Study of Vinyl Chloride Formation at Landfill Sites in California Final Report January 1987 Prepared for the State of California California Air Resources Board under Contracts A4-154-32 and 2311206978 + 4Battelle Pacific Northwest Laboratories ''PUBLIC HEALIN bibsoml BERKELEY LIBRARY UNIVERSITY OF CALIFORNIA LEGAL NOTICE This report was prepared by Battelle as an account of sponsored research activities. Neither Sponsor nor Battelle nor any person acting on behalf of either: MAKES ANY WARRANTY OR REPRESENTATION, EXPRESS OR IMPLIED, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, process, or composition disclosed in this report may not infringe privately owned rights; or Assumes any liabilities with respect to the use of, or damages resulting from the use of, any information, apparatus, process, or composition disclosed in this report. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply endorsement, recommendation, or favoring by Sponsor or Battelle. ''Dube "STUDY OF VINYL CHLORIDE FORMATION AT LANDFILL SITES IN CALIFORNIA" Peter M. Molton Richard T. Hallen John W. Pyne Materials & Chemical Sciences Center Materials & Chemicals Applications Department Chemical Process Development Section \ | BATTELLE, PACIFIC NORTHWEST LABORATORIES | ~ Richland, WA 99352 — January 1987 ''TB 4469.20 oe RC 268 V SuMGY (787 PUBL ''$0272 -101 REPORT DOCUMENTATION | }REPORT NO PAGE | | 2s | 3. Recipient's Accession No. 4. Title ana Suotitie — : - oO | 5. Report Date STUDY OF VINYL CHLORIDE FORMATION AT LANDFILL SITES Published Jan. 15, 1987 IN CALIFORNIA ‘. 7. Author(s) - - a Pertorming Organization Rept. No. P.M. Molton, R.T. Hallen, J.W. Pyne BNWL-2311206978/ %. Pertorming Organization Name and Address 10, Project/Task/Work Unit No. Battelle Pacific Northwest Laboratories | Battelle Boulevard L1..Contract(C) of Grant(G) No. Richland, WA. 99352 iy A4-154-32 (G) | 12 Soonsoring Organization Name and Adacress 13. Type of Report & Period Covered State of California Air Resources Board F 1102 "Q" St., (P.0. Box 2815) July 16, 1985 to Jan. 15, Sacramento, CA. 95812 a T3987; | 1S. Supplementary Notes “16, Abstract (Limit: 200 words) The Purpose Of this study was to determine if vinyl chloride (VC) detect ed in air above California landfills is produced in situ. Previous work indicated biolog ical action on chloroethylenes as the most likely source, although other emission routes are possible. Experiments were performed with N and S California landfill samples and anaerobic digestor sewage sludge. Test materials were incubated with various chlorocar- bons and with 13c- -trichloroethylene (TCE) to confirm biological production of 13¢-vc. Pyrolysis of PVC released residual monomer but no de novo VC production occurred. No VC was produced from TCE by photolysis or from heat- and radiation-sterilized landfill samples. VC formation was highest where landfill material already contained the greatest amounts of organic compounds. These experiments confirmed the biological dechlorination of chloroethylenes as the most likely route for VC emission from landfills, rather than chemical or photochemical routes, or PVC degradation. Leaching from PVC could be a minor source of VC, though there was less than 0.1% (estimated) plastic in the landfill samples, containing at most 330 ppm of VC monomer. A landfill sample known to produce VC was used to start an anaerobic chemostat using methanol as sole carbon source. The enriched culture resulting was homogeneous, and when incubated with 13C-TCE, produced 13c-vC, confirmed by GC/MS. Microbial action on chlorinated solvents dumped in landfills is the most probable source of observed VC formation. Anaerobic methanogenic bateria isolated from landfill samples have been shown to produce VC from TCE under laboratory conditions. 17. Document Analysis a. Descriptors Vinyl Chloride Landfill emissions Chlorinated Solvents Bioconversion Methanogens Anaerobic bacteria Dechlorination b. Identifiers/Open-Ended Terms ce COSATI Fleitd/Group J& Aveilability Statement 19. Security Claas (This Report) 21. No. of Pages Release unlimited Unclassified 132 7. RG ity Clags (This Page) 22. Price classified (See ANSi-Z39.18) See instructions on Reverse OPTIONAL FORM 272 (477) (Formerty NTIS—35) Department of Commerce ''''ABSTRACT Low concentrations of vinyl chloride (VC) gas have been detected in the air above California landfills, even where no material containing it has reportedly ever been dumped. The purpose of this study was to determine if vinyl chloride is produced in situ by natural processes. Review of previous work indicated that biological action on other chlorinated ethylenes present in the landfills was the most likely source of vinyl chloride. Other emission sources are possible, ranging from illegal dumping and leaching to thermal decomposition of plastics. (An evaluation of illegal dumping was beyond the scope of this study.) Experiments were performed with landfill material obtained at depths of 3-16 ft from a Northern and a Southern California landfill, and with control material presumably never exposed to chlorinated compounds (Richland, WA, anaerobic digester sewage sludge). Test materials were incubated with a range of chlorinated compounds and with ad labeled trichloroethylene (TCE) to confirm biological production of VC from other compounds, including TCE. Pyrolysis experiments with PVC showed release of residual monomer but no VC formation. No VC was formed from TCE by photolysis or from heat- and radiation-sterilized landfill materials. VC formation was highest in experiments using the landfill material already containing the greatest amounts of organic compounds (and the most adapted microorganisms), and least with the "uncontaminated" sewage sludge. These experiments confirmed biological production as the most probable formation route for landfill emissions of VC and not chemical or photochemical routes of formation. Leaching from old PVC could be a minor source of VC, although there was less than 0.1% (estimated) plastic in the landfill samples containing at most, 330 ppm of VC monomer. A landfill sample known to produce vinyl chloride was used as a source of microorganisms for initiation of an anaerobic chemostat. Methanol was the only carbon supplied and carbon dioxide and methane were the main products indicating the presence of methanogenic bacteria. The enriched culture was homogeneous when grown on agar based medium under strictly anaerobic conditions in roll-tubes using methanol as the sole carbon source. Under aerobic conditions no colony growth of any kind was observed. When a sample of the enriched culture was incubated with ''13 13 C-labeled TCE in the presence of methanol detected. C-labeled vinyl chloride was These experiments clearly show that microbial action on chlorinated sol- vents in landfills is the most probable source of observed VC formation in situ. Methods for minimizing future VC emissions from landfills are discussed. ''ACKNOWLEDGEMENTS The work of Emcon Associates, San Jose, CA, in obtaining landfill samples under subcontract #B-K8103-C-H to Battelle-Northwest, is gratefully acknowl- edged. This report was submitted in fulfillment of Contract # A4-154-32, "Study of Vinyl Chloride Formation", by Battelle, Pacific Northwest Labora- tories under the sponsorship of the California Air Resources Board. Work was completed as of January 16, 1987. ''DISCLAIMER "The statements and conclusions in this report are those of the con- tractor and not necessarily those of the California Air Resources Board. The mention of commercial products, their source or their use in connection with material reported herein is not to be construed as either an actual or implied endorsement of such products". ''TABLE OF CONTENTS SUMMARY AND CONCLUSIONS cscawcsteansee se cea 0 Oe DURST ED ERR RN KE Te we RECOMMENDATIONS cssccs cc ceee saree ee eset OCTET DETTE CRATE TT DSOT TEES ENTRODUCTION a cevassawesnvan ew ner sen eee end es eS E RRO RRR OMe eee Project Scope and Purpose ........... ceca ewe TT Trt Titre Survey Of PPIO® WOK cesccccacsoeseeseess eee eeeerensesnr eee ene List of Probable Vinyl Chloride Emission SOUPCES cen ccae seed dedns EXPERIMENTAL APPROACH ..... FLCC TN SOON sia ww TTT r rT Equipment Assembly, Calibration, and Procedure Verification ..... Experiments with Primary Sewage Sludge ......ceseccccsccescessess Obtaining Landfill Samples ........ 6 i eee ew Oe Stee ee mie we Description of Site A, Southern California ......... 4 if BERT Oe wee Description of Site B, Northern California .......ccccccecccccees Chemical Analysis of Landfill Materials ....... TTT Tr TTT Terr Microbial Metabolism Experiments .cccescanccssanccesnenseneeee nes Characterization of Vinyl Chloride- “Producing Organisms cssssscses MATERIALS AND METHODS ......cccccccccccccccccccccccccvccsccceesccccece Material Sources, Composition, and Purity .........eee. svenwwwwe Site Sampling Procedures ....ccececcccccsceccceccces 6+ DE GR S Sew ES Laboratory Instrumental Analysis Systems ......... cece eee c ce ccees Chlorocarbon Hydrolysis Experiment .xcnvcccssenevasewenenesis eomen Incubation of Sewage Sludge With Chlorocarbon Substrates ........ Incubation of Landfill Material With Chlorocarbon Substrates .... Incubation of Landfill Material With Labeled Trichloroethylene .. Isolation Experiments for Vinyl Chloride-Producing Organisms .... CONCIUSTON sccvsansmescnewse wns nawe 100 Cw wee Ewe EERO 8 809 omnes EXPERIMENTAL RESULTS ....ccccccccccccccccccccscccccvccccscessccvcececs Landfill] Sampling ReSults.....c.ccccccccccccccvccccccccceccvccees Sample Descriptions and On-Site Measurements pose seuursnee cece eee . Individual Sample Description «scscusenxnasenencnaewans cece eeeee Vinyl Chloride Formation by Chemical Routes ...........e00. TTTTT Hydrolysis Of ChlOYOCATDONS, sscccccsssmcvrsesaesanasemunsan “ PYFOLYSTS EXPETIMGNES scecescascenssassnvvens (CORR NT Photolysis Experiments ...csccccccccccccccccccce Sees eRTE TTS Effect of Steam and Gamma Radiation Sterilization Pbk ESSdsRER ES aR Chemical Analysis of Landfill of Material....... cc cece eee eee c nee Vinyl Chloride Formation by Biological Transformations .......... Chlorocarbon Incubation With Sewage Sludge ...........seseee Chlorocarbon Incubation With Site A Landfill Material ...... Chlorocarbon Incubation With Site B Landfill Material ...... Isolation of Vinyl Chloride-Producing Methanogen ...........eeee- ''DISCUSSION sscss.ccvccees eb cees asec tee eR eR CREE ET SDT EREN SED ET ETS CHES OR 97 Formation of Vinyl] Chloride ..... cc cece cece ccc ccc cc ec cccecccevens 97 Experimental Approach ...cercccccccccccccccccccccssecceeseccesens 98 Emission of Vinyl Chloride from Landfills (AHURA DR eR ER OE RS 103 Possible Remedial Actions and Additional Research ..........ee0e. 104 Achievement of Project Objectives ........cccccccecee Vaeeesuenss 107 Conclusion and Recommendations .....ccccccccccccccccccccccccceces 108 REFERENCES 9 au nisesisise sw einieie see ewiiiieiew ene ele a A WH iNT ES WW 8 HOTS WUT we 109 BIBLIOGRAPHY ssssicsscesnnssccccsimusneses CHAD TTCR DAT TTC TO HHTTES SC OH GETS 111 GLOSSARY OF TERMS, ABBREVIATIONS, AND SYMBOLS ........cccccccccccccces 113 APPENDICES. xis ieseie:ei 0s io isin: ei nie We leis ei alee Wineieisl a a6 688 WNT TM EN ORT TRS BUTTE SS 114 Appendix A: Suppliers and Usage of Major Chlorocarbons ......... 114 Appendix B: American Chemical Society Presentation and Press REIGASE b6 5450555605456 e8 Reese rene ee TERED EeeeNese 116 Appendix C: Inductively Coupled Plasma (ICP) Elemental Analysis PYOCECUTE .ocecec ccc ccccccceccccecececececes sim wi we 127 ''15. 11. 12. 13. 14. LIST OF FIGURES Sorption/Desorption Isotherms in PVC/Vinyl Chloride/Water at 22°C 25 Mechanism of PVC Degradation ..csccoseccnsccrencsncswovccveesnsess 26 Waste Products from Vinyl Chloride Production Process ........ee0-. 28 Landfill Site A, Southern California ....ccsnscsssscsssecscrscvevs 45 Landfill Site B, Northern California. cssssnnrsssvwennewseonnsnn wns 46 Gas Sampling Probe cacsmcccecsrwenrcemeneunsaesee ee HHS SH ROT ETS EES 58 Mass Spectrum of Vinyl Chloride Standard ......cceeecceeccceeccecs 60 Schematic of Purge-Trap Apparatus .......cccccccccccceccccccccvees 61 Total Ion Monitor (GC/MS) Trace for Gas Samples from Landfil] Sampling Sites ...cescssoscssesesessncvvwcces se ee ae wee eel wR 82 Production of Vinyl Chloride From 1,1-Dichloroethylene (Vinylidene Chloride, VI) and 1,1,1-Trichloroethane (TE), During Incubation With Sewage Sludge......csccccescccccccccsccsscsscvcvevsveene 87 Purge-Trap Analysis of Products From Labeled TCE Incubated With Richland Sewage SIUdge ...ccsccescssccscccccsecccvovesecceces 88 Landfill Samples Spiked with 13¢_trichloroethylene Jue HONOR ARR REE 93 GC/Mp, Trace Showing Production of 13¢_Labeled Vinyl Chloride from C-Trichloroethylene by an Isolated Methanogen Culture...... 95 Reductive Dechlorination Pathway for Chloroethylenes in Soil...... 102 ''10. iis LZ 13. 14. 15. 16. LF 18. 19, 20. 21% 22 23 LIST OF TABLES Physical Properties of Vinyl Chloride Monomer .........cecececeece Lees Lane Landfill Trace Gas AnalySiS ...cceccccccccccccccccccccees Port Washington Landfill Trace Gas AnalysiS .....ccccecccccceceene Water-Borne Contaminant Concentrations at Price Landfill Illegal DUMP SITE cttecsce. ows eee HCE KT SS OETA EROE EKER Rew Photodecomposition Rates of Chlorocarbons ......cececcccececcccees Biodegradation Half-Lives for Chlorocarbons by a Muck-Water Sample Components in Purgeable Halocarbon Mix 60/M2, Lot LA13373 ........ Chlorocarbon Standard Solution Concentration and Analytical Data . Anerobic Sewage Sludge/Chlorocarbon Incubation Experiment #1 ..... Anaerobic Sewage Sludge/Chlorocarbon Incubation Experiment #2 .... Landfill Site A Incubation Experiments .......cccccccccccccceccces Landfill Site B Incubation Experiments cicccccacccscasavcsnssencue Site A and B Refuse and Gas Sampling Conditions .........cceceeeee Volatile and Ash Contents of Landfill Samples .........ccccecceeee ICP Inorganic Component Analysis of Landfill Samples ............. Trace Organic Component Analysis of Gas Sample A-2-16 ............ Head-Space Gas Analyses From Site A and B Samples ........ceeceees Vinyl Chloride Formation From Chlorinated Ethylenes and Ethanes . Vinyl Chloride Formation From 1,1-Dichloroethylene ............... Gas and Purge-Trap Analysis of Sample A-2-16-16.........ccceeeeeee Vinyl Chloride Production from Site A Landfill Samples............ Vented Gas Analysis from Site A Samples After 1-2 Weeks........... Vinyl Chloride Production from Site B Landfill Samples............ 18 20 20 22 30 34 56 63 66 66 69 70 75 80 81 83 84 86 89 90 91 91 92 ''SUMMARY AND CONCLUSIONS This study was performed to determine whether observed emissions of vinyl chloride from municipal landfills in California are the result of in situ production by natural processes. Vinyl chloride is a gas that until about ten years ago was widely considered to be harmless and was subsequently identified as a potent carcinogen which gives rise to a rare form of cancer known as hemangiosarcoma. Exposure in ambient air in California is currently restricted to a 24-hour average of no more than 10 ppb. Emissions of vinyl chloride into air and groundwater in some cases have been found to exceed this limit, even where no material containing vinyl chloride was ever reportedly dumped. The sources of these emissions has been of increasing concern throughout the U.S. Initial work was directed at generation of a list of potential vinyl] chloride sources, and modes of formation based on published information. This initial list includes a wide variety of possibilities, some of which could be eliminated on the basis of available evidence, and others which required further experimentation and research. Possible sources initially considered included: Legal and illegal dumping of PVC production sludges Leaching of residual vinyl chloride from plastics Decomposition of PVC (polyvinyl chloride) Dechlorination of chlorinated solvents In situ synthesis from acetylene and chloride An experimental approach was designed to test these various possible mechanisms and identify those which are actually occurring. To make the laboratory study more relevant to actuality, material samples were obtained from two California landfills (one Northern and one Southern). This material was used in most of the experiments, and was obtained from the surface and at two different depths below the landfill covers. A further source of raw material used in some experiments was a "clean" (solvent-free) sludge obtained ''from the City of Richland, WA, sewage treatment plant. The landfill condi- tions at the sampling points were recorded and used to define experimental conditions. Detection and quantitation of vinyl chloride and other compounds was by means of a modified purge-trap technique based on EPA method 601. The major modification consisted of using a volatile compound chromatography column, with the first 1/3 of its length packed with Chromosorb 101 and the remaining 2/3 packed with Chromosorb 102. This modification permitted separation of water and more volatile compounds from vinyl chloride and gave complete separation of cis and trans-1,2-dichloroethylene (which is not achieved by the packed column used in the EPA 601 method). Landfill samples were analyzed for both inorganic and organic components, and especially vinyl chloride and other chlorinated solvents. The Southern California site (designated Site A) contained more organic contaminants than the Northern site (designated B), in agreement with the sites' histories of dumping. Vinyl chloride was present in the gas taken after core drilling at Site A at a level of 5-7 ppm, but was not detected in the gas taken after core drilling at Site B. Production of vinyl chloride from PVC plastics was measured directly by heating samples and monitoring the emitted gases by gas chromatography/mass spectrometry (GC/MS). PVC older than 10 years contained 330 ppm of residual monomer while a 1-year old sample contained only 3 ppm, reflecting industry attempts to reduce monomer levels. Pyrolysis to 650°C did not produce addi- tional vinyl chloride monomer, in accordance with published data which indicates that PVC decomposes thermally by a dehydrochlorination mechanism. This results in formation of a linear polyene which cyclizes to the major observed pyrolysis product, benzene. Further experiments were performed using 13-C-labeled trichloroethylene (TCE) as a marker compound. TCE is the most widely used solvent likely to be found in landfills and has been implicated as a vinyl chloride precursor. 1361 CE was used since it is non-radioactive and can easily be tracked by 10 ''GC/MS. Formation of 13¢ vinyl chloride and dichloroethylenes provided conclu- sive proof of their origin from 136 TCE. Photolysis of TCE to vinyl chloride at landfill surfaces exposed to sunlight was eliminated as a possible formation mechanism, by experiments in which labeled TCE was irradiated by UV light over landfill surface soil. In air, chloroacetylene was formed, while in argon no organic products containing labeled carbon were found. This is in accordance with literature reports indicating conversion of TCE to chloroacetaldehyde. Vinyl chloride itself is reported to have only a 4.5 hr half-life in sunlight in air. Labeled TCE and a range of unlabeled chlorinated ethylenes and ethanes were incubated with landfill and sewage materials under simulated landfill conditions. Vinyl chloride was first detected after a 1-6 week lag period. When the experimental systems ("microcosms") were initially sterilized by autoclaving or by gamma irradiation, no vinyl chloride was produced. However, dichloroethylenes were produced from 1,1,2-trichloroethane under sterile con- ditions, indicating a biogenic formation route for vinyl chloride but that the chemical conversion of chloroethanes to chloroethylenes also occurs. Landfill material was inoculated into an anaerobic fermentor and grown on methanol as a carbon source. Inoculation onto agar roll-tubes under anaerobic conditions resulted in growth of an apparently homogeneous species. The organism is a methanogen but an exact species identification has not been e-tee to made. Conversion of C-VC by this organism has been confirmed. These experiments led to the following overall conclusions: e The most probable route for vinyl chloride formation in landfills is by the action of anaerobic bacteria on solvents such as TCE. Even in care- fully controlled municipal landfills, some TCE would be present from paint products, duplicating fluids, and household materials. e A secondary background level source of minor importance may be resid- ual ("old") PVC plastics which can gradually be leached by water, especially 11 ''at the relatively elevated temperatures of active landfills, and release their load of vinyl chloride. e Other mechanisms (pyrolysis, surface photolysis, hydrolysis of TCE and related solvents) do not appear to be significant contributors to overal] vinyl chloride formation. e Interconversion of chloroethylenes and chloroethanes in landfills is a complex network rather than a simple forward pathway. Chloroethanes may be hydrolysed to chloroethylenes by abiogenic means. Vinyl chloride itself may be back-converted to 1,1-dichloroethylene. e Within the limitations of this study (small sample size, limited num- bers of samples, etc.), vinyl chloride appeared to be produced only where chlorinated solvents were present in the landfill. 12 ''RECOMMENDATIONS The most probable primary source of vinyl chloride emissions from land- fills is from the bio-degradation of trichloroethylene and related solvents, as determined during the course of this study. Monitoring of vinyl chloride emissions from landfills throughout the State of California is ongoing. Results to date support the hypothesis that vinyl chloride formation at land- fill sites is dependent upon the presence of chlorinated solvents previously dumped. As can be seen from the solvent use profiles in Appendix A, the major use for chlorinated solvents is in degreasing of metal parts, followed by use as chemical intermediates and components in adhesives, aerosols, etc. Household use accounts for only 5% of the total use of, for example, trichloroethylene. By contrast, 85% of vinyl chloride production goes directly into PVC manufacture, and virtually all of the remainder is exported. Vinyl chloride users are well aware of the hazards associated with the use of the monomer gas and currently release only very small quantities into the environment. This was not the case ten years ago, however, when vinyl chloride was generally considered to be harmless in small concentrations, and large amounts of PVC production sludges containing residual monomer were dumped. Because of these considerations, and the fact that biodegradation of chlorinated solvents is slow, emissions of vinyl] chloride may continue for many years after solvent dumping. The problem is therefore a long-term one, and will not disappear overnight even if dumping of chlorinated solvents is banned immediately. Remedial measures which can be recommended include a range from merely palliative, to final solutions. We found no evidence for the de novo syn- thesis of vinyl chloride from simpler constituents; traces may occur this way, but the major source is from added solvents. Currently, landfill gas emitted from hazardous dumps is collected and burned. This is an effective way of destroying vinyl chloride, and should be extended wherever practicable to non-hazardous waste landfills where vinyl chloride above a certain level (e.g., 10 ppb) has been detected. 13 ''The landfilling of all two-carbon based chlorinated solvents (ethylenes and ethanes) should be terminated forthwith. Preferably, all waste solvents of this type should be destroyed or recycled before leaving their site of use. Technical methods exist for doing this. Oxidation by incineration or electrochemical methods or pressurized high temperature aqueous alkaline destruction under development at Battelle-Northwest and elsewhere are quite effective. Alternatives to the use of chlorinated solvents in household products which eventually find their way into municipal landfills should be sought. These only account for 5% of solvents dumped, but can provide a background of vinyl chloride in air and groundwater. Research should be performed to investigate long-term, low cost methods for vinyl chloride trapping or destruction. One example would be to apply the technique of in situ bioconversion by developing aerobic bacterial strains which can degrade vinyl chloride, and seeding them onto landfill covers. Fortunately, the half-life of vinyl chloride in air and sunlight is quite short, and for this reason widespread health-threatening concentrations are unlikely to occur. (The State of California Ambient Air Quality Standard for Vinyl Chloride is 10 ppb, the lowest level that can at present be reliably detected.) Perhaps the most feasible control method is the collection and burning of landfill gases. Topsealing prevents random dispersal of the gases, but does however cause formation of positive gas pressure. This pressure is often relieved by lateral gas movement which can cause problems in neighboring buildings. Gas pressure must therefore be relieved. Treatment of the gases to remove toxic and odoriferous components is then necessary, and combustion may be the most economical method. Care needs to be exercised in order to maintain combustion conditions where complete oxidation of all toxic com- ponents occurs. A longer term approach may be to minimize the amounts of chlorocarbons deposited in landfill operations in the future. 14 ''INTRODUCTION PROJECT SCOPE AND PURPOSE This study was designed to identify the sources of observed emissions of vinyl chloride in the air over California landfills, both industrial and municipal. At the time the study was begun, various modes of origin of the vinyl chloride being considered included possible illegal dumping of PVC sludges, PVC thermal destruction in situ, biological conversion of chlorinated solvents, and chemical reactions. As a first step in developing suitable control and/or remediation approaches it was necessary to pinpoint the actual sources of the emissions by critically examining reports of research already performed. Vinyl chloride is a bulk industrial chemical. Almost 7 billion pounds were produced in the U.S. in 1985 in 11 plants in Louisiana, Texas, and Kentucky. It was used almost entirely in the production of PVC (polyviny] chloride). Until about 10 years ago it was considered to be a virtually non-toxic anaesthetic gas. It was subsequently identified as a potent car- cinogen responsible for a rare type of hemangiosarcoma. Restrictions on exposure were promptly applied, and at present the maximum 24-hour average ambient air exposure limit in California is 10 ppb. Concurrently with a gradual reduction in exposure limits to vinyl chloride, monitoring of trace organic emissions in air from landfills in California was widened to include chlorinated hydrocarbons. In some cases, vinyl chloride was identified as a component of these emissions, at concentrations above the 10 ppb exposure limit. It therefore became a matter of concern to identify the sources(s) of these emissions so that steps could be taken to control them. Emissions of vinyl chloride from landfills in California was particularly surprising in view of the fact that no vinyl chloride is currently produced in that state. A review of existing literature information was undertaken as a first step in meeting the project objectives. This allowed some probable routes for vinyl chloride formation to be identified, and also permitted tentative elimination of some theoretically possible routes based on negative experi- mental evidence. Following this an experimental plan was developed with the 15 ''intent of confirming the formation of vinyl chloride by the most probable routes, and eliminating other possible routes. This plan was based as far as possible on an examination and use of actual landfill material so that the results would be directly applicable to the real-life situation in California. This study was undertaken to meet the above objectives by the quickest research route and to provide information which can be used to help in solving an environmental problem. It was limited in the number of landfill samples we could obtain and the amount of analytical work which could be performed in working with them. We make no claim that the landfill material samples are "representative", although we hope they were typical. Similarly, purge-trap analyses for vinyl chloride are accurate, but at the expense of total recovery of higher molecular weight solvents from the vial samples. We have not per- formed detailed reaction kinetic determinations on TCE and other solvent bio- degradation as this was outside the project scope, but rather have identified the major routes for vinyl chloride formation. SURVEY OF PRIOR WORK The purpose of this introductory survey is to summarize available infor- mation on various potential mechanisms for vinyl chloride formation and emission from waste dumps, and to generate a list of mechanisms in order of their likely contribution to the overall emission. Key references or reviews are used to illustrate available evidence for or against a particular mecha- nism rather than a complete review of all available documents. The reason for concern about vinyl chloride emissions lies in recent evidence that the gas, long considered harmless, is in fact a potent car- cinogen and mutagen. For example, the entry for vinyl chloride in the Kirk- Othmer encyclopedia (Brighton, 1974) quotes Industrial Hygiene (1963): "A detailed study using rats has shown that repeated daily exposure to concen- trations of 2-5% of vinyl chloride causes no permanent damage. It has been confirmed from 5 minute exposures of human subjects to vinyl chloride in con- centrations of up to 2.0% that a concentration of about 600 ppm is required to produce minimum effects of anesthesia under continuous exposure". The 16 ''physical properties of vinyl chloride are summarized in Table 1. By compari- son, Hefner et.al. (1975) state that: "Vinyl chloride monomer.... has been associated with the development of angiosarcoma and portal cirrhosis of the liver as well as other untoward effects in workers exposed to unknown but undoubtedly high concentrations of VCM." Other more recent research resulted in the continued downward movement of vinyl chloride maximum permissible exposure limits, beginning at 10 ppm (EPA, Fed. Reg., 1976; Lamorte, 1978; Dimmick, 1981). The existence of vinyl chloride emissions from landfills has been known for some years, as has the presence of vinyl chloride in leachate water (vinyl chloride is slightly soluble in pure water). The major question to be answered in this study is the root cause of the vinyl chloride presence: Is it a result of past dumping of materials containing vinyl chloride, or is it being generated in situ, and if so, from which precursors? The potential approaches to identifying key experiments to decide this issue are many. We chose an approach which combines observational evidence with logical deduction. The conclusions reached in this brief review of existing evidence were used to produce an experimental plan to further define the exact mechanism for vinyl chloride emissions. Hopefully, further work (beyond the scope of this project) will then be done to identify ways of preventing or suppressing such emissions in the future. Much of the basic work performed to date has consisted in monitoring vinyl chloride and other chemicals in air and water in the region of produc- tion facilities and toxic chemical dumps. Information to 1980 on atmospheric trace organic contaminants over the continental U.S. has been compiled in one document listing 151 chemicals (Brodzinsky and Singh, 1983). A study of vola- tile organic compound concentrations in the air of the New Jersey/New York area has also been published (Kebbekus and Bozzelli, 1983), and lists data for 27 different compounds, including vinyl chloride and several other halogenated hydrocarbons. A number of states have developed programs for ambient air monitoring and analysis of trace organic contaminants. The State of California approach and some results are presented in two recent papers by Venturini et.al. (1985) and Kowalski et.al. (1985). 17 ''Table 1: Physical Properties of Vinyl Chloride Monomer Property Numerical Value Property Numerical Value M. Wt. 62,501 Critical Temp. 158.4 B. Pt. -13.37 Critical Press. 52.2 atm. F. Pt. -153.79 Dielect. const. 6.26 Flash Pt. -78 Refractive index 1.398 Liquid density Sp.Ht. liquid 0.38 -20 0.98343 Latent Ht. vap'n 71.26 -25 0.99176 Latent Ht. fusion 18.14 -30 0.99986 Ht. formation (cal/g.-mole) 9000 Viscosity, cp -10 0.248 Vapor Press., mm -20 0.274 25 3000 -30 0.303 -13.37 760 -40 0.340 -55.8 100 -73.9 30 Surf. tens. ,dyne-cm -87.5 10 -10 20.88 -109.4 1 -20 22.27 -30 23.87 All data in °C, grams, cal. unless otherwise stated. Dielectric constant at 105 Hz and 17.2°C. After Brighton (1974). Volatile organic compound emissions from municipal waste landfills have become a cause for concern in a number of states because of the potential hazard to local inhabitants from landfill gases. One example was described by Walsh (1984) for the Lees Lane (OH) and Port Washington (NY) landfills. Both are municipal landfills, and in the Port Washington facility particu- larly, no hazardous wastes were permitted and the restriction was reportedly stringently enforced. However, analysis of the landfill gases showed the presence of dichloroethane among other components at the Lees Lane site. (The isomer was not identified.) 1,2-Di-chloroethane is used mainly in the manufacture of vinyl chloride, but does have significant domestic uses. Until 1982, Stauffer Chemical produced 1,2-dichloroethane in the Los Angeles area. It was used in pesticides, solvents, gasoline, and in solid fuels as well as for vinyl chloride manufacture. The efficiency of screening of materials dumped at the sites may therefore be open to question. Other halogenated materials identified could conceivably come from paints, solvents, 18 ''etc. from non-industrial manufacturing uses. Vinyl chloride and other compound concentrations in monitoring well gases at these sites were determined over a 2 year period. At Lees Lane the mean vinyl chloride concentration was 28 ppm (range 0-188 ppm) and at Port Washington it was 15.84 ppm (no range reported). Controls applied to these landfills consisted of extraction and blower systems to divert the effluent gases to atmosphere away from homes and schools in the areas concerned. Gas incineration is being considered. This example illustrates the magnitude of the problem and the lack of efficient methods of dealing with it. Data on the concentrations of chlorocarbons measured in monitoring wells located in these landfills is presented in Tables 2 and 3. Data such as that presented in Tables 2 and 3 is indicative of the problem, but should not be considered as representing the situation in California landfills in any quantitative manner. Differences in sampling techniques, materials, and data interpretation can lead to quite different results from one landfill to another, regardless of variations in chemical composition. On-site monitoring by portable gas chromatographs can in prin- ciple eliminate many of the errors in sample collection, transportation, and handling. In the example given by Jacot (1983), a "typical" (unidentified) landfill was described: It had been in use since the late 1930's, with waste being burned at the site until the early 1960's. Landfilling began in the mid- 1960's and incineration was discontinued, and only municipal waste has been accepted since the mid-1960's. Vinyl chloride was detected (16 ppm) at one of the monitoring points near the center of the landfill, and is presumably being formed from solvents dumped years ago. However, the fact that vinyl chloride is also being emitted from municipal dumps supposedly free of hazardous material is significant, as is the fact that there are measurable amounts of dichloroethane, di-, tri- and tetrachloroethylenes present in association with the vinyl chloride. So far none of the literature examined has shown vinyl chloride production in the absence of detectable amounts of these other materials. (The term "-chloroethylenes" is used here rather than the alternative "-chloroethene" for consistency with other reports and common usage). 19 ''Table 2: Lees Lane Landfill Trace Gas Analysis (after Walsh, 1984). Compound Mean Concn.* Range # of Analyses OSHA Level Benzene 8 0-46 15 10 Butadiene 3 -3- 1 1000 Cyclohexane 2 0-19 15 300 Dichlorodi fluoro- methan 6 0-25 14 1000 Dichloroethane 6 0-22 15 100 Dichloroethylene 40 -40- 1 200 Ethylbenzene 6 0-27 15 100 Heptane 2 0-15 7 400 Hexane 6 0-37 16 50 Toluene 15 1-175 15 200 Vinyl Chloride 28 0-188 44 1 Xylene 10 0-45 7 100 * All data in ppm. + . Ca . eps Compounds named as in original report; isomers not specified. Table 3: Port Washington Landfill Trace Gas Analysis (after Walsh, 1984). Compound On-Site* Homes School OSHA Limit’ NYS Limit Benzene -- 5.1 3.0 10,000 33 Toluene 2,930 9.5 2.5 200,000 2000 Xylene -- - - 100,000 333 1,1,1-Trichloroethane 200 lal -0- 350,000 7000 1,1,2-Trichloroethane -- - - 10,000 33 Tetrachloroethylene 300 1.4 -0- 100,000 167 Trichloroethylene 80 - - 100,000 333 Vinyl Chloride 15,840 -0- -0- 1,000 0.16 * Wells * All values in ppb. Attempts are being made to establish standard techniques and methods for waste site analysis and monitoring (Ford and Turina, 1985; Ford et.al., 1985; Plumb, Jr., 1984; Pellizzari et.al., 1984), so that data from one site can be applied to another, but it is a difficult problem. 20 ''Information on vinyl chloride in leachate water from waste dumps can also be used to infer mechanisms for vinyl chloride formation, although in this case the relatively high solubility of vinyl chloride in water compared with the higher molecular weight chlorocarbons can "skew" quantitative results. An additional cause for confusion is the frequent error in the identification of cis-dichloroethylene as the trans- isomer, due to the inability of the standard EPA 601 analytical method (using a packed GC column) to separate the two (Cline and Viste, 1984). Three studies on groundwater contamination by vinyl chloride are indicative of the current state of research: Pennington (1983) performed a hydrogeological investigation at the aban- doned Price Landfill hazardous waste site (Pleasantville, NJ, west of Atlantic City). This investigation was instigated as a result of a lawsuit by citizens concerned about drinking water contamination, and consequently had to be per- formed quickly. Vinyl chloride was detected in the water leachate plume together with other chlorocarbons, and was also present (up to 33 ppb) in some surface water. Up to 650 ppb of vinyl chloride was present in drainage areas down gradient from an equipment storage site, and 380 ppb was found emanating from an illegal dump site. Table 4 lists the compounds and their concentrations identified in the latter samples. Highest concentrations of vinyl chloride were again associated with high concentrations of other chlori- nated materials. Sabel and Clark (1983) reported a compilation of data on organic compound contamination of leachate water from 6 Minnesota municipal solid waste land- fills, as well as 5 from Wisconsin and one from New York. Their conclusions were that dichloromethane (methylene chloride), trichloroethylene, 1,1-di- chloroethane (not 1,2-), and trans-1,1-dichloroethylene were ubiquitous. (There are no cis-/trans- isomers of 1,1-dichloroethylene, which error also casts doubt on the identification of 1,1-dichloroethane rather than the normally encountered 1,2- isomer). Vinyl chloride was detected in water from 3/13 of the sites sampled and from one of the 4 Wisconsin sites (at 61 ppb). 21 ''Table 4: Water-Borne Contaminant Concentrations at Price Landfill Illegal Dump Site (Pleasantville, NJ) (after Pennington, 1983). Depth (in.) Compound 0-9 15-27 43-47 56-62 Benzene _ 190* 100 <75 Chloroform =e <100 «75 <75 1,2-Dichloro- ethane we 570 <75 <75 Trans-1,2-dichloro- ethylene we <100 <75 K75 Dichloromethane ae 765 250 <75 Toluene ae am 230 100 Trichloroethylene we 150 <75 130 Vinyl Chloride _ 380 <75 <75 Petroleum Hydrocarbons (%) 32 0.5 0.9 1.1 * All values in ppb except for petroleum hydrocarbons (%). ** Target compound masked by large amounts of other VOC's. The third study is the most interesting, because it attempts to correlate migration and degradation patterns of the volatile organic compounds observed in waste site water leachates. This study (Cline and Viste, 1984) identifies the dechlorination pathway for formation of vinyl chloride from polychlorin- ated solvents (following page). The data presented includes municipal solid waste landfills, industrial] chlorinated solvent recovery facilities, and an industrial site. The concen- trations of primary compounds and breakdown products were generally correlated with distance from the landfill site, amounts of primary materials dumped, the presence of biodegradable material to support anaerobic digestion, depth, and flow direction. This evidence strongly supports the hypothesis that vinyl chloride is formed as a result of reductive dechlorination of common chlorinated solvents by anaerobic bacteria. However, before this conclusion is confirmed, it is necessary to consider all reasonable formation mechanisms to provide a basis for comparison. The evidence for vinyl chloride formation by biological and other mechanisms is summarized in the following pages. Ze ''Chloroethylenes Cl cl -¢l 86 C1 H | cl -Cl Cl Nes=co —4 ~~ C= C0 a, Sc=cl Space” crm = CI ci“ “el H~ H H H Perchloro- Trichloro- cis-1,2-Dichloro- Vinyl Chloride ethylene ethylene ethylene (2) (2) + Clo . trans-1,2-Dichloro H~ Cl ethylene (2) + Cl TH SC=Co 1,1-Dichloroethylene Cl H Ut Hen St riety 1 Chloroethanes _ _ -Cl -Cl Cl3C—CH; ——-“—___»(15CH—CH, ——-“&—--» C1CH»—CH 1,1,1-Trichloroethane (2) 1,1-Dichloroethane (1) Chloroethane (1) (1) Substantial degradation; (2) slow degradation Vinyl Chloride Formation Routes: The various means by which vinyl chloride may be formed in landfills (both toxic chemical waste and municipal dumps) can be classed as leaching, chemical, photochemical, and biological. Each is briefly defined below: Leaching Vinyl chloride emission may result from the physical transportation to the surface of vinyl chloride monomer already present in the landfill. Although vinyl chloride is a gas under normal conditions (b.p. -130C), it is present in solution or entrapped as the monomer in waste sludges from PVC manufacture, and in pre-1975 fabricated materials such as PVC pipe and film (e.g., Saran wrap). Recent legislation has considerably reduced the permis- sible levels of vinyl chloride monomer present in these materials, but significant amounts of older material are still present in landfills and could be contributing to the emission problem. 23 ''The problem of vinyl chloride monomer emission from waste PVC sludges has been addressed by Markle et.al. (1976), who took waste PVC sludges from a number of industrial sources and analyzed them for monomer content. The concentrations of monomer found ranged from 7 to 520 ppm in the wet filtered sludges (20 - 1260 ppm dry solids basis). These sludges were reportedly dumped in industrial and in municipal sites and could therefore be a con- tributor to observed vinyl chloride emissions from some municipal sites. Release rate studies performed using these sludges under simulated landfil] conditions showed a rapid initial release rate, followed by a slower release of the major amount of vinyl chloride over a much longer time. A background air concentration of 100 - 300 ppb of vinyl chloride was observed at landfills where PVC sludges have been disposed of for several years, although up to 1000 ppb could be detected 1.5 m above the ground for up to 24 hr after the sludges were covered. Although this amount of vinyl chloride is significant, and may account for part of the observed emissions, it does not account for all of the observed 50 ppm of vinyl chloride in air noted above some indus- trial waste landfill sites in California, some 10 years after PVC sludge dumping was prohibited. The equilibrium distribution of vinyl chloride monomer between PVC powder (0.15 - 0.25 um) and water has been determined (Kontominas et.al., 1982) (Figure 1), and is definitely non-linear with concentration in the polymer. At 22°C, and above 20 ppm, the desorption curve becomes almost exponential in nature. In PVC sludges, with high initial vinyl chloride content and small particle size, a very rapid desorption into the aqueous phase would occur. Exchange with gas evolved from anaerobic digestion of cellulosic materials in the landfills would then transfer the monomer to the air. This study supports the conclusions already noted, that vinyl chloride emissions from film and sludges is a contributor rather than the main cause of observed emissions. Chemical Chemical formation of vinyl chloride is considered to involve mainly the degradation of PVC (polyvinyl chloride) and chlorocarbons through the action of heat, alkali, or metal ions. An alternative route, in situ synthesis from 24 ''Figure 1. Sorption and Desorption Isotherms for PVC-VC-Water at 22°C. 40t- z = S Ss i £ a a & oO — 20 = oO > E 7 @ Sorption © Desorption 1 | 1 4 6 1.0 2.0 3.0 [von ppm (w/w) water (After Kontominas et.al. (1982). acetylene and hydrogen chloride in the presence of mercuric ion catalysts (based on the now obsolete process for vinyl chloride manufacture) is unlikely to occur in landfills, based on our present understanding of landfill condi- tions. Careful review of the available literature indicated that the chemical degradation route contribution to vinyl chloride emissions is at best, minor. Consideration of the various degradation routes which have been established for these substances shows the reasons for this. (a) Polyvinyl chloride: The processes involved in thermal, photolytic, and high-energy degra- dation of PVC have been reviewed up to 1975 by Close and Gilbert (1977). PVC degradation (either thermal or photolytic) involves loss of hydrogen chloride and formation of an unsaturated group. This general mechanism is illustrated in Figure 2: 25 ''Figure 2. Mechanism of PVC Degradation. (After Close and Gilbert (1977). ~CHa“CH-CHp-CH-+°CI Gl hv + e e keg bprenann “Han Han Ct +H Cl Cl Cl Cl ebtiget lebilat +°H cl Cl ~CHy“CH-CH>-CH- *°Cle—-CHo- CHeCH-CH- + HCl | Cl Cl Further loss of hydrogen chloride from adjacent sites, catalyzed by the HCl already lost, results in formation of a polyene. After approximately 6 conjugated double bonds are formed, chain stability has decreased to the point where scission occurs. In the presence of air, this is facilitated by oxidation. In the absence of air, ring formation and evolution of benzene occurs. Other aromatic hydrocarbons and chlorobenzenes are minor products. The overall composition of these minor products varies with the PVC compo- sition (additives, plasticizers, pigments, etc.). Levels of vinyl chloride occurring as products of side reactions have been reported as contributing no more than 0.25-0.7 mg/g of PVC between 200-600°C. (Boettner et. al., 1969). Lewis (1975) was able to detect no vinyl chloride monomer from pyrolysis at 225°C. A maximum of 35 ppm was found at 350°C for PVC degraded in air, and a maximum of 6 ppm was found at 500°C after thermal degradation in helium. The only places within landfills where these temperatures would exist is in regions where degradation of waste organic material has resulted in spon- taneous combustion. Under such situations, up to 35 ppm of vinyl chloride monomer could be formed from waste PVC. The PVC content of a typical landfil] is not very great (we estimate a maximum of 0.1% by weight), and combustion would be noted by smoke, carbon monoxide, benzene, and other emissions. Vinyl] chloride monomer emission has not been associated with combustion in any of the literature reviewed. Since such an association would almost certainly 26 ''have been noted, we conclude that thermal degradation of PVC is not a significant cause of the observed vinyl chloride formation. (b) Chlorocarbons. 1,2-Dichloroethane prepared from chlorination of ethylene is the major industrial precursor for vinyl chloride, from which it is produced by pyroly- sis at 400°C or higher. In 1974 over 86% of the dichloroethane produced in the U.S. was used for vinyl chloride monomer production (Drury and Hammons, 1979). Figure 3 shows the distribution of pollutants, including dichloro- ethane, resulting from vinyl chloride monomer production from dichloroethane pyrolysis. Dumping of chlorinated solvents could conceivably lead to vinyl chloride formation if pH and temperature conditions were right. In the absence of spontaneous combustion conditions, the maximum temperature likely to be achieved in any landfill is that of the maximum for anaerobic digestion. Composting (a technique used for digestion of sewage sludge and for killing of all pathogens) normally generates up to 55°C at maximum, and this is held for only a few days after addition. As soon as the cellulosic portion of the waste is digested, the temperature drops back towards ambient. However, under landfill conditions, with a heavily insulating overlay of soil, much higher temperatures can be generated and held for greater periods (years in some cases). Under certain conditions (i.e., addition of chlorinated solvents and alkali together) there could be some vinyl chloride formation. The rate is likely to be very slow. There is also competition with direct hydrolysis, which is favored in an aqueous environment. On the other hand, the persis- tence of 1,2-dichloroethane is high. Radding et. al. (1977) calculated a half-life for 1,2-dichloroethane at pH 7 and at 25°C of 50,000 years, based on a hydrolysis rate constant of approximately 5 x 10(exp-13) sec(-1). The presence or absence of alkali associated with dichloroethane in landfills may therefore have some influence on vinyl chloride emissions. Examination of the waste streams (Figure 3)(Drury and Hammons, 1979) from dichloroethane conversion to vinyl chloride reveals that alkali is in fact used and dumped to land together with some dichloroethane. Vinyl chloride 27 ''Figure 3. Hammons, 1979). Waste Products From Vinyl Chloride Production Process. (Drury and HCI SEPARATOR VINYL PYROLYSIS QUENCH uC « _ CHLORIDE FURNACE 0.6025 kg ww SEPARATION . = BASIS: 1 kg VINYL CHLORIDE MONOMER FS, (3 ) VINYL CHLORIDE = = 1.0 kg _ 2 & \) ” ETHYLENE 0.50 kg PURE 1, 2-DICHLOROETHANE 1 : DILUTE REFLUX HEAVY ENDS — apaneneer SOLUTION CRUDE 2 “ Sanne SCRUBBER E =| @ nae CHLORINE 1.22 kg ®@ sites HEAVY Ente CAUSTIC EFFLUENT wesHAGe RECYCLE PRODUCT CHLORINATION FEED FILTER LIGHT ENDS ~HEAVY ENDS REACTOR NEUTRALIZATION REMOVAL —sw REMOVAL @) CAUSTIC WASHAGE (WATER) (2) FILTER EFFLUENT (SOLID) DICHLOROETHANE —_ 0.00435 kg TARS TRACE SODIUM HYDROXIDE 0.00098 kg SOLIDS (AS CARBONS) —_ 0.00008 kg SODIUM CHLORIDE 0.00033 kg VINYL CHLORIDE —0.00098 kg FILTER EFFLUENT (LIQUID) METHYL CHLORIDE 0.00085 kg ETHYL CHLORIDE 0.00085 kg DICHLOROETHANE 0.0005 kg } SODIUM HYDROXIDE — TRACE TO WATER TO LAND @) VENT ON REFLUX CONDENSER (GAS) (@) HEAVY ENDS ETHANE 0.0049 kg DICHLOROETHANE 0.0024 kg DICHLOROETHANE 0.012 kg 1,1, 2-TRICHLOROETHANE 0.004 kg METHANE 0.0043 kg TETRACHLOROETHANE 0.004 kg TARS TRACE TO AIR TO LAND © HEAvy ENDS HEAVY ENDS 0.037 kg DICHLOROETHANE 0.0008 kg TARS 0.00005 kg SOLIDS ASH 0.0002 kg TO LAND 28 ''wastes are dumped to water - although more severe legal restrictions since 1979 will have reduced this considerably. The “heavy ends" waste stream would be the major problem in industrial landfills. Municipal landfills should not contain any of this material, however. (Also, measured pH at the collection sites for material used in our experimental work was never greater than 6.4.) Information on other halogenated solvents is more difficult to assess. Trichloroethylene and perchloroethylene (tetrachloroethylene) are mainly used in dry cleaning, an industry which practices stringent recovery of solvents for economic and environmental health reasons. Charcoal filters used for vapor absorption are regenerable, and require replacement only every 15 years or so. The total solid waste production by the entire industry was estimated as 132 tons by 1995 (Goodwin and Hawkins, 1980). Furthermore, these solvents are even more resistant to hydrolysis than dichloroethane, because of the presence of the double bond, and would not be expected to yield vinyl chlor- ide. Addition of water to the double bond in trichloroethylene followed by elimination of hydrogen chloride would yield dichloroethanal by rearrangement, while direct elimination of hydrogen chloride would produce dichloroacety- lene. Further conversion of these molecules would not produce vinyl chloride. Photochemical One possibility for the formation of vinyl chloride that must be con- sidered is that the vinyl chloride is formed by photochemical reactions occurring at the landfill surface, possibly catalyzed by the landfill cover material. For example, in principle the activation of 1,2-dichloroethane by sunlight could lead directly to vinyl chloride through elimination of. hydrogen chloride, by analogy with the thermal process for vinyl chloride manufacture. Examination of the detailed literature on photochemical reactions of chlori- nated hydrocarbons shows that this is most unlikely to happen to any signifi- cant extent in reality, provided that the photochemical degradation takes place in the presence of air. This is because the initially formed activated radical promptly reacts with oxygen, nitric oxide, hydroxyl radical, and other trace atmospheric components, leading to a degradation sequence which does not involve vinyl chloride. 29 ''Dilling et.al. (1976) investigated the solar photodecomposition rates of a number of chlorinated compounds in the presence of 10 ppm of nitric oxide. Their data is summarized in Table 5. The half-life of vinyl chloride was determined to be 4.3 hr under these conditions, which included a 2.6x solar UV intensity, compared to summer noon at Freeport, IX. Table 5. Photodecomposition Rates of Chlorocarbons (after Dilling et.al. (1976). Compound Hal f-Life* Compound Half-Life (hr) (hr) 1,1-Dichloroethylene 2.1 Dichloromethane €5%/24 hr. Cis-1,2-Dichloroethylene 3.0 Tetrachloroethylene 14.2 1,1,1-Trichloroethane €5%/23.5 hr. Trichloroethylene 3.5 Ethyl acetate 14.6 Vinyl Chloride 4.3 Toluene 6.8 * Time for 50% disappearance in simulated atmospheric conditions; 10 ppm each of compound and Nitric Oxide. Solar-assisted catalytic decomposition of trichloroethylene and chloro- form has also been reported, using a titanium dioxide catalyst in an aqueous environment (Ahmed and Ollis, 1984). The products were carbon dioxide and hydrogen chloride. This system offers promise as a method of cleaning con- taminated water. A similar sort of catalysis could in principle occur at the surface of a landfill, between sunlight, chlorocarbon vapors, water, and the inorganic soil matrix. However, as already noted, the products do not include vinyl chloride. If the UV frequency is lower (i.e., below the atmospheric absorption cut-off), vinyl chloride is produced from the radiolysis of 1,2- dichloroethane by an- a,8-elimination route: CH,CICH,Cl' CH,CICH,C!" —~ C,H, + 2Cl (or Cly* — 2C1) *. oHACl + HCl* (HCI* — H + Cl) —~ CH,CICH* + HCl** (HCl** — H + Cl) CH.CICH* — C,H, + HCl (Yano and Tschulkow-Roux, 1980). 30 ''At least part of the vinyl chloride is formed through molecular reactions and is independent of the presence of radicals such as nitric oxide. Salomon et.al. (1979) attributed formation of vinyl chloride and 1,1,2-trichloroethane from 147 nm photolysis of 1,2-dichloroethane to disproportionation of inter- mediate dichloroethyl radicals by chlorine atom transfer. Similarly, vinyl chloride can be formed from chlorocarbons by the action of laser light, which generates very high energies in the receptor molecules. The sequence involved in hydrogen chloride elimination from 1,2-dichloroethane to form vinyl chlor- ide has been investigated by Schneider (1981), who used a laser system. This study was done with the objective of replacing the present vinyl chloride production process with a laser-driven process instead of a thermal pyroly- sis. Caballero and Wittig (1983), using pulsed CO> laser light (infrared), showed that the primary reaction of trichloroethylene was elimination of hydrogen chloride to form dichloroacetylene, which further decomposed with loss of chlorine atoms. Vinyl chloride was not produced. Reiser et.al., (1979), using a similar system, obtained similar results, and also showed that vinyl chloride was dissociated to hydrogen chloride and acetylene. Confirmation of this route was obtained by Moss et.al. (1981) using a UV light source at 193 nm, and determining products by infrared emission spectroscopy. Using mass spectrometry and molecular beam techniques, Sudbo et.al. (1978) did find vinyl chloride as a product from trichloroethylene. However, this sort of reaction condition may pertain during lightning storms, but is not a normal condition in landfills. Finally, the gamma- and neutron radiolysis of aqueous trichloroethylene has been investigated by Sasse (1967), and found to liberate sym-tetrachloroacetone and 1,1,2-trichloro-1-propene, as well as chloride, hydrogen peroxide, hydrogen, carbon monoxide and dioxide, and glyxoylic acid. The formation of oxygenated products from application of ionizing radiation to halocarbons appears to be a normal process in the presence of air and water. No evidence has been found for in situ formation of vinyl chloride from acetylene and hydrogen chloride in the presence of a mercuric ion catalyst. Hydrogen chloride is formed from PVC degradation but would be expected to react rapidly with the landfill matrix to form inorganic salts. Acetylene production is not a normal event during anaerobic digestion, and acetylene emissions have not been noted from landfills. Finally, dumping of mercury 3] ''salts is strictly controlled and has been for many years. The presence of significant amounts of mercury contamination in landfills would long ago have been noted. Biological Routes: By far the greatest amount of direct experimental evidence favors forma- tion of vinyl chloride monomer from sequential reductive dechlorination of chloroethylenes by anaerobic bacteria. This evidence will now be summarized. Research performed since 1980 has proven that vinyl chloride may be produced biologically from chlorinated ethylenes (e.g., Barrio-Lage et.al. (1986); Vogel and McCarty (1985); Parsons et.al., (1984). The potential for production of vinyl chloride from chloroethanes has been suggested by Cline and Viste (1984). Another conceivable route for microbial vinyl chloride production is the breakdown of chlorinated aromatic hydrocarbons. Microbial degradation of aromatic compounds proceeds via an oxidative ring cleavage which yields small fragments of the original rings which can retain chlorine atoms (Alexander, 1965). The chlorine atoms are eliminated in subsequent metabolic steps. Other biological reactions which could also yield vinyl chloride are 8-elimination from a halogenated compound yielding a double- bonded chlorinated compound such as vinyl chloride, dehydration reactions, and 8-cleavage of aliphatic compounds. No evidence has been found in the literature for vinyl chloride production by any of these mechanisms. Another potential mechanism, microbial attack on PVC, seems unlikely since PVC is known to be resistant to microbial degradation (Pantke, 1977). The only documented mechanisms for biological vinyl chloride formation use two-carbon chlorocarbons as substrates. The other mechanisms briefly considered above do not really appear reasonable in the light of our current understanding of biologically catalyzed reactions. Thus, in the following discussion the only mechanisms that will be considered are those using chloro- ethylenes or chloroethanes as substrates. 32 ''1. Vinyl Chloride Formation from Chlorinated Ethylenes: Several research groups have shown that chlorocarbons are dechlorinated by a reductive dechlorination mechanism. One of the earlier studies is that of Wood et.al. (1981), who reported the presence of apparent by-products from biological degradation in groundwater. Specifically, vinyl chloride, 1,1-di- chloroethylene, and cis- and trans-1,2-dichloroethylene were observed where only tri- and tetrachloroethylene had been released to groundwater. These compounds were thought to represent biodegradation products. Laboratory investigations were conducted to confirm this. Samples were incubated in septum vials which had been seeded with bottom sediments along with samples of contaminated surface water from the same site. Anaerobic conditions developed in the vials because of the high total organic carbon concentration of the surface water (approximately 25-40 mg/1 TOC), and limited amounts of oxygen initially present. The initial concentration of chlorocarbons was about 3300 wg/1. Results showed that the disappearance of tetrachloroethylene was accompanied by a sequential increase in trichloroethylene, cis- and trans- 1,2-dichloroethylene, 1,1-dichloroethylene, and vinyl chloride (monochloro- ethylene). A decrease in 1,1,1-trichloroethane concentration was accompanied by an increase in 1,1-dichloroethane and chloroethane concentrations. Similarly, carbon tetrachloride degradation was accompanied by formation of chloroform and methylene chloride. Plots of tetrachloroethylene, 1,1,1- trichloroethane, and carbon tetrachloride degradation showed a linear increase with time. The biological half-lives calculated by Wood et.al. (1981) for various chlorinated contaminants are given in Table 6. It is interesting to note that the observations of Wood et.al. (1981) agree well with the results of the National Organics Reconnaissance Survey summarized by Dyksen and Hess (1982). These data show that 1,1-dichloroethane and cis-1,2- dichloroethylene occur in groundwater supplies at about the same frequency as 1,1,1-trichloroethane, trichloroethylene, and tetrachloroethylene even though the latter are far more commonly used and released to groundwater. 33 ''Table 6. Biodegradation Half-Lives for Chlorocarbons by a Muck-Water Sample. (Wood et.al., 1981). Compound Half-life (days) Methyl chloride 11 (Estimated) Methylene chloride 11 Chloroform 36 Carbon tetrachloride 14 Chloroethane 10 1,1-Dichloroethane Long 1,2-Dichloroethane Long 1,1,1-Trichloroethane 16 1,1,2-Trichloroethane 24 Vinyl Chloride Long 1,1-Dichloroethylene 53 Trans-1,2-Dichloroethylene Long Cis-1,2-Dichloroethylene Long Trichloroethylene 43 Tetrachloroethylene 34 Roberts et.al. (1982) studied the behavior of organic contaminants in groundwater following the injection of reclaimed wastewater at the Palo Alto Baylands, CA. Declining concentrations of several contaminants, thought to be due to biodegradation, were noticed after adsorption breakthrough. Aerobic processes appeared to be occurring since a decline in both dissolved oxygen and total dissolved carbon was noted at observation wells. Fairly rapid degradation of trihalomethanes was noted with half-lives on the order of 20 days. Much slower degradation of 1,1,1-trichloroethane, trichloroethylene, and tetrachloroethylene was noted. Apparent half-lives for these contaminants were approximately 200 days. Degradation of these chlorinated contaminants appeared to follow first order kinetics. Schwartzenbach et.al. (1983) reported the behavior of organic contaminants following infiltration of river water into groundwater aquifers in the lower Glatt Valley and the lower Aare Valley, Switzerland. They observed the apparent degradation of alkylbenzenes and 1,4-dichlorobenzene, but did not notice any degradation of chloroform, 1,1,1-trichloroethane, or tri- or tetrachloroethylenes. Both aerobic pro- cesses (e.g., nitrification) and anaerobic processes (e.g., denitrification, manganese reduction) were observed at different points in the aquifers. 34 ''Bouwer et.al. (1981) investigated anaerobic degradation of several organic contaminants including chloroform and tri- and tetrachloroethylenes. Laboratory batch experiments were conducted using 1670 ml serum bottles containing an anaerobic medium which was seeded with methanogenic bacteria. Chlorinated compounds were present at concentrations from 10 - 100 ug/1. Chloroform was rapidly degraded under these conditions, but there was no appreciable degradation of the other chlorinated materials over the 16 week study period. Further work by Bouwer and McCarty (1983a) involved the same substrates at up to 200 ug/l incubated with laboratory digester sludge at 35°C. Excess acetate was added to serve as the primary nutrient. Continuous flow experiments were also conducted in upflow columns containing glass beads. A biological film was established in the columns using a feed con- taining approximately 1 g/1 of sodium acetate. The columns were then switched to a feed containing chlorinated organic compounds (including chloroform, 1,1,1-trichloroethane, and tetrachloroethylene, 160 ug/l each), with 0.5 g/1 acetate as a primary substrate. The results of the batch study showed com- plete degradation of chloroform after a very short acclimation period. 1,1,1- Trichloroethane and tetrachloroethylene showed almost complete destruction after 8 weeks, though a longer acclimation period was required. Trichloro- ethylene was observed after 8 weeks as an apparent degradation product of tetrachloroethylene. The column study showed significant removal of chloro- form, 1,1,1-trichloroethane, and tetrachloroethylene after 10 weeks of acclimation. The rate of degradation was observed to be more a function of the organism concentration than of the rate of acetate utilization. Reductive dehalogenation was proposed as the degradation mechanism for chloroethenes. Bouwer and McCarty (1983b) performed similar experiments under anaerobic conditions with a variety of chlorinated and brominated aliphatic compounds, using ethanol and nitrate as primary substrates. No degradation of chloroform or 1,1,1-trichloroethane was noted, but carbon tetrachloride and brominated methanes did degrade. The kinetics of biofilm transformations of chlorocar- bons during laboratory simulations of groundwater flow were examined by Bouwer and McCarty (1984). A primary substrate (acetate) was found to assist in the bioconversion of a series of chlorinated aliphatic and aromatic compounds. Tracer experiments with 140_jabeled substrates demonstrated anaerobic conver- sion of chloro-, 1,3- and 1,4-dichlorobenzenes to carbon dioxide, eliminating from consideration one possible formation route for vinyl chloride discussed 30 ''previously. Further information on these anaerobic biofilm conversion experi- ments can be found in Bouwer's thesis (Bouwer, 1983). The aerobic degradation of chloroform and tri- and tetrachloroethylenes was studied by Bouwer et.al. (1981). One laboratory column containing granulated activated carbon and one containing glass beads were fed a nutrient medium containing sodium acetate as the primary nutrient, together with 10 - 30 #g/1 of chloroform, 1,1,1- trichloroethane, or tetrachloroethylene. Aerobic growth was maintained for two years, but no degradation of any of the chlorinated contaminants was noted. Tabak et.al. (1981) also conducted studies to determine the aerobic treatability of organic priority pollutants in municipal treatment plants. The study employed a static culture flask-screening procedure in which con- taminants were spiked into a yeast extract nutrient medium which was seeded with primary sewage and incubated for 28 days. Contaminant concentrations used were 5 and 10 mg/l. Blanks were run to account for the volatilization losses of volatile contaminants. Methylene chloride and carbon tetrachloride underwent slow degradation under these conditions. Considerable acclimation periods were required for degradation of 1,1-dichloroethane and 1,2-dichloro- ethane, 1,1,1-trichloroethane, and 1,1,2-trichloroethane. No biological degradation of 1,1,2,2-tetrachloroethane was noted and no attempt was made to identify possible biodegradation mechanisms. Laboratory studies of the biological degradation of a number of hazardous organic compounds were reviewed by Kobayashi and Rittman (1982). They re- ported that the critical step in biodegradation of many chlorinated organic compounds (including 1 and 2-carbon aliphatics) appears to be reductive dehalogenation (i.e., replacement of Cl by H). This process would be favored only under anaerobic conditions where there is not a more energetically favorable electron acceptor (i.e., free oxygen) available. Reductive dehalo- genation appears to be carried out by the membrane-embedded electron transport chain of whole cells, but will occur with soluble electron transfer mediators (flavins, flavoproteins, hemoproteins, cytochromes, etc.) isolated from a variety of organisms. Parsons et.al. (1983, 1984) found that surface water and subsurface muck actively degraded tetra- and tri- chloroethylene, 1,1,1-trichloroethane and carbon tetrachloride. They incubated muck and surface water samples anaero- 36 ''bically with small amounts of the polychloroethylenes. About 67% of the tetrachloroethylene was degraded over a 2 week period. At the same time there was an increase in the amounts of trichloroethylene, cis- and trans- dichloroethylene, chloroethene (vinyl chloride), and dichloromethane present. The amounts of chlorinated organics observed after three weeks was less than that which was present initially. Besides loss of chlorine there was also a loss of carbon, indicating that other metabolites must also be formed. This data did not define the relationship between tetrachloroethylene and tri- chloroethylene degradation products. Kleopfer et.al. (1985) examined the degradation of trichloroethylene and found that under anaerobic conditions, trichloroethylene can be converted to 1,2-dichloroethylene but not to 1,1- dichloroethylene. Their experimental techniques did not differentiate between cis- and trans- forms of 1,2-dichloroethylene, however. Recent work by Barrio-Lage et.al. (1986) demonstrated the conversion of tetrachloroethylene sequentially to trichloroethylene and then to a mixture of all three possible dichloroethylenes, terminating with vinyl chloride. The sum of the products formed, however, did not account for all of the substrate removed, indicating the existence of other conversion mechanisms which do not terminate in vinyl chloride. These conclusions were confirmed by Vogel and McCarty (1985), using 140_tetrachloroethylene. There was considerable mineralization of vinyl] chloride to carbon dioxide during these experiments. This result indicates that there may be a possibility for enhancing the biotransformation of vinyl chloride by adding a primary nutrient to groundwater (and possibly to landfills), a subject which was explored during a recent seminar (McCarty, 1986). The evidence summarized above indicates strongly that a major route for formation of vinyl chloride in landfills is the anaerobic dechlorination of tri- and tetrachloroethylenes. In view of this it is worthwhile to summarize what is known of the ability of anaerobic bacteria to degrade organic materials. The conversion of more complex organic material to methane and carbon dioxide is termed anaerobic digestion. The overall process consists of a large number of reactions, each carried by a different organism or set of organisms, forming what is defined as an anaerobic consortium. In the break- 37 ''down of cellulosic or other polysaccharide material, three distinct stages have been identified in the overall conversion process. In the first stage, complex polysaccharides are hydrolyzed into their component sugar units, which are then fermented to simpler compounds -- mainly low molecular weight organic acids. The major acid components found are acetic, propionic, and butyric acids, with minor amounts of formic, valeric, isovaleric, and caproic acids frequently found. Transitory intermediates in acid formation, ethanol and lactic acid, are converted to acetic and propionic acids, respectively. The second stage of the conversion process is the production of acetic acid from the other intermediate acids by bacteria termed acetogens. Acetogens also utilize one carbon compounds such as methanol, converting it to acetic and/or butyric acids. Methanol is produced during the breakdown of lignin. Although certain acetogens will produce butyric acid as well as acetic acid, others produce acetic acid from butyric acid with the concomitant production of hydrogen and carbon dioxide. Some acetogenic organisms will use monosacchar- ides such as glucose or fructose to produce acetate directly. The third stage in anaerobic digestion is the production of methane and carbon dioxide from acetate by methanogens. Other methanogens will reduce carbon dioxide to methane using hydrogen as the reductant. One problem with anaerobic digestion to produce methane is that the acid-generating microbes prefer an acidic pH and methanogens prefer a pH between 6.2 - 7.8. The pH incompatibility of the two types of organisms can cause methanogenesis to be cyclically active in an anaerobic consortium. Temperature is another important parameter for methanogenesis. Methanogens fall into two classes with regard to preferred temperature: 30-38°C for mesothermophiles and 45-60°C for thermophiles (McCarty, 1964). Methanogens require a typical assortment of metal ions for growth and metabolism and about a 30:1 ratio of C:N. A source of phosphate and nitrogen are also required (Wolfe, 1971). Based on the literature, the involvement of anaerobic digestion bacteria in formation of vinyl chloride from chlorocarbons is highly probable. Identi- fication of which class of organism is primarily responsible would be a sig- nificant advance. Identification of individual species could then follow. 38 ''List of Probable Vinyl Chloride Emission Sources. The primary purpose of this survey was to permit tentative identification of the most likely sources of vinyl chloride pollution. Our experimental work was designed to test the various hypotheses with a view to confirming the most important routes. It is apparent from the foregoing, that if vinyl chloride is not ini- tially present as an important landfill chlorinated compound, that it is formed biologically from the anaerobic dechlorination of tetra- and trichloro- ethylenes. Indeed, in terms of amounts produced or emitted from a landfill] site, the biological route is more likely to produce greater amounts than the leaching route, simply because of the volumes of solvent typically present in a landfill. Other routes appear unlikely, based on the literature. Chemical production from chlorinated ethylenes or PVC requires high temperatures and/or extreme pH, while photochemical production in the presence of air leads to oxidized products without involving vinyl chloride as an intermediate. Thus we are only listing two sources of vinyl chloride as in any way likely, al- though experiments were performed to confirm the absence of other mechanisms. The likely sources are: 1. Biological dechlorination of chloroethylenes 2. Leaching There are subsets of these two mechanisms which would be worthwhile investigating. In the case of the biological route, anaerobic dechlorination appears far more likely than an aerobic route; in which case, identification of the responsible organisms as cellulolytic, acetogenic, or methanogenic would be useful. In the case of leaching, the question of origin from "old" landfill sites containing larger amounts of monomer, or from newer samples, or even by leaching from plastic film, needs to be addressed. A further question is whether all chlorinated solvents are degraded by the same route. Some indications are that chlorinated ethanes may be degraded by a different process than chlorinated ethenes. These questions and others which arose during development of our experimental plan will be discussed in the next section. ae ''''EXPERIMENTAL APPROACH The following discussion of our experimental work is based on a Work Plan which was generated in consultation with ARB staff prior to beginning actual experimental work on the project. It provides the background reasoning behind the choice of experiments, which were designed to eliminate unlikely possibil- ities for vinyl chloride formation (pyrolysis, photolysis, etc.), leading on to the experiments designed to test more likely mechanisms (biological degra- dation of TCE). The designation of "likely" or "unlikely" in turn was based on the results of the literature survey presented in the previous section. Equipment Assembly, Calibration, and Procedure Verification Initial work consisted of equipment set-up and testing. After evaluation of various methods available for quantitative determination of vinyl chloride in gas and water samples, we decided to apply a modification of EPA method 601. This standard method uses gas chromatography, with a packed column con- taining 1% of SP-1000 on Carbopack adsorbent. The unmodified Method 601 suffers from two serious problems: Poor resolution of volatile components, and the inability of the column to separate cis- and trans-1,2-dichloroethy]- ene, two biodegradation products of trichloroethylene (TCE) on the pathway to vinyl chloride. In order to be certain that GC peaks observed did in fact correspond to the compounds identified, we decided that a simple measurement of retention time was insufficient. Therefore, we procured and installed a photoionizat- ion (PI) detector in series with the normal hydrogen flame ionization detector. The PI detector is especially sensitive to chloroalkenes such as vinyl chloride, as well as sulfur compounds, and by using the two detectors in series it was possible to improve the probability of a correct identific- ation of an unknown peak on the GC. For more critical experiments which were performed to show that vinyl chloride was formed from TCE, simple identification of vinyl chloride in gas and water from treated landfill material was insufficient. This is because the background level of vinyl chloride precursors was sufficient to continue 40 ''generating vinyl chloride even after addition of TCE substrate. Therefore, the TCE was labeled with a non-radioactive carbon isotope (13), Formation of labeled products from the labeled TCE was taken as a conclusive ident- ification of origin. The isotope was determined in products by mass spec- trometry (MS). The GC/MS technique was also used frequently to confirm identities of peaks observed in the experiments. In one case (methanethiol) it avoided a serious error in identification, as methanethiol and vinyl chloride have similar GC retention times, and both respond well to the PI detector. Experiments With Primary Sewage Sludge In the work sequence originally proposed, analytical equipment was to be assembled and techniques for vinyl chloride measurement tested prior to starting work on landfill samples. Completion of this work was intended to be coincident with receipt of landfill samples from 2 sites. This sequence required the availability of a substitute material for testing and verifica- tion of analytical techniques, one which was readily available and as closely similar as possible to landfill material. Therefore, we began measurement work with a sewage sample obtained from an assumed "clean" environment, City of Richland anaerobic digester sludge. The Richland treatment plant uses a combination of aerobic and anaerobic digestion steps to treat wastes. The material used in our experiments was taken from the second stage of two in series anaerobic digestors. After bleeding off several liters of material from the recycle line of the digestor, a one-liter plastic bottle was filled and then capped, carried back to the laboratory, and then used the same day. This material is collected from a city of 30,000 with no heavy industry and no solvent manufacturing or using facilities nearby, and is primarily of residential origin. Local industry is heavily concentrated on nuclear power plant operation and waste disposal, and hence has its own segregated and remote disposal facilities. For analysis of sewage and landfill material incubates under aqueous conditions, we used the purge-trap system described by Dressman and McFarren (1977). A series of 40 ml vials was used for each experiment, including appropriate controls and blanks. The vials were teflon-capped to avoid 41 ''problems with extraction of plasticizers (mostly phthalates) from the plastic caps. The entire system was checked using standards made from purchased chlorocarbons and a diluted VC gas mixture (100 ppm of vinyl chloride in nitrogen). The purge-trap system was set up on a Hewlett-Packard 5830A GC and standards were used to confirm that we could obtain measurable response for 2 nanograms of vinyl chloride. Using the apparatus described above, we analyzed the initial experiments performed with Richland sewage sludge containing methanol as a carbon source for anaerobic bacteria, and inoculated with various chlorocarbons. These experiments served to establish our analytical procedures and provided much useful information about the behavior of the experimental system prior to work with actual landfill samples. Obtaining Landfill Samples Previous work on vinyl chloride emissions has been focused on two areas: Measurement of vinyl chloride and other organic materials in groundwater from or in the air over landfill sites, and microbiological work with non-landfil] material (e.g., Florida muck; (Parsons et. al.)). Since the purpose of this project was to identify the sources of vinyl chloride emissions from Cali- fornia landfills, it seemed important to do most of the work with actual landfill material. Under ideal circumstances, a range of landfills would have been sampled to various depths and in various locations. In order to perform the project in a reasonable time, however, we selected just two sites, and arranged with our subcontractor (Emcon Associates, San Jose, CA) to perform sampling at two locations within each site, taking two samples from each drilling at depths of 3 and 10 ft below the cover, plus surface soil. To further vary the sampling conditions, Site A was in Southern California, had been used for solvent dumping, and did have measured concentrations of vinyl chloride air emissions; Site B was in Northern California, had no history of chlorinated solvent waste dumping, and had had only traces of vinyl chloride identified in the groundwater. With these widely varying conditions we hoped to cover the spectrum of vinyl chloride production conditions (although a hazardous waste dump site was not included). The penalty for this is that with so few samples covering such a wide range, any 42 ''sort of a "representative" sample could not be hoped for. In fact, samples actually obtained varied considerably in appearance. The identities of the sites selected for sampling is not relevant to the work, and one site requested confidentiality. Therefore, the two sites are simply designated A and B. These samples of landfill material were subjected to the experimental procedures described below to determine which of several classes of mechanism was responsible for vinyl chloride formation. These are: Microbial, chemical, thermal, and photochemical. A total of 12 samples was obtained from the two sites. These were labeled at the site, and gas analysis, pH and temperature measurements were made at the various sampling locations and depths. Because of the importance of maintaining any microorganisms in a viable state, arrangements were made with the subcontractor that transfer to Battelle-Northwest would be made as rapidly as possible, at above 20°C in sealed containers and in an atmosphere of landfill gas. We therefore presumed that the samples represented actual landfill site environments fairly closely on receipt. The initial measure- ments consisted of a void space gas analysis from each sample container (before opening in an inert atmosphere) and a microscopic examination to determine the presence of living organisms. The samples were then flushed with argon/carbon dioxide to remove residual vinyl chloride, split, and placed into vials for experimental work. Description of Site A, Southern California Apart from the sewage sludge sample, which was obtained from a local treatment facility, we felt it important to perform most of our work with actual landfill material, even though such material is highly heterogeneous and variable. We were able to obtain permission for sampling from two land- fill operators: One in Southern and one in Northern California. Since the purpose of the sampling was to obtain research material and was not intended in any way to be a site survey, the actual locations of the sites are not important. However, because the sampling procedures may have some bearing on the results and because other researchers may wish to compare their results with ours, general site descriptions, details of sampling procedures, loca- tions within the sites, and on-site measurements are presented here. 43 ''At Site A, in Southern California, refuse was disposed of from 1973 to 1983. Figure 4 shows the landfill shape and the sampling locations (black dots). The maximum depth of fill is ~ 135 ft. and the average fill depth is ~ 10 ft. Site A was permitted to operate as a Class III solid waste disposal facility by the California Regional Water Quality Control Board. These wastes consist of commercial and household refuse, construction and demolition debris, and non water-soluble, non-decomposable and inert solid industrial wastes. Disposal of toxic wastes and wastes with a moisture content exceeding 50% was prohibited. Most of the landfill cover soil material was derived from a conglomerate formation consisting of well-rounded siliceous gravel and cobbles in a silty sand matrix. Typically fine-grained, the matrix contains 25-30% silt and clay-sized material. The gravel and cobble inclusions range in size from 1" to over 1 ft in diameter, with an average size of ~ 3". The fine sand and silt size materials within the matrix reduce soil permeability to the 107° to 107! cm/sec range. The site is equipped with a landfill gas recovery system. To collect samples representative of the landfill, sampling borings were located a minimum of 1500 ft apart. Three factors influenced the choice of drilling locations at this site: (1) The desire to avoid interference with potential future landfill gas recovery activity involving the gas well system, (2) the accessibility of drill rig and other vehicles to the sampling loca- tions, and (3) the presence of active decomposition within the landfill loca- tion. Drilling locations had previously (April 25, 1986) been identified by Emcon in collaboration with ARB staff during a site visit. A portable VOC monitor was used to confirm point #(3), above. Description of Site B, Northern California At Site B, refuse disposal without burning began in 1973 and continues at this time. The sampling area (Figure 5) comprised ~ 6 acres with an average fill depth of ~ 30 ft. Site B is also a Class III landfill. The refuse soil cover consists of clay with silt and some gravel. The boring locations were chosen for similar reasons as for Site A, that is, (1) to collect samples as representative as possible, (2) to avoid interference with 44 '' suoije907 Buljdwes - 71 I]!ypue] yo Wut] ayewixoiddy aul Ajuadoig BLUdOJSL [PJ UMBYyANOS ‘Y AILS LLLFPUeT “py sun peoy ssso0Vv S8d!IHO suoiesiado ypuey 45 ''suoi}e907 Bulj|dwes 2 | BLUUOJLL CDQ UUBYZAON * wi asnyey 9861 Pay jesodsiq aay g@ atts LLEspue) “G JuNdI4 46 ''refuse disposal activity, and (3) from an area with active decomposition of refuse occurring, as measured with a portable VOC monitor. Drilling locations were identified on May 2, 1986, by Emcon in collaboration with ARB staff during a site visit. The borings were located a minimum of 500 ft from each other and a minimum of 300 ft from the active fill area. The samples obtained, as for Site A, included two surface soil samples, plus landfill material from depths of 3 and 10 ft. below the cover. Four gas samples taken at these depths were also obtained (the gas sample from the surface samples was included with the surface solid samples and consisted of surface air). At the time of sampling, temperature and pH measurements were also taken. Oxygen levels in the gas from buried samples was higher than the 1.2% background normally experienced, although Emcon measured zero levels at all depths. However, the Site B cover was sandier than the Site A cover, and this may reflect a higher oxygen permeability in the landfill. Chemical Analysis of Landfill Materials Confirmation of activity: Samples were incubated in vials without additions, for one week, with removal and analysis of a gas sample to confirm the con- tinued production of vinyl chloride from the material. Inorganic analysis: Approximately 2 g of each material was required for analysis of inorganic composition, including sulfur and phosphorus. Of the two techniques available at Battelle-Northwest (Inductively Coupled Plasma Spectrometry (ICP), and X-ray fluorescence spectrometry), we chose the ICP method (procedure described in Appendix C). Organic analysis by GC/mass spectrometry: A further 10 g of each material was used for head-space analysis and identification and determination of baseline levels of chlorinated solvents in the landfill material, using the standard techniques already developed. Irradiation sterilization samples: Approximately 10 g was taken from each active sample as determined above, for irradiation and subsequent incubation and demonstration that no vinyl chloride was produced. If vinyl chloride was 47 ''still produced, it would indicate that a non-biological route for vinyl chloride production existed in the sample. Samples for pyrolysis in inert atmosphere: About 20 g of two samples was pyrolyzed to 650°C and the evolved gases collected and analyzed by GC/MS. Because of the expected high organic content of the landfill material, a complex GC trace was expected (and obtained), with several hundred compounds present. The pyrolysate gases were analyzed for vinyl chloride and related chlorinated compounds. A further sample was 'spiked' with PVC powder (10%) and pyrolyzed similarly. The pyrolysis was performed in a tube furnace and evolved materials collected and analyzed after heating to 150, 300, 450, and 650°C. The 150°C sample was expected to contain water and most of the chlorinated solvents. Samples for photolytic treatment in presence of TCE: To confirm extensive literature reports that photolysis of chlorinated compounds does not contri- bute significantly to vinyl chloride production, we subjected two samples of landfill material to UV light in the presence of 13¢_labeled TCE, and con- firmed by GC/MS that in the presence of air there was no production of vinyl] chloride. One sample was purged with inert gas prior to the addition of the labeled material and the other was photolyzed in air. Based on the literature review already performed, we expected the above preliminary experiments to have negative results. The evidence already avail- able clearly implicates vinyl chloride production as a result of anaerobic microorganism action on chlorinated solvents. In the event that any positive results were obtained, additional experiments would have been performed as follows: 1. Vinyl Chloride Formation from Irradiated (sterile) Samples A positive result here could indicate a non-biological formation mecha- nism for vinyl chloride formation, or it could indicate radiolytic formation from other chlorinated compounds, or extracellular enzymatic dechlorination of these compounds (enzymes are generally not inactivated by this dose of radiation). To determine which is the case, a further landfill sample would 48 ''have been purged with inert gas and autoclaved with 15 psi steam for 30 min., and again incubated with labeled trichloroethylene. This treatment would have inactivated any extracellular enzymes. Autoclaving itself should not produce vinyl chloride, as the expected product from trichloroethylene is a hydrolysis rather than a dechlorination product. (Lack of vinyl chloride formation from the pyrolysis experiment would have confirmed that heat alone was not responsible.) If vinyl chloride was still produced, this would have indicated a purely chemical formation route unaffected by radiation, heat, or steam, and hence likely to be a catalysis reaction. This is an unlikely result, but if it had occurred, further experiments would have been designed based on the ICP and organic compound analyses of the landfill material, plus a series of incubations of labeled trichloroethylene with pure inorganic components of the sample. These experiments would be designed to indicate if the reaction was catalytic in nature. Since a biological formation route is already clearly established in the literature, work on catalytic formation of vinyl chloride would have had to replace some biological experiments. The degree to which this would have been done would have been determined through consultation with the sponsor, as it would have involved a change in work scope and direction. 2. Vinyl Chloride Formation During Pyrolysis Formation of vinyl chloride during pyrolysis of PVC is known, although the levels of vinyl chloride produced are low (several ppm maximum). The only rational way for vinyl chloride to be formed during pyrolysis of landfil] material is through local combustion. It is possible to make a rough calcula- tion of the maximum theoretical amount of vinyl chloride that could be pro- duced in this way, from analysis of carbon and chlorine contents of the land- fill sample. Theoretically, pure PVC could depolymerize to 1 million ppm of vinyl chloride; in practice, no more than 6 ppm has been found after anaerobic pyrolysis and 35 ppm after pyrolysis in air. Taking the higher value, assum- ing that all carbon in the landfill sample is in the form of PVC, and that no other gaseous products are formed (even excluding methane and carbon dioxide), 1 ppm of vinyl chloride would be formed if there were 1/35 of the total land- fill sample in the form of PVC (2.9%). This is equivalent to 1.12% carbon. Where the calculation falls down is in the assumption that there is no air 49 ''flow over the landfill and that all gases are collected - the 1 ppm vinyl] chloride value is the total formed in a closed system if all of the PVC decomposes in a very short time. Any air movement will dilute the vinyl chloride, as will any formation of methane and carbon dioxide. Hence it is unlikely that the observed vinyl chloride levels (maintained over a period of years) are due to PVC decomposition. This can be shown easily if the measured landfill sample temperatures are initially below about 400°C, below which temperature vinyl chloride formation from PVC is insignificant. This turned out to be the case. Formation of vinyl chloride by pyrolytic dechlorination or dehydro- chlorination of organic solvents is a further possibility, and could involve catalysis by inorganic components of the matrix. If we had observed vinyl chloride formation in excess of that expected from PVC pyrolysis, this possi- ble mechanism would have been checked by '‘'spiking' landfill samples with individual solvents and pyrolyzing them. 3. Formation of Vinyl Chloride During Photolysis If vinyl chloride had been formed during aerobic photolysis of a landfil] sample containing 13¢_jabeled trichloroethylene, the experiment would have been repeated (a) anaerobically, (b) with variations in trichloroethylene concentration, (c) with photolysis times of 1-12 hr., and (d) with other solvents including (unlabeled) tetrachloro- and dichloroethylenes and tetrachloroethane. The experimental matrix was continued with the assumption that vinyl chloride is not formed by radiolysis, pyrolysis, or photolysis. The available evidence strongly supports a biological formation route for vinyl chloride, from dechlorination of chlorinated solvents by anaerobic microorganisms. Microbial Metabolism Experiments Anaerobic bacteria are responsible for the fermentation of organic wastes containing cellulose to methane and carbon dioxide; they are also capable of generating a wide variety of other compounds en route to methane. In general 50 ''they consist of three types: Those which degrade cellulose to glucose and related compounds, those which take the fermentation further to acetic acid (acetogens), and those which decarboxylate acetic acid to methane (methano- gens). Each type has its own range of discrete organisms with its own growth requirements, which are not always compatible. Also, some organisms eiter into a kind of interdependence whereby an excreted product from one becomes necessary for the growth of another. Although a great deal of work has been done on the classes of anaerobic bacteria, there is still much to be learned about their nature and mode of action. Of particular relevance to this project is their ability to anaerobically dechlorinate TCE and other solvents to vinyl chloride. Hence, one of the goals of the project was to attempt to isolate and identify one or more of the vinyl chloride-producing anaerobes present in landfill material, after first demonstrating that anaerobic orga- nisms are in fact responsible for vinyl chloride production in the landfills. From the above comments it can be seen that this is by no means an easy task. In order to achieve the project goals, beginning with proof of biological formation of vinyl chloride, we decided on the following experimental approach: 1. Set Up a Standard Inoculum of Microorganisms for Biological Experiments A stock culture of anaerobes derived from sewage sludge or landfil] material was necessary to ensure uniformity in performing microbial incubation experiments. Ideally, the inocula for different experiments should contain the same number of the same species of organisms per unit volume of inoculum. This was achieved by maintaining a stock culture in a 2 liter fermentor. 2. Use Sewage Sludge to Test Experimental Methods and to Show Vinyl Chloride Formation A description of the use of City of Richland sludge to test experimental methodology has already been given. As part of these experiments, we incu- bated Richland sludge with various chlorinated solvents and labeled TCE to determine whether or not vinyl chloride was formed. A positive result from these experiments indicates that the ability of anaerobic microorganisms to form vinyl chloride is (a) ubiquitous, and (b) constitutive in these bacteria, 51 ''because it may be assumed that these organisms had never been exposed pre- viously to significant amounts of chlorinated organic compounds. This preliminary work was designed to provide information on whether or not vinyl chloride can be formed from TCE by common anaerobes, and whether adaptation and prior exposure to solvents is necessary for activity. We hoped thus to confirm the mechanism which appears from the literature to be of primary importance in vinyl chloride formation. We also hoped to learn which of the three primary classes of anaerobic bacteria (hydrolytic, aceto- gens, or methanogens) are responsible for the conversion, by enrichment on glucose or methanol as carbon sources. Bis Incubation of Landfill Materials with Solvents Experiments similar in scope to those to be carried out using Richland sludge were planned for landfill material. These experiments were to be carried out using a standard inoculum containing mixed landfill organisms maintained on glucose or methanol in the presence of 200 ppm of one of a series of various chloroethenes including tetrachloroethylene, TCE, and dichloroethylenes. A standard procedure for these experiments is described in the next section. Generally, these inoculated vials (termed "microcosms" in the literature) were incubated at 35 or 55°C for 1-6 weeks, and the vinyl chloride and other chlorocarbon materials determined by the purge-trap method. Use of labeled TCE as a substrate permitted confirmation of labeled vinyl chloride formation in critical experiments. Characterization of Vinyl Chloride-Producing Organisms Initial work with sewage organisms was intended to permit us to determine the nutritional and physical requirements for chloroethylene degradation. Temperatures of 45-60°C and 30-38°C favor different groups of anaerobic orga- nisms. Both temperatures are possible in landfills, the higher value being typical of freshly fermenting material, and the lower typical of established volumes where most of the cellulosic material has already been converted. Carbon in these experiments was supplied as methanol or acetate. Glucose will support a food chain of all three general types of anaerobes. Methanol 52 ''will support a food chain of acetogens and methanogens, and acetate will support methanogens only. An attempt to isolate the microorganisms responsible for vinyl chloride formation was included in the planned work, because of its importance, al- though it is a difficult procedure and somewhat “open-ended". Standard microbiological procedures will allow isolation of anaerobes, which sometimes can be identified as previously known species. However, these experiments are time-consuming, and subject to interruption from unknown factors which may cause cultures to fail to grow, or to "crash" after initial growth. Organisms in anaerobic environments are generally delicate and much less wel] understood than aerobic equivalents. The procedures followed are summarized thus: Plating sample cultures onto Petri dishes allows individual organisms to be separated. Petri dishes were incubated anaerobically using an anaerobic GasPak system. This system is preferred over an anaerobic incubator because it allows stacking of 10-15 plates in an individual container; observations can be made without disturbing the atmosphere over other experiments as would occur with a larger incubator. Also, this system offered a choice of using a pure carbon dioxide or a mixed carbon dioxide/hydrogen atmosphere. These small units can be filled inside an anaerobic glove bag to avoid air contami- nation, and then removed to the constant temperature incubator. Individual microbial colonies can be cultured separately and these cultures then screened for the ability to degrade chloroethylenes. Microscopically unique isolates found to degrade chloroethylenes can be typed using the procedures in Bergey's Manual of Determinative Bacteriology. If we found several organisms which could degrade chloroethylenes then a large number of isolates would have had to be examined microscopically or by biochemical tests to give an idea of the population distribution of active anaerobes in the samples. Initial experiments were performed with landfill sample microbial iso- lates to determine their temperature range (thermophilic (45-60°C) or meso- philic (30-38°C) and metabolic type (cellulose-degrading, acetogenic, or methanogenic). The temperature of the samples removed from the landfill was measured at the time of landfill sampling. Subsequent incubation of the samples with labeled TCE allowed correlation of vinyl chloride formation with specific sample temperatures and carbon sources. Streaking and individual 83 ''colony isolation of single species was then attempted on agar plates. Incuba- tion of colonies thus isolated with labeled TCE allowed us to define which isolates showed activity in producing vinyl chloride. These isolates were then cultured in liquid media to obtain information on their nutritional requirements. Because of the amount of work involved in this portion of the project, and the difficulty of estimating time required (due to the wide range of microorganism characteristics in landfill material) an accurate pre- diction of how many organisms could be isolated and identified was not possi- ble. We hoped to identify 2-3 of the major contributors to vinyl chloride production, with growth on two defined published media. Detailed nutritional studies on these organisms were beyond the scope of the project. One inter- esting question that can potentially be answered using an anaerobic reactor and a defined inorganic medium is whether or not the organism uses chloro- ethylenes for energy metabolism or as a carbon source for growth. In the latter case, the organism(s) can rather easily be isolated by enrichment on a mineral salts medium containing a chloroethylene. Initially, a good carbon source is added (e.g., acetate), and this is gradually depleted until only the chloroethylene is available as a carbon source. An aliquot of a growing culture is added to a continuous fermentor under conditions where only growing organisms are maintained. The dilution rate is then gradually increased until washout of the cultured organism begins to occur. At this point the dilution rate is stepped back to the last stable value and held until equi- librium is achieved. The organism metabolizing chloroethylene will thus have been isolated, or at least significantly enriched. The organism can then be typed. The above experimental approach for isolating vinyl chloride producing organisms was followed initially. Subsequent experimental work though led us to modify our techniques. The most important of these changes was the use of the roll-tube technique for the growth and isolation of strict anaerobes for isolation of single microbial colonies. A second modification was in the use of a stirred tank reactor to maintain anaerobic organisms at high levels for experimental work. Neat methanol was added to the reactor and gas formation monitored to maintain the organisms and determine their state of vigor. Details of experimental procedures used to perform the experiments outlined above are presented in the next section. 54 ''''MATERIALS AND METHODS Material Sources, Composition, and Purity Vinyl Chloride: Cylinders containing 1, 10, and 100 ppm vinyl chloride standard gas in nitrogen were obtained from Scott Specialty Gases (Ann Arbor, MI). The manufacturers' analysis is reproduced below: Mixture 70 0.9917 ppm VC Mixture 71 10.02 ppm VC Mixture 73 100.4 ppm VC Accuracy of the analysis was +/- 2%, traceable to NBS standards. The purity of the vinyl chloride in this standard gas mixture was confirmed by purge-trap analysis and GC/MS analysis of the purged (concentrated) product. Chlorocarbon Standards: These were obtained individually from Aldrich (Milwaukee, WI), and were of the highest purity available commercially (Gold Label grade). To check on calibrations for purge-trap analyses, standard mixtures of chlorocarbons were obtained from Alltech Associates (Deerfield, IL), and Supelco (Bellefonte, PA). The composition shown in Table 7 is of the standard mixture. Water: Water used in all experiments was deionized in our laboratories and was free of organic compounds, with a conductivity of >16 Megohm-cm. It contained no chlorocarbons to the limit of our instrument sensitivities, as determined by a purge-trap analysis of a water blank. The water was auto- claved for 15 min at 15 psig of steam for use as sterile water. Site Sampling Procedures Sampling requirements were specified as follows: At each of two locations at each site, one sample of cover soil and two landfill material samples were to be taken, together with associated gas, and measurements of pH and temper- ature. The landfill samples were to be as close to 3 and 10 ft below the cover, respectively, as possible, and were not to consist of a single material (e.g., rock or paper). In total, two soil samples, four refuse samples, and four gas samples were to be obtained at each site. This plan was followed for both sites, but practical difficulties made some variations inevitable. BS ''Table 7: Components in Purgeable Halocarbon Mix 60/M2, Lot LA 13373. Compound Concentration ~ Re (Lit.) (mg/m1) (min) Dichloromethane 0.60 6.6 1,1-Dichloroethylene 0.12 8.5 1,1-Dichloroethane 0.12 9.5 Trans-1,2-dichloroethylene 0.12 10.1 1,2-Dichloroethane 0.12 11.2 1,1,1-Trichloroethane 0.12 12.0 Tetrachloromethane 0.12 12.6 1,2-Dichloropropane 0.12 14.0 Trans-1,3-dichloropropylene 0.19 14.0 Trichloroethylene (TCE) 0.12 14.6 Cis-1,3-dichloropropylene 0.05 15.0 1,1,2-Trichloroethane 0.12 15.0 1,1,2,2-Tetrachloroethane 0.12 19.0 Tetrachloroethylene 0.12 19.0 Chlorobenzene 0.60 21.5 1,2-Dichlorobenzene 0.60 22+ 1,3-Dichlorobenzene 0.60 22+ 1,4-Dichlorobenzene 0.60 22+ + Stock # 4-8747. 1-!3c-Trichloroethylene (1-3c-Tce, 100 mg) was obtained from Merck, Sharp, and Dohme (Quebec, Canada). A CME 550 All-Terrain model drill rig with a 6" solid-flight auger was used to collect cover soil and refuse samples. The augers and bit were steam- cleaned before each borehole was drilled. The cover soil samples were collected from the flights of the auger as soon as the cover layer was pene- trated. The temperature of the cover layer at the refuse/cover interface was measured with an Electrotherm SH44 digital thermocouple sensor, pre-calibrated against a standard mercury thermometer, and the cover thickness recorded. Drilling then proceeded to the first refuse sampling point, at ~ 3ft. depth. The auger was retrieved and a refuse sample collected with a split-spoon sampler, which collects a sample from a refuse interval of ~ 18". The sampler also protects the sample from excessive aeration after retrieval. To provide an oxygen-free environment for the sample after retrieval the sample was immediately placed in a bag filled with landfill gas (55% CH,:45% CO.) 56 ''collected from the landfill gas test well system. The sampling procedure was chosen to minimize disturbances to the native anaerobic bacteria in the refuse by either oxygen infusion or high temperatures. Generally, a few seconds exposure to air is not considered harmful to most anaerobic species; also, collection of a solid from a tube bore sample keeps most organisms covered and thus protected from air during the short time required for transfer to the anaerobic environment of the sample bags. After the first sample was taken, a second refuse sample was retrieved from the boring immediately below the refuse sampling location, for measurement of pH. A Hach model 17200 porta- ble pH meter was used, after calibration in the field with pH 4 and pH 7 solutions. The temperature at the bottom of the boring was then measured using the digital thermocouple sensor. After the refuse sample was obtained at the first sampling point, drilling was resumed until the second sampling point was reached, where the sampling procedure was repeated. After refuse samples were collected, the resulting borehole was used to collect gas samples. Gas monitoring equipment used on-site consisted of a Gastech GX-3N instrument, with a capacity for measurement of combustible gas in the 0-5 and 0-100% ranges, and oxygen in the 0-25% range. It was pre- calibrated against 45% methane/55% carbon dioxide. Gas sample bottles at Site A were Whitey 1000 ml 304 stainless steel cylinders with Whitey 14-DKM 454 316 stainless steel shut-off valves. The manufacturer's data states that these cylinders were cleaned with TCE before shipment. To ensure cleanliness, the cylinders and valves were triple-rinsed with TCE and heated for 60 hr at 105°C to remove any solvent residue. At Site B, due to a late delivery of the gas sample cylinders from Site A to Emcon, different sampling cylinders were used. These comprised 3 Whitey 200 ml sampling cylinders with the same specifications described above, and 5 Kildee Scientific 500 ml glass sampling bottles. Approximately 2 ft of gravel was placed at the bottom of the boring, then a gas sampling probe tube was inserted into the boring to permit collec- tion of a gas sample at the deeper location. The gas sampling probe is shown in Figure 6. It consisted of 1/4 in 304 stainless steel tubes with a 0.065 in wall thickness, in 5 ft sections coupled with 316 stainless steel unions. After gas sample collection, the hole was then backfilled with cover soil to the elevation of the first refuse sample, where additional gravel was placed and the gas sample at 3 ft depth obtained. 57 ''FIGURE 6. Gas Sampling Probe with paimt. Shallow probe marked lastic vault box Piug(Typ.) E : % QQ COVER SOIL 9 “wo 1 ‘wo j ‘t _! - j 4 NN _ le wm) > = “se iL ° NS Ve COL oy SoG Ss ral. Aang mi PN an hy Nake Ls Soil and refuse bockfill 316 S.S. Union (Typ.) QO NOT TO SCALE oO fos ge 304 S.S. Tubing 4° OD 0.065" wall thickness Gravel Slotted end dae 5/86 fas’ EMCON ay Associates BATTELLE NORTHWEST LABS VINYL CHLORIDE LANDFILL METHANE RECOVERY ASSESSMENT STUDY Ee sy GAS SAMPLING PROBE CONSTRUCTION PROJECT WO 494 - 03.01 58 ''Gas samples were collected by connecting a 1000 ml stainless steel or glass gas sampling bottle directly to the sampling tube and then purging the sampling bottle. Purging continued until the methane content reached a con- sistent value as indicated by portable gas monitoring equipment. Duplicate samples were collected. The gas samples did not contact any plastic material after entering the stainless steel sampling tube. A plastic vault box was installed over the sampling tubes for protection in the event that additional gas samples were required. The location of the boxes on the site was clearly marked with lathe and marker tape. All equipment was cleaned and calibrated before sampling, as described above. Laboratory Instrumental Analysis Systems: Gas Chromatography: A Hewlett-Packard 5830A instrument was used for gas chromatography. It was fitted with a photoionization (PI) detector in series with a flame ionization detector (FID). The PI detector was obtained from HNU Systems, Inc. (Newton, MA). Tubing connections were 1/16 in diameter deactivated stainless steel to minimize void volume between the column exit and the detector chambers. Initial work was performed using the EPA 601 method, which requires a packed glass column (180 cm x 4 mm i.d.) containing 1% SP-1000 (60-80 mesh) on Carbopak B. This column diameter was used for purge-trap analyses; for direct injection analyses, a similar column but with an i.d. of 2 mm was used. A 180 cm x 4 mm glass column packed with 1/3 Chromosorb 101 and the remaining 2/3 with Chromosorb 102 was used for all subsequent analyses. This column gave good resolution of volatile compounds and separated cis- and trans-1,2- dichloroethylenes. Temperature and flow conditions were as follows: Initial, 50°C for 0 min, programmed at 10°C/min to 180°C; held for 20 min. Mass Spectrometry: A Hewlett-Packard 5992 GC/MS combination with an all-glass jet separator was used for GC/MS work. The GC was fitted with a 180 cm x 2 mm i.d. Chromosorb 101/102 column. A sample spectrum of vinyl chloride standard injected into this system is shown in Figure 7. 59 ''Figure 7: Mass Spectrum of Vinyl Chloride Standard. Retention Time = 4.8 minutes Scanned from 33 to 200 amu Number of Peaks Detected = 20 File Type = Linear Base Peak=61.90 Base Peak Abundance=349 Total Abundance=674 le om. i oT ie wv T | v qv q q | q qv q Tq | 0 50 100 150 Lower Abundance Cutoff Level = 0.0% Mass Abundance (%) Mass Abundance (%) Mass Abundance (%) 34.90 7.7 46.80 4.0 60.90 9.7 35.90 2.3 47.90 3.7 61.90 100.0 36.90 3.4 48.90 4.0 62.90 5.7 37.80 1.1 49.90 2.3 63.90 30.9 39.90 1.4 50.90 0.9 83.80 0.3 41.80 0.9 58.90 2.9 85.80 2.3 43.90 2.0 59.90 7.4 Purge-Trap Apparatus: A diagram of the purge-trap apparatus is shown in Figure 8. This is similar to the apparatus described by Bellar et.al., (1976). Operating procedure was similar to that described by Dressman and McFarren (1977). The packing for the trap used in the apparatus was modified somewhat to allow efficient trapping of vinyl chloride and elution of higher molecular weight chlorocarbons. Experiments with Tenac, Sperocarb (charcoal), or silica alone showed that none of these packings by itself would give the desired retention times. A 7 cm length of Tenac backed by a 7 cm length of charcoal was finally selected as the optimum. Operation of the purge-trap assembly was as follows (with reference to Figure 8): The purge assembly was mounted and system pressure released by actuating the three-way valve A. After insertion 60 '' Jaquosag snqeurddy deuj-abung 40 ILZeWaYIS deay se|pean—> aBung LJ __J Vv “8 3YyNDI4 IEIA ajdwes L J aajeA Aen-€ 61 ''of the gas purge needle into the sample vial (a standard 40 ml septa cap vial with ~ 10 ml head space), the three-way valve was turned to the purge position and a flow of 60 psi nitrogen was maintained at 20 ml/min for 10 min, to carry the volatiles into the trapping column. The purge nitrogen was a prepurified grade further purified by a molecular sieve drying trap. It contained no detectable chlorocarbons as determined by purge-trap analysis of a blank sample. The third position of the three-way valve is for post-purge, performed in order to remove excess water. Post-purge was not necessary when the Chromosorb 101/102 column was used. After absorption the head pressure on the column was reduced to zero and the septa quick-disconnect (B) was disconnect- ed. The purge-flow assembly was then mounted on the GC. Pressure on the desorber flow gas was vented by pressing the button on the quick-disconnect, which was then attached to the purge-trap. Desorption of vinyl chloride onto the GC column was achieved by a 3 min elution at 150°C, with a 60 psi, 20 ml/min carrier gas flow. During this period the GC column was maintained at room temperature. The oven door was closed and both detector integrators turned on. The water peak appeared first (negative peak on the PID). The trapping apparatus was then reconditioned for the next run. This was achieved by heating for 30 min at 180°C with a flow of 20 ml/min of nitrogen. The efficacy of the purge-trap technique for vinyl chloride analysis was determined to be 100 +/- 2%, using pure water standards. Chlorocarbon stan- dards were prepared at the same concentrations as used in incubation experi- ments. Spike solutions were prepared by dilution of 100 mg of the test compound in 10 ml of pure methanol. The actual concentrations resulting are given in Table 8, together with GC retention and GC/MS peak data. For the analysis, a ~ 10 ul sample of the diluted chlorocarbon was added to ~ 30 g of ultrapure water in a 40 ml septum-capped vial. Negative peaks on the resulting GC traces at 2.5 and 4.9 min were from water and methanol, respectively. The GC column (see above) was packed with Chromosorb 101/102, which separates vinyl chloride from the water peak and also separates all 3 dichloroethylene isomers. 62 ''Table 8: Chlorocarbon Standard Solution Concentration and Analytical Data. Compound Concn. GC/MS Ry Ions Used GC PID/FID Ry (ug/ml) = (min) (m/e) Col. 1* Col. 2 Vinyl Chloride 0.031 5 «1 62,64 6.5 Zul 1,1-Dichloroethylene 12.43 9.2 61,96,98 10.7 7.8 Trans-1,2-Dichloroethylene 9.18 10.3 61,96,98 11.7 9.9 1,1-Dichloroethane 9.73 11.1 63,65,83 12.6 9.1 Cis-1,2-Dichloroethylene 10.28 11.3 61,96,98 12.8 9.9 1,2-Dichloroethane 11.06 12.8 62,64,98 14.4 11.2 1,1,1-Trichloroethane 10.89 12.9 61,97,99 14.8 12.3 Trichloroethylene 10.81 13.6 95,97 15.6 15.3 3 130,132 1-"~C-Trichloroethylene 7.09 13.6 96,98 15.6 15.3 131,133 1,1,2-Trichloroethane 11.68 16.3 83,97,99 20.8 16.0 Tetrachloroethylene 11.05 16.8 129,131 21.0 20.8 166 1,1,2,2-Tetrachloroethane 12.41 >30 83,85, 168 3300 20.8 * Column 1, Chromosorb packing; Column 2, SP-1000 on Carbosphere packing. Elution of Chlorocarbons from samples with a Purge-Trap Apparatus: Mixtures of vinyl chloride and other chlorocarbons were injected, collected, and purged from the apparatus and % recovery determined over several purging cycles. The recovery of these materials was less than 100% on a single pass. More details on this experiment are given in the Results section. Systems were calibrated using pure water. Chlorocarbon recovery from samples including sewage or landfill material was much less due to adsorption on solid material. Head-Space Analysis of PVC, Saran, and Polyethylene for Residual VC Monomer: Samples of PVC, Saran, and polyethylene were obtained from existing laboratory supplies. The amounts of residual VC monomer were determined by an indirect headspace method. The PVC pipe was ground to a fine powder before headspace analysis. One sample was >10 years old and the second sample was estimated at <1 year old. The Saran wrap and polyethylene garbage bag were used as- received. The procedure involved placing 1.0 g of material in a septum-capped vial, heating the vial to 110°C for 1 hr, then removing the headspace gas by means of a gas-tight syringe. The headspace gases were analyzed by GC using 63 ''both PI and FI detectors and by GC/MS. The detection limit for residual VC monomer was 0.2 ug/g of polymer. Pyrolysis Apparatus: This consisted of a 0.5 x 6 in pyrex pyrolysis tube heated by a tube furnace. A sample of landfill material (~ 20 g) was placed in the apparatus, which was then flushed with argon at a flow rate of 1 ml/min. The sample was heated to 650°C over 45 min. Additional experiments were performed with landfill material mixed with PVC powder. Pure PVC without landfill material were also pyrolyzed. Gas samples were collected in a gas manometer trap, and were then analyzed by GC/MS for organic compounds, and by GC for major gas components. Photolysis Apparatus: The apparatus used for photolysis consisted of a 15 ml quartz cell containing ~ 5 g of test material. The test material consisted of landfill surface cover to which 10 ul of labeled TCE was added. The cell was flushed with argon for the inert gas photolysis experiment. Normal air was used for the landfill surface simulation experiment. Each matrix was irradi- ated by a 450 W Hanovia UV lamp for 2 hr. (These conditions are known from the literature to be sufficient to cause TCE photolysis and are considerably more intense than normal California surface sunlight over a 1-day period.) Inclusion of the surface soil was necessary to test for any photocatalytic effects. Irradiation of Landfill Materials: Gamma irradiation is a way of cold-sterilizing landfill material. Autoclaving in steam is the standard approach which was also used, but this has the disadvantage that at the autoclaving conditions (121°C for 30 min at 15 psig) chemical changes can occur which could conceivably lead to the hydrolytic formation of vinyl chloride. Gamma irradiation was performed on a total of 16 samples, which each received 2.5 megarad from a cobalt-60 source. The samples so treated were then incubated with 136 Tce to determine if there was any degradative activity. Inorganic analysis was performed by the Inductively Coupled Plasma (ICP) technique described in Appendix C. 64 ''Chlorocarbon Hydrolysis Experiment Sterilized blank samples were prepared to determine if the test chloro- carbons can react directly with water to produce vinyl chloride. The com- pounds tested were 1,1-dichloroethylene, 1,1-dichloroethane, 1,2-dichloro- ethane, 1,1,1-trichloroethane, 1,1,2-trichloroethane, and 1,1,2,2-tetrachloro- ethane. Deionized water and sample vials were sterilized in an autoclave, substrates inoculated into the vials, and the vials incubated inside a dark growth chamber at 35°C. Individual samples were removed periodically and analyzed by the purge-trap technique to determine the degree of hydrolysis under incubation conditions. Incubation of Sewage Sludge With Chlorocarbon Substrates A set of 48 vials was inoculated with 30 ml of sewage sludge diluted 10:1 with deionized, organic-free water and 10 U1 of chlorocarbon in methanol with methanol as a carbon source for bacterial metabolism. The source of the sewage sludge was the #2 anaerobic digester at the Richland Wastewater Treat- ment Plant. Sample identification and the chlorocarbon added to each vial are listed in Table 9. Care was taken in this experiment to exclude all oxygen from the test vials. Controls consisted of water and methanol only (no chlorocarbon or sludge). Duplicate vials were sampled by the purge-trap technique after 1, 3, 6, and 10 weeks of incubation at 35°C. The sample and experiment identification system used throughout this work is explained as follows: Experiments were designated according to the microorganism source (Richland, R), date of initial inoculation (1-20 = 1-20-86), an abbreviation for the chlorocarbon (for example, VC = vinyl chloride), and further abbreviations for water- and methanol- only controls. This system was followed throughout the project for QA purposes and for ease of sample identification. (Blanks were designated B). A second set of screening experiments was performed (designated as R-2-26-**) (Table 10). 65 ''Table 9: Anaerobic Sewage Sludge/Chlorocarbon Incubation Experiment Vial # Substrate Chlorocarbon Designation 1-8 1,1-Dichloroethylene 9-16 Cis-1,2-Dichloroethylene 17-24 Trans-1,2-Dichloroethylene 25-32 Trichloroethylene 33-40 1,1,2,2-Tetrachloroethane 41-48 Vinyl Chloride 49-56 Water blanks 57-64 Methanol blanks AD ADD Ao ed eo) 2 Ree eee ee ' RO RO PD PPD PD PD PO So ! + 7A 0-BMe R = Richland sewage sludge inoculum; 1 = Month #1; 20 = Day inoculated (1-20-86); last two symbols indicate chlorocarbon added (list below), or whether water (BW) or methanol (BMe) was added to the vial as a control blank. Chlorocarbon abbreviation list: VC = Vinyl Chloride; CD = Cis-1,2- dichloroethylene; TD = Trans-1,2-dichloroethylene; VI = Vinylidene chloride (1,1-Dichloroethylene); TR = Trichloroethylene PC = Perchloroethylene (Tetrachloroethylene); DE = 1,1-Dichloroethane; DC = 1,2-Dichloroethane; MC = 1,1,1-trichloroethane; TE = 1,1,2- Trichloroethane; PE = 1,1,2,2-Tetrachloroethane. Table 10: Sewage Sludge/Chlorocarbon Incubation Without Oxygen Exclusion. Vial # Substrate Chlorocarbon Designation 1-8 Vinyl Chloride 9-16 1,1-Dichloroethane 17-24 1,2-Dichloroethane 25-32 1,1,1-Trichloroethane 33-40 1,1,2-Trichloroethane 41-48 Lab. TCE 49-56 Tetrachloroethylene 57-64 Water blanks 65-72 Methanol blanks FD FI ZO AE A eS A oe WWMYNMNN DN PD P ! PRR M DM MDM DPD PP oO | o m * * Lab. TCE refers to 13¢_}abeled TCE in this experiment. Vinyl chloride was added from a stock solution made by bubbling 100 ppm vinyl chloride in nitrogen into methanol at -196°C and determining the VC concentration by GC. 66 ''The samples were prepared in the same manner as the first experiment. A fresh sample of waste water from the anaerobic digester at the City of Richland sewage treatment plant was obtained for inoculation of the test vials. The chlorocarbons used in these samples were 1,1-dichloroethane, 1,2-dichloro- ethane, 1,1,1-trichloroethane, 1,1,2-trichloroethane, tetrachloroethylene, and 13¢_labeled TCE. Separate samples were inoculated with methanol (control) and vinyl chloride, to determine their rates of degradation under these conditions. The results from these screening experiments were not as reproducible as expected, so another sample preparation procedure was employed. The sample preparation procedure employed sewage sludge diluted 5:1 with deionized water and no attempt was made to initially exclude oxygen from the vials. The vials were incubated for one week in which time the bacteria present had removed all the oxygen from the vials. They were then injected with the chlorocarbon diluted in methanol through the septa, inverted, and placed back into the growth chamber. Twenty-four vials (designated as R-3-4- **) were prepared by this procedure, 10 each with 1,1,1-trichloroethane and 1,1-dichloroethene spikes and 4 left as unspiked blanks. This sample preparation procedure resulted in much better reproducibility and was used for all subsequent sample preparation. Optimum incubation temperature was determined by incubating 110 sample vials at 35 or 55°C. For growth at 35°C, 44 samples were prepared (R-4-4-** -35) with 12 13 _¢-TCE spiked, and 8 each with cis-1,2-dichloroethylene, trans-1,2-dichloroethylene, VC, and water blanks. For growth at 55°C, 40 samples were prepared (R-4-4-**-55) with 10 each of 1,1-dichloroethylene, TCE, and VC spikes, and 8 each of 130_tCE-spiked, and methanol blanks. Another 20 vials were prepared using material from a 1 liter anaerobic reactor held at 55°C for 2 months and with methanol as carbon source. The vials (A-4-4-**) 13¢_tce, cis-1,2-dichloroethy1- ene, trans-1,2-dichloroethylene, and 4 water blanks. These vials were placed were spiked 4 each with 1,1-dichloroethylene, in a growth chamber with the other vials incubated at 55°C. Blanks (B-3-12-**) were prepared using sterilized water spiked with 1,1- dichloroethylene, 1,1-dichloroethane, 1,2-dichloroethane, 1,1,1-trichloro- 67 ''ethane, 1,1,2-trichloroethane, and 1,1,2,2-tetrachloroethane. The samples were incubated at 35°C for up to 6 months. In a further control experiment, a series of vials containing sewage sludge was autoclaved twice (1 day apart) and incubated for 2 months after inoculation with 10.0 wl methanol and 10 yl of labeled TCE or unlabeled 1,1,1- trichloroethane dissolved in methanol. Incubation was at 35 and 55°C for one-half of each subset. Incubation of Landfill Material With Chlorocarbon Substrates. Sample Receipt, Storage, and Vial Preparation: Upon receipt at Battelle, the landfill samples were immediately examined to make sure that they were undamaged and not exposed to air. They were then transferred to an anaerobic glove bag filled with 50% argon and 50% carbon dioxide. The headspace oxygen content and physical appearance of the samples in the individual sample bags were determined and recorded. (Oxygen content was determined using a Carle Series 400 GC system, by injection of 0.7 ml of headspace gas.) Gas samples were removed by syringe directly from the samples, for headspace analysis for chlorocarbons and other organic volatiles. Portions of the landfill material were removed from each sample and checked for visual homogeneity. (Highly unrepresentative samples containing large rocks or samples containing only one material such as paper were returned to their respective bags and more representative material taken). The appropriate amount of landfill material was then placed in a labeled vial (from which air had previously been flushed out by the inert gas mixture in the glove bag) and the vial was capped and sealed before being removed from the glove bag. Sample weights and original sources for each vial were noted. These samples were used for pyrolysis, photolysis, irradiation, and incubation experiments as noted in the Tables below. Unused landfill materials were stored in their original containers at 20°C in a dark room in case additional samples were needed. Landfill Material Incubation Experiments: About 10 g of each landfill sample was taken from the bulk core material in an atmosphere of 50% argon/50% carbon dioxide in a glove bag as described above, weighed, and used directly. 68 ''The assay system for determining chlorocarbon conversion is similar to that described above for sewage incubation experiments and is based on the methodology of Parsons et.al. (1984). Reaction vials containing the landfill sample and having a capacity of 40 ml were filled to 30 ml with sterilized water in a glove bag containing a mixture of 50% argon:50% carbon dioxide. The vials were removed from the glove bag and placed in a growth chamber held at 35°C. The chlorocarbon substrates, dissolved in methanol, were added by syringe through the septa after 1 week of incubation. The reaction vials were incubated without agitation at a temperature of 35°C). Incubations were terminated after 1, 3, 6, and 10 weeks. Each sample was prepared in dupli- cate. The solvent (methanol) was HPLC grade to avoid problems with contami- nants. The vial caps have an opening in them and a teflon/silicone septum (to eliminate extraction of plasticizers - phthalates - into the vial liquid contents). At the termination of the incubation period the chlorocarbon substrates and products were analyzed by the purge-trap technique. Tables 11 and 12 provide information on the microcosm compositions in vials used for incubation of landfill samples from Sites A and B (Not all samples incubated were analyzed, for reasons of time and economy and to avoid redundant data. The most active samples were determined by the amount of gas production). Table 11: Landfill Site A Incubation Experiments Sample/Vial # Vial Content Sample/Vial # Vial Content A-1-S-1 to 6 Blanks (BW, BMe) A-1-3-1 to 6 Blanks (BW, BMe) A-1-S-7 to 12 Not done A-1-3-7 to 10 VI A-1-S-13 to 15 ICP Analysis A-1-3-11 to 14 TCE A-1-S-16 to 17 Pyrolysis A-1-3-15 to 22 Lab. TCE A-1-S-18 to 19 Photolysis A-1-3-23 to 25 ICP Analysis A-1-S-20 to 22 Headspace gas A-1-3-26 to 27 Pyrolysis A-1-S-23 to 24 60-Co steril'n. A-1-3-28 to 29 Photolysis A-1-3-30-to 32 Headspace gas A-2-S-1 to 24 As above. A-1-3-33 to 34 60-Co steril'n. A-1-10-1 to 34 As above. A-2-6-1 to 34 As above. A-2-16-1 to 34 As above. S = Surface cover sample; A-1-, A-2- refer to core #1 and #2, respectively; A-1-3-** = 3 ft. depth sample; A-1-6-** = 6 ft; A-2-16-** = 16 ft depth. Lab. TCE = labeled TCE. 69 ''Table 12: Landfill Site B Incubation Experiments Sample/Vial # Vial Content Sample/Vial # Vial Content B-1-S-1 to 6 Blanks (BW, BMe) B-1-4-1 to 6 Blanks (BW, BMe) B-1-S-7 to 12 Not done B-1-4-7 to 10 VI B-1-S-13 to 15 ICP Analysis B-1-4-11 to 14. TE B-1-S-16 to 17 Pyrolysis B-1-4-15 to 22 lab. TCE B-1-S-18 to 19 Photolysis B-1-4-23 to 25 ICP Analysis B-1-S-20 to 22 Headspace gas B-1-4-26 to 27. Pyrolysis B-1-S-23 to 24 60-Co steril'n. B-1-4-28 to 29 Photolysis B-1-4-30-to 32 Headspace gas B-2-S-1 to 24 As above. B-1-4-33 to 34 60-Co steril'n. B-1-10-1 to 34 As above. B-2-3-1 to 34 As above. B-2-10-1 to 34 As above. S = Surface cover sample; B-1-, B-2- refer to core #1 and #2, B-1-4-** = 4 ft. depth sample; B-1-10-** labeled TCE. Lab. TCE = 10 ft; B-2-3-** = 3 ft depth. TE = 1,1,1-Trichloroethane. respectively; Incubation of Landfill Material with Labeled Trichloroethylene: An important experiment was designed to show that vinyl chloride is in fact produced from the added chloroethylenes and does not accumulate from an endogenous pool within the core sample. To demonstrate that exogenously supplied chloroethylenes are converted to vinyl chloride, carbon-13 labeled TCE was incubated with reaction mixtures derived from core sample material in water. Carbon isotope ratios of all chloroethylene products including vinyl chloride were determined by GC/MS. An accumulation of carbon-13 enriched products would demonstrate that the biological reaction is occurring in the landfill core samples. There was no plan to use other labeled precursors (e.g., dichloroethylenes) if the labeled TCE led to the formation of labeled dichloroethylenes and labeled vinyl chloride, since this experiment would have conclusively established the interconversion between these materials. An additional factor in this decision was the very high cost of labeled dichloroethylenes. The experiment consisted of 8 vials inoculated with material from each landfill site sample. 70 ''Isolation Experiments for Vinyl Chloride-Producing Organisms: Preliminary experiments showed that the biological formation of vinyl] chloride was relatively slow. To boost this rate it was necessary to have a large population of easily accessible microorganisms. Standard microbiologi- cal methods were not useful in this regard. Instead of isolating and select- ing single colonies on agar medium, we used a 2 liter anaerobic fermentor, which provided the necessary quantities of culture. For assay, a sample vial could be filled directly from the fermentor jar by means of a sampling device, with minimal exposure to air. The growth medium used consisted of the basal salts medium described for the roll-tube method (below) supplemented with 1 ml/1 of HPLC grade methanol. Gases produced in this fermentor were collected by water displacement in a liquid-filled gas meter with graduated markings. After consumption of each batch of added methanol, more methanol (1 ml/1) was added, approximately once per day. After sample removal from the fermentor, fresh deoxygenated liquid medium was used to make-up the volume. Deoxygena- tion was achieved by argon-sparging for 30 min. Culture samples were plated on agar either on Petri dishes in a GasPak anaerobic jar or by the Hungate roll-tube method. The anaerobic system used for strict anaerobes (the roll-tube method) was described by Hungate (1969). The medium used consists of mineral salts supplemented with 1 ml/1 of HPLC grade methanol. The mineral salts added to the medium are as follows: Disodium hydrogen phosphate 9 mM; Potassium dihydrogen phosphate 9 mM Ammonium sulfate 15 mM; Magnesium sulfate 2 mM Sodium chloride 0.1 mM; Calcium chloride 20 uM; Cobalt chloride 5 uM Copper sulfate 1 wM; Ferrous sulfate 1 uM Manganous sulfate 1 uM Sodium molybdate 4 uM Zinc sulfate 0.5 uM; Each tube contained 5 ml of medium supplemented with 2% agar (Difco). Cysteine (0.03%) and hydrogen sulfide (0.125 ml) per tube were added to the 7| ''tubes after inoculation, as additional reductants. Also, 0.0001% reasurin was added to the cultures as a redox indicator. The tubes were purged with a gas mixture containing 50% argon and 50% carbon dioxide, and incubated at 0 35°C. Conclusion: These experiments were designed to provide the maximum information on the cause(s) of vinyl chloride emissions from landfills in California. The experimental plan was designed to provide the flexibility needed to respond to unexpected situations and to information derived at earlier stages in the experimental work. The next section describes the results obtained in practice. 72 ''EXPERIMENTAL RESULTS Our experience in obtaining landfill samples and the results of experi- ments to determine the source of vinyl chloride emissions from landfills are presented here. The order of presentation is not always the same as the order in which the experiments were performed. Instead it represents a logical sequence from pure materials treated with a single agent through complex and contaminated landfill materials which may have been exposed to a wide range of different conditions leading to vinyl chloride formation. The description of experimental results is concluded with microbial isolation experiments which once again deal with defined systems. LANDFILL SAMPLING RESULTS Site A: The refuse removed from the 2 boreholes Al and A2 included common house- hold refuse components such as plastic, paper, cloth, etc. The organic refuse material removed from borehole Al was relatively dry and was uniformly brown. The organic material removed from borehole A2 was similar to that from Al to a depth of ~ 8 ft. Below that, the refuse became darker, with apparently more moisture. The unsatisfactory cover at sampling location A2 caused difficulty in sample collection. On the first drilling attempt, excessive cobbles in the cover material prevented penetration of the cover. On the second drilling attempt the refuse cover depth was found to exceed 10 ft; samples could not be taken at the 10 ft refuse depth (i.e., 20 ft total depth) with the equip- ment on-site. The boring location was therefore moved ~ 100 ft and the sampling was successfully achieved. Removing sample A2 at 4 ft was also difficult. Several attempts were needed to remove sufficient sample, and hence the total sample contained material from the 4-6 ft refuse depth. To provide refuse and gas samples and gas samples from sampling points separated vertically by a minimum of 5 ft, 73 ''and to sample the refuse in the lower, more moist refuse below 8 ft, this borehole was extended to a refuse depth of 16 ft before a second sample was collected. Samples of refuse and gas were forwarded to Battelle-Northwest by Federal Express on May 7, 1986, and were received at the laboratory on the morning of May 8. Site B: At Site B, the refuse removed from the boreholes included common house- hold refuse components such as plastic, paper, and cloth. The organic refuse material removed from both boreholes was relatively dry. No undue difficul- ties were encountered in obtaining the samples from Site B. The samples were forwarded to Battelle-Northwest on May 19, 1986, and received the following morning. SAMPLE DESCRIPTIONS AND ON-SITE MEASUREMENTS: On-site measurements recorded at both sites are presented in Table 13. Additional gas sample analysis was performed on receipt at Battelle to determine if any leakage of air into the samples had occurred. Oxygen levels in the gas from buried samples was at or less than the 1.2% background normally experienced. This indicates that problems encoun- tered in getting the samples from the sampling tube into the bags should not have significantly raised the oxygen stress on the organisms. The samples spent 2 nights at or below 20°C, plus one day in an airplane, and the effect of this on viability was unknown at the time. Individual Sample Descriptions: As commented earlier, there is no such thing as a “representative” landfill sample. However, the following visually determined observations on sample appearance are useful in determining the nature of the original material in the sample, and the approximate degree of decomposition. 74 ''Table 13: Site A and B Refuse and Gas Sampling Conditions. Sample ID # Description Refuse Temp. pH Methane Oxygen Interval (F) (%) (%) * A-1-S Surface cover -0- 78 NR -0- ely, A-1-3 Refuse + Gas 3-4 78 5.6 62 -0- A-1-10 Refuse + Gas 10-11 78 6.4 62 -0- A-2-S Surface cover -0- 88 NR -0- 21 A-2-4 Refuse + Gas 4-8 91 6.2 55 -0- A-2-16 Refuse + Gas 16-18 90 6.1 51 T# B-1-S Surface cover -0- 88 NR -0- 21 B-1-3 Refuse + Gas 3-4 88 5.4 50 -0- B-1-10 Refuse + Gas 10-11 96 7.0 54 -0- B-2-S Surface cover -0- 98 NR -0- 21 B-2-3 Refuse + Gas 4-6 98 5.2 50 -0- B-2-10 Refuse + Gas 10-11 119 4.9 50 -0- All depths in feet; soil cover depth at both locations was ~ 2 ft. Ambient temperature was 68 F., Refuse Interval refers to depth below cover base. NR = Not Recorded; = +/- 1%; " = Trace, <1%. Measurement precision: Refuse interval, +/- 0.25 ft; Temperature +/- 0.5°F; pH +/- 0.05 pH units; methane and oxygen concentrations +/- 1%. Cover samples A-1-S and A-2-S consisted of yellow, coarse-grained sand and rocks, as expected. The largest rocks in the received samples were about 1-2 cm in diameter. The cover samples from Site B consisted of black dirt with some small gravel. At the intermediate depth at Site A, the sample at location #1 (A-1-3-) contained much paper with print still readable (i.e., little decayed), while at location #2 (A-2-6-) there was some paper, foam, some rocks, and metal. Site B intermediate samples at location #1 (B-1-3-) contained sandy soil, paper, and plastic wrapping, and at the second location (B-2-3-) contained grass clippings in addition to these materials. At the greatest depths sampled, considerable refuse decay had occurred in all samples. At Site A (A-1-10-) there was decayed material resembling tree bark, plus dirt, rocks, and unrecognizable material. Sample A-2-16- also contained some decaying wood. Site B (B-1-10- and B-2-10- samples) 75 ''contained decayed grass clippings, plastic wrapping, metal foil, decayed cloth, and unidentifiable metal objects. The Site B samples appeared to be much wetter than the Site A samples at this depth. The original intent in obtaining samples at two depths was to obtain material at the beginning, and well into, the anaerobic zone. From the appearances of these samples, this objective was achieved. VINYL CHLORIDE FORMATION BY CHEMICAL ROUTES Hydrolysis of Chlorocarbons: Although not specifically investigated, one observation is of interest. Autoclaved vials to which 1,1,2-trichloroethane had been added after sterili- zation yielded 1,1-dichloroethylene as a product, in the absence of any detectable microbial growth. This result indicates that a non-biological conversion of chloroethanes is possible, leading to the formation of biodegradable chloroethylenes. Of course, this does not exclude the concurrent existence of a biological route between these two compounds. Pyrolysis Experiments: Because of its chemical nature, vinyl chloride polymer (polyvinyl chlor- ide, PVC) must initially be a prime suspect as a source of vinyl chloride emissions. If not the polymer itself, then perhaps the wastes from PVC production (PVC waste sludges), which contain large amounts of unreacted monomer. Information available from the literature, however, strongly indi- cates that PVC cannot be a source of significant amounts of vinyl chloride. The operators of sites A and B informed us that PVC waste sludges were never allowed to be dumped in either landfill. When landfill samples were visually inspected, only small pieces of plastic were noted, including some garbage bag material. Saran is a PVC-type material crosslinked with vinylidene chloride (1,1-dichloroethylene). In principle, depolymerization reactions could lead to vinyl chloride formation, but in practice, the decomposition takes place by a completely different mechanism leading to benzene as the principal product. One remaining possibility was that "old" samples of PVC 76 ''(>10 years) could contain larger amounts of residual monomer, since at the time of manufacture vinyl chloride was not recognized as hazardous. For this reason, and to confirm that PVC does not yield vinyl chloride, we performed pyrolysis experiments. Also, in case landfill material could exert some catalytic effect which might alter the PVC decomposition mechanism with the formation of vinyl chloride, we included landfill material/PVC mixtures in the experiments. The following results were obtained: Head space analyses were performed on PVC pipe, Saran wrap, and garbage bag material (mostly polyethylene). The sample containing the most residual vinyl chloride monomer was an 8-10 year old threaded PVC fitting (containing 336 ug/g). A 1-year old PVC pipe sample contained only 3 ug/g of residual vinyl chloride monomer. Neither fresh Saran wrap nor garbage bag material contained detectable levels of vinyl chloride by head space analysis. PVC, landfill material, and landfill material plus PVC were pyrolyzed to 650°C. Gas samples were collected for analysis at 150, 350, 500, and 650°C. PVC (2 g) began decomposing at 150°C, Vinyl chloride was detected in all gas samples up to 650°C. Yields of vinyl chloride were no greater than expected from residual monomer content. Landfill material (20 g) from A-2-16 was pyrolyzed. No vinyl chloride was detected in any of the gas samples. Fixed gas analysis showed the decom- position gas to consist of hydrogen (54%), carbon dioxide (31%), carbon monoxide (5.6%), and methane (9.2%). After pyrolysis, 13.8 g of char remained. Landfill material (18 g) from A-2-16 was mixed with PVC (2 g) and pyro- lyzed to 650°C. Gas samples at all temperatures contained vinyl chloride at concentrations similar to those from the PVC-only pyrolysis experiment. The product gas contained hydrogen (44%), carbon dioxide (22%), ethylene (1.5%), ethane (4.7%), methane (21%), and carbon monoxide (6.1%). The formation of hydrogen chloride from PVC pyrolysis greatly increased the production of condensable liquid from landfill material. After pyrolysis, 13.5 g of char remained. 77 ''Photolysis Experiments: Formation of vinyl chloride by photolysis of precursors in air is a possible mechanism which has been discussed previously, and found to be unlikely. Loss of chlorine radicals from TCE would be expected to produce chloroacetylene; in the presence of oxygen from air, chloroacetaldehyde is formed. Photolysis of PVC itself is known to lead to similar decomposition products to those formed during pyrolysis. However, in the presence of land- fill cover material, there exists the possibility that a photo-assisted catalyzed reaction could take place. Similar syntheses are known (e.g., formation of formic acid from CO and water over silica). The landfill cover is primarily sand and clay, both in theory capable of modifying the normal photolysis reaction sequence. Experiments were performed to check this hypothesis, as follows: Surface soil samples from the landfill sites were photolyzed in the presence of 13¢_jabeled trichloroethylene. Soil (5 g) was placed in a 5 ml quartz cell, spiked with the labeled TCE, and then photolyzed for 2 hr in air and in argon with a 450 watt Ace-Hanovia UV lamp source. Photolysis in air resulted in formation of labeled chloroethyne (chloroacetylene). In argon, the only compound detected was unlabeled benzene. This probably arose from photolysis of toluene, which was one of the major organic contaminants found in the landfill samples. Effect of Steam and Gamma-Irradiation Sterilization: Autoclaving and irradiation were used to kill any viable microorganisms in the samples, with the intent of showing that any vinyl chloride-producing activity ceased after treatment. This would strongly implicate viable micro- organisms as the causative agents of vinyl chloride formation from materials pre-existing in the landfill. Autoclaving was for 15 min at 121°C, the standard conditions for routine sterilization. A total of 16 samples were irradiated with cobalt-60 gamma rays to a total dose of 2.5 megarad. The sample vials were filled with sterilized water. Half were spiked with metha- nol only, and the other half with methanol plus 13¢_trichloroethylene. Bio- logical activity was monitored by headspace gas analysis, and vinyl chloride production by the normal purge-trap method. Disappearance of oxygen and the 78 ''appearance of Ho, CO, C05, and CHy was used to indicate the presence of bio- logical activity. (Radiation sterilization was not always fully effective, even at 2.5 MR.) Absence of vinyl chloride formation together with a lack of biological activity would have indicated that vinyl chloride formation is associated with viable microorganisms in the landfill samples. No vinyl chloride was detected from autoclaved sewage sludge samples after 5 weeks incubation. In no case did an autoclaved landfill sample that showed no biological activity also produce vinyl chloride. The autoclaving was only successful in sterilizing ~ 20% of the samples (a very low proportion, either demonstrating the ability of anaerobes to withstand heat or that they were protected in some way by the aggregate landfill material). Irradiation sterilization was successful in ~ 50% of the samples, and again, no vinyl chloride was produced from any vial that did not show biological activity. The association between the existence of viable microorganisms and production of vinyl chloride from labeled TCE is therefore shown by these experiments. CHEMICAL ANALYSIS OF LANDFILL MATERIAL A knowledge of the composition of the landfill material is essential in order to postulate vinyl chloride formation mechanisms. The presence or absence of a particular material in a given sample (e.g., TCE) coupled with the presence or absence of vinyl chloride constitutes good circumstantial evidence for or against specific formation mechanisms. For this reason, we performed detailed organic and inorganic analyses of our landfill samples. However, with so few samples taken from such heterogeneous mixtures, we do not claim that the following compositions are in any sense “normal" or "average". We also determined the head-space gas composition above the samples, to obtain an indication of the nature of any volatile organic compounds present and to determine whether or not there was active anaerobic digestion still occurring. Because of the large amount of laboratory work involved, not all measurements were made on all samples. The following information was obtained: 79 ''Volatile and Ash Analysis: Volatiles (primarily water) and ash determinations have been made for six of the landfill samples, including surface soil. Volatiles were deter- mined by heating the samples at 110°C under vacuum, and therefore include organic solvents as well as water. Ash contents were determined on the dry samples from volatile analysis, by heating at 750°C for 8 hr. The samples were extremely inhomogeneous, as expected for the type of material, so these analytical data should be considered indicative rather than representative. The results are presented in Table 14: Table 14: Volatile and Ash Contents of Landfill Samples Sample ID Sample Moisture* Sample Final wt.+ % Residue weight Wt. Wt.% Dry wt. (ash) A-1-1-25 7.5425 2.6881 35.64 4.8544 3.0335 62.5 A-1-S-14 10.4442 1.1022 10.55 9.3420 8.9955 96.3 A-2-16-28 15.2425 3.0351 19.91 12.2074 10.9635 89.8 B-1-3-27 11.7681 1.4983 12.73 10.2698 9.4839 92.4 B-2-S-16 19.0348 2.9061 15.27.) 16.1287 15.2232 94.4 B-2-10-27 3.4187 0.8923 26.10 2.5264 1.3479 53.4 * Moisture - includes volatile organics; + Difference between sample weight and moisture weight = sample dry weight. Sample ID Sequence is: Location (Site A or Site B); Bore #; Depth (S = cover); Sample reference # (1-34). Inorganic component analysis of landfill material was performed by the ICP (Inductively Coupled Plasma) technique (Appendix *C'), with results as shown in Table 15. Detection limits in Table 15 are given in ug/ml. Elements not listed were either not detected or are not detectable by ICP. A hyphen in Table 15 indicates that the element, if present, was below the detection limit. Sample labeling is the same as that used in previous reports of sample #. For example, B-2-S-16 = Site B, second core, Surface sample, tube #16. Gas sample analyses were also performed. Data for the location contain- ing the highest amount of vinyl chloride (A-2-16) is shown in Table 16, below. Figure 9 shows the total ion monitor traces from 33 to 200 amu for landfill gas 80 ''samples, determined by head-space analysis. The presence of vinyl chloride was confirmed by the presence of a peak at m/e 62. Table 15: ICP Inorganic Component Analysis of Landfill Samples Element Detn.* Sample # (Wt.%) Limit B-2-S-16 B-1-3-27 B-2-10-27 A-1-S-14 A-1-1-25 A-2-16-28 Al 0.03 6.45 7.80 6.33 6.71 7.29 6.23 B 0.01 0.04 0.04 0.04 0.04 0.08 - Ba 0.002 0.06 0.07 0.06 0.13 0.10 0.10 Ca 0.01 1.39 1.60 307 0.64 3.80 1.86 Cu 0.004 0.01 - - 0.02 0.03 0.03 Fe 0.005 3.78 4.03 3.29 3.38 4.04 4.88 K 0.3 1.43 1.95 0.99 2436 2.22 1.87 Li 0.004 0.02 - 0.02 0.03 0.02 Mg 0.06 1.21 1.21 1.07 0.62 1.00 0.83 Mn 0.002 0.07 0.08 0.07 0.07 0.09 0.09 Na 0.01 1.21 1.49 2.10 1.68 1.40 1.48 Si 0.02 31.9 30.8 31.0 32.6 28.5 31.5 sr 0.002 0.02 0.04 0.03 0.02 0.02 0.02 Ti 0.002 0.39 0.30 0.40 0.24 0.57 0225 Zn 0.02 - - 0.04 - 0.09 - zr 0.008 0.05 0.02 0.05 0.04 0.05 0.02 Head space gas analyses were performed by heating the samples in a septum-capped vial for 1 hr at 110°C. Gas samples of 1 or 5 ml were then directly removed by syringe through the septum and injected into the GC/MS. Head-space analyses of samples should be considered as qualitative only, as volatile components would have been lost during field handling. However, after sealing samples in plastic, head space gas analyses were useful indicators of organic compounds present in samples and of ongoing biological activity. Chlorinated ethylenes and ethanes were present in the Site A landfill head- space samples, but none in the Site B gas samples. Also, dichloromethane was present in the Site A but not in the Site B samples. Many of the compounds identified by head-space analysis of the Site B samples can be attributed to 81 '' \ | ea auajAyje0101401G- L’ a aplojYyD euajAujoj| \ ! aad *JQSHO @E==== 2, joueyiy s===== I~ - © _-- -~ o- ~ EO] cree mx ae O 7g uo} } z1I9%4957 ID?4HO SA J -" C . x x N \ ( \ \ ‘ } J 7 7 7 \ \ 4 ane < ~~, —_—— -—---” auanjol~7> < r < 2 hi n -----? = > a qr ex-~ oa wscira ---> ! } lon 63 — ; ' cs ii 4 it ral | ifs ' ' ot IN. i E.M. -200 4 . i a” 1,1-Dichloroethylene ' | pe NO aeenemamun en ee wo Me“ ® dag _// \/| Se ee 5 10 15 Time (min) 93 ''Figure 12b clearly shows the formation of 13¢_labeled vinyl chloride from labeled TCE after incubation with Site B landfill material. The ion at m/e 63 is characteristic of labeled VC, coupled with the VC mass spectrum obtained at the same time. ISOLATION OF VINYL CHLORIDE-PRODUCING METHANOGEN: Work with samples in vials has shown that methanol can be used as a substrate by the organisms which produce vinyl chloride from chloroethylenes. By using methanol as a carbon source, most of the landfill organisms were selected out as they cannot metabolize this compound. A 2 liter fermentor was used in the following culture isolation experiments. One of the advan- tages of this approach is that large amounts of the desired organism become readily available. The slow growth rate of methanogens and the slow conver- sion rate of TCE to vinyl chloride require large numbers of the organisms to be present in order to observe product formation in a reasonable time period. An anaerobic methane-generating population was established from vial samples grown on methanol. The methane: carbon dioxide ratio formed by this culture was 2:1, which is expected for methanogen cultures using methanol as the sole carbon source. After staining and microscopic examination of a portion of this culture, the cell population appeared to be homogeneous (i.e., single species) and were gram-negative cocci (spherical cells). The indications at this stage were that we had a pure culture, although with methanogenic species this assumption is less certain than with other types of bacteria. This culture was able to convert 130_labeled TCE to 130_labeled vinyl chloride, as confirmed by GC/MS analysis (Figure 13). From 100 ug of labeled TCE, 1.2 ug of vinyl chloride were produced in 1 week at 35°C. A BBL GasPak anaerobic system was used to monotor contaminant microorganisms. In this system oxygen is removed by reaction with hydrogen over a palladium catalyst, producing water. The hydrogen is generated from sodium borohydride. Since no other reductant besides hydrogen was initially present, the reduction potential was not great enough for methanogens to function. We did not obtain any colonies after incubation in the GasPak system when the culture was plated out on Petri dishes. We also did not observe any colonies after incubation on Petri plates in a normal atmospheric oxygen concentration. Both results suggested that 94 ''‘gunz[nj uaboueyzaW peze_osy ue Aq aud [AYZA040 [YOLA1-Joy WOAJ ApLuoLy) LAULA PPLEGP I-Ie] $0 UOLZONPOUd Hurmoys B9P4l SW/IH “EL JUNI (uon2ag spoyey| 88S -- pas SUOITIPUCD SW/D5 Psepuels) a i i | | aS == nee = =2 5 no 7 me OF =x —_— ———— — — - = 4 a Tr x a — —~_ er =~. wee 7 Ne ’ rye > we — oe — a / , ory L ia aS Fa . — / 1 Gf ¥ © ‘ i \ —s* ,o4 x= N 4 = | tor we V/ —) toa = ~~ . zi to i tow ae — 1 vt 25 ' vt 2 t Vt a \ v on ‘ 3 ‘ ' 0 \ 3 ' 2 t ° ' t ty ‘ 2 ‘ 2 ' 3 ' 2 : €9 e/uw . ' ‘ A Es apiso|yo |AUIA-D,, t : 1 1 ' ' ' ' ‘ \ ' ' : ‘ ; 1 ' \ ' | 301-9 ' eb ' ' ‘ = aa ‘ Vv 3 cs — — — ————_—___ > — = - — sO f YF L- \=a n =~ = —— oe Ne \= , ae \ \_—/ m oa XY ¢ 7 —— / ,.w rT \w f ed 1 wm. wo ge X ’ : —« 1 1 ad : vt 1 ‘zi y ¥ ¢ =a) 1 ' ‘ ’ usr 1 (=i 2 =o + i — | , oF 1 = 3 y 1 a s t : \_—! ton ’ Ae . t ie ‘ 4 t — 1 1 4 ' tu 1 twa 1 : —! ton , oF ’ rot x OUR —« og to ' tog . oa '_—! ot ior 1 to4 . - oa —) 4 tot ‘ . 4 x 1a \— 1 . os 14 1 io uw ce — wu ju i : . . z ‘ : [ 95 ''contamination was low and that mainly methanogens or strict anaerobes were present. Following these preliminary results, incubation under stringently anaero- bic conditions was performed to determine if single species which can produce vinyl chloride could be obtained. The anaerobic system used for these experi- ments was the roll-tube method (see Methods section for a description of this technique). Only one type of bacterial colony was observed to grow, demonstrating that the culture was apparently homogeneous (i.e., a single species, although this is not always true with methanogens). To our knowledge, this is the first reported case of isolation of a vinyl chloride-producing methanogen from an actual landfill sample (A-2-16-). The culture is being maintained and replicated in an attempt to obtain sufficient samples for distribution to others and/or to a recognized culture collection. 96 ''DISCUSSION FORMATION OF VINYL CHLORIDE During the course of the experiments reported here, we obtained and analyzed small landfill material samples, incubated them with labeled TCE, showed the formation of labeled VC and dichloroethylenes, and isolated a methanogen capable of causing this transformation. We also performed experi- ments to cause various chemical transformations on PVC and chlorocarbons, which uniformly failed to produce detectable amounts of VC beyond those already present as impurities. Killing microorganisms in the landfill samples which normally caused formation of VC from TCE resulted in the complete cessation of this activity. The conclusion from this research is definite: Vinyl chloride is produced from chlorocarbon solvents dumped into landfills through the action of methanogenic bacteria. This may not be the only route, however. Within the limits of error of our experiments we could not completely eliminate other possibilities, but these are minor contributors to the overall VC formation. This conclusion is true for the landfills we sampled, which were both municipal landfills taking no hazardous waste. The amounts of vinyl chloride present in the headspace gases above the landfill samples on receipt corre- sponded roughly with the degree of solvent dumping likely to have occurred, although due to loss of volatiles during a brief exposure of the samples to air prior to packing, this is only a qualitative observation. (Site A accepted some industrial waste, and Site B did not; Site A had 5 ppm of VC in the headspace gas, Site B had less than 1 ppm). Reportedly, gas samples taken above hazardous waste landfills where large quantities of chlorocarbons have been dumped contain much larger amounts of VC, in excess of legally mandated exposure limits. The following discussion is based on our results compared with those of others, and is in support of the basic conclusions stated above. 97 ''EXPERIMENTAL APPROACH The basic objective of the research as stated previously was to identify the source(s) of observed VC emissions from landfills in California. The simple nature of the problem belied the difficulty in obtaining a direct solu- tion. Various experimental approaches were considered. Previous and ongoing research by several groups (notably those of Parsons in Florida and McCarty in California) implicated microorganisms in the formation of VC, and more specifically, anaerobic bacteria. The ability of some microorganisms to dechlorinate chlorocarbons anaerobically has been recognized for some years (e.g., Kobayashi and Rittman, 1982). It would therefore have been logical to accept the conclusions of these groups, and to begin with anaerobic bacteria and use them to demonstrate VC formation from chlorocarbons. Since scientists like to work with defined systems, for obvious reasons, one approach would have been to obtain cultures of various anaerobes and incubate them with chlorocarbons. VC formation could then have been demonstrated and applied by implication to landfill conditions. The difficulty of this approach is that to obtain proof of the ability of the anaerobes to form VC in landfills, the identical species would have had to be isolated from landfill material. Com- plete and accurate species identification of anaerobes is notoriously diffi- cult, and for this reason we abandoned the approach of beginning with a defined system (anaerobe/salts medium/chlorocarbon) in favor of the undefined system of using actual landfill samples. Reversal of the approach described above would have led to isolation of a VC-producing anaerobe, which could then be identified. This approach of course was the one actually used, although identification of the exact species of anaerobe has not yet been achieved. Identification of one route for VC formation in landfill material is a positive step, but not the only one. From the point of view of California state agencies, it was also necessary to show the existence or approximate contribution of other mechanisms to the overall picture. A major consider- ation in this type of work is the fact that there is frequently a discrepancy between what is actually and what is reportedly dumped in landfills. The major concern in this regard is the possibility for illegal dumping in the past of PVC production sludges, at a time when the extreme toxicity of VC was not recognized, and when these sludges contained large amounts of residual 98 ''monomer (e.g., Markle et.al. (1976)). A further possibility is the obvious one of depolymerization of PVC itself with the formation of VC monomer. Research into these questions led to the conclusion that release of VC by these mechanisms in the landfills we sampled is inherently unlikely. For the first possibility (PVC production sludge dumping), reference to Appendix B shows that PVC production in the State of California and neighbouring States is small; furthermore, illegal dumping would most likely take place at hazardous waste landfills. It is difficult to imagine a clandestine dumping of such toxic materials at municipal landfills such as were sampled by us. As for depolymerization of PVC, considerable effort over the years has gone into identification of the decomposition mechanisms of this important plas- tic. The direct depolymerization occurs only to a trivial extent (parts per million of VC), and a dehydrochlorination and cyclization is the primary mechanism. Even this only occurs at elevated temperatures likely to occur only rarely under spontaneous combustion conditions. Hence we eliminated PVC and its precursors as likely sources of the VC we observed in our municipal landfill samples. A further possibility which we were not able to exclude based on our research is the slow leaching of VC monomer from "old" plastics. Samples from our laboratory contained over 300 ppm of VC monomer, while "new" samples contained about 1% of this level. There would have been no reason to refuse waste PVC in either landfill A or B ten years ago, and the same applies to other PVC-type plastics such as household wrap. To determine the degree to which leaching contributes to the overall VC production, long-term studies would be needed to (a) analyze landfills thoroughly to determine the PVC content, (b) analyze the PVC to determine the average VC content, and (c) determine the average water flow and leaching rates. This of course would allow an estimation of the VC contribution to groundwater; further work would have to be done to estimate the transfer to landfill gases, which is a complex function of gas generation rates, organic composition, and other factors. A best guess based on our results is that the contribution from PVC and other plastics is certainly less than 5% and probably less than 1% of the total VC measured. 99 ''While production of VC from PVC is unlikely for the reasons given above, the same is not true of VC production from chlorocarbon solvents. These are widely used and ubiquitous, and are found in degreasers, thinners, paints, duplicating fluids, dry cleaning solvents, and a host of other materials in common use. Such materials certainly find their way into municipal landfills in both household and industrial waste. In the latter case, the presence of metal fabrication, electronics, and similar industries in Southern California could have led to the observed higher levels of chlorocarbons in Site A than in Site B. The formation of VC from chlorocarbons seemed a priority to be the most likely and major source. However, there remained the possibility that some unsuspected mechanism could be implicated. Just because a mechanism appears likely does not necessary imply that it occurs or is the only contributor. Hence, we devised a series of chemical experiments based on heat and radia- tion, which could also possibly have produced VC from chlorocarbons. We did find that trichloroethane could be converted into 1,1-dichloroethylene, but found no evidence for the formation of VC by pyrolysis, photolysis, or catalytic formation from ethylene and hydrogen chloride. Having designed our experiments to eliminate formation of VC from PVC and by chemical means, the way was clear to investigate the biological route from chlorocarbons. As discussed in the text, sewage, landfills, and related anaerobic environments (not compost piles, which are aerobic) contain many representatives of three major classes of microorganisms. Generally, these include organisms which convert polysaccharides (cellulose, hemicellulose) to acids, acetogens which produce acetic and butyric acids from other non-polymeric compounds, and methanogens which produce methane and carbon dioxide from acetic acid. From the structural similarity between VC (chloroethylene) and acetic acid, (both Cy compounds) the methanogens seemed the most likely candidates. In fact, by using methanol as a carbon source for our experiments we selected for methanogens, and since VC was produced in all cases from TCE and other chlorocarbons, the assumption was shown to be correct. It does not preclude other non-methanogenic organisms from also producing VC. Had we used glucose or an organic acid as a carbon source we would have had a much wider variety of anaerobes growing in the incubation experiments, with correspondingly 100 ''complicated results. This is because the major metabolic product of one class becomes the growth substrate (carbon source) for the succeeding class. Since the beginning of the decade, when the ability of anaerobes to dechlorinate organic compounds was recognized, our understanding of the conversion route has been getting gradually more complex. Beginning with a linear conversion of tetrachloroethylene, for example, through TCE, all 3 possible dichloroethylenes, and then to VC, the route has been expanded. Figure 14 shows the current understanding based on the work of Parsons and Barrio-Lage (1985), and Barrio-Lage et.al.(1985). Vogel and McCarty (1985) have recently extended this by showing the "mineralization" of VC to carbon dioxide. In our work, the reversibility of the transformation of 1,1- dichloroethylene to VC has been shown by experiments in which VC gas was added, and 1,1-dichloroethylene formation observed. A further clarification of the literature relates to the ubiquity of the organisms responsible: By using chlorocarbon-free sewage sludge and showing formation of VC from a range of chloroethylenes and chloroethanes, we demonstrated that the ability of methanogens to metabolize chlorocarbons is inherent, not acquired (i.e., it is constitutive and not adaptative in nature). In going one step further and isolating a bacterium capable of performing the transformation of labeled TCE to labeled VC while growing on methanol, we have opened up the opportunity for other workers in the field to work with defined systems using this orga- nism. (This isolate is the property of the ARB and was derived from landfill Site A; we hope that it will be made available for further research.) In performing the experiments required to derive the above statements and conclusions, it was necessary to deviate slightly from strict scientific protocol and to make choices regarding which analyses to perform and in what degree of detail. We emphasize that this was done with the realization that the results would be indicative (although defensible) rather than comprehen- sive. For example, during the course of this study, we obtained a total of 12 landfill samples plus two sewage sludge samples. Each sample was subjected to various treatments defined in the Methods section, requiring a total of 34 vials for each sample. This total includes only TCE, 1,1-dichloroethane, or 1,1,1-trichloroethane and labeled TCE as the chlorocarbon substrates. A 101 ''\ H-9 7 7 \ 2H "(SG86L *°Le°3e8 abeq-oluueg > mmmonm e © e e aacaca J.F. Caballero and C. Wittig, 1983. J. Chem. Phys., 78, 7169. P.V. Cline and D.R. Viste, 1984. In: Proc. Munic. & Indust. Waste, Ann. Madison Waste Conf., 14. L.G. Close and R.D. Gilbert, 1977. Polym.-Plast. Technol. Eng., 8, 177. W.L. Dilling, C.J. Bredeweg, and N.B. Tefertiller, 1976. Envir. Sci. Tech., 10, 351. W.F. Dimmick, 1981. Envir. Hlth. Perspect., 41, 203. J.S. Drury and A.S. Hammons, 1979. Rept. #EPA-560/2-78-006, ORNL/EIS-148, NTIS PB 295865, p. 28. R.C. Dressman and E.F. McFarren, 1977. J. Chromatog. Sci., 15, 69. J.E. Dyksen and A.F. Hess, 1982. J. Amer. Waterworks Assn., Aug., 394. Environmental Protection Agency, 1978. Federal Register, 41, 46559. P.J. Ford and P.J. Turina, 1985. "Characterization of Hazardous Waste Sites - A Methods Manual, Volume I - Site Investigations". EPA/600/S4-84/075; NTIS PB 85-215 960/AS. P.J. Ford, P.J. Turina, and D.E. Seely, 1985. "Characterization of Hazardous waste Sites - A Methods Manual: Volume II. Available Sampling Methods - Second Edition". EPA-600/S4-84-076; NTIS PB 85-168 771/AS. D.R. Goodwin and D.G. Hawkins, 1980. In "Perchloroethylene Dry Cleaners - Background Information for Proposed Standards. Draft EIS", Rept. EPA 450/3-79-029a, NTIS PB81-100497, p. 7-10. R.E. Hefner, jr., P.G. Watanabe, and P.J. Gehring, 1975. Envir. H1th. Perspect., ll, 85. . Hungate, 1969. In: "Methods in Microbiology", eds. J.R. Norris and . Ribbons, Ch. 4, Vol. 3B, Academic Press, N.Y. . Jacot, 1983. Proc. Natl. Conf. Manage Uncontrolled Hazard. Waste Sites, 76. B.B. Kebbekus and J.W. Bozzelli, 1983. "Volatile Organic Compounds in the Ambient Atmosphere of the New Jersey, New York Area". EPA-600/S3-83- 022; NTIS PB 83-191 403. oo wz c=m 109 ''a uv Oo <_< Pon oe e nNwnVD oz =z So -— > = — =r Oo a= Oo Kleopfer, D.M. Easley, B.B. Haas, jr., and T.G. Deihl, 1985. Envir. Sci. Technol., 19, 277. Kobayashi and B.E. Rittman, 1982. Envir. Sci. Technol., 16, 170A. Kontominas, J. Miltz, and S.G. Gilbert, 1982. J. Food Sci., 47, 1208. . Kowalski, J.M. Shikiya, and G. Tsou, 1985. Presentation at 78th Ann. Meet., Air Poll. Control. Assn., Paper # 85-17.1, Detroit, MI. . Lamorte, 1978. “National Emission Standards for Hazardous Air pollutants. Inspection Manual for Vinyl Chloride". U.S. EPA, Contract # 68-01-4141. NTIS PB-289 778 . Lewis, 1975. Paper presented at 23 rd. Ann. Conf. on Mass Spectrometry and Allied Topics, Houston, TX, May 26-30. Markle, R.B. Iden, and F.A. Sliemers, 1976. Proc. Hazard. Waste Symp., EPA-600/9-76-015, NTIS PB 256 768, p. 186. . McCarty, 1964. Public World, Sept.-Nov. . McCarty, 1986. "Biological Approaches to Aquifer Restoration: Recent Advances and New Opportunities", Stanford Univ. Environ. Eng. & Sci. Seminar, June 17, 1986. . Moss, M.D. Ensminger, and J.D. McDonald, 1981. J. Chem. Phys., 74, 6631. . Parsons, G.B. Lage, and R. Rice, 1983. Am. Chem. Soc. 186th. Nat. Meet., Abstracts of Paper #ENVR 106, ACS, Washington, D.C., 1983. . Parsons, P.R. Wood, and J. DeMarco, 1984. J. Am. Water Assn., 76, 56. . Pantke, 1977. Biodeter. Invest. Tech., ed. A.H. Walters, Appl. Sci. Publ., Barking, England, pp. 51-76. . Pennington, 1983. In: “Role of the Unsaturated Zone in Radioactive and Hazardous Waste Disposal". J.W. Mercer, P.S.C. Rao, and I.W. Marine, eds., Ann Arbor Sci. Publ., pp. 211-227. . Plumb, jr., 1984. "Characterization of Hazardous waste Sites - A Methods Manual: Volume III. Available Laboratory Analytical Methods". EPA-600/S4-84-038; NTIS PB84-191048. .B. Radding, D.H. Liu, H.L. Johnson, and T. Mill, 1977. Rept. EPA-560/ 5-77-003. . Reiser, F.M. Lussier, C.C. Jensen, and J.I. Steinfeld, 1979. J. Am. Chem. Soc., 101:2, 350. . Roberts, J. Schreinder, and G.D. Hopkins, 1982. Water Res., 16, 1025. . Sabel and T.P. Clark, 1983. Proc. 6th. Ann. Madison. Conf. Appl. Res. Pract. Munic. & Indust. Waste, 108. Salomon, A.W. Kirk, and E. Tschulkow-Roux, 1979. J. Phys. Chem., 83, 2569. . Sasse, 1967. Health Physics, 13, 1015. . Schneider, 1981. Thesis, "Direct investigation of the Elimination of Hydrogen Chloride from 1,2-Dichloroethane by Laser-Induced Radical Chain Reactions", Max-Planck Inst. for Flow Research, URCL Translation 11803, Lawrence Livermore Nat. Lab., pp. 1-63. . Schwartzonbach et.al., 1983. Envir. Sci. Technol., 17, 472. . Sheldon and J.M. Tiedje, 1984. Appl. Envir. Microbiol., 47, 850. . Sudbo, P.A. Schulz, E.R. Grant, Y.R. Shen., and Y.T. Lee, 1978. J. Chem. Phys., 68, 1306. . Tabak, S.A. Quave, C.I. Mashni, and E.F. Barth, 1981. J. Water Poll. Control Fed., 53, 1503. . Venturini, W.V. Loscutoff, and C.B. Suer, 1985. Presentation at 78th Ann. Meet. Air Poll. Control. Assn., Paper # 85-28.4, Detroit, MI. Vogel and P.L. McCarty, 1985. Appl. Env. Microbiol., 49, 1080. . Wakeman and W.R. Johnson, 1980. J. Vinyl Technol., 2, 200. 110 ''J.J. Walsh, 1984. Hazard. Wastes Environ. Emerg. Mgt. Prev. Cleanup Control, Conf. Proc., 146. . Wolfe, 1971. In: Adv. Microbial Physiol., eds. A.H. Rose and . Wilkinson, Academic Press, N.Y., p. 107. . Wood et.al., 1981. "Introductory Study of the Biodegradation of the Chlorinated Methane, Ethane, and Ethene Compounds", presented at Ann. Am. Waterworks Assn. Conf., St. Louis, MO. T. Yano and E. Tschulkow-Roux, 1980. J. Phys. Chem., 84, 3372. uC DW oe e amnmM BIBLIOGRAPHY The following list or reports and articles provided useful source materials for this study, but may not have been referred to directly in the text: J.W. Bozzelli and B.B. Kebbekus, 1983. "Volatile Organic Compounds in the Ambient Atmosphere of the New Jersey, New York Area". EPA-600/3-83-022 NTIS PB83-191403. V.J. DeCarlo, 1977. “Multimedia Levels: Trichloroethylene". EPA 560/6-77 -029, NTIS PB 276535. B. Demian and E.D. Pellizzari, 1983. "Optimization of GC/MS-Based Tenax Collection Method for Toxic Organics". EPA-600/3-83-058, NTIS PB83- 229476. J.S. Drury and A.S. Hammons, 1979. "Investigations of Selected Environment- al Pollutants: 1,2-Dichloroethane". EPA-560/2-78-006, NTIS PB-295865. E. Edney, S. Mitchell, and J.J. Bufalini, 1982. “Atmospheric Chemistry of Several Toxic Compounds". EPA-600/3-82-092, NTIS PB83-146340. W.H. Fuller, ed., 1976. "Residual Management by Land Disposal: Proceedings of the Hazardous Waste Research Symposium". EPA-600/9-76-015, NTIS PB- 256768. M. Ghassemi, K. Crawford, and M. Haro, 1986. "Leachate Collection and Gas Migration and Emission Problems at Landfills and Surface Impoundments", EPA/600/S2-86/017, NTIS PB 86-162104/AS. D.R. Goodwin, 1980. "“Perchloroethylene Dry Cleaners. Background Inform- ation for Proposed Standards". EPA-450/3-79-029a, NTIS PB81-100497. W.F. Gutknecht, R.K.M. Jayanty, and J.T. Bursey, 1984. "Guidelines for Determination of Laboratory Acceptability for Analysis of Volatile Organic Pollutants Collected on Tenax GC Adsorbant". EPA-600/4-84-035, NTIS PB84-189638. M. Holdren, S. Rust, R. Smith, and J. Koetz, 1985. "Evaluation of Cryogenic Trapping as a Means for Collecting Organic Compounds in Ambient Air". EPA-600/4-85-002, NTIS PB85-144046. R.K.M. Jayanty, J.A. Sokash, W.F. Gutknecht, and C.E. Decker, 1984. "Performance Audit Results for POHC: VOST and Bag Measurement Methods". EPA-600/4-84-036, NTIS PB84-187889. R.N. Kinman, J. Rickabaugh, D. Nutini, and M. Lambert, 1986. "Gas Characterization, Microbiological Analysis, and Disposal of Refuse in GRI Landfill Simulators", EPA/600/S2-86/041, NTIS PB 86-179504/AS. R.M. Patterson, M.I. Bornstein, and E. Garshick, 1976. "Assessment of Ethylene Dichloride as a Potential Air Pollution Problem. Vol. III". EPA Contract # 68-02-1337, GCA Corp., Bedford, MA. NTIS PB 258 355. 11 ''R.M. G.W. G.W. Tels G.E. A. Riggin, 1983. "Technical Assistance Document for Sampling and Analysis of Toxic Organic Compounds in Ambient Air". EPA-600/4-83-027, NTIS PB 83-239020. Scheil, 1977. “Standardization of Stationary Source Method for Vinyl Chloride". EPA-600/4-77-026, NTIS PB 271 513. Scheil and M.C. Sharp, 1978. “Collaborative Testing of EPA Method 106 (Vinyl Chloride) That Will Provide for a Standardized Stationary Source Emission Measurement Method". EPA-600/4-78-058, NTIS PB 298775. . Springer, K.T. Valsaraj, and L.J. Thibodeaux, 1986. "In-Situ Methods to Control Emissions From Surface Impoundments and Landfills”. EPA/600/S2-85/124, NTIS PB 86-121365/AS. Walker, 1984. "Fate and Disposition of Trichloroethylene in Surface Soils". Ph.D. Thesis, Purdue University. Wilkins, 1977. "Industrial Process Profiles for Environmental Use. Chapter 10. Plastics and Resins Industry". EPA-600/2-77-023j, NTIS PB-291 640. Wisbith, E, Barnett, and L. Elfers, 1975. "Monitoring Vinyl Chloride Around Polyvinyl Chloride Fabrication Plants". EPA-450/3-75-084, NTIS PB 249 695. 112 ''GLOSSARY The following is a list of terms, abbreviations, and symbols used in the text and which may require explanation: Acetogen ARB FID GC GC/MS HPLC rep Methanogen NBS PI, PID TCE voc Type of anaerobic bacterium which produces acids from sugars. Air Resources Board, the sponsoring organization for this project. Hydrogen Flame Ionization Detector used in GC. Gas Chromatography. Combined Gas Chromatography/Mass Spectrometry (the mass spectrometer functions as a GC detector). High Pressure Liquid Chromatography (requires especially pure solvents). Inductively Coupled Plasma technique for inorganic element analysis. Type of anaerobic bacterium which produces methane and carbon dioxide from acetate. National Bureau of Standards. PhotoIonization type of GC detector. Trichloroethylene, a common dry-cleaning/degreasing solvent. Volatile Organic Compound. 113 ''''APPENDIX A: * U.S. Suppliers : APPENDICES SUPPLIERS AND USAGE OF MAJOR CHLOROCARBONS Company Location Capacity/Chemical~ VC PVC TCE EtC1l2 TCEt PCE Air Products Air Products Arco Borden Borden CertainTeed Diamond Shamrock Diamond Shamrock Dow Dow Dow DuPont Formosa Formosa Formosa Georgia-Gul f Georgia-Gul f BF Goodrich BF Goodrich BF Goodrich BF Goodrich BF Goodrich BF Goodrich BF Goodrich BF Goodrich Goodyear Keysor Occidental Occidental Occidental Occidental PPG Shel] Calvert City, KS - Pensacola, FL - Port Arthur, TX - Geismar, LA 700 Illinopolis, IN - Lake Charles, LA - Deer Park, TX - Convent, LA - Freeport/Oyster Creek, TX 75 Plaquemine, LA 800 Pittsburg, CA - Corpus Christi, TX - Baton Rouge, LA 420 Delaware City, DE - Point Comfort, TX 580 Delaware City, DE - Plaquemine, LA 1100 Avon Lake, OH - Calvert City, KY 1000 Deer Park, TX - Henry, IL - La Porte, TX 1000 Louisville, KY Pedricktown, NJ - Plaquemine, LA - Niagara Falls, NY - Saugus, CA - Baton Rouge, LA - Burlington, NJ - Pottstown, PA - Pasadena, TX - Lake Charles, LA 500 Deer Park, TX 840 200 200 350 320 220 240 280 340 180 120 50 300 250 200 700 600 190 800 2700 1700 560 1815 1000 1585 350 200 114 '' Company Location Capacity/Chemical~ VC PVC TCE EtC12 TCEt PCE Shintech Freeport, TX - 1000 - - - - Union Carbide Texas City, TX - 125 - - - - Vista Aberdeen, MS - 360 - - - - Vista Lake Charles, LA 700. == - 1150 - - Vista Oklahoma City, OK - 355 - - - - Vulcan Geismar, LA - - - 300 200 150 Vulcan Wichita, KS - - - “ - 50 Vygen Ashtabula, OH - 100 - - - * Information compiled from Chemical Marketing Reporter. VC = Vinyl Chloride; PVC = Polyvinyl Chloride; TCE = Trichloroethylene; EtC12 = Ethylene Dichloride (1,2-dichloroethane); TCEt = 1,1,1-Trichloro- ethane; PCE = Perchloroethylene (tetrachloroethylene). All figures are in millions of pounds annually. Uses: Vinyl Chloride - Polyvinyl chloride production (85%), co-polymers (2%); the remainder is exported. Demand in 1985 was 6.9 billion pounds. Polyvinyl Chloride - PVC is used in rigid pipe, tubing, and molded fittings (43%), flooring and textiles (10%), siding and accessories (7.5%), coatings and pastes (6.5%), wire, cable, film, and sheet (3.5%), other extrusions (4.5%), bottles (3%), other molding uses (1.5%), phonograph records (0.5%), and miscellaneous uses (6%). Demand in 1985 was 6,700 million pounds. Trichloroethylene - Vapor degreasing of metal parts (66%), chemical inter- mediates (7%), domestic uses (5%), the remainder being exported. Demand (1982) was 240 million pounds; 1987 prediction for use was 215 million pounds. Ethylene dichloride (1,2-dichloroethane) - used to make vinyl chloride monomer (90%), 7% is exported, and the remainder is used in manufacture of other chlorinated solvents, vinylidene chloride, and ethyleneamines. The demand in 1985 was 12.68 billion pounds. 1,1,1-Trichloroethane - used in cold cleaning (40%), vapor degreasing (22%), adhesives (12%), aerosols (10%), electronics (6%), coatings (1%), and miscellaneous uses (4%), 5% being exported. Demand in 1982 was 588 million pounds; 1987 use predicted as 680 million pounds. Perchloroethylene - used in dry cleaning and textile processing (53%), chemical intermediates (mostly Fluorocarbon F-113) (28%), industrial metal cleaning (10%), miscellaneous uses (4%), and the remainder is exported. The demand in 1985 was 595 million pounds. 115 ''APPENDIX B: AMERICAN CHEMICAL SOCIETY PRESENTATION AND PRESS RELEASE The following is the text of the ACS presentation and press release: TRANSFORMATION OF CHLORINATED ETHENES AND ETHANES BY ANAEROBIC MICROORGANISMS R. T. Hallen, J. W. Pyne, Jr., and P. M. Molton ABSTRACT The biological transformation of chlorinated ethenes and ethanes to vinyl chloride has been observed in experiments which simulate conditions found in a landfill or chemical waste dump. Experiments employed microbial samples obtained from a municipal anaerobic digester and which had not been adapted to chlorinated solvents. Samples contained in septa capped vials were incu- bated for one week then spiked with a chlorinated solvent diluted in methanol. Vials spiked with 1,1,2-trichloroethane and 1,1-dichloroethene produced the highest levels of vinyl chloride. Those spiked with trichloroethene iso- topically labeled with one C-13 atom produced isotopically labeled dichloro- ethenes and vinyl chloride. Autoclaved and unspiked vials did not yield vinyl chloride. However, autoclaved vials spiked with 1,1,2-trichloroethane produced 1,1-dichloroethene. INTRODUCTION The identification of vinyl chloride as a carcinogen has led to strict emission limits for all industrial uses of the material, which also apply to landfill sites. The existence of vinyl chloride emissions from landfills above the set limits has been known for some years, as has the presence of vinyl chloride in leachate waters. One example was described by Walsh (1) for the Lees Lane, Ohio and Port Washington, New York landfills. Both are municipal landfills, and in the Port Washington facility particularly, no hazardous wastes were permitted and the restriction was reportedly stringently enforced. However, analysis of landfill gases in monitoring wells at these sites over a 2-year period showed the mean vinyl chloride concentration at Lees Lane was 28 ppm (range 0-188 ppm) and at Port Washington it was 15.84 ppm (no range reported). The detection of vinyl chloride at landfills and in leachates where no vinyl chloride-containing wastes were reportedly placed led to initiation of research to determine the source(s) of vinyl chloride emission. One possible source of vinyl chloride emission from landfills is the biological degradation of chlorinated organic solvents. The anaerobic degradation of chlorinated ethenes and ethanes in groundwater systems has been observed by various workers (2-5). The formation of vinyl chloride in microcosms simulating a groundwater environment was reported to result from the reductive dechlorination of dichloroethenes (2). Dichloroethenes are presumably formed by a similar mechanism from tetra- and trichloroethenes 116 ''(3). The transformation of isotopically labeled trichloroethene to labeled 1,2-dichloroethene by soil microbes has been reported (4). In all of the studies, no transformation of chlorinated ethenes was observed when the soil samples or microcosms had been sterilized. The degradation of chlorinated organics by reductive dehalogenation has also been supported by data from studies conducted at solvent recovery facilities, solid/hazardous waste landfills, and solvent contamination near an industrial facility (5). The further away from the solvent sources the higher the proportion of degradation products. It was also noted that the higher the concentration of other organ- ics the greater the degradation of chlorinated solvents. The degradation scheme of chlorinated ethenes and ethanes via reductive dehalogenation is summarized in Figure 1. The existence of other degradation schemes was supported by formation of chloroethane from cis-1,2-dichloroethene and the fact that in no cases can the amount of substrate lost be accounted for by the observed products (2). Actively decomposing dump sites differ from normal soil conditions because of the higher concentration of degradable organic material and the larger populations of anaerobic microorganisms. The objective of this research was to study the transformation of chlorinated ethenes and ethanes in vials simulating landfills and chemical waste dumps. Because microorga- nisms from actual landfill core samples were not available at the time of this presentation, experiments were performed using microorganisms from the anaerobic digester at a municipal wastewater treatment facility in which no industrial solvents are added. Many of the organisms in this source should be the same as those found in a landfill since municipal landfills are used to deposit material from sewage treatment facilities. EXPERIMENTAL Two methods were initially used to prepare vials for experimentation. Initially, samples were prepared in a glovebag and solutions degassed to remove oxygen. Sample preparation involved placing 30 ml of a 10 percent anaerobic reactor fluid diluted with organic-free deionized water in a 40 ml septum cap vial and spiked with 10 ul of a solution containing a chlorinated compound dissolved in methanol at 10 mg/ml. The second procedure involved placing 30 ml of a 20 percent anaerobic digester solution in the septum cap vial without trying to exclude oxygen. The vials were then incubated for one week allowing the microorganism present to deplete the oxygen then the spike was injected through the septa cap and the vial inverted. Microcosms with no additions or methanol only were prepared in a similar manner. Heat treated microcosms were prepared by duplicate exposure for 20 minutes at 121°C on consecutive days. The vials were incubated in a dark growth chamber kept at 35°C or 50°C for periods of one to ten weeks, when duplicate samples were analyzed by the purge-trap technique. Microbial activity was monitored by analysis of the headspace gases in the vials. Levels of fixed gases were estimated using a Carle Series 400 AGC gas chromatograph set up for refinery gas analysis. Microbial activity resulted in the loss of oxygen and formation of methane and carbon dioxide. The autoclaved vials showed no loss of oxygen or methane formation and were considered sterile. 117 ''The highest available purity of tetrachloroethene, trichloroethene, cis- 1,2-dichloroethene, trans-1,2-dichloroethene, 1,1-dichloroethene, 1,1,2,2- tetrachloroethane, 1,1,1-trichloroethane, 1,1,2-trichloroethane, 1,1- dichloroethane, and 1,2-dichloroethane was purchased from Aldrich Chemical Company. Isotopically labeled trichloroethene containing one C-13 atom was purchased from Burdick and Jackson Laboratories. Vinyl chloride standards at concentrations of 1, 10, and 100 ppm were prepared by Scott Specialty Gases. Chemicals were used as received and purity checked by GC and GC/MS. The purge-trap analytical techniques used were similar to those reported by Bellar et al. (6). The entire contents of the 40 ml septum cap vials were purged using a needle purge apparatus reported by Dressman and McFarren (7). The traps employed contained 7 cm of Tenax TA backed by 7 cm of Sperocarb. The gas chromotograph employed a 6-foot glass column filled with Chromosorb 101 for the first third and the remainder with Chromosorb 102 (7). This column resolves vinyl chloride from water and the cis and trans isomers of 1,2-dichloroethene (not done on EPA 601 column). A photoionization detector and a flame ionization detector were connected in series, the photoionization detector responding to the chlorinated ethenes and not the ethanes. The gas chromatograph-mass spectrometer employed a similar column with a molecular jet separator for sample introduction. RESULTS AND DISCUSSION Initial screening experiments for vinyl chloride formation employed vials prepared by the first procedure and spiked with all of the chlorinated ethenes and ethanes used in this study (spike ~100 wg). It was found that by six weeks incubation at 35°C vinyl chloride (typically 0.03 to 0.01 yg) was formed from all of the chlorinated ethenes. The highest level of vinyl chlor- ide (0.06 ug) was found at three weeks incubation with 1,1-dichloroethene. At three weeks the transformation of C-13-labeled trichloroethene occurred producing 1,1-dichloroethene (0.22 yg), trans-1,2-dichloroethene (0.08 yg), cis-1,2-dichloroethene (0.16 yg), and vinyl chloride (0.02 yg) with some loss of the isotopic label. Little change in the concentrations of the chlorinated ethenes was noted after three weeks of incubation, which also corresponds to the time required for the disappearance of the initial methanol spike. Vinyl chloride levels actually decreased between three and six weeks for 1,1-dichloroethene. Incubation of the anaerobic microbes for three weeks with vinyl chloride showed a loss of 40 percent and the formation of 1,1- dichloroethene. Vinyl chloride (0.44 to 0.01 yg) was also formed from the chlorinated ethanes studied. The highest levels of vinyl chloride (0.44 yg) found for all compounds was after one week incubation with 1,1,2-trichloroethane. The transformation of chlorinated ethanes is exemplified by 1,1,2,2-tetrachloro- ethane. At six weeks incubation 1,1,2,2-tetrachloroethane was transformed to 1,1,2-trichloroethane (1.8 ug), trichloroethene (4.2 yg), cis-1,2-dichloro- ethene (8.0 wg), trans-1,2-dichloroethene (4.2 ug), 1,1-dichloroethene (0.42 wg), and vinyl chloride (0.05 wg). Vials incubated at 50°C with the chlori- nated solvent spikes also showed vinyl chloride formation after a similar incubation time but the levels were lower than those at 35°C. 118 ''The transformation of 1,1-dichloroethene and 1,1,2-trichloroethane was investigated in further detail using vials prepared by the second procedure. Microbial activity was much higher in these vials. Vinyl chloride levels from 1,1-dichloroethene were 1.1 yg at one week and 2.0 yg at two weeks incubation. From 1,1,2-trichloroethane the vinyl chloride levels were 4.3 yg at one week and 5.8 yg at two weeks of incubation. Blank and methanol only experiments did not produce vinyl chloride. Purge-trap analysis of blanks showed no change with time of incubation but the methanol only experiments produced methanethiol and dimethyldisul fide. Autoclaved vials were prepared with 1-(C-13) trichloroethene and 1,1,2- trichloroethane and compared with controls containing no_ chlorinated hydrocarbons. After five weeks growth at 35°C, autoclaved and unautoclaved vials spiked with C-13 labeled trichloroethene were analyzed by GC/MS. The chromatographs are shown in Figure 2 for two samples. Labeled trichloroethene was converted to labeled vinyl chloride (0.1 wg), labeled 1,1-dichloroethene (<0.5 yg), 80% labeled trans-1,2-dichloroethene (<0.5 yg) and 90% labeled cis 1,2-dichloro- ethene (<0.5u in the unautoclaved vials with no vinyl chloride formation in the autoclaved vials. A similar comparison is shown in Figure 3 for vials spiked with 1,1,2-trichloroethane. The unautoclaved vials yielded vinyl] chloride (2.4 wg) with 1,1-dichloroethene and 1,2-dichloroethane at similar levels. The autoclaved vial did not produce vinyl chloride but produced 31 ug of 1,1l-dichloroethene. The transformation of 1,1,2-trichloroethane to 1,1-dichloroethene appears to be independent of microbial activity. CONCLUSIONS Chlorinated ethenes and ethanes are transformed to vinyl chloride by anaerobic samples which simulate landfill conditions. The degradation scheme is more complex than that shown in Figure 1. The chlorinated ethanes are transformed to chloroethenes including vinyl chloride. The dechlorination reactions may also be reversible. Higher levels of vinyl chloride were found using 1,1-dichloroethene than from cis or trans-1,2-dichloroethene. Work in this area is continuing with microcosms prepared with actual landfill material. ACKNOWLEDGMENTS The authors wish to thank the California Air Resources Board for funding the research under contract #A4-154-32 and the Richland Wastewater Treatment Facility for assistance in acquiring digester samples. REFERENCES 1. Walsh, J. J. Hazard Wastes Environ. Emerg. Mgt. Prev. Cleanup Control, Conf. Proc., 1984, 146-153. 2. Lage, G. B.; Parsons, F. Z.; Nassar, R. S.; Lorenzo, P. A. Environ. Sci. Technol., 1986, 20, 96-99. ie ''Parsons, F. Z.; Wood, P. R.; DeMarco, J. J.-Am. Water Works Assoc., 1984, 76, 56-59. Kleopfer, R. D.; Easley, D. M.; Haas, B. B., Jr.; Deihl, T. G.; Jackson, D. E.; Warrey, C. J. Environ. Sci. Technol., 1985, 19, 277-280. Cline, P. V.; Viste, D. R. Municipal and Industrial Waste, Annual Madison Waste Conf., 1984, 14-29. Bellar, T. A.; Lichtenberg, J. J.; Eichelberger, J. W. Environ. Sci. Technol., 1976, 10, 926-930. Dressman, R. C.; McFarren, E. F. J. Chromatogr. Sci., 1977, 15, 69-72. 120 ''>» 1,1-dichloroethene tetrachloroethene —» trichloroethene —» t-1,2-dichloroethene —» vinyl chloride —™ c-1,2-dichloroethenes-™” X 1,1,1-trichloroethane —» 1,1-dichloroethane —»® chloroethane FIGURE 1. Pathways for Dehalogenation of Ch!orinated Ethenes and Ethanes in Anaerobic Environments 121 ''Total Abundance from 33 to 200 amu v v o tl ec! Full Scale = 30000 o ! | @ an | Unautoclaved ex c er 35 g ® or 4 Vinyl S= © = Qo! } Chloride ® s g oy ~ OO : 2 ‘. © ol | lon 63.0 Ow ano NS cy Full Scale=125 § § -65 “Ee Ty o> crc 2° ! | Ls 6.9 OO rot ' "s nv. eo '\ rtre4 \ --N\ — * Awe! * 1 \ iN t \ > { \ \ Autoclaved TT - o { 2 51 3 Poy lon 63.0 = * 1} Full Scale = 125 O o! | © Py ¢ 2 | ! £2 Ss! F = Benzene 1,1-Dichloroethene Dichloroethane -—— = Ce ) a ‘ \ \-—e P, iN D Saaan ‘ \, Cc Cc ( Tri § ( “N 4 . X Trichloroethene Dimethyldisulfide . 1,2- - Smee eH -- =m --5 - --oO X 4 ~-o-ocroOr 7 ne ee = mee ee eee Nw Autoclaved 1,1,2-Trichloroethane! lon 62.0 Full Scale = 125 Trichloroethene Benzene 1,1-Dichloroethene ? \ \ \ i i ! ! ! ! J \ 1 8 \ 1 4 \ : we ( 4 4 ( ( 1 ! \ ! \ \ > ” ‘\ é VL ( \ / \ zy j ! t 5 10 15 Time (min) FIGURE 3. GC/MS Chromatographs from Purge-Trap Analysis of Autoclaved and Unautoclaved Vials Spiked with 1,1,2-Trichloroethane 123 ''Non-technical Summary: The Transformation of Chlorinated Ethenes and Ethanes by Anaerobic Microorganisms Trace (parts per million) amounts of vinyl chloride have been identified in gases emitted from landfills, and in groundwater, even in places where the dumping of vinyl chloride-containing or other hazardous or toxic wastes has never been permitted. Although the amounts are small, vinyl chloride is a carcinogen and the problem appears to be nation-wide. ' The California Air Resources Board is sponsoring a research project at Battelle's Pacific Northwest Laboratories to determine the source of vinyl chloride appearing in groundwater and landfill gases. The research is being performed by Dr. Peter M. Molton, Dr. John W. Pyne, and Mr. Richard T. Hallen. Sample material was taken from a landfill site in Northern California, and one in Southern California. Samples were taken from 3 to 10 feet below the cover soil, and contained (among other unidentifiable household and office refuse at various stages of decomposition) dirt, newspapers, grass clippings, glass, and plastic. The samples were then placed in a plastic bag and flushed with gas to remove oxygen to ensure an anaerobic environment for experimental work. This is important because anaerobic bacteria which normally decompose refuse in landfills are considered a primary potential source of vinyl chloride when exposed to substances containing organic chlorine. Small amounts of chlorinated solvents in methanol. (a carbon source for the bacteria) were introduced into the experimental vials containing small amounts of landfill material. Various common solvents were used, including trichloroethylene, a dry-cleaning solvent. To ensure that any vinyl chloride detected was in fact 124 ''produced by bacterial degradation of the trichloroethylene, it was labeled with a naturally-occurring non-radioactive carbon isotope (carbon-13), and the presence of carbon-13 in the vinyl chloride produced was confirmed by mass spectrometry. Vinyl chloride formation from trichloroethylene was confirmed, after 1 and 3 weeks incubation. Prior to testing the California landfill samples, the researchers took samples of primary sewage sludge from the anaerobic digesters at the City of Richland, Washington, wastewater treatment plant. This material represents anaerobically digested sewage that presumably has never been exposed to any form of chlorinated industrial solvent, but may contain organisms similar to those in the landfill samples. When exposed to labeled trichloroethylene under the same conditions as the landfill samples, vinyl chloride was also formed, showing that the ability of microorganisms to degrade solvents is not restricted to those found in landfills. In both cases, vinyl chloride was formed by anaerobic dechlorination of common chlorinated solvents. Such solvents may enter landfills in items such as paints, paint thinner, plastic wrap, garden hoses, and phonograph records. Older landfills may contain more vinyl chloride, as this compound was not recognized as carcinogenic until recently, and permitted levels of the residual gas contained in PVC plastic used to be much higher than they are today. Decomposition of PVC (polyvinyl chloride)itself does not appear to contribute significantly to vinyl chloride release. In summary, vinyl chloride emissions from landfills appear to come from household and industrial materials, by conversions by microorganisms indigenous to the landfills themselves. 125 ''Although not currently considered a health hazard, at observed levels, these emissions deserve further study and monitoring. Vinyl chloride is not persistent in the environment (50% decomposes in sunlight in 4.5 hr), but can be a cause for concern in any amount. The problem may be alleviated by minimizing the amounts of chlorinated solvents dumped into municipal landfills, by collecting and burning the evolved gases, or by more innovative techniques currently under development. This project will help provide additional information from which useful and meaningful regulatory actions can be taken. Battelle's Pacific Northwest Division, with laboratories in Richland, Seattle, and Sequim, Washington, performs research and development for industrial sponsors and government agencies. The Division is a component of Battelle Memorial Institute, the world's largest independent research institute. Other major Battelle research facilities are located in Columbus, Ohio; Frankfurt, West Germany; and Geneva, Switzerland. 126 ''''10. APPENDIX C: INDUCTIVELY COUPLED PLASMA (ICP) ELEMENTAL ANALYSIS PROCEDURE Sample ignited in a platinum crucible using NayC03 as follows: Place about lg Na yC03 in a platinum crucible. Weigh between 0.36-0.40g sample. Place the sample in the platinum crucible. Add 1g NayC03 over the sample. Place sample in a muffle furnace. Cover with a platinum lid and set temperature of muffle at 150 C and heat for 15 min. Increase temp. at 50 increment for 15 min. to a temp. of 950 C. Cool to room temp. Leach melt with Hj0 in a 300 ml beaker. Acidify with 50 ml 1:1 HCI. Transfer solution into a 250 ml volumetric flask. Dilute to volume with deionized water. Pipet 10 ml of solution into a 100 ml volumetric flask. Dilute to volume with deionized water. Analyze the sample using an Inductively Coupled Argon Plasma/Atomic Emission Spectometer (ICP/AES). Calculation: Wt% element=Mg/ml element x 2500 x 100 Mg sample 127 '''' ''''U.C. BERKELEY LIBRARIES €005199084 ''''