Document EYX9MNMBwQKXJdmGwy67yn1x

EPRI Electric Power Research Institute Topics: PCB Chemical analysis Fires Transformers Insulating oils EPRI EL-5262 Proiect 2028-11 Final Report July 1987 Partial Combustion of Electrical Insulation Fluids Prepared by Westinghouse Electric Corporation Sharon, Pennsylvania MONS 218048 REPORT SUMMARY SUBJECTS HazardousJtoxic substances / TAD: Substations TOPICS PCB Chemical analysis Fires Transformers Insulating oils AUDIENCE Environmental managers / Distribution engineers Partial Combustion of Electrical Insulation Fluids In partial-combustion tests, five dielectric fluids being used as in sulating materials for power transformers left no evidence of the kinds of contaminants PCBs create. Although certain of the fluids decomposed into gases that could be toxic in confined spaces, procedures are available for the required short-term fire protection. cPRI EL-5262J BACKGROUND Until the late 1970s, askarels containing mixtures of polychlorinated biphenyls (PCBs) were the most common liquid coolants in transformers located in high fire hazard areas. After concluding that PCBs posed a significant environmental threat, EPA required the removal ot a large number ot PCB-contaming transformers from service. Replacement insur ing fluids needed evaluation to ensure that Ihey met electrical, thermal, anc physical requirements and that they did not decompose to produce residual toxic trace materials when combusted or pyrolyzed. OBJECTIVE To identify the chemical compounds--particularly the toxic compounds-- formed during partial combustion of dielectric insulating fluids used in power transformers. APPROACH Investigators studied five fluids presently used as substitutes for PCBs in power transformers--tetrachloroethylene, trichlorotrifluoroethane. polydimethylsiloxane (silicone fluid), transformer mineral oil, and a hightemperature hydrocarbon fluid. The work included a literature search, a computer analysis of theoretical combustion equilibrium expressions to characterize the potential combustion products, and an experimental study The testing included two sets of combustion reaction experiments--one em ploying a modified thermogravimetric analyzer and the other, a stainless steel reactor. The protect team analyzed combustion products either by gas chro matography or by coupling gas chromatography with mass spectroscopy RESULTS The experimental studies revealed no detectable production of such com pounds as polychlorinated dibenzodioxtns, polychlorinated dibenzofurans or other polycyclic aromatic hydrocarbon compounds. Researchers concluded HONS 2180*9 that long-term contamination, such as that caused by liras involving askareis, would not result from fires involving these substitutes. Under combustion, tetrachioroethyiene produced chlorine and. <n the presence of atmospheric humidity, hydrogen chloride. Trtchlorotrifluoroethane produced hydrogen fluoride and various fluorocarbons during combustion The production of these gases is a concern in the event of a fire. However, suitable fire protection is available to provide adequate safeguards for the firefighters. EPRI PERSPECTIVE To build a database on potential substitutes as extensive as that con structed for PCBs requires intensive analyses. By determining the trace combustion products ot the five dielectric fluids examined, this study has added significantly to that database. The report indicates that there is little risk of long-term contamination from these substitutes and iden tifies the combustion products that could prove toxic in confined spaces in the short term. EPRI projects RP2028-15, RP2028-16. and RP2028-17. now under way. address a larger list of utility materials (liquids, solids, and gases) and examine trace by-products of critical materials Re lated EPRI reports include EL-4407, Maintenance and Handling of Perchtoroethytene-Filied Electrical Equipment; EL-4497, Arc Products of Transformer Insulating Systems Containing Tatracdloroatlrylana; EL-4503. State-of-the-Art Review: Combustion and Pyrolysis of PCS Substitutes; EL-4939, Literature Review of Pyrolysis and Combustion Products of Selected Utility Materials; and EL-5143-SR, EPRI Workshop on Substi tute Insulation for PCBs. PROJECT RP2028-11 EPRI Project Manager: Gil Addis Electrical Systems Division Contractor: Weatinghouse Electric Corporation For further information on EPRI research programs, call EPRI Technical Information Specialists (415) 855-2411. HONS 218050 Partial Combustion of Electrical Insulation Fluids EL:5262 Research Project 2028-11 Final Report, July 1987 Prepared by westinghouse electric corporation Manufacturing & Materials Technology Department Utility Delivery Systems Business Unit 469 Sharpsville Avenue Sharon, Pennsylvania 16148 Principal Investigator and Author C. C. Claiborne Prepared lor Electnc Power Research Institute 3412 Hillview Avenue Palo Alto, California 94304 EPRI Project Manager G. Addis Trsnsmission Substations P-ogram Electrical Systems Qivisson HONS 18051 ORDERING INFORMATION Requests 'or copies of this report should oe directed to Research Reports Center (RRC). Sox 50490, Palo Amo, CA 94303. (415) 965-4081. There is no Charge tor repons requested by EPRl member utilities and affiliates. US. utility associations. US. goverhment agencies (federal, state, and local), media, and foreign ngarozaTtons wun which EPRl has an information exchange agreement, On request. RRC wilt send a catalog of EPRl repons gloctoc tawr flMMirri immuM ao EPRl *r# '*gw*r#d sorbet morn* of EMnc Poww ft*wirch iwum. 're Cooynght t '967 Eloanc Power Atstvcn inaiuM. toc AM r*g## rtiorv*a NOTtCE This 'too* sroowod Oy m* org*nitattonisl nuned mow si an MCOun ol *or* iDoniorad oy tn icrnc Powti R#*#trcr irwm*#. Inc |EPR<1 Nfiihor rnffnb*r* i PRI tr> ognmi'or.*l nantwo Mn nor iny O+rten #ettoQ on 0*n*j or any of mom rl niwf any warranty snont or m*on#0 wsn rotOKI to rn* uU O' a^v information Sooararu* m*tnoa. or (XOCftSS 3:KJ0Md m rhis raoort or th SuCh uM rmV not miring# prrwMatV Own#g rgrrs or (0| assumes any I'aOtMies .Mfi r*sOOC1 10 in# u of Of for damage* resulting Irgm lh gM ol any nlo/msiion aaparatuS mttftOO or C'^IU OiOOSOO n this rtOOh Pr#0r#O CS( 'Vtsi.nignOjtS Eactr e Catoorauon Sn*ron vinu MOWS 218052 ABSTRACT Recurring fire Incidence Involving transformer! have led to e desire to squire Information about the Combustion producte that may be generated not only in situations of complete combustion, but also In Incidents when limited oxygen Is present and combustion le Incomplete or partial. Thle report details some results from investigations designed to Identify the products of Incomplete combustion of five types of dielectric fluids Commonly used in transformers, as alternatives to eskarals. These fluids ere tetrechloroathylsne, trlchlorotrlfluoroethsne, polydimethylalloxans (silicone fluid), mineral oil, and a high temperature hydrocarbon. The investigation of these fluids was begun with a theoretical study of the thermodynamic equilibrium over e range of temperatures end pressures. The exper imental program wee carried out In two separate studies, one which involved heat ing ths fluid from room temperature to 1000 C and another which used injection into a furnace chat was preheated to 1000 C. Ths elr level In each chaaiber was chosen to provide 70% and 301 of the oxygen for stoichiometric total combustion. None of the combustions produced detectable chlorinated or polycyclic aromatic hydrocarbon particulates. Detectable quantities of chlorine end hydrogen chloride wars produced from tetrachloroethylane at well at e small amount of dlchloroacetylene. Trlchlorotrlfluoroethana produced e number of fluorocarbon compounds as well as hydrogen fluoride. Silicone fluid produced substantial quantities of a solid material, believed to bs silicon dioxide end like the two hydrocarbon fluids, produced several combustible hydrocarbon gases as well as carbon dioxide and carbon monoxide. HONS 21S053 Ill ACKNOWLEDGMENTS The author gratefully acknowledge* tha contribution* of Mater*. R. D. Draasar, I Kacrmerek. S. Studniarci and E. Walah, tha original projact manager, and othar cachnical paraonnal. all of the Waatlnghousa Elactrlc Corporation, for aitiitinc In chair apaclflc araat of axparclaa In furthering thl* project. The technical guidance end aupporc from EFRI In thl* work, through tha office* o Dr. C. Addle, Projact Manager, we* alao clncerely appreciated end 1* heraby acknowledged. MOMS 2ISO54 CONTENTS SlCSUn 1 INTRODUCTION Psrapactlvs and Background Alms and Objactivas of cha Program Mithod* Outlina Tharaodynaisle Equilibrium Calculations Combustion RasetIon Evaluations 2 LITERATURE SEARCH 3 EXPERIMENTAL STUDIES Katarlals Equlpmant and Proesduras 4 THERMODYNAMIC EQUILIBRIUM CALCULATIONS 5 EXPERIMENTAL RESULTS Tatrachloroathylana Trlchlorotrlfluoroathana Silicons (Polydlasthylslloxsns) Trans formar Mlnaral Oil High Tamparatura Hydrocarbon 6 DISCUSSION 7 REFERENCES Paga i.j_ 1.3 1.3 1-3 1-a 2-1 3.1 3-1 3-1 4.-1 S-l S-l 5*3 5-10 5*25 5-25 6-1 ^ -1 HONS 218055 vil ILLUSTRATIONS EiiUIi 3-1 Modified Thermal Analysis Combustion Apparatus 3-2 Stainless Steel Combustion Chamber 3-3 Plston/Cspsuls Injection Mechanism 5-1 Tetrachloroethylene Direct Injection With Chlorine 5-2 Tecrachloroethylene Direct Injection With Dlchloroacetylena 5-3 Tecrachloroethylene Particulate Chromatogram 5-4 Chlorotrifluoroethylene from Trichlorotrifluoroethane 5-5 Dlchlorotetrafluoroathane from Trichlorotrlfluoroethane 5-6 Unreacted Starting Material from Trichlorotrlfluoroethane 5-7 Remnant Starting Materials from Trichlorotrlfluoroethane 5-S Chlorine and Starting Material from Trlchlorotrlfluoroethena 5-9 Tetrafluoromethane from Trichlorotrlfluoroethane 5-10 Dichlorodlfluoroethylene from Trichlorotrlfluoroethane 5-11 Chlorotrifluoroethylene from Trichlorotrlfluoroethane 5-12 Trlchlorofluoroathylane from Trichlorotrlfluoroethane 5-13 Tetrafluoroethylene and Fluorochloroethylana 5-14 Benzene In Coabuatlon Products of High Temperature Hydrocarbon 5-15 Toluana In Combuatlon Products of High Temperature Hydrocarbon 5-16 Benzene to Toluene Relative Ratio Page 3.2 j-u 3-1* 5-S 5-6 5-7 5-11 5-12 5-13 5-14 5-15 5-16 5-17 5-18 5-19 5-20 5-35 5-36 5-37 lx MOMS 218056 TABLES Tab It 4.1 Thermodynamic Rasults for Tetrachloroethylene 4-2 Thermooynamlc Results for Trichlorotrifluoroe thane 4-3 Thermodynamic Results for Polydlmethylslloxane 4-4 Thermodynamic Results for Saturated Hydrocarbon 4-S Thermodynamic Results for Aromatic Hydrocarbon 4-6 Hydrocarbon Analysis of Mineral Oil 5-1 Combutclon Products from Tetrechloroethylene 5-2 Cas Chromatographic Detection Limits 5-3 Combustion Products of Trlchlorotrlfluoroethane 5.4 Polydlmethylslloxane Combustion Products 5-5 Transformer Mineral Oil Combustion Products 5-6 High Temperature Hydrocarbon Combustion Products 4-2 4-2 4-3 4-3 4-4 4-5 S-2 5-4 5-8 5.21 5-26 5-30 HONS 218057 xi SUMMARY The object of chla reieerch was eo provide lnforaeelon on the chemleel compounds formed during partial or Incomplete combustion of liquid dielectric Insulating materials used In power transformers. This Information was requested by the Environmental Protection Agency (EPA) ea a first step In prosiulgatlng a rule concerning transformer fires. Five fluids were investigated as a part of this project. All of these fluids have been considered or are being used ea dielectric fluids in placs of eakarela or polychlorinated biphenyls (FCB's) In power transformers. Thase fluids are as follows: tetrechloroethylane, trlchlorotrlfluoroethane, polydlmethylslloxane (silicone fluid), transformer mineral oil and a high temperature hydrocarbon fluid. The research was conducted In three separate phases: a literature search, a computer analysis of theoretical combustion equilibrium expressions and an experimental study. The experimental study Involved two types of combustion devices In two laboratories and the combustion products ware analyzed primarily by ges chromatography (CC) end by gas chromatography coupled with mass spectroscopy (CC/MS). The results of the research revealed no evidence of the production o such compounds as polychlorinated dlbenxodloxlns, polychlorinated dlbanxofurane or other polycyclic aromatic hydrocarbon compounds. Certain of the fluids generated gases that could be regarded as dangerous In confined areas, the Importance of which should not be minimized In the event of a fire Incident. However, suitable fire protection equipment should be available to provide adequate protection in the short term. No evidence was found chat long tarn contamination such as can ba caused In a fire incident with transformers which contain eakarela will result from a fire Involving these replacement dielectric fluids. S-l HONS 218058 Section 1 INTRODUCTION PERSPECTIVE AND BACKGROUND Power dlitrlbutlon eremfomeri ere uaed to convert electrical power from e state of relatively low voltage and high current at the generating site, to a atate of high voltage and low current for trensnission to the uaer. where it ia again converted to the lower voltagea commonly uaed in buildlnga and hone a . In each of the converalon proceaaea, an energy loaa la aianlfeated aa a heat loae. For reliable operation, aafety conalderetlona and damage prevention, a tranaformer must be Kept cool. While smaller transformers, such as those found in electrical and electronic equipment may be air-cooled, power distribution transformers need a cooling system which can transfer the heat generated at the center of the trana former to the outside and to radiators where it may be removed from the trans former. At the same time, the material used in the cooling system must provide insulation between the trensformer colls. Two types of materials can provide the necessary cooling and insulating properties in large transformers: Dry or gaseous and liquid. Gas cooled transformers are nonflammable but often more expansive then liquid coolad onee, and they generally sre noisier In use, which limits their use in Indoor locations. There are several liquid coolants currently in use. Until the late 1970's. the most commonly used liquid coolants in locations with relatively high fire hazard belonged to the class of compounds known as askarels which contained mixtures of polychlorinated biphenyls (FCB's). The determination that these compounds posed a significant environmental threat led to their prohibition in new transformers at that time. Alternative dielectric flulda which nay be used in place of eekarals in electrical transformers are of relativaly recent origin. In the final rule lseued by the Environmental Protection Agency, published in the Federal Rsglater of August 25, 1982 (1), that authorized the indefinite use of certain electrical transformers 1-1 MONS 218059 containing PCB', it wa* concluded "that adequate substltutss ixlittd for FCB' s ir, indoor transformer locations from th* perspective 0f firs safety and elactricsl efficacy". In 1982, although EPA waa aware of the February 1981 fire incident in Binghamton, New York, involving a PCS-tran*former, it believed that fire* involving *uch tranaformara war* rare, iaolacad event*. However, th* May 1983 incident in San Franciaco and th* September 1983 incident in Chicago brought thia aasuntptlon into question. As a result of chess incidents, EPA revised its planning end Issued a request for information as a first atop in promulgating a rule concerning trans former flras (Federal Register. March 23, 198*) (2). Although that rule waa directed at transformers which may contain PCS'*, there have also been incidents such as th* BART tunnel fir* in San Franciaco end an underground fir* in New York City in which partial combustion products of mineral oil were suspected of harmful environmental interactions. These suspicions, coupled with th* growing public concern over combustion products from plastic seat cushions end other commonly encountered items, apparently led to an expanded request by the EPA in the advance notice of proposed rulemsklng, mentioned above In the advance notice, th* EPA requested information on th* partial and complete combustion products of several transformer insulating fluids. Ths fluids Ident ified were chlorinated hydrocarbons. fluorocarbons, high temperature hydrocarbons, other synthetic insulating fluids, mineral oil and silicones. Although son* information is available on th* products of complete combustion (defined as occur ring when there is an excess of oxygen present). information la lacking on the products of partial combustion (where insufficient oxygen is present for stolchLometrlc completion of th* combustion) (lii). The nature of many fire* is such chat partial or incomplete combustion is the predominant mod* of combustion. In many firea, th* flow of air (oxyg*n) is limited by restricted air flow into a confined area such *s * building, tunnel or vault. Also, th* race of burning can be so rapid as to consume the oxygen more rapidly than it esn be replaced by convection flow. For many materials, partial combustion may result in th* forma tion of toxic substances such *s carbon monoxide end other subatsncea not normally produced in complete combustion situations. 1-2 MONS 218060 AIMS AND objectives of THE program Thl* study was initiated to inveatlgtte tha partial combustion product* of aavaral of tha more common initiating liquid* uaed in power dlacrlbution transformers. The program wn reatrlccad to eoobuitlon aituatlon* only, l.a. in tha abeanc* of decomposition product* produced by an alactrlc arc. The primary objective of this program was to determine tha chemical compound* formed during partial and complete combustion of liquid dielectric insulating materials used In power transformer*. The secondary objective* of tha program were to study the effect of reaction conditions (e.g., tamparatura or oxygen availability) on the combustion product mix, and to identify, in particular, toxic compounds produced under these conditions. METHODS OUTLINE To attain tha project objectives, an experimental program, outlined in tha fol lowing, was carried out after an axtanalva literature search. Tha literature search indicated that the information desired was generally unavailable. The references found In this search were transferred to SCS Engineers and are a part of a separate project (RP 2023-12) end report by that organization (2), Thermodynamic Equilibrium Calculations In order to establish some indication of the types and levels of potential combus tion products from the liquid dielectric materials, theoretical thermodynamic equilibrium calculation* war* conducted for various oxygen l*v*la over a rang* of reaction temperature* from room temperature to 1500 K. Data wa* assembled from various sources, including the *JANAF (Joint Army-Navy-Air Force) Th*rmoch*mlcsl tables'', and input into a mainframe computer utilising a program named "CHEHEQ", In the Meetinghouse R & D laboratories. This proprietary program, similar to a computer program developed by NASA (). but with a larger date base, ha* bean uaed extensively to calculate: 1. Concentrations of chemical specie* in multicomponent polyphaa* mix ture* at any specified pr*aaur*(a) and temperature(a). 2. Concentrations and adiabatic flam* temperature* of chemical products formed in the combustion process. 3. Thermodynamic and transport properties of gas mixtures at squillbrlum for speclflsd pressure(a) and temperature(s). 1-3 HONS 218061 Gomhuatlon Rotation Evaluation* Two asperate sets of combustion experiments were conducted for each material. In on* lie. thermogravimetric analyzer (TCA) was modified to conduct tha combustion raactlon . Reaction product! wara trappad In cold trapa and fad to a gas chrom atograph with an alactron captura datactor for subsequent analysis. This equip ment and tha tachniquaa uaad in tha analysis will ba dascribad in a aubaaquant aaction of this raport. In tha othar experiment aac, a atalnlaaa staal reactor wax employed for tha combustion, and tha reaction products ware fad co a gaa chromatograph/maxs spec trometer for avaluation. MOHS 218062 _. 4 Section 2 LITERATURE SEARCH A literature search was carried out to locate and examine any available published Information on combustion products of the Insulating liquids, with partlculsr emphasis on partial combustion products. This literature search was conducted through the use of the HERAC service of the University of Connecticut. Approximately 100 articles vers Identified which daslt with th combustion of the materials of Interest. However, few of the abstracts of these articles Indicated that they contained Information of Interest to this program. Several papers ware Identified which dealt with complete combustion end pyrolysis of the materials Involved In this study, but only two (3.41 dealt with partial combustion situations and vara, therefore, of use In conducting this program. These arclclea considered combustion products of a high temperature hydrocarbon <RTEmpR) and polydimethyls11oxana (PDMS) with limited oxygen present. In the first (}), simultaneous animal toxicity studies ware carried out to evaluate the presence of trace amounts of toxic prooucts. A comparison of the two materials showsd that PDMS exhibited significantly laaa toxicity than the high tempetature hydrocarbon. However, toxic products other than carbon monoxide ware not Identi fied. The second study (4) was a kinetic study of the oxidation of PDMS at temperatures ranging from 220-330C and discussed the evolution of formaldehyde and formic acid as a consequence of s chain mechanism of oxidation. One further study () discussed the evolution of phosgene (COCI2) from tetreehloroathylena (C2CI4) and concluded that COCI2 was far lssa likely to be formed at dangerous levels than it was from carbon tstrschlorlda. Hydrogen chloride, which waa assumed to be generated by rsactlon with water vapor In the air, was MOHS 218063 pradominanc. Tha author* arguad chat tha prasenca of HC1 would cauaa axpoaura to any ochar toxic material* to ba mtnimizad bacauaa it would drlva paraonnal from tha ar*e dua to It* pungant charactariatic*. Tha information obtainad from this aaarch vaa ralaaaad to SCS Enginaar* and la raviawad in chair final raport for RP 2028-12 (> . HONS 2X8064 2-2 Section 3 EXPERIMENTAL STUDIES MATERIALS One fluid of each of five type* of insulating fluid wee selected for thl* study; 1. Chlorineted hydrocerbon--tetrechloroethylens (CjCl^), WECOSOL*. Weatinghouse Electric Corporetion; original supplier: Diamond Shamrock Corporetion. 2. Fluorocarbon--trifluorotrichloroethane (C2F3CI3). Freon 113&, supplier: E.I. DuPont de Honours. 3. Silicone fluid-`polydlmathylailoxane, Dow Coming 551 Silicone Transformer Fluid; supplier: Dow Coming Company. 4. Mlnerel 0U--WEMCO C^, Weatinghouse Electric Corporetion; orlglnel supplier: Gulf Oil Corporetion. 5. High temperature hydrocarbon--RTEmp&; supplier: RTE Corporetion. In eddition, the elr used in the combustion experiments wee Dry Credo sir (< 3 ppn HjO) supplied by Linde Division, Union Cerblde. EQUIPMENT AND PROCEDURES The combustion epperetue used in the experiaente conducted by the Heterlsls end Menufecturlng Technology leboretorles of Westlnghouse Electric Co., consisted of a Dupont 990 Theraogrevlaatric Analyzer with a 1090 Thermal Analysis Module for control end date interpretation. This instrument wee slightly modified to conduct these experiments se Is schaMtlcally 11 lustreted in Figure 3-1. The liquid sam ples ware pieced in e small boat prepared from platinum foil. Air was con tinuously admitted to Che chamber through e flowmeter which was adjusted to pro vide the desired amount for the combustion experiments. A trap at the exit port of the combustion chamber wee filled with ice water to contain the reacting materials in the chamber. The cold traps consisted of a slurry of isopropenol and dry ice (69C) on the one hand, and a slurry of isopentane end liquid nitrogen (-160C) on the other. Two replicate runs at each sir level were carried out for each material. The samples collected in these traps were then analysed in a Hswlatc Packard Model 5730a gas chromatograph. 3-1 MONS 218065 Fig. 3`1. Modified Tharaal Analyaia Coabuacion Apparacua (not co acaia). HONS 218066 3-2 A Dupont 850 Liquid Chromstogreph was uiad in an attempt to idantify tha con stituents of tha mineral oil and high taaparatura hydrocarbon. For ainaral oil tha bast separations wars achieved using 90% aschanol and 10% vatar in isocratlc* mode at a flow rata of 2 al/ain of solvent. For tha high taaparatura hydrocarbon, a gradient profile of catrahydrofuran and asthanol at a flow rata of 2 al/ain, achieved tha beat results. Tha apparatus used in tha coabuatlon sxperiaenta at tha Research and Development laborstoriaa of Westlnghouse Electric Corp., was constructed froa a stainless scael cylinder, surrounded by an induction heater as Is schaaatically Illustrated in Figure 3-2. Tha reacting aaterlal was Injected with a aicrolltar syringe Into the preheated canbusclon apparatus. Tha aaount of reactant was varied to achlsvs tha daslred level of air for combustion. Tha products wars collected In a saaple loop, which was quickly attached to the Luer-Lok connector Immediately to tha left of tha Luer-Lok valve In place of tha syringe aseambly. Tha loop wee than used to transfer the produces Into e Hewlect Packard 5985a gas chromatograph/mass spec trometer for analysis. During tha course of tha work, a modification to tha R&D combustion apparatus was made, owing to the high vlacoalty of the high temperature hydrocarbon end the difficulty of transfer with a syringe. A piston assembly wee constructed which allowed small aluminum ampules of tha reacting materials to be transferred Into the combustion chsmbar. The design of this modification Is glvan In Figure 3-3 end In subsequent mention of this modification In this report. It will ba rsfarrad to as the 'platon/capsule* method. The temperature of combustion was chosen after consultation with Dr. Richard Cann of the National bureau of Standards (2.). Although temperatures of combustion can often reach very high levels whan sufficient oxygen Is present, limiting the oxygen can reduce Chase temperatures slgnflcently. It is generally accepted that partial combustion temperatures are on tha order of 1000 C or 1273 K. Conse quently. tha tharmogravlnetrlc analyser was hasted from ambient temperature to 1000C at 30 par minute and the RID apparatus was allowed to stabilize at 1000C prior to tha Injection. * lsocratic - the mobile phase or mixture of solvents is unchanged in the proportion of each constituent during the separation. M0NS 21806? 3-3 V*l<r* (Off forint Nun I ?I In# Ti*t* Fig. 3-2. Stalnlaaa Staal Coiabuatlon Chanbar (not to acala). Fig. 3-3. Plaeon/Capaula Injactlon Hachantan. A davica for lnaarcing aollda and vlacoua liquids Into tha acalnlasa acaal eonbuatlon chanbar (not to acala). MONS 218068 3-4 Section 4 thermodynamic equilibrium calculations Tha results of tha thermodynamic calculations ara given In Tablaa 4-1 through 4-5 for a combustion temperature of 1100 K. For tatrachloroathyIans, a chemical combustion aquation analogous to that of a hydrocarbon in pura oxygan was ussd to establish tha lsvals of oxygan to be addad for tha partial combustion situation, although this combustion doaa not nacassarily occur. Tha aquation davisad for this combustion reaction is: CjCl^ + 2 0j -> 2 CI2 + 2 COj (4-1) Similarly, for tha chlorofluorocarbon, tha theoretical combustion can be proposed to follow tha process: 2 C2CI3F3 + 4 O2 > 4 CO2 + 3 CI2 + 3 F2 (4-2) Polydlmathylsiloxana must be treated in this situation as separata monomer units of the polymer chain for whichtha appropriatereactioncan be considered to ba: -(CH3)2SIO- + 4 O2 > 2 COj + 3 HjO + SIO2 (4 3} For tha mineral oil and the high tamparatura hydrocarbon, the exact composition of tha reacting materials was impossible to determine. Therefore, calculations ware made for an aromatic hydrocarbon and for a saturated hydrocarbon, uelng tha rationale Chat thaaa materials had compositions lying somewhere between these two extremes. Indeed, carbon-hydrogen analysis of tha high tamparatura hydrocarbon revealed that its hydrogen to carbon ratio was 1.92, and analysis of data provided by Gulf Oil Corporation. Pittsburgh, PA. (presented in Table 4-6), gives a hydrogen to carbon ratio of 1.71 for the mineral oil: i.a. both reacting materials consist primarily of saturated hydrocarbons whose hydrogen to carbon ratio would ba naar 2.0 (long aliphatic chain approximation). MONS 218069 4-1 Tsbla 4-1 THERMODYNAMIC RESULTS FOR TETRACHLOROETHYLENE Probsbls Products (? 1300 K C02 C0C12 CO C (solid) n2 With SOt Air (Mol. *) 1.121 0.025 19.9 6.72 44.4 With 20% A I'Mols *1 3.16 7.69 9.77 Tsbla 4-2 THERMODYNAMIC RESULTS FOR TRICHLOROTRIFLUOROETHANE Prohibit Products @ 1300 K COC1F C0F2 CTU CC1P3 CC12F2 co2 CO H20 2 C0C12 C1F Cl2 Cl CC13F With SOt Air (Hols t) 0.269 5.25 8.30 0.262 0.007 2.59 13.2 0.007 47.7 0.006 0.001 21.8 0.583 With 201 Air (Mol# t) 0.04 0.157 26.2 4.11 0.531 3.75 ........... 7.92 56.1 0.936 0.103 1-2 MONS 218070 T*bl* 4-3 THERMODYNAMIC RESULTS FOR POLYDIHETHYLSILOXANE Prob*bl* Produce* @ 1300 K CHft H2O 2 CO co2 SIC (*olld) C <olld) S102 Colld) SI (solid) SI3N4 (*olid) Wieh 80% Air (Mol* %> 0.0415 0.211 50.1 10.5 0.0254 1.24 22.5 16.5 ........... Wieh 20% Air (Mol* %) 0.0711 0,0004 53.9 0.0268 33.3 9.91 0.013 1.00 Isbl* 4-4 THERMODYNAMIC RESULTS for saturated hydrocarbon Probobl* Produce* 0 1300 K a h2o h2 ch4 C02 Ml eh 80t Air (Mol* *) 65.0 14,3 9.52 11.2 Wieh 20% Air (Mol* %) 17.0 1.02 73.6 1.93 6.5 MOWS 218071 4-3 Table 4-5 THERMODYNAMIC RESULTS FOR AROMATIC HYDROCARBON Probable Product* @ 1300 K -------------------------------------------------------- n2 h2o h2 CHa C02 With 80% Air --MPlt *> 65.4 2.12 23.5 0,04 10.8 With 20% Air fMol* %1 33.4 0.592 59.8 0.315 5.82 Although quantities of reaction product* war* calculatad for othar taaparaturas than 1300 K, It 1* unlikely that thermodynamic equilibrium will ba praaant at lowar temperature* du* to kinetic conaldaratlona. Klnatlc information on th* combustion reaction* rapraaancad by th* aatarlsl* of chaa* experiments doaa not appaar to b* svallabl*. Thus any conaldaratlon of th* data require* th* assump tion that thermodynamic equilibrium la valid *t hlghar taaparaturas; an assumptton that appaars to ba accsptad at temperature* approaching 1500 K (fl) . At tha last, thermodynamic calculation* can ba uaad to provld* an approximate anawar In ragard to tha qua*cion of tha praaanc* or absanc* of a particular product apacla*. Tha Information can than ba uaad to provld* guldallnaa for th* experimental avaluatlon of a combustion raactlon. For axaapla, chaa* calculations war* uaad to aatabllah tha probability of formation of chlorophanol* and poly chlorinated dlbanzodloxlna (FCDD'a) In tha coabuatlon of tatrachloroachylan*. It was found that tha concentration of thaaa apacla* present In th* product atraam would ba vanishingly small, on tha order of 10-30% at all taaparaturas in th* rang* of conaldaratlon; In other words, thaaa apaclaa war* vary strongly thermo dynamically unfavored, HONS 218072 Tabla 4.6 HYDROCARBON analysis of mineral oil (HPLC saparstlon followed by aass apactroaatry) Compomnt________________________ Saturatas Paraffins 11.4 Naphthanaa 1 Ring 16.3 2 Ring 16.8 3 Ring 12.2 4 Ring 10.0 5 Ring 3.9 Total Naphthanaa 59.2 Total Saturataa 70.6 Aroaatlcs Mono Di 18.5 8.8 Trl 0.6 Tatra 0.3 Arooatlc Sulfur 1.2 Total Aronatlca 29.4 Wc. i Sourca: Gulf Oil Corporation, Pittsburgh. PA. MON5 210073 4-5 Section 5 EXPERIMENTAL RESULTS The experimental results will be created In flva separate (action*, aach con sidering ona of tha fluids Involvad In tha experiment, Any dlfflcultla* encoun tered in tha experimental arrangements dua to particular fluid charactarlitlc* or product* will alfo ba addrasiad at thli point. TETRACHLOROETHYLENE Tatrschloroathylen* 1* a ralatlvaly low vlacoalty, high vapor praaaura liquid at roon caaparatura, which necessitated axpadltloua walghlng and charging to tha combustion chaabar In tha caia of tha tharmogravlaatrlc apparatus. On tha othar hand, tha low viscosity. In particular, vaa affactlva In providing a raady aaans of transfer (via syrlnga injection) Into tha othar coabustlon chaabar. In tha MMT chaabar conaldarabla problaaa wara ancountarad whan tha coabustlon was carried out at an air level corresponding to 70t of theoretical total coabujtlon. Datsctbl* amount* of chlorine wara generated and In addition, since tha chaabar was hsstad froa aablant teaparstura upward, tha presence of water on the lncarlor walls of tha chamber caused tha production of small amount* of HC1. This In turn caused corrosion of soma of tha aatal parts In tha TGA aaasuraaant systaa such as tha sample reference thermocouple which broke down at approximately 600 C. Sine* this component Is not aaaantlal to tha operation of tha Instruaant, It was omitted In tha next run. However, the second run caused enough deterioration of the other coaponsnea of tha apparatus to require overhaul of tha Instruaant. As a result, cha runs at 301 air wara postponed and Indeed war* not completed during the project. The results of tha analyses of tha products contained In the trap* *rs summarized In Tebls 5-1. In this end succeeding tables, run 1 end run 2 represent the rapllcet* run* that war* performed with aech material. Other than the expected chlorine and the normal constituents of #lr, only HC1 wa* noted In small amounts In ch* products. As was predicted In the thermodynamic equilibrium calculations, HONS 218074 ; i Tabla 5-1 COMBUSTION PRODUCTS FROM TETRACHLOROETHYLENE Run 1 Isopropanol-Dry- lea Trap 704 Air Nltrogan 65.625 4 Oxygan & Argon 15.341 Carbon Dioxida 1.320 Chlorina 16.232 HC1 1.483 Sun 100.000 4 Run 2 Isopropanol - Dry- lea Trap 704 Air Nltrogan 64.452 4 Oxygan & Argon 16.302 Carbon Dioxida 0.896 Chlorina 14.776 HC1 3.574 Sun 100.000 4 Run 1 Isobutana-Llq, H 704 Air Nltrogan 70.743 4 Oxygan & Argon 21.013 C*rbon Dioxida Chlorina 8.242 HC1 Sun 100.000 4 Run 2 Itooucana-Llq 2 704 Air Nlcrogan 72.445 4 Oxygan & Argon 19.187 Carbon Dioxida Chlorina 7.206 HC1 1.163 Sun 10D.001 4 MONS 218075 phosgene was noc detected In sny of th* trsp*. No enrichment of product* gen erated at lover temperatures could ba substantiated. Howavar, unreected starting material vaa racovarad in significant quantltiaa, both In tha trapa and In tha combustion chamber, Table 5-2 lleta tha datection limits for various gaaia expected Co ba found in thaa* and subsequent experiments. The R 4 D combustion apparatus vaa uaed to conduct thraa analytical runs with tatrachloroathylana. The tvo direct Injection rune gave chlorine as the primary product (see Figure 5-1). The instrumentation of tha mass spectrometer causaa a "tic" co be inserted at cha normally unused m/s position 1 when any signal from the spectrum swamps tha Instrumentation capacity, such as la tha case for ^Cl'*'. It is believed that tha other csrbon-chlorlna peaks are an artifact generated from an interaction betvean chlorine and carbon 'dirt* which occurs in tha mass spec trometer. However, It Is also possible that sufficient carbon is ganarsted from tha combustion reaction to produce these small quantities. A low level (approximately O.lt) of dlchloroacacylane was saan (Fig. 5-2) In one of tha direct Injection runs. Unraactad tatrachloroathylana was also racovarad in both of these runs. An additional sample run with cha platon/cepaula method gsva essentially tha same results with dlchloroacacylane again being found. No particulates were dateetad on tha glass fiber disk trap (Fig. 5-3). With the sample quantities used In tha experiment, this Indicates <3 ppm of haloganatad or polycyclic particulate*. TRICHL0ROTRIFLU0R0ETHANE The physical characteristics of trlchloroerlfluoroechana and tatrachloroathylana are quite similar. Tha same considerations apply to the handling end placement of trlchlorotrlfluoroathane In tha cosfeustlon chamber as ware outlined for tatrachloroathylane. Few problems wars encountered when this substance was examined in cha MMT combus tion chamber. However, the presence of water on tha Interior walls did causa production of HF which In turn caused a minor amount of etching of tha chamber. Precautions were taken to remove as many mecal parte from the apparatus aa possible, similar to cha previous experiment*. Table 5-3 summarises the analysis results for this substance. In this particular case, however, noc *11 pask* In the chromatograph from the electron capture 5-3 MONS 210076 Tabla 5-2 GAS CHROMATOGRAPHIC DETECTION LIMITS Caa Spaclaa h2 02 2 CH4 CO co2 c2h4 c2hs c2h2 Cj'a C0C12 HC1 H20 (1) Dataction Llalt ___ _____ 1 0.04 0.008 0.009 0.003 0.03 0.006 0.002 0.003 0.002 0.001 0.0008 0.0008 0.001 HONS 218077 5-4 R e la ttv * In te n s ity Fi*. 5-1. Tatrachloroathylana Dlract Inj action With Chlorln*. Ha** *pactrun of chlorlna a* obtalnad froa coabuatlon of tatrachloroathylana. Saturation of dacactor lndlcatad by atica at n/z 1. MONS 218078 5-5 in imp<nTMgii )iii[iili ll<* nH iljilil hlljjtil'ithpiw njt^lh ini|iii1 l^tii mi)......n^in nii|innw^in Miqim'Hii^ la oiCM.oaoAeiin.CNf 11.13771 i oo *o to Cl CLl 40 10 u. 10 *0 SO *0 AA-I.J loo 120 1*0 ISO ...........................................a ip *m I Maaa co Charfa Ratio (n/z) Flf. 5*2. Tacrachloroathylana Diract Injacclon With Dlchloroacatylana. Maas apactrua of dlchloroacatylana racovarad froa ona of tha diract Injacclon runa of tacrachloroathylana. MQNS 218079 5-6 R e la tiv e In te n s ity Maaa Co Charga Ratio (/*) Fig. S-3. Tatrachloroathylana Parcleulata Chromatogran. Hat* apactrua of parciculataa from pyrolyala of tetrachloroachylana. MONS 218000 5-7 Table 5-3 COMBUSTION PRODUCTS OF TRICHLOROTRIFTUOROETHANE Run 1 Isopropenol-Dry- Ice Trap 70* Air Nitrogen 68.406 t Oxygen & Argon 29.464 Carbon Dioxide Peak 1 0.074 Peek 2 Peek 3 0.034 Peek 4 HF 2.022 Sua 100.000 Run 2 Iaopropanol-Dry- Ice Trap 70* Air Nitrogen 66.023 t Oxygen & Argon 31.032 Carbon Dioxide Peak 1 ........... Peak 2 Peak 3 Paak 4 ........... HF 2.945 Sua 100.000 t Run 1 Isobucane-Llq 2 ____ 70* Air_______________ Nitrogen 66.741 t Oxygan & Argon 32.871 Carbon Dioxide Paak 1 Faak 2 Peak 3 Peak 4 HF 0.388 Sua 100.000 * Run 2 Iiobutene-Llq 70% Air Nitrogen 64.234 % Oxygan & Argon 33.452 Carbon Dioxide 1.528 Paak 1 Paak 2 Paak 3 Paak 4 HF 0.786 Sua 100.000 % MONS 218081 5-8 Table 5-3 (Cone.) COMBUSTION PRODUCTS OF TRICHLOROTRIFLUOROETHANE (Cone.) Run 3 Isopropanol-Dry-Ice Trap 301 Alt Nitrogen 41.861 \ Oxygen 6 Argon 17.369 Carbon Dioxide 10.693 Peak 1 13.524 Paak 2 0.070 Peak 3 0.431 Paak 4 0.034 HF 16.018 Sun 100.000 t Run 4 Ifopropsnol-Dry-Ice Trep 3J?1 Mr Nitrogen 45.250 \ Oxygen 6 Argon 19.461 Carbon Dioxide 9 652 Peak 1 11.236 Peak 2 0.050 Paak 3 0.232 Paak 4 0.068 HF 14.052 Sun 100.001 \ Run 3 Isobucane-Llq n2 30% Air Nltrogan 61.710 \ Oxygan & Argon 33 228 Carbon Dloxlda 0.258 Paak 1 0.396 Paak 2 Paak 3 Paak 4 HF 4.208 Sun 100.000 6 Run 4 Isobutana-Llq - n2 301 Air Nitrogen 64.753 t Oxygan & Argon 33.212 Carbon Dioxide Paak 1 Paak 2 Peak 3 Paak 4 HF 2.036 Sun 10D.001 \ 5-9 HONS 218082 detector could be identified. Certain of theee peeks undoubtedly could be esso* elated with various fluorocarbon compounds that are noc readily available for confirmation of the peak identities It was decided that the results of the R&D combustion apparatus would be used to identify these other products. The results from the R & 0 epparatus are somewhat inconsistent. Two runs were made In the stainless steel reactor using the direct injection procedure. The first run. In which the wlthdrewal of products wes carried out after 10 seconds, showed chlorotrlfluoroethylene and dlchlorotetrefluoroethana (Figures 5-4 and 5*5) However, over 99s of recovered geseous material was unreactad starting material (Figures 5*6 and 5*7). A second injection was made and the gas sample was with drawn after 10 minutes, A little chlorine was detected (Figure 5-3) but unreacted material was predominant. Figure 5*8 showe CFj'1' and CFC1 + which may indicate a trace of active fluorine, although that saens unlikely. Fluorine would be outside of the scanning range. This run did not detect the other two aeterlals seen in the first run. Four runs were made on trlchlorotrlfluoroathana with the plston/capaule technique. In this technique, products detected Included tetrafluoromethane (Figure 5-9), dlchlorodlfluoroethylene (Figure 5-10), chlorotrlfluoroethylene (Figure 5-11), trichlorofluoroethylene (Figure 5-12), tatrafluoroethylene and fluorochloroethylene (Figure 5-13). Again, no particulates of interest were detected, even in the run with the greatest variety of products, with the same detection limit as with C2CI4, 1 ppm. Many of these products may well correspond to tha results of the other combustion technique. Since it was not possible to acquire these aeterlals, confirmation is limited to the observation that some of these materials were predicted by the thermodynamic calculations. SILICONE (POLYD1METHYLSILOXANE) Tha results of the combustion product analyses for this material are given in Table 5-4. For the first time, a relatively large variety of products were detected. Again in this case, a large amount of unreactad materiel was present in the chamber end in the combustion boat at the completion of the experiment. 5- IJ MONS 218083 R e la tiv e In te n s ity i]ii aaHw^mnwjwl lijiai titbit CM.MOia iriuoaoc TtnLtNC Ilt.*ft07 ti cl n M-IIM to o to ao 100 i to no i to a. ^miy&r2^~ ~^at Mata to Charga Ratio (n/t) ^a Fig. 5-4. ChlorocrlfluoroathyIona fro* Trlchlorotrifluoroathana. Matt apacerun of chlorotrlfluoroathylana ts racovared fron conbviaelon of trlchlorotrlfluoroachane at 1000C (ca. 10 tac). MONS 21808* 5-11 R e la tiv e In te n s ity Fig. 3-3. Dlchlorotatrafluoroathana from Triehloroerifluoroachana. Has* apaccra Including chat of dLchlorocacrafluoroathana at racovarad fro* conbuatlon of triehlorocrifluoroathana ac 1000C <ca. 10 aac). HONS 218085 3-12 Fig. 3-6. Unreactad Starting Material fron Trlchlorotr1fluoroathana. Meet apactrun of trichlorotrlfluoroachana 11 racovarad fron conbustlon at 1000C (ca. 10 iac). 5-13 HONS 218086 Response Fig, 5-7. Hannant Starting Materials from Trlchlorotrlfluoroathana . Analytical araa counts for trlchlorotrifluoroathana and chlorotrlfluoroathyl*na at raeovarad from combustion of trlchlorotrlfluoroathana at 1000 0 C (ca. 10 sac). HONS 218007 5-14 Fig. 5-8. Chlorine and Starting Material Cron Trlchlorocrlfluoroathana . Mata spectrum of (primarily) chlorine aa recovered from combustion of trlchlorotrlfluoroathane, second run only (10 min at 1000C>. HONS 21-8088 15 R elative In te n s ity i iiii|hii i 18* nihiNH mnim m wnimtii -tmp hi>imi jnrmfnffmtn'iinfiii! iinpnn iimii* nr 1 I2B MB 160 Mas* to Charge Ratio (n/z> Fig. 5-9. Tetrafluoronethane from Trichlorotrlfluoroathane. Mate apactrua of tatrafluoronsthene aa recovered from pyrolyala of trlchlorotrlfluoroachana (with argon at n/z &0). HONS 218089 ?-!6 R e la tiv e In te n s ity i 188.; Fig. 5-10. Dlchlorodlfluoroachylana from Trlchlorocrlfluoro Chan*. Han ipaeerun of dlchlorodlfluoroachylana aj racovarad fro* pyrolyiis of crlchlorocrlfluoroachana. HONS 218090 5-17 Fig. 5-11. Chlorocrlfluoroachylana from TrlchloroerlfIuocd- achana. Mas a spaccrua of chlorocrlfluoroathylena as racovarad froa pyrolysis of eriehloroerlfluoroathana. MOMS 218091 :-.a R elative ln te n s lt FLg. 5-12. Triehlorofluoroaehylersa fro* Trlehlorotrlfluoroathana. Hat* apaccru* of rrlchlorofluoroarhylana as ricovetad fro* pyrolyali of CrlehloroCrlfluoroathan*. MOMS 218092 3-19 R elative In te n s ity i-ruww-z CH.oMocTMn.rNc n.m>> cz a r M-iW'i Fig. 5-13. Tacrafluorotehylana and Fluorochloroathylana. Maai ipacera of tatrafluoroathylana and fluorochloroachylana aa racovarad froa pyrolyala of Fraon 113. HONS 218093 5-20 Tabl* 5 -4 POLYDrMETHYLS ILOKAN COMBUSTION PRODUCTS Run 1 Isopropanol-Dry- Ice Trep 70% Air Nitrogen 78.83 Oxygen A Argon 20.3} Carbon Dioxide Hydrogen 0.82 0,02 Carbon Monoxide Me thane Ethane Ethylene Acetylene C] Hydrocarbona ........... Water Sua 100.00 I Run 2 Isopropanol-Dry- Ice Trap - - 701 Air Nitrogen 78.68 % Oxygen & Argon 18.39 Carbon Dioxide O 93 Hydrogen 0.32 Carbon Monoxide 0.49 Me thane 0.35 Ethane 0.01 Ethylene 0.08 Acetylene 0.05 Cj Hydrocarbona 0.01 Water 0.70 Sub 100.01 * HONS 218094 Tabla 5-4 (Cone.) POLYDIMETHYLSILOXANE COMBUSTION PRODUCTS (Cone.) Run I Isobutana *Llq. n2 70* Air Nicrogan 76.07 t Oxygan & Argon 12.53 Carbon Dloxlda 9.57 Hydrogan 0.04 Carbon Monoxlda 0.30 Ma chana Echana 0.02 Ethyl ana 0.04 Acacylana Cj Hydrocarbons 0.04 Uatar 1.39 Sub 100.00 % Run 2 Isobutana-Llq. H .. 70* Air Nicrogan 71.34 % Oxygan & Argon 10.35 Carbon Dloxlda 15.54 Hydrogan 0.18 Carbon Monoxlda 0.35 Mathana 0.86 Echana 0.01 Ethylana 0.07 Aeatylana 0.02 C3 Hydrocarbons 0.03 Uacar 1.25 Sun 100.00 t HONS 218095 *-32 Tabl* 5-4 (Cone.) POLYDIMETHYLSILOXANE COMBUSTION PRODUCTS (Cone,) Run 3 Run 4 Isopropanol-Dry-let Trap Iiopropanol-Dry-Ic* Trap Air_______________________ ________ 10.1 Air Nitrogen 82.446 % Nitrogen 83.43 t Oxygan & Argon 14.963 Oxygen & Argon 14.17 Carbon Dioxida 2.024 Carbon Dioxide 1.81 Hydrogan ........... Hydrogan ........... Carbon MonoxIda 0,469 Carbon MonoxIda 0.49 Mathana ........... Machana ........... Ethana ............ Echana ............ Echylana ........... Echylana ........... Acetylene ........... Acetylene ........... C3 Hydrocarbon* ........... C3 Hydrocarbon* --------- Water 0.099 Watar 0.10 Sun 100.001 % Sum 100.00 t HONS 218096 5-: 3 Table 5-4 (Cont.) FOLYDIMETHYLSILOXANE COMBUSTION PRODUCTS (Cone.) Run 3 Isobutane -Llq. n2 30t_ Air Nitrogen 83.151 I Oxygen & Argon 3.200 Carbon Dioxide 6.940 Hydrogen 2.036 Carbon Monoxide 2.696 Methane 1.659 Ethene 0.028 Ethylene 0.159 Acetylene 0.066 Cj Hydrocarbon# 0.040 Uater 0,024 Sum 99.999 % Run 4 Iiobutane-Llq . 30* Air n2 Nitrogen 82.163 t Oxygen & Argon 3.953 Carbon Dioxide 7.062 Hydrogen 1.913 Carbon Monoxide 3.152 Methane 1.423 Ethene 0.024 Ethylene Acetylene 0.173 0.052 C] Hydrocarbons 0 035 Uater 0.030 Sum 99.980 t 5-24 MOWS 218097 Above *11, significant amount of solid material was found in *11 tubing leading from ch quartz combustion chamber, This material was whit* and fluffy in appear ance and wa* most likely silicon dloxid* which la known to raault from combustion of polydlmathylsiloxana <2). Th* axtant of this material may wall hav* Influenced ch* detection of certain gases such es carbon aonoxida, which would normally be found in ralativaly smell quantities in the traps, by essentially plugging the exit from th* trap. In the RAD effort, two runs with th* direct injection method war* eonpLatad. Since no volatiles ware detected and because of th* relative difficulty of trans ferring this rather viscous substance, th* piston method was developed. However, again no volatiles were deteccad. The furnace tube contained a large amount of black, fluffy material. Whan the material was hasted in a muffle furnace at 1000 C, ch* substance left a white ash. This led to th* conclusion that the material was probably SiOj mixed with carbon. No other particulates of interest war* detected on the glass fiber collector. In this case, detectability would b* on th* order of <15 ppm, du* to th* different mass of starting material. TRANSFORMER MINERAL OIL Th* results from th* combustion of ch* mineral oil are presented in Table 5-5. Significant quantities of combustible materials such as methane and ethane, possibly resulting from cracking of th* starting material, ware found in both traps wnan ch* amount of air was limited to 30%, Indicating chat th* combustion was quit* decidedly not complete in this case, A surprisingly low amount of carbon dloxid* was found with both ch* 70% and 30% air levels. Owing to th* limited time and funding involved in th* project. RAD wss not able to complata th* experiments with this materiel. HIGH TEMPERATURE HYDROCARBON Table 5-6 show* that th* quantities of combustible materiel* present in the pro ducts *r* ralativaly lass than for ch* transformer mineral oil. Again however, they are present, with significant quantities of ethylan* being present. Once again a large amount of unreectad material, comparable to th* quantity found with polydlmathylsiloxana. was present in th* combustion chamber of th* chermogravlmacric analyzer. 5-25 HONS 218098 Tibi* 5-5 TRANSFORMER MINERAL OIL COMBUSTION PRODUCTS Run 1 Isopropanol-Dry- lea Trap 70% Air Nicrogtn 77.842 % Oxygen & Argon 21.444 Carbon Dloxld# 0.139 Hydrogen 0,023 Carbon Honoxid# 0.311 Kachan* Ethan* EChylan# Ac*tyl*n* Cj Hydrocarbon* Wat*r 0.241 Sun 100.000 % Run 2 I*oprop*nol-Dry'Ice Trap 70% Air Nltrogan 75.444 % Oxygan & Argon 22.821 Carbon Dloxld* 0.142 Kydrogan 0.038 Carbon Konoxld* 0.351 Kachan* Ethan# EChylan# Ac acylan# C3 Hydrocarbon* tfatar 1.204 Sub 100.000 % HONS 218099 5-26 Table 5 -5 (Cont.) TRANSFORMER MINERAL OIL COMBUSTION PRODUCTS (Cont .) Run 1 Isobutene-Llq 2 _________ ISA Ms, Nitrogen 77.917 t Oxygen & Argon 21.866 Carbon Dioxide 0.066 Hydrogen Cerbon Monoxide Methane E thane Ethylene Acetylene ........... Cj Hydrocarbons Water 0. 171 Sum 100.000 t Run 2 Isobutene-Liq "2 70% Air Nitrogen 75.626 % Oxygen & Argon 22.960 Carbon Dioxide 1.046 Hydrogen Carbon Monoxide Me thane Ethane Ethylene Ace tylene ........... C3 Hydrocarbons Water 0.371 Sua 100.003 % HONS 218100 5-27 Tabl* 5-5 (Cone.) transformer mineral oil combustion products (Cone.) Run 3 Isopropanol- Dry - Ic* Trap 301 Air Nitrogen 61.676 1 Oxygon 6 Argon 12.370 Carbon Dioxide 0.793 Hydrogin 5.691 Carbon Monoxid* 4.527 Mathan* 11.856 Ethan* 0.256 Ethylene 2 .444 Acetylene 0.130 Cj Hydrocarbon* 0.076 Vicar 0.181 Sub 100.000 1 Run 4 Xaopropanol-Dry -1 c* Trap 30% Air Nitrogen 59.667 % Oxygan 6 Argon 13.240 Carbon Dloxida 0.929 Hydrogen 4.989 Carbon Monoxide 7.352 Methane 10.452 Ethan* 0.275 Ethylana 2.684 Acetylene 0.096 Cj Hydrocarbon* 0.096 Water 0. 172 Sub 99.999 1 HONS 218101 5-2 8 Tabla 5-5 (Cone.) TRANSFORMER MINERAL OIL COMBUSTION PRODUCTS (Cone .) Run 3 IsobuCan*-Liq, N2 30l Air Nlerogan 44.338 % Oxygon & Argon 4.900 Carbon Dioxlda 3.745 Hydrogan 6.060 Carbon Monoxlda 6.512 Ma thane 25.477 Echana 1.091 Ethylana 6.691 Acetylene 0.426 Cj Hydrocarbon* Water 0.740 0.020 Sub 100.000 % Run 4 Iiobutene-Liq "2 30% Afr Nitrogen 43.433 % Oxygen & Argon 5.420 Carbon Dloxlda 3.343 Hydrogan 8.050 Carbon Monoxlda 6.456 Machana 23.567 Echana 1.350 Erhylana 6.869 Acaeylana 0.492 Cj Hydrocarbon* 0.980 tfactr 0.040 Sua 100.000 % 5-29 MOWS 218102 Table 5-6 HICH TEMPERATURE HYDROCARBON COMBUSTION PRODUCTS Run 1 Isopropanol-Dry - Ice Trap 701 Air Nitrogen 74.136 \ Oxygen & Argon 16,053 Carbon Dioxide 5.023 Hydrogan 1.340 Carbon Monoxide 3.032 Mathane 1.118 Echana 0.012 Ethylana 0.171 Acetylene 0.025 Cj Hydrocarbons 0.069 Water 1.025 Sum 100.000 Run 2 Isopropanol-Dry- Ice Trap -70% Air Nitrogen 71.010 % Oxygen 6 Argon 15.746 Carbon Dloxida 6.632 Hydrogen 1.178 Carbon Monoxide 3.052 Methane 1.187 Ethane 0.011 Ethylene 0.160 Acetylene 0.023 Cj Hydrocarbons 0.075 Water 1.126 Sub 100.000 % 5-JO HONS 218103 Tabl* 5-6 (Cont.) HICH TEMPERATURE HYDROCARBON COMBUSTION PRODUCTS (Cone.) Run 1 Iaobutan*Llq n2 70* Alt Nitrogan 65.135 t Oxygon 6 Argon 7.123 Carbon Dioxid* 5.559 Hydrog*n 3.753 Carbon Monoxlda 6.694 Ha Chan* 3.916 Ethan* Ethylen* Ac*Cyl*n* 1.674 6.010 0.202 C3 Hydrocarbon* 0.715 Uac*r 1.419 Sum 100.000 t Run 2 X*obuCan*-Llq. n2 J70* Air Nitrogon 67.920 * Oxygon & Argon 7.310 Carbon Dloxld* 5.536 Hydrog*n 2.961 Carbon Honoxld* 3.932 H*chan* 2.525 Ethan* 1.485 Echylan* 6.425 Ac*eyl*n* C3 Hydrocarbon* 0.242 0.634 UaCr 1.028 Sua 99.998 * MONS 218104 5-31 Table 5-6 (Cont.) HICH TEMPERATURE HYDROCARBON COMBUSTION PRODUCTS (Cont.) Run 3 Isopropanol-Dry- Ice Trap 301 Air Nitrogen 74.164 1 Oxygen & Argon S .092 Carbon Dioxide 3.073 Hydrogen 1.660 Carbon Honoxlde 10.324 Me chane 2.243 Ethane Ethylene 0.018 0.274 Acetylene 0.046 Cj Hydrocarbon* 0.019 Water 0.085 Sum 100.000 1 Run 4 Isopropanol-Dry- Ice Trap 301 Air Nitrogen Oxygen 6 Argon 75.105 6.886 Carbon Dioxide 3.456 Hydrogen 2.185 Carbon Monoxide 10.077 Mechane 1.876 Ethane O.017 Ethylene 0.163 Acetylene C3 Hydrocarbons 0.036 0.07 5 Water 0.126 Sun 100.000 HONS 218105 5-32 Table 5-6 (Cont.) HIGH TEMPERATURE HYDROCARBON COMBUSTION PRODUCTS (Cone.) Run 3 Isobutene-Liq n2 30% Air Nitrogen . 55.156 % Oxygon & Argon 6.134 Carbon Dioxido 4.533 Hydrogon 6.731 Carbon Monoxide 10-985 MoChano 7.926 Eehano Ethylone 1.696 8.126 Acotylono 0.60S Cj Hydrocarbon* 1.917 Water 0.192 Sum 100.000 % Run 4 laobutane-Llq - *2 30i Air Nitrogen 45.050 \ Oxygen 6 Argon 2.390 Carbon Dioxida 2.556 Hydrogen 5.981 Carbon Monoxid* 10 322 Methane 12.515 Ethane 3.859 Ethylene 12.254 Acatylene 0.407 C3 Hydrocarbona 4.643 Water 0.023 Sun 100.000 t MONS 218106 5 -- 33 The high viscosity of the high temperature hydrocarbon was yet anocher reason fo tha development of the piston/tepaul* mechod for Che R & 0 procedure* of combus tion. In a run made with chi* mac*rial, only evo product* war* detected In tha gaseous product analysis by mass spectroscopy, benren* (figure 5-14) and toluene (Figure 5-15). Figure 5*16 illustrates the relative proportions of these substances. Nevertheless, In absolute terns only a small proportion of th* starting sample, on the order of l-2, seems to be represented by these products No particulate material was detected with a detectability limit of approximate 10 ppm. MONS 218107 5-34 Fi(. 5-14. Banian* in Combuielon Product* of High Tamparacur* Hydrocarbon. Mass spectrum of banian* a* racoverad from py roly* is of high camparatur* hydrocarbon. HONS 218108 5-35 Fig. i-lj, Toluana In Coabuacion Produce* of High Tamparatur* Hydrocarbon. Maa* apactrum of column* recovered from pyrolyeie of high temperature hydrocarbon. 5-36 HONS 218109 Benzana ft- 363108. Respon.se 1 ft- 24366. ______ Toluana Tlaa inCo Analyaia (minut*) Fig. 5-16. Banzana Co Toluana Ralativa Ratio. Gas chroatographic traca of gaaaoua produces fro* pyrolysis of high Caaparacura hydrocarbon. 5-37 MOMS 218110 Section 6 DISCUSSION In summation, none of che combustions produced detectable chlorinated or poly* cyclic aromatic hydrocarbon particulates. The production of chlorine (end hydrogen chloride In the presence of atmospheric humidity) from tecrechloroethylene must not be minimized, but rather should be balanced against che non flammable nature of the materiel. Much che seme statement may be made about trlchlarotrlfluoroethene and its production of hydrogen fluoride and the various fluorocarbon materials. The absence of phosgene in cheee experlmente should also be noted. For the other materials, the presence of combustible materials in the product stream indicates Chat even in events with only partial combustion, a potentially dangerous situation may remain even after the heat source le removed. Combustion could expend upon opening a vault or chamber with the consequent edmiaelon of oxygen if the alee has not had sufficient time to cool. The production of gases such as chlorine, hydrogen chloride and hydrogen fluoride from certain of the fluids does form e concern in the event of a fire Incident. However, suitable fire protection equipment should be available to provide ade quate protection In che short term. No evidence wee found that long term contam ination such as can be caused in a fira incident with transformers which contain askarels will result from e fire involving the replacement dielectric fluids investigated in this study. HONS 218111 6-1 Section 7 REFERENCES 1. Environmental Protection Agency. "40 CFR Part 761, Polychlorinated Biphenyl* (PCB*) Manufacturing, Proce**lng, Dlatrlbutlon In Commerce and U*e Prohibition*; U*e in Electrical Equipment." Fader*! ftaateifil, vo1, 47, Auguat 25, 1982, pp. 34342-34360. 2. Environmental Protection Agency. "40 CFR Pert 761, Polychlorinated Biphenyl* (PCB*) Manufacture, Proceeding, Distribution In Commerce and Us* Prohibition*; Dee in Electrical Tremformer*; Advanced Notice of Proposed Rulemaking." federal Raelscar. vol. 49, March 23, 1984, pp. 11070-11033. 3. J. Llpovltc- "Fire Safety Properties of Some Transformer Dielectric Liquids." J. of Fire A flammability, vol. 15, 1982, pp. 39-55. 4. V. S, Pepkov et el, "Investigation of the Thermal Oxidation of Polydimethyl*lloxanes." Polymer Science USSR, vol. 19. 1977, pp. 962-976. 5. B. SJoberg. "Thermal Decomposition of Chlorinated Hydrocarbons." Svenak Keelsk Jidskrift. vol. 64, 1952. pp. 63-79. 6. B. J. McBride and 5. Cordon. Computer Program for Calculation of Complex Chemical Equilibrium Compositions. Rocket PerformanceIncident and Reflected Shocks, and Chapman-Jouauec Detonations. Cleveland, Ohio; National Aeronautics end Space Administration, 1971. SP-273. 7. Personal communication with Dr. Richard Cenn, National Bureau of Standards. 8. H- C, MoCCel, C. C. Willlea*. C. N, Satterfield. Thermodynamic Charts for Combustion Processes. New York; John Wiley & Sons, Inc.. 1949, p. 6. 9. Staca-of-cha^Art ^Review: Pyrolysis and Combustion of PCB Sub stitutes . Palo Alto, Calif.: Electric Power Research Institute, March 1986. EL-4503. HONS 218U2