Document Zn2XNjGj258K5ampv3r0jXr0Z

Electric Power Research Institute Topics PCBs Chemical analysis Transformers Pollution control Insulating oils Waste disposal EPRI L 3581 Project 2028 Proceedings June 1984 Proceedings: 1983 PCB Seminar Prepared by Electric Power Reaearch Institute Palo Alto, California MOHS 21**68 Proceedings: 1983 PCB Seminar EL-3581 Research Project 2028 Proceedings. June 1984 Atlanta. Georgia December 6-8. 1983 Edited by G Addis. Electrical Systems Division R V Komai. Coal Combustion Systems Division Electric Power Research Institute 3412 Hillview Avenue Palo Alto, California 94304 EPRI Protect Manager G. Addis Transmission Substations Program Electrical Systems Division MONS 214669 t r ORDERING I NfORWAT ION Reouests lor coo<e& of this report should be directed 10 Research Reports Cents r (RRC), 80* r>0490 Raio Alto GA 94303, (415) 965-4061 There is no charge lor reports reQuested by EPRi member utilities end affiliates U S utility associations. U S government agencies (federal state and local), media ano foreign organizations with which EPRi has an information exchange agreement On request RRC wni send e catalog of EPRi reports i R*ircn c*wsi* H*t*reou*rtoc (uptime** TiO Subtmwnt Cooyrigni t iW Electric tower Re*reh Mtututt Inc M right* i*wv*u NOTICE twi reoorrwii txiwatM by in* EMotnc Pbw*i Anewcn Iminul* me lEPfllt Nn*t W member* EBW nor eny eetisn some on miir benen in mekei eny warranty *ipt*i ot intHee with rewecl to in* ue* or my mfoimenon uoamiu* twthoo or proc*n attcKMve m mi* (Kbitw mi ucn um mey not mlrng* pnvtitiy om*o ngne g> (Di eetumei my i>*m>Hm with r**p*cl to in* us* ot or tor a*mig* mulling trom m* us* ot nr mrorminon uomttut rrwthoo or orec*** AscKXKI m ihi* raporr or tot i* r*tportion* lor (iMmenn m*0* or opinion* *iiv*mm by nOtvrau*i utno'l HONS 21A*70 >* ABSTRACT _ The 1983 EPRI PCB Seminar wai held in Atlanta, Georgia on Oecember 6-8, 1983. It was sponsored by EPRI's Interdivisional PCS Task Force. It presented utility workers with existing and emerging tech nologies for lower cost compliance with EPA rules and PCB environ mental concerns. The papers in this proceedings are divided into topics as follows; Background; Detection and Analysis of PCBs: Retrofill; Spill Cleanup and Blo-Degradatlon; PCB Destruction; PCB Fires and PCDF; and Decontamination of Mineral Oil. Technology In some of these areas such as Analysis, PCB Destruction, and Decontamination of Mineral Oil, is maturing. However, other critical topics lacking information, such as polychlorinated dlbencofuran (PCDF) formation, spill clean up, and replacement fluids, have arisen to replace them. Fully 501 of the seminar attendees suggested more work on PCDF in response to a questionnaire filled out at the meeting. The seminar must be considered a success since 501 of the more than four hundred attendees considered presentations to be too technical and 501 said "too much sales talk" and future plans must take this dichotomy into account, possibly by better separation of the subject matter and perhaps provision for attendance at specific sessions rather then the full seminar. in. < HONS 214671 r r EPRI PERSPECTIVE PROJECT DESCRIPTION This EPRI PCB seminar was held In Atlanta, Georgia, December 6-8, 1983. It was the second EPRI seminar on this subject, the first hav ing been In Dallas, Texas, In 1981 (ED-2572). The papers presented by EPRI contractors and Independent speakers encompassed a broad range of PCB management subjects. The subjects covered ranged from commercial processing of contaminated mineral oil to plans for the possible use of gene splicing for biodegradation of PCBs In spills. PROJECT OBJECTIVES The main purpose of the seminar was to inform utility personnel of the currently available means for complying more economically with PCB regulations. In addition, research work In diverse areas was pre sented to delineate potential sources for future assistance. PROJECT RESULTS The program varied widely from practical answers to esoteric research, and the audience varied broadly in background. As a result, comments varied from "too technical" to "too much sales talk," but the overall audience reaction was overwhelmingly positive. A possible course for future seminars would be to better segregate the technical and sales talks and then provide means for dally registration, if desired by the individual. Gil Addis, Project Manager, Electrical Systems Division Ralph Romai, Project Manager, Coal Combustion Systems Division HONS 214672 v r r PRPTACE Much of the purpose of this preface is covered in ths first two papers of the proceedings, carl Manger of Baltimore Gas and Electric thoroughly covers the background of the PCB problem In the first paper. In the second, Gil Addis and Ralph Komai, the editors of this proceedings, summarise EPRI's efforts in the PCB area and thair rela tionship to other programs in the field. In the last few months, new plans for EPA rule-making on PCB trans former fires have been announced. This has resulted In added presaure to develop Information on PCDfa replacement fluids, and cleanup of both apllls and combustion products. A workshop to encompass material of significance to the utility industry and EPA in the rulemaking is planned for late 1994. HONS 214673 vii r r ACKNOWLEDGMENTS - The editors received help from many sources. Narain Hingorani, chair man of EPftl's Interdlvisional Task Force on PCBs gave the welcoming address and contributed In many ways to the success of the seminar. We thank Grady Baker, Jr. of Georgia Power, our keynote speaker. Additional people who assisted at the seminar or in preparation of the proceedings were: EPRI Staff Barbara Cole Brooke Bldredge Jim Gillies (Consultant) Jacques Guertin Luther Kolarik Ellen Lanun Bill Shula vasu Tahllianl Tasia Toombs Tom York Others Carl Brenner - Georgia Power Olin Compton - Vepco Harry Oniahi - Commonwealth Edison Corey Trench - Consultant HONS 214674 IX CONTENT:, Section Page PART 1: BACKGROUND 1. PCB Update - H. C. Manger ......... ............................................ .. i-i 2. EPRI Update - Emphasis on PCDF - Gil Addis and Ralph Komai............................................................ 1-7 PART 2i DETECTION AND ANALYSIS L)F PC&S 3. Fieio Determination of PCB in Transformer Oil "CLOR-W-OIL~ Kit" - D. J. Fisher, T. O. Rouse and T. R. Lynn......... .. 2-i 4. CLOR-N-OU" Field Test Program - vasu H. Tahiliam................ .. 4-s 5. Interpretation of PCB FielO Testing Kits - E. J. Walsh.............. 2-7 6. Detection of PCbs by infrared spectroscopy - Robert J. Nordstrom....................................................................................................... .. 2-11 7. Field Determination of Aroclors Using an Automated Electron Capture Detector Gaa Chromatograph - J. E, Picker and H. N. Colby.............. ........................ ................................................................ 2-19 S. The Semiquantitativo Detection of Polychlorinated Biphenyls (PCbs) in Contamlnatea soiLs by Thin-Layer Chromatography -BernardPlechalak........................................................ 2-35 9. Portable Infrared Field Monitor for PCbs: Phase II Mark a. Denton and Mark K. Walker........... ........................................ 2-49 PART 3: KETRUFILl 10. The Utility PCB Dilemma - RetrofilL or Replacement? Thomas L. Forrester and Thomas H. Mil by................ ............................... 3-i 11. Cleaning Askarel Transformers Using Silicone Ketrofill and Adaorption Techniques to Remove Residual PCBs - Daniel F. Meyer and Edward A.Raynaert..................... .......................................... 3-9 12. C2C14 as a Substitute for PCRs - e. J. Walsh................................. 3-23 PART 4: bPILL CLEANUP AND BID-DEGRADATION 13. Clean-Up of Soils Contaminated with PCBs - Judith F. Kitchens, Laura B. Mangoba, George L, Anspach and Edmund A. Kobylinskl,...................................... ................................................. 4-1 14. Claaning of PCM Contaminated Soils - Leo weitiman......................... 4-15 xi HONS 214675 Section 15. The Cost of PCB Spill Cleanup*. A Question ot Degree John P. Woodyard, W, Corey Trench. Peter McCormick and Thomas Juhasz................................................... ...................................................... 4-25 16. Innovative Design Concepts for Prevention of Airborne PCB Contamination - Brian W. West, warren G. Hansen and. Weston A. Fenner,................................................................................... ............. 4-S9 17. Biotechnology Potential as a PCB Disposal option - An Overview - Thomas 0. Peyton and Eric G. Terhorst........................... 4-71 16. Bacterial Degradation of PCBs: Evidence of Distinct Pathways m Corynebactenum Sp. MB1 and Alcaliqenes Eutrouhus HB50 - Donna L, Bedard, Michael J. Brennan and Ronald Unterman ................................ ........................ ,,...,...................... .. 4-101 19. Composting for Degradation of PCBs in Soil - Jeneflr D. Isbister, George L. Anspach and Judith F. Kitchens.......... ........... 4-119 2U. Bacterial Degradation of Polychlorinated Biphenyls in Sludge From an Industrial Sewer Lagoon - Walter S. Kim, Adrienne M, Takacs and David E. Kuivinen ........................................... 4-135 PART St PCB DESTRUCTION 21. Deetruction ot High Concentration PCBs in a Utility Boiler - T. E. Siedhoff, C, A, Zale and H. E. Morris.................. S-l 22. Arc Pyrolysis of PCBs - J. K. Wittle, C, H. Titus and S. 0. ....................................................................................................................... 5-29 23. Chemical Cleaning of PCB Capacitors - Leo Weitzman....................... 5-33 24. Mobile Plasma Pyrolysis - T. G. Barton and J. A. G. Mordy,,.. 5-49 25. Energy and Byproduct Recovery From Chlorinated Hydrocarbons: Chloe-Chimie's VRC* Incineration Process for Polychlorinated Biphenyls - Joost Galley.................................................................................. 5-63 PART 6i PCB FIRES AND PCDF 26. PCBs, PCDFs ano PCDDe Resulting From Transformer/Capacitor Firss: An Overview - Paul E. desRoaiers............................................ 6-1 27. Polychlorinated Dibenzofurans and pcB-Transformer Flresi Toxic Hazards and Determination of Decontamination Guidelines - Thomas H. Milby and Thomas L. Forrester.................. 6-21 2B. The Search for PCB Health Effects - John F. Brown, Jr................ 6-35 29. Chemistry of PCDDa/PCDFa - J. Rooney Marsh and Jasenka zbozinek...................................................................................................................... 6-41 HONS 214676 ii rr Section page 3u. An Update on Analytical Methods for Polychlorinated Biphenyls (PCB), Polychlorinated Dibenzofurans (PCDP) and Polychlorinated Dlbenzo-p-Dioxlns (PCDD) - Fred L. DeRoos, Marcus Cooke, Judith E. Gebhart, Laurence E. Si Ivon and Peter J. Mondron......................................................................... 6-59 Jl. Detection of polychlorinated Dibenzofurans and Other Chlorinated pyrolysis Products in the Soot Formed in PC6 Fires - C. H. williams, Jr., C. L. Prescott and L. D. Garretson....................................................................... 6-69 PART 7: DECONTAMINATION OF MINERAL OIL 32. Solvent Extraction of PCS From Transformer Mineral oil T. Rouse, G. Addis, C. Walker and P. Way........................................ 7-1 33. Pilot Plant Studies for solvent Extraction of Poly chlorinated Biphenyl (PCB) - C. w. Hancher and M. b. Saunders....................................................................................................... 7-7 34. Statistical Survey of PCB Contamination in Substation and Distribution System Equipment Containing Mineral Oil Mark D. Saperstein and Edward J. Feeder............................................... 7-19 35. Transformer Risk Assessment and Option Analysis Patrick R. Herbert.............................................................................................. 7-33 36. iunOhio PCb Detoxification Experience - M. J antes Kozak.............. 7-39 37. Design of a Fixed PCB Storage Facility to Produce the Most Value for Cleaned ull - Louis Centofanti.................................. 7-49 Appendix A - List of Participants...................................................................... A-l xiii HONS 214677 part 1 BACKGROUND HONS 214678 PCB UPDATE H. C. Manger Baltimore Gas Electric Company This is generalization of the Paper which Mas given as a slide presenta tion. !t should be made clear that when 1 speak of Askarels, It Is a generic term including both PCB's and mixtures of PCB's with tr1-/tetrachlorbenzene to reduce vl scoslty. Once upon a time, many years ago, PCB was heralded as a wonder fluid to be used In electrical equipment because it was: Chemically stable Hot soluble In water and Not volatile Early In the 1940`s, electrical equipment manufacturers, because of Its non flammable property, started to use PCB In applications where personnel and property protection were paramount. It has acceptable heat transfer characteristics close to mineral oil and Is a reasonably good dielectric. Chemically, it Is developed by starting with a biphenyl ring and substituting the hydrogen atoms with chorine atoms. There are 209 possible ways to chlorinate the biphenyl. The ask are 1 we use is made up of only a few of these possibilities. PCB can vary from crystalline structures to a clear viscous liquid, which Is the askarel used In electrical equipment. Before updating PCB regulations, I would like to give a little history of the regulations. The problems should be self evident. 1`ve long since stopped think Ing there is a cure. Because of growing awareness of hazardous substance storage and environmental problems, the Toxic Substance Control Act was passed In 1976. 1-1 HONS 214679 Based on anything but firm data, environmental advocates convinced Congress to amend the Toxic Substance Control Act to specifically cover PCB. you may recall PCB was the only substance that was singled out by the Congress to be regulated. They simply outlawed PCS; Its use, its manufacturer, and distribution In commerce. Nevertheless, Congress passes TSCA In 1976 and appointed EPA the watchdog. Since then, we have been In a continuing dilemma as to how to deal with PCB's. One must realize that, with the swipe of pen. Congress legislated a fluid we had used with no obvious 111 effects Into a hazardous substance. However, the Toxic Substance Control Act allows two ways for continued use of PCB's: first, and the most attractive, fs If the use can be considered totally enclosed. Totally enclosed Is defined as presenting Insignificant exposure. The second way Is use authorization. EPA can authorize a use if It can be shown that such use does not offer unreasonable risk. These two terms are pretty Important. If something Is totally enclosed. It does not fall under T.S.C.A., but you may also continue to use PCB that Is not considered totally enclosed by an authorization, l.e., If EPA deems that It does not offer an unreasonable risk. (EPA determines both totally enclosed and authorized use). Exemptions are applied for by the user, who must show no unreasonable risk. Indicated efforts for improvement, and reapply for the exemption every year. On May 31, 1979, the First Final Rule was passed. You should remember that EPA looked at the use activities of PCB and deemed the capacitors and transformers as being totally enclosed. EPA felt that with what they knew of the equipment, It posed an Insignificant exposure to the environment. EPA also authorized a number of uses, some of which were ftR transformers, small capacitors, carbon paper, and, If I remember correctly, even servicing other people's transformers. There ware a few exemptions applied for under that May 31, 1979 First Final Rule. Another Interesting side light Is that electrical equipment seemed to be the major focus of EPA's Interest. Looking back to the good old days and the first set of 100 questions (by the way, there Is an updated set). May 31, 1979 established 50 ppm and Introduced the under 5D, 50*500, over 500 concept. HONS 2146BO 1-2 r r It covered: marking - record keeping - restrictions on repairs - disposal - basically Incineration was the recommended method - storage Inspection _ It also covered Improper disposal and certainly addressed violations which can be and has already been hazardous to some peopled pocketboofcs. The Environment Defense Fund (EDF) challenged the May 31, 1979 Rule on three grounds (1) They felt that there was no record for PA to establish a 50 ppm cut off: (2) EDF felt that EPA had no data to back up the claim that capacitors and transformers were totally enclosed, and (3! the authorizations that were allowed were Improper. The court ruled that transformers and capacitors were not totally enclosed and that EPA lacked substantial evidence to support a regulatory cutoff of 50 ppm for manufacture, processing, distribution in commerce, or use of PCB. Had the court's decision gone into effect, EPA said It would have caused a major economic Impact - I think we would have had to turn the lights out. How ever, a stay requested by EPA was granted and the famous Interim Measures were established. As a result of this ruling, EPA published what Is referred to as Rule #1 on August 25, 1982. If this Infers to you that there will be more than one rule, you are right. Number 1 deals only with electrical equipment and authorizes certain uses of PCB in electrical equipment and set 50 ppm as a cutoff for electrical equipment. Some of the main points of Rule #1 are: Prohibits the use of PCB transformers and PCB electromagnets posing a risk after October, 1985. Authorizes the use of all other PCB transformers for the remainder of their useful lives, with inspections. Take note - It says ''authorizes", where May 1979 considered a transformer totally enclosed. 1-3 ' HONS 214681 / There are Inspection requirements of frequency, retention, and availability of records. These Inspections could be reduced If we contained 1001 or retrofllled to $0,000 ppm. Authorizes the use of large PCB capacitors located In electrical substations and Indoor installations. Prohibits the use of all other large PCB capacitors after October 1, 1988. Here we see a string tied to authorization. Authorizes the use of all PC8-conta1 nlng mineral oil-filled electrical equl pment. Presumes circuit breakers, reclosers, and cable to be less than 50 ppm. Expands definition of electrical equipment posing an exposure risk to food or feed. Allows storage of large PCB capacitors and PCB-contamlnated equipment outside of Qualified storage facilities. Requires records of Inspection and history be kept for 3 years after disposing of PCB transformers. Defines Hd1sposa1a as Including leaks and spills but does go on to say spills, leaks or uncontrolled discharges, resulting from the unauthorized use (or storage) of electrical equipment shall not constitute a disposal viola tion, provided adquate cleanup measures are Inltltated within 4B hours after notice of the discharge. Does not address required extent of cleanup of PCB spills. Since Rule #1 only deals with a certain part of the PCB problem, you must use a combination of the regulations to totally comply. Again, 50 ppm Is the cutoff level In this electrical equipment rule. HONS 214682 1-4 r W Since the August 25, 1982 Rule deals with electric eQuIpment, what about the other Items and systems that have PCB In them? Since the 50 ppm cutoff was struck down by court, there are a lot of non-electrical things that now fall In the cate gory, containing PCB (even a molecule of PCS). One of the biggest problems is the Incidental generation of PCB's, a manufacturing process can create as part of the process. Although It took a lot of effort and manufacturing Khow-how to develop the askarel we use In electrical equipment, CPA has stated, that any time you have a carbon, chlorine and some heat, the possibility exists that you can create a polychlorinated biphenyl. As 1 stated earlier, all you have to do Is replace one or more of the hydrogens with a chorine and you have yourself a PCS - It Is not the PCS In our transformers, but It Is a PCB. Now since the possibility exists that you can Incidentally generate a PCB, and the regulations say you can't, chemical manufacturers and processes fell under question by EPA, Since Industrial representatives explained that there are some systems In which PCB may be created, but no PCB's are released to the environment, EPA set about to write Rule #2 dealing with closed and controlltd processes, and published this 12 PCB Rule on October 21, 1982. Late this year (August or September), a number of policy statements were released by EPA dealing with PCB Separation Methods, Residual PCB and Enforcement Liability for violations of Disposal Deadlines. 1 think this Just about brings us up to date. What's ahead? - A proposed Rule 3 which was signed Decamber 1, 1983, addresses by-product PCB's that are not excluded from regulation by Rule #2. The final Is scheduled for July l, 1984. The August 25 Rule Is In effect now. 1 have to tell you that the rule may be challenged - EDF considered appeal since they thought the regulation was not strong enough In the areas of PCB transformers. EE1, USWAG and other Interests have considered appeal because they felt the rule was too strict In certain other areas. MONS 214683 1-5 r' EPA had early stated there was no need for additional electrical rules, but It appears pressures of PCB problems In Binghamton, San Francisco and possible changes within the EPA organizations have suggested additional rule making, and. In particular, PCB transformers. They have set a tentative schedule which should result In some decision In 1965. October 1964 Is an Important date since any information regarding PCB trans formers which could ease the regulatory burden can be presented at the Hearing. Proponents of stricter rules will certainly be there. It Is up to us to gather Information which will help our cause. You recall I mentioned earlier In connection with the major points of Pule #1 that It did not address required extent of cleanup of PCB spills. This must be dealt with somehow; probably with a rule - I certainly hope with more than a policy statement. So you see out of the original May 31, 1979 Rule have come many. We're up to 13 already. Mo one can be sure when we will see the final - Final Rule. To add to our burden, efforts are, at this time, afoot In Congress through a RCRA re-authorization bill, to list PCB's as hazardous waste under that act. Among other things, this could possibly add more manifesting requirements. Impact on burning less than 50 ppm contaminated Insulating oil, and Impose shorter storage time restrictions. The Department of Transportation Is proposing to Incorporate changes In Its hazardous material regulations which now apply to PFB's transported In one package or container greater than 10 pounds. Under this proposal, the regulations would apply to quantities greeter then 1 pound. the object of this seminar, to paraphrase the Invitation, Is to provide us with accurate data to permit knowledgeable decisions. In looking over the pro gram, I feel every effort hes been made to do Just that. But this certainly Is not the end of the regulations; It's not the end of our problems, and It cartalnly should not be the end of the search for data to somehow resolve this PCB dilemma. MONS 214684 1-6 EPRI Update - Enphaait on PCDF Gil Addis and Ralph Komai EPRI Much of the work included in this EPRI update will be covered in detail in individual papers presented during this nesting, it is appropriate, however, to point out the range of this activity, but since pressure for solutions to PCDF problems is becoming more intense, we will spend more time on a discussion of new contracts about to start in this area. He do not expect to provide immediate solutions, but hopefully in one case will provide the bases for future analytical gains, and in another develop data that may provide decision guidance in the retrofill or replacement of PCB equipment. EPRI'a INVOLVEMENT EPRI's involvement in PCB problems can be explained readilyr Why? - Tremendous potential cost to the utilities. - Can we hslp lower it? what? - Develop front-end feasibiliy and multiple approach funding. - Not competition to others developing processes. The work at EPRI is coordinated by the PCB Intsrdivisional Work ing Group of which Dr. Narain Hingorani is Chairman. Rsprssented are thrss divisona. Coal Combustion Systems (CCS), Electrical Systems (ES), and Energy Analysis and Environment (EAtE). The range and scope of EPRl's program is very similar to the breakdown of sessions for ths seminar. ANALYTICAL INSTRUMENTATION Rapid and simple anslytical methods are needed for portable use in ths field. The work has been dividad into two areas, analysis in mineral oil and analysis in soil. For mineral oil, a portabls instrument mads by floriba was modi fied for screening oil samples by determining total chlorine. The instrument, based on x-ray fluorescence measurement, has been field tested by a number of utilities for over a year. More recently, a simple disposable kit for chemical analysis of mine ral oil has been developed. The CLOR-N-OILTn kit is currently being tested in the field. Nearing the field trial stage in a second generation of instruments is an infrared technique to directly read PCBa. It is anticipated that completion of these projects will bring to a close EPRl's efforts to determine PCB in oil. . 1-7 MONS 214685 For analysis in soil (askarel spills)* where rapid field analysis is needed during cleanups* several analytical methods are being investigated. The leading candidate is a dedicated portable electron-capture gas chromatograph developed by 5-Cubed. A second technique that appears to be useful for spill samples is the refinement of infrared spectroscopy work by CS Associates; this work was begun at Oak Ridge National Laboratories. DESTRUCTION OF ASKAREL FLUIDS AND CAPACITORS In the past* EPRI had been actively involved in the early test burns of PCB fluids and capacitors* i.e.* the sampling and analysis for the ENSCO test burns. Incineration has been proven satisfactory* and a number of incinerators, both land based and at sea, are participating in the PCB destruction business. The preliminary design phase of a dc arc furnace (Electro-Petro leum) for destruction of capacitors has been completed. A con tract is presently being negotiated for a large scale demon stration unit. Also in a current EPRI project, Chemical Treat ment of Capacitors, laboratory tests to destroy PCBs from cut-up capacitors have been successful in reducing concentrations to below detectable limits. Details will be discussed in a paper by Acurex. RETROFILLING AND REPLACEMENT OPTIONS Retrofilling is being attempted by some utilities to substitute suitable non-flammsble or low flamability fluids for PCBs. Replacement fluids considered have been; RTEmp Silicones Perchloroethylane 751 Perchloroethylene* Freon 113 251 mineral oil mixture The same fluids plus solid insulation are being considered as rsplacements for PCBs in newly manufactured transformers. The decision to retrofill rather than replace must be based first on economics (in many cases the greater efficiency of new trans formers significantly reduces the relative lifetime cost); followed by accessibility for replacement* public emotion and hysteria. In many cases the latter two transcend economics. EPRI has had projects with Meetinghouse to develop transformers based on perchloroethylene or a nonflammable 75* perchloroethy lene, 25* oil mixture. These may bs substituted directly for PCB transformers. The perchloroethylene issue has been slightly clouded by a controversy concerning long term toxicity* although it has been in common use in the dry-cleaning industry for many years. EPRI also has a project just starting to determine arc and spark by-products of two systems* perchloroethylene and the perchloroethylene/mineral oil mixture. The objective of this projeot is to determine whether there are harmful by-product chemicals to which a utility worker in a maintenance situation might be exposed, over and above the perchloroethylene, to which 1-8 MONS 214686 dry cleaning workers have had relatively high daily exposures. In a project with General Electric, a two-phase Freon-cooled trans former has been developed. However, current economics of this Freon-cooled transformer make it noncompetitive. It should be noted that for some time after even the most effective initial retrofill, there is enough PCB in-the trans former so that it is still legally classified a PCB transformer. Hore will be said about this when we discuss PCDP. DECONTAMINATION OF MINERAL OIL A number of commercial processes using a sodium reagent for removal of PCB from mineral oil are currently operating. Following an investigation of alternative processes, General Electric, under EPRI contract, has developed a solvent extraction process for separating PCB from mineral oil, reducing the dispo sal volume of PCB-contaminated oil by a factor of about 100. The process has been tested at 10 gallons per hour, and scale-up is being implemented. SPILL CLEAN-UP Spill clean-up divides into two categories. In the case of small ruptures or spills with potential public exposure, immediate cleanup is needed. The present procedure is to excavate contami nated soil and drum it for landfill disposal. Development of a poultice to remove PCBs from concrete has been attempted with erratic results by Franklin Research. For long-term widespread clean-up in areas of limited public exposure, such as service shop areas, manufacturing facilities, or landfills, slower methods of PCB destruction can be used. An Acurex project is attempting to reduce the excavation/transporta tion/landfill costs by developing soil washing as an alternative disposal technique. Instead of burying 99+1 dirt, the PCBs will be removed for disposal, and the soil can be replaced in the excavation. A patent is being pursued for this process, which we believe will have application to many different kinds of chemicala, such as dioxin. Attaway and Associates has done a literature survey and inter viewed commercial biotechnology firms for EPRI to evaluate bio technological processes for PCB clean-up and diaposal problems and to outline a research plan for pursuing these technologies. RI6K ASSESSMENT Risk assessment, along with risk perception is an area of great importance in relating PCB problems to the interests and fears of the public. It is enormously difficult because each group, such as businessmen, scientiste, or homeowners, perceives an entirely different risk for the same event. One obvious example is a Californian with a home straddling the San Andreas fault, oblivious to the potential earthquake dangers, but worrying about the 10 kVA PCB transformer on the pole down the street. ' HONS 214687 1-9 1* EPRX hat initiated aeveral projects in the risk assessment area, described below: 1. A project is in progress to establish a methodology for exposure assessment of PCBt and PCB partial oxidation pro ducts. Progress to date has been in consultation with utility experts to establish the focus for these models. Selection of problem areas will then govern the choice of models. 2. Work is about to be initiated on risk management methodologies to establish profiles of risk for PCB spills and PCB fires. 3. Decision Framework concepts for PCB are presently being explored with utility experts. 4. An occupational health risk assessment program for PCB and its oxidation products is in progress. An advisory committee from utilities has been assembled to evaluate the progress of the contractors. PCDF/PCDD Not covered in the present EPA regulations, but undoubtedly of highest priority in the future, are polychlorinated dibenxofuran (PCDF) and polychlorinated dibenxodioxin (PCDD). These are by products formed under some conditions of partial oxidation from PCB and tri-tetrachlorobenxene present in askarels. Animal toxicology studies have found that certain by-product compounds are considerably mors toxic than PCBs. An initial critical survey of the literature (including fires, chemistry, and toxicology) has bean made for EPRI by SCS Engineers and reviewed favorably by international dioxin experts. This document is now available from EPRI's Reaearch Report Center (CS-330B). Because of the current intensity of interest in PCB substitutes and PCB fire hatard, EPRI has a short-term contract about to start. It is well known that when a transformer is drained for retroftiling, even under the best conditions, 2 to St of the original liquid in the transformer remains behind. Thus, no mattsr what the retrofill fluid, the immediate reault is still, by regulatory definition, a PCB transformer. There are some retrofilling processes which further reduce the PCB content and permit eventual reclassification over a period of time. In this new project, laboratory tests will be made to determine the pyrolysis and combustion products of PCB at two levels of contamination in each of several potential retrofill fluids such as perchloroethylene, silicone, RTemp, and mineral oil. Informa tion derived here may help acertain relative hatards of fires involving retrofilled transformers containing some residual PCB level. This may then be compared with the same event in an askarel transformer. 1-10 MONS 214688 r r In a second, longer term project, we are trying to determine the boundary conditiona under which PCDF or PCDD might form In askarel or contaminated mineral oil equipment. It has been made obvious to ua In several Incidents that PCDFa/PCDDa form as a result of a fire In an askarel transformer, what happens as a result of long term overload or corona discharge In an askarel transformer? is there any detectable PCDF formed In the case of a contaminated transformer fire? A number of these scenarios will be Investigated. As a preliminary step, an'attempt will be made to Improve the present method of PCDF analysis. Several competent laboratories will make the preliminary analyses and assess the methodology before testing the selected field samples. Also, In the course of this project, one of the organizations will simulate a number of extreme field conditions using model transformers in the laboratory. CONCLUSIONS If you recall the seminar In Dallas two yesrs sgo, you will see some of the same problems and in some few difficult areas very little progress. However in general you will note that many of the problems have progressed beyond resesrch and development and have resulted in commercially available equipment or techniques. Many new problem areas have assumed greater importance along with those that have so far resisted satisfactory solution. We expect you to find the next three days to be filled with a diversity of current and pertinent information on the various aspects of the PCB problem. 1-11 MONS 21*639 Part 2 DETECTION AND ANALYSIS OF PCBfi HONS 214690 r r FIELO DETERMINATION OF PCS IN TRANSFORMER OIL -CLOR-N-OILm Klt- 0. J. Fisher and T. 0. Rout* General Electric Company Plttiflald. Nattachutetts T. R. Lynn Oextll Chamleal Corporation Hamdon. Connatlcut Atkaral It a generic ton* for a group of non-flammable tynthatlc chlorinated hydrocarbon! utad at electrical Intulatlng fluldt in trantformert and capacltori. Mott contain poly-chlorinated blphenyit (PCS). PCS have been Judged to be harmful In the environment and their manufacture, ut* and dltpotal have been tubject to Government regulation tinea 1976. Thlt regulation extends not only to atkarel filled trantformert containing PCS, but alto to the much larger number of mineral ol 1 --f 111 md trantformert which may have been contamin ated with PCS during manufacture or tervlc*. Oils containing lett than 50 ppm when taken from trantformert are defined at non-PCS liquid! and have no unutual limitation! on their handling and dltpotal (except that ollt contamin ated with any detectable PCS may not be uted In widely dltperted application! tuch at dutt control). Ollt containing 50 to 500 ppm PCS are contldered 'contaminated* and thoie containing more than 500 ppm arc contldered to be totally PCS. Handling and dltpotal of ollt In that* lait categorlei require! tpeclal technique*. Ultimately, each of the 35 million oil-filled trantformert now In tervlce In utility and Induttrlal application* may have to be totted for PCS content to attur* proper dltpotal. Mill* adequate analytical method* exltt for determin ing the PCS content of ollt In the laboratory, thlt procett It time-consuming. Sample* mutt be packed, tent to the laboratory doing tho analytet and the analytlt mutt b* performed. The aim of the work reported her* hat been to reduce thlt problem by developing technique* for the measurement of PCS In transformer mineral oils by non-chemists In th* field. A member of Instrumental approaches were evaluated and the most promising for Immediate field ut* was X-ray emission using th* Horlb* MESA-200 analyzer (1). This Instrument determines th* total amount of chlorine present In oil end. hence, gives *n Indirect measure of th* upper limit on th* PCS content. Field testing of th* MESA-200 Instrument firmly established th* concept of using th* totel chlorine content In transformer oil at an Indication of th* PCS concentration. It appears that th* chlorine screening approach can eliminate 50-70X of th* transformers tested from further concern. HONS 21*691 2-1 T* Initial coit of tha x-ray emission analyzar is relatively high and tha Instrumant It best transported by automobile. Recant attention In this program has focused on developing a smaller disposable means of screening PC6 by determin ation of chlorine content of transformer oil. This has led to the ClOR-N-OIlm PCB screening kit. The kit developed here Is based on the quantitative conversion of the chlorine atoms In PCB In transformer oil to chloride Ions which in turn are extracted Into an aqueous solution and measured colorlmetrlcally. The conversion of chlorine to chloride Ions Is done by sodium salts formed by naphthalene together with the dimethyl ether of dlethylene glycol ('dlglyme") as stabiliz ing ligand {) The chloride Ions are extracted Into an aqueous buffer solu tion and reacted with a carefully controlled amount of dissolved mercuric nitrate. Olphenylcarbazone Is added. If the mercuric Ion content Is greater than that taken up by the chloride Ion In forming slightly dissociated mercuric chloride, a vividly blue complex Is formed by tho excess mercuric Ion and the dlphenylcarbazone. If the chloride Ion content exceeds that taken up by the available mercuric Ion, the complex Is not formed and the reaction mixture remains colorless to pale yellow. The reactions are carried out In soft plastic tost tubes containing premeasured reagents in breakable glass ampules. Proper control of tho sample size and of the quantity of mercuric nitrate allows the blue-colorless response level In the mineral transformer oil to be sot at levels from a few ppm to several thousand ppm. The test Is performed In two soft sided polyethlyene tubes. The conversion of PCB to chloride Ion Is done In the first tube by breaking ampules containing naphthalene-dlglyme and then sodium dispersion Into the oil. Buffer solution Is added from the second tube, the chloride Ion Is extracted Into the aqueous phase and any excess sodium Is eliminated. A minor amount of hydrogen Is vented. A measured amount of the aqueous solution Is returned to the second tube containing the mercuric nitrate end the Indicator In separate ampulules. The color development process Is done by breaking these last ampules In turn. A convenient way of transferring the aqueous phase beck and forth Is provided by a flip top cap on the first tube. The tube Is capped and allowed to settle after extraction of the chloride Into the buffer. It Is gently Inverted so that the heavier aqueous phase moves to the bottom over the cap. the flip top Is opened and a measured quantity of buffer Is squirted back Into the second tube. When done carefully, virtually all of the oil phase Is retained In the first tube above a residual aqueous layer. Only the aqueous phase Is returned to the tube where the Indicator Is finally contacted and the color observed. The lowest retie of chlorine to PCB found In an askarel Is 0.42. Therafore, an oil sample containing lets than 21 ppm chlorine cannot contain more than 50 ppm PCB. A color response level set et 21 ppm chlorine for tho kit should assure that no contaminated oil gives blue responses. The chlorine to PCB ratio for ether askarels Is as high as 1.34. Oil saaples containing as little as 15.7 ppm askarol contamination can result In colorless responses as can samples containing other sources of chlorine contemlnatlon-but no PCB. Samples giving colorless responses can be checked by ges chromatography to determine whether they actually contain more than SO ppm PCB. MONS 214692 2-2 r r The standard davlatlon on kits In prototypical production was found to be <2 ppm chlorine at a response level of 18.5. The kit then can be quite reliable. Actual field experience is needed to Judge the overall performance of the kit. A field test 1$ being conducted by EPRI to provide this experience. This work was sponsored by the Electric Power Research Institute as part of Research Project 1713-1. CLOR-N-OR Is a trademark of EPRI. (1) R. G. Hirst. Field Oetemlnatlon Of PCB In Transformer oil. Final Report Part 1. Chlorine Analysis by X-Ray Emulsion. EPRI RP 1713-1 (in preparation); also J. N. HcQuade, PCB Analysis Bv X-Ray fluoroionce. Proe. PCB Seminar. EPRI EL-2572 (1982); also A. L. Schwa lb and A. Marquez, Salt River Project Experience with the Horiba Sulfur/Chlorlne-in-oil Analyzer. Proc. PCB Seminar, EPRI EL-2572 (1982). (2) J. F. Brown, n. E. Lynch. J. C. Carnahan. J. S. Singleton, Chemical Destruction Of PCB In Transformer oil. Oetoxlfication of Hazardous Hastes, p. 201. Ed. J. H. Exner. Ann Arbor Scl. Pub. (1962).; also T. 0. Rouse. Removal Of PCB From Transformer Oil. Proe. PCB Seminar, EPRI EL-2572 (1982). MONS 214693 2-3 Discussion What Is tha shelf lift of tho kits? A, Kits have shown no sign of deterioration In the 6 months now In use. There Is no reason to suspect that shelf vfe will not be quite long, but tine Is needed to establish the shelf life. Has a national policy regarding test methods been addressed? A. EPA does not endorse any method, but there Is no reason why the method can not be accepted. What is the effect of moisture on the test results? A. The solubility Is approximately 60 ppm. There have been no attempts to dry the oil and no problems have been experienced to that level. What is the cost/test? A, $4 to $6 depending upon quantities. Has the method been tried on hydraulic oil? A. Some, but test results are not available. Additives could create a problem. There was a request for a show of hands regarding Interest In a test kit at the 500 ppm PCS level. 15 to 20 people Indicated Interest. HONS 214694 2-4 CLOR-N-OILTM Meld Test Program Vasu H. Tehlllanl EPRI Palo Alto, California Eighty-seven utilities have participated In the test program, and so far (as of the end of November 1983) we have received test results back from 40 of them. This report presents the summary of these test results. A total of 879 oil samples have been tested both with the CLQR-N-OIL test kit as well as on a gas chromatograph. Whenever a CLOR-N-OIL kit user recognized that he erred on his test procedure, either through breaking the capsules In the wrong order or spilling some fluid, the test results were discarded for this sumnary report. Of the 879 samples, 424 tested negative (with a blue color at the end of the test) indicating less than SO ppm of PCB. The remaining 455 tested positive (turned clear). However not all of these samples were drawn randomly, since many users wanted to explore this test method in a narrow critical range of PCB contamination. This led to a larger percentage of oil samples testing clear. If we were to choose only randomly picked samples, the test results show 372 (Sit) blue and 352 (491) clear, Of the 355 clear samples 161 (221 of the total) did contain more than 50 ppm of PCS and 194 samples (271 of the total) gave a false positive test. This In comparison to SIX of the samples being eliminated from any further tests Is quite small, Of course one would like this percentage to get even smaller but In order to safeguard against contaminated samples testing negative, this is the best we can do. A total of 18 tests showed false negatives, l.e., the CLOR-N-OIL test gave a negative (blue) test, but a subsequent PCB test Indicated a contamination level greater than 50 ppm. While we continue to examine the reasons and retest many of these samples, we do have an explanation for many of the tests, A user who reported four false negative readings In a batch of 20 tests subsequently found that the PCS tests carried out by an Independent outside lab gave high PCB readings. The PCS test reported by the lab gave readings between S3 and 126 ppm, whereas the subsequent tests carried out on these same samples at General Electric tested between 2 and 11 ppm. There were two other problems early In the test program that were rectified through appropriate changes in the Instructions. One change was to show through a photo graph that If a test sample has a substantial percentage of PCB (above a few - hundred thousand ppm) the test sample being heavier than water wfll sink below the water and this test method will not work, Another problem was eliminated through the removal of a photo showing a test result ing from a sample contaminated with 40 ppm of Askarel 1242. Here the photograph showed a slight purple ring at the top of the water layer which """'h innk tlmllar 2-5 MOMS 214695 rr to a test sample If some oil was transferred to tube #2 (In step 4 of the Instruc tions). Since the Instructions were revised only two false negatives have been reported. Cone1us Ions: Through the use of the C10R-N-0IL screening test one can typically eliminate half of the oil samples from any further testing. The general experience, especially since the Instructions were revised, has been exceptionally good. This test method can offer a significant cost savings to the utility industry. Number of Utilities Participating Number of Utilities that have Reported to Date 87 40 Total Oil Samples Tested Blue Clear 879 424 (48*) 455 (525) Random Sampling Slue Clear 372 (51*) 355 (49*) False Positives 194 (27*) PC6 Contaminated 161 (22X) False Negatives 18 (2.3S) MOMS 214696 2-6 INTERPRETATION OF PCB FIELD TESTING KITS E. J. felsh . Westlnghouse Electric Corporation Sharon, Pennsylvania 16146 Due to the increasing number of PCB-in-oil field test kits appearing on the market, it may prove beneficial for those concerned to know the bases upon which these kits operate. There are two major categories of test types. They are either instrumental, such as the Horiba x-ray fluorescence instrument, or a chanical based detection method, such as the Manleh, Chlor-N-oil , or Centec kits. Regardless of the category, all these kits are based on the same assumption: that a means of detecting total chlorine can be useful In determining the presence of PCS'$ In oil It 1$ Important to recognize that none of these test methods, neither instrumental nor chanical, directly determine the presence of PCB's. The PCB molecules are just a very small number of potential organochlorides. Fortunately, transformer mineral oils contain only very small (less than 1 ppm) of naturally occurlng organochlorides The most probable types of such chlorinated species present In transformer fluids would be the chlorobenzenes used to dilute the arochlors. ' The Horiba Instrument technique is based on using x-rays to excite a1_l_ the chlorine atoms and to use the resultant missions fran the chlorines as a quantitative measurement of the total chlorine content. This method Is being denonstrated to be reliable and useful. However, the equipment necessary to carry out such an analysis is costly andetille transportable in a station wagon, is not readily portable. The chanical methods rely upon a chemical reaction to cleave all carbon-chlorine bonds and produce either chlorine, Cl2 or chloride ion, Cl*, detectable by sane analytical technique less costly and more mobile than x-ray fluorescence. The chlorine or chloride detection methods most commonly used are colorimetric, a colored solution whose appearance or Intensity is dependent on the chlorine concen- 2-7 HONS 214697 tratlon, or a chloride Ion specific electrode system. As previously stated, these chemical methods can only be used to "screen" for the possible presence of PCB's. You cannot use these techniques to either quantltatlvly determine exactly how much PCB 1$ present, or can these techniques be used as a means to Identify which particular PCB's are present. Also contributing to the confusion on Interpreting these techniques, are the pre sence of chlorinated benzene solvents and less than satisfactory analtyical tech niques. These PCB-ln-oll tests are also designed to be most useful In samples containing only mineral oil- Scrap oil samples or soil samples may be so frequently contami nated with "other" chlorine containing materials that a direct gas chromatographic PCB analysis Is frequently the most economical method for these matrices. With this background Information, It Is now possible to discuss the possible types of results and how Interferences will affect tha Interpretation of these results. I would first like to discuss chemical methods which utilize a colorimetric method for chlorine detection. There are two kits now available In this classification, Chlor-N-oll and Manleh. Chlor-N-oll uses sodium metal as a reagent, and Manleh uses a sodium organic salt. As a first consideration, the Manleh reagent has a limited shelf life and Is best kept as a refrigerated reagent. Both reagents, when fresh, will completely and rapidly dechlorlnate all PCB's and any other organo chlorides present In transformer oils. The Manleh test sequence then uses an aqueous soluble reagent which develops an orange-red color, the Intensity of which Indicates the quantity of chlorine present. The operator coaipares the unknown oil sample with two samples of known PCB content, 10 ppm 1242, and 35 ppm 1242. Because of the small color difference involved and because the chlorine content will differ significantly from 1242 to 1260, the 35 ppm was chosen as a "safe" margin for requiring further analysis. HIth the Manleh kit, there are certain Internal checks which should be followed to assume that neither tha dechlorlrating reagent nor the color reagent have lost their capacity to perform adequately. Tha Chlor-N-oll kit uses a different colorimetric reagent systm which Is again designed to give a color dependent upon the chloride content of the aqueous extract. HONS 214698 2-8 t > As long as the sod Ian reagent Is sealed front air and moisture, the reagent should have a long shelf life. It Is possible, but Improbable, that excessive mol sture In the oil sample would cause Incomplete dechlorination. A visual check of the sample should be made to assure that the oil sample does not contain free water along with the oil. Again, due to the Inherent uncertainties of total chlorine methods, a safe limit of 30 - 40 ppm 1242 Is used to Interpret the result fron this test. _ Our studies have Indicated that both these kits are unaffected by aging of the oil, sludge, or other expected transformer related contaminants. Only the presence of organic chlorides can cause errors, and these errors will always be on the con servative side. The most serious type of situation which could occur Is If the reagent does not completely dechlorlnate the PCB's. A known PCB concentration should always be run when using these types of kits to check your reagent strengths. Another caution Is to never use tap water to dilute or mix your samples as the chlorine from this source will radically change your results. The other type of kit now commercially available Is sold by Centac. This kit uses a chloride specific electrode to determine the chloride Ion concentration. Because the dechlorlnatlng reagent Is an organo chloride salt, It too has a limited shelf life. Additionally, the data obtained on the final chloride determination Is a number, rather than a color. It Is Important to member that this number is just as Imprecise as the results of the colorimetric tests. In fact, the final milli volt reading Is subject to Interference by many more factors than are the color imetric reagents. Electrode conditions, total Ionic strength of the solution, and Interfering Ions from hand and wipe papers are all factors which were determined to significantly affect the analytical results. Additionally, the very accuracy of the chloride Ion electrode determination can cause serious errors unless the proper PCB standards are used to calibrate the Instrument. An example of the type of error which can occur Is, If one uses a curve based on Aroda1260 as a pass/all criteria, then the final chloride In concentration, based on the Centec dilutions. Is calculated to be 14X higher than the corresponding 1242 solu tion. The effect that this lower chloride Ion concentration has on the meter read ing Is very significant and could cause the type of Interpretation error which 2-9 MONS 214699 should be avoided. In summary, It Is Important that the users of these kits realize that they are de signed only to screen samples. A result Indicating that chlorine Is present does not necessary mean that PCB`s are present. A negative chlorine result means PCB's cannot be present. There are Instances where Interferences could cause erroneous results, and Internal checks should be done frequently. The proper use and Interpretation of these kits can result In reducing gas chroma tographic testing by 50 to 60S. The net savings for such a reduction can be very significant. Additionally, these kits provide a rapid and portable means of evaluating potential problems for spills of oils In the field. ClOR-N-OIl Is an EPRI Trademark. 2-10 MONS 214700 DETECTION OF PCBs BY INFRARED SPECTROSCOPY Robert J. Nordstrom BATTELLE Columbus Laboratories 505 King Avenue Columbus, Ohio 43201 Polychlorinated biphenyls (PCBs) have been used as cooling and dielectric fluids by electric power utilities In the United States since 1929. Recent concerns for PCBs In the environment have created the need for utilities to measure PCB contami nation In transformer oils. Currently, the need exists to determine PCB content In the range from 50 ppm to 500 ppm In transformer oils. The standard procedure for performing these tests Is the method of gas chromatography. This method Is costly and relatively slow. Attempts to reduce the complexity of analysis and to Increase the number of sam ples which can be analyzed In a given time have led to the development of a number of specialized methods for PCB analysis. Several of these techniques are being reported In this session of the PCB seminar. The possibility of using Infrared spectroscopy for the detection of PCBs Is sup ported by the fact that tha commonly used askarels In transformers have very strong Infrared absorption bands as shown In Figure 1. This makes the Infrared technique an attractive method for PCB detection and analysis. As a result of these preliminary spectra. Battelle-Columbus Division proposed to construct for EPftI a field-usable Instrument for detection of PCBs In transformer oil based on Infrared absorption. Operational parameters for the Instrument Include: e Ability to measure PCBs In the range 50 to 500 ppm a Measurement time of 5 to 10 minutes e Ruqged, portable operation a SI mol a ooeratlon requiring no special skills of tha operator 2-ll HONS 214701 i 1300 tOO WO Wmtwmbtn Figure 1. Spectra of various Arochlors: a) 1260, b) 1254, c) trichlorobenzene. Before going Into an analysis of these spectra and a discussion of the instrument developed, it Is Important to discuss the basic principles of Infrared instru mentation and the methods by which data are collected. A typical Infrared Instru ment If shown In Flqure 2. Radiation from a globar source Is focused Into a device which cataloges the radiation Intensity at each wavelength. The device could be a dispersive Instrument such as a prism or grating monochromator, or It could be a NicheIson Interferometer. For this discussion It Is not Important to understand the operating principle of the device. It Is only Important to know that the radiation is separated Into its various wavelengths. The radiation exits the separator device and passes through the sample. Absorp tion of the radiation at certain wavelengths occurs In the sample. The radiation which remains Is focused on the detector. The energy at each wavelength Is recorded and stored for later processing. 2-12 HONS 214702 r WAVELENGTH SEPARATOR RAOIATION IN RADIATION OUT Figure 2. Components of an Infrared spectrometer. Figure 3 shows a typical spectrum recorded by an Infrared spectrometer. The absorption process which occurred In the sample produces a spectral pattern which can be recognized and used to determine the components which are In the sample. 90 1000 mi 1250 Frequency (cm'1) -- --f --h 11 10 9 8 Wavelength (um) Figure 3. Form of data collected with a transmitting spectrometer. The rules for absorption are simple. The strength of the absorption Is the product of the pathlength of the radiation through the sample, the concentration of the absorbing material, and r quantity called the absorption coefficient. That is. 2-13 HONS 214703 where k is the absorption coefficient, c Is the concentration, and t is the pathlength . Unfortunately, the experimentalist has control of only two of the three variables. Mother nature dictates the magnitude of the absorption coefficient for any given sample. To detect low concentrations of a given sample, the experimentalist can increase the pathlength through the sample and hope that the absorption coeffi cient is sufficiently strong to produce an adequate absorption to be detected. For the specific problem of Infrared detection of PCBs In transformer oil, then, the first step Is to determine whether the absorption coefficients for the various Arochlor mixtures are strong enough to give good detection at low concentrations, figure 1 shows spectra of neat samples of various Arochlors. In the spectral region from a 8 un to 12 urn there are very strong absorption bands which are ideal for detection of PCBs. Not only are the absorption features very strong, but they are also unique from one Arochlor to another. This makes Identification of the Arochlor possible by Infrared methods. Thus, It appears that the absorp tion coefficients are sufficiently strong to support Infrared detection of PCBs even at low concentrations. Of course, the absorption created by the PCBs Is only one part of the total pro blem. The other part Is the absorption created by the transformer oil Itself Figure 4 shows the absorption spectrum of a typical transformer oil. Unfortu nately, the same spectral region where PCBs exhibited strong absorption shows strong absorption by the transformer oils. Thus, the problem of detection of the PCB absorption must be done In the presence of strong Interferences. Several methods were tried In an attempt to eliminate the problem of oil Inter ference. The simplest of this Is known as spectral subtraction. In this method, a spectrum of uncontaminated oil Is used as a reference and Is subtracted from the spectrum of contaminated oil. If all goes well, the resultant spectrum shows the absorption features of the PCBs only. However, It was quickly dis covered that variations In transformer oils produced spectral differences which made the use of spectral subtraction Impossible. Other methods for reducing Interferences were tried. They Included chromatograph ic separation, chemical alteration of the PCBs, and chemical clean-up of the oils. 2-14 HONS 214704 r r Wmnmbm Figure 4. Transmittance spectra through various pathlengths of transformer oil: a) 100 urn path, b) 200 um path, c) S00 um path, d) 1000 pm path. Of these techniques, the chemical cleanup of the oil samples held the most pro mise for several reasons. It was felt that the operational parameters Including simplicity of operation, rapid analysis, and portability would be better served by this clean-up method than by the other methods of eliminating the Interfer ences. The purpose of the chemical clean-up Is to eliminate the various components of the oil which cause direct interference with the PCBs and which cause the wide variations In the spectra of the oils. Since the predominant component causing spectral Interference was the aromatic component of the oil. It was decided to attack that component In particular. A sulfonatlon process was devised which eliminated a sufficient amount of the aromatic oil to detect PCS. However, the process was difficult and required some hazardous materials. This was judged to be less acceptable for field operation than a second method. 2-15 - HONS 214705 * As a result, a different chemical method was tried. This method is a modifica tion of a process developed by Union Carbide for recovery of clean oil from PCB contaminated oil. In this process, dimethylformamlde (OMF) Is added to the con taminated oil to strip the PCBs from the oil. The OMF Is heavier than the oil and is easily separated from the oil. After extraction, the OMF is boiled away leaving a small residue of PCB enriched oil. This residue Is then analyzed for total PCB content by the Infrared method. - This method satisfies all of the operational parameters required for the system. It Is fast, simple, and easily transported. Thus, It was decided to design an automated system for processing the oils by this method. Figure 5 shows the design lay-out for the automated Chemistry Processing Unit (CPU). The unit automatically injects the required amounts of oil and OMF, separates the fluids, evaporates the solvent, and Injects the resulting sample Into the infrared sample chamber. The entire process Is computer controlled. A commercially available Fourier transform spectrometer manufactured by Nlcolet Instrument Corp. Is used to collect the spectra. Spectra are collected at 4 cm'1 resolution over the range from 400 to 4000 cm*1 {\ Z5 urn to Z.5 um). After the spectrum of the sample Is collected. It Is analyzed by a least square method to determine the concentration of PCB. The data base contains spectral Informa tion on Arochlor 1Z42, 1248, 1254, and 1260. Also Included are spectra of tri chlorobenzene and OMF. The computer associated with the spectrometer Is used to control the Chemistry Processing Unit and to perform the spectral analysis. At the present time In the program. Battelle Is evaluating the efficiency of the automated chemistry and the programs for performing the least square analysis. Analyses of system operation Indicate that the 50 to 500 ppm concentration range can be achieved with an accuracy of t 1 OS or better with a 10 minute processing and analysts time. Final packaging of the system will take place during the next few months. It Is expected that a prototype PCB detector will be delivered to EPRI In February of 1984. 2-16 MONS 214706 r r A Syringe Sample Holder 0 Solvent Measurement Chamber C 011/PC8 Solvent Mixing Chamber D Lower Phase Liquid Measure* ment Chamber Solvent (OMF) Evaporation Chamber F Vacuum Valve G Pressure Valve H Solvent Valve I Vent Valve J Drain Valve K Sample Line Valve L To Vent M To Drain H To Vacuum Pump 0 From Pressure P IR Cell 0 To Solvent Pumo Figure 5. Design lay-out for automated Chemistry Processing Unit. 2-17 HONS 214707 rr Discussion 01. How mobile Is the Instrument? *. The instrument cannot be moved while treating a sample, but when properly prepared fos shipping, there 's no problem. 02. How about the sample thickness? A. The cell Is fixed and never dl ssassembl ed, therefore sample thickness Is constant. 03. How does the cost compare to G.C.? A. The author did not know the cost of G.C. 2-1B HONS 21470ft r r FIELD DETERMINATION OF AROCLORS USING AN AUTOMATED ELECTRON CAPTURE DETECTOR GAS CHROMATOGRAPH J. E. Picker and 6. N. Colby S-CU6ED Owing to existing Government regulations concerning PCS containing materials, S-CUBED undertook a search for a reliable and portable Instrument to Identify and measure A roc 1 ors. Several approaches to the problem were evaluated with respect to four major considerations; namely: 1. The instnment must be transportable In order to perform field analyses. 2. It must require only limited operator experience or training. 3. Expected regulatory procedures for the determination of Arodors should be Involved. 4. The range over which the instrument can monitor PCB concentrations should be as wide as possible. To meet these needs, S-CUBED, with partial funding from EPRI, decided to develop a field transportable gas chromatograph with an electron capture detector called the PC6A-102, Although the concept of a transportable gas chromatograph is not novel, the PCBA-102 Is unique In thet It carries an onboard microcomputer to perform the data Interpretation which usually requires a trained analyst. The computer monitors the output of tho electron capture detector. Identifies the Aroclor and determines the concentration In the sample. Little operator axperlonce or knowledgo of data interpretation Is required. The operator prepares the sample using a simple sample preparation procedure and injects it Into the gas chromatograph. All other functions are performed by the microcomputer. This report describes the PCBA-102, how It works and how It Is used, and provldts comparison data obtained using the PCBA-102 and ASTK/EPA procedures to dotenalna PCBs in oil and soil samples. 2-19 HONS 214709 INSTALMENT Figure 1 shows the PCBA-102 front panel. On the left-hand side Is a removable oven module wnlch nousts the Injection oort, the GC colunn and electron capture detector, as well as the necessary heating nerdware. In tne center of the front panel are toggle switches for carrier ;;s snutoff and diversion to a rotometer for flow measurement. Once flow Is set, no adjustment Is necessary. Also placed on the front panel are five lighted pushbuttons and one light which permit the operator to U) supply power to the PCBA-102, (2) supply power to the oven heaters, (3) monitor the temperature, (4) select fete instrument calibration mode, and (S) to begin sample Introduction. Calibration Is required each time the Instrument Is powered up and after every tenth sample has been analyzed. This provides continuing checks on Instrument sensitivity and maintains . accurate reported results. Also provided on the front panel is a paper tape printer for output of analysis results. A block diagram of the Instrument In shown In Figure 2. The Instrumvnc requires a small cyllndar of Argon/Hethane carrier gas and an oxygen trap to ramova any trace of oxygon from the carrier stream. The gas and trap art connected to the rear of the instrument using standard Swegelock Qulex-Connect' fittings. This provides a very smooth and secure connection. Carrier gas enters the rear of the instrument and flows through the oven assembly for preheating to the injector port whore samples are introduced. The flow then proceeds throufb the column Into the detector and finally through tha toggle switch and rotometer. Signal from the detector Is fed to the electrometer wnlcn If connected to the Interface board. The Interface bord directs data coming from and to tne electrometer, the front pentl controls and the output printer, and the microcomputer. Figure 3 shows sample chromatograms obtained In the laboratory using a gas chrumatogrcph similar to the PCBA-102. These traces are plots of 2-20 HONS 214710 r r F1aur 1. PCM Front Pinal 2-21 HONS 214711 rr GENERALIZED SCHEMATIC OF PCSA-lOa Phii*tR Ov*i> Ar/CH, Carrier Q* Veal Figure 2. Block Oltgraa of PC8A-102 2-22 MOMS 21*712 r r CHROMATOGRAMS OP SELECTED AR0CL0R8 ISM iato Figure 3. Chroutoya* of tht Four Targat Aroclor Mixtures 2-23 HONS 214713 IT til* detector response versus tine for cannon Aroclors fPCB mixtures). Analog plots can be obtained from the PCBA-102 by connecting the analog signal output on the back panel to a strip chart recorder. The PCBA-102, however, "records" chromatograms in Its microcomputer. Subsequent to recording the chromatogram, it performs the chromatogram pattern recognition and quantitation calculations a laboratory analyst would normally b* required to do. The first of these activities Is to locate and Identify the peaks In the chromatogram. The PCBA-102 performs this function automatically using Its peak finding program. Once an analyst has found the peaks and has determined their retention times and areas, he would decide whether or not the "pattern" represents an Aroclor. The PCBA-102 performs this function using a pattern recognition algorithm. The algorithm uses th* peaks In the chromatogram and compares their retention times and areas to those of standard Aroclor patterns which are stored In microcomputer memory. The algorithm then determines a "MATCH* value similar to the COS * function encountered In multlvarlant statistics. Th* function Is a measure of how closely the sample pattern matches that of an each Aroclor stored In computer memory. For convenience this "MATCH" masher 1$ based upon 1000 representing a perfect comparison. "MATCH" values above 800 are nigh and Indicate good correlation between the sample and the reference Aroclor In memory. A comparison of "MATCH" values for tn* Aroclors Is shown In Figure 4. Th* low values attained for most of the comparisons Indicate that this calculation can be used reliably to differentiate among th* Aroclors. Once th* Aroclor has been Identified the program then proceeds to determine how much of that Aroclor Is present. The program calculates conctntretlon based on Individual peaks by comparing them to corresponding peeks In the stored standard. A statistical evaluation Is then carried out to reject peeks which are outliers presueebly resulting from Interferences. The remaining peaxs are used to report concentration on th* printer. 2-24 HONS 214714 r r AROCLOR 1016 1242 1254 1250 1016 999 770 <1 <1 DETECTED MATCH 1242 1264 740 999 2 <1 <1 8 999 18 1260 <1 <1 12 999 Flgurt 4. Calculated "MATCH" Values for Standard Aroclor Mixtures 2-25 * HONS 214715 The PCBA-102 can d* readily transported Into the field In a van or statlonwagon. Its power requirements are such that It can be operated either from line voltage or through an Invertor connected to a vettlcles electrical system. It consumes about 300 watts when heating from a cold start but requires only about 3D0 watts once It reaches operatl-nq temperature. The PCBA-102 measures 10 x 19 x 20 Inches and weighs approximately 55 pounds. The unit requires a carrier gas cylinder but can operate for several days a on small cylinder. The required gas Is comprised of 5 percent methane In argon and should be of the highest quality available. A two-stage pressure regulator is used to reduce pressure from the cylinder to the unit. The second stage of the regulator should be regulated from zero to 100 psl and Is used to adjust flow through the Instrument. In addition to the gas requirement, an oxygen trap should be connected between the gas cylinder and the PCBA-102. One trap Is provided with the Instrument and It should last for several months If gas quality Is good. The PCBA-102 and the trap are provided with Swagelock Quick-Connect^ fittings for easy Installation. APPLICATIONS The PCBA-102 Is designed to measure four Aroclor mixtures In transformer fluid and/or soils sample extracts. Sample preparation Is required prior to analysis to clean the sample and adjust concentration to the range required by the Instrument. A kit Is available in order to help perform these simple manipulations. Saj^rle preparation Is based on EPA end A5TM methods which Involve partitioning wltn hexene and concentrated sulfuric acid. When power Is supplied to the Instrument, the printer re^onds with the message that the power has been turned on and that the Instrument ovens are heating. Hun the ovens have retched temperature, a second messaqe appears Indicating that the PCBA-102 Is ready end that calibration is required. The front panel calibration button Is lifted to remind the operator of this. At this point the operator presses the calibration button and sample button, and Injects one mlcrulltur of the calibration standard. The calibration standard is a 2-26 MONS 21*716 r r solution containing 1 ppm of Aroclor 1254. The Instrument chocks to make suro that this Aroclor Is found In tho calibration mixture. Aroclor concentration Is calculated and placed In memory. This value Is used to calculate a response factor for tho Instrument at the present time. This factor Is used In quantitation of the next ten samples. In addition, the retention times of peeks In the ~ chromatogram are checked against stored values to make sure that the instrument Is operating properly. If retention times are not correct, then a message appears asking for flow corrections. If the Instrument Is unable to calibrate due to Improper operating conditions, no sample analyses can be performed. If the instnaaent was able to calibrate properly the necessary "READY" will be printed and a sample can be Introduced Into the PCBA. for transformer fluids a 100 uL aliquot Is Introduced Into a vial containing 10 aL of hexane and 10 ml concentrated sulfuric add using a repeating pipette. The vial Is then dosed and shaken vigorously. The PCBs remain In the upper hexane layer while Interferences go Into the acid layer. When the two layers have separated, one mlcrollter of the hexane is Injected Into the PCBA-102 using a 10 wl syringe. The amount Injected Is a constant and does not have to be adjusted by the operator. To analyze soils, an optional balance Is provided In the sample . preparation kit. The operator placet a vial containing e premeasurtd quantity of methanol on tne balance and tens the balance to zero by pnsslng a button. The operator then adds 1.0 * 0.1 grams of soil to the vial. The vial Is shaken and the soil allowed to settle. One mL of the methanol solution Is then Introduced Into e vial containing hexane and sulfuric acid using a pi pete. This vlel Is then shaken and a 1 -L aliquot Introduced Into the PCBA as previously described for the oil analysis procedure. The PCBA-102 automatically acquires the resulting chromatogram. Identifies the Aroclor from the peek pattern end performs quantitation based upon peek trees. This all requires approxlmataly eleven minutes from start to finish. 2-27 HONS 214717 rr The output of the PCBA-10Z Is t peper ttpe from the printer. This tepe output Includes the recognized Aroclor, the "MATCH" value Indicating reliability of the Identification measurement, the concentration of Aroclor found In parts per million and an assoclatad error with the analysis. An example of the output tape Is shown In Figure S. ' PERFORMANCE TESTS The PCM-102 was tested for performance In several areas. The first was a check of measurement linearity. A constant current electron capture detector should be linear over a very wide range, approximately five orders of magnitude, however, owing to software considerations and a need to limit the amount of material injected on the column, the practical range of the Instrument Is three orders of magnitude. Triplicate Injections of Aroclor 1254 at a level equivalent to 1,000 ppm In a sample yielded an average reported concentration of 921 ppm with a standard deviation of 39 ppm. Triplicate Injections at 100 ppm yielded an average reported value of 104 ppm with a standard deviation of 3 ppm. Triplicate Injections at 10 ppm yielded an average reported value of 10 ppm with standard deviation of 1 ppm. Figure 6 graphically depicts these tests and shows that the Instrument Is linear over the range of 10 to 1000 ppm. Based on this It Is clear that nonlinearity errors are small compared to errors assoclatad wltn sample preparation and possible Interferences In the chromatogram. The ultimata test of any Instrument Is how It performs In comparison with others. To test this several oil and soil samples were prepared and analyzed using EPA/ASTM procedures and the results compared with those obtained using the PCM-102. Tne results of these analyses are depicted In Table 1 for various transformer fluids. In each case the PCBA-102 correctly Identified the Aroclor and quantitation was very close for all samples. 2-2B HONS 214718 Analytical result Analytical result Analytical result Calibration request . Instrument ready -- Temperature readout Instrument powered up 24ppm /MATCH - 034 PCB : 1260 2pp 23ppm */MATCH 90S PCB i 1242 lpp* 22dd*i *f-- MATCH 944 PCB : 1234 l DP" CALIBRATION REQUIRED PCBA-101 READY 206 188 221 INJ COL DET TEMPERATURES <CJ OVENS HEATING UP PCBA-101 ENABLED figure 5. Printer Output Tape 2-29 HONS 214719 PCBA-102 LINEARITY Triplicate tnjectlona at 10, 100, 1000 ppm Reported Concentration (ppm) Figure 6. PCBA-102 Linearity 2-30 HONS 214720 "O TAB.E 1 analysis of transformer oil containing pcb comparison OF EPA/ASTM PROCEOUAE VERSUS PCBA-102 (valuos In part* por Million) IDENTIFICATION A5TM/EPA PCBA-102 1260 1260 1254 1254 1242 1242 1260 1260 1260 1260 1254 1254 1242 1242 1260 1260 QUANTITATION ASTN/EPA PC8A-102 242 276 35 34 51 30 473 364 55 37 400 470 242 244 35 30 2-31 HONS 214721 The instrument was iho tested using PCB contaminated soils In conjunction w1t> a two-step dilution. Methanol Is first used to extract PCBs from tne samples. This extraction is followed oy the hexane-sulfuric acid partition cleanup used for oils. The method was compared to a procedure using exnaustlve hexane extraction In a soxhlet apparatus followed by exnaustlve acetone extraction with the combination of the two extracts being measured. Table Z lists the results from these tests. In all but two cases the PCBA-102 Methanol procedure correctly Identified the Aroclor and quantitation agreed well. In one case the PCBA did not detect any Aroclor because the actual concentration was below the Instrument's 10 ppm detection limit. The second dlsageement resulted for a sample which seemed to contain both Aroclors 1242 and 1016 In a combination. The PCBA-102 chose 1016 as the closest matcn Aroclor and quantitated accordingly. Quantitation for this sample was very close by both methods. Also provided In on Table 2 are results from two EPA performance evaluation samples consisting of PCB contaminated sediments. The "true value" Is that provided by EPA who Indicated that these samples contained Interfering materials and that quantitation and Identification would not ba straightforward. Tht PCBA-102 was not able eo recognize Aroclor 1242 In the first sample. Bather It selected Aroclor 1016 whlcu has a vary similar pattern. The quantitation was vary closa. In the second case the Aroclor was correctly Identified and again tha quantitation was reasonably closa. It should bt pointed out that tha standard deviation provided by EPA for the 24 ppm sample was 11 ppm bastd on thalr analysts. Consequently the 3 ppm difference between the PCBA-102 end the "true value" Is excellent. SUMMARY The PCBA-102 computerized gas chromatograph Is able to accurately end reliably Identify end quantify Aroclors In transformer fluids end soils. The Instrument performs analytical chamlstry functions in tha fltld which usually raqulrt a tralnad analyst and compltx 1 nstrumantatlon. MONS 2-32 r 214722 r r TABLE 2 ANALYSIS OF PCB CONTAMINATED SOILS BY PCBA-102/METHANOL PROCEDURE VERSUS EXHAUSTIVE HEXANE SOXHLET EXTRACTION FOLLOWED BY ACETONE EXTRACTION IDENTIFICATION SOXHLET EXTRACTION PCBA-102 1242 1016 1016 1242 * 1016 1016 1016 1016 ND 1016 1016 1016 1016 1016 1016 1242 . 1242 1016 1242 QUANTITATION SOXIR.ET EXTRACTION PCBA-102 5 ND 95 34 41 12 13 40 62 110 176 215 323 "tru* v*lu" 24* 48* PCBA-102 21 36 NO . Not DttCt4. *PA ptrfomonc* voluttlon taapUt. 2-33 MONS 214723 v-r Discussion 01. How dry must the simple be? A. Field samples taken after a heavy rain have come out well. Spiked, wet samples have also compared very well with dry samples. 02. How about Identification of various Aroclors? A. The Instrunent Is designed for utility spill analysis, not for combinations of Aroclors. 2-34 HONS 214724 THE SIMIQUANTITATIVE DETECTION OP FCLTCHLORINATEP BIPHENYLS (PCI#) [N CONTAMINATED SOILS BY THIK-LAYtt CHROMATOGRAPHY Barnard Piaehalak Conaonvaalth Edlaon Coapany Haywood, Illlaola 60133 Saeclon 1 INTRODUCTION A* a conaaquanct of thalr psralatanct, thalr ablllry to accuaulaea, and thalr distribution throughout tha anvlronnant, polychlorloatad blphanyla (PCBa) hava bacons a wall known anvlroonantal pollutant- Bird and aolnnl atudlaa hava ahown that PCBa can laduca vartoua toxic dlaordara (X, i)- PCBa hava alao bacon* a aajor coaearo to *aa bacauaa of hla raaldanca at tha top of aoat food chain* whar* ha can concantrata #ubatantial anounta of PCBa. PCB-f Iliad capacitor* ar* potential aourcaa for PCB raloaao to tha aavlroanoncIf a polo-top capacitor rupturaa, PCB ralaaaad to tha area baoaath tha pola nu*t ba cnllactad and coatanlnatad ground ranovad. Howavar, at tha praaant tlna aanplaa aunt ba aant to a laboratory and analyaad bafora tha axta&t of contanlnatlon or tha aaouat of PCB ranalnlag aftar a daan-up oparatlon can ba datarnlnad. Off-alts aaalyaaa ara tlna cooaunlag and can dalay aita raatoratlon for nany daya. Mora lnportaat daring thaaa tlna dalaya, coatanlnatad ground aay ba lafe axpoaad to tha opan aovironnant whara It could ba vaahad or carrlad away. Thln-Layar Cfaronatography (TLC) nathoda hava pravloualy baao daacribad (2* A) which ara capabla of dacarmlnlog tract ltvala of PCBa la biological aatrlcat- Tha aaln goals of ehla work wars to imrattlgatt tha dataceloa of PCBa in contaaiaatad aoila ualog thlo-layar chronatography and to avaluata lta us* *a an aid durlnp claaa-up oparatlona- Poaaiblt TLC Intarfarancas Iron chior In*tad paatieldta or high concaatratlooa of nlnaral oil wars alao invastlgattd- 2-35 HONS 214725 rr Saction 2 EXPERIMENTAL APPARATUS ft Baker-flax TLC plata*, typa "Silicon Gal IB", wore uaad. Plataa ware 20cn x 20en with a coatlnB thickness of 200 ailcrona, and wara activated at 150C for IS minutes. Germicidal lamp, capable of emitting strongly In the deep ultraviolet region. TM Chromlat indicator aolutlon apray unit. A perkln-Elmer Modal 3920B gaa chromatograph equipped with a Nl6^ electron capture detector waa uaad. A glaaa column (l.B m x 4 m l.d.) waa packed with 3Z OV-lOl on 100-120 BMah "Gaa Chron Q". The column and detector wara maintained at 200C and 325C respectively. Tha carrier gaa waa 5Z methane In argon, and the flow was maintained at SO cc/mlnuta. REAGENTS "Silver Nitrate Indicator Solution." ACS Reagent Crada silver nitrate (2.0 grams) wara dissolved In 223 ala of denatured alcohol and 23 mla of distilled water. Tha aolutlon la atored In an amber bottle. All acetone, acetonitrile, hexane, and methylene chloride uaad ware Burdick & Jackson'a "Diet111 ad In Glaaa" grade. Aroclor 1242, Monsanto Lot #194 waa uaad to prepare tha 23,30,100,230 and 1000 ppm (ug/ml) standards In hexane that wara uaad to obtain tha date shown In Figure 3-1. All other specif It aroelora and paatlclda standards wara diluted In hexana from 100 ug/g aroclor and paatlclda standards obtained from tha Folyaclence Corp., Kit #390A. 2-36 HONS 214726 r f PROCEDURE Soil Extractlonn Individual Maple* of contaminated soil (25 grams) ere placed Into separata 500 nl (lea* jars equipped with alualnun lined cap*. (If tha aoll la vary vet, one can add 10-20 |raaa of anhydrous aodiua aulfate to absorb tha rolstura.) Approximately 25 ala of hexane ara added to tha jar to saturate the sanpla. An additional 20 ala of hexana ara than added, and the soll/hexana mixture la shaken for two minutes. The extract 1* than decanted and auction filtered through a (leas fiber filter. Tha ease sample of soil la then extracted a second tins with another 20 mla of haxana. This extract la sleo decanted and filtered, and then combined with the flrat extract. Tha final voluna la adjusted to 50 ol in a traduated cylinder. Thia final solution represents 25 grams of aoll In 50 ml of hexana. TLC Determination An activated TLC plats 1* marked with a pencil line 5 cm from tha bottom. Each soil extract (40 mlcroliter* of sample) la plecad on this line, asperated by a minimum of 1-1/2 cm. Each extrace la spotted, applying only a 2-4 mlcroliter portion of tha sample et a tine and allowing tha epoe to completely dry batvean applications. To each plata 20 mlcrolltara of a 50 PPM arodor 1242 standard la also spotted to enaura complete plata development. Tha plata 1* developed In e chromatography jar in (Lt4) acetone/hexane, allowing tha solvent front to ascend 3/4 the length of the plata. Tha plata la removed, air dried, and sprayed with the silver nltrata Indicator solution. It la then expoaed for ten (10) minutes to the ultraviolet light of tha garmleidal lamp at e distance of approximately 5 cm, above the bulb. Any PCB present will develop aa a dark spot with an Rg value of roughly 0.50. If the 50 PPM standard la not raadily visible, tha plata ahould be eotpoeed to aMonle vapor, which lightens the background in preference to tha TLC spots; It la then ra-axpoaad to tha ultraviolet light for ten (10) mlnutee. 2-37 HONS 214727 ft The relative darkneas and alia of a davalopad apot la ralatad to lta PCB concentratlon. By ualni twlca tht volume of sample extract as atandard solution, one can compart the sample apota directly to tha SO PPM atandard spot. If the aempla apot alia (langth In an) la leas than tha 50 PPM apot alia, or not detectabla, tha tampla contains laaa than 50 PPM of PCB (on an aa received baeie). If the eaaple apot alia la larier than tha SO PPM atandard, chan one eatlaMtea sample concentra tion from a atandard working curve. The atandard working curve can be prepared by spotting and running 20 mlcrolltara of a 23,50,100,230,500 end 1000 PPM atandard*. 2-38 MONS 214728 r r Section 3 RESULTS TLC PROCEDURE - FIELD RESULTS A atandard working curve for PCBs In the concentration ranga of 20 to 1000 PPM for the TLC method wea conatructad aa a aamilog plot of PCB concentration (PPM) va. apot alia (length in aa). Thla aamilog plot waa found to be approximately linear over thla range (aee Figure 3-1). Baaed upon the working curve and an experimental uncertainty of 1 an for tha apot alia, the PCB concentratIona for ten poaalbly contaminated aolle from two iltei were atin*ted in the field following a capacitor aplll. Theae reaulta ware then compered to the ECD-gaa chromatography reaulta obtained In the laboratory (aea Tablee 3-1 and 3-2). The PCB coneentratIona were correctly eatlnated for all aoll aamplea. Furthermore, the In-fleld toll extracte of aamplee D and E from elte #1 were elao analysed uelng the ECD-gaa chromatograph to determine the recovery efficiency of the field extraction va. the eoxhlet extraction. Field extraction recoverlea for aamplee D and E were 91Z end 10IX reepectlvely, Indicating acceptable recoverlea. 2-39 HONS 214729 rr" Parts Per Million-PCB 2-40 HONS 214730 r r TABLE 3-1 PCB CONTENT OF CONTAMINATED SOIL - SITE t1 (TLC . ECD - CAS CHROMATOCRAPKY RESULTS) Sampla A B C D E 50 PPM Standard Thin - Layar Chronatosraphy Spot Slxa (-> 9+1 PPM PCB (aa rec') 200 - 600 15 + 1 > 1000 16 + 1 > 1000 15 + 1 > 1000 8+1 100 - 150 6+ 1 - ECD - Gaa Chromatography PPM PCB1 (aa rac'J 375 7390 7280 1990(2010)2 101(92)2 - ^Raportad aa arctclor 1242. Raaulta ara baaad upon aoirhXat axtractlona of ubaaaplaa fron tha originally contaainatad aolli. 2 Tha valuaa In paranthaala vara obtalnad fro* tha aoll axtracta parfontad In tha flald> Thay rapraaant a 101X and SIX racovary for aaaplaa D and E raapactlvaly whan co*par*d to aoxhiat attraction raaulta. 2-41 HONS 214731 rr TABLE 3-2 PCB CONTENT OF CONTAMINATED SOIL - SITE #2 (TLC va, ECD - CAS CHROMATOGRAPHY RESULTS) Sanpla Thin - Laver Chromatoarcohv Spot Sit* ^n) t-PM PCB (aa rac') A ND < 25 B ND 25 C ND < 25 D ND <- 25 E ND -r 25 50 PPM Standard 6+1 - ND " Not Detected. ECD - Gaa Chroaatoaraohv PPM PCB iSljrae',) ^ 10 <. 10 23 < 10 < 10 TLC PROCEDURE - INTERFERENCES I Chlorinated paatlcldaa have born Inown to interfere with th* TLC determination* lor PCS* (2) 6.)* Standard* of DDT, DDE, and Haptador and Aroclor* 1242, 1248, and 1254 (50 PPM concantratlona) war* run taparataly ualns ch* TLC aathod. Th* davaloptns aolvanta vara varlad In an affort to raduca th* intarfarancaa Iron tha paatlcldaa. Tha aolvanta choaan war* haxana, (1:4) acatona/haxana, aathyltna chlorida, and acatonltrlla (haxana saturatad) which raaultad In a wld* rant* of Rj valuaa for th* PCB*. Baaad upon th* R^ valua* llatad In Tabl* 3-3 for DDT, DDE. Haptador, and Arodor* 1242, 1248, and 1254, two conelualon* war* drawn; (1) Incomplete aaparation of th* chlorlattad paatlcldaa from PCB* can ba azpactad for all aolvanta taatad. Thia ovarlap will raault In poaltlv* raault* whan taatlnt for PCBa. Howavar, tha praaaaca of chlorlnatad paatlcldaa In aoll* at concantratlona near 50 PPM la highly unlikely. (2) Incomplete aaparation of tha Individual aroclora Indicate* that th* TLC nathod la a non-apacific t**t procedure. 2-42 HONS 214732 W Interference fro* high coneant ration* of mineral oil waa alao Invaatlgatad in the aolvants Hated previouely using mixtures of 1000, 10,000, and 100,000 PFM mineral oil and 50 PPM PCB in hexana. Table 3-4 llata tha valuta for various concantratlona of oil and of PCBt at a 30 PPM lcval ualng different developing aolvanta. For fixtures containing 10,000 PFM oil tha detection of tha PCBa v*e Inhibited vhan methylene chloride or acatonitrila waa uaad aa tha ^developing solvent. However, even whan 100,000 PPM of oil war* praaant which resulted in axtanaivs tailing, 50 PPM of PCBa ware atill detectable whan tha developing aolvant waa haxana. TABLE 3-3 SEPARATION OF PCBa AND VARIOUS CHLORINATED PESTICIDES BY THIN-LAYER CHROMATOGRAPHY Solvent Systaa Haxana (1:4) Acetone: Haxana Methylene chloride Acatonitrila (Haxana Saturated) Aroclor 1242 0,29 0.50 PCB 2 Aroclor 1248 0.29 0.50 R. Value* Paaclcld*3 Aroclor DDT DDE Kapcaclor 1254 0.26 0,18 0.26 0.25 0.48 0.45 0.48 0.51 0.64 0.89 0.64 0,87 0.64 0,85 0.68 0.68 0.89 0.87 0.68 0.85 ^Totel weight of each compound uaad waa 1 microgram (20 mierolitera of a 50 PPM aolution). ^ALl three aroclora tasted abovad approximately equal intaneltlea at tha 50 PPM level. Tha (o,p) and (p,p') Isoms of DDT and DDE wars not separated. 2-43 HONS 214733 rr TAJLI 3-4 SEPARATION OF PCBa AND MINERAL OIL BY THIN-LAYER chromatography Staple Mlxtura PCB-SO PPM1 + Mlnarai 011-1000 PPM PCB-50 PPM + Mlnarai 0il-lo.OOO PPM PCB-50 PPM + Minaral Oil-i00,000 PPM Rj Veluee for Different Solvent Svetaaa Hexane U s4)Acetona: Methylene Hexana Chlorida Acetonitrile (Hexane Sat.) 0.30 ND 0.49 ND 0.64 HD 0.85 HD 0.29 0.59+.OflJ 0.30 0.50+.16 0.50 0.63+.08 0.49 0.56+.16 0.66 0.71+.06 * _ * ^Arodor 1242. 2ND - Not Detectable. 3tfhen developed tha alneral oil eppeera ea an off-whita apot on a brown background. Tha "+" valua la tha ranga of R^ valuaa with which tha alneral oil ovariapa. L Ext ana iva tailing. 2-44 HONS 214734 Section 4 SUMMARY The thin-layer chromatography method inveatlgated vae found to be aenaitiva at the 50 PPM laval and yielded reeulta that are at leaat a good aatlmate of the PCB con centration of contaminated soli. The method can be eaelly executed In the field by a trained technician and la fairly rapid (5 aamplea/TLC plate/hour). Therefore, the TLC method could a area ae an aid during clean-up operatlona of aolla contamin ated with PCBe following a capacitor rupture. Aa an aid it can be uaed to expedite the clean-up operatlona, helping to minimize the amount and length of time con taminated aoll la left expoaed to the environmental. Furthermore, It can help to eliminate aecond and third coatly clean-up *fforte. However, it la not Intended to be ueed ae a final clean-up confirmation. The final confirmatory aamplaa ahould be analyzed uelng the ECD-gaa chromatograph procedure which la more accurate and aanaitlva. 2-45 HONS 214735 rr Section 5 REFERENCES 1. Nelaon. N. Polychlorinated Biphenyl: Envlronaantal lanact. Environaentel Reeeerch, Vol. 5, 1972, pp. 249-262. 2. Watolek, Petricie C. Envlronaental Science end Technology. Vol. 13, 1979, p. 1969. 3. Weetoe, Gunnel end Noren, Koldu. Acte Chilca Scandlnavlca. Vol. 24, 1970, p. 1939-1944. 4. Coillne, C.B., Holaea, D.C., end Jackaon, F.J, Journal of Chrouatoaraphy. Vol. 71, 1972, pp. 443-449. 5. Reinke, J., uthe, J.F., and 0' Brodovlck, H. Envlrotientel lettare. Vol. 4(3), 1973, p. 201. 6. Fehrlntei, N.J., and Ueatfall, J.E. Journal of Chroaatoaraphv, Vol. 57, 1971, p. 397. 2-4$ HONS 21*736 Ql. Why la Hexane uaad In vlav of it'i toxicity? A- Other aolvanta probably could ba uaad. What la tha effect of othar chloridea - particularly TCB? A. Hoat of tha aolvanta would be qulta volatile and would evaporate. Unlee* concentration waa axtraaely high, there ehould not be a problen. HONS 21473? PORTABLE INFRARED FIELD MONITOR FOR PCBs: PHASE II H- S. Denton. Ph.D. and M. H. Ht 1 ker R and D Department# C/S Associates# Inc.# 101 Midway Lana# Oak Ridge# Tannest** 37830 ABSTRACT Sine* tha Dallas# Texas EPRI Seminar of December* 1981# a gr*at daal of affort has been expended on developing and commerciallilng PCB destruction and cleanup from askarels and transformer oils. PCB-contamlnated oil Is still being stock piled at an alarating rat* and PCB spills contlnu* to occur. These problems have proven to be extremely challenging ones with no widely accepted# definitive and economically feasible answers. Unfortunately# regulations have become more stringent and are more widely affecting personnel throughout the electric power Industry, It Is hoped that this Phase II* study on refining and field proving an Infrared (IR) portable monitor# preliminarily developed In Phase I** for PCBs In Spill conditions# will be of assistance In each of the aforementioned problem areas. The largest PCB problems occur In the form of PCB-eontamlnated oil In field transformers and storage containers# and pure askarel In transformers and capacitor*. The most lanedlate need for a portable field Instrument Is still for use under PCB spl 11 conditions. A simple# rugged# reproduelble> and PC8speclflc Infrared analyzer and sampling protocol has been developed and field tested to fill this void. 2-49 , HONS 214738 Phasa II has sarvad to vartfy and raflna significant findings of tho prallmlnary study. All IR wavalangths hava baan optlmlzad by digitization and subsaquant absorbanca maximization. Tha salactlon of axtractlon solvants has baan finalIzad basad on tha application and physical propartlas such as volatility* toxicity* miscibility and PCB axtractlon afflclancy. Tha lattar two eharactarlstles hava baan axtanslvaly stud lad utilizing a 30 x 30 alsclblllty matrix and distribution coafflclants maasurad by shako out axtractlons and subsaquant alactron captura QC or radio! abal ad PCB counting tachnlquas. Tha amphasls has baan on tha davalopmant of a solvant/soll axtractlon procadura followad by a flxad-fllm application on a Horizontal Multlpla Intarnal Raflactanea (HMIR) staga coup lad with a Foxboro Hlran 960 IR analyzar. Tha raflnad protocol of flxad-fllm application using volatlla. PCB-ladanad solvants raqulras a slngla shaka out axtractlon In a standard scintillation vial using ralatlvaly small amounts of sampla (a.g.* soil or oil) and solvant. Using such volatlla axtractlon solvants* rasults In an actual praconcantratlon stap prior to analysis and al lows us to saa 25-50 v9 PCB/J Soil routlnaly* and down to 5 ug PC8/g Soil using additional sampla and solvant. TMs tachnlqua has baan aldad by tha davalopmant of a stalnlass staal "sampla contalnmant walr" which Is placad ovar tha HMIR window. Furthar studlas hava baan mada to mlnlmlza tha affacts of soil molstura on PCB axtractlon and potantlal Intarfaraneas by pastlcldas and harblcldas possibly prasant at spill sltas. A brlaf Instrwnant comparison study has baan car r lad out with tha Ml ran 980/HMIR (mleroprocassor control lad* basal Ina subtract* matrix cal Ibratlon* ate.) and tha slmplar and lass axpsnslva Ml ran IA/HMIR (matar raad out and manual wavalangth salactlon* 2-^0 HONS 214739 r r ca1Ibratlon, and baaallna subtract). A final IR Instrument aalactton and protocol will ba made. The final *taga of tha study Involvas the analysis of actual flald samples and flald tasting to establish tha utlllty-of tha unit and sample protocol for portabla or noblla usa. Present rasaareh affort sponsored by EPRI Agraasiant RP1263-10 with C/S Assoc1atas> Inc. of Oak Ridge. Tennessee. Preliminary research sponsored by EPRI under Interagency Agreement RP12630-5; DOE No. ERD-80-061 under Union Carbide Corp. contract tf-740$-eng-26 with tha U.S. Department of Energy of Oak Ridge National Laboratory. EPRI CS-2626* Project 1263*4 Final Report. January 1963. "Development of a Portabla Monitor for PCBs" with W.D. Bostick and $4t. Din snore. 2-51 HONS 214740 PCBIftBLE IffBAREP OELB MOHITQB FOR PCBti PHASE II INTRODUCTION Phssa I Summary Tha Initial phasa of thl* study assantlally layad tha ground work for tha currant Instrumant and protocol raflnamant projact <1>. Tha matrix Isolation tachnlquas currant! y usad wara pral Imlnarl ly davalopad In this phasa. Thasa Ineludad tha usa of radlolabalad (14C) PCB, batch adsorption studios, and axtractlon of PCBs fro* oil and soil. Tha faaslblllty studlas for datormlnlng tha utility of Infrarad apactroscopy 1IR) vara carrlad out using tha tachnlquas of Fourlar Transform Infrarad (FTIR) spactroscopy (Dlgllabs FTS-20C). All Initial IR flow cal 1 work was carrlad out using tha Mlran 1A 12, whll a al 1 Horizontal Mulllpla Intarnal Raflactanca (HMIR) studios wars parformad on tha Mlran 980 () (both Instrumants from Foxboro Analytical). With tha Initial faaslblllty of IR provan as a monitor for PCBs, a sari as of sxpsrlmanta was undortakon to naasura PCBs and askaral In oil and soil saaiplas. Flgura l dap lets tha dlraet raadlng of pyranol In 10^ Oil (mlnaral oil) using a 0.2 mm pathlangth BaF2 flow cal 1 mount ad on tha Mlran 1A. As ean ba soan from tha vary small absorbancs raadlngs, maasurlng PCBs In oil or oil spills at this vary critical coneantratlon rings Is vary difficult at bast. This Is dua primarily to tha Inharontly low Intanslty of typical IR sourcas and tha prapondaranca of ol 1. Flgura 2 again shows Pyranol InlO-COIl. Howavar, in this casa, tha oil was usad to axtract tha askaral out of soil saaiplas. Mlnaral 2-52 HONS 214741 r r oil It, In fact, on tdatl IR solvant for PCBs sines It 1> IR transparent from approximately 7.S to 12J microns. Tha axtractad oil inplos worm again run on tha Miran 1A, Out using a 0.2 mm pathlangth ZnSa flow cal 1. Tha cal Ibratlon curvat obtalnad In this mannar wars quits 1tnaar for concantratsd samp1 at. Tha simp last form of PCB/Soll msasurament was also attaaiptsd in Phsaa I. This Involvad tha dlract application of eontamlnatad soil to tha HMIR window of tha Mlran 960. Flgura 3 lllustratas tha application of varying amounts of Pyranol (0 to 1 g) to 5-g soil samp las. Tha top scan raprasants a soil blank. Tha corraspondlng calibration curvss wars again surprisingly llnaar. This dlract application tachnlqua provad usaful for coneantratad askaral spills and would ba usaful for an Initial scraanlng of a suspactad spl11 slta. For tha 1owar PCB concantratlons that might ba prassnt in a PCB/011 spill or aftar partial digging up an askaral spill* this tachnlqua had savaral limitations. Thasa includad 1neonsistant raadings rasultlng from tha soil typa* soil molstura and amount of oil prasant. Rathar than "dlluting" tha PCBs with oil (or oil mlsclbla solvant) In an aKtractlon or daallng with tha 1 neonsIstanclas of dlract soil application. It was dacldad that a praconcantratlon stop would tnstaad ba naadad. This would aspaclally ba nacassary to aaasura contamlnatad 150-500 ppm) and laftd-fl 1 labia (50 ppm) aamplas. For this purpoaa* a flxsd-fllm HMIR tachnlqua was davslopad which daposltad tha PCS onto tho horizontal stag# window at a volatlls axtractlon solvant avaporatad. Using this tachnlqua, which is tha basis of tho Phass II study, ona can aaa tons of micrograms of PCS from sol 1. Furtharmora, 2-53 HONS 214742 f* the method Is quit* reproducible and* as can b* t**n In Figure 4* very linear. Phasw II Ovarylaw As mentioned previously* the goals of the current phase of this oroject are to refine the necessary Instrumentation and sampling protocol* as sell as apply than to spill type sanples. While a great deal of work has bean carried out in 011 matrices* the primary thrust of this second phase Is to develop a rugged* portable and easy to operate Instrument for us* In askarel spills on land-based substances. Since the Initial studies were carried out* w* have had the "benefit" of further PCB Information* litigation* spills and* most Importantly* feedback from many potential utility users concerning actual spill conditions and possible Interferences. Several excellent and extensive reviews are also available covering the general PCB situation (* &)* PCB* In the environment (2)* and portable Instrumentation (fl). The scop* of work for Phase II Included the following tssksi * Verification and selection of the proper wavelength(s). * Selection of the proper solvent(s). * Refinement of the H4IR technique and development of a protocol. * Measurements of PC8>contam1nat*d soil. * Evaluation of interforencea, ' Instrument selection and completion of protocol. * Field testing of the Instrument and technique. 2-54 HONS 214743 f f INSTRUMENTATION Ml ranTM IA The simplest and moat rugged of the Mlran'" series of Instruments from Foxboro Analytical Is tha alngla baam IR photomatar cal lad the Ml ran 1A. This Instrumant shown In Flgura 5 can ba aqulppad with althar a flow cal 1 or HMIR taga. Its portability and ruggadnaas for flald work has baan attastad to In a NlOSH portabla Instrumant avaluatlon (fl), Wavelength salactlon Is mads with a circular Interference-wedge filter, which doas not glva as much spactral rasolutlon as a grating Instrumant but doas provlda hlghar sansltlvlty (dua to lass light loss). Flgura 6 schamatleal ly lllustratas tha optical path of tha IA. HI ranTM 980 Tha majority of our work has baan carrlad out on tha Hfran 000 programmable IR on loan to us from Foxboro Analytical for this study (see Flgura 7). This spectrosieter was equally rugged, had similar optics (and thus similar resolution and sensitivity, saa Figure 6 for tha 080 optical system), and could ba fitted with either a flow call or HMIR stage Ilka tha Klren 1A. What sets this Instrument apart, however. Is tha fact that It Is microcomputer controlled and Is capable of on-the-fly background-corrected spectral scanning. Thus, the black body curve, 10-C011 matrix, soil matrix, ate. can ba subtracted. This feature has proven Invaluable In method development. Tha Mlran 900 also provides autoantlc analysts of mixtures with complete data reduction. As can bs seen from Figure 7, tha 980 provides (1) a keyboard for operating and data 2-55 HONS 214744 reduction functions* (2) printer for "peak picking" and absorbance* transmittance or concentration readout* and (3) a cassette for parameter and baseline storage. - (ilUllfili Internal BtUgcilan SPSCtrMCQOV As mentioned previously* both the Mlran 1A and 980 can be edepted to a multiple Internal ref lection (MIR) cell, To common configurations of MIR cell are 11 lustrated In Figures 9-A and 9-B. The former depicts a vertical multiple Internal reflection (VMIR) cell while the latter Illustrates a horizontal multiple Internal reflection (HMIR) window. While the sample does see twice as many reflections In a VMIR cell than In a HMIR cell* the latter configuration was used for all our MIR studies due to Its convenience and simplicity in handling spl 11-eondltton staples (e.g.* askarel* PCB/011* or PCB/$o11). It Is very simple for a sample to be applied* scanned and wiped off (with an organic solvent) with no carry-over contamination or flowcell blockage. Due to Its avallabll 1 ty* Inertness (especial ly to water)* durabll tty and wide spectral range* a ZnSe-wIndow HMIR was used throughout Phase I and II. Figure 9C Illustrates the light path* typical angle of Incidence* and typical dimensions of auch a crystal. The development of a BaFj HMIR crystal would greatly enhance the light throughput and Instrument sensitivity. The principle of operation of MIR, also referred to as attenuated total reflection (ATM)* le based on the fact that infrared energy being totally* Internally reflected within a high refractive Index optical material (e.gM Zn$e or BaFj) will actually penetrate a short distance Into the applied sample. If 2-56 MOHS 14745 r f the sample absorbs energy from specific wavelengths. this attenuation of the beam. utually axprtaaod aa an absorbance. la related to concentration. Such MIR spectra cloeely reaemble abaorptlon apectra using a tranaelaalon cell. Figure 10 schematically repreaenta a MIR optical diagram from aource to detector. The advantage* of MIR, and In particular HMIR, for field analyala are elucidated In Table 1. LUlfl Sana Id Preparation JLLt Barring any exotic Interference* from herbicide*. Insecticides, pesticides, etc., the routine preparation of land-baaed material for analyala by HMIR is quite almple and rapid. Like the method, the and 1 lary equipment needed for aample prep. 1* almple. A alnglm email caae contain* the following! * Ohaus portable digital balance for soil weighing. Precision Is good to on* decimal place. Powered by an AC adapter or 9V battery. ' Spatula for transferring soil to vials. * Ewpty sctntlllatlon-typ* vial* for soil (or oil) extractions. * Vials containing the appropriate organic extraction solvent(s). * Pipette (0-1 ml) with disposable tips for transferring solvents from vials onto the soil and extraction liquor from the extraction vial to the MIR window. * A stainless steel "sample containment weir" which fits down over the MIR window. While It 1* open at the top to allow sample Introduction and volatile carrier solvent evaporation, it eorals the solvent over the window during fixed film application. 2 -5 7 HONS 214746 rf Disposable silica gal minicolumn* for the separation of Intarfaranca* *uch a* pattlclda* from PCS*. Additional scintillation vials of alutlng solvent can also b* Ineludad, Tha saparatlon/ purification of PCS* 1* a slapla and rapid procadura described In tha lltaratura ( ii>. RESULTS AMD DISCUSSION Ifaval snath StliCtlftn Tha first task to b* undartakan was that of varlflcatlon and final salactlon of analytical and rafaranca wavelengths. Wavalengths vara roughly datarmlnad In Phasa I which would min 1mlza tha affacts of Intarfaranca* (e.g.( TCB and 10-C Oil) whtla maximizing tha absorbanca contributions of PCS* (JJ. Saa Figure 11. Selectivity and sensitivity was thus built In. These Initial wavelengths wars optimized by digitization and subsequent absorbance maximization. Table 2 outlines tha final selection of wavelengths. Table 2 Optimized Wavelengths for Monitoring PCS* (in microns) Phase I Phase 11 7.2 7.169 7.5 7.447 8.5 6.551* 9.2 9.1S3* 12.3 12.254 These are tha wavelengths of choice for selectivity and sensitivity. 2-58 HONS 214747 r r Solvent Selection The second task of the atudy was the final selection of solvents for the extraction of PCBs from o11 and oil. Attractive altarnatlvaa to tha solvents chosen in tha prallmlnary study ara prasantad for usa In both matrices. Solvant salactlon was basad on toxicity, volatility, miscibility, and axtrsctlon afflclancy. Tha toxicity and/or hazardous propartlaa of aach solvant wara avaluatad using a variety of sources (12* 11* lit XL. 1&>. Volatilities (vspor pressure) as wall as densities snd molecular weights are given In Table 3 which was compiled from a wide variety of sources (X* 12* 11. ii. XX* 11* H>. Knowledge of tha specific grsvlty of a solvent is necessary In the extraction procedure, while that of tha vapor pressure is useful In designing fixed-film experiments. Solvent miscibilities were determined by ths development through Phase I and II of the 30 x 30 mlscibl 1 Ity matrix shown In Table 4, Although these are primarily common chromatographic solvents, this data Is not available In any handbook to date. Table 5 below lists several additional solvent mlscibl1Itlea found useful In solvent extraction development studies. Table 5 Additional Solvent miscibilities DMF M tepfictions M M Oil M $R M HjO S M M final ly. extraction efficiencies were determined by measuring sol vent/oll distribution coefficients and percent PCB removal from soil and soil/oil 2-59 HONS 214748 /- metric**, A uniformly labeled (^C) PCB Isomeric mixture. containing approximate!y 54* chlorine by weight (Naw England Nuclear. No. NEC-690) waa usad to spike sample*. Such radiolabeled PCB work ha* also been done by Seldl and Ba 1 1 schmlter (2fl). Counting was accomp 1 1 shed on a Packard Trl-Carb Spectrometer. courtesy of the ORNi Analytical Chemistry Division. A computer program was developed by tt. D. Bostick In Phase I (1> which took this spectrometer data and computed counting efficiencies and net disintegrations per minute (OPM). Distribution coefficients (D> were thus determined and are summarized In Table 3 along with selected values from the work of Napier et al (!> Once such distribution coefficients are determined. PCB extraction efficiencies can be readily determined for any combination of solvents and shake out volumes and Iterations ss shown belowi Calculation a CCB EatCICttflfl Ef.fifii.MCX * PCB 100 Remaining In Oil where. 1 + D D Distribution Coefficient /Cone, of Ext. Phase! \Conc. of Oil Phase / VA Volume of Extractant V8 - Volume of Oil n Iterations 2-60 HONS 214749 T r Sample Calculation*: 1. Four extraction* with acetonitrile (CHjCN) here. avg. D 0.29 *2 *0 n1 631 (remaining In oil) 37* (PCS extracted) 2 40* 60S 3 25* 75* 4 16* 64* 2. Four extraction* with dimethylformamlde (DW) here. D 1.43 2 26* (remaining In oil) 74* (PCS extracted) 2 7* 93* 3 2* 98* 4 0.5* 99.5* Based on the four criteria presented above, three solvent* were selected for PCS extraction from soli and other land-based substances. These solvents In order of utility were carbon tetrachloride (CC14). methylene chloride (NO. trichlorotrlfluoroethane (Freon 113) and hexane. Each of these solvents has two advantages over the previously selected solvent (1>. acetonitrile (AN). These are higher volatility for rapid fixed-film vapor deposition and. very Importantly. Immlaclbl 1 Ity with water. The distribution coefficient of AN In 011 of 0.28 drops to 0.11 when AN Is contacted with sol 1. The moisture In the soil mixes with the solvent and subsequently repels the PCBs rather than extracting them. Depending on the moisture content of the soil, the absorbance 2-61 ` HONS 214750 measured for PCBs varies. Dn the other hand* PCB extraction from oil or soil/oil. N.N-dlmethyl formamld* (DMF> and acetonitrile (AN) tro th* only 101 vontt which hv* proven effectlv* and ar* atl11 practical. Mo have not* unfortunately* found an organic solvent which 1i both totally 1--lsclble with oil and water. While DMF must be handled with considerably nor* car* then the other aol vents* It does exhibit a distribution coefficient som five tines that of AN. flefin--nt of Technlaue and Sane 11 no Protocol Develop--nt The nodest and lllary equipment required for sample preparation has been described previously In the Instrumentation! Field Sampling Kit section. A set of calibration scans* ss shown In Figure 12 can be run at the laboratory. From these scans (Mlran 990) or absorbance values (Mlrsn 1A>* a calibration matrix can be stored In memory for data reduction (990) or plotted on graph paper (1A) for use In the field. The 990 Is further capable of storing any number of background curves on cassette tape for subsequent baseline subtract. It will be necessary with the 1A to run backgrounds In the field. The scans In figure 12 were produced by making up PCB standard In a volatile organic (CCI4)* depositing the mixture on the HMlft window* allowing the solvent to evaporate* and scanning. An air baseline has been automatically subtracted for each. The actual Absorbance xs. Micrograas of 1260 calibration plots are quite linear* but there are three distinct regions* low (0-100 ppm) middle (100-500 ppm) and high 0500 ppm) concentration ranges. 2-62 HONS 214751 r r Upon arriving at tha site. If the ipl 11 la an obvious ona or If a rapid acraan la desired. ona can scan tha potantlally contamlnatad aoll dlractly on tha HMIR window, Instrument nolsa and air backgrounds can ba subtractad out to produca a PC6 or askaral scan. Claanlng of tha window batwaan scans Is exceptional 1y aasy using a cotton swab or tlsaua with an organic solvant such as haxana. No carryovar contamination has avar baan obsarvad ragardlass of tha saapls matrix, Praconcantratlon by multlpla film daposltlona as propoaad In Phase 1 provad Infaaslbla, As aubaaquant 1ayars wara deposited, tha Initial layar would radlssolva and flow. An axanpla of tha final sampling protocol davalopad which al lowad us to saa samp las > 50 ppm. 25-50 ppm and 5-25 ppm In 1260 will ba glvan. Actually, thraa slightly dlffarant samp la prap. procaduras wara davalopad for soil > 50 ppm. 25-50 ppm and 5-25 ppm In 1260. Tha first succaasful extraction and IR maasuramant of 50 ppm 1260 from soil was accomp 1 Ishad by adding 4 ml of CCI4 to 5 g of soil containing 250 pg 1260 (I.#.. 50 pg/g). Extraction is quits effectively carried out In a sealed sclntll latlon vial with 5 minutes In a vortex mixer. One (1) ml of the 4 ml Is then placed on the HMIR window, Whan tha solvant Is coup lately evaporated, tha fixed film Is scanned and a similarly extracted soil blank subtractad. It was found that a nearly Identlca 1 curve (shown In Figure 13A) could ba obtained by extracting tha same amount of soil with only 1 ml of CC14 and pipetting 0.2 ml of the 0.8 ml recovered from the extracted soil onto tha window. This greatly reduced tha film deposition time. As mentioned, extraction of 25 ppm 1260 sol 1 requires the former procedure of extracting with 4 ml and pipetting 1 ml of tha -3.8 ml recovered sol vent onto 2-63 AQNS 214752 the window. This procedure 1> volume dependent (s one would expect) at ahown In Figure 13B. The upper tcan It a 1 ml out of 4 ml film depoilted at deter 1 bed above. The lower tcan repretentt 2 ml of thla tame tolutlon. The abtorbance doublet, again at one would expect. Final ly. Figure 13C 11 luatratea a 5 ppm tcan, the loweat concentration of PC8 auccetafully extracted from toll by thla technique. A few changea are required to reach thla Halt, however. Five (5) ml of CC14 were uaed to extract 10 grama of aol1 and all of the 4.05 ml recovered after vortexlng waa applied to the window. It wet neceaaary here to have a anal 1 lecture bottle of nitrogen gaa to help blow down the aol vent. Thla waa carried out In the vial prior to application on the window. Mmurwnti of PCB-Contamlnatad Soil The next atep In the atudy waa to meaaure actual aakarel In soil aamplea. the Interaction of and problem* with PCS* or aakarela In aoll have been reported on numerou* occealon* (2, 21* 22), Some baala for the compoaltlon of aakarel muat be aaaumed alnce the make up In electrical equipment haa purpoaely been varied or mixed over the yeara (aee Table 6 ). A11 of our aakarel atudlea were carried out on Pyranol (45* 1260). Figure 14A repreaenta the acan of a 50 ppm Pyranol (aa 1290) In 66E CollegedaleRoek Outcrop aoll. The routine method of 1 ml CC14 extraction of 5 g of aoll (1.*., 250 yg total), 5 min. vortex, 0.2 ml appl led by flxed-fl 1m to the HMIR waa fol lowed. The upper acan of Figure 14A la the aoll blank, the middle acan la the 50 ppm film* and the lower acan for comparlaon la a 125 vg flxed-fl In atandard depoalted free CCI4 without any extraction. Figure 14B, C and 0 ahow 2-64 MOWS 214753 r r this same SO ppm Pyranol (as 1260) soil sample extracted with methylene chloride (MC), Freon 113 and hexane, respect1vely, Methylene chloride seems to bo tho most officiant oxtractant, whllo Froon >nd hoxano produce nearly Identical cant. While MC, Freon and hexane are al 1 more volatile than CCI41 hexane Is not a> good an extractant from soil and MC and Freon undergo very rapid evaporation which Is difficult to control. Figures 15A and B Illustrate 25 ppm and 5 ppm Pyranol (as 1260) In soil, raspectl vely. to show that the CC1 ^ extraction and flxed-fl !m procedure can reach these levels with askarels as well as PC8s. The two scans were the results of the special procedures developed for soils at these low levels. ^QitrfTWlCM The detella of our evaluation of possible Interferences to this method will be given In the Final EPRI Report on this project. However, It should suffice to say that there is a very real potential for Interferences from herbicides, pesticides and Insecticides at spill sites. Even pole preservatives such as pentach1orophenol could be a problem. PCBa were actually added to some herbicides to extend the kill life. If these were sprayed around power poles the problem of background Is further confounded, we have run scans on OOTj 2, 4-Oj 2,4, 5-T) dleldrlnt 2,4-dlnltrophenol, pentach 1 oropheno 1 j lindane, aldrln, and heptachlor. A HMIR fixed-film scan of a mixture of the latter three Is shown In Figure 16. We have eliminated the original interference of soil moisture with the current sampling protocol, however, the chlorinated, polar Interferences will have to be 2-65 MONS 214754  eliminated (or minimized) by one of two way*. Careful wavelength selection ualng the Ml ran 980 ora cl eanup step using a silica gel sample prep, column will be necessary (see Field Sampling Kit section). Fins! Instrument Selection The Miran 980 with its automatic background subtract and data reduction capabilities was an absolute necessity for the methods development portion of this study. Once the protocol was developed and specific wavelengths chosen. It became clear that the much lass expensive Miran 1A with a HMIR stage would possibly suffice. It Is recommended that a BaFj HMIR crystal replace the existing ZnSe for enhanced 1 Ight throughput and that the 980 be used In cases where soil night contain extra** Interference by oil* herbicides* otc. A field comparison of the 980 and 1A will be presented In the Final Project Report. Me have done* however* a simulated comparison by first scanning a 50 ppm 1260 in soil extractant using the 980 with a blank soil extractant background subtract. The same sample was then scanned versus air as would be the case with the 1A. The soil extractant reference was also scanned versus air and the absorbance for sol 1 (0.0095 All) was subtracted from that of PCB (0.0310 AU) to gl ve 0.0215 All. This value corresponds to 47 rnlcrogrsms/gram on our low range PCB calibration curve. Since the value Is very close to the expected 50 ppm and that obtained with the 960. it looks promising to use the portable Miran 1A for field work. CQHCLUSIQH Inherently* IR analysis Is a highly selective technique. Absorbance measurements at either 8.551 or 9.159 microns are selective for Aroclor 1260 In 2-66 HONS 214755 1' the presence of oil (an afkarel substitute) or trlchlorobenzen* (tho solvent In askarel fluid*). Tho chlof limitation of convontlonal IR analysis Is th* modest sensitivity afforded by the source, he have successfully overcome this limitation by the development of a fixed-film deposition of PCBs from a volatile organic pxtractant coupled with an HMIR Instrument. It has often been tempting during the study to go back to a flow cell IR technique (e.g.* BaFj) to enhance sensitivity, Howeveri any such flow-cel 1 work wouId Invariably have to be carried out in oil. or perhaps C$2> the on 1 y good windows available in this region. Since this would actual ly be a di lutton rather than our proposed preconcentration It is not as highly recommended as th* use of a volatile solvent and fixed film application. Solid particles In oil or solvent from sol 1 would also cause 11ght scattering problems not present In HMIR. A SaFj HMIR crystal Is further recommended for sensitivity enhancement. Mill* amaxlngly simple and convenient* th* direct application of contaminated soil on th* HMIR stag* is recommended for screening purposes or for obvious askarel spills only, Th* method gives Inconsistent results and Is frought with variables (*.g.> uniform contact with the window* soil moisture* and oil content). Th* sensitivity Is also limited her* and those samples below 200 ppm* or certainly below 100 ppm* would be extremely difficult. The most promising extraction solvents for PCBs In soil were found to be carbon tetrachloride* methylene chloride* Freon 113 and hexane. For PCBs In oil or soil/oil, dlmethylformamld* or acetonitrile are recommended. W* hav* developed 2-67 HONS 214756  DMF extraction procedure similar to that developed Indopondontly by Naplor at al (12) for PCD from oil cloanup on a much larger teal*. Tho Inherent aelectlvlty of IR analyala coupled with our fixed-film protocol and an available HMIR yield a simple* rapid* rugged* mobile (or-portable) and reproducible Instrumental approach for the measurement of PC8s under spill conditions In the field. ACKNOWLEDGMENT The authors wish to thank the Electric Power Research Institute and In particular Or. Ralph Y* Komal* Project Manager* for their continued support and guidance In this work. This research was sponsored by EPRI under agreement RP1263-10 with C/S Associates* Inc. of Oak Ridge* Tennessee. We also wish to thank Dr, william D. Bostick* Ms. Debbie Bostick* and Mr. Stanley R. Dlnsmore for their assistance In verifying Phase I work* standards selection* and radiolabeled PCB counting. Standards for many of the herbicide and Insecticide scans were greclously donated by Dr. Jurgen Exner of IT Envlroaclenee in Knoxville* Tennessee. Finally* we greatly appreciate the support of Foxboro Analytical* a division of the Foxboro Company* In Norwalk* Connecticut. In particular* thanks to to Mr. Louis C. Franconnl and Mr. Carl R. Brewer for supplying the Mlran 960 and backup support during Phase II. Comparison studies using the Miron 1A were done using an Instrument belonging to the Oak Ridge National Laboratory operated by Union Carbide Corp. for the U.S. Department of Energy. 2-68 MONS 21475 7 r r REFERENCES 1. W. D. Bostick, M, S. Denton and S. R. Dlnsmore, "Development of a Portable Field Monitor for PCBs," EPRI Final Report CS-2828. Research Project 1283-5, January 1993. 2. Ml ran 1A IR Spectrometer Operation 5 Maintenance Manual, Vol. I, 001-4201, The Foxboro Co., S. Norwalk, CT, 1962. 3. Ml ran 1-A: Applications Manual Vol. II, 001-4201, pp. 33-35, The Foxboro Co., S, Norwalk, CT, 1979. 4. Ml ran 980 Infrared Spectrometer Manual, MI 611-091, The Foxboro Company, S. Norwalk, CT, January 1963. 5. M. G, McGraw, The PCB Problem) Separating Fact From Fiction," Electrical World, pp. 49-72, February 1963. 6. J. Reason and tt. Bloomqulat, 'TOJ Replacements) Where the Transformer Industry Stands Now," Power. 121 (10), 64 (1979). 7. R. F. Addison, ^C8 Replacements In Dielectric Fluids." Environmental Science and TochiMloov. 12 (10), 466A (1963). 6. HEW Publication No. NioSH-76-166, "Evaluation of Portabla, Olrect Reading Hydrocarbon Meters, N10SH, Cincinnati, Ohio, April 1976 (U.S. Dept, of Commerce PB-266 439). 9, J. Armour and J. Burke, JAQAC. Si, 761-767 (1970). 10. Supelco, Inc. Bulletin 761A, "PCBs and Intsrferanee from Chlorinated Pesticides," Bellefonte, Pennsylvania, 1976. 11. R. J. McElhaney and M. L. Mann, "The Analysis of Selectsd Organic Priority Pollutants in Environmental Samples," Oak Ridge Y-12 Plant, Y/DK-255, Sept. 26, 1980. 12. R. H. Perry, C. H. Chilton and S. D, Kirkpatrick, Ch--ical Engineer* Handbook. 4th Ed., New York, McGraw-Hill, pp, 3-47 to 3-56, 1963. 13. laitoe*a Handbook of Chemistry. J. A. Dean, ad., 12th Ed., Vol. I, p. 61, McGraw-Hill. New York, 1979. 14. CRC Handbook af Chemistry and Phvslca. Weast ed., 56th Ed., pp. C-268 and D-192. MONS 214758 r* - 15. Merck Index, 8th ed., Paul G. Stecher ed., Merck and Company, Inc., 1968. 16. M. Slttlg, Haaardoua and Toxic Effect! of Induatrlal Chemicali. Noyea Data Corp., Park Ridge, NJ, 1979. 17. J, M. Napier, M. A, Travagl ini, . G. laggla and M. A. MaKareulcz, "Evaluation and Development of Polychlorinated Biphenyl Removal Proceeiea," Oak Ridge Y-12 Plant, Y/DZ-1, February 5, 1982. 18. D. W. MacKay and A. W, Wolkhoff, Environmental Science and Technology. 1, 611 (1973). 19. 0. Hutzinger, S. Safe and V. Zltko, The Chemlatrv of PCBa. Baton Roca, Florida, CRC Prest, Inc., 1974. 20. G. Seldl and K. Ballachmlter, "laolatlon of PCBa froai Soil! Recovery Ratea Ualng Different Solvent Syatems," QuaciCiHU, , 373 (1976). 21. G. J. Farquhar, et il, 1*06 Interaction! with Soil," 3. Environ. Scl. flMdttl, A14 (7), 547-557 (1979). 22. P. W. Albro, "Probleme In Analytical Methodology! Senile Handling, Extraction, and Cleanup," Annala of The Hex York Academy of Sclencea, W. J. Nicholaon, ed., Vol 320, 19-22 (1979), 23. Moneanto Aakarel Inepectlon and Maintenance Guide, Monaanto Functional Flulda Dept., $t. Loula, MO. 2-70 HONS 214759 r r MO 0 Ot i } 00^ 004 00* 0 00 Arettor 12*0 <w/w> pig, 3 FIG. 5 HMH Optical 5<hffttic n4 N*Jr Ct*pnc*t FIG. 6 2-71 HONS 214760 2-72 HONS Z14761 . ' . . . * 1* u --------- ;------*------r------r i r-r-T-^-l < TuanUM'm 't y ifnr wo* Mfl ^ H g. i3 l4 riG. --^ * V^Qd| Va tot . -V /-Vx " 4* \ .r-r-.r. . . L_ J-------1--' ` ` * fig. is fIG. is ~^V' "V--------- ----------- ------ ^ rtVH i ' ' 11' ifrrtitiwi* * ^'f*c***" \T ^ *,* w* it m * > .--* u t ii| (H* -hfcm , fit*, **r iw * i* HONS 214762 Part 3 RETROFILL HONS 214763 r r THE UTILITY PCS DILEMMA - RETROFILL OR REPLACEMENT? Thom** L. Forrester* and Thomas H. Milby, H.D."* DEFINING THE CRITERIA Tha utility industry has baan unjustifiably brandad as tha "bad guys" bacausa of PCBs. Onca considarad a banavolant public servant, tha alactric utility industry is now ragardad by tha public as anothar insensitive, profit-motivated corporata giant which must ba continually scrutinicad by citisan action groupa snd haavily ragulatad to pravant environmental pollution and harm to tha public from toxic materials. Ae a result, alactric utilitias ara confrontad with soma important dacisions about) (1) tha typa of alactrical insulating fluids which ara baat suitad to raplaca PCBei and (2) tha most affactiva manner in which to safaly and economically raplaca tha aiillions of pounds of PCBs in axiating equipment. Until recently, tha procass of choosing an alactrical insulating fluid has navar baan too compliestad. Tha propar salaction of a suitabla insulating fluid dapandad upon its ability to satisfy cartain operating and engineering performance criteria such est (1) dielectric strength, (2) viscosity, (3) flaaiMbllity, and (4) compatibility.1 Having passed these preliminary specifications, tha insulatingfluid would than ba evaluated against mors subjective criteria such ass (5) past field performance and experience, (6) availability, (?) tha supplier'a reputation and dependability and (8) costs. However, to coin an old adage, "the old must make way for tha new" so it goes for tha procees of selecting an alactrical insulating fluid. The task onca regarded as "business as usual" has now become more of a nightmare. To illustrate, at St30 a.a. on February S, 1981, a fire broke out in the first level baaament mechanical room of an 16-story office building in Binghamton, Hew York. Tha fire originated through a fault in tha building's secondary (460 v) switch gear, which "Pacific Gas and Electric Company ""Medical Consultant, Pacific Gas and Electric Company 3*1 HONS 214764 resulted in a cracked bushing on an ad^acant AiKarei-filled tranaformer providing service to the building. Approximately 180 gallon* of inaulating fluid spilled from the tranaformer. Ventilation intake ducta located adjacent to the transformer spread the oily smoke and soot throughout the building. Almost three years and more than 10 million dollars later, the building renaina clos'd.3 At 11:19 a.m. on Hay 15. 1983, in San Francisco, California, a transformer fire in an underground street vault at Steuart and Kission Streets contaminated a small area of a 18-story office tower with PCBs and forced the closure of the entire building. After 10 daye of extensive testing and dacontamination, floors 7 through 28 were reopened for occupancy, while floors 1 through 7 remained closed for six months. Out of these episodes have arisen more citizen action groups demanding the immediate replacement of all PCB transformers. The lesson that can be learned from these frightening experiences is that the criteria once thought sufficient to select a high quality performance insulating fluid are no longer adequate. In an scologically sensitiva cultura, the public's fser of environmental contamination demands that the electric utility industry broaden the scope of its evaluation process to include the fluid's potsntial toxicity and effect on the snvironmsnt. SELECTING THE BEST FLPID(S) Whet then are the alternatives? One concept becomes store end more apparent -- that the toxic properties of a proepectiva insulating fluid ar* equally, if not mora important, in terma of on industry's public imaga and financial health, than tha enginearing qualities, and that whatevar fluid it chosen must be at environ mentally accaptabla ee it it a good alactrical insulating fluid. Once the task of identifying all the criteria nacassary to properly choose an insulating fluid is completed, the next step it the selection of a suitabla insulating fluid. This is the decision area which is critical and rsquiras a thorough undsrstsndlng of tha lmportanca of ths toxicity and environmsntal critsris in making ths final fluid aalsction. Thera are several non-PCB insulating fluids commsxcially availabla which havs alrsady baen accsptad besad on thair anginaaring qualitisa.3 The question that remains to bs answered is ths sffsct that thesa alactrical insulating fluids will havs on ths public and ths environent. Minaral oil (Chevron Insulating Oil and Shall DIA1A Oil A) probably has ths longsst track record of any of tha non-PCB insulating fluids. It has dsmonstrstsd satisfactory alactrical insulating qualities and has naver to date been associated with any harmful effect# other than ae a flaamable liquid.3'6 However, it may be 3-2 MONS 214765 r r till too early to determine whether mineral oil may become a problem duo to tha increasing interest about polynuclear aromatic hydrocarbons and thair potential toxicity, and bioanvironnantal implication*.3 From all available information, polydimathylailoxana or ailicona fluid (Dow Corning 561 and Ganaral Electric SF 97 (SO)) appear* to be not only an acceptable inaulating fluid but represent* tha lowest potential for advarae affects to the public and tha environment.3 Tha only apparent problam eaaocietad with ailicona fluid ic its potential fire haterd. Perchloroethylane or tetrachloroathylana (WECOSOL) is being auggasted as a vary good PCS substitute insulating fluid because of its electrical qualities and more important, its non-flammability. 3,6 However, parchlorosthylsna, like PCBs , is another chlorinated hydrocarbon which may bioaccisnulate in tha environment and is regarded aa a toxic substance and suspactad carcinogen.^'5 Finally, tha last fluid to ba considered as a substitute for PCS la a paraffinbaaed, high molecular weight hydrocarbon oil (RTBnp). It hat good electrical propartiaa, but it is a flasaaabla liquid and, like mineral oil, may have some future problems due to trace amounts of polynuclaar aromatic hydrocarbons.3'6 RETBOrlLL OR REPLACEMENT As previously pointed out, tha electric utility industry faces two dilanmasi (1) ths selection of a a ui tab la PCS substitute fluid; and (2) what to do with tha PCBs on hand. Having discussed tha relevant issues surrounding tha first problem, lot us now examine what alternatives ars available for tha disposition of tha PCBs in existing aquipatent. Three options era available to tha electric industry to resolve tha PCB dilemma; (1) a retrofill program whereby tha concentration of PCBs is reduced in existing aquipswnti (2) a total replacement program whereby all PCB equipment it replaced; end (3) some combination of a retrofill and replacement program. However, before a coomitmant is made to any of theoo alternatives, tha advantages and disadvantages of each mutt ba carsfully weighed. Retrofill A retrofill program allows or accomodates the implementation of a low-key PCB removal program. The three most obvious advantages of such a program Include; 3-3 HONS 214766 (1) * reduced coemitment of immediate fund* and manpower resources* (2) possibly laaa attention from the media and the general public 1 and (3) the coata of retrofilling existing equipment are aotnewhat leaa than a replacement program, (Estimates of the coat of retrofill have been quoted by contractora aa 60-70 percent of the purcheea price of the new equipment,) Another attractive advantage of a retrofill prograat ia the Environmental Protection Agency'a recent prograat for recertification of FC8 and PCB-contaminated equipment. Thia ia the prograat whereby EPA will recertify equipment if it ia ahem through retrofilling that the concentration of PCBa can be reached and ataintained below the legal definitiona of 50 and 500 perta par million (ppm) for longer than 90 daya,4 EPA'a rnclessification of electrical equipatent ia beneficial to mdustriea becauae it may facilitate employae protaction and waate managemant programa by laaaaning ragulatory requiramenta. Finally, the ramoval of PCB transformers and ralated equipment mey require the complete reconetruction of the vaulte in highriee buildings or underground vaults. A successful ratrofill program uses existing equipment, resulting in a more timely, more economical and more convenient PCB-removal program. There are eleo tome disadvantages to the retrofill program which are important and muat be emphasised. One drawback to thia program focused on the safety of the insulating fluids. Although soma information ia availabla on the toxic and environmental affects of these fluids, the questions of long-term toxicity, cancer-causing effects and reproductive haaarde are still unanswered for some of the fluids.3 Another area of concern with thia type of program centers on it* alleged capability to reduce and more importantly, maintain the amount of PO* below certain levels. A successful retrofill program must filter PCBa fester then they leach from the windings and paper products in the core of a transformer end maintain PCB levels below EPA defined criteria or other prescribed levels. However, testa have shown that when the PCB-filtering and removal system is disconnected, the continued leaching of PCBs from the transformer core will, after some time, again raiae the PCB concentration in the equipment.3 Theoretical calculations Indicate that the residual PCB concentration muat ba maintained indafinitaly below 1 ppm in order to ensure that in the event of e transformer firs, PCBa and thair decomposition products are not found at or above the 3- i MONS 214767 r r preliminary sale decontamination lavala being propoaad by an expert panel of sclent!et> for the Binghamton fire. Studlee are under way by the Electric Power Research Inetitute (EPRX) and the Hew York State Department of Health to determine the minimum levele of PCIe necessary in siliconer psrcliioroethylene, mineral oil and high teetperature hydrocarbon insulating fluids to prevent the type and degree of contamination demonstrated by the Binghamton and San Francisco fires. Until such time as these studies are completed in March or April 1984, it cannot be ascertained whet minimum concentration of FCBe will ba acceptable in a retrofilled transformer to prevent the PCB episodes seen by tha Binghamton and San Francisco experiences. Replacement If the uncertainties of a retroflll program are too risky and unwielding, let us consider another possible alternative and discuss the advsntages and disadvantages of a total PCB equipemnt replscesmnt program. Perhaps the most obvious and important attraction of a PCB equipment replacement program is that once and for all, the risks associated with PCBs would be eliminated and the utility industry might begin to breathe a bit easier In the future. Unquestionably for a utility company, tha attractiveness of this course would be sufficient to pursue a PCB equipment replacement program. However, not unlike the retroflll alternative, there is also a dark elds to the PCB equipment replacement progrm*. One disadvantage of the program la the lengthy and disruptive construction phase necessitated by enlarging the openings into existing street and building vaults to remove PCB transformers and install new equipemnt. The increased costs associated with a replacement program is another drawback. In contrast to a retroflll program whera the cost* are derived from the purchase of new insulating fluid and any labor involved, the coats associated with a replacement program include not only the new equipemnt and insulating fluid, but the requisite reconstruction expenses of the program. Furthermore, since tha seme insulating fluids are used in both programs, the questions of toxicity and potential environmental effects are still unresolved. HONS 214768 3-5 SUM4ARY The coat*, tha conveniences and inconveniences, tha unanswered questions about toxicity and environmental affacta of the insulating fluid*, and the velu* of retrofllllng transformer# for recertification purposes era some of tha uncartaintiaa which make tha decision to ratrofill or replace difficult. Tha potential risks of either alternative are high and there arc no guarantees. 'The decision to ratrofill or replace reate with each individual utility and depends on its own particular circtmtances and goals. It has not bean our purpose to recoiend or condemn a pcb ratrofill or replacement program, but to provide son* insight into the type of coneideretione which are important and should be evaluated before a final deciaion is reached. REFERENCES 1. U. 5. Department of Commerce, National Technical Information Service (NTIS). Assessment of the Bee of Selected Replacement fluids for PCIa in Electrical Equipment, EPA 560/6-77-008, March 1979. 2. New York Office of General Services. Binghamton State Office Building Clean-Op, A Progress Report Update, January 1983. 3. Goan, R. L., Ryan, J. W., Pavlovich, A. M., Joiner, R. L,, Rogers, J. L., Mctwen, G. H., Rich, P. A. 3'ewmrv Data on Substitutes for Polychlorinated Biphenyls (PCBs). Stanford Research Institute, International, Contract No. 88-01*6016, February 1981. 4. Environawntal Protection Agency, 40 CPU Part 761.30A. Federal Register. Volvme 47, Ho. 165, Wednesday, August 25, 1982, Page 37358 5. Environmental Protection Agency, "List of Toxic Pollutantsi Petition to namove Aromatic Keloethars", Federal Register, Volume 44, Ho. 60, Pages 18279-83, March 27, 1979. 6. U. 8. Department of Commerce - National Bureau of Standards) u. S. Department of Energy - Division of Electrical Energy Systems) National Electrical Manufacturers Association (KEMA). MIMA Fluid Filled Transformer risamabilltv Study. Contract Ho. MB795BCA0024, July IS, 19B0. HONS 214769 3-6 wW DISCUSSION Th# questions basically war* in regard to the assumption* and limit* used in th tudy. For example, tha requirement that ratrofillinq ahould reduce and maintain 1 npm PCB. Thia requirement wa* bated on calculation* of the liaiita of furana in aoot. It wan pointed out, by teveral diacuaaora, that converaion to ppai furana from ppm PCB waa not linear, especially in low concentrations, Dry typa transformers were not conaidered for replacement because of coat, size, and lower efficiency. Question* were taiaed regarding trees contaaiination of silicone fluids with PCB and/or formaldehyde. It was reported that the silicone used was certified to contain less than 1 ppm PCB and that the formaldehyde problem had been solved. MONS 214770 3-7 CLEANING ASKAREl TRANSFORMERS USING SILICONE RETROFILl AND ADSORPTION TECHNIQUES TO REMOVE RESIDUAL PCBs Daniel F. Meyer and Edward A. Reynaert Dow Corning Corporation BACKGROUND Section 6(e) of the Toxic Substances Control Act (TSCA) of 1976 banned the use of PCBs after January 1, 197B. However, the statute permitted two exceptions: the Environmental Protection Agency (EPA) may allow PCBs to be used In a "totally enclosed manner" or If the use "will not present an unreasonable risk of Injury to health or the environment." In May 1979 the EPA promulgated a rule designating Intact nonleaking transformers, and other electrical equipment, as "totally enclosed", thus permitting their continued use (1). This ruling was challenged In the Federal Courts by the Environmental Defense Fund (EDF), The court, ruling In the EDF's favor, found In part that EPA had not shown that transformers were Indeed "totally enclosed," Thus the use of any transformer containing any amount of PCBs would constitute a violation of TSCA. An limedlate ban on the use of equipment containing PCBs would have disrupted electrical service and caused severe economic hardship. Therefore, the court granted a stay to permit the EPA to study the problem and reissue a new final rule. This "final" rule was published In the Federal Register on August 25, 1982 and became effective on September 24 of that year (2). With respect to transformers, the 1962 rule reaffirms much of the original 1979 rule, with a few Important additions. It continues the classification of transformers Into three categories according to the PCB concentration of their liquid coolant. These categories are: 1. "PCB transformer", 5DO ppm PCBs or more; 2. "PCB contaminated", 50 to 500 ppm PCBs; 3. "non-PCB", less than 50 ppm PCBs. 3-9 HONS 214771 Important additions to tht 1979 rule contained In the 1982 rule Include: 1. after October 1, 1985 no "PCB transformer" may continue In use where there is risk of PCS contact with food and feed; 2. "PCB transformers must be inspected regularly and records kept of these Inspections; - 3. recognition of the possibility of reclassifying "PCB transformers" by retrofllllng them with an environmentally acceptable dielectric coolant and further processing the new liquid to remove residual PCSs that leach Into It from the core, windings and Insulation; A. clarification of the procedure for reclassifying "PCB transformers* as "PCB contaminated" and "non PCB". The rules governing how a transformer may be used, serviced and disposed of depend upon its PCB classification and are summarized in Table I. Note that there is little significant difference between the rules for "PCB contaminated" and "non-PCB" transformers. With the exception of transformers that pose an exposure risk to food or feed, a "PCB transformer" may continue In use. However, It must be labeled, Inspected regularly and any service that requires the removal of the core from the tank Is prohibited. When disposed of, the fluid must be burned In an approved PCB Incinerator. The carcass may be disposed of In the same manner or burled Intact In an approved chemical waste landfill. A "PCB contaminated" transformer need not be labeled or Inspected regularly. Any service can be performed by the owner or an exempt service company. Its liquid coolant can be disposed of In an approved chemical waste landfill or burned In a high efficiency boiler or an approved PCB Incinerator. Finally, It may be used detplte exposure risk to food or feed, A "non-PCB" transformer may be used virtually without restriction. The only exception Is thet fluids containing any detectable amount of PCB cannot be used "as a sealant, coating or dust control agent". This Is the situation as It stands today. However, the current rule may not be final after all. The Environmental Defense fund petitioned the United States Court of Appeals to review EPA's August 1982 rule. And after lengthy negotiation, PA and EDF petitioned the court to resolve their differences In the following fashion. * 3-10 HONS 214772 r_ f EPA will Issue n Advance Notice of Proposed Rulemaking (ANPRM) covering PC6 transformers In buildings In March 1984 (2) and has scheduled the Notice of Proposed Rulemaking (NPRM) for October 1984, The rule will be Intended "to reduce the risk of future fires in PC8 transformers that are located In buildings," EPA's schedule calls for completing the rule-making process In July 1985, If a final rule is deemed necessary for building transformers. Against this backdrop of regulations and with a continuing, If not Increasing, public perception that PCBs pose an unacceptable health and environmental risk, many of the owners of the 10D.000 askarel-cooled transformers In the United States are looking for alternatives. Cleaning those transformers by using sllleone retroflll and subsequent adsorption techniques to remove residual PC8s is one such alternative. This paper will address this cleanup technique in a generic fashion to the extent that Is possible. In some instancy however, it will be necessary to refer specifically to the Dow Corning RetroSH System and to DOW CORNING* 561 Transformer Fluid, particularly when describing the system equipment, procedures and performance. THE PROCESS A silicone retroflll and adsorption approach to transformer cleanup has a number of distinct advantages. Total transformer downtime is typically less than 1? hours. Removal of residual PC8 that leaches out of the transformer core (windings, paper, wood, etc.) is accomplished while the transformer Is energized. This cleanup process Is also particularly well suited for transformers located in dlfflcult-to-reach locations, Removal and replacement of these transformers Is usually costly because of service Interruption and because building modifications may be required to provide access. At the heart of the cleanup process Is silicone transformer liquid which Is a dimethyl siloxane polymer combining the properties of low viscosity, high temperature stability, good dielectric performance, low toxicity, and low flammability. It Is used primarily In small and medium-sized power transformers (4). Silicone or silicone fluid or liquid as used In this paper refers specifically to the dimethyl polymer developed for and used In ' transformers. MOMS 214773 3-11 Silicone fluid Is tlso * good dielectric (.,,.2.) end has been used in more than 25,000 new transformers around the world since 1972, Its reliability and performance have been proven (8), Interestingly, the first application of 00W CORNING*561 liquid in the United States was as a replacement fluid for an askarel-filled transformer. Hundreds of successful applications In-similar transformers since that time have demonstrated convincingly that silicone fluid is a viable replacement liquid coolant for askarel. Two final points concerning slUcone liquid merit attention: low toxicity and low flaimablllty. The liquid Is chemically Inert and DOW CORNING* 561 fluid contains no thermal stabilizers, pour-point depressants, or other additives. Other chemically similar silicone liquids have been used in a variety of applications including many that involve contact with people. Personal care Items such as .antipersplrants, hair sprays, skin-care products, and even some edible products use silicone fluids as major Ingredients or additives. Extensive testing has demonstrated that silicone transformer fluid is nonirritant and not absorbed on skin, has no adverse toxic effect when inhaled, and has a very low oral toxicity (4.9,JO). SlUcone transformer liquid also possesses low fire hazard characteristics. It Is listed as a "less flamnable* fluid with a defined heat release rate by Factory Mutual and therefore meets the requirements of the National Electric Code (11,12,13). DOW C0RNING*561 liquid's flash and fire points and heat release rates are shown In Table II QZ.14). Even though silicone fluid has a very high fire point (350*C), It will burn when Ignited^ However, when burned It generates relatively low quantities of heat, smoke and toxic gases. In pool fires, silicone fluid eventually self-extlngulshes because a crust Is formed which smothers the flame. (4) A second Important part of the askarel transformer cleanup procej^ Is the people, knowledge, and equipment cosnltted to the program. The RetroSII process Is usually conducted by Independent service companies having expertise and a history of serving the electrical needs of the transformer owner. These companies have also been specially trained and certified by Dow Coming to perform the RetroSIITM process. 3-12 MOWS 214774 f r The two Min RetroSII equipment components are a control station and an adsorber unit. The control station contains the system pump, control valves, pressure gauge and safety Interlocks designed to shut the system down or prevent its activation m the event of problems. The control station also houses the adsorber unit which In turn holds the activated carbon and the liquid distribution system. The adsorber unit Is made from a standard 5S-ga11on drum that meets DOT requirements for shipping PCB wastes. Thus the adsorber unit or "filter* Is first shipped to the job site as a product, used to remove PCB from the transformer, disconnected from the system, bung closures are threaded into place, a PCB label is affixed to the drum, and it is ready for shipment to a disposal site. The adsorber units are shipped to an EPA approved Incinerator where they are shredded and Incinerated. This handling and disposal method minimizes employee exposure and assures that the wastes are totally destroyed. Dow Coming and the service companies also maintain a tracking system for each adsorber, thereby providing "cradle-to-grave" care. The third Important part of the cleanup system Is the ability of activated carbon to selectively remove PCB from silicone liquid. Carbon Is "activated" by high temperature heating to produce porous particles. Total surface area of some types of activated carbon Is estimated to be 1,000 m*/gm. In the process of adsorption, a nearly molecular layer of contaminate (PCB, In this case) binds to the carbon surface. An "adsorptive Isotherm* curve shows the constant-temperature relationship between the amount of contaminant absorbed per unit weight of absorbent and the equilibrium concentration of the contaminant In the bulk coolant (15.). figure 1 shows an adsorptive Isotherm curve for PCB In silicone liquid, using a typical activated carbon. Note that the capacity of the carbon to hold PCB decreases significantly as the concentration of the contaminant decreases. 3-13 HONS 214775 f RETROSIL1M PROCEDURES The RetroSilTM program consists of many essential steps. They Include: Preliminary testing Site preparation fluid exchange Equipment installation Final testing When preliminary testing Is completed, the Independent service company makes a recommendation as to whether or not the transformer Is a suitable candidate for processing. The following tests may be performed to aid In making that judgment: Performance review Insulation resistance Power factor Turn ratio Liquid dielectric strength Leak test . Polarization Index Site preparation Is directed at assuring that proper safety (eye wash, shower, fire extinguishing equipment, protective clothing, breathing apparatus, etc.) and spill protection plans and equipment (plastic sheets, drip pans, absorbing materials, etc.) are In place Including plugging or covering all sewer drains. 3-14 HONS 21*776 Dow Corning recommends a multiple solvent {1,1,1-trlchloroethane) and silicone flush to clean the transformer. The total flush volume Is approximately SOX of the total transformer volume. Effective flushing with draining between flushes can lower the residual PC8 content to less than IX, - With the transformer drained, all gaskets below the liquid level are replaced and any needed transformer repairs are made, The new PCB-free silicone fluid Is then introduced Into the unit. Vacuum degassing can aid In removal of any entrained gas. Next the control station and adsorber are Installed and connected and the silicone liquid Is retested for dielectric strength and water content. The electrical tests described In the preliminary testing paragraph above are rerun before the transformer Is reenergized. A sample Is also taken after reenergizing to establish the residual PCB concentration. PETROS IL PERFORMANCE A hypothetical 1500 kVA, 300 gallon, 13.B kV primary, ABO V secondary transformer having a residual PCB content of 1.0X after filling will follow a typical operating and adsorber changeout schedule as shown in Table II. Four adsorber units operating 105 days plus an additional 90-day waiting period will be needed to reclassify the transformer to "PCB-contamlnoted." The 90-dty waiting period during which time the cleanup system (control station) Is not operating Is required by EPA regulations to reclassify either to PCB-contaminated or non-PCB status. Table II shows that the weight of PCB adsorbed decreases from 21 pounds on the first adsorber to four pounds on eech of the subsequent three adsorbers. This decrease reflects the lowering of carbon capeclty with lowering PCB concentration In the fluid consistent with a typical adsorption Isotherm (Figure 1). The adsorbers holding only four pounds can be reused as the first adsorbers on other transformers. This reuse technique can provide significant cost savings on Jobs Involving multiple transformers at a single location. Table II also reveals Important Information about the nature of the residual PCB In the transformer. Adsorbers number one and two are predominantly removing PCB 3-15 MOWS 214777 that remains freely available In the fluid. That PCB can be quickly removed as evidenced by the 50 ppm PCB concentration In the fluid after only three days. By contrast, adsorber number three and four are predominantly removing PCB that is leaching slowly out of the transformer core (wood, paper, etc.). 'This leaching Is essentially diffusion controlled and is particularly evident In the PCB concentration gain from 40 to 450 ppm shown In the waiting period. Leaching will continue after the reclassification and additional adsorption Is therefore recoimwnded to maintain the PCB concentration In the fluid below 500 ppm. Additional adsorbers will be required to achieve non-PCB classification. The number will depend upon the Individual transformer and, more Importantly, upon adsorber utilization. Based on field experience to date, plus models from lab work, non-PCB reclassification would be expected In about two years following reclassification to PCB-contamlnated status. For that approximate two-year period, the PCB concentration In the fluid can be maintained below 50 ppm, or can be allowed to cycle from, for example, 500 to 50 ppm depending upon customer preference. The former method reduces any risk from spills or other accidents while the latter optimizes adsorber usage and reduces cost. SUMMARY Under present EPA rules, owners of askarel-fllled transformers that pose an exposure risk In food or feed plants cannot continue to use these units after Oct, 1, 1985. Other owners of such PCB transformers are seeking alternatives to the outright replacement of these units, particularly where the removal of the old transformer Is costly In terms of service Interruption or because It Is Inaccessible without major changes In the structure In which It Is housed. One alternative Is to clean the askarel transformer by using silicone liquid retroflll and adsorption techniques to remove residual PCBs. A typical 300-gallon 1500 kVA transformer can be retrofllled with fire-safe, environmentally safe silicone liquid with as little as 12 hours power Interruption. The transformer will then contain 1/50 to 1/100 of the original amount of PCB. After about 105 days, plus a 90-day wait required by EPA regulation, this typical unit can be reclassified as "PCB-contamlnated.* At this point, It will contain less than 1/1000 as much PCB as the original askarel unit. ' 3-16 HONS 214778 r r REFERENCES 1. Federal Register, Vol. 44, No. 106, May 31, 1979. 2. Federal Register. Vol. 47, No. 165, August 25, 1982. 3. Federal Register. Vol. 48, No. 201 October 17, 1983, p.47881. 4. "DOW CORNING 561 Silicone Transformer Liquid," Publication 22-846A-80, 1980. 5. Orbeck, T., "Replacement Fluids For PCB In Transformers: Service Experience and Safety Considerations," presented at the Ooble Conference, April, 1981. 6. Michaud, T. and Orbeck, T., "Report of Silicone Transformer Liquid Commercial Service Life Data Program," presented at the Ooble Conference April, 1979. 7. Page, w. and Orbeck, T., "Performance and Safety Capabilities of Silicone Liquid As Insulating Liquid For High Voltage Transformers," IEEE, Paper PCI-77-22, 1977. 8. Miller, R.E., "Silicone Transformer Liquid: Use, Maintenance, and Safety," IEEE Transactions On Industry Applications. Vol. IA-17, No. 5, September/October l96l, pp 463-468. - 9. Lentz, C.W., "It's Safe To Use Silicon Product In The Environment," Industrial Research A Development.. April, 1980. 10. Frye, C.L.. "Health and Environmental Aspects of Silicones," Soap/Cosmetlcs/ Chemical Specialities, August, 1983, 11. 1981 National Electrical Code. 12. Factory Mutual, Loss Prevention Data Sheet 5-48/15-8S, 13. National Fire Protection Association, "Tentative Interim Amendment 70-811." 14. Kanakla, M., "Characterization Of Transformer Fluid Pool Fires By Heat Release Calorimetry," presented at 4th Int. Conf, Fire Safety, San Francisco, CA, 1979. 15. "Activated Carton Evaluation and Selection," Carborundum Kennecott Corp., reprinted from U.S. EPA Technology Transfer. October 1973, Process Design Manual For Carbon Adsorption. EPA (>Z5/l/7t-002a. 3-17 MOMS 214779 rr DISCUSSION QUESTION: ANSWER QUESTION: ANSWER: QUESTION: ANSWER: QUESTION: ANSWER: QUESTION: ANSWER: QUESTION: ANSWER: QUESTION: ANSWER: QUESTION: ANSWER: Is. derating of the cleaned transformer required? No Is IS residual PCB level after retroflll actually experienced? Speclal 'technlques must be used to get to IS. A single drain, flush will result In St-61. Multiple processing Is required. The flushing solvent Is 60S trlchloroethane. How do you know when the filter unit Is saturated? Watch the PCB level In the fluid. When It starts to rise, the filter is saturated. What control do you have to ascertain that your contractors follow procedures? Dow provides training and technical assistance. Previous experience did not show the success ratio now clalned. Is this a new set of data? Yes. To date, approximately 1/3 of the PCB transformers treated have been reclassified as PCB contaminated. The process can take up to two years to reclassify to non-PCB. To date, only one transformer has been classified non-PCB. Have you examined the possibility of manifolding to speed up the process? Yes we are actively looking into It now. It could Increase the initial reduction. However, leaching is slow and manifolding will not speed up that process. How long does complete decontamination take? About two ytars. What happens to the trlchloroethane used for flushing? It Is drained and any remaining Is absorbed during the process. 3-1B HONS 214780 TABLE I SUWtARY OF CURRENT ERA RULES RECARO IRC TRANSFORMER CLASSIFICATION Rfjutrgannt KB concentration In dielectric liquid KB labeling Inspection progran Ml records Service or rebuild restrictions Carcass disposal Liquid disposal Satisfies `exposure risk to food or food* requirement Non-RCB Loss than $0 ppm () KB-Contanlnated 50 - $00 ppo (2> KB Transforamr 500 ppai or greater (]_) Not required Q) Not required (?) Not required 0) Not required (1) Required (1) Required {f) No rostrlctton (1) Any service by operator or eaaaat service company M) No service that requires coll core renova 1 No restriction (1J No restriction Approved landfill or Incinerator <j_) No restriction except no fluid elth detect* able KBs ten be used as a `sealant, coating or dust control agent1 (1) Approved landfill, high efficiency boiler, or approved incinerator Q) Approved Incinerator (1) Yes {) Yes {2) es (2) t-l HONS 214781 r' TABLE It FLASH ANO FIRE POINT RATINGS AND HEAT RELEASE RATES FOR DOW CORNING S61 Flash Point .................................................... >300*C Fire Point........................................................ >350*C Heat Release Rates Radiative ....................... .... 25 KW/et* Convective......................................53 KW/i* Heat release data generated by Factory Mutual Research in large-scale pool fire tests. Heat release rate values are used in the FM Loss Prevention Data Sheet 5-4S/14-8S 3-20 HONS 214782 r r TABLE III ADSORBER UTILIZATION SCHEOULE TO ACHIEVE PCB CONTAMINATED RECLASSIFICATION Adsorber Number Or Event 1 2 3 4 Waiting Period 4 PCB Concentration (ppm) In Liquid At Beginning End 10,000 1,500 50 50 40 1,500 50 50 40 450 450 75 Lbs. PCB Adsorbed Total Elapsed Time (days) Z1 4 4 3 Note 1 1 3 13 105 195 4 ZOO Note 1: The adsorber unit Is Inoperable during the waiting period. The transformer Is reclassified at the end of the waiting period. 3-21 ' MONS 214783 FIGURE 1 REPRESENTATIVE PCB/SILICQNE FLUIO AOSORPTION ISOTHERM PCB Concentration (ppm) 3-22 HONS 214784 C2C14 AS A SUBSTITUTE FOB PCBs E. J. Walsh Westlnghouse Electric Corporation Sharon, Pennsylvania 16146 When PCB's were forced out of the marketplace due to environmental concerns, there were no non-flammable replacement fluids available as a replacement. The alterna tives were to use dry type transformers or to use "less flamable" fluids. The Electric Power Research Institute, EPRI, and other utility companies Initiated several projects to Identify Insulation systans which would be both economical and heve fire resistance equal to or better than the Arochlors. One of these EPRI-ESSERCO projects resulted in the development of a C^Cl^ tetrachloroethylene based Insulation systm. As part of this original EPRI-ESSERCO project, two transformers were manufactured and subsequently placed In service at host utilities. As of this date, these trans formers are still operating satisfactorily. Additionally, Westlnghouse hes sold, and placed In service, war five kindred CjCl4 filled units since 1981. The intent of the presentation will be to discuss some of the maw variables which were considered In developing the CjCl^ based Insulating systems and also to briefly describe the verification testing which wes used to evaluate these systons. It will not be possible within the short time available to go Into all areas In depth, but additional Infoneatlon can be obtained fran either EPRI or Westlnghouse. Selection Of The Fluid There were four primary criteria used In evaluating candidate fluids for use In transformers - Functionality, Economics, FI amiability, and Toxicity. Each of these factors Is very Important, and a fluid must be capable of meeting all four require ments to be a candidate. 3-23 HONS 214785 Non-flanmable fluids come from only a few classes of chenical compounds - Inorganic polymers, phosphate esters, highly halogenated fluids, and specialty compounds such as water or liquid ammonia. These types of fluids were examined with regard to both functional and chemical re quirements. From this rather large body of potential candidates, only a few fluids met the project goals. FI ansna billty The obvious choice, water, sets the criteria by which all other fluids are judged on flammability. The major problen confronted in Judging fluids was what testing techniques were available for evaluation of non-flannable fluids. These were active programs at both Factory Mutual Research Corporation and National Bureau of Stand ards to evaluate the fire properties of dielectric fluids. FMRC efforts were direct ed at "less fl aura bit fluids", and the NBS project was directed at explosion resist ance of transformer fluids. The FMRC test was essentially a large pool burn whera heat release and weight loss were determined. The NBS tests were basically visual comparisons of the behavior of fluids In explosive situations. Both 100S C2C14 and 75* C2C14/25X oil were subjected to these tests. In neither test dld lgnltlon occur. In fact, In some of the tests, the fluid acted as an ex* tlngulshant for the Ignition sources. Although time does not permit a more In-depth discussion of this topic, details and further Information are available from either EPRI or Westlnghouse. 01 electric Functionality The dielectric performance of the fluid Is an essential requlrenent. Included In this area of evaluation are not only the performance of the new fluids txjt also what are the aging effects and the hydrolytic effects. An actual examination of the candidate fluids based on these functional criteria, eliminated many potential candidates. Only a small number of helogenated or specialty fluids met both the flammability and functionality criteria. Fluids were examined for dielectric and Impulse performance, for hydrolytic stabil ity, for thermal stability, for systems Interactions, and for processing require ments . As would be expected, no fluids could be used as a direct replacmnent for'Arochlors or for oil, It was, therefore, necessary to determine what system changes would be 3-24 HONS 214766 required to utilize each fluid. In the case of and C^C^/oll, it was deter mined that there were both advantageous and disadvantageous properties. Some of the more advantageous properties were the impulse strength of C^C^/oll solutions and the greatly Improved cooling and thermal properties of these C2C14 fluids. The major disadvantages were the low temperature performance of pure C^Cl^ and the materials capability with CjCl4 based fluids. The economic ramifications of using these systens will be discussed later In this presentation. Safety The one area of fluid evaluation which Is continually discussed but not often definitively measured Is safety. Frequently, It Is necessary to make a value judg ment as to which type of safety hazard must be tolerated. With the fluids we are discussing here, fl amiability continues to be the primary concern. In order to assume complete safety from a fire hazard, there has to be a tradeoff In sane other area. C^C14 was chosen as a dielectric fluid, not because It was a totally safe liquid, but because It was a fluid which had available a tremendous amount of both toxicological and handling Information. There Is a saying that "the greater the data, the greater the confusion", such Is the case for CjCl4 safety. In spite of all the conflicting data, there are certain bottom line observations which can be made. 1. CjCl4 Is not an acutely toxic material. . 2. C?C14 does require certain handling and processing techniques con sistent with good chemical practice. 3. The long term exposure data based on epldmlologlcal studies are con flicting In results and may neve* be resolved. 4. Ditto for carcenogenlc studies. 5. C-jCl4 should be Isolated from ground water systens. 6. There Is a lot of C2CI4 being released to the atmosphere, and govern ment efforts are directed at limiting this type of annlsslon. 7. When compared to Arochlors, C2CI4 Is much safer unde* both arcing conditions and fire conditions. I hesitate to make this last statenent because there are those who claim that Czd4, because It contains both carbon atom and chlorine atoms, should be put In the same category as PCS's. To a knowledgabla chmnlst, this statement Is ridiculous. It Is the tronendous chmical differences between the chemical classes of olefins and 3-25 HONS 214787 rr aromatics that make C2C1fl an acceptable alternative to PCB's. Economic s Because Westlnghouse Is a profit making business, the economics of use and process ing was a consideration for any PCB replacenent fluid. Qur goal was to develop a non-flammable fluid competitive with mineral oil. To achieve this goal, we had to balance the improved performance characteristics against the necessary materials and process requiranents. As the project progresses and additional information Is generated, there are continued Improvements being made in the cost of CjCl^ based units. It Is envision ed that the EPRI project goals are still possible. The following statements summarize the current status of transformers cooled with CjC14 based fluids. 1. C2CI4 based fluid Insulation systems are functionally superior to PCB's In many aspects, such as fire, cooling, and toxicity. 2. CjCl^j Is a safe fluid If used with commonly practiced handling tech niques, 3. C2CI4 based fluid Insulation systems are applicable over a wide range of transformer sizes. 4. Field performance over a three year period has danonstrated the superior performance of this type of unit. DISCUSSION QUESTION: ANSWER: what Is the status of the National Cancer Institute study? Both first and second studies have been discontinued. 3-26 HONS 214786 Part 4 SPILL CLEANUP AND BIO-DEGRADATION MONS 214789 r r CLEAN-UP OF SOILS CONTAMINATED WITH PCBs Judith F, Kitchens Laura 6. Mangoba George L. Anspach Edmund A. KobylInski Atlantic Research Corporation Alexandria, Virginia 22312 ABSTRACT Large tracts of land across the country are contaminated with PCBs. Because of the low solubility of PCBs In water, the migration of PC8s through soils Is not as rapid as observed for some other chemicals. However, as the PCBs approach property boundaries and water supplies, decontamination of the soil becomes necessary. The usual disposal method consists of physical removal of the contaminated soil to an approved landfill. Under Ideal circumstances, Impermeable clay barriers have also been used to stop the migration of soil contaminants. Both methods are expensive and do not remove the PCB contamination from the environment. Experiments have been performed that demonstrate the feasibility of extracting the PC8s from soil. 8y choosing a solvent that does not pose an environmental hazard, the soil can be returned to Its original site without Incurring the costs of long distance transportation required for landfilling. Once the PCBs are in the solvent, the lARC process can be used to dehalogenate the PC8s and the solvent recycled. The on-site extraction and destruction of PCBs appears very attractive when compared to alternative disposal/containment processes. MONS 214790 4-1 CLEAN-UP OF SOILS CONTAMINATED WITH PCBs INTRODUCTION Contamination of soils and sediments with PC8s has occurred as a result of transformer failure, accidental spillage and deliberate dumping, PCBs-that enter the environment migrate slowly through the soil eventually contaminating surface and ground waters. Little degradation of these molecules occurs end high concentrations of PCBs can be found in the soils around spi 11 areas 10 to 20 years after the Incident, The currently accepted spill clean-up procedures involves excavation of the soil in the spill area, transport of the contaminated soil to an approved landfill and backfilling and revegetating the spill area. Landfilling of PCB contaminated sons transfers the problem to another location. It does not solve the problem of removing PCBs from the environment. For examp1e< the State of North Carolina is currently facing the problem of what to do with a landfill containing PCB contaminated soils. North Carolina solved Its Immediate PC8 problem by removing 42,000 cubic yards of dirt from its roadsides and placing ft In a landfill. However, citizen reaction to the landfill has prompted the state to look into alternative technologies for decon tamination of the PCB landfill This paper describes the development of a process aimed at removal of PC8s from soil and their ultimate destruction. The process utilizes liquid extraction of the soil and degradation of the extracted PC8s via the LARC process. The extracted soil can be replaced in the original site and revegetated. LARC PROCESS DESCRIPTION LARC (Light Activated Reduction of Chemicals) Is a patented photochemical process (U.S. Patent #4,144,152) which uses ultraviolet light and an optimized reducing environment to dehalogenate various chlorinated, bromlnated and lodlnated organic compounds. Since the photochemical reaction Is Initiated by the absorption of light energy, the irradiation wavelength must match the absorption band associated with the bond of interest in the molecule and the solvent must not absorb significantly at the Irradiation wavelength. For PCBs. low pressure mercury lamps which emit approxi mately 95* of their energy at 2537A provide adequate Irradiation. Solvents which meet the optical criteria Include water, alcohols and hydrocarbons. However, these solvents have very different effects on the process efficiency and mechanism. In water solutions, the hydrogen gas provides the hydrogen source for the hydro deha logenatIon reaction as evidenced by the Incorporation of into kepone subjected to LARC In NaQD/D?Q solution. For PC8s In heavy hydrocarbons, the preferred reaction 4-2 HONS 21*791 is the polymerization of the biphenyl radicals to yield polyphenylenes. In basic alcoholic solutions, a stepwise dechlorination occurs In the LARC reactor. This photochemical reaction occurs via the triplet state and the reaction rate is dependent on the life time of the triplet state. The hydrogen for the hydro deha logenat ion initially comes from the solvent. In alcoholic solutions, the hydrogen gas Increases the reaction rate by removing triplet quenchers from the solution by combination with the solvent free radicals formed. In addition to the hydrogen, small amounts of sodium hydroxide are added to the alcoholic solution as a chlorine scavenger. These photochemical conditions lead to a rapid, highly con trollable reaction which yields only biphenyl and sodium chloride as the final products. No oxygenated derivatives, chlorinated dlbenzofurans or chlorinated dioxins have ever been observed In the gas chromatographs or the mass spectra of the intermediates or products In the LARC degradation of PCBs. METHOOS ANO MATERIALS The application of LARC to decontamination of soil contaminated with PCBs requires that the PCBs be efficiently removed from the soil medium. Studies were performed to determine the extraction efficiency of Aroclor 1260 from soil using Isopropanol as the extraction solvent. This solvent was chosen for evaluation in the laboratory because it Is a good LARC solvent, the solubility of PCBs In isopropanol Is greater than In methanol or ethanol and It Is relatively inexpensive. For the Initial extraction studies, a clay soil (2.OX organic matter; pH 5.6; Ca, Mg, P, and K levels were 660, 77, 56, and 20 mg/kg, respectively) was used. Air dried soli was spiked with Aroclor 1260 In acetone, thoroughly mixed and the acetone allowed to evaporate in a hood. Random subsanples of the soil were taken, combined and Soxhlet extracted using a mixture of SOX acetone and SOX hexane. The extract was diluted to the detector range and analyzed on an HP 5880 GC with computer controller/Integrator and autosampler. Separation was affected using a 10 ft. x 2 mn 1.0. glass column packed with 1.5X SP2250/1.95X SP-2401 on 100/120 mesh Supelcoport and nitrogen carrier gas at a flow rate of 25 el/mln. The oven temperature was held at 195C for 2 minutes and then programed at 10OC/mln to 225C (hold for 1 minute), then programed at 4<>C/m1n to 244C (hold for 15 minutes), The injector was maintained at 270C and the 3ni electron capture detector was maintained at 300C, This procedure yielded an Initial concentration of Aroclor 1260 In the dry soil. Weighed amounts of dried, spiked soil (approximately 330 g) were placed In two one liter beakers, One hundred mL of distilled water were added to the soil In one of the beakers and the soil and water thoroughly mixed to yield a wet soil containing approximately 23X water. Five hundred mL of Isopropanol were added to both the wet and 4-3 HONS 214792 dry soil and the mixtures stirred on a magnetic stirrer for 10 minutes. The liquid was then decanted and the volume measured. This procedure was repeated a second time. A portion of each extract was taken, diluted and chromatographed under the above conditions to determine the amount of Aroclor 1260 removed from the soil. The soils were then air dried, a random sample taken from each, Soxhlet extracted and analyzed using the above conditions to determine the amount of Aroclor 1260-remaining in the soil. The two extracts from the dry soil were combined, sodium hydroxide pellets added to form a 21 solution and a portion of the solution analyzed to determine the Aroclor 1260 concentration. The remainder of the solution was subjected to LARC using the single tube batch recirculation reactor shown in Figure 1. The reaction rates In this single tube reactor are significantly lower than a multiple lamp, high light densi ty reactor, however, only 000 mL of solution are required for operation as compared to 30 gallons for our 64 lamp pilot unit. Since the scaling factors between the two reactors are known, this single tube reactor Is a convenient unit for evaluating the degradation rates under differing conditions. The reaction was started by placing the solution into the reactor. The hydrogen gas was then turned on and allowed to purge the reaction mixture for 5 minutes. The ultraviolet lamp was then turned on. Samples (10 mL) were drawn from the LARC reactor after 20, 40, 60, gO, 120 and 180 minutes so that the degradation reaction could be followed and reaction rate determined. These samples were diluted to within the detector range and analyzed on a Varlan 3700 gas chromatograph with autosampler. Computer controller/integrator was provided by an HP 5660 GC. The separation was performed using a 12-foot by 2 mm I.D. glass column packed with 1.5% SP-2250/1.95* SP-2401 on 100/120 Supelcoport. The carrier gas was N2 at a flow rate of 20 mL/1n. The oven temperature was held at 180C for 1 minute, then programmed at lOOC/mln to 240C (hold for 20 minutes). RESULTS The extraction data for soil contaminated with Aroclor 1260 are sunsnarlzed In Table 1. Extraction of dry soil resulted In 70-75% removal In each stage with a solvent to solids ratio of 1,2/1 (wt/wt). The overall two extraction efficiency was 92* with 74* solvent recovery. The extraction efficiency for the wet soil was slightly lower, 55 60* In each stage. However, 12 mg of Aroclor 1260 or 7.5* of the Initial PCB concentration In the soil were unaccounted for in this extraction. The recovered solvent (77*) also contained significant quantities of water. Overall extraction efficiency based on the PC8s remaining in the extracted soil was approximately 90*. In a real-life situation, two solld/llquld extractors would be required-with each 4-4 MONS 214793 r r N1 Vt UV laiM an* IlMTl tM < Flgur* 1. Schematic of Single Lamp Lire Reactor HONS 214794 4-5 Table 1 EXTRACTION OF AROCLOR 1260 FROM SOIL Initial PCB Concentration {mg/Kg)* PCB Removed From Soil in First Extraction {mg/Kg)* Solvent Recovered From First Extraction (mL) PCBs Removed From Soil in Second Extraction {mg/Kg)* Solvent Recovered From Second Extraction {mL) Final PCB Concentration in Extracted Soil {mg/Kg)* Overall Extraction Efficiency Dry Soil 487 337 320 113 420 38 92X Wet Soil 486 268 360 131 410 50 901 All concentrations based on dry weight of soil. HONS 214795 4-6 extractor using a solvent feed free of PCBs. Solvent/solids ratios may also have to be Increased depending on the Initial PCB concentration in the soil. The results of the LARC degradation of Aroclor 1260 in the Isopropanol extract of the dry soil are shown In Figures 2 and 3. Degradation proceeds rapidly even in the presence of particulates with a pseudo first order rate constant of 0.052/mln. In a larger reactor with greater light density, the degradation rate can be expected to be 2 to 2.5 times that of the single tube reactor. DISCUSSION The laboratory data have shown that PCBs can be efficiently extracted front even a wet soil using a solvent such as Isopropanol. The PCBs in the extract can be successfully degraded by the LARC process. However, in actual spill situations, the concentration of PCBs In the soils will vary considerably with lateral distance from the source and with depth. Any process must be capable of handling extremes In PCB concentration and " still maintain preset criteria for removal from the soil and PCB destruction. The data on PCB extraction from soil and degradation with LARC were combined with similar data obtained on other chlorinated organics to design a mobile unit that can clean-up spills of PCBs from soils. The process schematic Is shown In Figure 4, As envisioned, this equipment will be mounted in a standard trailer and consist of two counter-current extraction units to remove the PCBs from the soil. After extraction, the soil will be vacuum stripped to remove and recover residual solvent. The soil can then be placed back into the excavated site. The extraction solvent will be sent to a stilt where the solvent will be recovered for reuse In the extractors and the PCBs concentrated. The extraction and concentration steps will be operated approximately 6 hours per day. The concentrated PCBs will be stored in a surge tank. The extractlon/dlstH latlon step will be run to produce enough concentrated PCBs to provide feed for the LARC reactors on a continuous 24-hours basis. The concentrated PCBs will be metered into the LARC units continuously over a 24-hour period. Thus, the PCB concentration entering the LARC units can be tightly controlled even though the concentration In the soil varies dramatically. This process design provides the flexibility to treat any type of soil at varying PCB concentrations and still maintain optimum rates In the LARC reactors. Costs for PCB removal from soil and destruction by the LARC process are dependent on the average PCB concentration In the soil. In most soils, high concentrations (e.g. 90,000 mg/Kg) will be found close to the spilled area. Once out of the Inaedlate spill HONS 214796 4-7 rCB* R tm la lH / Figure 2. LARC Degradation of Aroclor 1260 In Basic Isopropanol HONS 214797 4-8 Figure 3. ChroMtogrMs of PCBs lefore, During end After LMC Treetmnt HONS 214798 IMMihd -Q *|l atrMtM T -0--j 4 ftita lUtffW H l*H -ffi 0- - ry?-~ MW rChQn -CH> HZHOHD-CF ttftC MH lUtelfliHM) MM Vl*l ftfTBi B *V lf*u Figure 4, ScNttMtlc of LAAC Mobile Unit for Destruction of PCBs In Sol) HONS 214799 r r area. the PC8 concentration will drop off rapidly. However. If leaching has occurred, large areas contaminated between 50 and SOO mg/Kg can be expected. In our cost calculations, we chose an average concentration of 1500 mg/Kg of PCBs, Items included in the cost estimates are capital expenses, dally operating costs, labor, analytical, travel, per diem, and profit. At an average soil concentration of ISOOmg/Kg, total costs are S64.60/ton of soil. Landfill costs for PCBs are 170/tgn including tax. Transportation charges add 12.75 per loaded mile or approximately ISO/ton for a 400 mile trip yfeldlng a disposal only cost of J120/ton. Excavation costs, analytical costs and refilling costs must be added to the cost for landfilling the contaminated dirt. These costs could add an additional SSO-lOO/ton depending on the required analyses. Thus, LARC costs for clean-up of PCS contaminated dirt range from 38 to 50X of the costs associated with landfilling the contaminated dirt, but more important it removes PCBs from the environment permanently. 4-11 MONS 214800 discussion What throughput Is expected? A. LARC Is designed to degrade 6500 ppm/rln. It witf vary, but with 1500 ppm In soil, ft should be good for 15-60 tons per day. High tavals of organics will slow the process down. Have you published data on spill clean up? A, Ho. and don't plan to. Have you constructed a full scale unit? A. Ho. This work was with 21 organics. Any work with higher levels? A. Ho. What are the power requirements? A. lo provide power from Internal generator. Not a great problem, what is your estimate of cost? A. 184/ton Including excavation from site, but not Including return to the excavation. Is It a batch or continuous process? A. Could be either. 4-12 MONS 214601 r r 06. What H of PC8 is left in the solvent? A. Less than 21. 09. Dots EPA allow solvent extractions? A. EPA currently only allows landfill, but en EPA re|Tresentat1 v# present at the Seminar indicated willingness to review the process for possible approval. 4-13 HONS 214802 CLEANING OF PCS CONTAMINATED SOILS Lao Weltraan, PhD Acurax Corporation Cincinnati! Ohio Tha claanup of PCB apllla la a major concam to cha utility lnduatry, Thaaa occur If a capacitori transformer. switch or othar PCB-containing equipment develops a leak whils in aarvica and contaminates tha surrounding araa. A claanup coat of $30,000 la not uncooson for a leaking PCB transformer or capacitor, aapaclally if it had gone undatactad for a parlod of tiaa. In addition to tha claanup of PCB laaka from flald aqulpaant, tha utility lnduatry la facad with tha long-term problaa of tranaforaar yards, rapalr faeilltlaa, or subatatlona that hava outlivad thair usefulness. Thaaa oldar faeilltlaa can ba aavaral acraa In alia and bacauaa thay wara uaad at a tiaa whan PCBa wara not ragulatad, they ara oftan haavlly contaminated. Tha currant method of damning up a small aita auch aa a laaking tranaforaar, involvaa eiaply raaovlng tha aoil and taking It to a chemical waata landfill that haa bean approved for PCB disposal. Tha alta la back! 11 lad with daan dirt, Tha aajor expense of thia procasa coaaa fro* trucking tha dirty soil and disposing of It In a landfill. Whan large araaa ara contaoinatad, hauling Cha aoll to a disposal aita bacomae lapractical, Thara ia often insufficient landfill capacity available for lta dis posal and If it ia available, it la vary expensive. In thaaa cases, tha contami nated aita la isolated by eoae fora of groundwater barrier, auch as a grout curtain, and than capped to reduce rainwater Infiltration. Than, if necessary, groundwater la piaped froa this conflneaent into a treatment plant. This creates a net inflow of water into the contaminated aita and reduces tha possible migration of tha contaminant into cha environment. Thia approach Involvaa long-term expenses which can ba formidable. The program described hare examined aoll washing as an alternate PCB claanup procedure. Tha technique haa tha potential to remove PCBa from tha soil onsite and hence, eliminate both hauling, landfilling and backfilling for cleanup, thereby reducing tha coat substantially. Two aoll washing techniques wara examined, Tha first, solvent washing was found applicable to moat situations even where tha final 4-15 MONS 214803 PCB concentration in the soil must approach zero. The aacond technique, water washing proved to bo an Impractical procedure. the quoation of "how clean is. dean" will ariae in any type of a cleanup operation. In eomo caeee> any material containing leae than 50 ppm PCB la conaiderad clean. In othara, no measurable PCB la allowed. After diacueelng thla with a number of people at various Environmental Agenclea, we established 2 ppm as an acceptable level of cleanliness for PCB. THEORETICAL CONSIDERATIONS The experimental program described in the next section sought to determine the qulllbrlim constant that daacrlbea the actual concentration of the PCB in the liquid after eech wash. The process of washing soil is directly analogous to thst of adsorbing pollutants from e wastewater using, for example, activated carbon. In both cases, one first needs to determine the concentration of the pollutant (PCB in this case) in the liquid pheee that is in equilibrium with e given concentration on the solid phase. These values over the range of concentration ere coMonly referred to ee the adsorption isotherm for the system. In the case of soil washing, ue are actually dasorbing rather then adsorbing the pollutant but the process is entirely analogous from an engineering point of view. In addition to having a favorable equilibrium, it is Important that the rates at which the pollutant diffuses into the washing fluid ba fast enough ao that the overall process occurs in a reasonable length of time. This proved to ba a problem when using pure Freon with top-eoil. Tha equilibrium was favorable; however, it took up to 18 hours for the system to reach equilibrium. This would not raault in a practical solvent washing system. With the proper choice of solvents, the time required to reach equilibrium was reduced to 30-45 minutea, a very practical value. The design equations for tha removal of PCB from solid materials waa presented in graeter detail in an earlier report (1). Based on that derivation, it la possible to daflna a constant "K" relating tha PCB concentration in tha liquid coming from two consecutive waahea of the soil. It la defined as; K Wi Where yfl and y^ are tha PCB concentrations in tha liquid leaving the nth and (n-l)th washes re< pectively. On thla basis, it can ba shown that after N washed, tha amount of PCB raeldua left on tha aoll. S is: SN - L y^/U-K) 4-16 HONS 214804 rr Where S le weight of PCS left on the aaapla of eoll end L le the volute of liquid Meed for eech weah. L le conetent froe weeh to weeh. The leboretory lnveetlgetlon for thle eyetee, ehowed thet euch e aolvant weehlng eyetea cen echleve e velue of approxlaetely 0.5 for PCBe In varloue typae of eoll. Th;> velue proved to be raaeonebly independent of the type of eoll being cleaned and the PCS concentration of the eoll ltaalf. Thle neana that the weeh liquid leaving a weeh will have a PCB concentration of approxlaetely half of the preceding weeh. The aolvant uaad for the weehlng will Influence the econoelca end aefaty of the ayetea greatly. Ir nuat have the following propartlaa. 1. Mlaclble with PCB to high coneantretlone 2. Non-flovable for aefaty In the field 3. Low latent heat of vaporltatlon end low boiling point for eaae of recovery 4. Low toxicity 5. Be able to "cut through" water 6. Capable of wetting the eoll 7. Readily available The atudy axaalnad following aolvanta. rapraaantlng a range of propartlaa. 1. pure hexane 2. pure PC-113 (1, 1, 2 trlchlorotrlfluoroethana) 3. a proprietary eolvent blend . 4. FC-113/haxana blend It waa found that only the proprietary eolvent wee able to reduce the PCB concentra tion in the soil to below 2 ppe. In addition, thle aolvant aetleflaa all of tha other criteria Hated above. LABORATORY EVALUATION Tha firat step in the laboratory evaluation waa to contmlnate aoila with PCla. Tha following typea of aoila ware aelectad, 1. Pure sand 2. A high-clay eoll 3. Potting soil, rapraaantlng a high carbon aoll 4-17 ' HONS 214805 The solvent washing experiments ware conducted by placing 100 grama of contaminated aoll Into a Jar with a tight fitting lid. Tha aoll wea than covered with tha sol vent (100-200ml) and than tha Jar waa placed Into an ultrasonic bath. In the earlier testa of this aarlaa, aemplaa of tha solvent wars withdrawn at approximately IS minute Intervals so aa to determine tha time required to reach equilibrium. This showed that tha equilibrium la reached in essentially 15-43 mlnutaar Tha purpose of thaaa taste waa to establish whether these soils would be cleaned to laaa than two ppm PCS and how many washes this would require. It la recognised that tha number of waahaa required will ba function of tha volime of liquid uaad. Tha data can, however, ba convartad to a aarlaa of "daaorptlon laotharma" similar to the "adsorption isotherms" cononly uaad to alia edeorptlon coliwia for waatewatar or gaa cleaning. The teat ware conducted by placing 100-200 grama of the soil Into a Jar with a tight fitting lid. Enough solvent to cover tha aoll with a 1-2 cm excaaa waa than added to tha Jar and this volute of solvent determined. All successive weahad ware conducted using thla ana voltae of aolvant. Soma taata ware conducted by using ultrasonic agitation to facilitate tha cleaning operation. In thoaa taata which Included ultrasonics, tha Jar waa placed Into a small ultra sonic cleaning unit which was fillad with watar aa a transfer madlta. Tha ultrason ic cleaning waa found to ba a very effective aid In removing the PCX. In order to aatabliah how long a contact tine waa neceaaary to raach equilibria samples ware withdrawn from tha Jar at intervals ranging from 15 mlnutaa to over night. It waa found that with tha eld of ultraaonlca end (ha proper choice of sol vent, tha PCX concentration in tha liquid reached virtually a constant laval within 30 mlnutaa. Thla wea not tha case whan pure FC-113 wee used aa a washing madita. Evaluation of the rasulta of the potting aoll taata showed aa much aa three-fold In crease In the PCX concentration of tha liquid phase when kept overnight, This la contrasted with the waahaa using tha proprietary solvent which showed little or no change in tha liquid PCX concentration even over a weekend. Xaakar taata were elao conducted with combinations of these soils using water and a watar/laopropanol aolution. Thaaa proved very disappointing for all aolla axcapt, to a limited extent the pure sand, Thaaa liquids removed tha PCX from tha aoll In a linear rather than in a geometric manner. 4-18 HONS 214806 r r Mian claanlng aolla down to tha vary low lavela of PCB raqulrad hara. It la eaaantlal that aach successive waah remove a fraction of tha PCB rather than a flxad amount. Tha aoount of pollutant removed auat ba proportional to tha amount on tha aoll resulting in a geometric reduction In tha PCB concentration. Tha aquaoua baaad olvanta cauaad a linear reduction of tha PCB; aach waah resowed anywhere from 3 to 20 ppm. Reduction of tha PCB concentration from, aay 1000 ppm to 100 ppm could taka eevaral hundred waahaa. Thla la Inherently not Inpoaalbla; howavar, tha affluent froa aach waah auat ba treated raaultlng In thouaande of sallona of water for aach cubic yard of aoll procaaaad. It waa concluded that aquaoua baaad aolvant vaahlng of aoll doaa not appear to ba promising approach to achlavlnt tha low PCB concentration raqulrad. Tha bulk of thie experimental program investigated whathar It la poaalbla to remove PCBa froa aoll ualng waahlng tachnlquaa. Onca tha PCB la resowed, hovavar, it la atlll nacaaaary to dlapoaa of It, Tha overall achase raqulraa that tha aolvant be dlatlllad off and purified after the waahlng of tha aoll and tha PCB and other pollutanta left behind in a reactor. Tha concentrated PCB can than ba ahlppad to an In cinerator or destroyed ualng a eodlta baaad reagent. All of tha aolvant from thle experimental program waa treated In thla manner. It waa accwulatad in largo flaaka and dlatlllad. Tha PCB raaldua waa than auccaaafully daatroyad ualng the Acurex PCB daatructlon procaaa. Pinal PCB concantrationa of laaa than 2 ppm (the limit of maaaura) vara obtained In tha liquid. Thla la conaldarad to ba a non-PCB material by anvlromental aganciaa. CONCEPTUAL DESCRIPTION Whan atartlng thla type of raaearch program It la Important to eatablleh a design goal to work towards. Thla la a goal rathar than a dasign; further raaaaarch will undoubtedly cause major changes In tha final operating ayatem. Tha initial design la, however, useful to determine tha type of data that tha preliminary research pro gram nuat produce. Tha basic aoll waahlng ayatem la shown In Figure 1. Please note (again) that thla la a conceptual rathar than an acutal daaign. It la recognised that because of tha requirements of materials handling, structural strength and mobility, tha final unit will moat likely look significantly different. fundamentally, tha unit consists of aavaral (tha drawing shows thraa but that la not flxad aa yet) cella mounted on a trailer. Tha cella are eat up so that soil can 4-19 HONS 214807 FIGURE 1 CONCEPTUAL DRAWING OF SOIL CLEANING SYSTEM 4-20 HONS 214608 f readily b* put Into them ualng standard aarth moving equipment. The ayatam can be tlppad for aaaa of removing the cleaned aoll. Since the aolvant to bm uaad la vola tile, each cell can ba sealed tightly. The multiple calls raault In an aven distri bution of the aolvant. The laboratory teat found that ultreonlc vibrations greatly enhance the efficiency of tha washing cycles. The large aoll washing systems have vibrators damped to tha walls to duplicate this. Thaaa are similar to vibrators used in pouring concrete and are commercially available. It la anvlaloned to ba operated In tha following mannar: 1. Tha unit along with a portable aolvant reclamation system la transported to the alts of tha aplll or contamination and aat up. 2. Soil la placed into tha washing unit than It la aaalad closed, 3. Tha aolvant la pimped into tha waahar and raclrculatad for approximately 30 mlnutaa. 4. Tha aolvant la pimped out of tha waahlng unit into the aolvant reclamation system and claan aolvant la ones again added to tha washing unit containing tha partially cleaned aoll. 3. This cleaning cycle la repeated until a sample of tha solvent shows a sat isfactory PCX level. 6. At that point, tha residual solvent on tha aoll la raclaimed by applying vacuum to the waahar. It may ba naceaaary to also uaa a small amount of ataem. Aftar cleaning, tha aoll la dimped back into place, compacted and nutrients added ao that it will sustain vegetation. 7. The aolvant from each wash la paaaad through a aolvant reclamation ayatarn that includes a distillation column and a secondary stripping ayatam for concentrating the residual PCX pollutant rmaoved from tha aoll. 8. Tha PCX concentrate from tha aolvant ralalmar la sent to a chemical de struction ayatam which produces a PCX-free sludge and a PCX destruction leagent which la recycled to tha proeeaa aftar filtration. Tha ayatam schematic la shown in Flgura 2. CONCLUSIONS This study haa shown that It la possible to wash a wlda variety of soils fraa of PCX contamination ualng a aolvant ayatam that can ba used In tha field. It also showed that at this tlma tha uaa of water aa a cleaning mediim does not appear to ba prac tical In moat ceaea. Tha atudy further showed that the proprietary aolvant uaad for tha aoll cleaning satisfies all of tha criteria aat for It. It la relatively non toxic, easily recovered and non-flaMabla. This la vary important If a workable ayatam la to designed on tha basis of this work. MONS 214809 4-21 FIGURE 2 SOIL CUMING HWCBS 9CHPHTJC HONS 214810 quit* faaaibla to ua* th* Acurax FCB daatruction procaaa in th* field to daatroy tha pollutant* vaahad from tha aoll at th* alta. Thia complete* tha total claanup and will maka tha overall procaaa readily acceptable. In dealing with uch an oparatlon, political acceptability ia aa auch a factor aa are th* normal technical and aconoalc conaldaratlona. Tha coat of tatting local approval* or paraita can, at tiaaa oxcaad tha coat of doing tha Job ltaalf. ay making thla oparatlon totally aalf contained, tha chanca* of lea aeeaptanc* ara optimized. Having ihovn that tha aolvant vaahlng of FCB contaminated aolla la poaalble, it la nov worthwhile to determine lta coata. To do ao. It would b* uaaful to build a pilot acala unit, approximately 55 gallon* in volia* to vaah aolla at a graatar rateThla unit could ba uaad to anawor th* following quaatlon*: 1. What la tha boat ay#tea configuration to permit good contact batwaan the aoll and tha aolvant? 2. What liquid flow rataa can b* uaad ao aa to minimize aoll antralnaant In tha filter madlia? 3. What type of filter madiia auat be uaad comarclally to effectively prevent th* aoil from getting out of tha vaahlng ayatem with tha aolvant and yat not become clogged? 4. - What la the optima liquid to aolld ratio that need* to be uaad for each vaah? 5. What are tha Manpower requirement* for auch a comardal ayatem7 6. What at* tha overall coata of operation for varloua aisa cleanup*? REFERENCES . (1) Waitman. L.. Barth, R, and Faryan, S. "Feaalbillty Study of Chemical Detoxification of FCB Capadtatora, Phaae 2 Study" Draft report on EPRI Project 1263-7 (2) Waltaman, L. and Millie, J. "Faaelblllty Study of Chemical Datoxlflcatlon of Polychlorinated Biphenyl Capacitor*; Fhaae I Study", EFRI CS-2477 June, 1982 ACKNOWLEDGMENT Tha work daacrlbad heroin vaa aponaored by th* Coal Combuatlon Syatama Dlvlalon of tha Electric Power Raaaarch Inatltut*. Their aupport and tha technical aaalatance of tha XFRi project manager. Dr. Ralph Kernel la gratefully acknowledged. 4-23 HONS 214811 rr 01 scusslon Ql. Have you measured for solvent reslduels In ground water? A. U have not tested. However, the last step before returning the sell Is to steam clean. It comet out very dry. The materials used for clean up are generally not toxic. Q2. What reduction of dioxin can be expected? A. Initial concentration of 450 ppb reduced to 3 in & washings. Ut was noted In sampling, testing, and analysis at the 3 ppb level that It Is very difficult!. 4-24 MONS 214812 f THE COST OF PCB SPILL CLEANUP: A QUESTION OF DEGREE JOHN P. WOODYARD IT Corporation Wilmington, California W. COREY TRENCH Consultant for Environmental Health Management Westlake village, California PETER MCCORMICK THOMAS JUHASZ SCS Engineers Bellevue, Washington ABSTRACT The degree of PCB spill cleanup required by regulatory agencies, or "how clean Is clean", hat evolved Into one of the most prominent Issues surrounding PCBs today. Despite the continued occurrence of PCB spills throughout the U.S.. spill cleanup degree Is not explicitly addressed under TSCA. The degree of cleanup required to avoid a disposal citation varies with EPA region and state, to the Inconvenience and cost of Industry. This paper reviews the current state of regulation for PCB spill cleanup, and assesses the compatibility of these regulations with current practice. Several generic spill situations are presented for use In quantifying the cost of alternate cleanup procedures. Based on these situations, a cost/engineering model Is presented. Results of the model are shown. Including graphics to Illustrate cost sensitivity to cleanup degree and spill typa. Conclusions are drawn regarding degree>of'deanup policy, its economic impact and the future regulatory cl imate. 4-25 HONS 214813 r' Section 1 INTRODUCTION The August 1982 PCB rulemaking allows much of the utility-owned PCB equipment to remain In service through the end of Its useful life. While options for disposal of well-managed PCBs are now better defined, small PCB spills can still occur while the equipment Is In service or storage. Proper spill management and emergency response planning can therefore close the perceived environmental hazard loop. EPA has developed draft guidelines for PCB spill cleanup. The docunent has not yet been released for public dissemination, primarily because of the Inherent diffi culties associated with general guidelines geared tmard site-specific problems. Regardless, some form of guidelines may ultimately be published for suggested use by the utility Industry. More Importantly, there are no universal rules governing the degree to which a spill must bo cleaned up. While It Is generally recognized that the spilling of PCB fluid to soil or other surfaces constitutes Improper disposal If not cleaned up, the question of "how clean Is clean" often remains a mystery until the appropriate authority Indicates that the area Is clean enough. Each EPA regional office uses different cleanness criteria, some states are more stringent than EPA, and the criteria can change with time even If established by law. The purposes of this paper are to review the status of PCB spill cleanup regulations, discuss the nature of PCB spills and approaches to cleanup, and to attempt to com pare costs of cleaning spills to several levels of cleanness. The authors recognize the potential difficulty of modeling spill cleanup costs due to the unique nature of each event. Where appropriate, we have qualified our estimates and attempted to quantify the potential variability. 4-26 HONS 214814 Section l PCB SPILL CLEANUP TECHNOLOGY Despite the sincere efforts of scientists end engineers attempting to advance the state of the art, there are typically only two or three acceptable methods of clean ing up a given spill. For spills to soil. It Is excavation and removal. For spills to solid surfaces (houses, cars, concrete, etc.), solvent washing or high pressure steam cleaning Is sometimes acceptable In lieu of removal or disposal. For replace able and/or porous surfaces. It Is replacement. For discussion purposes, three generic types of spills can be Identified: trans former leaks, capacitor leaks and capacitor spraying (under pressure), with the proper training, the first respondent on the scene will attempt to visually Identify the extent of the spill, seal off the area (Including a buffer zone), mark the areas of expected contamination and proceed to use one of the above-mentioned cleanup techniques. Once the visible traces (and expected areas) of contamination are cleaned, sampling Is performed over a predetermined grid for residual PCBs. A second cleanup Is often appropriate If remaining contamination Is high In some localized areas. In the absence of a concentration-based standard, cleanup might often stop there. This Is not always the case. Attempting to estimate the cost of spill cleanup under these conditions Is difficult. Costs are nevertheless important to budgeting, risk assessment and bidding. It Is not unconinon for project budgets to be exceeded, sometimes by a factor of two or more, due to disagreement over completion criteria. Nevertheless, basic practice does not vary substantially between companies for a given type of spill cleanup. The practicality of an extended cleanup should there fore be a factor In determining completion, assuming an acceptable and prudent cleanup strategy Is proposed and Implemented. HONS 214815 Section 3 REGULATIONS GOVERNING PCB SPILL CLEANUP The cleanup of PCB spills Is addressed directly and Indirectly throughout the 1979 and 1982 rules. The August 19B3 "Question and Answer" book published by EPA* pro* vldes the most recent Interpretation of Its regulations. The booklet addresses three questions on the subject: spill reporting, what happens upon reporting a spill and what should be done to control or clean up a spill. The answers to these questions frame. In part, the agency's policy. It is open to interpretation, and thus explains the great variety of spill cleanup "standards" experienced today. On the reporting of spills, the agency cites the authority under TSCA Section 8(e) to require the reporting of spills "whenever the Incident poses a substantial risk to human health or environment." They note In the response that "substantial risk" cannot be precisely defined. However, they say that for any spill where people can come Into direct uncontrolled contact with PCBs, or the spill Is such that signifi cant numbers of animals are exposed, the spill must be reported to the National Response Center. The agency also cites the requirements of the CERCLA Section 103 (a) (48 CFR 23552 23605), DOT (49 CFR 171.15, 171.17), and the Clean Water Act (40 CFR 117.2). The present reportable quantity for PCBs Is 10 lbs. The agency has proposed lowering that quantity to 1 1b. Similarly DOT has followed suit In an advanced notice pro posing the reduction of the reporting requirement to 1 lb. Concerning the question of spill cleanup, the EPA suggests the following steps: (1) Report spill; (2) Control the spread of spill by damning or diking the leak (a high priority should be given to spills posing a threat to water); (3) "Once contained, cleanup can be simply the removal and subsequent disposal of contamin ated soil and debris." EPA further notes: "In some cases special filtration or * The EPA Regulations Under TSCA: Over 100 Questions and Answers to Help You Meet These Requirements. Revised Edition ho. 3, August 1983. Prepared by the iicA Assistance Office and Exposure Evaluation Division, U.S. EPA. 4-28 HONS 214816 PCB sorbents may be required"; (4) "Since the levels required for cleanup sometimes vary, depending on the region (state) In which the spill occurred, regional PCB experts should be contacted to obtain guidance on the extent of PCB spill cleanup." That Is the extent of what the agency has to say about the degree of spill cleanup. They do provide a list of experts and phone numbers. It Is Interesting to note that the June 1979 version of this booklet differs from the 1983 book In some of Its responses. For one. as a general rule, spills Involv ing a single capacitor do (did) not have to be reported unless PCBs threaten or enter a watercourse. However, transformer spills, other than minor leaks, should be reported. So where does all this leave the PC8 owner attempting to clean up a PCB spill? Since the promulgation of rules In 1979, many utilities have felt that cleanlnq up to 50 ppm would be an acceptable level of cleanup. This level could be found more from an Interpretation of the 1979 regulations than an explicit statement. Many EPA regions were advising, and continue to advise, utilities that this Is an acceptable cutoff. The EOF vs. EPA suit has undoubtedly had some impact on the state of the situation today, leaving a wide variety of approaches and standards, primarily because of the acceptance of the 50 ppm standard as Justified and supported. The cutoff issue relates more to Incidental production (and Is still pending a final rulemaking), but could still be considered applicable to those grouping for a numerical rather than procedural cutoff. Meanwhile, there have been a number of test cases with a bearing on the degree of spill cleanup. The agency has prepared an enforcement strategy and Interpretation of the PCB regulations that Indicates that the agency can enforce a cleanup to back ground standard, and Is endeavoring In some regions to do this. EPA Enforcement Headquarters Insists that the regulations provide them with the authority to proceed on a background standard, 1,e. any exposure to PCBs Is significant, however Incon sistently It Is applied. 4-29 HONS 214817 K Certainly, a major contributor to the confusion surrounding PCB spill cleanup is the variation In policy between EPA headquarters, regional offices and state agencies. Given the lack of a scientific basis for any "one number" cleanup cutoff. It would appear difficult to Justify the background PCB concentration as a cleanup level based on these arbitrary local limits. A straw poll of EPA regional PCB cleanup policies Is sumnarized In Table 1. Note the variation In both cleanup level and surface treatment requirements. The most stringent of those regions polled Is EPA Region 5 (Chicago); a quotation from their current internal policy memo stating the recommended practices for three types of spill locations (degrees of potential human exposure) 1$ as follows: (1) Limited Contact Area (l.c,. Indoor industrial Settings)? Cleanup of PCBs should be conducted down to 100 times background but not exceeding Bo uq/nWl cm*. cleanup down to l() ug/lOlT'cm^ wouli be preferrable, Control samplesarweVl as post cleanup samples would be required. Air sampling following cleanup should also be a mandatory requirement to ensure that the OSHA Tlv value for PCB of 1 ug/m3 Is not exceeded If unprotected Individuals will be present or working in post cleanup area (NIOSH Criteria Document for PCBs, 1977). (2) High Contact Areas (l.e.. Including Industrial Settings, Food, Feed, Schools): Cleanup should be required down to background levels. Post-cleanup samples and air sampling would also be Included. Any food prepared on surfaces pre viously contaminated with PCB should be tested for PCBs before human or animal consumption Is permitted. FDA has developed guidelines and limits for PCBs in various foods, feeds and packing materials (F.R., Vol. 44, No. 127, June 29, 1979, pp. 3B330-38340). (3) Environmental Spills: Contaminated soils should be removed until the concentrations of PCB In surface and core samples are less than twice that of background samples. The zone plot method of sampling should be utilized (TSCA Inspection Manual, March 19B1).* Memorandum dated November 13, 1981 from Milt Clark, Region V Health Effects Specialist, to Karl Bremer, Toxic Substances Coordinator, addressing "Interim PCB Cleanup Procedures." 4-30 HONS 214818 * This broad range of spill cleanup requirements raises an Important Issue: As long as no one can come to grips with a scientifically-based cleanup standard, perhaps a practical standard based on practice and performance Is most appropriate (assuming that a standard will Indeed be Instituted). Such a standard might address sampling, a practical field approach and perhaps a minimum cleanup level standard. The EEI Utility Solid Waste Activities Group (USWAG) PCS Comnlttee has formed a Spill Cleanup Task Force, to address the specific degree of cleanup Iseue. The goal of the Task Force appears to be the EPA consideration of a procedure standard for cleanup rather than using a concentration-based performance standard, and to determine whether a procedural standard Is feasible and accurate enough. In an effort to permit practicality to enter the decision process. 4-31 HONS 214819 f Section 4 GENERIC SP1U SITUATIONS The two principal cost components of a spill cleanup are (I) the Investigation and (2) the cleanup itself. The Investigation phase Includes engineering support, drilling/sampling and analytical requirements. The cleanup phase Includes excava tion, cleaning, disposal, transportation, closure sampling and analysis. Both phases also Include occupational safety and health support. The Identification of these cost components Is Important; doubling the amount of material to be removed not only Increases the excavation and disposal cost, but also adds engineering, analytical and safety costs. In-house cost data were used to develop an empirical cost model. The model essentially simulates the cost estimating process used by an engineer or con tractor to bid a spill cleanup job. The "model11 Is actually a cost estimating algorithm using several Inputs that define the scale of the cleanup effort. Input variables Include spilled fluid volume, type of spill (lead and spray), surface type, rainfall, degree of cleanup and location near obstacles. These variables, when added to several dispersion algorithms, define the vertical and horizontal extent of the spill, which In turn defines the removal/cleaning requirements. Using standard unit costs for each cleanup activity, a total cost can be computed. Varying any one or more of the Input variables changes the dispersion, which changes the cost. Graphics and tables can thereby be developed showing cost sensitivity to changes in these variables. Because of the number and variability of factors Involved In cost estimating, a general model with appropriate assumptions and simplifications was developed. The model does not address regional cost variations, multimedia contamination and a variety of other factors to be discussed In more detail below. Despite this, the model provides a reasonable general cost estimate, most useful for demonstrating the effects of varying parameters on cleanup costs. In addition, contractors' unit prices are used; differences with utility costs should be negated by comparing the relative magnitude of spill cleanup costs. 4-32 HONS 214820 The spill cleanup model Is based on several parameters critical to describing the extent of the spill and contaminate migration, Including the following* Nature of the spill. Volume spilled. Degree of cleanup required. a Surface material type. Indoor/outdoor location. Rainfall. Presence or absence of structures near the spill. - The assumptions, limitations and general sensitivity levels for each variable and their relative Impacts are discussed below. Nature of the Spill Two types of events are analyzed by the model: leaks and spray spills. For model ing purposes, spray spill Is assumed to extend outward 20 feet over an angle of 20<>. It Is further assured that the entire surface area beneath the spray pattern (70 sq. ft.) Is contaminated with fluid. For outdoor spills, the distinction between leaks and spray spills Is Important only for those spill volumes that would not spread over qn area of 70 sq. ft. In a flow spill. For Indoor events, spray spills are assured to contaminate walls and nearby equipment. Increasing the total area to be cleaned. Volume of Fluid Spilled The surface area or soil volume contaminated Is primarily a function of the volume of fluid spilled, Models are available to predict both the downward movement and horizontal mobility of fluids under a given driving force, given fluid volume and physical characteristics and a surface composition. These models are generally designed to deal with more sophisticated conditions than the model, and have been simplified accordingly. Using these models In their simplified forms, the volume of material contaminated can be estimated. Surface Material Both the downward and outward mobility of spilled fluid are dictated In part by the permeability of the surface material. As a result, the area to be cleaned or the volume to be excavated Is a function of the contacted surface. Fluid will spread over a much wider area on Impervious surfaces: In permeable soils, the outward spread Is reduced, but the depth of soil contamination Is Increased. The model 4-33 ' HONS 214821 rr addresses an Impervious surface and four permeable surfaces (gravel and three soil classes). The three soil classes are as follows: e Class 1 - Sandy gravel, less than 5 percent clay and silt. Class 2 - Sandy loam, less than 25 percent clay, e Class 3 - Sandy clay, less than 35 percent clay. ~ These soil classes represent most naturally occurring soils. The present model lacks the sophistication to address multi-soil spill events. Further, the model assumes a smooth surface with no slope. A leak Is assumed to spread out radially from the source until It Is all absorbed Into the soil or further spread Is retarded due to surface tension effects. In most Instances, the volume of fluid lost Is small enough that flow modeling Is not as critical as worker exposure. Indoor/Outdoor Location Outdoor spills are assumed to spread radially until constrained by soil absorption or surface tension effects. Indoor spills tend to be confined to a specific area, although contamination may extend to walls and adjacent equipment. Rainfall The availability of rainwater Is Important to the mobility of a spilled PCB fluid, because the fluid head above a PCB spill provides a driving force. As a result, any liquid chemical spill In a wet environment will be more widely dispersed than a dry environment (excluding Irregular geologic formations). However, because PCBs are not highly mobile In most soils, the magnitude of the effect Is not large. There will be some Increase In contaminated soil volume due to rainfall, and this effect can be Included In the model. Degree of Clean Required In the simplest sense, within a given contaminated soil volume the concentration of PCBs will )>e highest near the source, and will approach ambient levels at or beyond the extent of the contamination. The limits of contamination representing PCB con centrations of 50 ppm, 7 ppm and ambient levels can be most simply represented by concentric spherical surfaces, each one of larger radius. Consequently, cleanup to a lower level of contamination requires removal of more material. HONS 21*82* 4-34 r r Spills on concrete present a simpler picture due to the nature of the cleanup options available. Concrete (and other "Impermeable" surfaces) Is slightly porous, and will absorb some of the spilled fluid. Solvent scrubbing of the concrete surface will remove surface contamination and some of the absorbed material, usually sufficiently well to achieve the 50-ppm standard. However, some PCB will remain within the concrete. To obtain a 7-ppm or ambient level Is assumed to require com plete removal of the contaminated concrete. Spill Investigation Spill Investigation costs are very sensitive to cleanup degree, A typical investi gation of a spill on soil Includes surface soil samples to determine the area extent of the spill and subsurface soil samples (obtained by drilling holes) to determine the depth of contamination. Only enough samples will be collected to delineate the extent of the spill to the cleanup level desired. With an increase In the degree of cleanup required, the area to Investigate becomes larger. Consequently, more bor ings, surface and subsurface samples, laboratory analyses and labor will be needed to establish the bounds of the spill. Nature of the Fluid Spilled An important variable which could not be Included in the model Is the nature of the fluid spilled, l.e., the PCB concentration or mineral oil versus askarel. To Include this factor would require modeling the fluld/soll concentration gradients within the soil, an exercise not practical within the scope of this model. Con sequently, the SCS model Is strictly applicable only to PCBs, askarels and mineral oils with at least 0.5 percent PCBs. 4-35 HONS 214823 / Section 5 COST ANALYSIS OF SPILL CLEANUP As noted earlier, the cost of each and every spill cleanup Is different due to the unique features discussed on the preceding pages. Cost estimation Is nevertheless the first logical step In a comparative cost analysis. Figure l shows the logic of the spill cleanup cost model. Including the role of each technical factor In the analysis. Table 2 presents the unit cost assumptions used in the model. There are several mechanical assumptions Inherent In the model that would tend to make the cost appear lower than experience would Indicate. First, the model uses probabilities to estimate the cost of a second or third sampling and ' cleaning situation {that is, the cost of a second Iteration {BOX probable) would be BOX of the cost of the first). Second, the cost of mobllliatlon/demoblIllation for a cleanup crew is not included. In most real situations, slow lab turnaround and agency approval will require removal and reactivation of the field crew, at a cost of hundreds or thousands of dollars each Iteration. Third, the analytical require ments are based on a rational sampling grid system, which Is usually the minimum number required. Fourth, any analysis for TCDD/TCDF In a fir# cleanup will dramatically Increase the analytical {and possibly the cleanup) cost, but cannot be addressed here. The real value of the model Is to vary certain parameters over a range of values, and to show the relative cleanup costs. The following set of tables and figures illustrates some of the major cost comparisons possible with the model. Tables 3 through 8 present the costs of several example spill cleanup scenarios using the model. The costs are expressed as a percentage of the SO ppm cleanup cost rather than as absolute dollar values. Figure 2 shows the Investigation and cleanup cost elements separately for an example situation, an askarel spill on soil Class 2 cleaned up to 7 ppm, over a range of o to 10 gallons spilled. The relative magnitude between the two cost components is typical of other soil classes, slab surfaces and other cleanup degrees as well. While the total cost {not shown) compare* favorably with actual utility field a-36 HONS 214824 experience, the cost of investigation is thought to be too high for small volume spills. The stepwise nature of the investigation costs is the result of subsurface soil sampling. The costs exhibit a sudden upward rise whenever the Increasing areal extent of a spill demands that another sampling hole be drilled. Soil sampling holes are drilled at discrete Intervals, regardless of the size of the spill, and a larger diameter spill necessitates more holes. Figures 3 and 4 show the relative Investigation and cleanup costs at the 1 ppm and 50 ppm cleanup levels for soil Classes 1 and 3, respectively. As most PCB equipment spills are In the 0 to 3 gallon range, drilling will not be required. Figure 5 compares the total costs for cleaning up spills In each of the soil classes to 50 ppm. Class 1 soils represent a sandy gravel with less than 5 percent clay and silt; Class 2 is a sandy topsoil with up to 25 percent clay; Class 3 Is a sandy clay. The costs can be higher In the more clayey soils because of the more extensive dis persion of the material on the surface. More soil must be removed because of the greater soil volume contaminated In the more clayey soils, so excavation and dis posal costs are higher. Figure 6 compares the costs of cleaning up spills on open soil to the costs of clean ing up the same volume of spilled fluid next to a structure. Spills next to struc tures will migrate along the foundations and beneath the building through looser soils. There is Increased difficulty In excavating beneath a structure, and this Is reflected In the model by Increased excavation costs. Figure 7 shows the costs to clean up uncontained PCS leaks on Impervious slabs. In general, slab cleanup to 50 ppm represents labor-intensive triple scrubbing with solvents. Cleanup to 7 ppm represents slab demolition. This figure suggests that beyond a certain volume spilled (approximately two gallons). It may be more costly to scrub the slab than to remove It entirely. However, this generalization does not apply to more common spills to concrete Inside a building, where concrete removal Is Infeasible and scrubbing Is the only option. Figure 12 compares the costs of cleaning up leaks versus spray spills on slabs. The differences are most dramatic at the low volumes spilled, where a spray could result In greater surface area contamination than a spill. Above one to two gallons, a leak will eventually spread to contaminate an equivalent surface area as the spray spill. Thus, the lines become parallel. The difference between the two lines represents the costs of cleaning up walls, poles, nearby equipment, etc., all Items 4-37 HONS 214825 more likely to be contaminated by a spraying piece of egulpment. Obviously, spray ing spills of this type are the most costly per gallon, as they contaminate more surface and often result In replacement of contaminated Items. 4-38 HONS 214826 Table 1 SELECTED EPA REGIONAL PC8 SPILL CLEANUP REQUIREMENTS EPA Region Headquarters 1 5 6 g Contact (Kane, Phone) 0. Hannemann (202) 363-7649 T. Palermo (617) 223-5604 P. Porter (312) 353-2192 0. Mount (214) 767-2734 K. Glassel Soil 50 ppm+ 1-5 ppm 50 ppm 50 ppm Cement/Asohalt concrete: solvent asphalt: remove concrete: solvent (3) asphalt: remove concrete: wash asphalt: remove concrete: remove asphalt: remove Other* cbc* cbc cbc cbc *cbc - Each Instance evaluated on a case-by-case basis. + - Initial concentration. 4-39 MONS 214027 rf Table 2 UNIT COST ASSUMPTIONS USED IN PCB SPILL CLEANUP MODEL* Engineering Labor Labor Sample Analysis Or111 Rig Mobilization Drilling Costs S45/hr S35/hr S50/sample SI,300 S4S/ft Surface Sampling Productivity Frequency 6 samples/man-hour 1 sample/6 ft Subsurface Sampling Drill Depth Productivity Frequency 15 ft 2 ft/hr 2 sanples/ft planning & Engineering 25X of Total Labor Bulk Disposal Drum Disposal Drum Cost Equip. Mobil,/Demobll. S35/hr Sl45/yd3 S39/drum S25/drum S1.6D0 Equl patent Wheeled Backhoe/Loader Dump Truck S4.66/yd3 S7.BO/yd3 Slab Cleaning Triple Solvent Scrub Followed by Solvent Absorption Materials Cost Productivity S2.7D/ftz 25 ftz/man-hour 6-Inch slabs 1 Drum Disposable Equlpment/6 man-days Slab Demolition Equipment Productivity , S135/day 25 ftz/man-hour General/Admln1s t ra11ve Costs 5X of Total * These figures are typical of contractor costs and leased equipment. Sources: Means Building Construction Cost Data, 1962; Richardson's Estimator; 1982 Dodge Guide to Public Works and Heavy Construction Costs. 4-eo HONS 214828 r Table 3 SPILL CLEANUP - 1,5-Gallon Askarel Spill; Class 2 Soil; No Rainfall DESCRIPTION 1,5-gallon spill from a pole-mounted Askarel capacitor to a moderately permeable soil surface. Excavation by hand with drummed material disposal. Investl gatlon Sampling and Analysis Equipment Sample Disposal Plannlng/Reportlng Total Cleanup Excavation and Disposal Equipment Genera 1 /Adnlnl strati ve Total Relative of Costs Degree of Cleanup SO ppm 1.00 1.00 1.00 1.00 Lees. 1.49 1.00 1.35 1.46 1 PP"i 2.22 1.00 1.81 2.13 1.00 1.45 2.13 1.00 1.55 2.00 1.00 1.00 1.00 1.00 1.43 1.77 1,00 1-43 2.09 TOTAL 1.00 1.45 2.09 4-41 MOMS 214829 rr Table 4 SPILL CLEANUP - 1.5-Gallon Askarel Spill; Slab Surface; No Rainfall; Flow Spill DESCRIPTION 1.5-gallon spill from a pole-mounted Askarel capacitor to a concrete parking lot. Triple solvent scrub to 50 ppm versus slab demolition to 7 ppm (or less). Investigation Sampling and Analysis Equlpment Sample Disposal Planning/Reporting Total Cleanup Cleanup and Disposal Equipment Genera 1 /Admlnl strati ve Total TOTAL Costs (S) Degree of Cleanup 50 ppm 7 ppm 1.00 1.00 1.00 1.00 1.35 1.00 1.00 1.15 1.00 L& 1.00 1.00 1.00 1.00 1.00 1.84 0.45 1.31 1.31 1.26 4-42 HONS 214830 r r Table 5 SPILL CLEANUP ~ 1.5-Gallon Askarel Spill; Spray Spill; Slab Surface DESCRIPTION 1.5-gallon spill from a spraying pole-mounted Askarel capacitor to a concrete parking lot. Triple solvent scrub to 50 ppm versus slab demolition to 7 ppm (or less). Investigation Sampling and Analysis Equipment Sample Disposal Planning/Reporting Total Cleanup Cleanup and Disposal Equipment General/Administrative Total Costs (S) Degree of Cleanup 50 ppm 7 ppm 1.00 1.37 1.00 1.00 1.00 1.00 1.00 1.18 1.00 1.18 1.00 1.00 1.00 1.00 3.50 1.17 1.51 2.27 TOTAL 1.00 1.93 4-43 HONS 214831 r* Table 6 SPILL CLEANUP - 10-Gallon Askarel Spill; Class 2 Soil; Cleanup to 7 ppm DESCRIPTION 10-gallon spill from Askarel transformer to a moderately permeable soil. Cleanup to 7 ppm. Investigation Sampling and Analysis Equipment Sample Oisposal Plannlng/Repor ting Total Cleanup Excavation andDisposal Equipment General/Admlnl strati ve Tota! Costs (S) Degree of Cleanup______ Spill on Open Soil Spill Next to a Building 1.00 1.00 1.00 1.00 K00 1,35 1.00 1.32 1.32 TOTAL 1.00 1.08 4-44 MON$ 214832 r r Table 7 SPILL CLEANUP _ 1,5-Gallon Askarel Spill; Spray Spill; Class 3 Soil; No Rainfall DESCRIPTION 0.5-gallon spill from a spraying pole-mounted Askarel capacitor to a public park baseball field. Excavation by band with druimed disposal. Investigation Sampling and Analysis Equipment Sample Disposal Planning/Reporting Total Cleanup Excavation and Disposal Equipment General/Admln1s trat1ve Total Relative of Costs Degree of Cleanup 50 ppm 1.00 1.00 1.00 1.00 7 ppm 1.34 1.00 1.0S 1.46 Lean 1.80 1.00 1.08 2.13 1.00 1.D8 1.19 1.00 1.00 1.00 1.00 1.40 1.00 1.2--7 1.27 1.81 1.00 1.54 1.55 TOTAL 1.00 1.15 1.32 4-45 HONS 214833 Teble 6 SPIU CLEANUP 50-6*1 lor Spill from *n Asfcerel Tnmfomer Stored Outside on Connected Set) DESCRIPTION 50-9*1 ion spill fro* *n Asktrel tr*ntfor*er stored outside on coqpected soil Inv*stlBtton Sonpllng end An*lysis Equlpaent Senple Dispose! Pl*nn1ng/ftoport1n9 Tot*1 ___________________________ Costs ($)_______________________ jg-EES._______ After 10 Inches of Rein ________ Z.EES________ No Rtlnfell After 10 Inches of Rein - I PP" Mo Relnffll After 10 Inches of R*1n 1.00 1.49 7 70 1 00 1.00 l.oo 1.00 1.00 1.00 1.00 1.44 2 07 1.00 1.47 7.07 Eicevetlon *nd Olsposel Equipment Generel/AdMnlstretlve Tot*) TOTAL 1.00 1.00 1,00 1.00 1.00 1.10 1,00 1.10 1.00 1.04 i.as 1.95 1.50 1.50 1.63 1,93 1.63 1.93 1.56 1.(1 7.(9 7.00 7.65 7.66 2.21 2 79 7 00 7.75 7 75 7.31 HONS 214834 > Figure 1 FLOW CHART FOR SPILL CLEANUP COST MOOEL 4-47 HONS 214835 Figure 2 COST OF PCB SPILL CLEANUP TO 7 PPM, SOIL CLASS'2 SOIL CLASS 2 / , / / // _-" ' / ~! 1/ / /.......................... 7/ , r~ / ------'/ t: t t t c >: >: _ if TOTAL ______________________ investigation CLEANUP 0 I 2 3 * S 6 7 8 0 10 FLUID SPILLEO (GALLONS) 4*46 HONS 214836 Figure 3 COST OF PCS SPILL CLEANUP, SOIL CLASS 1 4-49 - HONS 214837 COST OF PCB SPILL CLEANUP, SOIL CLASS 3 HONS 214838 ^-s Figure 5 COST OF PCB SPILL CLEANUP TO SO PPM CLASS 3 CLASS Z cuss i _/ B*t --ir FLUID SPILLED <GALLOMS> to 4-51 HONS 214839 rr Figure 6 COMPARISON OF CLEANUP COSTS FOR SPILLS ON OPEN SOIL AND NEXT TO STRUCTURES B (7 pp) HEXT TO STRICTURE OPEN SOIL FLUJU SPILLED 'Hi*!.! Ol'CO *E-OI 4-52 HONS 2148*0 Figure 7 PCB SPILL CLEANUP COSTS, OUTDOOR SLABS HONS 2148V1 rr Section 6 CONCLUSIONS AND RECOMMEDNATIONS The preceding discussion of PCB spill cleanup costs has attempted to quantify the relative cost of cleanup to different levels of residual contamination. Recognizing the inherent weaknesses in generic modeling of such a site-specific event, some conclusions can still be drawn regarding relative cost. First, it has been shown here that a cleanup to 7 ppm Is at least twice as expensive as cleanup to 50 ppm. Cleanup to background 1$ still more expensive, with the actual Increase being a function of the background level definition and sampling protocol. Second, flow spills next to structures are up to 50 percent more expensive to remove than open flow spills. Spray spills next to structures are even more expensive and labor Intensive, varying with the material contacted. Third, contained spills are less costly to cleanup than open spills to slabs or soil. The model showed that spills to Impervious slabs of more than two gallons are more expensive to dispose of the slab than to clean it. More important than the specific costs is the probabilistic nature of both the model and the actual spill cleanup exercise. Some probabilities for second and third cleanup Iterations were conservatively assumed. in practice, a threeinteratlon requirement would show an actual cost increase of 200 to 300 percent (without the probabilities applied). A four- or five-iteration exercise with demobilization would be cost-prohibitive. This is a real, practical problem, not Just modeling philosophy. Some recooiaendatlons are in order here. A standard of some type Is needed. The choice appears to be between a procedural standard ("how to clean up"), a perform ance standard (clean up to a specified level), or a combination of the two approaches. A procedural standard must have a definable (yet non-concentration based) end point In order to be effective yet practical. A performance-based standard must be based on a practical concentration limit. 1-54 HONS 214842 f r in order to set a standard of either type, more research Is required, research into what is reasonably achievable by a trained, mature cleanup crew. Rather than per forming laboratory research, have some crews dean up a PCB-stalned area plus 6 In,, 12 In. and 24 In. and check the residual concentration. Or determine what level of cross contamination Is comnon using standard drilling and sampling equipment (hole wall collapse, for example). Then combine the two data sets to establish a practical standard to apply in all situations. industry has been subjected to double jeopardy In many cases, with application of both procedural and performance standards on the same spill. Those encountering PCB spillage from their equipment need to be prepared to write blank checks to their lab, cleanup contractor and attorney under the current regulatory situation. And using a standard numerical or "background" is not an Improvement over the current state of affairs. 4-55 MONS 214843 /* Section 7 REFERENCES Bennett, G. S., F. S. Feates and I. Wilder. The Hazardous Material Spill Handbook. McGraw-Hill, New York. 1982. Gilmer, H. and F. J. Freestone. "Cleanup of an Oil and Mixed Chemical Spill at Dlttmer, Missouri, Aprll-May, 1977." In Proceedings of the 1978 National Conference on Control of Hazardous Material Spills. U.S. EPA. pp 131-134. Hulbregste, K. R., R. C. Scholz, K. H. Kastman and J. Brugger. "Development of a Mobile System for Extracting Spilled Hazardous Materials from Soil." In Proceedings of the I960 National Conference on Hazardous Material Spills. U.S. EPA. 1980, p. 159. janelle, R. N. "Case Study: PCB Cleanup, Kodiak Island, Alaska." In Proceedings of the 1982 National Conference on Control of Hazardous Material Spills. U.S. EPA. 1982. Lafornara, J. P., F. J. Freestone and M. Pollto. "Spill Cleanup at a Defunct Industrial Waste Disposal Site." In Proceedings of the 1978 National Conference on Control of Hazardous Material Spills, U.S. EPA. 1978. pp. 152-155. Mackay, D. "Environmental Pathways of Polychlorinated Biphenyls." Vol. IV of Comments and Studies on the Use of Polychlorinated Biphenyls In Response to an Order of the United States Court of Appeals for the District of Columbia Circuit. Edison Electric Institute, Utility Solid Waste Activities Group. February 1982. Nadeau, R. J., H. L. Allen and G. R. Prince. "Hazardous Assessment and Criteria Development Methodology Applied to PCB Incidents." In Proceedings of the 1982 National Conference on Control of Hazardous Material Spills. U.S. EPA, 1962. Orser, p., w. Hansen and 0. Golden, "Spill Prevention Control and Countermeasure Planning for Polychlorinated Biphenyls (PCBs)." In Proceedings of the 1980 National Conference on Control of Hazardous Material Spills. U.S. EPA. 1980. pp. 327-332. Pastrovleh, T. L. D., Y. Baradat, R, Barthel, A, Chlarelll and D- R. Fussel. "Protection of Groundwater from Oil Pollution." Report nr.3/79. CONCAWE, Water Pollution Special Task Force No. 11. Oen Haag. April 1979. Resource Planning Corporation. "Report of the Study of PCBs In Equipment Owned by the Electric Utility industry," Vol. Ill of Conments and Studies on the Use of Polychlorinated Biphenyls In Response to an Order of the United States Court of Appeals for the District of Columbia Circuit. Edison Electric Institute, Utility Solid Waste Activities Group. February 1982, Robinson, J. S, "Hazardous Chemical Spill Cleanup." Noyes Data Corporation, Park Ridge, New Jersey. 1979. MONS 214644 4-56 SCS Engineers. "Engineering Services for West Sycamore Canyon PCB Cleanup and Abatement Plan at the Naval Air Station, Miramar, California." Western Olvlsion, Naval Facilities Engineering Command, San Bruno, California. July 1982. Stroud, F. 8., R. t. wllkerson and A. Smith. "Treatment and Stabilization of PCB Contamination Water and Waste Oil: A Case Study." in Proceedings of the 1978 National Conference on Control of Hazardous Material Spills. U.S. EPA. 1978. pp. 135-144. _ Sunden, R, L. "PCB Spill." Coast Guard Engineers Digest. April 1981. pp. 2-7. U.S. Environmental Protection Agency. "Follow-Up Study of the Distribution and Fate of Polychlorinated Biphenyls and Benzenes In Soil and Ground Water SMples After an Accidental Spill of Transformer Fluid." Division of Oil and Special Materials Control, Office of Water Program Operations, Washington, D.C. January 1976. MONS 214845 r INNOVATIVE DESIGN CONCEPTS FOR PREVENTION OP AIRBOWE PCB CONTAMINATION Brian W. West Warran G. Hanaan Waat A Hanaan Engineers, Lon? Baach, CA Inc. Waaton A. Fennar Yuen-Fenner, Inc. San Francisco, CA ABSTRACT Within tha past aavaral years, failure of alactrical equipment filled with polychlorinated biphenyl (PCB)-bearing dialactric fluids has resulted in extensive contamination of buildings in Binghamton. NY (1981); San Francisco, CA (1983) and other locations. Tha magnitude of these incidents was a result of "air-spills" of atomised FCBs and combustion byproducts of the dialactric fluids (furans and dioxins). Resultant cleanup coats and potential liabilities associated with these types of incidents and materials can easily run into the millions of dollars. Prior to the Binghamton incident, tha potential for airborne contamination was little appreciated, with the result that the scope of other past incidents can only be estimated. west a Hansen Engineers has recently completed surveys and design projects for seven California and Washington State agencies identifying potential sitae where airborne contamination might occur and devising cost-effective methods for isolating PCB equipment. Structures surveyed included office buildings, schools, hospitals, highway bridges and tunnels. Both transformers and capacitors were involved. This paper presents some of the design techniques used to prevent widespread structure contamination. These procedures include modification of ventilation systems, partitioning of mechanical and electrical areas in utility rooms, closed heat exchangers for air-toair equipment cooling and automatic fan shutdown devices. Considering the amount of utility and customer-owned PCB equipment in occupied structures throughout the country, it is probable that future airborne PCB contamination will occur. Identification and isolation of high-risk equipment installations can significantly reduce potential liabilities, health effects and property impairment caused by such episodes. MQNS 214646 4-59 rr INTRODUCTION There is presently a great need to develop techniques for the preven tion and control of airborne PCB contamination from electrical equip ment mishaps. Although these types of incidents are infrequent, their resulting costs in terms of cleanup, liability and property lose can be enormous, examples include the substation fire in the State of New York Building in Binghamton New York (1) and the sidewalk vault eruption adjacent to the One-Market Pla2a Building in San francisco (2). The types of airborne contaminants resulting from pressurized leaks or fires include aerosoled dielectrics such as dibenzodioxins and dibenzofurans. until now, spill prevention and control procedures (and regulations) for PCBs have emphasized fluid spills; i.e. dielectrics escaping from faulty valves, welds, cooling radiator fins, busnings etc, in a liquid state. Preventive and control measures for these types of spills have been developed and described previously (3,4). The Environmental Pro tection Agency's August 25the, 19B2 Regulations on PCBs |S| describe the required response to leaks "resulting in any quantity of PCBs running off or about to run off the external surface of the transformer..." (Section 7G1.30 (ill)). Secondary containment (curbs, dikes or sumps) as a fluid spill control method is also suggested in Section 761.30 (v)(A). More relevant to the issue of "air-spills" is the definition of "spills, leaks, and other uncontrol led discharges of PCBs" [underlining added] as constituting a disposal of PCBs (Section 76l.40 (d)(1)). However, specific methods for controlling airborne emissions from PCB equipment installations have not been well documented. SCOPE AND OBJECTIVE West t Hansen Engineers, Inc. and Yuen-Fenner, Inc. have been involved in a project sponsored by the California Office of the State Architect to identify and prepare risk reduction recommendations for State-owned PCB filled equipment installations. This paper presents results of those efforts as they relate to potential airborne PCB contamination. Topics include survey of installations to determine contamination potential, selection of a risk reduction approach, design of isolation improvements and areas requiring further research. The information contained in this paper is intended to assist member utilities contem plating or conducting similar risk reduction programs. PROGRAM ELEMENTS Risk Survey The State of California is similar to many member utilities in that they own hundreds of PCB-filled transformers and capacitors spread out over a large geographic area and housed in a wide variety of struc tures. Unlike the State, however, utility-owned equipment is most often on non-utility owned property adding another element of risk to those facilities. 4-60 HONS 214847 Only a small percentage of PCB installations are in what we terra "sensitive" locations with respect to potential airborne contamination. A sensitive location is one where substantial structure contamination, human contact or facility disruption would occur. A comprehensive survey, conducted in part by west b Hansen Engineers and Yuen-Fenner, Inc. was undertaken by the State to identify sensitive PCB equipment sites. for each site so identified information was collected to provide a Dasis for selecting a usk reduction approach. Particular importance was placed on identification of pathways leading from PCB equipment locations to other portions of the structure. Typical pathways include ducting, conduits, doors, windows, drains, hallways, ventilation shafts and utility corridors. Although PCB con tamination can migrate either actively, by forced ventilation or passively, via natural convection, neither appears to pose a more or less serious risk. Risk Reduction Alternatives Alternatives for reducing airborne PCB contamination risk fall into three categories, removal, relocation and isolation. Removal of PCBfilled equipment eliminates all spill risk entirely, however this option is expensive, and results in a PCB waste disposal problem. Equipment relocation is generally impractical for transformer equip ment, but works well in the case of capacitors. When technically feasible, relocation can provide a significant amount of risk reduction at less cost than outright replacement. Equipment isolation offers the advantages of low cost, minimal service interruption and no generation of PCB waste. A well-engineered and constructed isolation design will provide almost total containment of aerosols and gases. However, the risk reduction afforded by equipment isolation is lass than that provided by replacement or relocation. Equipment Replacement Some PCB equipment installations are so sensitive that the only accep. table level of contamination risk is zero. In such a situation cornplate removal is the only alternative. In some cases removal of the dielectric fluid only may be feasible, but generally outright replace ment is the only option. In other case*, equipment relocation or isolation may be as costly as replacement. Also, if the existing installation is aged, undersized, or has some other short coming replacement may be preferable to retention. In some cases relocation or isolation may be technically infeasible. Where feasible, or marginally feasible, isolation may not provide an acceptable level of risk reduction. In these cases, replacement may become tne preferred alternative. MONS 214848 4-61 Equipment Relocation The relative positioning of PCS equipment and ventilation structures has historically received little regard. in some cases relocation of either ventilation system components or PCB equipment can greatly reduce the contamination potential. This is particularly the case with capacitors since they can be moved so much more easily than transformer equipment. Where an acceptable alternative location is available, relocation is technically feasible, and equipment is worth retaining, relocation may be an alternative. Equipment isolation Due to its technical and economic advantages equipment isolation is the preferred risk reduction alternative. Although not providing 1001 complete protection from contamination, a well designed isolation job will restrict potential contamination to the electrical room vicinity, greatly facilitating structure clean up and rehabilitation. An isolation project consists of at least one and possibly as many as three elements: 1) blocking contamination pathways 2) providing alter native cooling 3) providing safe installation discharge. Element one is a factor in every isolation job, elements two and three are factors depending upon parameters specific to each equipment location. Blocking Contamination Pathways Preparation of an isolation design necessitates a thorough survey of all possible contamination pathways, both active and passive. Once identified, tne most efficient combination of improvements can be specified to seal the contamination routes. Isolation designs may be broadly categorized as equipment isolations or room isolations. Equipment isolation includes those cases where a containment shell is constructed around the equipment. Room isolation refers to the sealing of the existing equipment enclosure. Providing Alternative Cooling Provision is usually made for some form of ventilation cooling at high voltage equipment installations. An isolation project will often re sult in the removal of that ventilation. The resultant higher opera ting temperatures reducs efficiency, shorten service life and increase probability of equipment failure. A number of alternatives are available to provide for alternative coolina e.g. heat exchangers, air conditioners, or re-directed venting with our without forced circula tion, The decision to install equipment cooling is primarily based on economic considerations, i.e, capital cost of installation vs operating cost at higher temparatura. Safe Contaminant Discharge A transformer in failure mode can evolve substantial quantities of 4-62 HONS 214649 gases, potentially creating elevated pressures if contained within an enclosed space. Depending upon enclosure size, location and equipment type, an adequate isolation design may require provision of an emergency discharge to the outside of the structure. It is debatable whether there exists a safe discharge for airborne PCBs from a hot ruptured transformer. If aerosoled, a ateeh type mist arrestor can substantially reduce PCB concentration. Mist arrestors are inexpensive and require no maintenance. Gases will pass through such a device unaffected. Provision of a PCB gas scrubbing system is impractical except in some highly unusual situations. DESIGN TECHNIQUES An ideal airborne PCB spill containment design provides for effective spill control, equipment cooling, unrestricted maintenance and inspec tion access and ease of removal upon eventual equipment changeout. on the other hand, an ill-conceived design can significantly increase the probability of equipment failure and subsequent potential for PCB contamination. Effective Spill Control An effective spill control design provides for blocking every poten tial pathway leading to PCB contamination, common pathways were listed previously. Techniques for pathway blocking can be broadly classified as caulking, partitioning, sealing, weather stripping, enclosing, covering, and re-directing, in most cases the most important aspect of each design will be the materials of construction. Table 1 cross references design techniques and construction materials applicable to each pathway. The pathwaye listed in Table 1 were originally intended for ventila tion, access, drainage or lighting. Once blocked, their usefulness for these purposes may very well be reduced or eliminated and alternative service may be required. Alternative lighting can easily be provided, however providing backup service for the other 3 services is usually more complicated. Equipment Cooling and Ventilation Oil-filled power transformers installed indoor# are required to be placed in vaults with firawall isolation from tha other indoor rooms. PCB transformers are exempted from codes requiring vaults bacause of thair inherent fire resistant character. Ventilation systams from the oil-filled transformers were designed with fire isolation in mind. PCB ventilation systems were only designed to keep the ambient temperature at or below 30 degrees centigrade. The 30 degree ambient (by the standard transformer industry design) was used as the design basis for allowing a transformer to be loaded to name plate rating continuously without exceeding a 95 degree centigrade insulation hot-spot temperature. 4-63 HONS 214850 Pathway Ducting Conduits Doors Windows Drains Hallways Utility corridors ventilation shafts Table 1 AIRBORNE SPILL ISOLATION DESIGN TECHNIQUES FOR COMMON PATHWAYS Desiqn Techniques Sealing Re-directing Partitioning Caulking Weather stripping Sealing Partitioning weather stripping Sealing Partitioning Sealing Caulking Covering Partitioning Sealing Partitioning Sealing Re-directing Partitioning Material of Construction Sheet Metal Gypsum board (coated) Expanding silicone foam Butyl polymer gasket tape Sheet metal Concrete block Butyl polymer gasket tape Sheet metal Concrete block Gypsum board Metal Concrete Expanding silicone foam Epoxy Gypsum board Concrete block (compatible with existing materials) Sheet metal Concrete <-64 HONS 214851 r r The transformer industry then designed their cooling systems around this continuous hot-spot temperature of 95 degrees and further agreed that transient excursions above this hot-spot temperature were acceptable with the understanding that the loss of life of the insulation system would double for each 8 degrees in rise in hot-spot temperature- Therefore, a transformer operating at 103 degrees centigrade (hot-spot) would last only half as long as-one operating at only 95 degrees (based on the industry class A insulation system). Sophisticated operators realize that any transformer that is not operated continuously at full load and at a constant 30 degree centigrade ambient temperature could sustain transient insulation operating temperature incursions above 95 degrees) if offset with loading conditions below 95 degrees such that the normal operating life of the transformer is unchanged. Restricting the ventilation of the PCB transformers will limit the spread of airborne PCB compounds, but will also cause a rise in the ambient air temperature which in turn will cause a rise m the insulation hot-spot temperature. A conservative, safe operating ambient air temperature limitation is one that will not cause a transformer's normal life expectancy to be reduced. As a rule of thumb in the transformer industry, the insulation hot spot temperature on the average is 10 degrees centigrade higher than the transformer top oil temperature. The hot spots can not be readily measured, but the top oil temperature can be measured and often is brought out to a case-mounted temperature gauge. a first step to closing in a PCB transformer ventilation system is to first close the system with temporary seals and then monitor the top oil temperature. If the top oil temperature on a 55-degree rise tranformer averages below 85 degrees centigrade over a typical operating day then the seals can be made permanent and the transformer will have a normal operating life once it is closed in. If the average top oil temperature exceeds 85 degrees then heat exchangers or other cooling devices that do not have direct air-to-air contact between the traneformer room and exterior should be provided. Consideration must be given to other equipment in the transformer area that could be adversely affected oy the increase in ambient temperature created by shutting down the room ventilation system. Higher ambients will generally cause an increase in aging rates and an increase in electrical resistance with a corresponding increase in energy losses. Consideration should also be given to potential thermal loads applied to load-bearing structural members in high-rise buildings. The reduction of the risk of releasing airborne PCBs will probably override the disadvantages associated with the closing in of the ventilation system on a temporary basis. The overall advantage of auxiliary cooling should prevail over the long term. Maintenance and Inspection Access 4-65 ' HONS 214852 Recommended transformer maintenance operations include fluid sampling, fluid filtering, and cleaning. Access is also required for switcn operation and tap changing. Federal regulations require periodic inspection of PCB-filled equipment to check for fluid leaks. Additionally, it is good practice to monitor equipment temperature, pressure, and fluid level as indications of system performance. A spill isolation design must take into account personnel and equipment access for all the above listed activities. Further, electrical codes require minimum set backs around high voltage equip ment as do occupational and safety regulations. Blocking doors, windows, hallways, or utility corridors may result in a future access problem. In all cases provision must be made for personnel entry. Consultation with the manufacturer, maintenance supervisor, or operating electrician should define equipment access requirements. Equipment Removal An airborne spill containment enclosure must be both permanent and durable, yet removable to facilitate eventual pCB equipment removal. Since many member utilities have instituted accelerated PCB removal problems, equipment may be replaced long before the end of its normal service life. An evaluation of the anticipated service requirements of each transformer will have a bearing on the containment design. Drainage Maintaining dry conditions in the vicinity of high voltage equipment prolongs service life as well as promoting worker safety, when existing drainage openings are sealed for spill control reasons provision should be made for keeping equipment dry, although if the transformer is scheduled for accelerated replacement, or if water problems are insignificant no action may be taken. Possible techniques include curbing, sump pumps, alternate drama, or sealing to exclude water from the equipment enclosure. CONSTRUCTION METHODS AND MATERIALS Remodeling a high voltage installation for airborne spill prevention often involves working in cramped interior rooms with poor access. As a consequence, construction methods emphasizing small, easily manipulated, modular building units are preferable. Table 2 lists the advantages and disadvantages of four common construction methods as they apply to interior vaults. Besides indoor locations, PCB equipment installations can be outdoors, either pad or pole mounted. There are virtually no constraints on construction methods for outdoor installations. Materials utilized in a spill isolation design must be resistant to 4-66 HONS 214853 Table 2 ISOLATION CONSTRUCTION COMPARISONS Construction Typ Masonry units i.e. concrete blocks, bricks, stc. Cast In place concrete Pre-cut Assemblies Carpentry Advantage - Modular - easy to manipulate in cramped spaces. - Inexpensive - Durable - Relatively easy to remove. - Durable - Continuous unit - Minimum construction activity in vault. - Easy removal - Light weight - Inexpensive - Easy removal - Light weight Disadvantages - May not form a continuous unit. - Heavy - Expensive - Form work and pouring difficult to manipulate in tight quarters Difficult to remove. Heavy - Skill required for good fit. - Combustible materials of construction. 4-67 HONS 214854 rr noth chlorinated solvents and elevated temperatures. porous or adsoroent materials are to oe avoided as are tlammaole products. The interior races of all isolation structures, particularly tnose intended to impound liquids, are usually coated with a suitable protective sealant. Caulking, weather stripping and sealing compounds compaticle with tnese specifications are commercially^avallaole and nave oeen successfully used in State of California isolation projects. CASE STUDY Vault or transrormer room ventilation systems - either passive or active - are a primary cause of contaminant dispersion. In an actual transrormer room inspected by West & Hansen Engineers andYuen-Fenner, a forceo ventilation system was used to cool two 643-gallon Inerteennlleo Nestinghouse transformers (Figure 1). The wind tunnel was used to discharge air from the multi-story ouilding's interior to the outside. However, fan shutdown concurrent with a substation failure would permit the tunnel and adjacent vertical ducts to act as a passive contaminant route to the building's interior. 'The control program for this particular installation included recommendations that tne vent ports (influent and effluent) be sealed. At the same time, the ambient and equipment operating temperatures are Ding monitored to determine the impact of reduced ventilation. Should the ambient temperature exceed the design temperature of 40 degrees centigrade alternatives for modified ventilation will be considered. REFERENCES 1. G. A. Eadon. Particulate Contamination in the Binghamton State Office Building. Electric Power Research Institute, Proceedings: 1981 PCB Seminar, September 1982. 2. Pacific Gas and Electric Company. Underground Transformer Replacement Program. Exhibit submitted to'Bertram D. Patrick, Administrative Law Judge, Public utilities Commission, State of California. June 28, 1983. J. P. Orser, w. G. Hanaen and D. Golden. Spill Prevention and countermeasure Planning for Polychlorinated Biphenyls (PCBt). Proceedings of the 1980 Hazardous Materials Spill Conference, May 1980, pp. 327-332. 4. w. G. Hansen, B. w. West and P. Hypnarowski. PCB Spill Riaks: A Compranensive Investigation in California. Proceedings ol* the 1982 Hazardous Materials Spill Confer ence, April 1982, pp. 425-432.5 5. U. s. Environmental Protection Agency. Polychlorinated Biphenyls (PCBS) Manufacturing, Processing. Distribution in Commerce and Use Prohibitions; Use in Electrical Equip ment. 40 CFR~T?1 - Federal Register VoT7 47, No. 165, Wednesday, August 25, 1982. HONS 214*55 4-68 VO *0*S 0b V- r r BIOTECHNOLOGY POTENTIAL AS A PCS DISPOSAL OPTION An Overview Thome* 0. Peyton American Technology Management Consul tent* Eric G. Terhorit Lei and 0. Attaway 4 Associate* SUMMARY Electric utilities must comply with EPA requirement under the Toxic Substances Control Act for the eventual disposal of about one billion (109) gallons of utility equipment fluids containing polychlorinated biphenyl (PCB). EPA require* ments also address leaks and spills and the need for clean-up and disposal of contaminated environments under the Resource Conservation and Recovery Act and the Comprehensive Environmental Response, Compensation and Liabilities Act. PCB in electrical equipment is composed of over loo chlorinated biphenyl compounds all of which have been reported to be degraded by mixed cultures of natural bacteria. However, under natural conditions, the more highly chlorinated PCB compounds (those which contain a high weight percent chlorine) biodegrade far slower than compounds with less chlorine. Since the highly chlorinated congeners meke up a large portion of the utility PCB Inventory, rates of biodegradation awst be enhanced to make blotreatment techniques competitive with other disposal means. Methods such as emulsification and genetic engineering could accomplish this enhancement. In this paper we review electric utility's use of PCB and disposal requirements, existing evidence of PCB biodegradation, means of Improving the performance of biological systams M.e., genetic engineering), and blotreatment technologies. He conclude that blotreatment of both spills and the planned phase-out of existing equipment Inventories of PCB fluids are feasible given adequate research and development. He recommend that some of the existing data be validated, and a set of research objectives be Initiated in the areas of: 1) microbiology and genetic engineering; 2) system design and testing; and 3) cost-analysis. This biotechnology paper was prepared for the Electric Power Research Institute by Aaterlcan Technology Management Consultants, Lafayette, Indiana. Leland D, Attawey and Associates, San Rafael, California, provided Inputs on PCB uses and characteristics. 4-71 HONS 214857 Section I INTRODUCTION Polychlorinated biphenyl (PC8) has been In use In utility equipment since 1930. PC# Is a major component of capacitor and transformer dielectric fluids, and is a known contaminant of mineral oil used In transformers, voltage regulators, circuit breakers, reclosers and switches. Several million such units, containing about one billion (10> gallons of fluid contaminated with PCB, are present In the utility power distribution network. Due to the environmental persistence and possible adverse health effects of PCB, the complete, controlled destruction of PCB waste materials Is required by law. Thus, soils, and other materials contaminated by PCB spills from utility equipment, as well as fluids recovered from obsolete or recy cled equipment, must be properly eliminated. The nature of utility uses for PCB creates a diverse mixture of waste products particularly challenging to all means of disposal techniques. Inferences drawn on laboratory studies with pure PCB Isomers, though meaningful for destruction of the regulated PCB constituent, must also take Into consideration the total waste matrix before an Interpretation of Its practical application can be drawn. Of the 209 possible PCB Isomers, nearly 100 exist In comnerclal mixtures at concentrations ranging from 50 ppm to 100%. PCB solvents such as mineral oils predominate In transformers, and In askarel transformers, polychlorinated benzenes. (Capacitors are 100% PCB.) Trace contaminants such as polychlorinated dlbenzofurans can also be present. These utility PCB mixtures combined with wash solvents and contamlned environmental media pose complex problems to all waste management technologies. Nevertheless, all of the principal watte constituents tre hydrocarbon derivatives and are potential sources of energy and carbon for biological growth through enzyme catalyzed degradetlve reactions. In fact, all of the organic waste constituents, Including PCB, have been observed to undergo biodegradation by microorganisms. In this paper we review the results of an investigation on the feasibility of using biotechnology for utility PCB disposal (1). 4-72 HONS 214858 Section 2 UTILITY USE AND DISPOSAL REQUIREMENTS FOR PCB - Polychlorinated biphenyl (PCB) was first synthesized In 1881 (2) and during the 1920's It was manufactured In bulk. In 1930 General Electric became a major user of PCB for capacitor fluids due to Its thermal and chemical stability and electrlcal properties. Today it Is estimated that approximately 1.1 billion gallons of PCB-laden fluids exist In the electric utility Industries equipment inventory containing 62,000 tons of PCB (3). Due to PCB's potential toxicity and environ mental persistence, the United States Congress mandated the Environmental Pro tection Agency to control the use of PCB and to prescribe methods for PCB's phaseout and disposal [Section 6 (e), Toxic Substances Control Act - PL 94-469], PROPERTIES OF PCB MIXTURES Biphenyl has ten possible sites of chlorination, but of the 209 theoretical Isomers, only 100 are produced due to their thermodynamic properties, Aroclor Is the Monsanto trade name applied to coimnerclal mixtures of PCB, and other than Aroclor 1016, the designations 1221, 1242, 1254 and 1260 Imply, In the last two digits, the weight percent of chlorine in a Aroclor's mixture. For example. Aroclor 1260 will typically contain 100 different PCB Isomers, the mixture 60S chlorine. The above Aroclor mixtures are the primary types used In utility equip ment (4-8). Approximately 95 percent of 1260, 80 percent of 1254, 10 percent of 1242, and 1 percent of 1221 consist of PCB with five or more chlorine atoms per biphenyl molecule. PCB with higher chlorination tend to be more environmentally persistent, thus these percent distributions are Important considerations. The stable properties of PCB that make It nearly Inert and an Ideal fluid for use In electrical equipment also provide the properties that cause It to persist In the environment. PCB are nearly water Insoluble, non-polar, and lipid soluble. Being hydrophobic with very low aqueous solubility they are readily adsorbed to solid surfaces such as silt, clay and sand particles as well as to biological cells (9^ 10). PCB are volatile, and when not sorbed onto solids will transport the air-water Interface to the atmosphere (U_). Being hydrophobic and llpophylllc, PCB are also transportable across biological membranes and In the absence of 4-73 HONS 214059 enzymatic metabolism In the cell, will bioaccumulate. Often the octanol/water partition coefficient Is used to estimate this bloaccumulatlon potential, which is large for PCB (12). PCB USE IN UTILITY EQUIPMENT Seven utility equipment types are typically designed to use PCB or mineral oils (H). They are: e transformers e capacitors e voltage regulators e circuit breakers e reclosers a switches e underground cables Table 2-1 gives an estimated Inventory of the number of units, volume of fluid, and weight of PCB contained In utility equipment. PCB and mineral oil are used as coolants and dielectric fluids. Of the seven types of equipment listed, all but capacitors are designed to function with mineral oil. Capacitors are 100* PCB with Aroclor 1242 as the principal dielectric fluid. PCB OISPOSAL REQUIREMENTS Leaks and Spills The possibility exists that leakage or rupture of utility equipment can occur, resulting In contamination of environmental media by PCB. During a recent survey of utility use of PCB (7,), the Resource Planning Corporation (RPC) Investigated the Incidence of leaks and spills from transformers and capacitors. Table 2-2 sunmarlzes the expected frequency and size distribution of these spills. Table 2-3 provides estimates on the geographic distribution of spills for mineral oil transformers and capacitors. In comparison to oil transformers, percentages of spills from askarel units will be lower In "rural grasslands," higher In Industrial areas and public areas. The major concern for spills In each case occurs In residential neighborhoods. These spill locations reflect upon the Inwdlacy of decontamination requirements; and the method of decontamination depends upon the general terrain, soil type, rainfall and runoff, vegetative cover, and the presence of water bodies. 4-74 HONS 214860 r r Equipment Table 2-1 INVENTORY OF PCB IN UTILITY EQUIPMENT (7) No. Units Gal. Fluid - Askarel Transformers.............. .39,640 Mineral Oil Transformers........20,227,428 Capacitors........... ................... . .2,800,000 Voltage Regulators..........................145,159 Clrcult Breakers..............................180,939 Reelosers...................................... .....170,156 Switches/Sectional Izers............... .385,766 8,525,404 958,365,880 7.547.669 17,840,968 137,335,668 3.403.670 1,415,769 Lbs. PCB 74,597,263 262,230 87,552,460 6,707 12.685 410 329 Table 2-2 FREQUENCY ANO SIZE DISTRIBUTION OF PCS SPILLS (7) Unit Askarel Transformers Mineral Oil Transformers Capacitors No. of Leak/Spill Incidents 317 161,819 21,564 Size Distribution Lbs PCB/Incldent 4.4-3500 0.0-19 5.8-90 Average Lbs PCB 66 <1 17 Table 2-3 GEOGRAPHIC DISTRIBUTION OF PCB SPILLS (7) Waterways (lakes, reiervolrs, rivers, etc.) Rural Desert Land Rural Wetlands, Marshlands Rural Grassland (farmland, ranchland) Industrial Area Residential Neighborhood Public Area (shopping center, schools, etc.) TOTAL 4-75 Mineral Oil Transformers X of Spills 0.38 <0.09 0.1 9.2 13.2 70.6 5.9 lud.trx Capacitors t of Spills <0.00061 1.5 1.0 16.7 29.1 45.3 6.4 mot HONS 214861 rr Planned Olsposal Federal regulations under TSCA and RCRA (Resource Conservation and Recovery Act) require the eventual phaseout and controlled disposal of PC6, including all PCB contained In utility equipment (14.). This represents over 1 billion gallons. The data of Table 2-1 reveal that the most common form of PCB-contamlnated fluid Is slightly contaminated mineral oil. The Identity of Aroclors whfch make up the contamination Is not explicitly known. This low level of PCB watte will require large treatment system capacity to destroy the PCB. Fluids contained within electrical equipment must be removed by washing with an appropriate solvent (6), which further dilutes the waste. PCB disposal capacity depends on the rate at which the units In Table 2-1 are retired. Estimates (6) have yielded the following average annual waste production rates for the highly concentrated askarel transformer and PCB capacitors: For Capacitors: 5.25X of Inventory (approximately 140,000 units, 400,000 gallons PCB, 4.6 million pounds PCB) per year from 1984 1997. For Askarel Transformers: 2.631 of Inventory (approximately 1000 units, 224,000 gallons, 2 million pounds PCB) per year from 1984 2017. Therefore, a substantial portion of the Inventory will remain even with a 5-10 year lead time to develop blotreatment methods. 4-76 HONS 214862 Section 3 BIOTECHNOLOGY AND ITS POTENTIAL APPLICATION TO PCB DISPOSAL Biotechnology Is a field of the applied sciences which encompasses those engineered Industrial processes that use living organisms or their functional components (such as enzymes). For general applications to hazardous wastes, the reader is referred to a contract report to the Congressional Office of Technology Assessment (OTA) by AmTech Consultants (15) and In the proceedings of a recent symposium (1). The basic principle behind PCB biodegradation can be simply Illustrated In Figure 3-1. In biodegradation, enzymes which are coded for by ONA serve as blocatalysts .to cleave the biphenyl ring and dechlorlnate the substituent groups. In effect, PCB Is used as a source of food to the organism, and enzymes oxidize the hydrocarbons conserving the energy liberated while ultimately producing carbon dioxide, water and chloride. In this regard, the metabolism of PCB to inorganic end products can be equated to thermal oxidation (combustion and Incineration) but through naturally produced blocatalysts reacting at ambient temperatures. As Illustrated In Figure 3-1, there are many factors, genetic and non-genetlc, that control the production and activity of a blocatalyst as well as the host organism, usually bacteria. General texts In biochemistry, molecular biology and microbiology review these principles (17-20), Before presenting existing data on the metabolism of PCB, some of the characteristics of PCB which render it recalcitrant to biodegradation are presented. PCB RECALCITRANCE There are several factors that render PCB recalcitrant to biological attack. These Include Its Innate characteristics such as hydrophoblclty and molecular structure and complexity. Also possible are the presence of toxic inhibitory concentrations of chemicals and lack of a required degrading enzyme. e PCB Hydrophoblclty: PCB are hydrophobic molecules with very low solubilities in water. Thus insoluble In water, they will remain separated and form a relatively small surface area to volume Interface to the aqueous phase. Because degrading microorganisms live In primarily an aqueous environment. Inaccessibility of the PCB substrate Is a principal physical limiting Tacto^ contriouting to its biological recalcitrance and environmental persistence. This 4-77 HONS 214863 - <LO biioliculir lipfd bmd tMbrani non-ctnpfl* i t tv? Inhibition -^0=*/?P eonpatlitvt inhibition -0=i^> -ifWISf.I.fc GO J ft OC6tVt U producti IfAAtlltIon lodc! ton* rtf ftn Ion of fJfntli for ~jgj t Material Ion Inhuman________ (dtrcprtttion) mMA miHiiimmnnmiiiLiian tlfpntcri pi ton vJo MMojwji* uitij ii ...............................mini tr hiMl4 or Chromoicmol Bn (Boony ribonucleic ecld) UCtM0 S Subiirili t CfllJflK P PrMuci I llon-cempec 111 ve IrtMhltor Cl Competitive Inhibitor If Terminel (nop) coUe Op OperetOr e"* Pr Promoter ** Itpiluor * bepreitor Figure J-l. GMr(1 Scheabtlc of [nijmt Cttil/tli uUh Genetic end Hon-Gene tic Conti oil Source: AoFoch Contultenti HONS 214864 Inaccessibility, however, can be reduced by Increasing PC6 contact with the solution phase by forming emulsions of PC6. Enhancement of PCS degradation has been noted by several researchers using emulsifiers such as lignin sulfonate (2U and Tween 80 (22). Biologically derived emulsifiers, tenaed emulsans, have been found to form stable oll-ln-water emulsions and though they have not been tested with PCB, could provide a means for Increasing the exposed surface area of mineral oil or PCB for subsequent biological attack (23). Molecular structure of PCB greatly affects biological persistence, in is is due to "electronic and sterlc hindrance Imparted by the number and location of chlorine substituents on the biphenyl rings. The electronic factors contributing to PCB recalcitrance Include1, a) delocalization of electrons by resonance; and b) the high energy of the aryl carbon-chlorine bond. Resonance contributes to chemical stability by making electrons less susceptible to attack. Thus, Increasing the number of chlorine substituents can Increase stability. Chlorine atoms are quite large, especially In comparison to the hydrogen atoms they have replaced on biphenyl rings. Their size Inhibits effective nucleophilic or electrophilic attack on the carbon-chlorine bond or on nearby carbon-hydrogen or carbon-carbon bonds. Thus the location of chlorine substituents can also effect stabllIty. e Other chemical factors related to the recalcitrance of PCB are associated wlft the types of chemicals present, their concentrations and the medium In which the PCB Is located. Alexander (24) has enumerated many factors which contribute to the biological recalcitrance of xenoblotlcs In general, and some which are applicable to PCB. Comnercla) formulations of PCS are most often found In the presence of PCB compatible solvents such as mineral oils and chlorinated allphatlcs and aromatics. For biological degradation of PCB to occur, these materials must be simultaneously degraded by mixed microbial populations. Co-metabolIsm Is frequently observed (25) and as reported by Tabak (26) and Barth (27), a wide variety of substituted chemicals, Including chlorinated benzenes and phenols are relatively biodegradable. As PCB Is likely to be the most persistent material In a mixture, solvents should not pose a major problem. e Biological factors that are associated with the recalcitrance of PCB can be associated with either the absence. Inhibition, or deficiency of an enzymatic activity needed to catalyze their destruction. In some cases there may not exist In nature enzymes with the properties necessary In order to efficiently degrade certain higher chlorinated Isomers; however, over time processes such as plasmid evolution can be expected to develop degradatlve systems (26-30). and enzyme engineering may also be applied to correct catalytic deficiencies (31). To suanarlze, there are Innate characteristics of PCB as well as environmental factors that may cause Its biological recalcitrance. The majority of these may be controllable to achieve more efficient degradation. 4-79 HONS 214865 METABOLISM OF POLYCHLOR1NATEO BIPHENYL The only practical route of biological degradation of PCS Is through ring cleavage and dechlorination. Many types of unicellular organisms, primarily bacteria, have been Identified as being able to degrade to some extent the various PCB compounds (32-46). Both aerobic and facultative anaerobic bacteria have been Isolated from various environments, including the open ocean and sediments of the Hudson River. Thus, in nature, there appears to be no particular genus or species that is the major user of PCB. The two major chemical changes occurring in PCB metabolism are ring cleavage and dehalogenation. In no case has dehalogenation been reported to occur prior to ring cleavage. A complete pathway for the aerobic/anaerobic degradation of PCB is presented in Figure 3-2. . Ring Cleavage. Furukawa (35) presented a pathway which is quite consistent for the aerobic degradation of aromatic compounds, and provides confirmatory evidence that ring cleavage occurs prior to dehalogenation. Ring cleavage occurred on the ring with the lessor substituents at the meta position. Furukawa was unable to fully degrade the PCB metabolic product (chlorobenzoic acid) with his two cultures but other studies have demonstrated this to occur In mixed culture (47, 48). Ring cleavage of aromatic compounds is performed most efficiently in aerobic systems due to the ease of incorporation of oxygen by oxygenases, but also occurs anaerobically (49). A degradation product of PCB, dlchlorobenzoic acid, has undergone anaerobic ring cleavage to yield methane and carbon dioxide (50). Dehalogenation. Dehalogenation Is obviously an Important mechanism in the detoxification of PCB, but has not been specifically reported for the Intact PCB molecule prior to ring cleavage. Following ring cleavage dechlorination occurs quite readily (3S). Dehalogenation as an Initial step would be extremely important for the more highly chlorinated compounds of Aroelor 1260 or in those structures that appear highly persistent. The enzymatic reactions which catalyze the removal of halogens from hydrocarbons have been grouped into three classes by Bollag (51): a) hydrolytic dehalogenation, In which a hydroxyl group replaces the halogen atom; b) reductive dehalogenations, where halogens are exchanged with hydrogen, and c) dehydrodehalogenation, in which both hydrogen and chlorine are removed with the formation of a double bond. e Hydrolytic Dechlorination. Hydrolytic dechlorination is the most prevalent ot tne types observed. It Is the method of aerobic chlorine removal from PCB alkyl chains arising after ring cleavage. 4-BO HONS 214866 Figure 3-2. General Pathway for Total PCB Degradation. Source: AmTech Consultant* 4-B1 ' HONS 214867 rr' Reductive Dehalooenatlon. Reductive dehalogenatfon has been demonstrated for numerous microorganisms and Is most effective under anaerobic conditions, with the notable exception, until recently, of chlorines attached to aromatic rings. Direct reductive dechlorination has now been observed In the laboratory with meta-substituted dlchlorobenzolc acid (52,53). Rates of PCB Degradation _ Nearly all published data have Indicated that the rate of PCB degradation Is a direct function of the number and location of substituent groups on the biphenyl rings. As the weight percentage of chlorine Increases, the rate of degradation decreases (Figure 3-3). Table 3-1 presents a range of reported rates for PCB destruction according to the commercial Aroc lor forms/compounds/percent chlorine. As Indicated In the Table, there Is a general trend of decreasing degradation rate up to the higher chlorinated Aroclors. Data reported for the higher Aroclors are from experiments which focused on optimizing rate, or are reports by commercial suppliers of micro organism formulations. In two of the studies where nearly IDO percent of Aroclor 1248 (22) and 1254 (21) have been reportedly degraded, measures were taken to overcome some of the physical limitation factors on PCB solubility by adding emulsifiers. The results of converting the data In Table 3-2 to half-life form, assuming first order decay, Is presented In Figure 3-4. The data Indicates a biological half-life for PCB from 16 hours to one week for monochloroblphenyls and three days to one year for hexachloroblphenyls. 4-82 HONS 214860 X Chlorine w/w Figure 3-3. Btodegridatfon Rites of Various Aroclors (43).- 4-03 HONS 214069 rf Arpcl or Mixture or Set of Isomers Studied Biphenyl Monochlorlnated Monochlor Inated Aroclor 1221 Aroclor 1221 Dlehlorlnated Dlchlorlnated Olchlorfnated Trlchlorlnated Trlchlorlnated Trlchlorlnated Trlchlorlnated Trlchlorlnated Aroclor 1016 Aroclor 1242 Aroclor 1242 Aroclor 1242 Tetrachl orlnated Tetrachlorlnated Tetrachlorlnated Tetrachl orlnated Tetrachlorlnated Aroclor 1248 PentachlorInated Pentachlorlnated Pentachlorlnated Pentachlorlnated Aroclor 1254 Aroclor 1254 Aroclor 1254 Aroclor 1260 Aroclor 1260 Aroclor 1260 Aroclor 1260 Aroclor 1260 Table 3-1 REPORTEO RATES OF PCS DEGRADATION Reported Rate 100S/2 days >10 mg/l-hr 1001/5 days 81t/2 days lOOt/30 days 1 - i 10 mg/l-hr ave i 7 mg/l-hr 50-100X/7-15 days 99X/5 days 0-9 mg/l hr ave 6 mg/l hr 79-935/5 days 30-755/14 days 60-855/28 days 50-705/28 days 96-985/100 days 331/2 days 88-955/52-100 days 265/2 days 0-2 mg/l-hr 80-895/5-15 days 0-405/14 days 55-805/28 days 35-605/28 days 975/90-130 days 0.1 mg/l-hr 05/14 days 25-555/28 days 20-405/28 days 86-995/4-18 days 30-405/8 weeks i 505/8 weeks >995/16 weeks >995/4 days 565/2 weeks 9 110 ppm 245/2 weeks 9 380 ppm 205/2 weeks 9 600 ppm 145/2 weeks 9 860 ppm 4-84 _ Reference 43 35 54 43 45 35 33 54 35 54 55 55 55 33 43 33 43 35 54 55 55 55 22 35 55 55 55 2i 56 57 58 59 59 59 59 HONS 214870 -85 Source AaTcch ContultMU HONS 214871 rr DATA VALIDITY AND TENTATIVE CONCLUSIONS The scope of this study did not permit detailed review of each of the studies represented in figure 3-4. Most of the data on the lest chlorinated Aroclors were acquired from the peer-reviewed literature. However, all of the information in figure 3-4 on Aroclors 1260 other than the lignin sulfonate emuVelfler data came from proprietary field studies reported through direct cosmwnicatlon. Hits does not mean that these latter data are invalid, since proprietary trade-secrets are usually protected from public scrutiny. In order to establish the validity of such field data, specific case studies must be examined in detail or a controlled demonstration must be performed. For example, there have been no reports In the peer-reviewed literature of anaerobic degradation of PCB; therefore, validation of the bench scale thermophilic, anaerobic (compost! degradation of Aroclors 1248 and 1254 is required. Similarly for the other proprietary results. Nevertheless, Figure 3-4 suggests that biotechnology application to PCD appears to hold promise as a treatment technology and the following tentative conclusions can be reached: e Present results indicate PCD Isomers at all levels of chlorination are biodegradable. e The biodegradation rate for PCB appears to decrease with increasing molecular chlorination, all other factors being equal ( e.g., absence of emulsification). e This observed PCB biodegradation is performed by bacteria which occur naturally (i.e., have not undergone "recombinant DNA"). e The biotechnologies represented fn Figure 3-4 should be amenable to considerable improvement In efficiency. e Certain approaches represented In Figure 3-4 promise to Improve PCB biodegradation significantly, I.e., emulsification and selective cell cultures. e Although less promising than the cate for planned PCB disposal, ' Improved cell systems might actually permit rapid on-site treatment of PCB spills possibly with special cell lines. If not, It Is still possible that a mixture of (e.g.) soli removal for off-site biological treatment, and In-sltu biotreatment for low-level residuals would be attractive. IMPROVING BIOLOGICAL PERFORMANCE The overall performance of the biological component of a treatment system can be enhanced by several methods employing the relatively new techniques of applied genetics (I.e., genetic engineering). Several methods of Improving biological performance are discussed below. 4-66 MOMS 214872 f r Genetic Engineering "Genetic engineering" refers to deliberate manipulations of an organism's genetic code to a desired biological characteristic. The methods addressed concern: a) mutation and selection; b) recombinant OHA; c) plasmid evolution. * Mutation and Selection. Mutations can be Induced to dnpgm the chemical structure o7" OHA. Mutations In the genetic regions controlling promoter, operator, or regulator (MIA sequMCes can enhance the rate of transcription of specific proteins. Ionizing radiation and certain chemicals Induce mutations and have been coamerclally used for several decades In the pharmaceutical Industry to Improve yields of metabolites from production cell lines (60). It has also been applied to mutate microorganisms to preferentially degrade specific toxic pollutants (61). Selection Is required to conserve a beneficial genetic trait arid In the laboratory Is usually conducted through enrichment subculturing and plating techniques. e Recombinant QUA. The Industrial technique termed "recombinant OHA" implies the engineered transfer of OHA from one type of organism to another. In this manner a unique gene which codes for a specific degradatlve enzyme could be Inserted Into a rapidly producing bacteria, providing for enhanced production of the desired blocatalyst. Because numerous enzymes may be Involved, the utility of recombinant DMA In RGB degradation may be limited to producing whole cell lines capable of degrading many different PCB Isomers on single or multiple plasmid copies In one or more bacterial strains. Plasmid Evolution. As In the evolution of antibiotic resistance by bacteria, evolution of specific plasmid genetic codes for toxic pollutant degradation xcurs. Plasmids are not only transmissible between similar organisms but also between aerobes and anaerobes (62.63). Using techniques similar to routinely used enrichment/selaction procedures for Isolating cell lines, the knowledge of plasmid evolution allows for a more scientific approach In selecting the Initial Inocula. In this approach a wide variety of naturally selected bacteria (e.g., from contaminated dump sites) are inoculated with artificially selected bacteria harboring plasmids known to code for various sequences In a degradatlve pathway. Enzyme Enolneerlng Enzymes are the blocatalysts responsible for PCB degradation. Enzymes can be extracellular or Intracellular. Intracellular enzymes are more expensive to extract from the cell and can be unstable, though Immobilization often Increases their stability. There Is no evidence to Indicate that PCB degrading enzymes are extracellular. Because multiple Isomers of PCB usually are present the range of necessary enzymes Is large and the probability of using one or two enzyme systems to detoxify all PCB Is extremely remote. An enzyme's catalytic activity can be , 4-87 HONS 214873 rr Improved through various techniques using genetic engineering or by chemically synthesizing similar functional groups (3U. Bioaugmentation The addition of biological or other materials which Improve system performance Is frequently termed bioaugmentation. Additives can Include biological emulsifiers (emulsans), synthetic emulsifiers, nutrients, enzymes, flocculants, mad microorga nisms- They are used to overcome treatment difficulties In operating miotreataent systems, to maintain the population of a slow-growing organism, or in routine operations and maintenance. The main purpose of bioaugmentation Is to overcome factors contributing to the persistence of chemical wastes by providing the orga nisms with a growth complement. Their additions are no different, for example, than the making up of spent catalysts In chemical systems. 4-88 MOMS 214874 w r Section 4 610TREATNENT PROCESS SYSTEMS Many biological technologies used to trett Industrial wastes are applicable to PC8 destruction. System design depends upon waste characteristics (the waste matrix) and the degree of treatment required. The microorganisms utilized In biotreatment systems can be broadly classified as aerobic, anaerobic or facultative. Aerobic organisms require molecular oxygen for metabolism, and anaerobes function In the absence of oxygen; facultative organisms function In either the aerobic or anaerobic state. Aerobic organisms are usually considered to perform respiration, that Is the use of molecular oxygen as the ultimate electron acceptor. Anaerobic organisms usually perform fermentetlon, which Is metabolism using some other source than molecular o*y9" * the ultimate electron acceptor. The types of biological technologies applicable to PCB are Identified below, along with brief descriptions of their present applications, limiting factors and envlronmentel concerns. They Include: e conventional aerobic and anaerobic processes e land disposal techniques e enzyme systems e special Ized fermentation systems e spill clean-up end remedial techniques CONVENTIONAL AEROBIC AMD ANAEROBIC TECHNOLOGIES Most organic Industrial wastes are amenable to biological treatment and the degrad* Ing populations become specifically adapted to certain compounds. Higher chlo rinated PCB may, however, pose problems of recalcitrance and accumulation via adsorption onto the disposed sludge. In order to echleve maximum PCB depredation. It Is necessary to base system design upon the specific PCB degradatlve biokinetics, or upon a mode of bioaugmentation. MOMS 214875 4-89 f Aerobic Treatments Ring cleavage of halogenated aromatics such as PCD, is efficiently accomplished via oxygenases with molecular oxygen. Aerobic treataients of PCB are therefore very Important In their biodegradation. Types of aerobic treatments Include: e activated sludge e aerated lagoons e aerobic digestion e trickling filters e rotating biological contactors Anaerobic Treatments Anaerobic treatments are potentially applicable to the reductive dehalogenat ion of PCB and degradation of oxidative products. The following are types of conventional anaerobic treatment systems: e anaerobic digestion e anaerobic contact e anaerobic fixed-film treatment processes In general, these aerobic and anaerobic systems have been designed to treat miscel laneous soluble and Insoluble organic matter In the most cost effective manner reflecting minimum retention times and capital costs. As It appears that most PCB can be degraded over time -the length of time required Increasing with the percent of molecular chlorination -either extended periods of treatment time, special pretreatment, or bioaugmentation would be required for existing conventional system to degrade the more highly chlorinated Isomers. In adapting these systems as a primary means to treat PCB, aerobic processes are required because PCB Is aromatic and molecular oxygen must be supplied (via oxygenase) to efficiently cleave the aromatic ring. Secondly, PCB hydrophoblclty causes It to accumulate on sludges/sollds, requiring an aerobic system with high solids treatment such as aerobic digestion. On the other hand, complete anaerobic degradation of PCB might eventually be possible with the development of anaerobic cell lines, but would require strict microbial control. MOHS 214876 4-90 LAND FARMING AND COMPOSTING Sludges and solids contaminated with PCB can be biologically treated In various land-based systems, either on the soil (land farming), or In composting systems. e Land farming Implies that microorganisms are being gr%wn utilizing tne waste as a carbon and energy source, plant crops also can be grown. The Environmental Protection Agency Includes tand fenetmg In Its hazardous waste regulations as a disposal method (fd). #here are over ISO land farming operations In the United StatesTenglng up to 1,500 acres In size. The petroleum refining Industry uses the majority of these sites for the disposal of refinery wastes, such as API separator sludge. Land treating of toxic materials Is highly controversial mainly because It Is conducted In the open environment, but frequently It may be the only alternative In the case of low level contamination from spills. It can be conducted on a spill site as a primary treatment or as a polishing operation, or off-site for treatment of the removed contaminants. Polybac Corporation recently reported treating a quarter acre land site of PCI contaminated soils using land farming techniques (57). e Composting Is similar to land farming as It can employ a blended soil/organic matrix. Generally It Is a well aerated and mixed process occurring on the soil surface or In contained systems. To some extent It resembles a dry aerobic or anaerobic treatment with the particles of celluloslc and lignin based solids serving as the support medium for biological growth. Composting Involves blending dry organic material such as straw or agricultural residues with materials such as contaminated sludge or soil, and allowing the saprophytic organisms, bacteria, fungi and actlnosycetes to degrade the matrix over several weeks to months depending on the process. . Only one oroprletary report of PCI degradation by composting has been provided to this study which was conducted on a laboratory scale (55). PCB degradation was observed in aerobic and anaerobic compost systems, but not In controls. An unusual result was a reported 25 to 40 percent degradation in anaerobic systems. ENZYME SYSTEMS To date, use of either cell-free extracts or In-vitro enzyme systems for PCI degradation has not been reported. Applications of enzyme systems to PCI degrada tion would be most appropriate as an initial detoxification step, catalyzing either aromatic ring cleavage or dehalogenatlon. The use of enzymes In Industry has been reported extensively. How an enzyme system Is applied depends upon the characteristics of both the substrate, enzyme, and process. Maintaining enzyme stability and activity Is most l^orttftt as their makeup can be quite costly. Enzymes can be applied In a free or ImMbftlzed state, the latter on carriers in either a fixed or fluidized reactor comflgift'atlon. In 4-91 MONS 214877 their free state the enzymes become part of the substrate matrix and are wasted with the catalyzed substrate solution. Thus, Immobilization Is preferred In order to conserve the enzyme and allow It to be used many times over. However, for water Insoluble substrates (such as PCB), free enzyme systems would be required unless emulsifying could place PCB Into aqueous suspension. _ SPECIALIZED FERMENTATION SYSTEMS Specialized fermentation systems are developed for use on a specific substrate and are either aerobic or anaerobic processes with pure or mixed cultures. They usually take on an intensive approach to process control. Specialized fermentation systems employ specific cell lines which metabolize well-defined substrate streams and are optimized on scales much smaller than for conventional systems. Specialized fermentations use cell systems in either immobilized or in suspended states. The main difference between immobilization and suspension fermenters is In the method for contacting the bacteria with the substrate. Fermenters usually employ a high rate of mixing to force the biomass into suspension. Thus, care must be taken to avoid shearing and damage to the cells. Instead of mixing, Immobilized systems use an array of support media packed as columns or beds in the reactor. SPILL CLEANUP SYSTEMS Biological technologies can be used to clean up toxic comnerclal materials that contaminate either land or water. The principles Involved In land application also apply to accidents Involving spillage of organic compounds. The major difference is that very high loading rates are often associated with spills, but tha affected area is normally small in size. 4-92 MOWS 214878 r r Section 5 PROMISING APPROACHES BIOTREATMENT OP PCB SPILLS Biotreatment of PCB-contamlnated spill sites would be preferable to the commonly used technique of excavating and landfilling contaminated materials. A large demand exists for a rapid, easily applied treatment method for PCB spflls, and biotreatment may eventually be used In this manner. An Ideal blodegradatlve system would Include an organism and/or Its associated enzymes in a packaged form which could be reconstituted In the field and applied as needed to contaminated solids. In the near term. In-place treataient appears feasible for particular contamination situations where time Is available and a health threat Is not Imminent. Polybac demonstrated Its ability to treat on land Arxlor 1254 contamination of 230 ppm over a 5-6 month time period, but required the use of a proprietary reducing agent to dechlorlnate the higher chlorinated Arxlor forms. Hlth additional research and Improvement In cell lines and additives, It 1i likely that soil cleanup txhnlques could be considerably Improved. On-slte treatment with mobile composting facilities or centralized off-site treatment on land farms may also be applicable. BIOTREATMENT FOR UTILITY PCB FLUIDS Systems for planned disposal of bulk PCB-contamlnated fluids must be capable of treating PCB, mineral oils, polychlorinated benzenes, and solvents used to wash out utility equipment. Although these wastes will be diluted to PCB concentrations amenable to biological treatment, high organic loading will xcur. Continuous or batch flow systems could be used. Emulsification would play a significant role In all systems as well as bioaugmentation with genetically engineered cell lines. Types of coxepts Include: e coupled high-mix aerobic blorextors e sequential anaerobic-aerobic-anaerobic e {mobilized cell aerobic treatment with high pressure oxygen saturation e complete anaerobic treatment 4-93 MONS 214879 K Section 6 CONCLUSIONS AND RECOMMENDATIONS Biotechnology appears to hold considerable promise for the treatment of PCB for the electric utility Industry. However, studies to date have been conducted mostly on laboratory scales. Some have not been validated, and most have not considered the complex problems associated with actual waste matrices and practical treatment approaches. Thus, efforts are required to substantiate findings reported herein and to advance technology development. CONCLUSIONS From the review and analysis of available published and private data we conclude: 1. Efficient treatment of electric utility PCB wastes will become feasible and should be available In 5 to 10 years given adequate research and development. Z. Promising biotechnologies Include: --Improved PCB degrading species through genetic engineering. --Specialized bioreactors for aerobic and anaerobic PCB treatments. -Possible enzyme systems for select PCB decontaminations. --Systems for growing and harvesting cells for bioaugmentation, as well as In enzyme production processes. RECOMMENDATIONS FOR RESEARCH AND DEVELOPMENT If biotechnology Is to echleve its potential to treat electric utility PCB wastes, a well-designed research and development strategy should be developed and Imple mented. Such a strategy should Include validation of existing data as a first step to confirm and advance the state-of-the-art. Research and development should focus on three areas: microbiology and genetic engineering; system design and testing; and cost analysis. These recommendations are briefly listed below. 4-94 HONS 214880 Microbfoloqy and Genetic Engineering Inventory end collect PCB degrading cultures and isolate new strains e Identity degradative pathways, enzymology and genetics e Improve degradative performance through genetic engineering techniques on collected strains e Character!za and test emulsifiers, and transfer PCB degradative genetics to bioemulsifying strains e Test cell extracts for PCB degradative potential System Design and Testing e Process development systems are required based upon the performance of pure or mixed PCB degrading cultures above e Promising cell-free enzyme systems will require development of production reactors and application techniques ' e All charecterized waste matrices require laboratory and field test development e Successful systems require scale-up and pilot plant operations Cost Analysis e Conceptual design cost analysis should b conducted for candidate systems to screen for cost-effectiveness e Sensitivity and financial risk analysis should be performed in comparison with approaches alternative to biotechnology for those selected as serious candidates 4-9S MONS 214881 BIBLIOGRAPHY 1. AmTech Consultants and Lei and D. Attaway A Associates "Application of Biotechnology to PCB Disposal Problems." Draft Final Report EPRI RP-1263-16, July 19B3. 2. Versar, Inc. Durfee, R.L.; et al. "Polychlorinated blpenyls 1929-1979: Final report." EPA: 560/6-79-00*7 PF 296559, 1979. 3. Resource Planning Corp. "Consents of the Use of PCB In Response to an Order of the United States Court of Appeals for the District of Columbia Circuit." Submitted by the Utility Solid Wastes Activities Group, the Edison Electric institute, and the National Rural Electric Cooperative Assoc, to the US EPA February 12, 1962. 4. Versar, Inc. Durfee, R.L.; et al. "PCBs In the U.S.: Industrial use and environmental distribution." "EPIT $60/6-76-005 PB252012, 1976. 5. Hutzlnger, 0.; Safe, S.; Zltko, V.; The Chemistry of PCBs CRC Press. Boca Raton, FL, 1974. 6. Stearns, Conrad, and Schmidt Consulting Engineers, Inc. Disposal of PCB and PCB-contamlnated Materials Vol. 1 EPRI FP-1207, 1979. 7. ibid., 3. B. Personal CoanunIcatlon from Steve Hamilton, General Electric, to Eric Terhorst, Leland D. Attaway 1 Associates, April 7, 1983. 9. Steen, W.C.; et al., "Partitioning of Selacted Polychlorinated Biphenyls to Natural Sediment?* Water Reserach Vol. 12, pp. 655-657, 1976. 10. Hague, R., et al., "Aqueous Solubility, Adsorption, and Vapor Behavior of Pdiychl or Ina^ed "biphenyl Aroclor 1254," Environ. Scl. Tech. 6:139-142, February 1974. 11. Paris, D.F., et al., "Role of Physico-Chemical Properties of Aroclor 1016 and 1242 In Determfnlng their Fate and Transport In Aquatic Environments," Chemosphere 4:319-325, 1976. 12. ibid., 11. 13. ibid., 7. 14. u.S. Environmental Protection Agency; Federal Register 47:37342, 1982. 15. Peyton, T.O., "Application of Biotechnology to Hazardous Haste Disposal," Contract Report to the Congressional Office of Technology AssessmMt, February 1982. 4-96 HONS 214882 r r 16. "Genetic Control of Environmental Pollutants," Symposium, Seattle, 31 July - 3 August, 1983. 17. Lehnlnger, a.l., "Biochemistry: The Molecular Basis of Cell structure and Function," Worth Publishers, Inc., New fork, 1970. 18. Norton, C.F., Microbiology, Addlson-Wesley Publishing, Reading, HA, 1981. 19. Armstrong, F.B. and T.B. Bennett, Biochemistry, Oxford Un1ver*1ty tress. Mew York, 1977. ----------------- 20. Pelciar, M.J., Microbiology. KcGraw Hill, Inc., New York, NY, 1977. 21. Mu, 0.. "Enhancement of PCBs Biodegradation by Sodium Llgnlnsulfonate," Water Research Yol. 14, pp. 1467-1475, Pergamon Press Ltd., 1980. 22. Handy, M.K. and Y.C.Lin, "Biodegradation of Chlorinated Hydrocarbons," University of Georgia, January, 1963. 23. Wilson, G., personal coemunlcatlon, Petroleum Fermentations, Inc., Amelia Is., FL, March 1983. 24. Alexander, M., et al., "Nonbiodegradable and Other Recalcitrant Molecules." Biotech. I Bloeng.,*75:611-647, 1973. 25. Jacobson, S.N., et al., "Evidence for Cometabolism In Sewage," Appl. Environ. Microbiol. 40:5:917^21, 1980. ---------------- 26. Tabak, H., et al., "BlodegradabllIty Studies with Organic Priority Pollutants," jWtF *53:10:1503-1518, October 1981. 27. Barth, E., "Biodegradation and Treatability of Specific Pollutants," EPA Report No. 600/9-79-034, October 1981. 28. Pemberton, J.M., (t i[., "Evolution and Spread of Pesticide Degrading Ability among Soil Microorganisms/ In plasmids of Medical Environmental and Commercial Importance, Elsevler/Nortn MoiTand Tlbiefical Press. Amsterdam. wr------------- 29. Kamp, P.F., and A. Chakrabarty, "Plasmids specifying p-Chlorodblphenyl Degradation In Enteric Bacteria," In Plasmids of Medical. Environmental and Commercial importance. Elsevier/North Holland Biomedical Press, Amsterdam, 30. Kellogg, S.T., et al., "Plasmid Assisted Molecular Breeding Approach: A New Technique for tnhaffiftd Biodegradation of Persistent Toxic Cultures,* Scfence 214:1133-1134, 1981. ----------- 31. Ulmer, K.M., "Protein Engineering," Science, 219:666, 1983. 32. Ahmed, M., et al., "Degradation of Polychlorinated Biphenyls by Two Species of Achromobacter .'""tan J .Microbiol. 19:45-52, 1973. 33. Baxter, R.A., et al., "The Degradation of Polychlorinated Biphenyls by Microorganisms. ""Sc fTTotal Environ. 4:45-61, 1975. 34. Carey, A.E., et al., "Metabolism of polychlorinated Biphenyls by Marine Bacteria," BulfTEnyTron. Contain. Toxicol. 20:527-534, 1978. 4-97 ` MQNS 214M3 rr 35. Furukawa, K., at al., "Alcal Igcnes and AcInetobacter strains capable of degrading DOlvchTSnnated biphenyls. Agrlc.'Biof. 1'nem. 42:543-548. 1978. 36. Furukawa, K., "Microbial Metabolism of Polychlorinated Biphenyls. Studies on the relative degradability of polychlorinated biphenyl components by Alkallgenes sp-." Agr. Food Chem. 24:251-256, 1976. 37. Furukawa, K. and A.M. Chakrabarty, "Involvement of Plasmids (a Total Degradation of Chlorinated Biphenyls," Appl. Environ. Microbiol. 44:819-626, 38. Neu, H.J., and K. Ballschmlter, "Abbau von Chlorlerten aromaten: mlkroblolglslier abbaw der polychlorlerten blphenyle (PCS). II. Bfphenylole als metabolite der {PCB>." Chemosphere 7:419-423, 1977. 39. Relchardt, P.B., et al_., "Kinetic Study of the Biodegradation of Biphenyl and Its MonochlorlnaTed Analogues by a Mixed Marine Microbial Coneunlty." Environ. Scl. Technol. 15:75-79, 19B1. 40. Sayler, G.S., et al., "Growth of an Estaurlne Pseudomonas sp. on Polychlorinated tfTphefiyls," Microbial. Ecol. 3:241-255, lif/f. 41. Shlarls, M.P. and G.S. Sayler, "Biotransformation of PCB by natural Assemblages of Freshwater Microorganism." Environ. Scl. Technol. 16:367-369. 1982. 42. Sylvestre, M., and J. Fauteux, "A New Facultative Anaerobe Capable of Growth on Chlorob 1 phenyls," J. Gen. Appl. Microbiol. 28:61-72, 1982. 43. Tucker, E.S., et al., "Activated Sludge Primary Biodegradation of Polychlorinated BTplienyTs," Bull. Environ. Contam. Toxicol. 14:705-713, 1975. 44. Tulp, M., et al., "The Bacterial Metabolism of 4,4*" dlchloroblphenyl and Its Suppress 1 on""by""Altentative Carbon Sources." Champsphere d7:103-108, 1978. 45. Kong, P.T.S. and K.L.E. Kaiser, "Bacterial Degradation of Polychlorinated Biphenyls II. Rate Studies," Bull, Environ. Contam. Toxicol. 13:249-256, 1974. ------ 1------------------------------------------- 46. Vagi, 0. and R. Sudo, "Degradation of Polychlorinated Biphenyls by Microorganisms," J. Water. Poll. Cont. Fed. 52:1035-1043, 1980. 47. Horvath, R.S. and M. Alexander, "CometabolIsm of m chlorobenzoate by an Arthrobacter," Appl. Microbiol. 20:254, 1970. 48. Clark, A.R., et al., "Degradation of Polychlorinated Biphenyls by Mixed Microbial CulturesAppl Environ. Microbiol. 37:4:680, 1978. 49. Neufeld, R.D., et al., "Anaerobic Phenol Biokinetics." JWPCF 52:9.2364-2377, 1980. -------- -------- 50. Horowitz, A., "Reductive Dehalogenatton of Halobenzoates by Anaerobic Lake Sediment Microorganisms," (In press). Appl. Environ. Microbiol.. May 1983. 51. Bollag, Jean-Marc, "Microbial Transformation of Pesticides." Adv. Appl. Microbiol., 75-130, 1974. ----------- 52. Sulfite, J.M., et al., "Dehalogenatlon: A Novel Pathway for the Anaerobic Biodegradation oT~HaT*oaromat1c Compounds," Science 218:1115, 1982. 4-98 MONS 214884 r r 53. Sufllta, J.H., "Kinetics of Microbial DehalogenatIon of Haloaromatlc Substrates In Methanoqenlc Environments," Appl. Environ. Microbiol. In Press. May 1983. 54. Clark, R.R., E.S.K. Dhlan, and R.A. Griffin, "Degradation of Polychlorinated Biphenyls by Mixed Microbial Cultures." Applied Environ. Microbiol. 37:4:680. 1978. ---------------------------:-- 55. Atlantic Research Corporation, "Composting as a Method for Degradation of Polychlorinated Biphenyls (PCBs) In Soil, Proprietary Submission to EPR1, 1963. 56. Flow laboratories, personal comnunlcatlon, March 1963. 57. Nassef, N.A., personal comunlcatlon, Polybac Corp., April 1983. 58. Moe, Byron, "Test Data Confirmation Reports," Bloteknlka International, Inc., Springfield, VA. 59. Kopecky, Anne, "PCB Data," personal comtunlcatlon, Sybron Biochemical, Salem, VA. 60. Elander, R.P. and L.T. Chang, "Microbial Culture Selection," In Microbial Technology Ed. Peppier and D. Perlman, Academic Press, New York, 1979. 61. Nome, H.E., "The Activity of Mutant Microorganisms In the Biological Treatment of Industrial Wastes," Voordrtcht gehouden, op 7 maart 1972, Aqua Sana, Leute beurs van Vlaandered-Gent. 62. Farrell, R. and A. Chakrabarty, "Degradatlve Plasmids Molecular Nature and Mode of Evolution," In Plasmids of Medical. Environmental, and Commercial Importance, Elsevler/Nortn ho liana Biomedical rreis, Amsterdam, 63. Kamp, P.F. and A. Chakrabarty, "Plasmids Specifying p-Chloroblphenyl Degradation In Enteric Bacteria," In Plasmids of Medical. Environmental, and Commercial Importance, Elsevler/North Kouana Biomeoicat f'ress. Amsterdam,, TO------------- 64. Brown, K.W., Hazardous Waste Land Treatment, U.S. Environmental Protection Agency, SW-874, tinclnnetf, (IN, 19HU. " ' 4-99 MON$ 214885 f' Diacuaaion Qt, wouldn't many antywaa ba naadad to coaplata tha daatruetion of coaplax PC>? A. Yaa, it would naad a lot of davalopaant work, Uaa of ahola calls can produca aany ansyoaa. QJ. In pr a tr mating apilla, ian't thara a riak of groundwatar contamination? A. Buga ara not poiaon and thsy art in foil anyway. Howavar, if the aa proeaaaoa wr coauaarcialitad, it might ba nacmaaary to uaa aoaa oontainmant to pravant poaaibla mpraading. 4-100 MOMS 214886 r f ACmiAL DEGRADATION OF PCB: EVIDENCE OF DISTINCT PATHWAYS IN COklNESACIEEIMI SP. KB1 AND ALCALIGPM EPIEQP1PS 1150 Doaaa L. Bodard, Mlok**l J, Ituun, Roma Id Uatoraoo {(til Blootrlc Corporal* R****rok tad Dorolo^oat P.O. lot (, $ok*a*ot*dr> NT 12301 INTRODDCTION Nor* tkaa 100 fora* (coaaoaora) of polyoklorlaatod bipkoiylt (PCI*), dlffarlaa la tk* aalir lad podtioa of ckloriaoa, aoro coasoaly aood ovor tko li*t half coa tary. Ik* iui proportion vklek formed tko kaala for tkolr aao (tkoraal labil ity, okaiial tability) kav* alto ooatrlbatod to tkolr aotMalatloa la tk* oaviroaoat. Fartkoraoro, PCS* or* klakly laaolabl* la water aad arc oaly lifktly aobillaod by a'aooaa ty*t*at. Tka*. P3a kato aeoaalatod la aoll aad lak* aad river **dla*at at *lt*a of tkolr prodaotloa, oa*. *tor*|o, or dlopotal. Nlorooraoalw* aro t*a*rally rotpoatlbl* for tk* broakdow* of ctaplai ora*alc material a lato aiaplor fora a tkat oaa b* raoyolod. lmt*r, tyatkotlo coapouada, tpodally tkoao vkiok ar* oklorlaatod, frt^aoatly roaltt aioroblal do|radatioa (1). Thl* koa bo*a tram of PCB*} ooa|*aart oomtdaiat aoro tkaa tkro* aklorla** kavo booa aarkodly rooololtraat to biodotradatioa. Oa tk* etkor kaad, ovidoao* 1* oooalatia| tkat boot aria aoy b* adapt la| by ovolviai aacyao* oapabl* of dotradlas oklorlaatod orcaatia* iaoladiai f3a. It kaa booa wall d*om*atod tkat aay dlfforoat (*a*r* of baotoria ar* oopobl* of doiradiai aoao-, dl-, or trl- oklorobipkoayl* k*a proaoatod a* da|l* ooapoaada (i*J) or la a aaaaorclal aii- taro oaoh * Aroolor 1242 (10-14). lovovor, tkor* bar* booa f report* of dttradotloa of tk* aoro kifkly oklorlaatod P* la Aroolor 1254 . Fartkoraoro, la tkaa* *a*o* ao datoila kav* booa pabll*h*d of oltkor tk* or|aa- ia*> tka oaayaatl* aaohaalaaa, or tk* apoolflelty of d*|radatloa, Oa tk* otkor koad> Firaka** aad oolloaia** kav* aoavia*ia|ly daaaa*tr*tod tk* ability of too par* baotorlol oaltaro*. Aloalla*ao* T42 aad Aoiaatokaatar Pd. to 4-101 MOMS 214887 daproda fUi PCS HiMMti coatalalap Bp to fivo obloriaa atcaa (li.ll.il) , Tba ability of thaaa oaltaraa to daprada a broad raapo of PCS ooaioBara boa prcaptad qaoatioao aboat tba oripia of PCS dapradativa ocapatoaoo ia baotorla aad tba oaiyaotio path* ay a iavolvad. Ia roapoaao, wo bavo aadortabaa aa ia-doptb atady of tba bloobaalatry aad paaatiea of baotarial PCS doprodatloa alaod at providlap a aoaad baalo for aadorataadiap wbat ia bappoaiap la tba aaviroaoat. A aoabar of boy qaoatioao prowidad tba fraaovorb for oar atadiaa; Coaid otbor baotorla ba roadily foaad wltb a dapradativa ocapatoaoo aqaal or oaporior to tbot of Aolaotohaetor Pdf low rapidly caa aloroorpaaiaaa daprada PCBaT That ara tba doprodatloa prodaotaf Vhat atraetaral faotora liait tba dopradatioa of PCS ooapoaora (poaitloa, acabor of obloriao atcai)? Io tboro aoro tbaa oao pathway for dopradatioa of PCBaT Ia tboro aay rolationablp aaoap poaoa which apoeify PCS dapradativa oasyaot ia dlffaroat orpaaiaaaT Caa tbooo poaoa bo oxohaapod aaoap baotarial apooloa or pomoraT Ta doaoribo boro aoao of oar rooaat fladlapo wltb two baotarial apooloa oapobla of dopradlap a broad raapo of PCS ooapoaora iaoladlap poataobloroblpboayla. Ultiaotoly, it ia oar iatoatloa to obaraotorlao tbo bloobaalatry aad poaotioa of PCS dopradatioa la thoao orpaalaaa by otadylap tbo prodaota of PCS dopradatioa. olaoldatlap tbo patbtaay(o) of dopradatioa, aad iaolatiap aad atadyiap tbo poaoa raopoaolblo for tblo bioooavoroioa. SON and materials laolatioa aad Caltaro Motkodi All oaltaroo woro prova aad aolatolaod oorobicolly at )0*C la o pyrotary obabor oa a alaorol aalto aodim eoatoialap 1 ap/al of blpboayl oad aapploaoatod wltb 0.009% yoaat oztraot. Stroia 1150 vao laolatod frca o loadfill ooatoiaiap P(S oeataaiaatod drodpo opollal Approzlaatoly 5 p of ooll vao laeabatod la SO al of tbo oaltaro aodla dooorlbod above. Aftor approzlaatoly 12 traaoforo ladlvldaal ooloaloo woro lao latod oad okaraotorliod. Poar otraiaa of Aloallaoaoa aatrovbaa. oabibitiap oaly allpbt dlfforoaooo, woro ioolatod frca tbo oarlobaoat oaltaro aad It io oao of tbooo, otrala 1150, that vo daaorlbo borola. Tbo atraiaa of Aloalltoaat uttMlai voro parobaaod froa tbo Aaoriaza Typo Caltaro Collootloa (loabvillo, M>). Tbo typo otrala, ATCC 1707, baa boom oztoaalvoly obaraotoritod for diaoiallatloa of oroaatio ooapooada. (It) wbila tba oaooad otrala, ATCC 19597, io obarootoriaod oaly by ita ability to proa aatotrophloally. -102 NONS 214BS8 Preaaratloa of laetlna Celia Celle were |rova overaliht aa bipheayl, thea the cryatala of bipheayl war* r a*erred by flltretloa throa|h |ltn wool. Tba cillt vara harreated by eeatrifalatioe (3000 rp>) far 13 alaataa la a lonrall Oladal B0-2B) aalmj aa >534 retar. Tba aalla vara vaabad tvlaa la 0.03 k aodim pbaapbata baftar, ( 3.5, aad reaeapeaded la tba aaae baffar to |ive aa opt leal daaalty af 1.0 at 015 at. PCI Dairadatlaa Aaean Tbraa dlffaraat aabatrataa vara aaployed la dapradatlaa aaaayat iadivldaal PCB eeaieaere, a ayatbatla alztwre af eeveePCS aaapaaara, ar Areelor 1241. A 100X ataak af tba ayatbatla alitara vat preparad la aaataaa fra 1,4,4' aad 3.4.2' trleblereblpbaayla (1.20 p|/al each), aad 2,1,3,*; 2,5,2',5'; 2.3,2',3'; 2,3, *', 4'; aad 3,4,5`,4' tatraablereblpbaayla (1,4* |i/al aaab), Tba PCI ataak aalatlea la aaataaa vaa aired vitb tba call aaapaaalaa ta yield a flaal PIS eeeeeatretloa af appreriaately 10 ppa. Tba Areal ar 1241 aabatrata vaa added la (0 |il af aaataaa ta a 30 al eall aeepoeaiea, ta |ive * Oul eeaeeatratiea af appreriaately 10 ppa. After tberevdb airlay, a 10 al allaaat vaa raaavad aad traaafarrad ta a tvbe eeatalaiap 0.1 al af 70% parablaria aaid. Tbla tlaa aare (T#) aaatral aaapla vaa aired tharampbly aad freraa far later aaalyaie. Altaraativaly, aaltabla aaatrela vara prepared by beat iaaativatiaa (70*C, 20 alaataa) ef the eall aaapaaalaa prior ta iaaabatiea vitb P9a. Beth beat aad aaid pretreataeat ef eall aaapaaaiaaa daatroyad eetivity apint all Araaler 1241 PCS aaaiaaara. Tba raaaiadar ef tba eall aapaaaiaa vaa iaeabeted far 3 daya at 30*C vitb ebakiai, tbaa aaidifiad aad freeaa aatil aaalyaie. Vbaa iadivldval PIS aaaiaaara ar tba ayatbatla alxtare ef PIS aaaiaaara vara aaad aa a aabatrata, tba abare preeedara vaa aaad azeapt that iaaabatiea vaa Halted ta 24 bear*. M iiUlltiPl A C1#-fap-Pak aartrid|a (Vatere Aaaeaiataa, Milford, KA) vaa aativatad vitb aatbaaal aad tba aaidifiad ealtaraa applied ta tba eartrid|e by irarlty f litretiea, The aaaaaaa filtrate vaa diaeardad aad tba PISa vara elated frev tba oartrld|a vitb aatbaaal. Vbaa Araaler 1241 vaa aaad aa a aabatrata, tba aetbaeel artraat vaa aaad far |ee ebroaateirepbie aaalyaie diraatly vitboet farther aztraatiaa. Vbaa tba ayatbatla aoataaar alztara vaa aaad, vatar vaa added ta tba aatbaaal (fiaal aaaaaatratiea 25%), aad tba PISa vara raaztraatad vitb harare. 4-103 HONS 214889 Iki etltuit tin tij utid to pi 1-2 with perchloric told tid extracted twice with oae lil( toIwo each of aahydroai ethyl otkor. Tko extract wac dried, reaaapeaded lo O.C al of tetrahydrofuia. ood der'fltlxad *ltk 0.4 al of N,0-bla(trlaethylallyl) aoetaplde (ISA. Pierce Chao 1^*1 Cowpaey, Kookford, XL) for 30 laotoo at *3*C. 71a KSt aad tka allyl aatora of tko PC* depredatlM prodaot* vara dotootad aalap poo ohroaetopraphlc analyili. flu fttiiim> AmItiU laaplaa ooatalelap tka ayatkatlo fd alxtare vara aaalycod oa a levlett Packard Oao Chroaatoiraph (Medal 3S301A or 3AS0A) fitted vltk aa eetoaatlo lasplor aad aa aloetroo oaptaro dotaoter. A place col me (C ft x 4 aa) packed vltk 1,3% SP2230/1.9% SP-2401, oa 100/120 lapleoopert (Sapoloo, lao., Ballefoate, PA) woe need, Tko earner paa vaa 10.4% aothaaa la arpoa at a flea rata of <0 al/ala, Ckraaatoprapby aa raa leothamally at 192*C for aaalyala of PQi aloaa. For aaalyala of POa aad aatakolltaa, tka taaporatara *aa laaraaaad fraa 1(0*C to 240*C at 3*C/alaata aad kald at 240*C for 10 alamtaa. Aroolor 124* aaaplaa vara aaalyaad oa a Varlaa 4*00 paa okraaatoprapk nolap a faaod alllaa capillary oolaa (1 aad V, 30 a x 0,23 aa Id; IB #109*7). Tka llqald pkaaa vaa DB-1 aad tka film tklakaoaa vaa 0.23 p. Tka aarrlar paa wee keliw at a flaw rata of 30 aa/a. Tka aeke-op paa vaa altropoa at a flev rata of 23 al/ala. Orok typo lajaotloa vaa aaplayad (2 pi lajaotloa alxo). Tka eolai vaa raa at 40*C for 2 alaatee, tkaa tka taaporatara vaa laaraaaad to 00*C at a rata of 10*C/ala aad flaally to 223*C at a rata of C*/ala. Tko taaporatara vaa kald at 223*C for 10 alaataa. Tko data vara aterod oa dlaa aalap a Varlaa Vlata 401 loriee Bata lyataa aad raplottod to plva a direct aoapariaoa of aaok axparlaaatal eaaplo (T^) vltk tka oerroapoadlap aoatrol (Tc), Tka aoatrol to exporlaoatel ratio vaa oalaalatod for aaok of tko 13 PO poaka. Tko aaelpMaate far tka Aroolor 124* poake vara dataralaad by K.K, Vapaar aad J,C. Caraahaa (Maaaaarlpt la Preparatloa) bated oa pebllakad ldeatlflaatloaa <2) aalap paa ekraaatoprapfcle aolMaa vltk aaaparabla apoalflaatleaa. Filial* Pirllloatloa aad Aaalvaia Plaaalda voro daaeaatratod la 1*30 aad Ml aalap tka protoaol deeorlbed by laaaaa aad 01tea (2i>. Celia voro preva oa blpkaayl to aa optical daaalty of 0.4-0.* at *13 , filtered to raaewa tka blpkaayl aad karvaated by ooatrlfapatloa, Pellov- 4-104 MGNS 214890 lap a brief dlpaatlea wltb lyaotyaa, the cell* were lyaad wltb 2,3% aodlaa doda- eyl lolfata dmrlnp a rid of beat palaaa *t JJ*C. Cbraaoaaaal DM tad KNA ware ptrtUlly rworil b7 lUtllu liutiritlM. Cbraaoieaal DM va* raarid by dim dcdicpl iiltiti prtclpltitlcii til tbe pi*mid DM lm tba iipitutut * nicntfitid bp pxaclpltatloa wltb polyatbylaaa plyeol, -- Plaoeld DNA * aatlyrad by alaotropboraal* ! 0.79% aperoaa bortseatal (lab (1 apparataa. Tb* baffar aaad waa 19 all Trla, pfl (.3, 19 all barla aald, 2.3 all dlaodlm EOTA. After atalaiap wltb atbldim brealda, plaaalda aad DM frapaaati ware wlaaallcad by * DV traaaillmlaator, Hlmd III frapaaata of pbapa laabda DNA ware aaad aelaamlar valpbt aarkan. For growth atadlaa, blpbaayl waa parabaaad frea Aldrlab Chaaloal Co. (Mllvaakaa, VI) aad aoaoobleroblpb*ayla ware parabaaad frea Ffalti aad Baaar, lao. (Stoatord, (T). For dapradatloa atadlaa, aalytloal prada 9(Sa ware parabaaad fraa Aaalab*, lao, (North , CT). OSalaolw prada beaaaa, aatbaaol, aad aaatoaa ware parobaaad frea M(S Naaafaotarlap Cbaalcala, Iaa., (Claalaaatl, Qi). Naaoprado baiaa * parabaaad frea tba Aldrlab Cbaaleal Co, Baatrlctloa aadoaaolaaaaa vara parabaaad frea Naw Eaplaad Blolaba (Bawerly , MA) ad aaad par tba aaaafaatarar'a laatraetloaa, Blaotropboraala prada aparoea (Standard Low af) waa parabaaad frea Bio-Bad Laboratorlaa (Blobaoad, CA). USDLTS AND DIICDMIOM liolitloa aad Taaoaaala Cbaraatarltatloa of Baetarlal Strata* Strata BI90 waa laolatad by blpbaayl aarlobBaat frea a laadflll ooatalalap FCB aoataalaatad dradpa apolla aad aaat to tba Aaarlaaa Typa Cal tar a Collaotloa (ATCC) (loakvllla, MD) for ebaraotarlaatloa. It baa boaa Idaatlflad Aloall- ItMl lUHrtM aaaordlap to Bariay'i llilMi Dataralaatlwa Baotarloloty (U), Tba aalla vara praa aapatlwa ooaaal roda oeaarrlap alaply, la pair*, aad la balaa, aad aotlla by parltrlaboaa flapallatloa, Tba taata for oaldai* aad atalaaa vara poaltlwa, vbaraaa taata for bydrolyala of oaaala aad palatla ware aapatlwa. Tba orpaalaa la aapobla of prevtb batvaaa 23 aad J7*C bat aot at 4I*C aad aatabollaa la raaplratory, aawar fraaatatlwa. A vide variety of orpaala eeapoaada aaa ba atlllaad aa (abatratai by atrala BI3Q laaladlap fatty aald*, dlaarborylla aald*, bydr 0x7 aalda, aoa-altropaaoaa arcaatla aeapeaada, aad ali- pbatla aad araaatla aa lao aalda. Kovawar, fav aarbobydrata* aad ao paatoaa* ar dlaaaabarldaa era atlllaad. Frvotoaa, bat aot plaeoaa aaa aapport provtb of tba*a 4-105 MONS 214891 nil aaaaaal patters ii characteristic of AlaaUttm oatroohaa. This la the first report of growth oa blphaeyl or deiradetloa of PCB by tbla apoolaa although tbara have baaa several roporta of aaabert of tba teams Aloallraass with thoaa oboraotarlatloa ttrala Mil was lsolcied fro* a atoak of Aalaatebaotar ap. Pd (U> aad aabaaqseatly obaraetarlaad by the Aaarloaa Typ* Caltura Collaatloa, It tea boom olsaalflad aa bsloa|ia| to tka (aaoi CorvaabaotarlM baaad oa tka follovlai arltarla: Tka oalla ara some 111a, praa positive plalcaorpklo roda aad are aatalaaa posi tive bat oildaae aetatlve. Taata for ladola prodaetloa, aaaala aad palatln hydrolysis, oltrata atlliaetloa, aathyl rad reeotloa, aad Votsa-Proskausr reaottoa vara all aapatlTt. Vkola oall hydrolyala deaoaetrated tka praaaaaa of aaaodlaalaopiaelio aetd, a akaraotarlatlo typical of tkla paaaa. Only two otkar praa poaltlwa eriaatai kawa baaa akowa to daprada Kka. Tka flrat, Howard ii ap. NCIB 10603. vaa reported bp tester at al. <i>, Tka aaaoad waa laolatad aa a oaataalaaat of Aolaatobaatat ap. Pd aad vaa ldaatlflad aa Artkrobaatar ap. NS (JJ). Tka okaraotarletlea of Artkrobaatar aad Corviebaotatlv ara wary alallar, bat these leaera aaa be dietlatolthed bp sail wall asalpale. To aaaaot rale oat the possibility that tka MS strata of Farakava aad Chakrabartp (Jl> la Idaatloal to CorvaabiBtarlt ap MB1. Orovtk oa ItoaoaklotoblakaaTla loth Ml aad 1110 |r*v road Up oa blpkaayl, Corraabaatarii ap. tel |rav aodarataly vail oa 4-ehloroblphesyl bat aot oa 2** or 3-ohloroblpheayl. Aleallaaaaa aatroakaa 1150 atlllaad all thraa aoaochloroblphaayla aa a provth aabatrata bat to dlffarlap degreea. Orovtk vaa beat oa 2~ohloroblpbaayl, aodarata oa 1ohloroblpkasyl, aad poor oa 4-chloroblpheayl. Ta did aot taat tka ability of altkar of tkaaa atralaa to atillaa hitter oklorlasted PCB aoapaaara aa irovtk Daaradatloa of Aroalor 1241 Kaak aaltare vaa laaabatad vltk tenoral al trade Aroalor 1241 to deteralae Ita ability to deirade PCIa. Fitara 1 above |aa akraaatotrapka (O.C.> of Aroalor 1241 before (ooatrol) aad after a three-day laoabatloa vltk 1110 or Ml. Tka 41 ooaiesera la tka alrtare rea|e (ra di- to kaaaakloroblpkaayle aad ara resolved lato 11 peaks. Tka eataaalva akaapaa la tka Kl profile doaoaatratad by tkla ay ara ladloatlva of dapradatloa aad aot seleatlve adaorptloa for tka follovll raaaoati [1] la all aaaaa both tba aall aeaa aad tka aqaeoes pkaaa vara rtraetad. [2] Tka Aroalor 1241 profile recovered by axtraatloa of tka Ida 4-106 HONS 214892 DEGRADATION OF AROCLOR 1246 BY RESTING CELLS tttt t t ft t t Tfll TETflA PENTA HEXA P1|W* 1, Minlitlu of 1241 kr 'l*otla| CtlUi Oil* tin |lovi ii kiA'TL (uM> ii4 la fkufkiti Mtt. AiMln 1241 voo added te tiaal ir- eeatrotlea of 10 f>a. Ike itftl iksti uMllity |ii ikfiiti|fi|k of tke KS* estreated ftea n of oil* eeldifled inedlately otter oddltlea of the KSo. Hi alddlo ii4 lever feaale ikn tki KSi estreated frca tki SI u4 1150 ooll aaofeaeleaa of tor S daye laeako- tlea. frca oollo oo!41fio4 ot I(, lt (ta keat-iaaetiToted oollo ot T#, or of tor S deyo laeaketloa, lo idootiool te tke profile of aa Ax color 1241 otoadard. (11 A aatoat otrala of Ml tklik kae loot tke okllltj to (roe ca klykearl daoo aot alter Arcelor 114t> Siollarlr, two otrolao of Alooliooaco ittkn vklok do aot tree ca kipkeayl (ATCC otralao 17407 oad 2*5*7) kero ao offoot ca drooler 1241. [4] A roller color vklok lo icaarollr tokoa to ke ladleatlve of Kl aetekoliaa 4-107 MONS 214893 r Tabla 1 SHMDATION OF AKOCLOR 1241 BI BESTING CELLS Paak Usfeix PCS Coataaar ItSUUlf Aaalr--aat *1 Blit 1 2,3.2' 2 2,4,2' aat 4.4' P+ f f 3 2,3.2' aat 2.6.4' 4 2,3.4' 3 2.4.4' C 2,3,3' p f f f f P f 7 2.3,4' aad 2.4,2'.6' P f 1 Tatra p 9 3.3.3' f 10 2,3,2*.S' f 11 2,4.2'.3* aat 2,4,6,3' 12 2,3,4,6 + + 12 2.4,2',4' f + 14 2,3,2',3' aat 2.3,2',4' IS 3.4,4' f + 16 2,3,4,2' aat 2,3.6,4* p 17 2,3,2'.3' + 11 2,4,3,4* f 19 2.3,3'.4' P+ 20 2,4,3'.4' aat 2,3,6,2',4* Pp aat 2,3,6,2',3' 21 2.4,3,2'.6' 22 2.3,4,4' aat 2,3,4,2`,6' f p 23 2,3.6,2',3' aat 2,3,3.2',3' p aat 2,3,3,2*.4' 24 3.4,3*.S' aat 2,4.3,2',3' p 2 J 2,4,3,2',4' p 26 2.4,3,2'.3' Pp 27 2,3,4,2',3' p 21 2,3.4,2',4' 29 3.4,3',4* aat Paata p 10 2,4.3,3',3* + 31 2,4,3,3',4' aat P 2.3,6,2',4',3* aat 2,3,4,3,3' aat 2,3.4,6.2'.3' 32 Feata 33 2,3,4,2'.4',J' liii tikli riymiiti it iiiliUu! lulfiii af tk 4a t iktii la Fl|ra 1. Meat of tka ni|iun la kraalor 1241 Mere beta itittiiiii ii4 taatatlvaly a*l|a*4 te peaka la tka akreaatafrapk ftatila (J.C. Caraakaa iat t.K. Ta|*ar, Naaiaatlft ia Praparatlea) leak arpariaaatal aaayla vaa eaHtai ta ita na (T ) ooatrol lit tka rataatlaa la aaak tail kaijkt vaa aaat ta aaera lainlatiaa 2a altkar partial (JO-SO*; p) at aeaplata (> 10%, + ) oa tka beale of tka ratio of tka paak araaa, la aeaa aaaaa tkla oeaaarratlva aaerla* aatkot aap lira aateraatita4 tka aitaat af li|rilatlia, partlaalarlp vkara a alapla paak rapraaaata aevaral POO aoataaara. 4-106 NONS 2l*g9* (^;.9.13.1<"11|) appeara 1b Titbit emit art bat mot lm tba control a. Tbarafora, tbaaa ebamiaa camaot ba aiplalmad by aalaotira adaorptioa to tba oalla or tba plaaawara. Fiaally, tba fact that omly oalla abicb eaa mtlliaa bipbenyl altar tba PM proflla la fairly atroap arid# tee that tba obaaiaa ara dma to microbial dairadatioa. - T*a otbar poiata ara immediately apparaat; [1] both Ml amd ||M daprada tba commercial PCB mlitmra aitamaiyaly; Ml attaaba approximately 45% af tba P<3 oompamara amd 1150 dapradaa approximately <51. Im aaab aaaa Kla ablorimatad at ap to fiya poaitioma ara dapradad. [2] Tba depradetlre compete tee of tba two cal tor a a ia qmlta dlffaraat. Altboapb Ml amd 1150 daprada mamy of tba aaaa PCB oompamara, aaab atraia ia amparior to tba otbar ia ita ability to daprada eartaim oompamara, Tbaaa diffaramaaa ara imdiaatad by arroaa cm Flpmra 1 a ad ara awaa mora apparamt wbaa tba data ara tabmlatad (Table 1). tiiiilittfi ri-tattiUf, Ulitm pl M taumti Emaomrapad by tba abort raamlta, ra dor load am aaaay at lap a lymtbatio mlitmra af aaraa para PM oompamara to foama mora aloaaly om tba diffaramaaa im tba dapradatira oempatamoa of Ml tad 1*50. Tba raamlta, abeam im Table 2, wart atrlbimp. Table 2 DMBADATION OF A IIKHETIC XIX30U OF PM CONQINEKE PM Comxamor *1 1150 2.4,4' 55% 55% 5,4,2' 51 51 2.5,2',5' - 100 2,5,5,< -- 2,5,2',5' 2,5,1',4' 52 - 54 51 5,4,5* ,4' - - Tbia table abort tba paraamt dapradatioa of aaab of aaram PM aempamara of a aymtbatia mist are follorimp imaabatiom for 24 b with raatimp oalla of Ml or EI50. laab aompamar aaa pro teat at 5 Mol/ml (approximately 1.5 ppm). Paraamt detradetiom aaa aalamlatad by oemporimp tba area of aaab O.C, pool lm tba aoa- trol (T ) amd axparimoatal (T) (ample i. Tba ralmaa plraa ropraaamt tba m*am of two Xoparato axporimamta. 4-109 NONS 214895 Apaia, the two cmltarei deaoaatrated ideatioal owpeteace for aoa* p(J ooapeaera, two woro aot degraded by either atraio ood two waro readily degraded by both cultarea, Bowewer, HI20 completely di|radad two tetrachloroblpheayla which waro aot atabeliaad by MB1. Ia ooatraat. KB1 degraded 2,4,4' triehlorobiphaayl more raadlly thaa HI30. It la altar that both tha poaitloa aad tha aw bar of ohloriaaa art oritioal la deterwiaiag whathar a portlaalar PCB eoagoaar caa bt dagraded. Thaaa data lad at to paeatloa wharhtr tha poaitioa of tha ialtlal taayaatlc attaok oa tha arwatlo riap aipht dlffar la MB1 aad 1150. A awbar of leberatoriaa hawa poatalatad pathway! for PCI (2.2J..1&) or bipheayl (24-24) dapradatioa. Ia aaah eaaa, tha ialtlal eaayae ia thoapht to bo a 2,2~diocygeaaie. Sabeepaeatly, aete-eleawage oaoara batwaaa carboa attai 1 aad 2 (Fipara 2) (2.14.24) 2 3 Dihyaroiy Biphenyl Mala Claavapa Product CMorobanroala Fipara 2. Propoaad pathway for dapradatioa of PC)a by a 2,1 diecypaaaaa (2, or botwoaa oarboa atwa I aad 4 (21-11.), Ia thaory, a diecypaaaaa ooald alto attaak at poaitloae 1,2 or 2,4 (Fipara |1), Aa dapiatad ia Fipara 1, tha poeitlaa of tha ahloriaa atwa ia 2,2,2',2` tatraahlarobiphaayl woald Boat likaly hiadar tha approaoh of aithar a 2,2 or 1,2 dlacypaaaaa, bat oath of theao aaxywaa ooald raadlly attaak 2,4,4' trlahlorobiphaayl. Tha ooawana ia traa for a 1,4 diecypaaaaa. Pabatitatioa ia both para poaitioaa woald alaarly lapair tha approaah of a 2,4 diosypeaaee, bat 2,2.2',1' tatraahlorohiphaayl woald bo readily aaaaaaibla to thie eaayae. The diffaroaooa la daprmdatlwo owpotaaaa of thaaa two orpaaiwa dMaoaatratad ia tha 1241 aaaay aapport thia BOdol. Qeae rally Ml ahowi Boro actOBalwa dapradatioa of peake whioh laalada aoapaaara awbatitated ia tha 2,4,4* aoaf ipcratioa (Table 1, Pa aka IP,22,21) wharaaa aeay of tha paaka whiah are dapradad aora readily by 1220 iaaloda ooapaaara aabltltatad i 4-110 HONS 214896 Fi|r 3. Alloaad ftiitiMi of ttttek of ialttal ficifiiuH. FimI A show* tha dihydrodiol biphenyl iatamadiatae that weald reaait fres eaayaatie attack: of a dloxyieaaee at earboa poaltieae 3,4 (toy), 2,3 (aiddla) or 1,2 (bettM). Panela B aad C akn tki oerroapoadlai lateaedlatea for 2,4,4' ttlikUtikifkai;! a ad 2,3,2',3' tetreehlerebipeayl. Dotted llaoo ladlooto fitkiaja that aro theoreti cally peeeible bat tkiik km aot booa dMoaotreted. the 2,3,2* ooafl|aratloa (Tabla 1, paaka 10,11,23,24,23, aad 27). Baaod oa tkaaa oboe rta float, to propoae that tko laltlal lainfathi eaayae la Aloalkaaaaa antroehaa la o 3,4 diexyieaaee or poaalbly a 1,2 dlexyieaatt, Potradatloa Prodaota of ialootod-PO Coataaota To km reneatly bogaa to otady tko aotabolltoo of aolootod PCI ooafoaaro with tko altiaate |oal of olaoldatlat tko PCS depradative pathway la took of tkooo baotorla. Oar prellalaary fladlaio era ooaolotaat with tko aodol we have proyoaad. Tvo PCI ooagoaera whlek aro oabatltatod la botk port pooltloao, 4,4' dlokloroblpkoayl aad 2,4,4' trlakloroblpkoayl, were oxtoaoivoly aotabolliod by Ml ylaldlaj a loo aoleealar wei|ht aetabelite wltk tko aaao O.C. rotoatloa ttao ' aa d-ehloroboaioie aald la tko flrat oaao, aad lario aaoaata of foar dlfforoat ki|k aolooalar weight latoaiodlatoa la tko latter. la ooatraat, at ooald flad ao ovldaaeo tkat IS30 aotabollaod 4,4* dlokloroblpkoayl aad at datootod oaly Moll aaoaata of aotobolitoa fraa tko trlakloroblpkoayl. Coaroreely, IS30 prodaood ataaolao aatabolltaa {( PCS eeaiaaara aabotltatad at pooltloao 2,3,2' or 2,3,2',3' aklla Ml prodaood oaly Moll MOaata of dagradatloa prodaota ttm tkooo ooaioaara. Partkomoro, oltkoaik iai of tko aotabolltoo poaorated aoro oaaoa to botk ortaalMt, la aaay oaaoa tko aojor aotabolltoo prodaood by KB30 4-111 HONS 214697 *j , , , 1 DEGRADATION PRODUCTS OP 2,5,2' TRICHLOROBIPHENYL ( CONTROL Fl| 4, Di|tt4itiM r4ut if 2,S,2' littiii c*ll# iti iukiti4 titk 2,5,2' (50 pja) ( 21 k; aitabilltia tul PCS# wti n4 ih1tii4 bp ( cktaiti|t>rk7 ii 4iHtlk4. Ik* appir huI ikctl tkl llttlltill if kHt-iUltiTltl4 Mill vkilk till luikiti4 w4it tki hm niiitlni Ml i>4 1250, Tki nl7 fi*k ii tkl PCS, Ik* lint fiuli ikn tki PCS u4 aitabilltn ittriti4 frca Ml (lift) u4 1(50 (ri|kt>, la Mil tki Hint of PCS kaa 4iirim4 ii4 tkrii aetabilltii B.l. m4 P kivo ippiarK. Mitabillti I la ant likilj 2,5 4iikliribiaaoati. la 1150 tki PCS kai koia iwplitilp aita- billiiA, ippiriatlr ti 2.5 (likliribiaaMti. At liaat flvi aitabilltn 14iatlfii4 A-l, kai koia prolan 4. Mita- killtn S ami ktvi tki iaai ritiatlia tiai fir kitk kai- tirlal itrilaa. ulai, Aa isaapli of tkli li ikoia fir 2,5,2' trlakloroblpkiapl la Pl| 4. Tki if tki aitakilltii, B (aklik kn tki ritiatloa tlai * 2,5- (liklirikiuoati) ii4 1, win pri4aii4 by kitk Ml ia4 1(50. loairir. Ml pro- 4aii4 ill noaat of aitikillti P aklik an ait ma la 1(50, aa4 1(50 pro- 4n4 lr(* woaati if tki aitikilltn likilli4 C aa4 0 aa4 a laa .ailioalar aiiikt aitaboliti I4iatlfli4 Puk A. Noriorir, tki 2,5.2' Mapcair aai 4-112 HONS 214898 plitilj utikeliui >7 IIJO. iffimtlr to 2,5 diekloroboaioato. That tko dot* tapport tki kppotkooio tkit tki aijer roato of PM aotabeliaa diffort la Ml and H(50 tad art eoaolatoat oltk bb laltlal attack 7 a 2,5 dlexytOBaao la Corraahactari ip. Ml aad a 1.4 dlorystaatt la Aloalltoaat omtgoakat 1550. V# art la tki prooooa of idoatifplas tkeao aetakolltee by paa tkratosrapfcp oaablaad oltk aaoe apootraaotry la aa effort to olaoldato tko aajer PM dopradatlvd yatkaapi operative la BB50 aad Ml. Vo kavo aloe ba|u to lavootlpato tko (oaotloo of PM blodoiradatloa la Ml aad 10, Hoot (oaotlo laforaatloa la kaotorla rooldoa ea tko ehraaooaao, bat froqaaatlr a olsalfloaat aaoaat of poaotlo laforaatloa 1* aaaoolatad oltk aa oxtraokraaotoaal pitot of UNA ealltd a platald. la foot, plataldt art tko prlaaiy aoaao of (taotla oaokaaio la kaotorla, TJrpieally, pitta Ida karkor (oooo oklok oarry aoaoaooatlal bat aaofal laforaatloa oaok aa aatlklotlo or aotal roilotaaeo. Mart iaportaailp, tko dopradotloa of varioao araaatle hpdrooarkoat (tolatao, tal~ loplato, aapktkaleat) (21,li) aad ovea oklorlaatad araaatloa (ekloropkoaele, oklorokoaaoatao, oklorlaatad pktaoxpkorbltldtt) (22>1,H) 1* of too aooeolatod oltk plaaa Ida. Fiaallp, or Idaboo It aooaaalatlaf tkat tko |oaat raapoaalklo for PM doiradatloa aay rotldt ea planldt (2i>, Tktrafort, ti ooro aarloaa to loata If Ml aad 11)0 karkorod plataldt. At tkooa la Pl|aro 5, tkoro It a olaplo lotto platald la 000k orttalta, Vo iaediatolp aoktd okttker tkttt plataldt otto Idtatloal, ty atla| tptolflt eaapaoe ealltd rootrlotloa oadoaaoltaato to oat plaald DNA at opeelfla polata, oao eaa toaaroto a tpeolfle frarest at ioa pottara oklak la aalqaa to tkot platald. If too plataldt art Idtatloal, tkeir rootrlotloa pottoraa will ko Idtatlool. toatrlotloa aaapaee ooro atod to taalyao tko plaaaida frta kotk 1150 aad Ml (N. Broaaaa, data aot ikm, aad J.I. Tatoo, porooaal teaaaaioatloa). Tko rataltt olaarly iko> tkat tko plaaalda art aot idtatloal la tko too or(aaltat. loooror, it la alao pooeikle tkat tko too plataldt art aot oaapletely dlfforoat, kat kata at toaMoa Seattle laforaatloa. Oa* prolialaary fiadias* atlas krkridiaatloa aaalpeie ladloato tkat tko plialdi la tkoaa too kaeterie da aot tkaro tap oeaaoa |tatt (M. Brtaaaa, data aot ekooa). Vo do aot at pat kart tap firs orldoaoo tkat tko PM dopradatioo pothoap it aotoalatod oltk 0 platald la oltkor Ml or 1150. Vo do kaoo, kooooor, tkat for BB50 tko akility to sroo oa kipkoapl (aad proawoklp alao tko ability to doprada PMi) io readily loat if tko oaltaro it aot aalataiaod oa oolootioo aolla oltk 4-113' HONS 214899 fr OwR 1 Ii Figaro J. Lit|i |UaU la IMO 1*1 Ml. PlMll DNA tu ieolata4 tt (iittUil ill lulT'il bp (ret# |*1 eleetreghaaecle, Ik* 41raotloa of alaetropheroiia ! fta top to batta. Til flrct Im ahowa lla4 III (ripnt of phage la!4i ONA hi4 aa acliiiUt tiiflt aarkra. Tli paeltloa of tki plaaal4 DNA la la41aata4. Tha 4iffeaa baa4 of DNA which aigratea lightly behia4 tfea icrgoet laaMa DNA f repeat la iktailatl DNA. Tha largo haa4 of DNA which algrataa batwaaa tha oktMiaail tad plaaal4 DNA at llkaly aaatalaa plaaal4 DNA la a 41ffaraat phpalaal fata. blphaapl aa tha ala aarhoa aoaraa. Thla type of geaetie laetebllltp la aeat freqaaatlp aaaalata4 with plaaaU aaao4a4 faaatlaaa. la tha fatmre ia pica ta traaafar tha plaaal4 frcs Alaallaaaaa aatroahmi IgJO to that Alaallaaaaa aataaahma atralaa at wall aa ta other bacterial tpaalat which arc aot capable of growth aa blphaapl or 4a(ra4atlaa of PCS*. If tha raalplaata aaqalra tha abllltp to 4a|rc4a PCS a aa a raaalt of tha plaaaM traaafar. we will haw a ettabllake4 that PCS 4agradatlaa la caatrcllaA bp a plaaai4 la Alccllaaacc aitraahma 1(30. 4-114 HONS 214900 r r Fiaally. *a in omutlT aalag rtkluit DNA tickiiqui tc olea* fra^aata of (JO DNA u a firat atop toward Idaatlfylag aad okaraatarlalag tka gaaa(a) apaoi- fyiag tka iu^ia of tka KB dagratatlva pattern? la tkia atraia. Tkla >111 allow aa to dawalop aa aadar ataadlai of botk tka atraataxa aad ragmlatiaa af tkaaa aaiyaaa. - SDKMAIT Va kava laolatad a ad okaraetarliad two dlatlaet orgaalaaa. Aloallaawaa aatroahw (JO aad CorrwabaatarlM ap. Ml. wklak aitaaaival7 da grata Aroolor 124(, bat wklak axkiblt alaar dlffaraaaaa la tka partlaalar ooagoaara wklak ara dagradad. la gaaaral, PCI eeagaaara with tka 2.4,4' aakatltatloa pat tara ara aatabolliad rapidly by Ml. bat poorly by 1(30. la eoatraat, aaay ooagaaara aabatltatad at poaltlaaa 2,3,2' ara rapidly dagradad by 1(30. bat aot by Ml, Tka prodaotlaa af aatabolltaa aaofwlvoaally aatabllahaa tkat tkla la ladoad dagradatloa. tea da grata tl oa prodaata, iaaladlag aklarobaaaoataa, aro aoaaoa to botk orgaalaaa. * bat la ate aaaaa tko aajor aatabolltaa prodaaad by MJO ara aaigaa. Tkaaa data ara aoat taadlly aaplalaad by poatalatlag tkat tko aajor pathway of FCB aotabol* laa la IBJO atlllaaa a dloxygaaaaa wklak prafaraatlally attaaka at aarboa poaltioaa 2,4. If tkla la tka aaao. It will ba tko firat daaoaatratlaa of lawolva* aaat of a 1,4 dloxygaaaao la tko blodagradatloa of blpfcaayl u FCBa, la aowtraat. Ml probably aaploya a aora aoaaoa 2,t diary (aaaaa aoakaalaa. Flaally. tka two orgaalaaa karbor dlatlaat plaaalda wklak aay ba lawalwad la FC> dograda* tloa. To ara mow la aa axaallaat pooltloa to bagla roallalag oar loag*tatm goala, aaaaly, aadarataadlag tka bloakaalatry aad gaaatlaa of baatarlal PCS dagradatloa. ACZNOTUDQMNTS Ta tkaak J. C. Coraakaa, t,l, Vagaar, J. teltk, aad t. Kohl lag for gaa iktaatographla aaalyala, T.J. Tofflaalra aad 1,2, hrooka for kolp la oollaatlag PCS aor taaiaatod aavlromoatai aaaploa, 1, Farakawa aad A.H. Ckakrabarty for a at oak of Aalaotobaatar ap. N. I.l, Flakbolaar, aad T.M. (a for kalpfal dlaaaaalaa*. aad I. Sorooa for typlag tkla aaaaaarlpt, unwtxi 1. Alartadar, N. IMF. loll Biology, (Barlaw* of aaaarak, Natlaaal Baaoaroaa aaoarak. DNBB00) *:20F-240. 2. Atead, N. aad (,t. Faokt. 1F7S. Dagradatloa of polyoklorlaatod blpfcaayl* by two apoalaa of Aokraaobaatar. Cob. X. Nlaroblal, Jl;47-32. 4*115 HONS 214901 rr I. Barter, I.A., l.E. OUbort. It.A. Lid|*tt, T.l. Ndaprli*. *ad I.A. Voddoa. 1P7S. The d*|r*d*tlea of polpoklorlaatod biphsapl* bp alcreortaalta*. Sol. Total Eiriroi, :JJ-dl. 4. Trip, M. 1b.M., E. Ickoitx. tad 0. Bt*la|*r. 1*71. Tba baotorlal ottaboliat of 4,4'-dlokleroblphsapl, and It* *appr***loa bp sltoraotlvt otrboa tears**. Chaooophart A:10S-10B. 5. l**p, P.F. tad A.N. Chskxsbartp. 1*79. Flsnldt *po*lfpla| p-ebloroblpkanjl d*|redatloa la totorle baotorl*, pp. 275-215. Ia E.N. Tiaalt nd A. Fabler. (Id*.). Plataldt of Nadlcti, lirltoBiatil, tad Caaaorolal Import ease. Bitov lor/North Kolload Blostdlcol Pro**. Now Iork. 6. lolohordt. P.B., B.L. Ckodelek, M.A. Colo, B.l. lobortooa. oad D,E, tat tea. 1711, Uaotlo otadp of tko blodoiradatloa of blpkoapl tad It* aoaoeklorlaat*d oaalopao* bp * aired aorlao alorobltl ooaaoaltp. lavlroa. Bel. Took. 7. Bplvoatro, M. tad J, Faattaa. 1PB2. A aew faealtatlv* aaaorob* eapable of trovtk oa ekloroblpkoapl*. JT. Goa. Appl. MUroblol. 21 i<1-72. B. Koat, I.L. tad G.S. B*pler. IMS. poiradotloa tad total alaaralliatloa of aoaoh*lot*aat*d blpkoapl* la aataral **dla*at tad alrod baotorlol ealtar*. Appl. lavlroa. Mloroblol. l&i<M-<72. P. Parakawo, I. 1PB2. Nloroblal dotradatloa of polpoklorlaatod blpkoapl*, pp, SB-57, A.N. Ckakrabortp, (Id.), Blodoprsdatloa aad Dotorlfloatloa of Bavlroaoatal Pollataatt, <XC Pro**, Iao., Boor latoa. 10, Kaloor, K.L.I., aad P.T.B. loa*. 1P74. Baotorlol dotradatloa of polpohlorlaatod blpkoapl*. I. Idaatlf loatloa of *oao sotabollo prodeat* ft Aroolor 1242. Ball, lavlroa. Coataa. Torlool. Uilll-IN. II. Taokar, 1.5., V.V, too (or, aad 0. Blok*. 1575. Aotlvatod tlad|t prlaarp dotradatloa of polpoklorlaatod blpkoapl*. Ball, lavlroa. Coataa. Torlool. 11(705-712. 12. Clerk, t.l., I.B.K. Cbiaa. oad I.A. Grlffla. 1P75. polpoklorlaatod blpkoapl* bp alrod aloroblal ealtar**. Mloroblol. Ui<t0-dl5. Dotradatloa of Appl. lavlroa. 11. Tati, 0. aad I. Bade. 1550. Dotradatloa of polpoklorlaatod blpkoapl* bp aiorooriaala*. J. Bator Pollat. Coatrol Pod. 12,sl0B5-104t. 14. Baplor, 9.1,, M, Bkoa, aadl.t. Colwell. 1577. Growth of aa ottaarlaa IlUlmij op. oa polpoklorlaatod blpkoapl. liorob. 1**1. 1(241-255. 15. Lla, P. 1550. lakaaooaoat of PCI* blodo pr adotloa bp todlea 11 tala aalfoaato. Voter I* to orok 1:14<7-147S IS. Parakaw*. I., E. Termer*, oad A. Eaalbopathl. 1575. If foot of oklorla* tabatltatloa oa tka blodotradabllltp of polpoklorlaatod blpkoapl*. Appl. lavlroa. Mloroblol. Us22l-227. 17. Parakaw*. E.. N, Tnlaaka. aad A. Eaalbapaahl. 1575. Iffoot -of oklorlaa tabatltatloa oa the bootorlal aotabolln of varloa* polpoklorlaatod blpkoapl*. Appl. lavlroa. Mloroblol. |j|:S01-ll0. 4-116 HONS 21*902 r r 1*. Futkiii, I., N. Taiitkt- aad A. baikijitki. 1012. Baotarlal da gradation of folxcklcriutif biphaayla (Ml) ill tklit aitlktlitn, Adr. tip, Mad. Biol. 1141:407-411. 10. Jokatoa. B.F. aad K.T. Staalar. 1071. Dlaalallotloa of arMatlo oapouli bp Alotllrtaoi iimku. J. Baot. 107;441-471. 10. Iilliikiitit, I. tad M. loll. 1010. Aaalyala of polyohlorlaatod biphaayla by (lttt (ifUUtr cktaitoittpkf, politiPi of tookalaal Aroolor tad Clopkii - PCS alitmraa. Frttttlaa Z. Aaal. Cka, 101:10-11. 21. Itattt. J.B. tad K.l. Oltta. 107t. Ittlttloa of ltrao baotarlal pi a Midi a ad abaraatarlaatloa of tha P2 laooapt tlblllty iroap pltMld* pNB2 aad pNOJ. J. Baat. U2.:227-21 (. 22. CMBlaa. C.l.. I.A. Lai Hot. aad M. Bogota. 1074. Oaaaa 1. Corraabacpp, 402-417. la K.l, Baohaaaa aad N.B. Olbboaa, (Ida.). Borgay'a Maaaal of Dataniaotlao Baotariology, Bighth ldltloa. Tha lllllaa aad *11klaa Co., Balttaoro. 23. Farakawa, I. aad A.N. Ckakrabarty. 1012. InolvMoat of plaMida la total dagradatloa of ahlorlaatad blpbaarl*. Appl. Baalroa. Mlaroblal. 44:<1424. 24. Lwat, D. aad *,C. Bvaat. 1070. Thi altroblal aatabalin of blphoayl. BloohM. J. lU:I4p. 23. Olbaoa. B.T.. K.L. Kaborta, N.C. Valla, aad V,l, Kobal. 1071. Osldatloa of .blpkaayl by a oHorlaohla apaalaa, KlaabM. Blophyt. Kaa. Coaaaa. Ifi.lll- 210 24. Catalaal, D. > A. ColMbl, C. Oorllal, aad T. Traaaaal. 1071, HatabollM of blpkaayl. 2-By4rory-4-oio-4-pkaylha-2,4-dlaaota: tha ftftU~*l**T*l* prodaat fra 2,3-dihydroxybiphaayl by fatmdMoaat oat Ida. BlOabaa. J. 111:1041-1044. 27, Ckriatophor. F., 1. FraaUla, H. Bagdaaorlaa, aad K.N, Ttaala. 1011, Maalpmlatloa of da grad* tIta aaaaa of toll baatarla, pp 100-ISO. Ia T. Lalaiagor, 1. Battar. A.N. Cook, aad J. Naoaak, (Ida.), Mlaroblal Dagradatloa of Xoaoblotloa aad Kaoalaltraat Coapoaada. Aaadaala Praaa, Nov Tork. 21. Farroll, K. aad A.N. Ckakrabarty. 1070. Doiradatlva pi a aa Ida; aolaamlar mo tara aad aoda of ovolatloa, pp 07-100. Ia K.N TlaMla aad A. Piklar, (Bda.), FI a m Ida of Nodlool. laalroMoatal. aad Coaaaralal Iaptrtaaaa. Blaaviar/Morth-lollaad BlModlaal Praaa, Nov Tork. 20. Ckattarjoa, P.K., t.T. Ktllogg, D.K. Vatklaa, aad A.N. Ckakrabarty. 1011. Plata Ida la tko blodagrada tloa of ahlorlaatad arMatla ooapoaada, pp 510 321. Ia f.B. Lory, K.C. Clova a, aad K.L. Kooali, (Ida.), Nolaoalar Biology, Patkoioalolty, aad loolocy of Baotarlal PlaMlda. Ploaaa Praaa, Nov Tork. 30. PMbortoa. J.M., B. Coraay, aad K.l. Dob. 1070. Brolatloa aad apraad of poatlolda dagradlag oblllty aa| toll alorooriaalMa, pp. 217-200. Ia K.N. Tlaala aad A. Pvhlar, (Bda.), PlaMlda of Modlaal, SaviroMoatal, aad Coaaaralal laportaaao. Blaavlar/Mortk-Sollaad BlMtdloal Praaa, Nov Tork. 11. Ckakrabarty, A.M, 1012. Gaaatlo aoakaalMt la tha dlaalallatloa of ohlorlaatod OMpoaada, pp 127-130. Ia A.M. Ckakrabarty, (Id.), Blodogradotlea aad Dotorifioatioa of BaviroMaatal Pollotaata. <XC Praaa, lao., Boaa Katoa. (-11? HONS 214903 Discussion Isn't this procsss lieltad to a tpacific Aroclors? A. Ya?t the study to far has baan llvitad to elte-contaainoted with 12*1. Would naad to laolats elcrobacceria fro* other oltM to study other Aroclors. HONS 214904 f COMPOSTING FOR DEGRADATION OF PCBS IN SOIL Jeneflr 0. isblster Georg* L. Anspach Judith F. Kitchens Atlantic Reseerch Corporation Alexandria, Virginia 22312 ABSTRACT This study evaluated aerobic and anaerobic composting of PCB contaminated soil on a laboratory scale. Coagiosts were coagiosed of alfalfa hay, horsefeed and Lakeland sou spiked to contain 2000 mg/Kg of Aroclor 1242 (dry weight basis). Warai humidified air was drawn through aerobic coapost materials. Anaerobic congests were flushed with nitrogen. All coemosts were Incubated at 55C. After two and four weeks of Incubation, compost materials were oven dried, ground, subsaiapled and Soxhlet extracted for GC analysis. The aerobic coagiosts showed the greatest destruction of PCBs with 41.6 and 48.IS destruction in two weeks end 55.7 to 67.6% destruction in four weeks. PCB destruction In anaerobic coagiosts was 18.4 to 2B.0X after two weeks and 27.9 to 46.9X after four weeks of coagwsttng. Individual peek areas from the GC chromatograms of control and test coagiost extracts were compared to determine the decrease of different PCB Isomers. A significant decrease in the trlchloroblphenyls end one of the tetrachloroblphenyls was observed In extracts from the two week aerobic conposts. After four weeks of aerobic composting, significant decreases In all of the chlorinated biphenyls were observed. Anaerobic composting showed significant reduction of all chlorinated biphenyls after four weeks of composting. The mechanism of degradation appears to be rapid dechlorination of biphenyls containing three chlorines or less since no build-up of the mono- or dlchlorlnated compounds was observed. Dechlorination of the penta- and tetrachlorinated molecules Is slower but does occur at a useful rate. No chlorinated dlbenzo-p-dloxlns or dlbenzofurans ware found In the chromatograms or In the GC-mass spectra run on the composted sample extracts. HONS 214905 4-119 COMPOSTING FOR DEGRADATION OF PCBs IN SOIL INTRODUCTION Polychlorinated biphenyls (PCBs) re complex mixtures of verious chlorinated biphenyls which were widely used because of their physical, chemical and biological stability as well as their high dielectric constant. PCBs have entered the environment through spills and losses during manufacture, vaporization or leeching from PCB-contatnlng formulations, leaks from sealed or partially sealed systems and disposal of waste PCS (_1_). PCBs have been found In ocean sediments, the atmosphere and In many terrestrial and aquatic locations Including both polar Ice caps (Z). Because of their physical and chemical stability, these compounds are persistent environ* mental pollutants. Weathering and microbial processes degrade PCBs only to a limited extent m the environment (3). Evidence Indlcetes that PCBs may be biotransformed as well as metabolized by some microorganisms {4,5,6). Several Investigators have reported biotransformation of the lesser chlorinated biphenyls {mono- and dlchloroblphenyls) to chlorobenzolc acid (7.*.8.9.10). Tucker et al_. (6) showed that mono- and dlchlorobl phenyls were readily biodegraded by activated sludge but as concentrations of the higher chlorinated biphenyls Increased, degradation rates decreased. Kong and Sayler (9) simulated natural river conditions to demonstrate the degradation and mineralization of monohalogenated biphenyls by mixed microbial cultures from PCB contaminated river sediments. Furukawa and Chakrabarty (JJJ demonstrated total degradation of mono- and dlchlorlnated biphenyls using Aclnetobacter or Arthrobacter sp. containing a plasmid specifying conversion of chloroblphenyls to chlorobenzolc acids and a genetically constructed psaudomonad capable of metabolizing mono- or dtchlorobenzoates. This paper describes a study In which congesting was evaluated for decontamination of soil containing PCBs. This study Is part of a series of experiments to determine the applicability of coaeostlng techniques for the destruction of persistent hazardous materials contained In a soil or sediment matrix. EXPERIMENTAL MATERIALS AND METHOOS This study evaluated coagiosting of PCS contaminated soil on a laboratory scale. Both aerobic and anaerobic composting were studied to compare PCB biodegradation In the two systems. The experiments were carried out with soil that was Intentionally spiked with PCBS (Aroclor 1242). HONS 214906 4-120 r r The experimental apparatus used In the laboratory studies Is shown in Figure 1. Mason Jars were used to contain the compost materials. A ring of plastic tubing with holes drilled at 1/4 Inch Intervals was located beneath the compost tn each Jar. The ring was connected to a glass tube which extended through the stopper in the top of the Mason Jar, For the aerobic composts, the tube was attached to two trapsjn series, one containing sodium hydroxide and the other water, A second tube Inserted through the stopper was attached to a series of traps containing sodium hydroxide, sulfuric acid and activated carbon as shown in Figure 1. These traps were connected to a vacuum pump. Warmed air was first drawn through a sodium hydroxide trap to remove any CO; and then through a water trap to humidify the air and remove any caustic material. The scrubbed air was then drawn through the compost materials. The atmosphere above the compost was continually drawn out through sulfuric acid, sodium hydroxide and activated carbon traps to remove any gaseous biodegradation products. The dead traps prevent contamination of the composts or Individual gas traps In case of vacuum failure. For the anaerobic composts, the Initial NaQH scrubbing trap was eliminated. The anaerobic composts were Initially flushed with nitrogen to remove oxygen from the system and were flushed every 3-4 days thereafter to change the atmosphere In the composting vessels. Each compost was made up of 23.7 g of alfalfa hay, 24.7 g Purina Sweetena Horsefeed and 5 g bf Lakeland soil (a total of 50 g dry weight). The soil was spiked to yield 2000 mg/Kg of Aroclor 1242 In the total compost (dry weight basis). The compost materials were thoroughly mixed and the composting process Initiated by addition of sufficient water (containing primary effluent and horse manure) to give tha composts a 60S moisture content. The composting apparatus was then placed tn a SS<>C incubator and attached to the vacuum system. The experimental matrix for the composts Is presented In Table 1. The temperature of each compost was monitored dally by means of a thermocouple placed within the cosmos t materials. The composts were sacrificed as specified In Table 1. At the time of sacrifice, each compost was examined for Its appearance and smell. A sample (approximately 1 gram) was taken from the center of each compost and dispersed In distilled water (1:9 solids to water ratio) to determine the pH of the compost. The remainder of the compost material was dried in an oven at 60C for 24 hours. The dried compost materials were ground with a hamner pulverizer to 18 mesh (1 mm) and the ground powder thoroughly mixed. Two subsamples from each compost (approximately 3 to 4 grams) were taken, weighed and Soxhlet extracted with a mixture of 40% benzene, 40X acetone, 10* hexanes and lot methanol. Extraction efficiency of Aroclor 1242 from uncomposted material (positive controls) was 77X. Extraction efficiency from composted materials ranged 4-121 ' HONS 214907 til-' Air KOM Traa Flfurt 1, Laboratory Coopost Aaratlon Apparatus To Vacuum HONS 214908 r r Table 1 EXPERIMENTAL MATRIX FOR COMPOSTS Type of Compost SPIKED LAKELAND SAND A. Positive Controls B. Negative Aerobic Controls (no Aroclor 1242) C. Experimental Aerobic Composts 0. Negative Anaerobic Controls (no Aroclor 1242) E. Experimental Anaerobic Composts Number Time Sacrificed 2 D Meet 2 2 weeks 2 4 weeks 2 2 weeks 2 4 weeks 2 2 weeks 2 4 weeks 2 2 weeks 2 4 weeks 4-123 HONS 214909 from 83 89X. The extract was analyzed on a Varlan 3700 gas chromatograph with autosampler and electron capture detector. A Hewlett-Packard 5880 controller/ Integrator was used to control the Varlan gas chromatograph. The GC parameters were: Injection port - 280C; oven - 170 to 230C 8 15C/m1n; detector - 300C. Separation of the PCBs was accomplished with a 2 mm I.D. by 10 ft. glass column packed with 1.5X SP-2250/1.95X SP-2401 on 100/120 mesh Supelcoport using nitrogen carrier gas at a flow rate of 24 ml/mln. RESULTS Results of the laboratory-sc ale PCB composting studies are summarized In Tables 2 and 3. The aerobic composts showed the greatest reduction In PCB concentration with a 41.6 to 48.IX decrease in 2 weeks and 55.7 to 67.6X decrease in 4 weeks. PCB concentrations were not reduced as rapidly In anaerobic composts. A decrease of only 18.4 to 28.OX In the PCB concentrations of the anaerobic composts was observed after 2 weeks of anaerobic composting and 27.9 to 46.9X after 4 weeks. Aroclor 1242 Is a mixture of chlorinated biphenyls composed of approximately IX monochiorob1 pheny 1, 16X dtchloroblphenyl, 49X trIchlorobtphenyl, 25X tetrachlorobiphenyl, 8X pentachlorobtphenyl and IX hexachlorobtphenyl (J2). The normal method for GC quantitation compares the total Integrated area from all of the peaks In the sample to that of standards of similar chlorination. The methodology yields a comparison of PCB concentrations In the experimental samples with the positive controls but gives no Information on the effects of treatment on the Individual chlorinated biphenyl Isomers. To evaluate the effects of congesting on the Individual Isomers, areas of selected chromatogram peaks of the experimental samples were normalized and compared to corresponding chromatogram peaks of the positive controls. These comparisons for the 2 and 4-week aerobic and 4-week anaerobic coagwsts versus the positive control (100X) are shown graphically In Figure 2. Typical chromatograms are presented In Figures 3 and 4. After two weeks of aerobic composting, a significant decrease In the trlchloroblphenyls and one of the tetrachloroblphenyls was observed. No decrease in the higher chlorinated biphenyls was observed. However, after 4 weeks of aerobic composting, significant decreases In all of the chlorinated biphenyls occurred with less than 25X of the trlchlorinated biphenyls remaining. Anaerobic composting at four weaks also showed significant decreases In all of the chlorinated biphenyls. MQNS 214910 Tlbl* 2 AERO*1C COMPOSTING Of PCS CONTAMINATED SOIL Sy>1c Potlttv* Control Pot It tv* Control Nigitlve Control N*g*ttv* Control Eeperlnentil A Eeperlnentil 1 Negettv* Control Negettv* Control Eeperlnentil C Eeperlnentil D Ttn*1 0 0 2 Mtkl 2 weekt 2 Mtll 2 weekt 4 veeks 4 Meets 4 Meeks 4 weeks M. R.D. N.O. S.O 6.2 S.6 6.0 S.O S.4 S.O S.O PCS Concentritton2 u*/l............ 1S91 1517 17 11 901 107 17 IS 504 688 1 PCI Decrease3 41.6 41.1 67.6 SS.7 COMMItt Hit lux Mill. *t Mil COMOtted. Moderately deep, Atom, conpotted. Met luy. nodarately Imp. ytllow. Modern*lx dMp, brown. conpotted. Not very diap. not eelt conpotted. 014 tho* wall, fungal growth. Lett thin 60S no Ittore. Wet hiy ute Sweet tnell. <60* notitwre. yellow. UtM of ticrtftc* with rotpoct to cenpost itt-up. 2PC* concontrttton it tine of sacrifice tacroit* tn PCI concnntritlo* besed on the ever*9* control vein* of 1SS4 yg/g N.p. not deteratned. S2t HONS 214911 91T1 Table J mhermic COMPOSTING OF PCI contaminated soil Sample Positive Control Positive Control Ntgetlve Control Negetlve Control Experimental A Experimental 8 Negntlve Control Negotlvt Control Experimental C Experimental 0 Tim#' (neats) 0 0 7 Z Z Z a 4 4 4 21 N.O. NO. 5.7 4,8 S.O 4,7 6.0 7.0 5.5 6.5 PCI Concentration2 ug/5 1S91 IS 17 6 1768 1119 10 17 876 1170 De*crFeCaIse3* 18.4 78.0 46.1 77.9 Coanents Mildew smell, light fallow. Mildew smell, cohosted. Mildew smell, fellow, Musty smell, moderately composted Net hay smell, moderately composted. Net hay samlt, moderately composted. Sweet tamll, fungal growth, brown. Sweet tamll, dark color. 'Time of sacrifice with respect to compost tot.up, 2PCI concentration at time of sacrifice. ^Decrease In PCI concentration bated on the average control value of 1554 pg/g. *N.O. not determined. HONS 214912 ww Figure 2. Conner Ison of (formalized Peak Areu with Positive Controls <-127 HONS 214913 ff Figure 3. Chromatograms of Aerobic PCB Compost Extracts (1/20 dilution) NONS 214914 4-128 ^> imti w ihunt^ MM* M BuMTtM llltl lr< Figure 4. Chromatograms of Anaerobic PCB Compost Extracts (1/20 dilution) 4-129 < HONS 214915 K DISCUSSION The mechanism of PCD degradation In compost appears to be rapid dechlorination of biphenyls containing three chlorines or less since no build-up of the mono- or dlchlorlnated compounds was observed. Dechlorination of the tetra- and pentachlorlnated molecule Is slower but does occur at a useful rate. -No chlorinated dlbenzo-p-dloxlns or dlbenzofurans were found In the chromatograms or In 6C-mass spectra of the composted sample extracts - These preliminary laboratory-scale composting studies Indicate that composting may have potential as a method for on-site decontamination of soils or sediments contaminated with PCBs. Aroclor 1242 concentrations of 2000 mg/Kg In the compost materials (soil contaminated at 20.000 mg/Kg) were tolerated by the microbial populations present In the composts and significant decreases In PCB concentrations were observed. Composting for decontamination of soil or sediment Is relatively simple, requiring only that the soil or sediment Is thoroughly mixed with the compost materials. Composting can be carried out In various ways. The most popular methods currently In use are the windrow and forced aeration systems. In the windrow method, a pile or windrow of waste material is periodically turned by a machine to mix and aerate the organic matter. Anaerobic and aerobic decomposition occur In this system with aerobic decomposition resuming each time the material Is turned. Generally composting temperatures are lower In this system than In other systems and non-uniform biodegradation of organic materials may result. A second method, enclosed systems, requires a complex mechanical system which gives more rapid composting than the windrow system. These bloreactors are designed to provide operational control over environmental conditions such as air flow, temperature and mixing and to isolate the composting material from the surrounding environment. A third method, the Beltsvllle Aerated Pile Method was developed at the Beltsvllle Agricultural Research Center for composting of undigested sludges. In this forced aeration system, compost materials are mixed and bulking materials are added as needed. The compost pile is placed on a base of woodchlpt covering perforated plastic pipe. Air Is drawn through the pile via the perforated pipe In the base of the pile. The major advantages of this system are the production of a stable humus-like organic material, low capital Investment and low energy requirements. Composting for degradation of PCBs In soil or sediment should be further evaluated to answer the following questions: HONS 214916 4-130 f- r What It the maximum PCB concentration tolerated by the compost organ Unit? e What It the effect of the PCB chlorine content on composting? e What It the mechanism of PCB degradation In compost and what ara the byproducts? e What effects will different soils have on PCB degradation In composts? e What type of compost system is best suited for degradation of PCIs In soil? e What are the rates of PCB degradation for the different chlori nated biphenyls? e What Is the optimal composting time to achieve maximal PCB degradation? e Can the microbial populations responsible for PCB degradation In composts be Improved? Degradation rates as determined by laboratory-scale studies will provide a conserva tive estimate of rates In large-scale systems. Degradation rates observed In larger systems may be twice as rapid as those observed In the laboratory scale studies. The M4)or advantages of composting Include on-site decontamlantlon, production of a potentially useful organic material as a plant nutrient or soil conditioner, low capital Investment and low energy requirements. REFERENCES 1. Nlsbet, I.C.I. 1975. 'Environmental Transport and Occurrence of PCBs In 1975." In Proceedings of the National Conference on Polychlorinated Biphenyls, Chlcaoo, IL., November IT-2T7 Wb, U.S. EPA 5WVS7P0WT-------------- -- 2. Barton, T.B. and G.P. Arsenault. 1982. "Ultimate Olsposal of Pas." In J.H. Exner (ed.) Detoxication of Hazardous Waste, Ann Arbor Science, Ann Arbor, Ml. 3. Furukawa, K. 1982. "Microbial Degradation of Polychlorinated Biphenyls (PCBs)." In A.M. Chakrabarty (ed.) Degradation and Detoxification of Environ mental Pollutants, CRC Press, Boca Raton, fL. 4. Clark, R.R., E.S.K. Chian, and R.A, Griffin. 1979. "Degradation of Poly chlorinated Biphenyls by Mixed Microbial Cultures." Appl. Environ. Microbiol. 37:680-885. ---------------------------------- 5. Furukawa, K., N. Tomlzuka, and A. Kamlbayashl. 1979. "Effect of Chlorine Substitution on the Bacterial Metabolism of Various Polychlorinated Biphenyls." Appl. Environ. Microbiol. 38:301-310. 6. Tucker, E.S., V.W. Saeger and 0. Hicks. 1975. "Activated Sludge Primary Biodegradation of Polychlorinated Biphenyls." Bull, Environ. Contam. Toxicol. 14:705-713. ------ ----------- -------------------------- 4-131 HONS 214917 K- 7. Aimed, M. and 0.0. Focht. 1973. "Degradation of Polychlorinated Biphenyls by Two Species of Achromobacter." Can. Microbiol. 19:47-52. B. Furukawa, K. and F. Matsumura. 1976. "Microbial Metabolism of Polychlorinated Biphenyls. Studies on the Relative Degradability of Polychlorinated Biphenyl Components by A1call genes sp." J. Agrlc. Food Chew. 24:251-256. 9. Kong, Hay-Long and G.S. Sayler. 19B3. "Degradation and Total Mineralisation of Monohalogenated Biphenyls In Natural Sediment and Mixed Bacterial Culture. Appl. Environ. Microbiol. 46(3):666-672. 10. Shlarls, M.P. and G.S. Sayler. 19B2. "Blotransformatlon of PCS by Natural Assemblages of Freshwater Microorganisms." Environ. Scl. Technol. 16:367-369. 11. Furukewa, K. and A.M. Chakrabarty. 1982. "Involvement of Plasmids In Total Degradation of Chlorinated Biphenyls." Appl. and Environ. Microbiol. 44(3):619626. ---------------------------------------- 12. Hutzlnger, 0., S. Sate and V. Zltko. 1980. The Chemistry of PCBs. CRC Press, Ind., Boca Raton, FL. 4-132 MONS 214918 r r Discussion Ql. Have you looked at pH PufferIn 9? A. Hot In this study. QZ. Old you look at a carbon trap to see If temperature was actually driving the PCB out? A, Ho - we did not analyze traps> 4-133 MONS 14919 BACTERIAL DEGRADATION OF POLYCHLORlNATEO BIPHENYLS IN SLUDGE FROM AN INDUSTRIAL SEWER LAGOON by Walter S. Kim National Aeronautics and Space Administration Lewis Research Center Cleveland. Ohio and Adrienne M. Takacs Case Western Reserve University Cleveland, Ohio and David E. Kulvlnen National Aeronautics and Space Administration Lewis Research Center Cleveland, Ohio ABSTRACT A laboratory experiment was conducted to determine If polychlorinated biphenyls (PCB's) found In an Industrial sewer sludge can be effectively degraded by mutant bacteria. The aerated sludge was Inoculated dally with mutant bacteria In oroer to augment the existing bacteria with bacteria that were considered to be capable of degrading PCB's. The pH, nitrogen, and phosphorus levels were monitored dally to maintain an optimum growing medium for the bacteria. A gas chromatographic method was used to determine the PCB concentrations of the sludge Initially and also throughout the experiment. Results and discussion of the bacterial treatment of polychlorinated biphenyls are presented In this paper. 4-135 MOWS 214920 SUMMARY A 90-day laboratory sized experiment was conducted to determine the vfeblllty of using mutant bacteria to decontaminate an Industrial type sludge that was con taminated with polychlorinated biphenyls (PCB's). The contaminated sludge samples were obtained from an Industrial waste sewer lagoon system which Is an Integral part of a water pollution control facility, four representative samples of itudge were selected for treatment with mutant bacteria. The samples originally contained the PCB Aroclor 1260 In concentrations ranging from 135 to 232 parts per million (ppm). In addition to the four samples, a fifth sludge containing only three ppm of Aroclor 1260 was spiked with nearly 1000 ppm of Aroclor 1260. This spiked sam ple was used to determine whether high concentrations of PCB's In sludge would be affectd by the bacteria. The five samples were Inoculated dally with mutant Pseudomonas aeruginosa bacteria that had a potential for degrading PCB's. The nitrogen, phosphorus, and pH levels were monitored dally to maintain a growing medium for the bacteria, At the completion of the 90-day augmentation period, the concentrations of the PCB's In the sludge sample* were found to be higher than the original concentra tions. The cause of the apparent Increase In concentrations has not been fully understood. One possibility Is the biodegradation of the sludge In which the sludge Is metabolized and broken down Into lighter molecules. The lighter sludge Increases the concentration of PCB on a welght-to-welght basis. The apparent lower PCB concentration found around 20 days after Inoculation sug gests that the PCB was absorbed by the bacteria, thereby causing the apparent decrease In the PCB concentration of the sludge. After 20 days, the bacteria began to die, perhaps caused by the build up of toxic metabolic wastes, and subsequently released PCB back Into the sludge without having metabolized the PCB molecules. Thus, the experiment can be summarized as that the PCB In the sludge was absorbed by the mutant Pseudomonas aeruginosa bacteria for 20 days. After 20 days, the bacteria began to release the absorbed PCB's back Into the sludge which by now has been degraded Into lighter molecules. Thus, the PCB concentration appeared to decrease at around 20 days, then Increase after 20 days. Under the constraints of this experiment, the PCB was absorbed by V>* bacteria, but was not degraded. 4-136 HONS 214921 Section 1 INTRODUCTION Polychlorinated biphenyls (PCB's) are chlorinated aromatic organic compounds that are commonly used as dielectric fluids in capacitors and transformers. Their physical and chemical properties are such that they are chemically and thermally stable, fire resistant, essentially non-conductlve, and have a low solubility in water. Because of their chemical and physical stability, they are virtually indestructible when spilled Into the environment, whether accidently or discarded for disposal purposes. High temperature combustion processes and special chemical treatment methods are evolving which are used to destroy the PCB's or to degrade the compo nents Into non-toxic materials. The U.S. Environmental Protection Agency has established a level of SO parts-per-mllllon above which materials must be disposed of In a Federally approved landfill or Incinerator. Other treatment and cleanup methods are under Investigation, and one of the most appealing concepts Is bio logical treatawnt In which bacteria Is used to degrade PCB's In situ. Early studies In the use of bacteria for the degradation of specific chemicals reported In 1970, have shown that biphenyl can be degraded by gram-negative bacteria Isolated from soil (Ref. 1). In that study, biphenyl was converted to phenylpyruvate In a salt medium. Other studies of biphenyl degradation Involved bacteria such as Pseudomonas putlda. BeUerlnckla species, and a strain of Mucor. For the metabolism of pure chlorinated biphenyl (not PCB), Rhlzopus japonicus was used to convert 4-chloroblphenyl to 4-ch1oro-4-hydroxyb1phenyl. Two species of Achromobacter. Isolated from sewage effluent, degraded biphenyl to benzoic acid, and 4-chloroblphenyt to 4-chlorobenzolc acid. However, no natural bacteria was found to successfully degrade polychlorinated biphenyls, which are mixtures of many chlorinated biphenyls. If bacteria that can degrade PCB's are found, they most likely will be developed through genetic engineering. The use of such mutant bacteria would be particularly effective In the treatment of soils, sludges, and waste water for the decontamina tion of PCB's In those environments. Contaminated waste water lagoons would not HONS 214922 reaulre the subsequent draining and physical removal of PCB contaminated soil and sludge. The lagoon could be conveniently treated by adding a mutant bacteria >n situ and effect a clean up process at a relatively low cost and with little effort. To determine the feasibility and effectiveness of bacterial augmentation, a labo ratory experiment was conducted using PCB contaminated sludge saaiples from an Industrial sewer lagoon. By using the actual sludge saaiples from the lagoon, this experiment was to provide insight as to whether the lagoon can be docomtamlnated by mutant bacteria. This paper describes the experiment and discusses the resutts of using mutant Psuedoaonas aeruginosa bacteria with PCB contaminated sludge samples. 4-138 MONS r r Section l EXPERIMENT The Industrial sewer lagoon used for this study Is part of a six-lagoon water pollution control facility which receives decante waters generated from waste tanks that process rinse waters and liquid wastes generated by metal plating and tube cleaning facilities. This lagoon serves as an equalization pond holding about 3.5 million gallons of waste water. An earlier Investigation Indicated widespread, low-level PCB contamination in the lagoon. Few sites were found to be contaminated with PCB's In excess of 50 parts-per-mllllon (ppm) of Aroclor 1260. The EPA regu lated concentration Is 50 ppm. In order to fully determine the magnitude and extent of PCB contamination In the lagoon, an extensive sampling program was conducted. Sampling equipment was loaded In a boat and towed to the selected sampling points. Core samples of sludge and the sediment were collected. Average depth of the sediment was about six Inches with the base material of clay. These samples were analyzed for PCB concentra tions which ranged up to the highest concentration of 467 ppm. A definite water flow pattern was observed In which the Intet of the lagoon contained the highest concentrations and toward the discharge outlet the concentrations gradually declined. After contamination levels were determined, many alternatives and possibilities were considered for decontamination of the sewer lagoon. The alternatives ranged from closing the lagoon; permanently fixing the sediment Into concrete-Hke mate rial; biological and physlochenlcal methods; to physically dredging the sediment for disposal, treatment, or Incineration. Many other possibilities were also con sidered, but one of the most attractive alternatives was the In-place degradation of PCB by microorganisms. If this microbiological method proved valid, draining of the lagoon water and physically removing the sludge would not be required. Using the sludge samples that had been analyzed for PCB, a small laboratory experiment was carried out to determine whether biological degradation of PCB's would be feasible for the Industrial waste lagoon. Glass desslcators with approximately 2-llter capacities were used as experimental tanks. Five such sludge tanks ware charged with the first tank containing about 450 grams of sludge with 4-139 HONS 214924 f 205 ppai concentration of PCB. Niter was tdded to bring the equeoui level up to one liter. The second tenk conttlned 50 grams of sludge with 222 ppm PCB diluted to one liter. The third tenk was two liter tenk containing 100 grams of sludge in which 900 ppm of Aroclor 1260 was added to the sludge originally containing 3 ppm. This tank Mat started to determine whether high concentration of PCB In the sludge would be affected by the bacteria, and to roughly determine the degradation rate to see ho long the treatment would be needed to bring the concentration below the 50 ppm level. The fourth tank contained 50 greets of sludge with 135 ppm, and the fifth tank had SO gram of sludge with 232 ppm diluted to one liter. These experimental sludge tanks were Inocculated with the mutant Pseudomonas aeruginosa bacteria obtained through a biochemical firm that has considerable experience and expertise In genetic en9lneer1ng. The bacteria culture which was contained In a bran base was prepared for augmentation by soaking the culture in water for six to nine hours with a pinch of sodium bicarbonate. The Initial dosage rate was 0.1 grams of the culture per one liter, then the dosage rate was gradually decreased over 20 days to a maintenance level of 20 milligrams per liter. Along with dally supplementation of bacteria, a nutritional balance for biological activity was maintained each day. The dissolved oxygen was kept at about 7 ppm level by aeration of the tank and agitation of the slurry. The pH level was main tained at 7.5, and nitrogen as aanonia determined by the Nessleriiation method was over 5 ppm level. The phosphorus as ortho-phosphate was determined by the molyb date method and maintained at 1 ppm. The temperature of the tanks was kept at about 75* F. 4-140 HOMS 21492$ n T Section 3 RESULTS A.1D OISCUSSION _ In monitoring the sludge tenks delly, the general trend hts been the drop In the gH end nitrogen, while phosphorus concentretlon remelned above 1 ppm level. No pert leu Ur pettem was observed other than a need for almost dally addition of sodium bicarbonate to bring the pH up and an ammonium compound to supplement the nitrogen level. To condensate for nitrogen loss, ammonium hydroxide was used If the pH dropped to about 6.8, and amaonlum nitrate was used If the pH was near 7.5 level. During the augmentation period, biological activity In the tanks was monitored biweekly. An ordinary microscope at 100 magnification and a phase contrast microscope at 400 magnification were used to Insure the presence of active orga* nisms In the tanks. As activated sludge Is formed and eges, there is a successive predominance of protozoans and rotifers which correlates with the bacterial popu lation (2). In general, as the bacteria populetlon Increases, flagellates become predominant. When the becterla populetlon Is et the maximum, fret twinning dilates are the predominant organisms. As the bacteria population declines, rotifers become predominant. In the sludge tanks, the protozoans and rotifers were observed with the flagellates and free swimming dilates as the predominant organisms. At about the middle of the augmentetion period, the number of the higher organisms epptered to have sharply declined. Toward the end of the augmen tation parlod, the higher organisms were almost absent. In the last 12 days of augmentation, a large amount of bacteria was added to each tank along with supplemental food to provide an additional carbon source. The population of mlcroorganlsam correlates with the sludge conditions. In that the sludge starts to age and breaks down when the population of microorganisms Is at the maximum. The sludge becomes digested and the food source for the bacteria becomes diminished. With further digestion of sludge, the organisms eventually utilize the Internal material through endogeneous process, resulting In the diminished population. Near 12 days toward the end of the augmentation period, the sludge appeared emulslon-1Ike and the higher organisms were almost absent. 4-141 MONS 214926 This Indicated that the microorganism population hid severely declined due to lick of an external food source. Although i large amount of bacteria was added along with supplemental food, no Improvement was observed. The results of bacterial augmentation over a 90-day period are presented in the Table. These results were obtained by analyzing the sludge samples by a gas chromatographic method. A sample of sludge was weighed, then PCB was extracted with l;l acetone and hexane mixture using an ultrasonic bath. Extracted PCB was analyzed by a gas chromatograph equipped with an electron capture detector. This analytical Instrument was fitted with a twonaetar glass column packed with 3X OV-1 on 80-100 mesh Chromosorb W-HP and maintained at 200* C. The carrier gas was ultrapure nitrogen at a flowrate of 25 milliliters per minute. The Injection port was kept at 2S0* C. andthe detector was maintained at 300* C. The output of the gas chromatograph was recorded on a 1 mv strip chart recorder. The PCS chromatogram of the sample was quantitated from the standard chromatograms. As shown in the Table, PCB In the Industrial sewer sludge appears not to have been degraded by the bacteria over 90 days period. In these sludge tanks, the final concentrations of PCB at the end of 90 days are higher than the starting concen trations, although In Tank 2the concentration is decreased slightly. This small decrease In Tank 2 cannot be regarded as an Indication of bacterial degradation of PCB, since the fluctuation between the anlysls dates and the precision of sampling and analytical method could cause this small difference. An interesting observation from the results is the apparent dip in the concentra tion around 20 days from the beginning of the experiment. After this 20-day period, the concentration rises and remains at an elevated level. Although It Is possible that the sampling and the analysis may be In error, this Initial dip In the PCB concentration of the sludge might indicate an effective biological activity In the first 20 days, after which the biological activity appears to have ceased. The apparent lower concentration suggests that the PCB was degraded or absorbed by the organisms present In the tanks. Since the PCB concentration Increases after 20 days, the absorbed PCB may be subsequently released back Into the sludge. Perhaps, after about 20 days, the sludge tanks become saturated with materials such as the bacterial metabolites that may attack or cause the bacteria to de compose and release the absorbed PCB back Into the environment. This is difficult to substantiate because limited data Is available. With so many variables that can adversely affect the microorganism population. It Is conceivable that the 4-142 HONS 214927 r r organist** have decomposed since the microorganism population declined as the aug mentation period progressed. Another supportive factor is that the experimental sludge tanks are essentially batch reactors in which no water flows In or out of the tanks on a continuous basis. All added materials and metabolic wastes are accumulated in the tanks. If PCB was metabolized or degraded by the bacteria In order to p*1n energy and ennance growth, the PCB peaks in the gat 'chromatogram may show an Indication. Possibly, one or more gas chromatographic peaks may be de creasing faster than the other peaks since PCB contains a mixture of chlorinated biphenyls. This hypothesis could not be fully tested because of apparent rise In all PCB peaks after the 20-day period. The apparent rise In the PCB level cannot be explained at this point, but one possible explanation is that the organisms may be preferentially attacking the abundant carbon source of the sludge before attacking the PCB. This may have the effect of increasing the PCB concentration, on a weight-to-weight basis. In the biodegraded sludge. Therefore, when a sample of sludge is weighed out and ana lyzed, the concentration of PCB would appear to be higher as shown in the results. <*143 HONS 214928 rr Section 4 CONCLUDING REMARKS In each of the flv* experimental sludge tanks, the bacterial augmentation over 90-day period resulted In en apparent increese In the PCB concentretlon. At the end of 90 dys, the fine! concentrations of PCB In the Industrial sewer sludges were higher than the original concentrations. These results nay be due to the sludge being degraded prior to PCB and due to the sludge tanks becoming toxic to the bacteria by the bacterial metabolites that cause the bacteria to decompose and release absorbed PCB back Into the sludge. It appears that around 20 days after the augmentation, microorganisms released the absorbed PCB back to the sludge without metabolizing or breaking down the mole cules. In this case. If a continuous water system was used to allow a flow of water In and out of the augmentation system, the PCB concentration In the sludge would be decreased by the organisms with absorbed PCB being carried away with the water flow. At any rate, the mutant bacteria of Pseudomonas aeruginosa used In this experiment did not degrade the PCB's found In the sludge from an Industrial sewer lagoon. It is Important to test each PCB contaminated material to determine the feasibility and effectiveness of bacterial augmentation prior to an actual field demonstration. 4-144 HONS 214929 F r Section 5 REFERENCES 1. 0. Hutilnptr, S. Safa, and V. 21tko. Tht Chmlitry of PCB*t. tec* R*ton, Florid*: CRC Pro**, 197*, pp. 143-148. R.E. McKInnty and A. Gram. Protozoa nd Act1v*ttd Sludgt. Sdwao* *nd Induttrlal H*tt*. Vol. 28, No. 10, Oct. 1956, pp. 1219-1231. TABLE: RESULTS OF PCS CONCENTRATIONS IN BACTERIA AUGMENTED SLUDGE, ppa No. o*y Tank J__ Tank 2 Tank 3 Ttnk 4 Tank 5 0 21 2* 30 36 42 50 57 205 --94 145 --195 255 222 .6.5. 2--06. 1.5.4. 162 983 8.2.1. 9--62. 1449 --135 --no -1*31 1160 165 2..3.2 210 258 275 75 230 174 1443 82 253 157 1225 90 91 2..6.2 .1.81. 1.4.7.6 163 287 4-1*5 . HONS 214930 rr Discussion Ql. How or why did the Increase In ppe occur? A. The cauls Ion sludge Is auch lower density entf-lt contains both live end deed orgenlsas. 02. Why was thet particular streln of Pseudoaonas aeruginosa used? A. Thet streln hed been used In Jet fuel tanks, we o decided to use the seae on the sludge. Q3. Is 90 days enough tlae? A. Thet was the allowable test period. 4-146 HONS 214931 P*rt 5 PCI DtSTAOCTXOK NONS 214932 Vw OBSTRUCTION OT HIGH CONCENTRATION PCBa IN A UTILITY BOILER T. E. Sladhoff, C. A. Zala. Bud H. B. Horrla Union Eloctrlc Company INTRODUCTION Aa with oo<t electric utllltlta and othar lnduatrial concern*, Union Elacerlc coepany producaa elgnlfleant gutntltlae of wmeta oil and othar coabuetlbla pro* duct* during tha cooraa of nerval bualnaaa activity. Tha graataat Majority of thoaa veetoa con* from two primary aourcaa; 1) dlalactrlc flulda uaad In alactrlcal *qulpeent, and 2) lubrletting olla utod In actor vehicle* and large power plant nachlnary. location of pratrloua nonuftcturlng and eorvlclng procoduraa, aona alnaral oil dlalactrlc flulda will Inadvertantly contain tha chaaleal conpound polychlorinated biphenyl or PCS In varying but generally low concantratlona. Other dlalactrlc flulda, known geaarlcelly aa atkartl, Inten tionally contain about 60 percent PCBo for a apeclflc purpoaa and uea. In coeparlaon to nany other organic ehealcala tha oclantlflc Information available would aeon to Indicate that PCI* are relatively low In toxicity. However, becauae of their entreat otabllity and blooccueulatlve nature and tha Inexact nature of the aclance in thlo area, prudent action require* that opeclal earn be taken to tea that thoaa flulda are dlapooed of properly. On Hay 11, 1979 tha U.t, Bnvironaental Protection Agency proaulgated regulatlona (44 PR 11514) rigidly governing the handling of PCIa. A key feature of thla reg ulation la the grouping of PCI waata liquid* into throe categorlaa with differing dlopoeal optlona for each group. Tha claaalflcatlona and their dlapoaal option* erai HONS 214933 5-1  1) For thoaa liquid* containing 300 ppa PCI* or graatar, dlapoaal aunt talc* plact In tn incinerator which couplla* with the proviaiona of 40 CPU 161.70 ("PCB incinerator" raqulranenta). 2) For thoea liquid* containing between 30 and 300 ppo PCBa there era three dlapoaal option*: burning in a PC* incinerator; placement in a landfill which coapllaa with the provlalon* of 40 CP* 761.75 ("chemical waata landfill" requirement*), or incineration in a high efficiency hollar netting apeclflc requirement*. 3) Finally for thoaa liquid* containing laaa than 50 ppa PCI*, no direct reatrlctlon* are placed on thalr dlapoaal accept that they cannot be uaad aa eealanta, coating* or duat control agent*. With the promulgation of thie regulation Union Electric began to examine which option* beat eat the Company'a dlapoaal need* and objectivaa. Continuing and tubetantiml liabllltlea for the** wmata* even after proper dlapoaal in accordance with regulation* nede landfilling undeairabla. However, PCB liquid Incineration coata war* high. Therefore, Union Electric decided to puraua the high efficiency boiler dlapoaal option for contaminated oil. Thle alternative waa attractive becauea it allowed Union Electric to maintain control of the aubetanca through lta ultimata dlapoaal. Aa wall, the reeource waa not waated In that tha energy value of tha oil waa recoverad through generation of electricity. On July 7, I960 Union Electric raqueatad approval to dlapoa* of 30 to 500 pp* PCB llqulde in thalr Labmdla Plant Unit #4 located near Labadie, Hieeouri. During Hovow bar 1 HO public meeting* war* held and in early January, 1901 a aerie* of trial burn* were conducted <dille tha atack affluent, aeh, and ambient air in the aurrauadimg area were carefully monitored. Ho PCI* ware datectmd in any of than* ample*. On Hay 26, 1991 the Hiaaourl Department of Natural Reaourcaa certified tha bollar a a raaourca recovery facility and on tha following day approval under tha PC* rule waa granted by EPA Region VII. HONS 214934 5-2 r~ f With this approval Union Electric solved lta KB contaminated oil dlapoaal problaa, However, It did not aolva the problem of dlapoalng of thoaa liquids iraacar chan 500 ppo and especially sekaral. Por dlalaccrlc fluids contalnlni iraatar than 500 ppo PCM, tha refulatlone dlaallow thn hl|h efficiency hollar aathod of dlapoaal, and dilution to lower PCI concent rat Iona la alee dewed lllatal, Ac previouely noted, dlapoaal of thaae flnlda muet taka plena In an In cinerator which oeete ocrlnfent atandarda aatabllahed In tha retulatlona. However, thla requirement la not abaoluta. The Rational Adelnlatrator haa tha authority under 40 CFR 761,10(e) to approve alternative dlapoaal aethode If they can be shown equivalent to a PCI incinerator end If they will not preennt an unraeaonabla rlek of Injury to health or tha eavlronaent. Union Electric carefully examined the daalta and operational faaturea of tha Labedle Unit 64 and dacsrolned that ouch a deeonetretloa could be node for thla facility while burnlnt an oll/aekarel blend contalnlnt 5 percent Kla la conjunction with pulverltad coal aa tha prleary fuel source. Therefore la December, iMl Union Electric Initiated fornal dlacuaaiona with EPA Ration vii which ultlnataly raeultad in approval of the ayatee preaented herein. Thla paper deecrlbea the entlnearlnt anelyala of the deelfa feetnrea and opera tional paraaetere which lead to tha conclualon that the boiler can eeeentlelly destroy all PCRa introduced and tha raeulta of a carefully planned teat bum which wee conducted to confirm thaae concluelone. BOILER DESCRIPTION Labedle Unit #4 la the newest of four 600 eeiawett (MW) coal fired fenerstlaf units located on the eouch beak of the Missouri Rivar In Franklin County near Labedle, Mlasour1. The boiler (Fifure l), which la eeeentlelly Identical to the other three boilers at the site, le a teatantlolly fired unit manufactured by Combustion Entlnearlnt Inc. end was pieced In commercial operation In Aufttat of 1973. The boiler hat e axleib contlauoue rated (NCR) heat Input of 3367 million BTU per hour which corraepoede to o bun reto of aboot 240 tone per hour of pulverised coal. The HONS 214935 5-3 Approxtint# location of 0* and CO Monitor* Waaca Oil Burner (Elav. 564'1 Bpafe ! 1/ \ OI i.--dI =h mi>-- 1 IT 1 : <|e 1-- ^ir | i i \ -) 1 i j Figure 1. Sectional View, Ubedle Unit 14 Boiler SOURCE: COKIUSTIOff UKIHIUIIK MONS 214936 5-< r r pulverised coal 1* transported in air auapanalon froa the coal allla to ala levels of burners located at each of four cornara of tha boiler (Figure 2). The coal la lnjectsd Into tha furnace through tha burners In such a aenner ae to Intleatelr six with tha secondary air and to foru a large vortex about the boiler* vertical axis thereby creating a huge, rotating ball of flaee. Oil-fired wara-up guns are provided between tha bottoe two coal burner levels for boiler ware-up and to saelat la stable coal Ignition during start-up. One gue is located In each corner of the boiler. At full load the flaee taapereture within the fire ball Is about 3000*F and gas teaperatures at the superheater division panels ap proach 2AOO*F. SYSTEM DESIGN In preparing to burn PCle at La bad la significant codifications were eeds to the boiler and plant at considerable expense prior to the Initial test burn of contanlnatsd oil. The purpose In asking these changes was to provide e well-designed e vstea solely for PCI vneee oil which alnlalsee tha exposure of this oil to the envlronaent and Insures Its proper disposal. After approval to burn up to SI PCBs It wee decided to furehar codify the aye tea to Include additional storage capacity and blending facilities, the following le e core detailed description of the various coeponente of this ayetac (see Figure 3). Waste-Oil Storage end Heeding In order to store PCI eootealnated oil at the plant until sufficient quantities for blending end Incineration have been gsnerated, an existing tank located above ground and laaedlately south of the plant has been designated and prepared for this purpose. In addition an adjacent tank has been designated for blending 31 PCI batches for Incleeratloa. Both tanks were previously used to store No. 2 fuel oil, ware built to conform to tha Aeerlcen Petrol sue Ineeleute (API) Standard #30 end west ell applicable OSNA regulreaente. The oil tanka ere surroaeded by a 3 foot high concrete block dike designed to contain the entire volume of both tanks, la addition, the areas between the MONS 214937 5-5 UftMft WINPtOX WNQCOI OAMAftM 'KICQMOMV Aim MlffR OXtVl UNIT COIMlCTMa UttR rriLTHfQ IH0MAN4M COAtMOULl MTJUhCTMU OU 4UN WAct Tutu iCUANM R4NIU tON'TO* nOAM tUANfA NOttLf ICOAL CMWAftTMtftT) mNR MOttil IOIUAUX A1A COMAT I PUUH ICXMtfft rwrnu hao oil Aim t rigur* 2. Cutaway Vlav, Typical C-E Tilting Tangantiul Bursar At stably 5-6 i SOURCE) COHWJSTIOK ENGINEERING XONS 214938 f r riullf, t< oro iwtaatlr Hilt niatiuit DNA tnkii|Mi to (loci fripaati of ISO DNA tint otay tooorl Waatifylai oa4 okoroatarlolai tko ioao(o> ifool- fjiai tki iu^ii of tko PCB 4o|ro4itiTO yotkwiy la tklo itrola. Tkl till illo* to Aovolof (0 lititiiliii of botk tko otraatara oat roialitloa of tkooo amayaai. ~ S0BXA1Y To ktTO loolatol ill akoraatarltoA too Aiotlaot ot|iaioBi, Aloollroaoo oattoofcot SO ill Corraofcootol Of. Ml, oklok aataaiiToly 4o|ro4o Axoalor 1241. kat oklok asklklt cloot Alffotoatoi la tko yortlaalor aoaiaaoro oklok ora 4o|ro4o4. la laaorol. PCI ooaioatra oltk tko 2.4,4' takotltatloa pottara ora aatokollial raflAly ky Ml, kat poorly ky 1150. la ooatroit, aaay aoaaoaoro oakatl total ot pooltloaa 2.3.2' ora ropllly lo|tolil by IISO, kat aot by Ml. Tko pralaatloa of aotokolltao aaMaltooolly ootokllikaa tkot tklo la iaOoal 4o|ra4atlaa. Saaa _ AairoAotloa prolaota, laolaAlai aklarokiaaootao, ora Moan to kotk orioaloBt. kat la aoay oaaao tko aojor aatokolltoo prolaool ky MSO aro aal'ao, Tkooo Aoto ica aoat raallly aiyliiMl ky yoatalotlai tkot tko aajor yotfcvoy of fa aotokol1 ob la 1150 at 1 litoo a 41oay|amaoo oklok praforaatlolly attoako ot oorkoa pooitloao 1,4. If tkli la tko aoia. It olll ko tko flrot Aaaoaitratioa of laoolvoaoat ot i 1,4 41eay|imaoo la tko klo4oiroAotloa of kiphaayl or PCBi. la oootroat. Ml yrokokly aaployo a aora oaaoa 2,S 41oay|oaooo aoakaalaa, Plaolly, tko too or|ialma korkor Slot loot yliMilo oklok aay ko laoalooA la KB 4a|r*4atloa. To aro aoo la aa oaoallaat yoaltloa to ko|la roollolai oar loai-tara |oalo, BMaly, aaAorotaaAlai tko klookoalotry ill paaotloo of kootarlol 9CB AopraAotloa. ' AawwnwMBm To tkoak J. C, Coraokoa, K.B, Tafaor, J, toitk, oa4 I. tokllai for |ao okraaotoraffcia oaalyalo, I.J, Toffloalra ia4 t.B, Brook* for koly la oollaotlat PCI io*talaato4 oaoiroMoatal oasplaa, B. tarakova *a4 A.B. Ckokxakorty for o a took of Aolaotokootor of, P( B.L. Plakkolaor, aa4 T.B. la for kalpfal llaaaooloaa, *a4 I. Borooa for typlai tklo ataaoorlyt. imtBNCU 1. AloaaaAor, N, IMP. loll Biology, (Boolooa of Bataarok, Notlaaol Batoaraao to ao orok, TMBOD) l20P-240. 2. AkaaA, B. oat tb.D. Pookt. 1P7S. DopraAotloa of polyoklorlaataA kipkoaylo ky too oyooloa of Aokroaofcootor. Caa. J. Biaroklol. Hj47-S2. 4-115 HONS 214939 K 5, |astar. B.A.. ft.E. Gllbart. I. A. LU|att. J.B. Xalaprita, i*d I.A. Void.a. 1973. Tha dairadatlaa of palpchlariutad hlphaapia by sisrssrtasiaaa. Sal. Tatal Ea*lras. 1:55-61. 4. Talp. N. Th.X.. i. Sab* it*, *d 0. latiiaiar. 1971, Tha haatarlal aataballa at 4,4`-dlahlarabiphaspl, aad ita ppraaaiaa by altaraativa earbaa aaarcaa. Chaaaaphara 1:101-101. 5. Iaitip, 9.9. aad A.X. Chakrabartp. 1911. Plata Ida tpaalfpiaf p-ahlar*biphasp 1 dapradatlaa la aatarla baatarla, pp. 275-213. Ia E.K. Tlaala aad A. fihlar. (Ida,), Plaaalda at Xadloal. Esalramaatal, aad Caaaaralal lapartaac*. Elaavlar/Hartk Ballaad liaaadlaal Praaa, Naa lark. (. Ealahardt. 9.I., B.L. Ghadwlak. M.A. Gala. I.E. Eabartaaa. aad 1,1, Battoa. 1911. Elaatla atadp at tha bladairadatlaa at blpfcaspl aad It* aaaaahlarlaatad aaalaiaaa bp a alsad aarlaa alarahlal aaaaaaltp. Eavlraa. lal. Taah. U 175-79. 7. Iplaaatra, X. aad J. 9aat*as. 19B2. A aav tataltatlva aaaaraba aapabla at iravth as ahlarabiphaapla. J. 0*a. Appl. Klarablal. If: 41-72. I. Eaa|, I.L. aad O.B. Saplar. 19B1. Dairadatlaa aad tatal alaarallsatiaa at aaahala|taatad blpkaapla la aataral aadlaaat aad alsad haatarlal aaltara. Appl. Ear Iras. Mlarahlal. 11:444-472. 9. Faraka*a, E. 19B. Mlarahlal dairadatlaa at palpahlarlaatad blpkaaplt, pp. 15-57. A,M. Chakrabartp, (Ed,). Bladairadatlaa aad Dataslflaatlaa at EavlraaBaatal Fallataata, (SC Praaa, la*.. Baa* Eataa. 10. Ealaar, E.L.E., aad 9.T.B. Vaaf, 1974. Baatarlal dairadatlaa at palpahlarlaatad blpkaapla. I. Idaatlflaatlaa at acaa aatahell* pradaata trca Araalar 1242, Ball. Esvlraa, Caataa. Taslaal, 11:291-194. 11. Taakar, E.l., V,V. Ba|ar. aad 0. flak*. 1975. Aatlvatad alad|* prlaarp dairadatlaa at palpahlariastad hlphaapia. Ball. Ear1ram. Caataa. Taslaal. U**-7W* 12. Clark, E.E., E.B.E. Chlaa, aad E.A. Orlttla. 1979. Dairadatlaa at palpahlarlaatad blpkaapla bp alsad alarablal amitaras. Appl. Eavlraa. Mlarahlal. 11:400-415. 13. Tail, 0. aad E. Bad*. 1910. Dairadatlaa at palpahlarlaatad hlphaapia bp laraarBaalaB*, J. Vatar Fallat. Caatral Fad. 11:1035-1045. 14. Baplar, 0.1,, M, Bhaa, aad E.E. Calaall. 1977, Oravth at aa aataarla* tmlBHIII ap, aa palpahlarlaatad blpka^l. Mlarab. Baal. 1:241-255. 15. Lis, D. 1950. lahaaaaaaat at MS* bladairadatlaa bp **di llpals aaltaaat*. Vatar Eaaaarah 11:1447-1475. 14. Farmka**, E., E, Taaaaara. aad A. Eaalbapaahl, 197B. Etta at at ahlarla* aahatltatlaa aa tha hl*d*|radabllltp at palpahlarlaatad hlphaapia. Appl. Baa Iras. Mlarahlal. U'.215-227. 17. Farakaaa, E., N. TaaUaka, tad A. Eaalbapaahl. 1979. Etfaat at ahlarla* aahatltatlaa aa tha haatarlal aatahallaa at aarlaa* palpahlarlaatad hlphaapia. Appl. Esvlraa. Mlarahlal. 1:S01-S10. 4-116 HONS 214940 -H * II. luitnii I.. M, Taalaaka. iM A. EaalBayaakl. 1M1. luittKl dogrtdttloa of yolyoklorlaattd kifkoifli (PCI) tad tkair aotoBolltao. Ado. Eif, Nad. Biol. lJ4A;407-41t. 11. Jokatoa. 1,1. ob4 B.T. Stoalor. 1171. Dlaaialla tie* of ttaitlo oapoudi By AlMllMiu oatroakao. J. Boot. Ifll;44t-47S. 10. Ballatkaltor, E. ao4 H. Zoll. UN. Aaalyalo of yolyoklOTlaatod Blykoaylo By glaaa aaylllary |ao ekroattograyky. toayotltioa of tookaloal Aroolor aad Cloykoa - BO alstaroa. Proaoalaa Z. Aaal. Ckaa. 102 10-11. 21. Baaooa, J.B. ood l.t. Olaoa. 1171. laolotlos of largo Bootorial ylaaaidt oa4 oBorootorltatloa of tko PI laoaayatlBillty groay ylataida yM02 oa4 yMQS, J. Boot. UJ.;M7-11I. 12. Caaalaa, C.B., K.A. Loll lot, oa4 M. Bogota. 1174. Ooaaa i- CorraoBaotorlM. yy. <01-417. la B.B. Baakaaot oa4 N.B. OIBBobo. <Sda.), Borgoy't Moaaal of Do to ratat tiro Baatarlology, Blgktk Mltloa. Tko Vllllaaa oa4 111BIbo Co,. Baltlaora. 21. Parakavt, E. aa4 A.M. CkakroBarty. 118. lavolvmoat of ylanldo la total 4ogra4otloa of oklorlaatad kiykearla. Ayyl. Barrlrea. MltroBlol. <24. 24. Laat, D. oa4 V.C. Breaa. 1170. Tko aloroklal aotaBolloa of BlyBoayl. loakaa. J. U|:14y. 25, Olkooa. B.T., B.L. BoBorto, M.C, folio, aad Y.M. KoBol. 1171. Orldatloa of BlyBoayl By Boltorlaokla oyooloi. Bloofcta. Bloykyo. Boo. Coaaaa. 12:111 211. 24. Cotolaal, 0., A. Colatkl, C. Borllal, aai Y. Troooaal. 1171. MotaBollaa of Blykt^l, l-lydr o7'-4-eio-4-yfcooylkei-2,4-dleaoto: tko aota-oloooaoo yco4aot fraa 2,J-dlkydroxyklykoayl By tnilcaimi oatlAa. Blookaa. J. 124i1041-1044. 27. Ckrlatoykar, P., B. Praaklla, 8. ka|4aaarloa. oa4 I.N. Tlnlo. 1111. Naalyalatloa of dagrodatlvo goaoo of aoll Baotorlo, yy 101-110. lo T. Loiolagor, B. Battor. A.B. Cook. aa4 J. Kaoook. (B4a.), Mloroklal Dogradatloa of XoaOBlotloo oad Booaloltraat Coayoaada. Aoadaait Prooo, No* York. It. Porroll, B. oad A.M. Ckokrokarty. 1171. Dogradatloa ylaaaldo; aolooalor aataro oad aodo of ooolatloa, yy P7-10I. la E.N Tiaala oad A. Faklor. (Bda.), Naalfi of Modloal, SavlroMtatal. a ad Ctasorolal layortaaoo. Blaorlor/ltortk-Sollaad Blaatdiool Proa a, No* York. 21. Ckatrorjoo, B.C.# B.T. Eollogg, D.B. Votklao. aad A.M. Ckokrokarty. 1111. Noaalda la tko Blodogradotloa of oklorlaatod araaotlt oowyoaodi, yy Jiy52t. la t.B. Lory, B.C. Cloooa, oad B.L. Cooalg, (Bda.), Molooalar Biology, Patkogoalolty, aad Boology of Baatorlal Plaalda. Ploaaa Proai, Now York. JO. Ptskortoa, J.B., B. Coraoy, oad B.B. Boa. 1171. Brolatloa aad ayroad of yoatlolda dogradlag aBlllty aaoag toll a lor oar goal oat. yy. 2B7-2PI. la E.N. TIbIo aad A. faklor. <B4o.), Plataldt of Modloal, BaolroMOatal, aad Coaaorolal laportaaoo. Bloorlar/Mortk-Bollaad Bloaodloal Prooa. Nov York. Jl. CkakroBarty, A.M. 1IB2. Ooaotlo aookaalMa la tko dlaaiallatloa of tklorlaatod oaayoaadt, yy 127-lJf. la A.M. CkakroBarty, <14.), Blodogrodatlea aad Dotoalflaatlta of SavlroMtatal Pollataata. CMC Prato, lao.. Boot Batoa. 4-117 HONS 214941 Dlecuaaion Ql. Ita'c chit protean llalc*d to a epecific Arodora? A. Me, the etudjr eo far haa been Halted to elte-contaalnated with 12*2. Would need to leolete aicrobecterla froa ether alter to etudp other Aroclore , -- 4-118 HONS 214942 r r COMPOSTING FOR DEGRADATION OF PCBs IN SOIL Jeneflr o. isblster George L. Anspach Judith F. Kitchens Atlantic Research Corporation Alexandria, Virginia 22312 ABSTRACT This study evaluated aerobic and anaerobic composting of PCS contaminated soil on a laboratory scale. Composts were composed of alfalfa hay, horsefeed and Lakeland soil spiked to contain 2000 mg/Kg of Aroclor 1242 (dry weight basis). Warm humidified air was drawn through aerobic compost materials. Anaerobic composts were flushed with nitrogen. All composts were Incubated at 55C. After two and four weeks of Incubation, compost materials were oven dried, ground, subsanpled and Soxhlet extracted for GC analysis. The aerobic composts showed the greatest destruction of PCSs with 41.6 and 48.IX destruction In two weeks and 55.7 to 67.6X destruction In four weeks. PC8 destruction In anaerobic composts was 18.4 to 28.OX after two weeks and 27.9 to 46.9X after four weeks of composting. Individual peak areas from the GC chromatograms of control and test compost extracts were compared to determine the decrease of different PCB Isomers. A significant decrease In the trlchloroblphenyls and one of the tetrechloroblphenyls was observed In extracts from the two week aerobic composts. After four weeks of aerobic composting, significant decreases In all of the chlorinated biphenyls were observed. Anaerobic composting showed significant reduction of all chlorinated biphenyls after four weeks of composting. The mechanism of degradation appears to be rapid dechlorination of biphenyls containing three chlorines or less since no build-up of the mono- or dlchlorinated compounds was observed. Dechlorination of the penta- and tetrachlorlnated molecules Is slower but does occur at a useful rate. No chlorinated dlbenio-p-dloxlns or dlbenxofurans were found In the chromatograms or In the GC-mass spectra run on the composted sample extracts. * HONS 214943 4-119 COMPOSTING FOR DEGRADATION OF PCBS IN SOIL INTRODUCTION Polychlorinated biphenyls (PCBs) are complex mixtures of various chlorinated biphenyls which were widely used because of their physical, chemical and biological stability as well as their high dielectric constant, PCBs have entered the environment through spills and losses during manufacture, vaporization or leeching from PCB-contalning formulations, leaks from sealed or partially sealed systems and disposal of waste PCB 0). PCBs have been found in ocean sediments, the atmosphere and in many terrestrial and aquatic locations including both polar Ice caps (2). Because of their physical and chemical stability, these compounds are persistent environ mental pollutants. Weathering end microbial processes degrade PCBs only to a limited extent In the environment (3). Evidence Indicates that PCBs may be blotransformed as well as metabolized by some microorganisms (4,5,6). Several investigators have reported biotransformation of the lesser chlorinated biphenyls (mono- and dlchloroblphenyls) to chlorobenzolc acid (7.4.8.9.101. Tucker et ak (6) showed that mono- and dlchloroblphenyls were readily biodegraded by activated sludge but as concentrations of the higher chlorinated biphenyls Increased, degradation rates decreased. Kong and Sayler (9) simulated natural river conditions to demonstrate the degradation and mineralization of monohalogenated biphenyls by mixed microbial Cultures from PCB contaminated river sediments. Furukawa and Chakrabarty (JJ_) demonstrated total degradation of mono- and dlchlorlnated biphenyls using Aclnetobacter or Arthrobacter sp. containing a plasmid specifying conversion of chloroblphenyls to chlorobenzoic acids and a genetically constructed pseudomonad capable of metabolizing mono- or d1ch1oroben zoates. This paper describes a study In which composting was evaluated for decontamination of soli containing PCBs. This study Is part of a series of experiments to determine the applicability of composting techniques for the destruction of persistent hazardous materials contained In a soil or sediment matrix. EXPERIMENTAL MATERIALS AND METHOOS This study evaluated composting of PCB contaminated soil on a laboratory scale. Both aerobic and anaerobic composting were studied to compare PCB biodegradation In the two systems. The experiments were carried out with sol 1 that was Intentionally spiked with PCBS (Aroclor 1242). 4-120 HONS 214944 The experimental apparatus used In the laboratory studies Is shown in Figure 1. Mason Jars were used to contain the compost materials. A ring of plastic tubing with holes drilled at 1/4 inch Intervals was located beneath the compost In each Jar. The ring was connected to a glass tube which extended through the stopper In the top of the Mason Jar. For the aerobic composts, the tube was attached to two traps In series, one containing sodium hydroxide and the other water. A second tube Inserted through the stopper was attached to a series of traps containing sodium hydroxide, suJfur 1c acid and activated carbon as shown In Figure 1. These traps were connected to a vacuum pump. Warmed air was first drawn through a sodium hydroxide trap to remove any CO2 and then through a water trap to humidify the air and remove any caustic material. The scrubbed air was then drawn through the compost materials. The atmosphere above the compost was continually drawn out through sulfuric acid, sodium hydroxide and activated carbon traps to remove any gaseous biodegradation products. The dead traps prevent contamination of the composts or Individual gas traps In case of vacuum failure. For the anaerobic composts, the Initial NaOH scrubbing trap was eliminated. The anaerobic composts were initially flushed with nitrogen to remove oxygen from the system and were flushed every 3-4 days thereafter to change the atmosphere In the composting vessels. Each compost was made up of 23.7 g of alfalfa hay, 24.7 g Purina Sweetena Horsefeed and 5 g of Lakeland soil (a total of 50 g dry weight). The soil was spiked to yield 2000 mg/Kg of Aroclor 1242 In the total compost (dry weight basis). The compost materials were thoroughly mixed and the composting process Initiated by addition of sufficient water (containing primary effluent and horse manure) to give the composts a 60X moisture content. The composting apparatus was then placed in a 55C incubator and attached to the vacuum system. The experimental matrix for the conposts is presented In Table 1, The tenperature of each compost was monitored dally by means of a thermocouple placed within the compost materials. The composts were sacrificed as specified In {Table 1. At the time of sacrifice, aach compost was examined for Its appearance and smell. A sample (approximately 1 gram) was taken from the center of each compost and dispersed In distilled water (1:9 solids to water ratio) to determine the pH of the compost. The remainder of the compost material was dried In an oven at S0C for 24 hours. The dried compost materials were ground with a hammer pulverizer to 18 mesh (1 mm) and the ground powder thoroughly nixed. Two subsamples from each compost (approximately 3 to 4 grams) were taken, weighed and Soxhlet extracted with a mixture of 40% benzene, 40X acetone, 10% hexanes and 10% methanol. Extraction efficiency of Aroclor 1242 from uncomposted material (positive controls) was 771. Extraction efficiency from composted materials ranged 4-121 HONS 214945 -122 TuMng Ffgura 1. laboratory Cohost Aeration Apparatus MONS 214946 I i Table 1 EXPERIMENTAL MATRIX FOR COMPOSTS Type of Compost SP1KEO LAKELAND SANO A. Positive Controls 8. Negative Aerobic Controls (no Aroclor 1242) C. Experimental Aerobic Composts 0. Negative Anaerobic Controls (no Aroclor 1242) E. Experimental Anaerobic Composts Number Time-Sacrificed 2 0 week 2 2 weeks 2 4 weeks 2 2 weeks 2 4 weeks 2 2 weeks 2 4 weeks 2 2 weeks 2 4 weeks 4-123 MOWS 214947 from 83 89X. The extract was analyzed on a Varlan 3700 gas chromatograph with utotampler and electron capture detector. A Hewlett-Packard 5860 controller/ Integrator was used to control the Varlan gas chromatograph. The GC parameters were: Injection port - 280C; oven - 170 to 230C 0 15C/m1n; detector - 300C. Separation of the PCBs was accomplished with a 2 nn I.D. by 10 ft. glass column packed with l.SX SP-2250/1.9SX SP-2401 on 100/120 mesh Supelcoport using nitrogen carrier gas at a flow rate of 24 mL/mln. RESULTS Results of the laboratory-scale PCS composting studies are sumarlzed In Tables 2 and 3. The aerobic composts showed the greatest reduction In PCS concentration with a 41.6 to 48.IX decrease In 2 weeks and S5.7 to 67.6X decrease In 4 weeks. PCB concentrations were not reduced as rapidly In anaerobic composts. A decrease of only IB.4 to 28.0X In the PCB concentrations of the anaerobic composts was observed after 2 weeks of anaerobic composting and 27.9 to 46.9X after 4 weeks. Aroclor 1242 Is a mixture of chlorinated biphenyls composed of approximately IX monochloroblphenyl, 16X dlchloroblphenyl, 49X trlchloroblphenyl, 2SX tetrachloroblphenyl, 8X pentachloroblphenyl and IX hexachloroblphenyl {!). The normal method for GC quantitation compares the total Integrated area from all of the peaks In the sample to that of standards of similar chlorination. The methodology yields a comparison of PCB concentrations In the experimental samples with the positive controls but gives no Information on the effects of treatment on the Individual chlorinated biphenyl Isomers. To evaluate the effects of composting on the Individual Isomers, areas of selected chromatogram peaks of the experimental samples were normalized and compared to corresponding chromatogram peaks of the positive controls. These comparisons for the 2 and 4-week aerobic and 4-week anaerobic composts versus the positive control (100X) are shown graphically In Figure 2. Typical chromatograms are presented In Figures 3 and 4. After two weeks of aerobic composting, a significant decrease In the trlchloroblphenyls and one of the tetrachloroblphenyls was observed. No decrease In the higher chlorinated biphenyls was observed. However, after 4 weeks of aerobic composting, significant decreases In all of the chlorinated biphenyls occurred with less than 2SX of the trlchlorlnated biphenyls remaining. Anaerobic composting at four weeks also showed significant decreases In all of the chlorinated biphenyls. HONS 214948 llble 2 MHOS 1C COMPOST INC OF PCB CONIN'INATEO SOIL Swple Positive Control Poiltfve Control Negetlve Control Negetlve Control Expertnentel A Experlnentel Negetlve Control Negetlve Control Experlnentel C Experlnentel 0 Ttif) 0 PCB Concentration? I PCI . m/a DtcrttttJ N.D. 1591 0 N.O. 151? 2 atekt S.O 1? 2 weeki 6.2 31 2 xMki 5,6 908 41.6 2 weeks 6.0 ao? 48.1 1 MCOfci 6.0 17 4 vMki $.4 35 4 Noekt S.O 504 67.6 4 tki 5.0 688 55.7 Ct^nti Hit hey mil. flot Mil 1 coMpoited. Moderately do*, kroon, conpoited. Met hey, aodoretoly denp, yellow. Moderately denp. brown, conpoited. Not very dnop. not well conpoited. Old shoe inell, fungi1 growth. Lett then 601 noltton. Net hey see!) Sweet well. <601 nolstore. yellow. tTtae of tier I flee with respect to conpott let-up. ?PCI concentration it ttat of tier If Ice ^Decreite In PCI concentration hiiod on the ivortie control vilue of 1554 pg/g N.O. not lettrained. 125 HONS 214949 9 Z \-* Table 3 ANAEROBIC COMPOSTING OF PCB CONTAMINATED SOIL Sanole Potltfva Control PotHIvt Control Negative Control Negative Control ExpertMntal A EiperlMntal I Negative Control Negative Control Cxperlaental C ExperlMntal 0 TlM> (weofct) 0 0 2 i 2 i 4 4 4 4 J*L N.O. N.O. $.2 4.1 s.o 4,7 5.0 7.0 5.S 6.5 PCB Concentration2 U0/4 1591 Si7 6 e 1266 till to 12 825 1120 * PCI Oecreate' 16.4 26.0 46.9 27.9 CoMentt Mildew (Mil, light pel IonMildew taell, CMpotted. Mildew well, yellow. Nutty tMll, aoderately coapotted. Wet hay taell, aoderately coapotted Met hay taell, aoderately coapotted. Sweet taell, fungal growth, brown. Sweet taell, dart color. }T1m of tacrlflce with respect to ceapett tet-up> 2?CI concentration at ttM of tacrlflce. ^Decreeta In PCI concentration bated an the average control value of 3554 ug/g. *K.O. not detanelned. HONS 214950 Figure 2. Comparison of Normalized Peak Areas with Positive Controls HONS 214951 4-127 M (Mlkl* tlllMII tlllMt Figure 3, ChroMtogrms of Aerobic PCB Compost Extrects {V20 dilution) nqns 214952 4-128 mtIAfr MfrtSs Figure 4. Chromatograms of Anaerobic PCS Compost Extracts {1/20 dilution) 4-129 * HONS 214953 rf DISCUSSION The mechanism of PCB degradation In compost appears to be rapid dechlorination of biphenyls containing three chlorines or less since no build-up of the mono- or dlchlorinated compounds was observed. Dechlorination of the tetra- and pentachlorlnated molecules Is slower but does occur at a useful rate.-No chlorinated dlbenzo-p-dloxlns or dlbenzofurans were found In the chromatograms or In SC-mass spectra of the composted sample extracts. These preliminary laboratory-scale composting studies Indicate that composting may have potential as a method for on-site decontamination of soils or sediments contaminated with PCBs. Aroclor 1242 concentrations of 2000 mg/Kg In the compost materials {soil contaminated at 20,000 mg/Kg) were tolerated by the microbial populations present in the composts and significant decreases In PCB concentrations were observed. Composting for decontamination of soil or sediment Is relatively simple, requiring only that the soil or sediment Is thoroughly mixed with the compost materials. Composting can be carried out In various ways. The most popular methods currently In use are the windrow and forced aeration systems. In the windrow method, a pile or windrow of waste material Is periodically turned by a machine to mix and aerate the organic matter. Anaerobic and aerobic decomposition occur In this system with aerobic decomposition resuming each time the material Is turned. Generally composting temperatures are lower In this system than In other systems and non-uniform biodegradation of organic materials may result. A second method, enclosed systems, requires a complex mechanical system which gives more rapid composting than the windrow system. These bioreactors are designed to provide operational control over environmental conditions such as air flow, temperature and mixing and to Isolate the composting material from the surrounding environment. A third method, the Beltsvllle Aerated Pile Method was developed at the Beltsvllle Agricultural Research Center for composting of undigested sludges. In this forced aeration system, compost materials are mixed and bulking materials are added as needed. The compost pile is placed on a base of woodchlps covering perforated plastic pipe. Air is drawn through the pile via the perforated pipe In the base of the pile. The major advantages of this system are the production of a stable humus-like organic material, low capital Investment and low energy requirements. Composting for degradation of PCBs In soil or sediment should be further evaluated to answer the following questions: 4-130 HONS 214954 r r Whit Is the maximum PCB concentration tolerated by the compost organisms? What Is the effect of the PCB chlorine content on composting? What Is the mechanism of PCB degradation In compost and what are the byproducts? - What effects will different soils have on PCB degradation In composts? a What type of compost system Is best suited for degradation of PCBs In soil? a What are the rates of PCB degradation for the different chlorlnated biphenyls? a What Is the optimal composting time to achieve maximal PCB degradation? a Can the microbial populations responsible for PCB degradation In composts be Improved? Degradation rates as determined by laboratory-scale studies will provide a conserva tive estimate of rates In large-scale systems. Degradation rates observed In larger systems may be twice as rapid as those observed In the laboratory scale studies. The mejor advantages of composting Include on-site decontamlantlon, production of a potentially useful organic material as a plant nutrient or soil conditioner, low capital Investment and low energy requirements. REFERENCES 1. Nlsbet, I.C.I. 1975. "Environmental Transport and Occurrence of PCBs In 1975." In Proceedings of the National Conference on Polychlorinated Biphenyls. Chicago. IL.TTI575B5?'JT-Tf7 T97r,' O. EM BW/g^S^PWT--------------- 2., Barton, T.B. and G.P. Arsenault. 19B2. "Ultimate Disposal of PCBs." In J.H. Exner {ed.) Detoxication of Hazardous Waste. Ann Arbor Science, Ann Arbor, HI. 3. Furukawa, K. 1982. "Microbial Depredation of PolychlorinatedBiphenyls (PCBs)." In A.M. Chakrabarty (ed.) Degradation and Detoxification of Environ mental Pollutants. CRC Press, Boca Raton, Fl. 4. Clark, R.R., E.S.K. Chian, and It.A. Griffin. 1979. "Degradation of Poly chlorinated Biphenyls by Mixed Microbial Cultures." Appl. Environ, Microbiol, 37:680-685. ----------------------------------- 5. Furukawa, K., N, Tomlxuka, and A. Kanlbayashl. 1979. "Effect of Chlorine Substitution on the Bacterial Metabolism of Various Polychlorinated Biphenyls." Appl. Environ. Microbiol. 38:301-310. 6. Tucker, E.S., V.W. Saeger and 0. Hicks. 1975. "Activated Sludge Primary Biodegradation of Polychlorinated Biphenyls." Bull. Environ. Contam. Toxicol, 14:705-713. ------------------------------------------------- HONS 214955 4-131 7. Allied, M. and 0.0. Focht. 1973. "Degradation of Polychlorinated Biphenyls by Two Species of Achronobacter." Can. J, Microbiol. 19:47-52. B. Furukawa, K. and F. Matsumgra. 1976. "Microbial Metabolism of Polychlorinated Biphenyls. Studies on the Relative Degradability of Polychlorinated Biphenyl Components by A!cal 1genes sp." Agrlc. Food Chew. 24:251-256. 9. Kong, Hay-Long and G.S. Sayler. 19B3. "Degradation and Total MineralljatIon of Monohalogenated Biphenyls In Natural Sediment and Mixed Bacterial Culture. Appl. Environ. Microbiol. 46(3) :666-672. 10. Shlarls, M.P. and 6.S. Sayler. 1982. "Blotransformetlon of PCB by Natural Assemblages of Freshwater Microorganisms." Environ. Scl. TechnoT. 16:367-369. 11. Furukawa, K. and A.M. Chakrabarty. 1982. "Involvement of Plasmids In Total Degradation of Chlorinated Biphenyls." Appl. and Environ. Microbiol. 44(3):619- 12. Hutzlnger, 0., S. Sate and V. Zitko. 1980. The Chemistry of PCBs. CRC Press, Ind., Boca Raton, FL. 4-132 HONS 214956 01 scuff 4 on Qi. Hava you lookad it pH - buffering? A, Not In this study. Q2. Old you look at carbon trap to tat If taaparatura was actually driving tha PCB out? A. No - wa did not analyze traps. 4-133 MOWS 2X4957 r r BACTERIAL DEGRADATION OF POLYCHLORINATED BIPHENYLS IN SLUDGE FROM AN INDUSTRIAL SEHER LAGOON by Halter S. Kim National Aeronautics and Space Administration Lewis Research Center Cleveland, Ohio and Adrienne M. Takacs Case Western Reserve University Cleveland, Ohio and David E. Kulvlnen National Aeronautics and Space Administration Lewis Research Center Cleveland, Ohio ABSTRACT A laboratory experiment was conducted to determine If polychlorinated biphenyls (PCB's) found In an industrial sewer sludge can be effectively degraded by mutant bacteria. The aerated sludge was Inoculated dally with mutant bacteria In oroer to augment the existing bacteria with bacteria that were considered to be capable of degrading PCB's. The pH, nitrogen, and phosphorus levels were monitored dally to maintain an optimum growing medium for the bacteria. A gas chromatographic method was used to determine the PCB concentrations of the sludge Initially and also throughout the experiment. Results and discussion of the bacterial treatment of polychlorinated biphenyls are presented In this paper. 4-135 MOMS 214958 SUMMARY A 90-day laboratory sized experiment was conducted to determine the viability of using mutant bacteria to decontaminate an Industrial type sludge that was con taminated trlth polychlorinated biphenyls (PCB's). The contaminated sludge samples Mere obtained from an Industrial Maste seMer lagoon system which is an integral part of a Mater pollution control facility. Four representative samples of sludge Mere selected for treatment with mutant bacteria. The samples originally contained the PCB Arcelor 1260 In concentrations ranging from 135 to 232 parts per million (ppm). In addition to the four samples, a fifth sludge containing only three ppm of Aroclor 1260 was spiked with nearly 1000 ppm of Aroclor 1260. This spiked sam ple was used to determine whether high concentrations of PCB's In sludge would be affeetd by the bacteria. The five samples Mere Inoculated dally with mutant Pseudomonas aeruginosa bacteria that had a potential for degrading PCB's. The nitrogen, phosphorus, and pH levels were monitored dally to maintain a growing medium for the bacteria. At the completion of the 90-day augmentation period, the concentrations of the PCB's In the sludge samples were found to be higher than the original concentra tions. The cause of the apparent Increase In concentrations has not been fully understood. One possibility Is the biodegradation of the sludge In which the sludge Is metabolized and broken down Into lighter molecules. The lighter sludge Increases the concentration of PCB on a welght-to-welght basis. The apparent lower PCB concentration found around 20 days after Inoculation sug gests that the PCB was absorbed by the bacteria, thereby causing the apparent decrease In the PCB concentration of the sludge. After 20 days, the bacteria began to die, perhaps caused by the build up of toxic metabolic wastes, and subseouently released PCB back Into the sludge without having metabolized the PCB molecules. Thus, the experiment can be summarized as that the PCB In the sludge was absorbed by the mutant Pseudomonas aeruginosa bacteria for 20 days. After 20 days, the bacteria began to release the absorbed PCB's back Into the sludge which by now has been degraded into lighter molecules. Thus, the PCB concentration appeared to decrease at around 20 days, then Increase after 20 days. Under the constraints of this experiment, the PCB was absorbed by tjhe bacteria, but was not degraded. 4-136 HONS 2l**59 Section I INTRODUCTION Polychlorinated biphenyls (PCB's) are chlorinated aromatic organic compounds that are conenonly used as dielectric fluids In capacitors and transformers. Their physical and chemical properties are such that they are chemically and thermally stable, fire resistant, essentially non-conduetlve, and have a low solubility In water. Because of their chemical and physical stability, they are virtually Indestructible when spilled Into the environment, whether accidently or discarded for disposal purposes. High temperature combustion processes and special chemical treatment methods are evolving which are used to destroy the PCB's or to degrade the compo nents Into non-toxic materials. The U.5. Environmental Protection Agency has established a level of 50 parts-per-mlllion above which materials must be disposed of In a Federally approved landfill or Incinerator. Other treatment and cleanup methods are under Investigation, and one of the most appealing concepts Is bio logical treatment In which bacteria Is used to degrade PCB's In situ. Early studies In the use of bacteria for the degradation of specific chemicals reported In 1970, have shown that biphenyl can be degraded by gram-negative bacteria isolated from soil (Ref. 1). In that study, biphenyl was converted to phenyl pyruvate In a salt medium. Other studies of biphenyl degradation Involved bacteria such as Pseudomonas putlda, Beljerlnckla species, and a strain of Mucor. For the metebollsm of pure chlorinated biphenyl (not PCB), Rhlzopus .laoonlcus was used to convert 4-chloroblphenyl to 4-chloro-4-hydroxyblphenyl. Two species of Achromobacter. Isolated from sewage effluent, degraded biphenyl to benzoic add, and 4-chloroblphenyl to 4-chlorobenzolc acid. However, no natural bacteria was found to successfully degrade polychlorinated biphenyls, which are mixtures of many chlorinated biphenyls. If bacteria that can degrade PCB's are found, they most likely will be developed through genetic engineering. The use of such mutant bacteria would be particularly effective In the treatment of soils, sludges, and waste water for the decontamina tion of PCB's In those environments. Contaminated waste water lagoons would not 4-137 MONS 214960 r' reaulre the subsequent draining and physical removal of PCB contaminated soil and sludge. The lagoon could be conveniently treated by adding a mutant bacteria in situ and effect a clean up process at a relatively low cost and with little effort. To determine the feasibility and effectiveness of bacterial augmentation, a labo ratory experiment was conducted using PCB contaminated sludge samples from an Industrial sewer lagoon. By using the actual sludge samples from the lagoon, this experiment was to provide insight as to whether the lagoon can be decontaminated by mutant bacteria. This paper describes the experiment and discusses the results of using mutant Psuedomonas aeruginosa bacteria with PCB contaminated sludge samples. 4-138 HONS 214961 r r Section 2 EXPERIMENT The Industrial sewer lagoon used for this study is part of a six*lagoon water pollution control facility which receives decante waters generated front waste tanks that process rinse waters and liquid wastes generated by metal plating and tube cleaning facilities. This lagoon serves as an equalization pond holding about 3.5 million gallons of waste water, An earlier Investigation Indicated widespread, low-level PCS contamination in the lagoon. Few sites were found to be contaminated with PCB's In excess of 50 parts*per-miUion (ppm) of Arocior 1260. The EPA regu lated concentration Is 50 ppn. In order to fully determine the aiagnltude and extent of PCB contamination in the lagoon, an extensive sampling program was conducted. Sampling equipment was loaded In a boat and towed to the selected sampling points. Core samples of sludge and the sediment were collected. Average depth of the sediment was about six inches with the base material of clay. These samples were analyzed for PCB concentra tions which ranged up to the highest concentration of 467 ppm. A definite water flow pattern was observed In which the Inlet of the lagoon contained the highest concentrations and toward the discharge outlet the concentrations gradually declined. After contamination levels were determined, many alternatives and possibilities were considered for decontamination of the sewer lagoon. The alternatives ranged from closing the lagoon; permanently fixing the sediment Into concrete*!Ike mate rial; biological and physlochemlcal methods; to physically dredging the sediment for disposal, treatment, or Incineration. Many other possibilities were also con* sidered, but one of the most attractive alternatives was the in-place degradation of PCB by microorganisms. If this microbiological method proved valid, draining of the lagoon water and physically removing the sludge would not be required. Using the sludge samples that had been analyzed for PCB, a small laboratory experiment was carried out to determine whether biological degradation of PCB's would be feasible for the Industrial waste lagoon. Glass detsleators with approximately 2*1 Iter capacities were used as experimental tanks. Five such sludge tanks were charged with the first tank containing about 450 grams of sludge with 4-139 MOMS 214962 f 205 ppm concentration of PCB. Water was added to bring the aoueous level up to one liter. The second tank contained 50 grams of sludge with 222 ppm PCB diluted to one liter, The third tank was two liter tank containing 100 grams of sludge In which 960 ppm of Aroclor 1260 was added to the sludge originally containing 3 ppm. This tank was started to determine whether high concentration of PCB In the sludge would be affected by the bacteria, and to roughly determine the degredatlon rate to see how long the treatment would be needed to bring the concentration below the 50 ppm level. The fourth tank contained 50 grains of sludge with 135 ppm, end tne fifth tank had 50 grams of sludge with 232 ppm diluted to one liter. These experimental sludge tanks were inocculated with the mutant Pseudomonas aeruginosa bacteria obtained through a biochemical firm that has considerable experience and expertise In genetic engineering. The bacteria culture which was contained in a bran base was prepared for augmentation by soaking the culture in water for six to nine hours with a pinch of sodium bicarbonate. The initial oosage rate was 0.1 grams of the culture per one liter, then the dosage rate was gradually decreased over 20 days to a maintenance level of 20 milligrams per liter. Along with dally supplementation of bacteria, a nutritional balance for biological activity was maintained each day. Th* dissolved oxygen was kept at about 7 ppm level by aeration of the tank and agitation of the slurry. The pH level was main tained at 7.5, and nitrogen as ammonia determined by the Nesslerlzatlon method was over 5 ppm level. The phosphorus as ortho-phosphate was determined by the molyb date method and maintained at 1 ppm. The temperature of the tanks was kept at about 75* F, 4-140 MOWS 214963 1> Section 3 RESULTS ANO DISCUSSION In monitoring the sludge tanks dally, the general trend has been the drop In the pH and nitrogen, while phosphorus concentration remained above 1 ppm level. No particular pattern was observed other than a need for almost dally addition of sodium bicarbonate to bring the pH up and an amnonlun compound to supplement the nitrogen level. To compensate for nitrogen loss, amonlum hydroxide was used If the pH dropped to about 6.8, and ammonium nitrate was used If the pH was near 7.5 level. During the augmentation period, biological activity In the tanks was monitored biweekly. An ordinary microscope at 100 magnification and a phase contrast microscope at 400 magnification were used to Insure the presence of active organlsms In the tenks. As activated sludge Is formed and ages, there Is a successive predominance of protozoans and rotifers which correlates with the bacterial popu lation (2}. In general, as the bacteria population Increases, flagellates become predominant. When the bacteria population Is at the maximum, free swimming dilates are the predominant organisms. As the bacteria population declines, rotifers become predominant. In the sludge tanks, the protozoans and rotifers were observed with the flagellates and free swlamlng dilates as the predominant organisms. At about the middle of the augmentation period, the number of the higher organisms appeared to have sharply declined. Toward the end of the augmen tation period, the higher organisms were almost absent. In the last 12 days of augmentation, a large amount of bacteria was added to each tank along with supplemental food to provide an additional carbon source. The population of microorganisms correlates with the sludge conditions. In that the sludge starts to age and breaks down when the population of microorganisms Is at the maxlmue. The sludge becomes digested and the food source for the bacteria becomes diminished, with further digestion of sludge, the organisms eventually utilize the Internal material through endogeneous process, resulting In the diminished population. Near 12 days toward the end of the augmentation period, the sludge appeared emulsion-like and the higher organisms were almost absent. -141 HONS 214964 This indicated that the microorganism population had severely declined due to lack of an external food source. Although a large amount of bacteria was added along with supplemental food, no improvement was observed. The results of bacterial augmentation over a 90-day period are presented n the Table. These results were obtained by analyzing the sludge samples by a gas chromatographic method. A sample of sludge was weighed, then PCS was extracted with 1:1 acetone and hexane mixture using an ultrasonic bath. Extracted PCS was analyzed by a gas chromatograph equipped with an electron capture detector. This analytical Instrument was fitted with a two-meter glass column packed with 3X ov-1 on 60-100 mesh Chromosorb W-HP and maintained at 200* C. The carrier gas was ultrapure nitrogen et a flow rate of 25 milliliters per minute. The injection port was kept at 250* C. and the detector was maintained at 300* C. The output of the ges chromatograph was recorded on a 1 mv strip chart recorder. The PC6 chromatogram of the sample was quantitated from the standard chromatograms. As shown In the Table, PCS In the Industrial sewer sludge appears not to have been degraded by the bacteria over 90 days period. In these sludge tanks, the final concentrations of PCB at the end of 90 days are higher than the starting concen trations, although in Tank 2 the concentration Is decreased slightly. This small decrease in Tank 2 cannot be regarded as an indication of bacterial degradation of PCB, since the fluctuation between the anlysls dates and the precision of sampling and analytical method could cause this small difference. An Interesting observation from the results Is the apparent dip In the concentra tion around 20 days from the beginning of the experiment. After this 20-day period, the concentration rises and remains at an elevated level. Although It Is possible that the sampling and the analysis may be In error, this Initial dip In the PCB concentration of tha sludge might Indicate an effective biological activity In the first 20 days, after which the biological activity appears to have ceased. The apparent lower concentration suggests that the PCB wes degraded or absorbed by the organisms present In the tanks. Since the PCB concentration Increases after 20 days, the absorbed PCB may be subsequently released back Into the sludge. Perhaps, after about 20 days, the sludge tanks become saturated with materials such as the bacterial metabolites that may attack or cause the bacterle to de compose end release tha absorbed PCB back Into the environment. This Is difficult to substantiate because limited data Is available. With so many variables that can adversely affect the microorganism population, It Is conceivable that the 4-142 HONS 214965 1^ organisms have decomposed since the microorganism population declined as the aug mentation period progressed. Another supportive factor Is that the experimental sludge tanks are essentially batch reactors In which no water flows In or out of the tanks on a continuous basis. All added materials and metabolic wastes are eccumuleted in the tanks. If PC8 was metabolized or degraded by the bacteria In order to gain energy and enhance growth, the PCB peaks In the gas chromatogram may show an Indication. Possibly, one or more gas chromatographic peaks mgy be de creasing faster than the other peaks since PCB contains a mixture of chlorinated biphenyls. This hypothesis could not be fully tested because of apparent rise in all PCB peaks after the 20-day period. The apparent rise In the PCB level cannot be explained at this point, but one possible explanation is that the organisms may be preferentially attacking the abundant carbon source of the sludge before attacking the PCB. This may have the effect of increasing the PCB concentration, on a weight-to-weight basis, in the biodegraded sludge. Therefore, when a sample of sludge is weighed out and ana lyzed, the concentration of PCB would appear to be higher as shown in the results. 4-143 HONS 214966 rr Section 4 CONCLUDING REMARKS In etch of the five experimental sludge tents, the bacterial augmentation over a 90-day period resulted in an apparent Increase In the PCS concentration. At the end of 90 days, the final concentrations of PCS In the Industrial sewer sludges were higher than the original concentrations. These results may be due to the sludge being degraded prior to PCS and due to the sludge tanks becoming toxic to the bacteria by the bacterial metabolites that cause the bacteria to decompose and release absorbed PCS back Into the sludge. It appears that around 20 days after the augmentation, microorganisms released the absorbed PCS back to the sludge without metabolizing or breaking down the mole cules. In this case, If a continuous water system was used to allow a flow of water In and out of the augmentation system, the PCS concentration in the sludge would be decreased by the organisms with absorbed PCS being carried away with the water flow. At any rate, the mutant bacteria of Pseudomonas aeruginosa used In this experiment did not degrade the PCB's found in the sludge from an industrial sewer lagoon. It Is Important to test each PCS contaminated material to determine the feasibility and effectiveness of bacterial augmentation prior to an actual field demonstration. 4-144 MOMS 2149*7 - -N Section S REFERENCES 1. 0. Hutzlnger, S. Safe, and V, Zftko. The Chemistry of PCB's. loci Raton, Florida: CRC Press, 1974, pp. 143*148. 2. R.E. McKinney and A. Gran. Protozoa and Activated Sludge. Sewage and Industrial Wastes. Vol. 28, No. 10, Oct. 1956, pp. 1219*1231. TABLE: RESULTS Of PCB CONCENTRATIONS IN BACTERIA AUGHENTEO SLUDGE, ppm No. Pays Tank 1 Tank 2 Tank 3 Tank _4__ Tank 5 0 205 222 9B3 135 232 ...21 94 65 621 ... ... ...24 no 210 ...30 145 206 962 m+m ... ...36 -- 131 25B ...42 195 154 1449 ... ... ...50 165 275 ...57 255 162 1160 ***75 230 174 1443 82 253 157 1225 ... ... ...90 262 181 1476 91 163 2B7 4-145 . MOMS 214968 Discussion How or why did the Increese In ppm occur? A, The emulsion sludge Is a much lower density end It, contele* doth v live and dead organisms. Why was that particular strain of Pseudomonas aeruginosa used? A, That strain had been used In jet fuel tanks, we so decided to use the same on the sludge. Is 90 days enough time? A. That was the allowable test period. d-i'ie HONS 214969 part 5 PCB DESTRUCTION *ONS 214970 r r DESTRUCTION Or HICH CONCENTRATION RCl* IN A UTILITY AOIUR T. E. Sladhoff, C. A. lain. and R. E. Morrla Union Elaccrlc Company INTRODUCTION Aa with noat alactrlc utllltlaa and othar lnduatrial concarna, Union Elactrlc Company producae algnlflcant quantltlea of waata oil and othar combuetlble pro ducta during the ceuraa of normal buelnaae activity. The graataat majority of thaaa waataa coaa from two primary aourcaa; 1) dlaloctrlc fluida uaad In electrlcal equipment. and 2) lubricating olla uaad la aotor vehielee and large powar plant machlnary. Bacauaa of pravloua manufacturing and aervlcing procadurea, aoma nlnaral oil dlaloctrlc flulda will Inadvertantly contain tha chemical com pound polychlorinated biphenyl or PC> In varying but generally low concantratlona. Othar dlaloctrlc flulda. known fanatically aa aakaral. Inten tionally contain about $0 percent PCIe for a epeclflc purpoaa and uea. In coaparlaon to many othar organic cbomlcala tha eclentlflc Information avallabla would aeon to Indicate that PCIa ara relatively lew In toxicity. Howavar. bacauaa of their extrema atablllty and bloaccumulatlva nature and tha Inexact nature af the eclenco In thla area, predent action raqwlraa that apaclal cara be taken te ana that thaaa flulda ara dlaponad of properly. On May 31. 1973 tha U.8. Environmental Protection Agency promulgated reguletlona (14 Ft 31514) rigidly governing tha handling of PCBa. A key feature of thla reg ulation la the grouping of PCI waate llqulde Into three catagorlaa with differing diagonal optIona for each group. Tha claaalflcatlona and their diagonal optIona arei HONS 214971 5-1 1) For thooo liquid* containing 500 ppa FCEa or graotar, diapoaal nuat taka place In an incinerator which coapllaa with the provision* of 40 CFR 761.70 ("PCB incinerator" requirement*). 2) For thoaa liquid* containing between 50 and 500 ppe PCBa char* are three diapoaal optional burning In a FOB incinerator; placement In a landfill which coapllaa with tha provlalona of 40 CFt 741,75 ("chaaleal waata landfill" raqulraaanta), or Incineration In e high efficiency boiler mating apaclflc raqulraaanta. 3) Finally for thoae llqulda containing laaa than 50 ppa PCBa, no direct raatrlctlona era placed on thalr diapoaal except that they cannot be uaad ea aaelente, coatlnga or duat control ngante. With tha proaulgatlon of thle regulation Union Electric began to anaain* which optlona boat aat tha Company'a diapoaal naada and objectlvaa. Continuing and aubatantlal llabllltlaa for thaea waataa even after proper diapoaal In accordance with ragulatlona Bade landfilling uadaalrabla. However, PCB liquid Incineration coeta ware high. Therefore, Union Electric decided to pureua tha high efficiency boUar-diapoaal option for contaalnatad oil. Thia alternative waa attractive beeauea It allowed Union Electric to mlntaln control of tha eubatanca through lta ultlnata diapoaal. Aa wall, tha raaourca waa not waatad in that tha energy value of the oil waa recovered through ganaratlon of electricity. On July 7, I960 Union Electric raqueetad approval to dlepoaa of 50 to 500 ppn PCB llqulda In thalr Labedle Plant Unit #4 located near Labadla, Hlaaourl, During Novanbar i960 public metlnga wara held and In aarly January, 1941 a aarlaa of trial burn* war* conducted while the eteck affluent, eah. and anblant air In tha aurrouadlng area war* carefully aonltorad. No PCI* ware detected In any of the** aaplaa. On May 24, 1941 tha Hlaaourl Dapartaant of Natural laeoureaa certified tha boiler a* a raaourca recovery facility and on tha following day approval under the PCB rule waa granted by EPA Kaglon VII. MONS 214972 5-2 f r With this approval Union Elactrlc aolvad lta PC) contealnetad oil dlapoaal problem. However, It did not eolva tha problem of dlapoelng of thoaa liquid* greater than 500 ppa and aapeelally aafcaral, for dlalactrlc flulda containing graatar than 500 ppa PCBa, tha ragulatlona dleellow tha high efficiency hollar aathod of dlapoaal, and dilution to lower PCI concantratlona la alao daaaad lllagal. Aa pravloualf notad, dlapoaal of thoaa flulda auat taka plaea In an lneinarator which aaata atrlngant atandarda aatabllahod In tha ragulatlona. However, thla requirement la not abaoluta. Tha Regional Adulnlatrator haa tha authority uadai 10 CfR 761.10(a) to approve alternative dlapoaal oathoda If thay can ba ahown equivalent to a PCB Inclnarator and If thay will not praaant an unroaaonabla rl*k of Injury to haalth or tha environment. Onion Elactrlc carafully examined tha daalgn and oparatlonal faaturaa of tha Labodla Unit #4 and dateralnad that ouch a danonatratlon could ba mada for thla facility whlla burning an oll/atkaral bland containing 5 parcant PCBa In conjunction with pulvarltad coal aa tha primary fual aourca. Therefore In Docanbor, 1981 Onion Elactrlc lnltlatad formal dlacuaalona with EPA tagIon VII which ultimately raaultad in approval of tha tyatan praaaetad horaln, Thla papar daacrlboa tha anglnaarlng analyala of tha daalgn faaturaa and opara tlonal paraaetera which load to tha condualon that tha bollar can aaaantlally daatroy nil PCBa Introduced and tha raaulta of a carafully plannad taat burn which waa conducted to conflra thaaa condualone. BOILER DESCRIPTION Labad1* Unit 54 la tha naweat of four 600 nagawatt (MW) coal flrad ganarating unita located on tha aouth bank of tha Mlaeourl Rlvar In franklin County near Labodla, Mltaourl. Tha bollnr (fliura 1), tdileh la aaaantlally Identical to tha other three bo liara at tha alta, la a tangentially fired unit oenufeetured by Coafaoetlon Engineering Inc. and waa placed In commercial operation In Auguet of 1973. Tha bollar haa a maximum contlnuoua rated (NCR) heat Input of 5387 million BTU par hour tdileh correaponda to burn rata of about 240 ton* par hour of pulvarlaed coal. The MOMS 14973 5-3 Figure 1. Sect lone 1 View, Lebedle Unit 14 Idler SOURCEt COKIUlTION ENCIXEERINC MONS 214974 5-4 r r pulverised coal la transported In air suspension from tha coal sills to six levels of burners located at aaeh of four eornara of the boiler (Figure 2). The coal la Injected Into the furnace through the burnare In auch a nanaer aa to lntlnatali nix with tha secondary air and to fora a large vortex about the boiler'a vertical axle thereby creating a huge, rotating ball of flame. 011-flred ware-up guna are provided between the bottoe two coal burner levela for boiler warn-up and to aaalat In atabla coal Ignition during atert-up. One gun la located In aaeh corner of the boiler. At full load tha flaw tenperatura within tha fire* ball la about 3000*F and gaa tenperaturaa at tha euperheeter dlvlalon panala ap proach 2400*F. SYSTEM DESIGN In preparing to burn FCBa at labedla elgnlflcent nodlfleatlona were node to tha boiler and plant at considerable axpenae prior to the Initial teat burn of contaelneted oil. Tha purpose In naklng these changes was to provide a wsll-dealgned aystee solely for PCI waste oil which mlnieliee the exposure of thla oil to the environment and Insures Its propsr disposal. After approval to burn up to SX PCla It was decided to further modify the system to Include additional storage capacity and blending facilities. Tha following la a more detailed deecrlptlon of tha various components of this system (see Figure 3). Waste-Oil Storage and llandlni In order to store rci contaminated oil at the plant until sufficient quantities for blandlug end Incineration have been generated, an existing tank located above ground and Immediately south of the plant has been designated and prepared for this purpose. In addition an adjacent tank has been designated for blending SX PCB batches for Incineration. Both tanks sera previously used to store No. 2 fuel oil, ware built to conform to the American Petroleum Institute (API) Standard 6S0 and meat all applicable OSHA requirements. Tha oil tenke are surrounded by a 5 foot high concrete block dike designed to contain the entire volume of both tanka. In addition, tha areas between tha 5-5 ' MONS 214975 rr ukmim mmoi MMNIA KWMOX OAMHItt IffCONOAftV AIAI wmn o*ivi unit eOMKCTWO UMR mUMtt MCCNAMWI MTAACTAAL* OK. SUN MCBMMWI WtmMwBrtSK U*KACI WALL TUIIt ltU*C*AMILl IOMITOA NOAM umift HDlllI <COAL CMMAATMfNTI tumutiDozzu kxuaux am eoMrr i tCANMR MIAP L*VNTir Figure Z. Cutrr View, Typical C-E Tllttag Tngatl*l Junior Areoubly 5-fc somcct cohuistio* bhgikeminc 214976 Building CONTAINMENT TRIP VALVE HEAT EXCHANGER fM?-1 FORWARDING FLOW MEIER A BUMP JRECORDER DINED AREA rL(wr* )i UN4U ft* llmdlni ml lucliwrdloa FtclllilM ouncei union ILKCTK 1C HONS 214977 tr tank* and walla. Including tha Internal alda* of tha dlka hav* baan llnad with a 30 all dacron rulnforced chlorlnatad polyethylene liner. Tha liner la covarad with ala Inch** of aaooth atona to protact It froa daaag*. Whan plane paraonnal ar* ready to conduct a burn, contaalnatad oil la transferred froa the PCB contaalnatad oil stores* tank to tha blending tank using a paraanantly lnatallad tranafar puap and aatar located within the coatalnaent, Mounted In tha canter and fra* atandlng on tha bottom of tha blending tank 1* an addy jet mixer with a aarlaa of twelve double acting Jat notalaa. A positive displacement circulation pump adjacent to tha tank remove* fluid* froa tha tank and discharge* than back through tha Jet mixer. Whan tha aalactad quantity of PCI contaminated oil la pumped to tha blending tank, th* pump 1* snarglcad and circulation la established. A carefully calcu lated quantity of aakaral sufficient to produce a 31 PCI/oll bland 1* than lntroducad. After all th* aakaral la Introduced, circulation and mixing are con tinued until th* fluid* ar* fully blended. Upon completion, sample* are col lected at various depths to lnaura uniform mixing and proper PCB concentration. Hasting value, percent chlorine and flash point are also measured. When proper condition* have been confirmed tha batch la then reedy for incineration. West* Oil forwarding Tha vest* oil forwarding equipment la that equipment necessary to transport tha oil froa tha blending tank to the boiler during the Incineration process and In cludes the west* oil puap, totalising flow meter, strainers, a temperature regu lating valve, heat exchanger, and all associated piping, valves, gauges and control*. Piping la routed fro* th* tank to Just Inside the plant building where tha oil flows through e duplex basket strainer for solids removal to protact dovnatrean equlpmant from daaag* and enhance etomlcatlon. Th* oil then flows through a positive displacement puap nominally rated at 750 gallon* per hour. After dis charging from tha puap, the oil flow* through a heat exchanger to ealntaln a MONS 214978 5-8 r r coneletent oil vlecoelty regerdleea of ambient temperature extreme*. Tha haat exchanger la preaant to inaura optimum atomization and coabuatlon at tha burner. Hovavar, undat normal condition*, tha vlecoalty of tha PCS/oil bland la totally accaptabla and tha haat exchanger la not currantly uaad. Tha final placa of equlpeent la tha totalltlng flow aatar to racord the volume of amata oil being Incinerated. Thle entire aqulpeant package 1* aurrounded by 12 Inch high curbing to contain any oil which night look or otherwlee aecapa. A apaclal aaargancy raaponaa and tplll cleanup atatlon la located nearby, A written SPCC plan which Include* provlalon* for tha Incineration ayatan ha* alao bean prepared and Implanon tad. Want* Oil Burner Tha oil la than tranaportad through color coded piping up to tha aouthaaat corner of tha boiler where It pnaaaa through a trip valve and on to tha want* oil burner, Tha prlnary alanant of tha burner la tha oil gun. Tha waata oil and atonlalng air pace through tha gun and or* nixed at the burner tip *e they ar* Injected Into the boiler. Tor tha Mata oil ayatan, a apaclal Coabuatlon Engineering WRTK typo oil gun equipped with a *J* atyla noccl* haa boon purchaaad from tha hollar nanufacturar. Tha gun utlllta* an outald* nix arrangement wherein tha oil and air travel down tha gun In parallel. non-concantrlc pa***|*a and only nix at tha apray noctle. Contrary to otandord gun doolgn tha atonlclng air la placed at tha outald* of tha nixing plat* which promote* batter nixing. Tha nanufacturar ballavaa thl* gun will atonlaa tha fual to a flnar droplet alto, . will achieve greater flam* penetration and la nora tolerant of flame and fual fluctuatlona. Thl* gun la maintained aa a aaparata device which nuat ba inetellad In placa of tha normal Mrn-up oil gun whanavar Incineration of waata oil la required, A eacood apaclal aaargancy raaponaa and aplll claan-up atatlon la located near tha burner aa an additional precaution. Waata Oil gyatam Control* In order to Integrate the Mate oil burner Into tha hollar ayatan. appropriate controla have bean added and nacaaaary interconnection* have bean mad*. A trip HONS 214979 5-9 valve, identical to that ueed on the warm-up oil aupply line, la lnatalled In the veate oil piping at the point where It connecte Into the exletlng warm-up oil line near the burner. It la controlled by a eelector ewltch In the control room. Thle ewltch la ueed to eelect the normal warm-up oil burning mode or the epeclel weate oil burning node. In either poeltlon, tha ewltch locka tha trip valvm which wa not ealecced In the doeed poaltlon. All atert and trip elgoala effect only tha valve eelectad. The ayaten will operate ea It preeantly dene In the normal aode. In the weate oil burning mode, tha eoutheaet corner burner oparetee Independently of the other gune end dome not hinder their operation, Controla nuat confirm that certain operating condltlona e*let before they will allow tha operator to enter the weate oil burning noda. The axletlng trip eignele ere ueed to cloaa tha waeta oil trip valve If any of in the following evente occur. 1) air/fuel differential preaaure leae then 5.5 paid 2) oil preaaure leae then *0 pelg 3) emergency ehutdown commend due to boiler trip caueed by: a. elr flow laee then 25 percant b. loae of both I.D, or F.D. fane c. Inadequate water well circulation for morn then 3 eeconda d. high furnace preaaure a. turbine trip f, loea of AC power for eoro then 2 eeconda g, loaa of all fuel h, flMa failure 1. manual Margene y trip Additional trip elgoala are provided to trip tha weate oil valve wham 1) the exheuet gee CO concentration la greeter than or equal to LOO ppm; or 2) the axheuet gee Oj concentration fella below 3 percent; or 3) the unit'a generation fella below 300 MW. MONS 214980 5-10 r r Tha l>ttr trip algnal la provided to lndlractly control tha (aa taaparatura and raaldanca tlao i.o wall within llalta aatabllahad for eoaplata daatruetlon. Tha atoalclng air aupply during waata oil lnclnaratlon uaaa a aaparata praaauro ragulatlng valva (MV) for tha aouthaaat burnar rathar than tha comm PRV oaod for burning warn-up oil. Tha now FRY la aat to control tha air praaaura at 15 pal hlghar than tha oil praaaura. Tha atoalalng air piping haa a trip valva ldantlcal to that In tha oil piping, which cloaaa undar all of tha aaaa trip condltlona, thua, atopplng tha flow of air to tha gun. Coabuatlon Monitoring Equipment In ordar to nonltor and racord tha lndlcatora of high coabuatlon efficiency dur ing batch PCB/oll lnclnaratlon, aavaral dlffarant analytere and aaaoclatad nupport lnatruaantatlon ara uaad. Oxygon Oxygon lavolo ara normally required aa a hollar control Input for maintaining propar fual to air ration. Tharafora, aa part of tha hollar control daalgn a Haatlnghouaa Hagan Modal 218 oxygan analyear ayataa la lnotallad In tha hollar backpaaa downatraaa of tha aconoaltar, but prior to tha air haatara. Tha parcant oxygan algnal la raturnad to tha control roon whara it la lndlcatod on a contlnuoua atrip chart racordar. Carbon Dioxide Contlnuoua CO^ analytar raaulta froa tha two taat burna daaonatratad that COj 1avala in tha hollar ara eoaalatantly In tha ranga of 14 to It parcant aa would bn anticipated. lacauaa of tha high pradlctablllty and low variability of tha c2 lavala and thalr lnaanaltlvlty aa a coabuatlon indicator, contlnuoua aonltorlng of thla paraaatar on a routlna baala la not raqulrad. 5-11 i MOHS 214981 Carbon Monoxide In ordar to non!tor CO concentration*, a rapraeantatlva aaepla la extracted froe a location adjacent to th* oxygon analycar probe. Thla extractive aaepla flrat paaaa* through a aaepla conditioner where particulate* are filtered end aolatura la raeovod. The conditioned aaepla than paaaa* through a aaepla line co a Norib* Modal PIK-2000 Juft-type nondiaporaive infrared (NDId) CO enalyxer. fba algnal free the analyter la aant to th* control rooe for atrip chert recording. Sine* carbon dloxlda la an exhauat gaa conatltuent which 1* conalatently In tha rang* of 14 to 16 percent, and alnca thle gaa acta aa a low level Interference for NDIR analyaera, th* calibration gaa ueed with th* ayatae ha* a carbon dloxlda leval of 15 percent to affectively allelnata thla Interference. Toeparatura and ftealdenca Tine It la wall docunantad that the right coehlnatlon of coebuatlon teeparatur* and retention tine will lnoure conplete daatruction of PCBe. For thle raaeon tha FCB Incinerator atendard* require contlnuoue eaaaureeent of th* gaa taaparatura. Bacauae of th* very high temperature* and the abraalva and corroalv* nature of tha atack gaa envlronaent It would be very difficult to lnetall a contlnuoue teaparatura eaaaurlng device In a utility hollar. Metal probe* alon* cannot wlthatand the elevated tenpereturea. Cerealc probe* cannot conalatantly wlthatand large teeparatur* cycle* and eventually bacon* brittle and break. Beceuae tha boiler tube* are deelgned to atructurally act *a the boiler well and alnca they alvey* have a flow of water or ateen through then, a natal laplented tharnocouple In tha tube well would read netal taaparatura which la lower than gaa taaparatura. optical pyrometer* need a target from which to aeaeura ealaalvlty and require aona Interpretive Judgement In their reading. If boiler ilag 1* not preaent, which la uauelly the cat* at lower load*, the only aurfecoa availa ble are boiler tube*. Aa noted above, the tub* netal teaperature doaa not ac curately repreaant gaa teaperature but 1* generally lower. Therefore, the only MQNS 214982 5-12 r r technology available la water cooled thermocouple pro bee which wuld be very axpenelva to lnatall and vary ealntenance lntenelve. Later dlacueelon In thla paper clearly raveala that at norwal loede bath teaperetura and reeldance tlan alninun raqulreaente ere exceeded. Therefore, It logi cally follove that a reliable alternative to contlnuoue temperature neaaureaant le to eatabllah a elnlevu* load below which no PCB incineration would taka place. Baaed upon teaperature and realdanca tlea data, a load Halt of 300 MW will in aura boiler condltlone at all tleaa in axcaaa of eatabllahed elnleua raqulreeanta. Thla conclualon la conflrnad by the teat burn reaulta. Waate Oil Feed Rata While not a PCB Incinerator requlreaent, an additional precaution to lnaure conpleta FCB deatructlon la to llelt the feed rate of PCBe to a level ouch aeellar than the total fuel Input. High efficiency boiler requlreaenta for PCB contami nated oil incineration eatabllah thla level at 10 percent of the total fual. Cone 1 daring the aaxlaun capacity of tha waate oil puep and the eaount of fual required to achieve the nlnlnua load of 300 MW, It la eetleeted that the hlghaat ratio of actual PCBe to total fuel Input would be epproxlaately 0.1 percent. Operating Dealtn end Proceduraa tn addition to the phyelcol nodIflcetlona deecrlbed for proper PCB dlapoaal, well thought out operating proceduraa nuat be eatabllahed to lnaure proper coordina tion end underataedlng of the equipment by tha operating peraonnel. To lneure thla conalatent operation, written proceduraa have been prepared and are atrlctly adhered to during the incineration procaae. tn order to aialnlae expoaura of peraonnel to roelduel concontretlono of PCBe during operation end maintenance a flushing eystee la provided beginning at tha tank md flowing through tha ontlra ayaton and Into the boiler. Whenever the ayaten la ahut down It la fluahad with Ho. Z fual oil for aufflclont tlno to In aura at leeat thraa equlvalant voluaaa of tho ayaton hava baan punpad and burnad. HONS 21*93 5-13 During etartup the eye tea 1* Initially oparatad on fual oil. A coaplate walk down lnapactlon la aada to chock for lncorract valving or laaka and tha waata oil ayatan la not valvad In until tha ayatan coaplataly chacka out. Parlodlc lnepectlona ara alao aada during tha Incineration procaaa. Aa wall, varloua Important oparatlng parameter* ara audibly and vlaua)1/ alarmed in tha central control room. If an alarm occure, oparatlng poreonnel ara lamedlately dlapatcbed to lnapact and correct any malfunction. COMBUSTION EVALUATION During tha Initial avaluatlon of the bo liara1 ability to combuet FOB compounda a number of factora affecting thle combuatlon were carefully evaluated. On a purely Intuitive beaie the baalc daalgn would auggaat tha Labadla boiler aa an excellent FCB Incinerator. Tha boiler ltaalf la approximately the height of an alghtaan atory building (183 ft.) and hae a total furnace volume of 422tOOO cubic feat. Tha flame ball that apane tha burner alavatlona la approximately 36 foot In height with a eroae-eectlonal area of 2765 aquare feat. Tha Injection of tha PCB/oll bland oceura at an oil gun level which la located between tha lower two coal bumero In one corner of tha boiler. The oil nuet pace through the highly turbulent flame area and eplral up through nearly tha entire furnace volume. Under uorat caaa condltlona tha relative feed rata of the waata oil to coal la vary low and on a weight baala tha fraction of FCBa to tha total fuel In put la only 0.001. Beceuee of tha very large quentltlea of coal burned In utility bollora, It la aaeentlal that they bo dealgned with high combuatlon efflclencleo for economic raeaono. Minot rmductlone In combuatlon efficiency on a boiler tha alta of Labadla Unit #4 can anally add up to hundrada of thoueande of dollara par year in additional fual coeto. Obvloualy than, there la a atrong motivation to daalgn and operate tha unit at Ita naxlnun efficiency. Carbon monoxide and carbon diox ide data from the pravioue toot burnt dearly denonatrato thle fact. Tha teata ahowad that tha combuatlon efficiency of Labadla Unit #4 everagae 99.99 percent. 5-14 MOMS 214984 r r In order to aaka a a ora tachnlcal avaluatlon of tha potantlal for PCB daae ruc tion, tha datalla of tha varloua coabuetlon nechanieae auat ba examined. Tha Conbuatlon Procaaa Conbuatlon la on exothermic raactlon in which fuel la rapidly oxldlned thareby releaelng potantlal chealcal energy In tha torn of heat. Thara ara thraa baalc coaponanta of tha conbuatlon procaaa; fual, oxidant and dlluant. A fual la a compound containing anargy rich bonda auch aa carbon-carbon or carbon-hydrogon bond*. Whan an oildant, auch aa aolaeular oiygan, la praaant undar tha right taaparatura condltlona, tha conbaatlon raactlon occura, ralaaalng algnlfleant aaounta of anargy, Slnca air la tha aoat common aourca of oxygon and alnca It contalna other alaaanta In lta aolaeular coapoaltlon, (the Boat notabla of which la nitrogen)) thaaa other alaaanta era praaant during tha coabuetlon procaaa. lapurltlee in tha fuel ara aleo praaant. Thaaa additional aubatancaa ara conoidarad dlluanta and do not participate chemically In tha raactlon but rather In fluence It by haat abeorptlon thereby acting aa a thermal alnk. In addition to nitrogen, other coaaonly praaant dlluanta include exceea aaounta of oxygen beyond etolchioaetrlc requlresenta and fual contaalnanta auch aa aulfur, chlorine and aah. Tha praeanca of thaaa dlluanta are a practical reality to any commercial fual-burnlng procaaa, and they ara not conaldarad a limitation to highly offi ciant combuetlon. Coal and waete oil combuatlon In a utility boiler la an oxidation raactlon which muat have oxygaa praaant In order to occur. High taaparatura raactlona occurring within tha fual ltaalf whan thara la no oxygen praaant era known aa pyrolyole. Pyrolytic comdltioaa auat ba avoided alnca thay have potantlal to convert PCBe ' and othar organise Into aora toxic organic coapounde. Becauee of the Intlaato mix of fual and air in tha highly turbulent coabuatlou sons of a utility boiler, tha addition of sufficient axceea air and tha fact that thaaa potential bypro ducta ara deetroyed at much lower teaperaturee, thalr potential for ealaelon from tha Labadia boiler la oaaontially non-axlatant. S-1S MOMS 214985 rr On a macroscopic acala thara ara flva parameters which affact combustion; approprlata physical and chaalcal properties, availability of coabuatlon air, high (a taaparatura, long raaldanca tlaa and adequate nixing. Physical and Chemical Propartlaa Thorn ara two baalc physical propartlaa which affact coabuatlon af liquids auch aa uaaca oil; viscosity and tho alto and concantratlon of auapandad aolida. Tha lower tha viscosity tha aaalar It la to achlava flno otoalsatlon. Tha flnor tha drop! of oil tho groatar tha aurfaca araa axpoaad to tha haat of coabuatlon. Thla, In turn, allows for rapid vaporisation giving aora tlaa for aolacular con tact with tha oxidant and driving tha raactlon towarda aaxlaua coaplatlon. Tho viscosity of tha PCB/oll bland la alallar to No. 2 fual oil. In addition a haat axchangar haa boon lnatallad In llna In ordar to maintain propar viscosity prior to tha oil ranching tho bollor. Finally, a apaclal gun which provldaa flnar atomisation and which la aora tolerant of changaa In vlacoalty haa boon purchaaad for thla purpose. It haa baan shown that auapandad aolida In tho fuel can lapalr atomisation and cauaa plugging and buildup on tho burner tip. Experience above that tha blended oil haa negligible auapandad aolida. In addition, a duplex filter haa baan ln atallad ahead of tha puap to further protect agalnat auch occurancoa. Iaportent chaalcal propartlaa to be considered ara elemental composition, nolatura and heating value. Typical contaalnante for tha waota oil and aakarel ware characterised. It vae door from thla exaalnatlon that no contaalnenra or eoletura ara praoant at lovola of algnlflcant concern. Tha heating value of tho FCB/oll mixture la on tha order of 19000 Btu/lb from which it la apparent that aalf-auatalnlng coabuatlon can be eaally accoapllohad. However, by ayalaa daalgn, a coal supported flaae of auch greater magnitude will always be praoant. 5-16 MQNS 214986 r r Conbuatlon Air A* not'id earlier, although coneldarad a dlluant to tha procaaa, ftcaaa oxygon above ldaal etolchlonatrlc levelo la a nacooelty in tha roal world. This la tha caae bacauaa of tha phyelcal llaltatlona In eccoapllahlng complete aad iaetantanaouo nixing of fual and oxidant at a molecular level. Tha additional oxygon la praaant In ordar to lnauro that tha oxidation raactlon la aaxlolxad. However, dlluant acta aa a thermal oink to tha procaaa and If praaant in quentltlae too graat. It will limit tha raactlon taaparaturo and lncraaaa tha uaa flow through tha hollar, thereby raduclng tha roaldonco tlao. Therefore, an optlaltad balance auat ba atruck wharaln auffldant taaparaturaa, raaldanca tinea, and oxygon lavola ara all maintained. Tha requirement* for thla balance In a PCI Incinerator ara a minimum 3 pareant exceee oxygon level while Maintaining a gao taaparaturo of at laaat 2192`F (1200*C t 100*C) at a raaldanca tine of at laaat 2 eeconde. Normal oxygon level for tha Labedle boiler la batwean 3.5 on 4.3 percent. Tha controle ara arranged to trip tha waata oil feed ehould tha oxygen lavola fall below 3 percent. Therefore, It can ba safely aeeuaed that eufflclent oxygen will elvoyo ba praaant. Temperature and Boaldonco Tina Bacauaa raaldanca tine la ao dependant upon taaparaturo, they ara dlacuaaad together. Aa noted above, tha alnlaw acceptable regulatory lavola of taaparaturo and reaidanco time for complete PCB daatruction ara 2192*F and 2 eeconda raopactlvaly. Scientifically e taaparaturo of 2000*F and 2 aeconde roaldonco tlae ara coneldarad adequate. Heat tranafar, aaoa flow and raactlon rataa all lncraaaa with taaparaturo. Thorofora, all other factora being equal, tha higher the taaparaturo tha aoro HONS 214987 5-1? rr couple to the eoebuatlon. Obvlouely tnen, It la highly daalraabla to charactarlta the teeperatura reglaea within tha hollar aa thoroughly aa poaalbla. Eloawhare in thla papar tha difficulty In obtaining accurata eeaaarementa within tha eoebuatlon anvlronaant of a utility ataaa-ganaratlng bollar la dmacrlbed. However, a roaaarch project daalgnad to Invaatlgata bollar a lagging problama waa conducted on Labadla Unit #4 during 1977. Tha project waa conducted by Battalia Lahoretorlee wider contract to tha Electric Power Roaaarch Inetltute (EPR1) and Included an antlra aarlaa of taaparatura travaraaa at varloue polnta In tha bollar. Tha temperature aaaauraaanta ware aado ualng a ahleldad, high velocity theraocouple probe which waa water cooled. During thla aaaa period optical py rometer aeaaureaante ware tekan and correlation between the two waa ehovn to be excellent. All maaeureaenta ware taken between 540 and 560 MW with exceea oxygen ranging between 4.6 and 5.2 percent. Ovar 120 Individual obeervatlone ware eade. later, additional optical pyroeeter neaaureaenta ware made on eeveral Labadla bollara at four different loada ranging fro* 300 to 565 MW, Although the aeaaureaente ware taken on varloue Labadla unite, tha bollara are eeeentlally identi cal and the data ahould be repreaentatlve of Unit #4 condltlone. It ehould alao be noted that any Inaccuracy In theae eeeeureeente la conaarvetlve In that tha error introduced la generally conaldered to be on the low aide bacauea of the poaelblllty of experiencing the cooling effect phenomenon due to water end ataee flow through the boiler tubea. eeed upon thla temperature data and knowledge of the boiler elevetlona at which theae tampareturee ware taken, calculatlone were performed to determine raaldenca time at varloue polnta within the bollar. Cae flow at etendard condltlone and 4 percent emeeee oxygen wee determined ualng etolchlometrlc calculatlone baaed upon "aa burned* ultimate coal analyaee. Finally, ualng boiler dlmenalena and atertlag from the wuete oil burner level, e temperature end reeldence tlee profile waa determined at four different loade. Figure 4 auamarlua the reeulte of theae ealeuletloee. HONS 2X4988 s-ie RESIDENCE TINES t TEttEUTUDES MUM* A millUM Loins load 3W r* Hi m sgffr m--m- &r-m------M- &t-w--m- ag Wr ua--m- JWSL rf) 2A70 JLfifi- IA0 nuicts adapted rtox COHtUETIO* EMCI MAIN INC riiun A. DID# Elavttloo of Controlled Circulation SUM Caoorator for Labadlo Ualc# HONS 214989 r* Table 1 below above the temperature of the gaa after a two aacond realdanea time and tha average temperature of tha gaa during thla two aacond period. Thle in formation damonetratee that temperature and raaldenea tlaa raqulreaante for eoaplata PCI daetructlon era exceeded in thla boiler even at lower laada. Aa pravloualy doacrlbad, a minimum load laval of 300 MW haa bean eetebl lotted which lnauree that proper daetructlon condltlona will alwaya axlat wlthla tha boiler when Incineration to baa place. TAILS 1 CAS TEMPERATURES (*F> AT A TWO SECOND RESIDENCE TIKE Unit Unit Load Gaa Temperature 9 2 Sac. Realdanea Time Avaraga Temperature Durlnx 2 See. Time 1 300 4 375 3 400 2 460 3 365 4 500 2154 2251 2233 2183 2371 2351 2412 2484 2483 2480 2606 2634 Hlxlna The final parameter which la key to eoapleta daetructlon of PCIe la mixing. Aa pravloualy noted, In order for tha eoabuatlon reaction to occur, tha fuel and oxidant muat elx at a molecular level. Thla mixing la achieved by the creation of turbulent flow. Aa Inherent feature of the tangential firing uachanlom of the Lahad la boiler la tha turbulence which occura whan tha coal/alr mixture la blown Into tha boiler at high velocity. Tha vortex aetebllahed reeulta In a large awlrllng ball of flame. Action of adjacent fuel etreaaa on each other eugmenta tha highly turbulent condition and thla turbulence aa well aa oxidation, la fur ther encouraged by tha Introduction of Jota of oacondary air around each burner. 5-20 MONS 214990 r r The PCB/oll bland la Injected at a low point In the flane ball and therefore la axpoaad to tho aaxlnua elxlng poaalbla. TEST BURK ~ In tha early etegea of the projact EPA Region VII roqueatad a foraal taat plan ba ubaltted for ravlaw and approval. Raeognlxlng tha naad to ahow equivalency to a PCS lnelnarator and tha lnharant dangara In puahlng laboratory anelyeaa to thalr abaoluta aonaltlvlty llalta It vaa decided that tha aanpllng eethod aalactad nuat collact aufflciant eanple to danonatrate approprlata daatruction lavala and alnlalca potantial for falaa poaltlvaa. Aftar txtanalva otudy and avaluatlon and in conjunction with Envlronaantal Sclanca and Engineering, tha atack taatlng coneultant, a conprehenalve taat plan vaa praparad for aubalttal to EPA. Taat Plan Tha baalc acopa of tha taat progran conalatad of conducting four taat burna with tha flrat taat balng conductad without burning tha PCI/oil bland to aarva aa a baaallna. Tha next thraa taata would ba vhlla burning 5X PCSa In oil with bollar loada at or oaar 300 MW (l.a. worat caaa conditlona). During aach taat burn, aanplaa of atack anleelona, fly aah, and bottoa aah vara collactad aa wall aa varloua nupport aanplaa auch aa wnata oil, coal and rlvar watar uaad for alulclng aah. Oxygon, carbon dioxide, and carbon nonoxlda wara contlnuoualy and lndopendantly nonltortd In tha bollar backpaaa by tha atack taatlng conaultant. All aanplaa axcapt tha coal aanplaa wara analyaad for PCS content. Stack gea, fly aah, botton aah, and rlvar water ware aleo analyaad for polychlorinated dlbenao-F-dloxlae (PCDDe) and polychlorinated dlbanaofurane (PCDFe). Two aaparato aanpllng tralna ware uaad for collection of atack gaa aanplaa. Both tralna ware Identical with tho exception that tho organic tropa In tha PCS train ware packed with Florloll who race tha trapa In tha PCDD/PCDP train warn packed with XAD-2 raaln. Backup trapa ware provided to check for paoolble breakthrough. Tha approach aalactad for collection of atack gaa aanplaa conalatad of anploylng a high voluno lourco aanpllng (HVSB) vorolon of tho EPA Method 5 train 5-21 HONS 214991 appropriately Modified to collect PCBa end other chlorinated organics at high aeepllne rates. The train design (Figure 5) aa aaiantlally baaed upon the Methodology described by Beard & Seheuu In EFA'a lnterle policy guideline doeueent "Saepllng Hethoda end Analytical Procedurea Manual for PCB Disposal! Interla Report, USE PA office of Solid Wests, February 1978.' However the Method vea Modified by ecellng up the dealgn to collect aaaplea at a nominal 5 ofa flow rata as oppoead to the More atandard l cfa. Collected stack aaaplea uere extracted with hexane using a Soxhlet extractor fol lowed by concentration and aulfurlc acid cleanup. Quantitative analyala for PCBa wee performed using gaa chronatography with electron capture detection (CC/ECD) with confirmation aa nacaaaery by gaa chromatography naae epeetroeetry (CC/MS) In the selective Ion Monitoring (SIH) node. CC/MS-SIM wee also used for determina tion of PCDOs and PCDFe. Extensive quality aeaurence waa established both In the field and In the laboratory. For example elaborate glassware cleanup was conducted prior to eoch teat. A third sampling train waa pieced on the atack during each burn to provide a direct cheek for contamination. All equipment use calibrated to MBS tracoebla atenderdo lmeedlately prior to the teat burn. In the lab ealaeted samples ware analyied In duplicate. Splkae and blanks ware analysed as wars EPA Quality control Semples at three levels of detection. For a aore complete discussion of the sampling and analysis protocol, aae a paper by A. J. Polcyn, at el titled 'PCB Waste Destruction Study! High Efficiency Sollar* presented at the APCA Specialty Conference on Measurement and Monitoring of Non-Criteria (Toxic) Contaminants In Air, (March, IMS). TEST RESULTS The teat burns were conducted on Hay 19-2*, 1982. Pour teat burn runs uere orig inally planned. However, failure of one of the eteek sampling trains to east EPA Method 5 criteria for e poet-toot leak cheek voided the fourth run requiring that a fifth run be conducted. MQNS 214992 5-23 HONS 214993 r' PCS teat raaulta ere presented In Table 2. PCDD end PCDF teat raaulta era preeentad In Tabla 3. In auaaary, the teat reaulta how that no PCBa, PCDOa or PCDFa ware dotectad In anp of tho anlaalon aaaplaa nor In any of tha botton aah or fly aah aaaplaa collactad during oach of tha taat burne. Baaed~vpon thla lnforaatlon It can ba atatod that tha labadla Unit #4 bollar deaonatratad a PCS daatructlon afflclancy of graatar than 99.99999 parcant. All of tho raaulta glvan for tho PCBa leaving tho bollar aro praaontad aa loaa than tha atatod volua for tha raepeetlvt collactad aaaplo. Thla preaentatlon of raaulta lndlcataa that tha detection Halt of tha analytical method for the glvan aaaplo waa reached or that detection below tha atatod value wee Halted by Inter ference froa other organlce that could not ba raaovad without aloo atrlpplng tha organlce of lntaraat (PCBa). Where the latter eltuatlon waa true, ovary effort waa node to clean tha aaapla extrecta by eueceealve acid rtneee until no reeaone. bla further laprovaaent waa obaarvad. Thla nay be noted In tho cooo of tha PCBa reported In tha atack gao aaaplaa collactad during runo 1, 3, and 5 which era reported aa loaa than 600, 230, and 130 nanograaa, roapaetlvaly. note that al though Bun Pi waa tho background run, thla aaaplo waa obaarvad to have a higher level of Interfering organlce than thoea aaaplaa collactad during tha PCS burn rune. Although tha oil burned during tho background run waa off apac No. 2 fuel oil which whan analysed waa found to contain a low loval of PCBa (20 ppa) It la unlikely that tho quality of thla oil could have contributed algnlflcantly to tha total level of organlce In tha aaaplo. However, a large change In load waa ini tiated during tha latter portion of tha background run which could explain tha higher level of lnterferrlng organlce. All of tha control train oxtracte ware found to ba below tho eathod detection llnlt and therefor* deaonetrata that aaaplo train preparation, aaaplo recovery, extraction and analytical proeodurae did not Introduce any arrant eontaalnatlon or other interfering blaa. Aa wall, aaperata analyeee of each Florlell backup trap, which waro alao found to ba below tha nothod detection llnlt, euggaat that no breakthrough of PCBa occurred. 5-24 MONS 214994 SE-S TAILE 2 Avirnt* tiMriilni Lou* Unit* SMOcmi) tun #1 Run 12 nsinr (IimIIih) *00 "551---------- "HI---------- Rum ft* 313 uupllnl Fur Jo* Ml nut** 300 2*0 300 300 Avuragt (tuck Cu* Flou Rut* ACFM MCFH 1,*71,000 001,000 1,3)3,000 00,000 1.322.000 60,000 1,213,000 13,000 Totul tuapl* Cu* toluau BCF (iiurtl) 131.0 MCF 003.3 7.l 3*4.0 *30.0 702.0 ~/A M/A Totul Quantltp sf Hunt* Oil omul Culloaa 3,731 2,*11 3,200 3. *64 Run H 101 300 1,2*7,000 711.000 7*3.1 723.1 3.0*1 Kit In Uuttu Ml Totul FClt Intu lollur FClu In (tuck Cu* luarlu* FCI* In lack-up Ttuf* Totul FClu biting (tuck Dtatructlua CfflcluncF Klu in Hank (Cuutrul) Truln Kit 1* Flp Auh Klu In lettoa Auh ppa (/*3 *1 no *1 ! Ftrtuut no no/f* nt/** 20 .2S <600 <70 <202 M/D r30 <6 <0.* *7,000 *01.* < <30 <1* >.****7 <30 < <0.3 00,700 0*0.2 <230 <50 <3 > oo.ooooao <30 <6 <0.* /A m/a M/A N/A N/A M/A N/A <3 m/a 51,400 321.* <150 <50 <* >*.*91 <30 u <0.3 Klu in kluur Mutur____________________na/nl _____ <1L02 <0,03 <0.02 m/a * Ion F4 kti to 6* nlM Mu* to an unaccuptukl* p*t-t*ul luak chuck on thu Florlull (Kl^ truln, <0.02 **Thu fl]r auh unaptu for Run 13 wat lout Mu* to kru*k*|u la trunalt, thuruforu. Hu Run 1* ftp *ah ut Bplt Ut* unulpuud. Mutt that tout euaMitlua* Muring Run* #* ua* 13 *r* alallur. mtcui Ml. I M2. HONS 214995 TA*LR 1 tlMURir OF POLYCHLORMATID OUDttO-P-OlOXlH (FCDD) AMD POLYCHLORINATED DIIEMZOFURAN (FCOF) TEST RESULTS PiriMtir PC00 CaopaunOa In U) Train btract PCDf Coopaanda 1* XAD Trala btract PCM CoapsanOa la Pip Aah btract PCDF CaapouaA* la Pip Aah btract PC00 CaapaaaA* It Aottea Aah btract PCDF Coopo.nda la lottaa Aah btract PCM CaapouaAa la Pivar Uatar PCDF CaopouaA* la Rlaar Uatar Unit* B| a* /|t "*/* */* "/* ng/al *|/al Jr--Raa Vl --lu. Paa.lt* Ran FS <00 <00 <4i.J <42. J <*.4 <.* <0.A <0.4 <00 <00 <J4.2 <J4.2 <10.1 <10.1 <0.4 <0,4 _<M0 <00 <llj <l.0 <10.J <10.1 <0.4 <0.4 lourcn Ml, 1M1. Run #5 <00 <000 <14, J <14.7 <10.4 <10 4 <0.4 <0*4 MONS 2X4996 r r CONCLUSION >sa*d upon tha dail|n aodlflcationa, technical analytes and trlol_burn rsoulta, tha CPA Rational Adalnletraiur authorlted tha uaa of Labadis Unit t 4 far dascructlon of PCIa at a laval of SX In ainaral oil dlalactrlc on January 30, 1981, It was tha Agsncy'a conclualon that tha eyataa described "does not peasant risks of Injury to haslth or tha environment" and that tha 'disposal has bsau doaonstratad to provide PC* daatructlon squlvalant to on lncinarotor aastlnt tha ro quires onto of 10 CFR 761.70,* A* wall It la tha conclusion of Union Elactrlc that this meant of dlapoaol of PCBa provides a aonslblo and coat offactlva aaana of aolvlng a sensitive eovlronaental problaa whllo at tha saao tlao conosrvlnt a valuable snort? resource. 5-27 HONS 21499? Dlacueelon What wm the burn rat* u*a? a. 750 gpu. Wbat i tha local public reaction! A. Union Electric conducted eeveral public Meeting* and apeclal preee brief Inga, idilch reaultad In quite balanced reporting and general public acceptance* Have you done at coat-benefit analyela? A. Strictly epeaking, the proceea la Marginally beneficial aconoMlcally. Wbat la the plant tmrker attitude? A. The plant Manager net with the craw* to dlacuae procedure* uaed to Mlnleise expoaure. It haa been accepted by th* craw*. How Much tine have you actually been on line? A, There have been 5-10,000 gallon teat burn* and 5 at the 51 level. MONS 214998 r r ARC PYROLYSIS OF PCB* ELECTRO-PETROLEUM, INC. J. K. HITTLE EPI C. H. TITUS EPI S. 0. BLOIS PG&E Program Summary Electro-Petroleum. Inc.. In cooperation with Pacific Gas & Electric, has undertaken a three-year program to design, construct, demonstrate and commercialize a high temperature DC arc furnace for the destruction of PCBs and PCB contaminated wastes. The technology Is based upon the utilization of a direct current arc Impinging upon a molten metal bath. The system Is deslQned so that Initial decomposition takes place In the molten bath and the resultant gaseous products are passed In the vicinity of the high temperature arc. The system reduces the PCBs to HCL, CO, Hj, and carbon by thermal pyrolysis. When solids are added to the system, they will melt and slag In the molten pool. The arc furnace system Is made up of three major subsystems. They are. the electrical system (power supply), arc furnace, end gas handling equipment, `lajor effort Is underway to site the prototype system, provide for adequate environ mental protection and develop analytical procedures for evaluating the process, as well as obtaining appropriate permits. Prooram Objective The objective of this project Is to develop a mobile alternative technology to conventional Incineration which can destroy PCB solids with an efficiency accept able to regulatory agencies, at an acceptable cost for construction, operation and maintenance, and with environmentally acceptable solid waste and waste streams. Olscusslon PCBs have high thermal stability and Federal legislation requires combustion temperatures In the order of 1600 C. for total destruction. Conventional com bustion systems have several major problems which are: difficulty In maintaining 1600 C. temperature, generation of larqe gas volumes to be scrubbed, and difficulty In handling solids. 5-29 MONS 214999 A viable and highly desirable alternative method for the total and final disposal of all types of PCB waste Is by use of the Electro-Petroleum OC arc pyrolysis technloue. This technique, based on arc furnace technology and originally de veloped for municipal trash disposal, was Issued as U.S. Patent 3,612,620. This technology can be expanded to handle PCBs by designing a sealed system utilizing materials compatible with environment In the arc furnace. Basically, the process Is one In which the PCB will be Initially decomposed In a molten metal bath. In the reolon above the molten bath the gases, Ineludlnq PCB vapor, will be subjected to Intense UV radiation and heat from the arc plasma. The decomposition products will then be directed In the vicinity of the high current DC arc which will have a temperature In excess of 6000 C. There Is considerable evidence that PCBs are totally destroyed at 1600 C. and are converted to biologically acceptable qases. Analysis of the products created In electric arcs In PCBs have substantiated these claims. Our direct current arc pyrolysis approach decomposes liquid PCBs to CO, COj; Hj, CH4, and HCL. The conversion of the Inorganic materials in capacitors and other utility wastes will produce a harmless residue from which the metals may be recovered. Organic materials, such as paper and plastics, will also be totally destroyed along with the PCBs. This`progam Is Intended to be completed within a three-year period from start o* engineering effort to final coanierclallzatlon. In order to conmercfelite the process In the shortest amount of time, a Joint Venture Company Is being formed with a waste management company to manage the project, and offer to utilities a coamerclal service for PCB capacitor destruction. It Is the Intent of the Joint Venture to provide the number of units to meet market demands for capacitor dis posal services. System design A mobile arc furnace was sized to dispose of the largest PCB capacitors currently being removed from utility service and will be treated at a rate of 1.5 tons per hour . A novel feed system will permit the introduction of sealed capacitors to the system and minimize contact of workers to the PCB. This approach was selected after evaluating a number of alternative feed devices. A lining hes been selected which Is resistant to the envlrorment of the furance. The bottom electrode Is designed to be an Integrated portion of the furnace shell. The top electrode will utilize available arc furnace technology for positioning, moving and sealing techniques 5-30 MONS 215000 r r In Its construction. This electrode will be hollow so that gases generated Inside the furnace will be forced Into the vicinity of the arc plasma, thereby Insuring total destruction of the PCB materials. Existing technology will be utilized for the removal of metal and slag from the furnace. Gases exiting from the furnace will be treated In a fume handling scrubber system designed to meet the most stringent air emissions regulations as specified by California Bay Area and Southcoast Area Air Quality Management Boards. Power will be supplied by a DC power supply which was selected from seven potential candidates. The selected power supply proved to be small and more cost effective than other supplies while providing controlled power to the arc furnace and minimizing harmonics feedback to the utility power system. Three potential sites have been selected In the P61E power distribution area. They are located at three remote sites near Kettleman, California. Permitting agencies heve been Identified with contact having been made with those having regulatory control for PCB facilities. A public Information program Is also being developed. Cost analyses have Indicated that this is an economically viable process, which will be envlromentally more acceptable than current technology . The system offers several other significant advantages over existing technology: 1. Pyrolysis produces significantly less gaseous byproducts . to be treated. 2 Total decoaipotltlon Is assured at the ultrahlgh temperatures In the fumece. 3, The molten metal pool In the arc furnace solves the hearth problems of conventional furnaces. 4. Passage of exit gases In the vicinity of the high current arc exposes the gas to at least 6000 C. 5. The gas in the furnace will be subjected to Intense ultraviolet radiation from the arc. 6, The system generally uses developed technology. 7. Capital equipment costs are projected to be substantially lass than conventional Incinerators of comparable capacity. 8. The small size of the equipment suggests that It will be possible to move It from site to site. 9. The process handles sealed containers which need not be . opened or shredded prior to Introduction Into the furnace. 10 The process has the potential to recycle the metals fm . the capacitor cans. 11 Produces envlrormentally safe byproducts. momc 215001 5-31 r r CHEMICAL CLEANLHC OF PCB CAPACITORS Lao Weltzean. PhD Acurex Corporation Cincinnati, Ohio INTRODUCTION Itia dlapoaal of capacltora containing polychlorinated biphanyl (PCB) dlalactrlc flulda la a major problem for tha alactrlc utility lnduatry. Bafora March 1, 1981. the Environmental Protactlon Agency (EPA) had allowed chair dlapoaal in chemical waata landfilla; howavar. with tha atart of oparatlon of tha Energy System* Corporation (ENSCo) incinarator In El Dorado, Arkanaaa, EPA rulad that aufflclant capacity waa avallabla and landfilling would no longer ha paralttad. Thia prompted an affart by tha EPRI to Identify other tachnologlee with which to destroy thee. Thla paper daacrlbee a aarlea of atudlaa parforead by Acurex Waste Tachnologlaa which evaluated the chaelcal daatructlon of PCBe In capacltora and deelgned a ayatae to do thla. One of tha major difficulties In eatabllahlng new PCB dlapoaal fecllltlea la the objactlona voiced by lta neighbora. Thera la no need to elaborate on tha "not In ay beck yard" argtMnta voiced In oppoaltlon to auch facllltlee. Chaelcal procaeelng of capacltora eldaatepa thla laaue by allowing the operation to be eada nob tie. Mobile liquid PCB daatructlon ayataee have run Into no auch oppoal tlon in tha peat. Aa a reault, the nuaber of proceeeee built can be governed by tha requirements of the aarket rather than evallebllity of potential aitaa. In addition, chaelcal proceaalng ayataaa do not require aite-epecific air peralta aa they era not conalderad air pollution aourcee. Bacauaa of thaaa conalderatlona. chaelcal daatructlon of PCBe wee investigated through thla aultl-phaaa atudy. Tha atudy conaiated of the following phaaeo: 1. Phase 1, published in June, 1962 (1), waa e literature aaerch and theoretical evaluation of the concept. In addition, the EPA permits end epprovela needed for Phase 2 ware obtained. 2. Phase 2, completed in November, 1962, wee e laboratory atudy which resulted In: - design date for a full-eeela ayatae - a mad liar alas pilot unit - costa for tha process, both capital and operating HOMS 215002 - conceptual deeign of the full seals operation 5-33 3. Pheae 3. now in progress, is design of the pilot syetan. Tha bulk of the work waa performed undar Phase 2 and thia papar will daal aainly with thaaa resulte. Tha othar work will, howavar. ba lncludad aa appropriata. CHJEMISTAY OF PCS DESTRUCTION It ia known that aodlun can. in tha praaanca of naphthalana and tatrahydrofuraa. dachlorinata organic nolaculaa auch aa PCB'a. {fork haa boon dona la tUa araa by aavaral investigators alnca 1936. (1, _3. and 4). Tha procadura works aa followa t * Prapara a aodlun reagent by mixing aodlun natal with tha propar anounta of naphthalana and tatrahydrofuran. Thia nuat ba dona undar nitrogen aa air and watar will raact rapidly with aodlun and deactivate tha raasant. Tha matant ia rafarrad to aa aodlia naphthalida and la rapraaautad in thia rapart by tha aynbol Me*. * Tha reagent la addad to tha PCB-contamlnated notarial and tha PCI daatructlon raaction ia pamittad to occur. _ * Aftar tha raaction baa tafcan placa, tha excess sodiiai ia quanchad ' with watar, converting it to aodlun hydrosIda. PCBa ara daatroyad via a raaction of tha aodita naphthalida radical anion (Na*) with tha organic chlorine. Tha raaction procaada In atapa, ona chlorine rMoved at a tine. Tha producta ara aodlia chloride and a complex nixtura of nonchlorlnatad polyphenyla. aa wall aa tha aodita hydroxide fomad during tha quenching of tha axcaaa aodita with watar. Tha chaalcal raactiona ara llluatratad in einplifled torn below: C14H6(1P) + 2Na*-Cl3l6(lP)Na* + NaCl CljH6(BP)Na* + HjO--C13H7(BP) + MaOH or 2 a3E6(P)Na*--a^CBP) - (BPlCljHj + 2 Me* Thaaa raactiona illuatrete what happena to a tatrachloroblphanyl, identified aa C1^8^(BPr). Tha tarn (IP) refer* to tha biphenyl atructura. Aa ahown, tha product ia either a trlchlorobiphanyl or an undefined polyaar that Includea the chlorobipbanyla. The aodlun will continue to ranova chlorine fron thaaa eoapounda until tha nonchlorinated form la reached. At thia point, tha PCB nolacula caaaaa to axlet fron both a legal and technical point of view. With proper daaign of tha yatan. it la poaelble to achlova aaeantially eonplata daatructlon of tha PCB at laaat to tha llnlt of naaauranant. The baalc proceaa deecrlbed above suffera fron a nunbar of drawbacka. Tha noat aignifleant ia that naphthalana ia classified aa a "priority pollutant". Ita praaanca in any residua formed by tha process could subject tha raaldoo to disposal 5-34 MOWS 215003 limitation* under th* Resource Con**rv*tion and Recovery Act *RCRA). Thar* *r* number of chemlcele *v*ll*bl* to r*pl*c* naphthalan* in thi* process. Sine* th** variation* era. however, con*ld*r*d proprietary, th* subsequent discussion will focus on th* ua* of naphthalene. -- In addition to the choices available for replacing naphthalene, there la no reason why only tetrahydrofuran (THf) needs to be used. Th* THF can be replaced by any one of * number of ethers such ee dimethyl or diethyl ether, ethylene glycol dimethyl ether and diethylene glycol dimethyl ether, to name e few. THF was choaen for this purpose because when treating capacitors the ether acts ee both a component in th* formation of th* eodium reagent and * solvent to carry th* reegent to th* solid material. Th* volima of ether required to cover th* eolld material In the reactor 1* more than required by th* reagent. Th* THf on e per gallon (or per liter) beela la cheepar than the ethylene glycol ether* and safer to ue* than the dimethyl or diethyl ethera. Furthermore, the laboratory teat* found it to be satisfactory from tha point of vlaw of reagent preparation and PCI destruction. CAPACITOR DESCRIPTION Th* laboratory experiments were conducted on a 1SEVA Waatinghoua* capacitor of th* F? Inerteen type. The unit was obtained from the Cincinnati Gae and Electric Company with the approval of EPA Region 5. The capacitor had the following Name plate Specification*: Manufacturer: Meetinghouse Electric Type: FP Inerteen Normal Voltage: 2400 Maxima Voltage: 2640 KVA: IS Phase: 1 Cycle*: 60 . Style: 1176665A Teat Nimber: 44K 2974 Capacitor contains 1,1 gal. of inflammable fluid (inerteen) The unit*a external dimensions war* 4V deep by 13V wide by 13V high, Th* electrical connectors ere 7V high. It coealsted of * eteel caae, approximately 1/6" thick. A cardboard shall inside the case shielded it from contact with th* electrical packs. Th* capacitor packa were tightly wedged into the ahall, A ceramic insulating plate was on top of the packs. 1 cannot be cut with * saw, A few laportant observation* ere worth noting, first of all, the material ia wedged into the capacitor so tightly that it la Impossible to inject reagent or aolvant into the intact capacitor. Secondly th* surface tension and viscosity of the liquid la eo bigh thet any flat surfaces thet touch each other will tend to stick together. This prevent* the solvent or PCS destruction reagent from ranching 5-35 HOMS 215004 11 tha aurfaca areas. It la thua aaaantlal that tha capacitor ba proparly cut up prior to trsatmant. Tha unit waa opened, drained, cut up and tha cooponanta weighed and analyzed for PCS. Tha raaulta vara; Total weight of tha capacitor ........................ Weight of core notarial after draining .... Weight of ahall notarial after draining ... Weight of PCB liquid drained .......................... PCB content of drained ahall notarial ......... PCB content of drained core notarial ........... Total weight of PCB la tha capacitor ........... PCB content of total capacitor ........................ Total volim of PCB liquid In capacitor. Identified oa Aakaral 1254 (danalty 1.5) TECHNICAL APPROACH 29-B-fcg 13.38 kg 12.81 kg 3.58 kg .67X 22 X 6.72 kg 22.6 X 4.48 1 (1.16 gallons) Tha following two nathodo of treating capacitors ware axaalned: 1. Solvent extract tha PCB fron tha cut-up capacitor and concentrate tha PCB for aubaaquant chanlcal destruction or Incineration 2. Treat tha cut-up capacitor notarial, solid and liquid directly with tha PCB destruction chanlcalo. It la recognized that oavaral coaaerciel processors have attanptad to treat capacitors by extracting tha PCBa liquid and sanding It to Incineration; however, we had understood that tha cleaning of tha core notarial had net with varying degrees of auccaaa in tha past and further work In thia eras appeared warranted. Mora importantly. It quickly bacana apparent that tha chemical daatruecion sethod worked vary wall; If the PCB destruction chanlcalo could ba node to proparly penetrate tha solid notarial, they would destroy tha PCB. Tha solvent laaching testa, therefore, carved as surrogates to help determine tha beat ways of getting tha chanlcalo into tha naaa of tha solid notarial. Taata of the direct chamical destruction of the PCB with tha capacitor aollda wars conducted to determine the following: 1. chemicals uaaga 2. capability of rousing component* of tha chanlcal* 3. temperature changes during tha process 4. intensity of tha reaction 5. rata of reaction Tha successful and-polnt for all caaaa waa considered to be a PCB concentration of laaa than 2 ppn (mg/kg) on tha solid notarial and In tha liquid phasa. Solids war* analysed for PCBa by extracting overnight using a Soxhlat extractor with hexane and analysing tha resultant liquid with a GC/EC. Tha analytical method uaed waa croaa-chackad with other standard method#, Including extraction with nathylana chloride and found to give conalatent raaulta. 5-36 HONS 215005 / r Tha aost Important raault from thla aat of axparlmanta la that thara doss not appaar to ba any matarlal lnalda tha capacitor which Inhibita tha chamlcal daatruc tion of tha PCBa; that it la poaalbla to daatroy tha PCBa without separating tham from tha capacitor solids. What la mora. tha daatructlon la almoat completely tolchloaatrlc In sodium. Tha aodlum reagent la addad to tha ahraddad cora matarlal and than additional aodliai la addad to tha alurry until all of tha PCBa hava baan destroyed. Tha PCB daatructlon can ba followad both by analyala of aaaplaa of tha liquid and colorlmatrically; tha raagant ltaalf la a dark bluegreen In color. If tha aodlum la conaumad by tha PCB (or any othar raactlva aatarial) it turna brown or gray. Addition of mora aodlum raturna tha blua-graan color. Tha aconomica of chamlcal traatmant dapand on tha rausabillty of tha othar couponanta of tha raagant. -- othar than aodlum which la conaimd by tha chaatlcal raactlon. Tha laboratory taata ahowad that 93-97X of tha raagant la recovered whan a raactor la dralnad aftar tha PCB la daatroyad. Highar racovary rataa can raadily ba achlavad In a paoparly daalgnad ayatam. Thla maana that any azcaaa aodita uaad, along with tha tatrahydrofuran and naphthalana (or raplacamant chamlcala for thaaa) la largaly racovarad. Tha reaction occurs vary rapidly. Tha PCBa ara daatroyad in tha liquid phaaa almoat lnatantanaoualy; howavar. In order to allow tha aodlum raagant to panatrata into tha poraa of tha core aatarial. It la nacaaaary to keep it covered with raagant for at laaat two houra. In addition. It la important that tha aolutioa ba kept circulating to prevent tha formation of pockate of aodlum depletion where no raactlon would occur. Tha polyphenyl residue as wall as tha sodium chloride formed remains with tha solid aatarial rather than carrying over to tha next batch with tha liquid. This la Important aa no filtration of tha liquid is required prior to roues. Baaed on thaaa raeulta, it la poaalbla to describe a conceptual chamlcal traatmant ayatam. Tha first step in processing cepecltore la opening than to expose tha cora material. Tha capacitor packs ara typically about 4-6 inches in width. In order to assura that they become properly loose, it la nacaaaary to ehrad tham into places that ara laea than four inches. While a hammer mill cen ba used to achieve thla alia reduction, it is vary dusty and would require a large duet suppression system to ba uaad In this application. Thara ara ahraddars available that would fit thla appli cation nicely. They ara commonly used to shrad a variety of materials ranging from scrap mstal to tiraa and solid waste. They shrad tha matarlal by squeaking It . 5-37 HONS 215006 I bttvun hsrdaned disk* that contain cutting tenth on then. Such thraddsr would be mors then adequate for achieving the needed alee reduction. An alternative eathod of cutting up the capacitors at a lower initial Lnvaetnent la an induatrial aaw. Thla would require an additional operator but ita coat would be in the $10-20,000 range depending on the degree of automation required. A aew would be able to cut up approximately 500 to 1000 pouada of capacitor per hour; thla ie much doaer to the dealgn capacity of the reat of the treatment ayatee and la probably more eultable for a damonatretion pyetem. Ho matter which type of ahredding ayatan ia uaad. it will be neceaaery to ancloee it to prevent the releeee of fimae and duet. In addition, to minimise PCB fume formation, it will be neceeeery to cool the cutting eyatem. A recirculating atream of traneformer oil (PCB contaminated la acceptable) would be a good cooling fluid. In addition, at least a portion of the eyatem would have to be encloaed to prevent the raleaee of duet. The ehredded material along with the reeultant free liquid is trenaferred to a reactor for processing. The reactor muet have facilltlee for discharging eolld material ae wall ee for addition and removal of liquids. In additinn. it muet have provleloaa for the circulation of the liquid through the ease of aolid. It le very important that the reactor be air-tight, ha nitrogen blanketed and hava provialone for aefe preeeure relief euch ee e rupture disk or equivalent. After being filled with the ehredded material, the reactor la sealed and flooded with nitrogen to eliminate air. It ie then filled with sodium reagent and the circulation eterted. Then andium metal (either ee a dispersion or as activated metal) le added slowly to the reactor until approximately 1.2Z atolcblometric eodita hoe been added. A sample of the liquid la withdrawn. If it ia the proper color, the eyatem ia allowed to recirculate for at least two hours with samples of reagent taken periodically. If the reagent changes color, more sodlta ahould he added. After two hours, e sample of the aolid material ia removed from the reactor through an access port and quenched thoroughly with water. A sample of the soft core material la then maacereted and nixed with e measured volume of e solvent such ae hexane with rapid agitation. The hexane le then analysed for FCBe. If it contains leaa then the required amount of PCBs, the reagent le drained from the reactor to e holding tank. The reactor ia then flooded with water (slowly) to quench the excess eodita; quenching should taka et least 30 elnutea with good weter recirculation. The water ie then dreined into e second holding tank. Ths now-trsated solid mass of capacitor eatarlal ia than dumped for diepoeal. It la no longsr a PCB eatarlal and (if no naphthalana la ueed) it la a momhaaardoue wssta. 5-38 s HONS 215007 r FIGURE I REACTION SYSTEM SCHEMATIC FIGURE 2 QUENCH SYSTEM SCHEMATIC 5-3S HONS 215008 D-S Drain Una Service lube MONS 215009 T-S Shredder (shroud not shown) FIGURE 4 mmonshutiom system MOMS 215010 It la poaaibla to change tht through-put of a ayatam vary aaally. Tha throughput of tha design la limited by tha number of modulea that can ba leapt on tha truck while tha analyaaa for FOB ara being performed. A cycle only takes about two to thraa houra. If provlalona ara mada to remove tha dralnad, but chemically atlll active, druse froa tha struck and storing them prior to quenching, until tha analyaaa ara complete, It should ba poaaibla to raadlly lncraaaa tha amount of capacitor material dacontaminatad by aa much aa SOX. Conversely, by not install ing or not using all of tha nodulaa on tha system It la poaaibla to reduce tha capacity. In fact, for the purpoaea of a demonatration. It la only neceeaary to lnatall aa few aa two modulea on tha ayatan, SOLVENT WASH INC SYSTEM DESCRIPTION Tha aolvant washing ayatam baara a strong phyalcal reoMblence to the chemical traatmant ayatam. Thla la bacauaa tha phyalcal layout and the largo plecea of equipment ara all governed by the typo of eolid material that la balng procaaaad. Both ayatama have a ehredder or aona form of capacitor cutting ayatam and In both caaaa, tha capacltora ara treated In 55-gallon nodulaa or druna. Furthermore, each modulo alto in a coll that la of the aaaa configuration. Whlla tha procoaa looks alnilar, tha operation la vary different. In tha caaa of tha capacitor veahlng ayatam, each module la axpoaed to about 20 aeperata cyclaa of veahlng with aolvant, Slnco tha time par cycle la about two houra. It la expected that tha ayatam will ba able to procoaa, at beat, about 100 Iba of capacltora par hour, Tha alnplaat method of weehing the capacltora la to almply run fraah aolvant from tha reboUor Into tha proceaelng modulo for each waah and then run tha aolvant wtih PCB back to tha rebollar. Thla raqulrea an approximate 200.000BTU par hour (about 60 KV) rebollar ayatam which la difficult to beat electrically In a field location. It la poaaibla to reduce thla load In half or even lower by uaing the aolvant 2 or 3 times before reclaiming It, To illuatrace, consider Table 1. Thla tabla ahewa tha (aoeewhat hypothetical but adequate) values of tha PCB content of tha eolida In a givam waah and tha liquids leaving it, If ons vara to match tha liquid eompoaitiona of vaehee 1 with l;, 2 with 12, and eo on, it bacomae evident that if tha liquid leaving waah (for example) 11 ware directly piaapad into a nodule containing material about to begin It* firet wash (fl), there la little or no loee in cleaning efficiency. Tha same le true of each of tha washes paired In Tabla l. In thla way, tha pure aolvant antara only tha veehea going through tha final atagas of cleaning. It than goes directly into tha modulea containing material going through the first acagea of cleaning. Tha load on tha aolvant reclamation ayatam la thereby halved. These pairs of modulea ara referred to aa "mates''. 5-d2 MOMS 215011 r r WASH NO. 1 2 3 4 5 6 7 S 9 10 Table 1 ESTIMATED PCB CONTENT IN SOLIDS AND LIQUIDS LEAVING SUCCESSIVE HASHES WITH PUKE LIQUID SOLIDS PCB(M/kt> 207,000 143,500 73,500 37,000 19,000 9.500 4,800 2,400 1.200 600 LIQUID PC1(E/1) 100,000 80,000 40,000 20,000 10,000 5,000 2,500 1,300 650 330 HASH NO. 11 12 13 14 15 16 17 18 19 20 SOLIDS PCl(M/kt) 300 150 75 30 19 10 5 3 1.5 .0 LIQUIDS PCl(nq/l) 170 05 43 19 10 5 3 2 1 0.5 Note; Liquid PCD In |/llcar Solid PCD la (/k 5-43 HONS 215012 It ie possible to further subdivide the waahlnga Into three or even four matea by cascading up in concantration; howavar. In practlca the loglatica of thla bscoeas coo coaplax. Aa It la. It will ba nacaaaary to provlda tha operators with soma fora of computer aaalatanca to run tha system. Tha ayetam can work In tha follow ing manner; __ 1. Capacitora ara ahraddad or cut apart and paaaad over a almple ecremn whara tha fraa PCB liquid la allowad to drain from thaa Into a recmlving drtaa. 2. Tha drained capacitor aollda ara packed Into proceeaing aodulaa about 200 pounda (75kg) par aodula and placad Into a processing call whara tha ayataa la connected. 3. Relatively clean aolvent froa tha "mate" to thla particular process ing aodula than antara tha aodula and it la agitated for about ona hour. 4. Tha aodula la drained end tha cycle rapaatad for an additional nine tinea- 5. After tha tenth wash, tha aodula becomes the flret one of tha two in tha pair that ara 'bated"- Tha 'bate" to tha aodula In queatlon haa - coaplatad tha waah cycle. It haa bean drained, and tha axcaaa aolvent haa bean vacuua atripped off froa it. It la raaoved and a new aodula containing untreated aolld la placad into it. 6. Now fraah aolvent antara tha aodula under conalderatlon and at tha conclualon of the waah. it la puapad Into tha mate. Aa can be aaan. aultlplylng thla procedure by tan aataa gate to ba vary coapllcatad. In addition, it la nacaaaary to go through a noderate]y elaborate atartup proce dure ao aa to eetabllah tha achedule. Clearly, coaputar aaalatanca la required. ECONOMIC EVALUATION Aa waa atated In tha preliminary report on thla aubjact, there appeera to ba no eciantlflc or technical reason why PCB capacitora cannot ba detoxified chemically. Tha quaatlon la rather, la it feaelble to do ao. Thla section axploraa the coata of alternative routea for the rraoval or destruction of PCB in capacitora. Tha following achernes ara evaluatedi 1. Shredding followed by chemical treatment of the total capacitor material. 2. Shredding followed by draining of tha fraa PCB liquid followed by Freon washes to separata tha PCB froa the capacitor material. Tha liquid PCB la eant to aa incinerator for diapoeal. On tha baaia of tha following evaluation, tha coata of processing tha capacitora and tha capital investment involved ara ahown in Table 2, Aa can ba aaan. tha capital investment for tha capacitor washing achmaa la higher than that raqulrad for chemical traatmant. This la partially balanced by tha lower chemical coats but tha coat par pound of capacitor treated ie etill slightly higher. The difference is wall within tha bounda of tha error of thla analysis eo It appaera 5-44 MONS 215013 that tha coat par pound of capacitor traatad la aaaantially tha iiaa wbichavar routa ia choaan. Whila tha waahinp aathod done raquira an addad invaatnant, tha level of technical axpartlaa of tha oparatora can ba lowar, alnca they do not hava to work with aodlum. Thia way ba daairabla for aoaa conditlona. Tha waahinp routa ahould, tharafora, not ba alininatad aa a vlabla option at thta point. Thia atudy waa only lntandad to davalop a concaptual daaipn for tha two typaa of capacitor claanlnp ayatawa. Aa a raault, tha coata can only ba approximate at thia tine. Tha laboratory work pava raaaonabla valuaa for tha chemical coaca and thaaa ara baliavad to ba raliabla. Tha capital coata ara adaquata for datamlnlnp whathar tha baaic procaaa appaara to ba economical but bacauaa cto data 11 ad daalpna hava baan davalopad aa yet, thay can ba aa auch aa 30X on tha hlph aIda. Tha coata of tha varioua piacaa of equipment, wara pan*rally aatlmatad vary conaarvativaly. It waa fait that a faaaibility atudy auch aa thia would ba battar aarvad by uainp worat-caaa rathar than optlmlaitc coat eatlmatlnp tachnlquaa. Evan whan vandora of equipment providad raliabla coata, 10-20X contlnpancy factora wara addad for thia atudy. Tha oparatinp coata wara flpurad uainp tha baat aatlmata for traval expense*t haullnp, ate. that wara availabla. Whila thay do lncorporata a fair aaount of judpament. thay ara baaad on tha Author'a axparlanca in oparatinp fiald ayatana. Aa ahown in Flpura 2, on-aita chemical traatnant of capacitor* appaara to ba coat coapatltiva with incinaratlon. Tha hlphly conaarvativa coat aatlmata for tha chaanical treatment ia approximately 50c par pound dalivarad to tha lncinarator. Tha coat of ahlpplap capadtora variaa fro* about 3c to 10c par pound dapandlnp on tha diatanca thay auat ba chipped. Tha coat of ehanlcal traataant baa baan eatlaated uainp hiphly conaarvativa aathoda. Further procaaa rafinaaanta can reduce it aipnifieantly. Tha aolvant waahinp ayatea, if operated by a craw who travel* all over tha country la aa tin*tad to coat 58.2c par pound of capacitor traatad to operate. Thia malcaa it economically unattractive if a craf coaaa in froa a diatanca; howavar, if tha purpoaa of tha ayataa ia to tract a utility'a ln-houaa waata capacitor*, than tha differential batwaan chemical procaaainp and aolvant waahinp become* only A. 3c par pound of capacitor traatad. Thia ia wall within tha limit* of accuracy of thia analytic and for all intent*, tha two method* are of equivalent coat. It la fait that the aolvant waahinp ayatam can readily ba operated (if proparly daaipnad) by a utility'* own paraonnal. MGNS 215014 5-45 r r Table 2 COMPARATIVE COSTS OF TREATMENT FOR DIFFERENT SYSTEMS Capacity (lb/hr) (kg/hr) Capital Coat (4) Operatic* Coat Labor Chemicals Utilities Waste Diepoaal Maintenance Field Coata Depreciation (c/lb) Total Coat (C/lb) CHEMICAL TREATMENT 900 400 220,000 6.0 12.8 1.0 10.0 4.0 5.7 8.3 49.8 SOLVENT HASHINC 400 160 300.000 9.0 9.4 2.0 13.4 5.2 9.8 (0.9) 9.4 56.2 (49.3) Kota: c/lb, ceata/lb of capacitor treated The fltura la parentheses for the solvent washing ey*t eieia that the syetaa la operated by the utility for ita wi uaa, hancs, no field coat* for tha craw ia required. 5-4 HONS 215015 r r CONCLUSION Cheeical traetnent appear* to b* a reliable, coat-effective eathod for tha diapoaal of PCB capacitor*, it haa a nuabar of advantagaa over other nathoda of diapoaal, including; _ 1. can be Bade nobllc ao thare are no difficultlee aitlng it 2. low capital invaatnent 3. aaay to build more capacity to aeet aarkat deaand 4. environmentally acceptabla The proceaa haa been ahown to ba effactlva in tha laboratory end, indirectly, on a pilot acala. It ia now neceaaary to build a full-air* nodule to teat tha ayatem and then build e demonstration ayatem ao ea to obtain tha neceaaary coat data and panic* for it. unuNcis (1) Waitmen, 1 and Millie, "Feasibility Study of Chanical Detoxification of Polychlorinated Biphenyl Capacitor*, Phee* I Study", EPB.1 CS-2477 June, 1982 (2) Wait man, L., Barth, B.C., Peryan, S.J., "Feasibility Study of Chtaical Detoxification of Polychlorinated Biphenyl Capacitor*, Phase 2 Study", Draft final report subarttted to the EFK.I, November 1982. (37 Scott, N.D., J.F. Walker and V.P. Han*lay, J. Anerlcan Chanleal Society. SB 2442-2444 (1936). (4) Snith, C. and C.L. Bubbar, The Pee of Sodium NaphthalanIda ta Ch--jcellv Deetroy PCB by Dechlorination, ibid. ACXNOWUDCIMSn The work described her* wea eponeored by the Coal Combustion System* Dlvieion of the Ilectric Power keaearch Inatitut*. Their support and the technical MilitMca of the CPU project manager, Dr. Ralph gone!, la gratefully acknowledged. 5-47 MOMS 215016 r r NOBILE PLASMA PXROLVSXS T.G. Barton J.A.G. Mordy Department of Civil fcigineering Royal Military Collar* of Canada Kingston, Ontario, K7L 2H3 ABSTRACT A thernal p laana, proparly appliad to toxic watts destruction, providaa a pyrolytic environment with diatinct advantages ovar other competing technologies, Coaplata atomisation of organic fluida has baan daaonatratad to occur in laaa than ona third of a ailliaaeond. Kinatic racoabi nations of atonic antitiaa ara pradietabia. Chlorinatad wastes produoa a hydrogan chlorida byproduct which ia ooaplataly converted to salt in a caustic scrubber. Daetruction efficiencies in axoaaa of 99.9999999 percent have baan corroborated by external aonitoting agendaa for the destruction of Aakarel fluida with chlorine contents up to 96 percent. fines this process ia pyrolytio, the scale of the equipaent ia small for high throughput rates, fetergy requirements for the destruction of Aakarels ara typically laaa than ona hi Lowstt-hour par kilogram of waste, A mobile prototype unit ia being oonatructed for the Department of Environmental Conea rvation of the Sfeata of New tort for a throughput rata of ona gallon par minute of liquid waste. Die nobility and efficiency of ouch aystaaa Unit thn controversy eaaocieted with siting fixed facilities. 5-4 9 MOMS 219017 NOBILE PLASMA. PYROLYSIS INTRODUCTION The disposal of toxic and hazardous weetee continual u b controversial utttr. In North America, more than thirty five mege-tonnee of aueh waste requires diapoaal annually. Mora than ninaty percent of thia waata racaivaa lnadaquata attantlon and la ratumad to tha environment through leaking landfilla and lagoons. In North America, approximately fourtaan thousand landfilla hava baan identified aa contaminated with toxic natariala. and ona hundrad and eighty thouaand lagoone ara parcaivad by tha public aa poeing aignifleant threat to health and environmen tal quality. Imposing regulations which demand proper diapoaal of toxic and hatardoua natariala doee little for tha anvironaant unleaa coat affective diapoaal alternatives ara available to industry. All too often, public reaction to siting suitable technolo gy thwarts such davelopaant efforts. SiaLllarly, tha coats associated with permitting appropriate technology create financially unattractive business ventures for large industries or ara prohibitive for new growth industries. It must also be recognised that tha ragulationa required for permitting soma new technologies hava yat to be written. Tha and result in auch a scenario ia that midnight diapers prosper at tha expanse of tha environment. it has baan stated that wa do not own thia earth but wa hava msrely borrowed it from our children. Thia attitude ia profound whan it ia realised that 2,3,7,B TCDD ralaaaad to tha environment requires about two billion years to thermally dagrads to show a destruction and removal efficiency (DM) of 99.99 percent at ambient taparaturaa PC* and PCDP compounds raquira longer and hava thus samsd their reputations for persistence in tha environment. Batter procedures ara required to favour tha development and application of new technology if environmental quality it to improve. First, let's consider tha iaeuae associated with tha aiting and operating of a new technology. Centrally, three interest groups era involved and can be identified as either regulatory, enterprise or citizens. Tha regulatory repreeantativea ara concerned with tha maintenance of tha environmental quality. Ideally they would 5-50 MONS 215018 Ilk* to technology which destroy* compound* euch * PCS'* and produce* an affluent with zero raeidual toxic material*- However , technology he* not as yet provided a mechanism of measuring zero realduale and thua regulatory etandard* are aatabllshed. initial atandard* of ppm ware replaced with ppfa and autoeequently ppt level*, and before long we will be measuring ferntf-grama and atto-grama in the Inevitable pereult of zero, Enterprlae rapreaentatlvee are concerned With profit and lo** potential* which eurround a waate dlepoaal concept *uch aa a high tamperatura incinerator. The oonoept must be able to compete with alternative dlepoaal techniquea on both coat and performance acala*. However, a* regulatory standard* become more severe, monitoring and safeguard coats escalate to the point where a mega-scale complex is required to taka advantage of economy of scale. It is recognized that it ie difficult for any appropriate waste disposal technology to compete with often employed non-appropriata means of disposal and the question of corporate conscience la brought into play along with the third Intaraat group, the citizens. The citizens' representatives are the guardians of the heritage of our children. They demand a safe environment and are reluctant to aee a nags-scale complex sited in their neighborhood. Such a complex will attract a continuous float of tankers which pour pollutants from abroad into their vested domain. Tha dangers of transport accidents repraaanta one type of isolated threat. In contrast, whet goes into the cosplex must coma out at a detoxified affluent to tha dagraa that tha regulatory standards can be mat. This maana that a continuous threat it posed by the effluent stream which, albeit dilute in residual toxic material, la perceived to contaminate the neighborhood with weetae from abroad. Using today's standards for many regulated westee, of e DU of 99.99 percent, it can be readily shown that the processing of a mega-tonne of such waste will contribute one hundred tonnes of dispersed affluents to the neighborhood, a most unappealing proposition. If tha new facility can be demonstrated to give e higher DU, then tha affluents would have e smaller impact on the surrounding covrunity. similarly, if tha facility oould operate coat affectively at lower throughput rates, than tha rata of buildup of affluents in tha community would be proportionally reduced. in addition, if tha citizens' representative* could be assured that minlmw affluent* would only stem from processing local wastes rather than transported regional waataa, than a batter political environment would be created to deal responsibly with tha annual production of toxic and hazardous waataa. Plasma pyrolysis offer* high DH in a mobile configuration which ia coat affective with low 5-51 MONS 215019 throughput rates. TKB TECHNOLOGY Ths hsert of thl* technology centre* around tha plana arc device. a eo-llnaar alactroda arrangement la uaad to craata an alactrlc arc. Low pressure air ia uaad aa tha Radium through which an alactrlc currant ia paaaad. In passing through tha air, tha alactrical anargy ia converted to thermal anargy by abaorptlon by tha air aolaculaa which ara activated into ionirad atonic atataa, loaing alactrona in tha procaaa. A raeirculation vortex dynamically produced inaIda the eo-llnaar alac troda arrangement anauraa a atabla piaana core with predictable performance. Ultraviolet radiation ia emitted whan molecules or atone relax from the highly activated atataa to lower levele. Meats fluide can be injected into tha co-1inner alactroda apace where they interact with decaying plaaaa apeciae to become rapidly atomiaad prior to exiting tha electrode apace. An integral part of thin procaaa technology ia tha pyrolytic eisolation computer modal. New eompounde which ara created from tha recombination of atomic speciea ia predictable baaed on kinetic equilibrium. The ministration of Gibb*a free anargy ia uaad to determine tha equilibrium concantratione of product epodae over a wide range of aalected temperaturee and praaauraa, Thue appropriate operating conditiona can be predicted for any organic weeta fluid prior to ita deetruction, and undaairabla producta can be minimiaad or eliminated through altering tha character of tha fluid or tha operating conditiona. Changes in enthalpy between feedstock and output producta ara used to pradiot tha plasma energy required adlabatically for tha pyrolyeie of tha waete being treated. At the Royal Military Collage of Canada, a pilot unit haa bean operating einca 1979. Preliminary tearing on that initial Canadian unit waa oonducted with Aekarel which contained twenty percent each of aroclor 1254 and 1260 aa wall aa eixty percent of trichlorobanaene. The reeulte of tha initial taating showed a DU of 99.9999 percent, in 1962, tha ayatam waa modified to include tha co-linear electrode arrangmnt aa ahown in figure 1-1, aa wall aa increase tha throughput capacity. Tha oo-linaar pleMa device waa mounted on top of a refractory lined, etsinleae eteel veeeel with a volMa of approximately 0.5 cubic metres. A caustic soda liquid scrubbing ayatam followed tha veaael and an induction fan draws pro duct gaaaa through tha ayatam. Add gaa and graphitic carbon residues ara trapped by tha acrubber while fuel gaa producta ara burned on tha flare etsok. Figure 1-2 illustrates tha ayatam in operation, destroying Aekarel, and tha flare atack flame can be noticed through tha window of tha laboratory, Tha scrubbing syatam 5-52 HONS 215020 r r i .1U.7 * . tr Flyura 1*1, Pilot Seal* Schaaatic Fi^ura 1-2* Pilot Unit in operation 5-53 HONS 215021 had bean calibrated and shown to recover approximately 96 percent of inaolubla low volatile compounds. Measurements of PCB residual in the scrubber water ware externally corroborated to be lasa than 0.17 ppt, giving a destruction efficiency in axceaa of 99.9999999 percent. It ahould be noted that the scrubbing unit eaploya a water md liquid cauatic soda atomizing eystam which creates micron sited droplet that have a high trapping efficiency for acid gee and particulates. The amount of cauatic soda injection la controlled by pH Monitoring to assure slightly baaic scrubber effluent fluida. An assessment of approximately forty emerging technologies was conducted by the Environmental Protection Agency (EPA). Five were selected for demonstration sines they showed a high potential for successful application to toxic and hacardoua westas. Under a co-operative agreement between the EPA and the New York State Department of Environmental Conservation (NYCDEC) a demonstration program was established to further investigate plasma pyrolysis. A contract was awarded to Pyrolysis Systems Incorporated in Walland, Ontario in 1983 by NYSDEC for tha fabrication and daannatration of a mobile prototype. THE NOBILE PROTOTYPE The mobile prototype unit ia being constructed in Niagara Falls, Canada and ia composed of nine major aub-eeaambliaa which are completely contained inaida a closed forty-five foot moving van type drop-bed trailer, fiieae nine sub-systems link together to form tha plasma pyrolysis aystam. In addition, a fully instru mented control room complete with on-line, close to veal time monitoring ia also located inside the trailer apace. Tha unit ia being sized for a nominal throughput rata of one gallon par minute, yet has surplus power and capacity to that nominally required. Each sub-ayatam ia briefly described along with tha monitoring and failsafe considerations associated with the design of this unit. Plasma Sub-System The plssma sub-system ia comprised of a power supply and a plasma device. The power supply ia a six pulse thyristor water oooled unit rated at 500 kw and ia located in tha forward area of tha trailer. Primary 480 volt, three phase power feeds tha power supply from a comswrciai service line. Tha diraot currant power from tha power supply ia delivered to tha plasma device to provide variable plasma output from 200 to 500 kw. Tha plasma device ia horizontally mounted onto a rafraotory lined pyrolysis chamber which has an internal volume of approximately two cubic metres. Hhlla the eo-linaar electrodes serve as a plug flew atomisation zona for organic fluids, tha pyrolysis chamber serves as a completely mixed 5-54 . HONS 215022 r r recombination zona. Raaidanct times In tha atoadtation con* and tha recombination cona ara approximately five hundred micro-seconds and ona aacond raspectivaly. Tha racombination cont la aquilibratad at a tamperature of 900-1000C. Domaatlc Watar Sub-Systama A two inch watar connection ia required to faad domestic water to the trailer unit, in tha trailer, a header ia uead to vector domestic water to dear separata uaar aactiona. Mater ia fed from tha header through an air gap into a reaarvoir from which watar ia drawn for tha acrubbor. k aacond line feada water through a daionication unit and aubaequantly through an air gap into a aacond conditioned watar reaarvoir which provides cooling to tha plsaaa device, A third watar line vactora watar to a hast exchanger which ia used to cool tha conditioned watar loop. A fourth line takas water to tha onboard laboratory sink and hot watar tank ayatea. Conditioned Water Sub-System Tha conditioned water ayitem ia designed to provide adequate cooling to the plaama system with water of suitable quality. Thia system includes a 130 gallon tank, a high pressure, high flow circulation pimp, a heat axohanger and sufficient temperature, pressure, flow and quality measuring devices to ensure tha adequacy of tha system. Provisions ara made to automatically dump portions of tha conditioned water end add fresh deionised water to maintain tha desired resistivity within this water circuit. in tha event of a total power failure, tha system ia designed to be pnsvaaatically charged, thus forcing tha contents of tha watar storage tank through the plasma device and provide sufficient cooling. Scrubber Water 8ub-ystsm A variable speed pop is used to pwp water from the scrubber water reaarvoir. A portion of thia watar is breached off the main flow to pass through tha thyristor cooling loop prior to rejoining tha main flow. Water used in tha scrubber experiences a ones through criteria and a flow rata of approximately 20*40 litres par minute is anticipated. Additional temperature, pressure and flow instnenentatlon ara used to verify the operation of thia sub*ayatma. Previsions for pneumatic charging of tha scrubber water reservoir to provide cooling and cleaning of tha product gna is also included in tha design of tha ayatam. HONS 21SQ23 5-55 Caustic Sod* Sub-gyatam The caustic aoda sub-system ia designed to Mtar sufficient sodium hydroxide into tha acrubbar watar stream to anaura complsta nautralization of acid gas and rscovary of hydrogen chlorida. Cauatic soda raaarvaa nay ba contained in aithar a separata tank truck or an external raaaxvoir. provision for onboard storage of approximately S5 gallons of cauatic aoda ia made, A pH sensor is used to verify the condition of the spent scrubber water and provide operator signals for variation in spaed of the caustic soda pump. Gas Cleaning Sub-System A three stage spray ring is located at the producer gaa outlet from the plasma recombination zone. Water and cauatic soda are fad into this ring vhera an atomizing apray creates micron sized droplets. These tiny droplets impinge on particulate matter in the product gaa stream as well as react with the hydrogen chlorida to form salt water. The temperature of the producer gee ia reduced from the 900*1000c to leea than 60C in this ring. The scrubber solution end quenched product gas are drawn into a cyclone eyatem where particulate laden salt watar is extracted. A pmp on the aiap of the fluid extraction cyclone vectors the fluid to a sanitary sewer drain. Provision ia made for collection of e composite sample of this fluid which is than subjected to analyses. The quenched product gee is drawn off the cyclone by the induction fan aub-ayatem. induction ran sub-System The induction fan sub-system continuously draws on the reactor and acrubbar units to maintain atmospheric to slightly negative pressure in the eyatem. Producer gae from the induction fen passes through e three-way valve. During normal operation, this gee ia vectored to the flare stack where it is electrically ignited. Since the gas is mainly hydrogen, carbon monoxide end nitrogen, it burn# with a clean flame at a temperature of 1800-2100c. This flare stack serves ee an air pollution control device to prevent the releaae of fuel gas to the environment. The potential to vector this fuel gaa beck into industrial furnaces will ba considered later for site specific applications. In the event of a power failure, the thres-way valve vectors the product gae through an activated carbon filter to block the potential release of undeatroyed toxic material to the environment. Provision ia made to sample the product gaa prior to the flare. 5-5$ HONS 215024 compressed Air Sub-Sytaai Tha compressed air aub-syetsm auppliaa plasma air to tha plasma arc davlca as wall as comprassad air for operating air valves, prassuriring tha two pravlously described water reservoirs and blowing down tha aystam for Mintananca. A sarlas of temperature, praasura and flow sanaora ara uaad to snaura adequate air auppliaa for aach requirement. In tha avant of a total powar failure, tha air compressor la daaignad to hava a raaarvolr of aufficlant capacity to Mat all pressurization and flow requirements. Waste feed Sub-System Tha waata faad aub-aystaa contains two procaas rssarvoirs, a variabla apaad pump and a sariaa of automatic faaturaa which pravant aalaction of waata faad until appropriata raactor oparating conditions ara achisvad. Ona procaas rassrvoir contains organic fluid such as athanol and is usad during a startup sequsncs ss s pracursor faad and during a shutdown saqusncs to flush ths wasta faad aystam. Tha saeond reservoir would normally ba a storaga reservoir at an industrial sita. a flaxiblt connsction would normally ba usad to dalivar tha contants of that rasarvoir through tha waata faading ayatM to tha raactor. Provision ia mada for a 55 gallon onboard procsst rasarvoir into vhioh waata matarials can ba ptaaped. Failsafe Conaldarationa During normal opsration, tha plasM pyrolysis raactor raquiras approximataly thrss minutaa of warmup prior to cooaancing injection of toxic or hazardous vastas. Similarly, approximataly twanty minutaa ara required for tha cooling of key components prior to conducting poet-run maintananca. These short cycle times maka this unit basically an on/off ayatM which can respond rapidly to adverse conditions. Tha system hardware ia arranged to provide failaafa operation svan in ths avant of a total powar failure. Normal scheduled maintenance of pre-anginaarid components such as motors, fans, valves, matars and sensors should safeguard against unforeseen malfunctions. Thus, consideration for failsafe ia limited to monitoring essential conditions and providing protection in tha avant of tha development of an adverse situation. Three ecenarioa must ba considered which include controlled operations, loss of plasM and total loss of power. A MAC8YN 550 process control computer ia usad to control the total operation of tha aystam. Thus, under conditions which preclude total lose of power, tha computar is progressad to provide operator warnings as wall ss taka positive action in tha event of adverse conditions developing. Under total loss of powar. 5-57 *QNS 215025 tha irr*nfMnt of the hardware taka* over this function. Haata faad ahutdown eignala ara generated undar tha following conditional a loaa of plaaM arc a dataction of unacceptable producta a loaa of induction sub-system a loaa of cooling watar sub-systems Tha plasma arc nay ba loat aa a raault of alactroda burnout. Caloulationa hava ahown that undar thia condition, aa littla aa fiva micro-grama of undaatroyad waata nay paaa through tha pyrolyaia ayatan. Thia notarial would ba trapped on the activated carbon filter which forma one of the failaefe featuraa. Unacceptable producta nay include trace level dataction of principle organic hazardous conatltuenta (poHC'a) or axeaaaiva ralaaaa of hydrogen chloride from tha ecrubber, In addition to generating waata faad ahutdown aignala and taking appropriate action, the computer eyetan nleo provider operator warning aignala, which if not rectified in a preaet period will automatically ehut tha ayetan down. Operator warning aignala aten from monitoring tha following conditions which ara compared to oomputmr programmed limit rangest a reactor temperatures e scrubbed gee temperature a ecrubber water temperature a ecrubber water pH e plasms cooling watar temperatures a plasma air flow . e water pressure a caustic fluid Quantity a toxic waste fluid quantity a flare ignition Failsafe criteria ara instituted to safeguard the environment. Obviously, tha degraa to which tha environawnt ia protactad will bear directly on tha accuracy of monitoring aetivitiea. MOWS 215026 5-58 Monitoring Tschnlguss While w# My intuitively accept the theoretical concluaiona, that the conditions evolving from the plaama arc will diaaociate all organic compound* into their coaponante, there are other factora to be conaiderad under actual working conditiona. The Min onea are whether the feed atock ia totally or only partially deatroyed or diaaociated, or whether re-combination* following diaaocietlon will create more toxic coapounda. Theae concerns can be addraaaed through accurate and preciae monitoring and analyaia of product gaa, ita scrubbing aediue and minute quantities of deposits expected within the plaaM pyrolyaie ayetee. The prims objectives of the eaiaaien annitoring era to insure that the plate* reactor la achieving the required degree of deatruction efficiancy and that no now harardoua chemicals ere being created in the procaea. Theas objectives can bast be mat by analyzing the product gaa before it ia flared. Detection of trace chemical* la assisted by their being at their highest molar concentration at that point, where they have yet to be altered through combustion. Sampling at this point should preclude the need for trace analyaia of effluents further downstream. In the mobile unit, tree* analyaia will be performed by a Hewlett-Packard S792A gaa chromatograph coupled to e Hewlett-Packard 5970A mss aelactivo detactor (KSD). Tha MSD haa a da taction lavel of on* nano-gram while In the spectral scan sod* and 10 pleo-grams while in the aalected ion mod*. It ia proposed to use this equipment in e manner which will provide an analysis in laaa than tan ninutaa, as cioa* to real time monitoring for such trace analyaia as technically possible. The physical nature of tha product gaa and tha low concentration of chamicals bping monitored require a carefully designed aaapling system. A suggested analytical protocol would be that developed by tha National Research Council of Canada for a study of pesticide* in ambient air at concentration* ranging down to three pleo-grams per litre. A one hundred litre temple of product gaa will pass through a heat traced lint to e particulate filter to remove residual carbon. The cleaned gaa then paaaea through an absorber with s capture efficiency of 99 percent. The adsorber 1* then rapidly heated to release trapped organic*. A nitrogen stream carries tha organic* to the GC-HSD for analyaia. Tha mast detector in ita selected ion swtde should provide a detection level of on* pioo-gram of POBC per gram of waste feed. The KSD In the specific ion mod* will provide information on residual sum for aim aalected lone whose presence either indicate* the degree of toxic waste deatruction or the fontation of new toxic compounds. if the concentrations of 5-59 HONS 219027 that* chemicals axe*ad tha predetermined limits, automatic faad shutdown procedures will be initiatad. If tha concantration lijaita ara not exceeded, tha analyais cycle will than ba automatically rapaatad. Tha poaaibility of tha formation of chaadcale not specifically acannad for. by the above procedure, must alao ba conaidared. To determine whether such"chemicala are being produced, tha maaa apectromatar will acan from tan to aix hundred atomic maae unit* at appropriate tin* intervals. The detection of a significant concentration of an unknown compound whoee maaa is equivalent to that of a hatardoua chanuical will invoke a etanderd shutdown procedure. bulk gas analysis of tha scrubbed product gee will also be performed at this sampling point looated downstream of the induction fan. A Hewlett-Packard 5880A gaa chromatographic ayatam will provida on-line analysis of hydrogsn, watar, nitrogsn, methana, carbon monoxide, carbon dioxida, ethylene, ethane, acetylane, propane, propylene, 1-butana end hydrogen chloride. The monitoring of water will aeaiet in completing a watar balance for tha system. Tha analysis for hydrogen chloride will provida a means to monitor the removal efficiency of tha hydrogen chloride by tha scrubber. The operator will ba alerted upon tha detection of concentrations in axoeaa of 200 ppm. In this way, removal efficiencies in excess of the required 99 percent will be ensured for a waste faad containing thirty-five weight percent of chlorine. The detection concentration to signal tha operator can ba adjusted in accordance with tha chlorine content of tha waste being destroyed. SUHARY Production of the sobile plasma pyrolysis unit is well underway in Canada. Tha first unit should be ready for preliminary tasting early In 1984. In that the pilot unit has shown unprecedented destruction efficiencies for PCS compounds as wall as a high destruction efficiency for refractory eoepounda such ee carbon tetrachloride, the mobile prototype ehould be equally affective in dealing with toxic and hatardoua wastes. Mo permit to oonstruct has been applied for on this project since tha unit ia being built in Canada. However, other permit applications will be required prior to this unit operating in the United States. Permitting may be required under the Clean Air Act, the Toxic Substance Control Act (TSCA) and the Resource Conservation and Reoovery Act (RCRA). Preliminary diaeusaions indicate that tha unit will ba exaagt by virtue of scale and stack else (four inch disaster) under 5-60 MOMS 215028 r r the Clean Air Act- Under TSCA, a permit for da atruction of PCB compound* will ba required and tha technology will ba regulated aa alternative technology in contrast to combustion technology. However, waste PCB compound* do not exist as pure aubatratee. Commercial preparations normally contain solvents such aa trichlorobenxene and thus this technology requires a permit under ACM for wastes other than PCB. However, while ACM recognisea thermal processes such aa pyrolysis as distinct from incinaretion processes, regulations governing that permit application (ACM pert B sub-part X) have yet to be written. Technology Is taking an extremely serious approach to curb the Irresponsible disposal of toxic and hazardous wasts. It would ba most unfortunate if demonstra ted technology tied to sit idle until the present progrem of developing permitting regulations is completed. Perhaps it would be more appropriate to reconsider a mors efficient alternative. That is, rather than writs regulation# which attempt to encompass ell technologies, it would ba more appropriate to review each ' apecific technology and provide responsible guidelines under which it can operate. It is unlikely that all encompassing regulations will be effective and could conceivably restrict the coamarcial application of new technology to the global pollution problem. Tha mobile concept which ie represented in this report la highly effective in destroying wastes. It can operate cost effectively at small throughput ratsa. The parformanca of tha pilot unit indicates that it could operata in a c funity without any environmental impact and thus destroy toxic wattes at industrial sites where they were created. The attitude of citlsans1 groups in Canada indicates that such groups art prepared to allow the destruction of thair local haxardous and toxic waetee provided transportation of ouch wastes ie minimised end importa tion of such wastes to their coiunlty ie eliminated, conceivably, this approach could reverse tha trend favouring midnight duspora and tha development of new Love Canale. 5-61 HONS 215029 ENERGY AND BYPRODUCT RECOVERY FROM CHLORINATED HYOROCARBONS: CHLOE-CHIMIE'S VRC*INClNERATION PROCESS FOR POLYCHLORINATED BIPHENYLS by Joost Galley Pyrochlor, Ine. 10 Wickham Way Chatham, New Jersey ABSTRACT Chloe-Chlmle, subsidiary of the French Groupe Elf Aquitaine, has developed end successfully Marketed a heat and byproduct recovery process using waste chlorinated hydrocarbon residues. The prototype, full-scale incineration unit, termed the VAC* process, has been In operation In Salnt-Auban, France since 1974 and has a capacity of 16,000 tonnes of waste materials per year. It can accept 2,000 kgs/hr of chlorinated waste, is designed to operate at 1200% at a retention time of S seconds, and cen produce 4,300 kgs/hr of clear 33% HC1. Tests with waste PCBs were conducted In October 1979 and May 1981 and showed no dlscernable PCBs, PCDFs, and PCDOs In effluents et detection limits of 0,1 ppb (100 ppt) for etch entity. Borden Chemical Is currently operating a larger Chloe* Chimle VRC* unit at its Gelsmar, Louisiana chlorinated solvents plant. INTRODUCTION Chloe-ChlMle, as a leading European producer of chlorinated hydrocarbons, historically has faced the problem of disposal of highly chlorinated, refractory wastes, The research undertaken In the eerly 1960's led to the development of a series of Incineration units, culminating In the fabrication of the VRC-2 unit. The latter "VRC-2" (from Valorisation des Residue Chlores, Type 2) Is designed to burn more than 2 tonnes per hour of residues. The plant has an outstanding operational flexibility and recovers high quality hydrochloric acid solution of 33% w/w or gaseous hydrogen chloride, which can be either marketed or captive. The plant has been licensed seven times as follows: Spain Morocco USSR USA 1976 1977 1976/1981 1992 Perchloroethylene residues VCM wastes Four plants VCM residues t 5-63 MOMS 215030 In order to demonstrate the high efficiency of this type of Incinerator, tests were conducted on October 16, 1979 and from May 10-14, 1981, employing as feedstock used conmerclal PCUs returned from customers. The tests were carefully monitored to determine the possible presence of unburned PCBs and/or other oxidized harzardeus compounds such as PCOFs and PCDDs In the effluents from the unit, PA sampling and analytical methods were utilized (see Modified Method 5 Train, Appendix). CONCLUSIONS The Incineration tests performed In the Salnt-Auban plant on commercial, spent PCBs show clearly that this technology Is able to achieve essentially complete combustion of commercial PCBs without fuel addition with HC1 recovery as muriatic acid and comply with the recommendations of the PA published In the Federal Register of May 31, 1979; that Is to say, PCBs In emission gases were not detected at 0.1 ppb with a resultant destruction efficiency In the range 99.99997 99.999991. Oxidized byproducts of the PCBs combustion, such as PCDFs (polychlorinated dlbenzofurans) and PCDDs (polychlorinated dlbenzop-dloxlns) were not found at a detection limit of 0.1 ppb (100 ppt). More than 991 of the chlorine contained in the wastes can be recovered as 331 w/w KC1 solution or as 1001 HC1 gas. The VRC* unit can burn waste with a caloric value as low as 2,200 keal/kg (3,700 BTU/lb) without additional fuel. OlM costs for this technology range between 2.5-3.51 of capital Investment. This Is partly due to the minimal/almost lack of corrosion, but also to uninterrupted performance of the equipment over very long periods of trouble-fret operation. VRC* INCINERATOR AT SAINT-AUBAN "Historical Built In 191$ to produce chlorine, the Salnt-Auban facilities continue to be specialized In chlorinated hydrocarbons ($00 tonnes of chlorine per day). The heavy and light end byproducts must be disposed In ait envlromentally proper way and, since 196B, this has been accomplished by land-based Incineration In an Incinerator designed and constructed at the Salnt-Auban facility. This unit produces and recovers MCI and steam without corrosion problems. Recognizing its outstanding performance as an environmentally sound technique, the French authorities sponsored the construction of this plant and certified the tradename of: "Valorisation des Residue Chi ores". HONS 215031 5-4 r r Description (see Chert 1) The Incinerator comprises: * Furneee (F.101) ~ A verticil, cylindrical furnace Is designed to satisfy tlw previously defined temperature and residence time conditions (l.e., 1200% and S seconds). It Is lined with a refractory material specially selected for Its corrosion resistance and durabll Ity, * Burner (8,101) The furnace-burner has been developed and patented by Chloe-Chlmle. This burner can handle liquids, as well as viscous fluids, containing solid particles and gases. This system requires no liquid pressure (gravity feed) and has a maximum air pressure requirement of about 0.4 bar (58.S pslg). * Quench and Washing Tower (D.810) The quench (Chloe-Chimle's patent), a solid section made of graphite, was developed In cooperation with the French firm Vlearb-Grenoble. This very light, extremely reliable piece of equipment (more than 5 years of practically uninterrupted service at the Salnt-Auban plant) reduces the temperature of the combustion gases from 1200% to 60%. The temperature of the quench Is controlled by recycling cooled HC1 solution. The same solution Is also sprayed directly Into the washing tower. The cooling loop, connected to the washing tower. Is designed to operate In such a way that no HC1 absorption occurs. A purge withdrawn from the loop eliminates ashes, metallic Impurities, and/or other materials. This purge (approximately 251 w/w HC1 solution), can be led from the washing tower to a tank where It decomposes the hypochlorites resulting from the neutralization of effluent gases. The decomposition gases (chlorine plus Inerts) can be recycled to the furnace, and the resulting brine sent directly to the wastewater treatment system. * Absorption Equipment (ft.815) These units ensure the gas circulation and maintain a slight negative pressure In the furnace (venturi system). The Increase In the solution's concentration Is systematically accomplished through countercurrent mixing with the combustion gases. One, two, or three units are necessary, depending upon such factors as: - the chemical composition of the residues - the temperature of the available cooling water the desired concentration of the final HC 1 solution MOWS 215032 5-65 FLOW-SHEET OF VftC SYSTEM - CHART 1 XA 9S man i** CHLOt: Chimie MONS 215033 r r * Final Processing of Tall Gases (P.818) Without causing major changes In the pH and output conditions, a predetermined amount of caustic soda IsInjected Into the neutralization stage to react with any chlorine and HC1 that escapes absorption In the previous stages. Neither phosgene nor nitrogen oxides were observed In previous operations over the last six years at the Salnt-Auban VRC* unit. * Gas Effluent, Stack (F. 101) After the caustic soda treatment section, the gas Is practically chlorine* and HCI-free (l.e., less than 30 ppm), and easily meets the environmental protection requirements. Consequently, scrubbed gas Is discharged directly to the atmosphere. * Range* of Residues Processed The VRC* unit was designed to Incinerate waste byproducts from the following organic chemical manufacture: Vinyl Chloride Monomer (VCM) Trichloroethylene (Trl) Perchloroethylene (Perc) . Carbon Tetrachloride (Tetra) 1,1,1-T r1ch1oroethane Some residues (from VCM for instance) are mainly composed of light ends and have a chlorine content less than 60S w/w. Other residues (from Tetra/Perc) containing 70S or more chlorine, are very viscous with a melting point over 160*C, and resemble tars. Residues with up to 741 chlorine maximum, l.e.. 2,200 keel/kg, may be Incinerated without additional fuel, whereas other types of Incinerators can only accept residues with 3,500 keal/kg caloric value, and require fuel addition. * HCl and Steam Recovery The present Salnt-Auban plant, built before the 1973 oil crisis, recovers only NCI. Should this plant be designed today, a heat recovery waste heat boiler (WHB) must be added. Such a WHS. used In another plant of the group since 1974, has been licensed several times for use with waste Incineration units. Recovered HCl Is a marketable 331 w/w muriatic acid and the steam generated In the WHB Is of average pressure In the range of 16 bars (230 pslg). 5-67 HONS 21SQ34 INCINERATION of PC8s * Types of PCBs Incinerated The average composition of the spent PCBs trademarked PTRAI.ENE Tl* comprised a mixture of: ~ 601 OP chloroblphenyl, which In turn, consisted of; 601 hexachloroblphenyl and $01 heptachlorobl phenyl 401 Trl- and tetrachlorobenzenes Fifteen tonnes of this product, liquid at room temperature, were charged to the VRC* Incinerator In Salnt-Auban and totally Incinerated. * Operating Conditions * Preparation of the Plant No modification whatsoever was made to the plant for the Incineration tests using PC8s. The absorbers were emptied and rinsed with fresh water. Consequently, due to the relatively small quantity of PCBs burned, the concentration of the MCI produced was low: 201 solution at the termination of the test. Technical Oata for October 1979 Test Duration of the test ................................................ 1$ hours PCBs feedrate.................... ..................................... 0.8 tonne/hr Gas temperature at furnace outlet .............................. 1,160C Retention time In the furnace .................................. $ seconds The 3 tonnes balance, retained from the dally storage, was burned together with a new supply of aliphatic chlorinated hydrocarbons. Note: The furnace was formerly heated by ordinary wastes Incinerated at the Salnt-Auban facility (l.e., from VCM and chlorinated hydrocarbon production), so the startup with PCBs caused no problems as the caloric value of the PCBs was sufficiently high to allow burning without additional fuel. 5-68 HONS 215035 r r SAMPLING Four samples were taken: Acid solution In the washing tower just after the quench, 0.810 (see Chart 1}. - Muriatic acid production, R.815. - Neutralized solution, R.818. Gas effluent In the stack, according to EPA Modified Method 5 (see Appendix). RESULTS Results of the tests are delineated in Tables 1 and 2. No PCBs, PCOFs, or PCODs were found at detection limits of 0.1 ppb (100 ppt), respectively. Destruction efficiencies for PCBs of 99.99997-99.99999* were realized both for liquid and gaseous effluents. ANALYTICAL METHODOLOGY * Scope The aiethodology used to establish the amount of PCBs In the samples taken during the Incineration tests was developed in the Rhone-Poulenc Research Center of Declnes near Lyon as part of the Rhone-Poulenc research program concerning PCBs produced In Its manufacturing facilities since 1946. This analytical protocol is a modification of the German "DAPA 1976 Analytical zeltung 280-9-13-1976" and the method published by Drs. Han Rudolf Buser and Hans Paul Bosshard. Principle The RGBs present In the neutralized samples (aqueous) are extracted with a non-mlsclble solvent (hexane). The solvent phase Is dried and concentrated and the PCBs present are analyzed by gas chrenography, coupled with mass spectrometry and mass fragnentography (GC/KS/NF). The discrimination among PCBs, PCDFs, and PCDDs Is accomplished using both differences In retention times on GC columns and molecular weights and the Isotropic ratios of chlorine content per each type. Quantltetlon Is achieved by coeparing the peak areas to well-known standards. The linearity of the responses In the working range of quantities Is routinely checked. 5-69 MONS 215036 Tibi* 1. liquids - Itesults t*<H***l K1 MU ntm ku um *! ftorltttc *cl* ttm i.tlt o frw n ^ rt k U*Wm frw * Htrilln WIMld Ina t lit tit Ml *n VrM !) IV" It 1.1 l.f MCt l< <t frit ti.i i.i t* nt Lr_Kfa m * Ktot t*'*1 L.pvr tov K*l I* best KM !> m* w* 1 fft . .11! -- 1 ,1 ------- *. if *-H H.WH .1. in ________ M ii l.lt n tmf <i> m irm IM l* JM *1 m M n tmt ..1 IU ______ ______ t(rU|tH (SHTN* *rM Uw IMI. It l*n rtrt*U tlaa, It* |lt Mini ml** tM *!'* H IN **!"'"t * It* Kit It riKt It* 'Utf In* f CMCMtrtUo to Um twin f mifton. Ill Fm* tlllKMil ft-tr(l*l ns JjAf MOMS 21903 7 MN CMHMC Tafcli 2. Gastaus Cfflums - Results 4^-M fell 4m) Irtppri U tm HKt tl to* 4m "t (Hit Hr 1 1 l (4)Mh to* m Iria rut frn fr 11 i 1 ft to to U ,{f <" " Mn INtorM Kl d*i ** to* ut ifi to to* w* ftp i* *> <* S'* itai toK*l K*l n* nr C, nn;tl tli fctlnv ( Im tffkt* m */fcr to* * IJ l Mt f.tm 141 ICM P lil mu cmatu 9M itn n imt a4I1t Mttk4IIlMh IMMulll . *t.t Wrn> III NlM M4CtlH IMl f 4.1 H4 MOMS 21S0B8 Ml seellaneoui In solution, chlorine content, unburned particles, ashes, phosgene, N0X, CO, CO;, and excess oxygen have been determined following standard methods. BIBLIOGRAPHY Federal Register. Vol. 44, No. 106, Nay 31, 1979, p. 31551. Federal Register. Vol. 42, No. 160, August, IB, 1977, Part II. Environmentel Protection Agency. Methods for determining the total PCBs emissions from Incineration. EPA Contract No. 68-02*17S0. Department of Environment. PCB wastes. Haste Management Paper No. 6. Her Majesty's Stationery Office, London, B/79. Miller, Stanton (1983). Chlorinated hydrocarbon wastes. ESIT, Vol. 17, No. 7. MOMS 21.5039 r r APPENDIX Modified Method 5 Sampling Train Sampling train, modified following the EPA recommendations concerning isokinetic sampling, pressure drop across system, and position of sampling line in a stack (different position of two diameters in the linear part of the stack). The total volume taken was more than one standard cubic meter for each sample. AVMlMlI 5-73 MONS 215040 Part 6 PCS FIPSS AND PCDF MONS 215041 PCBs, PCnF*, and PHlOs Resulting from Transfoimer/Capacitor Fires: An Overview Paul E. desRoslers Environmental Protection Agency Office of Research and Development Washington, DC 20460 Abstract The purpose of this paper is to give an overview of what the author considers to be current Information concerning POP transformer/capacitor fires In the IJ.S. On December IS, 19B3, the Environmental Protection Agency Issued a national dioxin strategy fnr investigating. Identifying, and cleaning up sites contaminated by dioxin, within the framework o* this strategy Is a plan that calls for research to he conducted on PCD transformer/capacitor fires. This research will entail an Investigation Into the characterization of sources of PCDFs and PCDDs, that Is, the chemistry of PCR fires and theoretical products of thermal stress based on composition of askarel used; emergency response protocols for firemen. Including protective clothing; and practical remedial measures for building cleanup. The author also feels that only through the sharing of Information and full cooperation among electrical utilities, chemical manufacturers, firemen, Insurers, EPRt, EPA, and N10SH can "preventive maintenance" be performed to minimize or preclude future Impacts of PCB transformer/capacitor fires. PREFACE On January 10, 1979, a railroad tank car carrying 75,000 liters of crude o-chlotophenol ruptured near Sturgeon, Missouri, Forty-seven workers from the Norfolk 6 Western Railway Company were employed to clean up the spill and in the process were exposed to a variety of toxic chemicals. Including chlorinated phenols, phenol, and the toxic artifact, 2,3,7,8-TCDD (1). In fact, the concentration of this dioxin isomer amounted to only 22 ug/kg (ppb) (2), As a result nf alleged exposure and the manifestation of certain perceived health effects and physiological symptoms, e.g., rashes, dizziness, loss of memory, extreme fatigue. Impotence, upper respiratory difficulties, and in one case, cancer of the testicles and the skin, the workers brought suit in en Illinois court against the railroad company and the manufacturers of the tank car and the faulty coupling, including Monsanto, the manufacturer of the orthochlorophenol. On August 26, 19R?, the Edwardsville, Illinois Jury awarded the highest monetary judgments ever made as a result of a 6-1 HONS 215042 single tank car spfll - S57.95 mil Hon (1). Prior to the cnmmencemenr of the April 5 trial, Monsanto, (ieneral American Transportation, and Dresser Industries settled out of court for an estimated SB million. This court decision represents the first time a jury has concluded that dioxin caused permanent harm to humans and thus has highlighted the more than 30 waste disposal sites within the State of Missouri that contain polychlorinated dlbenro-p-dioxlns (PCPDs)(3). Recently, In November and December of 1983, three separate suits (A)(5) were filed in St. Louis Circuit Court: in one case, by S7 persons who contend they suffered general systemic effects from dioxin-tainted soils at six sites In eastern and central Missouri for (684 million; In a second case, by 25 dockworkers and a widow of a former dockworker of a trucking firm In St. Louis - for $620 million; and finally, by 183 people at Times Beach and Castlewood whn assert that they suffered serious health deficiencies as a result of dioxin exposure - for $1,8 billion, Therefore, this situation has catalyzed an acute public and governmental awareness of the potential severity and explosiveness o* the PCDO problem. PCS transformer 'Ires, accidental spills, and Inappropriate disposal have compounded the problem particularly (n regards to PCDFs (furans) and to some extent PCBPs (blphenylenes). INTRODUCTION The most toxic and extensively studied polychlorinated dtbenzop-dloxln (PCDO) and -furan (PCOF) Isomers are 2,3,7,8-tetr achloro-pdioxin (2,3,7,B-TCDO) and 2,3,7,8-tetrachlorodltoenzofuran (2,3,7,8TCDF), respectively. There are pronounced differences In biological and toxicological effects between different PCDD and PCDF Isomers. Those Isomers with the highest acute toxicity are 2,3,7,8-TCDD, |UUIiii<M |UU-m,KM -'fa* Figure 1. The most toxic PCDO and PCDF isomers (6). HONS 215043 6-2 f r Each of these isomers has its respective four lateral positions substituted for chlorine and each has LDco values in the range of 1-100 ug/kg for the most sensitive animal species (generally the male guinea plo) There exist 7$ PC00 Isomers and 13$ PCDF Isomers. Acute toxicity among the Isomers can vary up to 100,000 tfmes (See Table 1). I" comparison, the most acutely toxic PCR Isomer Is only about one-fourth as toxic as 2,3,7,B-TCDD (B). -- Table 1, Acute oral toxicity nf polychlorinated dlbenzo-pdloxlns as dose expected to cause death of 5(ir. of the animals within 30 days (7), Acute Oral LQcn. ud/kq Oody Weight PCOO 01 2,7 2.8 Guinea Plq >300,000 Rat >1,000,000 Mouse >2,000,00(1 Til 2,3,7 30,000 >3,000 Tetra 1.3,6, 1,3,7,9 . 2,3,7.8 >100,000 0.6 22 (H) 2,1 45 (F) 2 780 Penta 1,2,3,7,8 l.2,,7,8 3 1,100 340 >5,000 Hexa 1,2,3,4,7,8 1,2,3,6,7,B 1,2.3,7,8,9 Mixed Isomers Hepta 1,2,3,4,6,7,8 Octa 1,2,3,4,6,7,0,9 73 70-100 60-100 >600 100,000 825 1.250 >1,440 >1,000,000 >4,000,000 MOMS 215044 6-3 rr Recent Information arising from a 1983 National Cancer Institute carcinogenicity study of a mixture of 1,2,3,6,7,R- and 1,2,3,7,8,9-I1C0D In female rats led to the preliminary conclusion that thfs mixture of hexachlorodloxins was carcinogenic. Recause public comment criticized the manner in which the experiments were conducted, the National Institute of Environmental Health Sclences/Natlonal Toxicological Program (NIEHS/NTP) was asked to reexamine the results o* the study. It Is therefore the opinion of NIEHS/NTP that the hexachlorodtoxin mixture was indeed carcinogenic under the conditions r>* the test (9). Analyses of Commercial PC8 Mixtures PC8s contain a complex mixture of PCOFs with up to 40 isomers hemq present. Rappe and Buser (6) have analyzed a number of comterclal PCP-s and their findings are shown In Table 2. The highest concentration of PCOFs, found In a Mitsubishi PCS fluid used for over two years, amounted to 10 ug/g (ppm), with 1.25 ug/g being the 2,3,7,B-TC3F Isomer. When commercial mixtures nr well defined isomers are heated in air to a scaled glass container or ampoule, yields of PCOFs have been observed in the 1-si range and this technique has been successfully employed to prepare PCDF standards (10). Using the synthetic standards currently available, the major PCDF Isomers present in commercial PC8s have been Identified (See Figure 2). UV'lawt-Cor IUIHP lAJdM-ttapii CM Cl a Cl UM-ttHfCM UU*+Mli"CW IHUInticW FI guv PC0F Isomers identlffed In eonnerefal PCfls (6) HONS 215045 6-4 r r Table 2. Concentrations of PCOFs in commercial PCRs, ug/g (6). Sample 3--Cl 4-C1 5-C1 6-C1 7-C1 Aroclor 1248 (I960) Aroclor 1254 (1969) Aroclor 1254 (1970) Aroclor 1254 Aroclor 1254 (lot KK 602) Aroclor 1260 (1569) Aroclor (Lot AK 3) Aroclor 1260 Aroclor 1016 (1972) Clophen A 60 Clophen T 64 Phenoclor DP-6 Prodelec 3010 Mitsubishi (used) - n.io - 0.06 - * 0.10 - 0.41 2.13 0.5 0.1 0.2 0.25 0,05 0.1 0.2 0.30 <0.001 1.4 0.30 0.7 1.00* a.oah 1.2 0.2 0.4 0.70 0.10 0.4 0.3 1.0 <0,001 5.0 1.73 10.0 0.35 3.30 0.3 1.4 0.9 0.B1 0.02 0.5 0.3 1.10 <0.001 2.2 2.45 2.9 0.07 0.53 1.35 n.82 Total PCOF 2.0 3.7 1.5 1.9 o.2 1.0 O.R 3.8 8.4 5.4 13.6 2.0 10.0 * Major isomer 2,3.7,B-TCC>F b Contains 1.25 ug/g 2,3,7,B-TCDF Olscutslon of Various PCB Fires * Binghamton, New York On February 5, 1981 at approximately 5:3(1 a.m., an electrical panel In the basement of a 22 story office building located In a governmental complex was Involved in an Incident described as an explosion. A nearby electrical transformer containing about llfKl gallons o* Arcelor 1254 (657) and tt 1- and tetrachlorinated benzenes (357.) was Involved, leakage occurred and about 180 to 200 gallons of fluid leaked from the transformer. Hue to the ventilation system of the building, smoke contaminated with soot-containing chemicals, was spread In a non-uniform fashion to most work areas of the building as well as to interstices within air conditioning ducts, false celling arees, elevator shafts, and within desks and file ceblnets. nils appears to have been facilitated by air ducts or simply open shafts running from the basement area, up to the top of the building, with openings Into the men's and mmen's bathrooms on each floor. These bathrooms hed only metal gratings separating them from the open shaft area. In addition to this architectural arrangement, fire or smoke safety doors, which were designed to open In the presence of smoke as a health measure, did Just that during the series of explosions that occurred at the time of the Incident, Because the outside air temperature was below freezing at the time, the opening of safety doors on the roof, which were located above the building's two stairways, caused a small vacuian that sucked air containing smoke, soot, and as was later found, toxic chemicals from the bathroom areas on each floor and deposited the residue in a non-uniform fashion Into most areas of the office building later tested. MOHS 215046 6-5 Chemical analyses were performed at: Galson Laboratory In Syracuse, New York and at General Electric Laboratory In Schenectady, New York and revealed PCBs In soot and air , which was strongly suggestive o* PCDFs present In substantial amounts; The National Fish and Wildlife Laboratory In Columbia, Missouri, which established the presence of many of the dioxin isomers; and by the New York State Health Department's Division of Laboratories and Research, which established the presence-Of 2,3,7.8TCDD, A later examination of material from the underground garage that was contaminated and washed (twice with steam and detergent) was found still to have trace amounts (nanograms per square meter) of PCDEp end PCDOs in some areas by Wright State University. Total PCDF Isomers In the soot were Initially found to be as high as 2160 ug/g, PCDOs tO-20 ug/g, and PCBs 100,000-200,000 ug/g (See Table 3). It Is noteworthy that the most toxic Isomers are the major components within each group, viz*, 2,3,7,B-TCDD, 2,3,7,8-TCDF, 1,2,3,7,8-PCDO, 1,2,3,7,8-PCDF, and 2,3,4,7,8-PCOF (See Figure 1). Since the trl- and tetrachlorInated benzenes comprise 35X of the dielectric fluid, It Is not unreasonable to assume that this diluent mixture represents the precursor to the PCOOs produced as a result of the fire. Nearby Binghamton city Hall, which was used as a staging area during the Initial cleanup efforts In February 1981, was also contaminated. The County Building also next door to the Binghamton State Office Building was contaminated to a lesser extent. There have been $12 million allocated so far by the State of New York for the cleanup. The building Is self-insured so no Insurance funds are available. It has been closed since February 5, 1981, Over $1 billion In lawsuits have been threatened by some of the more than 500 persons who were, or believe they were, exposed to toxic chemicals, and who are themselves concerned with physical or psychological/medical damages. * Stockholm. Sweden In August 1981, a fire broke out In a 10 XV capacitor battery In an electrical power station In Stockholm, which was probably caused by a malfunctioning electrical system. Wipe tests were taken about one meter from the capacitor (13). Inspection of the chromatograms revealed several peeks. In addition to PCBs, with molecular weights corresponding to dl-, trl*, tetr a*, and pentachl or obi phenyl enes (PCBPs). It Is probable that these PCBPs were formed by a direct eye 11 ration of PCBs present. The level of PCDFs was found to bn 1375 ng/m< (See Table ), whereas a rough estimate for PCBPs gave levels of 25,000-30,000 ng/m*. Polychlorinated pyrenes (PCPYs) were also detected, hut not quantified. HONS 21*047 r r Table 3. Levels of PCDFs (ug/g) from accidental burning of PCB-contaimng electrical equipment. Isomer s Capacitor Transformer Skovde, Sweden(10) Binghamton(lO) 0.5m* 3m** Total PCDFs Total Tri-CDFs Total Tetra-CDFs 2,3,7,8-Tetr a-CDF Other isomers 0.8 0.6 0.1 0.5 0.2 0.1 0.02 0.06 2160 28 12 16 Total Penta-CDFs 1,3,4,7,8-Penta-CDF 1,2,4,7,8- " 1.2,4,7.9- 11 1,2,3.7,8- " 1.2.3,6,7- " 1,2.6,7,8- " 2.3,4.7.8- 11 2.3,4,6,7- " Other isomers 0.1 0.04 670 65 25 22 31" 60 25 48 12 110 Total Hexa-CDFs 1,2,3,4,6,8-Hexa-CDF 1,3.4,6,7,8- " 1,2.4,6,7.8- " 1,2,3,4,7,8- " 1,2.3,6,7.8- " 1,2,3,6,8,9- " 2.3,4,6,7,8- " Other isomers 0.04 0.04 965 50 125 50 510 150 58 10 250 Total Hepta-CDFs 0.01 1,2,3,4,6,7,8-Hepta- CDF 1,2,3,4,6,7,9- 1.2,3.4,6,8.9- " 1,2,3,4,7,8,9- 11 0.01 460 230 120 55 55 Octa-CDF 0.005 0.005 40 Tr ansformer Miami (11)# n.d.'lJ1) n.d.-0.18 n.d.-0.53 n.d. n.d.-l.0 n.d.-0.18 n.d. n.d. Transformer Boston (12)t 165 50 60 3 35 15 2 n .d * soot sample*taken 0,5m from capacitor on the floor soot sample taken 3m above capacitor on the wall f Detection limit: 10 ng/g t Detection limit: 100 ng/g MQNS 215048 Table 3. Levels of PCDOs (ug/g) from accidental burning of PCR-contalnlng electrical equipment, (cont'd) Isomers Capacitor Tr ansformer SkBvrle, Sweden(lO) Plnghamton(lO) 0.5m* 3m** Total PCODs Total Tr1-CODs Total Tetra-CODs 2,3,7,8-Tetra-CDFs Other Isomers Total Penta-CDDs 1.2.3,7,0-Penta-CDDs Other Isomers - " _ - - - _ * 20 1.2 n.6 0.6 5.0 2.5 2.5 Total Hexa-CDDs 1,2,3,4,6,8-Hexa-CDt) 1.2,4,6,8,9 " 1,2.3,4.7,0 " 1,2,3,6,8,9 " 1,2,3,7,8,9 " 1,2,3,4,6,7 H - - * - - 4.7 1.2 1.2 0.7 0.6 0,4 0.5 1.2,3,4,6,7,9-Hepta- CDD - 1.2,3,4,6,7,8- - * * 4 3 Octa-CDO *- 2 Transformer Miami (11)# n.d. n.d. n.d. n.d. n.d. n.d. n,d. n.d. n.d. Transformer Boston (12)t n.d. n.d. n.d. n.d. n ,d. n.d. n.d. n.d. n.d. * soot sample taken 0,5m from capacitor on the floor ** soot sample taken 3m above capacitor on the wall # Detection limit: 10 ng/g t Detection limit: 100 ng/g HONS 215049 6-B r r Table 4. Analyses of wipe samples from various PC8 fires. Isomer s 5kovde(14) ng/m2 Total PCDFs n.d.-873 2,3.7.8-Tetra-CDF n.d.-lOO 1,2,7,8-Tetra-COF 2,3,6,8-Tetra-CDF 1,4,6,9-Tetra-CDF 2,4,6,7-Tetra-CDF 3,4,6,7-Tetra-COF 1,3,6,7-/1,3,6,9-Tetra -CDF Other Tetra-CDFs Total Tetra-CDFs <1-600 2,3,4,7,8-Penta-CDF 1.2.4.7.8-Penta-CDF 1,2,3,6,7-Penta-CDF 1,3,4,7,3-Penta-CDF 2,3,4,6,7-Penta-CDF 1,2,4,6,8-Penta-CDF 1,2,4,7,8-Penta-COF 1,2,3,6,7-Penta-COF 1,2,4,8,9-Penta-COF 2,3,4,6,8-Penta-COF 1,2,3,7,8-/1,2,3,4,8-Penta-CDF Other Penta-COFs Total Penta-COFs n.d.-lOO Total Hext-CDFs n.d,-60 Total Hepta-CDFs n.d ,-8 Octa-CDF n.d.-5 Total PCDDs Total Tetra-CDDs 2,3,7,8-Tetra-CDDs n.d. n.d. n.d. StOCkhglia(l4) ng/m? 1375 150 isn 125 75 37 7.5 300 750 1200 45 38 15 11 7.5 3.8 3.8 3.8 3.8 2 15 19 175 <o.5 . Cincinnati (13) *9/m? n.d. n.d. n.d. n.d. n.d. n.d. MONS 215050 6-9 r' Skovde. Sweden In March 19B2, a 40D volt capacity battery, serving a high frequency oven In a casting line of a foundry In Sk&vde, experienced an electrical malfunction leading to a fire (13)(14). The dielectric fluid employed comprised either mineral oil or PCBs. The fire started in a mineral oil capacitor and burned for about two hours before being extinguished. Temperatures exceeding 1I00*C were realized because copper electrical wiring melted. The smoke spread from the basement into the building above, an area of 60 x 30 meters. The capacitor battery contained 21 capacitors filled with 5 kgs of PCBs each. Post-fire Inspection showed leakage from 12 of these capacitors. Both wipe tests and moot samples were taken. Results are tabulated In Tables 3 and 4. These show that high levels of PCOFs (>100 ng/m? and <800 ng/g) could only be found In an area near the fire. Concentrations of PCDFs dropped off dramatically with distance from the fire. The highly toxic 2,3,7,8-TCDF was one of the major tetra-CDF Isomers present. Other chlorinated compounds that may be present are the PCPYs, but verification could not b obtained due to lack of standards. No PCDDs were Identified in samples taken from this fire (there were no chlorinated benzenes used as diluents). No PCBPs were observed. Miami. Florida On April 13, 19B2, an underground electrical fire occurred in a transformer vault in Miami. City of Miami firemen were called In to extinguish the blaze. They also voiced their concerns about the possible contamination of fire equipment and protective clothing used during the performance of their duties. As a result, the International Association of Fire Fighters in Washington, DC, contacted the NIOSH Region IV representative and the latter agency conducted a thorough investigation. Table 5 contains the results of the analyses of surface samples taken during the Investigation. As expected, samples removed from inside the vault indicate the heaviest contamination levels (up to 27,400 ug/100 cm` PCBs). Results of prior NIOSH investigations Indicate that normal background levels of PCBs on uncontaminated surfaces should be less than 0.5 ug/lftO cm2. As Indicated by the sampling results, turnout coats, boots, helmets, and other personal protective equipment were not found to be contaminated from the fire. This Is probably due to the decision to allow the fire to self-extlngulsh, thus minimizing exposures to the smoke and soot. The only contaminated piece of equipment found at the station was the smoke ejector fan, which contained greater than 31 ug/100 cm2 PCBs because the area wiped for this particular sample was estimated to be less than the standard 100 cm2. ^ fan blade guard was found to be heavily coated with soot and dirt from the smoke exhausted from the vault fire. Results of PCDO and PCDF analyses are shown in Table 3, No PCDDs were detected, but trl-, tetra-, pent a-, and hexa-CDFs were found, ranging from non-detectabte (n.d.) to 1790 ng/g. No 2,3,7,8-TCDF was detected In any of the samples. However, the samples did exhibit high levels of PCBs through Clio and polychlorinated diphenyl ethers (PCDPEs) through Clg. MOHS 215051 6-10 r r Table 5. PCB residue from Miami transformer vault fire (11). Sample Location Inside transformer vault Wall behind transformer Top of primary cable above fire Primary cable support bracket Floor near Isolating switch Celling near fire location Secondary bus near vault celling Wall next to exit ladder Rung of exit ladder Above transformer vault: Sidewalk grating Water at curb near vault Firemen's clothing/equipment: Miscellaneous clothing Smoke ejector fan Normal PCB Background*** Type Sample^ PCBs* fug/100 cm2) (soot) wipe (soot) wipe (soot) wipe (dirt) wipe (soot) wipe (dirt) wipe (dust) smear tab--dry (dust) smear tab--dry 414** 3M 704 27,400 860 195 O4. 79 smear tab--dry wipe 2 3 smear tab--dry smear tab--dry wipe <0.1 >31 <0.5 * As Arodor 1260 (used as standard for quantifying samples). ** As mixture of Arodor 1254 (231 ug) Arodor 1260 ( 203 ug). *** As determined by previous KIOSK Health Hazard Evaluation. # Wipe samples were collected on sterile cotton pads soaked In n-hexane. 6-11 HONS 215052 Cincinnati. Ohio On December 3, 1980, a capacitor containing PCBs for a 1/2 HP electric motor overheated in a basement storage room of a private elementary school. This capacitor contained 2? ml of rfielectrlr fluid comprised of 21.9 ml of a biodegradable fluid and 0.1 ml of PCRs. Approximately one-half of the fifth-grade students using three classrooms in the basement area complained of itchiness of the skin. Poena to the electric motor was shut off and the exterior door was opened to allow ventilation of the smoke from the room. Since air Is locally recirculated, central distribution via a mechanical system throughout the school did not occur. PC8s were not detected in any air sample at a limit of detection of 0.05 ug/sample. which is equivalent to an airborne concentration of approximately 1 ug/m3. Table 6 summarizes the results of 63 wipe samples collected by NIOSH. Fifty-one of the 63 samples were taken from surfaces within the school including books, desks, walls, floors, ceilings, and other surfaces. Two were taken from surfaces within a mobile classroom and ten were obtained from surfaces at three locations In eastern, central, and western Cincinnati to establish background levels of PCBs for this area. The 20 samples obtained from Room 27, which contained jjhe overheated capacitor, ranged from non-detectable to 7200 ug/100 cm*PCBs. The remaining rooms ranged from non-detectable to 0.45 ug/100 cm*. The two wipe samples with the highest PCB concentrations were also analyzed for PCDDs and PCDFs, with particular emphasis on the 2,3,7,8-tetia-Isomers. Neither PCDOs nor PCDFs (See Table 4) were detected in either of the samples. The detection limits employed were 0.10 and 0.09 ug/sample, respectively. * Boston. Massachusetts In January 1982, there was an electrical fire Involving PCBs in a Boston office building. One bulk soot sample was analyzed for PCDOs and PCDFs. Table 3 shows that high concentrations of the tri- through the hepta-CDFs were found in the soot sample, rangiAg from 2-60 ug/g (detection limit * 0,1 ug/g). No PCDOs were found. A considerable amount of the highly toxic 2,3,7,8-TCDF was detected--3 ug/g. Levels of PCBs were also found at 10 - 100 times that of the PCDFs. MONS 2150S3 r r Table 6. Summary of PCBs wipe sample results. Our Lady of Visitation School, Cincinnati, OH (capacitor fire) (15). Sample Location Room 27 (capacitor location) Room 25 Room 24 Room 26 Ha 11 way Room 8 Room 22 Room 10 Room 11 Room 12 A.V. Room Pr incipal1 s Office Mobi1e Classroom Background Levels (3 Cincinnati locations) Mean PC8 lavel. ug/100 cm2 * n.d. - 7200** n.d. 0.14 n.d. - 0.06 n.d. n.d, - 0.07 n.d. - 0.08 n.d. - 0.29 n.d. - 0.20 0.07 - o.ii 0.06 - 0.09 n.d. 0.05 - n.<5 0.06 n.d. - 0.13 * Detection limit: 0.05 ug/sample (jjample quantity wiped per 100 cm*-). '* Two wipe samples with highest concentrations of PCRs were found tu contain no detectable PCOFs or PCODs. * St. Paul, Minnesota On June 22, 1982, a transformer fire occurred In a public high school In St. Paul. NIOSH took both surface wipe and air samples on June 23 (the day after the fire) and on July 7 (a*ter the area was decontaminated). Table 7 contains measurements of air contamination, pre- and post-cleanup, by PCBs, trichlorobenzenes, and tetrachiorohenzenes. The NIOSH recommended perarissable limit of 0.001 mg/m3 (B-hr TWA) for PCBs was exceeded in all areas saaipled except for two locations. The trl- and tetrachlorobenzenes found did not exceed any existing limits. Table S shows the effect of the decontamination methods employed. The transformer vault room PCBs (wipes) were reduced from a high of 4,000 U9/100 cm2 to 120 ug/100 cm? or less. No PCDDs or PCDFs were found in the liquid sample of Askarel provided (detection limit 40 ng/9 (40 ppb)). However, it is Interesting to note that NIOSH did not analyze for PCODs or PCDFs in the soot. The results may have shown significant concentrations of these isomers, given the areal concentrations of trl- and tetrachlorobenzenes detected. 6-13 HONS 215054 I* Table 7. Areal concentration of PCBs and chlorinated benzenes. Hi 11-Hurray Hioh School, St. Paul, MN (transformer)(1C). Sample Site Airborne Concentration, mq/m3 PCBs *1 T3CBs _ T4CRs Transformer vault, 5 ft above floor Transformer vault, 1 ft above floor Outside of transformer vault doors Hood shop Head of stairs Cafeteria A 0.05 R 0.DD7C 0.09 . 0.02 0.004 n.d. n.d. n.d. n.d. n.d. n.d. A 22.6 B 0.331 17.6 - 10.2 0.73 0.12 0.35 0.048 0.020 0.011 0.010 A IB.2 B ".651 25.7 - 11.7 1.21 0.26 0.27 0.229 0.0B9 0.037 0.032 8-hr time-weighted average exposure criteria 0.001* 40b none A June 23, 19B2; B July 7 , 19B2 * Detection limit : 0.D01 mg/m3 # as Aroclor 1260 a NIOSH recommended permissable exposure limit. OSHA permlssable exposure limit Is 0.001 mg/nr 6-hour time weighted average. b ACGIH threshold limit value. Neither KIOSK nor OSHA has exposure criteria. c sample taken 3 ft above floor 6-1.4 MONS 215055 f f Table 8. Analyzed PCBs* in wipe samples, Hi 11-Hurray High School, St. Paul, MN (transformer) *(16). Sample Site Vault room Ventilation dust Corridor outside vault room Cafeteria Gymnasium Rooms 106, 109, 138 Outside Room 106 Woodshop PCBs#, ug/100 cm2 A n.d. - 4000 n.d. B 2 - 120 - n.d. 0.22 - 0.24 0.29 0.26 2.1 5.B n.d. - n.d. n.d. O.B 0.8 * # Detection Limit 5 ug/sample A - June 23, 1962; B July 7, 19B2 t no analyses conducted for PCDFs or PCDDs. 6-15 ' MIMS 215056 r r' DISCUSSION The previous PCB tr ansformer/capacitor Incidents are but a representative sampling of cases where data are readily available. Other electrical equipment episodes include; Toronto, Ontario, Canada In 1977; and Norrtrflje and Surahammar, Sweden and Imatra, Finland In 19B2 (17). More recently, In the u.S., three such episodes have occurred: Date 5/15/83 9/2B/83 12/21/83 Location San Francisco, CA Chicago, IL Syracuse, NY PCB Fluid 1242 only 65t Aroclor/357, Chlorobenzenes B7X 1254 Because of the continuance of such Incidents, the Environmental Defense Fund and the Natural Resources Defense Council challenged the August 25, 19B2 PCB Electrical Use Rule Issued by EPA. As a result, a Start Action Request has been initiated hy the EPA Office of Pesticides and Toxic Chemicals, which Includes the following court*ordered schedule for the Issuance of an Advanced Notice of Proposed Rulemaklnq (ANPR) regarding PCB transformer fires (IB); ANPR: March 1964 Proposed Rule: October 1984 Final Rule: July 1965 The objective of an ANPR Is to present data that Indicate a reason for concern and, more Importantly, to solicit data for the Proposed Rule. The major elements of the ANPR include: e Description of risks associated with PCB transformer fires through analyses of Binghamton, San Francisco, Chicago, Syracuse, etc. events. a Description of number and distribution of PCB transformers, e Solicitation of data on number and location of such electrical equipment. e Presentation of Information on other lesser-known fires, e Presentation of estimates of frequency of fires, e Solicitation of data on other fires and frequency of other fires, e Discussion of PC0F/PC00 formation (and PCBP, PCCY (crysenes), PCOPE, PCPY, etc. where data are available on these compounds), t Solicitation of data on mechanisms of formation, e Presentation of possible regulatory options, e Substitutes e Retrofllllng e Fire hazard Inspections e Costs/benefits of various mechanisms to reduce spread of PCBs, PCDFs, PCDOs, etc. from fires - use of early warning protection devices HONS 215057 6-16 r r Currently, the EPA Office of Research and Development (ORD), in support of the Agency's Dioxin Strategy (18), released publicly on December 15, 1983, Is planning to conduct research on PCB fires, which will cover three areas: (a) the chemistry of PCB fires and a theoretical discussion of products of thermal stress based on composition of askarel used; (b) emergency response protocols for firemen, including protective clothing (this research will be conducted In cooperation with NIOSH); and (c) practical remedial measures for building cleanup. As a result of iqy participation in the preparation of the Dioxin Strategy document and in several Agency dioxin workgroups. Including the Dioxin Disposal Advisory Group and the PCB Transformer Fires Workgroup and based upon my expertise concerning the subject matter (20), I believe that I enjoy a perspective that few EPA technical staff have. Because of this, I can relate to you the following facts: Given that chemical production facilities In the U.S. no longer produce ?,4,5-trIchlorophenol (the most significant historical source of 2,3,7,8-TCDD), 1 can state that soot produced as a result of PCB transformer/capacitor fires contains the highest concentrations of PCDFs and PCDDs (the latter only If chlorobenzene diluents are present in the askarel mixture) found In the U.S. today; on a weight basis, however, municipal combustion devices account for the largest Quantity of PCDFs and PCDDs produced currently. The soot front the R1 ngriamton fire was evaluated by the New York State Department of Health In tests using chick embryos In 1981 (21). Results were positive. In other words, the PCDDs and PCDFs in the soot were not inactivated (as In the case of fly ash from an Incinerator or by activated carbon) as demonstrated by the chick embryo fetotoxlclty and teratogenicity tests performed. The soot can therefore be considered toxic and may contribute to "environmentally Induced or promoted cancers, reproductive abnormalities, Imnunologlc deficiency, and possibly premature death from infection, neurologic and gastrointestinal system pathology and possibly, by elevating serum triglyceride and cholesterol levels, excess or avoidable cardiovascular and cerebrovascular pathology" (22, p. 676). I thus stand firmly behind the convictions that firemen responding to such Incidents must have the most up-to-date Information at their disposal and that remedial methods employed must be practical, sound, and effective. As a result, EPA-ORD Is In the process of securing PCS samples from Binghamton, NY, Chicago, 11, San Francisco, CA, and Syracuse, NY. The EPA analytical laboratory In Cincinnati, OH will conduct a chlorobenzene isomer screen on the samples together with homologue analyses for PCBs, the purpose being an attempt to correlate PCDDs and PCOFs content of soot and wipe samples (in the vaults) with chlorobenzene and PCB content In the transformer fluid. The Information gathered will be shared among the electrical utilities, Insurers, firemen associations, EPRI and NIOSH. In closing, I wish to share some recent and quite pertinent Information. The NIEHS/NTP has concluded two animal studies Involving bioaccumulation of 2,3,7,8-TCDD from contaminated soils and has submitted these papers (23) to Science for publication. These papers conclude that 2,3,7,B-TCOO In soil from two sites In Missouri, namely, Times Beach and Mlnker-Stout, Is absorbed In a highly efficient manner (>501S) and that 2,3,7,8-TCDD 6-17 MONS 215058 -contaminated soil presents a hazard to humans if Ingested. Thus, further laboratory evidence has been amassed regarding potential risk to humans for 2,3,7,8-TCOO. I believe that this audience Is sufficiently astute to comprehend the long-term implications of these investigations. References 1. Chemical Week. September 1, 1982, p. 14. 2. Harless, Robert. Deposition In Lowe v. Norfolk l Western Railway Company, Circuit Court, Third Judicial Court of Illinois, Madison County, No. 79-L-81D. 3. Kleopfer, R.O., Runn, W.W., Yue, K.T., and Harris, D.J. (1982). "Occurrence of TCDD In Environmental Samples from Southwest Missouri." Paper presented at the 184th ACS National Meeting, Kansas City, MO, September 14, 1982. 4. Anonymous (1983), "57 Missourians File 664 million Dioxin Suit." Post Tribune. Jefferson City, MO, November 29, 1983. 5. Anonymous (1983). "Missouri Dioxin Suit Seeks $1.8 billion." New York Timet. New York, NY, December 17, 1983, 6. Rappe, Chrlstoffer and Buser, Hans Rudolf (1981). "Occupational Exposure to Polychlorinated Oloxlns and Dlbenzofurans" In Chemical Hazards In the Workplace--Measurement and Control. ACS Symposium wtesHonw.-- ----------------------- 7. Leng, M. L. (1979). "Comparative Toxicology of Various Chlorinated Dioxins as Related to Chemical Structure" In CIPAC Proceedings Symposium Series I. Heffers Printers Ltd., CemSridge, England, 8. Safe, S., et al. (1982), "PCBs: Structure-Activity Relationships." Paper presence? at the Symposium: "Recent Advances In Exposure, Health, and Environmental Effects Studies of PC8s," Rethesda, MD, May 12-13, 1982. 9. Letter of October 31, 1983 from Paul K. Hlldebrandt, O.V.M., Tracor Jltco, Rockville, MO, to E.E. McConnell, NIEHS/NTP, Research Triangle Park, NC, subject: Review of NCI Carcinogenicity Study of Hexachlorodlbenzo-p-dloxln in Rats. 10. Rappe, Chrlstoffer and Marklund, Stellan (1982). "Thermal Degradation of Pesticides and Xenoblotlcs: Formation of Polychlorinated Oloxlns and Dlbenzofurans." Department of Organic Chemistry, University of Umet, Umet, Sweden. 11. Letter of October 27, 19R2 from Stanley A. Salisbury, NIOSH Region IV to Richard Duffy, International Association of Fire Fighters, subject: PC8 Transformer Vault Fire, Miami, FL, 6-18 HONS 215059 r r 12. Memorandum of October 1, 1982 from 0. Choudhary and J. 0. Posner, NIOSH, Cincinnati, OH to Kevin McManus, NIOSH Region 1, subject: PCB Soot Analysis from Fire at U.S. DOL-OSHA. 13. Rappe, C., Marklund, S.t Bergqvrst, P.-A., and Hansson, M. (1982). "Polychlorinated Dioxins, Olbenzofurans and Dther Polychlorinated Polynuclear Aromatics Forming Ourlng Incineration and PC8 Fires." Department of Organic Chemistry, Unlversfty of Umel, S-90187, timet. Sweden. 14. Rappe, C-, Marklund, S., Sergqvlst, P.-A,, and Hansson, M. (1982). "Polychlorinated Dioxins (PCODs), Olbenzofurans (PCDFs) and other Polynuclear Aromatics (PCPNAs) Formed During PCR Fires." Chemlca Scripts. Vol. 20, pp. 56-61. 15. Health Hazard Evaluation Report: Our Lady of Visitation Elementary School, Cincinnati, OH. NtOSH Report No. HETA 81-237-915, July 1981. 16. Letter of July 26, 1982 from John R. Komlnsky, NtOSH, Cincinnati, OH to Darrell E. Anderson, State of Minnesota Department of Health, Minneapolis, MN, subject: PCB Fire at Hlll-Murray High School, St. Paul, MM. 17. Vuceta, J., Marsh, J.R., Kennedy, S., Hlldemann, L. and Hi ley, S. (1983). "State-of-the-art Review: PCDDs and PCDFs In utility PCB Fluid." Prepared by SCS Engineers, Inc. for the Electric Power Research Institute, Palo Alto, CA. EPRI CS-3308, Research Project 1263-11, November 1983. 18. Memorandum of October 11, 1983 from Martin P. Helper, EPA Exposure Evaluation Division to Irwin p, 8aume1, EPA Health and Environmental Review Division, subject: PCR Transformer Fires In Buildings. 19. Environmental Protection Agency (1983). "Dioxin Strategy." Office of water Regulation and Standards and Office of Solid Waste and Emergency Response In conjunction with the Dioxin Strategy Task Force, Washington, D.C., November 28, 1983. 20. desRoslers, Paul E. (1983). "Remedial Measures for Wastes Containing Polychlorinated D1benzo-p-d1ox1n$ (PCDDs) and Dlbenzofurans (PCDFs): Destruction. Containment or Process Modification." Ann. occuo. Hyq.. Vol. 27, No. 1, pp. 57-72. --------------- 21. Seely, Hart (1983). "The Building That Won't Come Clean." The Syracuse Herald, July 31, 1983. -------------------- 22. Schecter, Arnold (1983). "Contamination of an Office Building In Binghamton, New York by PC8s, Dioxins, Fur ans, and Blphenylenes after an Electrical Panel and Electrical Transformer Incident." Chemosohere. Vol. 12, No. 4/5, pp. 669-680. ---------- --------23 23. Letter of November 9, 1983 from Edward M. Brandt, Jr., M.D., Assistant Secretary for Health, to Lee Thomas, EPA, AA for Solid Waste and Emergency Response, subject: NIEHS/NTP Studies on Bioaval 1abl11ty In Guinea Pigs and Rats of Dioxin In Soil (with two ahtr?t submitted to Science for publication.) MONS 215060 6-19 r* QUESTIONS AND ANSWERS QUESTION: ANSWER- What can EPRI do to help solve the problem? Identifying the actual risk from transformers would be helpful. Several discussions questioned the source of PCDD and PCDF from fires, pointing out that, generally, aaterlals othar than from transformers are Involved. 6-20 HONS 215061 POLYCHLORINATED OIBENZOFURANS AND PCB-TRANSFQRMER FIRES: TOXIC HAZAROS ANO DETERMINATION OF DECONTAMINATION GUIDELINES Thomas K. Mil by. M.O.* * Thomas L. Forrester In 1970, Vos and his colleagues (U Identified polychlorinated dlbenzofurans (PCOFs) as toxic Impurities In European PCBs at the ppm-lftvel. Rappe and Buser (2) have extended this work (Table 1). Formation of PCOFs from the pyrolysis of PCBs has been accomplished In the laboratory (3.)- On February 5, 1981, a transformer fire In a state office building In Binghamton, New York demonstrated that conversion of PCBs to PCOFs and other toxic substances can occur under "real world* conditions and create a very serious problem of environmental contamination. The Binghamton state office building transformer fire was extensively studied and may be considered the model for the understanding of similar events. The transformer fire which occurred In May, 1963 In San Francisco confirmed the Binghamton finding that, under conditions of current usage as transformer dielectric fluids, PCBs can be converted to PCOFs In the event of a fire. Although It Is my Intention to cite certain data reported from the Binghamton fire In this presentation, to my knowledge very little data from this event have been published In scientific journals. Accordingly, I would refer those Interested to the Innovative and pioneering work of researchers from the Center for Laboratories and Research, New York State Department of Health, Albany, Medical Consultant, Pacific Gas and Electric Company t, Pacific Gas and Electric Company HONS -21 215062 New York 12201 for detailed information. At the time of this presentation, to my knowledge, no technical data on the San Francisco, California transformer fire have been published, although a draft document recommending clean-up levels has been released (^). I would emphasize, however, that there were very substantial quantitative and qualitative differences between the PCI-derived combustion products created by the two transformer fires. Therefore, It is Important to understand that although clean-up goals may be slmlalar where PCB-transformer fires occur, efforts required to meet these goals will vary considerably. At the time of the Binghamton fire, very little information on the toxicology of PCDFs was available. Information that was available was based primarily on experimental animal studies. However, because animal studies Indicate that the toxic effects of PCDFs and polychlorinated dlbenzodloxlns (PCDDs) are qualitatively similar, and because substantially more information is available on the toxicity of PCDDs, the Binghamton Investigators were able to make certain very useful assumptions about the toxicity of PCDFs. At the time of this presentation, we find ourselves In about the same situation. The significance of this state of affairs becomes extremily relevant to those facing the task of determining safe exposure levels to PCDFs present as residues in the air and on the surface of structures contaminated by smoke from PCDF-containlng transformer fires. To my own personal knowledge, two methods have been used for establishing standards for acceptable clean-up levels following transformer fires. One method assumes the major hazard to human health represented by the PCB-PCDF contaminant Is carcinogenesis. In this case, a comon approach is to extrapolate to low level exposure 6-22 MGNS 215063 In humans, using high dote carcinogenic bloatsay data from studies In experimental animals. This procedure calculates a dose which corresponds to a given lifetime cancer risk. A second approach, which does not consider the potential carcinogenic properties ef the contaminant mixture to be the major Issue, Is based on the establishment of an acceptable dally Intake (ADI), most often based on a no-observed effect level (NOEL) from animal study data. In connection with PCDFs, the decision on which approach to use, that Is to base the standard on carcinogenesis or on an established NOEL, Is one which does not enjoy a unanimity of opinion among those who consider the eiatter. The PCDF or PCDO for which the most toxicological data are available for risk assessment computation is 2,3,7,8-TCDD. Although this compound Is carcinogenic In laboratory animals. It does not appear to be genotoxlc. Thus, some would argue that It Is more appropriate to establish a safe standard based on the NOEL for this compound rather than on Its carcinogenic potential. According to the latest Information of which I am aware, Kim and Hawley, of the New York State Department of Health, acting with the approval of the Binghamton State Office Building Expert Advisory Panel, have selected the NOEL approach to calculation of risk assessment (5.). It is beyond the scope of this presentation to go Into detail on the Kin and Hawley calculation, however, It Is of Interest that the basis for the calculation Is the reported no-observed effect level for 2,3,7tB-TCDD of 1 nanogram/kg-day In rats reported in both a three generation reproduction study (6J and a two-year oncologic study (7J. An uncertainty factor of 500 was recosnended by the Binghamton Expert Advisory Panel and Included In the Kln-Kawley calculation. Accordingly, the acceptable dally intake (ADI) for humans was 6-23 HONS 215064 calculated to be 2 plcograms/kg-day. Assuming a 50 kg individual) the AOt would be 100 pg/day for 2>3,7,B-TC00. Further assuming a breathing volume of 10 m3 for an average 10-hour day, the calculated air guideline for 2,3,7,B-TCD0 would be 10 pg/m3 of 2,^,7 ,B-TCD0. Based on acute toxicity studies of transformer fire-generated toot from the Binghamton state office building, and adding a correction factor which takes Into consideration the ratio of 2,3,7,8-TCDF and 2.3.7.8- TCDD concentration In the soot, an acceptable air concentration guideline of B.4 pg/m of 2,3,7,8-TCOF was reached. Other assumptions Included In this final acceptable air concentration were that a worker would be exposed for 30 years, 250 days per year as a maximum possible exposure duration. Also Included was the assumption that the contaminant concentration would remain constant during the 30-year period. Kim and Hawley Included an alternate calculation of acceptable air concentration which Included the assumption that over a 30-year period, contamination levels would drop to IX of the value found on the day the building Is reoccupied. With this assumption, an acceptable air concentration of 39 pg/m3 of 2.3.7.8- TCDF was calculated. The Kim and Hawley calculation also Included provisions for Ingestlon/dermal exposure. This calculation made certain assumptions about surface area of the body of a SO kg female which might make contact with contaminated surfaces, dermal absorption, and amount of contaminant on the surface area of the hands assumed to be Ingested dally. Depending upon which set of assumptions on Ingestlon/dermal exposure Is selected, the range of acceptable surface contamination calculated was from 2.5 to 110 nanograms/m2 of 2,3,7,B-TC0F. Should a decision be made to establish an acceptable risk level based on carcinogenesis, several mathematical models are available 6-24 MQNS 215065 for performing cancer risk calculations. Mantel-Bryan in 1961 (8), defined a virtually safe risk for a lifetime as 1 x 10 . Since then, regulatory agencies have used risks in the 1 x ID*7 tc 1 x 10~5 range for setting standards for acceptable exposure to carcinogenic agents. The Carcinogen Risk Assessment Group of ERA perforated a risk assessment for 2,3,7,8-TCOO using a carcinogenic extrapolation procedure In which a 1 x 10" risk was found to correspond to a dose level of 0.00236 pg/kg-day (9) As might be expected, available statistics have been utilized to examine the risk of death in the United States from a number of causes which help put into perspective the acceptable risk standards used by regulatory agencies. Table 2 shows results of some of these risk assessment calculations. Calculation of a NOEL from Yusho data. As mentioned above, Kim and Hawley (S) calculated a no-observable effect level for 2,3,7,B-TCDF for application to human exposures based on studies In rats. An uncertainty factor of 500 was Incorporated at least in part to account for differences between rodent and man. It occurred to me that as another way of looking at the exposure-effect relationships It would be Interesting to examine the published Japanese Yusho data to determine whether information existed with which to calculate a NOEL for PCDFs based on observations in humans rather than in rodents. It Is to be recalled that "Yusho", Japanese for "oil disease" occurred as an epidemic In Southwestern Japan in 1968. The cause of this problem was traced to accidental contamination of rice oil by PC8s and PCB-derived materials, including PCDFs and polychlorinated quaterphcnyls (PCQs). Yusho was characterized by en abrupt appearance of severe dermatitis (chloracne), often confined with 6-25 HONS 215066 other skin, gastroenteric, and neurological manifestations. Kuratsune and his colleagues have received world-wide acclaim for their excellent Initial and continuing work on the clinical, epidemiologic, and toxicologic studies of Yusho (10). As It turns out, there Is sufficient information available to make such a calculation If we define "effect" as absence of skin lesions eased on clinical examination as did Japanese Yusho Investigators (11). In view of the observation that skin lesions constitute the most coemon positive clinical finding associated with overexposure to halogenated cyclic compounds such as PCBs, PCOOs and probably PCOFs- Indeed, Regglanl (1) calls chloracne "the most sensitive Indicator of overexposure to TCOO we have in the human subject"- there Is some support for this definition. However, since subcllnlcal effecs are likely to be more significant than chloracne, I offer these calculations as an Interesting exercise only. I do not suggest that a NOEL based on the absence of chloracne be considered as a human health standard for this class of compounds. Rather, I present these calculations because I found them to be of Interest. Having dismissed the Importance of these calculations as of personal Interest only, I will not dwell further on their weaknesses here. Rather, I offer the four following points which drew my attention to the Yusho Incident as the equivalent of a 135-day "feeding study" Involving humans: a) The comparison Is man to man, not rodent to man; b) PCBs, PCOOs and PCOFs share many basic toxic effects, the major differences are quantitative rather than qualitative; and HONS 215067 6-26 r r c) The clinical severity of Yusho disease was found to correlate fairly well with the total amount of oil consumed over a 1-6 month "latent" period, with the latent period defined as the period during which oil was Ingested before symptoms appeared. In sense, this corresponds to a "dose-response" relationship. Clinical severity of Yusho correlated with the total amount of oil consumed during the latent period but not with the amount of oil consumed per kg of body weight per day. Clinical severity of the Yusho response was graded from 0 to 4, with grade 0 designating a condition with physical complaints but without skin lesions, and grade 4, a condition with extensive distribution of acneform eruptions. In a table entitled "Data For The Six Yusho Patients Consuming The Smallest Oally Oose Of Oil During The Latent Period", Hayabushl, at al (11) included two patients graded 0 for severity of clinical response who Ingested a total of 314 and 353 ml contaminated rice oil, respectively, during a 135-dey latent period. Assuming, as stated by Hayabushl et al, (ll_) that the Yusho oil contained 633 ppm PCBs, 596 ppm PCQs, and 3.4 ppm PCOFs, and assuming further that all observed signs and symptoms were due to PCOFs only, none to PCS* and PCQs: kg/lOOOg 314 ml x 0.9 g/ml (Sp gravity) x 3.4 mg/kg x 1 0.96 no PCOFt or, for a latent period of 135 days, 0.96 mg/135 7.12 ug/day NOEL. 140NS 2150*8 6-27 rr Of course, this NOEL specifically refers to clinical grade 0 Yusho as defined by Japanese Investigators. However, [ believe that we can safely assume that a dose at or around this level did not result In a "epidemic" of birth defects or abortions In the Yusho population. Another way to examine this NOEL for comparison to animal-derived data Is to consider the 0.96 mg PCDF total dose over the 135-day period as a maximum lifetime allowable dose. Thus, spreading this 0.96 mg total dose over 30 years, assuming 250 days per year as the Binghamton Investigators did (5J , one divides 0.96 mg by 7600 (30 years x 260 days), for a value of 126 ng/day. (These data refer to two Japanese women, 19 and 46 years of age; It is unlikely that either weighed In excess of 50 kilograms.) In the case of TCOF, the advisory group at Binghamton argued that 50X of a NOEL dose is inhaled and 50X absorbed through skin contact (5). Therefore, for air, 50X of 128 ng 64 ng. Assuming, as Is usually done, that on Inhales 10 3 of air In an 8*hour work day, we can calculate a NOEL air concentretlon from the Yusho data to be of 6.4 ng/m3. Recall that the Klm-Hawley (5,) calculation, based on a NOEL for 2,3,7,8-TCDO In rats, applying the same basic assumptions, reached a recommended acceptable air level of 6.4 pg/*3 for 2,3,7,8-TCOF, a value approximately 1000 times lower than the calculated Yusho-based NOEL, but containing a 500-fold "uncertainty" factor. In view of the extensive list of assumptions applied to both the 1 ng/kg-day rodent NOEL for 2,3,7,8-TCDO based on chronic studies and the 0.96 mg human "clinical" NOEL for PCOFs based on 135 days of Ingestion, the calculated "acceptable" air levels are. In my opinion, suprlslngly consistent. MQNS 215069 6-28 SUMMARY When exposed to high temperatures, PC8s form PCOFs, This phenomenon has been demonstrated In the laboratory as well as In the "real-world" In connection with PCB-contalnlng transformer firms. One such fire, the February 5, 1981 Binghamton New York State Office Building fire has been well described In terms of environmental contamination by PCB-darlved materials, especially PCOFs. At least two methods have been used for establishing standards for acceptable levels of clean-up following such fires. One method assumes that the major hazard to human health represented by the PCB-PCOF contaminant mixture Is cardnogenlsis. Here, a comnon approach Is to extrapolate to low-level exposure In humans from high dose carcinogenic bioassay data from studies In experimental animals. This procedure calculates a dose which corresponds to a given life-time cancer risk. A second approach which does not consider the carcinogenic properties of the contaminant mixture to be the major Issue, Is based on the establishment of an acceptable dally Intake {ADI}, most often based on a no-observed effect level (NOEL) from animal study data. There Is no consensus among those who consider this matter on which of the two methods to use In calculating an acceptable level to PCOF follow ing a transformer fire. Based on data reported by Japanese Yusho Investigators, an alternative method of calculating a NOEL for human exposure to PCDFs based on the absence of chloracne as an effect is examined. Considering the inyrlad of assumptions Included in these acceptable exposure calculations, the Yusho-based NOEL for PCOF Is surprisingly close to the NOEL based on experimental animal studies. Nonetheless, the Yusho-based method Is not recoammnded for serious consideration because sub-clinical effects, as opposed to clinical MONS 219070 6-29 t* effects such as chloracne, may be more Important and occur at dose levels lower than those required to cause clinical dermatitis. 6-30 HONS 215071 ppm Levels of PCDFs In Connerclal PCBs Semple 3-Cl 4-CL 5--CL 6-CL 7-CL Total Aroclor 124B.1969 Aroclor 1Z54 Aroclor 1260.1969 Aroclor 1260 Aroclor 1016,1972 Clophen A 60 Clophen T 64 Pharwclor OP-6 Prodelec 3010 Mitsubishi (used) 0.10 0.06 0.10 0.41 2.13 0.5 0.25 0.1 0.30 0.001 1.4 0.30 0.7 1.08 4.00 1.2 0.70 0.4 1.0 0.001 5.0 1.73 10.0 0.35 3.30 0.3 0.B1 0.5 1.10 0.001 2.2 2.45 2.9 0.07 0.53 1.35 0.82 * 2.0 1.9 1.0 3.8 8.4 5.4 13.6 2.0 10.0 Modif'ted from Rappe 4 Buser () * FRG * Frence Japan MONS 219072 / TABLE Z Risk Assesment By Cause of Death CAUSE OF DEATH Risk of Oeath per Person Per Year Car racing Smoking ZO clg./day Car driving Industrial Air travel Orownlng ** Coffee, 1 cup/day Orlnklng 1 bottle/day ***2,3,7,8,-TC00 (0.00236 pg/kg-day) Earthquake (California) *01etary PCBs (3.3 ug/d.) Hurricanes Lightning Meteorites 1.2 x 10-3 5 x 10"3 1.7 x 10`4 Z x 10'4 1.0 x 10'5 3.4 x 10`5 4 x 10"5 7.7 x 10`5 1.0 x 10-6 1.7 x 3 x ID*7 4.0 x ID*7 7.7 x 10'7 6 x 10`11 Oa Source: no asterisk (13.) * (D ** (14) *** (9) 6-32 HONS 215073 r r REFERENCES I. Vos, J.G.. et al. Identification and Toxicological Evaluation of Chlorinated Olbenzofuran and Chlorinated Naphthalene In Two Commercial Polychlorinated Biphenyl*. Fd. CosmetT Toxicol 8: 625-633, 1970. Z. Rappe, C. and H-R Buser: Occupational Exposure to Polychlorinated Oloxlns and Olbenzofurans. ACS Symp. Ser. 149: 319-342, 1981. 3. Buser, H-R: Formation of Polychlorinated Olbenzofurans (PCOFs) and Olbenzo-p-DInxins (PCDOS) from the Pyrolysis of Chlorobenzenes. Chemosphere 6: 416-424, 1979, 4. Gravltz, N., et al. Interim Guidelines for Acceptable Exposure Levels In Office Settings Contaminated with PCB and PCB Combustion Products (Oraft). California Oepartment of Health Services, Berkeley, California, Appendix 1, p3. September 30, 1983. 5. Kim, M.K. and J. Hawley. Revised Risk Assessment Binghamton State Office Building (Draft). New York state Department of Health, Albany, N.Y.. June 7, 1983. 6. Murray, F.J., et al. Three-generation Reproduction Study of Rats Given 2.3,7,8-Tetrachlorodlbenzo-p-dloxln {TCDO) In the Diet. Toxicol. Appl. Pharmacol. 50: 241-252, 1979. 7. Koclba, R.J. et al. Results of a Two-year Chronic Toxicity and Oncogenicity Study of 2,3,7,8-Tetrachlorodlbenzo-p-dloxin In Rats. Toxicol. Appl. Pharmocol. 46: 279-303, 1978. 8. Mantel, N. and W.R. Bryan. Safety Testing of Carcinogenic Agents. J. Nat. Cancer Instlt. 455-470, 1961. 9. Kim, N.K, and J. Hawley. Risk Assessment. Binghamton State Office BulldlrtQ. New York State Department of Health. Albany. N.7,'. Rarcfi 29, 1982. 10. Kuratsune, M., et al. Epldemllolglc Study on Yusho, a Poisoning Caused by Ingestion of Rice Oil Contaminated With a Comaerclal Brand of Polychlorinated Biphenyls. Environ. Health Persp. 1: 119-128, 1972. II. Hayabuchl, H., et al. Consumption of Toxic Rice Oil by "Yusho" Patients and its Relation to the Clinical Response and Latent Period. Fd. Cosmet. Toxicol 17: 455-461, 1979,12 12. Regglanl, G. Toxicology of ?,3,7,B-Tetrachlorod1benzo-p-d1ox1n. Reg. Toxicol. Pharmacol 1: 211-243, 1981. MONS 21*07* 6-33 13. Anon. Environmental Impact Assessment of the Proposed Mateppa Gas Project. Western Research. Bow Valley Resources Ltd., November, 1982. 14. Mac Mahon, B. et al. Coffee and Cancer of the Pancress. N. Eng. J. Med. 11: 630-633, 1981. DISCUSSION QUESTION: ANSWER: Since other synthetic materials were Involved In these fires, couldn't the contamination come from something other then the liquid? It's possible, however, a large source of other materials would have to be required. 6-34 MONS 215075 THE SEARCH FOR PCB HEALTH EFFECTS Paper presented at EPRl PCD Seminar, Decamber 6-8, 1983, Atlanta, Georgia. John F. Brown, Sr, Biological Sciences Branch General Electric Research and Devalopment Center POB 8, Schenectady, New York 12301 INTRODUCTION Research on the chemical, physical, environmental, biological, and medical properties of PCBs has grown to a level sufficient to produce 1200 published scientific papers last year. Those of us who do research on PCB health effects keep gening some questions I'd like to address herei What's going on? Whafs being teamed? And how does It Impact on those national and international controversial regarding the risks ol PCB exposure? HISTORICAL BACKGROUND As most people In this audience are aware, the polychlorinated biphenyls, or PCBs, were widely used in this country for nearly fifty years, within the period 1929-1978, primarily in capacitors and fire-resistart transformers. During this period, about 1.3 billion pounds of PCB were used, end much of thet PCB is still in equipment that is In servlet today. Thus, the opportunities for human exposure remain. Mott of that PCB usage occurred after several toxicological studies at tne Harvard School of Public Health, and a fair bit of practical industrial axpariance, had Indicated that PCBs were of sufficiently low acute and chronic toxicity to permit handling without danger to the health of those involved. This longstanding perception was seemingly contradicted In 1961 by an outbreak of chlorecne and related illness in southwestern Japan. Chioracne Is a skin disease, first described in the 1190's, thet is characterized by enwtions resembling adolescent acne, often accompanied by hyperpigmentation, a general feeling of malaise, and a variety of less lrequant symptoms. It can be caused by chlorinated coal tar; by the polychlorinated naphthalenes, or Haiowaxes, that were once used in electrical cables) by certain polychlorinated dioxlnsi and by several other nature! and synthetic chemicals. The i 9(1 Japanese outbreak was traced to Ingestion ol rice oil (or "yusho") that had been contami nated with heat transfer fluid from a direct-fired heat exchanger that had originally been filled with a Japanese PCB, Kanechlor 900. By 1981, the toxic agents responsible for the yusho disease outbreak were generally believed to be 2,3,*,7,8-pentaehlorodlberuofuren 6-35 HONS 215076 r* and other polychlorinated dibenzofurans, or PCDFs, which may be formed from certain PCBs by heating m air to temperatures around or above 300 C. However, the initial Japanese reports implicated the PCBs themselves as the toxic agents, these reports appeared at a time when the PCBs were already of concern because of their environmental persistence, and were of major importance in stimulating the regulatory attack on PCBs that culminated in the 1979 ban, and the intensive search for biological effects that might have been over looked during the previous 40 years of widespread use. BIOLOGICAL FINDINGS The studies of PCB bioefleets have resulted in literally thousands of published scientific papers on PCB uptake by, persistence in, and effects on many life forms, including plants, algae, bacteria, tissue culture cells, invertebrates, fish, birds, mammals, and man. [ will not attempt to summarize this body of literature in its entirety, but 1 would like to note certain general findings that relate to the question of human health hazards. The first of these is that all PCBs are essentially water-insoluble, fat-soluble materials that are readily taken up from the food eaten or the air breathed. They can also be taken up by absorption through the skin, but the rates and importance of such processes are un known. Once inside the body, they rapidly distribute themselves, localizing In each tissue In proportion to its fat content, Thus, the route of entry Into the body has little effect on either the Internal distribution or the pharmacological effects produced. The second general finding is that the individual PCB isomers differ enormously in both persistence and biological activity. There are 209 theoretically possible ways of attaching between 1 and 10 chlorine atoms to a biphenyl nucleus, and about 100 of these possible isomers have actually been detected In the commercial PCB mixtures, generally in proportions that vary systematically with the mean degree of chlorination, or boiling range, of the parti cular PCB grade. Generally speaking, the more volatile, less highly chlorinated isomers, such as thoee that predominated In the capacitor-type PCBs, Arodors 1221, 1016, and 1242, are more easily metabolized and eliminated by mammals, and their pharmacological effects are limited to the Induction of drug-metabollzlng enzymes that include P450 cytochromes; that Is, the same enzyme group as is Induced by phenobarbltal. Conversely, the less volatile, more highly chlorinated Isomers, such as those that predominated In the transformer-type PCBs, Arodors 1254 and 1260, are usually only siowiy eliminated. Most of these "higher PCBs" also induce the P450 family of drug metabolizing enzyme*; however, a few of them are also weakly active at inducing the P44ft-type enzymes, whose induction In animals appears to be correlated with manifestations of a chloracnegenic toxic response 6-36 MONS 215077 The third general finding is that in sensitive animal species the manifestations of P44Scorrelated, or chloracne-iike, toxicity are independent of the type of chloracnegen usedi a minute dose of an active PCDF or a huge dose of PC8 may produce tha same symptoms. However, the relative sensitivities of different species to any of the ehloracnegenic agents very enormously, and with little regard for genetic relationships. Thus, chickens, guinea pigs, and rhesus monkeys are very sensitive to chloracnegens in general, while gisil, rodents, and humans are orders of magnitude less so, so that with weakly active agents, such as the commercial PCB mixtures, this type of toxic response may no longer be demonstrable. Until the biochemical bases for these large Interspecies differences are understood, it will be difficult to predict human risk from animat data on such substances. A different problem in risk extrapolation Is presented by the tumorigenicity testing data. The production of tumors by PCBs has required use of e species, such as the rat, which is pretty Insensitive so chloracnegens, and then feeding it a non-metaboiltable PCB such as Arodor 12(0 to a cumulative dose around 2.3 g/kg. Since such PCBs ere mostly retained In the rat, simple calculations show that tha PCB concentration tn body fats must rise to levels between i and 2%. At such itveis, purely physical affects, such as membrane disrup tion, wouid be anticipated. Conalstent with this is tha recent finding that the PCB loading is indeed acting upon the rat as a tumor promoter rather than a tumor initiator, What ali this means is that the usual rational* for linear dote extrapolation is lacking, and one has no basis for presuming that e low level of carcinogenesis might also occur at less extraordi nary PCB levels. In short, for tha PCBs there exist some unusual barriers to the prediction of human health hazards from tha usual torts of animal toxicological data. At the seme time, there exist tome unusual opportunities to make meaningful clinical and epidemiological observations on humans themselves! there was extensive exposure in the past) the degree of exposure can be determined by measurement of the levels of persistent PCB Isomers In the body) end there has been a sufficient lapse of time for the manifestation of any delayed health effects. RECENT CLINICAL FINDING5 Our own studies were initiated because of access to a 1?*-person group of capacitor workers that had direct exposure to Aroclors i254, 1242, and lOlfi liquids and vapors over the pwriod 1946-1977. We've had these people under medical surveillance since 1976) have collected enormous volumes of clinical data on them) and have run literally thousands of regressions on our computer In order to identify the statistical associations between their PCB levels 6-37 MONS 215076 and clinical measurement*. In addition, we've been working closely with federal, state, and academic researchers who've been running other studies on the GE capacitor worker population, and also keeping in touch with investigators looking at other PCB-expoied groups around the country. At present there is substantial agreement among scientists working in this field as to what happens to PC ^-exposed people, and 1 shall attempt to summarize that consensus in my remarks. The first thing we've all learned is that considerable quantities of PCB can be picked up, particularly by people like capecltor workers who were exposed to the more volatile PCBs. In our own study population, we estimated the mean minimal uptake up through early 1976 as I5g, and there were undoubtedly tome members of the group who absorbed severtl times that much, That 15 g mean minimal uptake, incidentally, represents 20 timet the mean PCB uptake ingested by the Japanese yusho victims. Nevertheless, neither in our study population, nor in any of the others reported in the recent literature, have there been any confirmed findings of chioracne or other PCB-reiated disease In workers exposed to ordinary, unpyroiysed, American-made PCBs. This probably should come as no surprise. Experience with other chemicals Indicates that a chioracne outbreak In a working group Is a pretty unmlstakeable event, and If there'd been any tendency for ordinary PCB to produce ehloracne, thet fact should have been well known by the rmd-1930*s, at the latest. The second thing that I believe most investigators in the field would agree on, although the observational evidence it mostly Indirect, Is that a weak Induction of drug-metaboiizing enzymes of the P450 family may occur. In workers with high levels of the lower PCBs, one can see some PCB-correlated elevations in the level of serum gamma glutamyl transpepti dase (GGTP); some decrease In antlpyrlne clearance times, implying increases in microsomal oxidases) and some decrease also in bilirubin levels. Implying Increases In microsomal glucuronyl transferase. The Implied changes in microsomal enzyme levels were, however, smaller than those seen In comparably dosed rats; the affected clinical parameters all remained within the normal ranges of clinical variation, and the apparent effects all disappeared after discontinuance of direct exposure. A very recant finding is that despite earlier reports to the contrary, PCBs in humans do not cause elevations of the serum lipids, that Is, the serum triglycerides and cholesterol, nor of the associated serum enzymes, alkaline phosphatase and serum glutamate pyruvate tranipeptldase, all of which had been interpreted at indicators of deranged liver function. The earlier reports had been based ipon observation of statistical associations between tha serum levels of these substances and those of the PCBs. It is new apparent that that particular statistical association it perfectly real, but that it hat a trivial explanation! because of the tendency of PCBs to distribute between body tissues and fluids In proportion 6-38 MONS 215079 r r to their lipid content, any elevation in serum lipid, from whatever cause, wiii automatically result In PCB redistribution, and hence an Increase in serum PCB. Two other research groups have now confirmed our discovery that when one avoids this pitfall by comparing the serum ilpld levels with those of the PCB* in body ilpids, no correlations are seen. Finally, the mortality studies that have been performed by NIOSH on tha entire exposed GE capacitor worker population, where the PCB exposures want beck to iM^pnd on a similar group at Aerovox, where the exposures went back to 1938, have drown no increases in standard mortality ratios for deaths due to cancer, cardiovascular disease, nervous system diseases, nor for alt causes combined, nor any significant Increases In any individual type of cancer. Nevertheless, It must be pointed out that the total numbers of deaths Involved in the mortal ity studies were still small enough so that a modest increase in a rare type of fatal disease could not have been detected. Similarly, tha total populations involved In all clinical studies to date may not have been large enough for the reproducible obaervatlon of a rare type health effect. In summary, although scientific studies can never exclude the possibility of unobservable phenomena. It does seem established that PCB exposures at the high levels provided by prolonged, direct occupational contact had no reproduclbly observed effects upon tha health of the vast majority of the exposed Individuals. These individuals had exposures that were approximately a thousand-fold higher than those provided by the kinds of PCB-contaminated environments being encountered today. Accordingly, there seems little basis for concern over the health risks presented by environmental exposure to residual PCBs at today's levels. 6-39 HOMS 219080 r r CHEMISTRY OF PCDDs/PCDFs J. Rodney Harsh and Jasenka Zbozlnek, Ph.D. SCS Engineers Since 1979. research has shown a potentil concern resulting from the presence of some chlorinated aromatic hydrocarbons associated with PCBs. predominantly polychlorlnated dl benzodloxlns (PCDDs) and polychlorinated dlbenzofurans (PCOFs). Traces of similar compounds, such as polychlorinated blphenylenes (PCBPs), poly chlorinated pyrenes (PCPYs), and polychlorinated chrysenes (PCCYs), have also been reported In some samples. These compounds occur In extremely low concentrations In PCB liquids. However, some POOD, PCDF, and PCBP Isomers are acutely toxic as deter mined by laboratory tests on animals and demonstrated in industrial incidents. Toxlcitles of some of these isomers are orders of magnitude higher than PCBs. PCDD and PCDF compounds have been found in the original PCB fluids manufactured In the United States, Europe, and Japan. Several PCB accidents In which PCDDs and PCDFs were prominent have also been documented, Including the Yusho Incident In japan fnvolvfng ingest ton of PCDF-eontaml anted rice oil, end e similar subsequent Incident fn Taiwan, more recently, a fire Iwolvlng electrical equipment containing PCB fluids In a Binghamton, Ntw York, office building produced soot which was found to contain PCDFs and PCDDs. These flndfngs indicate the possibility that PCDDs, PCDFs, end other related chlorinated aromatic compounds can form from askarel fluids at high temperatures. Polychlorinated Qlbenzo-p-dloxlns (PCDDsI There are a limited number of direct precursors and reactions by which PCDDs can be formed, in order to be a dioxin precursor, a compound must meet two conditions: The compound must be an ortho-substituted benzene ring In which one of the substituents Includes an oxygen atom directly attached to the ring. e It must be possible for the two substituents, excluding the oxygen, to react with each other to form an Independent compound. An example of the Ideal dioxin precursor is shown fn Figure 1. A reaction between two of these molecules could yield a dioxin and z XY molecules, as shown In Figure Z. 6-41 HONS Z15081 Although the number of direct dfoxln precursors Is relatively smell because of the conditions set forth above, there are many indirect precursors, l.e., compounds which do not fit the model, but which can degrade or be transformed into direct pre. cursors. This will be discussed in more detail later. Buser (_l) and Rappe, et al. (}, have Identified the following basic reaction for chlorinated dloxfn formation: e Dimerization of chlorophenols and phenoxy adds (pyrolytic and pho tochemical) - A bfmolecular reaction In which the relatfve ratio of isomers formed depends on the concentration of the chiorophanates; yields decrease fn more dilute systems, such as contaminated mineral oil s. e Cyclliatfon of predioxins (pyrolytic and photochemical) - Unlmolecular, and thus concentration-independent. e Dechlorination of higher chlorinated PCDDs Unlmolecular, and thus cone en t rat 1on-1nde penden t. e Direct chlorination of d1benzo--dloxfns. The dimerization reaction basically follows a two-step mechenlsm (Figure 3). The Intermediate diphenyl ether compound Is called a predioxin. Typically, a predioxfn will either cyellze to dioxin or break down to the precursors. For example, fn one test, when lrgesan DP300, a compound with a predioxin configuration (Figure 4) was heated to 980*C, the reaction yielded both dioxins and precursors (3). Because of such competing reactions, dioxins are seldom found in more than trace quantities. It has been shown that the higher chlorinated pradloxlns are amenable to self condensation, while the lower chlorinated predfoxlns apparently do not readily cycllze further (}. For instance. In one test, frradlatfon of 2,4-dlchlorophenol yielded a predioxin, but no further reaction (}. Temperatures for dioxin formation by dlmerlzatfon/eyclIzatlon have been reported at values from 18D* to 620*C. Dioxin formation from pentaehlorophenols has generally been reported in the 300* to 40D*C range, whereas using trl- and tetrachlorophenols as precursors requires temperatures from 400* to 600*C However, the formation of dfoxfns from chlorophenols is exothermic, although there are no published data on the amount of heat released (6, .7). This faeti combined with the need for heating to Initfate the reaction, suggests that one or more of the reaction steps In the dioxin formation sequence has a high activation energy. The 6-42 HONS 215082 r r high activation energy contributes directly to the low yields. In addition, the boiling points of many dioxin precursors ere lower then the reaction temperature range. Thus, there Is a loss of precursor molecules before reactions can occur (jt). Host laboratory syntheses of dioxins rely on closed reactfon vessels or pressure to keep precursors In the liquid state. Another major source of penta- and tetra-CDD Isomers Is photochemical dechlorination of higher isomers, mainly OCDD (). The principal toxic Isomers, 1,2,3,7,8-pentachloro and 2.3,7,8-tetrachloro, are not as readily formed as other penta- and tetrachloro Isomers. The major hepta-COD formed Is the 1,2,3,4,6,7,9-substltuted Isomer, Indicating a preferential loss of lateral (2,3.7*, or 8*} chlorine atoms (10. 11). The major hexa-CDD formed is either 1,2,4,6,7,9* or 1.2,4,6,8,9-hexa-COD (10). The major penta-COO Is expected to be the 1,2,4,6,9-substltuted Isomer, and the major tetra-CDO is 1.4,6,9-tetra-CDD (ID). These reaction products show a preferential removal of the lateral chlorine atoms. Lower Isomers (dl* and trlchloro) may be formed by irradiation, but they decompose via the same route, and thus do not accumulata (L2). Pyrolysis of chlorobenzenes at approximately 620*C in the presence of air also yields PCDOs. Since chlorobenzenes do not fit the conditions set down earlier for dloxfn precursors, Buser (_1, _13) has postulated that the alkaline hydrolysis of chlorobenzene to PCOD Involves a two-step condensation process from ortho-chlori nated phenoxy anions via a polychlorinated 2-phenoxy phenate (Figure 5). This dimerization is exothermfc. Below approximately 18D*C, usually very minor amounts of PCDOs are formed (<1 ppm), but significant quantities (1 to 10 percent yields) result at temperatures over 1B0*C (1_). Dioxins have also been detected In residues from a variety of combustion processes. Including municipal Incineration, the burning of fossil fuels, wood In wood stoves, and cigarettes (14). Precursors of dibenzo-*dfox1ns are also known to be formed after sfmply adding hydrogen chloride to the flame of a Bunsen burner (14). Combus tion research has been centered on municipal incinerators and Industrial heating plants. PC00 Isomers have been detected In municipal Incineration fly ash In Swit zerland, Italy, and the Netherlands in concentrations war 200D ng/g (ppb). The major Isomers detected In Switzerland were the same as those formed In the pyrolytic dimerization of 2,4-df-, 2,4,6-trl-, 2,3,4,6-tetra*, and pentaehlorophenates. HOMS 21S083 6-43 ** The mechanlsm(s) of PCDO formation in fly ash Is presently unknown. Two suggested sources are: e Condensation of chloropnenols. e Thermal synthesis from any organic materials and fnorganlc chloride (the "trace chemistries of fire" hypothesis) (15). __ Even ideal PCOO precursors, such as polychlorophenols, show very low PCDO conversion efficiencies. Thus, although every chlorinated organfe compound or organic cam* pound/inorganic chlorine mix might be regarded as a potential precursor, the range of likely precursors for formation of detectable quantities of PCODs is probably not that broad. It seems more likely that those compounds which can be pyrolyted to benzene, chlorobenzene, phenols, catechols, and/or Inorganic chlorine are more prob able Indirect precursors. The mechanlsm(s) by which PCDOs are formed during combus tion Is still not well understood; thus, the effects of combustion temperature, fuel characteristics, catalysis, and other parameters have not been determined. There has been some question regarding the role of PCBs In PCDD formation. In 1981, an electrical fire In the State Office Building In Binghamton, New York, resulted In the pyrolysis of PCS fluid from a transformer (l, T7, 18_), Soot from this fire contained PCDFs and PCODs, PCBs do not fit the precursor model discussed earlier, and PCODs would not be expected to form readily In a mixture of biphenyl compounds (19). Conceivably, PCBs could be oxfdfzed to polychlorinated hydroxy biphenyl s which could, in turn, cycllze to form dioxins (20). Another reaction that has been sug gested is degradation to a chlorinated benzene, followed by oxidation to a chlorophenol and dimerization to PCDD (21). However, neither mechanism quite fits the thermal decomposition pattern for PCBs. The PCB flufd at Binghamton was a mixture of 65 percent Aroctor 1254 and 35 percent chlorobenzenes, ftappe, et al. (22), have suggested that pyrolysis of the chloroben zenes is a more likely source of the PCODs than PCBs. Polychlorinated Dlbenzofurans (PCDFs1 while PCDFs and PCDOs can share some of the same precursors, PCBs are the most com mon source of PCDFs. Buser and Rappe (23) have Identified four basic mechanisms of PCDF formation from PCBs (Figure 6). 6-44 HOMS 18084 r r Reaction l Involves the loss of the ortho-Cl. Reaction 2 Is a loss of HC1 involving a 2,3-chlorlne shift. Reaction 3 has a loss of ortho-HCl. Reaction 4 involves a loss of ortho-H. PCDFs have been Identified as trace contaminants at the ppm level Iru number of Amerfean, European, and Japanese commercial PCB mixtures (24. 25. 26. 27). Concen trations and isomer ratios vary with the type of PCB and the country of origin, pro bably as a result of differences In manufacturing conditions (Table 1). For in stance, Bowes, et al, (27). detected PCDFs in American-manufactured Aroclor 1260 and 124B at total concentrations of 0.6 and 2.0 ppm, respectively; none were detected in a single sample of Aroclor 1016. The total quantities of PCDFs in German Cl ophen A60 and French Phenoclor DP-6 were estimated to be 6.4 and 13.6 ppm, respectively (26). PCDFs have also been detected In several Japanese-made PCBs at concentrations from 1 to 20 ppm, with the highest concentration in a sample of Kanechlor KC-400 (32. 33). PCDFs may also form during PCB use, The best known example of this Is the "Yusho" Incident In Japan, The PCBs, which had leaked from a heat exchanger, had PCDF/PCB ratios 200 to 250 times higher than that of the original Kanechlor KC-40D (32, _30). Furthermore, the PCDFs fn the Yusho oil contained the same Isomers as a sample of used PCBs (34). This suggested that most of the PCDFs had been formed fn service (6). Similar PCDF Isomer ratios have been detected In other used Japanese PCBs (!> Other sources of PCDF Include condensation of chlorophenates and chlorobenzenes, photochemical degradation of higher chlorinated Isomers, cycllzatlon of polychlori nated diphenyl ethers and polychlorobenzenes, and direct chlorination of dfbenzofurns (11). Chiorophenols and chlorobenzenes can form PCDFs via an Intermediate poly chlorinated diphenyl ether (PCDPE) {. ) (Figure 7). PCDPEs usually occur as contaminants In chlorophenols, and are known to form PCDFs by thermal, photochemical, or metal-catalyzed reaction (35), PCDPEs and PCDFs prob ably form early In the reaction when large concentrations of chlorobenzenes are still present, wnereas PCDDt are formed toward the end of the reaction (35). Fires in electrical equipment containing PCBs have resulted In PCDF formation. Jansson and Sundstrom (36) Investigated the aftermath of a 1978 capacftor fire in . 6-45 HONS 213065 Sweden. Where** the original PCB contemned 1.1 to 1.3 ug PCDF/g (amounts of mono-, dl-, trf-, *nd tetre- (saner* about equal), PCB sample* from Inside the exploded capacitors avereged 81 ug PCDF/g PCB (approximately one-half of PCDF dl- Isomer, one-third mono-, and the rest trl-, with a trace of tetra-), and samples from the inside surfaces contained 45 to 1D7 ug PCDF/g PCB (predominantly dl- and mono- Iso mers with some trl- and a trace of tetra-). Soot from a 1977 transformer fire In Toronto contained 5 ug PCDF/g, while the original PCB contained D.D5 ug PCDF/g (36). Similar results have been obtafned in the aftermath of other PCB fires. PCDFs have also been detected In fly ash samples from municipal incinerators and In dustrial heating facilities, incinerator fly ash from the Netherlands averaged 1309 ng PCDF/g (37). Incinerator fly ash from Italy averaged about 1300 ppb of combined PCDD and PCDF (36), Fly ash from an Industrial heating facility and municipal Incfnerator In Switzerland contained 0.3 and 0,1 ppm PCOF, respectively (Z5j. Thus far, the exact source of the PCDFs In fly ash fs unknown. TRANSFORHEft/CAPAC1 TOR CHEMISTRY Most PCB-f11 led electrical equipment Is designed to operate at temperatures of 100s to I30SC, and transformer "hot spots" are typically no greater than 135*C (31), All units may contain air, but most newer ones have an Inert atmosphere, usually nitro gen, Askarel units are sealed, but some seals can deteriorate with time, permitting the entry of air and water. Hydrochloric acid can be generated, leadfng to fnternal corrosion, water can cause arcing. Internal transformer malfunctions can cause bushings and winding connections to become white hot. During these occasional short-circuit episodes, temperatures reach 200* to 250*C for a few minutes. In cases of an arc where the transformer falls, maximum temperatures may reach 1000*0; these failures occur within fractions of a second (31). Capecltors typically oper ate below 100*0. In those capacitor failures resulting In rupture, heating and In ternal arcing may have occurred prior to rupture. Samples of used askarels frcm capacitors and transformers have been analyzed for PCDF with mixed results. One researcher concluded that concentrations will increase with time in service, although concentrations detected never exceeded 5 ppm. Other tests on 40-year-old askarels showed PCDF concentrations comparable to those In un used askarels. In general, PCB conversion to PCDF or chlorobenzene conversion to PCDD end PCOF might be possible In a malfunctioning transformer or capacitor, pro vided that there is sufficient time at optimum temperature and/or oxygen conditions 6-46 HOMS 215086 r r for their formation. Ordinary working conditions do not appear conducive to PCDD/ PCDF formation (31). Finally, It should also be noted that most of the research on PCB/askerel cowers ion to PCDD, PCDF, and other chlorinated species has been performed using pure PCds or askarel formulations containing at least 50 percent PCBs. Very little research has been conducted on dilute PCB systems, such as contaminated mineral oils. Conse quently, there is little data available on the effects of dilution on PCB chemistry. REFERENCES 1. H. R. Buser. Polychlorinated Dlbenzo-p-dloxlns and Dlbenzofurans: Formation. Occurrence and Analysis of Environmentally Hazardous Compounds'! Ph.o. Thesis, University of Umee. Umea, Sweden, 1978. 2. C. Rappe et al. Formation of Polychlorinated Qlbanzo-p-dloxins (PCODs) and pi benzofurans (PCOFs) by Burning or Heating Chi orophenates. Chemosphere. 7:2692B1, 1978. 3. M. P. Esposito et al. Ploxlns. Vol. 1: Sources. Exposure, Transport and Con, trol. EPA-600/2-80-158. lnOustrlal Environmental Research laboratory, Cincin. natT, June I960. 4. G. Jans son, G. Sundstrom, and B. Ahllng. Formation of Polychlorinated Dlbenzop-dloxlns During Combustion of Chlorophenoi Formulations, sei. Total Environ.. W:W17. IMS.--------- -------- ---------------------- 5. M. p. Esposito, T. 0. Tlernan, and F, E. Dryden. 01 oxlns. EPA-600/2-80-197. Industrial Environmental Rasaarch Laboratory, Cincinnati, November 19BO. 6. H. G. Langer, T. P. Brady, and p. R, Briggs. Formation of Qlbenzodloxlns and Other Condensation Products from Chlorinated Phenols and Derivatives. Environ. Health perspec., Experimental Issue 5:3-7, 1973 7, J. S. Stanley et al. Sampling and Analysis Protocol for Assessing Organic Emissions from Stationary Combustion sources in Exposure Evaluation Envision Combustion Studies. EPA-56Q/5BZ-oi4.1iriowost Research institute. Kansas City. Missouri, January 1962. 8. E. J. Duckett. Plant Emissions: Dioxins In Perspective: Knowns. Unknowns. Resolving the Issues. Solid Wastes Manage.. 24(M:Y6-57. 8B-B9. l9fel. 9. C. Rappe, H. R. Buser, and H. P. Bosshardt. Dioxins, Dlbenzofurans, and Other Polyhal oaenated Aromatics: Production. Use. Formation, and Destruction-! Ann. N.Y. Acad. Set., 320:1-18, 1975. ID. H. R, Buser and C. Rapp*. Identification of Substitution Patterns in Poly chlorinated Qlbanzo-p-dl oxlns (PCDDS) by Hass spectrometry. Chemosphere, Z: 199-211, 1978. HONS 21508? 6-47 rr 11. H. R. Buser. Preparation of Qualitative Standard Mixtures of Polychlorinated 01benzo-p-d1oxlns and Dlbenzofurans by Ultraviolet and Gamma Irradiation of the octacnioro compounTT j."thromatogr., 192:303-307, 1976. 12. D. G. Crosby, K, w. Mollanen, and A. S. Hong. Environmental Generation and Dearadatlon of Dlbenzodloxl ns and Olbenzofurans. Environ. Heal tfi Perspec.. Experimental Issue 5:25S-6, 1973.--------------------- _ 13. H. ft, Buser. Formation of Polychlorinated Olbenzofurans(PCOFs) and Dlbenzo-odloxlns (PCDDs) from the Pyrolysis ol Chlorobenzenes. ttiamospheTe. 4:415-424. wl.--------- ---------------------------------------------------------- 14. National Research Council of Canada. Polychlorinated 01benzo-o-dlox Ins: Cri teria for Their Effects on Nan and His Environment. Ottawa. 1981. 15. R. R. 8umb et a?, Trace Chemistries of Fire: A Source of Chlorinated Dioxins. Science, 210(446B):385-390, October 24, I960. ~ 16. G. A. Eadon. Particulate Contamination In the Binghamton State Office Build. 1 ng. Paper Presented at PCB Seminar, Dallas, Texas, December 1981. 17. Cleanlno Operations at the Binghamton state Office Building Suspended. New York State Press Release, February 26, 1981. _ 1 IB. E. J. Dionne. Fate of Binghamton's Sealed Tower Still Uncertain. New york Times, August 10, 1981. " 19. P. W. O'Keefe. Trace Contaminants In a Polybromlnated Biphenyl Fire Retardant and a Search for These Compounds In Environmental Samples. Bull. Environ. Con- tarn. Toxicol. ,22'.456-Ai5, 19V9. 2D. J. u. A, Lustenhouwer, k. Dlle, and D. Huntzlnger. Chlorinated Dlbenzo-odloxlns and Related Compounds In Incinerator Effluents^ diamosphere. 9:501 521." IW.------------------------------------------------------------------------- 21. ft. Huettar and M. Philippi, Studies on Microbial Metabolism of TCDD Under Lab oratory Conditions. In: Chlorinated oioxtns and Related Compounds, Impact on the Environment. IT Huntzlnger et al,, eds. Pergamon Press, Elmsford, New York, 1982. pp. 87*93. 22. C. Rappe et al. Polychlorinated Dioxins, Olbenzofurans and Other Polynuclear Aromatics Formed Hurl no incineration and pcB Fires. 'Preprint extended Abstract." Presented Before^ne invision or Environmental Chemistry, American Chemical Society, Kansas City, Missouri, September 1982. 23. H. ft. Buser and C. Rappe. Formation of Polychlorinated Olbenzofurans (PCOFs) from the pyrolysis of Individual PCB isomers, ctamosphere, 8:157-174. 1979. 24. G. W. Bowes et al. Gas Chromatograph1c Characteristics of Authentic Chlori nated Dlbenzofurans: identification of Two Isomers in American and Japanese' Poiy_ch 1 orinatmTbiphenyls. J. Agrlc. Food Cfiem., 23:l222-l!M, 1975. r 25. H. ft. Buser et al. identification of Polychlorinated Dlbenzofuran isomers in Fly Ash and PCB Pyrotyses. Cfiemosphere, 7:419-429, 1978. * 26. 1. C. T. Nlsbet. Criteria Document for PCBs. EPA-440/9-76-021. Massachusetts Audubon Society, Lincoln, July 1976. - 6-48 HONS 215088 27. G. w. Bowes et *1. Identification of Chlorinated Dlbenzofurans in American Polychlorinated Biphenyls, nature, 256:3Qb-3Q7, 19^. 28. P. W. Albro and C. E. Parker. Comparison of the Compositions of Aroclor 1242 and Aroclor 1016. J. Chromatogr., 169:161-166, 1979. 29. M, Horlta et al. Detailed Examination of Polychlorinated Dlbenzofurans in pcb Preparations and Kanemi Yusho Oil. Bull. Environ. Contam. ToxIcoT.. lB:67-73. ------------------------------------------ 30. C. Rappe and H. R. Buser. Chemical Properties and Analytical Hethods. In: Halogenated Biphenyls, Terph'enyis, Naphthalenes, Dlbenzodloxins and Related products. R, D. Kimbrough, ed. Elsevier/North Holland Biomedical Press. Amsterdam, I960, pp. 41-76. 31. National Electrical Manufacturers Association. Reply Comments on Notice of Proposed Rulemaking Concerning Polychlorinated Biphenyls ipc^s): Use in Elec trical Equipment. Washington, O.C.. June 1982. 32. M. Kuratsune, v, Masuda, and J, Nagayama. Some of the Recent Findings Concern ing Yusho. in: Conference Proceedings: National conference on Polychlori nated Biphenyls (November 1975, Chicago, Illinois) EPA-560/6-75-004, March 1976. PP 14-29. 33. J. Nagayama, M. Kuratsune. and r. Masuda. Determination of Chlorinated Dlbenzofurans In Kaneehlors and "Yusho Oil." Bull. Environ. Contam. Toxicol.. 15:9-13, 1575.-------------------------------------------- 34. H. R. Buser, C. Rappe, and A. Gara. Polychlorinated Dlbenzofurans (PCDFs) Found In Yusho Oil and In used JaoaneTe pcb. chemosphere. 7:439-149. 1978. 35. R. Lindahl, C. Rappe, and H. R. Buser. Formation of Polychlorinated Dlbenzo furans (PCQFs) and Polychlorinated Dlbenzo^p-dioxins (frEDDsrfran the Pyrolysis of Polychlorinated frfphenyl Ethers, ihemosphere, 9:351-361, 1980. 36. B. Jansson and G. Sundstrom. Formation of Polychlorinated Dlbenzofurans (PCDF) During a Fire Accident In Capacitors Containing Polychlorinated Biphenyls (pCbi. in*. Chlorinated Dioxins and Belated Compounds, impact on the Environ ment. 0. Hutzinger et al., eds. Pergamon Press, Elmsford, New York, 19B2. ' pp. 201-207. 37. K.-Olte, J. N. A. Lustenhouwer, and 0. Hutzinger, Polychlorinated plbenzo-pdloxlns and Related Compounds In Incinerator Effluents. In: Chlorinated Dioxins and Related Compounds, impact on the Environment. 0, Hutzinger et al eds. Pergamon Press, Elmsford, New York, 1962. pp. 227-243. 38, A. Llbertl and D. Sracco. Formation of Polychlorodlbenzodloxlns and Polyehlorodlbonzofurans In Urban*Tneinerotor ^missions. In: Chlorlnate^pfoxlns and Related Compounds, impact on the Environment. 0. Hutzinger et al., eds. Pergamon Press, Elmsford, New York, 1982, pp. 245-251. DISCUSSION QUESTION: How about furans generated In household cooklngT ANSWER: FOA has gone through risk assessment. Risk Is negligible. 6-49 MOWS 215089 LEVELS (ug/g) OF PCOFs IN COMMERCIAL PCBi Simple Aroclor 1248 {1969) Aroclor 1242 Aroclor 1242 Aroclor 1243 Aroclor 1254 {1969) Aroclor 1254 (1970) Aroclor 1254 Aroclor 1254 {KK 602) Aroclor 1254 Aroclor 1260 Aroclor 1260 (1969) Aroclor 1260 Aroclor 1260 {AK 3) Aroclor 1016 (1972) Clophen A6Q Clophen T64 Phenoclor DP-6 prodelec 3010 Kanechlor 400 Mitsubishi (used) Trl- . * * 0.1 * - 0.06 - 0.1 - 0.41 2.13 Tetra- Penta- Hexa- Hepta- Total 0.5 0.07 2.3 0.25 0.1 0.2 0.02 0.05 0.1 0.3 0.1 0.8 0.2 N.O. 1.4 0.3 0.7 1.08** - 4.00 1.2 0.03 2.2 0.7 0.2 0.4 0.2 0.1 3.6 1.0 0.4 0.9 0.3 N.O. 5.0 1.73 10.0 0.35 * 3.30 0.3 0.003 N.D.* 0.81 1.4 0.9 0.4-0.6 0.02 1.9 1.10 0.5 0.5 0.3 N.D. 2.2 2.45 2.9 0.07 m 0.53 2^0 . . 1.35 _ 0.15 4.5 1.9 1.7 1.5 0.8 0.2 5.6 3.8 1.0 2.2 0.8 _ 8.4 0.82 5.4 . 13.6 2.0 ca. 20.0 1D.0 Reference 27 28 29 30 27 27 31 30 29 30 27 29 27 27 27 30 27 30 31 30 * N.O. Non* Detected. ** Major 1 sneer 2,3,7,8-tetra-CDF. MONS 215090 Flgurt 1. Idctl Dioxin Precursor Molacult 6-51 MONS 215091 Ftgur 2. Slnplt OlmrUatton Rotctton to Form Oloxtn 6-52 MOHS 215092 6-53 hons 21M93 ia*It Cl MONS 21M94 ss- , Figure 5. Fomttlon of Dioxin Iron Chlorinated Phono*/ Anions MONS 215095 0v01***l Flgurt 6, Ittlc PCOF Fonutlon Ructions (run 2) (r*n 3) (rxn 4) HONS 215096 * Cl / \/Vci a v@-r C-- K> JSCf^. Cl ^ei ci^\/s<i ci NaCI + JMtCI HjO Figure 7. PCOF Formation \ MONS 215097 r r AN UPDATE. ON ANALYTICAL METHODS FOR POLYCHLOR1NATEO BIPHENYLS (PCS), POLYCHLORINATED D1BENZOFURANS (PCDF), AND POLYCHLORINATED OlBENZO-p-OlOXINS (PCOO) Mi reus Cooke, Fred L. OeKoos, Judith E. Gebhirt Laurence E. SI Ivon and Peter J. Mondron BattelTe's Columbus Laboratories DISCUSSION Evidence supporting the widespread distribution of polychlorinated biphenyls (PCB) In the environment has been documented as early as 1966, when PCB were found to Interfere with the environmental determination of pesticide residues. Due to the persistence and widespread geographic distribution of PCB, animal and human toxl cological studies were performed which subsequently led to the regulation of these compounds by the Toxic Substances Control Act In 1976, Until regulation was enacted, commercial mixtures of PCB, named Aroclors, were widely used In the electric power Industry as dielectric fluids In capacitors and transformers. During fires Involving transformers. Incomplete combustion of Aroclor containing oil produced chlorinated furan and dioxin compounds.U) Polychlorinated dlbenzo-p-dloxlns (PCOO) and polychlorinated dlbenzofurans (PCDF) are relatively stable under normal environmental conditions and tend to concentrate In biological matrices due to their lipophilic nature. Although limited data are available which demonstrate the effects of PCOO and PCDF on humans, significant data have shown the toxic, carcinogenic, and mutagenic effects of these compounds in mammalian systems. Toxlcltles for the specific Isomers vary by a factor of approxi mately 10,000. The chemical structures of two of the most toxic Isomers In this class of compounds, having four chlorine atoms In the 2, 3, 7 and 8 positions on the furan and dioxin molecular skeletons, are shown in Figures 1 and 2. Figure 1. 2,3,7,8-Tetrachlorodlbenzofuran [2,3,7,8-TCDF] 6-59 MONS 215098 Figure 2. 2,3,7,8-Tetrachlorodlbenzo-p-dloxln [2,3,7,8-TCDD] As t result of the persistent end hazardous nature of PCS, PCDF, and PCOD, considerable emphasis has been placed on controlling the distribution of these compounds Into the environment. A key Issue In the control of PCB, PCDF, and PCOD Is the analytical methodology. Although PCS have historically been quantified as commercial Aroclors, analysis of these compounds Is complicated when mixtures of Aroclors or compositional alterations through weathering occur. Accurate PCDF and P*CDD determinations are hindered by the large number of positional isomers that may exist. In total, there ere 21D PCDD and PCDF positional Isomers as shown In Table I. ' Table 1 PCDD ANO PCDF POSITIONAL [SOMERS Chlorine Substitution NonoD1TrlTetraPenta hexaHepteOcts- Number of PCDD Positional Isomers 2 10 14 22 14 10 2 I Number of PCDF Positional Isomers 4 16 28 38 28 16 4 1 Oesplte the difficulties In PCB, PCDD, and PCDF analysis presented by matrix complexity and Isomer configurations, reliable analytical methods are available for determination of these compounds. MONS 215099 6-60 ww Various screening methods have been reported for PCS. For quantification, the method of choice continues to be gas chromatography coupled to an electron capture (EC) detector. This technique, however. Is subject to Interferences from other chlorinated species which have the potential of Indicating elevated levels of PCB. As a result, most samples are subjected to some form of cleanup to remove Inter ferences prior to analysis. The cleanup procedure usually Involves liquid column chromatography using silica gel, alumina, or Florlsll. Complex environmental samples, however, may require a more extensive cleanup prior to analysis. Host PCB analyses are performed using packed column gas chromatography which is adequate for many applications. Recent advances In column technology making fused silica capillary columns widely available, have made the separation of complex mix tures easier for the average laboratory. The higher resolution of capillary columns and the relative convenience of fused silica capillary columns compared to glass capillary columns has led to Increased use of capillary columns In PCB analysis. Separation of an Aroclor standard using a packed GC column Is shown In Figure 3. The same Aroclor separated on a capillary GC column Is shown In Figure 4. 6-61 HONS 216100 rr Figure 4. Capillary Column SC Chromatogram of Aroclor Standard Emphasis In PCB analysis Is now shifting toward electron Impact gas chromatography/ mass spactromatry (GC/HS) and. In particular, capillary column GC/MS as tha tech* nlqua to provide the most selective and unemblgous determination of PCS, regardless of whether they are of Incidental origin or coenerclal Aroclors. Earlier efforts to quantify PCB using electron Impact GS/KS generally resulted In Identification of specific Aroclor patterns and quantification of these commercial products. A composite mess chromatogram resulting from the GC/KS analysis of an Aroclor 1254 standard Is shown In Figure 5, The extracted ion current profiles represent the molecular Ion of PCB having one to eight chlorines. A similar GC/MS analysis of a sediment sample containing a mixture of Aroclors as well as other chlorinated species Is shown in Figure 6. The elution pattern of Isomers having 5. 6, and 7 chlorines (m/z 326, 360, 394) In the unknown and standard clearly establishes Aroclor 1254 as a major constituent In the sediment sample, and provides a mechanism for quantification that Is relatively free of Inteferences. 6-62 MONS 215101 Figure 5. Mass Chromatogram of Aroclor 1254 Standard 01.U1, i .J - ."is "1 a.(i w '-ML a hat.Al jla xLL -wJLmXj. ."is * mu*. ,"!:S ' Iino-:rlna: i.i< Figure 6. Mass Chromatogram of Sediment Sample For PC00 and PCDF analyses, extremely high selectivity and sensitivity are needed to unambiguously determine these compounds at ultra-trace levels In environmental samples. These characteristics are achieved by high resolution gas chromatography/ high resolution mass spectrometry (HRGC/HRMS). A chromatogram from an HRGC/HRMS analysis of a ter liter water sample spiked with two TCOO Isomers Is shown in Figure 7. In this chromatogram, specific Isomer determination Is made for 5 parts-perquadrillion (W/W) 1,2,3,6.8-TCDO and 8 parts-per-quadrll 11on (W/W) 2,3,7,8-TCDO. This represents an analytical sensitivity of approximately 8 x 10*12 grams. 6-63 HONS 215102 irtus vatim uHixtit Figure 7. HRGC/HRMS Chromatogram From Spiked Water Analysis The high selectivity of HRGC/HRNS allows the analyst to distinguish close eluting Interferences as shown In the chronatogrem given In Figure 8. This chromatogram was produced from a spiked fish tissue sample which was extracted, extensively partitioned by column chromatography, and then analyzed by HRGC/HRNS for 1,3,6,8-TCDD, 1,3,7,9,-TC00 and 2,3,7,8-TCDD. Detection and quantification of 3 parts-per-trllHon (W/W) of 1,3,7,9-TCDO was achieved In this complex biological matrix. Figure 8. HRGC/HRNS Chromatogram From Spiked Fish Analysis The highest selectivity and sensitivity for PCOD and PCDF analyses Is provided by combining HRGC/HRNS with a technique called multiple Ion detection. With this 6-64 HONS 21*103 procedure, the mess spectrometer is repetitively jumped between the peaks of Interest. Since time Is not required to scan between peaks, the ultimate sen sitivity Is achieved. Unlike low resolution mass spectrometry that monitors a one atomic mass unit wide window, high resolution mass spectrometry monitors a much narrower window. Thus Interferences with nominal masses that are the same as PC00/PC0F are often resolved. The high resolution mass spectrometer Is typically operated at a resolution of 10.000. Resolution Is defined by the ratio M/aM where M Is the nominal mass and aM the difference between two adjacent peaks of equal height which have a valley of I01 between them. The exact masses that are monitored for selected PCOO/PCOF isomers are listed in Table II. Table II SELECTEO OATA FOR THE POLTCHlORINATEO OIBENZOFURANS ANO POLTCHLORINATEO OlBENZO-p-OIOXINS OF INTEREST Compounds Tetrachlorodlbenzofuran Tetrachl orodl benzo-p-dl oxl n Pentachlorodlbenzofuran Pentachlorodibenzo-p-dloxin Hexachlorodlbenzofuran Hexachl orodl benzo-p-dloxln Heptachl orodlbenzofuran Heptachlorodlbenzo-p-dloxln Octechlorodlbenzofuran Oct achl orodl benzo-p-dloxl n Accurate Mass Low Mass Hiqn Mass 303.9016 305.8987* 319.8965 321.8936* 339.8597* 341.8567 355.8546* 357.8517 373.8207* 375.8178 389.8156* 391.8127 407.7817* 409.7788 423.7766* 425.7737 441.7428 443.7398* 457.7377 459.7347* Theoretical Isotope Ratio .77 .77 1.54 1.54 1.23 1.23 1.03 1.03 0.88 0.88 * Quantitation Mass For each Isomar class, two of the molecular Ion peaks are monitored, thus Increasing the selectivity of the method. By requiring that isotope ratios are correct, false positive Identifications are greatly reduced. 6-65 HONS 21510* CONCLUSIONS The large number of Isomeric forms end the complex neture of environmental matrices pose a difficult problem for the chemists and biochemists struggling to determine the existence, distribution, and toxicity of PCB, PCOD, and PCOF in the environment. This problem can be met by the resolution and sensitivity capabilities offered by glass capillary gas chromatography, combined high resolution gas chromatography/high resolution mass spectrometry, and other sophisticated analytical methods. HOMS 215105 r r REFERENCES [1] 0. G, Barnes, "Dioxin" Production from Combustion of Biomass and ttastas," presented at the 1983 Energy from Biomass 4 Wastes VII Conference (Institute of Gas Technology) Lake Buena Vista, Florida, January 24-28, 1983. DISCUSSION There were a number of comments regarding the developing "standards" and the measuring techniques. 6-67 MOHS 219106 r~ r OETECTION OF POLYCHLORINATEO D1BENZOFURANS AND DTHER CHLORINATEO PYROLYSIS PRODUCTS IN THE SOOT FORMED IN PC8 FIRES C.H. Williams, Jr, C.L. Prescott L.D. Girretson Redlin Corporation 8501 MoPac Blvd Austin, TX 78759 ABSTRACT Uncontrolled electrical fires Involving transformer oils containing PC8s are found to form a complex and potentially toxic mixture of chlorinated products In the 'Soot, Including the chlorinated dlbenzofurans (CDFs), with degrees of chlorination from Cl, to Cl6. Concentration levels vary widely, depending on the specific con ditions of the fire, but levels of 2,3,7,8-tetrachloro dlbenzofuran (2,3,7,8-TCDF) In the range of 1-10 ug/g of soot were measured In some samples. In the same sam ples, no chlorinated dlbenzodloxlns (COOs) were detected, but several other classes of chlorinated products were found In greater abundance than the CDFs. These In cluded the chlorinated blphenylenes (Cl t-Cl6) and chlorinated PNAs, such as the chlorinated pyrenes (Cli-Cl5), as well as residual PCBs. Unchlorinated PNAs from anthracene to the benzopyrene (5 ring) group were also detected at levels compar able to the PCBs. INTRODUCTION The purpose of this work was a comprehensive characterization of the organic pro ducts found In the soot formed In PCB electrical fires. The samples, both bulk soot and surface wipes, were collected by NIOSH Industrial hygienists at the site of several fires Involving PCB-askarels. Radian was directed to Identify and quan tify any CDFs or CDDs In these samples, as well as the other major organic species, chlorinated or not. The analytical method was based on Soxhlet extraction followed by high resolution gas chromatography (HRGC) coupled to a mass spectrometer (MS) operated In either a limited scanning (SCAN) or selected Ion monitoring (SIM) mode, using only electron Impact Ionization. The results of these analyses are surmamed here. A more complete report of this work Is being published In Chlorinated Dioxins and Dlbenzofurans In the Total Environment. Volune 2, Butterworth, 1984. EXPERIMENTAL METHOD Bulk soot In quantities from 0.1 to lg and surface wipe samples on filter paper (100 cm1) were collected by NIOSK Industrial hygienists at the site of several PC8 fires. The samples were analyzed by Radian Corporation using the experimental method shown In Figure 1. Both sample media were Soxhlet extracted with benzene or toluene for 8-12 hours and analyzed without cleanup In order to characterize the complete sample. Samples were spiked before extraction with Internal stendards of llCia*2,3,7,B*TCDD, dj-3,3',4,4'-TC8 (dc-TCB) and dI2-chrysene. Sample extracts were concentrated to final volumes of 100-500 uL by low temperature (25-SO*C) 6-69 MONS 21510? solvent evaporation. The concentrated extract was spiked with another internal standard, d|2-anthracene, just prior to analysis. Samples underwent two successive HRGC-MS analyses: a screening analysis with the MS in a limited mass scan (150-500 amuj mode, and a confirmation analysis with the MS in a SIM mode, simultaneously measuring four characteristic Ions. Both these analyses were carried out on a Hewlett Packard Model 5985A guadrupole GO-MS system. The screening analyses were accomplished using a 50m WCQT capillary column with a bonded, neutral stationary phase, e.g., D6-5 or OV-1. Isomer-spectflc confirmation of the TCDFs was accomplished on a 60m silica capillary column coated with a more polar phase, Supelco 2340. Figure 2 shows an HRGC-SIM-MS chromatogram of a Radian standard containing all 22 of the TCDD isomers, plus the Internal standard 1}Cl22,3,7,8-TCDD, analyzed on the Supelco 2340 column. RESULTS AND CONCLUSIOHS Soma of the bulk soot samples collected at the site of PCB fires show a very complex pattern of occurrence of the PCDFs. Figure 3 shows the HRGC-SIM-MS chromatogram of a bulk soot sample analyzed for the PCDFs. At least 21 peaks were identified as PCDFs and as many as 26 of the 36 possible Isomers could have been present. It can be seen that one of the major TCDF peaks is Identified as the 2,3,7,8 Isomers (26.8 min). Three large peaks coeluting with the TCDFs were Identified as trlchloropyrenes. Figure 4 shows the characteristic mass spectrum of 2,3,7,8-TCDF in which the , loss of the CDC1 to produce the fragment Ion at mass 241 Is apparent. Figure 5 shows the mass spectrum of one of the trlchloropyrenes which has the same molecular weight but a different chlorine Isotope pattern and a different fragment Ion (234) due to the loss of Cl2. The occurrence of the most significant chlorinated products formed In the pyrolysis/ combustion of the PCBs In accidental fires Is Indicated In Figure 6. It seems clear that the mechanism of formation of these major products may Involve both pyrolysis and combustion and nay depend on other organic materials present In the fire In addition to the PCBs. Figure 7 gives a pattern of occurrence for the major pyrolysis/combustion products that were detected In this study of soot and wipe samples from PCB fires. The or dinate In Figure 7 Is given without units since the actual concentrations found will depend on the specific circumstances of each fire and the sites where samples are collected. However, In this study, maximum concentrations of the PCDFs In the soot were measured In the range 1-10 ug/g (ppm w/w), with corresponding levels on surface wipes falling In the range of 100-1000 ng/100 cm2. The most significant findings In this work are summarized as follows: e high concentrations of TCDFs and other CDFs can be found In the soot formed In accidental PCB fires-, e the acutely toxic 2,3,7,8-TCDF Is one of the most abundant isomers In the complex pattern of TCDF homologues extracted from the soot; e other potentially toxic chlorinated products and PNAs are found In even higher concentrations than the CDFs In the soot from PCB fires; and e no dioxins (CDDs) were found in this study above a detection limit of 1-10 ng/sample, i.e., 1-10 ng/g for soot and 1-10 ng/100 cm2 -for wipes. HONS 215108 6-70 lb w OISCUSSION QUESTION: ANSWER: Has the soot been tested for toxicity? Yes. At Binghamton they did toxicity tests with the soot It was toxic to animals 6-71 HONS 219109  CONFIRMATION: NKC-Sin-NS INTERNAL STANDARDS Hgur* 1. Block DUgrM of tbe Aiulytlc*) tttthod MONS 215110 HONS 215111 n -9 *jC*2370-TCDD (IS) IZfc* n<> Figure 3. HKC-SIH-HS CfcroaitogrM of hi)* Soot Seieple froa $ PCI Fire Analyzed for the TCDFi HONS 219112 Figure 4. Min Spectrwe of 2,3,7,8-TCDF fro* the Peek tn the ChrooHtogreei of figure 3 *ONS 215113 9L- ei Figure 5, Nets Spectrue of * Trlchloropyrene froa the Peek et 20 3 Minutes In the Chroaetogrea of Figure 3 MOWS 2LSL14 r r Cl* (PC8a) Polychlorinated Biphenyl* Cl?-Cl* ctr (PCBP) Polychlorinated Slphanylene* CIj - Cl* Polychlorinated Dlbanzofuren* Cl,-Cl* (PCPY*) Polychlorinated Pyrene* Cli - Clj Figure 6. Representation of the Structures of the Principal Chlorinated Pyrolysls/Combustlon Products Found In the Soot from PCBAskarel Fires. HONS 215115 6-77 25 Polychlorinated Biphenyls (Clj - Clg) 1 1 & V 1 t1 a tt 2 10 Polychiorlnated Blphenylenea (Cl2-Cle) 2 Polychlorinated Pyrenes (Cl0 - CI3) Polychlorinated PNAs (CI0-CI2) 1 Polychlorinated Dlbenzofurans (0,-01#) <0,01 Polychlorinated Dlbenzo-P-Dloxins Figure 7. Relative Abundance of Pyrolysis/Combustion Products Formed in PCB-Askarel Fires. 6-7S MONS 215116 r r Part 7 DECONTAMINATION OP MINERAL OIL MOMS 21511? SOLVENT EXTRACTION OF PCS FRON TRANSFORMER MINERAL OIL 6. Add1s(*\ T. Rousef6). C. Walker<6). P. My<c) () Electric Power Research Institute Palo Alto, California (b) General Electric Company Pittsfield, Massachusetts (c) Consulting Chemical Engineer Boxford, Massachusetts ~ PCS (polychlorinated biphenyls) can be removed from e transformer mineral oil by mliing the oil with an laenlsclble liquid In which PCS are soluble. PCS will be removed from the oil Into the extraction liquid until equilibrium Is established. If the partition coefficient Is one, one-helf of the PCB In the oil will be dissolved Into on equal volume of extraction liquid Initially containing no PCS. Ninety percent of the PCB will be extracted on mixing the oil with nine volumes of the second liquid - end to on, 011-llquld systems with higher partition coefficients require lest extrectlen liquid to reduce the PCB content of oil by a given amount. Unfortunately, large partition coefficients for KB-oll systems are unlikely on thermodynamic grounds and tha amount of extraction solvent needed to remove PCB will be large. Any practical process will require reuse of the solvent after removal of the previously extracted PCB. This con be done by using an extraction liquid which has a boiling point well below that of PCB (-2T5*C and up). The solvent can be recovered by distillation, leaving the accumulated KB In the distillation column bottoms for removal to final Incineration (1), ' The work reported here Is based on extraction of KB by the monemethyl ethar of diethylene glycol - methyl Cerbltol*" ('MeC'). The boiling point of MeC Is 1B3*C and Its density Is greater than that of oil (1.03 g/ml versus 0.1 g/ml d ZS*C). MeC was selected from half a doien candidates on the basis of preliminary testing at M*1ent temperature In the laboratory. The KB partition coefficient Is approximately one. Its solubility In oil Is -AX and the solubility of oil In MeC Is -8*. The Interfaclal tension between MeC and either unused or aged transformer oil Is rather low (<1 dyne/cm). The choice of MeC represents a compromise. The partition coefficient for dimethyl acetamide, far example, Is two to three times higher but so also Is Its solubility for oil. The solubility of oil In methanol Is much lower, but the partition coefficient for KB Is also substantially lower. A smell scale demonstration plant was designed and put In place to establish the technical feasibility of the extraction process and to evaluate the relation between controllable process parameters and efficiency In a continuous operation In order to estimate the magnitude of the process costs. A commercially available (York Process Equipment Company) agitated multi-stage countercurrent extractor was used. Oil containing PCB was fad Into the celum near the bottom end tha extracted product was removed at the top of the column. KB-froo MeC MOMS 215118 7-1 was Introduced Into the column near Its top and removed near the bottom. The extract was fed to a vacuum distillation column where the Mac was distilled from the dissolved PCB end oil The NeC was recycled to the extraction column and a PCS enriched oil fraction accumulated In the distillation column bottoms. Two unusual modifications were made to this process. If, for example, the flow rates of HeC and oil were equal In an extraction column of given design operating at ~25*C, a certain efficiency of PCB removal (PCB concentration In the feed streen minus concentration In the product streem/concentratlon In tha Feed) would be achieved, without modification, the process would low M of the oil feed to the PCB residue and hence to Inclneretlon. Cither e larger column or a higher solvent-to-oll ratio can Increase the efficiency. The first odds to the capital cost of the equipment; the second odds to the oil lost and the Incineration cost. However, If the oil Is recycled from the distillation column bottoms back into the bottom of the extraction column as extractor reflux, It contributes to maintaining the oil saturation In the extract streen. This reflux extraction reduces the oil loss to that deliberately removed for Incineration when the PCB concentration reaches some predetermined level. Loss of no more than IX of the oil was set as a goal In this study. This would result In a PCB concentration In the distillation column bottoms one hundred times that of the reduction In concentration between the feed and product oil. An extractor reflux concentration and residue concentration as high as $X PCB could be reached. It also Is desirable not to overheat the oil In the still bottom In order to reduce degradation of oil In the reflux stream, some of which mixes back Into the product oil flow. To minimiu the bottom temperature without requiring an extremely low vacuum, the solvent separation Is carried out In two stops, a partial separation by distillation, and then a final purification by phase separation. In the partial separation, enough solvent Is allowed to remain In the distillation column bottoms to permit operation of the column under economical vacuum conditions and acceptable temperatures. This mey correspond to a distillation column bottoms composition of between 20 and BO percent solvent. In the second separation, this bottom product Is cooled to a temperature below the limits of complete mutual solubility so as to form a PCB-oll phase end a solvent phase. After separating these two phases by decanting, the solvent phase Is returned to the distillation column, while the PCB-oll phase Is returned, for the most part, to the extractor as reflux. The pilot plant was operated with oil feed rates up to 10 gallons per hour and extraction reflux rates to 2 gallons per hour (the reflux rate cannot exceed the limit of the fractional solubility of oil In the solvent flow). A reflux retlo In the distillation of one to four was used and the PCB content of the returning NeC remained less than 1 ppm. Ratios (weight/weight) of solvent-to-oll (feed * reflux) rates ranged from l.T to 4. Unused oils with Initial concentrations between 300 and 1400 ppm of Aroclor 1254 and aged oils with Initial concentrations between 150 and 500 ppm of Aroclor 1250 were tested. The results were evaluated using the standard model for countercurrent solvent extrectlon; E<*) - 100 x [(mL/V)n+1 - (mL/V)]/[(mL/V)n+1 - 1) where E Is the efficiency of extraction; m, the partition coefficient; n, the number of theoretical stages in the extractor; L, the Internal mess flow rate of solvent and V, the Internal mess flow rate of oil (In this case, tha combined feed and reflux flows) In the column. The data Is an excellent fit for a value of n of 3.3 and values of n of 1.0 and 0.6 for Aroclcrs 1254 and 1260 respectively. MONS 215119 f r Aroelor 1260. the more heavily chlorinated PC6 mixture. It lett effectively extracted because of Its lower "average1 partition coefficient. The chromatography of the feed and product streams shows that the partition coefficients decrease for the more heavily chlorinated Isomers, Aroelor 1242, the least chlorinated of the PC8 mixtures used In askarels, should be the most effectively removed. . ... Variation of extractor temperature between 20 and 40*C had no effect. The column efficiency was not affected by agitator speed between l5 and 260 vpm. Above this foaming occurred, the phases failed to separate and the extractor coIumi flooded. The rather low Interfaclal tension Is presumably the cause.The case of ninety percent removal efficiency Is of Interest In that feed streams containing up to 500 ppm PCS could be reducedto below 50 ppm (At IX oil Incineration, the bottoms concentration would be less than 4.501). The case of 9SX efficiency Is also of Interest In that feed streams containing up to 100 ppm PCS could be reduced to less than 2 ppm. for estimating purposes. It Is assumed that three quarters of the PCS In a typical feed stream will be from Aroelor 1260 and the remainder will be due to a mixture of Aroclors 1242 and 1254; l.e., m~0.7. with an extractor of the present design, solvent to oil flow rate ratios of 2.2 and 4.1 would be required to achieve 90 and 911 extraction efficiency respectively. If the number of stages Is Increased by fifty percent, the same result can be reached with solvent to oil ratios of 1.7 and 3.0. The lower solvent-to-oll ratio and more efficient extractor design appears to be a more cost-effective approach. The costs resulting from several possible operating scenarios have been estimated using this approach. Costs wore estimated for operation at oil processing rates of 60 to 4B0 gal/hr and for one, two or throe shifts per day. Annual production then ranged from 100,000 to slightly less than three million gallons. In the base case. It was assumed that a fixed Installation site would be adjacent to an existing facility for treating, handling or burning oil. Labor and supervision wore expected to be shared with that of the existing facility. Available Indoor space for Installation of control and Instrument panels was assumed. Steam generation and cooling water facilities were expected to be In place. The costs for these scenarios are sunmerited on Table 1. The estimates Include transportation and the Incineration cost for 21 of the oil. The added costs for operation In a second siting situation also have been estimated, it was assumed that steam generation and cooling water facilities must be Installed together with housing for the weather sensitive equipment and, further, that Isolated Installation would be made where full minpower requirements must be met and that oil handling facilities would have to be added. The additional costs of these scenarios are listed In Table 1. It was assumed In the base case that transformer oil reconditioning facilities were available. Additional costs are listed for a situation where this Is not the case and filtering, degassing and claytreat equipment must be put In place. It must be reiterated that these are estimates based on the usual engineering basis and using typical costs for utilities. Interest rates and so forth. They are Intended merely to define the general range of costs and may not reflect the real cost. An estimate hai been made, however, on a comparable basis for a process described elsewhere (1,2) using sodium metal, naphthalene and dlglyme to destroy the PCS In situ In transformer oil. The results are similar. Cost of PCS removal by high energy electron treatment or by critical fluid extraction (1) are estimated to be higher. MOMS 215120 7-3 r'- Th results of thoto preliminary tests Indicate that reflux extraction of pcb from transformer oils can be done with high efficiency. Preliminary estimates suggest that It is potentially cost competitive with other processes. This work was sponsored by the Electric Power Research Institute as a part of Research Project 202B-1. (1) T. o. Rouse, Removal Of PCB From Transformer Oil. Proc. PCB Seminar, EPRI tm?%} Qf PCB FTM" Transformer 011-flnml Report. EPRI RP 2026-1 (In preparation). (2) J. F. Brown, K. E. Lynch, J. C. Carnahan, J. S. Singleton, Chemical Distribution Of PCB In Transformer oil. Detoxification of Hazardous Wastes, p. 201, Ed. J.H. Exner, Ann Arbor Scl. Pub, (1962). HONS 215121 7-4 r r TABLE 1 SUMMARY COST ESTIMATES RCB -SEP_PROCESS PRODUCTION RATE (OaVHr) Shlft/Dty ANNUAL PRODUCTION (KD*1) A,1 105 ME REMOVAL EFFICIENCY sttTipwA" inmiMiffl CAPITAL INVESTMENT (SR) PROCESS COST (S/Otl) 410 2.66 1.25 i1 840 2880 1260 1.03 0.60 0E REMOVAL EFFICIENCY CAPITAL INVESTMENT (SR) PROCESS COST (S/tel) 410 2.3T 1.12 1015 0.84 0.49 ADDITIONAL PROCESS IMS (APOTO OASE^CPST A8QVQ REMOTE LOCATION - ADA STEAM BOILER. COOUHS TOMER. 8UILDIN6. OIL TAHKAK IfMIL MMUMLR AOOEO INVESTMENT (SR) AOOEO COST (S/Gtl) 21$ 1.42 0.91 402 0.26 0.14 CLAY TREAT 6 DE6ASSIN6 AOOEO INVESTMENT (SR) AOOEO COST (S/8t1) ss 0.21 0.07 . 130 0.07 0.03 MONS 215122 7-5 Pilot Plant studies for Solvent Extraction of Polychlorinated biphenyl ;PCB/ C. M. Hancher M. B, Saunders Dak Ridge r-12 Plant Oak Ridge, Tennessee Nuclear Materials Processing and Waste Management Technology Department Development Division Operated by UNION CARBIDE CORPORATION for the U. S. DEPARTMENT OF ENERGY under Contract No. W-7405-eng-26 ACKNOWLEDGMENTS The authors would like to acknowledge the professional services of P. F. (Paul) Meredith and F. L. (Lynn) Yearwood for laboratory operating and construction advice; and K. R. (Kelly) Hopson. N. G. (Nell) Cannon, and R. J. (Ron) Mctlhaney for analytical chemistry assistance. SUMMARY The polychlorinated biphenyl (PCS) concentration system performed very satis factorily Mhlle using dlmethylformamlde (Wtf) as a solvent. The pilot-plant extraction system processed oil with a PCD concentration as high as 300 ppm at a rate of 0.8 L/mln using two columns which were each 10 cm In diameter and *,6 m high. Higher volume flows and lower PCS removal should be obtainable If larger extraction columns were used. The PCS removed from the BMP Is 16 to 20 times more concentrated than the PCS present In the original oil. The recovery of dry UhF (less than 1.5 vol % water) was accomplished using distillation at a rate of U.3b kg/h-cm2 (based on distillation column cross section). 7-7 MOMS 21S123 Section 1 INTRODUCTION A large quantity of mineral oils (used as transformer, machine cutting, cooling, and lubricating oils) In the Oak Ridge Y-I2 Plant* contain Ion concentrations (50000 ppm) of polychlorinated biphenyls (PCS). Because the Plant does not have a burial ground, these oils must either be stored or Incinerated. Also, as some of the oils are contaminated with uranium, they cannot be shipped off-site for Incineration or burial because commercial PCB disposal facilities cannot accept uranium-contaminated wastes. Long-term storage Is an option but, because of space and labor requirements, is expensive. To reduce storage or disposal costs, the PCB must be concentrated before off-site Incineration or on-site storage. The developed extraction process contacts N,N-d1methylfomam1de (DMF) with PCB-contamlnoted oil and transfers the PCB to the DMF phase. Meter Is then edded to the PCB-loaded DMF to destroy the polar bonding of the PCB and to transfer the PCB to a smeller volume of oil. *Operated by the Union Carbide Corporation's Nucleer Division for the Department of Energy. MQMS 218124 7-a Section 2 PILOT-PLANT EXTRACTION SYSTEM FOR PCB PROCESS The DMF-PCB extraction process removes PCB from oil into a OMF phase using a counter-current, liquid-liquid packed column (Figure 1). The UMF must be relatively water-free (less than l.S wt X) to ensure satisfactory extraction. The PLb is later released from the DNF phase Into a receiving oil phase by the addition of Mater, which Is soluble In the OMF. The Mater renders the PCB Insoluble in the UMF by reducing the poler bonding of the PCB. The resulting oil phase has a PCB concen tration factor of 16- to 20-fold, based on the PCB concentration In the original oil. The OMF-water can then be separated by distillation for recycling the DMF to the extraction columns. PROCESS CHEMISTRY Under ordinary circumstances, all PCB compounds are highly stable and unreactive due. In part, to the resonance stabilization offered by the chlorine substituents. Only extreme reagents end reaction conditions form new compounds. It would be expected that, because of the chlorine atoeis, the resultant dipole would render PCB soluble In polar solvents. However, experimental results revealed that the mineral oil, although nonpolar, was one of the best solvents. Furthermore, dipole moments measured for PCB showed thma to be considerably less than the analogous aliyl chlorides. An explanation for this Is that chlorine, although highly electro negative, Is releulng electrons to the ring by double bonding rather than attracting the electrons. This situation would allow the chlorine to Increase the resonance stabilization of the molecule and reduce the expected negative charge. The electron release by the chlorine would also make the carbon ring more negative. Thus, the intramolecular bonding In PCB is less polar and more van der MaeIs in nature; therefore, high solubility in nonpolar organics would be expected. The PCB compounds also show high solubility in polar organic solvents but low solubility in water. The solubility In polar solvents appears to be the result of a positively Induced carbon In the solvent that attracts the negative charge on the biphenyl ring structure. The insolubility of water Is possibly due to resonance inhibition caused by hydrogen bonding to the chlorine. HONS 215125 7-9 Ftflur* 1 PCS RECOVERY FLOWSHEET HONS 215126 7-i`o r r PROCESS EQUIPMENT The pilot-plant process equipment flowsheets (Figures 2 and 3) consist of three parts: (1} extraction, (2) transfer, and (3) distillation. The extraction consists of two 55-gal feed tanks and two 55-gal PCB-contamlnated oil feed tanks. Two packed extraction colimns are each 10 cm (4 In.) In diameter X 4.6 m (IS ft; high, with a 91-cm (36-in.) unpacked phase separation tone at each end of the columns. Two 5S-ga1 stripped oil product tanks, a DMF Interstage tank with a pump level controller, and a tank to hold PCB-loaded DMF were also part of the pilot plant. Each column had a bottom control valve actuated by a bubbler differential pressure (DP) cell, liquid-level control system to maintain a DMF continuous phase. The DMF transfer tank was a cone-bottom, 55-gal drum with a mixer. The columns were packed to a 4.6 m (15 ft) depth with 2.5-cm (1-In.) ceramic saddles. The distillation equipment consisted of the following Items: 1. A packed coluan 7.62 cm (3 In.) In diameter X 2.4 m (8 ft) high packed with perforated stainless steel packing. 2. A reboiler with a volume of 18 L and heated by a 0.46-m2 (5-ft2) steam coll heat exchanger connected to 120 pslg-regulated steam service. 3. A top condenser with a cooling area of 0.46 m2 (5 ft2 ). Including a reflux splitter. 4. Two 5S-gal feed tanks, two 55-gal water product tanks, and two 55-gal DMF product tanks. COLUMN EXTRACTION TESTS Using one or both of the columns, tests have been performed using recycled IMF from the dlstallatlon column and clean oil to which concentrated PCS was addedto a concentration of 200 to 400 ppm In oil (Table 1). The recycled OMF had a water content of 0.7 to 1.5 wt t water. About 2 wt X water Is the permissible maximum allowable for satisfactory PC8 extraction. The oil feed must also be water-free (l.e., less than l.S wt X). The mutual solubility of DMF and oil In each other Is about 5X. The extraction tests were performed using both DMF and oll-PCB as a continuous phase. Therm was no apparent dlfferece In extraction performance using the 7-11 HOMS 21512? OMF FEED OIL WASTE ~3-, ky OMF-PCS LOAOEO F)in2. PCS RECOVERY EQUIPMENT FLOWSHEET EXTRACTION SECTION MOMS 215128 F PlfunS. FCS RECOVERY EQUIPMENT FLOWSHEET TRANSFER ANO DISTILLATION SECTION 7-13 HONS 211129 Table 2-1 SDLVENT EXTRACTION TESTS OMF Oil Column OMF Test Flow Flew Length In Number (mL/mln) (mL/mln1 (ml (apm) OMF/Qil OMF Oil Oil {distribution (distribution Out In Out Coefficient Coefficient (aw) (ppm) (ppm) Too bottom 1 550 440 4.6 120 200 270 110 - - 2 370 430 4.6 5 150 160 50 1.6 1.4 3 230 1200 4.6 25 420 300 200 - 4 660 750 9.2 2 220 380 94 1.6 1.5 different continuous phases. Therefore, a DMF continuous operation Mas selected because the phase interface was at the top of the column, which was better for the Interface control Instruments. The flow rates were varied fro# 0.3*1.2 L/m for both phases to determine maximum operation rates. The phase separation capability at the top of the extraction coluan was the Halting operation. The continuous DMF-phase can be operated slower than the loaded oil phase. A typical set of flow rates are 0.60 i/m (7.7 cm3 /cm2 min) for OMf and 0.80 L/m (10.3 a?/cn min) for oil, based on empty extraction columns. Table 2 contains details of the extraction column tests. The phase distribution coefficient (DMF-oll) was determined using equal volumes entering and exiting each end of the extraction column (l.e., loaded OMF and loaded oil for one pair and clean OMF and dean oil for the other pair). The two phases were shaken for 1 min. allowed to separate, and then analyzed for PCB content and compared. The calculated value, using laboratory analytical data of i.b for the phase distribution coefficient, confirms earlier batch shake*out data.1 OMF-WATER, PCB-OIL TRANSFER The PCD can be released by the addition of water, which Is soluble In the OMF phase, reducing the polar bonding of the PCS to the OMF and allowing the PCS to be transferred to a receiving oil phase. Tests were performed (Table 3) to determine optimum water volume percent, oil volume percent, and mix time parameters for maximum PCB transfer. Using 22 vol t water and A and 10 vol X oil respectively. 7-ie HONS 215130 Table 2-2 SOLVENT EXTRACTION DATA ii4 DMF O.S - O.B Oil 0.8 - 1.0 Flow (L/mln) 0-10 200 * 300 PCS (ppm) 1 - Oil Hi Continuous Phase Discontinuous Phase Product PCS 011 DMF _LfifiSi-- ill J11 Lotted DMF 400 - 600 5 - Stripped 011 SO -5 Equipment 2 - Columns, 10 cm dlam (4 In.) X 4.6 m (15 ft) Filled - 1,9 cm (3/4 In.) Ceramic Saddles Phase Equilibrium Data (distribution coefficient) PCB ^ - 1.4 - 1.6 Table 2-3 PC8 TRANSFER PCB ImALeeal_ Oil DMF Volume 111 111 Esrsasi Loaded DMF 400 - 600 Water (recycle) S Receiving Oil 25-200 5 2 - 67 2 25 5B Nix - 30 min Discharge Oil DMF-Water Total Volume Percent PCB toon) DMF 111 7 4000 - 8000 5 93 10 - 30 95 Oil HI 95 5 Bttch Size - 40 gal Settling Tima - 4 h MONS 215131 7-15 the 10 vol X oil test removed the PCS from the DMF 120X better than the 4 vol % oil test. Tests using 15, 21, and 28 vol X water showed that the 28 vol % water tests phases separated better, and a better PCS removal was obtained. A test with 72 vol t DMF, 28 vol X water, and 15 vol X oil was mixed for 5, 10, 20, and 40 min and sampled after each time period. The PCS concentration In the DMF was o ppm in each time period sample, Indicating very fest equilibrium transfer time. Although 5 min should be sufficient, 30 min was selected because extre time was available. DISTILLATION The distillation unit Is a "high efficiency" unit 7.62 cm (3 In.) In diameter. The maximum water boll-up rate at 100 pslg steam Is 1200 mL/mln. The normal OMF-water lead Is 72 vol X OHF and 28 vol X watar (Table 4). Controlling the reboller at 151-153C, the water overhead distillate hes a OMF concentration of 5-10 vol X OMF. DMF In the water phase is not serious because it is recycled to the transfer tank along with the water used for polar bond breaking. The bottom DMF reboller product has a water concentration of 0.5-1,0 vol X. Table 2-4 01STILLATI0N Feed DMF-Hater PCB Oil Rate 1vol X) (PM) ivgl X) (L/mln) 67 - 26 10-30 7 0.5 - 0.9 Prfduft Top-Hater Bottom-OPF PCS Oil OMF Meter (DM) (voi xi (ysm (vol X) 15 5 10.0 B5.0 20 - 40 1 9B.5 0,5 Temperature Rate ( C) (L/mln) 103 - 105 0.2 - 0,4 151 0.3 - 0.5 Equipment Steam Heatad - 105 pslg Sat. at 160C Boiler - 5 gal and 5 ft2 Heat Transfer Column - 4 In. dlam - 5 ft PROCESS ECONOMICS The evaluation of the economic factors strongly favor concentration of the PlB In oil before incineration or other finel disposal operation. The resulting processed 7-16 HONS 215132 oil with less than SO ppm PCB can Incinerate In a furnace with heat recover/ and off-gas control. The ten- to twentyfold PCB-oll concentrate can be shipped to a licensed disposal center. The cost of final disposal is usually based on volume, not on PCB concentration. The container, shipping, and disposal costs will be decreased by a factor of two or more by the PCB concentration process. The DMF-PCS-solvent extraction process will provide the needed PCB concentration to be cost effective. Table 5 lists the cost factors coopered to Incineration of the original PCB-oll mixture. An attempt to concentrate the PCB more than twentyfold Is economically unsound because of base solvent extraction cost of S4/ga1. Table 2-5 COMPARISON OP COST FOR DISPOSAL OF PCB CONTAMINATED OIL (Basis * S/Gal of Original Oil) Solvent Extraction As Received ..... Orums Shipping 1.00 0,50 Incineration 11.00 Total Cost 12.50 Solvent Extraction 1/10 4.00 0.10 0.05 1.10 5.20 Cone. 1/20 4.00 0.05 0.03 0.55 4.53 CONCLUSIONS PCB can be removed from waste oils and concentrated for disposal of a concentration of about 300 ppm. PCB In oil has been removed from oil In pilot-plant equipment (1.e., two extraction colums having a 10 cm dlam), reducing the concentration to below SO ppm. The DMF extraction system consists of two 2.7-m <i5-ft) columns with an oil feed rate of Q.B L/mln (approximately 300 gal per day), which Is equivalent to a PCB rate of 0.2 g/mln. The PCB was water-transferred to a disposal oil phase with a PCB concentration up to 400 ppm. The DMF was separated from the water and added to the OMF in the PCB transfer-to-ol 1 storage step by slaple distillation at 7-17 MONS 215133 /* a reboiler temperature of 151-153C- The distillation rate is about 0.3S L/mln, and the water content In the DMF was reduced to less than L.5 vol X. REFERENCES 1. Napier. J. M., Travagllnl, M. A., Laggls, E. G., Nakarewlcz, H. A,, Evaluation and Development of Polychlorinated biphenyl Removal Process, Y-OZ-1, February 5, 1982, DISCUSSION QUESTION: ANSWER. QUESTION: ANSWER: questionANSWER -. Do you see my problem with DMF In contact with steel materials? No. How much OMF is in the oil? l%-2% which can be water scrubbed out, so essentially, none. Can the process get below 2 ppm? Pilot plant was based on getting below 50 ppm, but It can get below 2, If needed. 7-ie MOMS 215134 STATISTICAL SURVEY OP PCS CONTAMINATION IK SUBSTATION AND DIITKIBUTIOH SYSTEM EQUIPMENT CONTAINING HINEBAL OIL Mark D. Sapareteln* tad Edward J. Pander"" ABSTRACT Elaetrle utility aabatatlea tad dlatrlbutloa equlpawat coatalalai mineral oil eaa ba contaminated with polychlorinated blpheayle (PCI). Thla contamination may occur durla| tba manufacturing of tha equipment, or during eervlce aad rapalr If PCB contaminated aatarlala aad equipment ara aaad. Thraa acapline propra** have baaa coaductad to aiaalaa tha aitaat of PCB coataalaatloa la a larta alactrlc utility eoapaay'a aabatatlea aad dlatrlbutloa ayataa equipment. Tha apaclflc poale of each eervey aad tha aaapllaB technlquee aaad wara eomawhat differentDva to thaao dtffaraaeoa, tha nitrnpolatloa of eaaple raaalta to eetlmate contamination oa the aatlro eyetem la llaltad la aM reepacte. However, all thraa eurveya slva eoaalataat raaalta lodlcatlae that oaly about thraa to alz percent of mineral oil equipment la coataalaatad with PCB above tha 90 ppa level. The aaaa PCB cooeaatratloa la dlatrlbutloa ayataa traaaforaara waa approilaataly S ppa aad tha aadlaa level of coataalaatloa waa laaa thaa 2 ppa. Tha aaaa PCB eoaeaatratloa la aabatatlea equipment waa approilaataly 12 ppa with .a aadlaa level of eoataalaatloa of 2 ppa. INTRODUCTION Electric atllltlaa have pravloaaly had tha choice of purchaelaB aabatatlea aad dlatrlbutloa equipment eoatalatat either Aakaral (a aliture or PCB aad chlorlaatad benceaee) or alaaral oil aa aa laaalatloa or heat eichea|e fluid. la lBfP, a rasalatory baa waa lapoeod oa tha manufacture and *a of PCB* la coamerce, aad utllltlaa are now phaalag out tha uaa or cartel a PCB-coatalalai equipment- Thera la coacara over tha aitaat of PCB coataalaatloa of olaeral oil-filled equipment * KPRI Kaersy tlaaataoaat Intern. Eaarsy Anelyde and Environment &lvl*l*a, currently on loan to Southern California Edlaoa Compaay. ** Conaultlas lelaatlat, Southern California Edlaoa Company, Boeearch aad Development Department. 7-19 MONS 215135 manufactured before the ban of PCI* and equipment aervlead In tha same faellltiaa with PCI-contalnlng equipment. Undar currant fadaral regulations, liquid* con taminated with z 90 ppm of PCI ara con*Id*rad PCI contaminated liquid* and, If dl*po*ad, must ba handlad a* haiardou* wests. Furthermore, all mineral oil aqulpnaat muat ba HiiMd to contain z 50 PP* f PC* unJ-it otherwise verified. It 1* tharafor* of great Interest for utllltla* and regulatory agenda* to de termine th* aitant of PCI cootami nation In mlaaral oil aqulpmant *lacn til* will lnfluanc* waat* dlcpoanl method*, (pill cleanup praetlca* and aqulpmant headline procedure*. Th* extant of PCI contamlBatlon will, la part, determine the financial Impact of any futur* changes In th* regulatory definition of PCI contaminated liquid*. Th* southern California Bdlaon Company ha* 741,191 pleee* of aqulpmant on it* *y*t*m whleh contain mineral oil or lateral a* haat aiehang* or dlaloctrlc fluid. Th* great majority of til* aqulpmant contain* mineral oil. It la eetlmated that th* total amount of fluid In mlnaral oil aqulpmant 1* approximately 3t million gallon* with moat of thlc being In mlnaral oil tranaformar* (29 II gal) and mlnaral oil circuit braatarc (9 H gal). Th* remaining oil la contained la awltekea, ecttonallcara, voltage regulator*, raeloaar* and other equipmentTable l ahowa th* amount of PCI and mlnaral oil contained In varloua type* of aqulpmant. Over the peat two year*, three aurvaya have bean conducted to determine th* *x~ tent of PCI contamination In mineral oll-flllad equipment. Th* three aurvaya war* da*lgoad with different objective* In mind, and tharafor*, the campling technique* uaed In each enrvey are aomewhet different. The result* of eaeh nrvey are eubjeet to eertaln limitation* with reopeet to extrnpoletlon of aampl* tatletlce for decerlblng th* entlra population of tranaformera or other mlnaral oil equipment, lino* each eurvey wee eonductad In a different manner, combining th* data Into on* large sample for analyst* we* considered Inadvisable. Each survey will he deearlbed eeparately and then the reeults will be compared and contrasted. IAKPLE SELECT101 limplq li -Distribution Ttaniformer* at a Itoraae and Test Facility The methods end results of this survey have been described in n previous paper (1) and so they will only be briefly summarised her*. The objective of this eur- 7-20 MONS 215136 Table 1 ASKAREL, PCB AND MINERAL OIL ON THE SOUTHERN CALIFORNIA EDISON TRANSMISSION AND DISTRIBUTION SYSTEM JtouiPMnt Aekarel Traneforaera PCB Capacitora Total Fluid In Aakarel Equlpaent Hloeral Oil Traneforaare Htnaral oil Volta** tatulatora Mlnart1 Oil Circuit Breakera Mineral Oil aeloaera Mineral Oil Switches 6 Bactlonalliera Total Fluid In Mineral Oil Eqpt. Total AtKarel 6 Mineral Oil Cnllona of Fluid If,044 331.614 347.7TB |*1. 29,263.311 419,361 3.224,324 37,400 36.267,373 |tl. 36,613,301 (tl vey me to aaaple a saall portion of the 337,000 distribution traneforaare on the ape tea la aucb a way ao that an accurate eetlaate of PCB contaalaatloa la the entire distribution traneforaar population coold be aade. Distribution treaaforaara vary with reapect to aanufacturer> alee, deal*n and age- The flrat eyetea survey waa dee lined to obtain a aaaple which was repre sentative of the whole dlatrlbutloa traneforaar population. One aathod of obtalatni each a aaaple la to conduct atratlflad randoa aaapllni In which the population of traaaforaara arc cleat If led Into atrata with etallar values of knows chareeterlttlct. Unite are then eaapled within eaeh etrataa In a randoa aaaaer, uelug tables of randoa uuabore. Developla* a stratified aaaple with acre than one etratlfler can becoae cuaberaoae and It la alapler to stratify by one factor which la eipected to aaoount for the sreateat aaowat of variability within the aaaple. with reiarda to PCB contaalnatlon In alneral oil traaaforaara, It me felt that production eource of the unit (aanufneturer) me the key verlable. The aaaple waa therefore atratlflad on annufacturer. The nuaber of unite aaapled froa a given aanufacturer ma proportional to the percent of that aaaufaeturer'a unlta In the dlatrlbutloa traneforaar population. Age, alee, or aoae other faster any have also helped to predict PCB concentration, but It ma not poaelble to tabulate how the population varied with reapect to theae factors, and therefore, they could not be used in 7-21 MONS 21513? * stratification- Saapla dace war* developed for **ch aseufecturer and an Initial tralnlat eeaple of US unit* wat obtained. Although It would b* dee treble to acaple ttcneforaare which ar* oe-llne and la aarvlcai thla type of mpilng la quit* expenelve b*c*u** of th* tlM and wpowar Involved In de-energlitn* *|aanti of th* dlatrlbutlon ayatea. As an alternative, samples w*r* tak*a fro* transformers at a atorat* tad t*at fa cility . Only transformers which bad not b**a s*rvleed war* esaplad. Th* ataflint of th*a* traaafonaara was not truly random la that tha stapling personnel a*l*ct*d th* transformer* fro* aaeb aaoufacturar a* th*y wlibod. (Tb*y aa; her* cheaaa thos* unit* whleh war* *a*l*at to *mpl*.) It la also poaalbl* that transformer* at th* ahop and t**t facility war* aet representative of th* popu lation of transfers*TM with r*sp*ct to at* or to** oth*r factor. After th* Initial aaapl* was obtained and analysed, a second aaapl* wa* obtained ualat aa optlaua allocation strategy In whleh th* nuabar of unit* sampled fro* ach *aaufactur*r was a function of th* Internal variability la each strata (nuabar of cootaal**t*d unit*) a* wall th* ala* of the atrataa (2). This technique was uaad to minimis* the varlane* of the final aatlaat* of th* paraant coatselected. Tha total nu*b*r of unlta aaaplad la thla survey was 210. Two valuta whleh war* obtained fro* sludge realdues rsthar than oil war* dlaoardad. iotii i>--intiUjJrtmm imiitlaai Th* ascend survey was p*rfor**d In r**poaa* to KPA'a Interim Meeeure* regulations for the a*aat***at of PCB*. Th* ratulatlon* Mandated that *ln*r*l oil-filled tr*aafor**r* located In *r** wh*r* laah*(* conld cans* cont**ln*tloa of food or anlaal faod, *ost *lthor b* Inspected weakly for leek.* or b* toatsd to determine th* PCB content of the oil. There were 61* BCK distribution traniforawra which fall into the "food and food facility" category tad all war* taatad for PCB coo- toot. tfollka the previous storage yard sa*plat the** traaafonaara were la uaa and aaaplad on-alt*. Though tha aaapl* we* aot atratlflad with reapact to manufacturer, tha raaultaet coapoaltloa was slallar to the population la thla raspact. This aaapl* wa* alao reported to b* alallar to th* population with teepecl to overall ala# *ad ago; however, that* wars only th* awbjectlv* report* of tha personnel who aanpled th* equipment. and cannot b* v*rlfl*d via deeaaaate- tloo. . HONS 14138 7-22 ww Per sample* #1 and *2, the earn* laboratory wee ueed to determine PCS concentration*. Hathod* pravlouely have bean deacrtbad Q), Sample* ware periodically raeubmlttad to the laboratory for rateating to verify the laboratory'* accuracy No lncon*l*tanele ware found. ala V. Subitatlon taulomant The third capla cone let* of 1070 eampla* of mineral oil from *betatIon equipment Including tranaforaare (15 percent), circuit breaker* (10 percent) and eaall amount* of other aqulpawnt lncludini reactora, ra|ulatora. load top cheapen and oil etoroia tanka- Tha baelc approach In thle aurvey le to eaapla all large placae of equipment until the entire population of eubetetlen equipment hea bean aaaplad. Appro*laataly IS perceet of the larsa power treneforaara era rapraeantad In thle eaapla. lubatatlon paraonnal hare routinely aamplad alnaral oil and teatod for carbon contaalnatloa end dleeolved pacea which Indicate that aalfuactlene any be occurrlnp. Initially. PCI aaalyele waa added to thaea aaIntonance taata to determine If the equipment ahould be labeled aa contalnlnp PCI* at laval* * 50 or 500 ppm. Currently, equipment la teazled for the apaciflc purpoaa of dataralnlnp PCI concentration. While thle eaapla le larger than the flrat two combined. It hea aavaral ehortcoalnga which Halt lte uaafulnae*. Thar* war no euamary Information available to deecrlbe th* aubatatlon population with raapect to compoaltlon by varlou* aanufacturar* age and ala* of th* equipment. It 1* therefore difficult to datarmln* how wall th* aampl* rapraeent* th* population. The typo* of equipment aaaplad war* not recorded In a ayatamatlc manner and In a eaall percentag* of caaaa, It 1* not clear praclaaly whet type of equipment we* aaaplad (a.g,, th* oil In a circuit breaker or the currant traneforaar aaaoelatad with the circuit breaker). Thar* may have been aoma mining of olla from dif ferent equipment through th* ua* of aobll oil treatment equipment and atorag* tanka Th* PCI analyaa* war* conducted at a different laboratory than the on* uned le th* flrat two eurvay*. In order to eut analyaa* eoata, PCS concentration* below 1.0 ppm war* not accurately maaeured but .war* merely reported a* 1 ppm. 7-23 HONS 215139 ANALYSIS AnelysLs of dot* froa these thro* surveys wee designed to enswer three besle questions*. I) Ii alnorol oil found In electrlcel equipment eontoalnotod with PCBs, ond. If io, to whet oitontr 2} Whot li tho boot oittaoti of tho proportion of unlti eontoalnotod ot levels of > SO ppaf 3} Are ono or more annufectursrs lololy roiponilblo for tho PCI eonteal netlonT Imjj-flf-KL-ggfl.iialBAtUa1 toiulti froa Soaplo #1 Indlcoto thot PCI eontoalnotlon In almorol oll-flllod troniforaori woi o eoaaon ovont, elthough lovoli of eontoalnotlon worn vory low. Subsequent loaploi of alnorol oll-flllod equipment woro In agreement with thono flndlngi, Figure 1 ohowi tho froquoney distribution of eontoalmotod unlti by eoneontrotlon of PCI for loaploi pi ond 02 coablnod. Soaplo 03 exhibits o lallor pottorn. Tho froquoney of eontoalnotlon dropi droaotleolly botwoon tho eoneontrotlorn of 0 ond 2J ppa with only 3 to o percent of tho unite eontoalnotod ot lovoli obovo SO ppa In inch survey. This soao pottorn woi observed within inch of tho lndlvldunl aonufncturori which eomtltuto Soaplo 01 ond Soaplo 12, Figure 2 shows tho cuaulotlve froquoney distribution for Soaplos 01 ond 02 eoablnod. Soaplo 03 exhibits o lallor pottorn. Thli fl|uro olio doaomtrotoi thot tho greet aejerlty of eontoalnotloo occurs ot low eonceotretlou. Noon PCS eoneontrotlons wore eoleulotod for ooch soaplo. Sloeo tho eoapoiltlon of the distribution troneformer populotloo woo well known with respoet to aeoufeeturers, tho aeons for soapleo 01 ond 02 wore oorroctod to adjust for differ ences botwoon the soaplo strotlfleotloo ond tho populotloo. ^Although federel rogulotlons currently define PCS eontoalnotod fluids os those eontolnlng ot leost $0 but less then 300 ppa FCSs. In this study, PCS cootoalnotlon will refer to tho proooneo of ony cosceetretlon of PCSs aeoeured In alnorol oil. HONS 219140 7-24 r r Kozncowa-*! Pi|ur l. Pr*iHer distribution of PCS contaalnotlon In lmnrol oil flllod oqulpaaat. Data from taaplo tl and Saapla *2 coablnad. 7-25 ONS 2151*1 -iz n o w r# it< -- < rc lc n rr Fliur* 2. emulative percentate of PCS contaminate* dltfIbutlon traaafomen. Seta froa Sample pi an* Saaple *2 coablned. MOMS 2il^2 7-Z6 r r The composition of the population of aubatatlon equipment wan not known and tha Man for Sample *3 wai calculated by trouplnt all sample* to|*th*r without retard to manufacturer or typo of equlpMnt- Haan PCS concentration* calculated for Samples *1 and *3 aro somewhat conserva tive. Laboratory roaultt for idm sample* In Semple *1 worn reported a* "laaa than" a ylvan concactratlon. Thaat value* war* decreased by 0.1 ppm for ana* In calculation* (*.|., < 2 ppm bacoM* 1-0 ppm). For Staple *3 eoncantratlona balow 1 ppa war* reported aa 1 ppa. Clean tb* akawnaaa of tha dlitrlbutlon the Mao la not a toed Indicator of the control tandancy of the data, Altbouth tha Man for laaplt *2 la ,* ppa, on* la uch aor* likely to find a aaapl* below thla loval than above It. Median* and aodat war* alto calculated for each aaapl* to aor* coaplataly daacrlb* the data <** Table 2), If tha data froa the dlatrlbutlon trantforMra aaapl* are com bined, tha arnea laval of eontaalaatlon la S.3 ppa and the aodlaa la 0.1 ppa. To lnvaitliat* dlffaroneai In PCS eontaalaatlon la different type* of aubatatlon aqulpMnt, Sample *3 wai divided Into a tenoral eatotorlai of equlpMnt; 11 TranafotMra Ineludlo* power traaafocMra, tap chantara, currant tranaforaora and a few raectora. 2) Circuit braakora and aiaoelated currant traniforMra. 3) tesulatora 4} Oil Storat* tanka A few alecallaanoeo piece* of equlpMat could not be plaeed lato aey of theeo eatatorleo and war* left out of thla aaalyala. Simmery atatlatlca for each troup war* calculated aad era ehown In Table 3. ketulatori My be more often contaminated with hither level* of PCB then ethar aubatatlon equlpMnt. Horn samples aro currently bolnt collected to determine the ostont of eontamlnatloo In thla type of equlpMit. HONS 21*143 7-27 Table 2 SUNIAKY STATISTICS PCS THK tKKIK PCS SAMPLES Description "ffr*1* Distribution Transformers Staple Site Meet PCI Adjusted Keee Stenderd Deviation Skewness Median Mode SSI a. a .i is. a.a i.t 0.0 samole ss Distribution Transformers ait t. 7 t.t 31.2 7.2 o.s 0.0 s--i. m Substation Equipment 1,070 12.1 -- $2.2 12.1 2.0 1.0 Proportion 30 ppa PCI 3.5t *51 Confidence Interval for Adjusted Proportion i so ppm pcs s.n t 2.it 4.2t .n i.at 3.at 3.at * i.it Table 3 PCI CONTAMINATION IN SUISTATION EQUIPMENT Transformer* Number Sampled Kean PCI Standard Deviation Skavnese Median Proportion $0 ppm PCI *10 11.3 sa 13 2 3.01 ioa 10.1 1S.1 2.7 s 3.St 34 42.8 77.5 2.3 S.S 24.01 14 1.4 10.3 1.3 3.S o.ot 7-28 * MON 5 215144 wV Diffaraneaa Utwt.li HaautacturtfJ A comperlaon of sample atatlatlca tod a chl-aquar* anal?*la Indicated that there war* algnlfleant dlffaraacai between manufacturer*. A chl-aquar* anal?*la we* performed on dlatributioa tranaformer data ha asaalaa tha diffaraneaa la tha cumber of contaminated unit* batwaan ammufaaturar*. Call fraquoncla* vara tparae whan contamination waa daflnad na >0 ppm, rendering tha raanlta of tha taat unrallabla. Howavar, aa tha eat point naad to daflna contamination we* lowered (a.g., to 2} or 10 ppm), tha analjala lndlcatad algnlflcant diffaraneaa In contamination batwaan manufaeturara. Proportion Contaminated at Concentration! Over SO ppm Contamination of matarlala with PCBa at lavala greater than or equal to 50 ppm la of regulator? algnlflcaace. It waa therefore of Interert to attempt to eatlmate tha proportloa of tha compear'a total Inventor? of mineral oil-filled equipment which waa contaminated at thla level. The accnrac? of thaaa eetlmatee la dependent upon tha aampllag mathoda which ware uaad la each of the throe curve?*. Samplaa #1 and P2 provide tha boat aatlmataa of contamination In the dlatributioa traaaformavo alace tha compoaltlon of both tha population and aamplaa b? manufacturer could be daaerlbad occuretel?. To calculate thla eatlmate tha proportion of unlta which warn contaminated at lavala greater than or equal to $0 ppm worn tabulated for each manufacturer. Aa adjuatad eatlmate of tha proportion of the population contaminated at thla level and a atandard deviation ware calaulated ualng tha mathoda deacrlbed b? Coahraa for atratlfled aamplaa (2). Thaaa aatlmataa for gamplet *1 and #2 era 3.3 2.If and 4.3t 1.6t, raapaatlval? (93 pareaat confldanca limit*). tampla #3 provide* tha onl? avallabla data for aatlmatlag contamlaatloa In tha aubatatlon equipment population. Howavar, thaaa aatlmataa meat be made with caution bacauaa of the method* b? which tha aempla waa collected. On* can calculate a proportional aatlmat* of contamination over 50 ppm and than eatlmate tha uncartalnt? around thla aatlmat* ualag a binomial dlatributioa. Tha aatlmat* of contamination la 3.A pareaat which la In var? cloae agreement with the aatlmataa from tha other two aamplaa. Thaaa aatlmataa of contamination arc ahowa In Table 2. MOWS 2151%5 7-29 DISCUSSION Although tho thro* sets of mineral oil samples were collected uolng widely different technique*, the remit* of the enely*e* ere In general agreement with respect to the shape of the frequency distribution of PCI contamination and the proportion of unit* contaminated at level* above SO ppm. Caution must be exercised when attempting to u*e tha data In the** three samples to extrapolate to the Company'* entire Inventory of mineral oil-f Hied equipment er to equipment used by other utilities. When extrapolating to the total Inventory of equipment within the company, the potential sampling bias la osch sample must be considered. In Sample si, there may be some raeaon why transformers la the storage and service facility are not representative of the population with regard to PCI contamlaatlon. Sample #2 may have similar problem* bocauss all of the transformer* surveyed wore need In similar types of facilities (food and feed faallltles). However, Interviews with distribution system engineers did not uncover any reasons why such biases should exist. If these potential sampling biases caa be Ignored, we estimate the median level of contamination to be In the rang* ot 0,5 to 1.9 ppm and the proportion of units contaminated over 50 ppm to range from 1.2 percent to 5.9 percent (95 per cent confidence limits). Tha potantlal for bias In the substation equipment sample Is lerg* due to the sample selection methods. However, lack of Information on the characteristics of the substation equipment population precludes us from verifying or analysing these potential biases. A comparison of the results of this sample with the results of the first two Indicate* a good deal of consistency. It Is uncertain whether these results are applloable to other utilities. Since there appear to be significant dlfferoncoa among manufacturers, other utilities with a different mix of equipment from various manufacturer* would not be com parable. Perhaps the most Important problem In extrapolating these results to other utilities Is that of different malntenanae proaedures and the probability for cross contamination between Askarel and mineral oil equipment. According to geg maintenance procedures, mineral oil transformer* hove never had Askarel added to them to roplaae fluid losses. In addition, fluids from Askarel equlpmoat and mineral oil equlpawnt are net stored la the soma storage tanks. The proportion of Asfcarel-fHied equipment Is very small at gCK sad the possibility of acci dental contamlaatlon Is therefore reduced. Other otHit lee with a larger per centage of PCg equipment or which use different maintenance procedures may have a 7-30 HONS 215146 f r greater degree of PCI contamination la trenaalaslon and distribution equipment. It should bo noted that althosth tha aubatatlon sample was tha largest, It provldad tha laaat Information for extrapolation bacauaa of tba aan'.er In which tha samples wara obtained- Bach largo piece of aubatatlon equlpMnt la sampled ant 11 tha antlra population (or ooat of It) haa boon Maaurod. Little attantloa ha a boon *Ivan to tba order In which tha equipment baa boon stapled or the repratantatlvanoaa of tha eaaple at any point In time. Since tha potential tor (election blaa la large, any eatloatoa derived froa this sample mutt be need with caution until tha sampling hat boon completed. Thla type of Information hat bean utaful to SCB In Interactions with State regu latory aganclaa ratponalbla for haiardout waata dltpoaal and apt 11 cleanup proeaduroa It hat alto proved utaful In predicting potential regulatory lapaett reaultlng froa the lowering of the definition of PCS contamination below tha current ftdaral lovol of 50 ppa. IKPItnCKS 1. H. D. ceperatoln, t. J. Gordon and E. J. Faadar. "PCB Contamination la Dis tribution Transformers." Journal of Invlrweatil Science and Health AX7(2), 241-251 U9B2). 2. w. 6. Cochran nammllna techniques. third edition, 1977, John Wiley and font. 3. K. J- Cordon, J. delta and E. J. feeder. "Data nilnet Ion of Polychlorinated Blphanylt In Trantforoar Oils by Capillary Gat Chromatography.'' AatmUil gftwiltfy folnma 34, pp. 47B-4S1 (1912). DISCUSSION QUESTION 1. ANSWER) Do any other utilities have statistical data on their equipment? Several company repraaantativas indicated approximately the earns, i.e., 4% greater than 50 ppm (The atatlatica quoted were generally accepted). 2question . how did you find taat accuracy from different labe? ANSWER) Below 100 ppm, vary good agreement but not above. QUESTION 3. Any special procedure need for settling the tank? ANSWER) Ho. MONS 151*7 7-31 TRANSFORMER RISK ASSESSMENT AND OPTION ANALYSIS Patrick R. Herbert High Voltage Maintenance Corp. A Transformer Risk Assessment and Option Analysis Program is the first step in developing an action plan to cope with the po tential problems involved with the ownership and operation of PCB or PCS contaminated trans formes. Implementation of an inspection program utilizing the check list, approach sa outlined in this paper will provide an abjective viewpoint aa to the relative degree of risk with each individual transformer m your Electrical Distribution system. Action priorities can thw be developed to expediently implement a prac tical and economical approach to miniirize PCS transformer lia bilities. The inspection check list should be conpletad by only one person cn all wits within a plant or corporation so that a stand ardization of reoanmandations will result. The procedure of checking each applicable content under all topics and the substi tution of the nunerlcai values of each ucmimt into the equation at the bottom of page three will provide a risk value that cut then be evaluated by ecnparing it with the nunerical parameters an page one. The values assigned in this section have been developed through the ccnprehansive analysis of several thousand tr*tafar mers in various locations throughout the United States. 7-33 HONS 2151*8 f The option analysis on page four will point out the signifi cant characteristics of the inspected ixut that will assist with the decision to retain, replace, or retrofill the transformer. The Cost Estimate Outline on this page will strongly influence this decision. 7-34 MOWS 2151^9 PROTECT* TNFPBCTICNtt DATC TRAKSTORMFHt PCS KiSX ASSESSMENT Client__________________________________ _ .. Address . -- Transformer Identification . ___ Transformer location _. Transformer ManufacturerKVA Phase Serial Mo.ClassTypeVoltage fluid Vblune/gallons Fluid TyneFluid Weight___ Transformer Total Weight Length. ___ width weight Bottm Valve P/V Gauge Taro. Gauge Level Gauge_______ Ttop Valve Pressure Release Tap Changer Top 5idc_____ bushings Tbp Side Throat Transition Ccmartrrent _____ Bushings Gasketed Welded Marfarane lop weldedBoltedInspection Cover Fans Yes Mo Automatic Control Yea No_______ WAfRICAL RISK EVALUATION 401- 800 y>i-isoo 1MUMQQ >3400 Good - Inspect and maintain periodically Minimal Risk - Inspect Weekly otly.___ Potential Hazard - Inspect Weekly otly.___ Major Risk - Corrective maintenance required Immediate Problem - Repair Znrnediately jfalala.QamgiM= Inspection by:Analysis ty:. 7-35 , HONS 215150 irassmF wfianqgTQN ( 0 ) Non POJ (determined by analysis) A ( 1 ) PCS Contaminated (50 to 500 ppm) by analysis or nameplate ( 2 > PCB Filled t greater than 500 ppm) by analysis or nameplate TRANSFORMS? PAD IHTEGRITY ( 1 > Pad haii no drains, cracks or conduit openings ( 5 ) Pad has drain, cracks or conduit openings B ( 2 ) Pad ccrplexity would be difficult to decontaminate ( 3) Porous concrete bloc* or wood floor ( 5) Underground ped with snip purp LIQUID LEVEL ( 1 ) Full (as indicated an liquid level gauge) C ( 2 ) Slightly low ( 3 ) Very low res, yams auwniY ( 1 ) Leas than ten (10) gallons D ( 4 ) More than ten (10) gallons but less thanone hundred (100) gallons ( 8 ) More than one hundred (100) gallons but lessthan five hundred (500) gallons ( 10 ) More than five hundred (500) gallons FACTUTV TYPE ( 1 ) Nbn-food processing facility E ( 5) Fbod processing facility ncn-critical area ( 20) Food processing facility critical area ( 10 ) ( 20 ) ( 20 ) F( 2) { 55 ( 8) ( 10 ) ( 15 ) CTMBFCRMBt LOCATION Indoor vault or isolated area (minimal lea* hazard) Indoor people frequented area Indoor iper floor, balcony or roof Outdoor in locked vault Outdoor on open pad Outdoor in indergrotnd vault Outdoor cn pole Outdoor rooftop location tPSEXEEPIHS ( 1 ) Transformer locwticn free of debris and storage G ( 2 ) Transformer location adjacent to debris or storage ( 3 ) Traneformer adjacent to potentially damaging machinery SAFETY ( l ) Proper fire extinguisher available at transformer ( 1 ) Automatic Fire EKtinguiahlng System cn Transformer H ( 2 ) Fire extinguisher tnavailable at transformer ( 1 ) Protective clothing and breathing equipment available ( 2 ) Protective clothing and breathing equipment unavailable ( 1) ( 3) 0( 6) ( 10 ) acNTMFanir Trwnformer diked to contain total liquid Dike will not contain total liquid Dike deteriorated or danaged but functional No dike or contaminant SPILL RESPONSE K ( 1 ) Spill dean up dnans, rags, absorbent available at transformer ( 5 ) No spill clean i*> material available 7-36 MONS 215151 r r VW afiqHBRPY ( 1 ) Non susceptible to damage (vault or elevated platform) - ( 3 ) Minor - some protection provided (fence or curb) ( 6 ) Moderate - vehicular traffic damage possible ( 8 ) Highly susceptible - accidental damage risk TRANSFORMER VPTT SYSTEM ( 1 ) Vents - fires outside away frtn windows, food, people, water m ( 3 ) Vents - fires outside to possible critical areas ( 5 I Vents into applied buildings gmj. nrnsBnumc5 If nejor spill occurred, the liquid would ( 1 ) be totally confined or contained ( 3 ) contaminate pad and transformer only { S ) contaminate soil, gravel or other material ,, { 10 > contaminate processing or manufacturing area ( 30 ) contaminate floor bela* ( 20 ) contaminate occupied areas ( 20 ) contaminate water drains, ditches, sewers, etc. ( 20 ) contaminate bodies of water HttNSfqR leak status ( ) Treeformer not leaking ( ) Traneformer fins or tubes mstina ( ) No active leak but old stains an pad or casing ( ) New samll stains noted under valves ( ) New tmall stains noted inder bushings ( ) Large leak residual requiring immediate action ( ) Moderate leak requiring drip pan or absorbent ( ) Active spill contaminating pad and axes TSCA REMATICW OCM>LIAMCE ( ) Not required ( ) Labelled properly ( ) Labelling requited but absent ( ) Leak inspection required (weekly) (Quarterly) (not required) ( ) Inspection required (weekly) (quarterly) (not required) HTftT RFJfiftfiT Trsaformer to wall distance FR_ wall ttoteriai_____________ ______________ Transformer to celling distance_________ Ceiling Material Transformer Length width Dike Dimensions Lanoth width R. Height. RISK ASSESSMENT FORMULA A(B + G+H + L + M)+D(B + L) + 0 + F(G+H) +(A + D+F) <J + V + M2 + N) 7-37 MQNS 215152 RETAIN VS. RS'LACEMmr VS. RETROFILL Condition (Damaged) {Rusted) (Dctremely Dirty) (Average) (Clean)------- -------Is unit now leaking? (Yea) (No) (Possibly)---------- ------- ---------------- ---- Does location require high fire point liquid transformer (Yea) (No)-----Is location a potential environmental pollution risk? (Minor) (Major) -- Is location a potential risk to personnel? (No) (Minor) (h%Jor) - --------Is location (Wet) (Extremely Hot) (corrosive) (Extremely Dirty) (Dry)---- Is spill retention dike now in plaoe? (Yes) (No) (Inadequate)---------- -- Esee of physical replacement: (Easy) (Ocrrplicated) (Difficult) ------ Transformer characteristics (Standard) (Uhusual) (One of a Kind)------- -- Is transformer in food product sensitive location? (Yes) (Mo)---------------Client sensitive to adverse PCD publicity: (Low) (Average) (None)--------- Transformer bushing type: (Gasketed) (Non-Gasketed)---- -------------- `Transformer bushing location: (Top) (Sidewall)------------------------ ------------- Age Of transformer (5) (10) (20) (30) (40+)--------------------------------------------------History (No Problems) (Unit Rebuilt) (Existing (Jhexplsined Prohlan) -------Allowable downtime: (None) (8 hours) (24 hours) (Uhllmitad)---------------------Is a spare replacensnt available? (Yes) (Possibly) (No) -- Is transformer oarpatitle with future requirements? (Yea) (No)-------------Anticipated or desired rereining life: (5 yeare) (10 years) (20+ years) Transformer owned by: (Client) (Utility) (Other)--------------------------------------Transformer average tsiparatura: (Normal) (overheating) (Extremely Hot) Electrical testing program (Yearly) (2+ Years) (Nona)---------------------------Recent electrical test evaluation: (Poor) (Deteriorated) (Average) (Good) Leak inspection program: (Weekly) (Monthly) (Ouarterlv)------------ - -- COST ESTDA2E TO: Retain (Fineial liability - Minor - h%jor - Excessive) Remove, disposal and reolaoensnt Retrofill withFluid Rauomnandatlon Analysis By. 7-38 Hate. 4 MQNS 215153 r r SUN0H10 PCD DETOXIFICATION EXPERIENCE BT M. JAMES KOZAK, V. P. OPERATIONS EPRI CONFERENCE, DECEMBER 8, 19B3 I want to focus it on what we consider four key areas - quality, operational efficiency, environmental compatibility and reclassification capabilities. Quality is the heart of the Sunohlo process. We sell a unique service and that service Is the decontamination of PCB's In transformer mineral oil. However, if all we did was decontaminate PCB's. we would be the same as all the rest of the boys on the block. Our primary difference Is that we produce a decontaminated mineral oil of such Impeccable quellty that It can Instantly be reused In high voltage power equipment. We have come a long way In establishing a reputation as being a quality leader. We have now completed over 300 contracts and have treated over three million gallons of dielectric fluids. Part of our quality control program is rigid specifications and Intensive and accurate testing. Our field specifications are written Into and warranted by each and every contract. We have had our share of bumps and bruises while developing a sound quality assurance program. With five PCBX rigs and up to three shifts on each rig, ft's easy to understand why It would be difficult to obtain consistent field tasting results. When we first started, we had our share of oil quality problems, especially in tha area of high moisture content and low Interfaclal tension. These were due to extended use of the earth beds and Improper operation of the vacuum degasser. Our problems were complicated because we were getting Inaccurate testing. Our field technicians were analyzing the oil as good quality; but when the samples came back to the home office, the quality was unacceptable. With a little bit of a diagnostic effort wc soon resolved that the problems were related to sampling techniques and improper operation of the field testing equipment. With proper field Information we were then able to zero In and correct our operational procedures. HONS 216154 7-39 We took two steps to resolve the problems. First) we bit the bullet and told the customers Involved that there was a quality problem and returned at our own expense to reprocess the oil and make good on Our contractual obligations. Our coimitment to our contractual guarantee is Ironclad. Secondly, we re-evaluated our quality assurance program and took the proper steps to Improve field testing and to Implement a system which would tolerate a degree of test error In the field. Under our present quality assurance program, we utilize a fully equipped mobile laboratory for each PC6X rig. We monitor each and every transformer for both PCS level and electrical properties throughout the detoxification cycle and when we are satisfied, we have achieved completion, we draw a final set of samples, one of which we analyze in the field, a second of which we send for backup verifica tion at our In-house laboratory In ttavarre, Ohio and a third set which we will provide to the customer If he so desires. We utilize our In-house laboratory as our standard In evaluating our performance on a contract. To Insure that field operations meet contract specifications, we utilize a two tiered set of standards. (Exhibit 1 - comparison of oil quality specifications). By setting the field standards significantly higher than the contract specifications, we allow our selves a margin of test error when comparing field results with the final evalua tion of the home office laboratory. In evaluating PCB content, we use a similar duplicate testing structure. PCB standards are purchased from nationally known reputable sources and samples are analyzed on two different gas chromatographs. Field operations utilize a Varian 3700 Electron Capture Gas Chromatograph and these results are verified using a different set of standards with a different laboratory technician on a Hewlett Packard 5880 Electron Capture Gas Chromatograph In our Naverre laboratory. Our technique of employing a higher-than-required standard for field testing and a duplicate sample analysis is both time consuming and expensive and It has resulted In our returning to a few job sites to do warranty work, but overall It has resulted In en extremely Impressive record of high quality performance. We have come a long way In our detoxification capabilities and In the number of EPA regions and states In which we have been licensed to operate. When we first bagan commercial operations In the fall of 19B1, we had EPA approval In only two of the ten regions and we were limited to decontaminate oils with a maximum of 500 PPM of PCB's. By May of 1982, we had obtained approval in all ten EPA regions and by August of 1982 we had increased our detoxification license limit to 2500 PPM. While we technically have the capability to decontaminate in excess of 10,000PPM ... a 2500 PPM level is In concert with competitive economics. 7-4 0 HONS 21S15S We have experienced occasional difficulties with detoxification, in the early days, we had some significant problems with various aroclors and with various properties within the oil. 1 can recall a Job In early 1982 In which we were processing on a rather large transformer contaminated to a level of about 3000 PPM. I believe it was Aroclor 1242. It took us about 12 hours to make a pass on this transformer and In the first pass we reduced the contamination level to about 2000 PPM. For the next five days, we continued to process while the trensformer remained at the 2000 PPM level. We made a variety of process changes and modifi cations and tried alternate detoxification and filtration steps and slowly we began to reduce the contamination level. In all total, It took us over two weeks to fully decontaminate this transformer. Over the two plus years that we have been in operation, we have learned how to handle these types of problems. We have learned which adjustments to make for various aroclors and we have learned which adjustments to make for differing oil quality specifications. Today that transformer would have been completed In two to three days. Lately there has been a governmental policy shift to give the states more authority and responsibility In handling their own environmental affairs. Each state has or Is now developing. Its own permitting regulations. Each state now requires a separate application ranging from a simple submission of an EPA license to a com plete engineering and environmental Impact study. Managing field operations In 50 states Is a rather Interesting process. To complicate matters, California, for Instance, has 26 air pollution control districts which operate rather autonomously from the State Air Pollution Board. Even though It's not 1984 yet, the big brother Is always watching us. Our business Is like life In a goldfish bowl. The exhaust stream from the vacuun degasser was of major concern to several states which have active air pollution boards. Our first trip to California was rather Interesting. At that time, we had no control devices on the emission stream and the exhaust from the vacuus degasser looked very similar to the exhaust from a car on a cold day. A young engineer from the air pollution district came to visit the Job site with some rather unsophisticated testing equipment. He got very excited when he saw this exhaust stream and he took his hydrocarbon analyzer and stuck It up the exhaust pipe and said ... "Gee, you've got a lot of hydrocarbons here." I think It pegged his meter. Then he took his Opacity Meter out and tested It and said ... "It's more than 20X on the opacity scale and since It has hydrocarbons In It, you are In total violation of our standards and you must shutdown 7-41 HONS 215156 Immediately." Well, we quickly called the director of the district who recognized this was a demonstration run and since the exhaust stream was less than 1/10 of one pound per hour, determined that we could continue operations. He did, however. Inform us that we may be subject to a fine. The young engineer returned the next day wanting to collect some grab samples so that he could have those analyzed. He came armed with what appeared to be two Hefty trash bags and he proceeded to wrap these around our exhaust pipe until they blew up like a big balloon. He sealed these and carted them off to his laboratory. A month later we were told that if we wanted to come back Into that district, we would have to do something about the vinyl chloride which we were producing In our exhaust stream. If you have any familiarity with organic chemistry, you know It's pretty damn difficult to produce a vinyl chloride from a PCB contaminated mineral oil stream. However, It's not too difficult to pick up trace amounts of It from a vinyl trash bag. The next time we came Into California, we were armed with demisters and charcoal filters on our vacuum degasser exhaust stream. We Invited the Los Angeles Olstrlct Air Pollution Board, the California Air Pollution Board and an Independent testing organization to evaluate the air emissions from this vacuum exhaust. Of particular concern were emissions of PCB's, dioxins, dlbenzofurans and, of course, to satisfy the young engineer In another district, vinyl chloride. This second testing was considerably more formal than the early adventure. Each of the testing groups had a specific protocol and each team used the latest scientific sampling devices and were manned by trained engineers and laboratory technicians. They took sampling probes and monitored the exhaust both before and after the demisters and charcoal filters. These samples were then analyzed by three different laboratories. The Independent tests proved conclusively that the PCBX process emits no PCB's, no dioxins, no dlbenzofurans and no vinyl chloride. This Information was critical In allowing us to meet the permitting requirements of several states Including California, Texas and Michigan, We've stated that we can produce an oil of excellent quality, that we can legally and economically detoxify PCB's up to the 2500 PPM level, that we have no negative environmental Impact and that we have been accepted In 48 of the 50 states; but what about our ability to reclassify transformers? That's really the height of the matter Isn't It? Well, Just like the rest of the things we've talked about today, this has been a learning and maturing process for us. At the onset, we had great speculation about leach rates. We talked about reclassifying transformers In the 5000 PPM range. 7-42 HONS 215157 But we learned our lessons quickly. We have now developed a strong track record of enabling transformers to be reclassified. We guarantee that by using the PCBX process on a transformer of 1000 PPM or less, we will certify that the oil In the transformer will be less than 50 7PM after the EPA required operating time and temperature and that the owner of the transformer will then be able to reclassify It. Sometimes, when one of our salesman wants to sell a job very badly, we will stretch the limit of that guarantee and we have even been successful In doing that. How do we do 1t7 We carefully circulate on the transformer at a flow rate designed to provide maximum washing of the core and coll with an absolute minimum of turbu lence. We control the leach back by setting a PPM level for the oil leaving the bottom of the transformer and entering the PCBX unit. Although the oil leaving our unit has to be below 2 PPM, it Is critical to set an Incoming standard. The standard used Is dependent upon the original contamination, the type, site and age of the transformer. Normally, the standard Is the 6 to 10 PPM range and then we further insure costlflcatlon by continuing to process for an additional speci fied number of gallons before finally signing off on the transformer. The only way you can guarantee the customer's ability to reclassify a transformer at 10(30 PPM is to have a leach back rate that Is less than 53. That's about half the usual rate for retroflll. During the past two years we have reclassified over 400 transformers and have demonstrated an average leach back of less than 23. 1 want to make one point very clear. We define leach back rate as the Increase contamination level that has occurred after PCBXIng versus the Initial contamina tion level. (Exhibit II). Assume you have a transformer with an Initial contami nation level of 1000 PPM. Upon completion of the PCBX process, this transformer has been decontaminated to a level of 6 PPM. Upon resampling It after 90 days, the transformer shows a contamination level of 33 PPM. The increase In contamina tion level has been 25 PPM. therefore, the leach back rate Is 25 PPM divided by the Initial concentration of 1000 PPM ... or 2.53. I'd like to share with you the experience that we've gained over the past 24 years In estimating leach beck characteristics. (Exhibit III) This chart represents the leach back rates for all of the transformers that we have processed from January 1, 1981 through the first quarter of 1983 ... 344 transformers In total. I've divided the transformers Into four groups based upon their Initial contamina tion level. As you can see, the mean leach back rate for each group of trans formers is approximately 23. Extending this data Into a probability distribution we tee that for transformers under an Initial contamination level of 250 PPM, 983 7-43 ' MOMS 215158 9* of these have leach beck rates of less than lot. In the more critical groups, our performance Is more dramatic. 9BX of the transformers In the 250 to 1000 PPM group have leach back rates of less than 5*-, while 9BX of those above 1000 PPM yielded rates of less than 3X. Let me close by emphasizing that Sunohlo Is In the restoration business, wt restore the oil to a completely reusable condition and we restore the transformer to a non*PCB status, assuming It was originally In the 1000 PPM range. Restoring askarel transformers to a non-PCB status Is a formidable task, but when a working system Is announced, you just may find Sunohlo Involved. Thank you. 7-44 MOMS 21159 KEUT. MtMEA Asm 0-971 I.F.T. ASTH 0-971 DIELECTRIC STAEN6TH ASTH 0-A77 POWER FACTOR ASTH 0-924 9 2SC Exhibit I OIL QUALITY SPECIFICATIONS SUNOHIO FIELD SPECIFICATIONS .02 KAI 37 KIN 3S HIM .083 KAX NORMAL S4MOKIO CONTRACT SPECIFICATIONS .04 HA* 33 HIM 30 HIN -10S MAX S-| MOMS 21S160 Exhibit II LEACH PACK RATE COMPUTATION Initial ContMliutlon lerel Jvedlttely After PCM Retest *t to thy* Increase Since POST Percent Leech Sect 25 1000 1000 PP PP" 33 PP 25 PPM 2.St 9>-t MOMS 2151*1 Exhibit III LEACH SACK STATISTICS No. of Units Httn Roto MH Probability m Probability SO - 249 PPM 182 2.2* 6.91 10.SI 250 499 PPM 66 1.21 2.21 4.21 500 999 PPM 1000 - 1500 PPM TOTAL 20 1.41 26 1.01 344 1.8* 3.11 1.91 5.8* 4.21 2.61 4 01 moms 219162 DESIGN OF A FIXEO PC8 STORAGE FACILITY TO PRODUCE THE MOST VALUE FOR CLEANED OIL Dr. Louis Centofantl PPM, Inc. PPM, Inc. Is a waste management firm speclsllilng in PCB handling, chemical destruction and testing. PPM has offices and facilities In Kansas City, Missouri; Atlanta, Georgia; and Canada. PPM`s laboratory, located In Atlanta, Is equipped with the most modern equipment available to measure PCBs and other hazardous contaminants In oil, air and water. PPM, Inc. has incorporated Its chemical destruction process Into a mobile decon tamination unit that can destroy PCBs on the generator's site, thereby avoiding transportation liabilities. Mobile units are operating both in the U.S. and Canada.- During several of our utility clean-ups, we have worked with utilities to evaluate, design and site treatment facilities. This work has involved initial planning and evaluation of all disposal options and the optimum facility for such. A decision on the type of disposal used has a significant impact on the design of the utility's storage facility. Olsposal options available to utilities Include both on-site and off-site treat ment or destruction. On-site treatment Includes chemical destruction or the use of utility hlgh-efflclency boilers. Dff-sitt options Include chemical destruc tion, Incinerators or landfilling. In evaluating disposal options and facilities, general considerations should include: 1, Short- and long-term liability 2, Reuse of the decontaminated oil 3, Reliability of the disposal firm A. Economics. Both short- and long-term liabilities must be considered when reviewing options. 7-49 MOMS 2llt3 The most serious long-term liability exists with non-destruction options, especIt)ly landfills. PPM, inc. views landfills as nothing more than long-term stor age. Although It Is legal to use landfills for the disposal of PCS contaminated wastes, we recommend to our clients that only destruction options be used, since the generator may be required In the future to pay to have the landfill cleaned. Short-term liabilities exist for all options. The most serious off-sit* liabil ity exists with transportation of the hazardous material to the disposal facili ties. Leaky tankers, spills and accidents during transit can produce a signifi cant risk to the generator. This liability does not exist for on-site treatment. Liabilities do exist for on-site treatment and Include spills, accidents, etc. Liabilities associated with these are significantly less since spills can be easily contained and cleaned up when on site. Two other non-economlc Issues that should be considered are the eventual use of the decontaminated oil and the reliability of the disposal firm. If the oil is chemically decontaminated, the oil can be used either In transformers or as a fuel oil. Many users have raised concerns about the quality of chemically repro cessed oil. The properties of PPM's processed oil are shown below. PPM Transformer Oil Properties PPM Treated Oil Dielectric Strength, KV Color. ASTM Acidity mg KOH/g Specific Gravity Interfacial Tension Power 25*C Factor 100*C PCS Level 38.8 1.5 .045 .883 45,3 .016 .946 <2 ppm Suggested Limits >26 2 max .26 max .86 - .91 20 1.0 max 6.0 max He believe that the oil can be safely reused In transformers. The most Important factor working against reus* of transformer oil at the present time Is economics. Producing fuel oil from contaminated oil Is a more economic rout* for most utilities than going to transformer oil. 7rS0 HONS 21fl64 DISPOSAL ECONOMICS The most Important factor in determining the disposal option and facility design Is the economics of the various options. To demonstrate the economic options available to a generator, we have Included two very simple cases and generated the economics of each. In reality, the economics can vary significantly depending upon a variety of parameters. Case 1. 50.000 gallons of contaminated oil 300 ppm PCS 600 miles to nearest disposal facility (transportsIon costs of 13.25/mlle) Case 2. 6.000 gallons of contaminated oil 300 ppm PCS 600 miles to nearest disposal facility (transportation costs of 13,25/mile) ON-SITE DISPOSAL Three options are available to a generator for on-site disposal of the waste. These are: high-efficiency boilers, chemical destruction producing a fuel oil, and chemical destruction producing a transformer oil. On-site Incineration Is allowed In EPA-approved Mgh-afflclency boilers. Initial cost estimates Indicate that burning In these boilers Is the most economical method of disposal; however, most utilities have avoided this option. Public opposition to power plants becom ing hazardous waste disposal facilities 1$ Intense. Many utilities that have attempted taking this route have found that public opposition and ensuing ex penses to defend the burning rapidly make the project economically Infeasible. Cost estimetes to the generator to chemically treat PCS oils are summarized below: Chemical treatment (fuel oil) Fuel value Total 50.000(Jg/gaalloi)ns 1.70 .60 7-51 6,000 gallons "lJ7gair 2.50 U901 1.60 HONS 219165 / Chemical treatment (transformer oil product) Transformer oil value Total 50.000 gallons (JVU - 2.80 (1.60) 1.20 6.000 gallons Wgal) " 3.50 (1.60) 1.90 OFF-SITE DISPOSAL We recommend two options for off-site disposal: incineration. The costs, listed below, are similar. chemical treatment or Oil disposal (FOB facility) Transportation Total $1.10 .33 SI.43 Jn this option the materials are moved by tanker In 6,000-gallon lots. Trans portation costs are calculated at S3.25 per loaded mile, assuming 600 miles to the facility. Thus, of these two Idealized cases. If economics Is the major deciding factor, off-site disposal is preferred for the 6,000 gallons of oil, while on-site disposal producing a fuel oil product Is the preferred disposal option for the 50,000 gallons of oil. The various options are summarized below, and the preferred costs are underlIned. Disposal Cost Off-site On-site (fuel oil) On-site (transformer oil) 50,000 gallons SI.43 .80 1.20 STORAGE FACILITY After a decision has been made on the best method of disposal for PCS materials, a facility should be designed that best suites that option. Betides a facility, procedures should be Instituted to handle and store not only PCBs but alto other solvents and other wastes. Proper housekeeping can greatly reduce the size of the PCB problem by avoiding mixing of non-contamlnoted materials with contami nated ones. MONS 215166 7-S2 r r The itorege fac 111 ties should Include harmed storage tanks that, for off-site dis posal, hold approximately 10,000 gallons; for on-site disposal, these tanks should hold a minima* of 30,000 gallons. We found that the majority of spills of PCfi oils occur when loading and unloading tankers. We therefore recomnd that the loading and unloading area also be bermed, If on-site treatment Is antici pated, an operational pad should also be planned. A design we use that Incorporates all these features Is shown below; L Figure 1. PC6 Storage Design Forty-eight thousand gallons of tank storage are provided In the four 12,000-gal lon tanks. The tankers and the decontamination unit drive Into the diked area over the ramp. All loading and unloading Is done In the diked area. A chemical treatment unit should also sit In the diked area while treating. This would provide maximum protection against spills. This dtslgn was slightly modified and used for a recent clean-up at Virginia Electric Power Company (VEPCO). Five hundred thousand gallons of oil were treated by one of PPH's mobile units. A be rued operational pad 40 feat by 60 feet was constructed to hold two storage tanks, tankers and processing equipment. The unit processed 9,000 gallons a day of oil, containing approximately 130 ppm of PCBs. 7-53 MOWS 215167 tinkers -4- tank 1 drier i 12,000 gel. tank oil 6,000 4--......... -- gal. tank I decon. unit > sludge (1.01) Figure 2. VEPCO Process Flow TANK DECONTAMINATION After emptying the l,000,000*galIon storage tank, VEPCO was faced with the problem of decontaminating the tank. The EPA-approved procedure would require three flushes with 10 percent of the volim of the tank. This would require a minimum of 100,000 gallons of clean oil that would be considered contaminated. PPM was successful In receiving from EPA approval for an alternate procedure that greatly reduced the volim of oil needed In the flushing. The procedure Included first removing all liquid and sludge from the tank, producing a "dry 001100" tank. The tank was then flushed with 10.000 gallons of oil until the PCB levels for two- subsequent washes ware unchanged. This was done for a minimum of three washes. Our tests showed no changes after three washes with levels at 3 ppm. This oil was removed and the tank then rinsed with another 10,000 gallons of oil. The flushing and rinsing oil was treated as PCB contaminated oil even though it was less than 60 ppm. The procedure greatly reduced the volim of oil necessary to decontaminate the tank. After examining the procedure and PCB levels In the flushes, we felt that a better procedure would be to use several smaller volim batches of oil to flush the tank. For example, flushing the tank with five 1,000-gallon batches would have produced the same PCB-level reduction. This procedure was reconmended to EPA for any future large tank clean-ups. 7-54 HONS 215166