Document RaEk4KRO8qBBDjpyG3LvEo8Mz

1 -1 3 0 1 ] ] I I a i ] J j j j PHASE I REVIEW STUDY OF VINYL CHLORIDE NATIONAL EMISSIONS STANDARD DRAFT March 16, 1981 Prepared for: Environmental Protection Agency Office of Air Quality Planning and Standards Chemical and Petroleum Branch Prepared by: TRW, Inc. Environmental Engineering Division Post Office Box 13000 Research Triangle Park, North Carolina 27709 01644? GENC DISCLAIMER This draft report was submitted to the Emission Standards and Engineering Division of the Office of Air Quality Planning and Standards < of the Environmental Protection Agency by TRW Environmental Engineering Division, Research Triangle Park, North Carolina in fulfillment of Contract No. 68-02-3063. The contents of this report are reproduced herein as received from TRW. The opinions, findings, and conclusions expressed are those of the authors and not necessarily of the Environ mental Protection Agency. Mention of company or product names is not to be considered as an endorsement by the Environmental Protection Agency. ABSTRACT This VC NESHAP Review Study assesses the current VC regulation through an investigation of emission control techniques and techno logical developments in the industry. The study encompasses evaluations of existing and new control technologies, sources not regulated by the standard, and enforcement and compliance experience since promulgation of the standard. Information and data evaluated during this study were obtained through literature searches, plant visits, and interviews with industrial representatives and EPA Regional Office personnel. The results of this review study will form the basis for possible revison of the existing standard. * iii GENC 016449 ACKNOWLEDGEMENTS Thisistudy was performed under NSS Contract No. 68*02-3063 with the U.S. EPA Office of Air Quality Planning and Standards. The success of the study was dependent on information submitted voluntarily by the industries regulated under the VC NESHAP. Many representatives of those industries were very cooperative, and their contributions are cited throughout this report. The authors would like to acknowledge their assistance. Several regional EPA personnel also contributed extensive information and provided data necessary to evaluate the current status of emission control in the industry. The authors would like to especially recognize the guidance of Leslie B. Evans, the EPA Project Officer on this review study. GEWC 0U450 tv AAQS APRS ABBREVIATIONS AND ACRONYMS USED IN THIS REPORT Ambient Air Quality Standard Automatic Pressure Reduction System BACT BAT BID Best Available Control Technology Best Available Technology Background Information Document CAA CARB CFR CMA CTA CTG Clean Air Act California Air Resources Board Code of Federal Regulations Chemical Manufacturers Association Chain Transfer Agent Control Techniques Guidelines DOT DSSE Department of Transportation Division of Stationary Source Enforcement EDC Ethylene Dichloride EPA Environmental Protection Agency FDA Food and Drug Administration FID Flame ionization detector FR Federal Register HC Hydrocarbon KO Knock-out (vessel) LEL Lower explosive limit NAAQS NESHAP NIOSH National Ambient Air Quality Standards National Emission Standard for Hazardous Air Pollutants National Institute for Occupational Safety and Health v GENC 016451 NPDES NSPS National Pollution Discharge Elimination System New Source Performance Standards OAQPS OSHA OVA Office of Air Quality Planning and Standards Occupational Safety and Health Administration Organic vapor analyzer PMN Premanufacturing Notice ppm Parts per million PSD Prevention of significant deterioration PVC Polyvinyl chloride RACT RCRA ROL RVC RVD Reasonably Available Control Technology Resource Conservation and Recovery Act Reactor opening loss Residual vinyl chloride Relief valve discharge SCAQMD SIP SOCMI SPI SSEIS South Coast Air Quality Monitoring District State Implementation Plans Synthetic Organic Chemical Manufacturing Industry Society for the Plastics Industry Standard Support and Environmental Impact Statement TSCA Toxic Substances Control Act VC Vinyl chloride VOC Volatile organic compounds 6ENc vi TABLE OF CONTENTS Section Page 1.0 EXECUTIVE SUMMARY ..................................................................................... 1-1 1.1 Introduction.....................................................................................1-1 1.2 Process Developments .................................................................... 1-1 1.3 Control Technology Summary ................................................. 1-2 1.4 Primary Controls ............................................................................ 1-2 1.5 Relief Valve Discharges ............................................................ 1-4 1.6 Resin Stripping............................................................................ 1-4 1.7 Fugitive Emissions ........................................................................ 1-5 1.8 Reactor Opening Loss.................................................................... 1-6 1.9 Enforcement and Compliance Experience.................................. 1-6 1.10 Unregulated Sources........................................................................ 1-7 1.11 Impact of Other Regulations....................................................... 1-7 2.0 INTRODUCTION..................................................................................................2*1 2.1 Background Information ................................................................ 2-1 2.2 Scope of Review Study.................................................................... 2"2 2.2.1 Areas of Concern................................................................2*2 2.2.2 Review Study Methods....................................................... 2-9 2.3 The Vinyl Chloride EmittingIndustry ..................................... 2-9 2.3.1 Current Number and Geographical Distribution. . 2-9 2.3.2 Influence of the Standard onIndustry....................... 2-11 2.3.3 Industrial Trends ........................................................... 2-12 2.5 References for Chapter 2.............................................................. 2-14 3.0 PROCESS DESCRIPTION ................................................................................. 3-1 3.1 Introduction..................................................................................... 3-1 3.2 Production of Ethylene Dichloride...........................................3-4 3.2.1 Direct Chlorination ofEthylene ................................ 3-5 GENC 016453 vi i Section 3.2.2 Oxychlorination of Ethylene ...................................... 3-5 3.2.3 Purification of Ethylene Dichloride ..................... 3-7 3.3 Production of Vinyl Chloride .................................................... 3-9 3.3.1 Formation of Vinyl Chloride by Dehydrochlorination of Ethylene Dichloride. . . 3-9 3.3.2 Purification of Vinyl Chloride.................................. 3-10 3.3.3 Emissions for Typical EDC/VC Plants ..................... 3-10 3.4i Production of Polyvinyl Chloride ........................................... 3-11 3.4.1 Free Radical Polymerization ....................................... 3-16 3.4.2 Unloading of VC at PVC Plant Sites.......................... 3-17 3.4.3 Mixing, Weighing and Holding Vessels..................... 3-18 3.4.4 Suspension Polymerization ........................................... 3-18 3.4.5 Dispersion Polymerization ........................................... 3-21 3.4.6 Bulk Polymerization ........................................................ 3-23 3.4.7 Solution Polymerization ............................................... 3-28 3.4.8 Polymerization Reactors ............................................... 3-32 3.4.9 Emissions for a Typical PVC Plant .......................... 3-38 3.5 References for Chapter 3 ............................................................. 3-39 4.0 CONTROL TECHNIQUES USED TO COMPLY WITH THE EXISTING EMISSIONS STANDARDS .................................................................................. 4-1 4.1 Discharge of Exhaust Gases to the Atmosphere .................. 4-1 4.1.1 Introduction. . . 4-1 4.1.2 Incineration. . . 4-7 4.1.3 Steam Boilers . . 4-11 4.1.4 Flares...................... 4-11 4.1.5 Carbon Adsorption 4-13 4.1.6 Solvent Absorption 4-14 4.1.7 Refrigeration . . 4-15 4.1.8 Other Controls. . 4-16 4.2 Relief Valve Discharges. 4-17 4.2.1 Introduction.................................................... 4-17 4.2.2 Emissions From Safety Relief Valves . 4-18 4.2.3 Relief Valve Discharges From Reactors 4-23 4.2.3.1 Process Variations ................. 4-23 4.2.3.2 Causes of Reactor Discharges 4-25 viii GENC 016454 Section Paae 4.2.3.3 Prevention of Reactor Discharges . . . 4-28 4.2.3.3.1 Current Generic Preventive Methods..........................................4-28 4.2.3.3.2 Preventive Systems Currently In Use.........................4-40 4.2.4 Nonreactor Relief Valve Discharges.............................4-48 4.3 Resin Stripping....................................................................................4-51 4.3.1 Introduction............................................................................4-51 4.3.2 Suspension Resin Stripping..............................................4-52 4.3.3 Emulsion Resin Stripping.................................................. 4-56 4.3.4 Bulk Resin Stripping.......................................................... 4-57 4.3.5 Solution Resin Stripping..................................................4-57 4.3.6 Other Stripping Technologies..........................................4-57 4.4 Fugitive Emissions ........................................................................ 4-58 4.4.1 Introduction........................................................................... 4-58 4.4.2 Equipment Specifications.................................................. 4-61 4.4.3 Operational Procedures...................................................... 4-65 4.4.4 Leak Detection and Elimination Programs .... 4-66 4.4.5 Inprocess Wastewater.......................................................... 4-78 4.5 Reactor Opening Loss....................................................................... 4*80 4.5.1 Introduction............................................................................4-80 4.5.2 Solvent Cleaning................................................................... 4-82 4.5.3 Steam Piston............................................................................4-83 4.5.4 Water Piston............................................................................4-83 4.5.5 Reactor Purge Air Blower.................................................. 4-84 4.5.6 Steam Purge............................................................................4-84 4.5.7 Redox Catalysis............................................................1 . 4-84 4.5.8 Water Jet Cleaning............................................................... 4-84 4.5.9 Clean Reactor (Closed Cleaning) .............................. 4-85 4.5.10 Nitrogen Purge........................................................................4-85 4.5.11 Slurry Backfill ................................................................ 4-85 4.5.12 Calculated Emissions...........................................................4-86 4.6 References for Chapter 4............................................................... 4-88 ix GENC 016455 Section Page 5.0 ENFORCEMENT AND COMPLIANCE EXPERIENCE ........................................... 5-1 5.1 Introduction.........................................................................................5-1 5.2 Intent of the Standard................................................................... 5-1 5.3 Standards for EDC and VC Plants.................................................. 5-2 5.4 Exhaust Gases to the Atmosphere.................................................. 5-2 5.5 Inprocess Wastewater .................................................................... 5-3 5.6 Reactor Opening Loss.................................................................... 5-3 5.7 Relief Valve Discharge ................................................................ 5-4 5.8 Resin Stripping.................................................................................... 5-5 5.9 Sources After the Stripper ........................................................ 5-7 5.10 Fugitive Emissions ......................................................................... 5-7 5.11 Leak Detection and Elimination Programs................................. 5-8 5.12 Emissions Testing and Analysis ............................................... 5-8 5.13 Reporting................................................................................................. 5-10 5.14 Recordkeeping.........................................................................................5-10 5.15 NESHAP Applicability Determinations..........................................5-10 5.16 References for Chapter 5............................................................... 5-13 6.0 UNREGULATED SOURCES OF VINYL CHLORIDE ........................................... 6-1 6.1 Introduction......................................................................................6-1 6.2 Sources Identified During Original Study .......................... 6-1 6.2.1 Fabricating Operations................................................... 6-1 6.2.2 Miscellaneous Sources .................................................... 6-4 6.3 New Sources Identified During ReviewStudy ........................ 6-5 6.3.1 Mobile Sources of Emissions ...................................... 6-5 6.3.2 Nonplant Transfer Facilities.......................................6-6 6.3.3 Solid Waste Drying Facilities..........................."k . 6-7 6.3.4 Disposal Facilities (Landfill).................................. 6-7 6.4 References for Chapter 6............................................................ 6-8 7.0 IMPACT OF OTHER REGULATIONS ................................................................ 7-1 7.1 Introduction......................................................................................7-1 7.2 Clean Air Act (CAA)......................................................................... 7-2 7.2.1 Carcinogen Rule............................... .................................7r3 7.2.2 Prevention of Significant Deterioration (PSD) . 7-3 7.2.3 NESHAP Delegation to States ....................................... 7-7 GENC 014456 x Section Page 7.3 Resource Conservation and Rcovery Act (RCRA) ................. 7-10 7.4 Toxic Substances Control Act (TSCA)..........................................7-11 7.5 Clean Water Act.................................................................................... 7-12 7.6 Safe Drinking Water Act................................................................... 7-13 7.7 Hazardous Materials Transportation Act .............................. 7-13 7.8 Occupational Safety and Health Act..........................................7-13 7.9 Superfund Legislation....................................................................... 7-14 7.10 Food and Drug Administration Regulations.............................7-14 7.11 'Other State and Local Regulations...........................................7-15 7.12 References for Chapter 7............................................................... 7-16 APPENDIX A - Vinyl Chloride NESHAP APPENDIX B - Regional EPA and Industrial Contacts APPENDIX C - Current Industrial Sources xi GENC 016457 1.0 EXECUTIVE SUMMARY 1.1 INTRODUCTION This Phase I review study assesses the current National Emission Standard for vinyl chloride (VC) by investigating emission control techniques. This review evaluates technological developments in the jindustry and provides a preliminary basis for possible standard revision, Recommendations to revise the standard would be supported by a more detailed Phase II study that would develop a Background Information Document (BID). This review study was conducted under a contract awarded to TRW's Environmental Engineering Division by the U.S. Environmental Protection Agency (EPA). The impetus for the review was the proposed / Carcinogen Policy's requirement for periodic revievTof National Emission/' Standards for'Hazardous Air Pollutants (NESHAP) regulations' The review study 'focused on four areas which are summarized below: Technologies currently used for compliance. Existing sources identified during the original support study but not subject to the current regulation. Emission sources not identified during the original support study. Enforcement and compliance experience. Additional details can be found in corresponding sections1 of the text (indicated in parentheses). A description of the processes involved in VC production and polymerization is presented in Chapter 3. 1.2 PROCESS DEVELOPMENTS The sources subject to the VC regulation are ethylene dichloride (EDC), VC, and polyvinyl chloride (PVC) facilities. Recent modifications of processes in these facilities include: GEMC 016453 EDC/VC - The "balanced process" utilizing direct chlorination and oxychlorination is the most common process used for pro duction of EDC and VC. Many newer plants and some existing plants have incorporated oxygen plants as part of the EDC process. The change from air to pure oxygen as a feedstock ' has resulted in reduced emissions from the oxychlorination .vent. (Section 3.2.1 and 3.2.2) PVC - Newer plants are incorporating larger reactor systems resulting in increased capacity, fewer reactor openings and reduced emissions. These.large reactor systems have also accounted for a decrease in the production of specialty PVC ,:.' ' resins (Section 4.2.3.1) 1.3 CONTROL TECHNOLOGY SUMMARY Most of the discussion regarding VC control technology is focused on PVC plants because these facilities contribute proportionately more emissions. For this reason, most of the requirements in the VC regu lations pertain to PVC plants. The characteristics inherent in the batch processes of these plants account for the relative difficulty in emission control implementation. 1.4 PRIMARY CONTROLS Control devices, applied to reduce VC emissions when exhaust gases are discharged to the atmosphere, include: Incineration - This method represents the most prevalent means of primary control and is the only one used by both EDC/VC and PVC plants. (Solvent absorption and carbon adsorption are also used as primary controls in PVC plants, but in EDC/VC plants they are used on a smaller scale as part of the process or as a means to reduce fugitive emissions.) Although thermal incineration is most commonly used, two plants are using catalytic oxidation systems. In most cases, 1-2 GENC 016459 thermal incineration is reducing emissions below the required 10 ppm. Most units have scrubbers to prevent HC1 emissions. (Section 4.1.2) Solvent Absorption - This method is effective in reducing emissions below 10 ppm. It is used as primary control only in PVC plants. VC is recovered by this method rather than destroyed (as with incineration). (Section 4.1.6) Carbon Adsorption - As a primary control, this method is not as effective (in most cases) as incineration of solvent absorption in achieving the 10 ppm level. Usually,`carbon adsorbers must be supplemented by other primary controls (e.g. , incineration). Some plants that initially,selected -............... ....... ........ J*#1**' ........... ............... .1 ' up-mi----^ carbon adsorption as a primary control ^foreyhaustiqa'sesa --replaced the carb. -on .bed.s t. ottatll;--y with^anotner control device. % Carbon adsorption is effective on a smaller scale for reducing fugitive emissions or for recovering VC from some exhaust gases. (Section '4.1. 5) t Other Controls - These include steam boilers, flares, refrigeration systems and containment devices. Boilers are not usually used for primary control due to corrosion-problems. /Efficiency of VC reduction in flares has not been determined. Secondary pollutants (e.g., noise, smoke) have also been noted as a disadvantage of flares. Refrigeration systems are generally used as part of recovery systems or as back-up control in case of primary control breakdown. Contaihment"~devices are used to? reduce emissions to'the atomosphere ahd"Tnclucle"gasho1ders and/ ^e^fTze3'fioi'd1ng"vesse)sT These'dev ices collect vapors froiir^fio^^quipment vents and feed the VC recovery system and/or primary control device. They are also used in some cases to collect and hold emissions when the primary control is down for maintenance. (Sections 4.1.3, 4.1.4, 4.1.7, 4.1.8, and 4.2.3.3.1) 1-3 GENC 016460 1.5 RELIEF VALVE DISCHARGES Relief valve discharges cause short term excursions of VC emissions and, according to EPA regional enforcement personnel, represent the single most difficult enforcement problem. Contention is caused by the wording in the regulation that allows only "emergency |relief valve discharges. (Sections 4.2.1 and 5.7) Regional personnel indicate relief valve discharges continue to occur, but the frequency and magnitude vary throughout the industry. PVC reactors are responsible for the greatest frequency and largest quantities of emissions, but nonreactorrelated discharges contribute approximately 34 percent of the relief valve events and 20 percent of the total quantity discharged. (Section 4.2.2) Reactor releases are affected by the type of polymerization process and whether the newer reactors are employed. The newer reactors are larger and incorporate more instrumentation. They appear to provide better control over upset conditions that could result in a relief valve discharge. (Section 4.2.3) Procedures for prevention of relief valve discharges vary from plant to plant. Many of these procedures have reduced the frequency of discharges. 1.6 RESIN STRIPPING k Because stripping techniques are different for each type and grade of resin, a wide range of residual VC (RVC) content in the stripped resin has been noted. Suspension Resins - These resins represent the highest percentage of total PVC production. '^^MTr^lnethod^Ts^contTniw^Yteam^rfppTng. Many processors are attaining levels much lower than the required 400 ppm RVC - some less than 20 ppm. (Section 4.3.2) 1-4 GENC 016441 t Dispersion Resins - Dispersion or emulsion resins are usually vacuum stripped batchwise in the reactor or in separate batch stripping vessels. Continuous stripping technologies for dispersion resins are not as advanced as for suspension resins (probably because the demand for dispersion resins is not as great). Required emission levels for dispersion resins (2,000 ppm) are being met and, in some cases, processors are regularly achieving levels below 1,000 ppm. Latex resins, produced by the dispersion process and usually sold undried, are required to meet a 400 ppm level. These resins are sensitive to heat and shear stress, creating difficulties in stripping efficiently. (Section 4.3.2) Bulk Resins - The characteristics of these resins (e.g., uniform porosity and size) enhance stripping efficiency. Two ' bullcTresin processors reported 6-month average stripping level $ below 40 ppm. *Steam stripping (under vacuum) in the reactor is used for these resins. (Section 4.3.4) Solution Resins - Only one plant produces solution resins'. This process is unique among stripping procedures because'no particulate resin form exists. Stripping is a distillation process with a high efficiency averaging levels of 10 ppm RVC. (Section 4.3.5) 1.7 FUGITIVE EMISSIONS Fugitive emissions represent one of the larger contributions to VC emissions at EDC/VC and PVC plants. PVC plants appear to contribute more fugitives because of the batch process characteristics and the prevalence of plants with many old small reactors. One study done by a PVC pro cessor indicated that, after installation of required equipment to control fugitives and implementation of leak detection and elimination programs, a large reduction was achieved in fugitive emissions. Emissions from their old small reactor system are now 75 percent lower than the industry average fugitive emissions 1-5 GENC 016462 as estimated in the original standard support study. Those from their newer large reactor systems are now 95 percent lower. (Section 4.4.1) EPA Regional personnel indicate that almost all plants have installed the required equipment specified in the current regulation and are following the required operational procedures. Some plants have received approval for equivalent equipment. (Sections 4.4.2 and 4.4.3) 1 The leak detection and elimination programs represent the greatest variability among plants surveyed. Leak detection programs and routine surveys with portable monitors vary widely among the plants. In most cases, other requirements have been addressed adequately (e.g., area monitoring and plans to eliminate leaks) and fugitives have been lowered. (Section 4.4.4) 1.8 REACTOR OPENING LOSS (ROL) Control of this emission source is achieved through various technologies and process modifications. Several methods have been developed that are effective in reducing ROL emissions to levels below those required by the regulation. "Clean reactor" technology has also contributed to a reduction in emissions from this source. Selection of the type of control (reactor purging) is based on the polymerization process used. There is a problem in emission level determination for those processors whose resins are stripped in the reactor. Because actual . measurements cannot be made to ascertain RVC levels, determination is based on calculations. These may not always be indicative of actual emissions'. (Section 4.5) 1.9 ENFORCEMENT AND COMPLIANCE EXPERIENCE Industrial representatives and regional EPA personnel cited several areas of concern regarding enforcement and compliance experiences under the existing VC NESHAP. While many of these points were specific to either industry or the EPA, several areas reflected viewpoints common to 1-6 GENC 016463 both. Chapter 5 lists and discusses these concerns. Enforcement of relief valve discharges is the most common concern. (Chapter 5) 1.10 UNREGULATED SOURCES Many of the sources not currently regulated under the VC standard were identified during the original study. These include PVC compounders and fabricators as well as processors using VC as^a chemical inter mediate or producing it as a byproduct. *The-implementation of the OSHA workplace standard for VC has resulted in fabricators' receiving resins Wjjtlj^RVC\.levels of 10 ppm or lower. This has greatly reduced emission levels from fabricating facilities. Facilities manufacturing certain pesticides and trichloroethanes use VC as an intermediate, and emission information on these processes was not obtained during this study phase. Unregulated sources producing VC as a byproduct were contacted and they reported that the small amount of VC involved was either incinerated or recycled through recovery systems for use in polymerization processes in another facility. New unregulated sources include mobile sources, nonplant transfer facilities, solid waste drying facilities and disposal sites (landfills). Each of these represents a potential VC emission source and each has been identified as an area of concern by regional EPA personnel. (Chapter 6) 1.11 IMPACT OF OTHER REGULATIONS There has been substantial regulatory activity since promulgation of the current VC NESHAP. The new regulations that will have an effect in reducing VC emissions to the atmosphere are Prevention of Significant Deterioration (PSD) of Air Quality, plans for nonattainment review, and delegation of NESHAP authority, all under the Clean Air Act. These regulations call for the reduction of VC emissions below those previously required by the VC NESHAP. Other new regulations that will also have an effect on levels of VC emissions include the Resource Conservation and Recovery Act (RCRA) for the control of hazardous wastes, the Toxic Substances Control Act (TSCA) which regulates any new chemicals involved in polymer development, and the Clean Water Act requiring the development of effluent guidelines for VC as well as a possible VC drinking water standard. (Chapter 7) 1-7 2.0 INTRODUCTION 2.1 BACKGROUND INFORMATION The vinyl chloride (VC) standard was promulgated in 1976 under Section 112 of the Clean Air Act (CAA), National Emission Standards for Hazardous Air Pollutants (NESHAP), and is applicable to new and existing sources of VC - those plants producing VC and/or polymerizing VC into polyvinyl chloride (PVC). The proposed NESHAP "Policy and Procedures for Identifying, Assessing, and Regulating Airborne Substances Posing a Risk of Cancer" would require that emission standards promulgated under Section 112 be reviewed at intervals of no more than 5 years. These reviews would be used to determine the need for revision of the emission standard. VC was first implicated as a highly specific cause of angiosarcoma, a rare cancer of the liver, by evidence from occupational exposures. Following intensive study, the Occupational Safety and Health Administration (OSHA) promulgated a standard in 1975 to reduce occupational exposure to VC, and the EPA promulgated a standard in 1976 to reduce atmospheric VC emissions. Industry was allowed 2 years, until 1978, to incorporate the necessary controls. jfhe"existing" VC NESHAP/(henceforth referred to as the regulation) is one or the most complex air emissions standards promulgated by the , JiPA. The regulation-is applicable to three different types of facilities plants producing ethylene dichloride (EDC) by the reaction of oxygen and hydrogen chloride with ethylene, plants producing VC by any process, and plants producing one or more polymers containing any fraction of VC. Research and development facilities containing a polymerization reactor capacity greater than 0.2 cubic meters (50 gallons) but no more than 4 cubic meters (1100 gallons) are exempt from all parts of the regulation GENC 016465 except the 10 ppm emission limit. Reactors less than 0.2 cubic meters (50 gallons) are not regulated. Those reactors greater than 4 cubic meters (1100 gallons) are subject to all requirements of the regulation. Each of these facilities are subject to different standards at numerous points in the manufacturing process - numerical emission limits, equipment specifications, and work practice requirements (i.e., working procedures that must be followed by plant personnel). Table 2-1 lists each section of the regulation requiring a specific standard and the type of plant subject to that standard. The current regulation is reproduced in Appendix A. Compliance with the current regulation is determined through testing and monitoring results and extensive reporting and recordkeeping require ments (all of which are conducted by the plant). The plants are required to report to the responsible Regional EPA office and EPA enforcement personnel conduct standard compliance tests and review plant procedures and records periodically. Requirements for reporting and recordkeeping are summarized in Table 2-2. 2.2 SCOPE OF THE REVIEW STUDY Periodic review of regulations is an important part of the standards development program. The purpose of these reviews is to investigate the control techniques applied to industrial processes for reducing VC air emissions. 2.2.1 Areas of Concern The review study conducted to assess the current VC regulation concentrated on four areas of concern: Technologies being used for compliance, Existing sources identified during the original support study but not subject to the current regulation, New emission sources not identified during the original support study, and - Enforcement and compliance experience. Each of these areas served as focal points for research and appropriate study methodology. The original study supporting the current regulation (EPA, 1975) was thoroughly reviewed including all the information submitted 2-2 016466 GENC w-.-w L-~-,, I___ i___ fc2 i------1 6*3 Table 2-1. EMISSION STANDARDS IN THE VC NESHAP Section 61.62(a) 61.62(b) 61.63(a) 61.64(a)(1) 61.64(a)(2) 61.64(a)(3) 61.64(b) 61.64(c) 61.64(d) 61.64(e)(1) Applicability Exhaust gases discharged to the atmosphere from any equipment used in EDC purification. Emissions of VC to the atmosphere from oxychlorination reactors. Discharge of exhaust gases to the atmosphere from any equipment used In VC formation and/or purification. PVC reactor exhaust gases discharged to the atmosphere. Allowable reactor opening losses of vinyl chloride based on the amount of product produced between openings. Discharge to the atmosphere from any manual vent valve on a PVC reactor. Exhaust gases discharged to the atmosphere from a stripper. Exhaust gases discharged to the atmosphere from mixing, weighing, and holding containers which precede reactors. Exhaust gases discharged to the atmosphere from monomer-recovery systems. Emissions of vinyl chloride to the atmosphere from the combination of all sources following the stripper. Pertains to the requirements of residual VC (RVC) levels attained with the stripping process. (continuedT Standard 10 ppm 0.2 g/kg 100% EDC product 10 ppm 10 ppm 0.02 g VC/kg PVC none (except for an emergency) 10 ppm 10 ppm 10 ppm 2000 ppm - dispersion resins (excluding latex) 400 ppm - all other resins (including latex) Plants Involved EDC EDC VC PVC PVC PVC PVC PVC PVC PVC GENC 0 1 6 4 6 59 Table 2-1 Continued Section Appltcability 61.64(e)(2) Emissions of vinyl chloride to the atmosphere from the combination of all sources following the reactor if the plant has no stripper; or from sources following the stripper if the plant uses technology in addition to stripping. 61.65(a) 61.65(b)(1) 61.65(b)(2) 61.65(b)(3) 61.65(b)(4) Discharge to the atmosphere from any relief valve on any equipment in VC service (equipment contains or contacts either a liquid at least 101 by weight VC or gas 101 by volume VC). Fugitive emissions to the atmosphere from loading and unloading lines In VC service (I) opening of the lines (II) VC removed from these lines to meet (1) and ducted to a control system Fugitive emissions to the atmosphere from slip gauges (in VC service) used during loading and unloading operations. Fugitive emissions to the atmosphere from pump, compressor, and agitator seal leakage. Fugitive emissions to the atmosphere from leakage of relief valves on equipment in VC service. (continued) Standard 2 g/kg product from the stripper (or reactor) for dispersion PVC resins (excluding latex) 0.4 g/kg product from the stripper (or reactor) for all other PVC resins (Including latex) none (except for an emergency) Plants Involved PVC EDC, VC and PVC (I) 0.0038 m3 VC 9 STP (II) _< 10 ppm 10 ppm EDC, VC, PVC EDC, VC, PVC 10 ppm minimized by Installation of a rupture disk or by connec tion of discharge to a process line or recovery system EDC, VC, PVC EDC, VC, PVC 1------ L. L____ ~~J i zzm urn L--J ___ J CiJB ____1 Table 2-1. Concluded Section Applicability 61.65(b)(5) 61.65(b)(6) Fugitive emissions to the atmosphere from manual venting of gases. Fugitive emissions to the atmosphere from opening of equipment. (i) before opening 61.65(b)(7) (11) VC removed from equipment to meet (i) and ducted to a control system Fugitive emissions to the atmosphere from sampling. unused sample portions containing > 10% by weight VC sample containers in VC service 61.65(b)(8) VC emissions due to leaks from equipment in VC service. Standard < 10 ppm (1) total amount discharged per opening must be 2%, by volume VC or 0.95 nr (whichever is larger) (it) <_ 10 ppm return to process purge Into closed process system minimize via formal leak detection and elimination (LD i E) program, designed by operator and approved by the Administrator. Plants Involved EDC, VC, PVC EDC, VC, PVC EDC, VC. PVC EDC, VC, PVC : .. fi t ' : . Table 2-2. SUMMARY OF REPORTING AND RECORDKEEPING REQUIREMENTS IN VC NESHAP Section 61.69(a) 61.69(b)(1) (2) 61.69(c) Applicability Owner or operator of any source to which this subpart applies Existing source or new source having start-up date preceding effective date of regulation. New source having start up date after effective date. Owners or operators of above sources 61.70(a) 61.70(b)(1) (2) Owner or operator of any source to which this subpart applies Existing source or new source having start-up date preceding effective date. New source having start up date after effective date. Requirements Initial Report: Notification that equipment and procedural specifications required In Section 61.65 (fugitive emission control) have been implemented. Submittal of above notification within 90 days of effective date of regulation (unless waiver granted). Submittal of above notification within 90 days of initial start-up date. Notification inclusions: List of equipment installed for compliance Description of physical and functional characteristics of equipment Description of methods used for measuring or calculating emissions Statement verifying that equipment is Installed and procedures are being used Semi-annual reporting to Administrator (September 15 and March 15 of each year) containing the information described in subsequent sections. Submittal of notification within 180 days of the effective date (unless waiver granted). Submittal of notification within 180 days of Initial start up date. {continued) Category Initial Reporting Semi-Annual Reporting GENC 0 1 6 4 7 0 ta***j L-"-* L.~,. l. b*SM v-13 - _ J___ 1 no O s o 1&> -vj Section 61.70(c) (I) : Applicability Owner or operator of above sources Same (2) Owner or operator of PVC plants in which stripping 'operations are used . , Tv.-''' .0) Same Table 2-2. Continued Requirements Category Use specified test methods (from Appendix B) unless equivalency or alternative method approved. Semi-Annual Reporting Reporting of any emissions averaged over one hour which are in excess of limits prescribed for emissions from: EDO purification equipment Oxychlor(nation reactors Equipment used for VC formation/purification PVC reactor exhaust gases Control systems for reactor emissions for ROL . .1- Fugitives from loading/unloading lines *- Fugitives from slip gauges Fugitives from manual venting of gauges Fugitives from equipment opening emissions ducted through a control system .. Fugitives from Inprocess wastewater emissions ducted . through a control system . ..>V n 1* 7 Reporting of a record of the VC content in the PVC res/in' using method 107 as follows: '. '' If batch stripped, sample each batch of each grade of resin immediately after stripping -V-* 1' If continuously stripped, sample each grade of resin or at 8 hour Intervals (whichever is more frequent) ' Determine the auantity of materials processed by the stripper on a dry solids basis Report VC content found in each sample, averaged ,r< separately for each type of resin, over eacfi calendar * 1 day and weighted according to the quantity of each grade , of resin produced that day .(.;w 'V Retain records for at least two years as iv - > ` . Report record of Reactor Opening Loss (ROL) emissions Tcontinued) ! f- ~J > - v oa^- -L , . .. * . ' L GENC 0 1 6 4 7 2 Section 61.71(a) (1) (2) Applicability Owner or operator of any source to which this subpart applies Same Same (3) Same (4) Same 61.64(a)(3) PVC plants only 61.65(a) EDC, VC and PVC plants Table 2-2. Concluded Requirements Retention of the following Information (for minimum of two years): Record of leaks detected by the VC monitoring system Record of leaks detected during routine monitoring with the portable hydrocarbon detector and action taken to repair the leaks Record of emissions measured by the continuous VC monitoring of sources listed In 61.70(c)(1) Dally operating record for each PVC reactor Including pressures and temperatures Manual vent valve discharge. Operator of source must notify Administrator within 10 days of occurrence and submit a report containing the following: Source of relief valve discharge Nature and cause of discharge Date and time of discharge Approximate quantity of VC lost during discharge a Method used for determining VC loss Action taken to stop discharge Action to be taken to prevent future discharges Relief valve discharge. Operator of source must notify Administrator within 10 days of occurrence and submit a report containing the following: (Same as requirements for "Manual vent valve discharge" above). Category Recordkeeping Exception Reporting to EPA under authority of Section 114 of the Clean Air Act (CAA). Section 114 authority was also used to review Regional source files, but no Section 114 information requests were sent to any subject sources. 5 All responses were given voluntarily and cleared of any confidential material prior to inclusion in this'report. The results presented here will form the basis for preliminary recommendations for revision of the current regulation and will identify additional research that is needed to support these revisions. 2.2.2 Review Study Methods The following methods were used to conduct the review study. A literature review was conducted using published and unpublished information found in trade journals, EPA studies, EPA contractorconducted studies, other governmental agency studies, and other pertinent references. a EPA Regional personnel involved in enforcement and surveillance of the VC emitting industries from Regions I, II, III, IV, V, VI, and IX were consulted. Discussions were conducted during regional office visits, by telephone, and by mail. Representative plants in each region were visited (three EDC/VC plants and eight PVC plants). An effort was made to include old and new plants of each type as well as plants using the various process types. Meetings were held with industrial representatives to discuss control technologies currently used by their plants. Discussions were also held with representatives of two of the industry's trade organizations - the Society for the Plastics Industry (SPI) and the Chemical Manufacturers Association (CMA). The EPA OAQPS and the Division of Stationary Source Enforcement (DSSE) were also consulted. Appendix B lists the locations, dates, and purposes of the meetings held with industry and EPA personnel. 2.3 THE VC EMITTING INDUSTRY 2.3.1 Current Number And Geographical Distribution There are 40 operating PVC plants and 18 operating EDC/VC plants (1 producing EDC only) in 18 states and 7 EPA regions. Table 2-3 shows the 2-9 GENC 016473 Table 2-3. GEOGRAPHIC DISTRIBUTION OF OPERATING VINYL CHLORIDE-EMITTING PLANTS Region State I Massachusetts II New Jersey New York III Delaware Maryland Pennsylvania West Virginia IV Florida Georgi a Kentucky Mi ssissippi V Illinois Mi chigan Ohio VI Louisi ana Oklahoma Texas IX Cal ifornia Total Plants No. of PVC plants 3 1 2 1 1 1 1 1 2 1 2 1 2 5 2 4 3 40 No. of EDC/VC plants 0 0 0 0 0 0 0 0 0 1 0 0 0 0 11 0 5 1 18 2-10 GENC 01^474 distribution of these plants throughout the United States. (See Appendix C for identification of these plants). The number of EDC/VC plants has increased from 15 prior to promulgation of the regulation to 17 plants currently in operation. Two of the original plants discontinued operation and four new plants began operation. These new plants are identified in ' Appendix C. During this period the approximate VC nameplate production capacity increased from 3.1 teragrams (6820 million pounds) per year in 1974 to 3.7 teragrams (8200 million pounds) per year in 1980 (EPA, 1975; Chemical Week, 1980a). The number of operating PVC plants has remained consistent since -p-mromulgation of the regiu. lia' tion. Of the 41 original PVC plants,{4 plan1 tsVw- --have discontinued operation* and 3 new plants have begun operation. t*wi--- -- --- . - \ip-jWV-4' J These newer larger plants, along with extensive expansions at several other existing plants, have resulted in an approximate increase in PVC nameplate production capacity from 2.6 teragrams (5739 million pounds) per year in 1975 to 3.4 teragrams (7600 million pounds) per year in 1980 (EPA, 1975; Chemical Week, 1980b). 2.3.2 Influence of the Standard on Industry Promulgation of the regulation was followed by significant changes in the VC industry. These changes included plant and equipment moderni zation, process modifications, and redirection of some research and development resources from the product itself to the areas of environ mental control. Economically, these changes were felt most acutely by the older PVC plants that had to retrofit their processes with new controls. Several EDC/VC and PVC plants were still in the design phase during development of the regulation and the engineering was altered to accommodate the new requirements. The cost (of compliance) to the VC industry for a 10-year period (1977-1986) is estimated to be[$765.7 million (1977 dollars).-* This includes investments, capital, operating and maintenance costs for new and existing plants (EPA, 1979). * A survey of 14 PVC producers indicated a 10-to-12 percent average / ^ B~ |<a - -- _ ; loss in production capacity as a result of compliance requirements ' (Chemical Week,' 1979). The reasons for lost capacity were due mainly to the time needed to clean reactors and purge the different systems in an 2-11 GENC 016475 effort to reduce VC emissions prior to opening to the atmosphere. The PVC process may also need to be operated at a slower rate to strip residual VC (RVC) from PVC resins in order to reduce emissions. The production loss varies with the type of resin produced and stripping technology used. EDC/VC plants have expended capital to comply with the regulations, mainly for add-on control equipment and modifications to processes'. Furthermore, "state agencies regulating hydrocarbon emissions are stimu lating new technology for emission reduction. The main emphasis has been in changing from air processes to oxygen processes in` the EDC process. This makes control by combustion, for example, much easier. At the time of this review study, most of the EDC/VC and PVC plants have completed many of the modifications discussed above and are channeling their research and development resources back to product development. Relatively few plants have ceased production during the last 4 years. No EDC/VC plants have shut down; seven PVC plants have closed (four on a temporary basis). Construction of new or modified sources is currently underway in many regions. 2.3.3 Industrial Trends From the standpoint of process and control, there'is a significant ; trend in the industry towards automation and computerization. Among plants surveyed during the review study, processors utilizing these ^ types of advanced systems have attained a high level of compliance. There is a definite trend toward the use of larger reactors in the PVC industry. Economic and emission control advantages of these larger systems are discussed in Section 4.0. The tendency to minimize the number'of PVC resin grades has also been noted. The "grocery store" processor, with many small reactors producing multiple grades of resin, is leaning towards the processing of fewer grades. One reason for this change is the difficulty and amount of time required for stripping RVC from certain specialty resins (as mentioned above). Reduction of energy consumption at EDC/VC plants is also being achieved through various process modifications. Steam consumption has been greatly reduced by Stauffer Chemical who uses the EDC reactor heat 2-12 GENC 016476 in their purification reboilers and B.F. Goodrich who uses the heat of reaction for purification (McPherson, 1979). Many companies are also devoting more effort to make the "cracking" of EDC to form VC more efficient and eliminate unwanted byproducts originating from side reactions during the cracking. These trends will be discussed in more detail in later sections. The VC industry is currently experiencing a sales decline due to the recession that occurred in early 1980. This decline is due mainly to the depressed construction industry, a major consumer of PVC pipe. (Forty percent of the PVC used in the United States goes into the manufacture of pipe.) Because 96 percent of the VC produced by EDC/VC plants is used in the PVC industry, a domino effect occurs. Currently, EDC/VC producers are running their plants at an average 86 percent of the first-quarter 1980 nameplate capacity while PVC producers are running their plants at an average 65 percent of the first-quarter 1980 nameplate capacity (Chemical and Engineering News, 1980a; 1980b). 2.5 REFERENCES FOR CHAPTER 2 Chemical and Engineering News. 1980a. "Vinyl Chloride." July 7, 1980, p. 9. Chemical and Engineering News. 1980b. "Polyvinyl Chloride." October 6, 1980, p. 13. Chemical Week. 1979. "At PVC Plants, Compliance Curbs Capacity." March 28, 1979, p. 36. Chemical Week. 1980a. "PVC Growth Plans Lead to Big VCM Expansions." February 13, 1980, p. 26. Chemical Week. 1980b. "PVC: Big Plans Match Growth Forecasts." January 16, 1980, p. 33. Environmental Protection Agency. 1975. Standard Support and Environmental Impact Statement: Emission Standard for Vinyl Chloride, EPA-450/2-75-009. October 1975. Environmental Protection Agency. 1979. Report to Congress. Document #96-38. "Cost of Clean Air and Clean Water," Vinyl Chloride Air Pollution Control Costs. December 1979, p. 56. McPherson, R. W.; Starks, C. M.; and Fryar, G. I. 1979. "Vinyl Chloride Monomer . . . What You Should Know," Hydrocarbon Processing. March 1979, p. 36. 2-14 3.0 PROCESS DESCRIPTION 3.1INTRODUCTION This chapter describes processes used to produce vinyl chloride (VC) and to polymerize VC into polyvinyl chloride (PVC) resins. The most common method used for production of VC is dehydrochlorination (cracking) of ethylene dichloride (EDC). Currently, there are four polymerization processes being used to produce polyvinyl chloride resins from VC. Approximately 87 percent of the EDC, or 1,2 dichloroethane, produced in the United States is used to produce VC. Current commercial processes that manufacture EDC for production of VC use a combination of ethylene, chlorine, and oxygen (usually in the form of air) for feedstocks. Hydrogen chloride (HC1) recycled from the cracking of EDC is also used as feedstock. In general, the production of EDC is an intermediate step in a combination of processes known as the "balanced process" for the production of VC. Figure 3-1 shows a simplified process flow diagram of an EDC/VC plant using the "balanced process." Table 3-1 summarizes the potential emission points and regulation requirements for control of each point in each process step shown in Figure 3-1. The requirements of the regulation will be discussed in more detail as the different process steps are described. Polyvinyl chloride is one of the most versatile thermoplastics manufactured in the United States today. PVC resins are noted for their excellent chemical and physical properties. They are easy to process, cost relatively little to make, are self-extinguishing (when ignition source is removed) and can be compounded with other resins (Shreve, 1977, p. 589). Products fabricated from PVC resins can be either flexible or rigid. GENc 0164?? cn/itiijtL ei,_ CHjCNj-- tlmt OtlwrlMllM IffM fnif )Urfi Httry Mi llmfc i i ISCj- i \ i 1 Cm4t llfM Ml V NMijr Enii V 1 CO I ro CD r> o:* o L___J Ntfia Utm <7) tlmt Ot1<rrtl 0 lit Hrtf1c*lt* 0 *U Nmi* HltMUr Jtrtppir (?) kfcklvrtncUM 0 fC Cr*ht*f 0 K ItNfuf Mi Itmft Figure 3-1 EDC/VC "Balanced Process flow diagram. I-----1 L__ J .-il S3B LS J Table 3-1. POINT SOURCE EMISSIONS "BALANCED PROCESS" EDC/VC PLANTS Process step Potential emission points Regulation requirements 1 Direct chlorination Product condenser Not regulated 2 EDC purification EDC crude storage, light ends column condenser, light ends storage tank, heavy ends column condenser, heavy ends storage tank All emission points are required to be controlled to._<_ 10 ppm 3 "Inprocess" wastewater stripper Wastewater storage tank Wastewater stripper column VC removed from inprocess water is to be ducted to a control system from which concentration of VC in exhaust gas does not exceed 10 ppm i 4 Oxychlorination Water wash column Oxychlorination process vent Separator tank Emissions from reactor are not to exceed 0.2 g VC/kg of the 100 percent EDC product 5 VC cracking and purification EDC quench column HC1 column vent VC column condensor Concentration in all exhaust gases must not exceed 10 ppm 6 VC loading and storage Loading lines VC storage tanks Emissions from loading lines must be reduced so that upon opening of line to the atmosphere emissions do not exceed 0.0038m3 of VC at STP. VC removed from lines to meet this criteria must be controlled to <_ io ppm upon exhaust to the atmosphere. Concentration of VC in exhaust gases discharged to the atmosphere from storage tanks must not exceed 10 ppm. GENC 0 1 4 4 03 g Consumption of PVC resins totalled 2.6 teragrams (5.6 x 10 pounds) in the United States in 1978, and consumption of PVC thermoplastics was second only to consumption of low density polyethylene (Hatch, 1979, p. 176). 3.2 PRODUCTION OF ETHYLENE DICHLORIDE (EDC) In the "balanced process" (see Figure 3-1), ethylene dichloride is produced by two different methods 1) direct chlorination, and 2) oxychlorination. EDC is first produced by direct chlorination of ethylene (CH2CH2) with chlorine (Cl2)- Direct chlorination may be simply expressed as cich2ch2 + 2 ca^alyst-> ch2cich2ci. This process varies with the technology employed but usually yields a high conversion rate; approximately 95 percent to 98 percent of the ethylene reacts and 99 percent of the chlorine reacts (Nass, 1977, p. 20). Byproducts and unreacted chlorine and ethylene are removed during purification by scrubbing and distillation. The purified and dried EDC is then "cracked" (dehydrochlorination by pyrolysis) to produce VC. Approximately one mole of hydrogen chloride (HC1) is produced for every mole of VC formed during the cracking process. The HC1 byproduct from cracking is recycled as a feedstock to form EDC by a second process, oxychlorination. The reaction that occurs in the oxychlorination reactor may be simply expressed as CH2CH2 + 2HC1 + ho2 ----- --1yst -* ch2cich2ci + h2o. An oxychlorination reactor uses the available HC1 formed from the EDC cracking process in the "balanced process." EDC produced in the oxy chlorination reactor is usually routed through the same purification system used for the direct chlorination reactor. There are currently a number of licensed process variations for EDC/VC production. The most common licensors in the United States are B. F. Goodrich, Stauffer Chemical, Pittsburgh Plate Glass, Dow Chemical, and Ethyl. In some instances, combinations of these process variations are used. Section 61.60(a)(1) of the regulation states that EDC production by reaction of 02 and HC1 with CH2CH2 (oxychlorination) is subject to 3-4 GENC 016482 requirements in the regulation of EDC production (see Table 2-1). The direct chlorination step (or any other method) used to produce EDC is not required to meet stipulations of the regulation. Regulation of the oxychlorination step is needed because HC1 generated in the cracking furnace, which is recycled for oxychlorination feedstock, contains VC, which is also formed as a byproduct in the oxychlorination reactor. Therefore, VC may be released from the oxychlorination vent and may contaminate the EDC product stream. 3.2.1 Direct Chlorination of Ethylene Typidally, direct chlorination of ethylene is carried out in a liquid-phase reactor. Although vapor-phase reactors are available, better temperature control is realized with the liquid EDC medium, and dilution gas for safety reasons is not necessary. In the liquid-phase reactor, the reactants (ethylene and chlorine) are bubbled up through liquid EDC. Mechanical agitation may also be used to promote solubility (Nass, 1977). Operating temperatures normally range from 50C (120F) to 70C (160F) and pressures range from 400 kilopascals to 500 kilopascals (4 atmospheres to 5 atmospheres) (Nass, 1977, p. 20). Metallic chlorides may be used as catalysts for this free radical process. Iron chlorides seem to be most prevalant in commercial processes, although aluminum, copper, and antimony chlorides are also used (Milby, 1978, p. 16). The EDC product from direct chlorination may be contaminated with 1.1.2 trichlorethane and the metal chloride catalyst, both of which must be removed before the EDC can be cracked to VC. Although direct chlori nation has not been cited as being a source of VC emissions and is not subject to the regulation, the operations could possibly become contami nated by VC. If ethylene separated from vent or product gases in the oxychlorination reactor is recycled to the direct chlorination process (McPherson, 1979, p. 78), the ethylene could be contaminated by VC formed as a byproduct in the oxychlorination reactor. 3.2.2 Oxychlorination of Ethylene An important development in the "balanced process" was the start-up of the first large scale oxychlorination unit in 1958. The oxychlori nation process (see Figure 3-1) allows the production of VC from two 3-5 GENC 016483 chemical feedstocks, chlorine and ethylene, without coproduct formation (McPherson, 1979). The process can take place in two types of reactors, a fixed-bed reactor or fluidized-bed reactor. Fluidized-bed reactors are capable of better temperature control because of the excellent intermixing of reactants and catalyst. In the fixed-bed reactor, the catalyst is packed in tubes; however, hot spots can form in the reactor tubes if the catalyst migrates (usually in the direction of process flow). Heavy concentrations of catalyst in one or more areas accelerates the reaction rate and the subsequent heating may cause an increase in byproduct formation. Temperature control in the oxychlorination reactor is important whether fixed-bed or fluidized-bed configurations are used. If tempera tures exceed 325C (600F), an increase in byproducts such as VC is noted along with the burning of ethylene to form carbon monoxide (CO) and carbon dioxide (CO2). An increased deactivation of the catalyst may also occur at the higher temperatures (McPherson, 1979, p. 78). All oxychlorination reactors use copper chloride for the reaction catalyst (Albright, 1967, p. 219), Sodium or potassium chloride can be mixed with the copper chloride to lower the melting temperature of the salt mixture and to reduce the vapor pressure of the copper chloride, hence increasing catalyst life. Catalyst for both types of reactors is supported on a solid porous material such as alumina or silica. A new oxychlorination process has been developed by the M. W. Kellogg Company. In this process an aqueous solution of copper chlorides is used as catalyst and reactants are bubbled up through the catalyst. Some of the advantages of this process are high product yield, excellent temperature control, the use of aqueous solutions of HC1 as feedstock, and simultaneous chlorination as well as oxychlorination (Nass, 1977, p. 23). There are no known commercial producers currently using the M. W. Kellogg process of oxychlorination. Oxychlorination reactors incorporate large rupture discs as a safety measure. The rupture disc would allow pressure to escape to the atmosphere in the event of an explosion. Over-pressure due to an accelerated reaction is not an item of concern (DiBernardi, 1980). Thus, there are no specific requirements in the regulation for rupture discs on oxychlorination reactors. 3-6 GENC 016434 The oxychlorination process vent is subject to the current regulation. Emissions from oxychlorination reactors must not exceed 0.2 grams of VC per kilogram of the 100 percent EDC product. This emission limit can be met by an add-on control device such as an incinerator or by process modifications. The use of oxygen instead of air in the formation of EDC by this process greatly affects the quantity of inerts (e.g., nitrogen) that are vented from the oxychlorination process. When air is used as a feedstock material, emissions from the oxychlorination vent are large in volume and low in hydrocarbon content. Large amounts of inerts increase the work that must be done to condense the product stream in order to separate EDC and unreacted ethylene. Byproducts found in the oxychlorination product stream can include VC, vinylidene chloride, ^thyl chloride,^1,1-dichloroethane, 1,2-dichloroethylene, "'trichloroethylene, chloroform, "carbon tetrachloride, methyl chloride, methylene chloride, chloral, and high boiling compounds (McPherson, 1979, p. 78). These byproducts must be removed prior to cracking the EDC to produce VC. Acetylene in the HC1 feedstock recycled from the cracking furnace is used to form 1,1,2 trichloroethylene. This byproduct is not easily removed in the EDC purification step and reduces product yield in the cracking furnace. One method of eliminating 1,1,2 trichloroethylene formation is hydrogenation of the HC1 feedstock. Hydrogenation is a catalytic reaction that eliminates the acetylene by converting it to ethylene. 3.2.3 Purification of Ethylene Pi chloride In order to prevent fouling of the dehydrochlorination reactor (cracking furnace), the EDC cracked to form VC must be highly pure, at least 99.5 percent. Also, any moisture in the stream must be removed to prevent corrosion in equipment from the HC1 generated during the cracking process. Typically, process vent streams purified are those from direct chlorination of ethylene, oxychlorination of ethylene, and EDC recovered from the dehydrochlorination process (see Figure 3-1). Because the purification process is used to purify EDC recycled from the oxychlorination 3-7 6ENC 0U4S5 unit and the cracking furnace, VC contamination is possible. For this reason Section 61.62 requires that all exhaust gases discharged to the atmosphere from the purification process not exceed 10 ppm VC unless the equipment is out of service or opened. Prior to opening any equipment, Section 61.65(b)(6) of the regulation requires that the quantity of VC in the purification equipment be reduced to 2.0 percent by volume or 0.0950 cubic meters (25 gallons), whichever is larger, at standard temperature and pressure (STP). The first step of EDC purification is usually a water quench followed by caustic scrubbing. This step removes catalyst and unreacted chlorine from the direct chlorination process stream. Water that comes in contact with VC is termed "inprocess wastewater." This water is returned to the process or must meet "inprocess wastewater" requirements of the regulation before being discharged to other wastewater treatment facilities. The current regulation prohibits the discharge of "inprocess wastewater" containing more than 10 ppm VC. Water and low-boiling impurities such as VC, ethyl chloride, vinylidene chloride, chloroform, and methyl chloride, generated in the oxychlorination reactor, are removed by a light ends distillation column. Pure, dry EDC is taken overhead from a second distillation column which removes compounds of a higher boiling temperature. Gases taken overhead from the light ends distillation column are stored in a recovery tank and may subsequently be sold. Tars from the heavy ends distillation column can be fractionated to recover soluble components and the remains incinerated to recover chlorine as HC1 (McPherson, 1979, p. 80). All vent gases from processes and storage vessels in VC service (defined in the regulation as containing 10 percent by volume or more VC) are required to be controlled to 10 ppm or less VC. Ethylene dichloride condensed from the dehydrochlorination process stream requires purification to remove byproducts formed during cracking. Special treatment is needed to remove chloroprene which can polymerize inside the light ends distillation column and trichloroethylene which forms an azeotrope (mixture with constant boiling point and distilling off in a fixed ratio) with EDC. Trichloroethylene may inhibit dehydro chlorination if allowed to accumulate in the EDC. These two byproducts 3-8 GENC 016 are usually removed by chlorination prior to distillation (McPherson, 1979, p. 79). A method for formation of HC1 from the light and heavy ends distillation column byproducts has been developed. The method employs catalytic oxidation of the byproducts separated by the purification columns with air and other added reactants. The HC1 formed by this method is used as feedstock for the oxychlorination unit. 3.3 PRODUCTION OF VINYL CHLORIDE All VC is currently produced in the United States jointly with EDC using the "balanced process" method; however, VC manufactured by any process is covered by the current regulation (see Table 2-1). VC concentrations in all exhaust gases from the formation (cracking) and purification of VC cannot exceed 10 ppm. 3.3.1 Formation of Vinyl Chloride by Dehydrochlorination of EDC Thermal dehydrochlorination, commonly known as cracking, is the separation of hydrogen and chlorine from 1,2-dichloroethane (EDC) yielding vinyl chloride (CH2=CHC1) and hydrogen chloride (HC1) at about a one to one molar ratio. The cracking of EDC may be simply expressed as ch2cich2ci--------------------> ch2chci + HC1. The non-catalytic method seems to be preferred over the catalytic method (Nass, 1977). The typical cracking furnace operates at pressures between 2 megapascals to 3 megapascals (20.0 atmospheres to 30.0 atmospheres) and at temperatures between 450C to 650C (840F to 1,200F) (Albright, 1967, p. 223). Operating the furnace at high pressures results in an increased yield, fewer byproducts, and allows easier separation of the VC product from unreacted EDC and byproducts. Conversion rates are normally kept between 50 percent and 60 percent in order to minimize byproduct formation. Research is continuing in the development of cracking promoters and inhibitors of side reactions in pyrolysis chemistry. Considerable energy and cost savings could be achieved through increased conversion levels without concurrent losses of EDC to undesirable side reactions (McPherson, 1979, p. 87). The process stream from the cracking furnace is condensed to separate the VC product and unreacted EDC which is recycled back to the process 3-9 <*NC 0164S? (see Figure 3-1). Some systems quench the process stream with crude EDC to reduce formation of byproducts and to partially condense EDC from the product. Byproducts formed in the furnace reactor tubes in addition to HC1 are tars, carbon, chloroprene, butadiene, and trichloroethane. Carbon and tars tend to foul reactor tubes in the furnace so the tubes need to be opened and cleaned periodically. 3.3.2 Purification of Vinyl Chloride The VC in the product stream from the cracking furnace must be separated'from byproducts formed during cracking and unreacted EDC (see Figure 3-1). The first step in purification of the product stream as mentioned above is normally a quench of the hot effluent with liquid EDC or the condensation product from the cracking furnace. The product stream exits the quench column and is condensed and fed to a distillation column. In this column, HC1 is separated from the product stream. Acetylene and some ethylene byproducts will also come off with the HC1, and HC1 treatment by hydrogenation may be necessary if the HC1 is to be used as a feedstock for the oxychlorination step. The EDC, VC, and remaining byproducts are then fed to a second distillation column where VC is distilled. Methyl chloride and butadiene will come off with VC, depending on the efficiency of the fractional distillation system. Methyl chloride formation in the cracking furnace can be reduced by addition of chlorine or anhydrous HC1, or by selective hydrogenation (Nass, 1977). The remaining crude EDC is returned to the EDC purification step of the process. The VC taken overhead from the second column is stored in pressurized vessels for eventual shipment to PVC plants or other facilities using VC. In instances where the PVC plant is very close to the VC producer, VC can be delivered by pipeline. 3.3.3 Emissions for Typical EDC/VC Plants New emission data has not been obtained from EDC/VC plants since the regulation was promulgated. Prior to promulgation, EPA estimated VC emissions from the 17 existing EDC/VC plants to be approximately 11 gigagrams (24.2 million pounds) or approximately 15 percent of the nationwide VC emissions (EPA, 1975). The original study done in support of the existing regulation identified uncontrolled VC emissions from four areas 3-10 GENC 016488 within an EDC/VC plant producing 316 gigagrams per year (700 million pounds per year). These four areas were the EDC purification light ends vents, the VC finishing column, the oxychlorination vent and fugitive emissions sources (see Figure 3-1). Total emissions from this plant were calculated to be approximately 1.4 gigagrams (3.1 million pounds) of VC per year. 3.4 PRODUCTION OF POLYVINYL CHLORIDE Four polymerization processes are being used currently to manufacture PVC resins: suspension, dispersion (emulsion), bulk (mass), and solution. Figure 3-2 shows a generic flow diagram for PVC production by the suspension and dispersion processes. The potential emission points for each of the process steps shown in Figure 3-2 are listed in Table 3-2 along with a summary of regulation requirements for control of emissions. Many companies have licensed some phase of their particular process such as stripping technology or clean reactor technology which will be discussed in subsequent sections. The licensing of these process phases has evolved from the intensive research work done to reduce residual VC (RVC) levels in finished resins, limit worker exposure, and reduce emissions to the atmosphere. After polymerization, VC may be present in any of three forms in the polymerization reactor depending on the particular process. VC will always be present separately as a gas or liquid in the reactor and will also be trapped within the newly formed polymer resin. In the suspension, dispersion, and solution processes, VC will also be trapped in the process water. The resins produced by a process are categorized as to the "type" of resin. Those processes responsible for more than one variation of their type of resin are further subdivided into "grades." The resin grade is developed to allow compounding and fabrication to yield the desired product. Figure 3-3 shows the resin types, the compounds, and the various fabrication processes. PVC fabrication will be discussed in more detail in Section 6.1.1. 3-11 genc 01448? id i-1 H^O from Double Mechanical fs Seals on all Pumps, Compressors and Agitators H^O from Reactor Evacuating Prior to Opening Knock-out Pots PVC ind Storage Shipping l *-rj1" Cyclone or Screening _____________ ii!i Recovered VC i i i rt (nock-out Pot i i i L VC Recovery System H;0 to Vent to Hater Stripper Control Device Process Steps (D VC Unloading and Storage (D Mixing, Weighing and Holding (D Polymerization Reaction (D Resin Stripping (D Recovery System i i (D Blending i CD Drying ! PVC Storage "In Process" Wastewater Stripping GENC 016490 Figure 3-2. Suspension and dispersion processes flow diagram. Table 3-2. POINT SOURCE EMISSIONS TYPICAL OF SUSPENSION AND DISPERSION PVC PLANTS Process step 1 VC unloading and storage 2 Mixing, weighing and holding tanks before stripping operation 3 Polymerization 4 Stripping Potential emission points Loading lines, VC storage tank Mixing, weighing and holding tank vents Polymerization reactor opening loss (ROL) Polymerization reactor relief valve discharges Polymerization reactor rupture disc discharges Stripping vessel vent {continued} Regulation requirements Emissions from loading lines must be reduced so that upon opening of line to the atmosphere emissions do not exceed 0.0038nr of VC at STP. VC removed from lines to meet this criteria must be controlled to _< 10 ppm upon exhaust to the atmosphere. Concentration of exhaust gases discharged to the atmosphere from storage tanks must not exceed 10 ppm. Concentration of VC exhaust gases discharged to the atmosphere must not ; exceed 10"ppm. P ' ROL from each reactor is not to exceed 0.02 g VC/kg PVC products. Mo discharge to the atmosphere except for an^emergencyjrelief discharge Concentration of VC exhaust gases discharged to the atmosphere jnust not "gexceed^ldlppmPCNot required to be\ pre ported. __________ _ Concentration of VC exhaust gases discharged to the atmosphere must not ]exceed'r10'ppm$ I Table 3-2. Concluded Process step 5 Monomer recovery system 6 Blending (mixing, weighing and holding tanks after stripping operation) 7 Drying, sizing, screening of dewatered resin 8 PVC loading and storage 9 "In Process" wastewater stripper Potential emission points Recovery system exhaust vent knock-out pot Slurry blend tanks and holding tank vents Regulation requirements Concentration of VC exhaust gases discharged to the atmosphere must not exceed'10,ppm. J Not regulated.* Centrifuge vents, dryer vent stacks, storage silos, baghouse vents, screening operation vents Storage silos Waste-water storage tank waste-water stripper column Not regulated.* Not regulated.* VC removed from in process water is to be ducted to a control system from which concentration of VC in exhaust gas does not exceed 10 ppm. *If:-a-PVC plant is using stripping to control VC emissions, emission sources beyond the stripper * 1 are not regulated. r 3-14 GENC 0 1 6 4 ? to 1 Polymerization ProcesB_________ Resin Type Compound ------ r-------- Fabrication Process GENC 0 1 6 4 9 3 *Latices are usually sold directly to the consumer rather than being used in later fabrication processes. Figure 3-3. Polyvinyl chloride resins, PVC compounds and PVC fabrication processes. A plant manufacturing polymers containing any amount of polymerized VC is subject to the current regulation. The requirements of the regulation listed in Table 3-2 will be discussed in more detail below. 3.4.1 Free Radical Polymerization Polymerization is the chemistry of combining simple molecules (monomers) into long chains of repeating molecular units (polymers). There are two types of polymerization - addition and condensation. PVC resins are commerically produced by addition polymerization and, in general, polymerization is induced by the use of free radical initiators (Nass, 1977, p. 34). Thermal decomposition of the initiator, which is combined with the VC and other constituents in the polymerization reactor, yields free radical molecules having one unshared electron. This free radical reacts with additional VC molecules by removing an electron from the VC double covalent bond and sharing an electron with the free radical. The remaining unshared electron moves along the chain becoming the new radical bond site. Another VC molecule may then become a part of the polymer chain by reacting with the polymer radical, H Cl R* + C = C R - C - C- HH HH Radical + VC molecule ----------------------- Polymer radical Radicals are transferred from the polymer by a reaction with a VC monomer to yield a polymer and free radical. 'The transfer reaction, v,>r -- is*"iriintir: .-^rrr'f.-r i increases with increasing temperatures. This characteristic of theJ reaction kinetics allows a` desired molecular weight to be obtained by i controlling reaction temperature (Cameron, 1979, p. 39). Radicals are also transferred by reacting with hydrogen atoms obtained from VC or solvent in the reactor. Another way of transferring radical sites from the growing polymer chain is by using chain transfer agents (CTA). These CTA's are used to regulate the molecular weight of the polymer to a desired level. In all transfer reactions, the radical is released and able to initiate another polymerization reaction. The reaction is terminated when two polymers with radical sites are combined. Polymerization rate is controlled by choice of the appropriate initiator. 3-16 GENC 016494 The polymerization reaction is exothermic and the reaction rate must not be allowed to accelerate to the extent that the heat of reac tion cannot be removed by reactor cooling devices. Reaction rates tend to increase as the temperature increases. If the heat of reaction is not removed, a runaway reaction can develop and the rapid increase in temperature will increase the pressure inside the reactor vessel beyond safe limits. The monomer structure of VC is capable of several variations in chain structure, but free radical polymerizations generally produce atactic structures (random orientation of monomer unit in chain struc ture) (Nass, 1977, p. 46). Tendency towards syndiotactic structure (regular alternating orientation of the monomer unit in the chain struc ture) and crystallinity are both increased by reduced polymerization temperatures. 3.4.2 Unloading of VC at PVC Plant Sites Under the current regulation VC emissions from loading and unloading lines, which are opened to the atmosphere, must be reduced in the opened lines ,tcT0."0038 cubic meters (0.13 cubic feet) oPTless at standard fll'i >m ------------------------------------------------------ 'j' temperature and pressure. Also, the VC removed from loading and unloading lines in order to meet this requirement of the regulation must be vented to a control system from which the concentration of VC in the exhaust gas does not exceed 10 ppm (see Table 3-2). When PVC plants are not located close enough to receive VC by pipeline, it is shipped by railcar, tank car, barge or marine vessels. Some companies add about 100 ppm of phenol to inhibit polymerization during transport, but this practice is no longer common. The VC is normally unloaded at PVC plants by displacement. This is accomplished by pumping vapor from storage tanks into the transfer vessel which displaces the liquid VC from the transport vessel through the unloading line. When the liquid VC is displaced, the compressor line is reversed and used as suction to evacuate remaining liquid as vapor. The remaining liquid VC is allowed to boil during tank car evacuation for its removal as a gas. This vaporization of the VC usually takes 20 to 30 minutes (Mukerji, 1977, p. 155). Lines between tank car compressor and storage tank can be switched without disconnection by 3-17 GENC 016495 incorporating a 4-way valve in the pumping lines. Instrumentation of the system usually incorporates a turbine meter, a flow totalizer to measure the VC flowrate and quantities unloaded, and storage tank level indicators. Unloading and transfer lines may be purged with nitrogen to reduce VC to the required level before disconnection of the unloading lines. In some cases, the portion of the unloading line which is opened to the atmosphere has been reduced to diminish the amount of VC that will escape to,levels below those required by the regulation. 3.4.3 Mixing, Weighing and Holding Vessels Storage spheres, storage tanks, weigh tanks, gasholders, wastewater storage tanks, knockout pots, and surge tanks are representative of vessels covered under Section 61.64(c) of the regulation for mixing, weighing, and holding vessel (see Table 2-1). These vessels are used to hold liquid or gaseous VC and PVC slurry during various stages of the PVC process. Many of the vessels were open to the atmosphere prior to promulgation of the VC regulations, but have since been enclosed and are usually ducted to the recovery system and/or the primary control device. The number and types of vessels will vary from plant to plant depending on the size and type of process used to produce the PVC resin. Section 61:64(c) of the regulation requires the concentration of VC in all exhaust gases discharged to the atmosphere from mixing, weighing, and holding containers "in-VC-service" not to exceed 10 ppm (see Table 3-2). Any piece of equipment that contains or contacts a fluid that is 10 percent by weight, VC, or gas that is at least 10 percent by volume, is considered "in-VC-service." There are no requirements in the regulation for mixing, weighing or holding containers used in the process after the stripping of the PVC resin provided resin stripping is used. 3.4.4 Suspension Polymerization The suspension process for producing PVC resins (see Figure 3-3) is characterized by the formation of polymers in droplets of the liquid VC (or other co-monomers) suspended in water. These droplets are formed by agitation and the use of protective colloids or suspending agents. Protective colloids commonly used are water-soluble polymers such as modified cellulose or partially hydrolyzed polyvinyl acetate. The 3-18 GENC 016496 process is started by evacuating the polymerization reactor to remove oxygen and other contaminants that may inhibit the reaction initiator (see Section 3.3.1). Water and VC may be may be simultaneously added to the reactor with the protective colloids, or they may be added separately (water and colloids first, followed by the liquid monomer). The water used is deionized and deaerated in order not to inhibit free radical initiator formation. Water and other ingredients charged to the reactor must be carefully measured prior to charging because a level indicator for reactors has not been commercially developed. In some cases the reactor is1 on a scale and the amount of material charged is weighed in the reactor. More often, a separate weigh tank is used to measure materials charged to the reactor. A flow meter can be used to record the amount of water added. Reactor operators manually charge additives that are used in small proportions. The initiator is usually the last ingredient charged to the reactor. Initiators commonly used in the suspension process are peresters, peroxycarbonates, peroxides, or azo compounds. The initiators are soluble in VC and allow formation of PVC in the monomer droplets. After all materials are in the reactor, the batch is brought up to the reaction temperature by passing steam through the reactor jackets which allows free radical initiators to be formed. Reaction temper atures are varied in order to produce a resin grade of a particular molecular weight. Once polymerization is initiated, the reaction becomes exothermic and cooling water must be circulated through the reactor jacket to remove heat of reaction. In some instances reflux condensers have also been used to control reaction temperature and remove excess heat. After approximately 6 hours in the reactor, the batch temperature and pressure drop. This signifies that nearly all the VC has reacted (75 percent to 90 percent of the VC usually reacts) and the remaining or residual VC (RVC), which is in a liquid or gaseous state or trapped in ' the resin particles, must be stripped. This RVC is usually stripped with steam under vacuum. The suspension process yields a particle size distribution of a much wider range than the other polymerization pro cesses. Particle size may range from 90 micrometers to 130 micrometers (0.0035 inches to 0.0050 inches) with low to medium molecular weights. 3-19 GENC 016497 The regulation requires that RVC levels for suspension resins not exceed 400 ppm of the PVC product. PVC resin, unreacted VC (in the water, in the headspace, and trapped in the resin) and water are the constituents remaining in the reactor after polymerization. This polymer slurry may be steam-stripped of RVC batchwise in the reactor or in a separate vessel. B. F. Goodrich has developed a continuous stripping operation which strips the resin with steam running countercurrent to the PVC slurry. Most plants strip batchwise with steam in separate vessels or in the continuous stripping column. In all cases, water accompanying the PVC product is stripped with the slurry and then removed by centrifugation. Batch stripping procedures use temperatures up to 870C (190F) and a vacuum of 91 kilopascals (27 inches of mercury) or more (Nass, 1977, p. 83). Heat is applied by steam injected directly into the batch or by steam passed through the reactor wall jackets. If stripping is used to meet RVC levels, any downstream processes or equipment (e.g., water treatment, vents from mixing tanks, centrifuges, dryers, etc.) are not required to meet any other requirements of the regulation. After stripping, the batch is transferred to blend tanks which mix the batch with other batches to ensure product uniformity. The mixed batches are then fed to a continuous centrifuging operation that separates the polymer from the water in the slurry. Both mixing tanks and centri fuges are vented to the atmosphere if stripping is utilized. The water from centrifuging is not required to be stripped of VC because most of the VC is removed during resin stripping. Therefore, the centrifuge water is recycled back to the process or discharged to the plant's wastewater treatment system. The wet cake from centrifuging is conveyed to a rotary dryer for further removal of the remaining (usually 25 percent) moisture (Nass, 1977, p. 83). Most of the RVC not removed during stripping will be released during the drying operation. Counter-current air temperatures , in the dryer range from 65C to 100C (150F to 210F). Drying time is generally short, but large volumes of air containing RVC are released. After drying, the resin may be screened to remove agglomerates formed in the dryer. The resin is then bagged or stored in silos for bulk shipment by trucks or rail car. 3-20 GENIC 016498 Uniformity of suspension resin batches is dependent on control of variables such as: impurities (noncondensable gases present in the reactor prior to polymerization or impurities in raw materials charged to the reactor), rate of temperature increase to reaction temperature, agitation speed and schedule (speed varies as the slurry becomes more dense), charging rate of raw materials, and 'temperatures of raw materials charged. In some of the newer PVC facilities these variables are closely monitored by levels of computer control. In the older plants many variables such as cooling water flow rates, pump operation, agitator motors, temperature, and pressure are monitored and controlled manually from the control room panel or at the reactor. More versatility is possible with computer assistance. A reactor linked to computerized control elements can produce different grades of resin by using programs designed to yield specific reactor conditions (e.g., agitation, temperature, amount, and type of materials charged) that will produce the desired resin product. Computers can*also monitor operating conditions and respond to emergencies ......... ui 'i -- i win mi | (such as high pressure in the reactor).* They can be programmed to take necessary action to bring an upset condition under control, e.g., choosing the best compensatory action. The use of large reactor systems has increased quality control and allowed incorporation of equipment and procedures that reduce VC emissions. Large reactors, approximately 8,000 gallons (30 cubic meters) and up, offer lower plant cost, improved product uniformity, and increased productivity. A more detailed discussion of reactors is included in Section 3.4.8. 3.4.5 Dispersion Polymerization Although dispersion (emulsion) resins are formed in reactors similar to those used for suspension type resins, reaction kinetics of polymer formation vary greatly. In general, dispersion resins are of a high density with small particle size. The process is initiated by charging the necessary ingredients (water, liquid, VC, emulsifiers, and a free 3-21 GENC 016499 radical initiator) to an evacuated reactor. Proportions of these ingredients and other minor additives will vary depending on the type of resin and the resin grade. Emulsifiers (soaps and surfactants) are used to disperse VC in the water phase. Soap micelles (i.e., colloidal aggregates that are formed above a critical emulsifier concentration) also contain small amounts of VC and are the site of polymer formation. Initiators used are water soluble (versus VC soluble initiators used in suspension polymerization) and penetrate soap micelles to begin polymerizatioii. Commonly used initiators are hydrogen peroxide, organic peroxides, and peroxydisulfates. Two general types of resin are produced by the dispersion process latex and dispersion. If the latex type resin is to be produced, only a small amount of VC and initiator are normally charged. The reactor is heated by injecting steam into the reactor jackets to initiate formation of free radicals. Strong agitation is used to disperse the monomer and other ingredients in the water medium. Once polymerization begins, the heat of reaction is removed by circulation of cooling water through the reactor jackets. Small amounts of initiator and monomer are then con tinuously added during polymerization until the correct particle size is attained. Conversion of VC to latex polymer is almost complete. Latex resin particle sizes range from 0.05 micrometers to 2 micrometers. The residual VC is removed from latex resins by completely reacting the free VC with additional catalyst (post-catalysis). This is done in the reactor or in a separate vessel. The regulation requires that latex resins be stripped to 400 ppm RVC. Latex resins are usually sold in solution, and drying or separation of resin from the polymer slurry is not necessary. The process for formation of dispersion type resins follows that described for latex resins except that more monomer is added to allow particles to grow to a larger size. In some cases the process may call for "seed" latex. The "seed" latex is formed by starting a batch in a separate reactor. Before this batch reaches 60 percent conversion it is transferred to other reactors with more VC and emulsifiers. This allows formation of larger emulsion particles which are used in plastisol and stir-in resins. These dispersion type resins range from 0.2 micrometers 3-22 GENC 016500 to 5 micrometers in particle size and are more sensitive to heat and mechanical agitation than suspension resins. Because the dispersion resins are more sensitive, stripping the resin may take longer because lower stripping temperatures are used. The regulation requires that dispersion type resins be stripped to 2,000 ppm RVC. After polymerization is complete, the dispersion resins are usually spray dried. Spray-dried resins contain emulsifiers that inhibit the absorption of plasticizers. In many applications where heat is applied during fabrication, this property is acceptable because the heat melts the resins which allows plasticizer absorption. In some fabrication processes (such as calendering), plasticizers must be absorbed prior to fabrication because temperatures used are not high enough to melt the resin. If removal of the soaps from the dispersion resin is desired, coagulation of emulsifiers with salts is usually performed. The salts (calcium or magnesium) are added to the polymer slurry and the emulsi fiers precipitate and are rinsed out with water. The resin and water are then separated and the resin is dried in a rotary or pan dryer (Erdman, 1980). 3.4.6 Bulk Polymerization Bulk (mass) polymer resins are produced by a two-stage polymerization process. A simple diagram of the bulk process is shown in Figure 3-4 and requirements of the regulation for bulk plants are summarized in Table 3-3. The first stage is the formation of a "seed" resin in a vertical pre-polymerization (Pre-Po) reactor. The "seed" resin is transferred to a horizontal post-polymerization (Po-Po) reactor in the second stage of the process. More VC is added to the Po-Po which allows the "seed" resin to grow in size to the finished resin product. The bulk process differs from the dispersion process in that no water is used in the reaction - the process is anhydrous. The bulk process used in this country is licensed by Rhone-Poulenc of France. A new version of the Rhone-Poulenc process utilizing a vertical Po-Po reactor is now available, but is not currently being used in the United States. The Pre-Po reactor is charged with liquid VC and enough initiator to carry the reaction to approximately seven to twelve percent conversion. Initiators used are those commonly used in the suspension process - an 3-23 GENC 016501 Recovered VC --rt j=t ~l l i Pre-Polymertsatlor i f Reactor i ! J L. "Inprocess" __ Wastewater |n HjO from Oouble Mechanical Water| Stripper 4 Post-Polymerisation Reactor H H^O to Vent to ter Stripper Control Device OJ ! 2) Seals on all taps. i\> Condensed Steam from Reactor Evacuating 13) Condensed Steam from Reflux Condenser Process Steps Polymer Stripping CD_____________________ r JH.O to I i Pond I I Blending Tanks VC Unloading and Storage Mixing. Weighing and Holding Pre-Polymerisation ri 1 Product Storage * t and Shipping 1 * J11 1 i Coarse I Fine t I I Screening Post-Polymerisation Recovery System Blending and Screening "In Process* Wastewater Stripping PVC Storage Figure 3-4. Bulk process flow diagram. i M 1 j___ GENC 016502 p-rwe gpnrn r r --i s--i em ___ i ___ i cm csa l--i ,___ i mi c-a Table -3-3. POINT SOURCE EMISSIONS TYPICAL OF BULK PVC PLANTS Process step 1 VC unloading and storage 2 Mixing, weighing and holding tanks before ^_ stripping operation cn 3 Pre-polymerization 4 Post-polymerization (stripping in reactor} Potential emission points Loading lines, VC storage tank Mixing, weighing and holding tank vents Polymerization reactor opening loss (ROL) Polymerization reactor relief valve discharges Polymerization reactor rupture disc discharges Polymerization reactor opening loss (ROL) Regulation requirements Emissions from loading lines must be reduced" so that upon opening of line to the atmosphere emissions do not exceed 0.0038tri of VC at STP. VC removed from lines to meet this criteria must be controlled to <_ 10 ppm upon exhaust to the atmosphere. Concentration of exhaust gases discharged to the atmosphere from storage tanks must not exceed 10 ppm. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. ROL from each reactor is not to exceed 0.02 g VC/kg PVC products. No discharge to the atmosphere except for an emergency relief discharge. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppia^Not required to bi" (^TeportedT ROL from each reactor is not to exceed 0.02 g VC/kg PVC products. (continued) SOS?to DN30 Process step 5 Monomer recovery system 6 Blending (mixing, U> weighing and holding ro a\ tanks after stripping operation) 7 "Inprocess" wastewater stripper 8 PVC loading and storage Table 3-3. Concluded Potential emission points Polymerization reactor relief valve discharges Polymerization reactor rupture disc discharges Recovery system exhaust vent knock-out pot Dry resin blend tanks and screening operation baghouse vents Wastewater storage tank Wastewater stripper column Storage silos Regulation requirements No discharge to the atmosphere except for an emergency relief discharge Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Not required to be reported. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Not regulated.* VC removed from in process water is to be ducted to a control system from which concentration of VC in exhaust gas does not exceed 10 ppm. Not regulated.* *lf a PVC plant is using stripping technology to control VC emissions, emission sources beyond the stripper are not regulated. GENC 0 1 6 5 0 4 oil soluble, free radical catalyst (Nass, 1977). The Pre-Po reactor is brought to an operating temperature of 40C to 70C (130F to 184F) by injecting steam into the reactor jackets. Strong agitation is used to form small particles of approximately 1 micrometer (Nass, 1977, p. 75). These small particles provide the "seed" that grows in size to form resin beads in the Po-Po reactor. The remaining liquid VC and seed resin are pumped to the Po-Po reactor at approximately 7 to 12 percent conversion. Total cycle time for the Pre-Po is 2 hours (Dubec, 1980) which is normally dne-fifth the duration of the Po-Po cycle. Thus, one Pre-Po reactor can be used to feed as many as six Po-Po reactors. After the "seed" polymer and remaining liquid VC is pumped from the Pre-Po to the Po-Po, more VC and initiator are added to the Po-Po for completion of the reaction. Small amounts of additional additives may be charged to the reactor for heat stability and particle size control (Dubec, 1980). The number of PVC particles in the Po-Po is determined by the amount of seed resin charged from the Pre-Po. The "seed" resin absorbs the VC and at about 20 percent conversion the batch becomes solid and powdery, thus the agitator in the Po-Po must be of rugged construction. When the conversion to PVC has reached 70 percent to 90 percent, steam is injected and a vacuum is pulled on the reactor to remove RVC from the resin particles. This represents the bulk stripping procedure that takes place in the Po-Po reactor. Exhaust gases from this stripping procedure are vented to recovery and the primary control so that VC emissions do not exceed 10 ppm. VC recovered from the reactors is sent back to the charge tank for reuse. The regulation requires bulk resins to be stripped to 400 ppm RVC as determined on a dry solids basis. In producing PVC resins by the bulk process, temperature is the major control variable for determining the molecular weight of the polymer. One disadvantage of the bulk process is that there is not a good medium for heat transfer from the polymer reaction to the reactor wall. Reflux condensers are used to help control the reactor vessel temperature. The condensers remove gas from the reactor headspace, condense the gas, and then return the liquid VC to the reactor. Heat from the gas is removed by cooling water in the reflux condensers. 3-27 GENC 01 <6505 The Po-Po is opened after every batch for cleaning. VC concentrations that are emitted when the Po-Po reactor is opened must be measured in order to meet the ROL standard. Polymer removed from the Po-Po reactor (already dry) is pneumatically conveyed to screens that remove oversize particles. Batches may be blended to improve product uniformity. Bulk resins range in size from 0.1 to 1.0 micrometers and exhibit excellent qualities for dry-blending compounds. The beads are of uniform size and porosity which allows uniform absorption of plasticizer. Also, because suspending agents and surfactants are not used, the finished resin has a higher purity and therefore better heat stability than suspension or dispersion resins (Nass, 1977, p. 75). 3.4.7Solution Polymerization PVC produced by the solution process typically consists of copolymers of VC and polyvinyl acetate. Only one company is producing resins by this process in the United States. A simple diagram of the solution process is shown in Figure 3-5 and requirements of the regulation are summarized in Table 3-4. The solution process is continuous and liquid VC, vinyl acetate, solvent, and initiator in solution are fed to a polymerization reactor which operates at low temperatures. Conversion to copolymer is approxi mately 60 percent to 70 percent. The copolymer resin solution is removed continuously and fed to a stripping operation that removes the solvent. VC and vinyl acetate are soluble in the solvent. The monomers and solvents are mixed and charged to the reactor separately fr^ni-the initiator solution.^/fhecopoTymer~Tormed is not soluble in the solvent but forms homogeneous solution. ^Typical solvents listed in the literature Tor"" this type of process are N-butane; aliphatic alcohols, ketones, esters, and hydrocarbons; aromatic hydrocarbons; and chlorinated hydrocarbons. No resin particles are formed by the solution process and RVC is stripped by distillation. Distillation takes place in a conventional trayed column. Acetone vapors are used to strip VC from the copolymer resin solution. The acetone vapors taken overhead from the distillation column are sent to a VC recovery system and then recycled back to the process. 3-28 GENC 016506 tool l__i i__j .__j SL33 Mixing Solution (Solvent Ii Polymerization Solvent to u@__i_____ Reactor Solvent Recovery '-zi--____ i VC unloading lo___________ OJ ro u> r i I i i1_____ i L "1 1 VC i iiRecovery -- ------------- jii Proceii Step* 0 VC Storage and Unloading @ Mixing PVC Storage and Shipment L<2._ 1 -! (3) Solution Polymerization Reactor 01 VC Recovery | Drying Coagulation Drying 11 PVC Storage and Shipment Figure 3-5. Solution process flow diagram. GENC 0 1 6 5 0 ? Table 3-4. POINT SOURCE EMISSIONS TYPICAL OF SOLUTION PROCESS PVC PLANTS Process step 1 VC unloading and storage 2 Mixing, weighing and OJ holding tanks before 1 OJ stripping operation o 3 Solution Polymerization Reactor 4 VC recovery (stripping) Potential emission points Loading lines, VC storage tanks Mixing, weighing and holding tank vents Polymerization reactor opening loss (ROL) Polymerization reactor relief valve discharges Polymerization reactor rupture disc discharges Recovery condenser vent {continued} Regulation requirements Emissions.from loading lines must be reduced so that upon opening of line to the atmosphere emissions do not exceed 0.0038mJ of VC at STP. VC removed from lines to meet this criteria must be controlled to < 10 ppm upon exhaust to the atmosphere. Concentration of exhaust gases discharged to the atmosphere from storage tanks must not exceed 10 ppm. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. ROL from each reactor is not to exceed 0.02 g VC/kg PVC products. No discharge to the atmosphere except for an emergency relief discharge. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Not required to be reported. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. GENC 0 1 6 5 0 3 -- * t------- 1i-----_------- t. I.,||-I o&i mm mm u~i l i J______ 1_______ ET--T3 i _______ , i________ ___ Process step 5 Coagulation 6 Resin drying 7 PVC storage and loading Table 3-4. Concluded Potential emission points Resin solution storage tank condenser vent, resin precipitation vent Dryer vent Silo vents Regulation requirements Not regulated. Not regulated. Not regulated. U> CO / GENC 016509 After stripping, the resin in solution is recovered by coagulation. A constituent is added to the solution reducing the solution solubility of the copolymer resin and allowing the copolymer to precipitate. Copolymer resins are then separated by centrifugation. The resins are rinsed with water to prevent the particles from sticking together. The rinsed resins are flash dried and then are penumatically conveyed to storage silos or packaged for shipment. Solution resins are used by casting and coating fabricators to provide thin coatings for food and beverage containers. Solution resins are highly pure because emulsifiers or suspending agents are not used in the process. 3.4.8 Polymerization Reactors The design of the reactors used in the various polymerization processes is important for controlling resin quality. Polymerization conditions (e.g., temperature, pressure, and agitation) within the reactor are primary control variables in the production of PVC resins. These operating conditions along with reactive constituents charged to the reactor produce the various resin grades. Reactor emissions are subject to several sections of the current regulation (see Table 2-1). Requirements of Section 61.64 of the regulation specify the following emissions limits: Concentration of VC in all exhaust gases discharged to the atmosphere must not exceed 10 ppm except for emergency relief valve discharges. e Reactor opening loss (ROL) emissions are not to exceed 0.02 grams VC per kilogram of PVC product. Leakage from reactor agitator seals and relief valves are covered under subsections 61.65(b)(3v) and 61.65(b)(4), respectively, which require installation of double-mechanical seals on agitator shafts and installation of rupture discs upstream of relief valves; relief valves must be connected to a control device. Equivalency clauses for these equipment requirements are~TncTudedunder Sections 61.65(b)(3v) and 61.65(b)(4). Both suspension and dispersion (emulsion) type resins are produced in a similar reactor under similar conditions and thus will be discussed together. Bulk (mass) type resins are produced in a two-reactor system. Solution type resins are the only PVC resins commercially produced by a 3-32 GENC 016510 continuous process in a reactor similar to a distillation column. A discussion of various reactors follows. Reactors for Suspension and Dispersion (Emulsion) Type Resins The following reactor parameters are common to suspension and dispersion processes: Reactor Size and Number - each plant may have from 4 to more than 50 reactors. Reactors currently being used in the United States range in size from 11.3 cubic meters to 169.5 cubic -j- -- 1 *^TTT" "* '"-"1 . meters (3,000 gallons to 45,000 gallons) (Khan, 1978, p. 17). 'The larger reactors with a 23 cubic meter (6,000 gallon) capacity and larger are relatively new (early 1970's) and are being used to reduce variability in product quality and increase production capacity. With larger and fewer reactors, incorporation of sophisticated controls is less expensive on a cost-per-pound basis. Reactor Construction - reactors are of stainless steel, glass-lined carbon steel, glass-lined stainless steel or stainless steel-lined carbon steel construction. Choice of material is dependent on corrosion resistance required and desired lifetime of reactors (Khan, 1978, p. 17). In larger reactors where wall thickness approaches 0.025 meters (1 inch), a stainless steel-lined carbon steel reactor offers an advantage for thermal conductivity (Cameron, 1979, p. 45). This is necessary in order to dissipate heat of reaction from the larger amounts of polymer slurry. In small reactors, where heat conduction is not as critical, glass-lined reactors are used. The glass lining on the smaller reactors helps to prevent polymer build-up on the reactor wall which will normally be hotter than the reactor wall on a large reactor. Operating Temperature and Pressure - for suspension resins, temperature of the reactor is usually about 55C (Khan, 1979, p. 75) and is maintained as a function, of the resin properties desired. This normally produces reactor pressures of 515 kilo- pascals to 810 kilopascals (5.1 to 8.0 atmospheres). Dispersion resins are more sensitive to heat and are processed at lower temperatures (Lamorte, 1978, p. 23). 3-33 GENC 0165U Agitation - mixing blades are usually of either retreat curve or turbine-type and provide the agitation speed that directly affects product quality (Cameron, 1979, p. 45). In suspension polymerization, agitation is often operated at lower speeds than in dispersion polymerization because intense agitation of dispersion resins results in poor control of particle size (Nass, 1977, p. 94). Agitation speed is variable and is determined by the type of resin particle desired. The use of 'baffles in the reactor is also important to produce a better size distribution of particles. The tubular-finger baffle and single-blade baffle are used and in some instances their positioning can be set from outside the reactor. Power require ments of the agitator drive motor are usually monitored by an amperage meter. This monitoring instrument is important because loss of agitation reduces heat transfer to reactor walls and a runaway reaction could occur. Cleaning - in order to produce high quality resins, the internal surface of the reactor must be kept clean. After several batches, polymer build-up occurs where liquid polymer slurry contacts the reactor walls. Polymer build-up is also formed on areas of the reactor that are not in contact with the slurry. This formation on surfaces other than the walls is due mostly to a reaction of VC in the gas phase with oxygen, a common contaminant in the reactor, or possibly from HC1 formed by hydrolysis between liquid VC and water. Regardless of cause, if polymer build-up is not removed, small flakes of the polymer will contaminate the next batch. These flakes do not absorb compounding rtkI ns- and produce areas known as "fish eyes" in the finished resin product. The most common cleaning practice involves opening the reactor manways to allow plant personnel access to clean the reactor interior. This may be done manually by scraping the walls or by fitting a high pressure water cleaning system into the reactor manway. Prior to opening, the ROL requirement must be attained. In order to meet this requirement some 3-34 GENC 016512 method of reactor evacuation is required. The evacuation method used to attain the ROL is usually a time-consuming procedure and results in reduced production capacity. These procedures will be discussed in detail in Section 4.5. To reduce these reactor openings for cleaning purposes, high pressure water jets have been installed within some of the newer reactor vessels. Also, solvent cleaning systems have been used by several companies - a solvent solution is passed through the reactor several times to remove polymer build-up. abandoned the aDDroach because of high cost and potential toxicity of the solvent. The use of a chemical treatment of the the reactor walls after cleaning has also been successful in preventing polymer build-up. This chemical, which is usually proprietary, is applied to the reactor walls prior to polymerization. Relief Valves and Rupture Disks - relief valves and rupture discs are safety devices on the polymerization reactor vessel. These devices open directly to the atmosphere under abnormal high-pressure conditions. If the pressure in the reactor vessel increases beyond a safe limit, the relief valve or rupture disc relieves the pressure in the vessel. Without this safety feature a vessel could rupture. Causes of increased pressure conditions will be discussed in Section 4.2. Relief valves used for reactor safety are set at approximately 50 to 100 pounds over normal operating pressure. Conventional relief valves allow a reactor to depressurize and, if operating properly, the relief valve will close or reseat again after pressure is reduced. Reactors for Bulk Resins A description of bulk reactors follows: Reactor Size and Number - the bulk process uses two types of reactors for polymerization of VC - the pre-polymerization (Pre-Po) reactor where the reaction is initiated and the 3-35 GENC 016513 post-polymerization (Po-Po) reactor where the reaction is completed. The Pre-Po reactor usually has a 2,200 gallon capacity (Dubec, 1980) and cycle time is short (about 2 hours) in order to supply seed resins to other Po-Po reactor. The Po-Po capacity is normally 4,400 gallons (Holbrook, 1980). Reactor Construction - both Pre-Po and Po-Po reactors are of stainless steel construction with water jackets for cooling. The Pre-Po reactor vessel stands upright with the agitator ishaft entering at the top. The Po-Po reactor vessel is horizontal with the agitator shaft entering at one end. Operating Temperature and Pressure - polymerization temperature in the reactors is normally between 40C to 70C (120F to 158F) producing VC vapor pressures in the reactor of between 500 kilopascals to 1,200 kilopascals (5 atmospheres to 12 atmospheres). Temperature control is much more critical in the bulk process because there is no water to transfer heat to reactor walls. In addition to reactor jacket cooling water, reflux condensers are used to remove heat by condensing gaseous VC from the reactor and returning it as liquid VC. Agitation - agitation in the Pre-Po is by a flat bladed, turbine-type agitator and is much stronger than the agitation used in the Po-Po. This stronger agitation is necessary to produce small seed particles of the required size distribution (Goiran, 1980). Baffles are used to prevent formation of a vortex (whirlpool). Agitation in the Po-Po is by a ribbon blender which results in less extensive and slower agitation (Schoultz, 1977, p. 654). The particle size produced is dependent on the agitation history in the Pre-Po reactor. Cleaning - polymer build-up in tfie Pre-Po is slow because polymerization only reacts to 7 to 12 percent conversion. Cleaning may only be necessary every 5 to 50 batches depending on the resin product (Dubec, 1980). The Po-Po is opened for cleaning after every batch for manual cleaning because the dry, powdery resin adheres to the reactor walls and agitator blades (Dubec, 1980). 3-36 GENC 016514 Rupture Disc and Relief Valves - the same as those used on suspension/dispersion reactors. , Reactors for Solution Resins Polymerization reactors in the solution process operate on a continuous cycle as opposed to the batch cycle of reactors used in suspension, dispersion, and bulk processes. Heating to initiate the solution process is supplied by hot water that is passed through coils inside the reactor. The solution process is run at lower temperatures than the other processes and over-pressure problems are rare (Erdmann, 1980).' Heat of reaction is removed by reflux condensers. Agitation is provided by an external pump cycle that circulates the reactants through the reactor. Relief valves and rupture discs are used on the solution reactor for safety purposes. Rupture discs are set at pressures higher than the relief valve. Reactor relief valve discharges from an accelerated reaction are uncommon in the solution process, plost relief valve releases : ^that occur are due to premature rupture disc failure.- ' Polymer build-up is not a problem in the solution reactors. When cleaning is necessary, pure solvent is circulated through the reactors for cleaning. Reactors are only opened for maintenance and inspection procedures. 3.4.9 Emissions for a Typical PVC Plant The study done in support of the current regulation identified 41 existing PVC plants. The VC emissions from these 41 PVC plants totaled 85 gigagrams (187 million pounds) per year in 1974, which represented approximately 85 percent of the total nationwide emissions of VC. Emissions data submitted by PVC producers were used to calculate emissions estimates for seven areas within a typical PVC plant producing 68 gigagrams (150 million pounds) of PVC resin per year. These four areas of potential emissions from the typical plant were as follows: Reactor and Stripper Losses - this area includes safety relief valve and reactor opening losses. Monomer Recovery System - after recovery of VC from the process, the unrecoverable VC was discharged to the atmosphere (these emissions are now controlled with the primary control device). Slurry Blend Tank, Centrifuges. Dryers and Storage Silos these four areas are combined into sources after resin stripping, t Fugitive Emissions Sources. The total emissions from the typical PVC plant were approximately 2.7 gigagrams (6.0 million pounds) of VC per year. Emission data were also compiled for VC losses during equipment purges. These purges represent VC lost when equipment is taken out of service for maintenance or inspection. This equipment purging contributed an additional 833 kilograms (1834 pounds) per year. 3*38 GNc 01isl6 3.5 REFERENCES FOR CHAPTER 3 Albright, Lyle F. 1967a. Manufacture of Vinyl Chloride. Chemical Engineering (a). Albright, Lyle F. 1967b. Vinyl Chloride Polymerization by Suspension Process Yields Polyvinyl Chloride Resins. Chemical Engineering (b). Cameron, J. B. Lundeen, A. J. McCulley, J. H., Jr. 1979. Trends in Suspension PVC Manufacture. Hydrocarbon Processing. Chemical Week. 1976. PVC Rolls Out of Jeopardy into Jubilation. DeBernardi, James, Plant Manager, Lake Charles, La. Conoco Plant. Telecon with Matthew Boss, TRW, November 26, 1980. Dubec, Harold. 1980. Manager of Environmental Compliance. Trip report visit to Hooker Chemical Company, Ruco Division. Burlington, N. J. September 15, 1980. Erdman, J. F., Environmental Protection Coordinator, Union Carbide Corporation, Texas City, Texas. Telecon with Matthew Boss, TRW Environmental Engineering Division. December 14, 1980. Goiran, L. Polyvinyl Chloride (PVC) Manufacturing Equipment: Permit Application for Approval of Modification. Letter to David C. Hawkins, May 4, 1980. Hatch, Lewis F., and Matar, Sami. 1979. From Hydrocarbons to Petrochemicals. Hydrocarbon Processing. Holbrook, W. C., Director of Toxicology and Environmental Affairs. Trip report - visit to B. F. Goodrich Chemical Company, Pedricktown Plant. Pedricktown, N. J. September 17, 1980. Khan, Z. S., and Hughes, T. W. 1978. Source Assessment: Polyvinyl Chloride. Industrial Pollution Control Division, Industrial Environmental Research Laboratory. Cincinnati, Ohio. Lamorte, Michael F. 1978. National Emission Standards for Hazardous Air Pollutants Inspection Manual for Vinyl Chloride. Research Triangle Institute. Little, Arthur D., Inc. Vinyl Chloride Monomer Emissions From the PVC Processing Industries. Contract No. 68-02-1332, Task No. 10. August 1975. 3-39 6ENC 016517 McCulley, J., Process Engineer, Conoco. Telecon with Matthew Boss, TRW. November 24, 1980. Use of Relief Valves and Rupture Discs at PVC Faci1ities. McPherson, R. W.; Starks, C. M.; Fryar, G. J. 1979. Vinyl Chloride Monomer . . . What You Should Know. Hydrocarbon Processing. Mil by, Thomas H. 1978. Vinyl Chloride an Information Resource. Stanford Research Institute. Menlo Park, California. Mukerji, Asu. 1977. Unloading and Storage Technique for Vinyl Chloride Monomer. Chemical Engineering. 1 Nass, Leonard I. 1977. Encyclopedia of PVC. Society of Plastic Engineers, Inc., Vol. 3. New York. Marcel Dekker, Inc. Schoultz, Kenneth S.; Bochinski, Julius H.; Goeon, James A. 1977. Engineering Control Assessment of the Plastics and Resins Industry . . . Case Study: Manufacture of PVC by Bulk Polymerization. American Industrial Hygiene Association Journal. Shreve, R. N.; Brink, Joseph A., Jr. Chemical Process Industries. McGraw-Hill Book Company. New York, N. Y. 1977. Sorenson, Wayne R. 1977. A Close Look at PVC Today. Plastics Engineering. U.S. Environmental Protection Agency. 1975. Standard Support and Environmental Impact Statement: Emission Standard for Vinyl Chloride. Emission Standards and Engineering Division. Research Triangle Park, North Carolina. 3-40 GENC 016513 4.0 CONTROL TECHNIQUES USED TO COMPLY WITH THE EXISTING EMISSION STANDARD Tables 4-1 and 4-2 identify control technologies that can be applied to potential emission points from EDC/VC and PVC plants. Tables 4-3 and 4-4 show the estimated emissions reductions that result when the typical EDC/VC and PVC plants, developed during the original standard support study (EPA, 1975), comply with the applicable current standard. ^Actual ^emissions are lower,,..because the majority^of,,.plants,,.surveyed,_have.,,lower ^ . vr--iv;. .-- / l^ve2s.thap required by the current standard except for relief valve discharges'.* Primary control devices are reducing emissions, in most cases, well below the 10 ppm level for exhaust gases. EDC/VC plants using a pure oxygen (or combination oxygen and air) feedstock for the oxychlorination reactor and/or incinerating the oxy vent have reduced emissions below the standard. Fugitive emissions from new large reactor plants are 95 percent lower than those emissions from typical plants in 1975 (Holbrook, 1980a). Most PVC plants are stripping resins to lower levels than required and new purging methods have been developed to reduce reactor opening losses. The following sections discuss the control technologies that industry is currently using to accomplish the various requirements of the regulation and to achieve actual emissions reductions. Improvements and new i developments in control technology since promulgation of the regulations are also discussed. 4.1 DISCHARGE OF EXHAUST GASES TO THE ATMOSPHERE 4.1.1 Introduction The current regulations require that the discharge of exhaust, gases to the atmosphere be controlled to meet a set standard. The sources of these exhaust gases, the applicable standards, and control technologies GENC 01651? Table 4-1. POINT SOURCE EMISSIONS AND TECHNOLOGIES FOR CONTROL IN TYPICAL SUSPENSION DISPERSION AND BULK PVC PLANT Process step Potential mission points Regulation requirements Control technology 1 VC unloading end storage Loading lines, VC storage tank Emissions from loading lines must be reduced so that upon opening of line to the atmosphere emissions do not exceed O.OOMT of VC at STP. Purged to monomer recovery system VC removed from lines to meet this criteria must be controlled to < 10 ppm upon exhaust to the atmosphere. Concentration of exhaust gases discharged to the atmosphere from storage tanks must not exceed 10 ppm. Incineration, solvent absorption or carbon adsorption t Mixing, weighing and holding tanks before stripping operation Klxlng. weighing and holding tank vents Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Vented to monomer recovery system followed by Incineration, solvent absorption, carbon adsorption, or combinations of these 3 Polymerization -U Polymerization reactor opening R01 from each reactor Is not to exceed loss (R01) 0.02 g VC/kg PVC products. Solvent cleaning, sterna piston, water piston, reactor purge air blower, steam purge, etc., used before openlnq Vented to atmosphere or monomer recovery system Polymerization reactor relief valve discharges No discharge to the atmosphere except for an emergency relief discharge. Shortstop, contalnaent. Instrumentation, Improved operator training, etc. 4 Stripping Stripping vessel vent 5 Honomer recovery system Recovery system exhaust vents and knock-out pot Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. ( (Blending, mixing. weighing and holding after stripping operation) Slurry blend tanks and holding Controlled by stripping standards tank vents Vented to monomer recovery system followed by Incineration, solvent absorption or carbon adsorption Gasholders used in some Instances to collect all recovery vents and/or refrigeration to condense VC followed by Incineration, solvent absorption or carbon adsorption Stripping technology (contlnuerf) GENC 014520 u&t J --J Table 4-1. Concluded, Process step Potential emission points Regulation requirements Control technology Drying, siting, screening of dewatered resin PVC loading and storage "Inprocess* wastewater stripper Centrifuge vents, dryer vent stacks, storage silos, baghouse vents, screening operation vents Storage silos Wastewater storage tank Wastewater stripper column All of the above process steps Fugitive emissions sources Controlled by stripping standards. Controlled by stripping standards. VC removed from In process water Is to be ducted to a control system from which concentration of VC In exhaust gas does not exceed 10 ppm. Equipment specifications, operational procedures and leak detection and elimination programs. VC collected from equipment seals and operational procedures are to be controlled to 10 ppm upon exhaust to atmosphere. Stripping technology Stripping technology VC removed from wastewattr by steam stripping in column or batch vessel, vented to monomer recovery system followed by Incineration, solvent absorption or carbon adsorption Double mechanfcal seats, double outboard seals, rupture discs or equivalent equipment; closed systems and equipment purging to monomer recovery system; area monitors, portable monitors, routine leak surveys and maintenance programs Vented to monomer recovery system followed by incineration, solvent absorption or carbon adsorption II Table 4-2. POINT SOURCE EMISSIONS AND TECHNOLOGIES FOR CONTROL IN "BALANCED PROCESS" EDC/VC PLANTS Process step Potential emission points Regulation requirements Control technology 1 Olrect chlorination 2 [DC purification 3 "Inprocess" wastewater stripper Product condenser [DC crude storage, light ends column condenser, light ends storage tank, heavy ends column condenser, heavy ends storage tank Wastewater storage tank Wastewater stripper column 4 Oxychlor1 nation S VC cracking and purification s VC loading and storage Water wash column Oxychlorlnation process vent Separator tank [DC quench col urn HC1 column vent VCK column condenser Loading lines VC storage tanks 7 All of the above process Fugitive emissions steps Not regulated. All emission points are required to be controlled to <_ 10 ppm. Not regulated Incineration VC removed from Inprocess water Is to be ducted to a control system from which concentration of VC in exhaust gas does not exceed 10 ppm. [missions from reactor are not to exceed 0.2 g VC/kg of the 100 percent EDC product. Concentration in all exhaust gases must not exceed 10 ppm. Wastewater steam strtpped In column or batch vessel, VC can be recovered by refrigeration and exhaust gases Inctnerated Process modifications; Incineration; pure oxygen feed and Incineration Incineration [missions from loading lines (and any other equipment In VC service) must be reduced so that upon opening of line to the atmosphere emissions do not exceed 0.003Bar of VC at STP. VC removed from lines to meet this criteria must be controlled to < 10 ppm upon exhaust to the atmosphere. Concentration of exhaust gases discharged to the atmosphere from storage tanks must not exceed 10 ppm. Equipment specifications, operational procedures and leak detection and elimination programs. VC collected from equipment seals and operational procedures controlled to 10 ppm upon exhaust to the atmosphere. Closed systems, carbon adsorption and purge to monomer recovery system Incineration Double mechanical seals, double outboard seals, rupture discs or equivalent; purge to monomer recovery system; and area monitors, portable monitors, routine leak surveys and maintenance programs Incineration GENC 016522 ) Table 4-3. EMISSION REDUCTION FOR 316 Gg/yr EDC/VC FACILITY IN COMPLIANCE WITH CURRENT REGULATION Emission source Relief valve discharges Primary control Oxychlorlnation vent fugitive emissions Current standard_. ___, Is / Zero discharge J Uncontrolled3 emissions (kg/yr) Unknown 10 ppm 0.02 kg/100 kg EDC product Work practice and equipment standard 916,400 113,760 379,200 Regulated*1 emissions (kg/yr) 0 3,160 100,300d 37,920f Estimated actual emissions (kg/yr) C' 1.950c) 25,OOOe 10.0009 H r(i /O Total emissions 1,409,360 141,380 40,110 I cn *Based on EPA emissions estimates developed from emissions data submitted by Industrial sources. ^Represents emissions from EDC/VC meeting current standard; actual emissions are lower. cBased on relief valve discharge data fronrTaMt 4-22 and EDC/VC production data for 1977-1980 (Chemical and Engineering News, 1980a). ^ ^Assumes balanced process and 100 percent conversion during EDC cracking. `Estimate represents an average of emission levels ranging from plants using only air and not Incinerating the oxy vent to those using oxygen and incineration. This estimate Is based on very limited data (DeBernardl, 1981). fAssumes 90 percent reduction following Installation of required equipment and Implementation of leak detection and elimination programs. 9Based on results of a fugitive emission study done In an EDC/VC plant (Blacksmith, et al., 1980). GENC 016523 .Table 4-4. EMISSIONS REDUCTION FOR 68 Gq/yr PVC FACILITY IN COMPLIANCE WITH CURRENT REGULATION Emission source - Current standard lincontrol led emissions (kg/yr) Regulated emissions (kg/yr) Estimated actual emissions (kg/yr) Primary control Relief valve discharges Combined sources after resin stripping Fugitive emissions 10 ppm ^ -7ero discharge--. 400 ppm - suspension. latex, and bulk [2000 ppm - dispersion] Work practice and equipment standard 326,400 136,000 850,000 1,040,400 680 *~ 27,200 [136,000] 260,200e `0804,780C 1-3,6004^ to 2-6 O [74,800] 104,000f Reactor opening loss 0.002 kg/100 kg PVC product 312,800 1,360 Total emissions 2,665,600 ----------239 ,'440---------------- [398,240] [185,620] ___ g *Based on EPA emissions estimates developed from emissions data submitted by Industrial sources. ^Represents emissions from PVC plant meeting current standard; actual emissions are lower except for relief valve discharges. cBa$ed on relief valve discharge data from Table 4-21 and total PVC production for 1977-1980 (Chemical and Engineering News, 1978; Ibid. 1980b), Emission value includes 580 kg/yr for nonreactor relief valve discharges and 4,200 kg/yr for reactor relief valve discharges. ^Based on an average of stripping levels reported by Industrial sources. ^Assumes 75 percent reduction following Installation of required equipment and elimination of leak detection and elimination programs. ^Reflects a 90 percent reduction. Data Indicate that fugitive emissions are now 75 to 95 percent lower than EPA 1975 estimates (Holbrook, 1980a). 0 4-6 GENC 016524 `KPSi r frt> "~1 are presented in Tables 4-1 and 4-2. (See Figure 3-1 and 3-2 in Section 3.0 (Process Description) for locations of process steps). A complete copy of the regulation can be found in Appendix A. The most common primary control technologies currently used for exhaust gases from EDC/VC plants and PVC plants are incineration, solvent absorption, and carbon adsorption. Because only emission levels greater than 10 ppm are currently reported, there is no way to know how effectively exhaust gas controls are working. In fact, levels may be much less than 10 ppm most of the time. Results of the initial compliance tests required for primary control devices following the 2 year waiver period give an indication of the effectiveness of these devices. However, only one initial source test was required and, in most cases, the control device was operating at peak performance. The effectiveness of the control devices over long periods of time and under variable plant conditions is not known. The current standard requires^zero emission levels during periods of control equipment shutdown and for this reason most plants maintain back-up control equipment to provide emission control when primary exhaust gas control equipment is not operating. Some of these secondary devices are duplicates of the primary systems (e.g., parallel incinerators) that are capable of handling 100 percent of plant exhaust gas discharge emissions, enabling the plant to maintain full production status. Other back-up systems are potential primary control devices (e.g., incinerator back-up for a solvent absorption system). Some plants use temporary measures for back-up control such as short term storage vessels while others merely maintain an inventory of spare parts for the single primary control device. There are several plants that do not have any back-up control systems. In the original standard support document, several technologies were identified for possible control of VC emissions in exhaust vents. These control technologies are being used by the industry and are discussed in the following sections. 4.1.2 Incineration In the plants surveyed during this review study, incineration represented the most prevalent method of exhaust gas emission control. 4-7 GENC 0165 In those plants not using incineration as primary control, it is usually used as a back-up for the primary control system. Many of the plants maintain two incinerators so that control is provided if one is out of service. Waste streams containing chlorinated hydrocarbons such as VC are more difficult to combust and higher temperatures of combustion are usually required in comparison with non-chlorinated hydrocarbon waste streams. Combustion temperatures of 980 to 1,100C (1,800 to 3,000F) have been recommended for efficient destruction of halogen-containing hydrocarbons by thermal oxidation. However, many of the VC sources surveyed that use thermal oxidation are operating their incinerators at much lower combustion temperatures. A disadvantage to incineration is that it is a destructive control method. No VC is recovered and in those cases when waste heat has been recovered, no return has been shown. The methods of incineration used in the VC industry are thermal and catalytic. The incineration method used most often is thermal oxidation. Incineration temperatures for VC emission control range from 760C to 1,290C (1,400F to 2,350F) with residence times of 0.5 seconds to 2.0 seconds. One company has recently tested a back-up incinerator at a combustion temperature of 540C (1,000F). Results showed average VC concentrations of 0.26 ppm. The following data show the results of their tests with a range of incineration temperatures. These tests were performed to determine compliance and were observed by a Region IV representative as well as members of the Kentucky Division of Air Pollution Control. At the time of the tests, the plant (an EDC/VC facility) was operating under conditions stipulated by EPA compliance test parameters, i.e., at least 90 percent capacity (Holbrook, 1980b). Combustion temperature 540C (1,000F) 650C (1,200F) 870C (1,600F) 980C (1,800F) 4-8 Average VC concentration 0.26 ppm None detected None detected 0.18 ppm GENC 016526 One of the major VC and PVC producers has extensively studied incineration technology and they report that the large size and high temperatures (used in conventional thermal incineration) are not always necessary for efficient combustion of VC. This manufacturer uses a smaller (Brule) incinerator as a back-up control in its PVC plants. The incinerator and stack are lined with a ceramic material that allows the unit to reach optimum temperature within a few hours. A large surge tank contains the exhaust gases until the incinerator comes up to temperature and regulates the flow to the unit. The incinerator is heated to 704 to 760C (1,300 to 1,400F) using supplemental fuel. At this point the vinyl chloride waste stream is fed to the incinerator which raises the energy content of the stream, thus causing a reduction of the supplemental fuel feed. A temperature of 870 to 980C (1,600 to 1,800F) is maintained which burns the exhaust gases to less than 10 ppm VC (Varner, 1980). Auxiliary fuel usage in thermal oxidation systems varies according to processes. One EDC/VC plant adds methane during direct chlorination. This both enhances the combustion of vinyl chloride as well as provides a "fuel-rich" mixture in the reactor to avoid explosion. Some EDC/VC plants use pure oxygen instead of air as feedstock for the oxychlorination reactor. As mentioned in Section 3.1.2, this minimizes the venting of inerts, provides more efficient incineration, and greatly reduces auxiliary fuel usage. These plants view the reduced energy consumption as an economic advantage. The use of pure oxygen would seem to be a measure to attain compliance with other hydrocarbon standards (such as volatile organic compounds (V0C)) (Brittain, 1980a). There are several approaches to the control of the EDC oxychlorination reactor vent (oxy vent). Some EDC/VC plants do not incinerate the oxychlorination vent gas but, instead, meet the standard by process modifications (e.g., incorporation of vapor phase catalytic reaction) (DeBernardi, 1980). There may be high concentrations of VC emissions during start-up, shutdown and unstable operating conditions. Some plants incinerate the oxy vent during these periods (Brittain, 1980b). Oxy vent emissions are potential candidates for regulation (for V0C) by State Implementation Plans (SIP) and New Source Performance Standards (NSPS) programs. 4-9 ENC 016527 Incinerators are equipped with flame arresters or flash-back preventative devices necessitated by the hydrocarbon-oxygen content of the waste streams. Because combustion of halogen-containing hydrocarbons results in the formation of HC1, quench systems for cooling and caustic scrubbers for HC1 removal are incorporated when required by state regulations. Continuous monitoring of incineration stacks and compliance tests have shown VC levels ranging from "non-detectable at 0.1 ppm" to "less than 10 ppm" according to regional enforcement personnel. The following problems have been identified with thermal incineration systems. Incinerator overloading. When VC levels are too high the rich mixture of combustibles leads to elevated temperatures and results in maintenance problems. Some plants have solved this problem by using a surge vessel (such as a gasholder) to provide a constant feed to the incinerator. The thermal incineration process can be a source of secondary air and water pollution (e.g., HC1 and Cl2 generation from combustion and high total dissolved solids (TDS) levels from scrubbing). Supplemental fuel and high maintenance requirements represent additional expenses. One of the major maintenance items - "downcomers" (connections between furnace and quench system) - has been estimated to cost $6,000 to replace when corroded, and replacement may be necessary as often as once each month. Overall annual maintenance costs for incineration have been reported around $100,000 (for PVC plant incinerators designed to handle over 45 kilograms per hour VC). A PVC plant (with a production rate of 68 gigagrams per hour of PVC resins per year) uses 26.5 cubic meters (7,000 gallons) of No. 2 fuel oil per month to keep their dual incincerators hot (Dubec, 1980). Monitoring is difficult due to temperature fluctuations and moisture condensation. Some plants have solved this problem 4-10 GENC 016528 by using a system that removes moisture prior to analysis (Laundrie, 1980). Location of the incinerator requires "safe radius" considerations. Catalytic incineration systems are currently used by at least one plant for primary control, and at least one other plant is experimenting with them. Energy requirements for catalytic oxidation are approximately one-third of that used for thermal oxidation without heat recovery. The plant using catalytic incineration passes the exhaust gas vapors through a wash oil scrubbing system. This hydrocarbon oil contacts the stream in the vent condenser line countercurrently, recovers vinyl chloride, and provides a stable load to the incinerator. This is an experimental procedure serving more of a VC recovery function than an emission control purpose. Problems with catalytic incineration include: Surge capacity. Assuring even flow into the incinerator has presented a problem for some plants. t Catalyst pollution. Organic halides pollute and degrade the catalyst which is expensive to replace. Conversion to other chlorinated hydrocarbons. Efficiency. These units reportedly remove less than 60 percent of the VC in the oxychlorination process. Waste heat boilers can be used in conjunction with incinerators; they are used for steam generation for heat recovery. 4.1.3 Steam Boilers None of the plants surveyed use steam boilers as primary control. When used, they are maintained for back-up control. The long term reliability of these units is limited by the corrosivity of the vinyl chloride stream. One plant is considering the use of "expendable" boilers as a back-up for this reason. 4.1.4 Flares The use of flares is generally restricted to back-up control. Generally speaking, a flare would be installed to accommodate streams from a large chemical complex in order to reduce hydrocarbon emissions. One plant uses a flare as an equivalency to the use of a rupture disc jy-5 4-11 GENC 01652? upstream from the safety valve. (This represented an approval on the basis of an equipment standard as opposed to an emission standard.) In this case the relief valve would discharge directly to the flare. Parameters for usage were specified (Brittain, 1980a). Regional EPA personnel discourage the use of flares because they cannot be tested or monitored reliably. There are other objections to the use of flares. ' Flare efficiencies have not been determined. There is a need for a large capacity, low pressure vapor/ liquid separator. The safe radius restriction can be prohibitive for flare location. For example, this radius would be 170 meters (560 feet) for a 17,000 kilojoule per square meter (1,500 Btu per square foot) per hour radiation density at ground level (Finch, 1980). Secondary pollutants (e.g. , noise, hydrogen chloride, smoke) are produced by the flaring process, thus elevation or isolation of the flare is required. Capital costs. Elevated flares cost between $30,000 and $100,000; ground flares can cost as much as ten times as an elevated flare for the same capacity (Neveril, p. 5-76). Energy use consideration. For general consideration, the quantity of steam required can be assumed to be 0.4 kilogram steam per kilogram of hydrocarbon (0.4 pounds of steam per pound of hydrocarbon) (Neveril, 1980). Also, the dilute gas streams present in both EDC/VC and PVC plants cannot burn without the addition of natural gas or other fuel. All the V heat produced is wasted. One company has two flares - one servicing the large chemical complex and the other installed specifically for PVC process emissions. The latter was originally designed for emergency releases and would accommodate two simultaneously-discharging reactors. The stack is 99 meters (325 feet) tall, thus overcoming the "safe radiation distance" problem. Two vessels (one dry knock-out (KO) drum and one water-sealed drum) provide enough knock-out volume capacity to keep the liquid from 4-12 6ENC 016530 the top of the stack. The flare has never been used for that "worst-case" condition but it does handle periodic small relief valve discharges (non-routine emergencies). Ninety-eight percent of the rupture disc/relief valve combinations are tied into the flare at that facility. One disadvantage of the use of a flare for VC emissions is the smoke inherent in burning the emissions. Even with a discharge of 15 to 20 seconds duration, the smoke emanating from the flare lasts for 1 to 2 hours, creating problems with state opacity regulation (Kachtick, 1980). 4.1.5 Carbon Adsorption This method is used as primary control in only a few of the plants surveyed. More often it is used in conjunction with other control devices (usually incineration). Most of the regional EPA personnel and industrial representatives felt that carbon adsorption alone is not effective in reducing discharge emissions to below 10 ppm. PVC plants are successfully using carbon adsorption systems as primary control. (It is possible that the competition from hydrocarbons, other than VC, found in exhaust streams at EOC/VC plants prevent the use of carbon adsorption systems.) The most practical application of carbon adsorption control technology would be in PVC plant monomer recovery systems, closed slurry blend tanks, and storage areas because of the high VC concentration and low volume streams found in these areas. One EDC/VC plant uses a very small carbon adsorption system two 0.4 cubic meter (110 gallon) carbon-filled vessels run in series at their marine loading dock. This is a portable system which is regenerated by passing hot nitrogen through the carbon beds and into the incinerator (DeBernardi, 1980). Two PVC plants using carbon adsorption as primary control>for exhaust gases use a double bed system. When a probe at the outlet of the bed indicates an approaching 5 ppm level (as an indicator of break through), the waste stream is diverted to the other bed while the first bed is being regenerated. Continuous monitors on these units show the systems to be effective in reducing VC emission levels to below 10 ppm. Back-up control for one plant is a boiler. With the VC levels from carbon adsorption below 10 ppm, the boiler can also accommodate effluent from the adsorber without HC1 corrosion problems. 4-13 GENC 016531 The following disadvantages have been cited for carbon adsorption systems. Regeneration of the beds requires high energy usage. Treatment of regenerating gas streams requires another control device (e.g., incineration). Polymerization of VC on the carbon beds is a potential problem. With new regulations on hazardous waste treatment and disposal (RCRA), eventual disposition of the contaminated carbon could be a problem. Carbon adsorption represents higher capital costs than solvent absorption and incineration. 4.1.6 Solvent Absorption Four of the plants surveyed use solvent absorption as primary control for exhaust gases. (All of these are PVC plants although the principles of solvent absorption can apply to exhaust gas control in EDC/VC plants). The vent gas absorber system used by the major producers incorporates the same figure-eight solvent absorption technology described in the original standard support document, with proprietary modifications to improve efficiency. The function of the vent gas absorber is to strip and recover residual VC. The recovered VC is then reused in polymerization. The vent gas absorber system consists of two packed columns. In the first column the VC gas is absorbed by the lean solvent which enters at the top. The non-absorbables and a small quantity of solvent are then vented to the atmosphere from the top of the column. The VC-rich solvent is passed through a heat exchanger or\ its way to the strippingcolumn. In the stripping column the solvent is heated to remove the VC that comes off atthe top and is returned to the process. The lean solvent comes off at the bottom of the column. The warm, lean solvent passes back through the heat exchanger and another cooler before returning to the stripping column for another cycle. The vent gas absorber system described above was developed by B. F. Goodrich who plans to license the technology; it is commercially available. The only other solvent absorption system used by the plants surveyed in this study is a proprietary system designed by the company using it. 4-14 GENC 016532 The efficiency claimed by the plants using the vent gas absorber system is 99.99 percent. Capital investment is slightly more than for incineration but less than for carbon adsorption. Utility costs for the solvent absorption system are greater than for incineration but less than those for carbon adsorption; however, VC recovery, achieved in the vent gas absorber system, results in a substantial credit to the system. B. F. Goodrich regards the safety advantages of their vent gas absorber system as significant. The VC from the feed stream is absorbed in cold solvent and the non-absorbables (containing oxygen) and a small quantity of solvent are vented to the atmosphere at the top of the absorbing column. In this operation the cold solvent acts as a built-in heat sink. As the VC is absorbed by the solvent, it is removed from the absorber and further contact with any oxygen in the feed stream. Even though a small volume of gas passes through the explosive range, the cold solvent heat sink makes the operation safe. Because there is no flame associated with the operation of a vent gas absorber (as with incineration) it can be located close to other parts of the PVC process. In addition, B. F. Goodrich claims that the unit has a low environmental pollution potential even though some small amount of solvent is released to the atmosphere. The solvent is proprietary, commercially available, inexpensive, and reputed to be low in toxicity. 4.1-7 Refrigeration This method of controlling exhaust gases is used only in conjunction with other control devices to reduce the load on these downstream systems. Condensation of VC and water reduces the volume of gases to be handled by primary control systems. Use of refrigeration is usually limited to monomer-recovery systems and in some cases, installation of refrigeration units was in response to hydrocarbon control for SIP compliance. A typical application of this control technology in a recovery system can be illustrated as follows: Vents from compressor relief valves, pumps, transfer lines, weigh scales, condensers and knockout pots, and slurry and wastewater strippers would go to a common holding vessel. VC from this vessel would be recovered through compressing and 4-15 OENC 016533 \3*s ' condensing the monomer in a refrigeration unit. The noncondensable stream from the recovery system would be vented to a surge tank or gasholder which would vent to the incinerator (Laundrie, 1980). The above description represents a monomer-recovery system for a plant utilizing 18 reactors with an average capacity of 19 cubic meters (5,000 gallons). Equipment, including seven 0.2 cubic meters per second (400 cfm) vacuum pumps, two compressors (one 0.2 cubic meters per second (400 cfm) and one 0.1 cubic meters per second (200 cfm)) and two 0.01 cubic meters per second (30 cfm) vent compressors represents a $45,000 capital cost and a $1,000 annual operating cost (Laundrie, 1980). 4.1.8 Other Controls One EDC/VC plant in Region VI uses a separate fixed-bed oxychlorination reactor to receive only the oxy vent exhaust stream. The reactor converts VC to heavier chlorinated hydrocarbons which are used in other processes (Brittain, 1980a). In the past several years a number of procedures have been proposed for the removal of vinyl chloride from gas streams. Some of these systems that have been developed, but are not in current use by the plants surveyed in this study, follow (Sittig, 1977). Reaction with ozone. The disadvantage of this method is that it is slow and requires long residence times to reduce the VC content of the gas stream to 1 ppm or less. In addition, it is difficult to meter ozone into the gas streams in amounts that will destroy substantially all of the VC without leaving an appreciable amount of ozone in the effluent gas. There are also environmental problems arising from the presence in the effluent gas of ozonides formed by the reaction of vinyl chloride with ozone. Reaction with ozone in the presence of activated carbon. This process (patent assigned to Tenneco Chemicals, Inc.) is one in which VC is removed from gas streams that contain from 10 ppm to 1,000 ppm of VC by contacting the gas stream with ozone in the presence of activated carbon. The gas streams treated in this way contain less than approximately 1 ppm of vinyl chloride and no detectable amount of ozone or ozonides. 4-16 GENC 014534 Another process (patent assigned to Stauffer Chemical Company) is also one in which the stream contacts ozone but without activated carbon. The process can be used to treat gas streams that arise in the ethylene oxychlorination processes, ethylene dichloride cracking operations, VC polymerization processes in which VC is a monomer or comonomer, ventilation streams from areas in which VC is or may be present, processes for preparing vinylidene chloride, and polymerization processes in which vinylidene chloride is utilized as a monomer or comonomer. After reaction with ozone, the gas stream contains hydrogen chloride, oxygenated compounds such as carbon dioxide and water, and can contain phosgene and partially oxygenated hydrocarbons such as methanol. Contacting the treated gas stream with an aqueous medium is advantageous because products of the reaction are removed from the gas stream and partially oxidized hydrocarbons can further react with any unreacted ozone present in the gas stream. The aqueous medium also aids in hydrolysis of reaction products of ozone and the chlorinated hydrocarbons present. 4.2 RELIEF VALVE DISCHARGES 4.2.1 Introduction Pressure vessels, transfer lines, and other equipment in EDC/VC and PVC plants are equipped with safety relief valves, rupture discs or a combination rupture disc/relief valve assembly to prevent overpressurization which might cause a rupture to occur. The size of polymerization reactors, which can range from several thousand liters for the older reactors to the new reactors of up to approximately 190,000 liters (50,000 gallons) used by Huls of Germany (the largest reactors currently being used in the United States range from 35,000 to 40,000 gallons) plus this close proximity of the reactors to each other, represents the greatest potential for an explosion hazard. It is for this reason that existing safety regulations and insurance companies require and strictly enforce the use of safety relief devices on pressurized equipment. Relief valve discharges, which cause short-term peak emissions, ....... represented approximately 4 percent of total emissions from a typical 4-17 GENC 01*535 PVC plant prior to promulgation of the regulation (EPA, 1975, p. 5-6). Relief valve discharge emissions were not quantified for a typical EDC/VC plant. The major concern during the original standard support study was for relief valves located on PVC reactors because these relief valves represented a large source of emissions. PVC plants prevented reactor relief valve discharges by equipping reactors with instruments to warn personnel of an emergency condition. Preventive measures could then be taken such as injecting a short-stop agent to kill the reaction or manually relieving pressure to the recovery system. One plant used a gasholder to prevent relief valve discharges and EPA indicated that a gasholder could be sized to hold all the VC present in an entire reactor batch (EPA, 1975, p. 4-30). Based on this information, relief valve emissions were addressed in Section 61.65(a) which states, Except for an emergency relief valve discharge, there is to be no discharge to the atmosphere from any relief valve on any equipment in vinyl chloride service. An emergency relief discharge means a discharge [that] could not have been avoided by taking measures to prevent the discharge. When a relief valve discharge occurs, the plant is required to notify EPA and submit such data as the source (specific piece of equipment), cause, quantity, and measures that would be taken to prevent future discharges. 4.2.2 Emissions from Safety Relief Valves The intention of Section 61.65(a) was to eliminate relief valve discharges through the proper combination of control equipment and operating procedures. If a release then occurred, it would be considered either an emergency condition or one in which a plant had not implemented the proper combination of control techniques (e.g., instrumentation, short-stop agents, or gasholder). The plants with preventable relief valve discharges would then be subject to enforcement actions until relief valve emissions were eliminated. ................ ftJif^rkeg^aT enforcement..personnel -Indicate ..that ..releases continuing to occur, and the majority of violations jappear-.to-be `"able. As mentioned, the concern during the original standard support 4-18 GENC 016536 study was for relief valve discharges from PVC reactors that have the potential to discharge the entire reactor contents. Prior to the regula tion, relief valve discharges were not accurately measured but typically 2200 kg (5000 lb) of VC was released in a 5 to 10 minute period (EPA 1975, p, 4-30). With the change to larger reactor systems, the potential^) quantity of emissions from the relief valve is increased. _5 Relief valve discharge data from EPA Regional Offices substantiate the fact that releases are continuing to occur and reactors account for the largest quantities per release. Table 4-5 shows a compilation of reactor and non-reactor relief valve discharge data for 32 regulated sources (55 percent of all sources) for the period 1977 to 1980 (Brittain, 1980; Diem, 1980; Aronson, 1980; West, 1981). These^ plants`were3 responsible for 533 relief valve discharge events totalling approximately 450,000 kilograms (1 million pounds) of VC emissions over the 4 year/ period. PVC plants experience both reactor and non-reactor relief valve discharges. Table 4-6 shows that portion of the relief valve discharge data in Table 4-5 contributed from PVC plants, which is approximately 82 percent of the total events and 91 percent of the quantity of VC emitted. Based on these data, an average reactor relief valve discharge accounted for approximately 1025 kilograms (2275 pounds) of VC being released to the atmosphere. The time period for a reactor relief valve discharge event ranged from less than 1 minute to 115 minutes, with typical events less than 10 minutes in duration. The average PVC non reactor relief valve discharge was approximately 540 kilograms (1200 pounds) of VC emissions over a period of 10 minutes or less. EDC/VC plants experience non-reactor related relief valve- discharges. Table 4-7 shows that a portion of the relief valve discharge data in Table 4-5 was contributed from EDC/VC plants. Based on these data, the average non-reactor relief valve discharge from EDC/VC plants was approxi mately 430 kilograms (950 pounds) of VC emissions. The duration of the EDC/VC non-reactor relief valve discharges was not determined. Together, the EDC/VC and PVC non-reactor relief valve discharges represented 34 percent of the total events and 20 percent of the total quantity of VC emitted by relief valve discharge. 7^ 4-19 GENC 0US37 Table 4-5. TOTAL NUMBER OF RELIEF VALVE DISCHARGES AND QUANTITY OF VC EMITTED FROM 32 REGULATED SOURCES3 FOR THE PERIOD 1977 to 1980 1977*5 1978b 1979 1980c TOTALS Events 149 130 156 97 533 Reactor and non-reactor relief valve discharges kg VC (lb VC) 90,804 126,835 153,212 75,718 446,569 (201,787) (281,856) (340,472) (168.260) (992,375) aThe 32 regulated sources are 9 EDC/VC plants and 23 PVC plants. bRelief valve discharge data for 1977 does not reflect data from 1 EDC/VC plant and 5 PVC plants; these same plants reported data for only 3 months of 1978. cRelief valve discharge data for 1980 ranges from 8 to 12 months. i 4-20 GENC 01453s Table 4-6. RELIEF VALVE DISCHARGES FROM PVC PLANTS FOR THE PERIOD 1977 to 1980 PVC plants3 Non-reactor relief valve discharges Events kg VC (lb VC) Reactor relief valve discharges Events kg vc (lb VC) 1977 1978 b 1979 1980 18 23 23 23 TOTALS 9 6 50 26 91 17,550 1,742 11,912 17,998 49,202 (38,999) (3,869) (26,471) (39,996) (109,335) 118 65,028 (144,507) 102 111,375 (247,500) 77 129,689 (288,199) 53 52,174 (115,941) 350 358,266 (796,147) aThe 23 PVC plants represent 58 percent of the total number of PVC plants in the U.S.; 2 of these PVC plants reported no relief valve discharges for the 4 year period. ^Relief valve discharge data from 5 plants is for 3 months of 1978. cRelief valve discharge data' for 1980 ranges from 8 to 12 months. CD 2 o o CK cn 'tOo Table 4-7. RELIEF VALVE DISCHARGES FROM EDC/VC PLANTS FOR THE PERIOD 1977 to 1980 1977 197 8b 1979 198QC EDC/VC plants 8 9 9 9 TOTALS Non-reactor relief valve discharges Events kg VC (lb VC) 22 8,226 (18,281) 22 13,719 (30,487) 29 11,611 (25,802) 18 5,545 (12,323) 91 39,101 (86,893) aThe 9 1EDC/VC plants represent 50 percent of the total number of EDC/VC plants in the U.S. ^Relief valve discharge data from 1 plant is for 3 months of 1978. cRelief valve discharge data for 1980i ranges from 8 to 10 months. 4-22 GENC 01*540 Based on the relief valve discharge reports submitted to Regional offices, the rate of reactor and non-reactor discharges are not consistent throughout the industry. Many sources have reported few, if any, releases during the past several years, while other sources continue to have releases on a regular basis. The decline (or absence) of any relief valve discharges may also be due to the wording of the regulation that states only "relief valve" discharges are required to be reported. Plants using other safety relief devices (e.g., rupture discs only) are / not required to report discharges through these devices. The reported releases that continue to occur are more prevalent among the older sources using small reactor technology; however, some of these older sources have reduced releases through application of available control technology. 4.2.3 Relief Valve Discharges from Reactors Relief valve discharges are classified as either reactor releases or non-reactor releases - PVC plants experience both types while EDC/VC plants only experience non-reactor releases. The causes of reactor releases and measures that can be taken to prevent them are discussed in the following sections. 4.2.3.1 Process Variations. The frequency of, and ability to control, relief valve discharges from reactors vary among the different polymerization processes. In addition, the frequency of discharges is dependent on the age of the plant. Newer (and larger) reactor systems are replacing the older systems which usually consisted of many small reactors. These newer reactor systems provide better process control, fewer upset conditions, and fewer emissions to the atmosphere. Suspension/Dispersion i The newer and larger reactors are most often applied to the suspension and dispersion processes. As mentioned above, these newer systems provide better process control and reduce emissions. Newer) reactors have fewer piping connections, valves, and mechanicaf operations per kilogram of monomer transformed to resin. The fewer number of batches and reactors that need to be monitored when larger reactors are utilized results in a lessened probability of relief valve discharges. Also, relief valve discharge control equipment and procedures can be more economically applied to new plants with large reactors. 4-23 6ENC 016541 Older smaller reactors are usually operated closer to the rated pressure and when an upset condition develops it becomes more difficult to short-stop the reaction before the relief valve set pressure is attained. Construction materials in the newer reactor systems (e.g., stainless steel, carbon steel) contribute to better heat transfer and control of the reaction. The older smaller reactors are usually glasslined which has an insulating effect on the reactor and results in poorer heat transfer and less control over the reaction. Bulk Polymerization The bulk (or mass) polymerization process has some advantages over polymerization conducted in the aqueous medium used in the suspension and dispersion processes. Because the bulk process is anhydrous and no water or suspending agents are used, the process generates no foam and energy consumption is low. Exothermic heat is removed by water-cooled reflux condensers. The bulk process currently used in the United States utilizes a vertical pre-polymerization (Pre-Po) reactor and a horizontal post-polymerization (Po-Po) reactor. A new bulk process has been developed that utilizes a vertical Po-Po reactor, but it is not currently in use in the United States. These bulk processes have several unique aspects that affect relief valve discharges. Relief valve discharges from the Pre-Po reactor are unlikely because the reaction only goes to 8 to 12 percent conversion. If a high tempera ture or pressure is detected in the Pre-Po, the slurry can be dropped to an empty Po-Po reactor where the larger volume allows the low level of initiator charged to the Pre-Po to be used up and complete the reaction. However, reaction control is more difficult in the Po-Po reactor because the bulk reaction is anhydrous and this results in less efficient heat transfer to the walls of the reactor. Almost all of the liquid VC charged to the Po-Po is used during the e'-ly stages of the reaction and it is during this early stage that auto-acceleration of the reaction and a rapid increase in temperature can occur. Therefore, process control is more critical during this initial reaction stage. If an upset situation is detected in the Po-Po reactor, the conventional short-stopping techniques, effectively used in the suspen sion and dispersion processes, can not be used to terminate the bulk 4-24 GENC 016542 reaction because the contents of the Po-Po reactor are not fluid and mixing is not turbulent. As a result, the short-stop agent is not distributed rapidly or completely enough throughout the slurry. Rhone-Poulenc has developed a short-stopping procedure, but it is only applicable to their newer process technology utilizing vertical Po-Po reactors (Dubec, 1980). B. F. Goodrich operates two bulk plants in the United States and has developed a more effective gaseous short-stop system that may be available for licensing in the future (Dubec, 1980), Solution Process Union Carbide operates the only facility in the United States using the solution polymerization process. The solution process is a continuous, L homogeneous process - it is not a batch flow system like the other processes. The solvent, which is always present, provides a heat sink capable of preventing an accelerated reaction. In over 30 years of operating the solution process, Union Carbide has not experienced a reactor relief valve discharge to the atmosphere caused by an accelerated reaction. Reactor"relief valve discharges that have occurred were due to a premature failure of the ruptare disc under the relief valve / (Erdman, 1980). 4.2.3.2 Causes of Reactor Discharges. There can be many causes of reactor relief valve discharges ranging from total power failure caused by a natural disaster to mechanical failure of process equipment to operator errors. Two of the more frequent causes of relief valve discharges common to suspension and dispersion processes follow. The hydraulically full (hydroful) condition resulting from too much liquid or the presence of noncondensable gases in the reactor. This condition can be caused by instrumentation errors or an overcharging of the reactor and usually results in a small release to the atmosphere. High temperature in the reactor from an accelerated reaction. This condition is caused by inadequate cooling and can result in a discharge of the entire reactor contents. Both of these conditions can result in high pressure in the reactor and a substantial, multiphase discharge through the relief valve to the atmosphere. 4-25 L GENC 016543 Hydroful Condition The VC liquid charge expands up to-^'percent to fill the reactor when heated to reaction temperature. Therefore, an overcharging of the reactor may not be detected when raw materials are initially charged and the resulting increase in pressure above the normal operating pressure will open the relief valve when the slurry reaches the top of the reactor. Overcharging of the reactor can result from charge meter or weigh tank errors, leaking reactor valves, and wash-water or solvent cleaning solution incompletely drained from the reactor. The presence of noncondensable gases in the reactor will also cause high pressure to develop resulting in a hydroful condition. The operating procedure for a typical suspension or dispersion process is to charge a reactor with water, initiator, and other constituents (depending on the process and resin qualities desired) to about 50 percent of reactor volume. Pressure is then reduced to about 5 kPa (0.05 atmospheres). Vaporization of the water will sweep some noncondensable gases from the reactor if present. Liquid VC is added under vacuum to raise the slurry level to about 80 percent reactor volume. Liquid expansion from the reaction temperature then raises the slurry level to about 90 percent reactor volume. The concentration of any noncondensable gases present at 50 percent reactor volume can be increased by up to five times. Because the combined VC and water vapor pressure at normal operating temperatures can reach about 1,000 kPa (150 psia) and relief valve systems are usually set at 1,300 kPa (195 psia), the presence of only 5 percent noncondensable gases in the vapor space prior to VC liquid charging could trigger a release. The sources of noncondensable gases in the reactor include: gases not removed by initial evacuation during preparation for a new batch, gases dissolved or entrained in the VC liquid charge, and leakage from reactor valves into the reactor after vacuum treatment and before heating to reaction temperature. In the case of a hydrofuj condition caused by overcharging or the presence of noncondensable gases, the increased pressure causing the relief valve to open is quickly relieved by liquid and vapor flows 4-26 GENC 016544 through the valve. The relief valve will reseat itself under most circumstances after pressure is released. However, if the relief valve does not reseat properly or pieces of a blown rupture disc become lodged in the valve preventing complete closure, a larger volume of reactor contents will continue to be discharged until temperature and pressure are brought under control. Accelerated Reaction The polymerization reaction is exothermic and the reaction heat produced is controlled by circulating cooling water through the reactor jacket. The reaction normally takes place at a temperature and pressure of approximately 55C (130F) and 1000 kPa (150 psia). As the reaction proceeds, the polymer tends to coat the inner walls of the reactor reducing heat transfer effectiveness and, because reactors are no longer opened as often, more consecutive batches are run and polymer continues to coat the inner reactor walls. However, new clean reactor technology has reduced this polymer build-up even though reactors are not opened as often. In some of the newer systems, water jets clean the inner walls after each batch to prevent polymer build-up and loss of heat transfer effectiveness. The use of clean reactor technology or other proprietary methods for reduction of polymer build-up allows the most efficient heat transfer through reactor walls and thus reduces the danger of auto-acceleration due to high reaction temperatures. For suspension and dispersion processes, an auto-acceleration of the Reaction occurs with increased heat evolution at some point above th^SO- percent conversion level. Heat transfer effectiveness and control of the reaction during this period depend on vigorous agitation of the slurry and an adequate supply of cooling water applied to the reactor jacket. A malfunction of these systems will result in rapid heating and an increase in pressure that could cause the relief valve to open. The loss of agitation is the greatest concern because heat transfer is significantly reduced and liquid from the lower part of the reactor moving toward the slurry surface by convection currents will flash suddenly and cause a rapid increase in reactor pressure. Inadequate heat removal during the auto-acceleration period will usually result in a major discharge (possibly the entire contents of the 4-27 GENC o*6545 reactor) because of the rapid rise in temperature. An increase from the normal reaction temperature of 55C (130F) to 72C (162F) or a total increase of 17C (32F) can result in the sum of the VC and water vapor pressures reaching the relief valve discharge pressure setting. - However, if the reaction is in its later stages (greater than 80 percent conversion), most of the VC will have been used up and because the polymer density is higher (almost twice as high) than that of the VC density, the liquid level will be lower in the reactor. There would be a lesser discharge through the relief valve under these conditions. In the earlier stages of the reaction all of the VC would be vaporized, unless a shortstop agent were added to the reactor to kill the reaction. Addition of a shortstop agent usually requires that agitation still be available for distribution throughout the slurry. Conoco has developed a shortstop agent that is effective without agitation. This new development will be discussed in a subsequent section. 4.2.3.3 Prevention of reactor discharges. Proper instrumentation to detect upset conditions, gasholder equipment, and an automatic inhibitor solution (shortstop agent) addition systems can be used to eliminate many of these relief valve discharges (EPA, 1975). However, there are many other variables that must be considered for the above three controls to be successful. The following sections discuss other generic control measures, in addition to the EPA-recommended methods, used in the VC industry to prevent relief valve discharges. 4.2.3.3.1 Current generic preventive methods. Methods currently used by PVC plants to prevent emergency relief valve discharges follow. Shortstop systems * A shortstop or kill agent system can be used to stop the polymerization reaction when upset conditions develop. A shortstop system injects a chemical agent into the reactor which terminates the reaction by inhibiting the action of the initiator. The system is either manual, automated, or a combination of the two. The manual system generally uses high pressure water injection with the same equipment used to charge the reactor with initiator (Ledvina, 1980). Depending on the kill agent system used, success of the manual system is usually dependent on having agitation 4-28 GENC 016544 for complete dispersion of the shortstop throughout the slurry, the charge manifold being clear (this is a manifold used to charge ingredients to the reactor) and the availability of necessary personnel. Several variations of the automated shortstop system exist. The newer computer-controlled plants have built-in programs that recognize the upset condition by monitoring operating parameters and automatically injecting the shortstop agent. For example, one plant has installed motion sensors on the reactor agitator that sense when power has been lost and a pressure build-up is beginning. A shortstop agent is then automatically injected into the reactor (Ethyl, 1980). Under this particular condition, an alternative source of pressure would be needed to open the valve and inject the shortstop agent if power loss were the cause of agitator failure. Any shortstop system employed would have to be connected to other vessels that might receive the slurry under upset conditions because active initiator may still be present in the slurry even if it is dumped to a blowdown or holding tank. Instrumentation The degree of instrumentation is important in preventing relief valve discharges and varies greatly among plants. Those plants most successful in preventing discharges have several levels of back-up instrumentation. The instrumentation monitors reactor operating para meters (e.g., pressure, temperature) and either warns operators of an emergency condition or takes action automatically. Instruments monitoring the reaction can be tied into a computer system receiving data from instruments on and in the reactor or they can be locally mounted units on each reactor. Selection of operating parameters to be monitored is critical. For example, a temperature sensor mounted on a baffle can become fouled by polymer build-up. The temperature reported by the sensor can lag-behind the actual temperature of reaction. Therefore, pressure sensors may be a better indicator of the actual temperature. In most cases a combination of the two is more reliable. Other instruments, in addition to those monitoring actual reaction conditions, will contribute to the prevention of an upset condition. For example, overcharging a reactor is a common cause of relief valve i 4-29 GENC discharges. A metering system for charging exact amounts of liquid VC and other ingredients in combination with accurate weigh tanks can prevent overcharging and the subsequent hydroful condition. A level control on the reactor that is attached to an alarm system would indi cate overcharging. One company indicated success in using a radioactive source detected by ion chamber sensing to determine the level in the reactor prior to beginning the reaction. Reactors mounted on scales help to prevent overcharging. Agitator seal-water leaking back into reactors could also cause an overcharging condition. This leakage can be prevented with an electrically operated shut-off valve for the sealwater system. Other examples of instrumentation to prevent relief valve discharges will be discussed in more detail in the next section describing some of the preventive systems currently used by the VC industry. Auxiliary Power Supply The effectiveness of the above examples of instrumentation systems as well as other preventive systems is dependent on the availability of power to run these systems. Auxiliary sources of power are necessary to maintain agitation, cooling, and instrumentation in the event of losing the main power source to a plant. No auxiliary power systems currently found in PVC plants are designed to operate the entire plant -- enough power is usually only available to safely shut down the plant by allowing those polymerization reactions in progress to be terminated or finished. Most plants have dual power lines into the plant to provide primary power. The dual lines keep power constant and prevent sudden surges and dips in power or a complete loss of power. Emergency back-up power is usually supplied by diesel-driven generators. Back-up power may also be supplied in the form of an auxiliary source of instrument "air" that would be necessary to open valves to recovery or for the injection of shortstop agents. One plant uses high pressure, precharged nitrogen to provide pressure. All other valves needed for operation during an upset condition for an orderly shut-down are also nitrogen operated (Ledvina, 1980). Auxiliary Venting System An auxiliary venting system could be used to prevent the minor releases usually caused by the hydroful condition. The venting system 4-30 GENC 016548 would be connected to existing recovery systems or control devices. Variations of this type of system are currently used by some PVC plants. The auxiliary venting system is designed for two-phase relief and blowdown flows and can be used for minor events in which removal of a small quantity of material from the reactor will prevent over-pressuring and a more serious condition from developing. This auxiliary venting system could be used for the following conditions: t overcharge of the reactor, presence of noncondensable gases, moderate reaction rate increase, and L t heat transfer reduction. The vent line would be set at a pressure above normal operation pressure, but below the safety relief valve pressure. The vent line would then automatically open at this pressure or it could be activated from the control panel. i A computer program developed for two-phase flow through a relief valve is used in conjunction with the specific process design to select the relief valve opening pressure and size the necessary equipment. The program simulates the rate of pressure rise in the reactor and rate of venting from the reactor. Using this program, the relief valve and header system are sized and the pressure profile in the relief header is determined for the required relief flow rate so that all back pressure limitations at various points along the system are met. The program also determines maximum possible blowdown flow rate for given inlet and outlet pressures and for a given pipe header configuration (Richter, 1978, pp. 145-152). A typical auxiliary relief valve system would consist of >the following: polymerization reactor, relief valve (not open to atmosphere), knock-out drum for liquid/vapor separation, header system connecting relief valve and knock-out drum, blowdown tank, and blowdown header system. The auxiliary vent line would be connected into a knock-out tank to prevent a carryover of liquid or solid. Vapor from the knock-out drum 4-31 L GENC 01654? can be vented to the existing recovery system, a gasholder, or control device; slurry is pumped from the blowdown tank. It is assumed that pressure in the system is atmospheric and back-pressure from the vent to recovery or control is negligible (which may not always be the case (Richter, 1978)). Sizing of all equipment is based on the computer program results that also give a history of conditions in the reactor and determine when the relief valve will close. The estimated tkk cost of an auxiliary venting system for a 38,000 liter (10,000 gallon) reactor system, assuming availability of a blowdown tank, using December, 1979, dollars. These costs are shown in Table 4-8. Table 4-8 ESTIMATED COSTS FOR AN AUXILIARY VENTING SYSTEM Auxiliary venting system Eight 38,000 liter (10,000 gallon) reactor* Knockout tank Pump and motor Piping Instruments, control valves and safety devices Building and site development Total Physical Cost 419,300 15,800 224,500 65,200 95,700 $820,500 *The venting system proposed would accomodate an eight reactor line. Chilled Water System Many plants maintain a separate supply of chilled water other than normal jacket cooling water for upset conditions resulting from reduced heat transfer. The water is usually brine-cooled and temperatures range from 4C (40F) to 10C (50F). In the event of an increase in tempera ture, the chilled water can be pumped into the reactor jacket to slow the reaction. This replacement can be done in approximately 5 minutes. The chilled water can also be injected directly into the batch or into a blowdown tank where the batch might be dumped. Operator Training Programs A staff of qualified operators able to recognize a potential emergency situation and take appropriate measures to prevent the situation will 4-32 GENC 016550 help to minimze relief valve discharges. The different levels of computer control help to eliminate common operator errors and aid the operator in detecting potential problems, but the-computer,does.not.provide the decisiorrmaking capabilities that are only found in experienced operations personnel.! The right combination of operator experience and computer control can help to eliminate relief valve discharges. Operator training programs vary from company to company. Training programs range from several weeks to over a month with year-to-year retraining and refresher programs also offered. Polymer operators are trained to deal with run-away reactions - how to recognize them and what steps are necessary to bring them under control. At least one company has instituted a disciplinary system for those operators responsible for a run-away reaction and the relief valve discharge if it occurs (Laundrie, 1980). Flares A flare can also be used to help control a relief valve discharge. As mentioned in Section 4.1.4, one company uses a flare as an equivalency to a rupture disc in order to prevent fugitive emissions through the relief valve - the relief valve is connected directly to the flare. The flare is designed for relief valve discharges originating from upset process conditions and will accommodate two reactors simultaneously. A knock-out drum is installed between the relief valve and flare to prevent liquid entrainment. The flare has only handled minor discharges but could receive the entire reactor contents (i.e., worst-case condition). This method does not eliminate relief valve discharges, but only helps to minimize the emissions from this source. Safety and cost considera tions (e.g., supplemental fuel needs) must be evaluated before applying this control method to relief valve discharges. Gasholders and Other Containment Methods A gasholder is a cylindrical, variable-volume vessel. The most common type of gasholder is a vessel with a floating roof with either a water seal or a double inner synthetic seal that expands to accommodate the influx of gas. The water-sealed gasholder has a longer life because there is no seal failure, but the water seal is a constant source of VC fugitive emissions and freezing must be prevented in colder climates. 4-33 GENC 016551 The operating principle of a gasholder is based on piston displacement. A frictionless movable piston floats on the confined gas - rising and falling with changes in the volume of stored gas. As gas enters and builds up to the designed operating pressure, the piston rises and floats on the gas. Gasholders are currently being used as part of the recovery system to contain and store VC gas collected from various emission sources in the plant. The gases stored can be fed to the recovery system or the gasholder can serve as a surge vessel feeding the primary control device (incinerators must receive a near constant flow and concentration of combustibles for proper operation). Gasholders can be used to help prevent relief valve discharges without actually receiving or containing the entire reactor charge. Currently there is no plant that has connected a relief valve directly to a gasholder or uses a gasholder only for relief valve discharges. One plant relieves pressure to the gasholder when a batch is out of control (Brumbaugh, 1980). The plant identified several problems that can occur if the gasholder is used for this purpose. As discussed previously, a multi-phased discharge can occur and slurry can carryover to the gasholder. The plant installed a knock-out tank to prevent this carryover, but it does not always handle the relief valve flow. Also, this plant's gasholder is part of the recovery system and capacity is not always available to handle the entire VC charge to a reactor. The auxiliary vent system previously described for minor relief valve discharges would not be as effective in preventing a discharge caused by an accelerated reaction that could result in a major release or release of the entire reactor batch. Prevention of the major relief valve discharge would require two different types of control technology the auxiliary vent system and a containment system such as a gasholder. A gasholder for the purpose of containing a major relief valve discharge would have to be designed to accommodate a worst-case condition for one reactor (the entire reactor charge), and would have to be dedicated to that service only. Simultaneous discharges from many reactors would require many gasholders. The same knock-out tank described for the 4-34 0\6552 I auxiliary venting system would be a part of the gasholder containment system. There would be a separate line from the knock-out tank that would go directly to the existing recovery system or control device for the purpose of controlling minor releases. However, the knock-out tank would have a back pressure valve set to open to the gasholder in the event of a major release. This review study, Conoco Research Division (Ledvina, 1980) and the B.F. Goodrich Chemical Group (Holbrook, 1980b) have evaluated the gasholder potential for containing relief valve discharges only and the following discussion is based on this research. Typical specifications for a gasholder to contain a relief valve L. discharge from a 38,000 liter (10,000 gallon) reactor (assuming 1.4:1.0 water to VC charge ratio) is shown in Table 4-9. Table 4-9 TYPICAL GASHOLDER SPECIFICATIONS i FOR 38,000 LITER (10,000 GALLON) REACTOR Dimension i Volume Diameter Height Seal Specification 5,700 m3 (200,000 ft3) 24 m (75 ft) 15 m (48 ft) water The estimated capital costs for installation of the above gasholder for a 38,000 liter (10,000 gallon) reactor, assuming the use of an existing recovery system for the vapors in the gasholder, based on December, 1979, dollars is indicated in Table 4-10. The estimated total cost does not include the cost of the auxiliary vent system with knockout tank in Table 4-8 which increases the total system cost by approximately $821,000 to $4,802,100, Accuracy of these costs are 30 percent. Insurance, i taxes, recovery credits and other miscellaneous charges are not included. B.F, Goodrich estimated the cost of a similar gasholder system with 14,000 m3 (500,000 ft3 ) capacity to be approximately $3,000,000 4-35 GENC 016553 L Table 4-10. ESTIMATED COST FOR INSTALLATION OF A GASHOLDER Equipment Gasholder Piping Safety Site development Header extension from KO tank Engineering and construction Contingency Operation and maintenance Total System Cost Cost for 38,000 liter (10,000 gallon) reactor iw $ 883,400 382,200 70,400 232,900 600,000 1,070,000 648,000 94.200 $3,981,100 4-36 GENC 016554 (Holbrook, 1979). Their cost was based on earlier dollar values and did not include operation and maintenance costs. Conoco estimated a gasholder system alone without the auxiliary venting system (i.e. , knock-out tank and blowdown tank) to be between $2,000,000 and $4,000,000 for fabrication 5 and installation of a gasholder with synthetic rubber seal and 14,000 m (500,000 ft3) to 42,000 m3 (1,500,000 ft3) capacity. The life of the synthetic rubber seal is estimated at about 2 years and the replacement cost is 5 to 10 percent of the original gasholder cost. Replacement time is 4 to 6 weeks and seal delivery takes 20 to 22 weeks. Thus, a second gasholder would be required to prevent potential emissions from relief valves during downtime for seal replacement. Therefore, the estimated capital cost for installation of a gasholder system for containment of a relief valve discharge will approach $5,000,000. This system would be for containment of one 38,000 liter (10,000 gallon) reactor and does not take into consideration the possibility of multiple reactor relief valve discharges. Experience with gasholders indicates that the following conditions and considerations must be evaluated to ensure trouble-free operation: The gas from the reactor must be clean and free of suspended solids and liquids. Polymer carryover could cause line plugging which can result in backpressure to the reactor. The gas or mixture must be unreactive, non-corrosive, and non-explosive. There must be no condensation in the gasholder system or transfer lines could freeze. Water seals probably represent the best type of seal for trouble-free operation. ` Based on a computer simulation assuming adiabatic cooling conditions for the reactor, transfer of reactor contents at a limited rate that would allow only vapor to be vented would result in a gain of 0.1 to 1.5 minutes until the safety device to the atmosphere would open. Transfer is limited to this rate to avoid slurry (and foam) carryover. The maximum fill rate of a conventional gasholder, which is limited by expansion of the inner seal, will also limit the 4-37 GENC 016555 venting rate. The maximum piston velocity for a rubber-sealed gasholder is approximately 4.6 m (15 ft) per minute and would require about 6 to 8 minutes to fill - this seal expansion rate thus may limit the venting rate of the reactor contents. A venting system to the atmosphere must be present on the gasholder should the gasholder not be able to accomodate the discharge. 'i "3 For a gasholder with a capacity of 14,000 m (500,000 ft ) to 42,000 m3 (1,500,000 ft3) a horizontal knock out vessel with a capacity of 8,000 m3 (300,000 ft3) to 14,000 m3 (500,000 ft3) would be required. The length of the vessel would be 46 m (150 ft) to 54 m (175 ft). The pressure drop in the line from the gasholder to the knock out drum can not be controlled, therefore the K0 drum pressure must be atmospheric. The potential for back-pressure on the reactor relief valve must be evaluated. Conventional relief valves can take up to 10 percent of the set pressure before it reseats. A balancedbellows relief valve can take 40 to 50 percent of the set pressure. A pitot-operated relief valve can be opened on a differential pressure basis, but because of the small pitot tube diameter, this type is only intended for clean service. Relief valves to the atmosphere are designed to discharge directly to the atmosphere such that the explosion risk caused by VC flammability levels of 3 to 30 percent and worker exposure are minimized. Failure of the gasholder during an upset condition could result in a vapor cloud at low level because of the high concentration of VC at low pressure in tlhe gasholder and low velocity of the vapor. The above gasholder containment system could be designed into a PVC plant. Techniques such as elaborate monitoring and alarm systems, freeze protection, inert gas purging systems, strict personnel training programs, and inspection routines would be necessary to minimize some of the safety hazards. There are other systems that minimize relief valve discharges, including the use of a gasholder to help prevent these discharges. However, no matter what system or combination of systems is 4-38 GENC 016554 utilized, a safety relief device directly to the atmosphere will always be required. Another containment device that can be used to help prevent relief valve discharges is a blowdown or spare slurry-holding vessel. These vessels are usually used for stripping the batch of residual VC or for blending with other batches. A blowdown or holding tank is a constant volume, nonpressurized vessel (some may be pressurized) that would allow an influx of gas until the pressure of the tank and the source of pressure are equalized. The tank can also take the slurry directly and have chilled water and short-stop agent waiting for control of a run-away reaction. Discharge of the slurry to this tank can be manual or autoactivated. The timing of gas venting, slurry dump, or both, is critical in order to prevent the relief valve on this holding vessel from opening. These holding vessels can be used in combination with a gasholder or the recovery system to successfully control an upset condition--the reactor vapors are vented to the gasholder or recovery while the polymer slurry is dumped to the blowdown tank, or the entire reactor contents are dumped to the blowdown tank and pressure relieved from this tank to the gasholder or recovery system. Pressurized containment can also be used to help prevent a relief valve discharge. B.F. Goodrich replaced their gasholders with this type of pressurized containment which they refer to as a "burp" tank. The purpose of the burp tank is to prevent emissions to the atmosphere as part of the VC recovery system, but the tank can also be used to assist in preventing relief valve discharges (Holbrook, 1980a). This automatic pressure reduction system usually will open at a pressure between operating pressure and the pressure relief valve setting. The vapors from this system go to recovery or the primary control device. Other Methods for Preventing Relief Valve Discharges Prior to charging a new batch, pressure tests can be run on the reactor. T,hese tests will assure that rupture discs will not fail prematurely. One plant also monitors all plant devices (e.g., valves, pumps, flowmeters, instruments, controllers) before charging the reactor to verify their successful operation and ensure safe operation during the polymerization reaction (Holbrook, 1980). 4-39 G0AC Another newer preventive method is a multiple initiator system. This system maintains a high initiator rate early in the reaction cycle and a slower initiator rate during that part of the cycle when auto acceleration is expected to occur. The system thus maintains a near constant temperature and pressure during the reaction. 4.2.3.3.2 Preventive Systems Currently in Use. Several systems for prevention of relief valve discharges have been installed by PVC plants. These plants all use a combination of equipment and work practice procedures to prevent or minimize relief valve discharges to the atmosphere. Several of these systems are discussed in the following subsections. B. F. Goodrich Company-- B. F. Goodrich produces PVC resins by the suspension, dispersion, and bulk polymerization processes. The new, larger reactors as well as the older, smaller reactors are utilized in their processes. B. F. Goodrich used gasholders extensively as part of their recovery system, but have since replaced the gasholders with pressurized burp tanks (that provide direct recovery) with high volume liquid ring compressors. Following is a discussion of their relief valve discharge prevention systems (Holbrook, 1980c). B. F. Goodrich has reduced releases from reactor safety relief valves in their large reactor suspension PVC systems. This was accom plished through process .control with emphasis on early detection and analysis of abnormal conditions. A totally computer-controlled system was installed to detect these abnormal conditions and take the appropriate action with a minimum of operator involvement. A thermal model was developed to simulate the typical controlled polymerization reaction and a kinetic model was developed to compare the typical reaction with an actual reaction. When the computer recognizes a potential upset condition, preventive control methods can be implemented and the upset conditions are then brought under control. B. F. Goodrich identified three conditions that can result in a relief valve discharge from the large reactors - equipment failure, hydroful conditions, and an accelerated polymerization reaction. Equip-^ ment failure is prevented by an emergency generator, back-up instru mentation, and preventive maintenance follov The 4-40 GENC 016558 other conditions are minimized by computer control, but in the event of a high pressure situation in the reactor preventive steps are taken immediately. The large suspension reactor system is followed by a larger blowdown or flash tank which is generally 1.5 times as large as the reactor. Normally, the purpose of the blowdown tank is to receive the polymer slurry before the process is complete so the reactor can be prepared for the next charge. The blowdown tank has both an agitation and a shortstop system and thus can also receive the slurry under upset conditions so that an accelerated reaction can be brought under control. A recovery separator tank, preceding the recovery system, knocks out entrained liquid that would be vented during an abnormal condition. To relieve reactor pressure, if a hydroful condition exists, the computer immediately stops the addition of reaction ingredients. If the pressure is not due to a hydroful condition, but is due to a runaway reaction, the following control functions are triggered: Full cooling to the reactor and blowdown tank. Addition of reaction shortstop. Recovery of VC from the reactor. Opening of exhaust vents to a recovery separator to relieve pressure. Pump-out of reactor into a blowdown tank. In addition to the above process controls, all plant devices (e.g., valves, pumps, flowmeters, instruments, and controllers) are monitored before the reactor is charged to verify their successful operation and ensure safe operation during the polymerization reaction. The computer control system information display has been centralized and the following back-up systems have been installed: Dual power lines into the plant. Emergency power generation to maintain cooling and agitation should a primary power failure occur. Emergency power supply for computer and instruments. Manual control of instruments should a power failure occur. Instrumentation backup for computer. Installation of redundant equipment (pumps, compressors, and flowmeters). 4-41 GENC 01655 The above combination of control technologies for relief valve discharge prevention has resulted in no reactor relief valve releases for 31,000 charges at their large reactor facilities. B. F. Goodrich has also developed an effective relief valve discharge prevention program for their smaller reactors using many of the preven tive measures described earlier. Similar to the large reactor systems, there are three conditions that B. F. Goodrich feels can result in releases from the small reactor suspension and dispersion processes - equipment failure, hydroful conditions and accelerated polymerization. Equipment failure is prevented by an emergency generator, backup instrumentation, and preventive maintenance that follows a regular schedule. A hydroful condition is controlled in a manner similar to the large reactor system described above. When the computer detects excessive pressure in the reactor caused by a hydroful condition, all valves are closed to the reactor except cooling water, charging is stopped, the slurry is sent to a blowdown tank, and vents are opened to a separator recovery tank (knockout tank) to relieve reactor pressure. The accelerated polymerization reaction (high reactor pressure and temperature) prevention follows the same initial preventive steps for the hydroful condition. In addition, water is injected directly into the reactor to cool the contents. The reactor is then vented to the prevent or burp tank (pressurized containment vessel) to relieve pressure, and a shortstop agent is injected to kill the reaction. Conoco Chemical Division-- Conoco has done research at their pilot facility to develop a prevention for relief valve discharges. An elaborate kill system is the primary method for preventing a discharge tb the atmosphere. Tests were conducted in Conoco's large reactor pilot plant to determine the effectiveness of their kill system in stopping the polymerization reaction and the resulting pressure rise in a PVC reactor (McCulley, 1980). The tests were planned to simulate the concurrent loss of reactor agitation and cooling that would occur during a power failure or a worst-case situation. In each of the test runs, the simulated power failure was started 3 hours into the polymerization reaction to ensure that the reaction rate was near its maximum. Also, the reactor 4-42 GENC 016560 pressure was allowed to increase from its normal pressure of 118 to 119 psig up to 140 psig before injecting the killing agent. The 7 to 11 minutes required for this pressure increase provided sufficient time for the swirling inside the reactor to stop; this minimized mixing of the kill agent with the reaction mass. The kill agent was pressured through a nozzle (no spray nozzle or other distribution device was used) into the vapor space of the reactor. The investigators suspected that the kill agent simply ran down the sides of the reactor into the slurry and did not benefit from injection- caused mixing. The liquid kill agent used is soluble in the liquid VC. Results showed that after injection of the killing agent at 140 psig, reactor pressure continued to increase to a maximum of 150 psig over approximately 8 minutes (about 20 minutes into test). The reactor pressure then slowly decreased to 146.5 psig 70 minutes into the test run - the gradual pressure drop probably was due to heat losses from the reactor to the atmosphere. These tests confirmed that, if sufficient killing agent is injected, the polymerization reaction can be stopped and the rising reactor pressure can be controlled during a major power failure or other condition that results in total loss of cooling and/or agitation. Early detection of an upset condition and control equipment redundancy are the most important requisites to preventing an emergency situation from reaching the kill system stage of control. Conoco feels there are four major causes of a relief valve discharge: operator error, general equipment malfunction, events external to the system, and ' specific failure of the kill system. In order for any of the first three events to result in a release, the fourth event, failure of the kill system, must occur. Specific examples of operator error and equipment malfunction (overcharging reactor, cooling system failure, etc.) have already been discussed. Examples of external events are power failures, fires, or some natural disaster. 4-43 GENC 016561 The early detection program includes the following: Two sets of reactor temperature and pressure indicators on the control panel (from separate transmitters) are supplemented by pressure gauges on each reactor. (Use of temperature indicators represents a level of redundancy because an increase in VC vapor pressure will be accompanied by a rise in temperature.) Operators are always in 2-way radio contact with control panel personnel. t High pressure and high temperature alarms sound in the control room and, if acknowledged, turn off. If the pressure continues to rise, a second alarm, set at a higher pressure, sounds and cannot be shut off until the reactor pressure drops below its set point. This alarm also sets off a siren which can be heard throughout the plant. A temperature control instrument on the control panel regulates the water flow to a reactor. If this flow is improperly regulated, a panel switch can be used to override the controller, sending maximum cooling water flow to the reactor. The panel has reactor amperage indicators with alarms. A low amperage reading is used to confirm that the reactor has emptied after stripping to avoid overfilling the reactor on the next charge. Low amperage can result from a failure in the agitation system. There is a back-up power system for the control panel. Cooling water for reactors. Provisions are made to supply water to the reactor from cross-ties with city water or from the fire water pumps (diesel-driven). Each of these capabilities is checked once each week. Also, instrumentation prevents this backup cooling water flow from going to users other than the reactors. A severe-weather radio provides current weather conditions that may affect the plant (e.g., storm conditions that could cause a total power failure to the plant). A reactor can be vented to an empty vessel or the recovery system to reduce pressure. If the above early-detection prevention methods and redundant equipment are not adequate to bring the reaction under control, then the kill system is implemented. Two kill systems are maintained. The first kill system 4-44 GENC 016562 uses high pressure water injection. It is a completely manual system and is used to control reactor pressure increases during normal plant operation. It cannot be used in the event of a power failure. If agitation is lost due to equipment failure or loss of power, the second kill system can be used. This second system uses high pressure nitrogen injection (precharged), and is a remotely operated system backed up by a manual injection line. In this situation, valves are also nitrogen operated. This kill agent is effective without mixing. The system has been tested in Conoco's pilot plant and has proven to be effective under worst-case conditions including loss of reactor cooling or agitation. The kill systems for each reactor can be activated from the control board or locally at the reactor. The system from one reactor can be used on a different reactor, and each system contains enough shortstop agent for two complete kills in one reactors. Two racks of back-up nitrogen bottles are available for instrumentation use. To contain a plant fire, which could result in an an emergency discharge, a fixed-spray fire fighting system was installed throughout the VC areas of the plant. Hydrocarbon (HC) sensors detect explosive concentrations with an alarm set at 20 percent of the lower explosive limit (LEL) and watering systems are triggered at 40 percent of the LEU. These HC detectors are independent of the fugitive emission detection system required by the current regulation. Fire sensors throughout the areas where VC is handled monitor on a rate-of-temperature-rise basis. Back-up cooling is used instead of a back-up power system. The rationale is that if the emergency situation is due to a broken agitator shaft, motor, pump, etc., back-up power will not alleviate the `situation. The second kill system described above always accommodates these types of situations. Both of Conoco's plants use an extensive interlock system to prevent mistakes that could release VCM to the atmosphere, damage equipment, or otherwise seriously effect the plant operation. For example, there is an interlock which prevents the reactor dump valve from opening if the reactor is under pressure, thereby avoiding the accidental dumping of a reactor containing a large amount of VCM. The Aberdeen plant uses 4-45 GENC 016563 programmable logic controllers for its interlock control system. The Oklahoma City plant uses a hard-wired relay logic system. Hooker Chemical Company--As discussed previously, the bulk polymerization process presents some disadvantages for prevention of a relief valve discharge. The process is anhydrous resulting in poor heat transfer and agitation is slow which makes shortstop agents difficult to disperse throughout the dry slurry. Pre-Po reactor control is easier because the reaction only goes to 8 to 12 percent conversion and if a high pressure condition develops, the slurry is dropped to an empty, larger Po-Po reactor where the initiator is used up and the reaction brought under control. Also, the liquid VC charged to a Pre-Po provides some heat transfer and better temperature monitoring. Therefore, Po-Po reactor control is more important in the bulk process. Hooker Chemical Company prevents relief valve discharges through a combination of equipment installations and work practice procedures (Dubec, 1980). Pre-Po and Po-Po reactors are mounted on scales that provide back up control for overcharging of reactors. In order to better monitor temperature in the Po-Po, reflux condensers were retrofitted to the reactor so exothermic heat of reaction can be removed by the circulation of water through the condensers. If upset conditions develop in the Po-Po reactor, the following manual procedure is followed: The Po-Po reactors are equipped with pressure alarms that activate when the reaction pressure increases above normal operating pressure, thus alerting operators of the upset condition. t The reflux condensers are flooded with water to control the temperature. The Po-Po reactor undergoes degassing. Valves are opened back to the recovery system which is a brine-cooled system. Cooling water at 10C (50F) from chilled tanks (also a brine-cooled system) is pumped into the reactor jacket to slow the reaction. At this point if the above steps have not brought the reaction under control and time is still available, pressure will be manually released to keep the rupture disc from blowing. 4-46 GENC 016564 When the above steps are not effective, then the rupture disc will eventually rupture and the reactor will discharge to the atmosphere. As mentioned previously, the critical time for the bulk process is the first half of the reaction. Beyond this midway point, the unreacted VC has been substantially reduced and the reaction is more easily controlled. This particular plant does not use a shortstop agent in the procedure, but such an agent may be available in the future. Also, two power lines have been run into the plant to prevent a total loss of power. In the event that total power is lost to the plant, two emergency generators provide the power necessary to control the reactions in progress and safely bring the plant down in about 12 hours without any relief valve discharges. General Tire Chemical Pivision--General Tire has older, small reactor technology for the suspension process and uses a gasholder with a capacity of 1,400 m3 (50,000 ft3 ) to help prevent relief valve dis charges from the suspension reactors (Laundrie, 1980). The main purpose of the gasholder is for venting compressor relief valves, pumps, transfer lines, weigh scales, condensers and knock-out tanks. Emissions from reactor opening and the slurry and wastewater strippers are also vented to the gasholder. VC collected in the gasholder is recovered and returned to the process. When an upset situation is detected in a reactor, the following procedure is followed: Each reactor is equipped with a high pressure alarm which signals a potential upset situation. If possible, the reaction is vented to the gasholder to relieve pressure. A K0 tank prevents entrained slurry from reaching the gasholder, but this is not always possible. A chemical shortstop agent is manually added to the reactor. t The batch is dropped Into a stripper vessel or pressure is equalized to an empty reactor. Cooling water pumps are on separate electrical circuits to prevent a loss of cooling water during power supply outages. A back-up water supply Is available if primary cooling water is lost. 4-47 GENC 016565 Electrical power is supplied to the plant by two separate feeders. The system automatically switches to the other feeder if power fails on one feeder. If the above procedure is not successful, then the relfe-f valve is manually vented to the atmosphere. Because this is an older plant with less sophisticated instrumentation, more responsibility for prevention of relief valve discharges falls upon the operators. Polymer operators have been trained to deal with runaway reactions - how to recognize them and what steps are necessary to bring them under control. 4.2.4 Non-Reactor Relief Valve Discharges. Safety relief devices are found on all pressurized equipment in EDC/VC plants and PVC plants. These devices may be rupture discs, single or in series, or combination in-line rupture disc and relief valve. In most cases non-reactor relief valves discharge less than reactor relief valves. During the original standard-support study (EPA, 1975), emphasis for control was placed on reactor safety relief valves. However, Regional offices indicate that the non-reactor releases are equally Important. Tables 4-6 and 4-7 shows the frequency and quantity of VC discharged through non-reactor relief valves. The following section describes some of the typical causes of non-reactor relief valve discharges and some of the actions taken to prevent their occurrence. Some of the discussion pertains to reactor relief valve discharges where common equipment is utilized. Failure of rupture discs-- The premature failure of rupture discs alone, in series or in combination with relief valves is the most common cause of discharges. Section 61.65(b)(4) requires a rupture disc to be installed between equipment and relief valves to prevent fugitive emis sions from these relief valves. Rupture discs were installed for this purpose and failure of these discs was responsible for the majority of the non-reactor (as well as reactor) relief valve discharges reported initially after the waiver period. Some of the causes (and corrective measures taken) for rupture disk failure include: Overrated pressure capacity. Many rupture disc manufacturers rated their discs at a specific pressure, but failure sometimes occurred at a much lower pressure. This problem was usually 4-48 SENC 016566 solved by a reliable quality-assurance program that guarantees the rated capacities on the discs. Compatibility with process design. There are several types of rupture discs commercially available, but not all are compatible with a specific process design. For example, one bulk plant installed rupture discs on all pressure vessels and they continually failed below their rated capacity even though testing showed the disc capable of handling that capacity. After experimentation with several different types, reverse buckling discs were found to be successful, but only after the process piping was replaced. Another plant could not use knife-head discs because, with their process, these discs were susceptible to cracking at the welds. Finding the compatible rupture disc sometimes requires trial-and-error. Corrosion of rupture discs. Failure of many discs is caused by corrosion from process material coming into contact with the disc. Corrosion caused by plant location is also a problem. Plants located near ocean shorelines experience corrosion problems caused by the salt-water mist. Corrosion problems are usually solved by finding the right disc material such as nickel or stainless steel, or by coating the disc with a material such as teflon. Disc leakage. Leaks are caused by pin-holes and irregularities already present in the disc, or result from improper handling by plant personnel. The first cause is a function of cost. The more a plant is willing to pay for a disc, the more the manufacturer will test and guarantee their disc. A1though even with the most expensive discs, there is always the possi bility of a poor quality disc in a batch. Most plants maintain a rigorous inspection program for discs prior to installation. A potential stress problem that results in a leak can also be detected by gauging the space between the rupture disc and relief valve for pressure - many plants currently follow this practice even though it is not required by the regulation. The other cause of failure, poor handling, is minimized by instituting good maintenance and installation procedures. 4-49 6ENC 01656? Failure of rupture discs from most of the above causes can be reduced by following a good quality control program. Many plants have found it necessary to bring in a representative from the rupture disc manufacturer to train plant personnel in proper handling and installation procedures. In most cases this has eliminated failure of rupture discs. An example of Conoco's testing/maintenance program includes the following (Ledvina, 1980): The rupture disc is assembled in the shop between two flanges prior to installation below the relief valve on a reactor or other piece of equipment. This allows the disc to be pre-torqued in a safety head device and reduces handing during installation. After installation and before the equipment is put in service, discs are hydrostatically tested to insure reliability. The discs are tested up to 90 percent of burst pressure after installation. (Reactors are pressurized above normal operating pressure before charging). Upon receipt of a disc shipment a percentage of each lot is randomly selected and pressure checked for quality and actual pressure. The lot is rejected or accepted by this procedure. Discs are changed every six months and those removed are tested to verify actual rupture pressure. Relief valves are replaced every twelve months. Emergency procedures involving non-reactor and reactor discharges are updated once every year. t The rupture disc manufacturer's representative visits the plant each year for retraining purposes. Failure of level controls-- Slip gauges, which have a probe that moves through the gas/liquid interface in storage or transfer vessels indicating the level in the vessel by the physical state of the material discharged, have been replaced with more sophisticated level controllers. The new level controllers are used on charge tanks, storage spheres and tanks, rail cars, tank cars and marine transport vessels, all of which are under pressure and have some type of safety relief valve. However, the new level controllers are not always reliable and inaccurate readings on these controllers can cause an overpressurized condition during the filling operation resulting in a relief valve discharge. 4-50 GENC 016568 Many times these discharges are significant. One plant released approximately 17,200 kg (38,000 pounds) of liquid VC from a storage sphere during barge unloading (Battye, 1978). This was a combination operator error and level control failure. Most of the discharges are much lower and usually involve operator error caused by inattentiveness during the filling operation. High pressure in transfer lines and equipment-- An overpressurization can develop in the transfer lines during transfer of VC from storage areas to process units. The pressure surge (hydraulic hammer) may be caused by closing a valve too quickly and can be prevented by manually closing the valve slowly or instrumenting the valve to close slowly. This condition can also be prevented by installing an emergency high-pressure trip-switch or using a small, in-line surge vessel (Brittain, 1980a). The high-pressure-sensitive switches can be used to prevent relief valve discharges from other equipment such as recovery compressors. The compressor automatically shuts off when overpressured, a condition usually caused by clogging of the compressor lines (Moulthrop, 1980). Other causes of non-reactor discharges-- Cold weather can be a cause for non-reactor discharges. For example, one plant had an automatic valve (to the VC reclaim vent) freeze shut causing the secondary decant tank to overpressurize which caused a rupture disc to burst (Battye, 1978). 4.3 RESIN STRIPPING 4.3.1 Introduction Because one of the greatest sources of emissions is from those points downstream of the resin stripper, the effectiveness of the polymer stripping process is of prime importance in controlling emissions of RVC lost to the atmosphere. The release of unreacted VC from the polymer resin in the stripper is a function of time, temperature, and pressure of the stripping process, in addition to particle size distribution and porosity of the resin. Prior to the current VC standard, stripping was done by PVC manu facturers for economic reasons; i.e., recovery of VC for reuse in the process. Reduction of RVC concentration in the resin to meet regulated 4-51 genc 016569 emission levels now represents a primary reason for stripping to lower levels. In PVC plants using stripping technology to control vinyl chloride emissions, the daily weighted average of the residual VC (RVC) concen tration in the stripped resin must meet the following limits as required by Section 61.64(e)(1): 2,000 ppm for dispersion resins (excluding latex). 400 ppm for all other resins (including latex) averaged separately for each type of resin. Included in this category are suspension, bulk and solution resins. The prescribed emission levels are assigned without regard to grades of the resin types although different grades of resin have unique characteristics with respect to stripping efficiencies. Determination of the RVC concentration is to be made by sampling immediately after the stripping process and using a prescribed method (EPA Method 107). The quantity of materials processed by each stripper is determined on a dry solids basis. If batch stripping is used, one representative sample of PVC resin is taken from each batch of each grade resin. If continuous stripping is used, one representative sample of PVC resin is taken for each grade of resin processed or at 8 hour intervals. Results of the RVC analyses are submitted in the semi-annual report required from each processor regulated by the standard. Variables that affect stripping levels are: batch vs. continuous stripping, homopolymer vs. copolymer resins, and reactor vs. non-reactor stripping. Each of these determines resin stripping methodologies. Lower stripping levels, required by the standard, sometimes result in the resin being excessively exposed to high temperatures that adversely affect the resin's heat history. This heat history is a critical parameter for the fabricator. Table 4-11 shows comparative stripping levels reported by PVC plants producing various resin types and using different stripping methods. 4.3.2 Suspension Resin Stripping Of the 40 operating PVC plants in the United States, 32 of these use the suspension polymerization process. (See Section 3.3.4 for a 4-52 GENC 016570 Table 4-11. PERCENT DISTRIBUTION OF STRIPPING LEVELS BEING ACHIEVED BY INDUSTRY** Suspension Concentration of residual vinyl chloride In ppmb Range of dally average (ppm) Plant Method1" High Low 400 350 300 250 200 150 100 50 1 c 123 0.5 100 100 100 100 100 100 99.3 97.3 2 b 1,629 0.4 98.8 98.8 98.8 98.3 98.3 96.5 95.9 89.0 3 c 388 3 100 99.4 98.9 97.7 96.0 89.3 80.2 42.9 4 c 303 19 100 100 99.4 97.6 93.9 89.0 72.6 27.4 5 b 576 3 96.8 96.8 95.5 94.2 91.6 80.6 54.8 30.3 6 b 634 5 98.3 95.7 89.7 84.5 75.9 61.2 45.7 19.0 7 b 631 8 98.4 95.9 87.7 81.1 68.0 50.8 29.5 11.5 8 c 382 16 100 98.3 94.5 89.5 72.9 47.5 24.9 3.3 9 c 1,214 4 81.8 75.5 67.8 58.7 46.2 40.6 35.0 30.1 10 c 1,385 22 82.5 74.6 65.8 58.8 50.0 35.1 17.5 1.8 - FTtl 11 541 70 92.7 90.2 78.0 63.4 41.5 26.8 14.6 0 12 b 514 51 93.2 85.7 72.9 59.4 37.6 21.8 9.8 0.8 13 b 684 63 95.0 90.8 80.9 63.8 41.1 20.6 6.4 0 --L Orl A 14 b 1,099 71 93.8 86.9 81.8 68.7 51.1 15.9 2.8 0 15 1,056 96 75.9 60.3 46.6 37.9 20.7 6.9 0 0 Bulk 1 b 166 12 100 100 100 100 100 99.4 97.5 78.3 2 b 293 3 100 100 100 99.3 99.3 97.9 95.2 84.9 3 b 672 56 84.3 77.7 58.1 46.4 25.1 8.9 1.1 0 Latex 1 2 b b 15 0.2 100 100 100 100 100 100 100 100 310 8 100 100 99.4 99.4 98.8 96.9 74.5 24.2 aBased on EPA semi-annual reports for March through September 1980 obtained from Regional EPA offices. Data represents approximately 501 of suspension. 751 of bulk, and 331 of latex plants. ^Individual data are percentages of time that concentration falls below specified levels. Values represent dally averages weighted on a production basis. cMethod: b batch; c continuous (continued) 4-53 GENC 016571 Table 4-11 (Concluded) Dispersion Concentration of residual vinyl chloride in ppm^ Range of daily average (ppm) Plant Method High Low 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 too 1 2 3 6 tv-SX. --? b 1,245 15 100 100 100 100 100 100 100 100 99.4 98.8 98.8 98.8 98.2 97.6 97.6 95.8 89.3 75.0 38.1 9.5 c 1,949 132 100 99.3 93.7 90.9 84.6 79.0 71.3 61.5 54.5 42.7 39.9 32.2 20.3 19.5 14.0 9.1 6.3 2.8 1.4 0 b 2.364 418 99.4 96.6 92.2 87.2 82.7 70.9 62.0 50.8 43.0 33.5 27.4 21.2 11.2 8.4 3.9 1.7 0 0 0 0 b 5,515 49 85.6 83.5 81.4 79.4 76.3 73.7 73.2 68.6 65.0 61.3 55.2 51.5 45.9 36.6 28.4 18.6 9.3 5.1 2.1 0.5 b 14,092 297 80.3 78.0 76.4 76.4 74.0 72.4 71.7 68.5 66.1 59.1 51.2 37.8 23.6 15.0 7.9 3.1 1.6 0.8 0 0 b 4.129 89 83.8 80.2 78.4 74.3 71.3 67.1 64.1 56.9 62.7 45.5 39.5 31.1 24.0 16.2 10.2 7.8 4.2 2.4 0.6 0 c 6,065 644 70.7 62.6 58.5 55.3 45.5 37.4 26.0 22.8 14.6 10.6 8.9 7.3 4.1 0.8 0 0 0 0 0 0 _ 4a, 'Based on EPA semi-annual reports for March through September 1980 obtained from Regional EPA offices. Data represents approximately 40X of ^ dispersion plants. -t* k Individual data are percentages of time that concentration falls below specified levels. Values represent daily averages weighted on a production basis. cMethod: b > batch; c continuous GENC 016572 !--. w sa r-- r-- iwi process description.) These resins are homopolymers and copolymers and are stripped batchwise or continuously, in the reactor or in a separate vessel. Suspension resins have many grades with variable heat and shear-stress tolerances. One continuous stripping system for suspension resins has been developed by B. F. Goodrich and is widely used throughout the industry. In this process a pressurized stripping column feed tank, is used to release excess VC. This tank forms a transition from the batch reactor to the continuous, multi-stage stripping column. The hot slurry is pumped continuously from the feed tank to the stripping column through a vapor liquid separator. The liquid slurry enters the top of a counter- current steam stripping column and the vapor streams from the feed tank separator and column are sent to the VC recovery system. The system can be designed to handle porous or non-porous slurry feed with RVC content ranging from 5,000 to 200,000 ppm. This process uses high temperatures 80 - 90C (180 - 190F) and a short residence time (Varner, 1980). Levels attained in this process are often less than 10 ppm RVC in the stripped slurry. Disadvantages of continuous column stripping are (Fannin, 1981): Decreased scheduling flexibility. .Processors .requiring frequent- ,, ................ ......... '`n--i,ii>MinftawLl3agy --^'changes in batch sizes and/or recipe modifications cannot use j column stripping efficiently. j Difficulty in cleaning column. Due to the column's shape, internal cleaning is not as convenient as in an open vessel. Increased potential for burned particles. Crevices within the column, resulting from the contact of trays and relatively greater number of joints, tend to retain resin particles. This results in a long exposure to heat and consequent burned particles. e Mechanical Stress. Because of the pumping system used in continuous stripping columns, mechanical stress is imposed on the slurry. For resins with a high sensitivity to this stress, the particles can agglomerate resulting in poor uniformity and an increased tendency for those particles to be retained and burned. An extreme of this situation can be equipment fouling (i.e., plugging of filters, screens, and orifices). 4-55 GENC 0165 Complex process control. Continuous stripping columns must be fed at a rate controlled to maintain certain target temperatures and pressures. There are more variables to be controlled simultaneously than in an open vessel. Continuous steam stripping represents the most widely used method for suspension resins. Many of the processors using this technology are achieving less than 400 ppm stripping levels ranging as low as 25 ppm RVC (Pucci, 1980). One plant using a steam stripping method, but stripping batchwise, reports typical daily averages of 200 - 250 ppm RVC (Laundrie, 1980). Reportedly, this plant is unable to use continuous stripping due to the many (17) different grades of resins produced. Generally speaking, those plants attaining the lowest RVC levels have chosen to produce one or two resin grades with heat and shear stresses compatible with their continuous stripping technology. In some cases "specialty" has been sacrificed for increased production capacity. Firestone Corporation continuously strips suspension resins to levels less than 100 ppm RVC using their own proprietary technology. This system uses less steam than other known counter-current continuous strippers and may be marketed in the future (Schaul, 1980). 4.3.3 Emulsion Resin Stripping About half (21) of the PVC plants have dispersion polymerization facilities. Of these, 6 are latex resin processes. Dispersion resins are more sensitive to heat and shear stress than are suspension resins. Latexes are susceptible to these stresses and when exposed to shear the emulsion can degrade into an unstable latex with a poor heat transfer property. (See Section 3.3.5.1 for Emulsion Resin Stripping Process description.) ` Dispersion resin slurry is usually either vacuum stripped in the reactor or transferred to a separate vessel (blowdown tank) where it is steam sparged. Due to their thermal instability, dispersion resins are most often batch stripped under vacuum. Inert-gas sparging is sometimes used instead of steam when dilution of the emulsion has to be avoided. At least one manufacturer (B. F. Goodrich) does use a continuous stripping procedure for its dispersion resins. The process was developed 4-56 GENC 016574 by the company specifically for one of its dispersion facilities (Holbrook, 1980c). Daily averages of stripping levels for dispersion resins have been reported as low as 15 ppm for one plant (Gilmore, 1980). Typical RVC levels for dispersion resins in comparison to other resins are shown in Table 4-9. 4.3.4 Bulk (Mass) Resin Stripping Four of the forty PVC manufacturing plants produce bulk resins. Characteristics of bulk resins resemble those of suspension resins but the bulk resin beads are more uniform in porosity and size. These characteristics enhance stripping efficiency. Three of the bulk poly merization plants surveyed strip in the reactor using steam stripping technologies. In bulk resin stripping processes, steam must be injected under a vacuum to avoid contaminating the process with water because bulk polymerization is a dry process. One plant reports achieving daily emission levels of less than 150 ppm RVC and another, less than 250 ppm RVC. Because the bulk polymerization process is anhydrous there are no dryer emissions. However, those emission downstream of the stripping process are still governed by the level of RVC removal during stripping. 4.3.5 Solution (Solvent) Resin Stripping Only one plant produces PVC by the solution process. This plant operates a process for the copolymerization of vinyl chloride with vinyl acetate and other comonomers. The solvent process is unique, from the standpoint of stripping procedures, in that no particulate resin form exists and thus stripping can be accomplished by distillation. The efficiency of the distillation process results in average levels of 10 ppm RVC or less on a consistent basis. The stripper still operates on a conventional distillation principle using acetone and acetone vapors. The column is fitted with perforated trays designed to accommodate the viscous resin solution. This company is seeking approval for a less rigorous (and less costly) sampling/analysis plan to assure compliance (Erdman, 1980). 4.3.6 Other Stripping Technologies A short-residence-time device, adaptable to continuous processing, is a proprietary thin-film evaporator. This device is reported to 4-57 GEHC 016575 exhibit diffusivities of 1,000 to 10,000 times greater than those for simple molecular diffusion. In this system, heat is transferred through a metallic wall to a thin film of liquid. A mechanical agitator dis tributes the liquid evenly over the heat-transfer surface. Disadvantages of this device are: design criteria limits heat transfer area of each unit, more maintenance is required than for a non-mechanical device, and foaming may occur. Another device, the Parkson stripper, is a non-mechanical, plate-type, dispersed-flow contactor. The latex is dispersed into a controlled, high velocity stream of steam and the resulting two phase flow passes turbulently through a plate-type stripper. The stripped latex then discharges into a cyclone separator, usually operated under vacuum, where the latex is disengaged from the vapor. Multistage units may be used to achieve desired residual monomer level. This unit is applicable to foamy, heat-sensitive or viscous products. Its advantages are: complete absence of foam, lack of moving parts, low holdup, low residence time, and reduced surface fouling. The main disadvantage is that the plate design permits only a limited capacity range. One company's latex emulsion PVC facilities do not strip the resin but instead uses a proprietary post-polymerizer. This involves the use of a catalyst to react left-over VC. This process yields RVC levels less than 10 ppm, which is a level mandated by the company's internal standard for their particular product (Smith, 1980). 4.4 FUGITIVE EMISSIONS 4.4.1 Introduction Fugitive emissions represent one of the most difficult emission sources in EDC/VC and PVC plants to quantify and control. Most of the original fugitive emissions estimates were calculated by a mass balance. For EDC/VC plants a mass balance calculation is less accurate because .several gas streams into and out of the plant are not measured accurately (e.g., gaseous chlorine and ethylene feedstocks). PVC plant liquid and solid streams are somewhat easier to measure and fugitive emission estimates are more accurate. The accuracy of fugitive emissions estimates is also affected by whether the plant is open or enclosed, and new or old. An enclosed plant's fugitive emissions can be determined by using 4-58 GENC 016576 roof monitors and ventilation flow rates through the building. Fugitive emissions from an open plant are more difficult to measure because of the variable wind patterns causing dispersion. The newer sources have larger and fewer reactors and fewer piping connections which reduces fugitive emissions. These variables must be considered when assessing fugitive emissions losses from EDC/VC and PVC plants. Based on a recent study done by B. F. Goodrich (Holbrook, 1980a), fugitive emissions from their enclosed existing small reactor suspension and dispersion processes were determined (by actual measurement) to be c'b^0.36 kg/100 kg (0.36 lb/100 lb) of PVC produced or approximately 25 per cent of the EPA estimated 1975 ra.teo.vzj-N'ew large reactor suspension processes were determined to be OrOT-kg/lOO kg (0.07 lb/100 lb) of PVC produced or only 5 percent of the EPA estimated 1975 rate. This study was conducted following the implementation of controls required by the current regulation and reflects the effect of the regulation and new process technology on fugitive emissions reduction. Sources of fugitive emissions in EDC/VC and PVC plants include the following common equipment and operations: Pump, compressor and agitator seals, Loading, unloading and storage operation, Inprocess wastewater, Sampling and laboratory analysis, Equipment opening for cleaning and maintenance, Pipe and equipment flanges, Process drains and manhole cover seals, Process valves and pressure relief valves, and Open-ended lines. i In order to reduce the fugitive emissions from the above sources, Section 61.65(b) outlined the following three requirements for EDC/VC and PVC plants: Equipment modifications, Operational procedures, and Leak detection and elimination programs. These three requirements will be discussed in detail in the subsequent sections. 4-59 GENC 016577 In addition to the above three requirements for control of equipment leaks, fugitive emissions that could originate from water used in the various processes or water used to meet other requirements of the regu lation (e.g., reactor purging at PVC plants) are required to be controlled. Control of these inprocess wastewater fugitive emissions will also be discussed in a subsequent section. EPA has published an advanced notice of proposed rulemaking for generic standards for airborne carcinogens (44 FR 58662). These generic standards would reduce fugitive emissions of organic chemical carcinogens listed in the future under Section 112 of the CAA. The standards will provide a quick and simplified first step in regulating chemical air carcinogens by leak detection and repair programs. The generic standards are independent of process or chemical and are based on the similarity of operations and equipment throughout an industry such as the Synthetic Organic Chemical Manufacturing Industry (SOCMI). Depending on the nature of the listed organic chemical and emission sources of this chemical, the generic standards may require "tailoring" in certain cases to reflect unique and unusual situations. Generic standards would be followed in most cases by additional standards that would be developed under the proposed Policy and Procedures for Airborne Carcinogens. The draft generic standards need not be considered in any revision to the VC NESHAP. The standards focus primarily on reducing fugitive emissions through the use of an effective leak detection and repair program. This type of program is similar to the programs currently being used in the VC industry. In addition, EPA recently proposed regulations under Section 112 of the CAA that would limit benzene emissions from fugitive sources in new and existing petroleum refineries and organic chemical plants. The proposed regulation is similar in many ways to requirements that the VC NESHAP outlined for prevention of fugitive emissions (e.g., leak detection and repair program and equipment specifications). The following sections describe requirements of the regulation for control of VC fugitive emissions. These sections also discuss new developments resulting from the above draft generic standards and proposed benzene regulations. 4-60 GENC 016578 4.4.2 Equipment Specifications Valves, pumps, flanges and other pieces of equipment are used to move streams of liquid VC, PVC slurry, and VC-contaminated gases to and from various process vessels or control devices. Equipment incorporating sealed interfaces develop leaks after some period of operation, usually because of seal failure. The regulation requires specifications for the following equipment used in VC service: loading/unloading lines, slip gauges, pump seals, compressor seals, and agitator seals. Double mechanical seals are required on rotating pumps, rotating compressors, and agitators. Double mechanical seals are preferred over conventional seals to provide the greatest reduction of fugitive emissions because they have less leakage over a long service life. The inner and outer mechanical seals of this system provide double seal protection from leaks, where failure of either seal does not result in emissions to the atmosphere. Some plants are using tandem mechanical seals instead of double mechanical seals for better product control. Typically, EDC/VC plants are using tandem seals to prevent contamination of the VC product stream by water. The double mechanical seal allows seal fluid to leak into the product line while the tandem seal arrangement allows product to leak into the seal fluid. When water is used as sealing fluid, the water is required to be collected and stripped to 10 ppm or less VC. Other sealing fluids (e.g., oils) are not required to be collected and stripped / / pressure less than 0.4 kPa (0.1 lb/in^) at 20C (68F) or it can be \ a heavy fluid such as kerosene or diesel oil. No*$uch requirement is .______________ ..,,_______ 't- - r.HY..- reciprocating pumps and compressors, and rupture discs upstream from 4-61 6ENC 016579 relief valves. The double outboard seals provide double protection against leakage of emissions to the atmosphere. Relief valves can be a continuous source of fugitive emissions especially those that do not reseat properly after relieving pressure. Rupture discs installed under the relief valve prevent this leakage if properly installed. Slip gauges are no longer being used to measure the level of VC in storage, holding, and transfer vessels. These gauges have been replaced by more reliable level controllers that are not a source of fugitive emissions. Loading and unloading lines have been modified in most cases to reduce the VC remaining to required levels following purging and prior to opening. Section 61.65(b)(4) allows an equivalency to the rupture disc and Section 61.66 allows equivalent equipment to be proposed other than that required by the regulation. For example, instead of installing a rupture disc, leakage through the relief valve can be eliminated by connecting the discharge line from the relief valve to process equipment or the recovery system. As mentioned in a previous section (Section 4.1.4), .;.;.^ne^^^Ta^fcThas"corinected many of their safety relief valves on reactors to a flare as an equivalency for a rupture disc. VC emissions from the polymerization reactor leaking through the relief valve are combusted in the flare. Table 4-12 lists equivalency determinations approved or conditionally approved by the EPA since promulgation of the standard (Brittain, 1980c; Legro, 1977). The "0" ring equivalency determination for rupture discs listed in Table 4-12 can represent a reduction in emissions and equipment main tenance. In the normal operation of a relief valve without abrupture disc, the relief valve seat (metal to metal seat) lifts off the nozzle slightly as the operating pressure approaches the set pressure of the relief valve, and the relief valve begins to "simmer." "Simmering" -occurs with fluctuations in operating pressures, but many times the "simmering" does not cause the relief valve to fully open. The result can be a misalignment of the relief valve seat and continuous leakage when the operating perssure returns to normal. If properly installed and maintained, the use of an "0" ring between the relief valve metalto-metal seat prevents the continual leakage from improper reseating. 4-62 GENC 016580 Table 4-12. APPROVED OR CONDITIONALLY APPROVED EQUIPMENT EQUIVALENCY DETERMINATIONS Equipment Required By Regulation Equipment Equivalency Request Discussion Rupture discs under 1. RV's equipped with "0" all relief valves (RV) ring seat pressure (Section 61.65(b)(4)) seals 1. Approved under following conditions: Ethylene propylene rubber (EPR) rings must be used unless other material approved. Only can be used with RV having disc seat such that no leaks occur during simmering. Must be maintenance program for replacement (per year, each RV event and when leakage occurs). Maintenance program records (keep 2 years). Describe affected RV's prior to modification. Double outboard seals on all reciprocating f" compressors (Section 2 61.65(b)(3)(iv)) 1. Pressurized system -- the vent space between the two seals is pressurized with inert gas 1. yolume between the inboard and outboard seals will be pressurized with inert gas that will be continuously purged to recovery and primary control device. Leakage is into the compressor rather than to the atmosphere and less maintenance is required. 2. Packing rings in place of seals 2. Conditional approval if packing rings are vented at low pressure to a control device. Venting at pressures below atmosphere could lead to dilution with ambient air. Pressure between seals must be specified and flow rate monitored for leakage. Double mechanical seals on rotating compressors and vacuum pumps (Section 61.65(b)(3)(i) and (iii)) 1. Liquid seals with packing modified by adding two "boxes'1 placed over idle and drive ends of unit 2. Labyrinth seals 1. Conditional approval provided mechanical seal used on drive shaft. Idle end will not penetrate "box" which has a vapor tight seal. Drive end will have similar "box" containing mechanical seal. Both "boxes" are vented back into process at 1 psig to reduce emissions. 2. Used on centrifugal compressors. Oe0r0'! (continued) GENC l Table 4-12. Concluded Equipment Required By Regulation Equipment Equivalency Request Double outboard seals on reciprocating pumps (Section 61.65 (b)(3)(H)) 1. Reciprocating Hill-McCanna Type K pump with seals, shaft lubrication, regular leakage inspections and monitoring Double mechanical seals for horizontal agitators (Section 61.65(b)(3)(v)) 2. Double packing pumps with venting of water lubricant to a control device 1. Pressurized grease system Discussion 1. This pump equipped with four separate layers of packing material combined with a Merco Nardstrom lubricator. A sealed pump must be used to lubricate the stuffing box and lubricant levels checked on regular basis. A VC monitoring point must be close. 2. Water lubricant between the seals would be vented to the wastewater stripper which controls VC emissions to 10 ppm or less. 1. Would only be applicable to the bulk PVC process which employs horizontal agitation. 4-64 The purpose of the relief valve/rupture disc combination was to eliminate "simmering." If the rupture disc bursts, however, and the relief valve does not reseat properly, then "simmering" will occur until the rupture disc is replaced. The "0" ring does not completely eliminate "simmering" if it is improperly installed or not maintained, but it will eliminate the need to replace the rupture disc every time a relief valve opens. 4.4.3 Operational Procedures The regulation requires EDC/VC and PVC plants to use the following operational procedures to reduce fugitive emissions: Manual venting of equipment. All gases vented from equipment in VC service are to be ducted through a control device which reduces emissions to 10 ppm or less. Opening of equipment (including loading and unloading lines). Before opening equipment in VC service, the quantity of VC in the equipment is to be reduced to no more than 2.0 percent by 3 volume or 0.0950 m (25 gallons) of VC, whichever is larger, at standard temperature and pressure. The quantity of VC removed in order to meet the requirement is then to be ducted to a control device that reduces VC in the exhaust gases to 10 ppm or less. Sampling procedures. Unused samples containing at least 10 percent VC by weight are to be returned to the process and sampling techniques should employ closed-loop systems. Requirements of this section of the regulation have been met through modification of equipment that formerly discharged directly to the atmosphere. In most cases, the vents from equipment subject to these requirements have been enclosed and are ducted to the pi ant's,VC recovery system which is then ducted to the final control device to reduce VC emissions to the atmosphere to 10 ppm or less. All routine manual venting of equipment is to the recovery system. a positive-displacement unloading procedure and transfer lines are only open to the atmosphere when maintenance is required. When disconnection for maintenance is required, transfer ^ine; are purged to the recovery system. One plant has modified the transferlines so that after purging the lines to recovery, the volume remaining in the lines is less than 4-65 GENC oi6583 3 the allowed volume of 0.0950 m (25 gallons) (Laundrie, 1980). Plants surveyed during this study sample with closed-loop systems and return unused sample to the process. 4.4.4 Leak Detection And Elimination Programs The third requirement for reducing fugitive emissions from EDC/VC and PVC plants is instituting and implementing a leak detection and elimination program. As mentioned above, the major focus of the draft generic standards is a leak detection and elimination program. The VC sources were given the opportunity to develop their own program and submit it to the responsible Regional office for approval. The regulation listed the following six requirements for an adequate program: a reliable and accurate VC area or fixed monitoring system, a reliable and accurate portable hydrocarbon (HC) detector to be used to pinpoint leaks indicated by the area monitoring system and to make routine checks of the plant for small leaks, an acceptable calibration and maintenance schedule for the area monitoring system and the portable HC detector, an acceptable number and location of monitoring points and an acceptable frequency of monitoring, an acceptable plan of action to be taken when a leak is detected, and a definition of a leak which is acceptable when compared with the background concentration in the plant. In addition, plants are required to maintain records for at least two years. A record of leaks detected by the area monitors must include VC concentrations measured and recorded, and the location, date, and time of each measurement. A record of the leaks detected during routine monitoring by the portable HC detector must also include action taken to repair the leak in addition to the above same area monitor recordkeeping requirements. The intention of the fugitive emissions control requirements is to first establish a background level of VC in the plant following instal lation of required equipment and implementation of operational procedures described above. A leak definition would then be set based on the 4-66 GENC 016584 plant's background level of fugitive emissions. The purpose of the leak detection and elimination program would then be to monitor for leaks based on the definition and eliminate these leaks over time, thus con tinuously lowering the background levels as well as fugitive emissions. The background levels and leak definition would be reevaluated and redefined as the fugitive emissions level decreased over time. Following the initial two-year waiver period for compliance, the ERA Regions evaluated the adequacy of the leak detection and elimination programs (Battye, 1978; Battye and Hall, 1978). Evaluations were also made during this review study plant visits. These evaluations provide the background for the following discussion of leak detection and elimination program requirements. Many inconsistencies were found among the plants because each plant designed their own leak detection and elimination programs based on the above six requirements for an adequate program. The site-specific differences among the plants also contributed to inconsistencies. In most cases the requirements were addressed adequately with minimal changes recommended by the EPA Regions (e.g., number and location of area sampling probes). However, other requirements such as leak defini tions, background concentrations and routine plant surveys were not adequate in many cases. An example of the variability of the leak detection and elimination programs among five plants is shown in Table 4-13. Each of the requirements for an adequate program are discussed in more detail in the following subsections. Leak Definition A leak is defined on the basis of the plant's background concentration of VC which not only varies among plants but may vary among different areas of a plant. The background level for open plants or outdoor equipment is usually set at zero. The regulation requires that the leak concentration initially defined is to be reduced over time as background concentrations in the plant are reduced. It is for these reasons that leak definitions varied among plants and that one definition cannot be applied to all plants. In most cases the leak definition was regarded by the Regions to be adequate while In other cases the defined value was too high. 4-67 GENC 016535 1 Table 4-13 LEAK DETECTION AND ELIMINATION PROGRAMS ---------------- -- - - -- - --- .. . ___ _ Mint A O00t Type of area nonllor/ points Gat chromatograph Arc a mr*ni taring i^ttllng iittrrval taelground content ration level ES ppm Action level for are* *onf tor pHSpiirtse,' P *r\y responsible for repair Two consecutive readings greater than $ ppm foreman - nonwritten Hi Han 11 Infrared Analyters tta twelve-stream units. Total of St points. Each unit staples one stream per minute Area AYrray s- 1f CJ nPnPm* IT - S ppm lank farm - S ppm Three consecutive readings greater than 25 ppm. Monitor print reading once per week. foreman - written (OSHA WorthU) Calibration and maintenance schedules Wallthrough program Process equipment check program Area monitorspan charted daily, GC`s ca Mbrated weekly; equipment checked weekly, annual teardown. Portablecalibrated and checked week ly. No progran- areas checked dictated t>y area monitoring systen. Hone. Ail areas checked occasionally Area monitorcalibrated daily with 10 ppm standard. fortable100 ppm standard. Ho programweakly area checking as directed by area monitoring system. No pro gram. checked about once every two weeks. Action level for portable H.C. fl ppm above tackground level. A/A 6ENC 016586 Mi Han II Infrared Spectrometers One sii-strean unit Two twelve-stream units Areas gas chroma tograph with flO One sia-strvam unit lotat of A? point-.. Each unit samples one stream per minute "" Samples one sltvam per f-inotc S ppm One reading greater than ?$ ppm or four consecutive readings of 10 ppm above background level. Area monitor checked every shlft by portable IIC detection operation. Portahle HC detretor operator . nnn-writlen. Area aonltorcalibrated daily with IS.5 ppm standa r<f. Equipment checked weekly. Portablecalibrated w*M`k ly. No programareas checked three shifts daily at directed fey area annitoring system. lcontinue?] rn pm eras HWM WP Table 4-13 Concluded Action level for area monitor Hint Type of irei monltor/ point! Arei monitortng sailing Jnterva 1 Background concentration Jewel Response/ pirl y responsible for repair Cat Ibration and maintenance schedules WlHthrough program Procei! equipment check program Litton le*tt for portable H.C. D tOCOH Fourier COCOH - non-cycHc. 1-2 ppm 100 ppm conceit- tOCOH - Entire None. It/A Multiplex IJ. ay alx streams to trat ion over a plant iprctrometer find a group with 5 minute period. claims no three lltl to let! thin ] ppm. htfheit ppm, then breaks down this Respond with cal 1 bra lion Is needed. times each Incorporates It rocc*5ut**r group. Computes ititiiltcil proli- portable H.C. checking. Checked twice daily week. {Indoor irri). ahi11ty of an excur- with sion In a given area. foreman osc1lloicope. Two Kiri an l.R. Can analyze a sample Portibtr !P*ctromrter( in 20 seconds. (Century) (Outdoor im). catibrated Hlrtjn - continuous oix>> Total of 47 point* -- monl tor. electronically once daily. E Hirer Arcis SOS Each stream every lost than Two consecutive G.C. checked Thorough Reactors Single G.C.1! wttU ten minutes. 1.2 ppm readings greater daily and check oT are checked 2.5 ppn FIO. than 25 ppm for calibrated reactors daily at reading One screen per area monitor. twice a week daily; uni folds, Total of minute. with three other areas agitator 48 points Portable - 25 dl f rerent checked seals, rup ppm reading. standards. thoroughly ture disks, Respond with Portable - once per month. manways, con densors portable II.C. check. calibrated Oata Is and piping. Operator - for weekiy. recorded. minor repairs; shift foreman for more extensive problems. GENC 016587 A separate leak definition was usually identified for area monitors versus portable monitors. Table 4-14 shows the range of leak definitions for 12 PVC plants. As indicated by these leak definitions, there is a great deal of variability. Many plants had not compiled data from which to determine background levels, others had not defined a distance from the leak for the portable monitoring leak definition and many plants neglected the storage and handling areas completely. Area Monitoring System According to EPA Regional personnel an adequate area monitoring system should accomplish four purposes: monitor processes for leaks, protect employees from occupational exposures (OSHA requirements), identify process conditions (e.g., start-up, shut-down, upset conditions) that precipitate VC fugitive emissions, and provide background data base for relocation of sampling probes and redefinition of leak. The draft generic standards considered continuous area-wide monitoring to measure ambient concentrations of hazardous chemicals, but found this type of monitoring not as effective in locating leaks as a seal-by-seal routine inspection. Any added effectiveness from area-wide monitoring is minimal, plus it is a capital intensive technique. In the VC industry the leak detection and repair programs are based on area-wide monitoring for ambient VC concentrations. The effectiveness of this monitoring is discussed below. The most important variable for an adequate area monitoring system is the location of sampling probes. The number and location of sampling probes for the area monitoring system will vary among plants --'this requirement is too site-specific to set exact numbers and locations. Many plants already had area monitors to protect work areas as required b_y OSHA. As a result, this same system was utilized for EPA require* ments and many probes were located in worker breathing zones instead of fugitive emissions sources. OSHA probes were located where workers spend most of their time, thus many areas (which have a potential for fugitive emissions) are not monitored. Some plants completely neglected 4-70 GENC 016588 Table 4-14. VARIABILITY IN LEAK DEFINITIONS Plants Area Monitoring System Leak Definition (ppm) Consecutive Readings (or Persistent Time Period) Background Level (ppm) Portable Monitoring System ' Leak Definition 1 (ppm) Inches From Source (or Persistent Time Period) Background Levei (ppm) Storage and Handling Facilities (Area Monitor) Leak Definition (ppm) Consecutive Readings (or Persistent Time Period) Background Level (ppm) A3 >100 2 (10 min.) 1 0.50 >100 0.5 2500 B3 1 0.82 100 1 in. 0.82 5 C >5 >100 2 1 >50 6 in. >10 >100 D 25 60 mtn. None 50-300 3 in. (30 min) None Determined >300 Determined E 20-25 1 1 10 3 in. 1 50 F1 1 0.5 50 12 in. 0.5 5 G 25 2 1.2 25 1.2 25 H5 I 100 J 25 >10 K 25 L >5 2 5 min. 1 4 3 3 Confidential 1 to 2 5 1-10 0.5-5 25 100 * * Confidential , 5 1 to 2 * * \ 25 >10 ** 0.5-5 cr> 5o * Routine survey conducted only if area monitor printout indicates leaks. occno 0.1 i0 2 1 1 1 <0.5 2 1.2 2 Confidential 15 4 3 0.5-5 3 0.5-5 storage and handling areas which can be a significant source of fugitive emissions. For example, the area monitors are set at different concen tration levels which set off an alarm when this level is exceeded. Some alarm levels were set only for personnel safety, which can be higher than the leak definition for a particular plant, and the EPA leak defi nition level was neglected. In other cases where the alarm is set to detect fugitive emissions, the alarm is set at a level higher than the leak definition. Many of the plants had insufficient background data to determine if area sampling locations and background levels set were adequate. Some plants determine how often an alarm is sounded for a particular concentration in different areas of the plant over time and relocate sample probes based on these alarm frequencies. In some cases, background levels, as well as leak definitions, may vary in different areas of a plant. Some plants have located area probes close to exhaust fans and equipment near air ventilation intake fans. The probe near an exhaust fan responds to all leaks but there is a dilution effect, while the probe near the intake fan does not respond at all. A compromise between these two locations based on air flow patterns through enclosed buildings results in the proper placement of probes. The sampling location near an exhaust fan has been used in some cases to provide data for reevaluation of the program because VC concentrations at this point should decrease with time. Area sampling lines in some cases are manifolded to 2 or more sampling probes. The result is that a localized leak near one probe will often be diluted by air from other probes and thus, the leak will not be detected until the concentration is higher than the actual leak definition. Also, there is no way to determine if one probe becomes clogged or broken unless sample line integrity is checked on a regular basis. Probes close to reactors can be easily broken by reactor vibrations. Particle filters or metal shields on outdoor probes help to provide protection from probe and sample contamination and breakage. Continuous air purging of probes and sample lines is recommended to ensure that the concentration recorded is indicative of the current 4-72 GENC 016590 reading. Also, gases used to calibrate the area monitor provide more representative results when injected at the sampling probe. In a recent study (EPA 1980) done in support of the proposed standard for benzene fugitive emissions, these recommendations were followed. ..Calibration gases were injected directly into the sample probe to evaluate'the~T~~'. response of the area-wide monitor; Results of the calibration are indicated in Table 4-15. The concentrations measured by the area-wide monitor were much lower than the actual concentrations of the calibration gases. The gases were adsorbed to the inner walls of the sampling lines because concentrations continued to increase and then slowly decreased as the gases were desorbed from the sampling line inner walls. Another important aspect of the area monitoring system is the time interval used to cycle all sampling points. Most plants use more than one instrument and usually each instrument has 10 to 16 sampling streams. Analysis of a sample collected by a stream takes about 1 minute (20 to 90 second range), thus each stream is monitored once every 10 to 16 minutes. However, some plants have different systems, whereby cycle time selection is more elaborate. For example, one plant that produces other chemicals in addition to VC, uses a halogen analyzer and gas chromotograph which chose the point to be analyzed based on four classes of information collected - a halogen alarm level sounded, maximum time for a cycle, any alarm level sounded, and if no alarm level is sounded (this is most frequent choice of system and results in a sample once every 40 to 50 minutes). The maximum sampling time between each point is 1.5 to 3.0 hours with all points monitored at least eight times every 24 hours. Another plant does not cyclically monitor but instead uses a micro processor that computes statistically the probability of an excursion over the leak definition in each sampling area and selects the area most likely to be near an excursion. This selection is based on historical data and time elapsed since last analysis. This system also mixes streams to save time and if after mixing the VC concentration is high, the instrument immediately analyzes each stream separately. An operator can also select a stream to see if an excursion has been corrected and review an hourly printout showing time weighted concentration averages. 4-73 GENC 016591 Table 4-15. CALIBRATION RESULTS FOR AREA-WIDE MONITOR Calibration gases Vendor analysis (ppm) Laboratory analysis (ppm) Sample 1 Sample 2 Benzene Toluene (para)xylene (ortho)xylene 5.37 54.30 28.10 22.00 5.00 53.00 48.00* 6.00 52.00 45.00* The laboratory analysis and area-wide monitor measured total xylene. Area-wide monitor reading (ppm) Probe 1 Probe 2 0.04 18.16 0.00 44.39 21.29* 19.40* Z6S9I0 0N39 t-- t--" be--! i--* i-- r-'i r*i mm mm mr r'"' ' ) Plants surveyed use one or more of the following area monitoring systems: Areas 505 gas chromatograph with flame ionization detection (FID), Bendix 6000 series gas chromatograph with FID, E0C0M Fourier multiplex infrared spectrometer (FMS-7200) with minicomputer, and Mi ran II Infrared Analyzers with infrared spectrophotometry. The gas chromatographs use columns to separate the VC from other compounds. The FMS-7200 uses two wavelength bands and is therefore more specific for VC. Sensitivity of the FMS-7200 is less than 1 ppm. The Miran II is not as specific or accurate for VC and is usually used to monitor outdoor equipment only. There are other Miran instruments that are more appropriate for indoor monitoring. Routine Surveys With Portable Detector As mentioned above, the draft generic standards emphasize routine leak detection surveys with a portable monitor over area-wide monitors, or a combination of the two. The VC industry uses a combination of routine surveys and area-wide monitors. The portable hydrocarbon (HC) detector is used to immediately identify the leaking equipment source when an area alarm is sounded and to conduct leak surveys throughout the plant on a regular schedule. The following portable HC detectors were used by the plants surveyed: Century Systems Organic Vapor Analyzer (OVA) - the instrument measures total HC by flame ionization. H Nu Systems Photoionization Detector (Model P11Q1) - the instrument has three concentration ranges for total (HC. Some plants favor one of these portable meters over the other or use both, one as primary and one as back-up; both are regarded as adequate for leak detection. The routine surveys that are required to be conducted on a regular basis probably represented the least consistent area in the leak detection and elimination programs reviewed. The schedule for routine portable monitoring, other than following up an area monitor alarm, varied among 4-75 GENC 016593 the plants -- from a once per shift basis to a weekly or monthly basis or no regular schedule at all. In some cases, the routine survey was conducted on a regular basis and in other cases, on an irregular basis depending on area monitoring print-outs for a particular shift or day. For PVC plants, the most thorough routine programs consisted of monitoring the reactor daily and other equipment and areas of the plant on a weekly, monthly, or quarterly basis. Usually the weekly, monthly, or quarterly frequencies in most cases are determined by historical data and type of equipment service. Routine leak surveys usually follow an equipment checklist so that the same pieces of equipment are consistently monitored. This not only provides consistency, but also identifies those pieces of equipment that chronically leak and may require more frequent monitoring. Instrument Calibration And Maintenance Most plants surveyed follow the instrument manufacturer's recommended calibration and maintenance procedures. Standard gases were usually used for instrument calibration and the concentration of these gases was in most cases close to the leak definition. Some plants calibrated portable monitors with a Wheatstone bridge which only provides a check on the internal electronics. As mentioned above for the area monitors, calibration gases injected at the sample-probe location provided the best method for calibration as opposed to injecting the gas directly into the instrument. Also, there can be interference if other chemicals are being produced at a facility. This should be taken into consideration when developing a calibration and maintenance procedure as well as when locating area sampling probes and conducting routine leak surveys. Plan Of Action When Leak Detected 1 In most cases, when an alarm is sounded by the area monitoring system, the following actions are usually activated: a leak search is initiated with a portable monitor in that area of the plant, the leaking piece of equipment is identified, and the leak is eliminated. There is some variation in procedures taken once the leak is identified, but most plants follow similar actions for elimination based on the 4-76 GENC 016594 severity of the leak. Some leaks can be stopped immediately by turning a valve or tightening a flange or the leak may be severe enough to cause immediate equipment decommissioning, which requires that the equipment be immediately put on recovery and shut down at the earliest and safest time. The leak may also be somewhere between these two extremes, in which case the leaking equipment can be enclosed and ducted out of the building until the maintenance personnel can repair the leaking equipment. In all cases, it is important that an accurate log be kept on the leaking equipment - date, time, location, cause of leak, quantity of emissions (if possible), corrective action taken, and time until repair. Shift supervisors are usually made responsible for seeing that the leak is properly eliminated. The equipment usually is monitored again imme diately after repair to verify repair and elimination of the leak. An accurate record of these leak abatement steps is not only required but also identifies those pieces of equipment that chronically leak and may require more frequent monitoring. Recordkeeping Requirements As mentioned previously, detailed records are to be maintained at the facility for at least two years. These data are maintained for review by EPA personnel during inspections - no records on fugitive emissions are required to be submitted to EPA. Only the original proposed program was to be submitted for approval. Some plants maintain records for their own benefit that are not required by the regulation, such as the number of times a specific area alarm is activated or identification of equipment leaking below the leak definition. These extra data, in conjunction with required data, can be used to determine trends - the number of leaks found as a function of time or the background Tevels as a function of time. This information can then be evaluated periodically to identify problem areas and seek ways to eliminate the problems (e.g., replace chronically leaking equipment with new or different equipment). The information can also be used periodically to improve the program and reevaluate the leak definition. Adequate Leak Detection and Elimination Programs EPA Region VI, which is responsible for the majority of subject sources, evaluated the programs submitted and based on these evaluations 4-77 GENC 01,6595 and follow-up inspections, determined the following minimum requirements for an acceptable program (Ramirez, 1978): A sufficient number of sampling points must be utilized to detect a leak no matter which direction the wind is blowing. A leak patrol must survey the entire plant at least once a week. An accuracy test of the leak detection system should be conducted by introducing a known concentration of VC into a sampling probe. The complete system should be checked once a year and reported in the semi-annual report. The VC calibration gas cylinders should be analyzed when they are received by utilizing Method 106. There must be an alarm system (visual or audio) to notify the plant personnel of a leak. This alarm must continue until acknowledged. There must be an instantaneous printout showing location and concentration of leaks when they occur. The evaluation of each company's background level will be made on a case-by-case basis in order to evaluate their definition of a leak. The portable hydrocarbon detector must be calibrated at least once a week. An additional important requirement is the development of a data base for fugitive emissions through accurate recordkeeping. In some instances, this recordkeeping may require additional data collection other than that required by the regulation. The data base developed could then be evaluated on a regular basis to reevaluate a plant's background level and redefine the leak definition. 4.4.5 Inprocess Wastewater Water used in the EDC/VC and PVC processes can become contaminated with VC and be a source of secondary emissions if not contained and con trolled. Even though the above requirements of the regulation discussed for fugitive emission control (equipment specifications, operational procedures and leak detection and elimination programs) are not applicable 4-78 GENC 016596 to inprocess wastewater, it is still categorized as a fugitive emissions source. The regulation requires the concentration of VC in each waste water stream containing greater than 10 ppm VC measured immediately as it leaves a piece of equipment, and before being mixed with other wastewater, be reduced to 10 ppm VC or less. This concentration in the water must be attained prior to mixing with other inprocess wastewaters con taining 10 ppm or less VC, before being discharged to a wastewater treatment process or plant, or before being discharged untreated to a body of water. There are several sources of inprocess wastewater in EDC/VC and PVC plants. In EDC/VC plants, water from EDC purification and VC cracking and purification (as shown in Figure 3-1) and equipment seals is con taminated with VC at levels greater than 10 ppm. In PVC plants, VCcontaminated inprocess wastewater is generated from the following sources water used to evacuate reactors prior to opening in order to meet reactor opening loss (ROL) requirements, t water used as sealing fluid for double mechanical seals on pumps, compressors and agitators, water removed by knock-out pots used in the monomer recovery system, and water used to seal gasholders. Depending on the methods used to attain the ROL requirement (e.g., water piston which is discussed in Section 4.5), large quantities of inprocess wastewater can be generated. Inprocess wastewater from ROL methods, seals, and knock-out pots is usually collected in a vessel and then steam stripped of VC. Water used to seal gasholders is in contact with VC-contaminated gases and by definition is an inprocess wastewater. This water is always exposed to the atmosphere without treatment. Actual VC concentrations of this seal water were not available, but for one plant using a water-sealed gasholde emissions to the atmosphere from the seal were calculated to be approximately 0.79 kilograms (1.75 pounds) per year (Battye, 1978). Removal of VC dissolved in water is usually accomplished by a distillation column. The water collected by fugitive emissions sources, 4-79 GENC 0U597 and usually held in a vessel prior to the column, is close to saturation. Thus, separation in the stripping column is simple because the vapor pressure of VC is much greater than water. 4.5 REACTOR OPENING LOSS 4.5.1 Introduction The VC emissions resulting from venting a polymerization reactor to the atmosphere (other than an emergency relief discharge as defined in Section 61.65(a)) constitute reactor opening loss (ROL). The ROL standard is only applicable to PVC plants. These emissions are regulated under Section 61.64(a)(2) and are limited to 0.02 grams VC per kilogram PVC produced (on a dry solids basis). This regulation applies to any vessel used as a reactor or as both a reactor and a stripper. Determination of ROL emission levels is made by actual sampling and analysis of VC levels at the bottom, middle, and top of the reactor (methodology specified in the standard). A calculation option is pri marily used by processors stripping in the reactor and is based on number of evacuations, vacuum applied, and volume of gas. The calcu lation (if approved) represents a waiver from testing, not an equivalent method of ROL determination. When the reactor serves as the stripping vessel, the resin (after stripping is completed) contributes VC to the headspace concentration and consequently to the ROL measurement. Reporting of ROL emission levels is made in the semi-annual report submitted by regulated facilities to Regional EPA offices. (See Table 4-16 for representative reported ROL levels). To maintain product quality, reactors are opened for maintenance or inspection and for removal of residual polymer adhering to reactor walls. The frequency of openings is a function of resin types' and grades, reactor size and construction, location and method of the stripping process, and effectiveness of reactor cleaning methods. ROL emissions control can be achieved through various technologies and process modifications. Basically, the objective is to: t minimize the amount of VC in the gas phase prior to opening, and/or maximize the total PVC production per reactor opening. 4-80 SENC 016598 Table 4-16. REACTOR OPENING LOSS REPORTED BY REPRESENTATIVE COMPANIES (Data from September 1980 Semi-Annual Reports) 4-81 Company Resin (code) ___ type A suspen. 8 suspen. C suspen. D suspen. t disper. F disper. G disper. H bulk I bulk J latex Number of openinqs 72 47 757 215 GO 766 5668 1340 Number out of compllance 2 1 2 0 2 0 0 18 0 0 Compliance rate 97.21 97.91 99.7% 1001 96.71 1001 1001 99.71 1001 1001 Concentration calculated or actual Control technology actual calculation actual Water Piston (Displacement) Steam Sweep Technology actual calculation Solvent Clean!ng/Closed Charge Technology Steam Injection Steam Sweep Technology Redox Catalysis GENC 0165?? Reducing the amount of gaseous VC can be accomplished by evacuation to a low absolute pressure, by displacement with water or an inert gas, by steaming, or by a combination of these. Increasing total PVC production may be achieved by reducing reactor opening frequency. (This also has the advantages of reducing turn around time and minimizing employee exposure to VC). Technologies available for controlling ROL emissions have been developed by the various plants to suit their individual processes. The method identified in the original standard support document, water piston, is not applicable to all processes (EPA, 1975). The water piston method as well as other methods are discussed below. 4.5.2 Solvent Cleaning Solvent cleaning systems reduce the number of times a reactor has to be opened for cleaning. The solvent, circulated through the reactor, dissolves the solid scale present on the reactor walls, eliminating the need for manual scraping. Several solvents, suitable for reactor cleaning, have been tried by PVC processors. These include di-methylformamide (DMF), tetrahydrofuran (THF) and dichloroethane. However, there are serious disadvantages to the use of these solvents for reactor cleaning. These include: The expense of the solvents. THF is approximately $2.00 per liter ($7.00 per gallon). One plant experimenting with this chemical for ROL reduction lost 76,000 to 114,000 liters (20,000 to 30,000 gallons) of solvent per month through the process. This represents a loss of up to $210,000 per month. The creation of secondary sources of pollution (solvents appropriate for reactor cleaning pose environmental and work place hazards). > The requirement for energy-intensive stripping operations to remove PVC from the solvent so that the solvent can be reused. Nevertheless, a few processors using solvent cleaning are reporting low 'ROL levels. A solvent cleaning process using DMF has been patented by Air Products and Chemicals Incorporated and is probably available for license. This process involves filling the reactor with solvent under an inert atmosphere. The reactor is heated and stirred while a small 4-82 GENC 016600 continuous flow of solvent overflows the reactor top. After sufficient time elapses, the solvent is transferred to storage and the reactor is rinsed several times with water. The reactor is then ready for another polymerization cycle. Solvent regeneration includes precipitation of solids, centrifuging to remove solids, and distillation to purify the solvent. 4.5.3 Steam Piston One PVC manufacturer surveyed uses this technology to control ROL emissions for their suspension and dispersion reactors. The process involves draining the reacted slurry from the reactor and placing a puddle of water in the bottom of the reactor. A vacuum is pulled and steam is applied to the vessel jacket. Under vacuum the water boils and evaporates at low temperatures. The steam rises through the vessel in a "piston like" manner and is exhausted to the VC recovery system. This technique has proven effective in reducing ROL emissions. 4.5.4 Water Piston An application of this technology was noted in a PVC suspension homopolymer plant and identified in the standard support document. After the slurry is discharged to another vessel for stripping, the reactor is hydraulically filled to the nozzles with water, displacing VC vapors to recovery. Steam is applied to the jacket to heat the water in the reactor and vacuum is pulled from the top of the reactor. The vapors in the head space are exhausted to the recovery system by way of a knock-out pot (Laundrie, 1980). One problem encountered in this method is the generation of large amounts of water contaminated with VC. Another consideration is the type of polymer involved. Diffusion rates of VC out of the polymer are not only temperature dependent but also depend on the resin character istics. If the jacket temperature is too high, polymer degradation can proceed to the point where removal of polymer build-up is difficult after the reactor is opened. Jacket temperatures that are not hot enough result in decreased VC diffusion rates, thereby extending the time required for adequate VC removal and increasing emissions when opened. 4-83 SENC 016601 4.5.5 Reactor Purge Air Blower This technology is used by a suspension-polymerization plant manufacturing homopolymer and copolymer resins. This facility strips the slurry in the reactor. Prior to reactor opening, a purge air blower sweeps the vapor space above the slurry. The blower discharges directly to the incinerator. ROL compliance is then determined by calculation (Schaul, 1980). 4.5.6 Steam Purge (Sweep) This represents the most widely used technology for control of ROL emissions. It involves the injection of live steam under vacuum, con densing the waste steam, and subsequent stripping of the water. Less VC contaminated water is generated by this process. Parameters for this process, reported by a suspension and dispersion resin plant, consist of a 15-minute purge under vacuum. This procedure has consistently resulted in ROL emission levels well below the standard for this plant (Battye, 1978, Vol. II, p. 30). Vacuum pressures and temperatures vary from plant to plant depending on resin characteristics and process variables. One problem encountered by a plant using this steam sweep technology involves leakage by shut-off valves back into the reactor, resulting in ROL excursions. This has been attributed to scoring of the valves and subsequent failure to reseal properly. 4.5.7 Redox Catalysis One company, whose latex PVC facilities substitute a post-polymerization process for resin stripping, eliminates ROL emissions entirely. The method involves removing all equipment from VC service prior to opening by using a caustic wash and pickle rinse. These rinses are treated with a catalyst which scavanges unreacted VC and the rinses are monitored for VC until they are shown to be below 10 ppm prior to release to the sewer. The technology is proprietary (Ferrell, 1980). 4.5.8 Water Jet Cleaning A procedure reported for suspension resin reactor cleaning - Hydraulic Reactor Cleaning (HRC) - used 28,000 to 41,000 kPa (4,000 to 6,000 psi) water in a water jet cleaning system with the water jet collar fitted directly into the reactor manway. This equipment allowed 25 to 30 4-84 GENC 016602 batches to be run between reactor openings instead of opening between each batch. An identical system for dispersion resin reactors allowed 5 to 15 batches between openings. With some development expense, a spray system of the type could be developed by any PVC producer. The HRC technology described was developed by B, F. Goodrich and is probably available for licensing. 4.5.9 Clean Reactor (Closed Cleaning) Technology This technology encompasses a wide range of techniques used to keep reactor walls free of deposits. Included are wall coatings, recipe modifications and mechanical cleaning devices. In conjunction with large reactor utilization, which allows more room for internal cleaning devices, clean reactor technology affords abatement of emissions from other sources in addition to the ROL. High pressure spray systems and nozzle designs (for removal of deposits) and formulations of dispersants, catalysts and additives (for reduction of deposits) are incorporated in this technology. These are patented and some are available for license. 4-5.10 Nitrogen Purge Nitrogen, used to purge the reactor under a vacuum, is non-condensable and therefore this method is not easily adaptable to monomer recovery without increased expense. Primary applications of this ROL emissions control method are in Pre-Po reactor preparation for opening. The vacuum is broken three or four times with nitrogen - the final break being made with air. The nitrogen gas, which contains recovered monomer, is subsequently incinerated. 4.5.11 Slurry Backfill One plant is known to use this method to control ROL emissions for reactors used as strippers. The process involves dispersion resins which are stripped in the reactor. Following polymerization, the slurry is stripped under vacuum in the reactor vessel until the required resin RVC level is attained. The reactor is then backfilled with previously stripped slurry forcing headspace vapors out of the reactor to the recovery system, and thus also eliminating the reactor headspace. No further vacuum is applied because liquid slurry would be evacuated at this point. The ROL requirement is met by this method because the 4-85 GENC 016603 headspace vapors have been eliminated (Konter, 1980). A waiver of testing requirements has been obtained from the Region. Resin RVC emissions lost during transfer of the slurry to another vessel cannot contribute to ROL emissions because these emissions would be considered beyond the stripping operation and therefore not regulated (Section 61.64(e)(1)). The main advantage of the slurry backfill control technique for ROL is the small amount of water used, eliminating the need for a large VC-contaminated wastewater stripper. In addition to the pumps required for backfilling, a level sensor is required to insure that the slurry is not backed into the recovery system. This process would also appear to be adaptable to suspension processes. 4.5.12 Calculated Emissions In processes where resin stripping takes place in a separate vessel, the reactor is under pressure which allows the slurry to be evacuated without opening the reactor. The methods described above can then be applied prior to opening in order to attain the ROL standard. However, when the reactor also serves as the stripping vessel, the vacuum (pulled when the slurry is being stripped) must be broken prior to slurry dis charge and this constitutes a reportable reactor opening. The required testing procedure for ROL can not be applied until the slurry is removed. The slurry may contribute enough RVC so that the ROL standard is unattainable and these measurements cannot be made until the reactor has cooled. . The prescribed method for measuring ROL is impractical and unsafe for those plants stripping in the reactor. If the probe measurement is taken while the reactor is hot immediately after stripping, the high vapor temperatures present a safety problem to the person making the measurements. If the reactor is allowed to cool, this not only affects productivity (because several hours are required for cooling) but also results in measurements that are not representative. When ROL measurements are made according to the present regulation, the same VC is essentially "counted twice" (i.e., once in the resin and once in the vapor space). 4-86 6ENC 016604 Through waivers of emission testing for reactor opening loss, PVC plants whose stripping operations take place in the reactor, are con sidered to be in compliance with Section 61.64(a)(2) if the measured reactor opening loss added to the measured resin RVC is less than the sum of allowable ROL and the allowable resin RVC. For plants stripping below the RVC standard, the processor is not penalized for that portion of the RVC (removed by stripping) that contributes to the ROL. The ROL levels are allowed to be higher when the reduced stripping emissions are used as a credit. The calculation may present a problem even though the results as calculated are correct. The calculation will not include those emissions from leaking valves or broken equipment that may contribute to ROL emissions, but are only found by an actual test. It is the policy of some Regions, in keeping with the intent of the standard, to grant waivers of testing to those processors in the above category. These waivers are made on a case-by-case basis, and with the provisions that RVC samples must be taken on each batch. A calculation may then be used to establish ROL. Formulas developed for these calcu lations are generally treated as confidential by the plants using them because the formulas are considered to be potentially marketable and they can reveal some of the nature of the process. The basis for the ROL calculation is Raolt's Law. The general requirements for using calculated ROL1s are: (1) Calculations must be done for each grade of resin, (2) Parameters such as number of times a vacuum is pulled, must remain constant, and (3) Tests must be confirmed by standard methods. > 4-87 GENC 01*605 4.6 REFERENCES FOR CHAPTER 4 Aronson, Wayne, J. 1980. Enforcement Division, EPA Region IV, Atlanta, Georgia. Freedom of Information Request from J. W. Bodamer, Jr. , TRW Environmental Engineering Division. October 1980. Battye, William. 1978. GCA Corporation, Technology Division. "Technical Assistance to Region III for Enforcement of Vinyl Chloride Regulations," Contract No. 68-01-4143, Task No. 35, Volumes I, II, III, V and VI. December 1978. Battye, William and Hall, Robert R. 1978. GCA Corporation, Technology Division, "Technical Assistance to Region V for Evaluating Vinyl Chloride Leak Detection and Elimination Programs," Contract No. 68-01-4143, Task No. 16, Volumes I, III, V, VI and VII. July 1978. Blacksmith, J.R., G.E. Harris, G.L. Langley. 1980. Radian Corporation, "Frequency of Leak Occurrence for Fittings in Synthetic Organic Chemical Plant Process Units," Contract No. 68-02-3171, Task No. 001. September 1980. Brittain, Martin. 1980(a). NESHAP Coordinator for EPA Region VI. Meeting report - TRW visit to Region VI offices, Dallas. July 23, 1980. Brittain, Martin. 1980(b). NESHAP Coordinator for EPA Region VI. Telecon with J. W. Bodamer, Jr., TRW Environmental Engineering Division. November 7, 1980. Brittain, Martin E. 1980(c). Environmental Protection Agency, NESHAP Coordinator, Region VI. Information supplied during meeting between Region VI and TRW. July 1980. Brumbaugh, Gerry. General Tire and Rubber Company. Telecon with J. W. Bodamer, Jr., TRW Environmental Engineering Division. November 14, 1980. Chemical and Engineering News, 1978. "Key Polymers." September 4, 1978, p. 13. Chemical and Engineering News, 1980(a). "Key Chemicals." July 7, 1980, p. 9. Chemical and Engineering News, 1980(b). "Key Polymers." October 6, 1980, p. 13. DeBernardi, James. 1980. Plant Manager, Conoco Chemicals, Lake Charles, Louisiana, EDC/VCM plant. Telecon with M. A. Cassidy, TRW Environmental Engineering Division. November 18, 1980. DeBernardi, James. 1981. Plant Manager, Conoco Chemicals, Lake Charles, Louisiana, EDC/VCM plant. Telecon with M.A. Cassidy, TRW Environmental Engineering Division. March 6, 1981. 4-88 GENC 016606 Diem, Conrad. 1980. Special Enforcement Section, EPA Region III. Letter to J. W. Bodamer, Jr., TRW. December 1980. Dubec, Harold F. 1980. Manager of Environmental Compliance, Hooker Chemical. Trip report - visit to Hooker Chemical Company, Ruco Division, Burlington, New Jersey. September 1980. Environmental Protection Agency. 1980. "Emission Test Report Benzene, Fugitive Emissions - Petroleum Refineries," EMB Report No. 78-QCM-12B, October 1980. Environmental Protection Agency. "Proposed National Emission Standards for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer," Federal Register/Volume 44, No. 197/Wednesday. October 10, 1979. Environmental Protection Agency. 1975. Standard Support and Environmental Impact Statement: Emission Standard for Vinyl Chloride. EPA-450/2-75-Q09, October 1975. Erdmann, John F. 1980. Environmental Protection Coordinator, Union Carbide. Telecon with J. W. Bodamer, Jr., TRW Environmental Engineering Division. November 4, 1980. Erdmann, John F. 1980. Environmental Protection Coordinator, Union Carbide Corporation, Texas City, Texas. Telecon with M. A. Cassidy, TRW Environmental Engineering Division. December 4, 1980. Ethyl Corporation. 1980. Telecon with Joy Reed of TRW Environmental Engineering Division. July 1980. Fannin, James. 1981. B. F. Goodrich Chemical Division. Telecon with M. A. Cassidy, TRW Environmental Engineering Division. February 18, 1981. Ferrell, John. 1980. Union Carbide Corporation, Tucker, Georgia PVC Plant. Telecon with M. A. Cassidy, TRW Environmental Engineering Division. December 5, 1980. Finch, Walter. 1980. Senior Process Engineer, Conoco Chemicals. Meeting report - Conoco/EPA/TRW meeting at RTP. October 17, 1980. Gilmore, J. M. 1980. Plant Manager, Goodyear Tire and Rubber Company, Niagara Falls, New York PVC plant. Semi-Annual Report. September 10, 1980. Holbrook, W. C. 1979. Director of Toxicology and Environmental Affairs, B. F. Goodrich Chemical Group. Letter with attachments to Donald R. Goodwin, U.S. EPA. November, 19, 1979. Holbrook, W. C. 1980(a). Director of Toxicology and Environmental Affairs, B. F. Goodrich Chemical Group. Telecon with J. W. Bodamer, Jr., TRW Environmental Engineering Division. October 17, 1980. 4-89 GENC 016607 Holbrook, W. C. 1980(b). Director of Toxicology and Environmental Affairs, B. F. Goodrich Chemical Division. Letter with attachments to J. W. Bodamer, Jr., TRW Environmental Engineering Division. December 12, 1980. Holbrook, W. C. 1980(c). Director of Toxicology and Environmental Affairs, B. F. Goodrich Chemical Division. Trip report - visit to the Pedricktown Polyvinyl Chloride Plant. September 17, 1980. Kachtick, James. 1980. Tenneco Chemicals, Pasadena Texas PVC plant. Telecon with M. A. Cassidy, TRW Environmental Engineering Division. December 2, 1980. Konter, Ken. 1980. Senior Environmental Engineer, B. F. Goodrich Chemical Division. Telecon with Matthew Boss, TRW Environmental Engineering Division. December 5, 1980. Laundrie, Robert. 1980. General Tire and Rubber Company, Chemical and Plastics Division. Trip report - visit to General Tire's polyvinyl chloride facility in Ashtabula, Ohio. September 10, 1980. Ledvina, Joseph C. 1980. Director of Environmental Activities, Conoco, Inc. Meeting report - representatives from Conoco, TRW, and EPA. October 17, 1980. Legro, Stanley W. 1977. Environmental Protection Agency, Division of Stationary Source Enforcement. Letter to Regional Enforcement Directors containing determinations of equivalent compliance methods for vinyl chloride. May 26, 1977. McCulley, John H. 1980. Chemicals Division, Conoco Inc. Letter to J. W. Bodamer, Jr., TRW. December 16, 1980. Moulthrop, Samuel P. 1980. Enforcement Division, EPA Region II. Telecon with J. W. Bodamer, Jr. of TRW Environmental Engineering Division. November 10, 1980. Neveril, R. B. 1978. Capital and Operating Costs of Selected Air Pollution Control Systems, GARD, Inc. , 1978. EPA Contract No. 68-02-2899. December 1978. 1 Pucci, Michael. 1980. NESHAP Coordinator for EPA Region II. Meeting report TRW visit to Region II Offices, New York. August 12, 1980. Richter, S. H. 1975. Arthur G. McKee and Company, Cleveland, Ohio. "Size - Relief Systems for Two-Phase Flow," Hydrocarbon Processing. July 1975. Schaul, Peter. 1980. EPA Region III. Meeting report - TRW visit to Region III Offices, Philadelphia. September 16, 1980. 4-90 Sittig, Marshall. 1977. How to Remove Pollutants and Toxic Materials From Air and Water. Noyes Data Corporation. 1977. Smith, Cornelius. 1980. Chief Environmental Attorney, Union Carbide Corporate Headquarters, New York, New York. Telecon with M. A. Cassidy, TRW Environmental Engineering Division. December 5, 1980. Varner, Bruce. 1980. NESHAP Coordinator for EPA Region V. Meeting report - TRW visit to Region V offices, Chicago. August 19, 1980. West, Michael. 1981. Air Facilities Branch, EPA Region II. Letter to J. W. Bodamer, Jr., TRW Environmental Engineering Division. January 21, 1981. Wu, James. 1980. NESHAP Coordinator for EPA Region IV. Telecon with J. W. Bodamer, Jr., TRW Environmental Engineering Division. November 7, 1980. 1 4-91 GENC 01660? 5.0 ENFORCEMENT AND COMPLIANCE EXPERIENCE 5.1 INTRODUCTION Industry representatives and Regional EPA personnel presented their viewpoints drawn from experience with enforcement and compliance under the existing VC NESHAP. Many of these points were common to industry and EPA regional personnel while others were specific to enforcement problems or compliance experiences. These comments are reviewed below and have been categorized as they pertain to requirements in the regulation. (All of the following are opinions expressed by industry and/or EPA personnel.) 5.2 INTENT OF THE STANDARD Health effects: Nearly all-of the industrial representatives / 'contacted (and one Regional EPA person) expressed concern over mil. Him.I f IF..................pi-Awr* the potential revision of the NESHAP without a health effects / basis (Baise, 1980). j The intent of the standard (with regard to emergency releases) could be achieved by using a "bubble concept" total plant control as an endpoint (Holbrook, 1980). It is not always possible to achieve the 100 percent compliance required by the current regulation. There should be "malfunction language" written into a revision. A source should be allowed to offset emissions in another area of the plant until inoperative control equipment is repaired. For example, if the source's primary stripping unit goes down but stripping can also be done in the reactor, the slowing of production to allow longer residence time in the reactor is a self-imposed economic penalty and would allow the source to continue operating under a "short-term bubble." In California (South Coast Air Quality GENC 016610 Management District) a bubble or group of emissions from a plant is allowed. Their discharge limit (50 grams per hour) is based on a Dames and Moore study (Fannin, 1980). The California Air Resources Board (CARB) set an ambient VC stan dard of 10 ppb. This ambient level was based on the lowest possible detectable limit for VC at that time. Dames and Moore then calculated that an emission limit of 50 grams per hour would maintain the ambient level of 10 ppb. (See Chapter 7 for further discussion). PVC plants that are regulated by more stringent requirements should receive a blanket exemption from NESHAP (Holbrook, 1980). NESHAP regulations are designed to protect the public from hazardous air pollutants. However, in some cases, regulations designed to protect the public from nonhazardous air pollutants are more stringent than a NESHAP. For example, a new PVC plant constructed in a nonattainment area for hydrocarbons (an area of the country that does not meet the required criteria pollutant limits) may be forced to control VC emissions more strictly than the VC NESHAP requires. A PSD permit will not be issued unless the new plant can show that the lower levels can be attained. 5.3 STANDARDS FOR EDC AND VC PLANTS Industry feels that there should be separate sections in the regulation specifically for EDC, VC, and PVC plants (Oubre, 1980). The regulation should identify more specifically those requirements applicable to each plant. Regional EPA personnel and industry note that the classification of "in VC service" is a point of contention. 5.4 EXHAUST GASES TO THE ATMOSPHERE Industry feels that an allowance should be made for a reasonable amount of down time for control equipment. Emissions from shutdown may actually be greater than emissions that would occur if the plant were to continue operating (Ledvina, 1980; GENC 016611 5-2 Holbrook, 1980). The shutdown, as well as start-up, of an EDC/VC plant can create more emissions than are discharged by a plant that is allowed to run for a certain period of time without control devices. 5.5 INPROCESS WASTEWATER Under the current regulation, industry is required to measure VC concentration in wastewater following stripping only at the time of start-up of the stripper. Unless monitoring is required on a regular basis, the effectiveness of the wastewater stripper cannot be determined. One region felt that daily sampling of inprocess wastewater (following stripping) should be required to ensure proper stripper operation and maintenance. (One plant in that region, Borden Chemical, routinely tests wastewater stripper VC concentrations.) 'Compliance with the regulations has drastically increased the quantity of'iriprocess wastewater discharged from many plants (Varner, 1980). * The wording of the current inprocess wastewater definition includes that water used to seal gasholders. This seal water should be made exempt from the inprocess wastewater definition (Wyatt, 1980). 5.6 REACTOR OPENING LOSS (ROL) Industry and regional EPA personnel see a need for a separate standard for processors stripping in the reactor. The current regulation assumed the stripping operation was separate from the polymerization reactor (i.e., took place in a different vessel), which is not always the case. The resulting problem is that the test method developed for measuring the ROL is not applicable to plants stripping i_n the reactor. It is possible to use the headspace volume of gas above the resin slurry, but the slurry and reactor are too hot to measure immediately, and after the resin is stripped to the required limit, it continues to emit VC into the headspace (Wyatt, 1980). The proposed revision of the standard would be based on a combination of 5-3 6ENC 018612 ROL requirements and resin stripping levels (Ledvina, 1980; Brittain, 1980; Varner, 1980; Holbrook, 1980). Currently, processors stripping in the reactor show compliance by stripping to the required levels and calculating the ROL emissions. The only problem with this method is that the calculation does not show valves that may be leaking back into the reactor, causing emissions to be higher than calculated. Industry would like to see a provision for monthly or semi-annual averaging of the ROL (Holbrook, 1980). This would apply mainly to bulk processors and other plants stripping in the reactor. Concern was expressed about the prevalence of "ritualistic sampling," in which the reactor is opened before analytical results of the resin samples are available and/or slurry transferred to another vessel. The current regulation requires that the resin must sampled to determine compliance with stripping levels; but before the analysis can be completed, the slurry has already been transferred to the next vessel (Pucci 1980). Regional EPA personnel see a need for a prescribed method of testing for opening of the Pre-polymerizer (Pre-Po) in the bulk process. Because the Pre-Po is involved in only 10 percent of the polymerization reaction and is currently defined as a "reactor" under the standard, there are problems associated with determining compliance for these vessels; i.e., should equipment opening loss or reactor opening loss standards be applied. Post-polymerization (Po-Po) reactors are opened after every batch, whereas Pre-Po reactors are not opened as frequently (Brittain, 1980). The intention of the regulation was that the Pre-Po and Po-Po combined meet the ROL standard. Industry currently applies the ROL to each reactor. 5.7 RELIEF VALVE DISCHARGES The primary concern about relief valve discharges, shared by industry and regional EPA personnel, is how to determine what GENC 014613 5-4 is "preventable." This point was raised more,frequently than any other by EPA regions and industries. All of the industrial representatives stated that the zero concept for relief valve discharges was unrealistic. Many felt that frequency of discharge and the quantity of VC dis charged should be the basis for determining the level of 7compliance (Holbrook, 1980; Laundrie, 1980). Industry pointed out that controls for dispersion, suspension, bulk, and solution processes differ. They suggest that the prevalence of reactor discharges by resin types should be studied (Holbrook, 1980). Industrial personnel have cautioned against recommending gasholders as containment for relief valve discharges. They regard this as a safety and economic issue (Holbrook, 1980; Laundrie, 1980; Ledvina, 1980). An opinion expressed by industry is that 100 percent compliance 100 percent of the time should not be required and cannot be attained. They suggest that a malfunction clause be incorporated (Holbrook, 1980; Ledvina, 1980). Industry suggested that the VC standard could be made more consistent with other standards (e.g., proposed benzene) standards with regard to control of equipment breakdown that results in emergency discharges of a certain level. Two regions pointed out that due to substitutions of other devices in place of relief valves (e.g., double rupture discs) the standard should be reworded to change "relief valve (discharge)" to "relief device, including but not limited to ..." (Brittain, 1980). 5.8 RESIN STRIPPING Industry feels that the language used to describe resin "grade" is ambiguous and this affects the resin stripping regulations. Industry and Regional EPA personnel state because each resin grade has unique characteristics that require different stripping techniques, allowance should be made in the standard for 5-5 GENC 016614 stripping levels reflecting pertinent resin requirements (Holbrook, 1980; Laundrie, 1980; Shaul, 1980). Industry would like relief from the daily calibration and monitoring requirements of the current regulation (Wyatt, 1980). For example, the suggestion was made by industry that longer averaging periods should be permitted for determining stripping level compliance (Laundrie, 1980; Mercier, 1980). Daily compliance with resin stripping levels could be shown by reporting certain process parameters (e.g., vacuum pulled, temperature, time in stripper) that industry has proven achieve the required levels. ' Sampling could then be required on a less frequent basis to support the process parameters. (Note: The preamble to the regulation, published in the Federal Register, does make provisions for using process parameters to demonstrate compliance. The preamble states that, "For both reactor opening and improved stripping, it is possible that the relationship between the emissions measured and the corres ponding operating procedures used to attain the emissions measured can be established.11 (40 FR 59543.) One region suggested the possibility of applying a New Source Standard (under NESHAP) for stripping with retrofitting require ments for existing sources. This would encourage the industry to pursue the technology available for reducing emission levels below current requirements. Several plants are stripping suspension resins to less than 400 ppm by continuous stripping (Brittain, 1980). Sampling and analytical requirements for resin stripping in the solution polymerization process need to be reviewed (Brittain 1980). The stripping requirements (in the current regulation) were developed for the removal of unreacted VC from solid particles of PVC. However, there is no particulate resin form for solution resins. Solution resin stripping involves a distillation mechanism rather than a diffusion mechanism as in solid particles. Accordingly, solution resins can be stripped to an average level of 10 ppm or less. Therefore, the sampling 5-6 GENC 018615 and analytical procedures to ensure that the 400 ppm level is met for solution resins are relatively extensive. 5.9 SOURCES AFTER THE STRIPPER Regional EPA personnel suggest that dryer emissions be reevaluated. Based on a Prevention of Significant Deterioration (PSD) application, an average suspension plant producing 90 megagrams (100,000 tons) of PVC resin per year will emit 36 megagrams (40 tons) of VC per year out of the dryer stack. Indirect I9C drying (which would reduce drying air and make add-on control ^ more economical) may be a better way to reduce these emissions (e.g., one PVC plant uses steam coils that indirectly heat the resin in a rotary dryer) (Pucci, 1980). Blend tanks and centrifuges, the next steps following the stripping operation, have been suggested as potentially significant emission sources. One region's PSD Best Available Control Technology (BACT) analysis shows that 20 to 40 times as much VC is allowed to escape from process units following the stripper as is allowed from R0L and the 10 ppm exhaust emissions (Varner, 1980). 5.10 FUGITIVE EMISSIONS One region proposed that plants submit representative data on ambient (background) levels in annual report form, similar to the draft generic standards. This would provide a means to determine the plant's progress in lowering background levels and to establish criteria for evaluating fugitive control. A certain percent deviation would be permitted and if the plant achieved an average of a prescribed level, it would be in compliance (Brittain, 1980b). The submission of an annual report as mentioned above would also allow the responsible EPA Region to reevaluate the Leak Detection and Elimination Programs and, thus, reject a previously approved program (Flynn, 1980), An industrial source maintains that the levels for fugitives predicted in the Standard Support and Environmental Impact GENC 016416 5-7 Statement document for the original standard were considerably higher than actual levels. This company has determined the relationship of fugitive emissions to production rate and has measured suspension and dispersion process fugitives to be 25 percent of the predicted levels for their older small reactor suspension and dispersion plants (Holbrook, 1980). 5.11 LEAK DETECTION AND ELIMINATION PROGRAMS Many industrial sources, as well as regional EPA personnel, regard the definition of a leak in the standard as unclear. "Small" and "major" leaks need to be defined. The setting of a de minimus level was suggested (Yonge, 1980; Oubre, 1980; McNair, 1980; Pucci, 1980; Wu, 1980; Flynn, 1980). As previously mentioned, regional personnel have no way of determining the success of the programs in lowering fugitive emissions background levels in the plants (Brittain, 1980). One region commented that the location of monitors, while serving 0SHA purposes, may not be optimal for fugitive detection (e.g., many sources do not monitor storage areas) (Pucci, 1980). Several industrial representatives said that their Leak Detection Programs, submitted to Regional EPA offices, are rarely acknowledged (Yonge, 1980; Oubre, 1980; DeBernardi, 1980). 5.12 EMISSIONS TESTING AND ANALYSIS Region II commented that the units for allowable emissions from equipment opening are inconsistent. For larger pieces of equipment, such as a surge tank serving an incinerator, the allowable emissions should be similar to the R0L standard. Only the smaller pieces of equipment (e.g., loading/unloading lines) should have the 2 percent by volume or 25 gallon cutoff (Pucci, 1980). The continuous emission monitoring standard for the primary control device does not specify whether a "minute to minute" monitor (as for a NSPS) or a sequential leak detection monitor is required (Pucci, 1980). No performance specification GENC 016617 5-8 exists for these continuous emission monitors in the current regulation; this requirement of the regulation should be made consistent with the NSPS requirements so that continuous monitoring results can be used for determining violations of emission standards (Varner, 1980; Pucci, 1980). Sampling regulations for RVC are currently under the section (61.70(c)(2)(ii) for semi-annual reporting in the standard. It would be more appropriate to locate this regulation under testing (Section 61.67) and monitoring (Section 61.68) (Brittain, 1980). Sampling bag sizes, specified in methods 106 and 107, are too large for the sample (Ledvina, 1980). Methods 106 and 107 call for "zero-grade" gas for analysis. This method is very expensive to use on a day-to-day basis. Stack sampling should utilize zero-grade gas, but it is not necessary for daily analyses (Holbrook, 1980). Industry feels that instrumental calibration on a daily basis is not necessary. Drift history, etc., should be allowed to establish the reliability of the instruments (Holbrook, 1980; Oubre, 1980; DeBernardi, 1980). One region said that a source reported calibration requirements for a daily span check actually caused an increase in emissions (Pucci, 1980). Industry would like to establish its own program for calibration procedures with periodic checking of internal standard operating procedures by regional EPA personnel (DeBernardi, 1980). Some industrial representatives feel that the standard should not restrict analysis to specific instruments. There are questions as to reliability and availability of parts for some of the required instruments (Ledvina, 1980). Method 107 for R0L measurements is not considered to be reliable. Corrections for condensing water vapor in the sample cannot be made reliably (Ledvina, 1980). Region VI suggested the need for a monitoring performance standard for backup as well as primary control devices. When 5-9 VC emissions are bypassed to the atmosphere (e.g., in the case of a malfunction or shutdown of a primary control device), a backup incinerator, or a flare, the downstream monitor does not measure any VC. Region VI feels that monitors for VC should be located so that actual emissions of VC to the atmosphere are always measured during all process operating conditions (Harrison, 1979). 5.13 REPORTING Industry made several comments regarding the lack of uniformity in reporting requirements among the regions. Frequency of reports and level of detail required were given as examples of differences (Holbrook, 1980; Laundrie, 1980). Industry suggests that the possibility of consolidation of permits be considered (Holbrook, 1980). Many comments were made by Regional EPA personnel regarding semi-annual report requirements. JheTregions do not consider., the semi-annual reports effective for enforcement purposes. f They feel that excursions should be reported on a 10-day basis or immediately, with a minimum frequency of four times a year (Thompson, 1980; Aronson, 1980; Varner, 1980). Industry feels that semi-annual reporting should be replaced with exception reporting (Holbrook, 1980). 5.14 RECORDKEEPING Industry feels recordkeeping should be reduced to include only exception or noncompliance data for the pertinent time frame (Holbrook, 1980). This would reduce the amount of records V required to be maintained onsite. One region pointed out that the standard does not specify a time limit for retention of R0L records (Varner, 1980). -5.15 NESHAP APPLICABILITY DETERMINATIONS Since promulgation of the VC NESHAP, several inquiries have been made regarding the applicability of VC-use situations to the standard. The following are examples of the type of inquiries directed to the Division of Stationary Source Enforcement (DSSE) (Reich, 1980; King, 1980). 5-10 GENC 016619 Question Are duplicate continuous monitoring systems needed for vinyl chloride plants? Determination No Is EPA concerned with short-term monitoring malfunctions? Are relief valve discharges which are due to operator error considered violations of section 61.65(a)? Yes Conditional Can a demonstration of noncompliance of the reactor opening loss standard, using an unapproved test method, support an enforcement action? When is a relief valve discharge a violation of section 61.65(a)? No What measures can be taken to prevent relief valve discharges? Discussion Although backup moni toring equipment is not required, it is the responsibility of the source to ensure that the monitoring equipment operates continuously. The vinyl chloride stan dard does not provide for monitoring malfunctions of any duration - EPA is concerned about any malfunction. Relief valve discharges resulting from operator error are considered violations if the errors are preventable. The fol lowing are examples of preventable operator errors: 1. Errors due to lack of training 2. Negligence If a source is using a test method which has not been approved by EPA as an equivalent or alter nate test method, information pertinent to the method should be submitted to EPA for evaluation. A relief valve discharge is a violation if it could have been antici pated and preventative measures could have been taken or prevented by proper operating, main taining and inspecting equipment. Examples of measures that can be taken to prevent relief valve discharges include: 5-11 GENC 016620 Question Determination Discussion 1. Properly instru menting the reactor to detect upset conditions. 2. Injecting chemicals to stop polymerization reaction during upset conditions. 3. Venting reactor con tents to a gasholder and ultimately to a recovery system. 4. Maintaining a backup source of power. 5. Proper training of employees. Are vinyl chloride tank cars that are transported by rail to a vinyl chloride plant, "equip ment in vinyl chloride service"? No Tank cars are not subject to the vinyl chloride regulations, except to the extent that the requirements for purging of loading and unloading apply. Is a PVC sludge-drying facility, designed to accommodate sludge with some low residual VC content, subject to the VC NESHAP regulations? No The regulation applies to plants which produce EDC, VC, or PVC. The standard specifically places requirements on the reactor, stripper, con tainers (mixing, weighing, and holding), monomer recovery system, and sources following the stripper. If all dis charged material meets the requirements of this subpart prior to its exposure to the atmos phere, this subpart does not apply to the PVC sludge drying facility because the process is not considered part of the production processes of EDC, VC, or PVC plants. These are representative questions directed to many of the EPA Regions and are indicative of the need o clarify some points in the regulation. 5-12 GENC 016621 5.16 REFERENCES FOR CHAPTER 5 Aronson, Wayne, EPA Region IV, Air Enforcement Branch. Meeting report TRW visit to Region IV offices. September 3, 1980. Baise, Gary, Attorney, Beveridge, Fairbanks and Diamond. Meeting report TRW/EPA/SPI meeting at Durham, N. C. July 31, 1980. Brittain, Martin, NESHAP Coordinator, EPA Region VI. Meeting report TRW visit to Region VI offices. October 27, 1980. DeBernardi, James, Plant Manager, Lake Charles, La. Conoco Plant. Meeting report - TRW visit to Conoco Plant. August 7, 1980. Fannin, James, B. F. Goodrich Chemical Division. Meeting report - TRW visit to B. F. Goodrich Cleveland office. October 30, 1980. Environmental Protection Agency. 1975. "NESHAP Proposed Standard for Vinyl Chloride," Federal Register, Vol. 40, No. 248. December 24, 1975. Flynn, Peter M. Environmental Engineer, Air Facilities Branch, EPA Region II. Letter to J. W. Bodamer, Jr., December 2, 1980. Fradkoff, Steve, Region I EPA. Meeting report - TRW visit to Region I offices. August 13, 1980. Harrison, Arlene, Regional Administrator, EPA Region VI. Letter to Don Goodwin. January 29, 1979. Holbrook, W. C., Director of Toxicology and Environmental Affairs, B. F. Goodrich Chemical Division. Trip report - visit to the Pedricktown Polyvinyl Chloride Plant. September 17, 1980. King, J. A., DSSE, USEPA. Letter to Regina E. Thompson, EPA Region III, 1980. Laundrie, Robert W., General Tire and Rubber Co. Trip report - TRW visit to the Ashtabula, Ohio PVC plant. September 10, 1980. Ledvina, Joseph C., Director of Environmental Activities, Conoco, Inc. Meeting report - representatives from Conoco, TRW and EPA. October 17, 1980. McNair, Cathy, NESHAP Coordinator, EPA Region I. Meeting report - TRW visit to Region I offices. August 13, 1980. Oubre, Robert, Technical Manager, Dow Chemical Corp. , Oyster Creek Division. Trip report - TRW visit to Oyster Creek plant. August 5, 1980. Pucci, Michael, NESHAP Coordinator for EPA Region II. Meeting report TRW visit to Region II offices, New York. August 12, 1980. 5-13 GENC 016622 Reich, Edward, DSSE, USEPA. Letter to EPA Regional Directors regarding Summary of NESHAP Determinations, 1980. Schaul, Peter, EPA Region III. Meeting report - TRW visit to Region III offices, Philadelphia. September 16, 1980. Thompson, Jean, NESHAP Coordinator, EPA Region III. Meeting report TRW visit to Region III offices. September 16, 1980. Varner, Bruce, NESHAP Coordinator for EPA Region V. Meeting report TRW visit to Region V offices, Chicago. August 19, 1980. Wu, James, NESHAP Coordinator for Region IV. Meeting report - TRW visit to Region IV offices. September 3, 1980. Wyatt, Susan R. 1980. Office of Air Quality Planning and Standards. U.S. EPA Meeting report - TRW visit to 0AQPS. June 16, 1980. Yonge, John, Environmental Control Supervisor, Shintech, Inc. Trip report - TRW visit to Shintech Plant, Freeport, Texas. August 5, 1980. 5-14 GENC 016623 6.0 UNREGULATED SOURCES OF VINYL CHLORIDE 6.1 INTRODUCTION The current regulation is applicable to the following types of facilities: (1) plants producing EDC by the reaction of oxygen and hydrogen chloride with ethylene, (2) plants producing VC by any process, and (3) plants producing one or more polymers containing any frac. n of VC. There are, however, several categories of VC-emitting facilities that are not regulated under the current VC NESHAP. Many of these sources were identified during the original study. These include PVC compounders and fabricators, and processors who use VC as a chemical intermediate or produce it as a byproduct. New sources of unregulated VC emissions have been identified during this review study. They include mobile-mounted sources, nonplant transfer facilities, solid waste drying facilities, and disposal sites. Many of the sources not regulated by the VC NESHAP are subject to state hydrocarbon-emission control standards, specifically those plants located in nonattainment areas. 6.2 SOURCES IDENTIFIED DURING THE ORIGINAL STUDY 6.2.1 Fabricating Operations Following polymerization of VC, two major processes are involved in the conversion of VC to a finished PVC product - compounding and fabricating. Compounding involves the mixing of PVC resins with additives such as plasticizers, stabilizers, pigments, blowing agents and anti-oxidants. These additives impart certain properties that are required for handling the polymer during fabrication as well as in the final product. Compounding GENC 0L6624 may be done by the fabricator, the resin producer, or by independent compounders. Fabrication of the polymer consists of melting and shaping the compound by various processes. Some of the major processes are: Extrusion - This process consists of mixing and melting a continuous stream of plastic and generating sufficient pressure to force the compound through a die. Calendering - The compound is fed through sets of rollers to form continuous sheets of plastic. Molding - Several types of molding are used; injection, compression, vacuforming, and embossing constitute the more common processes. Bonding - The joining of two or more pieces of PVC can be accomplished using heat (or heat and pressure), adhesives, or hot gas welding. Most of the residual VC (RVC) in the resin is lost at the PVC plant. Following stripping operations (which remove RVC to levels at or below those required by the regulation) further losses of RVC occur in drying, bagging, and storage operations. More RVC can be lost while the resin is in transit. Thus, operations following the stripper (and prior to compounding operations) account for the reduction in RVC levels entering compounding and fabricating facilities as compared with those found in the stripped resin. The application of heat and pressure during fabrication can cause some of the remaining RVC to diffuse from the resin particles. A.D. Little (1975) prepared a report on VC emissions from PVC processing industries for the EPA in August 1975. This report quantified the VC emissions from compounders and fabricators of PVC resins. The conclusions drawn from the study were: There are more than 8000 PVC fabricating facilities in the United States, e Emissions from compounding and subsequent fabrication processes together accounted for less than one-half of one percent of the total United States emissions, and GENC 016625 6-2 r r r i. I E I I 0 r l r <LJ 3 r i t L # The most promising control technique to limit VC emissions to the atmosphere from compounding and fabricating facilities appeared to be further reduction of RVC levels in incoming resins. At the time of the 1975 study, average VC content of raw resin was reported to be 300 ppm (in 1974), with PVC g production rates of 2.0 teragrams (4.4 x 10 pounds). Total VC release from compounding and fabricating facilities during 1974 was 600 megagrams (1.3 x 10^ pounds). Since promulgation of the OSHA VC workplace standard (permissible occupational exposure level), PVC manufacturers have reduced the RVC content in the resins supplied to compounders and fabricators. JffcT?# ` control*Vc'*exposure in fabrication^faci 1 ities^theTpVC^industry'has / ^estabVisjed_a.j10 ppm VC concentration'1)imit in' dried PVC resins. / Surveillance and enforcement of this"requirement has been delegated to the Plastic Pipe Institute (Cameron et al., 1980, p. 43). In addition to the influence of the OSHA standard, requirements by the Food and Drug Administration (FDA) have contributed to the reduction of RVC levels in resins supplied to fabricators. PVC products that will come into contact with humans are required to have very low levels of RVC. Therefore, resins to be used for blood bags, food wrap, drug and beverage bottles, etc., enter the fabricating facilities with RVC levels well below 1 ppm. Bottle resin, for example, is currently provided at levels less than 0.05 ppm (Ter Haar, 1980.) Of the PVC plants surveyed, all report that average PVC levels in resin leaving the plant range from <0.002 ppm to 10 ppm. Many plants report levels of <0.5 ppm to 1 ppm. Typical reductions in resin RVC since promulgation of the VC NESHAP are shown by the following data from a PVC resin manufacturer (Ter Haar, 1980): Resin type Emulsion Resins Bottle Resins Other Suspension Resins RVC Levels (ppm) Prior to 1974 Current 3.0 0.5 <0.05 935 2.6 6-3 GENC 016626 ""N In 1979, PVC production rates were reported to be 2.8 teragrams (6.1 x q 10 pounds) (Chemical and Engineering News, 1980, p. 13). Assuming that the average RVC content of the resins supplied to compounders and fabricators is 10 ppm and that all of the VC is lost to the atmosphere during fabrication (worst case assumption), total emissions from these 4 sources for 1979 would have been 28 megagrams (6.1 x 10 pounds) VC. The emission levels from these sources would actually be much lower than those cited above since average RVC content of incoming resin is probably less than 10 ppm. In addition, VC migration studies indicate that a very low percentage of monomer is released during fabrication. During extrusion, for example, 10 percent of the monomer is typically released (Ter Haar, 1980). EPA conducted ambient air studies in which five PVC fabrication plants were monitored. There were no measured concentrations of VC emitted from three of these plants. The highest measured concentration was 0,006 ppm, 24-hour average (Padgett, 1980). 6.2.2 Miscellaneous Sources In a study done by Arthur D. Little, Inc. for the EPA (Lyman, 1976), miscellaneous sources of VC emissions were categorized as follows: Industrial processes in which VC is used as a chemical intermediate for the production of other chemicals, Industrial processes in which VC is used as a minor constituent (<50 percent by weight) for the production of resins, and Industrial processes in which VC is produced as a byproduct of the chemical reaction involved. The second category, use of VC as a minor constituent in resin production, no longer represents an unregulated source, as these processes are now subject to the current VC NESHAP regulations (i.e., any amount of VC used for polymerization constitutes a regulated process). Within the first category (use of VC as a chemical intermediate), two main sources were identified by the Arthur D. Little study: (1) production of 1,1,1-Trichloroethane (1,1,1-TCE) and 1,1,2-Trichloroethane (1,1,2-TCE); and (2) production of other special chemicals such as certain pesticides. GENC 016627 6-4 A more recent survey conducted by Conoco (McPherson, 1979, p. 75) attributes three to four percent VC usage to the manufacture of 1,1,1-TCE. Information pertaining to the current status of VC emissions from the manufacture of 1,1,1-TCE and 1,1,2-TCE was not obtained during the review study. VC emissions from pesticide manufacturing facilities were identified in the Arthur D. Little report. Total VC consumption was estimated to be only two percent of the amount estimated for TCE production. One of the two plants cited in the study, an insecticide manufacturing facility, was contacted during the current VC NESHAP review study. It appears that VC consumption, control technology, and emissions are essentially the same as they were when the Arthur D. Little study was done (Vines, 1981). At that time, VC emissions to the atmosphere were estimated to be about 0.14 kilograms (0.3 pounds) per day. The third category (VC as a byproduct) encompassed the following processes existing in the United States: (1) the manufacture of EDC via oxychlorination, and (2) the manufacture of ethylene amines and ethylene imines from EDC. The first process is now regulated under the existing VC NESHAP. Current information on VC emissions from ethylene amine production confirms that VC, as a byproduct in these processes, represents a very minor source of emissions (actual amounts not known). A representative of a plant manufacturing ethylene amines stated that the small amount of VC involved is directly vented to the incinerator in one plant. In another plant, VC is sent to the VC recovery column at their PVC facility. In either case, storage and handling of VC is not a factor. VC fugitive emission surveillance in these plants has shown no detectable VC levels (Wise, 1981). 6.3 NEW SOURCES IDENTIFIED DURING THE REVIEW STUDY 6.3.1 Mobile-mounted Sources of Emissions Three areas of mobile-mounted emission sources have been noted by some regional EPA personnel. These are rail''carsj tank cars, and marine unloading facilities. 'Department of Transportation (DOT) and Coast 6-5 GENC 016628 Guard regulations are primarily concerned with flammability and water pollution parameters. No designation of responsibility for VC emissions into the air has been assigned for rail car leaks (Aronson, 1980). Several companies have reported relief valve discharges associated with the unloading of VC from ships at marine unloading facilities (Brittain, 1980). Facilities used for cleaning rail and tank cars transporting liquid VC are often remote from regulated sources. There is a question as to whether these facilities or the VC supplier should assume responsibility for resultant VC emissions. 6.3.2 Nonplant Transfer Facilities Terminals for transfer and short term storage of VC from marine vessels are under the jurisdiction of the Coast Guard and are not regu lated by the VC NESHAP. The significance of emissions from these intermediate facilities is a concern of Regions I and VI (Pucci, 1980; Brittain, 1980). Because 16 of the 18 operating EDC/VC plants are located in Region VI and most of these plants are proximate to marine waters, this potential source of VC emissions is of major concern to Region VI as well as the plants who normally take responsibility for the emissions. In addition, many sources are clustered in the northeastern United States and VC can be transferred by marine vessels to transfer facilities near the sources. One of these transfer facilities was identified in Region II, but has temporarily been shut down. The unregulated source of emissions is usually from a safety relief valve on the marine vessel (barge or ship) that discharges when overpres surization occurs during loading and unloading. One plant in Region VI reported a total of 871 kilograms (1936 pounds) of VC emitted from marine vessel relief valves over a 3.5 year period since 1977. This is comparable to 58 percent of the total quantity of relief valve discharges reported during this same time period. However, in most cases, the discharges from marine vessels are not reported because this source of emissions is not regulated (Brittain, 1980). Tank farms, used for temporary VC storage and potentially for emergency stockpiling (due to rail strikes, etc.), are not currently regulated under the standard (Schaul, 1980; Varner, 1980). GENC 016629 6-6 I 1 l t L, < L L r f t- 6.3.3 Solid Waste Drying Facilities One facility is currently installing a PVC sludge drying operation, designed to accommodate sludge with some low residual VC content. Such facilities are considered to be potential VC emission sources (Schaul, 1980). However, the sludges to be dried at these facilities contain PVC- resins that have already met requirements-for-resin stripping levels. 6.3.4 Oisposal Facilities (Landfill) / &&S VC emissions were recently detected in vents from a landfill in 0 6- Region II. These vents were installed mainly for the release of methane from the landfill. It is thought that PVC wastes, generated prior to the VC NESHAP's stripping requirements, were disposed of in the landfill (Pucci, 1980). Measurements made at the landfill vents showed levels as high as 90.2 ppm. A review of the design proposed for the new venting system for the landfill included an evaluation of the number and spacing of the vents as well as an acceptable means for dealing with the VC emissions. The recommendation report stated that a manifold burner system would substantially reduce VC concentration, although the report did not state the level of reduction. The burner systems suggested included a Hirt Ground Flare and other waste gas burners. Interim control devices are being evaluated, and these include activated charcoal cannisters and activated charcoal tubes. VC has been measured in another landfill nearby, and Region II suspects that many more exist (Spatola, 1981) . Disposal of off-specification batches of PVC resin has been identified as a potential VC emission problem. These batches cannot always be stripped by conventional methods and thus may contain high levels of RVC. Some off-specification batches can be sold, but many are discarded. The State of California requires that all of these batches be stripped to levels appropriate to that resin type, but industry states that this is not always possible (Fannin, 1980). Ultimate disposition of the off-specification batches has not been determined. 6-7 GENC 016630 6.4 REFERENCES FOR CHAPTER 6 Aronson, Wayne. 1980. EPA Region IV, Air Enforcement Branch. Meeting report - TRW visit to Region IV offices. September 3, 1980. Brittain, Martin. 1980. NESHAP Coordinator, EPA Region VI. Meeting report - TRW visit to Region VI offices. October 27, 1980. Cameron, J. B., A. J. Lundeen, and J. H. McCully, Jr. 1980. "Trends in Suspension PVC Manufacture." Hydrocarbon Processinq. March 1980. Chemical and Engineering News. 1980. "Key Polymers." October 6, 1980. Fannin, James. 1980. B. F. Goodrich. Meeting with TRW in Cleveland, Ohio. October 30, 1980. Little, Arthur D., Inc. 1975. Vinyl Chloride Monomer Emissions From the PVC Processing Industries. Contract No. 68-02-1332, Task No. 10. August 1975. Lyman, Warren J., Arthur D. Little, Inc. 1976. Miscellaneous Industrial Sources of Vinyl Chloride Emissions in the U.S.. Contract No. 68-02-1332, Task No. 13 (Part 1-A, B, CjT"March 1976. McPherson, R. W., C. M. Starks, G. F. Fryer. 1979. "Vinyl Chloride Monomer . . . What You Should Know." Hydrocarbon Processing. March 1979. Padgett, Joseph. 1980. Director SASD, EPA. Letter with enclosures to A1 Montague, SAD, EPA. September 23, 1980. Pucci, Michael. 1980. NESHAP Coordinator for EPA Region II. Meeting report - TRW visit to Region II offices, New York. August 12, 1980. Schaul, Peter. 1980. EPA Region III. Meeting report - TRW visit to Region III offices, Philadelphia. September 16, 1980. Spatola, Joseph. 1980. EPA Region II, Air and Hazardous Materials. Telecon with M. A. Cassidy, TRW. January 13, 1981. v Ter Haar, Gary L. 1980. Director of Toxicology and Industrial Hygiene, Ethyl Corporation. Letter with attachments to Docket Officer, D0L - 0SHA. May 7, 1980. Varner, Bruce. 1980. NESHAP Coordinator for EPA Region V. Meeting report - TRW visit to Region V offices, Chicago. August 19, 1980. Vines, J. H. 1981. Manager Insecticide Production, Chemagro Division of Mobay Chemical Co., Kansas City, Mo. Telecon to M. A. Cassidy, TRW. January 6, 1981. Wise, R. C. 1981. January 7, 1981. Union Carbide Corp. 6-8 Telecon to M. A. Cassidy, TRW. GENC 016631 7.0 IMPACT OF OTHER REGULATIONS 7.1 INTRODUCTION The original study, which was done to support the current regulation, found that existing regulations as well as proposed regulations had little effect, if any, on the reduction of atmospheric VC emissions from EDC/VC and PVC plants. The Occupational Safety and Health Administra tion's (OSHA) regulations required a combination of ventilation techniques, engineering and work practice controls, and respirators to reduce worker exposure. Many of the engineering controls reduced atmospheric emissions, for example, portable and fixed point monitoring, improved sealing techniques, transfer line purges, reactor cleaning methods, and improved stripping, which not only reduced emissions from PVC plants but satisfied the demands of the fabricators (EPA 1975, p. 9-5). However, it was felt that compliance with the standard would not be uniform throughout the industry or significantly reduce VC emissions to the atmosphere. Some state regulations existed for hydrocarbons and new construction and they indirectly reduced VC emissions at some plants. Texas regulations required two EDC/VC plants to reduce hydrocarbon emissions from the oxychlorination reactor, thus indirectly reducing VC emissions. Louisiana had similar regulations. New Jersey and Texas both required what was specified by each state as best control technology for any pollutant, including VC, when a source was newly constructed or modified (EPA 1975, p. 9*10). Other regulations concerning water pollution, transport of VC, aerosol products, and food packaging had no effect on reducing VC emissions. However, since promulgation of the standard in October 1976, other regulations have evolved that can have a potentially greater effect on VC emitted to the total environment. The following is a list of the new GENC 016632 regulations, policies and requirements that have been established or proposed. Policy and Procedures for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer (Proposed Carcinogen Rule) Prevention of Significant Deterioration (PSD) Resolution of the South Coast Air Quality Management District Board (SCAQMD) adopting Rule 1005.1 - Control of Vinyl Chloride Emissions Resource Conservation and Recovery Act (RCRA) Toxic Substances Control Act (TSCA) t Toxic Pollutant Effluent Standards as required by the Clean Water Act of 1977 Proposed Primary Drinking Water Regulations Potential Revision to the OSHA Workplace Standard for VC Transport of Hazardous Wastes and Hazardous Substances Food and Drug Administration (FDA) regulations Other state and local air pollution regulations These new regulations have made the acquisition of permits necessary for construction and operation increasingly complex. The effect of the new regulatory requirements of the VC NESHAP will be discussed below with the federal and state laws responsible for implementation. 7.2 CLEAN AIR ACT The Clean Air Act (CAA) Amendments of 1977 provided a mechanism for instituting a program for Prevention of Significant Deterioration (PSD) of Air Quality and plans for nonattainment areas. These regulations provide for continued protection of the existing ambient air quality and, in some cases, have had an effect on reducing VC emissions from new and modified sources. The CAA recently proposed a rule for regulating airborne carcinogens. The proposed Carcinogen Rule directly affects any recommended revisions to the current standard that may result from this review study. In addition, the state of California has established the first ambient air quality standard for VC and the regulations necessary for its enforcement. GENC 016633 7-2 7.2.1 Carcinogen Rule The Carcinogen Rule, proposed on October 10, 1979, considered policies and procedures to: determine the carcinogenicity and risks for a specific pollutant, establish priorities for regulatory action, specify degree of control, and provide public input to the decisionmaking process. The proposed rule's requirement for periodic review of NESHAP regulations triggered this review of the current VC regulation. At least every 5 years, regulations would be reviewed for possible modification incorporating technological developments and health effects information. These reviews provide the opportunity to consider revising the standard (EPA, 1979a). 7.2.2 Prevention of Significant Deterioration The original, and more recently revised, PSD and related nonattainment regulations have been responsible for reducing VC emissions from EDC/VC and PVC plants, in certain respects, beyond the reductions required by the current VC NESHAP. The goal of PSD is to ensure that air quality in clean areas does not significantly deteriorate and yet maintains a margin for future industrial growth. Clean areas, or those areas meeting the National Ambient Air Quality Standards (NAAQS) for criteria pollutants, are classified as attainment areas. New construction or a modification to an existing applicable source may be subject to PSD review and specific required analyses. New construction or a modification to an existing source in a nonattainment area (an area not achieving the NAAQS) must be reviewed in accordance with the nonattainment provisions of the applicable State Implementation Plan (SIP). SIP's represent the plan of action that a state follows to restore non-attainment areas to attainment areas. EPA is continuing to publish control techniques guidelines (CTG) for those industries that emit significant quantities of air pollutants in areas of the country where NAAQS are not being achieved. CTG's provide infor mation to state and local agencies that can be used in maintaining air quality (i.e., for developing an SIP). The CTG identifies reasonably available control technology (RACT) that can be applied to the industries GENC 016634 7-3 to reduce emissions, taking into account technological and economic feasibility. No CTG currently exists that would directly affect a regulated VC source. CTG's are in the developmental stages for air oxidation processes, polymers and resins, which could affect VC sources in the future. PSD review is required for sources locating in PSD areas, in areas designated attainment, or in areas that are unclassifiable for any criteria pollutant. PSD areas, however, can also be designated non attainment for one or more pollutants. In such areas, significant increases in pollutants for which the area is designated nonattainment under Section 107 of the CAA are exempt from PSD review. If this is the case, the facility will still be subject to review according to that state's applicable SIP. Therefore, a newly constructed EDC/VC and/or PVC plant (or modification to an existing plant) may be subject to PSD or non-attainment review or both. PSD applicability is first determined for the new source or for the modification to an existing source. A new source is subject to PSD review if it is: (1) one of the 28 listed sources with the "potential" to emit 100 tons per year or more of a regulated pollutant, or (2) any unlisted source with the "potential" to emit 250 tons per year of a regulated pollutant. Regulated pollutants are the five criteria pollutants and nine non-criteria pollutants (of which VC is one). "Potential" emissions incorporate controls, any federal or state permit requirements and fugitive emissions. A modification to a major source is subject to PSD review if the physical change or change in operation results in a significant net emissions increase. The significance level above which PSD review is required is 1 ton per year for VC and 40 tons per year for volatile organic compounds. Once it is determined that the source is subject to PSD review, the following three analyses are required: Best Available Control Technology (BACT) analysis, air quality analysis, and additional impact analyses. The BACT analysis is the most important requirement and it provides the data for the other two requirements. Because NESHAP regulations do GENC 016635 7-4 not necessarily require BACT, PSD regulations can override NESHAP provisions and the resulting requirements for BACT may represent a more stringent emission control. NESHAP regulations require best available technology (BAT) which differs from BACT in the procedure used for its selection. BAT is selected on a nationwide basis considering economic, energy and environmental impacts, whereas BACT is selected on a plantby-plant basis considering economic, energy and environmental impacts. For this reason, BAT and BACT may not necessarily reflect the same level of control. Therefore, a new VC source undergoing PSD review may be required to implement a level of control (i.e., BACT) that is more stringent than the BAT required by the VC NESHAP. The primary purpose of BACT is to minimize consumption of increments (i.e., allowable growth within an attainment area) and thus expand the area's potential for future growth by addressing the interrelated impacts of energy availability, economy and environment. BACT determinations are made on a case-by-case basis, and the results form the basis for control strategy decisions. A BACT application may exempt a regulated pollutant from PSD review; for example, a company requesting a permit to construct an EDC/VC and PVC plant on the same site calculated potential hydrocarbon emissions to be approximately 5,000 tons per year thus making the source subject to PSD review (actual quantities of VC emissions were not determined). However, after application of BACT, the hydrocarbon emissions, including VC, were reduced to less than 50 tons per year which resulted in an exemption from PSD review. The BACT determination required more stringent control of specific VC emission point sources within the plant to levels less than those required by the VC NESHAP (Winkler, 1980). In another example an EDC/VC plant was allowed to discharge approximately 80 tons per year of VC from the oxy vent under NESHAP. However, the plant chose to incrementally remove 79 tons per year rather than go through PSD review (Brittain, 1980), The BACT process involves four steps. The first, pollutant applicability, has already been discussed. The second step is deter mining the emissions unit applicability for the source. All applicable emissions units must be analyzed. A chemical complex producing EDC, VC and PVC provides a good example. Each chemical and polymer is made in GENC 016636 7-5 separate processing equipment with each piece of equipment containing several emission units. Emissions from all of the units must be summed for the entire complex which then constitutes one source. Fugitive emissions are included in determining quantities of emissions from each unit. Secondary pollutants (i.e., those emissions associated with a source but not emitted from the source itself) are also included if they cause a potential air quality standard or increment violation. The third step in BACT review is to identify sensitive concerns (i.e., local air quality concerns and potential environmental impacts). These concerns should be quantifiable, if possible, so that various control alternatives can be compared. This step also encourages public involvement. The last step is to select alternative control strategies. A base case is first established in order to rank the alternatives and consider them quantitatively. The base case can be considered the case that would be applied in the absence of the BACT decisionmaking process. The choice of the base case is dictated by existing regulations such as New Source Performance Standards (NSPS) or NESHAP requirements. Selection of alternatives is usually based on technical feasibility -- previously demonstrated technology. Innovative technology can be selected also and PSD allows special consideration for its use. With the creation and analysis of the base case, alternative control strategies affording greater degrees of continuous emission reduction than the base case are ranked in order of control efficiency. The applicant then conducts an economic, energy and economic impact analysis for each alternative control strategy. Upon completion of these analyses, the information will be available to perform the final evaluations that will lead to proposal of BACT. The other analyses required after the BACT analysis are an air quality analysis and an additional impact analysis, both of which rely on the BACT results. The air quality analysis must demonstrate that NAAQS or PSD increments will not be violated. This is done for each regulated pollutant and is accomplished by projecting the air quality that would exist when the new construction or modification is operating. Dispersion modeling is used to project this air quality. The modeling 7-6 GENC 016637 \ ' *fJ \ takes into consideration the impact area, other sources in the areas, and existing ambient concentrations. The complexity of the analysis depends on location, and in some cases, ambient monitoring will be required. Finally, the additional analysis considers the impact on soils, vegetation, and visibility in the affected area from the increased emissions. The applicant now has conducted the required analyses and based on the results proposes BACT to the reviewing agency that has responsibility for approval. The reviewing agency's determination is made on a case-by case basis and the emission rates proposed as BACT may not necessarily be the rate ultimately specified in the PSD permit. 7.2.3 NESHAP Delegation to States Section 112(d)(1) of the CAA allows each state to develop and submit to the Administrator a procedure for implementing and enforcing emission standards for hazardous air pollutants for sources in their state. If the Administrator finds the state procedure adequate, he shall delegate to the state the authority under the CAA to implement and enforce the NESHAP standards. Currently, those states that have received NESHAP delegation that contain VC sources are Texas, Georgia, Illinois, and California. The state program must be as least as stringent as the Federal VC NESHAP. The state can also develop a program that is more stringent. California has developed a more progressive program that exemplifies the effect that NESHAP delegation can have on reducing VC emissions below those levels required by the Federal VC NESHAP. The California Air Resources Board (CARB) was delegated NESHAP authority and, in response, adopted a state ambient air quality standard (AAQS) for VC of 10 parts per billion (ppb). CARB chose the 10 ppb level as the state AAQS because it was the lowest detectable limit for VC at that time. Responsibility for implementing the standard was then delegated to the local Districts (county-wide areas) within California. The Districts' mandate is to attain and maintain the AAQS's adopted by California, although the Districts' rules and regulations apply specifically to the sources within their jurisdiction. All of the five VC sources in California are located in the South Coast Air Quality Management District (SCAQMD). In response to the 7-7 0ENC 0U638 10 ppb AAQS, SCAQMD adopted the Federal NESHAP (Rule 1005) as their authority for enforcement of the AAQS and proposed a resolution adopting Rule 1005.1 as their program for controlling VC emissions to 10 ppb. Some of the major differences between Rule 1005.1 and the Federal NESHAP will be discussed in the following subsections. Ambient monitoring. Designated plants, or those plants subject to Rule 1005.1, are not allowed to discharge VC in quantities that result in ambient concentrations greater than 10 ppb, 24-hour average, measured at any point beyond the property line of the plant where people reside and work. Sources are required to operate up to eight air monitoring stations in the vicinity of the plant to ensure that the 10 ppb level is being attained. These stations are selected and approved based on meteorological data, other available monitoring data, and location of populations around the plant. Meteorological data must also be monitored at these stations. Records of all the data must be maintained and monthly summaries submitted to SCAQMD. A plant can reduce the number of required stations if no violations occur in any period of 6 consecutive months. This exemption becomes void if a significant violation occurs. Minor, nonperiodic and infrequent breakdowns may be overlooked. Primary control device. All equipment containing more than 10 ppm VC is required to be vented to the primary control device. The control device must then be operated at an efficiency to limit total emissions from the stack to less than 50 grams per hour. Selection of the emission limit was based on worst case modeling indicating that this level of emissions would maintain the 10 ppb AAQS. The 50 gram per hour limit sets a ceiling on growth (which the VC NESHAP 10 ppm limit does not; plants can continue to grow and discharge larger quantities of VC under the 10 ppm Federal standard). Bubble concept. Rule 1005.1 allows a source to "bubble" or group their emissions to the degree that the 50 gram per hour limit is maintained. If a source requests this method of emissions reduction, all original construction and operation permits must be submitted to SCAQMD and new permits filed for approval. SCAQMD reissues these permits specifying emission levels or any other conditions necessary to ensure all emission limits are met. GENC 016639 7-8 Reactor opening loss (ROL). ROL emissions have been reduced to a 10 ppm concentration rather than the Federal standard of 0.02 gram VC per kilogram of PVC produced. Operational Requirements. All vent valves and relief devices (other than emergency relief valves) upstream of stripping must be vented to a receiving vessel. Off-specification polymer batches must be discharged to a sealed container or stripped to required levels. Failure of a rupture disc preceding an emergency relief valve is a violation unless vented to control equipment. Reactors and other equipment upstream of stripping must be equipped with automatic pressure reduction systems (APRS) that will open when necessary at a setting between normal operating pressures and emergency relief pressures. (Normal vapors from the APRS must be contained.) Management plan. A management plan must be submitted for the reduction of VC emissions. The plan should include, but is not limited to, A plan and schedule to locate and identify all emissions sources that may cause the AAQS to be exceeded; An outline of employee training programs for preventing emissions; A method for screening operating data to identify operators most often responsible for excessive emissions; and An outline of a special training program or other methods to eliminate excessive emissions. Leak detection. A leak is the detection of VC from any location, other than a stack vent or designed equipment opening, from which VC exceeds the background level of 10 ppm (measured 5 centimeters from source). All equipment containing or using VC shall be free of leaks. Equipment is to be inspected on a regular basis and records kept - all leaks are to be eliminated within 24 hours. Any leak detected during a SCAQMD inspection is a violation. New or modified plants. The builder must demonstrate that the ambient air quality will not exceed the 10 ppb AAQS as a result of any emission from a new or modified source. 7-9 GENC 016640 Relief valve discharges. Designated plants must install and operate pressure indicating and recording instruments (or approved equivalents) monitoring the discharge of emergency relief valves and manual vent valves. Data from these instruments must be summarized monthly and submitted to SCAQMD. Current status of Rule 1005.1. The above subsections represent a summary of the major differences between the Federal NESHAP and the SCAQMD Rule 1005.1. Rule 1005.1 is currently being challenged in court by Stauffer Chemical and the B.F. Goodrich Chemical Group on the basis that the Rule is unconstitutional. 7.3 RESOURCE CONSERVATION AND RECOVERY ACT (RCRA) RCRA was established in 1976 for the protection of public health and welfare by supplying guidelines to protect the quality of groundwater, surface water, and ambient air from contamination by solid waste. Draft regulations were issued in 1978 and final regulations in May 1980. These regulations control hazardous wastes from "cradle-to-grave" or from the point of generation through transportation, storage and ultimate disposal. This will be accomplished by a manifest system. If the generator produces hazardous wastes in sufficient quantities (greater than 1,000 kilograms per calendar month, except for some highly toxic wastes with lower limits of 1 kilogram per calendar month), he is responsible for their disposal. On-site disposal will require a permit with strict requirements for siting, operating, and monitoring the facility. Otherwise, the waste must be disposed of in a facility with a permit subject to the same strict requirements. The generator remains responsible for off-site disposal. RCRA will probably have more of an effect on EDC/VC plants than PVC plants. The following EDC/VC processes have been identified as sources of hazardous wastes (EPA, 1979b, p. 181), 1) heavy ends from distillation of VC in production of VC from EDC, 2) heavy ends from distillation of EDC in VC production, and 3) heavy ends from distillation of EDC in EDC production. These solid wastes have been specifically listed as hazardous but it will still be the responsibility of the generator to show that other 7-10 GENC 014641 solid wastes originating from his facility are not hazardous. VC has been identified as a pollutant that could cause a solid waste to be classified hazardous (EPA, 1979b, p. 175). Also, surface impoundments, utilized by EDC/VC and PVC plants to collect waste streams, represent disposal sites and may require upgrading to meet RCRA standards of performance (Hanrahan, 1979, p. 23). One of the more common methods being proposed for disposal of hazardous wastes is incineration. RCRA has extensive permit requirements for direct land disposal plus long term liabilities for generators using this method. However, incineration disposes of wastes without pretreatment and without incurring the long term liabilities of the manifest system. EDC/VC plants or PVC plants incinerating their hazardous wastes will be required to obtain a permit to operate the incinerator under RCRA. 7.4 TOXIC SUBSTANCES CONTROL ACT (TSCA) TSCA was enacted in 1976 and provides EPA with the authority to secure information on all new and existing chemical substances and to control the substances determined to be hazardous to public health and the environment. Basically, TSCA was instituted to control the commerce of toxic products and the regulations provide EPA with powers that do not exist under any other Federal toxics-related laws. The other environ mental laws (e.g., Clean Air Act, Clean Water Act, RCRA) are concerned with the control and disposition of gaseous, liquid and solid wastes and byproducts. OSHA focuses directly on worker exposure problems. However, TSCA deals with chemicals and products throughout their life cycle manufacturing, distribution, use, and disposal (EPA, 1979b). EPA will control the chemicals by requiring submission of a Premanufacture Notification (PMN) before marketing a new substance not listed in the TSCA 1979 Inventory of Toxic Substances. The PMN contains extensive background information including test data and literature predicting the effect that substances may have on workers, the environment, and consumer populations. Exceptions to TSCA are foods, drugs and pesticides that must be registered with specific agencies (e.g., FDA for foods and drugs, and EPA for pesticides). Any chemicals in the 1979 Inventory that are released to the environment by discharge 7-11 GENC 016642 must be listed in any state permits (to construct and operate), RCRA permits, and National Pollutant Discharge Elimination System (NPDES) permits under the Clean Water Act. EDC, VC and many polymers of VC are already listed in the Inventory, and public health and environmental effects have been well established for these substances. A PMN is not required unless the "old" or listed substance is used in a significantly new way (as determined by EPA). However, any new chemicals used in polymer development (e.g., new copolymers containing VC) will require a PMN prior to marketing. 7.5 CLEAN WATER ACT Section 307 of the Clean Water Act required the EPA to publish a list of toxic pollutants and authorized EPA to promulgate effluent standards for those pollutants. In addition to these listed toxic pollutants, amendments to the Clean Water Act in 1977 required that "Consent Decree Pollutants" be added to the list of toxic pollutants if effluent limits failed to achieve water quality criteria (EPA, 1979b). VC is one of these "Consent Decree Pollutants" for which effluent guide lines may need to be developed. If developed for VC, these effluent limitations would be listed on, and enforced through, the discharge permit required under the National Pollutant Discharge Elimination System (NPDES). Section 311 entitled "Oil and Hazardous Substance Liability," authorizes EPA to promulgate Hazardous Spill Regulations. Under these regulations EPA designated as hazardous those substances which, when discharged, present an imminent and substantial danger to the public health or welfare. An additional 28 chemicals have been added to the existing list of 271 hazardous substances published in 1978 --`ethylene dichloride (EDC) used to produce VC and vinylidene chloride (used as a comonomer with VC) are on the proposed list (EPA, 1979b). Any spill consisting of these chemicals would result in a penalty, which is determined on a case-by-case basis. Sources in compliance with effluent standards established for these chemicals in other sections of the Clean Water Act are exempt from these requirements. 7-12 GENC 016643 7.6 SAFE DRINKING WATER ACT The Safe Drinking Water Act of 1974 was established to ensure that the public is provided with safe drinking water. This protection of public health is accomplished by EPA through the adoption of National Interim Primary Drinking Water Regulations that specify maximum levels for certain toxic contaminants in public drinking water. Secondary drinking water regulations have been proposed as guidelines to the states to ensure non-health related qualities of drinking water. As a guideline, a list of chemicals indicative of industrial pollution has been published by EPA -- VC is one of these indicators (EPA, 1979b). In addition, the EPA Office of Drinking Water, Criteria and Standards Division, has drafted a criteria document that would be used to establish a drinking water standard for VC. 7.7 HAZARDOUS MATERIALS TRANSPORTATION ACT The Department of Transportation (DOT) under the Hazardous Materials Transportation Act of 1974 has promulgated final regulations on the transport of hazardous wastes and hazardous substances. VC, EDC, and several comonomers are subject to these regulations. The principal objective of this rule, as it pertains to the use of identification numbers for the regulated substances, is to improve the efficiency of civil emergency personnel (such as firemen and policemen) in the identification of hazardous materials, and to facilitate the accurate transmission of information to and from the scenes of accidents involving hazardous materials. 7.8 OCCUPATIONAL SAFETY AND HEALTH ACT OSHA recently published a request for information on VC and PVC in the Federal Register on December 18, 1979 (44 FR 74928). This request was for voluntary submission of data and information that could be used as part of a review of the current OSHA VC standard. OSHA is concerned primarily with PVC dust and that the dust might cause pneumoconiosis. OSHA is investigating to determine whether the disease is caused by residual levels of VC in the dust or by the actual dust itself. The progress of this investigation has not been determined to date. 7-13 GENC 016644 7.9 SUPERFUND LEGISLATION The Comprehensive Environmental Response, Compensation and Liability Act of 1980 (Superfund) was recently signed into law to further the control of hazardous substances in the environment. The purpose of superfund is to: establish a federal cause of action against those responsible for the release of a hazardous substance into the environment; create a $1.6 billion trust fund to be used for cleaning up hazardous substances released into the environment or for taking action to prevent a threatened release; and create a $200 million fund for surveillance, care and maintenance of RCRA-closed hazardous substance disposal sites, and for any damages or environmental cleanup costs associated with such a site. The $1.6 billion cleanup fund will be created from $1.38 billion in industry taxes and $0.22 billion in government revenues. The $200 million surveillance, care and maintenance fund will be created from industry taxes. These taxes will be on crude petroleum products, chemicals produced, and hazardous wastes generated. Regulations for implementing the superfund will be promulgated in the near future. 7.10 FOOD AND DRUG ADMINISTRATION REGULATIONS The Food and Drug Administration (FDA) establishes requirements for PVC resins fabricated into products that come into human contact (e.g., beverage containers, baby bottle nipples, blood bags, pharmaceutical products, and food wrap). A processor producing these resins must obtain a permit from the FDA. This permit specifies operational procedures, resin characteristics (e.g., clarity, purity), and residual levels of VC (RVC). The permit puts restrictions on all process steps from polymerization through fabrication. These resins are usually "small-batch specialty resins that are stripped under controlled conditions because of their heat sensitivity. Also, reactors must be maintained differently (e.g., cleaning requirements are specified) and frequency of reactor opening is dictated by the permit. 7-14 SENC 0V6645 7.11 OTHER STATE AND LOCAL REGULATIONS The state regulations having the greatest impact on reducing VC emission are those mentioned previously under the Clean Air Act. Prior to construction or modification of a source and subsequent operation, companies will be required to obtain permits to construct and operate. It is through these permits that states will set limitations and conditions for emission reduction in nonattainment areas. These limitations may require further reduction of VC emissions below NESHAP limits, either through a reduction in hydrocarbons or VC specifically. 7-15 GENC 016646 I 7.11 REFERENCES FOR CHAPTER 7 Brittain, Martin. 1980. NESHAP Coordinator, EPA Region VI. Meeting Report. July 23, 1980. Environmental Protection Agency. 1979(a). A Handbook of Key Federal Regulations and Criteria for Multimedia Environmental Control. EPA-600/7-79-175. August 1979(b). Environmental Protection Agency. 1979(a). "National Emission Standards for Identifying, Assessing and Regulating Airborne Substances Posing a Risk of Cancer," Proposed Rules, Federal Register, Vol. 44, No. 197. October 10, 1979(a). Environmental Protection Agency. 1975. Standard Support and Environmental Impact Statement:. Emission Standard for Vinyl Chloride, EPA-450/ 2-75-009. October 1975. Hanrahan, David. 1979. "Hazardous Wastes: Current Problems and Near-Term Solutions," Technology Review. November 1979. Holbrook, W.C. 1980. Director Toxicology and Environmental Affairs, B.F. Goodrich Chemical Division. Meeting ReportjTIB.F/'Goodrich;--f General^T i remand ^TRWj>represehtati ^s^SfftS6e?^!0^i?80 __ Winkler, Joe. 1980. Technical Support Section, EPA Region VI. Preliminary Determination for Formosa Plastics Company, PSD-TX-226. February 22, 1980. E L E [ L I E L GENC 016647 " [_ 7-16 L REVIEW OF THE VC NESHAP PART II 4/9/81 TRW/EPA RTP, N.C. The objectives of the "Part II: Vinyl Chloride Standard Revision" are to: Determine the Best Available Technology (BAT), and Obtain additional information required for revisions to those parts of the standard (identified during Part I) that need to be changed. To accomplish this, the following areas will be investigated: (A) Oxychlorination Vents. The objective will be to investigate existing- technology and compare emission levels from processors using: (1) air without incineration, (2) air and oxygen with and without incineration, and (3) oxygen with incineration. Cost analyses of these control systems will be determined. (B) Relief Valve Discharges (RVD's). An investigation is necessary to determine technically feasible equipment and procedures for RVD prevention, costs of these, and limitations of their application to older plants. (C) Resin Stripping. The objective in this area is to determine the advisability of requiring a more stringent standard. The following aspects of stripping technology will be evaluated: (1) Those factors involved in very efficient stripping processes and restrictions (if any) to using those processes for other resin grades/types, (2) Factors inherent in resin grades that are reputed to be resistant to low-level stripping, and (3) Correlation, if any, between newer plants and lower stripping levels. GENC 016648 Cost analyses will be included in the investigation of efficient stripping technologies. (D) Fugitive Emissions. New requirements may be needed for nqw plants (e.g., similar to benzene fugitive emission regulations for new sources). Fugitive emissions in existing plants will be addressed through contacts with the regions (1) Acceptable Urt'I^p|ogran4 and (2) Standardization of leak check frequency and leak definition. The elimination of fixed monitoring requirements is a possibility. "* (E) Malfunction Clause Provision. In order to determine the feasibility of plants operating without a back-up for their primary control device, the following information is needed: (1) How often do primary controls malfunction? (2) What options for control are available during such malfunctions and what are their costs? (3) For what periods of time do malfunctions occur? (How long is it before the primary control device can be repaired?) (4) What are the causes of malfunctions? (F) Unregulated Sources of VC. Several unregulated sources of VC emissions were identified in the original study for the existing standard, as well as in the Part I Review Study for this project. The objective for this area is to determine the significance (if any) of VC emissions from: (1) Fabricators -- Although OSHA workplace regulations have resulted in the incoming raw material residual vinyl chloride (RVC) levels' being lowered to 10 ppm or less, a survey of final RVC levels in resins leaving PVC plants is needed. The determi nation of these levels will allow a more reliable 6EMC 01664? estimation of the VC emissions from fabricating facilities. (2) Trichloroethane and certain pesticides -- These facilities were identified as using VC as a chemical intermediate. The extent of VC emissions / from these sources cannot be determined without process information. (3) Landfills -- During the Part I Review Study landfills were identified as a source of VC emissions. Off-specification batches of PVC are potential contributors to these emissions. Information required in this area includes: what are the roles of other regulatory agencies/branches in this matter? how are off-spec batches being stripped, and if they are not stripped, what are disposal procedures? (4) Marine loading/unloading facilities - Emissions of VC from these sources have been reported in two regions. Information is needed to determine: where the emissions are coming from, t If they are RVD's, can they be prevented, and what role other agencies (e.g.. Coast Guard) play in control of these emissions. (G) Miscellaneous Administrative Changes. The objective here will be to obtain the necessary information to: (I) Standardize definitions and requirements (e.g., Reactor Opening Loss (ROL) for those processors stripping in the reactor), (2) Consider certain exemptions from the standard (e.g., gasholder seal water from inprocess wastewater definition, solution and bulk processors from certain resin-stripping requirements), and GENC 016650 (3) Consider elimination of certain requirements (e.g., fixed monitoring for fugitive emission detection). The necessary information will be solicited from pertinent sources. GENC 016651 GENC 01665 DATES AVAILABLE TO DISCUSS PHASE II TOPICS FOR VC NESHAP REVIEW Air Products BF Goodrich Borden Certainteed Conoco Diamond Shamrock Dow Ethyl Firestone General Tire Georgia Pacific Goodyear Great American Hooker Chemical International Mat'ls. Pantasote Shlntech Stauffer Tenneco Union Carbide Relief valve discharge Resin stripping Primary control malfunction Other APPENDIX A VINYL CHLORIDE NATIONAL EMISSIONS STANDARD FOR HAZARDOUS AIR POLLUTANTS GENC 016653 TWt 40 P*MeiU*Mef CJw4ra**na<R CHAPTER 1--ENVIRONMENTAL PROTECTION AACMCY part si--natonai emission standM0$ FOR HAZARDOUS AIR POUUTANTS J4 Apptlaabfilty. 1 -04 Definition*. RR Ataintowi 44 Addrma. CUM prohibited HUTIUH, I1M DtWroalttaiion of eanetroction or medlficetlon. >47 application for Approve at whuud ub or modification. CR Approval by AdminMatter. no* Notification of rtamtp. U10 Rotirm reportln end iliv lequaa 01-11 Welter of OOtBpllAOOO. R-U Enaflop torti and botImbi01 J$ Wa4rar of amiaaien taata. Rij Rouioa toot end analytical mothoda. A4f ATtftftblllt; of Information. J State authority. 1J7 OlRUBfMttoa. 7 140 141 RE 14J 01 >4 1JM Applicability, Definition#. kaiaaion atandard. Alrel*Aninc< Eoportlnf. _ ww diopeoal Am7 o. 40 IJ1 0144 0149 1-44 AppUmbJltty. Daflnlttafia. feilmton atandard, Stack mmpllnf, Ar anapllnf. 0144 iHiiiBiun atnadord tar rtnyl abtarida plant*. 0144 lRriltadatlopniuatua.ndard tor polyvinyl ehto* 14* fenlaalon atandard for athylvn# diehiortd*. flnyl chlorlda and poly vinyl chloride plant*. 01.M Equivalent equipment and prooadurta. 147 femlaaton toata 41.M imiaaion momtorinc. 01.04 Initial report. <1 70 Semiannual repun. 01.71 EecordAeeplnf, pperulti A pnanpUanoa Rtatua Information. Appai)dli E--Tart Method#. Iflatnod 101--Refarenm method for dewrmi* nation of partieulate and faaeoua mercury amtaaiooe from ration# ry oourtM (air etrvami). Method >03 WafrroDoa method for detannt* nation of particulate and faaaou* mercury omiminna from ateiionary aourom (hydrofan rrraami) Method 100--nary!)turn anaoiac method. Method lo--nafaronoa mathod for detarmj. nation of beryllium --niaaiini from m* ttonary mow. Method loo... Method for determination of mercury in vwtewatar trontmant plant eewacetiudfm ' Mathod lO^Datrnnlnatlon of rtnyl eWorld# from fetadonary aoureaa. " Mathod 107--Datarmlnation of rtnyl chloride of lapraeaaa waetevatar aamplea. and nnyt ohlorlda oontant of polyvinyl chlorideH. main, alurry. vat mfce. nnd lata aamplm.** Avmavrr Ran 111, MKa) f th# Clana Air Art da ammdad (43 D4.C, Till, tOOXaU. nnlaaa otherv** nrTnl i 140 1.41 149 40 >44 4*fRd" feafeat Mala* FWtnf Applicability. Dafiniuoea. tolnlw atandard. Bniaaion teatlay wtrt ftrtaf or ptv* poUant dlapooai. 4he eampuof. MJO 141 149 149 144 144 AppUqabCUty. Definition*. ~ Rick mmpitnc. RvdnnatUBtl talMtomanviaal taw*! f WatiaaatjmnaaMmii^id tar 7Wy< 140 > 41 149 Appiinability. PiflnUlom mlaatoo atandard Mlarldi RuU. far athytane dl- l A-l QENC 016654 tertpert A ti-wml Rrmtaw* | *1.01 erating deelgn neparity of the stirinnei j (Hi An turret tn boom of operetta. (k) "New source- means any stationary <> PeMtarytiiw^ Ms* Tt* provision* of this part apply to the owner or operator of any stationary pure* for which a standard is prescribed u&der this part. source, the construction or modification of which Is commenced after the publi cation | the Fxskiul Racism of pro* posed national emission standards for hazardous air pollutants which will be <d> lfl.niadi emmifT | (102 Dcfinillenii applicable to such source. OP.I0HIMW 0) "Owner or operator- means any Ai used tn this part, all terms not de fined herein shall have the meaning fires them in the act: fa' "Act" means the Clean Air Act <42 U-S.C. 1*57 et *eq.). (b> "Administrator'1 means the Ad ministrator of the Environmental Pro* taction Agency or his authorised reptatentative, . (c) "Alternative method" means any method of sampling and analysis* for an air pollutant which is not a reference method or an equivalent method but which has been demonstrated to the Administrator's satisfaction to produce. In specific eases, results adequate tor his determination of compliance.2 (d) "Commenced'' means that an own er or operator has undertaken a con* person who owns, leases, operates, con ndsttaadard trols, or supervises a stationary source. <bcw 113 sad SOI (a) of U Clean Air <ra> "Reference method" means any aev a* aivsis jas Oac. ien-r method of sampling and analyzing for an ies?g<)).) air pollutant, as described in Appendix B to this part <n) "Startup" means the setting to 141.04 operation of a staUosary source for any purpose. <a) All requests, reports. applications, <o) "Standard- means a national submittal*, and other communications te emission standard for a hazardous air the Administrator pursuant to this part pollutant proposed or promulgated under Shall be submitted in duplicate and ad this part. dressed to the appropriate Regional Of (p) "Stationary source" means any fice of the Environmental Protection building, structure, facility, or Installa Agency, to the attention of the Director. tion which emits or may emit any air Enforcement Division, The rrglnns) of pollutant which has been designated as fices are as follows; hazardous by the Administrator. lUfioa I (Cenaeeucut. Main*. Raw Ramp* sbPe. Huarbuetti, Rhode IsUoA. T- tinuout profram of construction or modification or that an owner or operator has entered into a contractual obligation tn undertake and complete, within a Tea* enable ume, a continuous protram of construction or modification. (a) "Compliance schedule" means the date or dates by which a source or eatelory of sources Is required to comply with the standards of this part and with any steps toward such compliance which are Ise*t1f.1o2r.th in a waiver of compliance under (f) "Construction" means fabrication, erection, or installation of a stationary source. (*> "Effective date" Is the date of 141.03 Urtfiasdabhmtaian*.4* Used tn this pert are abbreviations and MiUili of units of maamre. Thmo ore defined es follows: <a> System International <EXi unite cf measure: istapm JSvwfeosistlw* i kem aur * !W* ms taaiiunifuri IS-* ifiw tegs pegacnm* ip (tsb meat), Jobs F. Kenaaey pidaral luildlog. boston, lUnifftliwta 093QS. Region O (Now Tor*. R*w Jcrvey, PuvrtQ Rice, Virtu lalaadi), Federal Ofle* SuU*lag, a* Federal Flag* (Foley Square), Rev York. |f,T, 10007, Region tB (Delaware. DUtnct of Columbia. Fsnnsylvanla, Maryland, Virginia. Wt Vtf* glniai. Curtis SuUdtsg, fiath an* Walnut gtreets, Fbllasalphia, Fennsyivanla Ifioe, Region IV (AUMbu, Fionas. Qaorgu, Wit* almippl, XentucST. Rorth Carolina. booth Carolina. Tbno--m*), Sulla 300, 14*1 reacD- Wae gtfeet. Atiaot*. Oeurgla S030S. Region V (DUnote. Indian*. Minnesota. Michigan, Ohio. Wisconsin \. 130 South-Dear born Street. Chicago., Illinois 606004 3*37 Region Vi (AikauM, Louisiana, Kew Oklahoma, Tegas), ieoq promulaatlon tn the Pntut Ri&sstu of an applicable standard or other recu* latlon under this part. <h> "Equivalent method" means any method of sampllnf and analyzing for an air pollutant which has been demon* strated to the Administrator's sails/ac Sms uaoneiwi )0-* meter fespoecal amostf WVsevoel*ill fuiss), 171# alrttttort street, gsnmi CUT. Ramon 63106. Region VJH (Colorado, H""*-tfta Rerih pa* beta. South SeadU, Utah. Wyoming). iH Ltnooln Tower*, 16*4 Unooln Street, Danvof. tion to have a consistent and quantita tively known relationship to the reference method, under specified conditions. (li "Exittine source" means any sta tionary source which is not a new source. <ji "Modification" means any physical change In. or change in the method of operation of. a stationary source which increases the amount of any hazardous air pollutant emitted by such source or which result* in the emission of any hazardous air pollutant not previously emitted, exeept that: (!) Routine maintenance' repair, and replacement shall not be considered physical chances, and (2> The following shall not he eon- qamwtecMtaM^pss Other mttte of mosiure: Csdegr-- OeMue QaoSgfJi) dmatiibc foot per miouw tanek PenWiener e*F*eW*r|tw Mmsm fVatquan Seat (Vansk feet ^ISl*singcahllon tiaszRtfoslnicsbdemi eKg ewoeMtetwy ttw pm m***t mmltsnasmiuiistnuemes IW* mm __________i.. y-mmiii-, a**tl>< Revada, Onam. American Samoa), 100 Cali fornia Street. San FmoetsoO, California #4111. Region X (Washington. Oregefi, Idaho. Alsaaa). IRC teeih Avenue. Paavtio, Wash ington MIOi. (b) Section 112(d) directs the Admin* tetrntor to deletete to each State, when appropriate, the authority to implement sad enforce the national emission stand* art* for hazardous air pollutants for sta tionary source* located in `such Bute. AH information required to be submitted to EPA tinder paragraph (a) of this e*etta must also be submitted to the ap propriate State Agency of any State to which this authority has been delegated (provided, that each specific delegation adored a change In the method rt operation: ri> An tnifeass in the production rate. If such tecreaee doe* not expert the op may exempt source* from a certain fed eral or State reporting requimnnt). The appropriate mailing address for thoee States who** delegation request has been approved le as followa: GENC 016455 A-2 (A) lRa*ii i il| (1) lutt erf Alibtmi, Air Pollution On- i Dtowan, Air Pollution Control Comnue- Wea, *43 8 McDonough Street. Montgomery. A-(fCuml e[sw--e1r0ve4d."] (D> Anxona Ptma County Air Pollution Control DU- noi.*nrtet, iSl Wpt Coogrw street, Turnon AZ (1 |Rwved| in Mum. Boy Aroo Air Pollution Control District. ** Bile Street. Son Praacisco. CA 84100 Dal Norte County Air Pollution Control ptsviet. Oourthouee. Creecent City, CA 99331. prmna county Air Pollution Control DuWWt. 113 1 Cedar Amu*. Prwno, CA 03703. umboldi County Air Pollution Control DaatMct, MOO S Broadway. Eureka. CA *9901. Kora County Air Pollution Control DtOrtct, 1700 Plower atroot (P.O. Box 007) BdAanSeld. CA 03309 Mmn County Air Pollution Control DteUriel. 133 W. Ycsamlte Artoui. Madera. CA Mandocino County Air Pollution Control District, Oounty courthouse, Ukiab, CA #*4*3. Monterey Bay Unified Air Pollution Con- mol District, 430 Church Street (F.O- Box 4m, Salinas. CA *3901. northern Sonoma County Air Pollution Odom District, 33 U Chanste Road, Santa BOSS. CA *3404 Sacramento County Air Pollution control District. J70J Branch Center Bond, Sacra* same. CA *3497 San Dtaco County Air Pollution Control DWtrltt. 3130 Choaopeaks Drive, San Diego, CA*91 San Joaouin County Air Pollution Control District, 1*01 E Haiti ton Street (P.O. Box 300*1. Stockton. CA *8301 Santa Barbara Air pollution Control Dis trict. 4440 Calls Rwd. Santa Barbara. CA mho. Btanistaus County Air Pollution Control District. M0 Scenic Drive, Modesto. CA *3390. Trtnitv Countv Air Pollution Control Dis trict. Box AJ. Weaverville, CA 0*0*3 Datura Countv Air Pollution Control Dis trict. *33 x, Santa Clara Street, Ventura. CA M001. (Ol State of Colorado. Colorado Air Pol* lotion Oenmot Division. 4910 test llib Ave na*. Denver, Colorado *0330-* (Hi Suto erf Connecticut. Deportment Of tevtronmental Protection. State Offlos ButIdles, Hartford, Connecticut Milt.* (J> Buis af Delaware (for foosO feel*fired mana *snsiatoti; Mneratort: nitric add ptanu asphalt oonerete plants: dsnif veodota for petrel*um liquids. and ooera treaw Mont plants only): Delaware Department erf Matural Resources and tRvvenmenul Con sul- Stfaard Tatnall 13*01.** Dover, DaL (*|*(H) {Rooorved) (1*1 Stau of Oooriia. Bnvireamentpi proteetfon Division. Department of Natural Reooereee. tto Washington Street S.W, At lanta, Oootfia SM34 W <M|-<0) 1 Rsoerved | (P) State of Indiana. Indiana Air PoUtitton control Board. 1330 west Michigan Street Indianapolis. Indiana 4*90*.*" (Q)-(B) [Swerved] rKoln^4mMem01a.lf>ioutUsii. CA 1*7. Pteotiert (V1 r erf Maine. rmi>i leisii orf te- vtronmeotai Protection, sum Bom. Au- (W(VU). M[Raeisne*>rvo0e4ad3a|3a0o."" (W1 Mamachueerts Department of Bnvl* mnmtntil Quality Engineering. Dlvtsioti of Air Quality Control. *0* Washington Street, Boston, Massachueette 09111.*' (Z1 State of Michigan. Air Pollution Oan* Uol Division. Mlchlcan Department of Natu ral Reoourum. Stevens T Mason Building, th floor, Lansing. Michigan 43938.'* (T) Mlnnwoaa Pollution Oeamol Agency, Divuson grf Air Quality*, l*B8 West county Bead B-3, Iiamui, Minn, 34U3.44 iZl | OB) BMW M liuiasns. Deperteewt M H--ith and teiimiiiimiwl B'Iichwl Com* wan --------- Manx. MSOl.41 [tmsrvsd). (ZX1 Now Hampthire Air PoUutlon Con trol Agency. Department or Health and W*i far*. State Laboratory Bulldinf^Baaen Drive. Connor*. New Hampshire 03301.77 (ir) Spokane County Air teiluttoa DqW Authority. Ncetb 3ll Jstferoon. flpokan* WaahiAfun **901. (v) Yakima County Cioen Air Authority. Coust^^ourtbouse. Taklma. Washington (vi) Ctymple Air Pollution Oontml Au thority. 190 Bast State Avenue. Olympia Washington **30i. (vii) Southwest Air Pollution cnso| Au thority. Suite 7*01 H. NT Rami Doll a**uu< Vancouver. Waahlnfton MM9.1) (XX) IReaerved] IVY) Wisconsin--Wisconsin Department of Natural Resources PO. Box 7*31 Mae*, ton Wisconsin 33707 If <ZZ> (Reserved) (AAA) |Reserved| () COmmnna'itlth ( Puerto Rieo commonwealth of Puerto Rico En\tfo-.. meat*! Quaht) Board. PQ Bov 117*4 turee. PR *0*10 ICCC) DR Virfin Islands- US virgin Islands Department of Conservation and Cultural Aflain. PO- Box 87*. charlotte 0A0m*a0l1ie. St. Thomas. UR. Virfin island* (Sees. 101. no, in. ns and 301 of the Clean Air Act. a amended. 49 UR.C. 1S37 !*3?cS. 4. T and l37g ) partment t m3*Jnhn Pitch Plaaa, PA. (OO) |Re**rvtd| (KH) New York: New York State Departmeat of Znelronmental Coaeervstion, M Wolf Road, Albany, New York 13933. attention*. Division of Air Resource* (111 North Carolina Environmental Men- a*ement Commiwion. Department of Natural and Economic Resource*. Division of Envi ronmental Manafement, P.O. Box 37637, Ralelfh. North Carolina 37611. Attention: Air Quality Section ** (JJl State of North Dakota. Stau De partment of Health, stau Capitol. Bismarck, North Dakota uui 2' (XX)-(LLi (Reserved] tMM) State of Orefon. Department of Environmental Quality. 1934 SW Morrison Street, Portland, Orefon *7704(NN> (a) Commonwealth of Pennsylvania (except for City of Philadelphia and Alle* fheny County) Pennsylvania Department of Environment*) Resource*. Bureau of Air Quality and Noise Control, Post OAce Box 90*3. Harmburf. Pennsylvania 17)90 <bt City of Philadelphia. Philadelphia De partment of Pubiie ffeaun Air Manafement Service*. *ol Arch street. Philadelphia, Penn sylvania 1*107. 33 (OO) [Reserved] <PP| Slate of South Carolina, Ofle* of En vironmental Quality Control, Department of Health and Environmental Control. 9*00 Bull stmt. Columbia, south Carolina 9*3oi?v (QQMYTt (Reserved| (UU) State of Vermont A*ency of tevivonmentaJ Protection. Bax 4*9, Montpelier, Vermont 0**09.33 (W) Commonwealth of Virtmia, Virginia State Air Pollution Control Board. Ream 110*. Ninth Street OOeeBUUdln*, Richmond, Vlfflhla 33919*3 (WW) (l) WaaUhfwn; Stau of Waahinf- ton. Department orf Reolery, Olympia. WaahMftOtt *M04. (U) Northwas* Air Pollution Authority. 07 Pioneer Butldlnf, Bscond and Pine Btreeta. Mount Vernon, Waahinfteo 9*973 (Itt) Puet Sound Air tellution Control Afwcy, 410 Wool HirfTls-- Sweet, Smitia Weablafto* Hill. A-3 GENC 0166-56 161.65 (ft) After the effective dale of any itimlinl prescribed under this part, no owner or operator shall construct or mod ify any stationary fouret subject to uch eundard without flnt obtaining written approval of the Administrator to accord ance with this subpart except under on exemption granted by the president under section 112(e)(2) of the act ffoureci, the construction or modification of which commenced ifter the publlea* tloo dftte of the standards proposed to be ftpplicftblc to such source, ftre subject to this prohibition. <b) After the effective dftte of fthy eundard prescribed under this pert, no owner or opentor shell opermte any new source in violation of such standard itoept under in exemption mated by the President under section 113(0 (2> of the act, (e) lflncty dftjt ftfter the effective dete of fttu standard prescribed under this port, no owner or opentor shell operate sny existing stationary source in viola tion of such stead*rd, except under ft waiver mhted by the Administrator in accordance with this subpert or under an exemption mated by the President wider section H2<e> (2) of the act. (d) No owner or operator subject to the provisions of this pert sheJl f*U to report, revise reports, or report source tost results es required under this pert. 161*06 DtKmlnatiM mi w aMf!A(tiion. Upon written application by fth owner or operator, the Administrator will moke o detertninetion of whether setlons token or Intended to be token by such owner or operator constitute construction or modification or the commencement thereof within the meaning of this port. The Administrator will within 30 dors of receipt of sufficient information to OToluote on ftppllcoUoa. notify the owner or operator of his deterrain*tlon. I61.U7 AppUestiftw for appraval of WMintdiw or modiffcftUeo. (0) The owner or operator of cap new source to which o standard proscribed under this port Is ftppUeoble shell, prior to the dote on which construction or modification is plenned to commence, or within 30 dftjs ftfter the effective d*U hi the cte of o new source that already has commenced construction or modlfl* cotton and has not begun operation, sub mit to the Administrator an application for epprova] of sueh construction or modlffcatlon. A separate application shall bo submitted for each stationary source. Each appUcatlon shall iaclude: (1> The name and address of the ap plicant. (3) The location or propwd location of the source. (1) Toebaked information deacrfbftg the proposed nature, sin. ri--lrn operattof datign capacity, and method of oper- atton of the source, tnrhuttnq ft dmrrtphoc of ny equipment to be used f<* control of emissions. Such )****<-*) in formation ih*U Include calculations of omission estimetea la sufficient detail to permit imTtmfD* of the validity of such calculations. of initial startup of Urn days after such date. <43 DAC H14R I 41DI Afpnnl by Afolnlilniwr, (a) The Administrator will, wlthia 66 days of receipt of sufficient information to evaluate an application under I 01.07. notify the owner or operator of approval or Intention to deny approval of ccmstruct!on or modification. If the Administrator determines that a stationary source for which an application pursuant to I 61.07 was sub mitted will, If properly operated, not cause emissions in violation of a stand ard. he will approve the construction or modification of sueh source. (e) Prior to denying any application for approval of construction or modifica tion pursuant to this section, the Admin istrator will notify the owner or operator making sueh application of the Admin istrator's Intention to Issue sueh denial, tocether with: (1) Notice of the information and findings on which such Intended denial is baaed, and (2) Notice of opportunity for such owner or operator to present, within such time limit as the Administrator shall specify, additional Information or arma ments to the Administrator prior to final action on such application. (d> A final determination to deny any application for approval win be in writ* Inc and MU set forth the specific frnunds 60on which sueh denial Is based, ftuch final determination will be made within days of presentation of additional Infor mation or arfuments, or 60 days after the final dau specified for presentation. If no presentation Is made. <e) Neither the submission of tn ap plication for approval nor the Admin istrator's (rantine of approval to oonsfenict or modify shall: (l) fuller* an owner or operator of local responsibility for compliance with any applicable provision of this part or of any other applicable federal. State, or local requirement, or <2> Prevent the Administrator from tmplementinc or enforeinc this part or taking any other action under the ecu | 61.06 lUdlWadMi sf i*i rap. (a) Any owner of operator af a source which has an initial startup after the effective date of a standard prescribed mder this part shall furnish the Admin istrator written notification as follows: <1) A notification of the anticipated daU of initial startup of the source not more than 60 days nor less than 36 days prior to such date <3) A notification of the actual dale 6 61-10 8wm rfforWif mmd *iKr fe rn*--** () The owner or operator of any existing source, or any new source to which a standard prescribed nnder this part Is applicable which had an initial startup which preceded the effective d*U of a standard prescribed under this part than, within 60 days after the effective date, provide the following Information in writing to the Administrator: (1) Name and address of the owner or operator. (3) The location of the source. (3) The type of hazardous pollutants emitted by the stationary source. (4) A brief description of the nature, else, design, and method of operation of the stationary source including the op erating design capacity of sueh source. Identify each point of emission for each hazardous pollutant. (I) The average weight per month of the hazardous materials being preroeee d by the "urae, over the last 12 *>** preceding the date of the report. () A description of the Mn. tool equipment for each emission point. (I) Primary control device(e) for sac* hazardous pollutant. (II) Secondary control devko(s) for each hazardous pollutant. (ill) Estimated control efficient (per cent) for each control device. (7) A statement by the owner or oper ator of the source as to whether he ean comply with the standards prescribed in this part within 36 days of the effective dato, <b) The owner or operator of an exist ing source unable to operate tn compli ance with any standard prescribed under this part may request a waiver of com pliance with sueh standard for a period not oxeeedlnff 3 yean from the effective data. Any request shall be in writing and hall include the following information: (1) A description of the omitrole to ba Installed to oompty with the standard. (3) A compliance schedule, Including the date each step toward compliance will be reached, Sueh Uet shall Include as a minimum the following dates: (1) Date by which eontracts for emie- Non control syitomi or pracs-- modifica tions will be awarded, or date by which orders will be issued for the purchase of component parts to accomplish nis- Non ooutrot or progees modification: (ID Oats of initiation of onsite eon* etruction or Installation of unieelon oon- trui equipment or preceee change: (Ui> Date by which multe construc tion or installation of emlatioo control equipment or process piwillkatlnn is to he completed: and <lv> Dal* by which final compliance is GENC 016657 A-4 r r r L L L I 1 l tr L iL L_ i L, L L to to eehteved. (t) A description ef mterlm rnitolrm centre! step* which wffl to totoo MW tto waiver period. (c) Chaniw to the Information pro vided under paragraph (a) of toil taction tool! to provided 1o the Administrator within JO din after such change. txeept But If changes will reeult from modifiesMoo of the source. u defined to I (1.03 (J), the provisions of I (1.07 tod I <1-00 on applicable. <d> The format for reporting under Bila tectlon la Included at Appendix A of this part. Advice on reporting the atatui of compliance may to obtained from the Administrator. Me. 114 t the Ohs AJr Act m mWM tea OAfi 1414H. 4M7 |1.11 Waiver *f eouplleoe*. (a) Bated cb the mfonnatlcB provided in aar requeet tinder I (1.10. or other In formation. the Administrator mar grant a wairir of compliance with a etandard far a period not excecdlni 1 Jean from the effective date of euch standard. tb) Such valvar off] to m writing and will: (1) Idtoitlfj the (tatlarurT eowoe aovered. 43i Specify the termination data of the waiver. The waiver may to termi nated et an earlier dale If the conditions epeclfivd under paragraph (b) <J> of this Mctloo iw not met. (1) Specify datee by which ttepe to ward compliance are to to taken: and tmpoee such additional condition! aj the Adminlttrator determinat to to nocaamry to atjure Irutallatlon of the neeeeeary eontroU within the waiver period, and to atiure protection of the health if petaetu durlnv the waiver period. <e> prior to denytns any request far a waiver pursuant to thli feelion, the administrator will notify the owner or operator maklni euch request of the Ad ministrator's Intention to Issue euch Stoilal. together with: ill Notice of the Information and findings on which such Intended denial Is based, end (3) Notice of opportunity for eueh owner or operator to present, within each time limit as the Administrator specifies, additional information or am monia to the Administrator prior to final action on such request. <d> A fine] detetmtnetlon to deny any lequaet for a waiver will be In writing end will set forth Uu specific (rounds on which such denial I* baaed. Such final determination wtll to mad* within (0 days after presentation of additional in formation or arguments, or (0 days after tha final date (pacified for euch presen tation. If no presentation ts made. tel The granUnf of a waiver undtr Idle eeetSon shall pot abroett* the Ad ministrators authority under section 114 ei the act. |(1.11 Feetoleo Htot ad Imlag (a) Kmleelon torts and monitoring hall to conducted and reported as est forth m this pert end Appendix B to this part. (b> Tlte owner or operator of a new source subject to this part, and at tbs or operator of an etlsttof source sub ject to this part, shall provide or aaus* to to provided. testing facUl- ttee as follows: (1) BampUnc port* tdequate for test methods applicable to euch source. (31 Safe sampling platform ts). (1) Baft access to plat form ts). (4) UttUttee tor --and testing equipment. (to 114 *f tee Qas Ah Am m mnM 141 DJtC. 141411.4*41 |1.1> fftewslmlatotoll. (a) tests may to waived upon written application to the Admin istrator If, in hit judgment, the source Is meeting the standard. or If the source is operating under a waiver of compliance or has requested a waiver of ---pii.n.. (bi If application tor waiver of the emission test Is nude, such application shall accompany the Information re quired by I (1.10. The appropriate form la contained in Appendix A to this part (c) Approval of any waiver (ranted pursuant to this section shall not abro gate the Administrator's authority under the act or In any way prohibit tha Ad ministrator from later eanotUni such waiver. Such cancellation will to mad* only after nolle* Is given to the owner or operator of the source. tSea. 114 ef the Q--p Ah Act is 44J CA-C. 141411. (M7 | *1.1* fiearee sees eod isiljtllef meth od* (a) Methods 101. 103. and 104 In Ap pendix B to this part shall to used for all source tests required under this part, uniats an equivalent method or an al ternative method has been appioied by the Administrator. (b) Method 103 In Appendix B to this part Is hereby approved by the Admin istrator as an alternative method for sources subject to I (1J3ta) and I (1.43 (b), <*> The Administrator may. after no tice to the owner or operator, withdraw approval of an alternative method (rented undtr paragraphs (a). (b> or Id) of this section. Where the teat rteulte uelng an alternative method do not ade quately Indicate whether a source is to compliance with a standard, the Ad minlitretor mar require the use of the reference method <h Its equivalent. ' (d> Method 1M In Appendix B to this pan It hereby approved by the Admlnls- Bator a* an alternative method for sum us* subject to I (lJJtb)! tarn 114 to the Qw AST AM to ttotedto <43 DAC. 1414)1.4*47 | *1.15 Availability af Minullen.11 The availability to the public of In formation provided to. or otherwise ob tained by. the Administrator undtr this part shall to governed by Pan 3 ef this chapter. (See. 114 to the Omo ibteu------ (43 BAX. 1414)1. *M7 | *1.1* State aaibae lly(a) The provisions of this part shall not to construed In any manner to pre clude any State or political subdlvlslai thereof rrom: (1) Adopting and enforcing any omledon limiting regulation applicable to a stationary source, provided that such mission Uniting regulation It not lass stringent than tha standards prescribed under this part. (3) Itaqulrtng the owner or operator of e stationary source, other than a sta tionary aource owned or operated by the United States, to obtain permits, Ucauet, or approvals prior to Initiating construc tion, modification. or conation of such source. t0fie^e1. 4U1S(. ^C4ls*a4p7Ah Act to toaM (43 I *1.17 OrcwtovtoUla*.7 No owner or operator sublect to tha provisions of this part shall build. Meet, Install, or us* any article machine, equipment, process, or method, the use of which conceal! an emission which would otherwise constitute a violation of an applicable etandard. Such concealment Includes, but It not limited to, the use of gaseous dilutants to achieve compliance with a visible amissions etandard. and the Piecemeal earning out of an opera tion to avoid coverage by a standard that appllts only to operations larger than specified els*. A-5 GENC 016458 I(14i (> nu nbport ppll** to ptonto which product: < 1) Btiiytotw dlehlcrtd* by yctlcp *i mm and hydrocon cbtoridt wh Mhykn*. (1) Vinyl dtlorid* to W toWto, (1) Oh or non potynwn wwiUtntn W ttoctua al palynwlowl vlnrl eblottdt, <b) Thi* wbpait tom not *Wl to qulPtMit UMd In noterch ud dovaloptont U Uh ntetw iwd to polymulo* b* rtnyl chlotld* pmr--id In th* *Qulb- toont ho* a c*parity of no not* than 0.11 m* iMnl). ic> action* <d thh nabpart othr than IIM1: lint (a)(1), (b), (ci.and <d>; Ml; al.: till; 01.10: and 11.11 da mot applr to ooulpmat uaad to rmaftb and d*i*lopant u the naetor uaad to poirmanat tba Tiny] eblodda pm e mid to U aoulpaunt baa a capacity of motor than 0.11 m1 (M cal) and no toon than 4J1 m1 (1100 oal) | till DHUKtoa Tarmi uicd in thli nbpart an dtflnnl to tha Act, to Oubpart A at tola part, or to thli Motion u follow.: (a) "Ethylene dJchlorlda plant" toahidat any plant which producaa athylao dlchlanda to medon of axyfan and hydro*an ehiarld* with cthylon*. (b> "Vinyl ehlondt plant" (nrlmtoa any plant which producaa itoyl chlorld* toanypracaai. (e) "Polyvinyl chlorld* plant" toolbdaa any plant whtr* ytoyl chlorld* toon* or to combination with other material* n polymerised. (d) "Slip caur*" caaani a you** which ha* a proha that nora* throach the foe/ liquid Interface to a atomot or trmnefar rwaial and Indicate! tha lent of vinyl chloride to tha cacaal to tha phyototo ototo of tha malarial tha oaao* diaeharta*. (t> "Trpa of main" moan* tba broad (toolIdeation of recto tofenlno to the took monuficturtnf proccoa for productoe that rmln. tocludlnc. but not lhnltod to. tba auapentoon. dlaperaian. latex bulk, and oolution ptocaoam. If) "Qrada of note" maani tha ufeatrlaloo of main -1---W-W-r which daoenbaa it oo a unique raaln. La., the tooet enact daacnptlon of a raato with no fur- thar cubdlTlalon. (*) "Dtapenloo raato" ntaana a mtoi toannfactured to ouch away aa to fmoa tdd dlapanlan* whan dleparead to a r plaatlclaar/dltueit mtx- (h) "Loteq leeto' maon* a recto which k preduced to a pelymertmflrwi pcmmto which toftlatae from free ledlenl totolyut duimtkiMOi 01 "Sulk min' `maani a min which ll produced by a polymcniatloo proem to which no water I* uaad. til "Inprocaa* wietewater" maani any water which, durln* manufacturto* or paocmatnr. comm Into direct contact with Tlnjl chloride cr polyrtoyl chloride or reeulte from the production or uec of any raw material, intermediate product, finlahad product, by-product, or weete product con(aInin* ytoyl chloride cr polyrtoyl chloride but which hoe not been dlecherpcd to a weeteweter tmetmant proeore or dlechiiicd untreated aa (k) "Wastewater treatment pi mow" toclude* any process which modifies characteristics such * BOD. COD. T88. nnd pH, usually for the purpooe of meettoy eAuest yuldeltoes And standards: it does not tochidt Any proems the pwpoa# of which U to remora vinyl chloride from water to moot requirement* of thk (l) "In vinyl chloride aarrlca' that a place of equlpmeot contain* or oantoete either a liquid that h at iaaat 10 percent by waleht vinyl chloride or e oae that U at Imat 10 pmcent to volume vinyl chloride. (m) "Vtandard operattof procedure" meene a fennel wrltteh pmecdun offi cially adopted to the plant owner or operator and available on a routine baala to thaea pareaa lerponalbla for carryto* out the pmoadure. In) Him" maani the net period of too* durln* which an ctnlmlon eampl* I* collected. (o) "Xthylm* dkhlarld* purification' toeludea any part of the procaea of cthylHa dlchlorlde production which follow* othylen* dlchlorlde formation ud to which ftntatted *thyl*n* dlchlondc I* (p) "Vinyl ehlorlde purification" to- cludet any pert of the proton of vtoyl chloride production which follow* Ttoyl chlorld* formation end to which finished Ytoyl chloride la produced. <q) "Reactor* todtidea Any twhI to which vinyl chlorld* l* perttolly or toto&y polytoerisad Into polyrtoyl chloride. <r) "Reactor opealny loaa'* meani the --kitokiMik Ytoyl ehloride oecuntoc when a reactor la ranted to th* atmoa- phere for any pmpoe* other than an >|*ln.Ci |>t(nai):y. rellad dlechciye aa defined to () "BUIppei" toeludea any vaaael to which rwdduai Ytoyl chMda k removed from polyrtoyl ehloride raato, except hulk raato, to th* alurry form by th* ua* cf beat and/or vacuum. In th* eaa* of bulk imto. atitpper toehidm any veaael which h uaad to ramor* raaldual vinyl chloride from polyrtoyl chloride recto fr**11*^ foUowtoc th* polymeria*- ttai atop to th* plant proem* How. ft) "Standard temperaiur**4 toaane a temperature of JO' C <#9* F).5* "Standard peaaaura" mrana a praantra of TtO aam of H| 3 to. of >** | *1Xd1ldfcFhr-tloefaptela.Mtehw31fcrd fer aibylm (a) Ethylene dlchlorlde purification: concentration of vinyl chlorld* m Oil exhaust fua dkchsifrd to the it* Biosphere from any equipment used to ethylene dlchlorlde purification k not 1e saesud 10 ppm, except u provided to I 41,CA(a>. Thk requirement dom not Apply to equipment that ha* been opened. I* out of operation, end met th* require- it to 141.45(b) <>(i) before bttoc (b) Oxychlorinatlon reactor: Ezctpt a* provided to fdl.f&fa), eml?*iom of rtoyl ehloride to the atmoaphere from aach oxychlorination raactor are pot to oxoaed OJ f/t| (0.0003 lb/lb) of th* 100 Percent ethylene dlchlorlde product from the oxyehiortoatiou proeaaa. 14I.4J Ea^eetoa kenderd fer vtart ftliiriili pleat*. An owner or operator of a vinyl chlo ride plant ahall comply with th* require* menta of thk aeetlon and f 11.00. (a) Vinyl ehloride formation and puri fication: Th* concentration of vinyl Chloride to all exhaust yaaea dlacharyed to the atmoopbera from any equipment deed to rtoyl chloride formation and/or purification la not to exceed 10 ppra. ex cept at provided to 101.45(a), Thk r*- qulirmetJt does not apply to equipment that has been opened, k out of operation, and met th* requirement In | 41.45(b) (4) (1) before both* opened. | 41.44 tnWos aiwdsd fer palj itoyl chloride pieaU. An owner or operator of a polyrtoyl chloride plant shall oomply with th* re quirement* of thk section and 141.45. (a) Reactor. The foUowtoc require ments acmly to reactor*: (1) Th* concentration of vinyl chlo ride to all exhaust cases dkcharyed to the atmosphere from each raactor k not to exceed 10 ppm, except as provided to parayraph (a)(3) of thk aecuon and 141.45(a). (3) Th* raactor opening loss from each reactor k not to exceed 0.43 y rtoyl Chtoride/ky <0.00003 lb rtoyl chloride/ lb) of polyrtoyl chlorld* product, with th* product determined on a dry solids beak. Thk requirement applies to any vessel which k used a* a reactor or as both a reactor and a stripper. In the bulk process, the product means the frees produet of prapolymerkaUon and poetpolymerttation. t (3> Manual rent ralra dkchary*: Ex- eept for an m*ryocy manual rant rale* dkchary*. there is to be no dkchary* to the atmosphere from any m*11"1 rant valve on a polyrtoyl chlorld* reactor to rtoyl chloride service. An m*ryency manual rant ralra dkchary* means a dkdiarf* to th* atmosphere which could Ml have been avoided by takiny m*as- wm to pf< vent the dkchary*- Within 10 6ENC 01665? A-6 r r F F t L L r< Li i c L L 4ay of any dtacharg* to tb* from enj manual tt valve, the wu er operator of the sowoe fnxn which the discharge occurs shall submit to th* Ad* unutf mr * report to writing contain- tog toformetioo on the eouroe. nature and ctuM of toe discharge, the date and Mm# of the discharge, the approximate total vtoyl chloride km during toe dis charge, the method \wsd for determining the vinyl chloridt tom, the action that m taken to prevent the discharge. and mmsurm adopted to prevent future dis chargesfb> Strlpptr, The conceu(ration of vtoyl chloride la n exhaust fleet die- harced to the atmosphere from each Shipper ta not to exceed 10 ppm. except m provided In |6i.65(a>. ltd* require* meat does not apply to equipment that haa been opened. Is out of operation, and met the requirement to | tl.tolb) <#) ti) before being opened. (e) kftrinp. wighing, and holding container*. Hie concentration of vinyl chloride In all exhaust cases discharged to the ionosphere from each mlxlnc, weighing. or boldine container In vinyl chloride serviee which precedes the stripper (or the reactor if the plant has wo stripper) to tbs plant pmn flow is hot to exceed 10 ppm. except as provided to 141.01(a). Hill requirement does not apply to equipment that has been opened, la out of operation, and met the requirement to 141.49(b) (4) U) before being opened. (d> Jfouomer recovery system. Hie eonoentraUoo of vinyl chloride in all ex* haust cases dlacharfed to the atmos phere from each monomer recovery sys tem Is not to exceed 10 ppm. except as provided to f 41.45(a), This requirement does not apply to equipment that has been opened, la out of operation, and met the requirement la 14145(b) (4) (1) be fore belnc opened. (a) Pourees /oJZovtng the strfppcr(t). The foUowlzif requirements apply to emissions of vinyl chloride to the at mosphere from the combination of all sources following the strlpper(s) (or the reaetor(s) If the plant has no strip* p*r(i>] to the plant nrnrnss flow in- aludlnc but not u*"<**< to, eentrlfugas. ooneentratora. blend tanks. Alters, dry ers, conveyor air dlacharces. baggers, storace ccntolnen. and Inpimiss waste water: addition to stripping, --i--of vtnyl frmn the slip gauge through a control chloride to the atmosphere may not earned: (1)3 g/k# (0.009 lb/lb) product from the stripper (a) (or reactor(s) If the plant has do stripper(a)) for dispersion polyvinyl ehlorlde mins, excluding latex aystmn from which the concentration of vtoyl chloride to the exhaust gssea does not exceed 10 ppm. or equivalent as pro vided in I 61.44. (3) Leakage from pump, compressor, and sgltotor seals: resins, with the product determined on a <0 Jtotating pumps- Vtoyl chloride dry basis; emissions from avals on an rotating (U> 0.4 f/kg (0.0004 Ib/lb) product pumps to vtoyl chloride service are to be ftom the strippers for roactor(s) If the minimised by Installing samli-ntt pumps, plant has no stripper(s> 1 for all other pumps with double mechanical seals, or polyvinyl chloride resins, including latex equivalent as provided in 141.66. If resins, with the product determined on a dry ootids basis. double mechanical seals are used, vinyl chloride emissions from the seels art to be minimised by maintaining the prw- sure between the two seals so that any leak that occurs Is into the pump; by |1.* Emtittoa rtsudard far ribylm ducting any vinyl chloride between the SkfcbrMi. vtoyl chlgrki sad paly* two seals through a control system from vieyl chloride pUath which the concentration of vinyl chlo An owner or operator of an ethylene dichlortdc. rtoyj ehlorlde. and/or poly vinyl chloride plant shall comply with the requirements of this section. (a) Atlitf odfpc discharge. Except for an emergency relief discharge, there is to be no discharge to the atmoephere from any relief valve on any equipment in vinyl ehlorlde service. An emergency relief discharge means a discharge which could not have been avoided by taking measures to prevent the discharge. With in 10 days of any relief valve discharge, the owner or operator of the source from which the relief valve discharge occurs shall submit to the Administrator a re port in writing containing information on the source, nature and cause of the discharge, the date end Ume of the dla- charge. the approximate total vinyl chlo ride loss during the discharge, the meth od used for determining the vinyl chlo ride loss, the action that was taken to prevent the discharge, and msssurge adopted to prevent future discharges. (b) fagittoc cotfstto* sources. (1) tending and unloading totes.* Vtoyl chloride emissions from loading and un loading Unoa to vinyl chloride service which are opened to the atmoephere af ter each loading or unloading operation are to be minimised as follows:* (I) After each or operation and before opening a loading ride In the exhaust gases docs sot ex ceed 10 ppm; or equivalent as provided to 141.44, (II) Jtoetprocsttag pumps, Vtoyl chlo ride emissions from seals on all recipro cating pumps to vinyl chloride servlet are to be minimised by tracing double outboard eeais. or equivalent as provided to f41,44. If double outboard eaals are used, vinyl chloride emissions from the seals are to be minimised by m*t*mirlrir the pressure between the two wait eo that any leak that occurs is into the pimp; by ducting any vtoyl chloride be tween the two seals through a control wstem from which th concentration of vinyl chloride In the exhaust gaaes doea not exceed 10 ppm: or equivalent as provided to I 41.44. (III) Rotating compressor. Vtnyl chloride emissions from seals on ell ro tating compressors to vtoyl chloride sennet are to be minimlied by *"-**ii<"g compressors with double |~i1 seals, or equivalent as provided to 161.44. V double mechanical eeais are used, vtoyl chloride emissions from the seals are to be minimised by aMint<wiy the pree- ure between the two ***lt so that any leak that occurs is Into the compressor; by ducting any vinyl chloride between the two seals through a control system from which the concentration of vinyl chloride to the exhaust gases does not or unloading line to the atmosphere, the exceed 10 ppm; or equivalent as provided quantity of vtoyl chlorldf In all parts of to | 61.44. each loading or line that are (tv) Jtedprooattng compressors. Vinyl (I) In polyvtnyl chloride plants usInc to be opaned to the atmosphere is to be chloride cmiistnrks from seals on all re stripping technology to oontrol vinyl reduced so that the parts combined con ciprocating cempraman to vtoyl chloride Miloride cmlasions, the weighted average tain no greater than 0.0034 m* (0.13 ff> service are to ba minimised try residual vinyl chloride concentration to all trades of polyvinyl chloride resin processed through tbs striding opera tion on each calendar day. maaeurud hnmedlatoly after the stripping opera tion Is completed, may not exceed: (!) 9000 ppm for polyvinyl chloride dUpentoc mine, excluding latex restos; (II) 400 ppa for all other polyvinyl ehlorlde resins, including latex rectos, of vinyl chloride, at standard temperatore and pressure: and a to(II) Any vtnyl chloride removed from loading or mdoaHiwf line accord ance with paragraph (b)(1) (1) of this section is to be ducted through a eontrol system from which the concentration of vtnyl chloride to the exhaust gases does not exceed 10 ppm, or equivalent as pro vided in I 41.44. double outboard stall, or equivalent as provided in I 41.44. If double outboard seals are used, vtoyl chloride omissions from the seals are to be minimized by wintaming the pressure between the two seals so that any leek that occurs is Into the compressor; by ducting any vtoyl ehlonde between the two seals through a control system from which the concentration of vinyl chloride in the averaged separately for each type of rmto: or (3> to potyvlnyi chloride plants otmtioQJnf vinyl chloride s-iisdons with torfinology other than stripping or to (3) SUp geuget. During loading or un loading mwrsumi, the vtnyl chloride tinlislnni from each slip gauge to vtnyl chloride service are to ba minlmtosrt by duettag any vtoyl ehlorlde discharged exhaust games doss not exceed 16 ppm: or equivalent as provided in I 61.46. <v) AgitMtor. vinyl chloride emlesions from seals an A agitators to vinyl chlo ride service are to be axlnlmi--r| by to* A-7 GENC 01 <6660 staDtnt sqltsion aitta tfoublo irhsiH* eal Mil. or squtralem is osvfldid hi II M a oeublt morhsnical Ml air mai, vtnjl ehloxMe sailMlms from tha Mil an to b* mixiliniaad by .**,.t....!.., the prwun bstnea th* two Mil ao that UT lot that oceun Is Into tba ssltatsd Tin ill by duetto* any ytayt chlo rate batwacn tin two Mil tbrouth a octroi system tram which tha ecriomtratlon of Tlnyl chloride hi tha exhaust hw doaa not exceed It pom: or equiTM1st 11 presided hi 1Sl.es. (1) Laakaft from rtUtf taint. Vinyl Mortdc immioni dui to leadate from each rolllf yalvt on equipment hi Tlnyl Monde lerrtee ere to be minimised by fc--T*'IH'r a rupture dlih between the Mttpmest and the relief Talte. by coo- Biethif the relief ealve dkcharvi to a Pio.m line or recorery system, or squlr- aleat ai provided hi I (l.M. (I) Manual pewfine of pom. Ezoipt ai prorlded hi ttl.M(ai(l), all PM which are manually Tented from equip ment in etnyl chloride eerelce are to he ducted throuch a control tyetem from which the concentration of Tlnyl chloride to the exheuit caeca doea not exceed 10 POm; or equiralent ai provided tn I <1.41. Ill Opceinp at Mwipemf. Vinyl chloride emlerUmi from ~"of equipment flneludhic loadlnc or unloadtnj line* that are not opened to the atmoaphere after each loadlnc or unload- hif operation) are to be mtnlmlaed ee fallows: (II Before openlnc any aquipoent for any reason, tha quantity of vinyl chlo ride h to be reduced eo that the equlpaaent nontains no more than 3.0 prroent IT volume vinyl chloride or 0.0#10 m* (SI cell of vinyl chloride, whichever te larva', at standard tmnporature and meeeure: and (hi Any vinyl chloride moved from the equipment th accordance with para* craph (b) (C) (1) of thie eeetloo la to be ducted throuch a control lyetom from which the concentration of vinyl chlo ride In the exhaust gaiee dom hot exceed 10 ppm, or equfvalent ai provided to I ci.ec. <71 Samtltt. Onuaed portion* of earn* plea eentainttlc at laaet It percent by etchl vinyl chloride are to be returned o the procaet, and aampUnc techniques are te be such that sample cneitelucre hi etnyl chloride service ere purved Into a ehwad process system. <ti took detection end tihmtnttion. Vtnyl chloride wnlesVms due to had* from equipment tn vtnyl chloride eervtoe in to be mlnimliid by lnitltutl&c end muuementinc a formal load deteetteo and sUminauon proeram. The owner or operator ahaD eubnut a dm iliillnii of the proeram to the Admtnutrator fee approval the proeram U to be mb- Tinted Within 4* day* of the effective dale of these reculattoni, onlew a waiver of compliance is evented under | Cl,11. If a waiver of compliance u cyanted. the mudiatn la to ho submitted on a data scheduled by tbs Admhtbrtrator. APbrwval of a proeram win bo crantod by the Administrator provided be Bade: (1) B mdudm a raltabli and mam ala vbryl cfalortde mooltortn< system for ds- taction of malor loads and Idssttfleattoa M tha (aneral ana of the plant when a load Is located. A vinyl chloride menltormc system means e devlcs which obtains air samples from one or mere points an a continuous sequential basis and ana- tyias tbs samples with eas ehromatoctaphy or. If the owner or operator asnna thit ill hydroetftefii Dtuund arw vinyl chloride, with infnrtd spectrvpbotm&ttry. flAaf loo detection, or to equivalent or alternative method. It Includes * reliable sod seeurato portabit hydrocarbon detector to be uoed routinely to Had small iMk* and to pin point the major loti indicated by the rtnyl chloride monitoring system. A portable hydrocarbon detector means a device which measures hydrocarbons with a sensitivity of at leait 10 ppm aad la of uch design and dee that tt can be uaed to measure emission* from local- laod points. (ill) It provide* for ah acceptable cali bration add mhitoww aebcdule for tha vinyl chloride monitoring cyvtem and portable hydrocarbon detector. l*w the vtnyl ehlorldt monitoring system. a dally span check to to be eonduotod with a concentration of vtnyl chloride equal to the concentration defined aa a leak ac cording to paragraph <b) (I) (vl) of thla action. The calibration la to be done with either: (A) A calibration cm mixture Pre pared from the gases epactfied In eocttona BJ.1 and 3-2J of That Method 104 and In accordance with section 74 of Test Method lOd. or ai <B) A calibration gaa cylinder standard mntemlng the appropriate concen tration of vtnyl chloride. The cm earnpineltlon of the calibration gmi cylinder standard to to have been certified by the manufacturer. The manufacturer muet have recommended a maximum atoelf life for each cylinder eo that the oonoentratkm dom not change greater than ml percent from the certified value. The date of cm cylinder preparation, oaiUfled vtnyl chloride concentration ami recom mended maximum itaelf fife muet have been affixed to the cylinder before ahip- ment from the manufacturer to the buyer. Zf a cm chromatograph to need ae the vtnyl ehlorldt monitoring system, theee gee mixture* may be directly uaed to prepare a chromatograph calibration curve m deeerfbod in eeetloo 7J of Tbit Method 104. The requirements tn eecOoq I4J.1 and 444 of Te*t Method IP4 for certification of cylinder stand* ard* and for ectabltohment and verifies* tfanef calibration etondards are to be followed. * <tv> The location and number of potato to be monitored and the fraouency to monltorlne orovtded for tn the program are aecmtable when they tn compared with the number to piece* of equipment to vinyl chloride service and the its* and physical layout to the plant <v) It eontain* an anweptabto ptan to action to be token atom a leak to de- tmtod. <vt) It enetatnc a definition of i--y ^hieh to aeceptahto when eomparod with the background ecnoentratlon* of vtnyl chloride In the etoe* of the plant to be monitored by the vinyl chloride monitor ing tyvtem. Measurement* of toekgrou&d ooncentratlon* of vtnyl chloride In the aroM of the plant to be monitored by the vinyl chloride monitoring y*tem are to be Included with the dmcrtptlon of the program. The definition of leak for a given plant may vary among the differ* mt areae within the plant and to alao to change over time a* background con centration* tn the plant arc reduced. <*> Inproctu ipectoiMter. Vinyl chlo ride emtoalom to the atmoephere from taproce*! waetewiUcr are to be reduced m followv: (I) The concentration of vtnyl chlo ride in each inprocea* wactowater stream containing greater than 10 ppm vinyl chloride measured Immediately as It leave* a piece of equipment and before being mixed with any other taprocm* wastewater stream Is to be reduced to no more than 10 ppm by weight before being mixed with any other inproeess wastewa* ter stream which contain* leas tn*" io ppm vinyl chloride; before being exposed to the atmoshere; before being dis charged to a wastewater treatment proc ess; or before being discharged untreated as a wastewater. This paragraph does apply to water which Is used to displace vtnyl chloride from equipment before it to opened to the atmosphere to accord ance with f 01.04(a) (i) or paragraph <) of this section, but does not apply to water which to used to wash out equip ment after the equipment has already been opened to the Atmosphere In ac cordance with |fl.*4<a><2> or para graph (b) <) of this section." (II) Any vinyl chloride removed from the inproeess wastewater to accordance with paragraph <> (1) of this section to to be ducted through a control system from which the concentration of vinyl chloride in the exhaust gases does not exceed 10 ppm. or equivalmit as provfctod to f ei.w. <c> The requirements to paragraphs (bMl), (b>(2), (b)(9), (bMS). (b)(7) and (b> (I) of this station are to be In corporated into a standard operating procedure, and made available upon re quest for Inspection by the Administra tor. The standard operating procedure Is to tadude provisions for measuring the vtnyl chloride Is equipment fee.75 m* <1430 gal) In volume for which an emis sion limit to prescribed to I Sl.*3(b><> (t) prior to opening the equipment aad using Test Method 104. a portable hydro* oarboo detector, or an equivalent or al ternative method. The method of meas urement to to meet the requirements In |41.7(()(|>(1MA> or <*><<)(1MB) " 9&rOh. 114 M the I4t OAC, 74141k Ato Aet e GENC 016661 A-8 f tlM ih4im tmd pH- Upon written xpphcatioc from 4a otur or operator, the Administrator may approve uh of equipment or procedures wttich have been demonstrated to hi* satisfaction to bo equivalent In tamo of reducing vinyl chloride emissions to tba atmosphere to those praaertbod for com pliance with 4 specific paragraph of this support. For 4A existing source, any re quest for using an equivalent method as tb initial----- '-- of oootral Is to ba submitted to the Administrator within 66 days of the effective data. Fbr a haw oource. any request for using an equiva lent method la to be submitted to the Administrator with the application for approval of construction or mortification required by f 61.07. | 61.67 Emimhm teeu. (a) Unless a waiver of mission testing Is obtained under I 61.13, the owner or operator of a source to which this subpart applies shall test ttnlmlnns from the source, (1) Within 60 days of the effective date m the case of an emetine source or a bow source which has an initial startup date preeedlnc the effective date, or (3 > Within 60 days of startup In the case of a new source. Initial startup of which occurs after the effective date. (b> The owner or operator shall pro vide the Administrator at least 30 days prior notice of an emission test to afford the Administrator the opportunity to have an observer present during the test. (c> Any emission test Is to be cooducted while the equipmmt being tested Is operatint at the masifflum production rate at which the equipment will be op erated and under other relevant condi tions as may be specified by the Adminis trator baaed on representative perform ance of the source. <d> [Rwr*dl" <) Whn t ill peaalblc. h hoibI* is to be analysed within 34 hours* but in no case in excess of 73 hours of sample collection. Vinyl chloride emissions are to be determined within 30 days after the emission test The owner or operator hall report the determinations to the Administrator by a registered letter dis patched before the doae of the next but!mbs day followinf the determination.11 <f> The owner or operator shall retain at the plant and make available, upon request, for Inspection by the Admlnistutor. for a minimum of 3 years records of emimlon test results and other data needed to determine emissions. <c) Unless otherwise specified, the owner or operator shall use test Test Methods in Appendix 8 to this part for each test as required by paragraphs (|M1), (f)(3), (gxl), (f)(4), and <i) (l) of this section, unless an equiva lent method or an alternative method has been approved by the Administrator. If the Administrator finds reasonable fiounds to dispute the results obtslnsl tten: It an equivalent or alternative method, be may require the use of a reference c^ = |C*(*JO)Q IP--1 | ICO) method. If ths results of the reference and equivalent or alternative methods do not acres, the results obtained by the reference method prevail, and the Ad ministrator may notify the owner or operator that approval of the method previously considered to be equivalent or C*i=k| etnyt chlorl4/100 k| product. C=Ta* concentration of vinyl chloride oe meaeured by Test Mstho* 10*. iao= Density of vinyl chloride u one sMnmfrhtft and aO* C m kg'm. alternative is withdrawn. U) Test Method 106 is to be used to determine ths vinyl chloride emissions from any source for which an emission Unit is prescribed in II 61.63(a) or <b) | 61.63(a), or 11 61.64(4X1), <b). (e). or Volumetric flow rate in m * 'hr u *. temined by heference Method a or Ap pendix A to Part *0 of this chapter 10--= conversion factor for ppm. = Production rato(fcf/hrj, M (3) Test Method 107 is to be used to <d>, or from any control system to which determine the concentration of vinyl reactor emission* are required to be ducted In I 61.64(a) (3) or to which fuel- ties emissions arc required to be ducted h f 61.66(b)(1)(U), (b)(3)* (b)(6), (b)(6)(11),or (b) (6)<11). (1) For each run, one sample is to be collected. The sampUnc site is to be at Mast two stack or duct diameters down chloride in each Inproeess wastewater stream for which an emission limit is prescribed In | 41.46(b) (6) (1), (3) Where a stripping operation U used to attain the niMtnn unit in | 61.- 64(e), emission* art to be determined using Test Method 107 as follows: (I) The number of stripper* and sam stream and one half diameter upstream ples and the types sod grades of min to from any flow disturbance such as a be are to be determined by the bend, expansion, contraction, or visible Administrator for each individual plant flame. For a rectanvular cross section an at the time of the test based on the equivalent diameter Is to be determined plant's operation. from the following equation: (II) Each sample U to be taken imme (iMftb) (width) equivalent SUmtss] ~lifciftb+rwidth1*" The sampling point in the duct is to be at the centroid of the cross section. diately following the stripping operation. (ill) The cormponding quantity cf material processed by each stripper 1* to be determined on a dry solids basis and by a method submitted to and approved The sample is to be extracted at a rate by the Administrator. proportional to the fas velocity at the (tv) At the prior request of the Ad sampllnc point. The sample is to be ministrator. the owner or operator shall taken over a of one hour* and provide duplicates of the samples re is to contain a minimum volume of 60 liters corrected to standard conditions. (tl) Each test is to consist of quired in paragraph <c)(3Hl> of this section. (4) Where control technology other than or in addition to a stripping opera three runs, Tor the purpose of determintnf mwImImm, the averafe of results of all rani is to apply. The averafe 1* to be tion is used to sUain the emission limit in |61.64<e>, emissions are to be deter mined as follows: computed on a time weighted basis.1* (ill) For cas streams containing more (i) Test Method 106 Is to be used to determine atmospheric emissions from than 10 percent oxygen the concentra all of the process equipment simultane tion of vinyl chloride a determined by ously. The requirements of paragraph Test Method 106 is to be corrected to 10 (g> U) of this section are to be met. percent oxygen (dry basis) for determi <U) Test Method 107 Is to be imed to nation of emissions by using the follow determine the concentration of vinyl ing equation: chloride in each inprocess wastewater C,,* II*HIH**SCh*,, ioaI m stream subject to the emission limit pre Mi-pnwsl O, scribed in f 61.64(e). The matt of vinyl chloride in kg/100 kg product in each In process wastewater stream is to be de MBMBtntion of viayt ehlcne* la ths Uuu*t gsess, corrected is tO-percent oxygvnCmThe Moetntiauon of vtoyi chloride ee measured by tmt Method 10*. termined by using the following equa tion: 1C* *10"*] 11001 Ciiw --- '' 60J m Percent oxygen in the ambient Ur at standard eandluoas. lOJ= Percent oxyiea la the ambient air at etandard conditions. Minus the 10,6-per- mi naygen to which the uMTcriinn la being made. Percent 0,sPrent oxygen in the exhaust fas as measured by aurerence Method S Cat-kf vinyl ehfaridr/tes be pnitart. Ca-lUbytvTn*tvwnirMvUrOvMrt*v1#*7*.lnyl ttoUrt** a wr #* n4 m 1/hr. **mh1m* is wvUm vltb 1 RHtM vhtrtl Hu b*M NlMNIKl M n*uynwl by Ih* jimliUUralf IT*-CviHifn Ufivr tot ppm. g-frvdenwi TV* <M/Vi. SvUfttiwd In etui- in Appendix A of Part 00 of this chapter mv vltb 4 Mta Ntb IMS b**n VatoMUf* (hr) For those emission source* where the emission limit is prescribed in terms * mass rather than concentration, mass (t) The reactor opening lorn for which aa snilUnn limit is prescribed in 161 64 missions tn kg/160 kg product are to be (a) (3) Is to be determined. The numbet determined by using the following equa- of rweeton for which the determination A-a GENC 01666 to to be mad* k to be if mu) by the Administrator for mch Individual plant at the time of the determlnaaoo bawd ec tht plant's opcrmtiOA. a toactor feat is also used w a stripper. the deter mination may be made immediately fol lowing the stripping operation. (t> Cseopt as proTldad to paragraph <*)() <U) of thk seettoa. the reactor opening teas b k be determined using Um following equation; _ W (SJQ) (lO-*) (Cb) Cm------------ n------------ C*IHTl7ltfctoi4tltouWtoeradBcL tdWtom-DCa*npa*dtrtfWMHiNtywWawtotein* etm*. VCm tym*. to4* Conmonlaciflf lor m, CMpvn to ntamo ttftrl ebMrtS* or 4rtlMfl to Ten KtiSod met pa-uate hydmrtoa Oeteeter vkirt M*nm ferdraartoe* vita a wnoitiritf ( at laot io hi I'vHwabv ai bntcha gnor UM fadMr ee Met fin"dttoo tnb*aooiniywvOinpyaime.hii n Um oumtw ai aneo im nmut (A) tf Method 106 k used to determine the concentration of vinyl chloride (Cb>, the tample k to be withdrawn at a constant rate vith a probe of sufficient length to reach the vessel bottom from the manhole. Samples are to be taken for S minutes within laches of the ves sel bottom, 5 minutes near the teasel center, and 5 minute* near the vessel top. (Bi If a portable hydrocarbon detec tor k used to determine the concentra tion of vinyl chloride <Cb), a probe of sufficient lenrth to reach the vessel bot tom from the manhole k to be used to make the measurements. One measure ment will be made within 6 laches of the vmstl bottom, one near the vessel center and one near the vessel top. Measure ments are to be made at each location until the readtof Is stabilised. All hydro carbon* measured are to be assumed to be vinyl chloride. <C) The production rate of polyvinyl rfilortde (2) Is to be determined by a method submitted to and approved by the Administrator, (11) A calculation based on the number of evacuations, the vacuum Involved, and the volume of res in the reactor Is hereby approved by the Administrator as an al ternative method for determining reac tor opening loss for poetpolymcrlxatlon reactors In the manufacture of bulk rosins. tke. 114 ef the CVeap air As* to imoM <41 DAC. 1414a | 61.66 Emission meahovtog, (a) A vinyl chloride monitoring syitom to to be used to monitor on a con tinuous basis the emissions from the sources for which emission limits are pre scribed in <1.03(4) and . I 61.63(a). and I 61.64(a)(1). . (c).and (d).and for any control system to which reactor emissions are required to be ducted to | |l.<4<aj (3) or to which fuflUve emis sions arc required to be ducted to I 61.fi <b)UKll), and (b)(3), (b)(1), (> (tt).aad (b))(11).* (b) The vinyl chloride monitoring syt9*n(s> Used to meet the requirement to paragraph (a) of this section is to be a device which obtains air ssmpels from one or more points on a continuous sequential basis and analyses the samples with cat chromotofraphy or. If the owner or operator assumes that all hydrocar bon* measured arc vinyl chloride, vtth infrared spectrophotometry, flame too detection, or an equivalent or aiternafive method. The vinyl chloride monitorInc y*tern used to meet the requirements in I 61.65(b) > (1) may be used to meet the requirement* of this section. (c) A dally span check Is to be con ducted for each vinyl chloride monitorin* system used, Bor all of the sources listed In paragraph <*> of this atotlon, except the one for which an ml*> don limit to prescribed in I <l.<3<b), the dally span cheek k to be ecnedueted with a concentration of vinyl chloride equal to 10 ppm. Bor the mmion source for which an emission limit to prescribed In | <1.63(b). the dally span check to to be conducted with a concentration of vinyl chloride which to determined to be equivalent to the emission limit for that source based on the era lesion test re quired by f 61.67, The calibration to to be done with cither: U) A calibration *ms mixture pre pared from the cases specified In sections 14.1 and 5.3.3 of Test Method 10* and in accordance with section 7.1 of Test Method 10*. or9* (3) A calibration fas cylinder stand ard contaInin* the appropriate concen tration of vinyl chloride. The ems com position of the calibration *as cylinder etandard to to have been certified by the manufacturer. The manufacturer must have recommended a maximum shelf life for each cylinder so that the canceltraUon does not chance treater than m5 percent from the certified value. The date of *as cylinder preparation, certified vinyl chloride concentration and recom mended maximum shelf life must have been affixed to the cylinder before ship ment from the manufacturer to the buyer. If a cas chromatocraph to used as the vinyl chloride monitorlnf system, these <as mixtures msy be directly used to prepare a ehromatocreph calibration curve as described in section 7.3 of Test Method 10*. The requirements In sec tions 5.24.1 and 4.24.2 of Test Method 10* for certification of cylinder stand ard* and for establishment and verifica tion of calibration standards are to be followed.9* ter(Sec. tl* ef the <43 OAC 7444U | 41.69 tataial report, <a) An owner or operator of any source to which this subpan applies shall submit a statement In writing notifytnc the Administrator that the equipment and procedural specifications to ll.<5 <b)(l), (b)(3), <3>. (b)(4), <b)<>. (bxu, (b)(7). and (b>> are totplonented. (b)(1) tn the case of an source or a new source which hM an initial startup date precedln* the effec tive date, the statoicat to to be submit ted within 90 days of the effective date unless a waiver of compliance to xranted under 161.11, alone with the Informa tion required under I 61.10. If a waiver of compliance to created, the statement to to be submitted on a date scheduled by the Administrator. (3) In the ease of a new source which did not have an initial startup date pre- eodln* the effective date, the statement to to be submitted within 90 days of the Initial startup date. (O The statement Is to the followin* Information: (1) A lilt of the equipment Installed for compilenet. (2) A description of the physical and functional characteristics of each piece of equipment. (3) A description of the methods which have been Incorporated Into the standard operatine procedures for new- urln* or ealculatto* the emissions for which emission limits are prescribed in f fl.65 (b> <l)(i) and <b)(6)<l>. (4) A statement that each piece of equipment to Installed and that meh piece of equipment and each procedure Is bain* used. <>. 114 a* the Ctou AtiKM------ ---<41 9AC 1414a 9 61.70 Scmtoenaal report* <a> The owner or operator of any source to which this subpart applies shall submit to the Administrator on Septem ber 15 and March 15 of each year a report in wrltln* contalnln* the information required by this section. The first semi annual report to to be submitted followin* the first full month reportin* period after the initial report to submitted.90 <b) <1> In the case of an exlitmc source or a new source which has an Initial startup date precedln* the effective date, the first report to to be submitted within 160 days of the effective date, unless a waiver of compliance la granted under 161.11. If a waiver of compliance to granted, the first report to to be sub mitted on a date scheduled by the Ad ministrator. (3) In the ease of a new source which did not have an initial startup date pre ceding the effective date, the first report to to be submitted within 190 days of the initial startup date. <c) unless otherwise specified, the owner or operator shall use the Test Methods m Appendix B to this part to conduct emission tests as required by paragraphs (c)(3) and (c><3) of thto action, unless an equivalent or an alter native method has been approved by the Administrator. If the Administrator finds reasonable grounds to dispute the results obtained by an equivalent or al ternative method, he may require the use A-10 GENC 01666 - of a ref*reooe method. B tbe naol* of ef resin biujimni by the stJrtwx> tot reference and equfralmt w elmnxs- that calendar day. aeoerdtog to the foi* IH methods do oot agree. the raeulm tearing equation: obtained by th reference method pra- mad the Administrator may notify tbe ^wner or operator that approval df lb method previously considered to bo A ^`V*'-^ ^P*1 **?+ ` ' ' +P.V. oquiraient or alternative Is withdrawn. U> The owner or operator shall ta- dude Is the report * record of any emli- where: kt&j which averaged over any hour period (commencing on the hour) ore Id excess of the emission limits pre- eertbsd in H 91.93(d) or (b). 191.93(e), A s M-hour a*nttnUoB of type, r i me la ppm (ary weight beau), dxTotal production of type r i mm orw tbe M-hour period, la kg. MxOoaeeatrouoo of nnyi cdortd* in u aaaple of gride a . mm. In ppm. PsProdticuoa of grad* <3 , min repretented by tb* sample. m kg or III.M(o) (1). (b). (c). or <d>, or for 0zi7 control system to which reactor emimioas ort required to be ducted to f iType ef iwaia: 1= IJ ... where * ta total dumber of min types prodooed dunag tbe Is Sour parted. O (xOrad* of min. *4.. 0 ,, 0,, 4*4 o r axTotal number of gradm of mjn pro- duood during tb* M-feaur period. ** I |i.*4(s) (3) or to whieh fugitive emls* tons are required to be ducted In I *1.91 (bXlXli). (b)(2). (b)(1), (bX9xU>. v (vl) The owner or operator ahell re <V) <U). The emissions ere to be mens tain at the source and make available ured In Accordance with f 91.99. for inspection by the Administrator for (2) Is polyvinyl chloride plents for a minimum of 3 yean records of all data which e (tripping operation Is used to attain the emission level prescribed to |91.94<). the owner or operator shell include in the report e record of the vinyl chloride content m the polyvinyl chloride resin. Test Method 107 is to be used to determine vinyl chloride eontent es follows: needed to furnish the Information re quired by paragraph (eX3)(v) of this action: The records are to the following Information: (A) The vinyl chloride content found to all the samples required In paragraphs (c) (3) (1) and (c> (3) (U> of this section, (1) If betch stripping Is used, one reti identified by the min type and grade resMtatiTi sample oC polyvinyl chloride and the time and date of the sample, and min Is to be token from eech betch of (B) The corresponding quantity of each grade of resin bnmedlswly follow* polyvinyl chloride rasin processed by the tog the eocnpleuon of the stripping op etripper(s>, identified by the resin type eration, end Identified by resin type end and grade end the time and date It trade end the date and time the batch represents. b completed. The corresponding quad* tity of material processed In eech strip* per batch 1* to be molded and tdenU- flod by resin type and grade end the^ date and time the batch la completed. (h) If continuous stripping is used, one representative sample of polyvinyl chloride rosin Is to be taken for each grade of rosin processed or at Intervals of 9 hours for each grade of rosin which Is being processed, whichever Is more fre quent. The sample Is to be taken as the (2) The owner or operator shall In clude to the report e raeord of the mis sions from each reactor opining for which an emission limit Is prescribed to I 91.94(a) (3). Emissions sre to be deter mined in Accordance with I 91.97(g) (5), accept that emissions for each reactor are to be determined. For a reactor that Is also used as a stripper, the determination may be made immediately following the stripping operation. I *1*71 B*c*rihi*plwf. <) The owner or operator of any source to which this subpart applies shall retain the following Information at the oouree and make it available for inspec tion by the Administrator for a mini mum of two yean: (1) A record of the leakj detected by the vinyl chloride monitoring system, s required by |91.93(bxB>, including the concentrations of vinyl chloride measured, analysed, and recorded by the vinyl chloride detector, the location of each measurement and the date and ap proximate time of each measurement. (3) A record of the leaks detected dur ing routine monitoring with the portable hydrocarbon detector and the action taken to repair the leaks, w required by I 91.95(b)(9), todudlng a brief suument explaining the location and cause of each leak detected with the portable hydrocarbon detector, the date and time efthf leak, and any action taken to attmlnste that leek." (2) A record of emission* measured to accordance with | fl.w * 14) A dally operating record for each polyvinyl chloride reactor, pressures end temperatures.** ream flows out of the stripper end Iden tified by resin type end grade end the , 114 ef tbs CUsp A* AM to 1 tia. H4 sf tbs Oma A , dete end time the sample wee taken. The corresponding quantity of material proofed by each stripper over the time period rspi minted by the sample during the eight hour period, la to be recorded and identified by min type and grade and the date and time It repressnu. Oil) The quantity of material proe* mstd by the stripper I* to be determined en a dry solids basis and by a method submitted to and approved by the Ad ministrator. tlv> At the prior request of the Ad ministrator, the owner or operator shall 1 provide duplicate* of the samples re quir'd to paragraphs (0(2X1) and <c> It n *A2(, 4/6/73 (1) (2> 01) of this section. () The report to the Administrator by the owner or operator is to Include the vinyl ehlorlde content found to each 41 FI 44440. 10/21/7* (2*1 ample inquired by paragraphs (e)(2) 41 FI 41017. 12/1/74 (10) (l) and (o (2) (U> of this section, aver 42 Ft 2*004, 4/7/77 (It) aged separately for eech type ef reato. over each calendar day end weighted according t the quantity ef each grade 42 Ft 41424, t/17/77 (40) 41 Ft MOO. 1/1/71 (47) A-ll GENC 016464 mnmtih fetlwl Mn1 ttontorfc f*r "--Air Nllvtuti Colima tut* i. q<r jmucTlOB: Onm *r ntnun ff Ktw *f kniOrn ptllftwu atojact * tto totiiml MmIm tUMr tor Mazarfeu* Air toliftoftts in ftotrirto t* Mil th* lafaimtlto cittototo to toctlto I to tto itomtHito U.J, tudfwwul fotoetia Afwcr toiiaal Offle* jrtor to SO toy* ptwfvtot*rtoMMtittoiitcoktf.tro**ctrtiitno ttotot*ttfofa*liuvftiota#tof rMtoettr oato- A tilt #f ttotiMl offltoi Is pmlfctf to IT.W. . tm mrnw 1. l^rtlfiatlwAwHM . Matt tkMX U*u --t mm fcf. ki W *Sqcnp SfetsE o---------------- hnh.-----------~m i i-i a Knrf Mfrtu U**U T Plat] It I O-----------Etylw----------- X ICR IT" 4E lau miiT kaif"--n wmsE nn ew m A W1 w l*M4 Vm SnI> I4Ei > Wftoto tin M arf Platon* ^p tf tin Mr r ptottfr rtofWikto tfflclil Stoma; CMtoct toiwmln Uli opr*. i VM 11 __to H IkaN H*r *# -maim 9m*).I. Itw pticHrtip* Miflji itm to* tom ff to* inn* (.. ^CM*rtr n " cwi^mm " / 4* Attfimn tollto* tomi - ttofato M iluwtln f*T7Twf ttortu II c*r*litotom* It U t) #1r*et* to 4 tocitta* 4iffr**t Utf mat aftclftoi **. * *9 toll 4 1 _____________________ 1i n nTCBf stmt *r * m*~ i toll J_!_ IU7 nre oni A CmMioct Ititw Tii* RtiitM tnm toll wm tm M *fUH** llMtotton* (PUl<M u to* NatMl {?*! ... |H*r to so ton iftir tto affactlaa fata #f a*jr itimm ar i tola r**#1r* M* MKula vf aa lufarmtlaa. E,,[; i?to*wrhim rni ttlilwi*UTMrtwinfJuWtw.i^BCyfaHClH^aarUTiOMarl-l tWIkotfilTTafMMttootlol*Hi1Tl1lt*-ngTm^l Huilma SItotowntorif** far mitm--i Hr tott*ca*u. to* i*ra rtll to la ftolattan ato MartFart to fatoral aafarc^m arttan* ata*i pm>m* i toi*r af eatollM** *f to* Itort^riittiur r to f.l. totniwu; towtm Apay, tto toMtlp a**o* tor aiM to llato* to toetto* II ff toto tor*. A-12 GENC 016665 to. fr ai r ViiIm n H(iuuf#a paltMt, jlii \m i^jwl H ! -tz(1j ^ pit mwt 4* N )>U TT*"5* Ww m~ "JC" W JT u ur I. Pollutant wta< * Wtcata tte tyna *f tezaHaus pollutant aafttai w tte **ncats. UK I cats ' fr astasias, *! ter teryUlia, aa far treur] _M pollutant MfUiatlOA q IT tc ton - Proviso a Mat assertatIon of oart oroenti (a.a., A ana tax' la asrooty cklop-alkslt plant, 'grinding naeMno* la Ilia aocatM step). Usi additional steats if nacotaaiy. *0 Itaaaaa Dnsa-iption n E I* VI* 1 11--B 11 ID 1 tea Mi f 2 11 B XI ' 75 m W fl 7b IT i fmount Of Mlvtmt - Jndlcit* th* iwfapi night gf th* hazardous ftatarial 14 14m i hmcH aatart th< KQcttl 1a pounds pit noth (hand m th* previous frelvt moths of oporat1on)f #te Mi 1_1 i 1B 11 1H./no. .........'17 zS 3t VT I?4. Thdfctti tbs typt of pollution control IHcm, If any, n*4 to roduco tho missions fm tit* proctii (o.f.* vontwrl *Ctvb6*r, bophouso* wit Orient) end th* oitlmtad pareont of thrpollutant which ttm divles fra th* pnciu gu stress. Ml 14 11--B ftIMUtT CWT10L CCYlCf: fl------ 45 41 mrnry Doric* Jim a un H POrc*rtt Amovtl n-- Iffl clancy 75 n1 to Ml o s 1 n"ro *C8**it emoi Ktfctsi * 47 liawiry Pwrlet mm MM nne. TO 71 ft W fffltlaftc/ \ A-13 GEHC 016666 n. A. MlvTl Of C0*ir*wcr. OMn ar aoeratera af iipcw raahli to operate In OPWlfenct wUh tM MHWl Cnlllten for HUir^qt Air HHwUntf prior to #0 day? after tM effective data ef 1*7 ttanderdt or *eiPiW**M| mich tvovira tM tvaoiiilen ef such Inferwatier */ rawit a Mlver of caviiivi fw tM A#M*iitrataa af tM U.S, fttvlreiwimil fratactlon Ayncy far tM tlra period nacaiotry to lotto!t appraoriate central devices *r Mi Mdlflcetlom to achieve cowl lane*, TM AMiniitrater mtj front a ^tf*r tf cao1lance with M standard far a period not eiceedlM t** yeari fra* tta effective date of the huardevi pelleted standard, if M fled* that pitch aarlod it nactfsary far tM irstallatien af contrail m* that itaoi trill k* Uta* dwrinj tM parlorf of tM Mirer to aisura mat tit* health f pan**** trill ha protected fro* loolnert mdanycnvAt, TV raoe-t Information provided la Sactlar I an! acconpaar tkl* application* Applltatieni mowU a* tarn to tM aorrapriata Cfa rational afflca. I, Pratvtiet Involved * lodleat* the protest or procoirat ootKlnr Miirfew* pallbtaau to M1O1 aalszlaa centra*; art ti h apfV4. L. Cratralt Baicriw tho trap tied type af control device to he added or pndlflcatlon to ha ewdt to tho fractal to radvea tM nliitohl of Mzerdews pellvtenu to on occopuOla level. (Use additional (Matt If MtHUry.) L DcHhe tV BUwrai that will ho token durlnf tM waiver parted to oitura that tha haalth of penoni trill he prottetod free iMlnont onaanparamt. (Use Additional oMets If Mcesury.) S, ' '' M dates ip which tM followfnf Data hy which contract! for rtlllon control lyitM or procot! odlflcations will M Mrdadi or fato r which ordon will ha Ittwod for tM pwrahaio af tM coMonant parts to hcta^llsh 1*1an control or pracoss wllflntloA. M 1-1 T 0 1 7 r n r To *i--mm7n----u Dot* of Ihltlatlan of on^flto ca*ftradian ar Instnllotlon of fission control ofwlF*nt or pracaii cMnfo, : OaU hy which en-flu ceostrvctlen ar Installation af vllllra cantral quipwit ar pracau ndlficatlsn 1i to h* caapletod. 0 17 3 1 r fT'u *r to (i1 mwm--a c Doto hy which final ca*11*nti It to ho achieved. DM 1-1* 0 < 1 __ 17 II O IS t5o0 11--rnmm--a w I. MAIffl Of tmxiow Ttrts. A -elver af vliHon Ustlnp nay ha fronted ta pitrt ar awnlrn af tawtti af horylHt* or wrewra pellvunti If, in tM Jwdywnt Of tM Aooialitrator of tM tnvlranwowtai Protection Aftncy tM Million* frgn (M tawet ca1y with tM appropriate stanceM r if tM owner* or operatort of tM towrea Mra requested a wolvar af conpllance #r kin Man fronted a wolvar af cwllract.' Thil application should acca^any the popart infprwatlon provided In Soctlon I. 1. Melon lute tM roetori for roovettinp a ralver of 0*11*1*1 toitinf. ir tM Melon ttotad 1* tMt tM Mlttton* fra* tM lowrc* era wHnln tM preicrlMd Haiti, decantation Of tnli coMHIer oust ha attoCMd. V "Bt7 I'ifiotwra wf tM a*eier ar eparatar A-14 GENC 016467 4.141 flto tato--Typ* * (or aqditotoat). ablartd* la aitrofeo cyUadml fur vhicb to* OlfLOUB non VfittDlMT a* rhad a to* proto *o that ta* Bampllnf pa* eompoauian u*s b**a oertlfled by to* Bev r*w aaa to rfui*t*4 proporaoaal to manufacturer Th* manufacturer must ban rmTiirmanri of this method should net bo ntmwptad 07 ptnooi Hiif.inin.j vita tb* ^wwuoa of a pj chromatograph, dot by ttw who on unfamiliar with sourm ms* Ht as them on mtay details that on tojiiiiil to* ooopi of uiio prMinmion. Ctn to* nock |u T*loclty 44 Oampl* raco**ry. 44.1 Ttibiaf--Taflo*. 4.4 pa oomld* fi*m*t*r. to ooaaaet tof to f** ebrooiatograpb tompl* loop. A n*v iirniatfl pt*s i* ploytd for acb **n of b*f tampla* tb*t aoartituta* *a ml*loa ta*t, *ad la to to dto- r*nomrn*nd*d * muimum tnelf lift for **cb cylinder *o tost to* cone*atr*tlo& dci not ebanfe fr**tr to*a =1 percent rrom to* e*rtifl*d T*iu Tb* dst* of (u cylinder prep aration. ecrtld*d rtnyl ebiende oonearu*Uos and recemmnd*d maximum to*lf lift tat to exercised to prevent exposure of oard*d upod eooclualoa af analyala of Uw muct h* been *fBt*d to to* cylinder before sampling p*moah*l to vinyl eWorld*. a eor- tof*. shipment from to* fa* manufseturrr to tha 44 Abolyvl* buyer The** fa* mixture standards may be I. Principle oaf Applicability. 1.1 As integrated tof ouapu or stack ^o rniptamtof vinyl eWorld* (cWotvetoaae) to eublected to chJwmatocrspWt analysis. toteg tome looltotioQ 4totot." 1J Tb* method l* aopllmbl* to to* m**a- wtoMDt of vinyl eWorld* la at*el pi* from ethylene diehiortd*. vinyl eWorld* u4 poly* otoyl eWorld* manufacturing prooooooo, t- aipt wtort to* vinyl eWorld* la oodtiltol in 44.1 Oa* cbroaiatofrapa--Wtth flin todleatloa d*tetof. pot*ntlom*tJlc ftrlp edart r**ord*r aad id to S4 ml &*at#d h* pQn| loop ia automatic laapl* 4J4 CAromatOfrapktc column. Staidly ataal. 3 m x 34 mm. eoatai&inf e0 MOO b*o& Cbromaaorb 103 A toeondary eoluma of OB BP<-04.30 pareut oa OO/AO m*ak AW c&roB*- wrt P. *t*JW**a *t**l. 3 m v f4 mm or Para* pad T. ao/loo aiaaa. ttaioiaM it**l. 1 mxt J directly used to prepare a cbrcmatofrspb calibration eurr* as deacrltod in section 7 3** 044-1 Cylinder t*ndsrd< cerri^ct<on Tbe concentration of elnyl thlorldt in nitrof*u la aacb cylinder must hare torn certified by tbe manufacturer by a direct uis!ti:i of each cylinder usiaf an aaalyticsl procedure that toe manufacturer bad calibrated on the day of cylinder analysis Tb* calibration of particulate matter. mm U raquirad if er*taidhyd* u praaant. tr too analyttcal prooodurt slmll, a* a minimum. S. Boo* and fleasitmty. aaod. a eacoodary eoluma la placed after lb* ba*e utlllaed a torm-polnt nilbration curee no lover limit of detection vJ vary t^ Cbromuorb 103 eoluma. Tbi eomblna^ It la recommended that tb* aianufacturtr onrdtnf to ta* chromatograph used. Valum eoiu&uu aaouid tb*n a* operated at 130* cr maintain tvu eabbrauoa stsadards and us* topui t*a iaeiuto i x lo** mg and x Id-* 444 fiov a>tn (3)--RotamtUr typ*. tom* ftaadard* is to* follow!nf way. 111 A 0 to 100 m] 'mia opacity, vita flov eoawl blfb ooaoeatrauoa standard (totvecn so and I. /etr/ptto Acetaldehyde, vfeteh odd eaiaa*. 100 ppm) for preparation of * calibration ooov m eoBo vinyl eWorld* saurem. win mIowh vita ta* vinyl eWorld* pool from ^ ehrumacorb 1091 column. #* iiftHM 4JJ tad 0.4. If ttoOltttioa Of ta# vtnyl otoiorid* pool is autl aot sattufactory for a particular sampt*. toon chromatograph pa* nMtin c*a to further altered vita prior fyr~--U of tb* AdmiUltrator. Zf alteration of tb* chromatograph panm*t*n falls to vwatev* ta* nay] eWorld* peak. ta*o sup 44.4 Oa* r*fulatqr*--For required gaa eyUadori- 444 TTvrrmittnaiiar- rtmiraT* to oa* dtfTve ea&tlftad*. to meaaurv tamperaturc of heated eemplt loop at Um* of tempi* tn)*c- 44 4 Baronxtor Accurat* to ft aim Hf. to ma*ur* atmoepherle pmeuta areuad f* ebromatofraph durtof aampi* abalyvu. 4.7 Pump toaf-frt* **" capae* cure* by an appropriate dilution technique. (3) a )ov concentreUda standard <tot*en 0 and 10 ppm) for ecnAcation of to* dilution technique uwl- ^ 0444 rf*6H*Amvir and ocei^eatton 0/ ceiibrerum standards. Tto concentration of each caiibrsUot) etandard must bare tots established by the manufacturer utlnf tollable procedures Additionally, tech Calibration standard must bare been veri fied by tb* manufacturer by one of to* plemental confirmation of ta* nay! rtiwtii toy 100 ml/mia. ptob through oa abooluta analytical tceh- 4.4 Calibrauoa. foilowlnf procedures, and to* aftoeme&t between tb* initially determined concen ffininod.**to^u*. nub h Boat spactreoeopy, Buat to 4.4.1 Tubiof--Taflaa. *.4 mm cwtJdd* tration value and toe verification eonctn- dlamtter. pvparat* place* marked for each wmtlon value must to within = S pereenf A Apparatus. m .l lampMtof (Figure 104-1). 4.1.1 rmto tHilulM ataal. Pyruf CM**, ar Tbflan tubing according to stack i*mp*r- iw*. each equipped vtta a float *m plug a remove particulate scatter. 4.14 ampi* tin* Teflon, 0.4 mm outside ammeter. of auoeiant length to connect proto to tof. A orv unnmd pi*o la empio^d far aacb ** of tof aamplm that constltutm aa emission laet. 4.11 Mai* (0) aa4 ftatli (0) etalnlaB toot gulf t mnneri* vita toll check* (ova ealibrauod ooaor&tratioa. 4.44 Tedlar tof* mrtcea-lnea equarv toe, peparmM baf marked for each calibra tion concentration. 4.44 Byrtaft--O.ft ml. fa* tlfbt. 4.4.4 iyiinf*--ft0l. fa* tlfht. 4.44 Flow mater--detain*ter typ*. 0 to 1000 ml/mia r*cf* accurat* to 1%. to m*tr aitrofed to preparation of vtabdard faa aitxtur**. A,ft Otop vateh^y fenova accuracy, to um fto Bov in preparation of *t*adard fa* tolfturea. ft. Btofe&t* It to d*ccai7 that all tea- (l) vertiflcetion value determined by com parison with a calibrated vinyl eWorld* permeation tub*, (9) vettflcation value determined by cemp*rt*on with a fas mix ture prepared la accordant* with to# pro cedure deacrltod In cectlon 7.1 and mint .a-*- percent nayl* chloride, or (9' verifi cation **lu obtained by bavtnr tb* calibration standard analyzed by to* Na tional Bureau of Btnadsrds. All calibration standards moat to voo*w*d on a tim* interval eon*1*tent with the shelf Uf of tb* cylinder *landards ftH * 4. Procedure. pair without) located a* abovn la Flgur* feata be of chromalofrapblc frade. d.i Bampitof. Amembl* to* sampt* train ft.l Aaaiyai*. as la Plfur* 104-1. Perform a bay leak check 4.1.4 Tadiar top. loo Utar cnpadltT--m 0.1.1 Ballum faa or dltrofea fa*--Zero aocordlBf ta fiction 7.4. Obecrv* that all contain mmpi*. Tbflcn bog* aro sot accept nd*. for chromatofrapblc earner (a*. ooueetton* batmen th* baf and th* proto able. AluialnMvd Mylar tof* stay to nato, 0.14 Bydrofvd ***--Z*f> frad*. aro tlfht. Place th* end Of tb* ptoto at tot provided that (to aampi** aro nnaiyued 0.1 J Qcyv*a fa*, or Air, aa required by centroid of to* stack and atari to* pump vttbin M boon of collection. IBa 4*toctor~oro frad*. vtto tb* hiidlt valve adjusUd to yield s 414 Bifid tovfcproof eaotalficr* for 44.4. 14 CafiOrattoa. Oa* oaa of tba foltovtof Bov of 04 lpa. Aft*r a period of time ua- vltb ootonaf to pnitact oodtanu from aun- optica*: ntb*r ft4.1 and 844, or ft44.M altnt to purfv to* lln* atvvral tlm*s ha* 4.1 r<yl eWorld*. fJ+ perecet. Pur* l*fd. eocmect tb* vacuum lln* to to* 4.1.0 BtodJa valvw--To adjust i DPMI Ttayl ebiorida fa* eertiflad by tb maauf*c> baf snd *vacuat* to* baf until to* rotam* toper to contain a mtolmupi of 004 parwdt tar indicates ho flov. Tb*a reposition to* f.l.l rUB^UH tin. otoyl eWorld* for uo* la tb preparation of aampi* aad vacuum lln** and toftn to* ac Ity 0 liton par mmtiu. ftaadard fa* miftuf** la Bectlott 7.1. If to* tual sampling, keeping to* rvu proportional 4.1 J Cbaretoj tuto--To prmnt aimk. |n BbDUfaeturer malnttln* * bulk cylinder to th* stack velocity Direct th* <as *kitinf Ma of nayl eWorld* to atmoapboro to neto* Upply af H4+ percent nayl ehlond*. to* to* imimrttT away from sampling p*r*onncl Ity of aampton. 4.14 ftov **<>--^to itto*mnf aaBpto Bov toa; aapaoi* af awntflai a Bov finfi from 0.10 to l.oo ittar par Btauta. A1.10 Cratailff hiMuf. Ttflto. 04 IB ootoda fUB*in. to taaiablt oBpla tovlto (Hfuro 100-1).* ecruflcatioD abalyai* may bare bean par- At to* end of to* aampi* penpd, shut of to* farmed oa tola supply ratoer than on each pump, diaeoaaect to* eampt* Ua* from to* 1*4 cylinder prepared from tola bulk supply. The data af faa cylinder preparation and the bag. and dlacosnect to* vacuum lln* from too bag oostainar. Protect to* bag container from sunlight. ,,nm*ii enaiyata suet have bean afliad ta 4 Jcvtpl* sforepr fiampl* bag* must b* tbe cylinder btfor* abtpcncat from tb* faa bspt out of direct sunJifbt When at all manufacturer to to* buyer.11 penalhi* aaa]yet* is to b* performed within 044 JMedfm fa*. Zero frad*. for pwpn. B4 hours, but la no me* is *%mm of 73 Kaittov af into a i ai i av aptrtf i prvdvoto daaa pot Btoatuvto avdonaBant by too tatton af standard fa* misturea " bows of aalicctica * 44 Bample recovery With c p*oea of T*r- town otobdortoa (M. 10. and 0 ppm nfl ftm tubing M--tided tm that bag. connect a A-15 GENC 016668 L mtet vote* te M* flee ^nsutegriih Tm. flvttoh th* *SW to Withdrew |M CMS Ml Ml*flb ttt totepll iMf. ftusb (M >wnt s the smpte fs PMII CMS tlM M>U ValV* t* tfc* l**k*frv* pump. u4 tbea to * (hirnnil Mb*. follow^ by 0-100 ml/mm ntibiur with flow ten* toot filw. M AAalyvl*. flat the column ttimmiuin S 100* c, th* IfMor twptntun to 140* C. end the sample loop temperature to TO* CWh*a optusum hydrogen end oryg*D flow rates hero b#n determined verify and mam* Mis them flow fftiao during all chromate* ^aph npararimn U*ing a*ro helium or hMrogaa a* th* carrier gaa, ooteblUb a flow iau in th* rang* eonaiateot with th* menuMotors'* faqufrvsaati /or stltfaetory do- meter operation. A flow rmto of approxi mately *0 ml/mia should produce ad*qu*u eparatioos. Qiwerra th* boo* tine pcriodi* caUy and determlha that th* ootaa level bw MabUlaafl and that baaa Um drift baa hh4. tvp the eampla loop for thirty teeoodi at th* nu of 100 ml/min. than kumu tha toapl* vmiv*. Tirt tha injection Us< (tha pnaitinn of tha pen on tha chert at tha time of aempi* injection), tha cample number, tha smpu loop tatapaiatiira. tha column tem perature, eamar faa flow rata, chart ip-d kb4 tha atttowtar setting. flatord tha lob* saury praaaur* From tha Chart, aatact tha peak having tha retention tlpaa cocreapoadtac to vinyl ablerIda. aa datanninad to flee- toen 7j. Moaeum tha peek esa, Am. by uee of a dtae iniegmor of a planlmcter. Measure tha pact Might, M.. Beoord A*. Ha, and tha ratantion time. Hepeot tha Injstion at Meet two tlmaa or Ohtil twp consecutive vinyl hiorid* pwn do not vary in itv more than *%. Tha avaraca value for tha* two um wtu ^ uaad to compute tha baa concentre* Compar* tha ratio of M. to a for tha vinyl amend* *wpl* with tha aama ratio for tha Sendarg paat which ta els*at tn balfhL Aa a guideline. if thaw ratio# tttflar by mora than io%. tha vtnyi chlertd* p**h may not ha pum (poetebly acetaldehyde u praoant) ad tha secondary aotumii should ba as* ployad (aaa flactlon 444). 14 Measure tha aablant temperature and barometric pressure near tha be*. (A tM relative humidity to ba 100 potoant.) prom a wator mturation vapor praoaura tabla. determine and >award tha water vapor ooocat ot tha beg.JO T. Calibration and Standard*. 7J Prcpereteoe of vdayl eblertdc cteodsd pw mfefers. Kvecunw a oirtaaa-lnoh querv Tbdlar bag that hs pamad a laal bach (described tn flactlon 7.4) and motor s I lltara of nitroan. Whfla tha bag U tiling, us th* 04 ml ryrtno* to intact M0*1 of Hi+ poreant vtnvl chlonda through tha wall of tha bag. Open with drawing tha vyring* ooadla. Ismadlataly fr tha raaultlng hela with a plan of adhesive tap*. Tba bag now rootalna a amyl ahlonde aooaantoatlOfl of 0 ppm. tn a like manner use tha othar iyruit to prapara faa mixture* having 10 and pps vinyl ahlonda aoneantvntlOOA place aach bag on a Mooch aerface and alternately daprs oppoalta tedm Of tha bag M time* te fwtbar mlh tha gaaaa. Thos faa mtftura ataadarda may ba uaad for 10 flay* from tha data af ptopoMOon, after which Oas pvnperatien d new gaa mlxtuis la ragnirad. (OiTTWiir --ffiontamlnetlirn may ba > lam wbao a bag I* rouaad if tha now gas eaotoation than t^ praviow gaa atenoarddid.)* 14 Oatennnation af vinyl ahiartfla vottotba time Thia aactton ann ba perrersod taaujmneouaiy wtth flvms 7d. Mtobitah oMtttona Maatleol with thoa* In fl*etton 14. abo*. fl*t aitenuatcr to X 1 paction. Flueb tha aamplmg loop with sro hllum or nltrogan and activate tha aampl* valve. Kacora th* in)eetion tun*, the euaple loop temparatur*. the column toanpMature, the cam*r fa* flow mu. tha chart epeed and th* attenuator anting. Herord peak* and detector mpente* that occur tn th* abeenee of vinyl chloride. Meintejn sndltlon*. With th* equipment plumb ing arranged identically to flection gj, flueh th* aampl* loop for >0 *aeonS at the rate of 100 ml'Oiin with on* of the vinyl chloride calibration mixture* and activate the aampl* vnlv. fleund the in)acUoa time, fltiact tha peak that eorveaponda to vinyl chlortd*. Mesurv tha dhtanee on the chart from the inlection time to th* time at which the pack occur*- Thia quantity, divided by tha chan p*ed. la defined a* tha retention 74 Frapenartoa of ohrontefoprepA eailbvwrlow aunw. Make a gaa chromatographic uresant Of each gs mixture atandard (flaacrlbed in aactton 444 or 7.1) ualng w&* dltfon* Identical with thorn lifted in aectleoe dj and 4.4. Fluah the smpllni loop tor 40 eaeond* at th* rate of 100 ml/mih with each tehdard gs mixture and activate tha cam ple valve, flaeard c. tha concentration of vinyl ahiond* injected, the attahuator st- ttng. chart apaod, peak ars, aampl* loop temperature, column temperature, carrier gaa flow rate, and retention time. Hound the laboratory prseure. Calculate At, th* peak erae multiplied by the attenuator siting. Hopeat until two Injection ares are within percent, then plot tbeec points v. c. Wban the other aonovntretioh* haw bsn plotted, draw a smooth eurw through the point*. Perform calibration dally, or befora and after each eat of hag aampls, which*w* u mere frequent.^* 7.4 Bag teak chsk* White performance of thi* eecOoo la required eubavquent to bag to*, tt to also advtead that tt m pertsmed prior s bag wa*. After each ua*. sake aura a bag did not develop teaka s fuUu *. To teak hack, mnnact a water meneseter and prea- tha bag to t-io cm M,0 (4-4 tn 11,0). AUow s Rud fs id minute*. Any dliplessent In the water manometer indicate* a leak. Ateo check the rtgid siaiwit fs teake (gn An aisrnatlw teak check method fc S piesurme tha bag to 4-io m h,0 ot -4 |a. H^> and Stew te mini overnight. A deflated beg fdteeteo a teak.) P aach ample Mg in tm rigid container, place a rotameter la-Une between the bag and the pump inlet. Bvecuete the bag. Failure ef the rouseur to reglsttf atro flow who the bag appear* to be empty indicate* a leak. I. Calculation*. i.l Det*rmin* tha aampl* peak area a* follOWl! A'-AmA, wteri, X'-Tte wBibte S*k A*-T4 niBUiJ peak *m. A~Tbt muwiton helm, Equktlon 104-1 44 vinyl chloride eonce&tratlona. Ttas the ealibrauon cum deecnbed In flection 74. above, sleet th* value of C, that as* locpotirt* to Ar, the aampl* peak ana. Cal culate C, s followe; r CtP.T. Wbm: SgUAtloo 164-3 4t*Tb* w**w maw matmi af tb* tea Ntotn. w *B*ir *4. CvwTIh toBvatiads f vttryi chteite la tha teg C*Th^ewBMum f vtayl chlwis Hiflwiag by tte gw eftnsetecraph. l> ppm. P*Tte imwrrat- prttour*. tte kteiley pvwaar* wriliil datis caMbfvUce, ton tt*. TipTW ** hep i*top*nt*f* ia* item* S *t ite tS* M UAiyai. *X. F.-Tb* nbwiiary wen u um* *f wwlyWa. s Hi. r,-Tte ribwe** wasmart, ite wwpb tea* isaei m ns tea snac mkteauaa, *. Hefarcnos. I, Brows. D. Loy, I. w. and itephan- os. u. H. "Vinyl Chlortd* Monitoring ?*ar . P. Ooodrich Chemical Company in Lnutevdte. Kentucky " kegtoo IV. UJ. Bnvi- meminiil Protectmo Agency, BumUlaos and Analyte* tHvlalon. AthCA*, Oeorgla. ^un* 14. 1*74. fl. "Bvaluattefl of A OoUeCtlOfl and Analy tical Procedure fs Vinyl Chteted* tn Air,- by O. D- Clayton and Iwnriatee. Dii--iir tfl. 1*74. EPA Coetract |4o. B-d9>l40fl. TM Order Bo. 1. SPA Bepert aB. 7I-VCD-1. I. "fliacdardistion of itatiotury floune tateeton Method fs Vinyl Chlortd*.** by Mid* wst Bssrch xnetttute, i*7d. wa contract no. m-on-iood. nok Ovds no. i. (4m. 114 or the Ctenn Ah Ate s smtei 441 OJ.C. 74141*. 4M* r L L L C i 5 Lm L % t. r * A-16 GENC 016669 "5 Mm mo* 107--DtrnicxmeM o* Vmn. cwlopb owtwt or Imooai Wuttv41s lutrtA, n Tom rwiiocme Corrcrr or PoLTOTL CHLOUDI KtMtH. Sinii, Wn Tm fm msnns of this method should mi be attempted by person* uniuni]1a/ with th* operation of 4 gas chromatograph. nor by tko who f* unfamiliar lu> sampling, u tb*tt tfi mu7 details that in beyond the oops of this pnHauilfiD Cor* must be exorcised to prevent expo*ur* of sampling personnel to vinyi chiortd*. t osjctnogea. 1. Principle and Applicability. 1.] Th* buu for this method relate* to tlM vapor equilibrium which is miabiitnod between flVCJI, PVC. rmm. water. Ud sly in o oloood system, it fau (wis demonstrated that tb* RVCM in 4 PVC min will equlltbroM la 4 eJCMd vul quite rapidly. pro* vld*d that the temperature of tbo PVC min to maintained abort the glass transition temperature of tbat specific min. 14 This procedure la suitable for dator* mining tb* vinyl cboridt monomer (Vcmi content of Inprocoaa wastewater samples, and tbo residual Ytnyl chloride monomor (VtVCM) contest of polyvinyl chloride (PVC i mine, vet alt, slurry, and latex samples 2t eannoc be umd for polymer lit fused forau. such aa aboet or cubes. If a moluuon of tbe rtnyl chloride poak la not otiafactory for a particular sample, then chromatograph parameter! may be altered provided that the pmtaiou and reproduci bility of the analysis of rlnyl chloride cylin der ftudarda are not impaired, if there la reaeon to believe that eome other hydro carbon with an Identical retention time la paint in the aample, then supplemental ooaflmaUon of the vinyl chloride peak through an absolute analytical technique, ouch mjm*4 apoctroacopy. abould be per* formed.3* 3. Mange ud denaltlvliy. The lower limit of detection of vinyl chlo ride will vary aooordlng to the chromato graph need. Values reported include 1 x 10-T mg ud 4 x 10-T mg. with proper calibration, the upper limit may be extended aa needed. I. precision ud Reproducibility. An Interlaboratory eomparieon between seven labor*torive of three reels samples, each split into three pane, yielded a standard deviation Of 3.43 "i for a aample with a mean of a.M ppm. 4.io^ for a sample *uh a mean of i.oc ppm. and 64**3 for a aample with a mean of 43.44 ppm. 4. Safety. Do not release vinyl chloride to tbo labora tory atmosphere during preparation of stud ents. Venting or purging with VCM/alr mix ture* must be held to a minimum- When they art required, the vapor must be rwvted to owteid* air. vinyl Chloride, even at low ppm levels, must never be vented inside the laboratory- After vials have been analysed, the pressure within the vial bust be vented prior to removal from thV Instrument turn* table. Vials must be vented into an activated charcoal tube using a hypodermic needle to prevent releeee of vinyl chloride Into the laboratory atmosphere. Tbo charooal must be replaced prior to vinyl chloride break* throughA. Apparatus J Rampling >1.1 Mottles 40 ml (1 w|. with waaed toned screw on tope, for PVC samples. 1.14 Vials-->0 ml Hypo-rials.1 aled with Teton faced Tuf-Rond diets foe water earn* ' i.l 4 Beetnoal tape--e equivalent, to prevent loosening of bottle tope. |J mmpie recovery. >44 vial*-->WUh aeale and eape. Ferklnttoner corporation He. IM-oil*, or equiva lent. gj_2 Analytical balance `Capabls cf weighing to =0.001 gram. >44, syringe. 100 al Ft a.ielnn Mcrlce preparation of a oahbratlon curve by an ap propriate dilution wcnmque. (31 a lew conanno-iuo* standard (between >0 and Aoo ppm) for irnifleation of the dilution tech nique used." 4.14 ritebiukftw*! end oeri/tcattea of Hhbratlen ftsadard# The concentration ef each calibration standard must hav* been mtablished by the manufacturer using reli A" Ho. 010033. or equivalent. able procedures. Additionally, each callbrs* >4.4 Vial Sealer, Perkln-Smer Ho. 106- toon standard must have been verified by the 0104 or equivalent. manufacturer by one of the following proce AJ Analysis dures. and the agreement between the ini AJ.l Oas chromatograph--Pmkln-Bxner tially determined coneenbwuon value and Corporation Model 7-40 haad-epaoe ana- the verification concentration value must be lymr. No. 104-0001. or equivalent. within s3 percent: (1) Veriflmtion value de* |jj cbremetoyrcpMc oolvms. flTalnlrn tarmlned by comparison with a gas mixture steel, 1 n X II &m containing 0.4 percent standard generated in a similar manner to Carbowax 1300 on Carbopak A, Perkln-ttmer the procedure described tn section 7.1 of Onrporatlon No 103-0113, nr equivalent. Method 104 for preparing gas mixture stand Carbopak C can be used in place af Carbopak ards using 991+ percent vinyl chloride, or A. If methanol and/or acetaldehyde is pi ee (3) verification value obtained by having the ent lb the aample, a pair cf Poropak Q col* calibration standard analyiod by the Nation twn m eerie* (l m x 44 mm followed by al Bureau tf standards. All calibration stand inxll sup) with provision for hackflosh ards must be renewed on s time interval af the Arvt column baa been shown to pro^ consistent withthe shelf life ef the cylinder vide adequate wparetlon of vinyl chlondfP Standards sold * 64.3 Thermometer--0 to 100* C, accurate 7. Procedure. to =0.1* C, Perkih-dmar Ho. 106-010* m T.1 Rampling. equivalent. 44 4, hample tray thermostat sysleiu 7.1.1 pvc sampling--Allow the reetn or slurry to flow from a up on thr tank or sue Prkin*rimer Ho. 106-0103. or equivalent. >44 hepta-^udwich type, for auto- matic dosing, 13 tan. partm-Uinar Ho. 1051000. or equivalent. >4.4 Integrator recorder -- Hewlett Packard Model ssaOA. r equivalent. A4.7 Tllter drier amenably (j)--Perkin* Xlmer Ho. 3330117, ox equivalent. >44 Soap Aim flowmeter--Hewlett Pack* #d No. 0101*0113, nr equivalent. A.4 Calibration. 4.4.1 Regulators for required gaa cylin ders. until the tap line has been well purged- Ex tend a 40 ml sample bottle under the tap, All, and immediately tightly eap the bottle. Wrap electrical tape around the cap and bottle to prevent the lop from loosening. Place an Identifying label on each bottle, and record the date. tome, and sample location both on the bottles and in a leg book. 7.14 Water sampling--Prior to use, the >0 mi vials (without the diemt must be capped with aluminum foil and mufiled at 4Q0*C for at least ooe hour to destroy or A. Reagents. remove any organic matter that could in 1.1 Analysis. terfere with analysis. At the campling loca 4.1.1 Hydrogen gas-eero grade. tion flu the vials bubble-free, to overflowing 4.14 Nitrogen gaa--aero grade. g.14 Air--aero grade. so the: a conve* meniscus forms at the top. The excess water U CUplsced ss the sealing 44 Calibration. disc is carefullv placed. Teflon side down, cn 44.1 Cylinder stenderda (4], Qae mixture the opening of the vial, plaee the aluminum standards (40. 600. 3.000. and 4.000 ppm vinyl chloride in nitrogen cylinders) for which the gas composition has been oerttAcd by tbs manufacturer. Lower eooeontmtoOd stand seal over the disc and the neck of Uit vial and crimp into place. Affix an identifying label on the bottle, and record the date. time, and sample location both on the vials and In a tog book AH samples must be kept re ards should be obtained If lower concentra frigerated until analyzed. tions of vinyl chloride samples are expected, 74 Sample recovery. Samples must be run as the intent la to bracket the sample con within 34 bourn centrations with standards. The manufac 74.1 Resin samples--The weight of th* turer must have recommended a maximum reals used must be between 0.1 and 44 grams abelf life for eacn cylinder so that the con Aft exact weight must be obtained (oooi centration does not Chengs greater than =4 percent from the eertlAed value. The date of gas cylinder preparation, certified vinyl ahlorldc eoncentreline and recommended gram) for each sample. In the case of sus* pension mint a volumetric cup can be pre pared which wtil hold the required amount Of temple. The aample bottle It opened, and enelf life must have been affixed the cup volume of min is added to th* tared to the cylinder before shipment from the sample etal lincluding septum and alumi manufacturer to the buyer 34 num cap). The vlii is tmmedlstcly etaled 4.1.1 Cylinder etaadards eevTtfleeMan. and the exact sample weight is then obtained The concentration of vinyl chloride tn nitro gen in sech cylinder must have been noti fied by the manufacturer by a direct analyets af each cylinder using aa analytical proce* Report this value on th* data sheet as it u required for calculation of RVCM In the e* of relatively dry min samples (water content <04 weight r-l, 100 .1 of distilled dure that the manufacturer had calibrated watr must be injected into the vial, after on the pay of cylinder analysts. The calibra skiing and weighing, using a 100 1 tynnge tion Of the analytical procedure shall, as a In the can of dl'psnlon re*lns. the cup inmiuni have utilised s three-point call* eannot be u**1. The sample is instead brsUon curve. It is reuomminded that the weighed approximately in an aluminum dt*h. manufacturer maintain tvs calibration transferred to she tared vial and weighed standards asd use these itandarde in the accurately in the vial The aample is then following wayt (l) A high ooneentration placed in the Perkln-Dmer bead space ana- fsaderd (between 4*ooo and moo ppm) to lybsr (or equivalent) and conditioned for one hour at 00C. 1 Meouon of trade isirn an speciAe prod* Norx: mam* aluminum vial oape have * wets does not aowetituse endorsement by th* cantor mum which must to r*mu*ii prior tovironmental Flutes mm Agency. so placing into esmpu irmy if not removed. A-17 6ENC 01647Q HrtWil <lW|f 10 th* kfcfSCUaO Mtt Will MW. TAJ Bucpensioa ream slurry aod wot 0U1 wrln ftt-irrt must be filtered using * til Buchner funs*) with nevus to yield wtt au*. Tb* filtering promt must be con tinued only u long " t steady *tram */ nttr t axitiar froft; th* fun-*l Excw-iv* filtration tiw could mult in tome loot of ycW. Th# wot cake samp)* <0.10 to 44 gramsl 10 044*4 to o tared viol (Including septum u4 aluminum mp) on* immediately hoi*4. Bampl w-lgM It then determined to 3 deel- moJ pioon Th* eampl* It then pliced In tut Perkin-Blnicr hood tpoo* xnalywr (Or equlvs- loot) od4 oondltionef for oo* hour it *0*C. A --r`` of wot ha it u*ed to determine Tf (totol solids i. This is required for oalcu- feting th* xveu. TAJ Dispersion min slurry sample*--- This matariaJ should not' be filtered. Bampie pun o* thoroughly mixed. Using s tortl viol (including septum nod ilwustm oap) odd approximately drops (04S to 04# grama) cf slurry or lata* using 4 median* drupptr. Thu should be dooo immediately after mixing. seal tho viol os soon o pooslblv D*t*rm;ut sample night ucumi to 0.001 grunts Totol somplv night mutt not exceed 40 groins condition *.n* viol for on* hour at tO*C in th* onolyior. Determine th* TB eft tho slurry sample (Section '1441. 74.4 Inprocctt vuU*ittr Mftiple*-- Utlng o torod viol (including reptum ohd Itimlnum cop) quickly odd approximately 1 oc of water using o medicine dropper Beal th* viol os too i os postibl* DrurnlM somplv night accurate to 0 001 gram con* dltlon th* viol for two hour* ot tO'C in th* onoiyur 74 Anolytlt. 74.1 Preparation of gut chromatograph-- Instoil th* cbromotogruphlc column end con* dltlon overnight ot ]&o*C. Do not connect th* tit end of th* column to th* detector while conditioning. 74.J.1 Plow rut* adjustment* Adjust how ntn ii follows a. nitrogen wrier gao- get regulator oh cylinder to r*sd #0 pug. Art rugulstor on chromatograph to 14 kg'em*. Hcrmal flow* ot this pressure should b* S3 to 40 ct/mlout*. Cheek with bubble flow meter b. Burner tir euppiy---Set reguiotor on cyl inder to reed SO pslg Bet regulator on OhmmsrnfTsph to cupply sir to burner st o ruts between JM end 900 ce/mlnute. Check with bubble ftowm*ter c. Hydrogen suoply--Bet roguiotor on eyi- teder to md SO pslg. Bet regulator on chromotogruph to supply approximately |l^l oc.'mlnute. Opttmtr* hydrogen flow to yield th* most eentitiv* detector rsiponee without extinguishing the flam* Check Bow with bubble meter uad record thi* floe 74.14 Temperature adjustments 8*t temperatures ee follows: a. Oven (chromatographic column), 10* c. b. Deane line. i*o* c. c. Injection block. 140* C d. Bampie chamber, water temperature. 00* C; 14* C. 74.1 j ignition of flame lootration iofec- tor--Iffuit* the detector according to the monufoctvrer's tnetructione 74.14 AmpUfler balance Balance the amplifier according to the manufacturer's metraetlons. 744 Programming the ehiomsiogysph Program th* ehromatofrapb as follows' a. 1--Posing time--Tb* normal setting Is fl mwoad*. b. A--Aftalyd* lime The sirmil setting w g minutes. Certain types of samples eentaia bigb bolting meteeisis which caa new intertsfenes wtih the ebsyl tbierifis peak an sabHUUsat analyme. In ish caeas the analysis time must be adjusted to eliminate the interference. An automated bscfcflush system can also be used to Hlv* tbls preb* lam. C. * --77m normal setting is 04 minute*. d. V--gtabthsaticu note. Th* normal set ting is 04 minutm a. X--Humber of analyse* par sample--1Th* normal setting Is 1. 744 Preparation of sample turntable--B*- fer* placing any sample into turntable, be certain that th* center section of th* alu minum cap has been removed The numbered simple bottle* should b* placed in th* cor responding numbered positions in the turn table. Insert sample* in the following order: Positions 1 k 3--Old 3000 ppm standards for conditioning. These are accessary only after th* analyser has not been used for 34 hours or longer. position 9--60 ppm Standard, freshly pre- porod. Position 4--eoo ppm standard, fmhly pre pared. position I 3000 ppm standard, freshly prepared Position g--4000 ppm standard, freshly pre pared. Position 7 Bsmpls Ho. 7 (This ts the first sample ftf the day, but is given as 7 to be con sistent with the turntable and th* integrator printout.) A/Ur all sample* have been positioned, in sert th* second sat of #0. #00, 3000. and 4000 ppm standards, flamptes, including stand- ardi must be conditioned in the bath of O' C for 1 hour (hot to exceed 4 hours). 74.4 Start chromatograph program-- When all samples, including standards, bav* been conditioned at *0* C for l hour, start th* analysis progrsm according to the manu facturers' instruction* These instruction* must be carefully followed when starting and stopping program to prevent damage to the dbaing assembly. 744 Determination of total Solid* (TB). Per wot cake, slurry, rssin solution, and PTC latex samples, determine TB for eich mmpl* by accurately weighing approxim- auiy 3 to 4 grains of sample in an aluminum pan before and after placing in a drift oven (to# to 110* C). Bampie* must be dried to constant weight- After first weighing re turn the pan to the oven for a short pe riod of time and then rsweigh to verify am- pieu dryre**. TB is then calculated as the a*i sample weight divided by initial sam ple weight. A Calibration. Calibration is to be performed each eight- hour period when the instrument is used Each day. prltf to running samples, th* col umn should b# conditioned by running two of th* previous days 3000 ppm standards. i.I preparation of Standards. Calibration standard* are prepared by fill ing tho vial* with the vinyl chlortde/nttro- g*n standards, rapidly Mating the septust and eeailng with the aluminum cap. Vh a stainless steel lift* from th* cylinder to the vial. Do not use rubber or tygon tubing. Th* ample line from th* cylinder must be purged (Into hordi for several minute* prior to filling via* After purging, reduce th fio rate to sopraumaieiy MO-IOOO ee/trur. piacs end of tubing into vial (near bouami and after one minute slowly remove tubing Place Mptum in vial a* sota as pcstible to minimice mixing ajr with tar.ole After th mne* aid vial* ar* Haled, inject 100^1 of distilled water. 4 Preparation of chromatograph calibra tion curve Prepare two to ppm, two too ppm. two 3000 ppm, and two 4000 ppm standard samples Run the calibration samples in txattiv the earn* manner a* regular samples, plat a., th* integrator area counts for cs:h standard sample n C,, the concentration of vloyl chloride in each standard sample. Draw a Un* of best fit through the points. fi. Calculation*. J Response factor Prom th* calibration cure* desmoed in Boction g4, above, eelect the valut of C> that correspond* to A, for each simple Com pute th* response factor. Bt. for *cn sample, as follows; Equation 107-1 4 Rmldual vinyl chloride monomer con centration, or vinyl chloride monomer con centration. Calculate as follows: C.,,- A,P, */T, where: Equation 107-2 C,,, Concentration of vinyl chloride P. In the <*mple. in ppm. Laboratory atnicwphrre pres 7Y sure, mm Hf. Boom tempernturf, *K. A/, Mo1l0e3c.3u)l.ar weight of VCM V fcVolum* of vapor phase (vtal volume less sample volume i. m i b Weight of sample, gram*. AbOsi constant ju.340 icc-mm-mol- degtfH Kelvin) | XsHenry* Law constant, por VCM In PVC at 40 C. A=43 X 10'*= X. Por VCM in I cc (approximate l wastewater sample s: PO'C. X-SO > 10`*3X P# Equilibration temperature *1C. If the following conditions ar* met. Equa tion 107-3 can b* simplified as follows* 1. T ,= 22 C (394 Kf 3. r<=*0 <343 Ki 9. P = 7M nun Mg. 4. v,*v,-: when V ,aVUI volume, cc (3341 |. Bampie less than 04 percent wfiteg. C-'T, ') Eqvuii*. 10T-3 Tho foUowtng gcnoral equation eon be need for any sample which contain* VCM, PVC and r*sr. c._-^ X [ts7'+x.( rs )T,-rKM-T3) r.] Equation 107-4 A-18 GENC 01671 r*=*wi bh^- Mon: t . mmx b* dotmiaod far mb|>I vtttt * npor mum* us liquid rauo to l. ratio eu a* ob* toad by wijurtf** tb* mpu waifbt tbrauqh ftviBt onnfidwatioa w th* tntai o4l4a and d*n*ity of m FVC. jbMdtfti oaiouiotad mdnf Squats itn-4 ippi Mm~ omumiwttqu buw on t totaJ mapI Tp obtain rwum boaod on dry rvc oontoat, dl*1d* by 73. f^jf %, i-oc *v<ncr oaapl* (tint E. Bl to i ompor folum* to liquid tIuo mUo). I*IW* 10-*. Thu*. equation 1014 04o iM ropUdod to tn* following * <1 i. iu*nd>tt(a) C*M*" 7T ftb *an kL *et."ai CJAC. 04^ v Ittfo-T **d J Equation 107-3 l* 10. lilttVOM. 4 I4ddui Vinyl CUortd KowDtf Oon* Mat of polyrlnyi Chlottdt H*'& and <Ml Samplw. E. F. Qoodrtcn Cbmic*J Co standard Twt Froeadura Ho 100A-T. B. H- I Ooodrttb Taeanical OOBtor. Aton Lada, Ohio. January jo, illl, a. |ni. A- E "Tfc* BalubUlty f Vinyl Cmirulrt* in Folyrtnyl CBlartdo," ACB-Dtri* ton af F0ly#*r Cd*UCTT. Hoiyum ff* pnnta JJ (3): m,int d. Parana, A. Em * DlffualOft of Vinyl CElorldt ta folplnyl C&lond." AC*-d*l- *ra of Py*r Cb#mEtt7, Hoiymor rr*Mtnt* 11 (It: W, 1074. . Baron*. A. E. L. . Ctldar, C. /. Ttena- p** and J. U. Wbitnoy. Anniyna for vinyl CJUorid* is FVC PodT* by Eoad-ApaOo 0a* CEraasotofrapby.- to bo pubUabad ttf<Eac 114 af Ua <41 TJAJC, 1414U, Air A* l V l I A-ia GENC 014472 L APPENDIX B REGIONAL EPA AND INDUSTRIAL CONTACTS 6EWC Table B-l. REGIONAL EPA AND INDUSTRIAL CONTACTS RegIon/Company/Division EPA/DSSE Region I (Boston) Region II (New York) Region III (Philadelphia) Date of meeting 6-19-80 8-13-80 8-12-80 9-16-80 Place Washington,DC Boston New York Philadelphia (continued) Contact Purpose of meeting Rich Biondi Cathy McNair Michael Pucci Jean Thompson Obtain input from 0SSE personnel as to areas of concern to be incor porated into the VC Review Study Discuss the VC Review Study and obtain input from the Region to the review study's areas of concern. Discuss the VC Review Study and obtain input from the Region to the review study's areas of concern. Discuss the VC Review Study and obtain input from the Region to the review study's areas of concern. GENC 016474 Table B-l Continued Region/Company/Divisi on Date of meeting Place Contact Purpose of meeting Region IV (Atlanta) 8-3-80 Atlanta Wayne Aronson Discuss the VC Review Study and obtain input from the Region to the review study's areas of concern. Region V (Chicago) fC\I>D Region VI (Dallas) 8-19-80 Chicago 7-23-80 10-27-80 Dallas Bruce Varner Discuss the VC Review Study and obtain input from the Region to the review study's areas of concern. Martin Brittain Discuss the VC Review Study and obtain input from the Region to the review study's areas of concern. South Coast Air Quality Monitoring Division (SCAQMD) 10-28-80 El Monte, CA Doug Newton (Region IX has designated authority to SCAQMD) Discuss VC Review Study; discuss SCAQMD1s Rule 1005.1 which supercedes the VC NESHAP. Society of the Plastics Industry 7-31-80 EPA, Durham, NC Robert Laundrie Discuss the scope of the VC Review Study. <9910 0N39 cn (continued) p m mm mm tern trs$ r~* t ' Region/Company/Divisi on Conoco Chemical Company Conoco Chemical Company Diamond Shamrock Corp. Table B-l. Continued ----------------- Date of meeting Place Contact -' Purpose of meeting 8-7-80 10-17-80 8-6-80 Lake Charles, LA Joseph Ledvina TRW, RTP Joseph Ledvina Deer Park, TX Alex Evins Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss containment devices. Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. (continued) 6ENC 0166 'J I Table B-l. Continued Region/Company/Division Date of meeting Diamond Shamrock Corporation 8-6-80 Dow Chemical Company 8-5-80 Dow Chemical Company 9-11-80 General Tire & Rubber Co. 9-10-80 Place Contact Purpose of meeting Independence,TX Oyster Creek, TX Midland, MI Astabula, OH (continued) Alex Evins Become familiar with new and existing air pollution control techniques currently being used to control VC emissions: discuss plant processes. Robert Oubre Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. Robert Ammons Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. Robert Laundrie Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. GENC 01667? sa cea t-- r*-? *--i mn Table B-l. Continued Region/Company/Division B. F. Goodrich Chemical Div. Date of meeting 8-20-80 B. F. Goodrich Chemical Div. 9-17-80 B. F. Goodrich Chemical Div. 10-30-80 Great American Chemical Corp. 8-13-80 Place Henry, ILL. Pedricktown, NJ Cleveland, OH Fitchburg, MASS (continued) Contact Purpose of meeting W. C. Holbrook Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. W. C. Holbrook Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. W. C. Holbrook Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss SCAQMD Rule 1005.1. Russ Mercier Become familiar with new and existing air pollution control techniques currently being usd to control VC emissions; discuss plant processes. I t Region/Company/Division Hooker Chemical Co. Shintech, Inc. Table B-l. Concluded Date of meeting 9-15-80 Place Burlington, NO 8-5-80 Freeport, TX Contact Purpose of meeting Harold Dubec Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. John Yonge Become familiar with new and existing air pollution control techniques currently being used to control VC emissions; discuss plant processes. 3 -6 GENC 016679 r-- bees r"" r1 J ' sm Ssjtl fcxtt APPENDIX C CURRENT INDUSTRIAL SOURCES GENC 016681 OPERATING ETHYLENE DICHLORIDE/VINYL CHLORIDE PLANTS Region IV VI VI VI VI VI VI VI VI VI VI VI VI VI VI VI IX Plant B. F. Goodrich Borden Chemical Conoco Chemical Diamond Shamrock Diamond Shamrock Dow Chemical Dow Chemical Dow Chemical Ethyl Corporation Georgia Pacific ICI Americas Monochem, Inc. PPG Industries Shell Chemical Shell Chemical Vulcan Materials Stauffer Chemicals Location Calvert City, Kentucky Geismar, Louisiana Westlake, Louisiana Deer Park, Texas* LaPorte, Texas* Plaquemine, Louisiana* Oyster Creek, Texas Freeport, Texas Baton Rouge, Louisiana Plaquemine, Louisiana* Baton Rouge, Louisiana Geismar, Louisiana Lake Charles, Louisiana Norco, Louisiana Deer Park, Texas Geismar, Louisiana Long Beach, California Began operation since promulgation of the regulation (October, 1976). Process EOC/VC EDC/VC EDC/VC EDC/VC EDC/VC EDC/VC (2) EDC/VC EDC/VC EDC/VC EDC/VC EDC/VC EDC/VC EDC/VC EDC/VC EDC/VC EDC only EDC/VC I 1 OPERATING POLYVINYL CHLORIDE PLANTS Region Plant Location Polymerization process I Borden, Inc. Leominster, Mass. I Great American Fitchburg, Mass. I International Materials New Bedford, Mass. 11 B. F. Goodrich Pedricktown, N.J. Suspension, latex Suspension wC Suspension Suspension, dispersion, bulk II Goodyear Tire & Rubber Niagara Falls, N.Y. Suspension, dispersion II Hooker Chemical Burlington, N. J. Bulk II Pantasote Passaic, N. J. Suspension II Tenneco Burlington, N. J. Suspension, dispersion II Tenneco Flemington, N. J. Suspension II Tenneco Piscataway, N. J. Suspension, dispersion* II Union Carbide Sommerset, N. J. Latex III Diamond Shamrock Delaware City, Del. Suspension, dispersion III Firestone Perryville, Md. Suspension, dispersion III Firestone Pottstown, Pa. Suspension, dispersion III Pantasote Point Pleasant, W. Va. Suspension III Stauffer Delaware City, Del. Suspension, dispersion IV Air Products Calvert City, Ky. Suspension, latex IV Air Products Pensacola, Fla. Suspension IV , B. F. Goodrich Louisville, Ky. Suspension, latex IV Conoco Aberdeen, Miss. Suspension IV Union Carbide Tucker, Georgia Latex Began operation since promulgation of the regulation (October, 1976). I GENC 016632 r~^ ar? r~~' r r-~> ra *--1 Jr--) easel ttrrr OPERATING POLYVINYL CHLORIDE PLANTS (Continued) Region Plant Location Polymerization process V B. F. Goodrich V B. F. Goodrich V Borden V Dow V General Tire VI B. F. Goodrich VI Certainteed VI Conoco VI Conoco VI Diamond Shamrock VI Ethyl VI Firestone VI Georgia Pacific VI Shintech VI Tenneco VI Union Carbide IX B. F. Goodrich IX Stauffer IX Union Carbide Henry, 111. Avon Lake, OH Illiopolis, 111. Midland, Mich. Ashtabula, OH Plaquemine, LA Sulphur, LA Ponca City, OK Oklahoma City, OK Deer Park, TX. Baton Rouge, LA Addis, LA Plaquemine, LA Freeport, TX Pasadena, TX Texas City, TX Long Beach, CA Long Beach, CA Torrance, CA Suspension, dispersion Suspension, dispersion, latex Suspension, dispersion Suspension, dispersion Suspension Bulk Bulk Suspension Suspension Suspension, dispersion Suspension, dispersion Suspension * Suspension Suspension Suspension Solution Suspension Suspension Latex * Began operation since promulgatiori of the regulation (October, 1976). 6ENC 016683