Document 7ObOQe6j9b46VZe57wXDRVMVo

c iff 3M General Offices 3M March 17, 2000 3M Cerner Y E fj t-0 3 0 0 - 0 3 1 3 ..St. Paul, MN 55144-1000 612 733 IU0 AR13605HS RECEIVED OPPT CBiC 2USMAR 2 G M !! i3 Document Processing Center (7407) (Attn: Section 8(e) Coordinator) Room G-099 OPPT US EPA 401 M Street, SW Washington, DC 20460 & E P A -OTS 00D1177SX RE: TSCA 8(e) SUPPLEMENTAL NOTICE: Sulfonate-based Fluorochemicals Docket Nos. 8EHQ-1180-373; 8EHQ-1180-374 To Whom It May Concern: Enclosed you will find a white paper entitled "Sulfonated Perfluorochemicals in the Environment: Sources, Dispersion, Fate and Effects'' prepared by 3M. 3M is submitting this document as a supplement to our original 8(e) submissions on sulfonate-based fluorochemicals. If you have any questions, please feel free to contact me. Best regards, RECEIVED 3FPTHCIC William A. Weppner, Ph.D. Director Environmental, EEealth, Safety & Regulatory Affairs Specialty Material Markets Group 3M Center, Bldg. 236-1B-10 St. Paul, MN 55144 C: Enc. 00000* rx> cn / Q f p J "~ 0 0 ? ~ O O f y -- 0 O S " RECEIVED DP?T CBiC 2000 HAR 2 0 A M i i : i > 3 Sulfonated Perfluorochemicals in the Environment: Sources, Dispersion, Fate and Effects Prepared by 3M March 1,2000 i 0 ^ 0 0 0 2 > Table o f Contents 1.0 PREFACE 2.0 EXECUTIVE SUMMARY 3.0 INTRODUCTION TO FLUOROCHEMICALS 4.0 PHYSICAL-CHEMICAL PROPERTIES OF FLUOROCHEMICALS 5.0 ANALYTICAL TEST METHODS FOR FLUOROCHEMICALS 6.0 SOURCES OF FLUOROCHEMICALS 6.1 M anufacturing Waste Streams 6.11 W aste Stream Characterization 6.12 A ir 6.13 W astew ater . 6.14 Solid Waste 6.2 Supply Chain Waste Streams 6.3 Releases from Waste Treatment and Disposal Methods 7.0 ENVIRONMENTAL TRANSPORT AND DISTRIBUTION 8 .0 ENVIRONMENTAL SAMPLING FOR FLUOROCHEMICALS 8.1 Environmental Levels 8.11 Historical Data 8.12 Recent Analyses o f Wild Birds and Fish 8.13.T estingofFishm ealU sedinR atStudies. 8.14 Plant Site Analyses 8.15sBiosphere Sampling ' 8.2 Human Exposure Levels 8.21 M ulti-cities Sampling 8.22 Carpet Use Studies 8.23 Paper and Packaging Studies 8.24- Exposure Scenarios 2 4 5 9 13 17 19 20 22 22 22 23 24 24 25 27 27 27 27 31 31 32 32 33 33 33 33 <!' 1 9.0 ENVIRONMENTAL TRANSFORMATION/DEGRADATION OF FLUOROCHEMICALS 9.1 H ydrolysis Studies 9.2 Photolysis 9.3 A tm ospheric Studies ' ... 9.4 Biodegradation Studies 9.41 M icrobial Studies cm Perfhioroclieuiigals 9.42 Biological Transformation 9.43 Optimizing Conditions for Biodegradation 10.0 ECOTOXlClTY TESTING OF FLUOROCHEMICALS 11.0 COMPREHENSIVE PLAN TO ASSESS ENVIRONMENTAL EXPOSURE 11.1 Plan Overview 11.2 Com ponent 1: C haracterize Fate and T ransport Properties 11.3 C om ponent 2: Estim ate Releases . 11.4 C o iu p o u cu t 3: C h aracterize D istribution lu th e E nvironm ent 11.5 Component 4: Estimate exposure 12.0 ECOTOXICITY DETERMINATIONS 13.0 ECOLOGICAL RISK EVALUATION 14.0 REFERENCES 34 ' 36 37 37 33 33 39 39 .4 0 45 45 48 48 48 49 49 50 51 3 ; 1.0 Preface This paper provides an overview o f 3M's current knowledge about the sources, dispersion, fete and effects o f some o f its fluorinated chemical products. It specifically addresses sulfonated perfluoronated chemistry and products, w ith the major focus on those compounds w ith an eight carbon chain structure. There are other fluorinated chemical products but these are not covered in this white paper. The paper presents the past testing o f these chemicals for environmentally relevant properties and assesses the quality and adequacy o f past testing. It also presents recent results o f environmental sampling, estimates o f quantities of wastes generated at m anufacturing plants and from product use, and new data on physical, chemical and ecotoxicological properties o f sulfonated perfluorochemicals. It describes in detail the comprehensive exposure assessment plan currently being implemented. This plan is aimed at providing a better understanding o f the transport, fete and effects o f these chemicals in the environment and will help the company determine appropriate future actions. As these studies return data, test plans will be revised to incorporate new information. For this reason, the results o f present testing should be treated cautiously. Some data represent first attempts at characterization o f complex chemicals in very difficult and dynamic environmental test matrices. The program incorporates new analytical technology, complex models and many variables. Theso initial findings arc subject to change as results from currently planned testing on degradation, biological receptors, wastes from manufacturing facilities and other exposure data are obtained. This paper should be read in conjunction wife previous submittals about the health and environm ental issues associated with 3M's sulfonated perfluorochemical product line. In January 1999,3M submitted to fee Environmental Protection Agency (EPA) a report, Perfluorooctane Sulfonate: Current Summary o f Human Sera, H ealth and'Toxicology Data, that provided details o f analyses o fpooled blood sera samples that demonstrated the presence o f perfluorooctane sulfonate (PFOS) at very low levels. In February 1999,3M provided a comprehensive review, The Science o f Organic Fluorochemistry, describing the health effects and background chemistry associated w ith PFOS. Another report subm itted to EPA in May 1999, Fluorochemical Use, Distribution, and Release Overview, describes how 3M produces sulfonated perfluorochemicals, which product lines incorporate them, and the uses for these products. Finally, various Section 8(e) submissions have been forwarded to EPA relative to these sulfonated perfluorochemicals. 4 2.0 Executive Summary 3M produces sulfonated perfluorochemicals by an electrochemical fluorination process. This process creates a complex and variable mix o f chemicals in which fluorine atoms replace hydrogen atoms on the organic feedstock and carbon-carbon bonds are rearranged. Because o f the carbon-fluorine bond formed by this process, the compounds created are considered to be very stable. Perfluorochemicals have complete substitution' p f fluorine for hydrogen. Fluorochemicals can repel both water and oils, reduce surface tension dramatically, act as catalysts for oligomerization and polymerization, and function under extreme conditions. Major uses for sulfonated perfluorochemicals are surface protectors and surfactants. Fluorine's high electronegativity confers a strong polarity to carbon-fluorine bonds, contributing to the stability and nonreactive character o f perfluorochemical molecules. They are unusual as that perfiuoroalkyl chains are both oleophobic and hydrophobic. The addition o f charged moieties to the chain may affect the water solubility o f the shorter chains. The highest volume sulfonated perfluorochemical produced by 3M is perfluorooctanesulfonyl fluoride (POSF). After synthesis, it is used to create several product lines. During their life cycles, POSF and POSF-based products may degrade. If degradation occurs, current research suggests perfluorooctane sulfonate, (PFOS) and a few other perfhiorinated forms ore degradation products. Timeframes for degradation arc variable, w ith some polymeric products apparently stable for very long periods o f time. The identification and quantification o f sulfonated perfluorochemicals pose difficult analytical challenges. Reliable methods for extraction, separation and identification o f sulfonated perfluorochemicals in tissues and environmental matrices have evolved and have been developed only in the last few years. New analytical technology is providing capabilities o f detection in wide varieties o f matrices at parts per trillion (ppt) levels and identification o f metabolites and breakdown products. As fully described throughout this paper, completion o f a comprehensive exposure assessm ent and related scientific studies w ill require many years o f intensive research. 3M is pursuing an aggressive program to reduce releases to the environment w hile that scientific research is being conducted. It is not the purpose o f this paper to describe the nature and extent o f that undertaking. Readers should be made aware, however, that 3M has initiated a wide range o f activities to utilize available opportunities for reductions in releases. These have included installation o f new controls to reduce waste stream s in 3M manufacturing facilities. They have also included product stewardship efforts to communicate to customers and downstream users, information regarding fluorochemicals and the need to exert careful management over these substances. In addition, 3M has rtrWfalran m ajor efforts to reinvent its products through the use o f alternative chemistry to reduce the volume o f fluorochemicals used in those products. All o f these efforts w ill 5 be continued w ith intensity while the scientific research described in this paper is carried out. 3M is examining the life cycle o f its sulfonated perfluorochemical products to identify releases to the environment from manufacturing processes, supply chains, product use and disposal. First it is determining waste streams generated throughout the life cycle. This inform ation w ill be used to estimate environmental releases. This approach is necessary since not all waste produced is released to the environment. Manufacturing waste studies are underway at the 3M plant in Decatur, Alabama on POSF-based processes. PFOS-based waste streams generated from these manufacturing processes are conservatively estimated to be about 1.1 million pounds per year, about 90% as solid waste, m ost o f which is incinerated and destroyed. Recent wastewater controls have reduced amounts o f PFOS actually discharged to the river by half since 1998. Data from business units have identified key products that contain the majority o f the fluorochemical solids sold in the United States in 1997. Using this sales information, 3M estimated customer and end user waste streams. M ost o f the waste generated from these sources is in the form o f solid waste. Releases to the environment from product disposal to landfills, wastewater treatment plants and incineration are all being investigated. S ev eral d iffe re n t fate and tra n sp o rt mechanisms have been identified as im portant to study. Initially models are being used for screening-level assessments o f potential fete mechanisms, M ulti-media fugacity models are under development to incorporate the unique properties o f fluorochemicals. - Sulfonated perfluorochemicals have been detected at low levels in some species o f eagles and w ild birds. Low levels were detected in bird plasma and bird livers. 3M believes that these sets o f data are insufficient to draw conclusions w ith any statistical merit. In screening sampling o f the river and sediments near the Decatur manufacturing plant, PFOS was present in a few samples collected near the outfall. All this inform ation was used in fee design o f a more comprehensive program o f biosphere sampling. The goal o f the biosphere sampling plan i3 to screen for PFOS across a range of species, habitats ami geographic locations and to identify areas on which to focus scientific investigation to develop a better understanding o f any potential environmental effects. A m ulti-cities study will determine environmental distribution and potential sources o f human and ecological exposure. The multi-cities study pairs cities w ith significant manufacturing or commercial use o f fluorochemical products w ith cities o f the same size without significant use. Levels o f PFOS and its precursors w ill be measured in food, air, water, sediment, and disposal facilities. Additionally, levels are being measured arising from carpet use, product uses and potential migration into food from packaging. The role o f hydrolysis, photolysis and biological processes in fee degradation o f sulfonated perfluorochemicals is being studied. Research suggests that the biodegradation o f fluorinated sulfonates requires the presence o f hydrogen at fee alpha .6 1 carbon on the fluorinated chain and that perflunrinated molecules are susceptible to breakdown only at non-fluorinated side chains. Degradation o f sulfonated perfluorochemicals is not complete but results in production o f other fluorochemicals. Studies suggest that compounds made from POSF, a commercially important . perfluorochemical product and intermediate, are transformed during metabolism to another sulfonated perfluorochemical, PFOS. PFOS does not appear to further degrade except by incineration. .' . ,, Several sulfonated perfluorochemicals have been subjected to basic screening tests for environmental toxicity. Different species varied significantly in their response to the same chem ical even when using the same laboratory procedure. New testing is underway using purified sulfonated perfluorochemicals. measured test concentrations, and a wide variety o f test organisms. Results o fthese studies are reported in this white p ap . The research projects that are yielding new information on sulfonated perfluorochemicals are part o f a comprehensive plan to assess the potential pathways of environmental exposure associated with the manufacture, use and disposal o f sulfonated perfluorochemical products. Figure 1 portrays the plan components. Work on the plan is now underway using a combination o f 3M resources and outside experts. Recent analytical advances and this extensive research effort are expected to contribute significantly to a better understanding o f environmental fete and effects. The findings resulting from the comprehensive plan, along w ith new ecotoxicologicat test d a ta , w ill be used to evaluate ecological risk. While this evaluation is underway, 3M is im plementing actions to reduce generation o f waste in manufacturing processes and to reduce releases o f sulfonated perfluorochemicals into the environment through process improvements, waste reduction and engineering redesign. 7 i Figure 1. Diagram o f Florochemical Assessment Plan. 8 3.0 Introduction to Fiuorochemicals Fiuorochem icals arc components o f several important 3M product lines due to their unique and useful properties. They are stable, chemically inert and generally nonreactive. A s components o f products, they repel both water and oU, reduce .surface tension much lower than other surfactants, act as catalysts for oligomerization and polymerization, and function where other compounds would rapidly degrade. 3M has produced fiuorochemicals commercially for over 40 years. 3M produces fiuorochemicals by combining anhydrous hydrogen fluoride w ith hydrocarbon stock in the presence o f electrical energy. The highest volume sulfonated fluorochemical produced by 3M is perfluorooctanesulfonyl fluoride (PGSF). C*H17S 0 2F + 17H F 4.5-7;0 V --------> C8F ,,S 02F + 17H 2 1-Octanesulfonyl fluoride Perfluorooctanesulfonyl fluoride (PO SF) The fluorination process overall yields about 35-40% straight chain (normal) POSF, and a mixture o f byproducts and waste of uncharacterized and variable composition containing: -higher or lower straight chain homologues, n -C Ji^ S O jF , o f various chain lengths (7% o f process output) e.g. C6F13S 0 2F, C7Fi5S 02F, C9F19S 02F -branched chain perfluoroalkyl products o f various chain lengths (18-20% o f output) CF3 CF3 CF3 I e.g. CF3CF2CF2CF2CF2CFCF2S 02F II. c f 3c f c f 2c f 2c f c f 2s o 2f - gtraight chain, branched and cyclic perfluoroalkanes and ethers (20-25% o f output) e.g. CF4 C^Fi, CjFg, C4F10, C jF n , cC4Fg -"tars" (high molecular weight fluorochemical byproducts) and other byproducts, including m olecular hydrogen (10-15% o f output). Because o f slight differences in process conditions, raw materials, and equipment, the m ixture produced by the electrochemical fluorination process varies somewhat from lot to lot and from plantto p lan t Numerous process steps are used to convert the fluorinated m ixture into final products. , 9 f The largest production o f fluorochemicals occurs at the 3M manufacturing plant in Decatur, Alabama, and this plant is the focus of current studies. During production, m any byproducts and waste products are formed. The volatile waste products have been vented to the atmosphere in the past but improvements are underway to capture and destroy these releases by thermal oxidation. The tars are disposed at hazardous waste landfills or treated by incineration. The byproducts, many o f which arc incompletely fluorinated w ith hydrogen atoms still present, are recycled back into processes or partially degraded in stabilization processes and discharged to wastewater treatment systems. The treatm ent sludge is landfilled. Some o f die non-FOSF-based byproducts are recovered and sold for secondary uses. The product o f the-electroehemical fluorination process is thus not a pure chemical but rather a m ixture o f isomers and homologues. Perfluorochemicals have complete substitution o f fluorine for hydrogen. The commercialized POSF derived products are a m ixture o f approximately 70% linear POSF derivatives and 30% branched POST derived im purities. POSF is used as a product and is also an important intermediate in the synthesis o f substances used in many other 3M products. To a lesser extent, homologues o f POSF, [CnF<2n+|)S02F where n - 2-9, exclusive o f 8], are also components used in the formation o f other 3M products. Some o f the POSF derived products are surface active materials and monomers o f relatively low molecular weight (~500 daltons). These monomers are used as low m olecular weight surfactants or are joined with other monomers to form higher m olecular w eight oligomers and polymers with a mix o f fluorinated and unfluorinated portions. Fluorochemical monomers can also be joined to phosphates, to polymeric and oligomeric urethane, or to acrylate backbones through ester and other linkages. The m ajority o f 3M 's sulfonated perfluorochemicals produced are used in polymeric form for treatm ent o f surfaces and materials. For example, fluorochemical containing polymers (urethanes, acrylics and esters) can provide soil, stain, and water resistance to personal apparel and home furnishings. Some products synthesized from POSF and its homologues are sold as raw m aterials to customers who use them as intermediates or components o f their products. The interm ediates can be covalently bound to a variety o f polymeric hydrocarbon backbones. The 3M product lines that use sulfonated perfluorochemicals are summarized below. (Product'lines using fluorochemicals that contain no sulfonyl'groups are not listed.) r-1-rrfci.rv i n m n i W I 10 Surface Treatm ents Fabric/Upholstery Protector (High molecular weight (MW) polymers) Carpet Protector (High MW polymers) Leather Protector (High MW polymers) Paper and Packaging Protector (High MW phosphate esters or high M W polym ers) ... . . S urfactants (Low MW chemical substances) Specialty surfactants Household additives Electroplating and etching bath surfactants Coating and coating additives Chemical intermediates Carpet spot cleaners Fire Extinguishing Foam Concentrates Mixung and Oil Surfactants O th er Uses Insecticide Raw Materials (Low MW chemical substances) Typically a fluorochemical product contains a small amount o f fluorochemical residuals: unreacted o r partially reacted starting materials or intermediates. Residuals which are common to formulations o f sulfonated perfluorochemical products include: perfluorooctane sulfonate (PFOS), N-ethyl (or N-methyl) perfhiorooctane sulfonamide (N-EtFOSA or N-M eFOSA), N-ethyl (or N-methyl) perfluorooctane sulfonamidoethyl alcohol (N -Et FOSE alcohol or N-MeFOSE alcohol) and perfluorooctanoic acid (PFOA). Table 1 identifies some sulfonated perfluorochemicals, their acronyms, chemical nam e, and formulas. 11 Table 1. Perfluorochemical Glossary D esignation POSF PFOS PFOSH PFOS.NH4 salt PFOS.DEA salt PFOS.K salt PFOS.LLsalt FOSA . PFOSAA PFDS PFHS N-EtFOSA N-MeFOSA N-EtFOSE alcohol N-M cFOSE alcohol N-EtFOSEA. N-EtFOSEMA N-MeFOSEA PFOA Name Form ula perfluorooctanesulfonyl fluoride c *f ,7so 2f perfluorooctane sulfonate Q F ^ O j- perfluorooctanes'ulfottic acid \ ammonium perfluorooctanesufonate C ,FnSOjH CjFpSOjNH * Perfluorooctanesulfonate diethanolamine salt potassium perfluorooctanesulfonate C,F17SO,NH(CH2CHiOH)2 c 8f ,7s o 3k .......... lithium perfluorooctanesulfonate CgK17S 0 3Li perfluorooctanesulfonamide CjF17S 02NH2 perfluorooctane sulfonylamido (ethyflacetate perfluorodecanesulfonate C ,F I7SOaN(CH1C E 3)CHJCOOCioFI3S 0 3' perfhinrnhwrane sulfonate CfnSO,- N-ethyl perfluorooctanesulfonamide N-methyl perfluorooctanesulfonamide c ,f I7s o 2n h c h 3 N-ethylperfluorooctane sulfonamidoethanol C,F|7S 02N(CH2CHj)CH;,CH2OH N-methylperfluorooctane sulfonamidoethanol CjF,7S 02N(CHj)CH2CH20 H N- ethylperfluorooctanesulfonamidoeth yl acrylate Nethylperfluorooctanesulfonamidoeth yl methacrylate N-methyl- - perfluorooctanesulfonamidoethyl acrylate perfluorooctanoic acid C jF pS O iN fC H iC H jJC I^C l^O C C X ai-C aij C sFpS O jN iC jH siC H jC H jO C O C fC H ,)^!!, C2F,7S 02N(CH])CH2CH20C0CH=CH2 c 7f 15c o 2h , 12 4.0 Physical-Chemical Properties of Fluorochemicals Fluorinated organics are less well described in the science literature than organic molecules bearing other halogens, i.e. bromine and chlormej which have been more thoroughly investigated by many researchers in published reports. To understand the properties o f fluorinated organics, it is necessary to describe the properties o f fluorine. Fluorine has several characteristics that differ from the other halogens and contribute to the unusual properties o f fluorochemicals. _ Fluorine has a van der Waals radius o f 1.35 A, more comparable to that o f oxygen than other halogens, and isosterically similar to a hydroxyl group. Fluorine has the highest electronegativity (4.0 -Pauling scale) o f all the halogens, indeed the highest in the periodic table. This confers a strong polarity to the carbon-fluorine bond. The carbon fluorine bond is one o f the strongest in nature (~110 keal/mol). This very strong, high energy bond contributes to the stability o f fluorochemicals. The high ionization potential o f fluorine (401.8 kcal/mole) and its low polarizability leads to weak inter- and intramolecular interactions. This is demonstrated by the low boiling points o f fluorochemicals relative to molecular weight, and their extremely low surface tension and low refractive index. The partitioning behavior o f perfluoroalkanes is unusual. Some perfluoroalkanes when mixed with hydrocarbons and water form three im miscible phases, demonstrating that perfluorinated chains are both oleophobic and hydrophobic. A charged moiety, euch as carboxylic acid, sulfonic acid, phosphate o r a quaternary ammonium group, when attached to the perfluorinated chain, makes the m olecule more water soluble because o f the hydrophilic nature o f these charged moieties. Therefore, such functionalized fluorochemicals can have surfactant properties. Typically, the presence o f these charged groups on short chain perfluorinated compounds (<C6) noticeably increases the solubility o f the compound in water. Physical data available on fluorochemicals at 3M have been principally those param eters needed for quality control use and material handling. Table 2 summarizes the physical data for low molecular weight, POSF-based fluorochemical products that have been developed for use on M aterial Safety Data Sheets (MSDS). Some o f these perfluorochemical products are primarily used as surfactants; others are prim arily used as intermediates in the formation o f polymeric or oligomeric products. Some o f these low molecular weight fluorochemicals are also likely intermediates in the degradation o f polymeric compounds. Some can also result from environmental transform ation o f other low molecular weight fluorochemical products. It is im portant to remember that these data were obtained using products that were not highly refined, and products may have more than one fluorochemical component. Some may have nonfluorochemical components that enter into d eterm in atio n o f th e v alu es. B ecau se o f improvements in analytical techniques and product refinement, these data are in the process o f being replaced by better quality data. 13 Table 2. Physical Data on Fluorochemical Products (Developed for Use on MSDS Sheets) Source: MSDS Sheets Abbreviations: N/D: not determined; N/A: not applicable; -.-approximately measured at 1mm Hg measured at 2 mm Hg Product Use Interned. Interned. Interned. Surfactant Interm ed. Interned. Surfactant Surfactant Surfactant surfactant Surfactant Surfactant Surfactant P rin c ip a l Fluorochem ical POSF N-MeFQSE alcohol N-EtFOSE alcohol N-EtFOSA N-EtFOSEA N-EtFOSEMA PFOS NH.T salt PFOS Li salt PFOS K salt PFOS DBA salt PerfluoroCIO sulfonic acid, NH,+salt G lycine derivative o f FOSA N-EtFOSE alcohol, ethylene oxide adduct boiling p t(b ) m elting Pt,(m ) C 154 b 75-95 m ~118 b* -110 b# ~ 90 m -150 b* -150 b* - 82b -100 b N/A -98 b -96 b -100 b 210 b vapor pressure mmHg calc. @20C <10 N/D <10 <10 vapor density cole. @20C A ir= l > 1.0 N/D > 1.0 >1 <10 > 1.0 <10 > 1.0 -3 4 - 1.0 N/D N/A N/A -31 -0.62 -1 6 -1.08 -18 -0.87 -18 0.64 evap rate solubility BuOAc in " 1 w ater Specific pH G ray. W ater" 1 <1.0 N/D < 1.0 N/D < 1.0 <1.0* < 1.0 <1 N/A . < 1.0 <1 < 1.0 <1 neglig neglig neglig neglig - 1.8 -1 .7 -1.7 . - 1.6 N/A N/A N/A N/A nil neglig moderate complete slight complete moderate -1 .5 -1 .5 - 1.1 - 1.1 - 0.6 - 1.1 1.08 N/A N/A -7 6-8 7-8 -7 8.5-9.5 complete -1 .3 -11 apprec 1.31 1.34 5.S-8.4 Additional physical data were developed in the mid-1970s and early 1980s on a few, high volume products. Typically these data are related to developing an understanding o f environmental fate, e.g. data on soil mobility and partitioning coefficient:. They are summarized in Table 3. 3M has evaluated these data for reliability and the reliability codes are included as part o f the table. Progress in analytical techniques has significantly im proved the reliability o f current data compared to the reliability o f these historical data. Current pbysical/chemical data are found in Table 4. 14 ' " ' rffscar" Table 3. Historical Physicai/Chemical Data on Fluorochemical Products Related to Environmental Fate The reliability code w hich follow s the test value in parentheses is Interpreted as follow s: (1A) Study used published test guidelines or well-documented procedures. Where applicable, concentrations were measured. All quality control data were acceptable. (IB ) Results were obtained hy mathmatical estimation. (2) Study meets all the criteria for quality testing, but has one or more deficiencies. A. Concentrations NOT measured-parameter determined via indirect measurement B. Analytical methodology questionable. (3) Study does no t meet criteria for quality testing due to A. Demonstrated weaknesses in experimental procedures. B. Imufficiem methodology description. C. Unacceptable quality control. (4) Study data are available only as summaries. Original reports unavailable. N/D= not determined Product Solubility P rin c ip le in FC water mg/L octanol/water partition coefficient lognoctanol/water partition coefficient soil adsorption coefficient (K) PFOS K+salt N-MeFOSE alcohol N-EtFOSE alcohol N-EtFOSEA POSF N-EtFOSA 1080 (2A) 0.82 (2B) 0.03 (2B) 0.89 (2B) 1 est (2A) N/D 10 (2B) 1 (2B) 56,800 (2B) ND 6,600,000 (4) N/D N/D N/D 3.60 OB) >6 (2B) N/D N/D 0.99 (2B) 77 (2B) 330 (2B) N/D N/D N/D organic carbon adsorption coefficient (K J 66 (2B) . 3,500 (2B) 17,800 (2B) N/D N/D N/D Vapor pressure N/D N/D 1.22.uunH g(ll5 0.5Pa@20C (1A) 6.0 x l 0-3P a (lA ) 1.6 torr@ 20C (4) 0.16Pa@ 20C (IA ) H istorically, formulated products containing other components and residuals rather than pure perfluorochemicals were used to collect physical/chemical data. While m ost o f the products above consist largely o f one active fluorochemical component, the values obtained for the product are not likely those for the purified fluorochemical alone. Computer models used in conjunction with empirical sampling can be used to predict environm ental fate and transport o f these substances. Existing models can require the following physical/chemical data for operation: molecular weight; boiling/m elting point, pKa, octanol/water partition coefficient, vapor pressure, solubility, Henry's law constant density, evaporation rate, heat o f vaporization, bioconcentratioii factor, and degradation mechanisms in air and water (hydrolysis, photolysis, and biodegradation). Precise values for the parent fluorochemical compound, its intermediates, and the end degradation produces) are essential for comprehensive predictions about environmental fate and transport. 15 \ 3M is developing the missing physical/chemical data on individual fliiorochemicals -with the assistance o f several consultants. While laboratory studies are underway on physical/chemical properties o f PFOS, EtFOSE alcohol and MeFOSE alcohol, models are being developed to estimate tire physical/chemical properties o f other sulfonatcd , perfluorochemicals. The data at being determined using the Guidelines for die Testing o f Chemicals developed by the Organization for Economic Co-operation and Development (OECD) for physical/chemical testing (3) where available. W here possible, melting points, boiling points, vapor pressures, dissociation constants, water solubility, noctanol/water partition coefficients, air/water partition coefficients, and soil adsorption/desoiption will be determined. This information is needed for both environmental fate models and manufacturing emission models. Current modeling efforts are hampered by lack o f data on physical/chemical properties. Data are being collected according to Good Laboratory Practice (GLP) standards. The air/water partition test is non-standard This test protocol was developed j ointly by 3M and an outside expert. Results will be reviewed by several technical experts, both w ithin the 3M Environmental Laboratory, and outside the company. 3M is generating the information on soil sorption/desorption characteristics as non-GLP, screening studies. These data w ill aid in the evaluation o f the transport process and partitioning. For example, 1will a fluorochemical be retained by the soil m atrix or rem ain in the w ater phase? The bioconcentration potential o f PFOS and EtFOSE alcohol w ill be examined through empirical testing that determines the extent o f the uptake o f these chem icals by fish. Work on degradation including hydrolysis, photodegradation and biodegradation is described in another section. (See Environmental T ransform ation/D egradation.) The physical/chemical testing is proceeding in order o f PFOS, EtFOSE alcohol, and MeFOSE alcohol. The results to date are reported in Table 4. The inability to determine an octanol/water partition coefficient makes it difficult to do predictive modeling. \- - Table 4. New Physical/Chemical Testing Results on PFOS, potassium salt P aram eter Results Solubility: pure water 570 tng/L Solubility: fresh water 370 mg/L* . Solubility: unfiltered seaw ater ...... ......., 4-5 mg/L* estimated Solubility: filtered sea water 25 mg/L* Vapor Pressure 3.31 x l O 4 i>o @20C Melting Point >400C Boiling Point not calculable Octanol/W ater Partition O C.) not calculable; three phases Air/W ater Partition Coefficient 0(<2 X 10-4) Data developed in support o f other studies; not developed using GLP standards. 16 The methods used in these current physical/chemical tests will provide values reported in consistent formats that are internationally familiar and accepted. This standardization will aid in the review and comparison of data on individual fluorochemicals and in model operation and prediction. The data will contribute to analytical method development and overall improvements in sample handling, shipping and storage as well as mamribehiritug 5.0 Analytical Test Methods for Fluorochemicals i Procedures for detecting and identifying fluorochemicals in the environment require a very high level o f technical expertise. Most general analytical methods do not.providc enough sensitivity or selectivity. The complex mixture o f possible components in a product the m ultiple matrices in which they could reside (e.g. the atmosphere, soils, surface w ater, groundwater, wastewater, different animal tissues, different animal species, plant species, foods, etc.), and trace level detection require selective extraction and diverse analytical techniques. Each fluorochemical requires a unique analytical methodology. Separate methods may be needed for every matrix. Validation of each method is tim e intensive. Often, standards are not available. Reliable quantitative methods for extraction, separation and identification have been developed only within the last few years. Prior to that, relatively in sen sitiv e and n o n -sp ecific analytical m ethods, such as "to ta l organic flu o rid e (T O F )," were used. The analytical technology used in extraction, separation, identification and quantitation includes combinations of: - High Performance Liquid Chromatography (HPLC); - High Pressure Solvent Extraction (HPSE); - ElectroSpray Tandem Mass Spectroscopy (ESMSMS); ' - Gas Chromatography (GC) w ith a Flame Ionization Detector (FID), a Mass Spectrometer (MS), a Photo Ionization Detector (PID), or an Electron Capture Detector (ECD) - HPLC-Quadrapole- Time O f Flight-mass spectrometer (QTOF) For example, analysis o f PFOS extracted from tissues requires ESMSMS analysis. This technique focuses quantitation on three secondary ions o f one primary ion at a specific HPLC retention time. To provide positive identification o f target analytes in complicated matrices, the 3M Environmental Laboratory uses a quadrapole time-of-flight mass spectrometer. The instrument provides high mass accuracy (to 0.0005 amu) and so is useful in identifying 17 ' ' r\-rrcrr fluorochemical metabolites and intermediates for which standards are not available. Compound identification is based on reasonable HPLC retention tim e as compared to standard compounds o f sim ilar structure, reasonable interpretation o f fragment ions associated with the primary ion, interpretation o f the accurate mass spectrum, and agreement between the experimental and theoretical molecular weight (+0.0005 amu). The addition o f new technology has permitted 3M analysts to increase the numbers o f sulfonated perfluorochemicals that can be identified, expand the matrices in which they can be detected, and lower the levels at which they are detected. The technology has expanded the volumes o f analyses that can be done. Nonetheless, capacity limits require analyses to be prioritized. When samples cannot be analyzed soon after collection, care is taken to store the samples appropriately for the matrix and the analytical method, both to prevent sample deterioration and contamination. 3M now has in place several methods for analysis o f sulfonated perfluorochemicals in several m atrices. The methods produce data o f varying quality. They may be used in com bination to produce test data The method performance can be categorized as follows: 1. Quantitative methods that have been validated by studies conducted according to Good Laboratory Practices (GLP). These exist for analyses o f samples o f blood, liver, and several animal tissues o f certain species, drinking water, and certain types o f food. 2. Quantitative methods that typically are based on methodologies that have undergone significant analytical characterization during development. These methods are validated by extensive quality control testing, but validation studies may not have been conducted according to GLP requirements. These exist for wastewater, sludge, and air, for example. 3. Semi-quantitative methods that typically are based on the quantitative methods but for which validation studies arc lacking or quality assurance cannot be demonstrated because, for example, standards are unobtainable or sample matrix is extremely limited. . 4. Screening methods that typically are under development or a result o f exploratory studies. These methods yield only qualitative data, Le. they reliably detect the presence or absence o f an analyte. M ethod development is continuing, not only at the 3M Environmental Laboratory but also at independent laboratories in consultation with 3M Environmental Laboratory scientists. For some matrices, the detection limits sought are at lower levels. Method validation o f low level analyses may be confirmed at a university or other contract laboratories, as appropriate. When samples re sent to consulting laboratories, 3M supplies the methodology or shares expertise to develop the method. Quality assurance is 18 required, along w ith method validation and oversight at levels comparable to those mW in the 3M Environmental Laboratory. Continual improvements are sought in analytical methods as the ability to detect trace quantities is essential for a number o f reasons such as: screening laboratory supplies and environments prior to initiating toxicity testing, for detecting environmental exposure, for determ ining sources o f perfluorochemicals, and for understanding perfluorochemical . metabolism kinetics. 6.0 Sources of Fluorochemicals A few fluorochemicals occur naturally in the biosphere, produced by biological and geochemical processes. Several green plants produce monofluoroaceticUeid (CH2FCOOH). Some fungi produce monofluormated organics. All fluorochemicals produced biologically contain only one fluorine atom. Volcanoes and other geological processes produce tetrafluoroethylene, sulfur hexafluoride, perfluoromethane and some chlorofluorocarbons in small quantities. M ost fluorochemicals in the environment are present as a result o f human manufacture and use. Releases o f fluorochemicals into the environment can occur at each stage o f the fluorochemical product's life cycle. They can he released when the fluorochemical is synthesized, continue during incorporation o f the fluorochemical into a product, during the distribution o f the product to users, during the use o f the product by consumers, and during disposal practices at all o f these stages. 3M is using a two step approach to estimate environmental releases o f fluorochemicals. The initial efforts have focused on determining waste generated; the second step w ill focus on determining releases. This two step approach is necessary since not all waste produced w ill result in a release to the environment. Much o f the waste that is generated is destroyed through treatment or otherwise actively managed to prevent release into title environment. Efforts are also being made to further tighten such controls. 3M has estim ated waste generation from each o f the following life cycle stages: the manufacturing processes, the supply and distribution chains, customer uses and product/waste disposal. ,, - For ease in comparing w aste stream data, wastes are described in term o f "PFOS equivalents." PFOS equivalents are the weight o f CgFX7S 0 2present in a sulfonated perfluorochem ical product. It is the mass o f PFOS molecules that would be formed in.the breakdown o f the product. The assumptions o f complete breakdown to PFOS o f each sulfonated perfluorochemical product, in the year in which the product was sold, are unlikely "worst-case" assumptions. Various degradation testing finds a broad range o f 19 product degradation rates. Some polymeric products appear to be quite stable in the environment, w ith long half-lives; other polymers hydrolyze quickly. 6.1 Manufacturing Waste Streams The assessm ent o f the release o f sulfonated perfluprochemicals into the environment begins w ith manufacturing waste generation. Some waste streams, such as wastewater discharge o r disposal o f off-spec products, can be anticipated and controls provided. Other w aste can be generated during any o f the steps required to produce the fluorochemicals and manufacture the product. The greatest production o f the parent fluorochemical product, POSF, occurs at the Decatur, Alabam a p lan t Here POSF is created in electrochemical cells and undergoes num eroussteps to convert it into final products. Salts o f PFOS are also m anufactured at the facility. Because o f its production volume, the Decatur facility has been the focus o f m anufacturing waste stream studies. Understanding waste generation and how wastes are managed and disposed o f provides a better understanding o f potential releases into the environment. That understanding will help to identify opportunities for reductions in such releases. The manufacturing process for sulfonated perfluorochemicals is complicated. There are more than 600 intermediate manufacturing steps associated w ith the production o f POSF and POSF-based products. This translates into hundreds o f process steps that require venting or that generate wastewater or solid waste. Although the manufacturing process attempts to capture, reuse, and recycle most fluorochemicals as desired product m aterial, until recently, the unique chemistries created in each step o f the process could not be analyzed precisely to confirm composition and to quantify amounte. The manufacturing process is dynamic, with rapidly changing matrices and many process steps. Ongoing process optim ization activities continuously change the waste stream profile. Progress has been made in analytical techniques. In 1997, analytical laboratory , techniques and methods could quantitatively identify the presence o f only one fluorochemical analyte in a wastewater matrix. In 1999, improved analytical techniques and m ethods were developed for additional fluorochemical analytes in a wastewater m atrix. . Advanced field monitoring technology has been developed based on Fourier Transform In fra re d spectroscopy (FU R ). This field tool has been used to detect where emissions to air are occurring during the manufacturing process andto evaluate whether a process change or a control technology can decrease the release. 20 As better analytical techniques become available, efforts are being made to: - characterize the major manufacturing processes generating fluorochemical w aste streams; - evaluate the effectiveness o f fluorochemical removal technologies; and .... - provide better estimates o f the amounts and kinds o f fluorochemicals released to the environment from manufacturing processes and from waste treatm ent and 1 disposal. Inform ation currently available on waste streams generated during manufacturing processes a t Decatur is derived from engineering calculations, air emissions m odeling, and lim ited testing. An overall site materials balance was developed in the mid- 1990's lining the amount o f POSF-based solids initially created in the electrochemical cell and the amount o f POSF contained in final products sold. The difference was an estim ate o f total waste streams generated during processing. The emission factors derived from this balance are used to calculate waste streams from production throughput. They are the basis for the estim ates in Table 5. These estimates derived from the m aterial balance are not precise, as this methodology can produce only rough approximations. The estim ates in Table 5 reflect the most current information available and combine data derived from several sources: information from the mid-90s site balance, wastewater testing, w aste disposal records, process models and supplemental information from 1997, 1998 and 1999. Several changes in waste disposal and processing have been implemented since the mid-1990s in order to reduce potential releases to the environment. Wastewater sludges that were once land applied on site are now sent to a m unicipal landfill for disposal. Off-spec materials that were discharged to w astewater are now shipped off-site to be incinerated. Table 5 helps to demonstrate the vast difference between volumes o f wastes generated and volumes o f releases to the environment, since the vast majority o f wastes sent to incineration are destroyed in the incineration process and m ost material sent off-site to landfills w ill be effectively managed to prevent release to the environment. Table 5. Estimated 1998 Wastes Generated (in PFOS Equivalents) a t the Decatur Manufacturing Plant W aste Type A ir Emissions Wastes sent off-site to Incineration Wastes sent off-site to Landfills Discharge to River after Wastewater Treatment Total Wastes .... Estim ated PFOS Equivalents, lbs 19,000 657,000 380,000 10,000 ' 1,066,000 Note: The 10,000 lbs/yr of PFOS equivalents in the discharge to the river are estimated releases to the environment after wastewater treatment, not the lbs/yr generated prior to treatm ent 21 mntr" More explanation o f the estimates and efforts currently underway in air, wastewater and , w aste management follows. . 6.11 W aste Stream Characterization Updating m aterial balances for the manufacturing process is an ongoing effort. Today process engineers use a model o f process steps to calculate air emissions. New inform ation is being compiled to aid with model operation and waste calculations. The effort to determine physicai/chemical properties for sulfonated perfluorochemicals w ill improve m odel inputs and waste stream calculations. Analytical technology is improving understanding o f process chemistry ' D ata from the process engineers'available material balances in the plant's reporting system have been used to supplement the earlier site balance in estimating air em issions. in itia l reports from this system indicate that most site waste and air emissions result from fewer than 10 key steps in the early stages o f POSF production. Process experts are examining these steps for ways to reduce or eliminate the impurities and wastes generated in the steps. ' In 1999, the Decatur plant installed a discotherm unit which heats the process m aterials, vaporizing and capturing the fluorochemicals. It will significantly reduce the organofluorides in the wastewater. This technology will operate to reduce emissions and waste at the source. It will make it easier to segregate waste streams and recycle fluorochemical wastes back into the process. 6.12 A ir V 3M engineers have reviewed specific process steps to determine what air emissions testing is feasible and appropriate. Testing o f complex batch-processing systems is difficult due to quickly changing process conditions, venting pressures, and difficulty in isolating processes; however, characterization testing may be possible. The technical feasibility o f performing this testing for two major processes is now under evaluation. Any emissions testing w ill require modifications to process vents and m itigation o f potential safety hazards. About 80 separate venting points are associated w ith the equipment used to make sulfonated perfluoroqhemicals. 6.13 W astew ater Analytical methods have been developed during the past year to better characterize the wastewater discharge from the site. The first testing o f wastewater before and after treatm ent for specific fluorochemicals occurred at Decatur early in 1998. The testing was 22 lim ited and reflected operating conditions for a relatively short period of time (24 hour composite samples o f influent and effluent for one w eek) Some o f the compounds that were identified in the wastewater were: a diester o f EtFOSE alcohol, EtFOSE alcohol, MeFOSE alcohol, PFOS, FOSA, PFOSAA, PFOA and PFHS. ^ In 1998 an interim carbon adsorption treatment system was installed as part o f wastewater treatm ent D ata for the effluent estimate in Table 5 reflects this change. This treatm ent system treats the largest single source o f fluorochemical-containing wastewater in order to remove PFOS ami other shlfonated perfluorochemicals from the wastewater. Comparison o f the results from sampling done in February 1998 with sampling done in the end o f 1998 indicates the quantity o f PFOS discharged to the Tennessee River declined by about half. Tn addition to the carbon ad so rp tio n system, in-process operational changes were made in off-spec product discharge procedures that also contributed to the reduction in PFOS content o f the discharge to the river. The carbon system has been incorporated as a permanentupgrade o f the wastewater treatm ent system. M onitoring indicates that with proper operation, carbon adsorption removes better than 99% o f PFOS. Removal efficiency o f other sulfonated perfluorochemicals varies, but the treatment appears to provide a high degree o f removal for most. A number o f wastewater streams currently going to sewers are in the process o f being diverted to thermal treatment facilities for disposal. This w ill result in a reduction in the values listed in Table 5. 3M has conducted an extensive review o f state-of-the-art technology for wastewater treatm ent Various upgrades are currently being evaluated. The long term goal o f wastewater treatment at the plant is to utilize source control and end-of-pipe treatm ent to remove nearly all sulfonated perfluorochemicals from wastewater prior to discharge to the river. 6.14 Solid W aste An effort to identify all waste streams and their disposal methods is underway. Existing waste tracking is done on a site basis, so it is difficult to distinguish the particular stream s w ith POSF chemistry. The mid-1990s emission estimates did not distinguish final disposal o f the material lost from production, so site records were used in combination w ith the existing emission estimates to create the current picture o f potential releases resulting from disposal. ' A review o f plant records for 1998 has been completed to determine primary waste disposal locations for the site. According to Decatur plant records, 63% o f the fluorochemicai containing wastes are sent to incinerators, 33% of the wastes are disposed in hazardous waste landfills and 4% in non-hazardous waste landfills. 23 6.2 Supply Chain Waal Streams Using sales data, 3M identified key products that contain a majority o f the fluorochemical solids used in products. These products represent 89% OfPFOS-equlvalents Sold by 3M in 1997 in the United States. Most commonly, these products were sold to commercial users who applied them or incorporated them into their products. Using the inform ation developed from sales, 3M estimated customer and end user w aste streams (Table 6). th e se estimates are imprecise and based on several assumptions, but provide qualitative information. Using the chemical formula for PFOS, the fluorochemical solids were converted to "PFOS equivalents" for ease in estimating and comparing total losses o f sulfonated peiflunrochemicals and in comparing losses. The assumptions o f complete breakdown to PFOS of each sulfonated perfluorochemical product, in the year in which the product was sold, are unlikely "worst case" assumptions. Product w aste stream estimates are based on conservative, worst case assumptions about the generation o f waste streams at supply chain facilities. These are often based on 4 operator experience or engineering estimates rather than laboratory tests and can result in wide ranges in waste stream calculations. In estimating wastes, these data do not include loss o f product residuals in the waste streams because information on the properties o f residuals and processes at supply chain facilities and end user locations is inadequate to estimate this loss. Initial estimates associate waste streams generated from uses and disposal o f the products by customers of each business unit. These estimates are helping to focus efforts in improving customer stewardship practices and 3M product reengineering. As is evident, most o f the waste generated is in the form o f solid waste. Table 6. Customer and End User Waste Stream Estimates, PFOS equivalents, lbs in 1997 W aste Stream A ir W astewater . Solid W aste Supply C hain Use 2,600 112,000 . 59,000 3,300 181,000 . . 377,000 D isposal 0 0 1,262,000 6.3 Releases from Waste Treatment and Disposal Methods 3 M and it3 c o n su lta n t are g ath erin g in fo rm atio n o n treatm en t and waste handling at several landfills and wastewater treatment plants which receive wastes containing sulfonated perfluorochemicals from the supply chain facilities and 3M manufecturing 24 1 ... facilities. Inform ation is also being compiled on some o f the largest wastewater treatm ent facilities and landfills in die United States in order to Estimate the potential perfluorochemical releases to the environment from municipal disposal facilities not associated with the supply chain or manufacturing. Incineration is a favored disposal method because o f its high rates o f destruction o f sulfonated compounds. 3M and its consultant arc fu rth er evaluating the effectiveness o f incineration for this purpose, th e basic bond breaking chemistry o f thermal destruction o f POSF-based fluorochemicals, the destruction efficiencies o f various technologies/situations such as municipal incinerators, and the products that could result from incomplete combustion are elements o f the study. The study involves a review o f 3M and external literature to compile information on the formation and properties o f therm al transform ation products o f sulfonatedperfluorochemicals. M odeling w ill be used to determine to the extent practical, the releases to the environm ent from the amount of material sent to incineration, wastewater treatm ent plants, and landfills. The goals o f the life cycle release studies are: ' - to identify important fluorochemicals based on volume o f release, mode of release and chemistry; - to provide values for use in modeling the distribution o f fluorochemicals in the environm ent; . to determine sampling sites and substantiate sampling results; - to predict which fluorochemical releases may result in exposure to humans and tire environment; and - to identify fluorochemicals that require further study as to their transport, fate and exposure potential. 7.0 Environm ental Transport and Distribution The transport and fate o f chemicals in the environment depends on many factors but principally on the interaction between environmental conditions (e.g. water, temperature, sunlight), and chemical properties (e.g. partitioning and reactivity). In the environmental area, eleven im portant fate and transport mechanisms for sulfonated perfluorochemicals have been identified for further study. These are: 1. Partitioning between air and product, i.e. volatilization from product to air; 2. Indoor air deposition; 1234 - 3. Accumulation on airborne particulates; 4. Fate and transport to the stratosphere; 5. Accumulation at the surface water microlayer; 6. Degradation (includes hydrolysis, photolysis and biodegradation); 7. D issociation in water; 8. Uptake in plants; 9. Uptake in fish; 10. Uptake in birds; 11. Efficiency o f wastewater treatment systems. .. ._ .. _ A ll o f these fate and transport mechanisms have been linked to models. M odeling uses mathematical equations to simulate and predict real events and processes. Many types o f models w ill be considered for use in this effort to evaluate sulfonated perfluorochemicals. Simple models o f ecosystems, indoor air, and treatment systems (wastewater, landfills) ' are being used to screen for possible fate mechanisms, possible exposures, and possible sample detection lim its. For example, one preliminary screening model suggests that top trophic level species such as fish eating birds and sea mammals should be examined. This finding was incorporated into the design o f the biosphere sampling plan. Chemicals differ greatly in their behavior. The major d ifferen ces in behavior o f organic chem icals in the environment are due to physical-chemical properties. Although laboratory studies are underway on physical/chemical properties o f PFOS, EtFOSE alcohol and MeFOSE alcohol, models are being developed to estimate the physical/chemical properties o f other sulfonated perfluorochemicals. This w ill reduce the timR and testing required to gather these data for use in environmental fate models. Fugacity is a concept that is used to describe the tendency o f a compound to m igrate in and between one environmental medium and another. Different media include air, water, soil, sediment, and biota, all o f which together compose a dynamic, interactive systemTM an ecosystem. Predictions about movement o f a chemical m ust incorporate both its physical/chemical properties and the environment the chemical is in. For example, a low vapor pressure does not mean a chemical is not present in air. It may evaporate appreciably from water despite a low vapor pressure if it has low solubility in water. By entering the physical-chemical property data on a chemical into a fugacity model o f a generic or specific environment, it is possible to estimate general features o f a chem ical's likely behavior and fete. The output o f these calculations can be presented numerically and pictorially. (6) Fugacity models will be used to predict fate and transport o f sulfonated perfluorochemicals. Existing fugacity models typically are based on experience w ith chlorinated organics. An internationally recognized modeling expert is developing/ adapting models to consider the unique properties of fluorochemicals. The goal o f this m odeling effort is to have a multimedia model or models to predict the fate o f sulfonated perfluorochemical products and associated byproducts in a variety o f ecosystems. 26 8.0 Environmental Sampling for Fluorochemicals 8.1 Environmental Levels 8.11 H istorical Data In the late 1970s, 3M conducted a very limited number o f studies to assess the distribution o f fluorochemical constituents in the environm ent Several freshwater fish species were tested for a number o f fluorochemical compounds. In reviewing the data obtained from these studies in context of the current knowledge o f the behavior o f these m aterials, 3M has concluded "that these historical data arc highly questionable and may be m isleading. Therefore, they are not included in this paper. The sections following present m ore reliable data and information collected using validated sampling and analytical methodologies. 8.12 Recent Analyses of W ild Birds and Fish In analysis in 1999 o f the plasma o f ten fish eating birds, albatross nestlings at M idway Island in the Pacific Ocean and eagle nestlings in Minnesota and M ichigan, PFOS was detected in each o f the samples from eagles. The samples were collected in 1989, 92, and 93 by Dr. John Giesy o f Michigan State University as part o f other surveys. Three o f the albatross adults showed no detectable levels o f PFOS (< 1 ppb detection level). D etectable, but not quantifiable levels o f PFOS were found in the remaining albatross samples, both collected from birds less than a year old. All albatross samples were collected in 1992-93. See Table 7. These data are semi-quantitative, screening quality. As only a small amount (< lm L) o fplasma was available to conduct the analyses, no m atrix spikes were possible to estimate the method's recovery efficiency, but the methods used have been characterized in other, similar matrices. A fter the initial, screening results on wild bird plasma, the plasma from a second set o f w ild birds was examined for the presence o f PFOS. (See Table 7.) The source o f th e, plasm a was three sea eagles collected from the Baltic Sea and seven bald eagles collected from N orth America. The samples were collected in 1992-93 and again by Dr. John Giesy. PFOS was detected in all o f the eagle plasma screened. These data are semi q u a n tita tiv e , screen in g q u ality . Two matrix sp ik es (250 ppb) prepared from eagle plasm a were extracted and analyzed. Both showed >80% recovery. 27 ( Table 7. Levels of PFOS in the Plasma of W ild Birds Species Bald Eagle Bald Eagle Bald Eagle Bald Eagle Bald Eagle A lbatross A lbatross Albatross A lbatross Albatross Sea Eagle Sea Eagle . Sea Eagle Bald Eagle Bald Eagle Bald Eagle Bald Eagle Bald Eagle Bald Eagle Bald Eagle Collection D ate 5 Jun 93 3 Jun 93 1989 1989 17 Jun 92 13 Dec 92 18 May 93 13 Dec 92 13 Dec 92 18 Iday 93 28 May 93 27 May 93 23 May 93 26 Jun 92 28 Jun 93 23 Jun 92 5 Jun 92 26 Jun 92 22 Jun 92 8 Jun 92 Location Lower Penn, MI Lower Penn,MI Upper Penn, MI Upper Penn, MI Voyageurs,MN . Midway atoll Midway atoll Midway atoll Midway atoll Midway atoll , Baltic, Sweden Baltic, Sweden Baltic, Sweden North America North America L. Superior ONT North America Devil's Is., WI M udC reek,O H Carroll Twp, OH ' Age, Gender 163 days, F 228, F unknown unknown 82 days, M 6 years 0 8 years 15 years 0 nestling nestling nestling nestling nestling nestling, F Adult, F nestling, F nestling, nestling ' PFOS, ppb 30 34 77 31 34 BUD BLQ BLD BLD BLQ 125 93 215 165 198 494 1047 226 371 374 ... BLQ= Below L im it o f Q uantitation (10 ppb) BLD= Below Limit o f Detection (approximately 1 ppb) Following the bird plasma studies, sixty liver samples collected by the U.S. Fish & W ildlife Service from various species of birds were analyzed. The dead birds were collected at a variety o f sites across the United States. They were not part o f a controlled research study, but were selected for their location and d iet All but sandhill cranes are fish eating species. The sandhill cranes are an insect eating species. The purpose o f the analyses was to determine if the presence'of PFOS could be detected in these sample matrices. 3M believes that these sets of data are insufficient to draw conclusions w ith any statistical merit. The PFOS data in Table 8 are semi-quantitative, screening quality, w ith a margin o f error estimated at + 30%. The lim it o f quantitation for PFOS is 6 ppb. 28 Table 8. Analysis of Wild Bird Livers BLQ=* Below lim it o f quantitation <6 ppb) Sam ple No. Spedes Location 1 2' 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 . 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Sandhill Crane Kearney, NE Sandhill Crane Kearney, NE Sandhill Crane Kearney, NE Sandhill Crane Kearney, NE Sandhill Crane , . Kearney,NE Sandhill Crane Chochise Co., AZ Sandhill Crane Chochise Co., AZ Sandhill Crane Chochise Co., AZ Sandhill Crane Chochise Co., AZ Sandhill Crane Chochise Co., AZ White Pelican Calipatria, CA W hitt Pelican Calipatria, CA White Pelican Calipatria, CA Whit Pelican Calipatria, CA White Pelican Calipatria, CA Brandt's Cormorant San Diego, CA Brandt's Cormorant >San Diego, CA Brandt's Cormorant San Diego, CA Brandt's Cormorant San Diego, CA Brandt's Cormorant San Diego. CA DbL Crested Cormorant St. Martinville, LA Dbl. Crested Cormorant S t Martinville, LA Dbl. Crested Cormorant S t Martinville, LA Dbl. Crested Cormorant St. Martinville, LA Dbl. Crested Cormorant St. Martinville, LA Brown Pelican Miami, FL Brown Pelican Miami, FL Brown Pelican Miami, FL Brown Pelican Miami, FL Brow n Pelican M iami, FL Sandhill Crane Valenica Go., NM Sandhill Crane Valenica Co., NM Sandhill Crane Socorro Co., NM Sandhill Crane Socorro Co., NM Sandhill Crane Valenica Co., NM Dbl. Crested Cormorant Naples. FL Dbl. Crested Cormorant Naples, FL Dbl. Crested Cormorant Naples, FL Dbl. Crested Cormorant Naples, FL Dbl. Crested Cormorant Naples, FL Brown Pelican Calipatria, CA Brown Pelican Calipatria, CA Brown Pelican Calipatria, CA Brown Pelican Calipatria, CA Brown Pelican Calipatria, CA 29 PFOSppb 41 BLQ BLQ BLQ BLQ BLQ BLQ BLQ BLQ . BLQ 35 1293 29 15 153 53 46 46 80 2055 59 145 333 76 170 106 134 125 159 48 BLQ BLQ BLQ BLQ BLQ .212 10 52 100 152 16 36 "T O 1 6 32 < Sam ple No. Species 46 Great Blue Heron 47 Great Blue Heron 48 Great Blue Heron 49 Great Blue Heron 50 Great Blue Heron 51 W hite Pelican 52 W hite Pelican 53 White Pelican 54 White Pelican 55 White Pelican 56 Brown Pelican 57 Brown Pelican 58 Brown Pelican 59 Brown Pelican 60 Brown Pelican Location ist. Maitinville, LA S t Maitinville, L StM airtmviile,LA St. Martinville, LA S t Martinville, LA . Fallon, NV Fallon, NV Fallon, NV Fallon, NV Fallon, NV F t Lauderdale, FL F t Laiuderdale. FL F t Lauderdale, FL F t Lauderdale, FL F t Lauderdale, FL PFOS ppb 188 59 "M I 261 173 141 . 362 927 i~133 291 194 75 71 31 91 " . In addition to w ild birds, some fish from the wild were tested for the presence o f PFOS. The fish were collected in 1997-98 from sites in Michigan as part o f surveys conducted by Dr. John Giesy. They were stored frozen and analyzed in 1999. Six species were tested. Low levels o f PFOS were detected in four o f the twelve samples. Since no sample matrices were available for matrix spike studies, these data are o f screening qualify only. No clear meaning can be drawn from the data. They are being used to develop sampling programs. Table 9 reports the findings. Table 9. PFOS Screening in Fish. BLD= Below Limit o f Detection (approximately 7ppb) BLQ=Below Limit o f Quantitation (approximately 70 ppb) Sam ple No. 1 2 3 4 5 6 7 8 9 10 11 12 Species Carp Lake Trout Lake Trout Lake Trout Lake Trout W alleye Ciscowet Brown Trout Brown Trout Channel Cattish Channel C attish Channel Cattish L o catio n Pinc River, Ml Siskiwit Lake, Isle Royale, MI Siskiwit Lake, Isle Royale, MI Pine River, MI Lake Superior Detroit River, MI Lake Superior, Marquette, MI Detroit River, MI Rouge River, MI Lake S t Claire, MI Lake S t Claire, MT Lake S t Claire, MI Test M atrix whole body whole body whole body whole body whole body whole body muscle muscle liver . muscle egg egg T est Result BLD BLD BLD BLQ BLD BLD BLD BLD BLQ BLD BLQ BLQ 30 r m tr 8.13 T estin gof Fishm eal Used in Rat Studies W hile performing human health toxicity studies (see Perfluorooctane Sulfonate: Current Summary o f Human Sera, Health and Toxicology Data. January 1999), 3M found, "endogenous" levels o f PFOS in some o f the naive rats used in the studies. The levels found in the rat livers ranged from 29 ppb to 300 ppb. Livers o f rats from one supplier showed no PFOS above the detection limit o f 15 ppb. Further investigadon revealed fishmeal to be an ingredient in the rat chow fed to the rats in which PFOS was detected. Fishmeal was not a, dietary component o fthe rats that had no detectable levels o f PFOS. 3M developed a complex analytical method to analyze fishmeal samples collected from different fish stock. A t a detection lim it o f 2 ppm, PFOS was detected in three sam ples o f fishmeal and not detected in three samples. A t this time, these data are not conclusive. 8.14 Plant Site Analyses In March o f 1998,3M conducted screening level sampling for PFOS around the D ecatur plant. The outfall o f the Decatur wastewater treatment plant is at a bay near the m outh o f Baker's Creek. Baker's Creek flows into the Tennessee River, a large river that supports barge traffic. About 25 miles downstream is Wheeler Dam. The samples tested were o f w ater surface film, subsurface water and sediment. A goal o f the sampling was to experim ent w ith sampling techniques and analytical methods. Therefore, the analytical data are o f screening quality only. Data on PFOS from the sampling'are in Table 10. Table 10. Sampling Near the Decatur W astewater Discharge Sam ple Locations: . UP1 & UP2: Tennessee River, upstream of discharge BC1: Baker's Creek below outfall Q 1 & Q2: Baker's Creek, downstream of discharge, in quiet waters near Tennessee River WD1 & WD2: Tennessee River below Wheeler Dam UP1 UP2 n e t Q t Q2 Sub-surface water, in ppm PFOS <.010 <010 0.44 0.025 0.012 PFOS homologues <.010 <010 0.10 <010 <010 Siafacefilm s, in ppm PFOS ' N/C N/C 1.60 1.00 0.28 PFOS homologues N/C .. N/C . 0.02 <010 <010 Sediment.in ppm PFOS 0.177 <050 31.1 N/C N/C PFOS homologues <050 <050 <050 N/C N/C WM <010 <010 N/C N/C <050 <050 WD2 <010 <010 N /C N /C <050 <050 N/C = not collected . Surface film samples were skimmed from the top o f the water, at the air/water interface. Sedim ent sam ples w ere collected from the river bed using an Ekman Dredge. Sam ples w ere token atth e w ater collection point or, if sediment was lacking there, as close as possible to it.' 31 Based on this initial sampling, a more extensive sampling was conducted. Sampling locations extended from about 10 miles upstream o f die facility to 25 miles below the facility. A s a result o f analytical techniques being developed to lower detection lim its, analyses o f these samples is pending. . 8.15 Biosphere Sampling 3M is building on recent information with advances in technology to design a program that could detect traces o f sulfonaled perfluorochemicals across a range o f species, environmental habitats and geographic locations, including soil, water and organisms. 3M's approach is to use existing, scientifically recognized, sampling and data collection programs in order to minimize the time needed to obtain information. The goal is to set some bounds on the geographic regions where sulfonated perfluorochemicals are currently found, identify areas that should receive more investigation, and eliminate some general environments from further sampling in the immediate future. Key ecosystems and species o f concern surrounding manufacturing plants are being tested as well as ecosystems remote from manufacturing and use locations. W here possible, synoptic samples o f soil, sediment, air or water are also being taken, but the prim ary focus o f initial studies is tissue samples from biological receptors, especially those in upper trophic levels. The information obtained in the initial studies w ill be used to determine appropriate studies for ascertaining critical pathways. , 8.2 Human Exposure Levels Studies to investigate human exposures take several approaches: 1. Environmental exposure o f the general U.S. population will be assessed i n ' phases through a "Multi-Cities Study." This involves field investigation o f . paired cities, one with significant manufacturing or commercial fluorochemical use, matched with a city without known significant use. The study will involve direct sampling for dietary and environmental presence. 2. Residential exposure will be assessed through a product's use and controlled measurements o f the product's releases. This study will measure releases o f fluorochemical residuals and total PFOS from carpets. 3. The migration o f sulfonated perfluorochemicals used in food packaging to the food is being quantified for several foods. 32 rii 8.21 M ulti-cities Sampling The m ulti-cities study pairs a city having significant manufacturing or commercial use o f fluoroohemical products based on customer sales with a city that does not. Initially six cities, (three pairs) are being examined. This may be expanded, depending on initial results. The multi-cities sampling w ill yield environmental distribution data as well as data o n potential sources o f human exposure. The cities were selected to represent urban locations w ith various levels of fluorochemical releases and various types o f m unicipal w ater supplies. The samples to be obtained, where possible, are: urban air, surface w ater column and surface microlayer, sediment, river fish, drinking w ater intake, treated drinking water, tap water, the influent and effluent to publicly owned waste treatm ent works, sludge, and municipal landfill leachate. Additionally a "market basket" o f several food products will-be sampled. These include: beef, pork, chicken, hot dogs, catfish, eggs, m ilk, bread, green beans, apples from three grocery stores and, if possible, produce from local fanners' markets. 8.22 C arp et Use Studies ' The carpet study w ill estimate any loss o f fluorochemical from normal use o f carpets. I f a pilot study o f carpets finds significant releases, then the study w ill assess human exposure that may occur via inhalation, dermal and ingestion routes. 8.23 Paper and Packaging Studies Results o f past studies on the migration o f fluorochemicals from packaging into food have been submitted to the FDA, and FDA has cleared the use o f paper and packaging protectors for food as indirect food additives. Current work focuses on the development o f new methodologies to extract various fluorochemicals from paper and several foods, then perform quantitative, low level analyses (< 1 ppb). 8.24 Exposure Scenarios These scenarios w ill be developed using data from release, fate and distribution studies. Their purpose is to prioritize exposure pathways for fbrther study by developing quantitative estimates o f specific exposures under known conditions in a specific location. ... 33 9.0 Environmental Transformation/Degradation of F itio ro ch em icals There are many physical, chemical and biological mechanisms that operate in the environment to transform or degrade molecules. They include abiotic mechanisms, e.g. hydrolysis and photolysis, and biotic mechanisms, especially microbial metabolism. Because the carbon-fluorine bond is one o f the strongest in nature, with high bond energies, its cleavage requires large amounts o f energy. M ost chemical and physical processes naturally occurring in the biosphere lack the required energy. In the laboratory, perfluoroalkyl chains are not degraded in the chemical oxygen demand (COD) test, nor in total organic carbon (TOC) analyzers that use very reactive chemical and ultraviolet degradation mechanisms. Combustion docs destroy organic fluorochcmicals nrl degradation is found in high temperature TOC analyzers. In perfluorinated molecules, the fluorines surround the carbon chain completely, shielding the carbon-carbon bonds from attack. The fluorine atoms confer a "rigidity" to the conformation o f the molecule. This rigidity could make it difficult for the molecule to jo in w ith enzymes, thereby blocking biological attack o f the carbon-carbon bond. As a molecule becomes m ore fluorinated, carbon-carbon bonds, carbon-hydrogen and carbon fluorine bonds all typically increase in strength. Early work w ith perfluorochemical products using standardized screening tests for degradation found little susceptibility to degradation. (See Table 11.) Fluorochcmicals lacking nonfluorinated organic portions produced essentially no biochemical oxygen demand (BOD). Those with ionically bonded organics showed BODs near those expected from their non-fluorinated portion alone. Fluorochemical surfactants w ith ' covalently bonded organic portions produced mixed results. The early data on these degradability studies has been given a reliability code that follows the test results. '" - r r ir 1 34 Table 11. Historical Results o f Standard Degradation Tests on Fluofochami&Us if P roduct Principle Fluorochem ical POSF N-MeFOSE alcohol N -EtFO SE alcohol 500-720 (4) 163,000 (1) 260,000 (4) BOD 5-day mg/Kg nil (4) N/D nil (4) BOD 10-day mg/Kg nil (4) N/D N/D , N-EtFOSA N-EtFOSEA N-EtFOSEMA Perfluoro CIO, sulfonic acid, NH4 salt K salt o f carboxylic acid analogue o fNEtFOSE alcohol PFOS Li* salt (4) PFGS K salt 1,800(4) 240,000(1) 80,000 (1) 1,000,000 (2A) 462.000 (1) 54,000 4,000 (4) nil (4) 12,000 (2A) 800 (2A) 67,000 (2A) nil (4) 19,000 (2A) 2,000 (?A) 600,000 (2A) <39,800 172.000 (2A) (2A) N/D N/D nil nil (4) (4) PFOS DEA salt N-EtFOSE alcohol (CiH40 ) 1+H adduct PFOS NH, salt 78,000 (2A) 1,070,000 (1) 44,000 (2A) 0 (4) N/D 412,000(4) 232,500 267,500 (4) (4) BOD 20-day mg/Kg n il(4 ) n il(l) N /D BOD 28-day mg/Kg N/D N/D N /D Photo degra dation N/D n il(l) nil (1) nil (4) 23,000 (2A) , 11,000 (2A) 720,000 (2A) N/D N/D N/D N /D N/D N/D N/D N/D 179,000 289,000 N/D (2A) (2A) - N/D N/D N/D ail N/D nil . (4) (4) 82,000 <2A) 107,000 (4) N/D N/D 292,500 N/D (4) N/D N/D N/D O ther N/D N/D 0 2 uptake" 3% o f ThOD; No deg in 6 m onth shake flask studies or 7 day activated sludge studies (2B) N/D not readily (1) biodegradable N/D ' N /D N /D - N/D no degradation in Warburg 3 hr study or 2.5 month shake flask study (2B) N /D 40% removal BiAS (3A) N/D GOD means Chemical Oxygen Demand. It is a measure o f the oxygen equivalent o f the organic matter content o f a sample that is susceptible to oxidation by a strong chemical oxidant such as potassium dichrom ate. BOD means Biochemical Oxygen Demand. It is the amount o f oxygen consumed by microbial processes while breaking down a known amount o f a test substance. . 35 ThOD m eans Theoretical Oxygen Demand. It is die theoretical quantity o f oxygen used when die test compound is fully mineralized. This value is calculated using the structure o fthe test chemical. BiAS means Bismuth Active Substances. These are materials, such as water soluble polyethoxylates, that precipitate with barium teuaiutlobismuihate. ' N/D means N ot Determined. Code meanings are: (1) Study used published test guidelines or well-documented procedures. Concentrations were measured, and all quality control data were acceptable. (2) Study meets all the criteria for quality testing but has a deficiency A. Concentrations NOT measured. B. Analytical methodology questionable. (3) Study does NOT meet criteria for quality testing; data have one or more flaws. A . Demonstrated weakness in experimental procedures. B . Insufficient description o fmethod. C. Unacceptable performance o f controls. (4) Data are available only as summaries; original reports not found. 9.1 Hydrolysis Studios H ydrolysis is a m ajor mechanism contributing to abiotic degradation o f organic molecules, although it rarely is responsible for complete degradation. The hydrolysis o f sulfonated compounds is described below. O R -- SII -- X +, H20 II O -- OH -f~ HX X = halogen, OR, NR 3M is evaluating the potential for hydrolysis o f fluorochemicals using EPA guidance [Fate, Transport and Transformation Test Guidelines, Hydrolysis as a Function o fpH and Temperature ](2), and is conducting pH dependent studiesof PFOS and MeFOSE alcohol, as well as on fluorochemical monomers, to estimate half-lives. Selected fluorochem ical products are being subjected to a single temperature (50C), variable pH screening process. For those that demonstrate hydrolysis or a deviation form first order kinetics, m ultiple pH, multiple temperature studies are planned. Hydrolysis test data are being reviewed by an outside expert. , 36 9.2 Photolysis Like hydrolysis, photodegradation is a major abiotic mechanism contributing to the transform ation o f organic molecules, but rarely responsible for complete degradation. Photodegradation occurs primarily in air, in shallow water, on soil and vegetative surfaces. It is likely an important factor in foe fete of soluble and volatile compounds, less so for insoluble and sorbed compounds. Products and intermediates m ost susceptible to photodegradation are those most likely to be used outdoors in sunlight Initially, foe degradation that might occur in products dissolved or suspended in w ater is under investigation. The first studies are using foe PFOS precursors such as EtFOSE alcohbl, MeFOSE alcohol, MeFOSA, EtFOSA, and FOSA. Later studies w ill use POSFbased polymers. I f simple methods can be found, gas phase photolysis o f volatile and sem i-volatile fiuorochemical degradation intermediates w ill be investigated. The preliminary results suggest that PFOS is unchanged as a result o f light exposure. However, EtFOSE alcohol, MeFOSE alcohol, EtFOSA and MeFOSA as well as a surfactant and foamer product all appeared to undergo photolysis to FOSA, PFOA, a hydride, and olefins. PFOS was not detected. One product, an aromatic perfluorooctane sulfonate, did photodegrade to form PFOS. 9.3 Atmospheric Studios Although PFOS has a low volatility, several PFOS precursors are volatile. These include: EtFOSE alcohol, MeFOSE alcohol, MeFOSA, EtFOSA, and FOSA. When present as residuals in products, these precursors could evaporate into foe atmosphere when the product is sprayed and then dried. Once in foe atmosphere, foe compounds can rem ain in the gas phase, condense on particulates present in foe atmosphere and be carried or settle out w ith them, or be washed out with rain. The measured vapor pressure o f Et-FOSE alcohol is sufficiently high that essentially all o f it is likely to be in the gas phase and not condensed on particulate matter. Gas chromatic data suggest that other precursors are even m ore volatile. The low water solubility o f these compounds makes it unlikely they washout from foe atmosphere in rainwater. Thus the rate o f removal o f these precursors from the atmosphere w ill likely depend on their photochemical, reactivity, e.g. their reaction w ith hydroxyl ions in the atmosphere. How widely distributed they are locally, regionally or globally depends on foe rate o f photochemical transformation to more soluble or less volatile products. 3M is examining atmospheric lifetimes o f these PFOS precursors. Initially EtFOSE alcohol and MeFOSE alcohol will be tested for reactivity with the OH radical in the gas phase. M odeling w ill determine their atmospheric lifetimes and analytical work w ill " determine their gas-phase degradation products. Then those properties o f foe degradation 37 products that affect removal rates from the atmosphere, e.g. solubility and vapor pressure, w ill be determined. This information will be used to predict distribution o f these compounds resulting from atmospheric mechanisms. 9.4 Biodegradation Studies Biodegradation is essential to the functioning o f living systems. Natural systems rely on living organisms, especially microbes, to breakdown complex organic molecules to simple inorganic molecules that can be recycled back into die ecosystem. Some microbial communities have demonstrated the ability to degrade some xenobiotic compounds. During biologically catalyzed degradation o f these compounds, die degradation intermediates produced are frequently o f a molecular structure that naturally occurs. Particularly important environments for biological breakdown are: sewage treatm ent systems, soils/sediments, estuaries and wetlands. Both aerobic and anaerobic organisms play important roles in degradation. 3M is studying biodegradation using several approaches. 9.41 M icrobial Studies on Perfiuorochem icals W ork at M ichigan State University by Blake Key (6,7) under the direction o f Dr. Craig Criddle used a laboratory isolate of a bacterium, a Pseudomonas species, to investigate the potential for biodegradation of fluorinated sulfonates. The researchers used model fluorinated sulfonate compounds: difluoromethane sulfonate (DFMS), trifluoromethane sulfonate (TFM S), 2,2,2-trifliiornethanesulfonate (TES), PFOS and H-PFOS (1H, 1H,2H,2H-perluorooctane sulfonate). Criddle et al. demonstrated that the microorganism degraded those fluorochcmical compounds containing hydrogen and used them as sulfur sources for growth under sulfurlim iting, aerobic conditions. They later found that such degradation occurred in soil even when s u ite was not limiting. The organism completely defluorinated DFMS. It used DFMS as the sole source o f sulfur, but.not as a source o f carbon or energy. TES and HPFOS were partially defluorinated. Six volatile products were detected for H-PFOS, all containing oxygen and fluorine but not sulfur. Where the carbons were fully fluorinated, i.e. TFMS and PFOS, no degradation was found. Criddle et al; concluded that the transform ation o f fluorinated sulfonates required the presence o f hydrogen at the alpha carbon on the fluorinated alkyl chain. They theorized that when hydrogen is present at the alpha carbon, a site for attack is provided and the carbon-sulfur bond becomes more accessible. Perfluorinated compounds have a rigidity conferred by the fluorine substitution and no structures that are susceptible to electrophilic or nucleophilic attack. 38 9.42 Biological Transformation When perfluorinated organic molecules do biodegrade, it is not the fluorinated portion that is affected. Enzymes attack at non-fluorinated side chains. Rather than complete degradation, i.e. degradation to inorganic compounds, another fluorinated molecule results from biodegradation processes. Existing studies o fmetabolism appear to indicate that for POSF-based compounds; the biological degradation halts when PFOS is formed. C*F,,SOa-F. + HaO ------ -- * C JM C V POSF PFOS C8F 17S 0 2-R. ----------------------- ----------- CaF 17S 0 3` POSF derivative PFOS Once formed, PFOS has not been shown to degrade any further under any natural conditions except combustion. Because PFOS is resistant to physical, chemical and biological degradation, it persists in the environment, but the mechanism o f accumulation is under study. 9.43 Optim izing Conditions for Biodegradation Past studies on fluorochemicals with hydrocarbon portions have demonstrated resistance to biodegradation under standard test conditions, i.e. aerobic microbial degradation using a wastewater inoculum. These studies did not examine all combinations o f conditions that could be optimized to favor the degradation o f partially fluorinated chemicals. 3M is conducting new screening studies for biodegradation. These w ill determine if aerobic and/or anaerobic degradation o f key fluoirochemicals occurs using activated sludge, anaerobic sludge, aquatic sediments and soil; If degradation occurs, the studies w ill determine to w hat extent it occurs and the nature o f degradation products. It w ill also provide information on the degree o f fluorochemical sorption onto microbial sludges and toxicity to microbes. New studies are being designed to promote degradation. They w ill use enriched environments that support biodegradation, e.g. sewage, soil, sediments, and cultures o f microbes selected for biodegradation capabilities. 39 10.0 Ecotoxicity Testing of Fluorochemicals Ebotoxicology is the extension o f toxicology to the ecological effects o f chemicals. Ecotoxicological studies measure the effects o f a chemical substance in the environment on indigenous populations o f organisms. They provide a mechanism to estimate hazard. Ecotoxicological data are appropriately interpreted with knowledge o f the ecosystem where the organisms live. In aquatic ecntnx studies, what may be toxic tinder conditions created in the laboratory, may be more or less toxic in the aquatic environment due to factors present in the aquatic ecosystem which affect bioavailabilily. Also the chemical itaelf may be transformed as a result o f physical and biological mechanisms, including metabolism. An accurate evaluation o f toe toxicity o f a chemical requires knowledge o f these factors. Sulfonated perfluorochemicals appear to produce a variety o f responses in single species tests o f aquatic organisms. Different species have varied significantly in their response to the same chem ical even when using toe same laboratory procedure. In ecotoxicology, environm ental concentration often substitutes for knowing toe actual amount or dose o f a chemical entering an organism, but concentration and dose may not be directly related and their relationship varies from species to species. Basic environmental toxicity screening data arc available for many sulfonated perfluorochemicals (see Table 12), although their quality is variable. In considering the toxicity test results, it is important to note the year o f the test. Test protocols typically were developed considering water soluble, stable and well-dispersed compounds. Compounds such as sulfonated perfluorochemicals challenge test protocols due to their insolubility, polymeric, or surface active nature. The older data may reflect these test lim itations. Older test protocols are not comparable to recent and current bioassays that follow accepted, standardized test methods (OECD/USEPA). Almost all previous testing used products which are complex mixtures and not purified perfluorochemicals. In old tests, toe sulfonated perfluorochemical product used was likely m ore variable, with more impurities because manufacturing processes and product purity have significantly improved over time. Several tests were hampered by toe insolubility o f toe perfluorochemical and results are expressed as greater than toe measured solubility. Two sulfonated perfluorochemicals have more toxicity test data than others because o f their use as insecticides in ant and roach bait stations. These perfluorochemicals are N EtFOSA and PFOS Li salt. Toxicity data on these compounds may be found in toe disclosures filed by other registrants under toe Federal Insecticide, Fungicide and Rodenticide Control Act (FTFRA). 40 3M has evaluated the reliability o f its aq u atic to x icity te s t d a ta b ase. The n u m erical descriptor is modeled after the reliability coding used by EPA's Office o f Toxic Substances for the AQUIRE (Aquatic Information Retrieval) toxicology data base. Table 12. Ecotoxicity Testing on Sulfonated Perfluorochem ical Products Pimephalesprom elas=*Fathead minnow Salmo gairdneri= Rainbow trout Seienastrum capncom utum Green algae Lepomis macrochirus=* BluegiU sunfish Daphnia magna*= Water Flea Micxatax=Photobacterium phospharevm R eliability Codes: 1. Study used published test guidelines or well-documented procedures. Control performance was satisfactory. Toxicant concentration was measured. Test water temperature, pH and dissolved oxygen w o e m easured. 2. Study meets all the criteria for quality testing but has one or more o f the following deficiencies: A. Nominal test substance concentration; actixal concentration not measured. B. Test water quality variables not reported or incomplete. C. A water accommodated fraction (WAF) was used. D. Analytical methodology was questionable. 3. Study does no t meet the criteria for quality testing. Characterized by one o fthe following: A. Demonstrated weaknesses in experimental procedures. B. A static test with unmeasured concentrations was conducted in the presence o f precipitate or some undissolved chemical . C. Insufficient description o f methods. D. U nsatisfactory control m ortality; P roduct's Principal Fluorochem ical POSF N-MeFOSE alcohol N-EtFOSE alcohol N-EtFOSA T est Organism Pimephales prom elas Lepomis macrochirus Daphnia magna S. capricomutum Pimephales promelas ' Daphnia magna Pimephalespromelas Ceriodaphnia dubia Daphnia magna P im ephales prom elas Study Type R esu lts mg/L Y ear 96hrL C 50 96h rL C 5 0 48hrL C 50 14dayEC50 30 day hatch, growth,survival histopathology NOEC LOEC 4 8 h rE L 5 0 48 hr ELIO 48 hr NOEL 9 6 h rL L 5 0 06 hr 1X10 96 hr NOEL 4 8 h rE L 5 0 48 hr ELIO 48 hr NOEL 48h rE C 5 0 96 hr LC50 >1000 >solubiliy >solubiIiy >1800 .020 .020 >.020 14.5 7.3 5.8 206 115 130 328 184 216 3.2 34 84 79 79 81 78 78 98 98 98 98 98 98 98 98 98 84 84 R elia b ility Code 2A 3B 3B 3B 2D 2D 2D 2D 2A,C 2A ,C 2A ,C 2A,C 2A,C 2A,C 2A,C 2A,C 2A,C 3B 3B 41 Product's Principal Fluorochem ical Test Organism Study Type R esu lts mg/L N-EtFOSEA N-EtFOSEMA PFOS NH* sa lt, PFOS L i salt PFOS K salt PFOS DEA salt perfluoroClO sulfonic acid, NH4+salt . K salt o f carboxylic acid analogue o f N-EtFOSE alcohol , N-EtFOSE alcohol ethylene oxide adduct Pimephalespromelas P im ephales prom elas P im ephales prom elas P im ephales prom elas Microtox P.phosporeum Daphnia magna Pimephalesprom elas M icrotax Daphnia magna Selenastrum capricornutum P im ephales prom elas Lepomis maerochirus Salmo gairdneri Daphnia magna Pimephales promelas Pimephalespromelas ~ Lepomis maerochirus Daphnia magna Pimephales promelas M icrotox Pimephales promelas Selenastrum capricornutum Daphnia magna M icrotox Pimephalespromelas Lepomis maerochirus Daphnia magna 96 hr LC50 96 hr LC50 96hrL C 50 96 hr LC50 30 min EC50 48 hrEC50 48hrN O E C 96 hr LC50 30 min EC10 30m inEC50 48hrE C 50 28 dayNOEC 4 day EC50 cell count 14 day EC50 cell count 30 dayNOEC 30 dayLEC 96h rL C 5 0 96 hr LC50 96h rL C 5 0 48 hr EC50 96h rL C 5 0 96 hr LC50 96 hrLC50 96 hr NOEL 4 8hrE C 50 9 6hrL C 50 30 min EC50 96brL C 50 96brN O E C 96 hr EC50 96hrN O E C 48hrE C 50 48hrN O E C 30 min IC50 96 hr LC50 96hrL C 50 9 6 hrL C 50 48 hr EC50 >1000 >1000 85 100 >1000 ' 210 100 19 45 >280 27 7 82. 95 1 1.9 38 68 11 50 29 32 31 18 1 44 4.8 330 97 54 600 216 9.1 3.9 270 518 15 285 ' 1.5 Y ear 84 84 74 74 94 ... 94 94 94 91 91 84 84 82 82 78 78 77 78 78 79 74 73 79 79. 92 92 92 97 97 97 97 97 97 97 81 74 78 78 R elia b ility Code 3B 3B , 2 2A 2A . 2A 2A 2A 2A 2A 2A 2A 2A 2A 2D 2D 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 2A 3B 3A 2A 2A T able Key E C 5 0 - Median Effective Concentration. It is the concentration of a test substance that causes a 50% effect on a specific characteristic of the test organisms (e.g. immobilization o f 50% o fthe Daphnia, reduction in algal cell growth by 50% as compared to the controls) after a specified exposure period. It is the usual endpoint in a toxicity test with Daphnia and other small organisms where death is hard to determine or in tests where growth is measured. 42 1'T ''.p fiifijr. LC50= M edian Lethal Concentration. It is the concentration o f a substance that kills 50% o f the test organisms exposed to it in a specified time. It is the usual endpoint in an acute toxicity test with fish. IC 50= M edian Inhibitory Coaccatratioo. It is the concentration o f a test substance that Inhibits a biological process o f a test organism by 50% (e.g. light production, respiration) after a specified exposure period N O E l> N o Observed Effect Level N O E O N o Observed Effect Concentration ELHBffective Loading, LL=Lethal loading. These are used where the test substance is not completely water soluble. A water accomodated fraction (W AF)is prepared. The test substance is loaded into water at different loadings to prepare each test concentration. The solutions are mixed and the liquid fraction is decanted to use as tite test water. Additional studies are underway on ecotoxicity using established OECD/EPA methods. Initially purified PFOS and EtFOSE alcohol are being tested to determine acute and chronic toxicity to a wide range o f species. The results to date are found in Table 13. 43 Table 13. New Ecotox Studies on PFOS, potassium salt Param eter W astewater Bacteria (OECD 209) Sem nastrum capricormttu (green algae) Daphnia magna (freshwater flea) . hfysidopsis bahia (marine shrimp) Freshwater mussel > Study Type 3 hr NOEC 3 hr. EC50 Inhibition @ highest cone (1Q00 mg/L) 96 hr NOEC (growth rate) 9 6 h rE rC 10 96 hr. ErC50 Acute 48 hr NOEC .Acute 48 hr EC10 A cute48hrU C S 0 A cute48hrEC90 21 day semi-static life cycle NOEC 21 day semi-static life cycle NOEC Acute 96 hr NOEC Acute 96 hr EC50 35 day flow thru life cycle NOEC 35 day flow thru life cycle NOEC Acute 96 hr NOEC A cute96hrLC50 R esu lts, 1.0 mg/L >1000 mg/L 39% ~ 48 mg/L 65 (59-69) mg/L 138 (125-449) mg/L 36 mg/L 57 (<12->99) mg/L 66 (36-99) mg/L 69 (<12->99) mg/L 13 mg/L 26 mg/L 1.2 mg/L 4.0(3.3-5.0) mg/L 0.28 mg/L 0.6 mg/L 22 mg/L 65 m g/L - Pimephcdes prom elas (fathead winnow) , Oyster Shell Deposition Avian Dietary Toxicity Testing r Acute 96 hr. NOEC A cute 96 lir LC30 47 day early life-stage toxicity NOEC 47 day early life-stage toxicity LOEC Acute 96 hr NOEC Acute 96 hr EC50 (Solubility limits precluded EC50) Inhibition @ highest cone (3.3 mg/L) Acute Mallard DuckLC5Q Acute Mallard Duck, no mortality Acute Mallard Duck NOEC Acute Bobwhite Quail LC50 Acute Dobwhxtc Quail, no m ortality Acute Bobwhite Quail NOEC 3.6 mg/L 10(8.8-12) mg/L 0.33 mg/L 0.65 mg/L 2.1 mg/L >3.3 mg/L 28% 730 (532-1059) mg/kg 160m g/kg 40 mg/kg 214 (163-260) mg/kg SO m g/kg 80 mg/kg Data in italics are from draft reports. A ll o f the results shown in Table 12 suffer ftom limitations in the reliability o f the data, nnA there is a olear need for high quality ccotoxicity data using established OECD/EPA methods. Testing on purified PFOS and EtFOSE cohol is in progress. However, the available new data (Table 13) and the historic data are consistent in that almost all toxicity values for PFOS and related sulfonated perfluorochemicals are greater than 1 mg/L and m ost are greater than 10 mg/L. An exception is the fathead minnow results reported on Table 12 for N-EtFOSE alcohol where apparently there is no acute toxicity a t 44 or above the w ater solubility level The available new'data on PFOS itself s u r e s t that it has sim ilar aquatic toxicity to that o f other anionic surfactants (9). Few other conclusions can be reliably drawn at this tube. For instance, ecological risk assessment typically relies on chronic toxicity values, but there are too few data on th to draw any . conclusions. 11.0 Comprehensive Plan to Assess Environmental Exposure The ongoing activities described in the previous sections o f this paper are being carried out as part o f a 3M developed comprehensive plan using a combination o f 3M resources and outside experts. This plan, summarized in Figure 1, is designed to assess the potential pathways1o f environmental exposure associated w ith the manufacture, use and disposal o f its sulfonated perfluorochemical products. . 11.1 Plan Overview The plan structure consists o f four components: 1. Characterize the properties critical to understanding the fate and transport o f sulfonated perfluorochemicals. 2. Estimate the releases o f sulfonated perfluorochemicals. 3. Characterize the distribution o f sulfonated perfluorochemicals in the environm ent. 4. Estimate human and ecological exposure to sulfonated perfluorochemicals. Several individual research projects feed information into each component. The early components provide information needed to complete the later ones. Thus the information base expands when one goes from component 1 to component 4. This section provides an overview o f the plan and its research projects. Specific descriptions o f how and why the research projects are being conducted can be found in the preceding sections o f this document. The results generated by this plan will be combined w ife the ecotoxicological studies to develop an assessment o f risk. A tentative initiation date for each o f the research projects is found in Figure 2. It is expected that the studies w ill continue over several years. 45 Figure 1. Diagram of Fluorochemicai Assessment Plan. 46 Figure 2. Schedule for FC Exposure Plan Components. lHFirst Quarter, 2Q= Second Quarter, 3Q=Third Quarter, 4Q=Fourth Quarter Characterize Fate & Transport P ro p e rtie s Estimate Releasee PFO S Phye/Chem Properties............. ... _. . EtFOSE alcohol Phys/Chem Properties MeFOSE alcohol Phys/Chem Properties Hydrolysis Photodegradation and AtmosphericTransport Biodegradation (aerobic and anaerobic) Sorption.Processes " PFOS Bioconcentration EtFOSE alcohol Bioconcentration initiation Dates Complete 1Q2000 1Q2000 1Q 1999 1Q 1999 1Q1999 2Q 1999 2Q2000 4Q2000 3M Plant Effluent & Process Waste Analyses Estimate Mfg, Supply Chain, & Use Waste Streams EstimateWaste Releases FC Thermal Destructability 1Q1999 Complete 1Q 1999 2Q 1999 Characterize Distribution in Environment Bird & Fish Analyses U.S.Bird Livers Analyses Biosphere Sampling Pan Multi-media Modeling Multi-cities Study Complete Complete 1Q 1999 1Q2000 1Q 1999 Estimate Exposure Carpet Study ' P ap er and Packaging Studies Exposure Scenarios 2Q 1999 1Q 1909 2Q 1999 Ecotoxicity Determinations PFOS Acute Ecotoxicity PFOS Chronic Ecotoxicity EtFOSE alcohol Acute Ecotoxicity EtFOSE alcohol Chronic Ecotoxicity FOSA Acute Ecotoxicity FOSA Chronic Ecotoxicity 1Q 1999 2Q 1999 2Q 2000 3Q2Q00 1Q2000 2'Q 2000 47 ) 11.2 Component 1: Characterize f ate and Transport Properties This component was developed in three steps. First, the important fete and transport mechanisms were identified. Next, priorities were set for testing, with the likely " degradation products having the highest priority for testing. Finally, methods and laboratories were selected to do the testing. For sulfonated perfluorochemicals not tested, models are being developed that will predict physical and chemical properties. The specific research projects underway are: ' 1. Physical and chemical properties testing. 2. Hydrolysis testing. .' 3. Photodegradation and atmospheric transport testing. 4. Anaerobic and aerobic biodegradation testing. 5. Soil and sediment sorption testing. 6. Bioconcentration testing. 11.3 Component 2: Estimate Releases This component was also developed in several stages. F irst product sales data from 1997 were used to identify a study set o f products. The study set was based on volume o f use and waste streams, mode o f release and product chemistry. Next, evaluation efforts o f the waste streams generated focused on those study set products sold by 3M in the greatest quantities in the United States. Commercial and residential uses o f these products, including transportation, handling and application during the supply chains that lead to product use, were examined. Additionally, the releases likely to result from disposal during these portions o f the products' life cycles were also estimated. The estimates included disposal via incineration, landfilling and wastewater treatment. W aste streams generated at the start o f the products' life cycles, i.e. the manufacturing process, were also examined. Better estimates are continuing to be developed o f waste streams occurring during the manufacturing process. 11.4 Component s: Characterize Distribution in the Environment This component is distinguished by iterative interaction between modeling and field sampling. M odels are being used to suggest sampling locations and detection lim its. Field sampling is planned to obtain empirical data to validate model output and improve predictions. As new data become available, research projects become more refined and focused. The research projects completed or planned to characterize environmental distribution include: 48 1. Field sampling o f environmental m edia near the Decatur manufacturing plant. 2. Screening o f eagle and albatross plasma and fish tissue from archived samples. 3. Analysis o f w ild bird livers. 4. Biosphere sampling plan to determine levels in biota o f different geographic locations. 5. Development o f a.multuriedia model for predicting distribution. 6.... M ulti-cities studies in which cities o f a sim ilar size are paired, one demonstrating sigrfficantm anufacturing or commercial uses o f sulfonated , perfluoroqhemicals, the other having no identified use q f sulfonated perfluorochem icals. 11.5 Com ponent4 : Estimate Exposure When data from the release and distribution components are available, hypotheses w ill be developed about important exposure pathways. Iterative sampling and modeling w ill be used to test these hypotheses and to determine the important exposure pathways to be used in risk assessment. The research projects planned to estimate exposures are: 1. GJarpet releases and links tam gestion,inhalation and dermal exposure. 2. Paper and packaging studies and ingestion exposure. . 3. Exposure scenarios which combine information from the release, fete, and sampling studies. 12.0 Ecotoxicity Determinations In conjunction w ith the studies described in the four component plan described above, 3M is conducting ecotoxicological studies. Ecotoxicological studies are used to estim ate hazard. Initial ecotox testing is focusing on PFOS. EtFOSE alcohol, and FO SA The results o f the release, distribution and exposure assessments may provide reasons to test more substances. The following research projects on ecotoxicity are planned or underway: 1. Aquatic acute toxicity studies: sewage microorganisms, freshwater and m arine algae, duckweed, daphnia, mysid shrimp, freshwater mussels, fathead m innow s. -- 2. Terrestrial acute toxicity studies: mallard duck and bobwhite quail dietary exposure studies, earthworm toxicity studies, and green plant growth and uptake studies. 3. Aquatic chronic toxicity studies: oyster shell deposition, daphnia, mysid shrimp, frog embryo development and fish early life stage studies. 49 t '.I 4. Terrestrial chronic toxicity studies: mallard duck and bobwhite quail reproduction. - 13.0 Ecological Risk Evaluation Evaluating ecological risk is more complex and more uncertain than assessment o f human health risks where a clearer connection can be drawn between dose and response. As this comprehensive science and exposure assessment program progresses, a framework in which ecological risk can be evaluated will be developed. The evaluation o f ecological risks and human risks will both use information about distribution in the environm ent and exposures generated by this comprehensive exposure plan. Because o f . the scope and magnitude o f the overall program, aspects o f the knowledge gained w ill be com partmentalized into discreet elements to create this framework. For example, as data on ecotoxicological properties, fate and transport mechanisms, and environmental distribution are developed, they will be used to evaluate ecological risk within a certain geographic area or locality. The science data and environmental sampling results w ill be applied to a very specific area and set o f species to evaluate relative risks in that area. Building a number o f these compartmental evaluations w ill result in a much more complete picture o f ecological risk. This evaluation w ill identify additional actions 3M could take to minimize the releases o f sulfcmated perfluorochemicals. . 50 14.0 References < 1. Cohrssen, J.J. and Covello, V.T. Risk Analysis: A Guide to Principles and Methods for Analyzing H ealth and Environmental Risks, U.S. Council on Environmental Quality, Executive Office o f the President, 1989 -............. 2. Fate, Transport and Transformation Test Guidelines, Office o f Prevention, Pesticides and Toxic Substances (OPPTS) 835.2110, Hydrolysis as a Function o f pH and Temperature. EPA 712-C-98-057, January 1998. 3. Flynn, R.L. in "Industrial Applications o f Organochlorine Compounds," Proceedings o f the Symposium on Electrochemistry in the Preparation o f Fluorine and Its Compounds, Childs and Fuchigami, Eds, The Electrochemical Society, Inc.: Pennington, NJ, 1997; 97-15: 51. 4. Guidelines for the Testing o f Chemicals, vol. 1, Section 1, "Physical Chemical Properties," Organisation for Economic Co-operation and Development (OECD), Paris, France. 5. Guidelines for the Testing o f Chemicals, vol. 1, Section 2, "Effects on Biotic Systems," Organisation for Economic Co-operation and Development (OECD), Paris, France. 6. Key, B .L.; Howell, R.D.; Criddle, C.S. "Fluorinated Organics in the Biosphere," Environ. Sci. Technol, 1997,31:2445-2454. 7. Key, B .L.; Howell, R.D.; Criddle, C.S. "Defluorination o f Organofluorine Sul&r Compounds by Pseudomonas Sp. Strain D2," Environ. Sci. Technol, 1998,32:2283 2287. 8. M ackay, D . Shiu, W.Y., and Ma, K.C., Illustrated Handbook o f Physical-Chemical Properties and Environmental Fate for Organic Chemicals. Volume 1: Monoaromatic Hydrocarbons, Chlorobenzenes and PCBs. Lewis Publishers, Chelsea, MI. 1992. 9. Scholz, N ., "Ecotoxicology of surfactants," Tenside Surfactants and Detergents, 1997, 34:229-32. 51 TTTT r "t--