Document pmpb056B3rpyYj4enxeLy18Qk

FPRI tnc Power nstearcFl Institute Topics: PCS Chemical analysis First Trantformtrs Capacitors Electrical insulation EPRl EL-4503 Pro|ect 2028-12 Pinal Papon Marcn 1986 State-of-the-Art Review: Pyrolysis and Combustion of PCB Substitutes Praparsd by SCS Engineers. Inc. Long Btacn. California RONS 215982 REPORT SUMMARY SUBJECTS H*roou/toic luostancssr T&D: Substations TOPICS PCB Ghtmical analyst* First Trsnatormars Capacitors Elactncai insulation AUDIENCE Environmamti managars i Distribution engmaars Stete*of*th*Art Ftevtew: Pyrolysis and Combustion of PCB Substitutes Utilities are concerned about the possible long-term toxic effects of trace combustion products of potential PCB substi tutes. This literature review turned up much information on the bulk combustion products of replacement compounds that are already in wide use but few data on either the trace or bulk prod ucts of relatively new materials. BACKGROUND ,, Several compounds are under consideration at substitutes tor polychlori nated biphenyls iPCBs) in transformart and capacitors. TO tnsurt that mass insulating materials do not doss problams simitar to those encountered with PCBe utilities need to determine not only their electrical, thermal, ano pnysteat characteristics but also tht bulk and tract maianais they product whtn pyrolyzed orcomoutted. OBJECTIVE To review the literature on the materials producad during ms oyroiyus tnd combustion or ttvarat solid, liouid. and gaseous tubttitutts tor PCBs. APPROACH Rettarchert focused their review on is potential PCB replacement materials--2 solids (epoxy resins ano polyvinyl chtonda). 2 oases (Freon 113 and sulfur hexafluoride). and 12 liquids, including phinsists tttert. beniyineocaprata. silicone, chiorobentenea methylated diphenyietnene. and oarchtoroethyiene. In particular, protect workers sought information on trace products related to those found during PCB thermal decomposition--tor example, poiyaromanc hydrocarbon*. polychlorinated dlbenzofuran. end poiychionnated dibahzedioxtn. RESULTS The literature contains t great deal o> information on tha bulk pyrolysis and combustion products of PCB tubttitutts that have had wtda use in tne dec trlcity industry end other industries. Exempiee of such substitute* ere epoxies, sulfur hexefiuonde. silicone, pnthatete esters, methyl tied dioneoyi ethane, perchloroethylene. and polyvinyl chloride. However, there it Mttit information on the trace products of concern. Literature on both the bulk end trace combustion product* ol tom# of the lees familiar replacement compounds is spare* and. in some caaes. nonexistent. For example researcher* found no information on tn# matariais ganersied during thermal HONS 215983 decomposition of benrylneocaprate and alkyibiphenyts. The study con cluded that further laboratory work is needed to identity additional trace materials of the better known PCB substitutes and both bulk and trace products of the newer compounds. EPRl PERSPECTIVE ' Concern over the toxicity of PCBs has focused on both ns bulk and trace combustion products. To bund a database as sophisticated as that lot PCBs. researchers must perform intensive analyses of all potential sub stitutes. EPRl is currently conducting laboratory studies to enlarge the present database. Pour such studies are project RP2028-11. which is fo cusing on the bulk pyrolysis and combustion products of several liquid PCB substitutes, and projects RP2026-15, RP2028-16, and RP2028-17 which are analyzing the bulk and trace products of a group of liquid, solid, and gaseous PCB substitutes. " PROJECT RP2026-12 EPRl Protect Manager: Oil Addis Electrical Systems Division Contractor: SCS Engineers. Inc. For further information on EPRl research programs, call EPRl technical Information Specialists (4iS) 855-2411. HONS 215984 State-of-the-Art Review: Pyrolysis and Combustion of PCB Substitutes EL-4503 Research Project 2028-12 Final Report March t986 Prepared Dy SCS ENGINEERS. INC <*0i4 Long Seacn Bourevaid long Beacn. California 90607 PrinciDal investigators J. V ZPozme* J R Marsn D Gutn A Bonmerud Prepared for Electric Power Research institute u-*i2 Hitiview Avenue Paio Aito. California 94304 EPRi Proiect Manager G Aoais Transmission SuDStahons Program Electrical Sysiems Division . MOHS 215985 ORDERING INFORMATION ------- - Requests for copies of mis report snouid be airecioa to Research Reports Center IRRCI. 8ok 50*90. Peio Alto. CA 9*003. (*15> 965-*08i. There is no charge tor reports requested by EPRI member utilities ana affiliates. U S. utility associations. U.S. government agencies (federal, state, ana local I. media, and foreign organizations with which EPRt hais an miormation exchange agreement. On request. RRC win send a cetetog ol EPRf reports. c ''mm.:# *rvc ai qntt ffif/vM NOTICE r"* rt?drt Ml C'*D***0 Cv `*4 5' *3w#' -itiren i*tii^rt fine S' - *"'*`3 0**6w || r> tcCOv"! O' ..C'* CC MOre; 'tCI<< ** Wi 'r*0'sa*'/j,-o,in|i '*^oo#ow fo OOfW li'ing o" & >r-v ol -frr> .sjri*ri* O' mcufo A'P rO*o#c: !0 g$#0t i** -mso^lIrOfl tDDi'ftigt m#moo O' O'OCtM 9tK`Qf3 * -- ^ *4W' O' "`4' ! .; .4* Tttv *Ql 0<rvtOW ;wr:#0 'Qf't* 0' l*v f'lOtfi'tfi *tin etc :C :*4 ,i< C ; 'C' j'-r; ?'<* l*t v*t Of ny ir.rjKmoiron 8008'tiwf '^if'OC J' 0'tlt OifCOtvO r :-ij. -*>DQ'f ^'Difigov SCS i-yntt*'* me lcv"? #tte C'omio MONS 215986 abstkact An extensive literature search wis conducted to Identify tne combustion ina pyrolysis by-products of severe! PCB substitutes. These compounds incluoeo pnthalate esters, benzylneocaprate, siloxines, chlorineted benzenes, uixylyletnene, alkylblphenyls, pereff'ntc hydrocarbons, perchloroetnylene, cnlorof1uoroctrbons, sulfur hexafluoride, epoxy resins, end polyvinyl cnlorloe. M> Informetion wes found on the thermal decomposition of oenzylneaceprete or the alkylblpnenyls. Siloxenes decompose to silica. cyclic siloxenes, ena hydroceroons of reletlvely low toxicity. Pereffinic hydrocarbons decompose Into e series of shorter hyarocerbon compounds. Epoxies yield both alkyl end aryl hydrocarbons. Sulfur hexafluoride yields a series of sulfur and thtonyl fluoride compounds eng unidentified toxic PICs. The,phthelate esters produce phthellc anhydride, napn. thalene, and biphenyl. Methylated diphenyletnene ana phenylxylylethane decompose Into a series of polynuclear aromatic hydrocarbons. Perchloroethylene produces predominantly hydrogen chloride and chlorine and sometimes phosgene. Chlorofluorocarbons yield these compounds ana their fluoride analogs. Polyvinyl chlo ride decomposes Into phosuene, naphthalene, benzene, and chlorinated benzenes. Chlorinated benzenes end possibly butyl at ed monocnlorooi phenyl ether can produce PC00s and PCOFs. MONS 215987 HI acknowledgments It Is Impossible to do an adequate state-of-the-art review without consulting those individuals end organizations currently engaged In research. Published information is often months to years out of date, and only contacts with re searchers can produce a timely review. Consequently, it was necessary to request assistance from many organizations. There is no way to list every company or individual talked to during the course of tnis effort. Out several, because of the extent of their cooperation, deserve special mention. Among the organizations and companies which provided extensive assistance and/or material were EPAl, Westinghouse Electric Corporation, General Electric Corpora tion, Dow Chemical, Monsanto, ATE Corporation, ISC Chemicals Limited, and Cioa(ieigy. . Individuals deserving special thanks Include A. Crespi and B. Kueng (Clba-Geigy); E, welsh, C, C. Claiborne, and D. L. Mandelcorn (Westinghouse); h. Burns and S. J. Parkinson (ISC Chemicals Limited); D. Duvall, F. Hileman, n, taghes, and L. Parts (Monsanto); A, Ristow, E. Feuerstein, and T. L. Mayes (Gen eral Electric Corporation); k. kinnebrew (General Electric Transformer Service Group); J, Hearse (National Industrial Transformers, Inc.); P. Gervason (Prodelac); T. Lovkvist (ASEA Kobel); H. M. Wortnington (The Micanite and Insulators Company, Ltd.); C. P. McShene (RTE Corporation); ht W. Stacy (Southwest Research Institute); D. C. Wilson (Harwell Laboratory); G. A. Gauger (Thomas A. Edison Technical Center, McGraw-Edison Company); k L. C. Meuzelaar (University of Utah); W. Aubey (University of Dayton Research Institute); C. T. Olsen (Air Force Aeronautical Medicine Research Laboratory, wrlgnt Patterson Air Force Base); M. Saperstein (Southern California Edison); and C. Hi 1 ado. Thanks also to the EPRI project Manager, G. Addis, for constructive guidance and assistance in formulat ing and executing the project. *0*S 215988 V CONTENTS Section 1 INTRODUCTION Background Objectives Report Organization 2 data ASSESSMENT Introduction Confcustlon/Pyrolysls Chemistry Toxicity end Exposure Limits Synergism 3 PR0P0SE0 PCB SUBSTITUTES: LIQOIOS introduction Esters SI loxanes Aromatics Perefflnlc ttmrocaroons 4 DIELECTRIC cases Cniorofluorocarbons Sulfur hexafluoride 5 PROPOSED PCB SUBSTITUTES: SOUQS Introduction Epoxy Resins Polyvinyl Chloride (PVC) 6 REFERENCES APPEWiX A GLOSSARl Ap ABBREVIATIONS APPEttlX B GENERIC AfC TRADE NAMES v 11 Page 1-1 1-1 1-1 1-3 Z-l 2-1 2-1 2-3 2-4 3-1 3*1 3-1 3-8 3-13 3-24 4-1 4-1 4-? S-l 5-1 5-1 5-6 6-1 A-l B-l HONS 215989 ILLUSTRATION lias.re 3-1 Phthalate Esters 3-2 Typical POKS Pyrolysis Products 3-3 Chlorinated Beniene Compounds used in Dielectric Fluids 3-4 Typical Chloropentene Thermal Reactions 3-5 Formation of PCOFs/PCDOs from PCOPEs Page 3-3 3-10 3-14 3-17 3-22 HONS 215990 ix TABLES Table 2*1 Toxicity and Exposure Limits Oiti of Common Toxicants '' 3-1 Chemical and Physical Characteristics of Selected Phthalate Esters 3-2 Effect of Reaction Temperature on the Thermal Decomposition of Selected Phthalate Esters 3-3 Toxicity and Exposure Limits of 01 -(2-Ethylhexyl )-Phthal ate, o-Dloctylphthalate, and Their PICs 3-A Chemical and Physical Characteristics of a Typical PDMS Transformer Fluid 3-5 Chemical and Physical Properties of Selected Chlorinated Benzenes 3-6 Chlorobenzene Decomposition Temperatures 3-7 Formation of PCDFs and PCOOs from the Pyrolysis of Chlorobenzenes at 620 *C 3-t) Tox 1 cl tie* and Exposure Limits of Chlorooenzene Combustion/ Pyrolysis Products 3-9 Physical and Chemical Properties of Selected Oiaryl Alkanes 3-10 Physical and Chemical Properties of Selected Alkyl Biphenyls 3-11 Chemical and Physical Characteristics of PAO 13CE 3*12 Chemical and Physical Characteristics of RTEmp 3-13 Chemical and Physical Characteristics of Perchloroethylene 3- 14 Toxicity and Exposure Limits of Perchloroethylene and Decomposition Products 4- 1 Toxicity and Exposure Limits of Freon 113 Decomposition Products 4-2 Chemical and Physical Characteristics of Sulfur hexafluoride 4*3 Toxicity and Exposure limits of Sulfur hexafluoride Decomposition Products - Page 2-J 3-2 3-5 3-7 3-9 3-12 3-13 3-16 3-18 3-19 3-24 3-25 3-26 3-2B 3-30 4-3 4-4 4-6 HONS 215991 X1 TABLES (continued) Table ' Page 5- 1 Volatile Products from Pyrolysis of Epoxy Resin . 5*3 6- 2 PICs Pound In Fire Oases from Two Cast Resin Transformer Fires- 5-4 5*3 Toxicity and Exposure Limits of Epoxy Decomposition By-Products 5-5 5-4 Decomposition Yields of Some Aromatics Obtained by the Pyrolysis of PVC (Soon-103 EP) at Various Temperatures ' -5-8 5-5 Toxlcltles and Exposure Limits of PVC Thermal Degradation By-Products 5-10 HONS 219992 x 11 EXECUTIVE SUMMARY Since 1976. tut manufacture ana comercUl use of polychlorinated biphenyls (PCBs) have been restricted by law. Proposals have been made to remove existing PCB-containing transformers and capacitors from service. A number of chemicals nave been and are being considered for use as PCB substitutes In electrical applications. Although all nave been tested for acceptable dielectric, heat transfer, and flamaolllty properties, few have been extensively analyzed for conbustlon by-products. PCBs were not restricted for problems with their dielectric properties or heat transfer capabilities, but rather for suspected toxicological properties that were not considered when PCBs were first commercially produced, These toxicolog ical properties are attributed not only to the PCBs themselves, but also to the presence of various toxic chlorinated aromatics In the by-products of PCB/askarel fires. it is thus necessary to examine the chemical behavior of proposed PCB substitutes to identify tne potential for similar problems. The purpose of this report was to conduct a state-of-the-art literature review to determine what is known regarding the chemical and toxicological nature of the conbustlon byprod ucts of various PCB substitutes. DATA ASSESSMENT Until very recently, the major concern in any fire or explosion episode was fire fighter safety. As a result, tests neve focused almost exclusively on major gaseous conbustlon products such as carbon monoxide, hydrogen cyanide, oxides of nitrogen and sulfur, hydrogen chloride, and so forth. These are the Immediate life-threatening conbustlon products with which one must deal, and they are gen erally tne simplest to detect and measure. towever, it has become evident in tne aftermath of several PCB Incidents that tne toxic trace conbustlon product problem Is concerned more about soot deposits tnan the vapors formed. Trace products of Incomplete conbustlon (PICs) are produced MQNS 215993 S-l in lucn low levels tnat vipor emission* dispersing In tne atmospnere rapidly raacn IncomtQutntlil levels. Theie same PIC* may concentrate in soot, however, and loot provides a reservoir for continued emissions or exposure long after a Tire has been extinguished and the gases dispersed. furthermore, analysis of soot or even gas samples for trace PICsat the levels typically produced requires very sophisticated analytical Instrumentation and highly trained specialists. Even then, complete Identification of PICs in a par ticular sample may take years of work. As the level of sensitivity of enalytlcal instruments increases, wen more toxic PICs may become evident, for Instance, continued research on PCB ctwtustion byproducts has identified not only chlori nated dioxins and dibentofurans, but alto polychlorinated chrysenes, pyrenes, xantnenet, terphenyls, quaterphenyIs, and others, This level of research has been directed at PCB because of its negative publicity and the perceived crises In regards to several PCB Incidents, to such crises exist for tne proposed PCB substitutes, As a result, with a very few exceptions, compareble research has not been conducted Into tneir PICs, As noted above, existing data tend to focus on the macro cntoustion products present In the vapor state rather than trace PICs in the soot. Consequently, there is relatively little data on trace PICs comparable to that on dioxins or dibentofurans. These types of by-products simply have not been identified and studied in most cases. PROPOSED PCB SUBSTITUTES: UQUIOS Two types of esters, pnthalate esters and bentylneocaprate, have been considered as PCB substitutes. In general, pyrolysis of esters will produce gases similar to and no more toxic than those obtained from hydrocarbons. One major decomposi tion product is phthalic anhydride which can react further to form naphthalene and biphenyl. In the presence of chlorooentenes, PCB and chlorinated naphthalene may be produced. Polydimethyl slloxane has bean widely used as a retrofill fluid. During tharmal decomposition, amorphous silica, several cyclic siloxanes, and a variety of sim ple hydrocarbon* are produced, host of these compounds are not highly toxic, ana animal inhalation tests on pyrolyieo siloxenes have generally iraicated a low toxicity. HONS 215994 S-2 Chlorobenzenes hive generally not been considered PCB substitutes s much is waittvs to- PCBS end otner dielectrics. Decomposition or cnlorooenzenes cin produce e variety of highly toxic compounds. Pyrolysis of monochlorobenzene can yield mono- end dlchioroblphenyls, chloronaphthalene. end vinyl chloride. In the presence of air, chlorobenzenes cen produce chlorophenols which cen. In turn, dimerize to PCOFs end PCODs. Many of these comoustlon/pyrolysts products are very highly toxic. . Methylated diphenyl ethane end phenyl xylyl ethane ere used es components In dielec tric rluld mixtures, ho Information was found on the confeustlon/pyrolysls prod ucts of either compound, but pyrolysis of chemically similar dlphenylmethane art tphenylethane produces a nuneer of polynuclear aromatic hydrocarbons. Few of these are considered to be highly toxic, but several are known or suspected carcinogens. There Is no Information available on the thermal decomposition products of butylated monochlorodlphenylether, another PCB substitute. However, both the unsub stituted dlphenylether and polychlorinated dlphenylethers nave been studied. Dlphenylether can break down Into phenol, benzene, and dlbenzofuran. Polychlori nated dlphenylether can produce PCOFs and PCOOs. There is also no Information available on the thermal decomposition products of the alkyl biphenyls, specifically isopropylblphenyl and n,-propylb 1 phenyl. Paraffinic hydrocarbons, such as polyatphaolefln and RTEmp, tend to decompose into a variety of short-, medium-, and long-chain saturated and unsaturated nydrocarbons. In the presence of air, organic acids, alcohols, and aldehydes are formed. Perchloroethylene Is being promoted as a transformer fluid neat, mixed with min eral oil, or mixed with fluorocarbons, Chlorine or hydrogen chloride are the principal decomposition products to be expected, under certain conditions, phos gene or trichloroacetic acid may be produced, DIELECTRIC GASES Chlorofluorocarbon! are used as dielectrics both neat and mixed with percnloroethylene. Although generally tnermally stable, at nigh temperatures or during HONS 215995 SO arcing, decomposition Is possible, Chlorine, hydrogen chloride, hydrogen fluo ride, phosgene, end carbonyl fluoride ere among the by-products which can be expected. Phosgene and carbonyl fluoride, in particular, are highly toxic. Sulfur hexafluoride can be used alone as a dielectric or in a mixture of gases. It produces a nuncer of long-lived by-products under fire and arcing conditions. The principal by-product is thionyl riuorlde*, other products Include sulfuryl fluoride, thionyl tetrafluoride, and sulfur dioxide. Toxicity tests on'sparxlng gases have yielded a higher toxicity than the above compounds would indicate. Thus, there are one or more unidentified trace decomposition products which con tribute significantly to the toxicity of the decomposition gases. PROPOSED PC6 SUBSTITUTES: SOLIDS Two solids were reviewed: epoxy resins and polyvinyl chloride (PVC), Complicat ing an assessment of their decomposition is the fact that neither is a pure com pound or even a mixture of a few discrete compounds. Rather, each Is a mixture of resin, hardener, filler, and various other materials, many of which may have a variable composition. Thus, the decomposition products of these solids can vary extensively, depending on the nature of the additives, Pyrolysis of epoxy resins can produce a variety of organic compounds Including bentene, methyl chloride, ethyl chloride, acetone, propylene, ethane, pentane, toluene, cresol, phenol, and others. Several of these are known or suspected carcinogens. Decomposition of PVC can yield bentene, toluene, xylene, aliphatic hydrocar bons, p-nethylnaphthalene, several chlorinated bentenes, and pnosgene. Because of the production of chlorinated bentenes, tne formation of PCDDs and PCDFs is a distinct possibility, although none have been identified in PVC pyrolysis gases to date. HONS 215996 S-4 Section 1 INTRODUCTION BAUGROUK) The U.S. Congress pissed the Toxic Suostmces Control Act in 1976. under re quirements of the Act, the tPA issued restrictions on tne minuficture end cornierciil use of polychlorlnited btpnenyls (PCBs). PCS production in tne United Stltes wis hilted in 1977, ino tne production of PCB-contllnlng tr ins formers md cipicltors nis been limited since then; however, there wis no order to tike such units out of service. The service life of these units is meisureo In tens of yeirs, so thit the replicement rite is low (ebout 0.7% per yeer) (117). However. Should the CPA order PCB-contlining units removed from service (FR <9(198):3996639969, 11 October 1984, Proposed Rules), the need to find icceptible PCB substi tutes will become crltlcil. Severs! possible substitutes for PCBs r* reisily nii'iblt (97, _97). PCis were not restricted for problems with their dielectric properties or heit trintfer cipibl11 ties, rither for suspected toxicologies! properties not considered in 1929 when PCBs were first commerciilly produced (117). These toxicologies! prop erties ire ittrlbuted not only to the PCBs themselves, but ilso to the presence of polycnlorlnited dIbtmodioxlns (PCODs), polychlorlmted dlbenzofurins (PCDfs). ind otner cnlorimted iromitlcs In utility PCB fluids ind In the by-products of PCB/isxirel fires. It Is thus necesstry to eximine the chemlcil ind toxlcologicil oenivlor of proposed PCB substitutes ind to Identify potentlil problems for eich iltermtlve cnemlcil option for trinsformers ind cipicltors before Its use becomes wlaespreid. OBJECTIVES The ivitlible PCB substitutes ire of widely viriint chemlcil composition. The toxicology ind environment!! hizird evilultlon of most of these .compounds his been reported (97), The objective of this study is to review the stite of the irt regirding the tnermil degnoitlon cnemlstry of PCB substitutes, the formitlon MOMS 215997 1-1 of hazardous products of Incomplete con&ustlon (PIC*) from these compounds. and the toxicity end environment*! hazard of those PlCs. The liquid, gaseous, ind solid dielectrics examined In this study are oesenoea below. Liquids e Phthtllc esters dl-U-ethylhexyl )-phth*l*te end dloctylphthalate. e Benzyl neoetprtte. ' e Polydimethyl si lox*n*. e Chlorobenzenes, e Butyl*ted monochi orodl phenyl ether, e methylated dlphenylethtne. e Phony Ixylyl ethane, e Isopropyl bl phenyl, e Propyl biphenyl. * Poly(l-octene), . e RTEmp, e Perchloroethylene. Gases e e Freon 113, Sulfur hexafluoride. Solids e Epoxy resins, e Polyvinyl chloride. The risk from exposure to these compounds Is modified In this study by the assumption that the hazard from the compounds studied occurs only when the com pound is in a transformer or capacitor which has failed catastropnlcal1y (rup tured) due to Internal arcing, which Involves temperatures of up to 6D00 *C for periods of less than a second, or due to snort-circuits which cause internet "hot 1-2 MONS 15998 spots,'' involving temperatures of up to 250 *C For a few minutes before evolution of gas From the fluid ruptures the transformer case, or wnicft has suffered fluid loss as a result of an external Fire Any unit which completes its ser vice life without losing fluid to the environment iv considered to pose no hazard in the context of this study. Such units may pose a disposal proolem; this dif ficulty is not considered here. REPORT ORGANIZATION ^ Section 2 of this report discusses proper assessment of the data. Sections 3, 4, and S cover the thermal degradation of the liquid, gaseous, and solid dielec trics. respectively, and the toxicology of their PICs. MOHS 215999 1-3 Section Z DATA ASSESSMENT INTRODUCTION ' If comparative toxicologic action of the combustion products of several materials is to be used, even In part, as a method of preferential selection, available Information must Include: e The nature and amount of the toxicologic agents generated when the materials undergo combustion or pyrolysis under specified decomposi tion conditions, ' e The dose-response characteristics or these agents. a The additive, antagonistic, or synergistic stress from the products generated by the degradation (72). The available data for each of the above criteria are highly variable, In terms of both quantity and quality. Some chemicals have been itudltd extensively, some not at all. Before addressing the specific data for each PCB substitute, ft Is necessary to examine the types of data available and their limitations. COMBUST ION/ PYROLYSIS CHEMISTRY The ideal way to Identify the expected tnermat degradation products of a given substance Is to coebust or pyrolyte It under conditions nearly Identical to those whlcn might be encountered In the field, yet in a tightly controlled environment, followed by a full-spectrum analysis of the vapors and soot produced. This Is virtually Impossible to do for a variety of reasons, not the least of which Is the cost of such experiments. In general, available data are based on laboratory tests and Incomplete analyses. Part of the problem Is that, until very recently, the major concern In any fire or explosion episode was firefighter safety. As a result, tests have focused almost exclusively on major gaseous combustion products such as carbon monoxide, hydrogen cyanide, oxides of nitrogen and sulfur, nyorogen chloride, and se forth. HONS 216000 2-1 These ere the immediate life-threatening combustion products with which one must a eel,'ena they ire generally the simplest to detect end measure. towever, it has become evident in the afte^inath of several PCfJ intid'ents that the toxic trace combustion product problem is concerned more about soot deposits than the vapors formed. This is not to suggest that a compound like a dioxin is less toxic in the vapor state. however, trace PICs are produced in such low'levels that vapor emissions dispersing in the atmosphere rapidly reach inconsequential levels. These same PICs may concentrate in soot, however, and soot provides a reservoir for continued emissions or exposure long after a fire has been extin guished and the gases dispersed. Furthermore, analysis of soot or even gas samples for trace PICs at the levels typically produced requires very sophisticated analytical instrumentation and highly trained specialists. Even then, complete identification of PICs in a par ticular sample may take years of work. As the level of sensitivity of analytical instruments increases, even more toxic PICs may become evident. For Instance, continued research on PCS combustion by-products has Identified not only chlorlneted dioxins and dibeniofurans, out alto polychlorinated chrysenes, pyrenes, xenthenes, terphenyls, quaterphenyls, and others (113). This level of reseerch has been directed et PCS because of its negative publicity and the perceived crises In regaros to several PCS Incloents. fo such crises exist for the proposed PCS substitutes. As a result, with a very few exceptions, compareble reseerch has not been conducted into their PICs. As noted above, existing data tend to focus on the macro combustion products present in the vapor state rather than trace PICs in the soot. Consequently, there is relatively littia data on trace PICs comparable to that on dioxins or dibenzofurans. These types of by-products simply have not been identified and studied in most cases. A major fraction of the available data was not generated with a view toward util ity dielectric applications. Experimental conditions were not designed to match a transformer failure or similar occurrence. The fact that a particular dielec tric can, in a laboratory, pyrolyze or coeoust to a dioxin or dlbenzofuran does not, in Itself, indicate that a similar reaction will occur in an actual Inci dent. The experimental temperatures to which the dielectric was exposao and th residence time at those temperatures, must be compared to the conditions expected to occur during normal operation of a transformer or capacitor and during cata strophic failure of such a unit. HONS 216001 2-2 Et Is not the purpose of this report to specuUte on potential PIC*. Rather, the objective is to report the state-of-tne-art. In some cases, where there Is lit tle comoustlon beta on a particular compound, there are more data on a structur ally analogous compound. Such data are included as guide!ines - for -any .future research, although even ,ry closely analogous compounds may not behave identi cally. TOXICITY AM) EXPOSURE LIMITS ' There is no tingle, universally accepted standard for expressing chemical toxi cides; rather, there Is a series of terms depending on the nature of the data and how It was obtained. Lethal dote, lethal concentration, toxic dose, and threshold limit value are examples of the types of toxicity data available. Fur thermore, within each type there may be variations to describe the lowest re corded value or the value at which a certain percent of a test population died or showed adverse reactions. The data may oe expressed in terms of liquid concen tration, atmospheric concentration, or concentration per unit body weight, often with a time factor included, ..Finally, the data are generally IdenttfJed by the types of organism from which they were obtained, either human, rat, mouse, guinea pig, or other animal. These different types of data are frequently unrelstable, High oral toxicity, for instance, may or may not accompany an equally nigh Inhalation or dermal toxi city. Toxic dose, letnal dose, and threshold limit value express three different types of data which, even within the same animal species, may be unrelated for a particular cnemical. Finally, due to morphological differences, extrapolation from animal toxicitles to human may not be possible (.22, 63). Unfortunately, every particular type of data is not available for every chemical. This can make toxicity assessment and comparison a difficult task. There is no way to take an oral TC^_^ for rats and a dermal LOjq for guinea pigs, and project an Inhalation TLV for humans. Consequently, it is necessary to establish a few guidelines to the application of the data to be presented later In this report. In general, this report tries to present Inhalation or dermal toxicity data for humans, wherever such were available. As they frequently are not available. Inhalation data for rats, oral date for humans or rats, or Inhalation or oral data for other animals were used, in the aoove order of preference. Some attempt was made to qualitatively assess the data. HONS 216002 2-3 It It also recogniied that discrete toxicity data values presented in isolation prooaoly mean little to most readers. Sax (95J attempted to describe toxicity data as fol lows: -. a High toxicity: LDgg < SQO mg/kg. e Moderate toxicity: LDj0 SOO to 7,500 mg/kg. e Low toxicity: LOjq 7,500 to 15.000 mg/kg. e fo toxfcity: LDS0 > 15,000 mg/kg. These are general descriptions and are not to be relied upon dogmatically. For Instance, synergistic effects (to be discussed below) are overlooked entirely. Table Z>1 presents toxicity data for several common or well-known toxicants. Data comparable to that presented elsewhere in this report are presented in order to provide a general basis for comparison with PCS substitute by-products. SYNERGISM The tonicity data, as presented In this report, are specific for Individual com pounds. In a fire or catastrophic failure situation, a variety of compounds will be produced. Depending on the nature of the toxic effects, a mixture of two equally toxic compounds may be twice as toxic as either compound alone. On the other hand, if the action of the two compounds is sufficiently different, there may be little evidence of this "additive" effect. Furthermore, there is a poten tial for synergistic effects whereby a mixture of two toxic chemicals is far more toxic than the sum of the individual toxicitles. These questions generally have not been studied in regard to PCS substitute thermal byproducts; consequently, little data relating to coot) In ad toxicitles or synergisms are presented. The potential for such should not be overlooked, however. HONS 216003 2-* i Table 2-1 TOXIC ITT AMI EXPUSUMI L WITS DATA OF COMUh TOXICANTS (83) EtHanoi ffydroyen Cyanide OUT LC!>0 (ihl-rat) 2D, 000 ppa/lor 484 ppm/ SN Hydrogen Sul t ide Carbon Monoxide 444 ppa 1807 ppa/ 4H 2,3,7,8-telrachlorodibenzo- p-dioxin Benzene 10,000 ppa/7H lD50 (orl-rat) 113 aq/kg 22. 6 ug/kg 4804 a9/k9 LOS0 (orl-aus) 7800 ug/kg 136 9/*9 114 ug/kg TDLU (orl-boa) 1430 ug/kg 130 9/* 9 l0lO (orl-Hua) 2000 ag/kg 570 ug/kg TCLO lCL0 TlV-TUA STEL 120 ag/a3 IK luoo PI* 10 ppa (TlV-Cl) 1 ag/xi3 j ay/a3 650 ppa/ 45M 600 ppa/ 30H* 4U0U ppa/ 3UH 20 ppa 50 ppa 400 ppa 100 ppa 20.000 ppa/5H ' 10 ppa 25 ppa See Glossary for definitions of abbreviations (Apperdix A). *H - (burs. N * Hinotes. HONS 216004 Section 3 PROPOSED PCS SUBSTITUTES: UQUlUS [KTSODUCTION The liquid PCB suostitutes cm oe divided Into the following groups Desea on their chemicel structures: Esters: Phthelic ecid esters (d 1 -Z-ethylhexyl phthelete end dloctylphthelete). --Beniyl neoceprete. e Slloxenes: -Polydimethylsiloxene. e Aroeutlcs: --Chlorobenzenes (mono-, trl-, end tetre-chlorooenzene). --Butyleted monochlorodlpnenylether. --Methylited dlphenylethene end dlxylyletnene. -- Isopropylblphenyl, PropylDlpnenyl. e Perefflnic hydroceroons: --Poly (1-octene). --RTEmp. e Perchloroethylene. ESTERS In generel, pyrolysis of esters unoer high tnermel end disruptive cmrge condi tions will produce geses simller to those opt lined from the hydrocerbons. The presence of oxygen in the molecule results in the formetion of higher levels of cerbon oxide geses. The products of disruptive discherge end the volume of ges 3-1 HONS 216005 produced par joule of energy dissipated are reported to be less than those pro* duced from mineral oils, askarels, and silicones. The deposit formed is also less carbonaceous, and thus causes less switch contact fouling. The product's-ofdecomposition are reported to be no more toxic than those'produced by hydrocar bons (119). The structure of the ester molecule greatly Influences thermal sta bility. The stability Is greatly Improved by the absence of hydrogen atoms on the carbon In the 0-position to the nyproxyl and carbonyl group of the alcohol and acid components, respectively. Phtnallc Acid Esters General. 01-{2-ethylhexyl )-phthal ate (Figure 3-1) was marketed In the United States by General Electric under tne trade name of Dlelektrol II, which is a mix ture of dl-(2-athylhexyl)-pntnalate, 1,2,4-trlchlorobenzene, and small amounts of 1,2,3-trlchlorobenzane. It Is also sold in Japan under the trade name Selectrol I, Tr I chlorobenzene was added to raise the corona extinction voltage (35). Chemical and physical properties of these two compounds are shown In Table 3-1. t 1* Table 3-1 CHEMICAL Ai PHYSICAL CHARACTERISTICS OF SELECTED PHTHALATE ESTERS (97) Molecular watgnt Melting Point Soiling Point Flash Point Ignition Temperatura Specific Brevity Vapor Pressure Viscosity D1-(2-ethylhexyl )phthalate 390.6 -55 *C (-67 *F) 231 *C (4*8 'F) it S mm 215 *C (420 *F) 390 *C (735 *F) 1.32 mm 14 (200 *C) 81 cps (20 *C) dioctylphthalate 390.6 -25 *C (-13 *F) 220 *C (428 *F) it 4 an h) 0,978 <0.2 mm 1^ Therm*! Decomposition of Phtnallc Esters. Phthalate esters are most widely used at plasticizers in various plastics such as polyvinyl chloride. Consequently, most of the detailed decomposition data relates to decomposition In plastics under high-temperature processing conditions with a view toward the Impact on tne stability of tha plastic. The degree to which these data apply to neat phthalate esters or phthalate ester/cnlorobenzene mixtures Is unknown, HONS 216006 3-2 3-3 HONS 216007 Table 3-Z presents some of the degradation products of dl-(Z-etnylnexyl)-pnthal. etc aru dloctylphtnalate end the effects of temperature on tne therms! decomposi tion of these esters. It should be noted thit tnese are not trace .P.|s*_ The smallest weight percent shown is 0.1, whlcn is equivalent to _ 1000 ppm. Fo attempt was made to identify any PICs present in the ppb to few-ppm range. This same type of analytical procedure, used on askarels, would fail to detect dioxins or dibenzofurans. to data were found on the decomposition of either ester at higher temperatures. The thermal decomposition of these compounds Is believed to proceed via a clselimination that takes place according to a two-step mechanism. In the first step, a six-member intermediate is formed from tne abstraction of the C-carbon in the side chain by the carbonyl oxygen. The second step, which is regarded as therate^etermlning step. Involves caroon-oxygen cleavage from the intermediate with a ii-carbon radical (47). A (CN|JjCHj CM|wCHj ICH,|,.C, C-tHf-CH) (3-1) Phthilic anhydride, a major decomposition product of Dotn ci-(Z-ethylhexyl). phthalate and o-dloctylphtnalate, can pyrolyze to produce tne highly reactive intermediate benzyne (or dehydrobenzene). t (3-2) In the presence of an excess of benzene, phthalic anhydride reacts to form naph thalene and biphenyl (). The major pathway to biphenyl is probably by overall Insertion of benzyne into a C-h bond of benzene; and naphthalene Is formed by either 2*2 or 2*4 cyclo addition of benzyne to benzene, followed by elimination of acetylene. HONS 216008 3-4 I Table 3-2 EFfECT If REACT ION TEMPERATURE ON THE IKRHAL OECCHPOS 11 ION . if SELECTED PHTHALATE ESTEHS (87. 88) OI-(2-ETHTHtf XU)-PHTHALATE _____________ Composition of Decomposition Products, wtl, at: Component 250'C 280*C 300'C 360*C 380*C 4U0*C 420*0 450*C 480*C 500*C 2-ethyl -1-hexene 2-ethyIhexanal 2-ethyl-1-hexanol phthalic anhydride Methyl, 2-ethyIhexylphthalate* Oi-(2-ethyIhexyl)-phthalate 0.1 0.4 - O.S 0.3 98.; 0.1 0.3 - 0.2 0.1 99.3 0.2 a6 0.1 0.2 0.1 99.2 l.S 0. 7 0.4 0.2 1.7 95.8 3.3 a9 0.6 0.8 4.2 90.2 8.9 0.3 3.2 4.2 7.1 76. 3 19, S 0.2 2.2 7.5 8.1 55.5 41.9 0.1 18.5 7.4 7.2 24.8 SI.O 0.2 35.5 10.5 9.6 6. 7 52.0 U. 1 31.4 18.1 U.2 0. 1 DIOCITt PHTHALATE Component l-octene Oct anal l-octanol Phthalic anhydride Methyl, octylphthalate* pimethyIphthalate* OioctyIphthalate Commsition of Decomposition Products, wtl . at: 250"C 280*0 300'C 320*C 340*C 350'C 360*C 380*C 400*C 420*C 450'C 480'C 500*C 0.4 - 0.4 ai 0.3 - 99.8 0.3 - 0.5 0. 1 U.3 0.1 9b. 7 0.2 - 0.4 0.1 0.5 U.l 98.6 0.5 0.1 0.4 - 0.4 0.1 98.4 1.7 0.4 as 2.5 ai 94.7 2.1 - 0.5 a2 1.3 ai 95.7 2.0 0.4 0.6 - 3.2 0.1 93.7 5.2 ai 1.0 as 4.0 0. 1 89.1 8.4 ai 3.4 3.0 5.0 ai 79.9 20. 3 - 10.0 6.3 9.4 a4 53.6 43.9 :- 22.1 4.3 'ao 0.6 2U: 4 / 0. 4 - 24.5 4.5 8.9 0.2 1.4 5/. 2 , 22.6 8.8 9.8 0. 1 1.6 Components of low heat stability were converted to methyl esters for analysis. The actual decomposition products were 2-ethylhexyIphthalate (methyl, 2-ethylhexylphthaIate), octyl phthalate (methyl, octylphthalate), and phthalic acid (dimethyl phthalate). HONS 216009 sine* tn* oleleetrle fluids Olelectrol [I and Seleetrol I ilso contain substan* till mounts of chlorobenzenes, It Is eonc*1vibl* that tnts* reactions may occur, possibly producing PC6 ind chlorinated naphthalene. Decomposition ttsts on PVC plastics containing phthallc acid *st*rs hive yielded basically tn* same products, ilthough som* *r chlorinated, Apparently 1CI, pro* duced from tn* PVC, r*ets with som* of the estar decomposition products to give chlorlmtad sp*c1*s (B8). Th* sun* typa of Cl* audition ruction Is' conceivable In th* pr*s*nc* of chlorobenzenes. Toxicity and Exposure Limits of OecomposItlon Products, Toxicity and axposure limits data for d1-(Z*ethylh*xyl )-pnthalate, dloctylphthalate, and th*1r Idcntlf 1 d PICs ar* shown In Table 3*3. This tabl* dots not Includ* th* cniorlnatad compounds nantionad abov*, as non* of these PICs has actually b**n Identified. In genaral, the Identified PICs appear to hav* th* same d*gr* of low to moderate toxicity as thalr precursors. tow*v*r, an assessment of th* toxlcltles Is olffl* cult, as all of th* available data relate to animals, and, In most cases. Inhala tion data ar* not provided. It Is still worth noting th* low TLV for phthallc anhyorld* (TLV-THA 1 ppm). Phthallc annyorld* is a pyrolysis product of both Z-(d1*thy1hexyl )-phthalat* and oloetylphthaiata, and Is formed In considerable quantities at th* higher temperatures. HONS 216010 3-6 I Table 3-3 TOXICITY AN) EXPOSURE UNITS OF 0l-(2-ETHTlxn)-PMlHALATE. O-UIOCTYlPHimiATE, AN) TKlK Pits (83)* LCLq (thl -rat) LOw (thl -ns) di-(2-ethylhexylphthalate dioctylphthal ate phtltalic anhydride 2-ethyl*l-he*ene 4000 pp*/4K 100 *9/119 2-ethyl-l-hexanol 1-octanol phthalic acid ICjy (thl-ms) tolo (orI -man) M3 *q/ig SO (or 1-rat) (or I ill 31 *g/kg SOOU *9/*' 6SI3 wg/kg 4020 *g/kg 2 g*/kg 2460 *g/kg 1790 *y/kg 7900 *g/kg ILV-IWA S ng/*3 1 ppa SIEl 10 *g/*^ pp* *Md dU were jvjildble for ptttluNtc, benefit, octyl phtNltet or l-octe* / HONS 216011 Benrylneocaprata General. There Is little published information available on benzylneocaprate. It nit been marketed by Asei of Sweden is Faradol IDO, but the manufacturers hive recently discontinued production. The only diti founo on the chemical ind physl- cil chiricterlstlcs of this compound ire Its moleculir weight (262) end vipor pressure (0.D003 mm Mg it Z5 *C). - Therein Oecomposltlpn of Benzylneociprite. to fnformitlon wis ivillible on the thermil degrioetlon of benzyl neociprite. Benzylneocaprate would be expected to be thermilly very stale. The cls-ellmlmtlon ruction with the slx-memoered trinsltlon ring, by which the phthallc esters decompose, would not be expected to occur. The lick of hydrogen itoms ivilUbl* for eostrictlon from the 3-carbon results In "hindered" 3-carbon. The degred it Ion of such esters presumably fol lows a free redlcal mechanism which requires high energy levels, and therefore high thermal degradation temperatures (11B). SIIOXANES General The principal slloxane used as i transformer Insulating liquid Is polydimethylslloxane (POMS), PONS is produced commercially by Dow Corning as DC561 and 0C200, General Electric as SF97, Union Carbide as L30S, and SWS Silicone as P101 and F10. Table 3-4 lists the Important properties of a typical POMS transformer fluid. HONS 216012 3-8 Tib 1 e 3-4 chemical and physical c haractew 1ST ICS of a typical pdms tkansformer fluid (li?) .Molecular weight Bolling Point Fire Point Flasn Point Specific Gravity 2826 250 *C 360 *C 300 *C - 0.961 at 15 *C Thermal Decomposition of PQHS POMS begins to thermally decompose at 200 *C. The thermal degradation occurs rapidly at 300 *C with a breaking of the S1-0 bond and formation of lower mole cular weight compounds such as trl- and tetracyclic slloxanes <D3 and 04. respec tively) (Figure 3-2) (119). These are volatile, low flash point materials that boll at 150 to 170 *C. Their presence In excess of 2.51 reduces the flash point of the silicone liquid to that of a similar viscosity mineral oil. In addition to the cross-linked volatiles, both gaseous and solid (l.e., amorphous silica) decomposition products are found. Silicone transformer liquids, after being aged for 4 hours at 350 *C, have been found to generate several cyclic slloxanes, notably O3, D4, 05, and Dg (in mat order, quantitatively) (110). These same compounds can b* produced at lower tem peratures if the thermal aging process Is extended, for Instance, 5 months at 240 *C. with Increasing time, the rate of formation of D4 begins to exceed that of D3 at the lower temperatures. in one experiment, POMS was pyrolyied ovar me temperature range of 600 to 1200 *C; a rapid temperature rise (60 *C/msec) and a residence time of 20 sec were used (61.). The following decomposition products and tendencies were observed: e 600 to 800 *C - Predominantly cyelle slloxanes arising from slloxana rearrangement. The relative amount of cyclic slloxanes decreased regularly wltn Increasing molecular weight, with me exception of a greater abundance of 06 compared to D$ at temperatures below 800 *C. 800 to 900 *C - Compared to the lowest temperature range tested, total volatiles Increased and tne relative fraction of nign-vol atl 1 Uy material Increased. Metnane ana nyarogen were tne major compo nents of tne hlgh-voletl 1 Uy fraction; C? nydrocaroons (CjHj, C2H4. Cjhg) nava also baen Identified. 3-9 HONS 216013 CN,''' I -CH, cn/ nCn, MtuMuricrtittrtiiitiint [0]) CN) CN) w 0 CN, / 0 ; - V CN, )t < CN) \ / CN, 00 s Jf' t\ CN, CM, QcUlNjrf<)i It un* (0> 10, tltrwfti 0, lt MUttlt) OKtMtNrlC)tlH*NUttUM (Oj> CN, 1 HjC-- SI -- N I TrtattNrlliUw CN, CN, CN, II N,c-*1 - CNj --n -- 0 -- It--CN, 1 CN, CN, CN, Fl9r 3-2. Typtcil PDMS Pyrolysis Products MOHS 216014 3-10  900 to 1200 *C - A number of trece component* ippeir it 900 *C end become more abundant et higher temperatures. The Identified trece compourts Include tnoie mown in Figure 3-2. They ip peered to irlte vla'C-Hind SI-C bond claavage, and contain *1 lylhyonae, sllmethylene, art monomethyl slloxene itructure*. a 1200 *C - At thl* temperature, the 1 gtn-y o 1 a til 1 ty fraction accounted for tne majority of the volatile*. In tha absence of oxygen art under arcing conditions, a variety of hydrocarbons can be produced. Including methane, acetylene, ethylene, ethane, and propylene (21. 61. 112). Tnermal oxidation of POMS leads to the formation of paraformalde hyde and formic acid, as well as carbon oxides (77). Thermal degradation Is catalyzed by trace Impurities of materials such as water, metals, and Ionic compounds. The manufacture of the slloxanes, via an organic chloride with tha production of hydrogen chloride. Introduces the possibility of chlorine contamination, although chlorine concentrations probably do not exceed 1 ppm (119). Toxicity art Exposure Limit* of POMS Decomposition Products There are no available data on the toxlcltles of the various cyclic and cross1 Inxad organotl 1 oxanas (Figure 3-2) formed during POMS decomposition. Of tha organic compounds produced, only formaldehyde and formic Kid are particularly toxic, and neither Is considered highly toxic (formaldehyde L0^ (rat-orl) - 800 mg/kg; formic Kid LDjq (rat-orl) 1100 mg/kg), Both are Irritants, and formic acid can be dangerously caustic to the skin (69), Both are liquids at ambient temperatures, art may therefore be found In the soot as contaminants on surfaces following a flra. The silica produced Is amorphous. Crystalline silica is a known cause of sili cosis, art Is considered to be far more toxic than tha amorphous form. The TLV for crystalline silica (crlstobalIte, 0.0S mg/m3) Is several orders of magnitude lower than that of amorphous silica (precipitated silica, 5 mg/m3) for resplraole dust. Silicon carblda, formed under arcing conditions, is considered to be a mild Irritant when Inhalad. Tha toxicity of the pyrolysis gas of POMS (Oow Corning 561 silicon* transformer liquid) was evaluated using a method developed by Hlledo, where rats ara sub jected to Inhalation of tha pyrolyzad compound (44), Tha toxicity of POMS was 3-11 MONS 216015 I Table 3-5 CKNICAL AN) PHTSICAL PROPERTIES OF SELECTED CM.OR1MIED BENZENES (97) Molecular Height Melting Point Boiling Point Flash Point Vapor Pressure Specific Gravity Mono Chi oro 112.56 -65. 2 *C '132 *C 29 *C 8. 8 mo N) 8 20 *C 1. 1 1,2,3 Trfchloro 1.2.4 1.3.5 181.45 1181.45 181. 45 16.95 'C 63. 5 *C 52.6 *C 213 *C 208.5 *C 221 *C III) *C 107 *C 113 *C 1 MO It! 10 m H) 8 38.4 *C 8 78 *C 1 -- Hg 8 40 *C 1.446 N1 1.69 1,2.3,4 215.9 47.5 *C 254 *C Nl* Ml Nl Tetrachloro 1.2.3.5 215.9 54.5 *C 246 *C Nl N) Nl 1.2,4,5 215.9 140 *C 243 eC 155 *C <a 1 mo 1 8 25 *C 1.858 *Nl * No information. 3-12 HONS 216016 compared t0 thit ot *noth*r transformer liquid. * paraffinic hydrocarbon. under the specific test conditions, the toxicity of the pyrolysis 91s from POMS wes found to oe less then that of the pereffinic hyorocaroon. The silicone tnns- former liquid wes among the leist toxic of over 270 suOsttnces tested by HI lido under the sime conditions. . _. aromatics rhlnrobenienes . fteneral. Chlorobenzanes generally were not used independently as dielectric fluids. Rather, they were mixed with other fluids (e.g., d1-(Z-ethylhexyl). phthalite or PCBs) to lower pour points. The principal chlorobenzenes used are the trl- and tetri- congeners. These compounds typically make up about 301 of a dielectric fluid mixture. Figure 3-3 snows the different trl- and tetrachlorobenzene isomers possible. Monocnlorobenzene his also been considered for use In dielectric fluids. Table 3-5 lists some of the chemical and physical properties of these chlorinated benzene compounds. Thermal Decomposition of Chlorobenzenes. The chlorobenzenes are relatively ther mally stable compounds. Table 3-6 lists the temperatures for the onset of ther mal decomposition and 99% decomposltlon for some of the compounds. Table 3-6 CH.0R0BENZENE DECOMPOSITION TEMPERATURES Monocnlorobenzene 1,2,3-trichlorobenzene 1,2,4-trlchlorobenzene 1,2,3,4-tetrachlorooenzen* 1,2,3,5-tetrtchlorobenzene Onset *C 540 645 640 660 655 99% Decomposition *c 710 780 750 BOO 760 As with other chlorinated organic compounds, tht major combuStlon/pyrolysts prod ucts are HC1, hydrogen and metnane (no air), and carbon oxides (In air). How ever, consldereble research his been focused on the trace PICs produced* 3-13 HONS 21601? Flours 3-3. Chlorlnittd 6tnz*nt Compounds Usd In 01 loot Me Fluids HONS 216018 3-14 Monocnlorooenzene, pyrolyzed it 690 *C for 24 sec, can produce cMorobi phenyl s, dichlorobiohenyls, naphthalene, ino chloronapnthalenes At 770 to 800 *c ena 86D to 830 -*C , the following reaction his been ooserved: <1 * J * CM, Ullf MrMucit (3'6> Between BSD ana 88U *C, yields of HCl, hydrogen, methane, and chlorinated olpnenyl decrease, but considerable quantities of vinyl chloride (CHj CfCl) are pro duced. Pyrolysis of chlorobenzenes in the presence of air yields chlorophenols which can, in turn, dimerize to PCDFs and PCDDs (J7, V, J_13). The degree of chlori nation in the chlorophenols tends to be the same or higher than that of the par ent chlorobenzene (16), In other words, monocnlorobenzene will produce mono-, di-, or trlcnloropnenols; trlchlorobenzene will produce trl-, tetra-, and pentachloropnenols. Buser (J_6) pyrolyzed trl-, tetra-, and pentacnlorobenzenes at 620 *C in sealed quartz mini -ampul es in the presence of air. The pyrolysis gases Included higher chlorinated benzenes, chlorophenols, PCOFs, PCDDs, and. In tone cases, polychlo rinated naphthalenes, polychlorinated styrenes, and small quantities of PC8s. Polychlorinated diphenyl ethers and polychlorinated bipnenylenes were identified in other experiments (17, 113). Table 3-7 lists the Quantities of PCOFs and PCDOs produced. The known toxic com pounds .(those with chlorine atoms in the 2, 3, 7, and 6 positions) were identi fied, but these were not major components. In general, tne degree of chlorina tion in tne PCOD/PCOF product seems to follow the pattern below. rcor (3-7) w Wif 1 * f I Im 3- IS (3-8) HONS 216019 Following this pittern, i monocniorobenzene would not generilly product hlgner fun t d ichlorodlbenzodloxln or dlchlorodlbenzofurtn, Trt- or tetrtchlorobenzenes could 'product up to hexl- tnd octt* Isomers, respectively (set Ttble 3*7). Figure 3** tunmtrlzes the vtrlous chlorobenzene rttctlons discussed tbove. -- Tib It 3-7 FORMATION OF PCOFs Al PCOOs FROM THE PYROLYSIS OF CHLOROBENZENES AT 630 *C (16) Pi rent Compound Trl chlorobenzene Tetrtchlorobenzene Tetri2000 <10 Trichlorobenzene Tetrtchlorobenzene ISO <10 PenttSSOO 2S PCOFs (ppm) HexiZ 7 SO 800 Heptt2S0 2250 PCODs (ppm) 100 <2S <25 26 700 800 Octt<2S 1000 <25 ISO Toxicity of Chlorobenzene Otcoaposltlon Products. Mifiy of tht combustlon/pyrolysls products of tht chloroOtnztnes trt very highly toxic (Ttblt 3*8). It Is gtn* trilly believed thit, on t molecultr bills, 3,3,7,8*tttrich)orod1btnzod1ox1n Is tht most toxic synthetic molecule known to men. Although tn Identified chloro benzene PIC, 2,3,7,8-TCOO bis not been found except In trice quintltles. How. ever, the other PICs, while less toxic then 2,3,7,8-TCDO, ere it so very toxic. In gthtril, toxicity teem to be relited to chlorine content ind position. Chlorlntted trowttlcs with three or fewer chlorine ttoms tend to be slgnlflctntly less toxic then higher chlorlntted congeners. Thus, PICs from monochlorobenzene would probebly be less toxic thin those from tr 1 - or tetrtchlorobenzene. How ever, they ere still moderttely to highly toxic ts discrete compounds. The com bined toxic effects could be synergistic end gretter, or possibly tntegonlstlc ino less. A further problem with these PICs Is thit, with the exception of vinyl chloride, ill ire solids or liquids it normtl imblent temperttures. -Thus, they tend to be found es surflclel contwilntnts with toot In the efteratth of t fire. The soot provides t reservoir for continued exposure unless completely removed. 3*16 MONS 216020 Figure 3-*. Typical Chlorobenzene Thermal Reaction* HONS 216021 3-17 Tabic 3-8 ' TOXIC IT 1ES Al EXPOSURE L IK ITS OF CHLOROBENZENE CBUST I ON/PYROLYSIS PRODUCTS (83) Chloropntnoit m-chloroph*nol 0-chloropn*nol p-ch1orophnol 2,4^1 Ichlorppnanol 2,641ch1orophenol 2.4.5-tr1ch1oroph*no1 2.4.6-trlchlorophtnol Z,3,4,6-t atrichloropncno1 NiphthiUnts Nipnthalpn* 1-chloronaphthtlene 2-chloronipnthtlene tr1 chi oroniphth al on* ttcrachlorsnaputhiisns pontichloronophthtI on* o ct ac h 1 oron apnt halana Dloxlm 2.3.7-trlchlorodlbanzodloxin 2,3,7,6-tetrachlorodlbanztfloxln 1.2.3.6.7.8-haxacfllorod Ibanzod 1 oxl n 1.2.3.4.7.8-haxachlorodlbanzodloxln Hlxturo of 1,2,3,7,8,9- and 1.2.3.6.7.8-haxaehlorodlbanzodloxlns L0SU. orl-rtt S70 mg/kg 670 mg/kg 261 mg/kg S80 mg/xg 2940 mg/kg 6Z0 eg/kg 820 mg/kg 140 mg/xg 1780 mg/kg 1540 mg/kg 2078 ng/kg TLV-TWA 10 ppm TIV-TWA 5 mg/m3 TLV.TWA 2 ma/m3 TLV.TWA 500 ug/m3 TLV-TMA 100 ug/m3 29.4 mg/xg (LD50, orl-gpg) 22.5 ug/kg 0. 5 ug/kg (LOso, orl-gpg) 1250 ug/kg (L050, or I-mu*) 825 ug/kg (LOjq, orl-mut) 800 ug/kg 2,3,7,8-tatrachlorodlbanzofuran 5 ug/kg (C050, orl-gpg) Vinyl chloride 500 mg/ kg TLV-TMA S ppm HONS Z16022 3-18 Methylated D1 phenylethane and ^enyIxylylethane General. Thp chemically similar methylated dlpnenyI ethane (1,1-dixyiylecmne) ana phenyl xylylethane (see below) are used is components In dielectric fluid mix tures. Ed 1 so 1 II, (iwnuflctured by McGraw Edison, is i mixture, of the 1,1-bis (3,4^1methylphenyl) ethane ind propyldlphenyl. _ Phenyl ^yJyTethane is found in Olelestrol III (Gentril Electric), Selectrol ll (Sangtmo), ino SA5-29S ino SAS- 296 (Mssetl Compmy; mainly the 1-phenyl-l-( 3.4-xyly 1)- and l-phenyl-l-(2,J- xylyl)- Isomers). >, . 11 .k <i ty'MeiM* ******* Very little Informition is evailible on the pnysical ind chemical properties of these compounds (see Table 3*9). Table 3-9 PHYSICAL AM) CHEMICAL PROPERTIES OF SEIECTEO DIARYL ALKANES Molecular Weight Bolling Point Flash Point Specific Gravity Vapor Pressure Mwthylatari 01 phenyl ethane 238 316 *C 1.62 *C 0,97 210 0.0006 mm Hg at 25 *C Thermal OecomoosltIon Products of 01 xylylethane and Phenylxylylethane. No Infor mation was found on the combustion/pyrolysis products of either of these com pounds. The thermal decomposition of chemically similar compounds, diphenylmethane and dlphenylethane (bibenzyl), has been studied, however (_) 3-19 i.i>oi marietta** HONS 216023 Pyrolysis of met* compound* at 700 *C nas yielded benzene, toluene, ethylben zene, styrene, stllbene, pnenanthrene, and fluorene, If dixylylethane and one* nylxylylethane follow the tame general decomposition mechanisms, then products could Include benzene, toluene, xylene, trir>ethylbenzehe, ethylbenzene, etnyixylene, methylated fluorene, metnylatea phenanthrene, methylated Styrene, ana so forth, Coitoustlon in the presence of lir may produce phenols, cresots, ana other oxygen-containing species. It Is possible that methylated dioxins or dlbenzofurans could be formed. All of this is supposition, however. Utthout further study, It is Impossible to determine with certainty what trace PICs might oe formed. Toxicity of Dlxylylethane and Phenyl xylylethane Decomposition Products. Since the specific products cannot be Identified with certainty, toxlcities cannot oe determined, tone of the possible PICs discussed above Is considered to be highly toxic, although several are known or suspected carcinogens. Most are liquids or solids at normal ambient temperatures, and could be expected to deposit on nearby surfaces with soot and other conventional fire byproducts, Butylatmd Honochlorodipnenylether Qenmr |1 4 Rijtu lafaj mnnnr h 1 Qrnrt 1 nh*nu 1 h*r It arbf^gr) hv [)nu ThMiral Tnann^ny under the trade name of Dielectric Fluid C-4 (formerly XFS-41691). The name spe cifies neither the extent of butylatlon nor the Isomeric form of the compound. Among the isomeric forms which have been found In dielectric fluids are: m+wtri The major components of one sample of XFS-4169L were (wt %): 4-ch I orooi phenyl ether (19,6); 2-sec-butyl -4' -chi orodl phenyl ether (27.2); 4-sec-buty) -4' -chlorodiphenylether (24.0); and small amounts of dl* and trl -butylated chi orooi phenyl ethers (2). Physical and chemical property data were not available on Individual 1 somer s. Thermal Decomposition of Butylated Monochlprod iphenyl ether. There Is no Informa tion available on the thermal Decomposition products of butylated monochloroatphenylether. towever, both the unsubstituted diphenylether and polyeniorlnated diphenyl ethers (PCDPEs) have been studied. HONS 216024 3*20 Vipors of dlphenylether passed through a glass tube at red heat have oeen shown to break flown into phenol, benzene, and dlbenzofuran (46). At 600 *C, the thermal conversion of PCDPEs into PCDFs is of tne,same order of magnitude as that from PCBs Lindahl et ali (_|9) conducted a series or tests with tri- to octachlorodlpheny lathers at (00 *C, and concluded that PCDFs form primarily via loss of ortho-Hj or ortho-HCI with a minor fraction formed by loss of ortho-Clj, A few PCDDs were also observed, formed presumably via loss of orthO'Clg. The PCDDs formed showed lower cnlorlnation than the PCDPEs, Figure 3-S shows these reactions, If butylatad monochlorodl phenyl ethers follow the same mechanisms, the following reactions could result. 1K41 (3-11) Assuming that the chlorine is always m tne 4' position, no reaction will occur via C1 j toss. This Is Important as Lindahl et at. (59) observed dioxin formation only vie Clj loss. Thus, butylated-4' -chloroolpnenylethers would not bt expected to degrade to dioxins. Also Important Is the fact that the PCDPE to PCDF reaction Is monomotecular. In contrast, the PCB to PCDF reaction Is blmotecular, requiring PCB and oxygen mole cules, while the chlorobenzene to PCDF reaction Is trtmolecular, requiring two cnlorobenzenes and oxygen. These multlmolecular reactions can be significantly less frequent In oxygen-starved or dilute systems. Being monomolecu 1 ar, neither oxygen content nor PCDPE concentration will affect the rate of PCDPE conversion to PCDF. '' 3-21 HONS 216025 *Marks tho ortho- positions. Mguro 3-5. Formotlon of PCDFs/PCDDs from PCDPEs (31BJ HONS 216026 3*22 No other PiCs h*ve been Identified or suggested for butyleted monochloroalpneny 1 ether. Toxicity of Butylatad Mpnochlorodlphenylether Oecowpotltlon Products. As noted ebove,' there Is no definite Information on the combuttlon/pyrolysls products of butyl eted monoenlorodi phenyl ether. Thus, no toxicity Information^ 1* available. Assuming thet the reectlons discussed ebove ere possible, .then reectlon products could Include monoehlorodlbemofuren (from the 4-chlorodlphenylether) end buty- leted (mono-, dl-, end trl-) monoehlorodlbemofuren. There Is no information available on the toxlcltles of these compounds. - Alkyl Biphenyls Generel. Isopropylbiphenyl Is manufactured by Uestlnghouse under the trede name of Memcol. Uencol Is described by the manufacturer es being predomlnently 4-isopropylbiphenyl with consldereble amounts of the 3* isomer end traces of more highly substituted biphenyls. towever, an Independent analysis (107) revealed the following composition (wt X): 3-1 sopropylbiphenyl (60.31); 4-lsopropylbiphenyl (38.41); 3,5-dl isopropylbi phenyl (0.31); S.S'-dllsopropylblphenyl (0.41); 3,4'-d1lsopropylb1phenyl (0.21); end 4,4' -dl 1 sopropyl biphenyl (0.11). ItttrMrWitMv1 n-propylblphenyt is mixed with dixylylethene in d 1 so 1 11, menufectured by HcGrew Edison. Several mono* and dipropyl Isomers ere possible, but there is no Infor mation available on which isomers ere found in Ed 1 so 1 II. Table 3-10 presents physical end chemical properties for these compounds. HONS 21602? 3-23 Tab It 3-10 physical aw) chemical properties of selected alkyl biphenyls (s?) Molecular Weight Bolling Point Flash Point Ignition Temperature Specific Gravity Itopropylblpheny 1 196 270 *C 141 *C 435 *C 1.0 n.propylblphenyl 196 260 *C >100 *C 445 *C Thermal OecompotItlon of Itopropylblpheny! and n.propyl pi pnenyl. to Information was found In the literature on the coffbustlon/pyrolysls products of either of tntst compounds. Assuming that tnty degrade by C-C bond fission llkt other alkyl aromatic compounds, tht pyrolysis products could Includt blphtnyl, propylbtnztnt, proptnt, and similar compounds. In tht prtstnct of air, phtnolt, alkyl alcohols, unsubttltuttd and alkyl substltuttd dlbenzofurans, and to forth might bt potsiblt. All of this It conjtcturt, however. Toxicity of ItopropylblPhtnvl and n-proovlblphenyl Decomposition Products. Information is available on the combustion toxicology of these compounds. to PARAFFINIC HYDROCARBONS Pplyalphaoleflnt Generalv A polyalphaolefln (PAO), or poly 1-aUene, Is a polymeric chain of alktne monomers with a tingle double bond in the 1-potltlon. "|C . (CIO. . CXj i-4ium *ljr UIUl Among tne polyalphaoleflnt marketed for use as transformer oils are PAO 13CE arm PR, manufactured by Unlroyal Chemical and Gulf Oil Chemicals, respectively. 13CE it derived from 1-octene, and nas an average molecular weight of 600 PAO HONS 216026 3-24 to mention of tnt degree of cross-linking was found. chemical and physical properties of MO 13CE. Table 3-11 presents the Table 3-11 CHEMICAL AMI PHYSICAL CHARACTER 1ST ICS-OF-PAD 13CE (117) Molecular Weight Flash Point Fire Point Specific Gravity viscosity (Centlstokes) 600 200 *C 307 *C 0.84 14 at 100 *C Thermal Decomposition of Polyalphaolefln. Information on the thermal decomposi tion products of poly (1-octene) Is very limited; no specific PICs are named. PAO is reported to undergo practically no thickening or sludge formation when exposed to temperatures of 175 *C for 100 hours (101). '' Random chain scission along the backbone chain is the dominant mecnanlsm in the pyrolysis of polyolefins. As the term Implies, scission Is a random event, and polymer molecules are first broken into large macrorad leals. There is a rapid decrease in molecular weight, and almost no monomer Is formed in the early stages (39). The expected decomposition products would Include a variety of short-, medium*, and long-chain saturated and unsaturated hydrocarbons. mainly olefins, paraffins, and cyclic hydrocarbons (.106). In the presence of air, various oxyspecles could also be expected. For instance, pyrolysis (approximately 300 to 350 *C) tests on poly (4-methyl-1- pefitene) In vacuum yielded Isobutene, propane, isobutane, 4-methyl-i-pentene, isopentane, 2,3-dlmethylbutane, ^-pentane, pro pen e, and a series of to C18 branched saturated and unsaturated hydrocaroons (4). In the presence of air, peroxides and hydroperoxides are formed Initially, leading to the formation of such products as isobutyraldehyde, acrolein, acetic acid, acetaldehyde, acetone, isobutyrlc acid. Isovaleric acid, isopropanol, ethanol, crotonalydehyde, metha nol, isobutanol, propionic adc, oxalic acid, proplonaldehyde, butyric acid, butyraldehyde, and ^-propanol. 3-25 , HONS 216029 -- CM - Cnj -- C* -- Cm t -CM 11 CNj 1 1 CHj I /\ /\ / \ CNj CHj Cj CM3 CHj CH3 ft* **(<* > Toxicity of Poly 1-Octene Decomposition Products, to Informetlon wes evelleble on the combustion toxicology of poly 1-oetene. The types of hydrocerbons men tioned ebove ere generelly of low to moderete toxicity, but the presence of more hlgnly toxic compounds cennot be ruled out without further testing. Generel. RTEmp Is * high moleculer weight, purified ptrtffln menufectured by RTE Corporetlon. It Is refined from crude oil end solvent extrected so thet there ere no erometlc or polyerometlc compounds present, RTEmp is el so cherecterlzed by Its leek of helogens end inorgenlc compounds. The exect composition Is difficult to escerteln. According to meterlel supplied by the menufecturers to the Occupetionel Sefety end Heelth Admlnlstretlon, it is listed es ning e boiling point eowe 427 *C, corresponding to en n*elkene of cneln length C^. Teble 3-12 presents chemteel end pnyslcel properties of RTEmp. Teble 3-12 CHEMICAL AW PHYSICAL CHARACTERISTICS OF RTEmp (117. 121) tolling Point Flesh Point Ignition Tempereture Specific brevity Viscosity (Centlstokes) 427 *C 264 *C 312 *C 0.67 120-140 et 40 *C; 12-14 et 100 *C Thermel Decomposition of RTEmp. The leek of knowledge of the exect cnemicel neture of the compound nes meoe it impossible to describe tne thermel oegreoetlon HONS 2L603G 3-26 products of pyroiysls/combustlon. The "fir* naiand properties" of RTEmp have been studied under simulation of e "worst case" scenario for a catastrophic transformer'failure. The major arcing product reported was hydrogen, with lesser amounts of acetylene, methane, ethylene, and other gaseous nyorocRfbons not spe cifically named (6ZJ, All of the above gases are flammable and are believed to contribute to the fireball observed in catastrophic electrical failure tests. Toxicity of RTEmp decomposition Products. Since combustion products, and parti cularly PICs, have not been identified, it is Impossible to assess their indivi dual toxlcltles. Kbwever, the toxlcltles of tne gross pyrolysis gases were eval uated using a method developed by Hi 1 ado where rats are subjected to Inhalation of the actual pyrolysis gases (_1). Exact toxlcltles were not given, but the toxicity Of the pyrolysis gases from RTEmp was compared to that of another trans former liquid, a polydlmethyIslloxene (Dow Corning $61 silicone transformer liquid). Under the specific test conditions, tne toxicity of the RTEmp pyrolysis gases was greater than that of tne silicone transformer liquid, Rerchloroethylene General. Perch)oroethylene (tetracnloroethylene; CjCle) is a nonreactive, non flammable liquid which has generally been used as a solvent. Table 3-13 lists Its chemical and physical properties. It Is manufactured as Wecosol* transformer fluid by Uestlnghouse, and may be used neat or mixed with mineral oils. Per ch loroethyIene Is also major component In a relatively new British Import, For mal If (90), This fluid is a mixture of percnloroethylene ano three fluorocar bons (111;, 112a, and 113). ci Cl ct \ Cl NrcalffMihritw MOMS 216031 3-27 Tibi* 3-13 CHEMICAL AM) PHYSICAL CHARACTERISTICS OF PERCHLOROEThTLENE Molecular Weight Melting Point Bolling Point Flesh Point Fire Point Vapor Pressure Specific Gravity 165.83 -22.* *C 121.2 C et 1 atm None None 14 mm et 20 *C 1.62 Thermal Decomposition of Perch loro ethylene. The amount of date available on the thermal decomposition of perChloroetnylene is quite limited, It Is a very stable compound due to high molecular bond dissociation energies and a lack of hydrogen atoms available for abstraction or formation of hydroxy radicals. As a result, few decomposition products would be expected. Claiborne heated perch I o methylene In varying levels of air In J stainless steel tube to 1OD0 *C, The principal products were C02 and chlorine. A small amount of HCl was found. Phosgene was not detectable either by gas chromatog raphy or mass spectrometry. Low energy (3U KW-sec Input energy for 30 min) arcing experiments generated small quantities of CO, CO2. chlorine, and HCl (60), 10 other products were reported. Decomposition tests on Formal NF-flllefl transformers produced only traces of chlorine and several chlorofluorocarbons, other than the fluid constituents (96). No HCl or phosgene were detected. The halocarbons were probably produced from the halocarbon components of the fluid mix (see Section 4, Freon 113), There are reports that perch)oroethylene may decompose to phosgene or trichloro- acettc acid in the presence of arc welding (_25). The tests cited above do not confirm this, but the experimental conditions did not simulate arc welding. It does not appear that these highly toxic compounds are generated under Simple pyrolysis or low-energy arcing conditions. * HONS 216032 3-26 Toxicity and Exposure Limits of Perchlproethylene Oeeompos i tlon Productt. Taole 3-1* Hit* the toxicides Tor various known ana suspected perenloroethy1ene aecompoi.ltion products. The most toxic of tne PICt formed ore chlorine, pnos- gene, ana trichloroacetic acid, which appear to be clearly more toxic than tne parent compound, however, with the exception of chlorine, these PICs seem to be formed in trace quantities, If at all. Furthermore,.of the PICs listed, only trichloroacetic acid Is likely to be found on surfaces In tne aftermath of a fire. Tne others are gaseous and would readily disperse. ; HONS 216033 3-29 1 Tattle 3-14 TOXIC ITT AMI CXPOSWU l iPIITi Of PEHCHLOROC THUE It AM) DECOMPOSITION PRODUCTS (M3 > lC50 ihl-han "l0 Perchloroelhylene 4000 ppm/ 4H Chlorine S/3 ppm/ 30 ain Carbon Dioxide lUX/1 ain Phosgene (Carbonyl Chloride) 3^00 ay/ Trichloro acetic Acid Hydrochloric Acid 1300 ppm/ 30 am Ifrdrogen Chloride (Gas) IOOO ppm/ 1 ain "to 9b ppm/ 7H ft Ml/ 30 ain ihl- rat LCy, lCL0 293 ppm/ IH 651/ 15 ain 50 ppm/ 30 ain 4/01 ppm/ 30 ain 3124 ppm/ IH or 1 -rat lD50 unk -man lDL0 TtV-TMA 50 ppm 0.SX 5(100 ay/kg 1 ppa SI ag/kg 5 ppa STE1 ?00 ppa l.SX 3-30 MONS 216034 Section 4 DIELECTRIC GASES CH.OR0FIUORQCARBONS General Freon is OuPont`s registered trademark for carbon compounds coot elf log fluorine, end is used only for fluorocarbons mat are nonflammable and low in toxicity. Freon 113t used as a dielectric gas, refers to l,l,2-tr1chioro-l,2,2-tr1fluoro- etnane. ' Cl -- c -c -1 II Cl Cl 111 In addition to being used individually as a dielectric gas, chlorofluorocarbon 113 Is a component of the Forme! NF mixture. This fluid contains two otner fluorocarbons, l,l,2,2-tetrachloro-l,2-difluoroethane (chlorofluorocarbon 112) and l,l,l,2-tetracftloro-2,2-difluoroethane (cnlorof 1 uorocarbon 112a). These two fluorocarbons are not discussed elsewhere In mis report, but many of the follow ing comments on Freon 113 decomposition apply to them generally. Thermal Decomposition of Freon 113 The chlorofluoroalkanes, of which Freon 113 Is a member, are thermally stable up to 230 *C (119), At higher temperatures or during arcing, the fluorocarbons can decompose. Sparking tests on 1,1,1- and 1,1,2-trichlorotrifluoroetnane produced the negative ions Cl*, F", Cl2*, and CC13~ (93_). All of these could continue-to react witn the parent compound. In this test, analyses were not carried forward to identify the final decomposition products, - HONS 216035 4-1 Partial concustion of CjCIjFj at 1000 L in air produced Six tract compounos ten tatively identified as CF^, Clj, COFj, CO, CCIFj, and COC12 (Phosgene) (_!_). In addition, the quart! reaction cnameer was noticeaoly etched, suggesting the pot- S10111 ty of If formation. '. Tests on Formel NF (which contains percnloroecnylene as well as the fluorocarbons mentioned above) yielded only Cl2 ano several unidentified halocarbons i98). no pnosgene or HCI were detected, tydro9en chloride would not generally be expected from these compounds because of the lack of hydrogen. Many fluorocarbon decomposition reactions depend on the presence of at least one hydrogen atom on the molecule. At nigh temperatures, a hydrogen halide can form, resulting in an unsaturatad helocarbon or a halocarbon radical leading to the formation of longer chain molecules and even polymerlaatlon. These are fairly common reactions, but would not be expected with Freon 113 or Formel NF, Toxicity and Exposure Limits of Freon 113 Decomposition Products Table 4-1 lists the toxlcities of the Identified Freon 113 decomposition prod ucts. Most of these products are more toxic than the parent molecule, and sev eral (phosgene, carbonyl fluoride) are hiunly toxic. .411 are generated as vapors, however, and tauld be expected to disperse with adequate ventilation. Carbonyl fluoride decomposes In the presence of water. Thus, there would be few, if any, highly toxic residues (of the Identified PICs) remaining after an inci dent. There is a potential for highly corrosive and irritating residues of hydrogen chloride and hydrogen fluoride on surfaces, but these present a low tox icity hazard. SULFUR HEXAFLUORIDE General Sulfur hexafluoride (SF^) can be used alone as a dielectric or in a mixture of gases such at perfluorocarbans. Its physical ana chemical properties are listed in Table 4-2, HONS 216036 I Title 4-1 TOXIC ITT AM) EXPOSURE LIMITS OF FREON 113 DECOMPOSITION PRODUCTS (Kl) Freon 1)3 IC10 ihl-han lC50 Cirbon Monoxide 4,000 ppa/ 30 ain (ain) Phosoene (COCIj) 3,200 mq/mr l^droyeA Chloride (Ois) 1,000 ppa/ 1 ain t^droyen Fluoride Cirbon*1 Fluoride (COFz) Telrif looroaethine (CF^) rcLo 6S0 ppa/ 45 ain () 25 ppa/ 30 am ihl-rat LC10 lc50 8/,000 ppa/ 0 ain it*I-ms lL0 25 pph/ t. 5 ain 50 ppa/ 30 ain 4.701 ppa/ 30 ain TIV-TUA stel 1,000 ppa 1,250 Ppa 50 ppa 400 PPa 0. 1 ppa 5 ppa 110 ppa/ 1 ain SO ppa/ 30 ain 360 ppa/ 1 hr 095,000 ppa/15 ain 3 ppa 2 ppa \ 6 ppa 5 ppa i i i 2^6037 HONS Tab Ie 4-2 CHEMICAL AW PHYSICAL CHARACTER 1ST ICS OF SULFUR HEXAFLUORIDE (97) Molecular Weight Melting Point Boll ing Point Speel fic Gravity Vapor Pressure 146.07 -50.8 *C -63. 5 *C 1.79 (-39 *C) (liquid ) 5.11 (20 *C) (vapor) 21 mm Hg 20 *C Thermal Decomposition of Sulfur Hexafluoride The thermal decomposition of sulfur hexafluoride has been studied under condi tions simulating arcing, arc welding, and fires (20, 35^, 4, 111). A numoar of long-lived by-products heve been identified. The principal by-product In most tests was thlonyl fluoride (SOFj). Under spark ing at 36 kJ, the typical concentration of SOFj deposited Into the test cell fol lowing 24-hour storage was 1.21 (36). Other Identified products (and their con centrations at the condition: above) Include suiruryi fluoride (SO^F^ ' 0. ISl), thlonyl tetrafluoride (S0F( - 0,221), silicon tetrafluorlde (S1F4 0.031), and sulfur dioxide (SOj - 0.0021). A suggested generalized reaction sequence is (94): srt Sf4 SOfj tr \ nil) S8j * Mf sir, . ntjO This reaction sequence Indicates that sulfur tetrafluorlde (SF4) and hydrogen fluoride (HF) are also decomposition by-products. 8oth are highly reactive, how ever, and were not found during analysis. HONS 216038 4-4 Silicon tetrafluorloe it not really an SFg decomposition by-product. Rather, it is t product of the reaction between HF (a probable decomposition by-product) and s' 11 ca-contelrring components of the test cell (or, by Inference, the transformer) sucn as ceramic insulators (36). ` The presence of impurities Influences the types and quantities of decomposition products formed C35. 111). These Impurities can Include production Impurities in commercial grades of gases, volatilized electrode material, and volatile frac tions of other equipment components (e.g., greases). It has been found that the addition of only a few percent of a hydrocarbon impurity can Increase SOF^ pro duction by two orders of magnitude (35,1. In the presence of hydrocarbons, addi tional by-products formed can Include C^Fg, CF4, and SlfCHjJjF^ (111). Finally, as will be noted below, tne compounds Identified above may not account for all of the toxicity noted in SF6 decomposition gases (J)7J. It has been sug. gestad that tnere is a potentially large array of PICs at trace concentrations which have not been Identified. For Instance, several researchers have hypotnes 1 zed the production of $2Fj0, although it has not been found (36,, 21) However, the analyses to date have concentrated on the major PICs; sulfur dioxide at 0,002X (20 ppm) was the lowest PIC concentration. Toxicity and Exposure limits of Sulfur Hexafluoride Decomposition Products Table 4-3 lists toxicity data for sulfur hexafluoride decomposition products. The major products formed during arcing (SOFj, $02F2, SfF4, SO^) are all more toxic than the parent compound, SF&. The sparxlng decomposition products have been found to exhibit strong cytotoxic effects on mammalian cells. The cytotoxicity of the sparking gases was tested using an In vitro cell culture system (Chinese hamster lung cells) (87). The cytotoxicities of the Individual decomposition products were tested to Identify tne most potent cytotoxlctnt. There were, nowever, no conclusive results. The sum of tne toxlcltles of the individual products could not account for the toxi city of the sparking gas. The toxicity of a synthetic mixture with a composition similar to that of the Identified sparking decomposition products exhibited a toxicity of a lower degree tnan that of the sparking gas. Thus, It 1$ reasonable to assume that there are perhaps several as-yet-unidenti fled decomposition products which, although present In trace amounts, contribute 4-5 MOWS 216039 t 1at>le 4-3 TOXICITY AN) EXPOSUHE LOUTS OF SULFUR ICXAFLOWIOE DECttlPUS IT ION PWOOUCTS (83) *6 sufz su?fz S.f so? Vio ihl-han ICL0 ihl-rat LC10 LCS0 orl-rat l050 ihl-us LCtO 260 ppat/ 1H 30Z0 ppm/ 100 ay/kg 111 TlV-TWA STEL 1000 ppa 12M> ppa 2 b ag (f)/ S PPM 10 ppM V *9 <H/ 1000 p(M/ 1000 ppm ION 2 PPM 1W ppa/ 2b ppt> 1H 5 ppM 7b ppb i \ NONS 216040 slgmflcantly to the cytotoxicity of the decomposition gases. One compound that has seen suggested Is SjFjq, l highly toxic compound with a TLV In the severs! ppt> ring*. The identified oecomposltlon products ere ill gases at normal ambient tempera tures, tnd would present l problem only in the tbsence of proper ventilation. Hydrogen fluoride may be found on surfaces. out the emount genereted does not seem to be signlflctnt. The physlcil state, and thus tne persistence,am long term hazard, of the unidentified decomposition products are unknown. HONS 216041 4-7 Section 5 PROPOSED PCS SUBSTITUTES: SOUOS INTRODUCTION Solid dielectrics can be cast or Impregnated Into the colls of a transformer, used as films In capacitors, or used as insulation for various parts of trans formers, capacitors, or other electrical equipment. There are several types of solid materials wnleh have been considered, including silicone resins, epoxy resins, and a variety of polymers. The two solids to be considered within the context of this review will be epoxy resins and polyvinyl chloride (PVC). EPOXY RESINS General Epoxy resins are those polymers whose monomeric units are linked or cross-linked through an epoxide-hydroxyl dehydration reaction. Cured "epoxy" resins contain few epoxide moieties. The monomeric units may be Simple or complex. based on blsphenol A. The most commonly used resin is This molecule Is linked Into long chains by a chlorine-containing epoxide com pound, eplchlorhydrln. HONS 216042 S-l The result It a long cnaln polyether uhich terminates with epoxides. cnJ1vcx - ex,--a-(C))-1 - CHj 0 - , - '*-ft* o -(o)-1-(Q)-O-a%-ciT-^*, -I" CJ Most commonly, this compound Is then "cured" with a cycloaliphatic anhydride or amine as a "hardener" to effect cross-linking (96). The largest producer of resins for use In casting transformer colls used In the United States reports that roughly 801 of their production for transformer colls may be approximated by the above generalltleSi The remainder use cydoallpnatlc compounds in place of blsphenol A. The principal constituent of the material used to impregnate trans former colls Is not the resin (typically 25x of the mixture) or the hardener (up to 251 of the mixture) but the filler material, a fine siliceous material (usu ally $102 SOt to 751 of the mixture). Thermal Decomposition of Epoxy fteslns All epoxy links are ether 1c, of weaker send energy than the adjacent carbon-car bon bonds. Uh# might then expect to find the monomer (e. g., blsphenol A) as an initial degradation product. This is not the case. The extensive cross-linking and the presence of the filler, both specifically Improve heat resistance and prevent any degradation below temperatures at which the monomers are stable. Pyrolysis of a tovolac epoxy (based on a chain of phenol-methyl moieties rather than blspnenol A) yielded no detectable monomer, but did yield significant amounts of benzene at temperatures above 500 *C (_M_) (Table 5-1), The major por tions of the volatlllied fraction of the epoxy is designated Vpyr, an otnerwlse unidentified material that is volatile at pyrolysis temperatures but liquid at room temperatures. The only data available for the 800 *C material Is an average molecular weight of 350. Lum and Felnsteln (3_, 64) studied the thermal degradation of Novolac epoxies at relatively low temperatures (0 to 500 'O, Release of significant amounts of bentene, phenol, and cresol was detected In the 300 to 500 *C range. MQNS 216043 5-2 Table 5-1 VOLATILE PROOUCTS FROM PYROLTSIS OF EPOXY RESIN (66)* Componentt 360 *C Hydrogen CO Methane Acetyl ene Ethylene Acetone Propane Propylene Ethane Cyclopent ad lanes Pentanes Benzene Methyl Chloride Ethyl cnlorlde Others Vpyr# Volatl Hied Part of Simple; % -- 4.7 16.2 1.0 0.3 -- 0.9 - 6, 5 --* --- S.l 1.7 -- 63.6 38 ' 500 eC 800 *C -- 3.1 6.0 0.8 -- -- 2.2 1.1 2.3 --- 0.5 1.3 --* 0.7 82.0 - (*)............................ 0.8 11.2 3.7 1.6 -- 3.0 -* -- 2.2 1.6 --2.8 --*- 73.1 75 86 1200 * 2.1 25.9 1.8 4.3 2.5 3.0 -- -0.6 -6.1 -- -* . 0.8 SU.9 87 'percentages are based on weight of total volatl Iliad fraction. Components In amounts of less than 0.Si are not shown. #A heavy fraction, volatile at pyrolysis temperature, but not room temperature. MONS 216044 5-3 Ox id at iv degradation of a complete 800 kVA GEAFOL east resin transformer was carried out by the Trifo-llnion Corpontion of the Federal Republic of Germany ($)i The experiment consisted of sequentlelly (not simultaneously) burning a wood fire (1000 C) under one coil of a transformer, and l propane fir* 11200 *C) under tne other coll. The filtered combustion gases were fed continuously to a GC/HS apparatus, which accepted a 10-microliter sample each 20 seconds for the duration of the fire. Mass spectra in the range of m/e 12 to m/e 650 were recorded each second. The results are shown in Table 5-2. ft) peeks above m/e > 92 were found. fbwhere are the ceroon dioxide concentrations In the smoke re ported, so tnere Is no way to exactly convert the peak Intensities to absolute concentrations. A further problem with these tests was the failure to separate the wood and propane contoustlon gases from the resin combustion gases. Conse quently. it is Impossible to say with certainty which of the produet gases are attributable to the epoxy resin. Table 5-2 PICS FCXJK) IN FIRE GASES FROM TUO CAST RESIN TRANSFORMER FIRES (5) PIC Acetylene Aromatic Fragment Seniol Toluene 3. 6 ppm 0, 4 ppm 1, 2 ppm 1, 6 ppm Propane Fire 3.4 ppm 0.4 ppm 1, 2 ppm 1.2 ppm As part of tnis same series of tests, a sample of resin was burned at 600 *C, and the product gases were analyzed for 2,3,7,8-tetraehlorod1b*nzo-p-d1oxin and 2,3,7.8-tetrachlorodlbenzofuran. Neither was detected down to 50 ppb. fto other dioxin or dibenzofuran was analyzed for. Toxicity and Exposure Limits of Epoxy Resin Decomposition Products Table 5-3 presents toxicity data for most of the decomposition by-products men tioned above. Few are highly toxic. Several of the oy-products, however, are known or suspected carcinogens. These compounds are generally solids or liquids MONS 216045 5-4 i Acetone Acetylene Benreoe Cresol -cresol o-cresol -cresol Cyc lopentadiene, 1,3Ethylene Rteool Table 5-3 TOXICITY AN) EXPOSURE UNITS OF EPOXY DECONPOS1II OH BY-PROOuCTS (83) tclo (ihl-han) 500 ppa 100 pp l0to (orl-han) "lo (Ihl-rat) lb.000 Ppn/4H LC50 (ihl -rit) L0M> jor1-rat) LCS0 (Ihl-nus) LC50 10,000 PPN//H 500,000 Pf/5M 2*2 ag/ky 121 ag/kg 207 ay/kg TLY-THA 5 ppa (skin) 75 ppa 140 ay/ky 316 mq/m'* 95 ppfi 5 pp* HONS 216046 it normal undent temperatures, end are risk via dermal exposure as well as oral or Inhalation. The toxicity of tne unidentified Vpyr fraction has not been asses sea. finally, few of the tests conducted to date nave been concerned with identifying byprod ucts at trace levels. POLYVINYL CK.ORIOE (RVC) General RVC is a synthetic thermoplastic polymer. It is generally resistant to weathering and moisture, and to most acids, fats, petroleum nydrocarbons, and fungus. It is dimensionally stable with good dielec tric properties. The first Important use for plasticized RVC was as electrical cable insulation sheathing. The electrical industry Is still a major consumer of PVC, with most being used as flexible cable Insulation or rigid electrical con duit, PVC resins are always compounded with a variety of additives. Stabilizers are added to act as scavengers for free radicals produced by heat or light, metallic salts, oxides, and salts of fatty acids are common stabilizers. Fillers such as calcium or magnesium carbonate can be added to retard the catalytic Chain daily* drochlorinatlon effect of released hydrogen chloride. Plasticizers may be added to impart or Improve flexibility. Plasticizers may be up to 301 by weight of the compounded plastic. Typical plasticizers include pnospnates, dioctylpnthalate, and low-molecular-weight propylene glycol polymers. To Increase flexibility, vinyl enloride may be copolymer 1 zed with vinyl acetate. The vinyl acetate can constitute 3 to 15% of the copolymer. There Is no single compound formula used ubiquitously in tne electrical industry. MOMS 216047 6-6 Thermal Decomposition Products of PVC Data on tne thermal degradation of PVC Iff fairly extensive. 'The major problem with using the data it the variety of Pvc formulations available. A given test sample may be stabilized or unstabitlzed. There are several different staoilizert available. Other additives (plasticizers, fillers, dyes,.etc.J can vary in composition as well. The extent to which tnese additives affect the tnermal degradation of the polymer and the extent to which the additives themselves con tribute to the final by-product mix are unresolved. Furthermore, most o'f the published research is not highly specific regarding the formulation or composi tion of the PVC being tested. Consequently, there Is some question regarding the applicability of a particular test result to other PVCs. As a result, the fol lowing discussion must be taken as being only generally Indicative and not neces sarily specific for any particular PVC formulation. In the absence of oxygen, PVC pyrolyzes by dehydrochlorlnatlon (loss of HCI), followed by cycllzatlon of polyene (^9, 39, 76). The reaction proceeds via 1,4* elimination of chlorine from cU_ double bonds, generally from groups at or near chain ends. /IAS \ The principal products formed are benzene and hydrogen chloride Benzene is bmlleved to form prooably by one of the two following mechanisms (76j: MQNS 216048 5-7 Tne presence of tCl may catalyze tne thermal degradation, but tnls has been dis puted (_19_, 76). Benzene is not tne only decomposition product. Table 5-4 11st? thp-decompositlon yields of several more aromatic Compounds at various temperatures from 300 to 800 C. in addition to tnest compounds, aliphatic hydrocarbons, d-metnyI naphthalene, monochlorobenzene, o^lchlorobenzene, jMSlcnlorobenzene, 1,3,5-tr(chlorobenzene, and 1,2,4-trlchlorobenzene were also Identified in tne pyrolysis gases (19), Table 5-4 DECOMPOSITION YIELDS* OF SOME AROMATICS OBTAINED BY THE PYROLYSISt OF PVC (GE0N-103 EP) AT VARIOUS TEMPERATURES (19) Product Benzene Toluene Ethylbenzene o-Xylen* Styrene Naphthalene 300 *C 3.50 -- -* -- ** 400 C 4.67 0.21 -- -- 0.13 0.47 Yield, wt t 500 *C 600 *C 5.00 0.67 0.16 0.16 0. 22 0. 52 5.23 0.82 0.18 0.16 0.37 0. 73 700 C 6.30 1.55 0.18 0.23 0.74 1. 50 800 *C 9.94 2.10 0. 23 0.25 t. 06 3.00 'Calculated using gas chromatography with -d (chlorobenzene as an Internal stan dard. rSamples (powder), 1-3 mg, were pyrolyzed in helium gas; flow rate of helium being 25 ml/mln. Ah ling et al. (_3) tombusted PVC over the range of 570 to 1130 *C. They found from dl- to hexechlorlnated benzenes, octachlorostyrene, and PCBs. Ourlng tne combustion experiments, PVC was mixed with wood chips to help sustain combustion. The combustion of PVC can also yield carbonyl chloride (phosgene) (J^3). Yields are on the order of 0.0$ to 0.15 mg/g (or a concentration of 0.5 to 1.3 ppm from 4 to 7 g decomposed In a 200-1 animal exposure chamber). Subjectl ny 'the PVC to an electric arc can Increase yields of phosgene 10 to 20 times. HONS 216049 5-8 Because of the production of chlorinated benzenes, tne formation of PCDOs ano PCDFs Is a distinct possibility {see Section 3, Chlorobenzenes) U9)> The pres ence of otner chlorinated polycyclic aromatics, such as chlorinate^ naphthalenes, Is also feasible. To date, none of these have been identified in PVC pyroLysis gases, but PVC is suspected to be one of the potential sources of the PC Oils and PCDFs found In fly ash and flue gases from municipal and Industrial IncTnerators (.19). ' Toxicity and Exposure Limits of PVC Thermal Depredation Products Table 5-5 lists the toxlcltles for the Identified PVC degradation products. Speculative byproducts, such as PCDOs, PCDFs, PCKs, etc., are not included, although the clrcumstantlal evidence for their production is good. Their toxlcl tles are discussed elsewhere In this review. The Identified PVC byproducts are of variable toxicity, but generally fall Into the moderate to high toxicity range. In addition, several are known or suspected carcinogens. Host of these compounds are solids or liquids at normal ambient temperatures, and could be found in soot or on surfaces after a fire. Phosgene, a gas, is the notable exception. HONS 216050 5-9 i OT-S Table 5-5 TOXICIIIES AM) EXPOSURE LIMITS UP PVC THERMAL DEGRADATION BY-PKOOUCTS (B3) Mtosyene LC50 (ihl-ft*n) 3$00 *9/ Benzene Toluene Ethyl benzene -Xylene Styrene Naphthalene Methyl naphthalene Honorh1orobenzene o-Dlch1orobenzene -01 chlorobenzene 1,2.4-Trichlorobtnient tclo (ihl-h*n) n ppm/ 30M IDO pp* ,DlO (orl-hu*) 2U0 pp* 1U0 pp*/ OH 600 pym 300 *9/kg LCLO (TNI -rat) SO pp*/ 30M 4000 pp*/ 4H 4000 pp*/ 4H 612S pp*/ 12H SOOO pp*/ 8H 821 pp*/ 7H - LCSO (ihl-rat) 10,000 PP*/7H LDSO (or 1-rat) ldlu (orI -rat) 1780 *g/kg 5000 *g/kg 2910 *9/fc9 500 *9/kg ; 1 500 *9/1(9 756 *y/ky .' ILV-IWA 1U0 ppb 1U pp* 100 ppai 100 ppm 60 pp* 10 pp* 75 pp* 75 pp* HONS 216051 Section 6 REFERENCES 1. Addison. R. F. PC8 replecements in dielectric fluid*. Environ. Scl. Terh. no I,. 17:486A*494A, 1903, ------------------------ -- 2. Addlton, R. F., M. L Zlnck. 0. E. unlit, end 0. C, Derrow. Induction of hepetlc mixed function oxideses in trout by polychlorlneted biphenyls end butyleted monochlorodlphenyl ether*. Toxicol. Appl, Phermecol.. 49:245* 248, 1979. 3. Ahllng, B., A. BJorteth, end G. Lunde. Formetlon of chlorineted hydrocer* bon* during conbuttlon of polyfvlnyl chloride). Chemotpnere. 7:799-806, 1978. 4. Allen, ft S,, ed. Dearedetion end Staollitetlon of Polyolefin*. London: Applied Science Publishers, 1983. 5. Altaenn, G,, K, Mertln, R. Pfeiffer, end K. Schreymr. Fire end the Cast* Retln Dry-Type Transformer. Stutt9*rt, Fed ere 1 Republic or uermeny; Trefo-union, 1984, 6. Anderson, R, C,, end . C. E. Alerle. An attempt to trentlete toxicity of polymer thermel decomposition product* into e toxicological hazard index end discussion on the epprotches selected. J. Combust. Toxicol.. 5:476* 4B4, 1978, 7. Bedoer, G, M, Pyrolysis of hyorocerbons. Prog. Phys. Org. Chem.. 3:1*40, 1965. ----- --------------------------------- 8. Bed9er, G, N,, R, G. Buttery, R, u. U Klitter, G. E. Lewis, A. G. Moritz, end 1. M. Nipler. The formetlon of erometlc hyorocerbons et high tempers* tures. Pert t. Introduction. J. Chem. Soc.. 1958: 2449*2452, 9. Badger, G. K, S. 0. Joled, end T. M, Spots wood. The formetlon .af erometlc hyorocerbons et high temperetures. XX. The pyrolysis of [l*1 C]nephthe* lene. Aust. J. Chem.. 17: 771-777, 196*. 10. Sergmen, A., A. Hgrnon, S. Jecobsson, 8. Jensson, end M. Ah 1 men. Therms! degredetlon of polychlorlneted elkenes. Chemotpnere. 13:237-250, 1984, 11. Bresleu, A, J. Electrlcel end electronic eppllcetons: seelents end foems. In: Epoxy Resins: Chemistry end Technology. C. A. Mey end f, Teneke, eds. New York; nerceiUekker, 1973. pp. 485*583. 12. Broedhurst, K G. Use end repleceeblllty of polychlorlneted biphenyls. Environ, heelth Persoec. Experiment*! Issue Z:B1*102, 1972. HONS 216052 6*1 13. Brown. J. E. , ind M, M. Birky. Phosgene In the thermal decomposition prod ucts of polyfvlnyl chloride): generation, detection ano measurement. J, Anal. Toxicol., 4:166-174, I960. ~ 14. Brown, H. F. C. Pyrolltlc Methods In Organic Chemistry:, 'Application pf flow and Flash Vacuum Pyrolytic Techniques. Qrq. Chem.T *1. Mew *orx; Academic PretT.liSGT "p. 5U6.------------- --- ------------- 15. Bur rage, L. M., and G. A. Gauger, Problems facing manufacturers1; dielec tric fluids for power factor correction capacitors. IEEE Trans. Electr. Jnsul,. 13:370-388, 1978. 16. Buser, H A. Formation of polychlorinated dlbenzofurans (PCOFs) and dibenzo-p-dloxlns {PCDDS) from the pyrolysis of chlorobenzenes. Chemosphere. 0:415-424. 1979. 17. Buser, H, A. Polychlorinated Dlbenzo-p-Ploxlns and Qlbenzofurans: Forma tion. Occurrence and Analysis or Environmentally hazardous compounds. Ph.O. Thesis, Urne a, Sweden: university of Umea, 1978. 10. Castonguay. G., A. Theoret, and A. Gilbert. Arc degradation of sulfur hexefluorioe in the presence of polymeric Insulating materials. Conf. Aec. IEEE Int, Symp. Electr. lnsul. , 1970:162-166. 19. Choudhry, G. G.. K. Olle. and 0. Hjtzlnger. Mechanisms In the thermal formotion of chlorinated compounds Including polychlorinated dibenzo-p-dlox- ins. In: Chlorinated Oloxins and Related Compounds, Impact on the Envi ronment. 0. Hitzlnger et al., eos. ElmsfordV fork: Pergamon Press, 1J8JL 'pp. 275-301. 20. Chu, F. Y., and R. E. Massey. Thermal decomposition of sulfur hexafluoride and sulfur hexafluoride-air mixtures in substation environments. Gaseous Dielectric Proc. Int, Symp.. 3:410-4ig, 1981. 21. Claiborne. C. C. Products of Incomplete combustion of dielectric fluids. Presented at EPAI PCOF Workshop. Palo Alto, California, Oecemoer 19B4. 22. Croucn, E. A. C., and A. Wilson. Inter-risk comparisons. In: Assessment and Management of Chemical Risks. J. V. Aoorleks and R. 6, Tareur, eos. ACS Symp. Ser. 2lT. wasnlngton, O.C.: American Chemteal Society, 1904. ppT 97-112. 23. Crummett, W. 0., and 0. I. Townsend. The trace chemistries of fire hypoth esis: review and update. Chemospnere. 13:777-700, 1984. 23a. Dellinger,B., j. 1. Torres, w. A. Rubey, D. L Hall, and J. L. Graham. Determination of the Thermal Decomposition Properties of 20 Selected Haz ardous Organic Compounds, EPA-600/ 2-84-130. Industrial Environmental Research Laboratory, li.S. Environmental Protection Agency, Cincinnati, Ohio, August 1984. 24. Deroos, f.L., M. Cooke. G. Addis, R. Komel, and J. Guertln. Formation of PCOO and PCDF in eskerel and contaminated mineral oil equipments Presented at EPRI PCOF Workshop, Palo Alto, Cellfornte, Decemoer 1904. 25. Oow Chemical U. S. A. Perchloroethylene SVG - A Vapor Deoreaslno Solvent That Beats Corrosion. Hid 1 ano , Michigan, 1981. HONS 216053 6-2 26. Dow Corning Corporation. Dow Corning 561: Silicone Transformer Llould. Retrof 111. or Replace four PCB Transformer, tne Choice is four*. ntdland- Michigan, ivbj. -- . 11. Oryden, F. E., H, E. Ensley, R. J. Rot*, and E. J. Westbrook. Oloxln*. vol. III. Assessment of Oloxin-forminq Chemical Proctttt*. Final deport,. EPA-600/ 2-tiU-1walk, Haycel and Associates, lnc.", New Orlean*. for Industrial Environmental Research Laooratory, Cincinnati, Ohio, June 4980. 28. Duvall, 0. S., u. A. Rubey. ana J. A. Mescher. High temperature decomposi tion of organic hazardous weste. In: Treatment of Hazardous waste. Proceedlngs of the 51 nth Annual Research Symposium, Chicago, Marcn 1980. EJT Shultz, ed. EPfc-60d/4-8o-lJll. Cincinnati, Ohio: Municipal Environmental Researcn Laboratory, March 1980. pp. 121-131. 29. Eadon, 6. Pyrolysis and combustion of mfnerel oil, tetreehloroethylene and silicone oil mixtures containing Aroclor 1254. Presented at EPR1 PCOF Workshop, Palo Alto, California, December 1984. 30. Erickson, M. D., C. J. Colt, J. 0. Flora, Jr., P. 6. Gorman, C. 1. Halle, G, 0. Wnshaw, F. C. Hopkins, S. E. Swanson, and 0. T. Heggam. Formation of PCDFs under "optimized" combustion conditions from low-level PCS dielec tric fluids. Presented at EPRI PCOF Workshop, Palo Alto, California, December 1984, 31. Esposito, M. F., T. 0. Tlernen. and F. E. Oryoen. Oloxlns. EPA-600/2-80- 197. Cincinnati, Ohio: Industrial Environmental Aeiearch Laboratory, November 1980. 32. Cara, A. Studies 5f rulyehiorinated pioenzofurant. Ph.o. Dissertation, Umea, Sweden: university or umea, 1982. 33. Gifford, L. It, and T. Orbeck. Evaluation of the long-term thermal capa bility of a hlgh-temperature insulation system using silicone liquid as a dielectric coolant. IEEE Trans. Indus. Appl., 20:349-384, 1984. 34. Grassle, It. and t. G. Maefarlane. Degradation Reaction* in Silicone Poly mers. Glasgow, Scotland: University of Glasgow, Department of Chemistry, December 1976. 35. Griffin, G. 0.. C. E. Easterly, and P. j. Walsh, Assessment of potential health effects of dielectric gases. Gaseous Dielectric Proc. Int. Symp.. 2 : 439 -4 50, i960, 36. Griffin, G. D., 1. Sauers, C. E. Easterly, H. W. Ellis, and L. G. Christo pnorou. Decomposition of sparked SFfi and cytotoxicity of SF- decomposition products. Presented at International Symposium on Gaseous oielectrlcs, Knoxville. Tennessee, May 1984, 37. Griffin, G. D. . I. Sauers, L. G. Chrlstophorou, C. E. Easterly, and P. J. Walsh,On the Toxlcltv of sparked 18:551-552, lg83. SF*. IEEE Trans. Electr, tnsul.. 6 --------------------------------------- 38. Hartzel 1 G. E., S. C. Packham. and w, G. Switzer. Assessment of toxic hazards of smoke: toxicological potency and intoxication rate thresholds. Fire Mat.. 7:128-131, 1983. MONS 216054 6-3 39. HtMklns. w. L. Polymer Degradation ana Stabilization. Berlin. hew rork: Sprlnger-Verleg, 19$ 40. hbad, J. 6.. P. S. Gale, and w. G. Lawton, Effects of high temperatures *nd electric ttrett onthe degradation of oil-filled, cable fnsdtation. [nt, Conf. Oleleccr. Wet., Measur, , Appllc.. 3iS6-60, .1979. " 41. Heltmuller, P. T,, T. A.Hollister. and P. A. Perrlth. Acute toilclty of 54 Industrial chemicals to Sheepshead Minnows (Cyprlnoapn varleqatus). Bull. Environ, Contam. Toxicol.. 27:596-604, 1981. . 42. mi too, C. J., and P. A. Hitttlnger. A bibliography of published papers on coiiDustion to*1co1o9y. J. Combust. Toxicol.. 7:119-123, 1900. 43. MTedo, C. J.. and P. A. Hittllnger. Some relationships between relative toxicity and total and lethal weignt loss of polymers, J. Combust, Toil- col.. 7:194-199, i960. " 44. MlaOo, C, J., C. J. Casey. 0. F. Christensen, and J. Llpowltz. Toxicity of pyrolysis gates from sllcone polymers. J, Combust. Toxicol.. 5:130-140. 1970. 45. Hlzlroglu, H, A, , M. A. Selnz, 0. M, Coleman, and A. H Ouersni, Spectro* metric analysis of sulphur hexafluoride during an electrical discharge. Conf, Pec. IEEE Int, Symp, Electr, Insul,, 1982:150-152. 46. Hird, C 0. The Pyrolysis of Carbon Compounds, fmw York: Chemical Cata log Company, 19*9. 47. Hird, C. 0. The pyrolysis of esters. Am. Chem. J.. 60:2419-2425, 1938. 40. Inchallk. E. J. ENJ-2065 an electrical Insulating fluid. In: national Conference on Polychlorinated 81pnenyls, Chicago, fOvenber 1975. f. A. Ayer, tal--E?A-560/f-75-004.--Research Triangle Park, forth Carolina: Ae* searcn Triangle Institute, Marcn 1976. pp. 334-337. 49. ISC Chemicals Limited. Forme! HF, a Hew, Safe, rpn-flemmaole Dielectric Coolant for Transformers. Bristol, England, 19&3. 50. Isobe, S., T. Mori to, S. Mlura. T. Uchlyama, F. Ohyama, n. Fukuoka, 1L Tsu- k.ioka, E. Ohe, and M. Yamagucnl. Silicone oil filled transformers and Its silicone oil characteristics. Proc. Electr./Electron. Insul. Conf,. 12:215-219, 1975. 51. Johnson, A. T., Jr., and A. M. Blefeld. Electrical conductivity and ther mal decomposition of high temperature phenolic- and silicone-based mate rials. Annu, Aes. Conf. Electr. Insul. D1 electr, phenom., 1981:85-92. 52. Josephson, J. Phasing out PCBs. Environ. Scl, Techno!.. 1B:43A-44A, 19B4. 53. Kaplan, K L. , A. F. Grand, and G. E. Hartzell. Combustion ToxicologyPrinciples and Test Methods. Lancaster. PennsylvanlaT" lecnnomic Publish* ing Company, 1983. 54. Kelly, J. J. Predicting transformer Incipient faults by dissolved gas In oil analysis. Proc. Electr./Electron, Insul. Conf.. 14:15-21, 1979. HONS 216055 6-4 55. Kotzlit. 0., u, Klein, and F. norte. Photochemlsnces vernalten elncs PCB- Subsmuten In waessriger Loesung und <uf Traegermaterial, Chemosonere. 1 :99-102, 1977, ------ 56. Kovalenko, V. V,, M, L. Fusty!'nfk, T. S, Berchuk, M; \ Mwafakova, v, y, Zherdev, V, 0. Nedorosol, and k. A. Andrianov. 'Structure of the monomeric unit of crosslinked polymer and Its relation to the thermal stability of polymethylslloianes, Polym, Sc 1.. 20:1824-1833, 1979. 57. Krevlt, H. h,, 0. H. Oamon, J. J. Johnson, and S. V, heyer. Analysis of volatile components In polymeric dielectric Insulation. Conf, Pec, IEEE Inter. 5ymp. Electr, Insul,, 1982:179-182. 58. Kuwahare, H,, K, Tsuruts, T, isnll, and K. Yosnlnga. Study of explosion and fire hazards of silicone liquid under arc conditions. Conf, Rec. IEEE Int, Symp. Electr. Insul.. 1976:186-194, 59. Lindahl. R.. C. Rappe, and H. R. Buser. Formation of polychlorinated dlbenzofurans (PCOFs) and polychlorinated dlbenzo-p-dloxlns (PCOOs) from the pyrolysis of polychlorinated dlpnenyl etners. Cnemospnere. 9:351-361, I960. 60. Llnsley, K, inflammable small power transformers. Presented at American Power Conference. April 29, 1981. 61. Llpowltz, J. Flammability of poly(dlmethylslloxanes). I. A model for combustion. J, Fire Flammability. 74:482-503, 1976. , 62. Llpowltz, J., J. E. Jones, and M. Kanakia. Fire safety properties of some transformers dielectric liquids. Proc. Electr,/Electron. Insul. Conf,. 11.911.^111 IOTA ' '' 63. Lum, R. M., and L. 6. Felnsteln, investigation of the molecular processes controlling corrosion failure mechanisms in plastic encapsulated semicon ductor devices. Mlcrpelectrpn, Rellab.. 21:15-31, 1981. 64. Lum, K. M., and L <*, Felnsteln. Thermal degradation of polymers for molded integrated circuit {IC) devices: tne effect of a flame retardant. In: Polymer Materials for Electronic Applications. E. 0. Felt ana C. w. Wilkins, Jr.. edS. no Symp, ^er, 184. wasnl noton. D.C.: American Chemi cal Society, 1982. pp. 65. Madorsky, S. U, S, Straus, D. Thompson, and L. Williamson. Pyrolysis of polylsobutene {Vlstanex), polyisoprene. polybutadiene. GR-5, and polyechy. lene In a high vacuum. J. Res, hatl. Bur. Standards. 42:499-514 (Research Paper 19891, 1949. 66. Kadorsky, S, L, and S. Straus. Stability of thermoset plastics at high temperatures. Hod, Plast.. 13*+, February 1961. 67. McCHntlck, B. R. PCB transformers: replace or retroMll? Hazard, hat, waste Manage. 2(5):30-34, I9B4, 68. Mantel, D. B., and E. D. Smolko. interspecies extrapolation. "in: Assess ment and Management of Chemical Risks. J. V. Rod ricks and R. G. raroi rr, ads. ACS Symp. Ser. 239. washtngton, 0.C.: American Chemical Society, 1984. pf. /J-Jb. HONS 216056 6-5 69. The Merck Index. 9th Edition. Merc* 6 Co. , 19/6. M. Wlndholz, . Rahway, few Jersey. 70. Montgomery, R. H, Tne use of 0o Corn1n9 02-1090 Olelectrlc Liquid in power trensformers. In; National Conference on Polychlorimcea j)(pBenyls. Chicago. hovemeer 1975. F. A. Ayer, EPA-S6Q/&-76-004. - Research Trt- m;1e Pint, North Carolina: Research Triangle Instttote, March 1976. pp. 312-313. 71. National Electrlcil Minufecturers Association. Reply comments on Notice of proposed rulemaking concerning polychlorInittd biphenyls (PCBs)x' use In eleccricel equipment, 47 Federal Register 17.246 (April 22, 1982), wish* Ington, 0. C., June 1982. 72. Nitlonel Hiterlels Advisory Soira. Smoke end Toxicity (Combustion Toxicol ogy of Polymers) Fire Safety Aspects of Polymeric Materials. volume""!/ Uasnington, D. C.; Titlonal Academy of Sciences, i9/?. _ 73. Nonf 1 immeple substitute for PCH Introduced In U.K. 62(13): 31 , 1984. Chem, Enq. lews. ------------ 74. Norstroem, A., K. Andersson, end C. Rippe. Formation of chlorodlbenzofurms by irradiation of chlorlmted diphenyl ethers. Chemosohere, 1:21-24, 1976. 75. Novik, X. M. Toxicity of Si 1 Icon Compounds. 8NL-S1509, Upton, New York: Nit tom I Center for Analysis of Energy Systems. Srookhaven Nitlonel Labora tory, June 1981. 76. Owen, E. 0., ed. Degradation and StH> 111 sit Ion of PVC. London; Elsevier Applied Science Publishers. 1984. 77. Papkov, V. S. , A. F. Bulkin, A. A. 2hdmov, G. L. Slonlmskif. and X. A. Andrianov. Investigation of the thermal oxidation of polydlmethyliiloxanes. Polvm, Scl.. A19:830-841, 1977. 78. PCS substitute Is safe for environment. Chem. Eng. News. 54(49):7, 1975. 79. Peirano, w. 8. health Assessment Oocument for Chlorinated Benzenes. EPA600/8-84.015. Cincinnati, Ohio: Office of 'Health ana Environmental Assessment, Environmental Protection Agency, April 1984. 80. Pellenberg, R. Environmental poly(organoslloxanes) (silicones), Environ. Scl. Toxicol.. 13:565-569, 1979, -------------- 81. Phillips, P. J, Morphology and molecular structure of polymers and their dielectric behavior. In: Engineering Dielectrics. vol. 2A. rtSTM STP 783, Philadelphia, PennsylvanTJfi ASTM, 191)3. pp. 119-237. 82. Physiological and Toxicological Aspects of Combustion Products: Inter national Symposium^ Sait Like utv. March 19M. Washington, u.l. ; National Acaaamy ot Sciences, 1976. 83. Registry of Toxic Effects of Chemical Substances. 1981*82 Edition. R. L. Tatxen ana R. J. Lewis, Sr., eos. Clnctnnat 1, Ohio: U.S. Public realtn Service, June 1983. MONS 216057 6-6 8. Hollins, R. L. PC8*t in capacitor application*. in: National Conference on Polychlorinated Biphenyls., Chicago, tovemoer 1975. T; K," (kyer. . EPA-S6U/6-75-004. Aesearch Triangle Parle. torth Carolina: R#*#arcn Tr 1 - angle Institute, Karen 1976. pp. 306-308. ....... . 85. HTE Fluid*. RTEmp Dielectric Fluid: UL Classlf~1j, Your A**urance of Greater Safety, waukesna, tJtsconsin. 198*. 86. Rubey, W. A., and . A, Carnet. Review of laboratory systems developed for studying gas-phase thermal decomposition. Proc. tat. Waste Process'Conf., 11:495-506, 1984. 87. Saido, X., T, Kuroki, T. lkemura, and M. Xirisawa. Studies on the thermal decomposition of phthalate esters, IX, Thermal decomposition of bis(2ethylhexyl) phthalate, J, Anal. Appl, Pyrol,. 6:71.181, 1984, 88. Saido, X., T. Xuroki, T. lkemura, and N. Kirisawa, Thermal stability of pnthalic esters, J. Am. Oil Chem. Soc,, 61:945-949, 1984, 89. Salto, F, Evaluation of the toxicity of coabustion products, J, Combust. Toxicol., 4:32*51, 1977, ' 90. Santoro, R. J., and 1. Glassman, A review of aromatic compounds. Combust. Sci. Techno 1.. 19:161-164, 1979, -------------- 91. Saperstein, M, 0., and E. J, Falder. Environmental toxicology of PCB sub stitutes for capacitors. Reoulat. Toxicol. Pharmacol.. 2:238-246', 1983. 92. Saperstein, M., and J, Afong, Environmental Toxicology of PC8 Substitutes for Power Capacitors. 81-RD-lua. xosemeaa. California: Southern canI`.0"* fl ,! , tflSevil, JiXl/bVWCLV_f_l_<n6ItOi 93. Sauers, 1., L, 6. Christophorou, L. C. Frees, and K W, Ellis, Aspects of environmental effects of dielectric gases. Gaseous 01 electric Proc, lnt. Sympi. 2:429-430, I960. ---------------------------------------------------- 94. Sauers, I., L. G. Christophorou, and S, m, Spyrou. Negative 1o formation in SFc spark by-products. COIf-840403-6, Oak Rtdge, Tennessee: Oak Ridge National Laboratory, 1984. 95. Sak, It 1, Dangerous Properties of Industrial Materials, 5th Edition. New tork: Van rcstrano Keinhoid, 1979. 96. Sill art, ft. W, Electrical Insulating Material* and Their Application. IEEE Monograph Ser^ jl, London, Peter Peregrlnus Ltd., 1973. pp.^l-T/g, 152-198, 205-2ZB. 97. Simmon*, w, S. Toxicity Profiles of PCB Substitutes, Final Report, EPR1 EA-3567. Systems Applications, Inc., San Rafael, California, for Electric Rower Research institute, Palo Alto, California, July 1984. 98. Stevens, C, Environmental Impact Assessment of Formal tf. AERE-G 3247, Harwell, England: AERE Harwell, Environmental and Medical Sciences Divi sion, Septenber 1984, ~ 99. Sumlno, X, mass fragmentation determination of diisopropyl naphthalene and phenyl, xylylethane, and the environmental contamination from them.- Arch. Environ, Contain. Toxicol.. 6:365-369, 1977. HONS 216058 6-7 100. Suzuki, T., S. Murekaml. S. Kobayashl, K. Klyotakf, and M. Limed a, Oecompotitlon product* of silicone liquid due to electric discharge. IEEE Trans. Electrl. ln*ul.. 18:152-157, 1983. 101. Synthetic luorlcent* will dlsplece. Can. Chew. Proc. . August 1977. 102. Ttntka, J.. end K. Wolter. Composition end structure of dielectric solids. In: Engineering Qlelectrlcs. vol. 2A, R. 8*rtn1ke* end R. M. Elchhorn, ed*. AS in Mt1 783. mi iadelphi a, Pennsylvania: ASTK, 1983. pp. 521-618. 103. Terrill, J. B., R. R. Montgomery, md C. J. Reinhardt. Toxic gases from fires. Science. 200 {4348); 1343-1347, 1978. 104. TLVs: Threshold Limit Value* for Chemical Substances In the Work Environ ment. Adopted by AtilH for IU84-B5. 2nd Printing. Cincinnati, Ohio: American Conference of Governmental Industrial 1984. 105. Tracy, S. New transformers: nontoxic, nonflammable. EPR1 J. , 9(8): 16-19, 1984. 106. Troltzsch, J. International PlasticsFlammability Handbook. Macmillan, 1983. New York: 107. Tulp, M. T. M., G. C. van Stein, and 0. Kitzlngtr, Environmental chemistry of PCS-replacement compounds Vl - Composition and metabolism of WEMC0L, an Isopropylblphenyl formulation. Chomosphere. 1:747-759, 1978. 108. Tulp, M. T. M., G. Sundstroem, L. 8. J. M. Martron, and 0. MUtllnger. Metabolism of chiorodlphenyl ethers and lrgasan DP 300, Xenoblotlca. 9:65 77, 1979. 109. U.S, nvIrsnmefttal Protection Agency, Office of Mit'ib and Environmental Assessment. imalth Assessment Document for Polychlorinated Dlbenzo-o- Qloxlns. EPA-600/8-84-014, Cincinnati, Ohio, May 1984. 110. Urabe, X., T. Masuda, and H, Tamaguchl. Thermal degradation In silicone 11quid/aramld paper Insulation system. Conf. Rtc. IEEE lnt. Symp, Elec tric. Insul., 1980:11-14. 111. Vouros, P. The analysts of arc by-products of SFg by comblntd gas chroma tography /mass spectrometry. In; Workshop on Arc By-Products In Gas-Insu lated Equipment. Chicago, June 198$ Final Report. tl-30-B-LD. PaTo Alto,' tallrornta: Electric Power Research Institute, 1980. pp. 9-10. 112. Voytlk, 0. E., and H, A. Paarca. Liquid-filled transformers having hlgn temparature oparatlng capabilities. Proc. Electr./Electron. Insul. Conf.. 15:203-211, 1981. ---------------------------------------------------------------- 113. Vuceta, J., J. R. Marsn, S. Kennedy, l. Hlldemann. and S. Ullay. State-ofthe-Art Review: PCQOs and PCDFs In Utility PCB Fluid. Final Report. EPR I Li-JJUB. 5t5 Engineers, Long eeacn, Lafiforma~for Electric Power Re search Institute, Palo Alto, California, tovambar 1903. 114. Weddlngton, F, B. A new synthetic ester fluid for transformers. Proc. Electr. /ElbCtron. Insul. Conf.. 14:211-213, 1979. HONS 216059 6-8 115, Will, L. A. M*ss*spectrometr1c Investlgttlon of trie therm*I decomoositlon * of polyiMrs. J. Wet. h*t. 9ur. 5t*no*ros. 41:315*322 (Res**rch Piper 1928)., 198. 116, Well, L. A,, ed. The M*ch*n1sms of Pyrolysis. Qxldttlpn, *na ~8urrrtna of 0ron1c Meterlels. ~ Rfct, Bur, btiWr :>pec~. Kuo. j&/. wVsnindton. p.c. U.I. Governmenrprlntlns Office, June 1972, - 117, westtn, R. A. Assessment of the list of Selected RepUcement Fluids for PCBs In Electric*! Equipment^ Fin*) Report. tPA-SSO/4-77.QaS, ' verier, Inc,, ipringMeufTVtrgin: t, tor Office of Toxic SuOsttnces. H*rehl979. 118, Wlll1*ms, S. J., tnd F, 8. Cltrke, Combustion product toxicity: depen. dene* on the mode of product 9*ner*t1on. Fire H*t,. 7:96-97, 1983. 119, Wilson, A. C. N, Intul*t1nq LSquids; Their uses, M*nuf*ctur1nq, *nj Prop. ertles. IEEE Electric, tiectron. ~fo*l*r. Lev. i*r. i. London: Peter Per*. linwrnq:: tsto-------------------------------------------------------- 120, Wilson, 0. C. , tnd C, Stevens. Development of Tot* 11 y Non.FI*mut>le Trniformer Fluid. Oxfordshire. Engfina: H*rw*ii L*OOr*tory, 1983, 121, Y*sufuku, S.. tnd T. Ueemur*. A new study of fleaentblllty of dimethyl sil icone liquid. Proc. Electr./Electron, Insul. Conf,. 15:212*216, 1981, HONS 216060 6-9 Append 11 A GLOSSARY AH) ABBREVIATIONS 0P9 Kim ini "10 LC50 LS0 Letnel Concentration Fifty let Ml Concentration low letnel Dote Fifty Letnel Dote Low MuS Or T PC 00 PCOF PC N PIC ppn ppn STEl Guinee pig Kiinan Inhalation lethal concentration low Lethal eoncentr:. on fifty Lethal oote low Lethal dote fi fty A calculated concentration of a tuottance In air, exposure to which for a specified tine (H - hours, m minutes) is expected to cause the death of SOS of a defined animal population The lowest concentration of a substance in air which has been reported to have caused death in humans or anl malt A calculated dose of a substance which is expected to cause the death of SOI of a defined animal population The lowest dose of a substance reported to have caused death In humans or animals Mouse Oral Polychlorinated dlbenio-p-dioxln Polychlorinated dibeniofuran Polychlorinated naphthalenes Product of Incomplete combustion Parts per hundred ' Parts per million Short term exposure limit; maximum al lowable exposure concentrations for iS-minute exposure A-l HONS 216061 Apptndlx A (continued) TCOO TCOF io TQlo TOjq Thrtlhold limit Valut Tttr ch I orod 1 b tmo-p -d 1 ox 1 n Tetrtchlorod tbtniofurtn Toxic conctntrttlon lot* Toxic doit lot* Toxic doit fifty - Tht ACGlH-rtcommtndtd conctntrttlon of t lubittnct to which most worfctn can bt txpoltd without tdvtrtt ifftct; mty bt tiprtittd at tlmt-wtlgnttd tvtrtgt (TVA), ihortttrm txpoiurt limit (STEl), or calling valut (Cl) TIV Toxic Conctntritlon lot* Toxic Dost fifty Thrtthold limit vilut Tht lowtit conctntrttlon of t lubittnct In tlr to which human! or tnlmtli havt bttn txpoltd for tny glvtn ptrlod of tlmt thtt htt productd tny toxic tfftct A ctlcultttd doit of t lubittnct rtporttd to ctutt toxic tfftct! In SOX of t dtflntd populttlon Toxic Oott Lot* TUA Tht lowtit doit of t lubittnct rtporttd to product tny toxic tfftct in huntnt or tnlmtli TlM-wtlghttd tvtrtgt; tvtrtgt allowable txpoiurt ovtr tr> 8-hour workday or 40-hour work wttfc A-2 HONS 216062 Appendix 8 GENERIC AW TRAOE NAMES Benzylneoceprete Butyl eteo monochlorodiphenylether FARAOOL 100 (Asee) DIELECTRIC FLU 10 C-4 (Dow), XFS-4169L (Dow) DC 200 (Dow) DC 561 (Dow) 01-2-ethylhexyl pnthelete Polydimethylslloxene Polydlmethylslloxene Component of OIELECTROL II (Genere) Electric) end SELECTROL I (Sengemo) 01 ELECTRIC FLUID C-4 (Dow) Butyleted monochlorodtphenytether OIELECTROL II (Generel Electric) 01 -2-ethylhexyl phthelite OIELECTROL III (Generel Electric) Olxylyletnene 01xylylethene OIELECTROL III (Generel Electric). SELECTROL II (Sengemo), SAS 295 Mssekl), SAS 296 (Mssekl) EOISOL 11 (HeGrew Edison) Mixture of methyl eted dlphenylethene end propyldlphenyl Epoxy GEAFOL (Trefo-Unlon) F 101 (SWS Silicone) Polydlmethylslloxene F 190 (SUS Silicone) FARAOOL 100 (Asee) Polydlmethylslloxene Benzylneoceprete FORMEL If (ISC Cneatlcelt) Mixture of perchloroethylene end freons Freon GEAFOL (Trefo-Unlon) Component of FORMEL tf (ISC Chemlcels) Epoxy Isopropylblphenyl UEMCOL (Westlnghouse) L 305 (Union Cerblde) Methyleted dl phenylethene Polydlmethylslloxene Component of EDISOL It (McGrew Edison) PAO 13CE (Unlroytl) Poly 1-octene HONS 216063 B-l Appendix B (continued) Perchloroetnylene Poly 1-oetene Polydlmethyltlloxene PR (Gulf Oil) Propylb 1 phenyl SA$ 29S (Mtiekl) SAS 296 (Nmekl) SELECTROL I (Stnomo) SELECTROL It (Sengemo) SF 97 (General Electric) Trichlorobenzene UECOSOL (We*tfnghoute) WEHCQL (Wettinghaute) XFS-M69L (Dow) WECOSOL (Uettlnghoute), component of FOKNEL ff (ISC ChemlciTs) - PAO 13CE (Unlroyel), PR (Gulf Oil) DC 200 and DC 561 (Dow Corning), F lOl ind F 190 ($W$ silicone), L 305 (Union Cirbidt), SF 97 (General Electric) Poly l<octene Component of EDISOL II (McGrew fell ton) Ulxylylethane 01 xylylethme Mixture of dl-2*ethylhe*yl phthalate end trlchlorooenzene 01 xylylethane Po lyO Imethy 1t11 oxene Component of OIELECTROL II (Genertl Elec* trie), SELECTROL 1 (Sengamo) Perchloroethylene liopropylbl phenyl Butyleted monoehlorodfphenylether HQNS 216064 8-2 EPA Hearing on Proposed Regulations on PCB Transformer Use Statement of Dr. John Craddock, Chairman Chemical Manufacturers Association -- PCB Program Panel . _ - January 15, 1985 . Good afternoon, I am John Craddock, Chairman of the Chemical Manufacturers Association PCB Program Panel. CMA appreciates the opportunity to appear here today. CMA has actively participated in many EPA proceedings concerning PCBs. Following remand from the Court of Appeals in 1980, CMA conducted surveys for the Agency of PCB use in electrical equipment in the chemical industry and of inadver tent generation of PCBs in chemical manufacture. CMA has filed extensive comments in each of EPA's many PCB rulemak ings. We also worked jointly with the Environmental Defense Fund and Natural Resources Defense Council to develop the consensus proposal that EPA employed as the baBiB for its final rule on inadvertent PCB generation. Most recently, the Panel has submitted comments in response to both the ANPR and the current proposal concerning risks should PCB transformers be involved in fireB. As our comments detailed, CMA believes the incidence of fires involving PCB transformers has been greatly overesti mated because a few unusual incidents in recent years focused MONS 216077 2 public attention on this issue. Indeed, CMA's survey of transformer use in the chemical industry over the past 40 years found no fires involving PCB transformers. Our comments thus conclude there is insufficient information to justify new regulations on PCB transformers., especially those in use under controlled conditions in the chemical industry. This memiag -- rather than restating CMA's previously detailed position -- I would like to address the issues in this proceeding by posing several questions to the Agency and to other interested parties. The fundamental question underlying this proceeding is whether any of us can conclude with confidence that any of the wide range of measures being proposed by EPA or other parties will in fact reduce the number of transformer fires or the risks associated with such fires. CMA suggests the answer is no, and, accor dingly, concludes this regulatory initiative should be abandoned until considerably more information existB. EPA believes that a health risk exists when a fire occurs in the vicinity of a PCB transformer. It is con cerned that PCBs, dioxins and dibenzofurans will be released and spread through nearby buildings. It has thus proposed a variety of controls that EPA believes will make it less likely that fires will get out of control or that their combustion products will spread. As we have determined from talking to companies within the chemical industry, and as HONS 216078 3 EPA's economic consultants recognize, the expense involved in EPA's proposals may lead many operators to choose retrofill or removal even if such actions are not mandated. Thus, CMA believes that the practical impact of the regula tion will be a forced phaseout of many, transformers cur rently in use. Serious questions for which the record currently offers no firm answers are raised: -- Will fires involving transformers be more, or less, likely should any of these controls be imposed? -- Will the consequences of such fires be more, or less, dangerous? -- Will combustion products should a fire occur be more, or less, dan gerous? -- What risks will be posed by the storage and disposal of PCBs forced out by regulation? Without answers to these questions -- answers that are not at all clear today -- it is possible that the new efa con trols will create risks that are greater than any riskB EFA hopes to reduce. 1 would like to review briefly each of these questions to explain why CMA believes EPA may be embarking on a potentially counter-productive rulemaking. First, the record in this proceeding is woefully lack ing in evidence that any particular EFA solution will make it less likely that fires will occur. As is well known, PCBs were initially chosen for transformer dielectric use HONS 216079 -4- precisely because of their fire resistant characteristics. Further, electrical specialists have for many years been aware of the risk of fires with electrical equipment and have thus taken numerous precautions through the National Electric code and other guidance to minimize fire risks. For EFA, whose expertise in electrical matters is limited, to attempt to impose new regulations intended to reduce fire risks creates the distinct possibility that wise choices may not be made. As the many comments received by the Agency over the past two months demonstrate, EFA has in fact chosen electrical protective devices whose availability and efficacy are questioned bydfl*tfxperts in the field. At the same time, little data exist on the risks that alterna tive transformer fluids may contribute to the initiation or spread of fires. In other words, the best protection against fires involving transformers may well be continued use of FCB transformers as they are used today. At a minimum, the record does not indicate whether there will be fewer, or more, fires if new EPA controls are promulgated. Nor does the record indicate whether the isolation requirement will adequately contain the fire and combustion products for any period of time; nor whether there will be more, or possibly different, toxic combustion products as a result of EFA controls. Furthermore, some of the control measures suggested by the Agency could increase other fire risks. For example, HONS 216080 5 many building emergency systems depend on continued flow of electricity to allow occupants to evacuate. Deenergization measures that might be useful in limiting fires might at the same time hinder building evacuation. Experts in electrical design and fire safety should be making decisions in this area. - Third, EFA has focused all its attention on the pos sibility of FCB, dioxin and dibenzofuran contamination. Even if the proposed regulations were to reduce spread of these chemicals -- either through controlling fires involv ing FCB transformers or removal of PCBs from transformers -- the possibility remains that other combustion products will be formed during fires and pose their own toxic risks -- risks that perhaps will be more significant than those presented by dioxins and by dibenzofurans. We do know that the record demonstrates FCBs are not the source of dioxins, but we do not know whether the regulations would reduce dibenzofuran combustion products while increasing production and dissemi nation of other toxic compounds. Until all such risks are assessed, no assurance will exist that removal of FCBs will be beneficial. Finally, if the EFA regulations do in fact cause sub stantial PCB removal from transformers, we must be concerned with the fate of this fluid. As EFA's economic report recognizes, a three-year phaseout is not a practical alter native. Current approved FCB ininceration capacity in this MOMS 216001 6 country is already overburdened. What risk will be created by the necessary handling and storage of FCBs until such incineration capacity exists is yet another unexplored issue. In sum, the record reveals an alarming l.ck of data indicating whether the proposed controls will in effect reduce risks. We all can agree it is undesirable to have any transformers involved in fires and that it is desirable, should such incidents occur, that the fires be quickly con trolled and their combustion products handled. But, we must also recognize that EPA should not impose new measures to prevent such fires unless it is sure that those measures will in fact reduce overall risks, if alternative wiring or isolation protection creates new risks, or if alternative fluids increase the risk of fire or the production of toxic ~ combustion products, the nation will not be better off. As in EPA's prior PCB proceedings, the focus should be on development of reasonable controls to limit human or environmental PCB exposure. In this proceeding EPA has ventured far afield into complex questions of electrical system design, building emergency procedures, relative risks \ of a wide variety of chemicals under unusual fire condi tions, and, indeed, of fire itself. The resulting regula tory proposals thus raise more questions about potential new risks than they answer about the risks initially perceived. HONS 216082 7 CMA is concerned that the Agency may issue a regulation whose risks are greater than its benefits -- without even considering the costs imposed. We urge the Agency to recon sider and withdraw its proposal until it obtains a more substantial base of information upon which to compare the risks of various fire control alternatives. HOHS 216083