country , area C hina, Beijing C ord S N 6:2 ClPFESA 8:2 ClPFESA (0.12.6) 98% 3.6 (<0.1NR) (<0.004 -0.08) 31% (<0.1NR) 96% 1.6 (<0.1NR) 100% 1.75 29% (<0.1NR) 69% 0.01 1.2% 100% 4.2 (0.019.9P95) 25 100% 2 8.6 100% 0.06 (0.010.2P95) 70% 0.06 37% (<0.14 -1.7P95) 15 93% 0.42 ⱡ (<0.0261.4) 13% 13% 20% (<0.0380.31) (<0.017 -0.09) (<0.031 -0.08) 60% 0.22 ⱡ 0% 0% 0% (<0.026 -0.31) (<0.038 ) (<0.0 17) (<0.0 31) 100% 1.6 (0.45- 100% 0.017 (0.005- 11 7 11 7 C hina, Shenyan g 12 28 2015 2016 S 61.8 ±14. 4 (F,M) 2016 S 53 C hina, 48 (mean) Huntai (F,M) Near FPP 2017 S C hina, 19Tianjin 87 (F,M) Worke C hina, rs at Wuhan FCM 2017 S 2018 2019 Adult s ref. grou p C hina, Wuhan Adult s (F) C hina, Beijing 15 28 ADON A HFP ODA PFECH S C6/C 6 PFPiA 3:3 FTC A 4:2 FTSA /FTS Other PFASs Reference (Li et al., 2020c) 17% 0.5% (Yu et al., 2021c) 4:2 Cl-PFESA: 98% 0.04 (<0.01-0.1P95) HFPO-TA: 98% 2.9 (<0.1-53P95) 2% (Pan et al., 2017) (Duan et al., 2020) (Gao et al., 2018) HFPO-TA: 25% (<0.05-0.48) (Kang et al., 2020) 491 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Sam ple S Age (gen der) 2019 2020 S 0-96 (F,M) Country , area C hina, Huantai Near FPP N 97 7 6:2 ClPFESA 8:2 ClPFESA 4.3) 0.10) 100% 2.2 ¥ (0.0629.9) 70% 0.03 ¥ (<0.021.8) ADON A HFP ODA 8% (<0. 050.65) PFECH S C6/C 6 PFPiA 3:3 FTC A 4:2 FTSA /FTS Other PFASs Reference 4:2 Cl-PFESA: 4.5% (<0.01-1.4) Nafion by-prod. 2: 99% 0.095 ¥ (<0.01-24) HFPO-TA: 99% 1.8¥ (<0.05-429) PF4OPeA/PFMPA: 17% (<0.01-0.38) PF5OHxA / PFMBA: 0% (<0.02) PFMOAA: 100% 12.7¥ (0.42-158) PFO2HxA: 70% 0.033 ¥(<0.01-12) PFO3OA: 30% (0.05-5.9) PFO4DA: 96% 0.14¥ (<0.02-3.4) PFO5DoA: 100% 0.98 ¥ (0.03-43) 3,6-OPFHpA: 0% (<0.02) (Yao et al., 2020) ⱡ Arithmetic mean, ¥ Geometric mean 492 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.93. Breast milk concentrations (ng/mL) of PFASs (PFAAs, precursors, PFAEs and PFECHS) that are not covered by existing or proposed restrictions, described by detection frequency (% or x/xx), median (bold) and range (within brackets). Grey cells indicate sampling sites near known PFAS point sources. P95=95th percentile, w pp=weeks postpartum, P=pool. Year Sampli Countr N PFBA ADONA 6:2 Cl- 8:2 Cl- Other PFASs Reference PFPeA PFHpA PFBS PFHpS PFDS ng y, area PFESA PFESA 20% 2004,2 Primipa Swede 50 0% 12% 0% (Kärrman et al., 007, ra n, (<0.03) (<0.015- (<0.4(<0.009) 2013) 0.39) 2009,2 Uppsal 0.02) 011 a 0% 2007 France 48 17% 0% 2% 0% (Antignac et al., (<0.07- (<0.07) (<0.07- (<0.07) (<0.07) 2013) 0.13) 0.074) 2010 Primipa Spain, 10 C atalon ra 5-9 w ia pp 40 w Spain, 20 pp Barcelo na C zech 50 Republi c France 30 20102013 <1 w pp 0% (<0.04) 0% (<0.04) 0% (<0.04) 20102013 6 w pp Slovaki 238 2.5% a (<0.01NR) 28% (<0.010.16P95) 1.7% (<0.01NR) 2012 Primipa Spain, 10 Valenci ra. a 18 month pp (mean) 0% 2007 2008 2010 France 61 0% (<0.02) 0% (<0.02) (Karrman et al., 2010) 6/20 (<0.0040.07) 0% (<0.006) 0% (<0.02) 0% (<0.02) 1/30 (<0.01NR) 0% 0% (<0.002) (<0.01) 0% 0% (<0.001) (<0.001) 0% (<0.04) 100% 0.035 (0.0060.15) 0% 0% (<0.006) (<0.006) 0% (<0.03) (Llorca et al., 2010) (Lankova et al., 2013) (Kadar et al., 2011) 0% (<0.04) (C ariou et al., 2015) 0% (<0.005) (HBM4EU, 2022) 3/10 7/10 2/10 1/10 0.067 (<0.0006 (<0.0001 (<0.0006 (<0.0003 -0.009) (<0.0006 -0.025) ) -0.7) -0.45) PFHxDA: 0% (<0.0001) PFOcDA: 0% (<0.0001) (Lorenzo et al., 2016) 493 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year 2016 2016 2017 2019 2013 20102016 20182019 Sampli Countr N ng y, area 2-14 w Swede 10 pp n, Stockh olm PFBA PFPeA PFHpA PFBS 1/10 (<0.010.011) 7/10 (<0.0010.021) PFHpS 6:2 Cl- 8:2 Cl- Other PFASs PFESA PFESA 0% 1/10 0% 0% (<0.001- (<0.003) (<0.002) (<0.002) 0.003) PFDS ADONA Reference (Awad et al., 2020) 0% Primipa Ireland 92 (Abdallah et al., (<0.1) in ra 2020) 16P 3-8 w pp Primipa C zech 232 0% 0% 0% 1.3% 0% 0% 0% 0% HFPO-DA: 0% (C erna et al., 2020) ra Republi (<0.006) (<0.006) (<0.003) (<0.003(<0.03) (<0.015) (<0.015) (<0.015) (<0.015) 2-8 w c NR) pp Primipa USA, 50 0% 74% PFPeS: 8% (Zheng et al., 98% 0% 4% ra Seattle (<0.0001-0.002) 2021) 0.006 0.001 PFNS: 58% (<0.0007 (<0.0002 (<0.003) (<0.0003 (<0.0010.0004 ) -0.046) -0.007) 0.001) (<0.0003-0.001) 4:2 FTS: 14% (<0.0006-0.035) PFHxDA, PFPrS, PFECHS, 4:2 FTOH, 8Cl-PFOS: 0% 67% South 264 82% 42% 40% (Kang et al., 0.05 0.03 2016) Korea C hina 30 10% 43% (<0.010.017) 0% (<0.0030.011) 0.6% 1 w pp C hina, 174 63% 68% 0.002 Hangzh 0.003 ou (<0.003- (<0.002- (<0.003) (<0.010.03) 0.11) 0.02) ⱡ Arithmetic mean, ¥ Geometric mean 0% 27% 100% 0.09 (<0.002) (<0.003- (0.0170.008) 0.98) 100% 0.016 (0.0080.20) 60% 0.003 (<0.0020.09) 20% (Awad et al., 2020) (Jin et al., 2020a) (<0.0010.014) 494 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.94. Concentrations of PFAAs in urine, follicular fluid, semen and cerebrospinal fluid (ng/mL) and hair, nail, placenta and internal organs (ng/g) that are not covered by existing or proposed restrictions, described by detection frequency (% or x/xx), median (bold) and range (within brackets). Grey cells indicate sampling sites near known PFAS point sources . P=pool, FCM=fluorochemical manufacturing facility Reference Year Age Country, N PFBA PFPeA PFHpA PFBS PFPeS PFHpS PFDS Other Sample Gende Area PFCA/PFSA/P FPA r 2010 20-88 C hina, 86 67% (Zhang et al., Urine (F,M) Handan & 0.0008 2013d) Shijiazhu (<0.0002ang 0.019) 2012 Urine 2012 Urine 20132014 Urine 20142015 Urine Hair (matche 5-13 South Korea, Dae-gu 120 0% (<0.17) 1% (<0.110.49) 0% (<0.11) 0% (<0.14) 7% (<0.121.1) 1/9 0% (<0.11) 0% (<0.11) 0% (<0.14) 1% (<0.171.7) 7/9 0.054 39 39/39 0.94 39/39 1.1 29/39 0.016 36/39 1.6 7 7/7 0.29 7/7 0.22 4/7 0.012 4/7 0.25 0% (<0.1) 1.1% (<0.10.3) 4% (<0.0410.06) 0% (<0.1) 2% (<0.0580.064) <0.1% (<0.10.1) 9% (<0.0560.077) 30% (<0.41- 0% (<0.80) 0% (<0.28) 57 Refere C hina, Hubei nce group Fisher y emplo yees Fisher y family >6 USA, (F,M) National 9 South Korea, Seoul & Busan 36% (<0.124.4) (<0.2212) 25% (<0.2217.6) 9/9 0.35 22-77 2-82 70% 0.35ⱡ 268 13% 2 (<0.13.4) 94 84% 1.7 (<0.139.3) 59% 0.42 1/9 (Kim et al., 2014) (Zhou et al., 2014) (C alafat et al., 2019) 0% (<0.12) (Kim et al., 2019b) 2% (<0.42495 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age Country, N Sample Gende Area r d) PFBA PFPeA PFHpA 1.0) 98% 0.016 (<0.0000 3-0.05) (<0.362.3) 93% 0.014 (<0.0000 2-0.06) 95% 0.0042 (<0.0000 1-0.02) 95% 0.014 (<0.0000 07-0.04) Hair (matche d) 34% (<0.093.8) 10% (<0.116.9) 17% (<0.032.9) Nails (matche d) 0% (<0.22) 0% (<0.34) 0% (<0.019) 63% 0.24 (<0.0122.7) 85% 0.80 (<0.053.2) 8/11 0.003 ⱡ 10/11 0.0004 ⱡ 2015 Urine 2016 Urine 3-68 (F,M) 25-46 (F,M) 41 C hina, Shijiazhu ang Austria 11 2016 Urine NR USA 50 20172018 Urine 4-6 (F,M) Hong Kong 53 Hair (matche d) 2018 Urine 56% 0.2 (<0.10.8) USA, Bladen county 30 0% (<0.1) 0% (<0.1) PFHpS PFDS Reference Other PFCA/PFSA/P FPA (Wang et al., 2018b) 0% 75% 0.0006 (<0.0000 1-0.013 70% 0.023 (<0.0020.080) 0% (<0.1) PFPeS 7.4) (<0.0005- (<0.00020.008) 0.001) 0% 0% 0% (<0.1) (<0.1) (<0.1) 27 14-79 (F,M) PFBS (Hartmann et al., 2017) 0% (<0.0003) (<0.0003) 0% (<0.1) PFPrS: 0% (<0.1) (Kato et al., 2018) (Li et al., 2021c) 0% (<0.1) (NC DHHS, 2018) 496 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age Country, N PFBA Sample Gende Area r (NC ) Near FCM. Bottled water last 414 months. 2018 24-61 Sweden, 17 Urine (F,M) Arvidsjau Airpor r t Cont. worke drw rs 2019 8-12 C hina, 189 Urine (F,M) Shainghai 20192021 Urine 6-10 (F, M) 2020 Urine 50.3±9 C hina, 80 .65 Shanxi & (F,M) Shandon g 49 19-26 C hina, (F,M) Guangdo ng 2009 Hair 2013 Hair NR Austria Belgium 85 30 77% 0.21 ⱡ PFPeA PFHpA PFBS PFPeS PFHpS PFDS 100% 0.072 (0.0210.12) 0.025 (<0.010.08) 99.5% 87% 0.046 0.021 (<0.005- (<0.0121.0) 0.062) 99% 96% 0.0015 0.0035 (<0.0005- (<0.00020.007) 0.037) 12% 44% (<0.02(<0.01NR) NR) (Xu et al., 2020c) (Li et al., 2021a) 0% (<0.01) 1.2% 0% 0% PFNS: 0% (Kaiser, 2021) (<0.00040.0014) (Ji et al., 2021) 40% (<0.4118) 44% (<0.02212) 30% 0% 100% 0.055 ⱡ 0% 0% 3% (<0.017) (0.030.12) (<0.008) (<0.004) (<0.0060.03) (<0.019- (<0.0170.086) 1.5) Reference Other PFCA/PFSA/P FPA (Liu et al., 2020a) PFDoDS: 63% (Alves et al., 0.046ⱡ 2015) (<0.0030.091) 497 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age Country, N Sample Gende Area r PFBA PFPeA PFHpA 16-85 (F,M) India, 14 cities 39 0% 15% (<0.1/0.4 (<0.4/17) -0.59) NR Hair NR Spain, Seville 6 0/6 (<3.2) NR Hair 4-90 (F,M) Spain, Seville 42 14% (<1.015.5) NR Hair NR Italy 11 3/11 0% (<0.1-15) (<0.1) 0% (<0.1) NR Hair Adults 2009 Nails 19-26 (F,M) 47 South Korea, Busan 39 C hina, Guangdo ng 36% (<0.361.5) 58% 5.5 (<0.4144) 19% (<0.410.74) 63% 9.8 (<0.2265) 2012 Nails Fisher C hina, y Wuhan emplo yees childre Spain n& adults (F,M) 2012- NR age C hina, PFHpS 0% 0% 0% 12 7/12 (<1-15) 11/12 4 (<1-7) 9/12 5 (<0.76) 100% 100% PFDS Reference Other PFCA/PFSA/P FPA PFNS: 0% (<0.015) (Ruan et al., PFOcDA, 31% 2019) (<0.19/0. PFHxDA, PFEtS & 83-1.2) PFPrS: 0% (<0.02/<0.09) (Martin et al., 2016) 3/6 5/6 (<0.9-13) 1.9 (<0.2-10) 26% 86% (<0.92.4 13.3) (<0.210.1) 8 103 100% PFPeS 21% 0% (<0.1/0.4 (<0.02/0. 08-0.23) ) 20172018 Hair NR Nails PFBS (Martin et al., 2019) 1/11 (<0.020.50) 0% (<0.80) 0/11 (<0.02) 0/11 (<0.02) 0% (<0.28) (Piva et al., 2021) (Kim and Oh, 2017) 2% (<0.420.42) (Liu et al., 2020a) 8/8 4.1 (0.15-7.7) (Wang et al., 2018b) (Martin-Pozo et al., 2020) 94% PFPrA: 100% (Song et al., 498 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Year Age Country, N Sample Gende Area r 2013 Guangzh Semen ou 29.4±5 C hina, 664 2015.4 Nanjing 2016 Semen 20-45 UK, 59 2015 Hull Follicul ar fluid PFBA PFPeA PFHpA PFBS 3.1 1.7 0.36 0.076 0% (<0.075) 0% (<0.05) 0% (0.2) 0% (<0.3) 3% (<0.010.06) 15% (<0.10.53) 6% (<0.010.094) 21% (<0.20.44) PFPeS Reference Other PFCA/PFSA/P FPA 0.95 2018b) 0% (<0.1) Australia, 97 Queensla nd 38% (<0.08NR) Belgium, 38 Antwerp 0% (<1.3) 20182019 Follicul ar fluid 2016 Placent a C hina, Beijing 28 79% 0.012 ¥ 29% (6 (F, M) USA, National 26 82 0% (<0.1) 3-68 (F, M) C hina, Shijiazhua 41 95% 0.0020 0/8 ( carboxylic acid group > alcohol group (Qiu et al., 2020) In vitro: ­ short-chain PFAS had no cytotoxic effects on rat thyroid cells and did not interfere with thyroid-stimulating hormone (TSH)-dependent cyclic adenosine monophosphate (cAMP) production (C roce et al., 2019) ­ FRTL-5 cells exposed to increasing concentrations of GenX displayed both genotoxic and cytotoxic effects (C operchini et al., 2020) In vivo: 506 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) ­ GenX caused a dose-responsive up-regulation of 28 different genes involved in the PPAR signaling pathway (C onley et al., 2019) ­ After exposure to 0.5–50 nM of 8:8 PFPiA in zebrafish larvae, an increase of T4 and T3 was observed. In addition, corticotropin-releasing hormone (C RH) and TSHb were downregulated and uridinediphosphateglucuronosyltransferase (UGT1AB) resulted upregulated. The authors suggested that this should be regarded as a compensatory response to the hyperthyroid status (Liu et al., 2019b) ­ C ontrary to above study by Kim et al., showed an up- regulation of corticotropin releasing hormone b (C RHB), thyrotropin receptor (TSHR), and thyroid transcription factor-1 (NKX2.1) genes which was suggestive of a negative feedback in response to decreased circulating thyroid hormones (Kim et al., 2020b) ­ study performed on dams exposed to PFHxS reported no statistically significant effects on thyroid gland weights and Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference 507 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance 6:2 FTOH, 8:2 FTOH, 8:2 FTOAcr, PFOA, PFNA, PFDA, PFUnDA, PFPeA, PFHxA, PFHpA, PFOS, PFDS, E2 + others Activity/Effect (in some cases direct quotes from the referenced source are used) histopathology (Ramhoj et al., 2020) In silico: ­ computational model based upon crystal structure from human Erα: PFOA, PFNA, PFDA, and PFOS all efficiently docked with Erα from different species in a similar manner to BPA and nonylphenol In vitro: ­ PFOA, PFNA, PFDA, PFUnDA, and PFOS significantly enhanced human Erα-dependent transcriptional activation at concentrations ranging from 10– 1000 nM ­ All PFAAs tested weakly bound to trout liver ER with IC 50 values of 15.2–289 mM In vivo: ­ PFOA, PFNA, PFDA, PFUnDA all potent inducers of vtg in vivo (at high concentrations of 50 ppm) ­ Structure-activity relationship for PFAAs was observed, where eight to ten fluorinated carbons and a carboxylic acid end group were optimal for maximal vtg induction Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference In silico: ­ IC M-virtual ligand screening (VLS) procedure (Molsoft IC M v3.5-1p) ­ proteins were built based upon 1ERE as the 3Dtemplate with Molsoft IC M v3.5-1p 6:2 FTOH and PFHxA were recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* (Benninghoff et al., 2011) In vitro: ­ ER saturation binding assays and competitive binding assays with trout liver cells (cytosol fraction of liver homogenate) ­ Erα reporter gene assay with HEK-293T cells co-transfected with XTEL luciferase reporter plasmid containing a consensus estrogenresponsive element (ERE) sequence from the Xenopus Vtg promoter human Erα expression vector In vivo: ­ Rainbow trout (Oncorhynchus mykiss): Dietary exposure of juvenile individuals (5 months old for VTG analysis, 11 months old for other) 14 day PFAS mixture exposure (in the ratio of 1:1:1:1, mix A 508 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) PFOSA Hepatocellular fatty acids content, gene expression (lipid metabolism, oxidative stress), PFAA mixture (PFOA, PFOS, PFBS, PFNA) C ellular level changes in the endocrine organs, including follicle cell degeneration in male fish and follicle cell atrophy Test System / Methods (in some cases direct quotes from the referenced source are used) composed of 5 mg/L individual PFAS (PFOA, PFNA, PFDA, PFUnDA), mix B composed of 50 mg/L individual PFAS, mix C composed of 250 mg/L individual PFC ) ­ For VTG: PFOA and PFDA exposure, 0.026, 0.128, 0.64, 3.2, 16, 80, 400, 2 000 mg/L PFOA or PFDA; DMSO vehicle (0.05 mg/L); In vivo: Atlantic salmon (Salmo salar) hepatocytes, Static exposure (12 h and 24 h); 0.01% DMSO (solvent control), 2, 20, 50 µM PFOSA Remark Reference Only abstract available (Wågbø et al., 2012) In vivo: Japanese medaka (Oryzias latipes), 238 d exposure at 0.5 and 5 µg/L (nominal) mixture ratio 1:1:1:1 Effects with relevance on population level (Lee et al., 2017) significant increase in vtg expression relative to the control in the F2 generation (but not significant for F0, or F1 generation) at 5 µg/L reduced fecundity: suppression of hatching rate in F1 generation survival rate of F1 and F2 generation >80% for all treatment groups 509 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance 6:2 FTOH, 8:2 FTOH, NFDH, PFOS, PFOA Activity/Effect (in some cases direct quotes from the referenced source are used) sex ratio: PFAA high concentration caused the shifting into male portion Treatments with 6:2 FTOH, 8:2 FTOH and NFDH dose-dependently induced hER-mediated transcriptional activity with interaction between the hERα or hERβ ligand binding domain and TIF2 Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference In vitro: yeast two-hybrid assay: modified by incorporation of hER isoforms (hERα or hERβ) 6:2 FTOH was recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* 6:2 FTOH was recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* (Ishibashi et al., 2007) estrogenic effects of FTOHs on hERα were higher than those on hERβ, indicating a differential responsiveness of hERs to FTOHs 4:2FTOH, 6:2 FTOH, 8:2 FTOH, PFOS, PFOA estrogenic effects for hERα and hERβ descended in the order of estradiol17b>>>6:2 FTOH > NFDH > 8:2 FTOH Dose-dependent vtg induction after exposure to 6:2 FTOH but not 4:2 FTOH or 8:2 FTOH Significant vtg induction after 12 h (6:2 FTOH) exposure, and 72h exposure (4:2 and 8:2 FTOH (but not dose dependent)) C o-exposure with E2 inhibited E2induced hepatocellular VTG production in a dose-dependent manner except for 4:2 FTOH suggesting an anti-estrogenic activity In vitro: non-competitive enzyme-linked immunosorbent assay (ELISA) investigating vtg induction in in primary cultured hepatocytes of freshwater male tilapia (Oreochromis niloticus) (Liu et al., 2007) C o-exposure with known estrogen receptor inhibitor tamixofen inhibited 510 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) the ability of test compounds to stimulate vitellogenesis: estrogenic effect of PFAS may be mediated by the estrogen receptor pathway Test System / Methods (in some cases direct quotes from the referenced source are used) PFBS No treatment-related mortalities or effects on body weight, weight gain, feed consumption, histopathology measures, or reproductive parameters evaluated in the study when compared to the control group PFBA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTeDA, 7H-PFHpA, 6:2 FTUA, PFBS, PFHxS, PFOS, L-PFDS, L-PFOSi, 6:2 FTOH, 8:2 FTOH, N-MeFOSE, N-EtFOSE, FOSA, N,N-ME2FOSA, N-MeFOSA, N-EtFOSA binding potency decreased in the order: perfluorohexane sulfonate > perfluorooctane sulfonate/perfluorooctanoic acid > perfluoroheptanoic acid > sodium perfluoro-1-octanesulfinate > perfluorononanoic acid In vivo: Northern bobwhite quail (Colinus virginianus) reproduction study: adult quail were exposed to nominal dietary concentrations of 100, 300, or 900 mg PFBS/kg, ww feed for up to 21 weeks In vitro: radioligand-binding assay testing for binding capacity to human TTR maximum potency at a chain length of eight carbons (PFOA). The binding potency is clearly associated with the degree of fluorination of the alkyl chain For PFASs with a carbon chain length of four to eight, TTR binding potencies were significantly higher for compounds containing a sulfonate functional group than for those containing a carboxylic acid functional group QSAR models indicated the dependence on molecular size and functional groups Remark Reference (Newsted et al., 2008) Modelled binding to human TTR. Of relevance for ENV due to high conservation of HPT axis (Weiss et al., 2009) 6:2 FTOH and PFHxA were recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* 511 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) (but only preliminary result, more detailed description of chemical properties and data for validation needed) Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference In vitro: primary cultures of avian neuronal cells of two avian species: the domestic chicken (Gallus domesticus) and herring gull (Larus argentatus). Measurement of mRNA levels of thyroid hormone (TH)–responsive genes D2, D3, TTR, Oct-1, myelin basic protein and RC 3 after exposure to PFBA, PFBS, PFHxA, PFHxS, PFHpA, and PFHpS at five concentrations: 0.01, 0.1, 1, 3, and 10lM PFHxA was recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* (Vongphachan et al., 2011) Of the six PFASs with the same fluorinated carbon chain length (C 8) but with different sulfate-based functional groups, highest binding potency was observed for the sulfonate (PFOS), followed by the ulfonate (perfluorinated octane ulfonate) and the sulfonamide (perfluoro-1-octane sulfonamide) PFBA, PFBS, PFHxA, PFHxS, PFHpA, and PFHpS Test compounds with the sulfonamide functional group protected by an alkyl group had no TTR binding potency themselves but seemed to cause a slight increase in T4-TTR binding at high test concentrations C hicken: ­ PFAS < C 8 altered the expression of TH-responsive genes (D2, D3, TTR, and RC 3) in chicken embryonic neuronal cells to a greater extent than PFAS > C 8 ­ chicken: D2 upregulated after exposure to PFHxA, PFHpA, and PFNA ­ D3 mRNA expression increased twofold to fivefold following exposure to several PFC s (PFBS, PFHxA, and PFHxS) 512 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) ­ PFHxS treatment significantly decreased TTR mRNA expression ­ PFHpA exposure increased TTR mRNA levels ­ PFBS and PFHxS treatment increased RC 3mRNAexpression ­ MBP mRNA expression was upregulated at 3 and 10 µM following PFHxA treatment ­ no changes to Oct-1 mRNA levels Herring gull: ­ upregulation in RC 3 mRNA expression (Fig. 4A) following PFHpA exposure ­ RC 3 mRNA expression did not change ­ increase in Oct-1 mRNA following treatment with PFBS, PFHxA, and PFHxS General: ­ study provided evidence that the brain may be a target organ for PFAS effects and that these contaminants have the potential to alter TH homeostasis in birds Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference 6:2 FTOH, 8:2 FTOH 6:2 FTOH & 8:2 FTOH induce cell proliferation at 10 μM In vitro: proliferation-promoting capacity of 6:2 FTOH and 8:2 FTOH with an Escreen assay of MC F-7 cell lines 6:2 FTOH was recently assessed for its endocrine disrupting properties in the (Maras et al., 2006) small but relevant up-regulation of the estrogen receptor as a consequence of exposures to 6:2 FTOH or 8:2 FTOH. 513 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) Test System / Methods (in some cases direct quotes from the referenced source are used) PFOS, PFBS significant decrease of the transcription of VTG1 in exposed organisms transcriptomic and cellular results indicated that exposure to PFEC HS reduced VTG in D. magna no effects were observed on the survival, the frequency of molting, the number of neonates produced or the growth of exposed organisms neither PFOS nor PFBS had a significant effect on the survival and growth caused hepatohistological impairment at higher concentrations (100; 1 000 ug/L) PFBS had no effect on the sex ratio and gonadal histology. PFBA, PFHxA, PFOS and PFBS promoted expression of estrogen receptor (ER) and androgen receptor (AR), but not affected aromatase expression in the brain PFOA, PFOS and PMOH enhanced Reference course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* Up regulated genes: TFF1, PGR, ESR1, PDZK1 Down regulated genes: ERBB2 PFEC HS (acyclic perfluoroalkane sulfonic acid used as an erosion inhibitor in aircraft hydraulic fluid) Remark In vivo: sublethal exposure (12 d) of Daphnia magna to PFEC HS (0.06, 0.6, and 6 mg/L), microarray and quantitative real-time PC R (Houde et al., 2016) In vivo: growth and sexual development of Xenopus laevis tadpoles exposure to series of concentrations of PFOS and PFBS (0.1; 1; 100; 1 000 ug/L) as well as 17-betaestradiol (E2, 100 ng/L) and 5 alphaandrostan-17-beta-ol-3-one (DHT, 100 ng/L) from stage 46/47 to 2 months postmetamorphosis (Lou et al., 2013) In vitro: PFHxA was (Behr et al., 514 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance PFOA, PFBS, PFHxS, PFOS Activity/Effect (in some cases direct quotes from the referenced source are used) 17beta-estradiol-stimulated estrogen receptor beta activity PFOS, PMOH, PFHxA and PFBA enhanced dihydrotestosteronestimulated androgen receptor activity H295R steroidogenesis assay, PFOA and PFOS slightly enhanced estrone secretion, and progesterone secretion was marginally increased by PFOA. PFHxA, PFHxS All effects were only observed at concentrations above 10 µM, and none of the PFAS displayed any effect on any of the molecular endocrine endpoints at concentrations of 10 µM or below Pipping success (pip = first break in eggshell) was reduced to 63% at the highest dose of PFHxS; PFHxS exposure (38,000 ng/g) decreased tarsus length and embryo mass. PFHxS and PFHxA accumulated in the three tissue compartments analyzed as follows: yolk sac > liver > cerebral cortex. Type II and type III 5’-deiodinases (D2 and D3) and cytochrome P450 3A37 mRNA levels were induced in liver tissue of chicken embryos exposed to PFHxS Test System / Methods (in some cases direct quotes from the referenced source are used) C ell Lines: H295R, HEK293T, LNC aP, MC F-7, MDA-kb2 Exposure Duration : 24 h – 6 d Exposure Range : 0.001 – 500 µM Types of Endpoints: Androgen related, C ytotoxicity, Estrogen related, Steroidogenesis, C ell proliferation Remark Reference recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* 2018) In vivo: C hicken (Gallus gallus domesticus): determined in ovo effects of PFHxS and PFHxA exposure (maximum dose = 38,000 and 9700 ng/g egg, respectively) on embryonic death, developmental endpoints, tissue accumulation, mRNA expression in liver and cerebral cortex, and plasma TH levels PFHxA was recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* In vivo effects of PFHxS but not PFHxA (C assone et al., 2012) 515 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) Test System / Methods (in some cases direct quotes from the referenced source are used) no effects were observed for PFHxA. PFBS exposure decreased the levels of 3,5,3’-triiodothyronine (T3) in F0 female blood; increased T3 or thyroxine (T4) levels in F0 brains, in which hyperthyroidism suppressed the local transcription of 5’deiodinase 2 (Dio2). Reference Effects with relevance on population level D2, neurogranin (RC 3), and octamer motif binding factor 1 mRNA levels were upregulated in cerebral cortex. Plasma TH levels were reduced in a concentration-dependent manner following PFHxS exposure PFBS Remark In vivo: Exposure of F0 marine medaka (Oryzias melastigma) eggs to PFBS at different concentrations (0, 1.0, 2.9, and 9.5 µg/L) until sexual maturity. The F1 and F2 generations were reared without continued exposure. Effects with relevance on population level! (C hen et al., 2018b) Decreased T3 was transferred to F1 eggs, although the parental influences were reversed in F1 larvae. Delayed hatching was coupled with elevated T3 levels in F1 larvae. F1 adults showed comparable symptoms of thyroidal disruption with F0 adults. Slight recovery was noted in the F2 generation, although F2 larvae still exhibited thyroid disruption and synthesized excessive T4. Results suggested that the offspring 516 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) suffered more severe dysfunction of the thyroidal axis albeit without direct exposure Test System / Methods (in some cases direct quotes from the referenced source are used) 6:2 chlorinated polyfluorinated ether sulfonate (F-53B) In silico ­ F-53B binds to transthyretin (TTR) by forming hydrogen bonds with Lys123 and Lys115, thereby interfering with thyroid hormone homeostasis In vitro ­ F-53B enhanced cell proliferation in a dosedependent manner, indicative of thyroid receptor agonistic activity. In silico ­ Based on the homologymodeled structure of zebrafish TTR, F-53B was docked automatically into the binding site of zfTTR using AutoDock Vina 1.1.2 In vivo ­ In zebrafish larvae, F-53B exposure induced significant developmental inhibition and increased thyroxine (T4) but not 3,5,3’-triiodothyronine (T3) levels accompanied by a decrease in thyroglobulin (TG) protein and transcript levels of most genes involved in the hypothalamic-pituitary-thyroid (HPT) axis In vivo ­ Zebrafish (Danio rerio) embryos (2 hpf) exposed for 5 days to 0, 0.5, 20 and 200 μg/L F-53B followed by depuration in clean water for 5 days Reduced surface area of swim bladder (3 dpf) and significant changes in gene expression patterns (3 dpf) In vivo: Zebrafish (Danio rerio) embryos exposed to PFOA, TBBPA, TDC PP, DOPO, PFBA. Sub-chronic (0-6days post fertilization (dpf)) and chronic PFOA, TBBPA, TDC PP, DOPO, PFBA Remark Reference (Deng et al., 2018) In vitro ­ Molecular docking study to TTR C ell proliferation assay with GH3 cell line (Godfrey et al., 2017) 517 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) TDC C P, PFOA, PFBA Females displayed significantly larger swim bladders (which are under thyroid hormone control) after exposure to all chemicals with the exception of triiodothyronine, which caused the opposite effect PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFBS, PFHxS, PFOS IEFs for the PFAS was as follows: PFOA (IEF: 1)>PFHpA (0.89)>PFNA (0.61)>PFPeA (0.50)>PFHxS (0.41)>PFHxA (0.38) approximately PFDA (0.37)>PFBA (0.26)=PFOS (0.26)>PFUnDA (0.15)>>PFDoDA and PFBS (not activated). SAR analysis showed that PFC As having more than seven perfluorinated carbons had a negative correlation (r=-1.0, p=0.017) between the number of perfluorinated carbons and the IEF of PFC As, indicating that the number of perfluorinated carbon of PFC As is one of the factors determining the transactivation potencies of the BS PPARα. Test System / Methods (in some cases direct quotes from the referenced source are used) (0-28dpf) exposures at 1% of LC 50 In vivo: Japanese Medaka (Oryzias latipes) embryos were exposed to sublethal concentrations of TDC PP, 0.019 mg/L), PFOA, (4.7 mg/L) and PFBA (137 mg/L). Exposure from 0 – 10 dpf In vitro reporter gene assay Transactivation of the Baikal seal (Pusa sibirica) peroxisome proliferator-activated receptor alpha (BS PPARα) by PFAS (C 4-C 12) estimated the PFOA induction equivalency factors (IEFs), a ratio EC 50 of PFOA to the concentration of each compound that can induce the response corresponding to 50% of the maximal response of PFOA Remark Reference (Godfrey et al., 2019) PFHxA was recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* (Ishibashi et al., 2011) PFC As were more potent than PFSAs with the same number of perfluorinated carbons 6:2 C l-PFESA and 8:2 C lPFESA (PFOS 6:2 C l-PFESA and 8:2 C l-PFESA bound to PPARs with affinity higher than PFOS In vitro: fluorescence competitive binding assay (Li et al., 2018a) 518 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance alternatives in C hina) HFPO-TA, HFPO-DA Activity/Effect (in some cases direct quotes from the referenced source are used) showed agonistic activity toward PPARs signaling pathways with potency similar to (6:2 C l-PFESA) or higher than (8:2 C l-PFESA) PFOS C l-PFESAs fitted into the ligand binding pockets of PPARs with very similar binding mode as PFOS receptor binding experiment showed HFPO-TA exhibited 4.8-7.5 folds higher binding affinity with PPARγ than PFOA, whereas HFPO-DA exhibited weaker binding affinity than PFOA. Agonistic activity toward PPARγ signaling pathway in HEK 293 cells in the order of HFPO-TA > PFOA > HFPODA. Test System / Methods (in some cases direct quotes from the referenced source are used) luciferase reporter gene transcription assay (C ell Lines Used: 3T3-L1, HEK293T) Remark Reference In vitro: investigation of receptor binding, receptor activity, and cell adipogenesis effects (C ell Lines Used: 3T3-L1, HEK293T, Preadipocytes) to compare potential disruption effects of HFPO-TA, HFPO-DA, and PFOA on peroxisome proliferator-activated receptor gamma (PPARγ) via the (Li et al., 2019) In vivo: Assessed plasma concentrations of (Nøst et al., 2012) Molecular docking simulation indicated HFPO-TA formed more hydrogen bonds than PFOA, whereas HFPO-DA formed fewer hydrogen bonds than PFOA. HFPO-TA promoted adipogenic differentiation and lipid accumulation in both mouse and human preadipocytes with potency higher than PFOA. Adipogenesis in human preadipocytes is a more sensitive end point than mouse preadipocytes PFHpS, PFOS, PFNA + others PFAS dominated the summed HOC s concentrations in both species (77% in 519 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) kittiwakes and 69% in fulmars). Positive associations between total thyroxin (TT4) and PFAS (PFHpS, PFOS, PFNA) were reported by the authors for both species. The authors qualify that “Although correlations do not implicate causal relationships per se, the correlations are of concern as disruption of TH homeostasis may cause developmental effects in young birds” 6:2 FTOH, 8:2 FTOH, 10:2 FTOH, PFBS, PFHxS, PFOS, PFBA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTrDA, PFTeDA, Most of the tested PFASs bound TTR with relative potency (RP) values of 3 × 10−4 to 0.24 when compared with that of the natural ligand thyroxine, whereas fluorotelomer alcohols did not bind Structure-binding analysis revealed that PFASs with a medium chain length and a sulfonate acid group are optimal for TTR binding, and PFASs with lengths longer than 12 carbons are optimal for TBG binding Molecular docking showed that the Test System / Methods (in some cases direct quotes from the referenced source are used) halogenated organic contaminants (HOC s) in chicks of two seabird species: black-legged kittiwake (Rissa tridactyla) and northern fulmar (Fulmarus glacialis) to investigate possible correlations of HOC s with circulating thyroid hormone (TH) concentrations. Plasma chicks were sampled in Kongsfjorden, Svalbard in 2006. Samples were analyzed for thyroid hormones and a wide range of HOC s (polychlorinated biphenyls (PC Bs), hydroxylated (OH-) and methylsulphoned (MeSO-) PC B metabolites, organochlorine pesticides (OC Ps), brominated flame retardants (BFRs), and perfluorinated compounds (PFAS) In vitro: fluorescence displacement assay was used to determine the binding affinities of 16 PFASs with two major TH transport proteins, transthyretin (TTR) and thyroxine-binding globulin (TBG) Remark Reference Not included in overview table as no new information regarding substance activity but interesting to describe trends (Ren et al., 2016) 6:2 FTOH and PFHxA were recently assessed for its endocrine 520 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance PFBA, PFPeA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTeDA, PFBS, PFHxS, PFOS, FOSA Activity/Effect (in some cases direct quotes from the referenced source are used) PFASs bind to TTR with their acid group forming a hydrogen bond with K15 and the hydrophobic chain towards the interior. PFASs were modeled to bind TBG with their acid group forming a hydrogen bond with R381 and the hydrophobic chain extending towards R378 Test System / Methods (in some cases direct quotes from the referenced source are used) C ellular concentration of PFC As increased with perfluorocarbon chain length up to PFDoDA. PPARα activity of PFC As increased with chain length up to PFOA. In vitro: investigated the relationship between PPARα activity and cellular concentration in HepG2 cells: C ellular concentrations were determined by high-performance liquid chromatography–tandem mass spectrometry and PPARα activity was determined in transiently transfected cells by reporter gene assay Maximum induction of PPARα activity was similar for short-chain (PFBA and PFPeA) and long-chain PFC As (PFDoDA and PFTeDA) (approximately twofold). PPARα activities were induced at lower cellular concentrations for the shortchain homologs compared to the longchain homologs Remark disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* Partly included in overview table as no new information regarding substance activity but relevant to describe trends In line with results from Zhang et al 2014 PFHxA was recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More Reference (Rosenmai et al., 2018) 521 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference details are provided below the table* PFBS PFBS-exposed embryos had significantly increased caudal fin deformities, delayed swim bladder inflation, and impaired yolk utilization Incidence of fish with significantly stunted growth and truncated exocrine pancreas length was significantly increased, although these two effects occurred independently. (Sant et al., 2019) In vivo: Dechorionated zebrafish (Danio rerio) embryos from two different transgenic fish lines (Tg[insulin:GFP], Tg[ptf1a:GFP]): exposed to 0 (0.01% DMSO), 16, or 32 µM PFBS daily from 1 – 7 dpf. Were examined using fluorescent microscopy for islet area and morphology, and exocrine pancreas length Islet morphology revealed an increased incidence of severely hypomorphic islets (areas lower than the 1st percentile of controls) and an elevated occurrence of fragmented islets. 6:2 C l-PFESA (F-53B) RNA-Seq data (4 dpf) also identify disruptions in regulation of lipid homeostasis F-53B accumulated in the F0 gonads and transferred to the F1 generation via maternal eggs, and even remained in F1 adult fish and their eggs (F2) after 180d depuration In the F0 generation, F-53B exposure significantly inhibited growth and induced reproductive toxicity, including decreased gonadosomatic index and egg In vivo: Adult zebrafish (Danio rerio) (F0 generation) were chronically exposed to different concentrations of F-53B (0, 5, 50, and 500 µg/L) for 180d using a flow-through exposure system, with F1 and F2 generations reared without exposure. The reproductive toxicity endpoints were assessed in F0 and F1 adult fish Effects with relevance on population level (Two follow up studies investigated the ED effects further to identify possible MoA ((Shi et al., (Shi et al., 2018a) 522 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) production/female, changes in the histological structure of the gonads, and increased serum testosterone levels Test System / Methods (in some cases direct quotes from the referenced source are used) Remark Reference 2019a; Shi et al., 2019b) not mentioned here) serum estradiol and vitellogenin levels were significantly increased in 5 ug/L F53B-exposed adult males. Transcriptional levels of several genes along the hypothalamic-pituitarygonadal axis were altered in F0 generation fish. Testis transcriptome analysis revealed that F-53B exposure disrupted spermatogenesis in F0 male zebrafish. Maternal transfer of F-53B also induced adverse effects on growth and reproduction in the F1 generation. PFHxS, PFPeA, PFHxA, PFHpA + others Higher occurrence of malformation and lower survival in F1 and F2 embryos indicated that parental exposure to F53B could impair the embryonic development of offspring no estrogenic responses to PFASs In vitro: MMV-LUC cell line reporter gene assay was used to assess estrogenic activity of PFAs (C 6-C 12) and other chemicals PFHxA was recently assessed for its endocrine disrupting properties in the course of the substance (Wielogórska et al., 2015) 523 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance 6:2 FTOH, 8:2 FTOH, PFBA, PFHxA, PFHpA, PFOA, PFNA, PFDA, PFUnDA, PFDoDA, PFTeDA, PFBS, PFHxS, PFOS Activity/Effect (in some cases direct quotes from the referenced source are used) binding affinity was strongly dependent on their carbon number and functional group. For PFC As the binding affinity increased with their carbon number from 4 to 11, and then decreased slightly. Test System / Methods (in some cases direct quotes from the referenced source are used) in vitro: binding of 16 PFASs to human PPARγ ligand binding domain (hPPARγ-LBD) and their activity on the receptor in cells were investigated using HepG2/C 3A cells For PFSAs binding affinity was stronger than their PFC A counterparts. relative epididymis and testis weights decreased in the 1.0 mg/kg/d group evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* Partly included in overview table as no new information regarding substance activity but relevant to describe trends Reference (Zhang et al., 2014) 6:2 FTOH and PFHxA were recently assessed for its endocrine disrupting properties in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA. More details are provided below the table* No binding was detected for the two fluorotelomer alcohols (FTOHs) 6:2 C l-PFESA Remark In vivo: subchronic exposure study to Study with higher relevance (Zhou et al., 2018) 524 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Substance Activity/Effect (in some cases direct quotes from the referenced source are used) compared with the control. No changes were observed in the serum levels of testosterone, estradiol, folliclestimulating hormone (FSH), or luteinizing hormone (LH), nor in the histopathological structure of the epididymis and testis and sperm count. Test System / Methods (in some cases direct quotes from the referenced source are used) investigate the reproductive toxicity of 6:2 C l-PFESA exposure (0, 0.04, 0.2, and 1.0 mg/kg/d body weight, 56 d) in adult male BALB/c mice Remark Reference to HH but interesting to compare to other studies with 6:2 C lPFESA 56 d of consecutive gavage of 1.0 mg/kg/d of 6:2 C l-PFESA did not affect male mouse fertility. RNA sequencing showed that no genes were significantly altered in the testes after 6:2 C l-PFESA exposure * Results of a FSDT (OEC D TG 234) performed with 6:2 FTOH as one main degradation product of FTA/FTMA in the course of the substance evaluation (SEv) for 6:2 FTA and 6:2 FTMA indicate an estrogenic MoA for the substances: increased VTG levels in male fish as well as changes in the secondary sex characteristics (significant decrease in the number of anal fin papillae in males) were observed. Additionally, there are hints (not statistically significant in the presented study most likely due to discrepancy between nominal and measured concentrations) for an influence of FTOH on the s ex ratio. The ED expert group supported the conclusion that considering all data available, there is sufficient evidence on the estrogenic modality to iden tify FTOH as ED for the environment. During the substance evaluation of FTA/FTMA also an AMA (OEC D TG 231) assay was performed for PFHxA the other main degradation product of FTA/FTMA. The result of this test clearly shows a thyroid agonistic activity with a dose -dependent acceleration in metamorphosis of the test animals. During the ED EG discussion the view was expressed that the thyroid mediated adversity could already be concluded based on the effects observed in the submitted AMA study. A summary report of the discussion at ED EG 21 can be found here: https://echa.europa.eu/documents/10162/1459379/flashreport_edeg21_en.pdf/e530deb9-5baf-7fd4-dc35-8a4cd7c73f33?t=1639059043393, date of access: 2022-09-29. The SEv for FTA/FTMA was recently concluded, stating “The assays requested for 6:2FT-OH and PFHxA during the substance evaluation of FTA and FTMA, together with available in vitro data, leads the eMSC A to the conclusion that the WHO/IPC S criteria for an en docrine disruptor in the environment are fulfilled for both substances under evaluation.” “For 6:2-FTOH it is concluded that both in vitro and in vivo data in different fish species provide sufficient evidence to prove an es trogenic activity of the substance that can be linked to adverse and population relevant effects like femininization of male fish and the loss of male specif ic secondary sex characteristics, possibly leading to an impairment of mating success.” “For PFHxA it is concluded that in vitro and amphibian in vivo data provide sufficient evidence to prove a thyroid agonistic activity of PFHxA that can be linked 525 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) to adverse and population relevant effects (accelerated development in amphibians).” Once published, the conclusion document can be found here: https://echa.europa.eu/de/information-on-chemicals/evaluation/community-rolling-actionplan/corap-table/-/dislist/, date of access: 2022-09-29. 526 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.3. Textiles, upholstery, leather, apparel and carpets Table B.97. TULAC – emission estimates (ERC assumptions). Life cycle phase ERC Applicable Total release PFAS into the environment Manufacturing phase Comments Manufacture of non-woven membranes for use in textile applications. ERC 12A Processing of articles at industrial with low release. Fluoropolymers and Side C hain Fluorinated polymers. Emissions to air: 2.5% w.w Emissions to soil: 2.5% w.w Releases into wastewater system: 2.5% w.w TOTAL: 7.5% w.w Applied ERC 12A to the production of membranes. Fluoropolymers are highly stable, but assume emissions relate to dusts and any cutting activities. Manufacture of mixtures for use in textile applications. ERC 2 – Formulation into mixture. Non-polymeric PFAS (PFAA precursors C 2C 3 and PFAA precursors ≥C 4) and PFPE. Applied the standard ERC for mixtures. The manufacture of solid articles where the PFAS is used as processing aid. PFAS is not intentionally retained in the article. ERC 6B Use of reactive processing aid at industrial site (no inclusion into or onto article). Non-polymeric PFAS. Emissions to air: 2.5% w.w Emissions to soil: 0.01% w.w Releases into wastewater system: 2% w.w TOTAL: 4.51% w.w Emissions to air: 0.1% w.w Emissions to soil: 0.03% w.w Releases to wastewater system: 5% TOTAL: 5.13% w.w The treatment of textiles with commercial textile mixtures. ERC 6A: Use of intermediate meaning use of chemical building blocks (feedstock). Non-polymeric PFAS (PFAA precursors C 2C 3 and PFAA precursors ≥C 4) and PFPE. Emissions to air: 5% w.w Emissions to soil: 0,1% w.w Releases to water: 2% w.w TOTAL: 7.1% w.w Applied the standard ERC for mixtures. Data from the C fE suggests that use or processing aids for production of textile materials is carried out within closed systems. Therefore applied ERC 6B 527 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) ERC Applicable PFAS Total release into the environment Comments Indoor use – frequent cleaning/ wetting, washing for a higher potential of release. Assume 100% used indoors with high release. ERC 11b widespread use of articles with high or intended release (indoor) to all PFAS. Non-polymeric PFAS (PFAA C 2-C 3 and PFAA ≥C 4) and Side-chain fluorinated polymers. Emissions to air: 50% Emissions to soil: n.a. Releases to wastewater: 50% TOTAL: 100% w.w Used the ERC from the C ommission study but amended for fluoropolymers. 100% release for fluoropolymers is highly unlikely. Indoor use – infrequent cleaning /wetting. Assume 100% indoor use with low release. ERC 11a Widespread use of articles with low release (indoor). Fluoropolymers and PFPE. Emissions to air: 0.05% Emissions to soil: n.a. Releases to wastewater: 0.05% TOTAL: 0.1% w.w Used the ERC used in the C ommission study. Outdoor use – low release- This could include for example outdoor wear. Assume 100% outdoor use. ERC 10a Widespread use of articles with low release (outdoor). All PFAS: Fluoropolymers and side chain fluorinated polymers and non-polymeric PFAS (PFAA C 2C 3 and PFAA ≥C 4) and PFPE. Emissions to air: 0.05% Emissions to soil: 3.2% Emissions to surface water: 3.2% TOTAL: 6.45% w.w Used the ERC used in the C ommission study. Life cycle phase User phase 528 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Figure B.80. Default worst-case release factors. 529 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Table B.98. Assumptions for backward facing usage rates. PFAS substances General rules 1990-2002 2003-2015 2015-2020 Non-polymeric C2 – C3 substances No usage identified for medical or ‘other’ textiles. Assume usage rate is zero for all years. Assume a steady growth rate of 2% annually for home textiles, consumer apparel and technical textiles. The oil repellence properties of ultrashort chain PFAS is considerably weaker than C 8 chemistries. Assume that while the phase-out of PFOS sparked an uptake in C 4, C 5, and C 6 PFAS. C 2-C 3 was unchanged and continued to grow at 2% per annum. Non-polymeric C4 substances No usage identified in professional apparel; medical textiles; or ‘other’ textiles. Assume usage rate is zero for all years. Assume a steady growth rate of 2% annually for home textiles, consumer apparel and technical textiles. Assume that the voluntary phase-out of PFOS accelerates uptake for short-chain in home textiles and consumer apparel. Growth rate of 4% annually between 2003 and 2012, after which assume growth becomes static due to continued tightening of regulation. For technical textiles, assume growth rate remains steady at 2% annually due to more limited application for oil repellence. Assume that while regulatory pressure applied on C 6 technologies may slow growth in C 4 and above, the demand for PFAS substances and difference performance profile of C 2 and C 3, means ultra-shortchain is unaffected. C ontinues to grow at 2% per annum. Assume growth rates for home textiles, consumer apparel and technical textiles become static (i.e., no further growth after 2012). This is expected where tightening regulations on PFAS suppress the market and drive uptake of nonfluorine alternatives. Non-polymeric C5 substances No usage identified for professional apparel, medical textiles, or ‘other’ textiles. Assume usage rate is zero for all years. Assume a steady growth rate of 2% annually for home textiles, consumer apparel and technical textiles. Assume the same rules apply as C 4 Assume the same rules apply as C 4 Non-polymeric C6 substances No usage in medical textiles, assume usage rate Assume steady 2% annual growth for home textiles, consumer Assume that the voluntary phase-out of PFOS accelerates uptake of short-chain alternatives. Unlike C 4 Growth for use in home textiles, professional apparel, technical textiles, and other 530 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS substances General rules is zero for all years. Non-polymeric C9-C14 substances No usage identified in professional apparel or medical textiles. Assume usage rate is zero for all years. Other nonpolymeric No usage in medical 1990-2002 2003-2015 2015-2020 apparel, professional apparel, technical textiles, and other textiles. assume that C 6 has greater application in professional and technical textiles. Therefore, assume all uses grow by 4% annually year on year from 2003 – 2012, after which regulations tighten further. Assume from 2012 growth in all except consumer apparel continues at 2% annually. C fE suggested demand for C 6 in home textiles has been strong throughout. For consumer apparel growing use of fluoropolymer means C 6 use becomes static. Assume steady 2% annual growth year on year. textiles continues at 2% per annum. Use for consumer apparel remains static, i.e., no annual growth in use of C 6 PFAS (but also no decline). As with the above assume the voluntary Applied a standard 2% growth rate per The information from TUV suggests that C 98-C 14 may be impurities present as an impurity within other PFAS mixtures, particularly fluoropolymers. But lack of further evidence in the C fE to that affect. Therefore, the standard growth rate of 2% per annum, is applied for the whole backwardlooking timeseries. Assume steady 2% annual Assume steady 2% annual growth year on year. 531 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS substances PFAS General rules textiles, assume usage rate is zero for all years. 1990-2002 2003-2015 2015-2020 growth across all application sectors. annum to all application sectors. All fluoropolymers aggregated Use in all application sectors identified. In lieu of better data assume a standard 2% growth annually year on year to all application sectors. phase-out of PFOS drives diversification. Home textiles and consumer apparel use of other PFAS substances assumed to grow at 4% per annum year on year to 2012, and 2% per annum thereafter. Assume professional apparel, technical textiles and other textiles are more specialist, growth assumed to be 2% per annum year on year. Use of fluoropolymers was already ongoing before the voluntary phase-out of PFOS. It is assumed that the selection of nonpolymeric/polymeric is driven by the specific application and production processes. Therefore, it is assumed that other non-polymeric PFAS substances would be more likely to replace PFOS than a fluoropolymer substitute, at least in the short to medium term. Therefore, while C 4, C 6 and other nonpolymeric increase during this phase, assume the use of fluoropolymer continues to grow at a steady 2% per annum year on year. Side-chain fluorinated polymers No usage for home textiles, technical textiles, or medical Assume that growth for consumer apparel, professional apparel and As with fluoropolymers, assume that the use of nonpolymeric/polymeric is driven by the specific application and More recently the demand for fluoropolymer has started to accelerate, particularly PTFE and PVDF (EEA, 2020 report). Market data 59 suggests between 2019 and 2025 annual growth rates of 5 – 8%. Therefore, assume demand for all fluoropolymers in the third movement phase (2015 – 2030) is a minimum of 5% annually year on year. Assume that due to the growing awareness and concerns around side-chain fluorinated 59 https://www.statista.com/statistics/732029/ethylene -tetrafluoroethylene-market-volumeworldwide/, date of access: 2022-12-21. 532 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) PFAS substances General rules textiles. Assume usage for these applications is zero, throughout. 1990-2002 2003-2015 2015-2020 other textiles is a standard 2% annually. desired physical properties, along with compatibility with the existing production processes. Assume therefore that other non-polymeric PFAS are more likely to replace PFOS than a fluoropolymer. At least in the short to medium term. Therefore, assume growth continues at a standard 2% per annum year on year. polymers as a source of nonpolymeric PFAS emissions, that growth becomes static. I.e., no further increase or decline based on the 2015 year. 533 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.4. Food contact materials and packaging Manufacturing emissions: Paper and board packaging Very limited data have been identified that enable a comprehensive estimation of the emissions of PFAS from the manufacturing of products. However, some data have been identified that has allowed an estimation of emissions from a representative use, namely the manufacture of paper and board packaging to be used in contact with food and feed. For the manufacturing of the other uses not sufficient data was availabe Quantities of PFAS (Surfactant) used in paper and board Packaging in 2019 were between 827 and 6 036 t. Various environmental assessments produced for the purposes of US FCS Notifications have estimated releases to the various environmental compartments for paper and boardmaking processes (FDA-US, 2021). These assessments have been done for different PFAS and for different manufacturers in the same way and using the same assumptions. The assumptions are as follows: - Proportion of copolymer fluorochemicals going to wastewater for treatment is 12% (with 88% incorporated into the finished P&B product) At least 90% of the copolymer going to the wastewater is expected to be retained in either the filtered solids or sludges during wastewater treatment i.e. recovered from the wastewater Up to 10% of the copolymer will remain in the wastewater and be released to the environment. These assumptions are reported to be based upon substantial experience (FDA-US, 2021). An estimate of the total amount of PFAS used by EEA P&B manufacturers in 2019 was derived as follows (assuming that the US data applies in the EEA). If 88% of PFAS are incorporated into the finished P&B product, then total starting quantity of PFAS is derived by 100/88 x quantity of PFAS are incorporated into the finished P&B product = 1.14 x 827 to 1.14 x 6,036 = 943 to 6 881 t i.e. Total quantity of PFAS used by Paper-making mills in 2019 was between 943 – 6 881 t Consumer and industrial cookware and food and feed manufacturing equipment Insufficient information has been obtained to allow a quantitative estimation of the emissions during the manufacture of domestic and industrial cookware in which fluoropolymers are used either as a coating for cookware such as baking tins or frying pans, fluoropolymer conveyor belts or used in components for industrial food processing equipment. To further refine qualitative estimations presented more detailed information would be required for each manufacturing process. Because of the absence of such information, even a default calculation by the REACH methodology has not been possible. Service life emissions:  ERC 10a: widespread use of art icles outdoors with a low release. This was chosen as the best fit ERC to reflect the service-life for paper and board packaging and for 534 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs)  uses of fluoropolymers and side-chain fluorinated polymers for example in consumer or industrial cookware in which the release of the PFASs is not intended. ERC 11a: widespread use of articles with low release (indoors). This was chosen as the best fit ERC to reflect the service life for paper and board packaging and fluoropolymer and side-chain fluorinated polymer uses indoors in which the release of the PFASs is not intended. The default release factors (ERFs) associated with ERCs 10a and 11a are shown below (Table B.99 and Table B.100). Table B.99. Default ERFs for ERCs 10a and 11a. Release to Environmental Environmental Release Factors (%) Compartment ERC 10a ERC 11a To air 0.05 0.05 To water 3.2 0.05 To soil 3.2 Not applicable* * Emissions to soil indoors are not considered relevant for soil in this ERC (EC HA, 2016a). The annual release is driven by the stock of an article in use. The annual release is derived from the ERF divided by the service-life (T service) of the article. To be representative of the uses in the scope a 1-year life for paper and board packaging was assumed and a 3 year life for fluoropolymer applications (consumer and industrial cookware, and industrial food and feed production equipment). • • One year service-life ERF = ERF(air/water/soil) default x 1/20 Three year service-life ERF = ERF(air/water/soil) default x 3/20 Other uses: Lacquers & ink, beverage can coating and car wrapping Tonnages have been multiplied by Adjusted Environmental Release Factors (ERFs) for ERCs 10a and 11a. (The ERF is an ERC corrected for the lifespan of the article). Emissions were finally based on outdoor use, ERC10a, because of a precautionary principle. Table B.100. Emission factors for “Other uses” during service life . Type of Use Generic plastic packaging (processin g aids) Residues in packaging # Lacquer s and ink Car wrapping (protective sheet during transport) Drinking and beverage Service life stage / Type of PFAS Service life /outdoo r nonpolymeric PFAS nonpolymeric PFAS nonpolymeri c PFAS Fluoropolyme rs Fluoropolyme rs Drinking can lubricant coating (PTFE wax/ micropowder PTFE) Fluoropolyme rs ERC 10b ERC 10a ERC 10a ERC 10a ERC 10a ERC 10a 535 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.8. Ski wax Table B.101. Summary of assumptions and factors applied to (emission) data . Component Value Assumption and justification Assumption: There is reasonable market data to suggest that 60% of global manufacture of ski-wax occurs within the European Union. However, data on imports/exports has not been identified. Imports = 0 t/y. Import/export Exports = 0 t/y. All manufacture remains for use in the EU. Working concentrations of PFAS in skiwax Average concentration of 7.6% w.w Use ERC emission scenario for ‘formulation into a mixture’ Emission factors during formulation of ski-wax Losses during storage. 2.5% w.w to air; 2% w.w to wastewater; and 0.01% to soil. Furthermore assume 0.5% w.w as solid waste residue for waste management. Assume 0% from spillages/accidents. Assume 0% from volatisation to air. Justification: The justification is that the EU is a primary location for manufacture of ski-wax and can be largely assumed to meet all EU demand. Therefore, the Dossier Submitters do not expect any imports. It is possible that the EU is a net exporter but data on exports has not been identified. It is therefore assumed all manufacture and use remains inside the European Union. Assumption: Industry responses indicate working concentrations for PFAS between 0.2 – 15% w.w in ski-wax. However, specific concentrations may vary further (for example the call for reference makes comment on ‘pure fluoro’ waxes which are 100% PFAS mixtures). Based on the range, a middle value of 7.6% w.w has been used to calculate total quantities of all PFAS (which includes both fluoropolymers and non-polymeric PFAS) in use for ski-waxes in the European Union. Justification: The range provided by industry is relatively broad. In lieu of better data it is possible to justify the use of a middle value recognising the potential high uncertainty in the final outputs. Assumption: Data on monitoring of emissions for formulation of ski-wax has not been identified. The production of ski-wax is expected to be completed at elevated temperature using both liquid and solid processes. Therefore, emissions to air (from thermal processes), wastewater (from liquid processes) and generation of solid waste residues (liquid and solid processes) are expected. Trace amounts of PFAS are expected in these wastes. In lieu of better data values from the ERC for formulation of mixtures have been used (see ERC No2). Justification: Potential emission pathways can be readily identified, however assigning emission factors is more challenging. For consistency against other PFAS-based Annex XV dossiers the current values have been assigned from the ERC . Assumptions: Storage is a potentially important life-cycle stage for other sectors and so has been included within the emission model. Industry feedback suggests that storage may cover a period of up to 2 years (including both warehousing and stocking within retail outlets). The product is sold as both a solid (wax) and liquid formulation. Therefore, accidental release is possible but 536 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Component Value Assume 80% is lost during application. However, feedback from industry and NEA is that application sites are typically well maintained and lost material may be recaptured (e.g. through vacuuming/sweeping)and consigned to waste (typically thermal destruction/landfill). In use emissions (application) In lieu of better data assume 50% of the lost material is recaptured, and 50% genuinely lost to environment. The lost fraction is then evenly split between water and land. This gives the following fractions: Recaptured and consigned to waste = 40% Lost to land (true emission) = 20% Lost to water (true emission) = 20% Retained on the skis/snowboard and pass to next life-cycle stage = 20% In use emissions (service life) 90% over the course of active use. Split: 50% to water 50% to land Assumption and justification perhaps less likely. In lieu of better data releases are assumed to be zero in this life-cycle phase. Justification: For completeness, this life-cycle has been included within the model. Potential accidental losses could be assumed to be very low and therefore to avoid increasing the uncertainty in the estimates and simplicity of the model the emissions calculated to be zero. Assumption: The in-use phase is split in two, with application the first sub-phase. The interview notes from the consultation (see Annex E.2.7.) comment that application efficiency can be quite poor with a range of values quoted. The average of these values is 80% loss. The interviewees also commented that application can take place directly over snow or indoors. Further feedback suggests a significant proportion of lost material may be recaptured and consigned to waste (e.g. vacuuming). Therefore, it is assumed that 80% is initially lost, with 40% (half) recaptured and consigned to waste and 40% (half) truly lost to the environment, with an equal split between water and land. Justification: The consultation interviews with industry provided good data on the application efficiency which allows estimates to be made. Realistically application will occur indoors and outdoors. As a worst case assume 100% occurs outdoors but recognise that this may overestimate direct environmental losses (note even losses indoors will reach environment either through tracking on footfall or washing to wastewater). Assumption: Emission rates for the active use of ski-wax during skiing have not been identified. Where wax is applied to the underside of skis or snowboards and is eroded during use (requiring further application) it is assumed 100% of the wax used is emitted to environment. In lieu of better data it is assumed 10% of the wax purchased is not used and enters the waste cycle (this is where 537 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Component Landfill/ incineration. Value 100% of formulation wastes are incinerated. Recaptured waste from application and from inuse phase (left overs) is discarded in general waste. 50% of general waste goes to landfill 50% of general waste inuse goes to incineration. Assumption and justification the ski-wax bars are worn particularly thin, or liquid products are residue amounts in containers). Emission will be directly to snow/ice. Assumptions about melted ice and run-off to rivers would require significant data and is a high complex process. Therefore, for high level estimates an assumption is made that equal amounts reach surface water and soil, respectively. Justification: Upon use of ski-wax the emission to environment is absolute. i.e., the function of use causes the wax to be emitted. It is possible that close to 100% of the ski-wax purchased is used, however, where winter sports are seasonal, it is more likely that some remaining product does enter the waste cycle (i.e., end of season/end of vacation, end of life of ski, etc). An expert judgement of 10% has been applied for remaining fraction to waste. Assumption: Data on wastes for ski-wax has not been identified. Feedback from industry suggests that it should be incinerated, but what happens in practice is less clear. It is assumed that manufacturers will have greater control over the formulation of ski-wax, so all solid wastes from this life-cycle stage will be incinerated. The ‘in-use’ phase covers a wide set of activities, with the most likely pathway being disposal to public bins or household waste. It is assumed that equal portions are landfilled and incinerated Justification: Data on waste has not been identified and is a very challenging element of the model. Feedback from industry does suggest incineration would be preferred. It can be justified that manufacturers have greater control over wastes from formulation. Wastes from in-use phases, are more likely to be typical of other wastes discarded to public bins or household refuse. 538 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.11. Transport Table B.102. Annual emissions to air, water and soil, calculated using ERC and release % as shown below. Tonnage [t] (0.35 kg/ vehicle): Tonnage [t] (0.8 per vehicle): min max 48 607.94 111 103.87 10a 48 607.94 111 103.87 11a a ERC service life [a] a Environmental release [%] air water soil 11.95 0.05 3.2 3.2 11.95 0.05 0.05 Average age of a vehicle in Europe (AC EA, 2021) Emissions were calculated using: Fluoropolymers used in 2019 (t) x compartment specific release factor from ERC (%)/service life (y)   50% of the fluoropolymer volume was multiplied with each compartment specific release factors from ERC 10a (Widespread use of articles with low release (outdoor)) 50% of the fluoropolymer volume was multiplied with the compartment specific release factors from ERC 11a (Widespread use of articles with low release (indoor)). 539 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.12. Electronics and semiconductors In Table B.103 the selected ERC are included. In Table B.104, the release factors per ERC are included (ECHA, 2016a). As the data made available for electronics and semiconductor industry is limited to a single section in the production, it is not possible to differentiate between ERC based on, if it is the formulation or industrial use. The ERC in Table B.103 describes the production (industrial use) and the use stage. Table B.103. Selected Environmental Release Categories (ERC). Production Use Fluoropolymers Perfluoropolyethers 3 5 11A 11A Side-chain fluorinated polymers Fluorinated gases 5 4 11A 9A Ionic non-polymeric PFAS Non-ionic non-polymeric PFAS 5 4 11A 9A Table B.104. Calculated one compartment based on default release factors (%) per ERC . ERC no 3a 4 9A b ERC description Formulation into a solid matrix Use of non-reactive processing aid at industrial site (no inclusion into or onto article) Use at industrial site leading to inclusion into/onto article) Widespread use of functional fluid (indoor) 10 11A Widespread use of articles with low release (indoor) 0.1 5 One compartment (%) 1.3 100 50 aERC 3 one compartment has been amended in the use for fluoropolymers, as emission to air is not relevant. The amended ERC 3 is 0.3%. bERC 9A one compartment has been amended in the use for non-ionic non-polymers and fluorinated gases, as emission to soil is not relevant. The amended ERC 9A is 5% A study by the Beu and Raoux (2019), on greenhouse gases in the electronics industry, indicates that 89% to 99% of fluorinated gases is captured, and 1% and 11% is emitted. Based on these results, a calculation factor of 0.05 (5% will be emitted, 95% captured) is used to estimate the total emissions of PFASs used for cleaning or as solvents (non-ionic non-polymers and fluorinated gases). It is likely that larger emissions happen during production of day-to-day electronics than for semiconductors, hence the emission results for these sub-categories may likely be an underestimation of the emissions. Other groups have a calculation factor of 1 (100%) as all tonnages not emitted are expected to remain in the products until the use stage. In Table B.105, emission factors for each PFAS group in production and use stages are included. The values are a result of combining Table B.103 with Table B.104. Table B.105. Emission factors (%) to the environment based on a one compartment model, corresponding with emissions scenarios from ERC. Production (%) Use (%) Fluoropolymers: Perfluoropolyethers 0.3 50 0.1 0.1 Side-chain fluoropolymers 50 0.1 Fluorinated gases 100 5 540 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Production (%) Use (%) Ionic non-polymer 50 0.1 Non-ionic non-polymers 100 5 a For fluorinated gases and non-ionic non-polymers a combined release factor for production and use was used. 541 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.13. Energy In Table B.106, the ERC are included. In Table B.107 the release factors per ERC are included (ECHA, 2016a). As the data made available for the energy industry is limited to a single section in the production, it is not possible to differentiate between ERC based on, if it is the formulation or industrial use. The ERC in Table B.106 describes the production (industrial use) and the use stage. Table B.106. Selected Environmental Release Categories (ERC). Production Use Fluoropolymers Perfluoropolyethers 3 5 11A 11A Side-chain fluorinated polymers Ionic non-polymeric PFAS 5 5 11A 11A Non-ionic non-polymeric PFAS 4 9A For fluorinated gases and non-ionic non-polymers a combined release factor for production and use was used. a Table B.107. Calculated one compartment based on default release factors (%) per ERC . ERC no ERC description One compartment (%) 3a Formulation into a solid matrix 1.3 Use of non-reactive processing aid at industrial 100 site (no inclusion into or onto article) Use at industrial site leading to inclusion 50 5 into/onto article) 9A b Widespread use of functional fluid (indoor) 10 Widespread use of articles with low release 11A 0.1 (indoor) aERC 3 one compartment has been amended in the use for fluoropolymers, as emission to air is not relevant. The amended ERC 3 is 0.3%. 4 b ERC 9A one compartment has been amended in the use for non-ionic non-polymers and fluorinated gases, as emission to soil is not relevant. The amended ERC 9A is 5%. A study by the Beu and Raoux (2019), on greenhouse gases in the electronics industry indicates that 89% to 99% of fluorinated gases is captured, and 1% and 11% is emitted. Based on these results, a calculation factor of 0.05 (5% will be emitted, 95% captured) is used for the total emissions of PFASs used for cleaning or as solvents (non-ionic nonpolymers) in the energy sector as these short chain substances have similar properties to fluorinated gases or degrade to such gases. Limited information on emissions was received from stakeholders, and therefore the emissions are based on the electronics and semiconductor sector. It is likely that production of specialised energy equipment such as fuel cells occurs under similar conditions to specialised electronics (e.g. semiconductors). It is likely that these emission scenarios will lead to an underestimation of emissions from the energy sector, as some production may occur under less controlled environments. All other PFASs groups have a calculation factor of 1 (100%) as all tonnages not emitted are expected to remain in the products until the use stage. In Table B.108, emission factors for each PFAS group in production and use stages are included. The values are a result of combining Table B.106 with Table B.107. Emission factors (%) to the environment based on a one compartment model, corresponding with 542 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) emissions scenarios from ERC. Table B.108. Emission factors (%) to the environment based on a one compartment model, corresponding with emissions scenarios from ERC. Production (%) Use (%) Fluoropolymers: Perfluoropolyethers 0.3 50 0.1 0.1 Side-chain fluoropolymers Ionic non-polymer 50 0.1 50 0.1 Non-ionic non-polymers 100 Emissionproduction = Volume production x C alculation factor 5 Emissionuse = Volume use x C alculation factor With: Volume production = the volumes in section A.3.13.2. 543 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.14. Construction products Assumptions for the assignment of ERC and release factors Split of volume between articles and construction mixtures. Based on the categories ‘Building and construction’ and ‘Coating and paints’ in Glüge et al. (2020) for fluoropolymers a split of 19%/81% in articles/construction mixtures is estimated. For non-polymeric PFASs a split into articles/construction mixtures is estimated to 81%/19%. All side-chain fluorinated polymers is assumed to be used in construction mixtures Processing aids Based on feedback from one stakeholder the use of processing aids in the production of construction products the processing aid is captured within closed systems and 99.9% re-used. Application. A 50:50 split between outdoor and indoor applications is assumed in lieu of better data In Table B.109, the selected ERC for the total release to the environment are presented. Table B.109. ERC for the total release to the environment. Total release to Life-cycle stage ERCs the environment Formulation 2 4.51% (coatings) Formulation (articles) 3 1.3% Processing aids* 4 100% Application (coatings) 5 50% Use (outdoors) 10A 3.25% Use (Indoors) 11A 0.1% *99.9% is considered by industry to be recycled. As can be seen from Table B.109, the same ERCs are used across the different PFASgroups. 544 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.15. Lubricants In Table B.110, the selected ERC for the total release to the environment are presented. Table B.110. ERC for the total release to the environment. Total release to Assumed ERCs Life-cycle stage the splits environment Formulation stage Use Stage-Sealed equipment Use Stage-open application C losed/sealed 95% 2 4.51% Open 5% Indoor 70% 9A 10% Outdoor 30% 9B 10% Indoor 70% 8A 90% Outdoor 30% 8D 90% Indoor 50% 8A 90% Outdoor 50% 8D 100% Pre-use stage (use of cleaning agents) Other assumptions ERC 8A and 8D with amendment. Assume 10% remains as residue on the article ERC 8A with amendment. Assume emission control technology is in place for 10% 545 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.16. Petroleum and mining Flow diagrams for estimated emissions from tracer products, anti-foaming agent products and fluoropolymer products (low and high scenarios) can be found in Baseline year: 2020 Notes on ranges of data: [1] Scenario 3; [2] Scenario 1; [3] Scenario 1 and 2; For a full description of each scenario, see NEA (2021). * C alculated using Environmental Release C ategory (ERC ) emission scenario no.2 for ‘formulation into a mixture’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 2.5% w.w to air; 2% w.w to wastewater and 0.01% w.w to soil. Figure B.81 and Baseline year: 2020 Notes on ranges of data: [1] Scenario 3; [2] Scenario 1; [3] Scenario 2 and 3; For a full description of each scenario see NEA (2021). * C alculated using Environmental Release C ategory (ERC ) emission scenario no.2 for ‘formulation into a mixture’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 2.5% w.w to air; 2% w.w to wastewater and 0.01% w.w to soil. Figure B.82. A source flow diagram for estimated emissions from fluoropolymer products (low scenario) is given in Baseline year: 2020 [1] C alculated using the total calculated residual monomeric PFAS content of the fluoropolymer and the total emissions across all use stages * C alculated using Environmental Release C ategory (ERC ) emission scenario no.3 for ‘Formulation into a solid matrix’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 30% w.w to air; 0.2% w.w to water (before STP) and 0.1% w.w to soil. Figure B.83. The high scenario picture for emissions from fluoropolymer products is given in Baseline year: 2020 [1] C alculated using the total calculated residual monomeric PFAS content of the fluoropolymer and the total emissions across all use stages * C alculated using Environmental Release C ategory (ERC ) emission scenario no.3 for ‘Formulation into a solid matrix’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 30% w.w to air; 0.2% w.w to water (before STP) and 0.1% w.w to soil. Figure B.84. 546 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Baseline year: 2020 Notes on ranges of data: [1] Scenario 3; [2] Scenario 1; [3] Scenario 1 and 2; For a full description of each scenario, see NEA (2021). * C alculated using Environmental Release C ategory (ERC ) emission scenario no.2 for ‘formulation into a mixture’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 2.5% w.w to air; 2% w.w to wastewater and 0.01% w.w to soil. Figure B.81. Source flow diagram for estimated emissions from tracer products. 547 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Baseline year: 2020 Notes on ranges of data: [1] Scenario 3; [2] Scenario 1; [3] Scenario 2 and 3; For a full description of each scenario see NEA (2021). * C alculated using Environmental Release C ategory (ERC ) emission scenario no.2 for ‘formulation into a mixture’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 2.5% w.w to air; 2% w.w to wastewater and 0.01% w.w to soil. Figure B.82. Source flow diagram for estimated emissions from anti-foaming agents products. 548 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Baseline year: 2020 [1] C alculated using the total calculated residual monomeric PFAS content of the fluoropolymer and the total emissions across all use stages * C alculated using Environmental Release C ategory (ERC ) emission scenario no.3 for ‘Formulation into a solid matrix’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 30% w.w to air; 0.2% w.w to water (before STP) and 0.1% w.w to soil. Figure B.83. Source flow diagram for estimated emissions (low scenario) from FP products (fluoropolymer products). 549 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Baseline year: 2020 [1] C alculated using the total calculated residual monomeric PFAS content of the fluoropolymer and the total emissions across all use stages * C alculated using Environmental Release C ategory (ERC ) emission scenario no.3 for ‘Formulation into a solid matrix’ from EC HA (2016) Guidance on information requirements and C hemical Safety Assessment, C hapter R.16: Environmental exposure assessment. 30% w.w to air; 0.2% w.w to water (before STP) and 0.1% w.w to soil. Figure B.84. Source flow diagram for estimates emissions (high scenario) from FP products (fluoropolymer products). 550 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Appendix B.9.21. Human exposure Table B.111 and Table B.112 summarise concentrations of PFASs in indoor air and dust and include all measured PFASs, including precursors and PFAAs covered by existing and/or proposed restrictions. Table B.111. Concentration of PFAS in indoor air (pg/m 3 ) described by detection frequency (% or x/xx), median in bold and range within brackets. Site Countr PFCAs PFSAs FTOHs FOSAs FOSEs FTACs 6:2 Refere (N) y, year FTMA nce C Homes (10) Spain, C ataloni a 2009 6:2FTOH: 100% (2.0-47) 8:2FTOH: 100% 42 (7.5170) 10:2FTOH: 80% (<0.6-47) ΣFTOHs: 100% (13-234) ΣFOSE/FOSA s: (0.02-263) PFTA: 5% PFSAs FTOHs FOSAs FOSEs 6:2FTOH: 97% 1 040 (<722 890) 8:2FTOH: 100% 2 720 (660-16 080) 10:2FTOH: 100% 980 (2208 160) MeFOSA: 98% 21 (<1.4812) EtFOSA: 100% 19 (3.1-910) MeFOSE: 100% 320 (54-9 520) EtFOSE: 97% 56 (<0.9-1 000) FTACs 6:2 FTMA C Refere nce 0.7 (<0.22.9) lin-PFOS: 88% 1.2 (<0.55.0) PFOS: 0% (<0.02) (Shoei b et al., 2011) 554 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Site (N) Countr y, year PFCAs PFSAs FTOHs FOSAs FOSEs FTACs 6:2 FTMA C Refere nce (<0.03-3.7) Table B.112. Concentrations (ng/g) of PFASs in indoor dust described by median in bold and range within brackets. DCC=daycare centers, CR=classrooms. Locat Coun PFSAs PFCAs MonoPA DiPAPs FTCA/ FTSAs FTOH FOSA FOSE FOSAA Other Referen ion try Ps FTUCA PFASs ce (N) Year (Goosey PFOA: 240 FOSA: PFHxS: MeFOSE: DCC & UK and (<0.02660 2007- 700 (16- (18-1 700) CR Harrad, 750) 34 000) (<0.02-8 08 (42) 2011) MeFOSA: 400) PFOS: ND (<0.1) EtFOSE: 840 (22-3 700) 370 EFOSA: (<0.1230 13 000) (<0.07640) MeFOSE: PFOA: 31 FOSA: Home Franc PFHxS: 130 (15-220) 3.0 77 (54s (10) e (<0.22(<0.022007- 320) 610) 300) PFOS: 09 MeFOSA: EtFOSE: 160 (54-1 (<0.1-31) 140 700) EtFOSA: (<0.12130 (23- 550) 320) PFOA: 300 FOSA: 47 MeFOSE: Home Germ PFHxS: 38 (<0.02s (10) any 150 (16- (19-730) (<0.22130) 2007- 790) MeFOSA: 700) 09 PFOS: EtFOSE: <0.1 170 (47-1 (<0.1-16) 120 (11000) EtFOSA: 180) 555 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Locat Coun PFSAs ion try (N) Year PFCAs MonoPA DiPAPs Ps FTCA/ FTSAs FTUCA FTOH FOSA FOSE FOSAA Other PFASs Referen ce 120 (36730) PFOA: 190 PFHxS: Home UK s (45) 2007- 210 (7.6-6 (<0.98-4 100) 100) 09 PFOS: 140 (3.5-7 400) PFBA: ND Home Norwa PFBS: PFPeA: s (41) y 0.40 2008 (0.17-9.8) 3.0 (1.529) PFHxS: PFHxA: 28 0.60 (0.21-142) (4.3-96) PFHpA: PFHpS: 9.4 (4.50.19 (0.10-2.1) 28) PFOS: 3.1 PFOA: 18 (6.2-56) (1.2-94) PFDS: 1.1 PFNA: 23 (0.15-42) (3.9-92) PFDA: 4.1 (1.1-12) PFUnDA: ND PFDoDA: 19 (1.4- 6:2FTU 6:2FTS: CA: 10 4.8 (2.2(4.353) 301) 8:2FTS: 8:2FTU 5.3 (2.2CA: 3.6 99) (1.078) FOSA: 20 MeFOSE: 93 (0.22(<0.022 500) 300) MeFOSA: EtFOSE: 34 <0.1 (<0.12-3 (<0.1900) 110) EtFOSA: 40 (<0.07840) MeFOSA: MeFOSE: 0.51 9.0 (3.5(0.27-1.1) 90) EtFOSA: EtFOSE: 0.61 ND (0.26-33) PFOSA: (Haug et 0.45 al., (0.22-41) 2011b) 556 ANNEX XV RESTRICTION REPORT – Per- and polyfluoroalkyl substances (PFASs) Locat Coun PFSAs ion try (N) Year Home Germ PFOS: 20 s (31) any (3.3-1 2008- 046) 09 Home Spain PFBS: s (10) 2009 0.36 (<0.0016.5) PFHxS: 0.44 (0.17-5.3) PFOS:2.4 (1.1-12) PFCAs MonoPA DiPAPs Ps FTCA/ FTSAs FTUCA 78) PFTrDA: 6.8 (1.146) PFTeDA: 3.3 (1.135) PFOA: 39 (6.1-676) PFBA: 19 (7.2-25) PFPeA: 0.36 (<0.0130.93) PFHxA: 1.0 (0.402.9) PFHpA: 1.2 (0.46- 6:2FTU CA: (<0.00 2-0.08) 8:2FTU CA: (<0.00 2-0.08) FTOH ∑ FTOHs: 26 (4.8734) 4:2FTOH: ND 6:2FTOH: 3.7 (