Document X7azVkba77gy7Mk3eQmVOE47y

EPA-450/3-82-003 Vinyl Chloride - A Review Of National Emission Standards Prepared by TRW, Inc. P.0. Box 13000 Research Triangle Park, North Carolina 27709 EPA Project Officer: Fred Porter Emission Standards and Engineering Division U.S. ENVIRONMENTAL PROTECTION AGENCY Office of Air, Noise and Radiation Office of Air Quality Planning and Standards Research Triangle Park, North Carolina 27711 February 1982 SPl-05084 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. ii SPI-05085 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. SPI-05086 ACKNOWLEDGEMENTS This study 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. 1v SPl-05087 ABBREVIATIONS AND ACRONYMS USED IN THIS REPORT AAQS Ambient Air Quality Standard APRS Automatic Pressure Reduction System BACT Best Available Control Technology BAT Best Available Technology BID Background Information Document CAA Clean Air Act CARB California Air Resources Board CFR Code of Federal Regulations CMA Chemical Manufacturers Association CTA Chain Transfer Agent CTG Control Techniques Guidelines DOT Department of Transportation DSSE 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 SPI-05088 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 mill ion PSD Prevention of significant deterioration PVC Polyvinyl chloride RACT Reasonably Available Control Technology RCRA Resource Conservation and Recovery Act ROL Reactor opening loss RVC Residual vinyl chloride RVD 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 VI St*50** 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 ............................................ 1.6 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 Emitting Industry ............................... 2-9 2.3.1 Current Number and Geographical Distribution. . 2-9 2.3.2 Influence of the Standard on Industry................. 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 of Ethylene ........................... 3-5 1-4 vi i SPI-05090 Section Page 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 3.4 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......................... 3-37 3.5 References for Chapter 3................................. * . . . . 3-39 4.0 CONTROL TECHNIQUES USED TO COMPLY WITH THE EXISTING EMISSION STANDARD .......................................................................... 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 fromReactors ................. 4-23 4.2.3.1 Process Variations .................................. 4-23 4.2.3.2 Causes of Reactor Discharges ............... 4-25 vi ii SPI-05091 Section 4.3 4.4 4.5 4.6 Page 4.2.3.3 Prevention of ReactorDischarges . . . 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 Non-Reactor Relief Valve Discharges ............... . 4-48 Resin Stripping.........................................................................4-52 4.3.1 Introduction................................................................. 4-52 4.3.2 Suspension Resin Stripping....................................... 4-54 4.3.3 Emulsion Resin Stripping........................................... 4-57 4.3.4 Bulk (Mass) Resin Stripping ................................. 4-58 4.3.5 Solution Resin Stripping........................................... 4-58 4.3.6 Other Stripping Technologies....................................4-59 Fugitive Emissions .............................................................. 4-59 4.4.1 Introduction................................................................. 4-59 4.4.2 Equipment Specifications........................................... 4-62 4.4.3 Operational Procedures...............................................4-66 4.4.4 Leak Detection and Elimination Programs .... 4-67 4.4.5 Inprocess Wastewater.................................................. 4-80 Reactor Opening Loss.......................................................... 4-81 4.5.1 Introduction................................................................. 4-81 4.5.2 Solvent Cleaning..........................................................4-82 4.5.3 Steam Piston................................................................. 4-84 4.5.4 Water Piston................................................................. 4-85 4.5.5 Reactor Purge Air Blower........................................... 4-85 4.5.6 Steam Purge.................................................................4-85 4.5.7 Redox Catalysis..........................................................4-86 4.5.8 Water Jet Cleaning......................................................4-86 4.5.9 Clean Reactor (Closed Cleaning) Technology. . . 4-86 4.5.10 Nitrogen Purge............................................................. 4-87 4.5.11 Slurry Backfill ....................................................... 4-87 4.5.12 Calculated Emissions.................................................. 4-88 References for Chapter 4....................................................... 4-90 ix SPI-05092 LIST OF TABLES Table Page 2-1 Emission Standards in the VC NESHAP........................................ 2-3 2-2 Summary of Reporting and Recordkeeping Requirements in the VC NESHAP.......................................................................... 2-6 2-3 Geographic Distribution of Operating VC-Emitting Plants. . 2-10 3-1 Point Source Emissions - "Balanced Process" EDC/VC Plants............................................................................................ 3-3 3-2 Point Source Emissions Typical of Suspension and Dispersion PVC Plants......................................................................3-13 3-3 Point Source Emissions Typical of Bulk PVC Plants.............. 3-25 3-4 Point Source Emissions Typical of Solution Process PVC Plants............................................ 3-30 4-1 Point Source Emissions and Technologies for Control in Typical Suspension, Dispersion, and Bulk PVC Plants. . . . 4-2 4-2 Point Source Emissions and Technologies for Control in "Balanced Process" EDC/VC Plants ............................................. 4-4 4-3 Emissions Reduction for 316 M kg/yr EDC/VC Facility in Compliance with Current Regulation ......................................... 4-5 4-4 Emissions Reduction for 68 M kg/yr PVC Facility in Compliance with Current Regulation ......................................... 4-6 4-5 Total Relief Valve Discharges for 32 Regulated Sources From 1977 to 1980 4-20 4-6 Relief Valve Discharges from PVC Plants................................... 4-21 4-7 Relief Valve Discharges from EDC/VC Plants ......................... 4-22 4-8 Estimated Costs for an Auxiliary Venting System.....................4-32 4-9 Typical Gasholder Specifications for 38,000 Liter (10,000 Gallon) Reactor................................................................. 4-35 4-10 Estimated Cost for Installation of a Gasholder.....................4-36 4-11 Percent Distribution of Stripping Levels Being Achieved by Industry.................................................................................... 4-55 4-12 Approved or Conditionally Approved Equipment Equivalency Determinations .............................................................................. 4-64 4-13 Leak Detection and Elimination Programs....................................4-70 4-14 Variability in Leak Definitions.................................................. 4-72 4-15 Calibration Results for Area-Wide Monitor................................4-75 4-16 Reactor Opening Loss Reported by Representative Companies. 4-83 xi i SPI-05093 LIST OF FIGURES Figure Page 3-1 EDC/VC "Balanced Process" Flow Diagram ............................................ 3-2 3-2 Suspension and Dispersion Process Flow Diagram ............................. 3-12 3-3 PVC Resins, PVC Compounds, and PVC Fabrication Processes .... 3-15 3-4 Bulk Process Flow Diagram.........................................................................3-24 3-5 Solution Process Flow Oiagram................................................... 3-29 xi i' SPl-05094 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 industry 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 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. e Emission sources not identified during the original support study. e Enforcement and compliance experience. Additional details can be found in corresponding sections 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) produced by oxychlorination, VC, and polyvinyl chloride (PVC) facilities. Recent modifications of processes in these facilities include: SPI-05095  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 oxychlorination plants as part of the EDC process. The change from air to pure oxygen as a feedstock can make combustion more feasible and therefore could result 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: e 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 SPI-05096 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 or 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 carbon adsorption as a primary control for exhaust gases have replaced the carbon beds totally with another 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) 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. Containment devices are used to reduce emissions to the atomosphere and include gasholders and pressurized holding vessels. These devices collect vapors from various equipment 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 SPI-05097 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) a 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) e 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 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. e Suspension Resins - These resins represent the highest percentage of total PVC production. The most widely used stripping method is continuous steam stripping. Many processors are attaining levels much lower than the required 400 ppm RVC - some less than 20 ppm. (Section 4.3.2) gpi-05098  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. 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. 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 as estimated in the original standard support study. Those 1-5 SPI-05099 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 proce dures. Some plants have received approval for equivalent equipment. (Sections 4.4.2 and 4.4.3) 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 primarily on operating preferences and economics. 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 x*C SPI-05100 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 with 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. Other 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) SPI-05101 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 Administra tion (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. Waivers of compliance were granted, in some cases for up to 2 years, allowing industry to incorporate necessary controls. The existing VC NESHAP (henceforth referred to as the regulation) is one of the most complex air emissions standards promulgated by the EPA. 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 SPI-05102 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 SPI-05103 s. g. Q. Q. o-- OO O V| o' Vi v| Li/t 0x)9 x 2L O O 9 c9<o-- s9 _9 i 9 M=* -- U 5. 9 x a. e o -- v| Tcontlnuedy 01 <0 e o>S a + u or 9 3 a. I 01 & 9 o o 3 *4 4-> O IA ^ 9 /* A 3 99W e 3& 9-- 33 X O* UJ 9 >o w w s (/ -- ew oo 4/1 3 i/l o -- > Ex UJ o i =<o f2 9-- W *9-- VT I 1 4-* -- -- 9 1/> m 99 9 <* (A O3 9 <M 9e w ft OE 3 5. o-- 4/1 3 --a O9 X 9 w 4oJ o <0 9 w w > a O 0) 2u wi -- .2 l''S >9 of 03 e <*JL S W3 3g I -- C Vi O9 Ss --1t/! "0=) -- 43 & <3 9k>>U <9 0. 3 9 1/1 e OO 9 w 9 9 i/ 3 3W Q. O. 9 -e I 13 I C o 349J ua e o-- 3>g 4- ^0) i/i Oi 3 &. *-- 3W xw UJ i/ 1/1 -- 9 i/ 9CW --o <0 9 0 en) w 41S-/>1 C9O99 3O x--9U x-- w u Ea i/i (4 9> 9O (4 u <0 9 9w I W Li .9 fi .:a. WWW C /l3 4- > I/iI 1g/1 9c--<--<3o 09u W<0 J09--W0. *-- 4/1 9--99 .3 O W C 4- U- <- w. a. O V) o & W9 -- o*w 9 W W */ -- 099 3 C 4/1 W i/l 9 o Sr ---- w w 3---- 0 9 33 9*J -- 4- 9 -- > O w % -- C 99 >039 -- +> e <l w -- O O4 C 4- 4-> 4/1 -- 4- 1/1 ie 5 8.5 1/1 4/10 0 -- IS) W 9 --9W> E3 9 9 lj *-> a. -- to vC 2-3 SP/-05104 2-4 SPI-05105 T able 2 -1 . Concluded 2-5 SPI-05106 T a b le 2 - 2 . SUMMARY OF REPORTING AND RECORDKEEPING REQUIREMENTS IN VC NESHAP (co n tin u e d ) 2-6 SP/-05107 T able 2 -2 . C ontinued 2-7 SPI-05108 T able 2-2. C oncluded se'*5109 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. 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. 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 39 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 SPI-05110 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 Georgia Kentucky Mississippi V Illinois Michigan Ohio VI Louisiana Oklahoma Texas IX California TOTAL PLANTS No. of PVC plants 3 6 1 2 1 1 1 1 1 2 1 2 1 2 5 2 -4 _3 39 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 SPl-siii 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 promulgation of the regulation. Of the 41 original PVC plants, 4 plants have discontinued operation and 3 new plants have begun operation. 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 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 SP1-05112 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 results in a lower volume emission to be combusted. 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 SPI-05113 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'1 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). 2-13 SPI-05114 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 SPl-05115 3.0 PROCESS DESCRIPTION 3.1 INTRODUCTION 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. SPI-05116 3-2 SPI-05117 Figure 3-1. EDC/VC "Balanced Process" flow diagram. Table 3-1. POINT SOURCE EMISSIONS "BALANCED PROCESS" EDC/VC PLANTS discharged to the atmosphere From storage tanks must not exceed 10 ppm. 11 ; i 1 1m i *-> c OJ S p 3 <T o> p c o 4J rc 3 d Qj c X Qj P_ 3 5O' o. 01 3*. O o> Pv <0 1 to o 4P 4P c 1- X o QJ D-f-- X o> c o +-> O P fO 4P to 3 to O du QJ p a3 oi XI 4P oo z < 4P = G P '-0 CO 0- 3 *-- ro 5X 1- o x a-p D +-> to >,c J lA -p -- to*-- O Cl to O > C. OJ P U 4-> tL- o oc o i- o Clu c o o QJ ro *f QJ 4P U 5 O <0 X op i. QJ pw <4- X C +> OJ 0) O op u 0) u c >30 to O X qj QJ --O a) oc -Q Wj3 (J to O o X ro > 4-> X cn 4-J QJ u o sQJ CL 4P cc O --i QJ QJ ro . 4-1 u o 4-J c QJ ro Ci QJ lx: ^ O E> C J- 4- 4P CM U to 3 c o -o oo 1- > itO QJ Cl tO QJ *- o o XQ UJ Qj UJ to QJ to ro 4-> to 3 ro X QJ E a *-- c. -- ro O f--i C r- X QJ QJ OU P X 4U QJ ro S- 4P 4-> O C QJ U 4- tO O3 OS QJ QJ o to ^C --1 . QJ F- 4-> QJ 00 M F- C O v P ro tO 3> a? d 3 <4- to p* EOO *-- C a. 4-> X 4- to O CL 4U QJ C to h- X -- to P C) --co QJ _ f-- Q) .F- 4-5 QJ p- CL to rc O o o to 0. C P" U OP ^ E > 4P X O QJ O ro Cl P tO QJ (O w'; C3 = -O 4P X <o X QJ a; X 4P *P* oo 4P 2> O 3 QJ O QJ *-- S- C QJ 4P P- +J Qj m f- ^ E ro . c ^ O -C Q.C0 i- 4-5 on co E to H- O 0 0 3 O E O L- E to O 3 ro XO X *p QJ 4P (0 to ^ * P- C ro O ro OX X *p- p- QJ QJ X QJ L. to qj x; QJ > QJ to 3 4-5 QJ OP *F-- -O O E **- E QJ O X QJ S- LU i- 4P QJ i- QJ ro CP O 4P ac 3 QJ V C Cl 3 CLQJI to X to to cx O' C X c c CO qj O' E 4-> 4-> to 3 C X 4-> X >> -X i-- cn-c c > co o P OIQI ro ro U Cl 'P 01 4P r-- > x P C > o p o> * to P 01 d *- a> p QJ QJ Cl o. (0 & to to ro 0) X 0) to c p to to p op 'e 0) o c * 4P 4-5 X 4P 0) J* 0) -X to to qj c lO'D C T3 C o <0 C ro pp ro QJ 0) o 4P O 4-> X QJ QJ QJ QJ 4_) 4-> 4-> 3 QJ a; ro ro 4P o &. C C CD S3 c 3 qj E <o 5 <o QJ 0J QJ X 3 P 3p +J 4-5 4-5 O O f-- O I-- o to to o a. P O o 4-> O 4- ro ro O. UJ QJ tO U to 3 3 4U QJ > to to QJ U O s CL s 3 O -X O U *J (O <0 --> jr c tO *f- irO O 5 O 4- -- ro W Z SQJ U ro 4-> >> O. ro X QJ 3000 O to EC 3P OJ r-- C C O QJ <J > o (J JC c <J E c 3E QJ *-- 3 3 O-- O' u o u o -- QUO U> to 10 ro QJ 4U M- QJ p- ro d sco 4- X to ro Oo ^> e o Q. 0) c W rO o to c F- <S) X 4-> /O to o u Qj r U X 0- o QJ p PP Q. 4-> 3 QJ CL oi pO F-- a Q UJ CM t P QJ QJ d 4-5 ro ro L. 3 QJ C 4U 4-> to to ro 2 r to to 01 p QJ QJ o a. p a. C o p 4U ro W o -C XJ <0 cn o p* *P- 4-> ^ ro U (J ro *f- tK X c ro d C p X ro O Q..F- V u - P i-- 4-> s to >> L. X O3 O o > Cl > m <0* on QO 3-3 SPI-05118 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 <C12). Direct chlorination may be simply expressed as ch2ch2 + ci2 cata^^ 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 + h02 ---- Cata1yS--- CH2C1CH2C1 + H20. 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 SPI-05119 requirements in the regulation of EDC production (see Table 2-1). The direct chlorination step 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 Typically, 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-t; SPI-05120 chemical feedstocks, chlorine and ethylene, without coproduct formation (McPherson, 1979). The process can taice 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 (CQg). 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 SPI-05121 The oxychlorination process vent is subject to the current regulation. Emissions from oxychlorination reactors must not exceed 0.2 grams cf 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, ethyl 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 Dichloride 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 SPI-05122 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 tne equipment is out of service or opened. Prior to opening any equipment, Section 61.65(d)(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 2-8 SPI-05123 are usually removed by chlorination prior to distillation (McDnerscn. 1979, p. 79). A method for formation of HC1 from the light ana heavy ends distillation column byproducts has been developed. The method employs catalytic oxidation of the byproducts separated by the purification columns with air and otner 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 SPI-05124 (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 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 within an EDC/VC l-l 0 SPI-05125 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 SPI-05126 ? & mk 5 i w T8 rt 3 iNk k > & If k td sS' t Z5 i I O- gk f ot s r @@@@@@ Recovered VC Figure 3-2. Suspension and dispersion processes flow diagram. 3*12 SPI-05127 Concentration of VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Stripping vessel vent Table 3-2. POINT SOURCE EMISSIONS TYPICAL OF SUSPENSION AND DISPERSION PVC PLANTS a; 01 Xc p 4-) 4-> r-- o V) c 3 4. E oo ct V) V) CD a. 4-5 0) c VI hC c p C LO 0> p- c o s 01 -p- 4-> oi CL v> re i- cn O V) c r- LJ 3 C E> O' -o o 0> 0) re CL 4- s- o 3 o> o J- c 4-4 arn o E= re x e O 4= CL CO 4-5 L. 4-5 VI CO re 4- QO p" o E O 3 V) V) 4-> CT) c re O 0) o -o os r-- 01 01 *CT VI L> x oi V) 3 4-> o> r- "3 u E 0) O x LU i- 4-5 01 o V) *--1 01 p- L. E CL o. cr X V 01 OO 4J 41 S- ^ o CL VI <4- 4-> +4 V) 01 "CT 01 O <A 01 0) 0) -o E ^ i- oi E a> 4-4 CT> a? u p" re xx o <-- 4-> CL 0) 4-> o 01 VI V) s- X 3 2 **- V) 4-> 4-> re E O oi c x lj e c o O X re r- U 44 01 4-5 r-- 0) VI 01 4-5 4- X 3 E JO V) O 4- o3 L. 4-5 re e O tA 4- V) r- Q 4-> 3 x *- C "CT oi 44 re? re oi > re c re u o> 45 cn lr o -> cl e sre 01 cn <D oi 3 01 x re L- 4-> L> u s- E 3 V) O LJ i- CL O 4-> > <j CL LJ ^ V) 4-J o c V 4-1 01 VI V3 rcen E 0) 4-> L- Vl 01 3 -C re cl V xO 01 E 4J o re > 01 4- .C O M coE O 4-> CL p CL 4-5 re? re s~ ocni o ^4 4-> U c re ^ Ol -C 01 u u 01 C V) u o r- X o re? 01 "CT 01 01 D X 01 O 4-J 4-> o c V) VI 4-> P- u 3 s- re? oo M &- U CL re oi o L. > CL U CT) re -x 01 ^ o E 5> O s- cn 4- CS1 --1 O o cc o Cl a> u X OI oi cn Oi re U 4C 01 u -C V) CO-fV *CT O E 44-4 oi re p* p 01 01 X L. 4-> ou 4-1 C 01 01 CT> CT) &s- 01 re E X oi o iA C re "CT oo Z 4- 4-5 o c 01 V 4-5 X 3 V) V> 3 O re E 44 CT) 0) "CT 4-5 U 01 V) 3 i3X re a- 3 X Vi CT" X O 01 01 E J- 4-5 lj re 4-4 >O 01 z 4- X O 4-5 COE 0+5 0. *P- CL 4-5 -CT re oi o S- CT)pH 4-5 U c re -3 01 X 01 u a oi C V) (J o *-- x LJ *CT 01 .X C re 4-^ +-> ac?n c re w oo CL 4-> V) c o LJ > VI V) * p* V) E oi oi c p- re cn 4-5 c e reoi "O 4-4 oo CL cn c -o o -C re? c re cn c f-- OI 01 V) S 4J c * 01 o> > c p ^ Xc p* re S -*-> CT) c r" c 01 CL O So 4-) U re 01 u c o 4-> re -- N1 _J p- o L. QC fl) >> (A p- V) oo CL p- 4- 01 p-- 01 S- io 4-5 u re 01 U V) 01 C CT) O ip* re 44 X re u N V) -f* L. "3 a I.SJ p* f- o re a. > 01 s3 44 CL 3 So 4-> LI re o> w V) C 01 O CT) -p- L44 re re x Nu p* V) L. 0) "CT ^ V) o -pCl -CT 01 cn re o +> (A a. a; ct 4-> c v> re </) cn tn c 01 U "CT O re S- o Q. c 3 LJ > re? oi c s- c re o o 3) 01 4-> C XI re JZ V) 01 ai^ a. f- c o oi re S CT) C cn p" CT) c CL C *" CL r- -CT r X ^ Sp- o 4-1 Z -C V) c o p 4-5 re N r-- S01 !> O Q. CSJ ro Stripping *J SPI-05128 I oc in Ol in V5 3 I cn 4-> WOi in 01 3 -C Cl C in x01 4Q=-5 Regulation requirements > 01 O* OC-OM ED. 4*-4- "O o. #0 01 o i- 05^-4 4-1 j0C> -C *00) UemO u01 O f- X o *o 0) *o 40*14 to 3 00l-15 4-) o z O 4-1 t*n- L. t3n rO S01- E01 -xC 4-) 4-4 oi *0 tn 5 >> tn c-- in E s- m -- o G- 01 (U o> a uO 4i-- <4- o S- c o -* Cl a. T" ao U CD *> tn * cr- fO 4o*-> 0OU1l 4t E c to x 0J 4-> -o 40->1 to o 4O-5l r<--0 4O&-- 44 -O 4LC-5 01 4-5 XI 40-15 u01 oC 01 u c CQ O 01 3 3 > 3 o tn 05 O) o -O U 0) 0s14-5 L01. Eoi airxO 4-5 t- .a u tn tn Ol u i. zo zO LJ O -C fO > 4-5 5 05 3 O tn Potential emission points 4-> in 3 to .CX 4-> 01 o Q. 5E 4-5 4-> 3 </> o i>n) -X1 u s>- oc 01 4* > O 4-) uc 01 01 Q > -o c <0 in m .X 4-> cc 4<-0 >01 "Oc -cX i--01 4t-o> -O 05 >> C i3- "--O f(--S) zo tn 01 4-5 tn C 3 01 4-> O > C JZ c 01 05 c c > t. .IoO o 4*->- 3 ^ i-- co 01 * <9 >) tn &- (O u 4-5 i. O 01 L. -o Q. f- O * tn in 05 44 0) C e 05*^ 0O1) 0a). to o. 1- 4o-1 4&-5- 01 to c tn tn tn > L. 01 O O 01 r-- L J- 01 4-> u oi tn u tn 40-11 40-11 4s- - tn tn * 0015 to to 2i 2t 4i-.> .(xJ 4t-n> #s0 40-14 40-14 0C) 4<-0> c01 4o-4 i<nQ i<no u <n > </> C o ttnn E 01 > tn C o f-- in in E ai o > 4os->c o u 4o-4 0c5 Process step 05 c 44E0-0>1 i>/1> >w 01 05 oC> TC- 1Q-. -o o. 0c5*o- fur- Z 4-4 c 01 0L1. uc tn tn- 05 L. 00tio-15 t40t-o1 4o-5 in 402-14 tn "O to c > o u 0L1. x in --E' Cfl iO. )^ 4-1 C 05 054- o C *- -o Nf- &01tn oi 4-1 ns -- 05 c min 01 o s- t01 oE tC- !C-- to ^4-4 "O -C /> fT3 C 05-X S- <0 oi 5 ci- -031 *3 O 01 o S- Q. Cl Or-. c o 01 fO-J <ctJ o01. >> &- 4- CO 5 4-5 o O O o > ^c 4&-5- c_ = tn Q. Cl 4S-4in 05 c tn 3 tn ^ 4041 4-4 c pt--o 3 05 a. 0s_) o > 4-4 o. c c 1/5 iO 00 05 L- 0i-1 10 4t 3-14 SPI-05129 P olym erization Process R e s in Type Compound F a b ric a tio n Process 60ac0 C30 in j* y-V C3Q CO CO *2CrO. --cci c 0 LCO tHfi a; -i ft. 3 rr E --i faJ C^ 3-15 totlCO/oo> Ci- c acoci 4<u->z 3 00 c u 7c X4) o e> CL X>04ft-*1l *fccc "D(cO3/> w0) CO- 1M coU o>a. Ci) to C 4J r- oe Cto i- Tc3y o auu> OL. T3 (J 0X CO la b rlc a tlo n processes. m3 VU oc c00n) Vo3 <0 * SPI-05130 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 H Cl I Il R- + C = C --------------------- R - C - C- I) HH Ii 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 increases with increasing temperatures. This characteristic of the reaction kinetics allows a desired molecular weight to be obtained by 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. 2-16 SPI-05131 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 to 0.0038 cubic meters (0.13 cubic feet) or less at standard 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. 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 incorporating a 4-way valve in the pumping lines. Instrumentation of SPI-05132 the system usually incorporates a turbine meter, a flow totalizer to measure the VC flowrate and quantities unloaded, and storage tank level i ndicators. 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. Some of these 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 spi.05l33 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 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 is 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. SPI-05134 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 87C (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. The resin is then bagged or stored in silos for bulk shipment by trucks or rail car. 3-20 SPI-05135 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), t 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 (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 2-21 SPI-05136 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 poly merization. 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. 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 SPI-05137 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 SPI-05138 3-24 SPI-05139 R e c o v e re d VC Figure 3-4. Bulk process flow diagram. Table 3-3. POINT SOURCE EMISSIONS TYPICAL OF BULK PVC PLANTS (continued) 0) 0) -O C o to c 3 EOo "O to CL 3 c to H- c c - c 00 3 r- C o s *-- 3 0) Cl to ro U O to c O 3 f- c E > O' -3 o 3 O) O Cl u O 3 0) O u c Qpo c E to -C e o -C Q.C0 U to co iO O o gj 3 to to CD c to o 0) QC O "O *- 0) 3 T3 to u -c to 3 r- *3 E3O LU u 3 0) u X 3 cs CO --1 3 U a. s C. -C v CJ -C oo i- - o 4-4 4-4 a. 1/1 >*tO QJ 3 0) O to QJ 03 3 "O E 01 4E-> O10* W QJ QJ V <0 -C X o ^ 4L O. QJ o O) to to U 50 -C +4 3 fO Oe 4-4 C 0) c .c 43 c c o O X rO 1" u 4-4 Q4 4-4 QJ CO 0) 3 E to c 4-4 s o3 W 4-4 to C o to 4- to -C o 4-4 3 3E 3 > o Q * U X *f- c 03 4-4 lO 3 QJ io 4-4 cu O U 03 Q. C ro 3 0) U 4-4 U >u 3 03 .c ro _ U u i- E c to O Cl O 4-4 0.0 3 to 4-4 O c to 4-> 0) to tO 3 ro E 4-4 3U to QJ 3 .C ro Q.e to xO QJ 4E-4 ro > QJ .C C 4-4 co E O 4-4 a. Cl 4-> -O <Q 030 W ^ 4-> &. C ICO QJ 03 (J U QJ C to (J o f- X "O 04 QJ QJ U X 03 O o c to to f o 3 i- "o oo &u cl <o 04 i- > D. O ro ^ 03 ^ E> O i- CSJ a: QJ U X QJ 03 S_ QJ ro ^ a; (j -C to CL-ito -o O E 4 03 0 03 0) -C s. >5 ou c 03 03 &S- 03 ro E -C 03 U to C t-- <o i- zo o o c QJ to -Q 03 to to 3 O ro 5 03 -3 i- QJ tO QJ ^ 3 -C *i" ro O. 3 ^ CO X c QJ 03 E S- O ro >O QJ Z O COE O CL Q. *3 ro OJ fc. --* "3 03 C <00 Q3 03 fcU U 03 O C to u c. o x OJ (J -3 03 QJ QJ U X 3 O O c to to o 3 i- "3 OC So o. r0o4 (j) S* > CL -c u ro -X 3\ O E> o t- C\J cc c to to 3 4-) C *3 r-- U oO Q. to C oO > to CO *> to E 0) 3c fv ro c c 0) "3 fO Qo> o -J c -D o -C "3c ro C JZ F- 03 tO 2 4-4 c 03 > c Xc f-- <0 c c QJ a. o l. o u ro 03 U c o ro --^ Nl S*-. Q. 0E) ^ >*, to -- to OO a.-- 03 r 03 S- So u ro 03 L. (0 03 C O s- ro ro O N -ft--O t- -3 03 E 03 >> O ro a. > 03 s3 CL 3 S- to u ro 03 L- co C 03 O f- & ro ro -C wU f* to &- f03 *3 Su to O Q. *3 c c 3 CL o V o u ro 3 U C o to --' N f- W CL 3i '--' ^>i to to GO. FO- 3 <0 u o Q. 3 tO to tO CD to 01 Uo au. o roO 3 *3 3 cue ro O O f- 3 C -O ro f- U r to ^ aai. r- C O 3 ro 5 C ^ CO. C Q. -3 *- X -- u o Z -C to c o ro N " U E >Y Oa i 3 U CL CM m 3-25 o c ou - ro 3 ro U Nl *- C U i >Y C p-- *rO CL Q- CL 1 U to O to CL SPI-05140 Process step Potential emission points Regulation requirements Cl X cn re . <j . 00 .00 7 E 44-5 re *- 0) 6 . +4 >> O +-> c a; 0) 71 cn 6 JC a 00 c co 7 oo Z 4- 4-5 o c to 4-4 -O oo to 3 o 3 E 4-5 CT O *7 4-> L CJ to 3(q .aC s*- -C to O' x o S U fu >o 0) z 4- jz o +> COE 44 a. f- Cl 4-5 -3 re a>^ 7 4- a. E re 7 4-> .c ouo c to u o. C-r- x 01 UO (D L 4-4 o c to 4-4 W tO CO 3 <w = cn a; +4 i. to 3 re o. JC to X E 44 ; re > >4- . 4-> c E O 4- . o. 4-> 7 m HO a> 4-> C re 7 -C u C 00 U X 5 7 4c 7 4-5 <e 3 1 4-> z E 4-5 to to ^-4-3 re c: .C X 4' 4-5 re to J xc to to E to ^ o Cl > Ou +4<4-o E ^ a. C7 c re 4-5 5 O re x O 4-5 4- * 4-5 C 4-> o 7 4-5 c C > 3 to 07 0 EO -C 7 JC to u CX nj >4-550") 4t 7 4-5 re 3 cn 4-5 z0 4- 4o-5 u re 00 0) c a> o re 4-> . re U INI 10 *7 0) >) > O re Q. > 3 4-5 a. 3 o 4-4 U re to C O O) 44 re re -C N1 o **" to a !0) 7 5, -- to O *D- *7 4-> to 3 <0 -C X 4-5 G. E 4-5 4-4 3 to O > 1 to .X U >o c -x > O 4-5 UC Q > * 7 C <e to ce re o 4-5 4-5 7 re c to 4-5 Cl c J3 > C 1 *-- c to f- to c 3 O >, s- JC1 re o to . c E 3 .X f-- c0 re 4-5 cn Cl re 0. *0 4-5 4-5 to to 4- 4-5 re re 4-5 4-4 to to re re 33 to 0 (-- F* to cn re 0 +4 c/o E 5 4-5 5 fO to cn c 4-5 >> *- re 0 to f- a. $ 44 *7 a. to >> F--' 4-5 CO to 7 1- -C 44 re c > X to $ re *- 7 EC = 1 ^ re --* to 4-5 C to ) 14- O 7 c c re *- re F-- t- 4-5 O Q. E 7 -C to re c cn-x Cl -f- ^ re Cl c F-- +J 0 > z C 5+J O = to Cl in to CO 3-26 Cl Q. CO C 4-4 7 c 0 >0 -c to 3 0 to c 0 to to F-- E to 0 to to E 0 > 4-5 c 0 0 +4 >> F-- . 4-5 1 C. CL 4-5 co 1 pIO 3 CO 7 - 4-> 4-5 re 1-- re 3 F- > Cl > 4-5 CL O re 4- m re + SPI-05141 oil soluble, free radical catalyst (Mass, 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 one-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 tne 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 SPI-05142 The Po-Po is opened after every batch for cleaning, VC concentrations that are emitted when the Po-Po reactor is opened may be determined by actual measurement or by calculation as approved by the Administrator. 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). 3.4.7 Solution 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 from the initiator solution. The copolymer formed is soluble in the solvent and forms a homogeneous solution. Typical solvents listed in the literature for 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 o w u 3-29 SPI-05144 Figure 3-5. Solution process flow diagram. Concentration o f VC exhaust gases discharged to the atmosphere must not exceed 10 ppm. Recovery condenser vent Table 3-4. POINT SOURCE EMISSIONS TYPICAL OF SOLUTION PROCESS PVC PLANTS to 4-5 c cu a) k p3 C" & k c o 4-4 tO 3 d QJ 1 Q qj a; -C e -lj r-- O tO C 3 U- E OO ID fcO C". O. O C tA -- c *- e on - c o 1-- O' 4-> Cl to ro ao ^ C *r- O f* c E > -oca; <o a. 4- osa; O _u 4-5 QfO cr <1 O -c o. W 4^ iA co 4- O o O Eo (A V14J C ro o O "O - a; Q) -o V) U -C OJ 00 3 4-4 a; *F- "O o S 0) o X LU k 4-) OJ E </) s--/ OJ C. Cl k= V a.' 4-5 .3 5O L. MM O O- 00 4- 4-5 4-J 1/5 C "O a; c -a " -- qj -*-* c oo QJ r O'1 OJ -C 0) 4) OJ X o p a. a; +-> o OJ 1/5 00 k . 3 O 4-> to 45 4-> ro s o QJ e .C 4-5 c C o C X <0 p-- u -*-> a; i-- OJ 00 QJ 5 -Q o k 4-5 4- .c oo o 4-5 3 3C o 3 s oo 4- v> -C o 4-5 -X 3 X *r-- c <0 E QJ 4-5 T5 ro a> _ tO OJ 4-> > c s. CD o O 4-> k a; k a c tO 31 a> qj i. 4-> fM ok 3 QJ 03 _u u k E c 00 o cx f- 4-5 >> o 0.0 "3 to 4-2 o c C/0 4-5 CJ vo '-0 3 03 Cl QJ L CO QJ 3 -C ro Cl J3 XO OJ E 4-> o <o > QJ 4- .C O -*-> CCE o 4-5 D. *r- Q. 4-> *D 03 QJ O S- O*-^ 4-5 U C <0 "O 0; 0) U U QJ c to U O " X U -O QJ 3 QJ QJ U X 0) c 4-> 4--2 r~ to CO 4-J f- 4J 3 S- "3 oo 4-5 sV Cl 3 QJ O S- > D. -C U Cl ro -X 0) O E> O U Cl 4- CSJ -J o o ca c 4-2 Cl OJ U x a; 3C SQJ HD S- -C O 4J -C io to *3 o E H4-5 QJ IO f-- QJ OJ J4-5 >> ou 4-5 C QJ QJ C C tS- QJ ro E = QJ U i/5 C *r- ro *D Soo 2 4- 4-5 o C OJ tn 4-> -C OJ to 3 O rO E 4-5 C OJ "3 i) i a to QJ C. 3 -3 *" <0 CL 3 r / O' X O OJ O S s- -j 'O ro 4-5 >O qj z 4- JC C 4-5 COE O 4-5 cl r- a. 4-> -O ro QJ O * i- CH -O 4-5 L. 05 C r0'O4J 0J OJ L U U QJ C C to U Oo !- X QJ U*D O) L c <o 4-5 to 4-5 a; c e ro S- o CL 4o-5 VI c O <-> C* </! 00 VI E QJ OJ C r-- r- ro c c C Oi -o 4-5 ro oO a. -J c c *3 p-- o -C XJ c <o c c -C c p-- QJ V) J 4-5 c * Q) c> c I- ^ Xc <0 2: +j c c c a; CL o u o +4 u ro 0) U c o 4-5 ro N -J - O U Q a?'-- E % V) r-- to oo O. e-- 4QJ a> S- t. 4o-4 u ro O k. V) 05 CO Ok f-- ro 4-5 f to u rsj vt s- *3 $ E 05 >> p> C ro a. > QJ i3 4-J CL 3 i- S- o 4-5 U ro 05 S- 00 C OJ O O) F k 4-5 to ro -C NT U 00 k -f- 05 -3 E >> u r-* VI o O. "3 k0OtO5 o 4-5 v> to "3 <0 co o V0u1) 3c o k a. o a; ck re o a> a; -*-> ic- .o iko rOO.iXa aOi. J0f-) +E<J0 O Cl oicOVap* c- -*r3- aF-. x-- k f- o -M Z JZ VI c o 4<K-o>i Si |> o a. o*- ko i44--3O-04 Q4U<0-Q04) co 3-30 VC recovery (strip p in g ) SPI-05145 Table 3-4. Concluded to 4-5 0) c 0; Sr---. 3 CT QJ * S- 01 4-) c ro o 3 44 cn tO 0> im 2 o3> 4J QJ a: z -o -3 01 01 4-> 44 (O ro pM 33 05 O 01 01 s. u ow z 4-1 zo to to 44 4-> c p o QJ CD Cl. fO S_ Vi C o 01 44 O 4-> L. to * a>> p- e o 4-> <D O QJ 44 .* 44 4-< (0 PM S- f0 +4 O 01 4-1 01 44 V) VI - a. > 0>1 c Oi u 0> TJ o 01 o 44 Q_ Vi OQQJ o O 0u1 CL >> o pLO a 4QJJ irt O 44 V) VI <0 0<1J 3QT oi. lO o Cl o ui o> rOo Oi <0 01 Oi >L.> r&Q -o O 4-> V) to 01 Q>J a. to f^s 3-31 SPI-05146 After stripping, the resin in solution is recovered by precipitation. 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: o Concentration of VC in all exhaust gases discharged to the atmosphere must not exceed 10 ppm except for emergency relief valve discharges. o 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; connecting relief valves to process lines or recovery system or equivalent as approved by the Administrator. Equivalency clauses for these equipment requirements are included under 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 3-3? SPI-05147 continuous process in a reactor similar to a distillation column. A discussion of various reactors follows. Reactors for Suspension and Dispersion (Emulsion) Type Resits 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 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 kilopascals 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 SPI-05148  Agitation - mixing blades are usually of either retreat curve or turbine-type and provide the agitation speec 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. 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 ingredients (e.g., plasticizers, stabilizers) 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 SPI-05149 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 nave been used by several companies - a solvent solution is passed through the reactor several times to remove polymer build-up. Some have found this technique successful while others have abandoned the approach 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 SPI-05150 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 (aoout 2 nouns; 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 shaft 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. t 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 the 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 SPI-05151  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. Most 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). 3-37 SPI-05152  Slurry Blend Tank, Centrifuges, Dryers and Storage Silos these four areas are combined into sources after resin stripping. 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 SPI-05153 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, Oh'o. 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. Auqust 1975: 3-33 SPI-05154 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. 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-A? SPI-05155 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 levels than required by the current standard with the possible exception of 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) feed stock 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 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 SP1-05156 S trip p in g technology S lu rry blend tanks and Bolding C ontrolled by s trip p in g standards tank rents Table 4 -1 . POINT SOURCE EMISSIONS AND TECHNOLOGIES FOR CONTROL IN TYPICAL SUSPENSION DISPERSION AND BULK PVC PLANT i f**woe>kee s-se --k * >euk C5 w ^wk w til1 !ir **i! j!h C--0^w 5 : * * "S K rcOki c w Jw w* & tS*tcIIl Slzi Sk e t 1-*m1 two st 2s JK -c ?i* sw &o: h II li J!g "8 t .fe Hflii |ij _C islJl ^SShe **5w_*S*oSw oaki ce SI_*'Tk>ft fWeS~Wo c W >o o wa S*-8 = i W -- *d k l5l*t2i*1l I| i 4 w e I I f ils *Ko* fpkt"e5wO9t'saweS^lk> |i:il8 I1 li 2 S- I j * Slve -- W 9 * w Eib |H|el8 k w C5 O w i tslf is* i!lns sill lltfll w-*w^ w 1U O e t K "Z2f m I!UW - f I t S| |C s W I* -1 si & jc w S| s- SK tl *1 isi Is! II*II els in k2k ? ww < its IS* l l I m * t 18 ! 11# 1_o> III K^ k SIS e 5 * i -- r & s ts> 4-2 (Blending, n ix in g . neighing and holding a fte r stripping operation) Sp\-05157 P rocess s te p P o te n tia l e m is s io n p o in ts R e g u la tio n i*n u in n o l . I n n t r n l 1m h n o lo g y Xk * ft ft --o e k-- jrs I I*-* W *V Mft *f*t >. N*s kO 5w 21-til ii?' -k 8 ft*n e" Ea *1* = a. Nk *> *m t* a5i k >e e* Seu> v k { o-- l-SSSS o t> i i] H:issf fit ~t* c- *w**Tew -o I c S I ,**25 ? l?2 8ft SA * k>t k | ?t (2 c w ft* vo 5S *m* ~ -2if ft .1 lit 5 ** k 28 ft mnt I Illll .=2.i"r1s.5c* iUNweS kk m *^ u PI J k ^-I^sa * ill 1 J>i 2 N Sse-Uj ? ft f:i! |om* I? m II ** si x 282 |il j ft ;j|i t*"szs !e w iff S j S2i ft Hi ill -S-os* !S:*il ill Ip. faf llll si2 Cl s2 I t-r *21 o * U cm w m i 5k a ?* II *2 kk 22 n 2282 s v v 22S2f ft 2 4N N w 2 e VI r L fst 2 2 M ! t? ft s si r m f] I11| &O * 2 4-3 SPI-05158 Incineration VC co VIk ted froo equipment scats and operational procedures controlled to 10 ppm upon exhaust to the atmosphere. BALANCED PROCESS" EDC/VC PLANTS 4; c Q Oo I! = o Cl. CSJ I -rQo i; -- 4> Ok/ U tl tl 5 3 3. ^9 4w> * |k 41 9 <e 5= O b*/->-- -(oI i:nsi44/ v4* k --^ W 41 41 I*osc. U 4) II =O 1 <= sk. 5S 3 . I w i * *I u c* ss -- 4> cSjSf ii'ii **|rf J- :*Ii "c -* S|2 _A *Sko. > ss n2 2& i| 14l t/ 9 2 ^& Ij'.a m > c ^ --c -- */ /*g C 2/ w = i fS * we w a. vw o ; 'in. 4.1_ I* 60 * e o ** wST -- /4 v w sSt < W 4 SS-Si* t e a o * si:il8 So * -W. %** ii! i::i W4 u S2S5. e si SIsl s-Is Jts:= --* a ^ 1 StSoS U / II -- s * iio- i bSeo -iSish- i*2 flflit ral k;*S- ~W cs|a 3 Sl! = 8 S 32j;s 4-4 SPI-05159 Table 4-3. EMISSION REDUCTION FOR 316 Gq/vr EDC/'/C FACILITY IN COMPLIANCE WITH CURRENT REGULATION Emission source Current standard Relief valve discharges Non-preventable discharges only Primary control 10 ppm Oxychlorination vent 0.02 kg/100 kg EDC product Fugitive emissions Work practice and equipment standard Total emissions Uncontrolled* emissions (kg/yr) Unknown 916,400 113,760 379,200 1,409,360 Regulated*6 * * 9 emissions (kg/yr) Non-preventable discharge 3,160 50,150e 37,920 91,230 + non-preventable discharges Estimated actual emissions (kg/yr) 1,950 3,160d 25,000f 10,000h 40,110 * Based on EPA emissions estimates developed from emissions data submitted by Industrial sources. 6 Represents emissions from EDC/VC meeting current standard: actual emissions are lower. c Based on relief valve discharge data from Table 4-7 and EOC/VC production data for 1977-1980 (Chemical and Engineering News, 1980a). Production for the 9 EDC/VC plants was estimated to be 6,325 million kg for the 4-year period. An emission factor of 6.2 kg VC per million kg produced was used. 6 Due to the relatively small amount of emissions Involved, a new estimate of emissions was not made in this study. Data on hand Indicate that emissions may be lower than those shown. e Assumes balanced process and 100 percent conversion during EDC cracking. f Estimate represents an average of emission levels ranging from plants using only air and not incinerating the oxy vent (and still meeting the standard) to those using oxygen and incinerating. This estimate is based on very limited data (DeBemardl, 1981). 9 Assumes 90 percent reduction following Installation of required equipment and Implementation of leak detection and elimination programs. n Based on results of a fugitive emission study done In an EDC/VC plant (Blacksmith, et al., 1980). 4-5 SPI-05160 Table 4-4. EMISSIONS REDUCTION FOR 68 Gg/yr PVC FACILITV IN COMPLIANCE WITH CURRENT REGULATION Emission source Current standard Primary control 10 ppm Relief valve discharges Non-preventable discharges only Combined sources after resin stripping 400 ppm - suspension, latex, and bulk [2000 ppm - dispersion] Fugitive emissions Work practice and equipment standard Reactor opening loss 0.002 kg/100 kg PVC product Total emissions Uncontrol led* emissions (kg/yr) 326,400 136,000 850,000 1,040,400 312,800 2,665,600 Regulated15 emissions (kg/yr) 680 Non-preventable discharges 27,200 [136,000] 108.80C 1,360 138,040 [246,840] + non-preventable discharges Estimated actual emissions (kg/yr) 68Cc 4,780d 13,600e [74,800] 25,500f 1,360 45,920 [107,120] Based on EPA emissions estimates developed from emissions data submitted by Industrial sources. d Represents emissions from PVC plant meeting current standard; actual emissions are lower except for relief valve discharges. c Because of the relatively small amount of emissions Involved, a new estimate of emissions was not made in this study. Emissions may be lower than those shown here because plants are presently controlled lower than 10 ppm. d Based on relief valve discharge data from Table 4-6 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. Production for the 23 PVC plants was estimated to be 5800 million kg for the 4-year period. An emission factor of 70.2 kg VC per million kg PVC product was used. e Based on an average of stripping levels reported by Industrial sources. f Estimation quoted from SPI, based on B.P. Goodrich fugitive emission study (delaCruz, 1981). ^ Because of the relatively small amount of emissions Involved, a new estimate of emissions was not made. Emissions may be lower than those shown. 4-6 SP*.0516A 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 allows no excess emissions 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 SPI-05162 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 SPI-05163 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 1C 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 SPI-05164 Incinerators are equipped with flame arresters or flasn-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 rangirg 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 Clj generation from combustion and high total dissolved solids (TDS) levels from scrubbing). e 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 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 -r" iJ SPI-05165 by using a system that removes moisture prior to analysis (Laundrie, 1980). Location of the incinerator requires "safe raaius" 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 tne 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. 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 4-11 SPI-05166 upstream from the safety valve. (This represented an approval cn tne 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. A destruction efficiency of 90 percent can be calculated and therefore it is Agency policy to allow their use only on exhaust gas streams containing 100 ppm VC or less (Ullrich, 1981). 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. Installed capital costs for elevated flares range between $40,000 and $700,000; ground flares cost between $30,000 and $900,000. 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, 1978, p. 5-76). 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 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. The stack is 99 meters (325 feet) tall, thus overcoming the "safe radiation distance" 4-12 SPI-05167 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 the top of the stack. The flare has neve*' been used for that "worstcase" 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 EDC/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 4-13 SPI-05168 (Battye, 1978). 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. 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 on its way to the stripping column. In the stripping column the solvent is heated to remove the VC that comes off at the 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 aoove was Developed by B. F. Goodrich wnc plans to license the technology; it is commercially available. The only otner solvent absorption system used by tne plants surveyed in this study is a proprietary system designed by the company using it. 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. SPl-05170 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 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 includes 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. 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 SPI-05171 carbon. The gas streams treated in this way contain less than approximately i ppm of vinyl chloride end no detectable amount of ozone cr ozcnides. Another process (patent assigned to Stauffer Chemical Company) is also one in which the stream contacts ozone but without activated carpon. 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 HOls of Germany (the largest reactors currently being used in the United States range from 35,000 to 40,000 gallons) plus the 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. 4-"` SPl-05172 Relief valve discnarges, which cause short-term peak emissions. represented approximately 4 percent of total emissions from a typical 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 reduce 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. EPA Regional enforcement personnel indicate that releases are continuing to occur. As mentioned, the concern during the original standard support study was for relief valve discharges from PVC reactors 4 SPI-05173 that have the potential to discharge the entire reactor contents. Prie to the regulation, relief valve discharges were not accurately measured but typically 2,2GG kg (5,GGG 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. 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 32 plants were 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 1,025 kilograms (2,275 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 (1,200 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. A i rj SPI-05174 Table 4-5. TOTAL NUMBER OF RELIEF VALVE DISCHARGES AND QUANTITY OF VC EMITTED FROM 32 REGULATED SOURCES3 FOR THE PERIOD 1977 to 1980 1977b 197Sb 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. ^Relief 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. 4-20 SPI-05175 Table 4-6. RELIEF VALVE DISCHARGES FROM PVC PLANTS FOR THE PERIOD 1977 to 1980 R e lie f valve discharge data fo r 1980 ranges from 8 to 12 months. oic/is o > -soUtCn- CNJ co CO CNJ lD ro^s. CO in cn vo CSJ r-. CO r-* VO o ro VO ^h CSJ Lf> pH os CSJ CO VO pH CNJ m in pH pH ro CD CSJ r- roi o pH OpH fn. ml m | ro />*S /PH /--s os os pH VO m os VO os ro os GO os ro 00 ro ro 'w' <D CSJ os os ro o SH s-' pH O CNJ CSJ 00 CSJ m pH os o in OS os CSJ fH. pH pH fH OS pH pH o*i vo O H lO CNlt OS CO CO pH CNI v/S rCoM ro l<O-- CSJ -C r- GO fH. rn Cs os pH pH U os o rn. CD O' O' pH pH 4-21 c CNI m => <u .c H-> C 4UHS c D. >4-> o. 4o s- -0C> "oO e3 **&""C QJ CL <C i+-> <0 o a; H-> >> QJ c -C h-> -aCi 4- 4-> o iHC-> 4OQJ U vs i- QJ QJ OS c. i- rC CO -C in u US HH .pC "O 0us) QJ QJ > S- PD. fC 0s-) > 410 0) H-J *p* C p03 QJ p- W CL O oc D> C- ^ 0J ro -w CNI i- o JZ0.' Cl QJ 1-- i. co fCHO. OS 1--1 4o us c H-> c o E ro &. 4Oups 10 H-> C p<0 CL m oEi4ra 4-> fTJ o 0J OS L. fO -C L> UpS o QJ p>-- fO > 4QJ p- OQZJ -Q SPI-05176 Table 4-7. RELIEF VALVE DISCHARGES FROM EDC/VC PLANTS FOR THE PERIOD 1977 to 1980 1977 1978b 1979 1980c EDC/VCa plants 8 9 9 9 TOTALS Non-reactor relief valve di scharges 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 EDC/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. ""Relief valve discharge data for 1980 ranges from 8 to 10 months. 4-22 SPI-05177 Based on the relief valve discharge reports submitted to Regions1 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 folTowing 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. 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 mechanical 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 SPI-05178 Older smaller reactors are usually operatea 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 early 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 T .1U1+ SPI-05179 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, 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 rupture 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. SPI-05180 Hydroful condition. The VC liquid charge expands up to 13 percent 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 win 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 noncondensafcle 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 hydroful 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 SPI-05181 through the valve. The relief valve will reseat itself under most circumstances after pressure is released. However, if the relief valve aoes 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 1,000 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 nc longer opened as often, more consecutive batches are run and polymer continues to coat the inner reactor walls. However, new clean reactor tecnnology 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 around the 50 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. Inadequate heat removal during the auto acceleration period will usually result in a major discharge (possibly the entire contents of the 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 4-217 SP1-05182 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. This decreases the potential quantity of emissions through the relief valve. 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 i>e 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 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. 4-28 SPI-05183 Several variations of the automated shortstop system exist. The newer computer-controlled plants have built-in programs that recognize tne upset condition Dy 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 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. Dual metering in series for both VC and water can help prevent overcharging. 29 SPI-05184 Upstream filters can help maintain meter accuracy (Ullricn, 1981). A level control on the reactor that is attached to an alarm system would indicate 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 seal-water 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 SPl-05185 wcu'a be connected to existing recovery systems or control devices. Variations of this type of system are currently used by some PVC plants. Tne auxiliary venting system is aesigned for two-phase relief ana 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: overcharge of the reactor, presence of noncondensable gases, moderate reaction rate increase, and 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. 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 A-' SPI-05186 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 the 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 Auxi1iary venting system Eight 38,000 1 iter (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 accommodate 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 4-22 sw-6187 nelp to minimze relief valve aiscnarges. 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 decision-making 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. 4-23 SPI-05188 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 manually relieves pressure to the monomer recovery system 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 SPI-05189 auxiliary venting system would ce a part of the gasholder containment system. Tnere would be a separate line from the knock-out tank that would go directly to the existing recovery system or control device tor 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 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 FOR 38,000 LITER (10,000 GALLON) REACTOR Dimension Volume Diameter Height Seal Specification 5,700 m3 (200,000 ft3) 24 m 15 m water (75 ft) (48 ft) 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, 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 SPI-05190 Table 4-10. ESTIMATED COST FOR INSTALLATION OF A GASHOLDER Equipment Cost for 38,000 liter (10,000 gallon) reactor Gasholder Piping Safety Site development Header extension from KO tank Engineering and construction Contingency Operation and maintenance Total System Cost S 883,400 382,200 70,400 232,900 600,000 1,070,000 648,000 94,200 $3,981,100 4-36 SPI-05191 (Holbrook, 1975). Their Cu=t was based on earlier dollar values a^a o`a net include operation and maintenar.ee costs. Conoco estimated a gasr.c'icer system alone without the auxiliary venting system (i.e., .<nock-out tank and blowdown tank) to be between $2,000,000 and $4,000,000 for fabrication 3 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 (and several gasholders may be required). 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 SPI-0S192 venting rate. Tne maximum piston velocity for a roober-seaiec gasholder is approximately 4.6 m (15 ft) per minute and would require about 6 to 8 minutes to fill - tms sea'i 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 accommodate the discharge. For a gasholder with a capacity of 14,000 m3 (500,000 ft3 ) to 42,000 m 3 (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 the 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 SPl-05193 utilized, a safety relief device ci reedy to tr.e a ones ore re will always be required. Another containment device tnat can oe usea to neip 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 Diending 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 cnilled 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. Periodic pressure tests run on the reactor will decrease the likelihood of premature failure of rupture discs. Other preventive maintenance areas pertaining to rupture discs include (Ullrich, 1981): Premature disc failure can be caused by defective metal. This can be minimized by paying the disc supplier to pre-test a fraction of the lots before shipment. -1 SPI-05194  Premature disc failure is likely if me maximum ooeratino pressure is exceeded. This pressure "anges frcm 70 to 90 percent of the burst pressure, depending on the type of disc. Once the maximum operating pressure is exceeded, the disc is deformed and must be replaced to ensure that it does not fail below the burst pressure. It is believed that in a computer-controlled plant, the chance of exceeding the maximum operating pressure, and thus the frequency of premature disc failure, is reduced. Rupture discs in vinyl chloride service, particularly on PVC ractors, are subject to pressure cycling, and therefore must possess the best fatigue resistance. The discs should also be replaced at a frequency determined by operating experience. 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). 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 hign volume liquid ring compressors. Following is a discussion of their relief valve discharge prevention systems (Holbrook, 1980c). 5P\.05195 E. F. Goodrich has recucea releases from reactor safety re;ievalves in their large reactor suspension PVC systems. This was accom plished thrcugn process control witn empnasis on early detection ana analysis of abnormal conditions. A totally computer-controlled system was installed to detect these abnormal conditions and take the apDropriate 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. Some equipment can be kept in service with an emergency generator and back-up instrumentation. Equipment failure can be minimized by preventive maintenance following a regular schedule. The 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. 4-41 SPI-OSI*  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). 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 4-42 SPI-05197 slurry is sent to a blowdown tank, and vents are opened to a separator recovery tank (knockout tank) to relieve reactor pressure. The acce'er2ted 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 shortstoo 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 to 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 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 injectioncaused 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. 4-43 Sp\-05198 These tests confirmed that, if sufficient killing agent is injected, the polymerization reaction can be stopped and the rising -eactor 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. 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.) e Operators are always in 2-way radio contact with control panel personnel. 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. e A temperature control instrument on the control panel regulates the water flow to a reactor. If this flow is improperly regulated, 4-44 SPl-05199 a panel switch can be used to override the controller, senGing maximum cooling water flow to tne reactor. a 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 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 4-45 SPI-05200 agent for two comolete kills in one reactor. Two racks of Dack-uc 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 LEL. 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 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. 4-46 SPI-05201 Hooker Chemical Company prevents relief valve disc."arses thrcucn a combination of equipment installations and wor< practice procedures (.DuDec, 1980). Pre-Po and Po-Po reactors are mounted on scales tnat 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. 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. 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. 4-47 SPI-05202 General Tire Chemical Division. General Tire has older, smal1 reactor technology for the suspension process and uses a gasholder with a capacity of i,4C0 m" (50,000 ft") to help prevent relief valve dis charges from the suspension reactors (Laundrie, 19S0). 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 KO tank prevents entrained slurry from reaching the gasholder, but this is not always possible. A chemical shortstop agent is manually added to the reactor. 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. t 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 reactor 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, SPV0S205 single or in series, or comDination in-line ruoture disc and relief valve. In most cases non-reactor relief valves discharge lass t^an 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 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 4-49 SPI-05204 susceDtibie to cracking at the welds. -;nding the compatibly rupture disc sometimes requires trial-and-er^cr. 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 proD^m. 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 Dy 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. Although 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. 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. ~r - U SPI-05205  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 6 months and those removed are tested to verify actual rupture pressure. e Relief valves are replaced every twelve months. Emergency procedures involving non-reactor and reactor discharges are updated once every year. 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. 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 SPI-05206 caused by closing a valve too auickly and can be orevented by manually closing the valve slowly or instrumenting the vaH-e to close s'cwly. This condition can also be prevented by installing an emergencyhigh-pressure trip-switch or using a small, in-line surge vessel (Brittain, 1960a). 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 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). SPI-05207  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. Molecular weight and porosity of the resins influence stripping rates. Stripping VC from PVC resin involves (Ullrich, 1981): VC migration to resin particle surface, VC dissolution in slurry liquid, and VC vaporization and evacuation. The first step is rate controlling. Suspension resin particles are porous while dispersion and latex resin particles are not. Dispersion particles are smaller than suspension, and latex smaller than dispersion. For suspension resin, particle porosity enhances migration and allows stripping below the 400 parts per million (ppm) residual vinyl chloride (RVC) standard. For latex resin, the short migration distance allows stripping below the 400 ppm RVC standard. For dispersion resin, the longer migration distance and the particle non-porosity, in addition to thermal and mechanical stress sensitivity, make stripping more difficult. 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. 4-53 SPI-05208 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 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 countercurrent 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 changes in batch sizes and/or recipe modifications cannot use column stripping efficiently. 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. 4-54 SPI-05209 Table 4-11. PERCENT DISTRIBUTION OF STRIPPING LEVELS BEING ACHIEVED BY INDUSTRYa Suspension Concentration of residual vinyl chloride in ppm Range of daily average , (PPr)____ Plant Method' 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 Z 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 CM 8 8c 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 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 14 b 1,099 71 93.8 86.9 81.8 68.7 51.1 15.9 2.8 0 15 1.0S6 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 `Based on EPA semi-annual reports for March through September 1930 obtained from Regional EPA offices. Data represents approximately SOU of suspension, 75X of bulk, and 33X of latex plants. ^Individual data are percentages of time'that concentration falls below specified levels. Values represent daily averages weighted on a production basis. Stethod: b > batch; c continuous (continued1 4-55 SPJ-05210 SPI-05211 Dispersion Concentration o f residual vinyl chloride in ppni  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). 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'1 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.) 4-57 SPI-05212 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 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 4-11. 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. 4-53 SPI-05213 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 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 4-59 SPI-05214 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 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 0.034 kg/100 kg of PVC produced or approximately 20 percent of the EPA estimated controlled 1975 rate (de la Cruz, 1981). New large reactor suspension processes were determined to be 0.0085 kg/100 kg of PVC produced or only 5 percent of the EPA estimated controlled 1975 rate (de la Cruz, 1981). 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, e Sampling and laboratory analysis, e Equipment opening for cleaning and maintenance, e Pipe and equipment flanges, e Process drains and manhole cover seals, e Process valves and pressure relief valves, and 4-60  Open-ended lines. 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. 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 4-61 SPI-05216 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. 4.4.2 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 4-52 SPI-05217 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 of VC contamination. Regulations proposed for benzene fugitive emissions specify properties of the seal or barrier fluid. The fluid must have a vapor pressure less than 0.4 kPa (0.1 lb/in2) at 20C (68F) or it can be a heavy fluid such as kerosene or diesel oil. No such requirement is specified by the VC regulation. Other equipment requirements are double outboard seals on reciprocating pumps and compressors, and rupture discs upstream from 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), one PVC plant has connected 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). 4-63 SPl-05218 2. Used on c e n trifu g a l compressors. en O) o> cj tn w 3 a0). <X> 3WO X >> f- VO f- 3 0) c o f o 33 M C 3 P 4-> vo c co VO 0) 3 0) *- c 3 S- X <o 01 CL E 4-> co 3 3 s to to x > in O co 4o-> 3XXX s 4-> QO VO 0> o c p to F-- o F- X U p co c *- 3 0) CO X *f > P 4-- (0 0)f GO VO a> +-> *f a to o > C 01 cr>*-- *- vo co 3 cn <j 0) co W 01 c 3 u w 3 X *f w 0) F- X 3 tn z o v> CL LU o v> fa01 tT3oO>CNJ *To--O s- .* a e i- tCoO t0o> O O -- X 3 0> 4-> 3 S_ -C -- CO O Of L w0) ct'or X f" X *3OJ3f <0CJ X |Q vo v-- 3 CO >w> oW eu X c tn co E vo C 3W CO vo *f to coo> Wa wa Xc o CO P C Q. C O vo voo rqj X OJ 3X W O COJO*- X3W 0) ^cf- *- o e 0>> X 10 30 co <J 3 OCO f 3*to0 f- CtoO i--c0o) O f- vo 4- S- W co O to C c 0) CtoO X co f 0) vo c a 0) in oCO4->w>f Foo- auVOi o> e o o E cw f- co a 3 F- to 3 3 fC-O 0) Qu) E >o 3c vo 1 o> W 0) X C . to>- o F- *F> 0) x E a 3 0> w 3 w u c co o C 0> 3 c to W 3 +JT3 cuc O u sE i-- 0) CO F* X 3 C O c f-- 0) x *- 0 3 0) a X *- u 0) oo c CO F *F X 0) P 3 3 CO > OO 0) .c vo co O O 3 3 CO 3 X co tor" f-- co OJ C 0) 3 X o >, E S vo W S3 x > W (J a to 0) * >>x cn .X S- X 0) X X to > o x to O CO 0) s- W 0) X o > -- X3 o w c to s- o o X vo OS- u C tt X X>o a o s- x a ao c c w w to to o F 0) x c *- 0X3 ax to to - ax CO <D O o ax x o CO X U CO 0) 3 o c c co X W C CO X > s > *F* to tocr uu 3 p Euc C 0) o e vo co CO 0) 0) 0) *c-- CO 33 w 0) a 3 o 1- L (/) VO E CO X vo r-- X co o 0) VO 3C CO X c X X 0) vo 0) CO 3 O VO co V) X W T- 01 VO (J CO X 0) !- to n o w c vo 0) to c vo s (0 E _ _ CO 0) co a. co f- > X W s O' Xw w v- a 3 a O 3 OC 3O *--OWfOf <P f- X CO 4- **3 o O i3 c O u *Xr- 3 co v0o) X fE- J. oO v>: a u 3 3 0w) OaCal <c & z o c x o) 0) CO XW f3- 3vo oo >3p- WO a* a> LU 3 3" in 3Q0). 0aW) CL fX oa: Xc a0:) CD 3ato* CaO; vo o ao- - C co >> f- 0) QC S- VO CL CaL. < I 3ttu-o c 3O W' Ol 3 V0---* 0) -3 x<o Ov 0) x a co >' r- tro VO co to ^-- u > X3 VO 1--1 C 31 FXlD 0) QJ 3 to r- C 0) P- o f >5 L 0)f 3 CD 3 WX 3 Xo X Of 0) 3 F- tn a co'--1 cX I ato 3 w oo c EX '0) to X X V) X VO *i-- X >> to vo p vo 3 CO f- T3 Q. rB T3 a> vo a; 0) N VO N -F- X wco w 3 QJ 3 3 01 o CO a c f-- vo OO c p w in 30 FC co 0)010) to vo S- X X w co a x x a oo <0 X ao CSI V) r-- 30 C CO C O 0) *t-- V-- v> X X CO U- w o tn > CO oa ^ X *f-- vo P--1 X <j w m 3 0) O --" o W VO f-b V) X to F- 01 ^ f Lin X co a to 3 E o c o --< o o ux 4-&4 3 > c X S CO x X vo t-- X 3 to 0) e i- to x 1-- 3 *3IS 3 , X VO f s o f-- 3 w co O O to vo to E 3 > 3 V) x o c OO to 3 C 3>3 O" O 3 CO *F" F oo F- W j a co ao c o X T3 f- UCO 35 *fC- 3 <t/o0 f- X* C- C <o X tuo E 0) r-- co x o s- e o VO cO __ to --' Q.^ E co 3a^' X E3 --tn"- X f3 CO O to a s vo Eo iuo O VO Lab^ in th seals SPI-05219 Table 4-12. Concluded E C- o O S- M 00 tO **-> tn c ---- QJ O > CL w *o qj QJ i- O -- 0> >>(0 4-) C TO Z +-> -- -- "O c u o oj to o Ol U (4 u >p M S- 3 -- -- <0 OJ S- C S- Z 0) -Q O TO 0 3 Q- TO r-- Q) +> CP 1 (4*D> -*- 3 C W -- E TO < 3$ O O- X 4- *U E O QJ 3 -O to C C Ql -- 4-> f- 3> CO f" -O -o c TO 5 E QJ -- -O 0 -- 4- ^U s. QJ QJ 3 TO O-p-- </) 4-> -- C. TO to 3 -- --< o> 3W OOP O" QJ JC L- CO 4) +J 4-> O to fc- c-- Q. E O QJ O O -*-> 4-) 3 CO TO 10 T> QJ CL C U U QJ Q -- -- -- -X <4^ L L U^> r U00 0) (4 -C TO 3 3 -C 3 H- Q.---- U E o QJ 4-> C QJ > CP > qj .O co T3 O -- C. 3 4-> C 5o o CO -- -C TO o *P" CO -C 2 QJ CO x: S-- co 4-> a) a; Q- -- c a. qj -- t QJ w o 2 M 4J (4 E QJ O. -O S- Q. QJ 4-> -P O C TO -- to 2 O QJ O e 4J 4-> &- CO -O TO CO 33C --O QJ S- JO CO Q> 4-> CO JJ -- TO O E 3-MO) CO CO aj o O Jc. CP > Q. -- 3C JZ o QJ x; to 4^ 4-> O CO 4-> TO Q> -- -- TO ^ P> TO C OO --K -- O. &CL O TO QJ CO -O >> O >0 -- -- O. CE O QJ -o -C --U 3-- O -C 23 *H CNJ H > o c QJ TO f>-- +J 3 CO O' QJ LU 3 0" eoQJ QJ K f-- 3 a* LU TO ie to c-- TO 3 T3 cp co o) e O -- QJ TO STOW 1 QJ co -- 40 C -- co -- o-- 2C 4J -- 4-> f +J u O) 2 TO QJ C CJ CL -- O.-- . to p E l c cn TO 3 -O -- c u a. 3 -- O -- QJ &C ^ 0)0 a. 4P TO 4-> O Q. TO TO C QJ >-C QJ O QCI- CO -- g # -o QcJ. t- c 3O t0t) 4T-O CO --IQ. >. 3cr CO <o TO m CO to --i o>vo "Du --c c Ofl 4fUJl1*O-M'-' 4-> O U --* O Q.C/J -- QJ CJ --* iJO--3O3 QU> OC Q3EVQ)>*.C--wOO*^ 4-> -C C 4-> TO --U 3-- tco .O as qj E-- u 3-- CL W > QJ QJ D)4-> *0 C TO -- *-- O CJ 4- STO O <P CL C O) o aieu lO 3C O QJ O O> E OJ +> CO >> CO QJ CTOO QJ W o> "O QJ fM fc. 3 CO CO QJ sa. CSJ H TO *-> c -- CO too -- CJ N +> --* 5I> -(uCO .OC*l--/)--sr->'* qj i- tow^ Eo *. O -Q qj +> '-- -- /> LO .3OO --<00) --4o->> C--O 0 14 (0\O 4-55 SPI-05220 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 a rupture 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. 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: e Manual (non-emergency) 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. 4-66 SPI-05221 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 plant'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. Most plants use 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 lines are purged to the recovery system. One plant has modified the unloading lines from railcars so that after purging the lines to recovery, the volume remaining in the 3 lines is less than 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: e 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, e - an acceptable calibration and Maintenance schedule for the area monitoring system and the portable HC detector, e 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 4-67 SPI-05222  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 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. After any initial waiver period for compliance, a few of the EPA Regions evaluated the adequacy of the leak detection and elimination programs (Battye, 1978; Battye and Hall, 1978). Evaluations were also made during these 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 4-68 SPI-05223 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. 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: e monitor processes for leaks, e 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 4-69 SPI-05224 Table 4-13. LEAK DETECTION AND ELIMINATION PROGRAMS II Ka ^* *7 < -- a. x *I rg <9 A 3>*" ?-.S= *i- foe, I; i&ua| ,510 ft ** C S ?w 1 W w** MW * ** * ft* ii j v e > ft .5 -- ft o >> a* u - . U c wo Zi9,Cfs<t.ijWfWSt 3O'DfUCtOw>ft t29lf^.tf*tli* ( U ft* w j*ft?:f=t *ft* Qe *x ^ a* * .** ft* ** >s U> a. a w t* w - * ftO->Kftl m~ s JXftftftoV33svi iiii?35kxEft ^(i >**- - 5c V-- '. <e -- -- >J.m3i ~ g ~ ^ ; a* * -- --*3a*ft*c. ^ * S ft a < * ft eta a* c j Jft ^ft* -->. o+ft* ---- ^c .fffttt ^ u < X --C 41 8*Aft --W X >* C_G."'fmfOtt -A i"fi. e -- -- a c C0S W*ftX-- CsaT.3fCItft^>ji C w ftft! -- --ft^*^*>e3u*fut Xfftt ^*ft-* < -l. -- ft 3S ' * - S'= ft o i 5 O ft ( g : cA M9> O. O cm w? c | ;s 22 er ss :s> .: 4ft* f+* X. *f I .22. -- O ft a ft ft ft ft ft > ft U > ft WOftjgft^.fi'O o'fft'fl^ o ** w & v g c*< e ** as e d o taeCtngf*tau*Oa*^^f.OtprSfs*f>*t*Ttja`|--f-fc>teSJ* iu^eS3S<SaSaIo. --i ft C ^C w ft ft -.w<O- *of*t fe| ft. o "O 3<D C , ,^6 Sis -- -- m m 4-70 SPI-05225 A c tio n level for portable H.C. Process equipment check program Walkthrough program C alibration and maintenance schedules minor repairs; s h if t foreman fo r more extensive problems. Response/ party responsible fo r repair - ,, . & Oe -* mk ^t4l* )| OV> -- >W.**/ tl --A 3 U 1 E A * -- * A --<e---- S w --U Xw --u t -- !! I" OU 1 U *>' X ^ --* W e . </ fsj D Q. i/i iA i < i/ 3 be .> w b b A O ^ EJ >.w -- /i x c --O. ^ ~ * -- X > C4 9-- 41 b3 Cr `C s e * i i si n ^ObwMUtb>O*X9^*o*b*wiJi_o<*3_wg/C_0^c.-.5QC* o--/_.bb_ * mC * wI bb 43 U *|? t/lo. ---- .--*- -w*- A+->x+- <e --^ i eo S ut w S iaa*oc* 2ea --g5w i> 0> -- C C > ~ o * .* o-. m u,, i > w ob -- X * +* b 3W bO' B *O *-- Sw --encm! f :js r'eCr li CSEti 4i -efc I Si -- I M M K V <! g ^ .* S3Wge*1}sbX*J*52.i85tAetbiS2b>'*5**w**:^~2~xo-c*cMrs5--xUC uo_oi* be si / b3 we iw . e OC -B2 ^k *b<*0#y<> 9> mC us. S Ha _cTr*-~Ifaeesyt:oo w * v* o u Si *^ c* ^* oc v6-: --e' e 5| ~b Jb --o e o <j ** a. 6 b- W U. e :I I*--> V Background concentration le v e l Area m onitoring sampling in te rv a l Type o f area m onitor/ p o in ts P la n t 4-71 SPI-05226 ib le 4-14. VARIABILITY IN LEAK DEFINITIONS Routine survey conducted only I f area moi .or printout Indicates leaks. L O>) 5E. O) oj a > U C *-* #-- c ! >*>O--'h --* 4-* 4- -O 9 *A C w cn *OJ C j L </>**!/> C **0 i*/>- CL U aCcj u e*- 5g. -J wQ. s OE 'V.i > Q.. 00 o - *-- CsJ %fi fcO I Uft CSJ E . O^ s> US wW *c OU U. + u 4- V) O - 0 U L t/l e 3k -- e VC & -J & **->-- < | sil 8 8 CO o s - d - -S 3 I * *> u s 3u VI6 --O og^e --***u* ow0. ae & > e-4 * M ' *+ E C fi. _i - a. m 4-72 SPI-05227 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 by 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 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 4-73 SPl-05228 of the program because VC concentrations at this point should decrease with time. Area sampling lines in some cases are manifolded to two 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 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 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. Host 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 4-74 SPI-05229 CM OL. jQQOJ SQ. CL Oi "D C f5OI ^o<>o OsJ. w a; o Id. cron o \D o 00 o a^4v * CT) CM CM Table 4-15. CALIBRATION RESULTS FOR AREA-WIDE MONITOR 4-75 SPI-05230 for a cycle, any alarm level sounded, and if no alarm level is sounded (this is the 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. 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, EOCOM Fourier multiplex infrared spectrometer (FMS-7200) with minicomputer, and e Miran 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 4-76 SPI-05231 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 P1101) - 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 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, 4-77 SPI-05232 i 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. 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 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 4-78 SPI-05233 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 levels 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 and follow-up inspections, determined the following minimum requirements for an acceptable program (Ramirez, 1978): a' 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. e 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. e 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. 4-79 SPI-05234  The evaluation of each company's background level will be made on a case-bycase 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 to inprocess wastewater, it is still categorized as a fugitive emissions source. The regulation requires the concentration of VC in each wastewater 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, e water used as sealing fluid for double mechanical seals on pumps, compressors and agitators, 4-80 SPI-05235  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 gasholder, 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, 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 4-81 5PI-05236 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: minimize the amount of VC in the gas phase prior to opening, and/or maximize the total PVC production per reactor opening. 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: 4-82 gpi-05237 Table 4-16. REACTOR OPENING LOSS REPORTED BY REPRESENTATIVE COMPANIES (Data from September 1980 Semi-Annual Reports) <4BUOuJUt9X~'? 4-83 Company (code) Resin type suspen. suspen. suspen. suspen. disper. dlsper. disper. bulk bulk latex Nuaber o f openings 72 47 757 215 60 766 5668 1340 Nuaber out o f compliance 2 1 2 0 2 0 0 IS 0 0 Compliance rate 97.2* 97.9* 99.7* 100* 96.7* 1001 100* 99.7* 100* 100* . Concentration calculated or actual actual calculation actual actual calculation Control technology Hater Piston (Oisplaceaient) Steam Sweep Technology Solvent Cleaning/Closed Charge Technology Steam Injection Steam Sweep Technology Redox Catalysis SPI-05238  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. The hydrocarbons present due to solvent cleaning may require that ROL be measured by Method 106 to demonstrate compliance, or that many more batches be produced between reactor opening. This is because the hydrocarbon detector (alternative to using Method 106) will measure these hydrocarbons along with VC (Ullrich, 1981). 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 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 4-84 S PI-05239 "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. 4.5.5 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 4-85 SPI-05240 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 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 4-86 SPI-05241 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 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. 4-87 SPI-05242 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). 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. 4-88 SPI-05243 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 Raoult's Law. The general requirements for using calculated ROL's 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. 4-89 SPI-05244 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. 4-90 gp\-05245 de la Cruz, Peter L., 1981. Assistant General Counsel to the Society of Plastics Industry, Inc. Letter with attachments to Don R. Goodwin, EPA. May 18, 1981. 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. 78-0CM-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-450/2-75-009, 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. 4-91 SPI-05246 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. 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. 68-02-285)9. December 1978. 4-92 SPI-05247 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. 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. Ullrich, David A. 1981. Chief, Air Enforcement Branch EPA Region V. Letter to Jack R. Farmer, EPA. April 28, 1981. 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. 4-93 SPJ-05248 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 the potential revision of the NESHAP without a health effects basis (Baise, 1980). 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 SPI-05249 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 NESMAP (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 nohhazardous 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 e 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. a 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 a 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; 5-2 SP1-05250 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 concen trations.) Compliance with the regulations has drastically increased the quantity of inprocess 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 in 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 SPI-05251 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 e The primary concern about relief valve discharges, shared by industry and regional EPA personnel, is how to determine what 5-4 SPI-05252 is "preventable.11 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 compliance (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)11 to "relief device, including but not limited to ..." (Brittain, 1980). 5.8 RESIN STRIPPING e Industry feels that the language used to describe resin "grade" is ambiguous and this affects the resin stripping regulations. e 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 SPI-05253 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." (40 FR 59543.) One region suggested the possibility of applying a standard for new sources (under NESHAP) for stripping with retrofitting requirements 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). e 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 SPI-05254 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 gigagrams (100,000 tons) of PVC resin per year will emit 36 megagrams (40 tons) of VC per year out of the dryer stack. Indirect drying (which would reduce drying air and make add-on control more economical) followed by a control or product recovery device, 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 (Pucci, 1980). e 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 ROL and the 10 ppm exhaust emissions (Varner, 1980). 5.10 FUGITIVE EMISSIONS e 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). e 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). e An industrial source maintains that the levels for fugitives predicted in the Standard Support and Environmental Impact 5-7 5PI-05255 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 OSHA 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 e 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 ROL 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 5-8 SPJ-05256 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). e Method 106 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 SPI-05257 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. The regions do not consider the semi-annual reports effective for enforcement purposes. 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 required to be maintained onsite. One region pointed out that the standard does not specify a time limit for retention of ROL 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 Spl-05258 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 SPI-05259 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 SPI-05260 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 SPI-05261 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 SPl-05262 6.0 UNREGULATED SOURCES OF VINYL CHLORIDE 6.1 INTRODUCTION The current regulation is applicable to the following types of faci1ities: (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 fraction 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 SPI-05263 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. e 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: a 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 6-2 SPI-05264  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 0SHA VC workplace standard (permissible occupational exposure level), PVC manufacturers have reduced the RVC content in the resins supplied to compounders and fabricators. To control VC exposure in fabrication facilities, the PVC industry has established a 10 ppm VC concentration limit 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 0SHA 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 SPI-05265 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: e Industrial processes in which VC is used as a chemical intermediate for the production of other chemicals, e 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 0. 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. 6-4 gpl-05266 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 cars, tank cars, and marine unloading facilities. Department of Transportation (DOT) and Coast 6-5 SP1-05267 Guard regulations are primarily concerned with flammability and water pollution parameters. Ho 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). 6-6 SP'*6268 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). 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 Disposal Facilities (Landfill) VC emissions were recently detected in vents from a landfill in 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 (and therefore unregulated), 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 accept able 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 SPI-05269 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 Processing. 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. 66-0Fl3327"Task No. 13 (Part 1-37 B, CT~^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. Ter Haar, Gary L. 1980. Director of Toxicology and Industrial Hygiene, Ethyl Corporation. Letter with attachments to Docket Officer, DOL - 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. Union Carbide Corp. Telecon to M. A. Cassidy, TRW. January 7, 1981. 6-8 SPI-05270 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 SPI-05271 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) e Toxic Substances Control Act (TSCA) Toxic Pollutant Effluent Standards as required by the Clean Water Act of 1977 e Proposed Primary Drinking Water Regulations Potential Revision to the OSHA Workplace Standard for VC Transport of Hazardous Wastes and Hazardous Substances . a Pood 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. 7-2 SPI-05272 7.2.1 Carcinogen Rule The Carcinogen Rule, proposed on October 10, 1979, considered policies and procedures to: t 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 7-3 Spl-05273 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, a 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 7-4 sp\.0527* 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 E0C/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 7-5 SPI-05275 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 SPI-05276 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, 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 SPI-05277 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, npnperiodic 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. 7-8 gpi-05278 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. 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. Hew 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. 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. 7-9 SPI-05279 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. PVC sludge is not designated as a hazardous waste. These solid wastes have been specifically listed as hazardous but it will still be the responsibility of the generator to show that other 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). 7-10 SPI-05280 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 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 7-11 SPJ-05281 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 were developed (see Federal Register, Nov. 28, 1980; 45 FR 79318). These effluent limitations are 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. 7.6 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. 7-12 SPI-0S282 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. 7.9 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 7-13 SPI-05283  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 meet requirements specifying operational procedures, resin characteristics (e.g., clarity, purity), and residual levels of VC (RVC). The FDA 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. 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. 7-14 SPl-084 7.11 REFERENCES FOR CHAPTER 7 Brittain, Martin. 1980. NESHAP Coordinator, EPA Region VI. Meeting Report. July 23, 1980. 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. 1979(b). A Handbook of Key Federal Regulations and Criteria for Multimedia Environmental Control. EPA-600/7-79-175. August 1979(b). 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 Report: B.F. Goodrich, General Tire and TRW representatives. October 30, 1980. Winkler, Joe. 1980. Technical Support Section, EPA Region VI. Preliminary Determination for Formosa Plastics Company, PSD-TX-226. February 22, 1980. 7-15 sp'-5285 APPENDIX A VINYL CHLORIDE NATIONAL EMISSIONS STANDARD FOR HAZARDOUS AIR POLLUTANTS SPI-05286 CMAFTtK 1 mCNVWONMCMTAL mOTtCTION ANCY IVKNtfW C <MO *POlMM FART l--HATIONAl EMISSION STANDAltOS rOS HAZARDOUS AUK SOLU/TAffTf IAS 01IAjMt AA6 ti.oo OAT AppttOftMUtr OoftttittQQ*. iIiImMrIiWrH.UBM KtbiMM MttvtlMB. Dttcmiaouoa or wuitfisimi or IMMlUlKHWl 6|Ult1rtn <r opptfM Of om*> OaAJti AMoocparamoBlttAfAtotttomip. OAt Sow fopnctiny tad Oil Uttar ot nnoftfitliom* 041 AUJ1t fVeu4n0r8olMi Oftt no 046 loom toot QOS ODOlyHaol QttStSt 041 AvofiobCOT ot Utfi--Wii fSJf ttiu uitoriQ. tlJT OBOtUMMlM.' OJS Dfgrtri<IO AircltontBf. out -- ------ f 01 IIJ4 IAS l At l AT oi .co ti.tt 1 TO l.Ti plOBU. Bmuoion rttadord for poifrtart eMo' TO* pioati C&iMton cundird for otfcyloor dl- cfelerMr. tinyi chloride tad poly* noyi ebioridt piuiu tquiroJont oquipoost oatf proordurotSmitsioo toots BBumon noaiionaf. Initial nport. SoaUeoutl report. sooorqrwpLnt, AAppppooooiQooi ts--OTr--ootpamtontbfosssu. twfft--lino ttstbot iOi--Soforooat aotlod for Aennu* oourooo <oirssues oi poamiMe sad foaout ooroury --itmno froo tuttosorr Method iat Mortiot for so-- or aoreury la watotacor goomoot ptoat oowoco atadfm.7 Method top,. Potonainatiaaat oiayi otwoido tot ottUeaWT aeoreos.n Method iot--Pwormlnoaoo of nayl ehiond* of ttprcta nnmur staples. im wayi cfciortd* eoatoat `of poiynnyi ehlortd*,. ms. ahon. ms Aka. aaf lotos ! TWUaiLt . ui mt) ttat A-l SPl-05287 it Qanaril Pruyf eraunt dasttn capacity of the etaSSonary eource: (il) An increase in boon of operation. 161.01 Applicability. The provisions of this put apply to the owner or operator of any stationary eource for which a standard li prescribed under thle part. (k> "New source" meins any stationary source, the construction or modification of which Is commenced after the publi cation in the fmu Kmmi at pro posed national emission standards for hazardous air pollutants which will ba | <1412 Definition*. At uied In thli part, all term not de fined herein ihall have the meanlna (Iren them in the act: >a> "Act" meant the Clean Air Act (42 O.S.C. inlet tea.). ib' "Administrator" meant the Admtnletrator of the Environmental Pro tection Agency or hie authorized repre sentative. . (c) "Alternative method" meani any method of eampllng and anairtinr for an air pollutant which ti net a reference method or an eoutralent method hut which ha been demonstrated to the Administrator's satisfaction to produce. In specific eases, results adequate for his determination of compliance.1 id) "Commenced" means that an own applicable to such source. (!) "Owner or operator" means any person who owns. Masts, operates, con trols. or supervises a stationary source. (m> "Reference method" means any method of samplinf and analrung for an (Secuom ns i AMe1l tto>!->* Ono1<*> cf l*t Air | SAC. idle-7. air pollutant, as dmerfbad in Appendix B to this part. (n) "Startup" means the aettmx m I 41.04 operation of a stationary source for any (a) AH requests. rtporta applications (oi "Standard* maans a national tmission standard for a aaaardous air pollutant prop osid or promulfatsd under this part. submittals. and ocher coomunlcations to tho Administrator punuant to this post Shan ba submitted in duplicate and ad dressed to tho appropriate Regional Of p) "Statlouary source' meins any fice at tho environmental Protection building, stmcturs. facility. or mstalla- Aftncy, to the attention at tho Director tton which emits or may emit any air fcifortmnem Division. Ths riglimel atpollutant which has been iMilgnatsd as fleas are as follows. hazardous by the Administrator. Baglon I (Connecticut. Slalaa. Maw Hamp- er or operator has undertaken a con tinuous procram of construction or modification or that an owner or operator has entered into a contractual oblisation to undertake and complete, within a rea- aoncble time, a continuous prafram of construction or modification. (e> "Compliance schedule" means the date or dates by which a source or cate gory of sources is required to comply with the standards of this part and wtth any steps toward such compliance which are at forth lc a wairer of oompUaooe under f 41.03 Hah* and sbbm iallasu 48 Oaad In thla part are abbreviations and bull at units of measure. Timm an daflnad as follows:' damn" totemstsmal (SI) unto sBneotwuta).. JUoabMs MPb. uSaavafatiateiynP.adarai Suliaiaa. Stagloa n (We* Tor*. New Jemr. Puart Blco. Virgin Islanos). PMsral Odtoa Buliotog. se Peaami PUu (Poley Square;. rlsv Tara. ITT. 10007. negiou m (Delaware. Dsstr.es or OolirmDla Mauritania. Maryland. Virginia. Wees Vlr(>nui. cum* auuetnt. atzta and Waians nuesia Pluiamiphu Pvnaayleaaia 1II0S Bagiae IV (A'aoama. PlorsOa Qaorgla up. taUPPl- Sastuaky, Harm Carolina aoata Oesaiins Tanneassn. nuts* 100. itl PaaabOw SMat Atlanta. Oaaegta SOM. 14l.ll. if) "Construction" maaas fabrication sncbijan. Onto twaaonaai. SS0 gwy^paar- erection, or Inatallatkui at a stationary sowcc. if) "Mactiva data" Is tbs Oats or promulcatlon m the Pmast 11inis ns of an applicable standard or other recuIhMhW wtMf tliH haglim VB (leva. Vsnsai. : mrnai. tig* r mm sina ib) "equivalent method" mmns any method of aampllnx and analrzins for an air pollutant which has been demon awamvaaalsa strated to the Administrator's satisfac hsglnn rm (Ooiarado.1______________ ~~ mssmOabca*. Utah. TThmIiii i im tion to have a eonststest and euantlta- Uveiy known relattonehip to the reference method, under specified eondittona (I) "EztsUnr source" miens any sta tionary source which is not a mw sauna. <J> "Modification" moans any Physical chants in. or chance m the method of operation of. a stationary mums which tnereasu tho *"'** of any hnmrdous air pollutant emitted by such source or which units m the amllon at any Paiaaim Pcbiaaaui X rw I. moo ~ toioi. (b) aaetion 112(d) diraeto the Admm- appeoprtato. tho authority to tmpiement and enforce tho Mason stand ards for haosdoua air pollutants for stadmmiT saurom Monad to such mate. AO Infannasioc nqufsed to bo sntaulttod to KPA under pangmph (a) of this toe- (1) Roattne mahitenanoa repair, and don. must also bo submitted to the ep- lupMomaant shall not ha osnsdmud physleal ahanam. and whiefa thM authority has baan itslswfl i from a certain fad~nl nr firsts nrnrilng iiiiiiIiuiimiIi Tin approprtata malting addnm for thorn Sasm whom italsgetliui raquaetha* boon A-2 gpi-05288 iA) iRmrrnJ! 9/*) suit of Alabama. Alt Pollution Con- i Division. Ait Pollution Control Cosuoi** Alabama MICH * (Cl |Bmsrv*d] (Di Amons Pitt* County Ait Pollution Centro) Dis- *j5\7t(0\I.)) ]- IB___ _ _ ,*> c^ifcwm. **.*.** II Bar Atm Ait touuuon Control District. 339 CUU Burnt. Son Praawseo. CA 94103. Del Notts County Mr Pollution Control District. OourtneuM. crmeont City. CA 95631 Prmne County An Pollution Control Dts- tnet. Sift S. Coder Aoonut. Prmno, CA 33703. Bumbeldt County Alt Pollution Control District. Moe a. Broadway. Burma. CA BMOl. Korn County Alt Pollution Control Dte* tBnaekvsnS17a0m0. cnamf3*roaBerest <p.o tot *37> MaAers County Air Miutfon Control Du* trtct. 1U W. Tenniu Amour. Modor*. CA Miatccao County Air Psnutwra costroi District. County courthouse. Uhlan. CA 9B4B3 Monterey Bay Umflotf Air Pntluttee Con* otrmol.aDoiisutrwicAt,C4A30tsaCohiu.rch Street (P-O. Sot Northern ionom* County Air Pollution Oontroi oumet. 3313 Cbaaate Bond. anntn Km CA 00404 Bactimento County Air Pollution Central Diemet. 3701 Brtnen Center ten*, Beers* monte. CA3SB37. Ban Otago county Air Pollution Control cAosiaiDtttriev 4100 rnmapmf Orta*. Bna Oman. One Tnoema County Air Prtiooao Oontroi fifstnet. iaoi % BnaMton Sweet fp.o. Bon 0003!. ttoebtoa. CA *6301, Innto Barbara Air Penman Oontroi DM* *t*. 4M0 Calls deal. sum Swims ca bub. aomuitui County Air PoDvttos Control Barn BN Bronir firm. Moiaeta. CA MSM. Trtnitr Cvonw akp footmen Oontroi Dm* (TD tuu of Hunt fiounmnont of In* vtoment*l Ttotmyoa. Btnto Doans. Au* gust*. Must 04390- (V) fBmsrmd) CW> Massachusetts Department of fine!fonmonui duality tnttneertn*. Dictate* of Air duality Control. BOO Washington Boost. Beaton. Mamaeheaetca Qam.'7 (Xi But* or Michigan. Air PoDunoo eon* trol D)vision. Michigan Popnrtmont or Katu* r*i Bmourem. Stereo* T Maape Building, tth floor. Lancing. Michigan 4B3tt.17 (T) Minamata PoUutteo Control Korney. fitvtiMo or Air Qualify. Ml Boat Conwy tend B-3. BtaarUls, Man. Mill44 ill (Bmtrcadl At |*i <B) Mov Bampttitro Air Pollution Con trol Atoney. Dapartmast of Saalth and Wa)> Cv. Btata Laboratory BuUdinc^Haaan Drtrt. Concern U* Bampanir* ouoi.17 Iff I Mad of tlao Janoy: Don Ammf fin* (ir> Spokane County Air MUtfBon Con trol Authorty. Berth tU Jaffarmo. Sp*rt Washington snoi. (v) Yakima County CVoos Air Authority, ^untj^^ourthoum, TAkimn Wmtuagtoa (tt> Olympic Air fnuuttm Control Au* thorny, iso taat Buta Aranas, Olympia, Washington 0*501. tTii) Bouthvmt Air Pollution Codtrot Au* thertty. Brnta TOO! K. KB Kami Can itnw. fa&snmr, Vnuaim mm.1) iXX) iKaimd! *VT> Wisconsin--Wisconsin Omartmen' or Natural Resource*. PO Box 7031, Maat- *onl.ilWtisc(Bonmsainnr6s3dTQi 7 *7 (AAAi i Smarted j trtsita t Puerto Bicc Puarta Rico Knricon> amotai duality Beam PC Bet tlT*4 sa tttrea.PJt 00!0*5 (CCC VM. Vtrcin Island* on Viin ialAnds Oapartmaut of Conservation ana Cultural Affair*. PO Box 37*. Ctoariotu Amalie, at. Thomas CA. Virgin uiaoeto 00301. *3 tBacs. loi. lio. m. 113 and 30) of th* Claan Air Act. m Mnandad. 43 OA.C. i7. iBSic* 1.3.7 and lB7g. > iCO) )>oorvart) <KHi Non Torn Kao Tor* State Depart* meat of Bnrtroomantal CaAaarvatien. SC Wolf Dead. Albany. No Torc.iaa. attontmn: fimaioa of Air Baaomma.* mi North Carolina Bnnronmotal Man* Afamant Comauaiea.'Dapanznant of Natural and Economic amourtm Dmaion of final* tnnnuntai Manaeamant. P O. fine 7T*7. fia* Msa. Karth Caniis* 37311. Attantaon: Air duality Section.** (33> Btata of Berth Dahom, Btata Da- taooi.vm. BM s. Bants CUA tamt. Vanm*. ca <v> me or nninruna cwwm air Ptt- bsdon Qttiott^Pmiaun^m^nat ltd dam of fcrtruftimnmi^awiailm tBrnTSS MMd Dartfard. nnwstimi atXW.14 ID tan of firtauma (for tmfl taltaf (sanafi for City of Philadelphia and Alia* Many County) Penney!vama fisinHinani of fiuvifnnmoDtAi Bmruirum. Burma f Air duality and Koua Control. Past Ode* Bon BBSS. Jumaoum Pmmsyivama itiao. (hi City of PhDsiMphi* Phiisddphf Pa partmant of Puhlie Dsaith Air Manacnnons Bar num 3QI Arh straat. PMisdMfTita. Pann* aytraata 13137. M (OO) rfimerrodl (PP> Btataaf BouttCarotma.Odeoaf Bo* Ttfonmantai duaaty Control, Department mt Hmith and Bnvtfonmautal Control, aaoa Doll Straat. Cnlumtta. South Cwouna 3*901?' (dd)-lTT) .............. .. (CTT) Stats of varment. Affaney of anti* VS2Z&2?W"~ *" - --' (TV) coaomanvaatih of tlrsttfa. Tuinn tnaas*rn. mAairtsPsoemt mo-s- - -Oonorm Same " - Virginia 93313. A-3 SPl-05289 -- .raaMNMlfilb. sta(nd)arAdftperrestcoraibeedffuecntdiveer tdhaistepoafn.asno? owner or operator shall construct or mod' Uy any stationary source subject to such standard without first obumtof written igMn <u outtaj i after vivrf within ; approval of toe Administrator to accord* aaeo with asamottof' thgirsansutebdpebn,yactoceep,tPursedsoidreanst under ecuon il2tc>(3) of the act ourees, the construction or modification J5aar^^iSsy53Ls of which commenced a/Ur toe publica tion data of the standards proposed to | <LN iifpvnl by MuWamsr. lo.ii tottuoaowtaponincaePoens<aoadfacctbonedryiCaowte(ptvreaeua.e)prstnr*rese>drntafobtaaecsNnrniotau>dydfNmpeetrdr(ononaetpUrcssoipcisdrannAlotsutQoeidittritcoetneiowaocowfotaMarvnttpcnhwumnnyepbniiardoMersedilede1nsdoetgalretharra2btatrfrtaaobeotrtahyl2dototintyntrtso*knhe(ouhrhdb0eBospntsrtsiseeiaoscuorsfp.es(adodertpt3pchfdeiterspoffhaU)eyua,ebcrnAnausronclorsytalwasurbidhoto1dbfotstoicpmvteou1reecpgtuhhaaehrr3sedererrichernh(sceapfrtsetaedfeatiahunPeut.s!)aaoa)ttnitca(btet)nnscuetrr3odftejrotadtoiedaae.)ovpeupatdcvneiodotarnoslrttiaoyorarddrdufftrtattlhnaabon,ottatoatobhattijreanessenaeec-oyttimioatfsmaicomo^toatarrdroyr<nbattuianaad.eh,ststwptn)op(Ieyimtehoefgtaugi*prmdfecPrrUye)rsrioaswswotumrwtBtiuiuwtfioUvnohoiiectiaotiraeis)ttnfhrnheeenMnuinlt:,oictotnpanosooiAeatwupfntupotfiicndrpipo*pdmsfsotmrytUnuhueatoptnoocanJtcipvirstsbnohneofye--opvt_iasrirtpssaaintttu_eoiocohtoeusctgrc_clepuowraeaSatttlor_eiiayatttcooncIoiooron*aonneuaofnr-oyns,irtnnt1owpouhtou.staoorerc0rernftptom?rihoofnLodapwApatAfwfotedbeowehdaddaertesdrtimnaisimncmoft,ftatihthinsci(taliannoonnuiialn.od-d,aerobdar-o-ttteo.Jetteoodortdsmatesoptr<owssaoh((tt(oa,iw3iar3tst1ahUn>aaitpdr<)r)steli)rg,tlacte%/es,itduArtthitrtpsItM!indwhiahspaeniorbgbegatoTaanetgewnoyanstrmvcdhharsipi,thaightt.elaoedttaoryapooinaoycfreuppUcntednahhdoraeoscdtcednesewopoptaaiainodatrmsuorpbAnas.edcrtfrnUaoeetddaccraoioftcicotrdeemetipyhfprawiornddtttsiotoriiytcibronaetheaootdanraaeiodwristnndsacyrtofdtecttytusiotthohtaroroetstrpsiasmegbtou,,offrhonoeoutteue*onpteoaudtrtchrdhfnarewgdercptfhcef:eefteorrooueoortaacs,shn&.tntarrnttohMnoiaeidufotavuouiftetsteeneatroutwroo&ecfrccrtndppWneontetettettaiaaehvisf,atanrrIetieJsoViy report, revise reports, or report source (l) Notice to tbs Information and hasardous poButant tost results as reouhad undar ttds part OMnip <**, wtddb *Mt> istoBdad drtial la based, and (l> the avaraga weight pm mento to tot hasardous materials bring pfpreosrt W MM or opportwtfty tor mel> by too rmret. over the last 13 months maroroperatorto mat, Mtdb neti Mm Ualt m ttw BilHiUiJouotct tball wtmtOt. MtMoa*1 tBfnrmtrtm or tmmm. to ttar Adaimtntar pnor to Sul aothm on such application, <4> A ttul dowailiiaoan tpdaaaar imuiramai lar appmol 1B bo lo ntt. tar tad WinMttortb ttw apatiteoraaote ob which wfh fiatoal is beseff Such ftul dtosrmlMhrw wifi be mads wtthto to days to praaiofticc to *j***Mto"*1 infor mation or aiffuamnli. or to day* after toe deal data mantled for pramntw-- preceding toe date of too resort. <) A dewutptian to toe atostinr eon* troi eotupttaat for each emission point. hau(Ia)rdParnims aarriytuctoanntLrol dorioao) for each aMabbmbh>aauamIuownoMMDataOeirnMrrUyMoaamMatttsraomkll 4nlaM) Mr MB) totrttaUuubueoastnaltMnl tmO.i (T) AMUMMMInttemnarirwM Mr at tMaraaa a. rbatbar baaa mhiiOt Mtb tbt Bangarti gr--mni t Mttu.put MM w wt at tba MMaa* |fU7 A--Htasioa for aspcuvto to eoMtrwctie*e* medtAeaue*. (a) the owner or operator of ear ses source to which a standard pr--srtbsd to toobdhatepaornt iswehpicphttooaebinesttornaUettmprioorr tf no pr--nistimi is MM(.o..)MNMaatittMaorarratsobaoanaaaablaOnoMrMteba. tMaasats>Mntora (rnttac at approni ta cootract or taadlIt tbaU: (1) b&m as araar or aoMMar at Mao! maoculbmtr tar ermpiunor atca aar awatralUr pramaD at tbi. part ar at vo attar aasUeMa raOMit. Hut. taa(..r.a) t--rxanca--oa--tomaboUraMroPflpMarOaMtaromfoiamoiMiM- atwtb m ctanMra pr--nrtbaa mbOm tbte pan awr laanccl a wanar at m- pMnw wttb roeb rtandart tar a ptrtoa oat ocoooaat i roan Iran u rSMm 4a. bar MMOban ba ------ Matt m--*--* -------- Mtofieation |g to eoanasnoa. or arlaeal roaatraciaut. ar witoto to days after the effective dale htoasto--t-c-a--s-e--t-o--a- now soon* to** *--^ (J) rimst tbr Mtnibiirtratar fna larplaaiantlnt ar aatarclac tIM port or totta aar attar aatbai aaaar tba act. |ttn nMMOlMtlaaap __ lai Aar atRHraratantar at araaror --~TT'? **** Mob bat aa Mittal itortap attar tbr --2 tSwcttn data at a rlaadaM pcaMibad <B> aaaaMoa < a*Mt mm. . Mdtrtbs panMatt Itantab tbt AMaM- MaaMct r laataBrtaai at MOM *-- Mrator rritMa aatmactlan at tailaat: ** toMMant ar praam ebaoaa: CMatatAtaBmnoH)tkoatottnnnop<a4 ttbtbrrarsataMraptaotaaat ttatatt>arBrantme nbartintt,toabt ramatlmtanri,moimIiamat ant* tbos 0# dan oar Mar tbas M 4apa iqubimn ar praam MOdinmOtr ta priorUaacb data ba namptclt: and (* a aondtanm at tba aacaal data a?) oata br Meb (ml aoMpD^m * A-4 SPI-05290 tntar u u altsrnettre method lar bMnwium------- oarora sum to l M4t(b). tob(et)aeAbMdarstcdr.iptlao of totals oinltrtfm astral Dm which MU M taM dura* <ai fcniraton tMt> ud raonltortnx has b eceduoMd mu npomd m t*t forth to this port and Appmdlx B to toil a(MmMliCt.aMuMIaI.Vsr a ab m the(ewi aClhmoipwtrratodS. the mtarmattta sraetdod mOer paraymph ta> of UttoteeUaa mall ba prertdad to tht Admlatotraior MthM W dartfar men chans*. aswt that it eltanat MU malt Ins modiftraScb of the source. os debited m 111.03 (Ji, the pransooi of I <1.03 mu I tlM pu0t).1 TSt owner or operator at a saw aospoe wbltet to thto part. Md at the or operator o( to extottot eaane ob ject to tbtf part. Moll pnrtdt at eeuee to be prartdod, emuelnr teetlnt loom. tln(aI)aldSleomwpeU:Dt parte adesste tar test | OUS AraUMUitrat MarmaOtn." Tbt MtitoWUty to Um public of la* faomrmoarttitovn, pUrmoviAdretdicmtoM. otrrotootrhuenrdme*r oOtftu port thail be gomaort by Pan 2 of ttm choptor. ina<pd)plTicOa*bltea.nsat far Mporuot under ml^e(tJlhj)omdSsaelaepepaliccacasobUsleat sttoopieaeatlcuohsrmpeUecabsnlf.e.plot- S__ftr*..L. ,,tu_5__J_&_ this taction to tseludad u Appendix A af Uito pert Adriee on rcpartln* ttot statue ot oampltooco may M obtain<uins tfe* Ititwam tor esbipant *a< | Al.lt (tan ssabtoiTj. ta) Ths pnntolaus at thto part shall S. Tc..iaiim um C"tot* As am ee tmmOot --,l4 - a* ago as m tSMC.Mteu.VT sat ba cosetrusd to asy manner to pr*. abide asy data or poUUeal aabdtetnas thearetoAf mdompt:tn* and mUorctsf asy aaUa> |6X*1S W*wWi-- .___ i> imurinn taott mar bo vafrad Sss UmiOnx rodolatioB applicable to a uttonaty eaane. prortded that each (o> tooted ea the tofenaaaenpiwnaad Is oar realist asder I Sl.lt. or other is* boon vittbs to um Adsd&* Mrator it. m hie lodfseat. the source senimno Umltlsc letutettos to net torn atnsteat tbas the standards prescribed Uoc. Admtototretor meforms the y sreel a wairar at eempUeas with a atasderd to moottsf the etasdard. or It the aource tooponttnd under o wairar at aeapbano as(dSe)rtIhtouanplarirBtd the owner or operator (ram(or a period pot excotdlnx 2 years the(heidsSeatitroawdaatiaraartweiallebbesitnanedrratrtda.tt ahd at hat TCdtieatad a wairar at eompttaaoa. Cb) a applloatlon tar wairar at the oBdndcB toot to BMtee oppUcottoo at a itstimer? source, other than a steUoaory eouroo owned or operated by the United hulas, to obtels permits. Uooosee. MB(\1:)i Imdi m--u--t-ywf|lJf)ld*.mwtfw the twametwattam dOtO Of the wairar. The welter say be tormi. flbOU confllBI' tbt ISfaRBlOOB Ike aalnd by l U0. The approprtese farm to contHnad m Appendix a to this part (e) Appnral ot any wairar sraatad --to thto aeetlas shall not abro or oppraesto prior to lsitutlne construc tion. mortlfloenon. or easnxtou at auob taea at om/ ** 1 salad at as earlier data It the craiitltwis ' `^rarae* seitliuml^ VMtof jc.Mtsit*M Wiirtnm usher paridnph (b) (SI at thto aae(1tt)aa(pareerotlayt sdiatet.s bp which etepe to* wairar. (web rannetunnn wtU ba made weed orennHencc are be be tabes: tad tapooe additional **^0^ u tdo Odmtmstrqlqr dtteralam to be seeee. oolr attwr sauce to ftna to the aaaer or operatorot theeeuroa tdt.IT trail. - , Mo owner or operator subJeet ta ths prartoiaea of thto port abas build, tract ary ary Is secure ecstrots Installation at the wtthm the waiter SpseeMonecse.' end is secure eraseoar at the her'*' MmenstaiecastattsaeitOWne A* Aet as amatool MpaaiwMdarfhaeuauiOnctbmp&h.nrsanaoOpnPct epsQrotoMooarna0.8eom?rtmoOretrctthofocwdll ,htifhcffhevaw*1o*tu"al*dt at pertaes carats the wstrar period. (e) Mar to denytsf oar remeet lor a wstrar pvrsusat to this ssextan. the eipthpsUrwgllsUsi oOcugefltOitepgta. a elalaOUOaDsOMatllMaPsi | *1.14 Saaeaa wa rad saatyMai erath" boctadoA but to oat limited to, tbc uas at Admtnletratar win notify the owser or operator mehtotf eucb raaaeet ol the Ad. ssuetnbors tsteeaoe to tons each Omtti. ... tteorontsh*erawt itthhe- info-rm- e-*--tlo n4msa|sd| ado. <s> Methods m. in. tad 104 m apemtdtt a to this--p-a--r-t-o.b-o--SrbartSnMoetentaoltr. rataoui dOatanta ta actalara oooipllaiwe wttb a etolbto em--t-o*a"tca"e e*tcabs# dUa)rdM. MuAd* at Mo act A-5 SPI-05291 far Map! IA14B (i) nui asapart appum to Vhld>(R4lM'. d) tVtfUn* tfftebiartt* to rmtrtm C (1) *Ba!k mu' Tutat * mm which B produce* br t polnnermtlon pracam I *14* . _deader* lor etbrWae dbbUrede plaaliue. ." m which as water u uM (1) "Inproeam waatawatcr* maaaa aar mar which. dunne manufacturini or amt min*, aaa late direct entaet witt marl chloride or pairmrl chloride or rerelte from the praduetlon kwH to >*w material. Intermediate product. Bobbed product. br-prednet, or wnte product coatamm* rtnrt chloride er potrtmrl chloride bat which baa aot bean Cbcharped is a waatewetar treat- la) Ethjlene dichionda permeation: The ceaeaBtratioB of rtnri chloride ta all rrhaait loam diacbaraM to tba atmoophata tram aar aqutpnwt uaad m ttrme dlehlonde purification a not la exceed 10 ppm. axcant u prodded m 1*140(1). Thu taatumaant deaa aot applr taaompmnt that baa been opened, a oat a( caarebon. aad amt tba reoutrrmam ta | l4*(b> <> U> baton bam* ib) nil eobpert Boa not applr ta nulpmant taod ib march and dotelewaaant tf tta meter uaed te palrmartae tta marl eaiaMda emM a>she eatdaaant baa ni at ao man than Ula'lHM). <e> BaetlaheMttbattpart otterttaa I*141: *144 liXl). <b). <c`.u4 (4); *141: *141: *14: *141: in* 11.71 o not ibpIT *o agtapawat aw* In march aad dcreloptnmt a tb< meter end is aainanae the *mrl chloride mw* la the ecmpmant bii oapecicr at Bitiir thee`a.i*t (m11***(PtoBnl)l) aad ao Itm on* a thb aokpart an Mas* it the Act, ib Bataan * it ttb part. *r tnthUaeetiaBic toUevt: amt ptacaat or *behara* uatnata* aa waetowater. <ki "Waatewattr Bwetateat preome* tBdudaa aar prorata which modOct cbarecteratua cuch ar BOS. COD ***. aa* pH. tauaBr far tha porpoac af motto* eaueot ruidatlnaa aatf ataadaida: It data aot Include aar pnaam tba popaae at which u to romor* nnrl eWorld* tram water ta awat regmremanti aI Bill (1) "In atari chlartda aatrtea" mmm that a placa at adulptnnt acataiaa er antaete attbor a baud that a at baat 1* parent Or watebt Marl Chlondt at a pa> that a at laaat 10 parent br ratoma mart chloride. <m) 'tteiulaid ataratttit rtwailm >" maanc a formal arum prooodura ottatattr adopted br tba pint ownar ar aptntar aad arallabia aa a routtor bnb ta tboaa paraoca raapnmlbla tor aanpm* <b) OareblormaBoa meter: Except *"**,* !-*<> ambelene of marl ehtetda m tta atawapbere from aach oxyehlcrmatm iwatiia an not to anaad 04 /k* (C.eooi am at tha loo parentathrln. dmiartdr product Iroa Mt-M. Falaalia Bilal far atari liloatdi piaati Ah owaar ar operator of a marl chlortda plant ahall coraplr with tba raemramte af tbb aaotm and | i.*e. <a) vmrl chloride formetied aad pntfBeatm: The cenentratioB of marl abiortda m an aabanat ama dtecharred *a tba atmoaphare tram enr aaidpamt on* m marl chloride formation and/or wirtBcatm a not to axeaod 10 ppm. naapt aa promdad In |i.M(a>. Tbb r*atamont deal aot applr to aanlpmnt that Baa bon opnad. b ow of cpartBan. in mmi Bichloride pi*nt~ te> etuaac nr mat M prtidiiiae r ' eae dlchlartde to reeeUoe at am end mat tba mtdrmaat m I ll.aoib) B*(dbw> ae--n chl.o.r.id..e.w..i.th al'ume aa* mat Wc* pradeam ikrl ahlattta triuima <e> Tctmnrlchtorm plant*______ Bar pint abort rtny! tBMB Han er la wanbltiettan wBh MB* mtMrble b aftbbt (*) has ft prob* that ___ am* latarttce m a Mm* er treaaftr meal tad BiiBmn tta Imi M marl man*! ib tta* m* kr the mm mu <f the BBtarbl ia* mm BO- ahlonda formetam aad m which! (t<r|l)eHhkanMdetaafprtaaddoMeaiddi.i **t mm h aeb mam cbleetda b partaOrar aamitr po<trtm) a--rtaad Into potrmayl chloride. of ttao* eblcrldo aaeaRM* ----- b anaad to tho <a> applr to i 1) no oOBerntratton at marl able* Ma m an aabanat caaat dbebanad m the ettntaphort frcca tech raaeur a not toaagoad 1* ppm. anapt aa provided m para*ra*B lax*) at tba aeeuon and I 4aia). <*) Tba reactor apnlnr Iraa fram aach nacter b aot to meaad *4* ( atari aftantattt <040001 ib marl chlonda/ > af pelrmar) eblortde product. wh if) -Ondo of dtmakmef raam Mk iwaa. M tba ahatr form br tho an aim aad/ar taanmta tha earn at mmadaa an aamal tta product nwannlnail n a drr aollda name aa raaummnt apphea to nr natal which b and ac a reactor ar ae bb a motor and a atnppar. Id tba Mi preen, tta product arena tta man product at prapolmcnoatloB and mote rebdaal marl . . Mama* tba pabmertm* tin bap m tha pint preen Hew. ft) "Otaadar* tamaretare-nreua a tUBpowan a* ** C t*' B). fa) "Bandar* pamtore* mane a prana ar IN mat M a* (1*4* m. r <*) Mnoai rent tain dbebarec: Bx> tapt fern bubano manual tnttalre Abebarea. than b to ta ne dbebant to relre n a potmen cblaclda reactor m mnrt chlorlda aanha. An ammner areal rent tain dbebarea mana a am bat* bare aaatdad kr term* taaaerea to preretit tta dbahart* Wltbtn l* A-6 SPI-05292 dan at any dhetmrn to tin aaDomhmv tram any nral vnt valve. the owner or operator ol the souree from which the discharge ocean ahall submit to ttM Ad ministrator report id writing cantatahii talormatlon on tns oouroe. nature and came of the discharge. tbt dot* ond ttsM of the discharge. ttM approximate fcotol vtnyl chloride loss durtn* ttM dMcharge. the method used for dttermintag the Tiny] chloride lorn, the action that os token to present the discharge. ond massures adopted to present future die* oburfM. <b> Stripper. lb* eeneentrotton at Ttnyl chloride In ell exhaust gases dis charged to the ataoapheie from each atilptMl le not to exceed 10 ppm. except a* provided In I tl.eaio). Me require ment dee* not apply to equipment that hoe bean opened. Is out of operation, and met the requirement In I gl.M(b> <> (I) before belnc opened. <c> Mixtot. wclyMsy. and fteldtay aeatetnert. The ooneeotratlon of Ttnyl chlorld* m all eehaaet exes* dlacharesd to the aonoephen from each mlxlnq. wemhlng. or holdlne eontatner In Ttnyl chloride earnot which pracadm the tripper (or the reactor if the plant baa no etrtpper) la the plant praam flow I* not to exceed 10 ppm. except aa provided In I (l.U(a) - Ms requirement does not apply to equipment that has bean opened. Is out of operation, and mat the requirement in | glAi(b) (> U) Mtoee betas upmiad. <dl Jfoaomer recoeery system. Me eoeuonfritlon at Ttnyl chloride In all ahatat (aeee dtaohargad to ttaa etmotpfiwe from each monomer teeoeeiy entam is net to eaeead 10 ppm. except a* pen mm m | tl-t<a> Mia nquuatawt the requirement m f 81A4(b) (> til beferebelsd opened. addition to ftrlpptnp. mnlmkmt af ttnyl from the slip rauet tnraupb a control chloride to the atmoopbere Buy pot system from which the concentration ol vinyl chloride in the exhaust rasas does (I) 3 p/kp (0.003 Ib/Ib) product from not exceed io ppm. or aqulralem as pro the (tripper(s> (or reactor(!) If the vided Is 181 .M. plant has no atrtpptr(t) ] for dlepeielne potyrtoyl chloride rmtae. emdudlnp lata md(3) Leatokargme firlo: m pump, oompramor. resins, with the product determined on a (1) Roiaanr pamps. Vinyl chloride dry aolkUtaHs: emlielnns from seals on all routine (II) a.c p/kp (0.0004 lb/lb > product pumpe In vinyl chloride semee are to be from the etrlppen (or reactcr(e) If the mtnimleert by seelleti pumpe. plant hae no stripper (e) ] for all other polyetnyl chloride mint, twinning lata reslnx with the product determined oo a dryeollda baela pumpe with double mechanical male, or equivalent as peevtded m IPI.P4 If double mechanical seals art used, vinyl chloride cnucttceu from the avals are to be mttiimlaed by mairtaltanr the pra wn between the two seal* ao that any leak that occurs is hue the pump: by HAS Ei dicMeride. i sssIsiS far eebylene vmrl eklacide ead paly- duettap any vinyl chloride between the two seals throupb a control system from riayl dslaeMe ptaou. which the concentration of vinyl chlo An ownar or operator at an athylmw (Uehlerlde. rlnyi chloride and/or polyTtnyl chloride plant shall comply with Ua noulmati of thk mcQcsl (a) Astie/ eales dhekarpe. Except for an mnaipency relief dhcharge. them la to be no discharge to the atmosphere from any relief calve on any equipment m etnyt chloride samca. An amagwey relief dheharpt mean* e dMchart* which could not have been avoided by taklnp measure* to prevent the dkeharge With in 10 day* of any relief vale* discharge, the owner or operator of the eeura from which the reltat valve dlacharp* secure ride tn the eehauet cases does not ex ceed io ppm: or equivalent as provided tn | pi.ee. ill) AacigroeetMg psnssps. vtnyl chlo ride missions from seals on all rampreenttag pumps m vinyl ehlaride service Art to be ainiaimd by outboard seals, or equivalent as prevtded tn III.M. if double outboard seals ere uood, vtnyl chloride amiastons tram the seals are to be muumtcsd by melntamlnp the pressure between the two teals ao that any leak that occurs it Into the pump: by duettap any vinyl chloride be tween the two seals through e control oatam from which the concentration of port In wrmap --InfniiiMimei an the aouree. nature and oanee of the dheharpt. the date and tune of thedheharpe. the emu inliiMle total Tlnylehloride lorn durinp the dheharpt. the meth od imad ter determlnmp the vinyl ehloride lam, the action that was taken to prevent the dheharpt. and ateaaara vinyl chloride In the gaem dem not (weed 10 ppm: or equivalent se provided ta IP1AP. (Ill) Eotgtaey cempramor. Vtayl chloride ecnMebew from seels on all ro tating nwnpraeon ta vinyl ehlaride mrriet are to be mtntmtiwl by tneuiltrip ooBOrMBon with doutali mehiBieil male, or tqulvttaBt as provided WI Pl.dP. (b> wa(sia)i n double meebanlcal seals an naad. vtnyl chloride emhebew from the seals ere to aoureaa faDowtat the atnppvu) [or the anisna* eiiiiweeu irom emnmp ami milaadlnc ltam ta vtnyi thtaeldt eerrlue which are opanad te the atmamham af- an bstutmi the two sash eo that any leak that ocean Is Into the ninipi team by ductsng any vtnyl chloride ktteqeri reactor(e) if the plant hae no atripparts) 1 la the plant praam flow toetadtag but dm llmttad to. resiti ifupes oonuesiimat. Hand tanka. tutors. dry- anfl)taAbflehbmihnhwnhiltaIfrttaaitalMfainner** the two (sale through a oootrol system from which tbs eanewtrattcB af vinyl amwekMnttiaoMnMatantdWwbheffoenretttsnpaatBlMacraaa|MloMadMin*g* chloride ta the --cases dam nee eireeri 10 ppm; or equlvmlens e* provided quantity of vtnyl ehlaride la all parte of talR.PI. each loading or unloading IBM that an (tv) Bactpraettag oontpreseon. vinyl ta he opmed te the atmnwhan h to be chloride from eeah on an n- fadnead aa tfait ttM nni AonbtDBd aon* chloride ----- the eelilitesl assssp tata oo aieeter than coosa art (0.U fr) earricq are te ba nrmunlmrt by WWfling raeidual rinyl chlorld* ronnmwiumm m all pndm ad polyrmyl ohMrld* ram :and pewldad ta I P1.P4 If double outboard (It) Anar vinyl ebleridt i seals an and. vinyl shlorlda cmbtancu A iMtftW AT UDlOldtaff M** is |MDrt* from the aeale an te be mlwlnUnd by mm vttfa pttMnph (b)(1) a> of ttato --w--w the ptemure between the etM to to bt Oku* tbroatfa muni two seals eo that any tank that nocun k system from which the niasicnfntlnn at tata tat oempraeor; by duettar any nan chloride ta the i vtnyl chloride betwemi the two aeale net eeeied Upp through a oantrol eyetem from whtch the eidad in I RAC. concentration af vtayl chloride ta the ) SUpguam _______ eqhanet team dam dot nceed 18 ppm: ta; or taedtai npevetlcnt. the vinyl ehleclde or equlralwt as provided m I tl.PP. <3) In iMlyebiyl chloride plants sen- wltpkew from each attp gang* la vtayl (V) Agitator. Vtayl chloride wrmnne mutag Ttnri chloride ahmtaM wMta chloride aerrie* an te be minimInrt tap Item male w all intainri ta vtayl chlo taabBoloqy athsr than strlppme or m any vtayl rhtartite ilta harped ride semee an le be mmimmed By m- A-7 SPI-05293 --*-- akttaton wtth double morhonteal ttelt. or equireiem at prodded IB IU.M. If double mochontcal eoett ere uood. rayl chloride omintone from We wit an to be mtmmtoort by maimainlnp toet ear task tbat occurs B late ta epltoted remet] by daetat tar rays ehlo(Mo bbtvwn tot two awB throukh > eerttrol xjxtam item which tba ecneeotntuc of rayl chtorMo m tot exhaist lank at pweMtd In 1ti-ec 14) LeeJcooe from redo/ mJmi. Vtayi chloride wntwbuw duo to Bakapc trow oocb tenet ralee oc equipment is dart chloride Mrta an to bo talnlmtood br **-*-"`-- a rupton ditX between tor wapaont aad Dm raUaf (alto. Or ooaaactlBt Wo nUtf wire dtochorpo to a prumnUnoorrocorwry orrnm, orogwralwt at proratal a t f-M <) Kamel oratMo of KM. Inapt at prorldad a |lJ4(aK). aO (toot Baled an manually Teotod tsoa eqiapaent a ran eblerldo aandoo art a bo daetod tdroued a eoatrol intern tram Wed tdo oeaewtratlaB of day) chloride a tdo oabaan caoao dote not tontoil to ppm: or oquirmtant at prorldad a 111-04 (*> OveefV of ooWpawaf. Vinyl eblerldo ornament tarn oponlao of eempnent (Inelodtnf madmt or tmloadtat liDto that an not opoaod to the rtaoapbaro after teed Inarttny or tmloodfeit opontaor an a bo mmtmloai at (i) Mon (Mo B a bo rodaood ao tbat bo equtpamt oontaiw ao aero than tdpmont bp rahaao rayt chlorMo or JM* m pal) at ran edlondt. whhihora B luwr at tm day rap! tbletMt neh (tblv)(o0u)s<bD %of(ImhBinit ] wfcfcfr tte edBMftlnttte rtot te tbt csfcfttut i cfato* I MM. (3) fpatpte. Ban pardons at taw* pita centamtnp at Beat It pmtt Ip obtw rtayl chlerldt an to bo totunwi w d {1) It 1 Miiibk tad iccuftto marl chloride monitortnk nna for de tection of major lwba bad laetmflceaoo of the ttoan) m of tor plant when a Bek a iBwnd A rlnrl chloride mocltcr- tar --"'t*-- from ot* or mar* potato on eoatteaooi mpaontto) bons ind on** 2ysot tbo imptoi with (u obroaftfioc* rapahanpontrtoaoarmyoteblrt.caryhUt.loadtrtliooldomeh,oyewdwlreinqthueerqIdnrobefrrtqaeuocrprotiodomrnmmt.atoootcrurtr*raaoda-- waWaiaataaltaraatfro method. (U) ft aebMoo o mtahB aad aoeana portabla hydrocarbon detector a ba met nottaaly a bad maS laakt aad to Ptapaat tht aiajor lotH latbrotod by tba rtayl eblorMo PMaitoRa* amtOBL A portabla hydrocarbon detector maaao a dortot which moooum hydrocarbons rrtth o aanatttettr of at Beat tb ppa aad B of saeb anion bad Ba tbat It oaa be aaad to mattun omltalono treat Ideal* (tt) Itprondaferae onorfMacaU- bratten aad mtimantnoe achadnlc tar the rayl chlondt monitorial mttm and portable bydraoarbea dotaetor for the rtayl eblorMo montterat ayataa. a dally apaa chock a to ba eeodoeted with a concentration of vayl ehlondo aqual to tht cancantrsUco defined ao o laak ao* eordai to perairapb (b) (t> <rt) of tbit Motion. Tht callbraOna B to bo dtao adtbtotbar: _____ pond txta tbo taoto tpoddod a a 111 pad 1X3 of Twt Vetoed lb* aad a accordance with aaettea TJ of Tea MBedMd.tr>* i of tot oaMrtttan toadard B to dan aawtfaetanr. The bpta neownnandad a for aaob cyUadar to tip IBb daw pat ahaapt ! Ufa 1 mob taxed to Wo eyttaltar t nnr, u |M wn Bm rayl eblorMo teetod <n> ft eontaint a dednltlae of leak bleb B acceptable whan compand wtth the bactproimd ecataBOmtioai at rayl chlorid* a tho anaa of tba pleat to ba monitored by the rayl ehlondo mcmltorIna ayoum. Mtomrtmontt of bockmuBd eaneaBtruioBt of rtayl chloride in tbo aroat of tot blast to to aonltortd by tbo rtayl ehloridt aanttortac ayatoa an to to taehidad with too Caacrtpticc of tht Pfaftam. Tbo dadattMB of leak far a ftrai plant aay raiy aaoaif the differat anaa wttha tba plant and B alto to cheapo orer ttaa at baekeramd conowtratione a Wa pleat an reduced (> faproetn meetemater. Vinyl chlo ride ammxni to tot atoMcphan tram tnrmnaaa waatowator an to bo foducod oalaUom: <li Tbo oorewitmiflB of rayl chlo ride a web agirottat wattawatar atnaa ronmntnr amtor than 10 ppa rayl ehloridt aoaaond tnanadtataly at It taarw a place of oqtapwant and bmort beta* mixed with any ether intirm om waatowator town B to bo reduced to oe won than 10 ppm by walpbt hofon beat Bdmd wtth any other inproctrt waatowattr stream which coatauw Ian than 10 ppa ray) chloride: before brink expand to too otmoohort: briar* beInk diechorptd to a waatewatar tnatatat proc* aa; or before boat dBchsrpbd untreated at a waatowator. Thie paratnph don apply to wptar which B wad to dtaploro rayl ehlondo from equipment baton it B earned to too otaoopban a oaoordaaco with lixtotaxi) or aamcriBb (b)(1) of tbit aaetton. KB doat pot apply to wator which Bund to with act oculpmhde|Mot Mawftenar ^totatw equipatmatoeanatatahdoPttaatlRiwWdyw ecrdaaco with (lldtiaxll or pqroinpbfbXf) of thB aacttoB.3* ttM<ut>BpdfapyMrfttaWylftfctWhloWrMfttoV 1nfot aftnOr0oOdfttrfta&wflft with paracracS (b) <01 (I) of thB aaettea B to to ductad tbretab a control am from which tba oamBBtaattaa of rayl chlondt a too wbaaot paote dew oat wamioai.da1t0. ppa. or aqtoratoBt ao proeMed Cc) Tbo nqqtnwaata a puapapbo (b)(1), (b) (ii. (bxo. (bx). (bxi) aad (b> ((> of WB taction are to bo a- wrparatod ate o ataadard oparatat i eatrai www uea w meme wMpwtBt a rayl tblotMt ttrrln a be bibibbbpd dr InMdnlWt aad -- or evee w doosraoe Motood lot. Tho Bdataa 1X1.1 aeaadt lUU of Taot _ Id tar nrttoooaen of eytadw i trda tad far totobld' _____ &*"** *>..** >* -- ___j B to ba tdd tehodelod ml of dr tot _______ a prwiaw arm ba _"*" r prortoed bo M4 IP (frlTtotoeMtatadBaBbtrafpad to to menttorod tad tbo fnmaaop atadatao orondtd tar ta tho pnpn on imiwtotli wbte Wop on copppor with tbo saabar of paaow of opnluiua a rtoyl eblerldo aorrioa aad thoaBo a) BbyakallbyealofWaitaat (r> B aaaaaa aa aaowdatw pita watee ta ba tohw ban a lab a a earl tokcMa Ja oquaunoat wo.n a* tlJMgeJ) la raluar for which aa waioPBB bait B prwathad a I L(h> () (11 pater to maito toa rqiiiimaht aad oaPM Twt Method ltd. a portable hydiwaartoa Catoter, or aa aoptnlaat or eltoratora motood Tbo aotoad of mooewowait B to awe* tbo imoMwou a I 47() (0) (II (A) or (pH (0) (1) (*). MmoIHkCdNwUaiiwb*dtddbabW A-8 SPl-05294 I 61.M, EqaWaie aqaipmewt erimi m- Open written eppllcatoon Inn an own er or opentor. the Admintotrator but approve use of equipment or praeoduno which bon boon demonstrated to bis satisfaction to bo equivalent to tamo qf (educing vinyl chloride otoiaalona to tbo atmosphere to tbaot prescribed tor compUoncc with o (pacific paragraph of tbio oubport. For on or1stIns source, oaf raouoot (or uslnc an aqutvahnt method os tbo Initial measure of control to to bo submitted to tbo Odmtntotrbtor wlthm 30 dors of tbo affsetlet data. For a naw sount. any nouast for using an squlvslant method to to bo submitted to the Administrator with tbo application tor approval of i minis I Inn nr ninilinnatlnr required by 101.or | dl.6? Eadsoion soots. (a) Cbtosa a waiter of omlmiaB taottna Is obtained under | ei.U, tbo owner or operator of a source to which this sub- part appllaa abail test entlosWms from tbft MUTC4. <l> Within 00 days of tbo affective data in tbs case of an aatottng source or a new source which bos an initial startup date rrsrsritnr tbo eSeetlee data, or (2) Within M days of startup in tbo caee of a new taunt, initial startup of which occurs after tbo affective data. (b) Tbo owner or operator shall pro vide the Administrator at toast SO days prior notice of an amisoton taat to afford tbo Administrator the opportunity to have an observer present during the teat- (e> Any omtoalop toot to to be eon- dusted while tbo squlpmmt bam* tooted to operating at tbo production rate at wbicb the oqutpmant win bo op- aratad and under otbsr ratovent eondl- toons as may be maeUMd by the AdmbUa- Detor based on rataseenWtlvo perforta- anee of tbo source. <d> [ltsasrvodl* (a) Whan at all peadbie. aacb ----r** to to beanalyaad srttbin 24 bom, but m no cast m rvnsas of 72 boon of maple eoboetlOB. vinyl cbkwida mitoaliaa am la be datwamsd wturin M dam aftor tbo smtomon toot. Tha own* or operator Administrator by a nats day folluwlm tbo t (f) Tbea at tbs plant and i request, far bnaittoei by the , tmlor. fbr a mtntanm of a yean raoords ofjsmjadei.............................. grounds to dispute the results obtained by an equivalent or alternatlvt method, be may require tbe urn of a raference c,,,IC.Il-IOl^lO-l 1100) method. If the results of tta reference and equivalent or alternative matboda do not agree, tbe rosulta obtained by tbo reference method prevail, and tbe Ad ministrator may notify tbe owner or operator that approval of the method previously considered to be equivalent or alternsuve to withdrawn. vinylC-s=kg colon(U loc kg product. CtsTbs osaqaotrenon of vinyl ehlortae sa murif By Test Method 100 aaOsDvaklty of vinyl chlande as oas resa nsatosaepben sad 20* C la kf/m' QsVolumstnc flow tn '/Or se a*, (I) Taat Method 1M to to be used to determine the vinyl chloride mriim from any source for which an emission bout to prescribed in II l.S3(a> or (b) tarsuasd by lafsroaev Matboq a of *o- pepdix a to Part so of tea chapter ta-s Oobvarcoa taesar for ppm. gspsooucuea sou (kg/br). M I ll.gg(a), or it gl.64(a) (1). (b). (c), or (2) Taat Method 107 to to ba mad to <d). or from any control system to which determine tbe concentration of vinyl reactor amissions are required to bo eblonds m each Inprotess wastewater dueled m l gl.04(e> (2) or to which fugi stream for which as *wtavu lumt is tive amlesions arc required to be ducted prescribed tn I 01.00(0) (I) (!). to |gl.tt(b)U)(Ui. (b)(2), (b)(1), (21 Where a strippins operation to (b>(gxui.cr (bxyxui. aaad to attain tta amualcn limit in I i.- (11 For each run. oat aample to to bo ita). wntoslona an to b* datarmlnad collected. Tha sampling site to to be at using Test Method 107 as follows: toast two stack or duct diameters uown- (I) lbs number of strippers and sam seaam and ana ball diameter upstream ples and tta typm and grades of resin to from any flow dtoturbanca such as a be sampled an to be determined by tbo band, expansion, contraction, or vialbis flamt. For a rectangular cress section an equivalent diameter la to be determined tram tbe toUoerlng equation: equivalent dlaasettrwt-^g. ) Administrator for aacb Individual plant at the -time of tbo tost baaed on tbs plants operation. (II) Bach sample is to b* taken imme diately following tilt stripping operation. (lb) The corresponding qugntlty of Tbe sampling point m tbe duct to to amurlal processed by aacb stripper to to be determined on a dry solids basts and be at the centroid of tbe ernes section. by a method submitted to and approved The mmpto to to be sktraetad at a rata by the Administrator. proportional to tta gaa velocity at the (iv) At tta prior request of tbs Ad mmpltng pomt. The sample to to be ministrator. tb* owner or operator shall tekm ever a of one boor, and provide duplicates at the ---r^ re to la contain a minimum eotauas at H quired tot paragraph (gxaxi) at this Ufa uwiaclpd to standard oondittoua. (it) Bub emtoMoa Mat to to consist of thrcs rasa. Fbr tta pnrpom of itaurmbi ine wntoalntii tta averse* of seaulto of an runs to to apply. Tb* averse* u to b* (4) Where control technology otter than or m addition to a itnpplng opera tion to ussd to attain the --*i-- limit ta 101.04(e), tmlwinm an to be dstar- || (oDofl; wenpuled on a Urn* weighted baMa* (lb) For gaa streams containing more than |0 percent oaygan tb* eoneantmtton eg vinyl chloride as datarmlnad by That Method 10# la to b* corrected to to percent oaygan (dry basis) for determi nation at wntooiiwii by aWna tta follow- (1) Toot Method IN to to be uaed to determine atmoapbaric miatnn. from aS of tb* procam equlnmant siimiltaii*oualy Tbe nqulramanta of paragraph (g) (1) of this section are to be met. (ill Teet Method 107 is to ba mad to the concentration of vtoyl C IIMIIIIIfliftsoJ-pioa m each Inprnnsw wastewater subject to tbe amtootim limit nreaeribad m I l.S4(e). Tbe maos of vinyl flbisrido M kg/IOC kg product in aacb loprcewn wastewater stream to to be de- tarmtobd by nnng tbe following aqua- ic,gio-i ttool Csew------ ---------- (g) Bntom Mhsrwlos sportotod. the owner or water obeli am tsM Tam Methods tot Aopandtx B te ttto pert far mefa tam m rctuirad by pamoapba <g)(l>, (g><2). <gX>. (g><4), and (g) (g) of this sort)on, untoaa an eqnteelent method or an attarnatiee matted bat bean approved by tta Admmiatrator If tbs AdmmtoBator fintt rasanewbit ttt waimtsn limit to prmoriaed m Ur (0) Tbo reactor -p-->* tom for which afmssitattmttaaeooMntraltaa.auim an amlinn bmlt is preaarlbad in I gl.04 mono m kg/100 kg product are to ba ta) (2) is to ba Oatarminad Tha number ileiwmliiiil by oong tbe feUswhig equn- of lesctars for which ttt dotoraunatlon a-a SPl is to be Bad* M to be specified by the Administrator tot each individual plant at tht time of the determination bated on the plant's operation. For a reactor that u alao used as a stripper, tbs deter* mutation may be made Immediately foltowlni tbe itrtpplnq operation (1) Except at prsrtdad in panenph ()() (U) of this eeetlim. the reactor openlnt Mat Is to be determineil Hama the followms equaaon: C_m-w---u--c--o--i n(to----)--(-C--*-) amC*--DkCesseetsteestlrrtircshisrlsimtssMrmitlceeeimsessieIBseih*t.see Mssmbee set Fl--NSedrsseIBs'esitmtrtbaastsoape|irrt^tsarseseaesrst(listlistamtsStaalmsasaspascMBtaeib--thsmiiarsorSttbeyneeasSrssIlstsssssafssea.tMeblasaesststssisOnlbsasbatatsrtssealliteyttreii1beet0atmlsassttiseseWliaeSss.ir (A) B Method 1M U taed to detarHat the cooceotratloB of rtnyl chloride (Cb), the aa--ipia is to be withdrawn at a constant rate with a probe of sufficient tanfth to reach the tssssI bottom from tae manhole, hemnlee art to be taken for ( mlautee wtthm mehee of the nstel bottom. 1 mlnutee near the rated eenttr. and S mlnutat near the metal top. (B> If a portable hydrocarbon detec tor It used to determine the eonotntratton of rtnyl chloride (Cb). a probe of sufficient lencth to reach the mmol boatom tram the manhole le to be mad to i win be made wtthl^l taebm anm* M|| ffiJra th hn |A| ffit Mk lOBftllflB until the rmrttnc It tiehllteed AO hydrotarbant measured ankle ammntd la ba rtayl ehldMda (C) The production rate of potyrthyl chlartdo (Ei b te be determlhtd by a method submitted to end appeomd by the of eraeimttoaa. the racism Involved. and the TOlume of iai tn the reactor Is hereby opprorsd by the ArtmlnMtratsw at an tlteraattre mathod for detsrmtntnr reac tor opaline Mae tor poetpolymematlen macton Is tbe manufactura of bulk reatna (IdaeeaclxitcrMaimaeu.OMpe*n' (b AM maM ' on a acnthc t tar whkffi mimm limits an promrffied m t (lC3(t> and (h). I ClAlla). and t UA*(a>(l>, (k>. (O.aad (d). and matoas in rtonmd to to tfnstod to ( lJ4(a> (2) or to vblcfc futottw Mtotans an icdulrad te ba ductad m 1 me (bxixil), and (bxs>. cbXS). <bxi> (bxti. (b)(7). and III III an amf (U).and ibxexii).M knplememed. (b) The rtnyl chloride mcnltorhif eye- (b) d) Is the earn of cn exJstlnc tamit) used to meet the requirement in source or a new eource which he* tn perecraph (a> of this section is to be a kittle 1 startup data preeadtac tbe effec device which obtains air simpsIs from tive date, the statement M to be submit one or more points on a continuous ted within *0 deys of tht effective date, toawntial bull and analytes the nmplm uniats t waiver at compUcnce b (ranted with rat ehramotocraphy or. If the owner under I (1.11. alonf with tht informa or operator assumes that all hydrocar tion required under I (1.10. If e waiver bons measured ere rtnyl chloride, with of compliance Is created, the statement infrared spectrophotometry, flame Ion It to be submitted on a date scheduled detection, or tn cqulraicnt or alttrna- by the Administrator. ttre method. The rtnyl chloride monitor- (2) In the cue of a new eource which lac system used to moot the requirements did not hove an Mittal startup cute prr- In | ciJS(b) <l) (1) may ba used to meet eartlne tht effective data, the statement the requirements of this section. It to be submitted wtthm (0 days of the <e) A dally span chock It to be con taittol furtup did. ducted for each rtnyl chloride monitor- (c> The statement le to contain the Inc system used. For oil of the emlseVin toUewmc tnfoemethm: Bournes listed in poracraph <a> of this (1) A list of the oqulpaunt metalled taction, except the one for which in mole for compliance. skin limit It pneerlbad In I (l.(3(b>. the (3) A description of the physical and dally span efaack n to be eenedueted with functional characteristics at each place a conctntraUon of rtnyl chloride equal of equipment. to 10 ppm. For the emimlon eource for (3) A description ef the methods which an minimi limit It prescribed m which hare been Incorporated into tbe I 61.(3<b). the dally span check It te be standard operatlnc procedural for meat- conducted with a concentration of rtnyl urtnc or t>*f om>tiiiwt for chloride which It determined to be equlralont to the emMelon limit for that ffiOUTO taMd OB th* anMm toft It* qulrrd by 161.(7. The caUbrattcn It te be dent with either: (1) A calibretton pat mixture pre pared from tht caste tpoclftod tn sect)one which.wntoloa limit* ere prcrtbfd to II1.C5 ibi >1X1) and (bx(xii, (4) A statement that each place of equipment M metalled end that erh Mace of equipment and each procedure hbatacueed. 1X1 and (XI of Text Method lee cad Mae 114 at tbe Ann Cb cm m MmCed m accordance with eectton 7.1 of TUM <4* OXC T414IJ. 4M7 Mettnut 1M. or* (3) A calibration cat cylinder standBfri MBfffitffiiffid ftkd dBffiMffiMdftd Mffid^ffie trattsn of rtnyl chloride. The me comat thf efUtoftton cm evtoto ctenderd It te hare been oartblad by the | (1.70 ............................ (O Tha owner at operator of any couraa to which this eobpert appllm ehall manufacturer. The manufacturar meet submit te the Administrator cn Septan- bar IS end Much 1( of each year a report lltt tor each cylinder ao that the oooemi- m wrtttnc --'Vivryif the mformatton tratkm dote not chanc* creator than ml pcromit tram the oertffiad raha. The data at cat cyMndtr preparation. mnmort rtnyl chloride concentration and raeommendad maximum ehalf Ufa man hare been affixed to the cylinder before ship ment tram tbe manufacturer to tbe buyv. If a cat ehnmalcciapb It used ee the rtnyl chloride monltortsc eyettm. these cat mixtures may ba directly mad to prepare a chramatoeranh calibration cum as dmcribad in aactton 7.J of Tmt Method 10*. The requirement* m sec tions (.2X1 and 6XCJ of Tmt Method 1M for certification at cylinder stand ards and for settoltshment and rarlfleaban of calibration amndardt are to ba followed.* required by this eeeOon. The flnt enl- annual report It to be euhmltted fallowInf the flrct full t month report!ne period after the Initial report is submitted* (h)(1) Inlbecaatafiattoit&f tourca or t new eeurae which bee in Initial atartup data peaeedtof tha effaettv* data, the ffnt report M to be eubmlttod within 1M daye at tha Mtaettra data, imlaei a waiver at compUanee le brantad under t (1.11. If a waiver at rnmpllanre la (nnted. the am report le to be mb- mltted on a date echadultd by tha Ad ministrator. (3) In tht cnee at a now source which did not hare in mitlel startup data pracadmq the effective date, the first report M to be nbmtttcd within Me days of tha ttadrnUlibt Nef tubsaQwmrn cm Am oe maid tBtttoi itutap dfth. (e> Unites oUianrtee muffled tha ewnr or operator shell urn the Teat Methods m AppendIT B to thfe part to | (1X9 labial repeet- omtdoet emtmtwi Mate te required by paracraphs (c)(1) end (e)(3) at thle (a) An owner or operator of any eectton ualam an (quireloat at in alter- source to which this subpart appliae shall nattre method has hots enprneed by the aubmlt a statamest tn writlnc ncttfytnc Administrator. B tha AdminMtrator tht AdmlnMtratcr that tht equipment Bade rmemetili craundt to dbputo tha and procedural iparlflrotten* m IC1CC raeulti obtamad by in equtritost at alibid), (b)(3). (bi(3>. (bxi). (bXD. larnattr* mathod. ba may nqubo tha urn A-1Q SPI-05296 at a nfmci method. If the results of the reference end equivalent or eltemstme methods do not agree. the reeulte obtained by the reference method pro mt. tod the Administrator may notify the owner or operator that approval of the method previously considered to be equivalent or alternative Is withdrawn. (11 The owner or operator shall In clude m the report a record of any emis sions which averated over any hour period (commencing on the hour) are m excess of the emission limits pre scribed In 1161.12(a) or (b). 161.13(a), or 161.64(a) (1). ib). (c), or <d). or for any control system to which reactor swissteisis qre required to be ducted In 161.64(a) (2) or to which fugitive emis sions are required to be ducted In I 61.66 (bXlHU).(b)(2).(b)(6), (b)(6)(U), or (b) (6) (11). The smminnt are to be meas ured in accordance with 161.66. (2) In polyvinyl chloride plants for which a stripping operation is used to attain the emission level prescribed m 161.64(e). the owner or operator shall Include in the report a record of the vinyl chloride content m the polyvinyl chloride ream. Test Method 107 is to be used to determine vinyl chloride content as follows: (1) If batch stripping Is used, one rep resentative sample of polyvinyl chloride ream Is to be taken from each batch of each grade of ream Immediately foilsw ing the ramnletlnsi of the stripping op eration. and Identified by ream type and grade and the date and time the batch Is cnmpleted The corresponding quan tity of material processed in each strlp- by reem type and grade and date and time the batch Is i (U) If continuous stripping Is need. chloride ream Is to be taken for each grade of ream processed or at intervale of g hours for each grade of ream which Is bamg processed, whichever Is man fre quent. The sample k to be takan as the of ream procmeed by the strtpper(a) that calendar day. according to the fol- ^-lFe.Jfe, Afi' ----- 57----- Pe,Mei+Pe|6fs,+ +F*.Wv. 57 isU-bow average wnsioentTallnn of type, T , rvaln us ppm (Cry weight basis). q kTotal production or type 7 i rvaln ever the 34-hour period, in eg. y teTypa of ream: iwU . . . at waste as a total number of team types piodaaag Sunns tbs at bees------- SSwOdaeaBWatloB of vinyl chiondv la oao r*aaPmropdlvuoctfiognrasosr 0gr,aladaemO. la ppm. roam repre sented by the sample la eg. O .wOrade of raala; eg.. O,. o r sad O r asTmal number ar gmgrn of raaiapro- gutss sunag the at-aour evrtoo. J* (vt) The owner or operator than re tain at the source and make available far Inspection by the Administrator ter a minimum of 2 yean records of all data needed to furnish the Information re quired by paragraph (c) (2) (v) of this taction: The records ere to ---* the following information: (A) The vinyl chloride content found mall the samples requiredm paragraphs (e) (2) (1) and (c) (2) (ill of this section. Identified by the ream type and grade and the time end date of the simple, end (8) The corresponding quantity of polyvinyl chloride reem processed by the strtpper(s>, identified by the raala type and grade and the Urns and date it represents. (3) The owner or operator than In clude m the report a record of the emie- which an emleelcn Unfit It prescribed tn 111.64(a) (2). BatMeno art to be deter mined la accordance with 161.67(g) (I), inept that emissions for each reactor are to be determined. For a reactor that h also used ao a otrtpper. the detarmmatton may be made immadlatety following the |dl.Tl unfit......... (a) The owner or operator of any aourot to which this subpart applies shall retain the following information at the source end make It available for inspec tion by the Administrator tor a minimam of two years; (1) A record of the leaks detected by the vinyl chloride men)terms system, as required by I 61.6*<b>(6>. including tht concentrations of vinyl chloride measured, analysed, end recorded by the vinyl chloride detector, the location of each measurement and the date and ap proximate time of etch measurement. (2) A record of the leaks detected dur ing routine monitoring with the portable hydrocarbon detector and tba action taken to repair the leaks, ae required by IIIJ6(b> (I). tnrludlnt a brief etatamtm explainmg the location and cause hydrocarbon detector, the date and time of the leak, and any tenon taken to eBrntnete that leak.* (I) A retard ef < m accordonce with I 01.0k1 (4) A dally operating record tor i poirRnn chloride reactor tada <6ta 114 ef the <41 SAC 64141k Ml SCC 94Mlk < The quantity of material proaaaaad by each attlpper over the time period raprooonted by the sample darmg the eight hoar period, n to be recorded and Identified by reem type and grade (ill) Tbs quantity of material proeoamd by the tripper a to be determined tn e dry solids bant and by a method qulied tn l (2) (ID ef l <T> Tht by tbe i the nyl ohlmKIe content found m east) sample required by paragraphs ()( (1) and (ei (2) (fi> of Sue motion. aver aged venerately far each spa ef i M Ft MM. 4/6/73 (1) it mended 41 n 46660. 10/21/76 (26) 41 F* 63017. 12/3/76 (30) 62 FI 26006, 6/7/77 (36) 42 n 41424, 6/17/77 (40) 43 MOO. 3/3/76 ( 47) A-ll s?v mOB'.l A lit1n>1 tafulaa IMv* far horM Mr MlatMti tetellct ttatu Mmtin nmnCRCM! aan tr Mntan tftiteM tf larla aallaaaaaa aatjact at M attlMl Mulaa Staattr* frr miirMi Air Hllmuiti an raairtM t Ml tea lafanaaiaa a-- la lactua > at tat annaHatt ii.t. tenia--tal teatactfaa hmma teaiMt Offlea arlar la M text after ana affactlat laaa af ate atiterte ar --M teaaa natch raaalia tea a--aalaa af -- k Kit af railMl afflaai <a pwltel la Ml.M. 1. IMMflaMaaAaaatlaa .laMaatetea -- late-- afaatel W liutte tiwvi a n--n itinla 17i*R wi W SUB IT-------- atim tf nail" 11-11 a------ --M II tejll. MM HR w AH 1harP v a. Ira IwctiitlM > ftp*ffW im i IWn FU>r tr timMm ftp*). ->vu irta n---------- `smiti *1---------------------------- (aaUteti----------- * rnr^riMTn'm U taotm iff*rwt ttH Hut apartfl tor > ir^a *r >-te a Strut arteal w " W ( --ItteehStnn tea MialM frm tttt aara _ tea _ MM Mt M aatutaa IWaatlM caauiaaf la M antMl irnlalhiaNi at ar an* at la hat after tea anwtiia am af aar IBM --m M a--aaua af i -------------- tt te. w. yc^itiiiSF^1 ` Ml f* Tlf-- Air AnUM. tt* M *1 ti U rUlttUA * torrtl uUUrtir.r--imt iawctlm--i*NUnucfTtiM*UiptfaApM. lwItrtTmmmarHmMiimftpMttmHr A-12 SPI-05298 [ U)Ct 11.11(1) Mr M*tft, fItmiltillIwIWfvrIIIMiiT<I--3rWtoNrHt i| iMaW t*. ()M %mr*u\y *fw' "^*** M*t ff 4. mirnu} "u -- tWi4 v *r U SIf A, HUrtmt farttfd Miuta tfca tyM ftf hiTtrftrur MlliitMt wtttftt by Mm ftmM. 2WT5& *W* fir uMt V Mr fttryntw. *r far mw: a a r_ iTTSEmt IJT" "BirntToo- T IT EC Htw BaacHatloa . krmrIO i krtef Oserlatln ef nek .mi ntyOrMtri end Mi1 i nreurj cMor-allaM .lent, *piiM1n| ncklM* la etoiyfllia aaettlm thee). IM aottlcral ikMtt if Mcnwr. tT^joaietU^Sa vm ef n^ettn central Men, If ngr, enf tto Militant fra the mens (a.|., rantarl senator, to*>eate> toi oelane) iM tto otlotto nrcert of ttopeUvUtt nick tto Orica la Ml iiI *kp zi DMf owtml ana: im hw me B H farcanc anml n h tmclaac? --------------------- tmctaaqr A-13 SPI-05299 tl. ttittt maxm A. MfcTvn Of CQMllAWCt. tem o? ooomtom tf mwch oMklt t o^fiti In cam i'me* vitn-tM Notional Cnitston StiMirA f*r KtranOoof Air MMutmti prior to >0 oays aftor tM offoctim Oat* of my stonOorOi or noonmantf Mch -eoeulr* tM tuM'tilon tf sue* Infcmatlon My Mount Mlmr of umllanca fro* tM Aminutrour of tM U.S. (nvironwontol Pmtoetlon Aftncy for tM tim mrloO nocotaary to Initill ooorooriat* control lovleot or HU Mlflcatlmi to ocMovt comllanco. TM AMlniatrator ay front 1r Of comHone* olta tM staMarO for 4 parfoo not oaeooOfM toe yiin fan tOo offoetlvo Oot* of tM MtorOowt pollutant ttoMares, If ho find tMt one* ptriod It nocottary for tM installation of eontroU im tMt itm aril! 00 toton Owrtnf tM otrlM of tM Mlmr to assort tMt tM Mol til of nfMM *111 to protoctoO fro tMinont ooeanfrnvnt. TM rooo-- tnfo*Mtion froviMf In Sort im T mrst otcommy tNs on^ettfoA. Applications tMu>0 00 tmt to tM OOOMfHoU CfA MflOMT offleo. 1. OrotottOt 1*olm0 Indents tM OMCoti or pmeotoot Mitt loo bttsrfeot oouwuntt u mien Mission emtnn* am to a* ocr'*M. t. Controls 0. ftoocHM tM pMQBiof tyo of cMtrol Mirteo to M IMM or Mflfleotlm to M mOa to tM proeon to roOoct tM onliilont of MzoMout pollutant* to an occopttOlo loml. (Dm oMItlonol MU If ninotary,) A. toterlbo tM Mura tMt *111 M tokm Ourlnf tM mlmr port* to UIUM tMt tM Molts of POMOAI *111 M OMtOCtOd ffM iMlMftt OMMfiMMt. <UM OMItlMOl MoU If MCOtliry.) Dot* Oy Midi contract* for odulflo control oyttoM tr proem oiiflcotloM *111 M OMMOO; or fata by Mid orMrs win m IsmoO for tM purcMoo of tM CfWM port* M Million cootMl or procats mdfintloo. T# AT B/BV/flT U C Mtt of Inltlotloo of OMttto unitrictln or iMtollotlM of obUsIm oantrol ooolp--t or pneon cMnyo. *B 1-M * * 7n r "ib si1 mmm um Mtt by Midi on-ilto 1 MtMCtloo or MottllotlM Of oMuloo COOtMl PiflCitlSO It to M BMplttM. Mp 1.11 8 3 y _ lT Tl mi _ 10 _ IT w/w/wr Otto by dick final cadllMen hum ocnitmo. M 1-11 an* vr *> t--rnmm--u w I. WITH Of Wratio* TCT8. A mlmr of Million tottlnf nay bo yrootoC so OMOrs or oooritort QTiomcM Of PorylllM or nort'iM pollotoou If, In tM JMpMPt Of tM AdHolotrotor of tM CnolMMOnUi Protoetlo* Afaney CM MlltM fvon tM oooru CMfly wltn tM looroorioto itmoorf or If tM won or ooorotort of tM oowreo Mo roeuoatoO 1 aoloor of cowpHonco or Mo boon yrontoo Mlvor of camnanca. TOIl MfHoitlon MotilO Soctlon t. tMtoport Inforaotlon protrlMC In Miw - Stott tM mount for roouoitinf a wlvor of mHiIm totting if cm rooooo 1totOC ft CMt tM Millions fmo too ooiirco am wltfiin IM proocrtMO IMiu. totwnlotion of Wi contfitlon Mit M tttoeM*. Hywino or im 1 (<CsobO. Atiee.0t1osthoeh.*r A-14 SPI-05300 imiauc ini fttimmasma m Tom Cbmm non VtineiiuT lmu>f rtnrmim > of this method should set bo attempted tor pirmi upfeiolUar with tha ^wood of i |M chromatograph. nor by them who tfi with source mm- puag, a* than an mtay details that an beyond tbe aeopt of this prmaatatloa. Caro rat ba mreMd to provost exposure of sampling personnel to vinyl chloride. a aw* naogea. l. Principle and Applicability l.l An integrated hoc aaapla of Rad gas <****"*"r vinyl chloride (ehioresthane) la subjaetad to ehroratogmphie analysis, tat a fiame loaimnon detector. s 14 Tha method la aopUoafale to the mra- urraeat of vtnyt ehlortde in ataek gras from ethylene vinyl chloride end palp* flays ehlartd* manufacturing prooaaaaa. m* aapt where tha vinyl afeiertta Is eonfamart ta paatteulata matter. S. hangs sad hewlttatyTbs lower Unit of Cataatlaa wtB wry e> mrdiag to tha chromatograph used. walum ropanid taetude l x 19' mg nad 4 x lb* award 103 > < i t Zf raolutloo of tha vtayl i paak la cUU sot for a > ean ha ftuthar altered with prior paak thiwucb aiqw. ne^a Cl CaipPOf (newt 40.1 Final gmialim pair w^bout) Inested at Haws m new* To adfwt 4.1.t pf| y> tty t Ilian par strata nwr 9 4.1.11 Rtot tube -< Typo 8 (or equivalent). ehlondt lo aitregro ryllndmi for which tbr attached to tha proto to that tha sampling fee compotman Uas baas cart*4ad by *J> lov rata eas ba regulated proportional to tha ftaek fat valoetty 4.2 8aapla moocery 44.1 Tubtng--Tefioa. 8.4 diameter. to oonnart bag to f graph aaapla loop. A now unused piaoa la employed for aacb aartaa ef baf aaapla that ra&ufactum. Tha tnanufaeturar nun bar* renomaiandad a maximum uialf life for aacb cylinder ao that tha concentration docs not change grater than s5 percent from the emitted value. The date of ga* cylinder pr*paration. certified noyi ehlond* concectra> floattitutat aa ernstum taat. tad la to ha dis uoc and reoommended maximum shelf life tapcarded opon conclusion of analysis at thoaa must have baas affixed to the cylinder before ahlpment from the gaa manufacturer to the 44 Analytts. 4.1.1 Oat chromatograph--With flame Ionisation detector. potentiomuric atrip than raeerdar and 14 to 54 al heated sam- plicf loop a automatic aaapla valva. 444 Chromatographic column. Stalattm atoal. I axil aa. containing 10/100 aanh Chromaaorb 103 a secondary ooluan of OS P-M. 30 parcant on M/30 mesh AW Chroma sort F. ataialaaa steel. 3 a x 34 am or Fora- buyer Three gaa mixture etandards may b directly ueed to prepan a chromatograph calibration curve ac deaenhed in section ?.3M 844.1 Cylinder standard* certification The concentration of vinyl ehloride in nitro* gaa ta each cylinder must have been certified by tha manufacturer by a direct analysis of each cylinder using an analytical procedure that the manufacturer had eaUbratad on the pak T. 80/100 atth. atalnlaa ataal. 1 axil Say af cylinder analytic. The calibration of aa Is required if eeetoldebyde la praaant. If the aaalytiaal prccadurt shell, as a minimum, wad. a aarnnrtarr eotuan la plaoad altar tha have utlllacd a three point calibration curve. Chromaaorb in coluan. Iti combined^ It is Tenmnimwirtad that the manulacturer .......... in should than ha operated at 130* CT maintain two caUbreoon standards and uss 444 Flow meters (3)--Rotameter type, thsee standards la the followt&g wsy at A to 100 al/alfi capacity, with Sow eoattol high concentration standard (between 50 and 100 ppm) for preparation of s calibration 44.4 Oaa isgalston For raqulrad gaa cylinders. 444 Thermomew Aocurata to on* de- fraa centigrade. to aaaauw taaparstura of haatad aaapla loop at time of aaapla lajac- sum by an appropriate dilution technique; (3> a low concentration standard rbetwaen 8 and io ppm) for verification of the dilution technique used * 8444 fetebUsfcmsut tad oertfloetlon of eeftbrartoft standards. The concentration of 4J Barometer Accurata to 6 am B|. to aaaaun ttmomhenc prwaura lround gaa ehrometogreph during --mini analyaia. each calibration standard must hare been established by the manulacturer using reliable procedures. Additionally, each calibration standard must have been veri 44.7 tap too ai/aih. fied by the manufacturer by one of the following procedures, and the agreement 4.4 OaUbrauem. between the initially dstsrmisad concen 4.44 Tubing--'Teflon. 8.4 aa w tration value and the verification concen pieces marked for trate value must be within 8 percent: U) vertlflcatloe value determined by com- with s calibrated vinyl chloride (3) vertfieauoe value by campartsea with a gas aux 444 ynnfv--04 mL gw tight. in eocordaoce with the pro- 44.4 tyrtaca Bflal. gas ught. 8 la camion 74 sad using 444 Plow raw Rotameter typt. o to 884+ parent vtnyle ehlonds. or (9) vertfl- ~r8 hl/htt range aaowratt to si%. ta ohon value obtained by baring the aragan a prapaiailaci of ftaadard mUbraooa standard analysed by tbe Ns- humu of htaadards. All aaUbratton be renewed ea a time with the maU life of the cylinder standards sold/*1 . _ -- tt a hi wry that ah na* 8. tooeadure. .------ata ho af chmaairnraphiw pW. 8.1 8.1 gampllng. Aramble tbe sample tram as la Figure 10^1. Fcrfatm a bag leak check 8.1.1 acnming to hactloa 7.4. Obawvt that all aaoaaetloos between tbe bag and the probe 8iJ siiHugw a**1 ____ . re tight. Flaec the end of the probe at the 8.14 Oaygah gaa. or Air. aa required by sentroid of the stack aad start the pump *ttb the needle valve adjusted to ytaid a 4 Cattbrg&oh. Cat oat of tha foUpwtng how of 04 ipea. After a period of time sutt- Sow: althar 844 and 84X or 844 * cleat to purge tha Una several times has 844 PtapJ edJortda. Mi* percent. Fun alapaed. connect tbe vacuum line to tbe yl chlenda gw oartthad by tha manufae- bag and evacuate the bag uatu the rotam rar ta eeatata a lamimup of 884 paimt eter tadlcara ao how. Then ropemtion the yl chlarlSt for wa in tha prapawtlwi of Mplc aad vaeuua uam aad begin tbe ac hhdaid gaa txturaa lh Itcdoo 7.1. Zf tha tual templing, keeping the rata proportional iiwmifnaiiinw matntataa a bulk cytimtar to the stack valoetty. Street the gas exittng vply af 884+ parent eisyl rmortar. tha the rotameter away from ~ftll,y personnel. rusahtm ahaiyeh ray have both pw> As the end ef tbe sampis period, shut off the rrao n tom euppey riwn > -- pump, diranpctt the use from tbe i cylfakw preptrad baa thh bulk aippty. hag. aad lUsnrmmrr tbs vacuum line from tha bag eenuuwr. Frotaet tha bag aontaiasr amtlhod aaaiyms must have hew cm 84 Mem tha eyttadw before iMpmnt from tt ' ** kept out sf direct sunlight. Whan at. all raaufaiiiuici ta tha buyer. * pmmbla aaalyms is to he performed within 111 JHmfaa ft. In (no*, for | *pb 88 bourn, but m ao era m caeam ef 73 iovs cf mmole ejection-* 84 fiample ........... .. with a place of Tef- A-15 SPI-05301 IK mist nsn pi kMMr^k -inp*- twits. Pvtteb tte rain to vitbAnv M 4NM tbs bftf ttfWOflP tiM SSMplC loop- Piwmb the oamptomt m tte map* pm |Mm mat um hwm> nin is tt* iaak.frwt ftmp, >M tteo to * eternal tubs, ttuowafl by O-JOO al/tttt iwiBiur wltt tew ms* vet TOlM. . Analysis- tet tte estnab <iaiuwi is 100* C, tte Ootoetor tempsnson to loo* C. ate tte MBiHi loop Mbparatiu* to To* C. Wteo aptittoa. hynrafts ate aayitn flow meat ten teoa doesratted wsrtfy ote wU* tala item tew aim Ptmbp oil obremato* prspb teftwai CM& am teiiua <r awtoqsb m tte amor too ortobttob o flow uo i& tte cost* noniinwit with tte aoou* mliy ate PstsrmiM test tte mm total Om atsbuimfl ofld ttet te-- lias Orgt te> mm fwpt tes ate loop lor thirty aseowte ot tte na of 100 ai/aio. than aetttvto tte sampto t1*s. Booorc tte laloMte ttao (tte pesmoa at tte poo ob tte oten it tte tte* poab tevi&f tte iwto&tioa urn* oarmcpoM. lag to nan shtortd*. o tetaraiMfl tt tw ite 7Jl Moaours tte pook mm, a*, fly i or o doe tstofrocor oi % pio&Utetor. Mmsms tte pook telffrt. X*. teoarO te. fc*. ote tte rwtoociaft tunc. tepooi tte lajoction at Moot two bam or natol top octearnttoa myi sfefenPa pate te am wi> in an* mars tana *V no wip mat ter tteat two anas win gg "*"* witepmi tte m% coooootrt* OtMpara tte vatto of a* to a ter tea step! ablates otiBpli wtvi tte Mato tatto *er tte ototeart pate wfcttb to etonss to batgpv te a rtoteHm tt tew Tates Otter 0? am teas io%, tte mu tteirtte pate mm m te pom (foteiry aataiiOteyOi to pro--at) pStoMomParWari SST* TtiTT l ** *** tte Mattes bwifllty ta te tt* i 14 Prtpmmoa tt step) toUscwic Mate, ate ter mtatewt. teoauaf a totem lam team iwiar teg mat tea paamo a late abate (Pssanbsfl ta pseteb t.> ate mw l ttton or oftrapM. Vfcte tte am to . not tte 0J ipM ta tstote asooi to tS+ ptoaaat mto rblntes tbraogb tte vaO tt tte teg. Bps* vtteteawtac tte oyrtafs aaaPlo. tettspbtotoy awnr tte wumpg tela wta a paste m tehten mp. Tte tea wow evtaa* a > tt ft ppm. so Oaf a aaaath < tepraat appwtoto taflm tt tte tap If tent ta twit to ttu tea pmm. Tteaa pm muamp ---------------------11* i pays no tte tteaa ip Pwrrtwi 0J. oteva. ass snaawatsr to X l patttioo. nasb ta map130* loop vim am htUvuc or aitrofta aafl aetlwats to* oampio ralva. teeora tte laisctioa uai. tte ssaipto loop tsmporaturo. too ooioam toatpteBSun. tte samsr pas Saw ru. tte apart spate aaa tb stesauator taaf> ooeur tn tte atea&es or tiayi eteerMo- teas* tsia saaamote wita tte rumpmant Muato* tat smapte Mooueauy to psctiaa 4. paafl tte asapis leap for te aseotoas at tte taas or 100 al/aus vita am of tte myt chlorite aatttoratiea miacum ate aettteM tte aate>* watte, teowa tte latostooa toms, atom tte ta mn afllante tainp't af tt us ntt asmaiM. piooo a fwtatostsr ttUw Mtvtao tt bap ate tte pvap lolst. twaeuats fbt bof Pamirs or tte tetaatotsr te ropistto at*o flaw wteo tbt bat appaart to te satpty uuueotos s tote, fl Calculations. .l Ostaratts tte mmpts pate ana as fatlavs: AfW AmAf Soufltm 104-1 f1. atewa. stoset tte talos of C, ttee oor- ittotoli totopto pttb ana. Cbl' salsto C, as tollewt: (pmaribte a --etiaa tAi or T4> ttottf am* ---- --aaoal sttt tPaM Itototf ta mettom too. Rap tte sainpuiif loop /or op i o* tte ram or ttL/mia vim M samarwi pas atmorw ate astomr tte saatpis telte- Ptoaste O*. tte rmrsniriiinii oC wmn fiplonos ttjortsfl. tte sttenasitr sM* tmMttaaepsratsMaiowtrtybitttpoprltotMonmtbiynt-jtaOterutistsentitianmniaa4sw,airtntsevstcpftitbontipae liroit (tea plat most potato r. O'- ttten .tte otter t---- nrTnrUTaii bsos tern ptattofl. Prow s aswetb avm fftrwipb tte potato. *-rMAzs BpppttttlM-P 4te*Ttootob>sMnwtmwttotoMto.s CfTW totetotoabsa to onto ibbtoli to tte tov C^TffsSmwSa Mjyr^sbtortos iMWtoi by iww?5<fMste'telSStoWtei pnotot wavoM asitiv tobtoMwa. Be. T<*Tte Mtotos bos nwfwisii as ite item toab to is* tM to wwiytb. *t P/-T-SH*aMS. VIIWT --------------------- rTib rtowtaa l. Prawn. O. f. Lwy. S. W. ate ptsptew. am* M- B. -Wtoyt CteortPs MsBitorttf tear `tte 1. P. " `Oootoneb "CtteepattoootoatTTO.atmtjfpl.stwernt-t i tte pap ta P-10 m Up tt MbVito allow s state atenipbv AsalMto TbTttowiii libsns, Onrpti, On Pa. ivii. a. "BmmiTiso or a Oaiissnan ste APaiytlaal Poitovi far Vinyl Cbiartts tt Air.** ttet. 4li.t*B..Sdkaommanamate# Mtonw. tstb-atafl.-iBttaflo.sTttwMkr otter te. A BPS JtepMt Mr. n-VCL-1. fl. ~PtatevPttnttoe or Pttttotoary Psmw ttustoan MatbaP far Tttyt Cbiartto.'* by Mttwan isitofoo Btottwm. sort. BPS Osatract too. te^p-lflpA 1PM Otter te. 1. Wm ll tt tea <jgf Ato i A-16 SPJ-05302 Mirmon I07 farm mum-- or Tom Cbloibi CoKTurr or Imoaai WaarswATm ftMtns. ut Turn Chlomm oirrsKT or Poj.mj*vi. Cwvow Ibo>. IbOUT, War Con. awe lotb SiKnu Psrfrrttspo* of this method should not b# attempted 07 parson* uatMiuhsr with the oparmuoD of a ya ebromatoyrapn, nor 07 those who or* with enmplt&f, as ttoare or many detail* that or* beyond tb* eopa of this pramatation. Cora moat Pa aurcieetf to prevent exposure of wmpMnp l--suum) to nnyi ehlonda. o aardsopan. 1. Principle aad Applicability 1.] TP* bools for this method relates to tps aopor equilibrium which B established betwsaa BVCM. PVC. now. water. oad oir in a dosed system. It boa Paan demonstrated that tpa BVCXI in o PTC rasa win aquUl- p*ota a 0 eioaad aaaaai quit* rapidly. pvo- ndad tbot to* temperature of tt* PVC room la mainrunwi oPota tba ytoes traaslttoa tamperstur* of that specific Main. 14 This proeadura u suitsPI* for dstsr- minim the vinyl ebortda monomer (YCM> content of laprooaa* wastewater samples. and tp* raoiduoi nnyi chloride monomer flPTCM) contest of poiynnyl chloride (PTC) resins. wot con*, slurry, and latex asapia*. It es&not p* used for polymer in fuaao form*, such aa ahaot or eupo*. if a resolution of tba vinyl cblorida pass la not satisfactory for a particular sample, than enromatogmpb parameter* may ba altered provided tba* tba pracialou and rvprodud- 0411ty of tb* analyaia of nnyi chloride eytin- dar standards art not impaired. Zf tbar* u rewot to boltov* that m* otbar hyqro- earboa vltp on Idtntical retention am* u praoent in tba sample, than supplemantal ewifirwatton of tba vinyl ohiond* peak through on nnaaiuto analytical tacbnlqua, such ujua spicwosoopj, should bo par* S. Bang* and fionmtmty Tba lowar unit of dstaction of amyl cfllortda will vary aannrdtng to tba chromate- frapp used. Value* rapertad Inatuba ixifi* m aad x io* my. Witt proper ealihcmuoa. tba upper ttaut stay ba axtandad m needed, a. Praamon ana lapwwiui minty, bn tbfitnonrory ooapnnaon batuisa aaaan laboratory** of tbraa mm miaplaa, aacb spilt tnu> thro* parts, yiatdod a 1------ ' daaiadoa of 343% for a mmpia with a mass f 2nd ppm. .ld% for a moipto with a mean of l.di ppm. and 44*7 tor a nmpto wttb a moan of cant ppm. 4. safety. Do aot ralaaaa amyl rhlorta* to tba libera tory atmoapbar* durinf pvparatlon of stand arm Vanany or purging vim VCWoar 1 turn must Pa bald to a (bay art required. tb* anpor mi to outattu off. Vinyl eblortda. ppm Mvsto. muat asaor a* ma UPonmvy. Af-- alaH baa* bo tbs praastua wltbm tba vial ma pnar to ramoval frm tama. Tints must bo 1 Pattiml tuba aaftnq a P: l.U dactneai tap* or squivatoct, to prevent loowaia# of Oottia tope 84 bampl* recovery. 84.1 viais--1With aaai* and caps- Parkin- Baar Corporation No. 108-C118. or aqutve- last. 844 Analytical Paianfla <sapaP1a of wtiyhinf to so401 from. 044. Byrtnfa. 100 "A" No. 010028. or aquiaaitnt. 84.4 Vial Sealer. Parkin-Emar No. 108- 010e or equivalent. 84 Analysis. 84.1 Oa Corporation Modal P-40 limit spans lymr. No. it* 0001. or equivalent. prsparauou of a ealibmioB nm by an ap* preprints dilution taebniqu*. (3) a low con centration standard (bovwaan 10 aad Mo ppm) for yanfleatiea of tba dilution teenuaad. 4.14 IrUbttaAmcat and sanfleattoa of aaUbratlo* ttcadordr Tb* eoneanoatton of acb calibration standard must bav* baan aatabUabad by tb* manufacturer nsuif railabl* procedures. Additionally, aacb caUPraboa standard muat bav* bean --lflad by tb* manufacturer by one of tb* foUowinf proo*- duM. and tba aproament batwaaa the ini tially determined concentration value and tba sanfleattoa concentration valu* muat be withins8 peroant: (l) Tarlflaatton valued*> 844 Cbro--togfcpAic eoluma. Bminlma taraunod by compancon wttb a pas mixture stael.In x U mm. cnntslntny 0.4 pcrea&t standard fsnoratad in a maanat to Carpowaa !I00 oa CarbopaP A. Partla-Bmar tba proeadura dearnbaa m aeetica 7.1 of Corporation Mo 108-0128. or equivalent. Method 108 for propartbf pm mixture stand CarbopaP C can Pa uaad in plae* of OarPepaP ards uamp 884+ pareeat vmyi ebionea. or A. If msthaaol aad/er soetnhiabyde la pres <3) sarlftaatien value abtainad by nanny tba ent is tb* sample, a pair ot PoropaP Q aoi* oaiibratiob standard aaaiyaad by tba Nation umn* in atria* <i m x 84 mm loltoved by al Bureau ef tanrtarda. An eaUbratton stand 2 a x 24 mm) with prevision for haekfluab ards muat ba renewed on a time interval of tba am column baa baas shown to pro^ consistent wttt\xba abetf Ufa of tb* cylinder vide adequate separation of nnyi ohlocidsr mandarda aoifl * 844 Thermometer--0 to 100* C. accurate ?. Mrooadur*. to sO.i* C. Psekin-Bmsr No. l0*-0l0d or TJ flamplt&f. equivalent. 744 pvc aaaapUnf--AUvw the ream or 844. bampl* tray thermostat system-- lurry to flow from a tap en tba tana or 1U0 Mba-Kmer No. 108-0102. or equivalent until tb* tap line baa baan wail purfad. Ex 844 fiepts 1 Sandwich type, fer auto* tend a to ml sample bottle under tb* tap. flu. matie doatne. 12 mm. Psrktn-Bmor Mo. 108- and immediately tiybtly cap tba bottle. Wrap 1009. or aquivaiaat. 844 Intafrator Packard Modal S2SDA. or oquivalant. aloctneal tap# around the cap and bottle to prevent tba top from looaasinf. Plae* as idtntifyint label on aacb pcttla. aad raeord 84.7 PUtsr drier amaably (2) Ins Me. 22201 IT. or equivalent. 844 Boap Aim flowmeter 8**pw aid No. 0101-din. or agutsalemt. 84 Calibration. 1.44 Patniatiaa fra luquaod gaa eytm< tba data, time, and +!> location both on tb* bottle* and m a lot book. 7.14 Water aamplinf--Prior m ua*. the 80 ml nals (without tba di--) must ba oappad wttb aluminum fed aad muAad at WC for at toast 00* bow to ia--ey or raBM.it any etfinic matter that could in- tartar* wttb analysis. At tha aampunf loco, tiaa flu tba stale bubbie-fre*. to overdovmi 0 that a convex mt&iaeua forma at tb* top. Tb* mm water la diaplacad a* tb* toaUny dim is carefully placed. Teflon aid* dews, on tb* opeataf of tb* vial. Plae* tba aluminum eaal ever tba disc aad tb* beep of tb- vial pad crimp into place. Aftx an idanttrytay land on the bottle, and record the data. time, and inapt* location botb on tb* rial* and m a toy boon. All samples must be kept ra- myeratad until analysed. ar nnyi < aa tb* miens to bn mm tb* 1 ?4 2ampl recovery. Bamptoa must ba ran wttbin M bows. eaatrauo&a with turn must bar* 74.1 Beam mmpia*- Tba welybt of tb* mam used muat bt between 0.1 and 44 frame. bdf life fer aacb cylinder so that tba con* An exact watynt must be obtained < -0.001 --tranoti Boos not abanf* fraatar than m# ------- > for aacb sample. In tb* earn or au- paeaaat from tba amtiBad vatua. Tba data penaiou raatsa a voiueaatnc cup ean ba pro- iaeem--ad*uof a cylinder preparation, --uflad amyl pared which wui bold tb* raquirsd amount sblonda waBwtauoa aad of sample. 77m aampi* bottle la opened, and maiimtiBi abetf Ufa muat ban baaa afiaad tb# cup solum* ef roam la added to tb* tired to tba eyttadm befon ample vui (ineiudi&y septum aad alumi maaufmmtrer to tba buyer. num cap). The vui a lmaaediamiy sealed 84.1.1 Opbiif itcudavN* and tba asset aampi* watynt u than obtained. Tba anaeanmaiicn at amyl antenia in um* Beport tbM valu* on tb* data abast aa it la pm m aacb cylinder muat pave bo-- caru- requited tar calruiinon of BTCM. in to# fled by the manufacture by a direct anaiyob earn ef relatlvsiy dry vemn aampi** fwater ef aaab eyimdar u--f aa abaiyttoai prece ooBtobt <04 wotynt 7,). ieo l of distuiod de* that tba maaitfbesu-- bad mllbramd mis mum be tapetod into tba vuL after an tb* pay ef cylinder analysis- Tb* calibre- aaoany and wotyhiay. uaiay a 100 ^ aynnya. ban ef tba analytical pfoaadiira ahalL aa a Zb tb* earn of dispersion rsnnv. tb* cup Tb# mmyie is taetaad wmybed appvokimately in aa aluminum di*b. fib* sarotf'vtal and watyhad eoauratoly in tb* vial. Tba mmpii m than niaoad in tba Parkm-B--or baan ---- **. lyaer (or equivalent) and conditioned fer oaa flow St 0O*C. vial cap* baa* bs raenoi toy. U not 1 A-17 SPI serums *r to the injection noui win eeeuf ?4.S fluepcnsion resin slurry *t cak* minpiM filurrp m\m be filtered uib| * m*U VuchBer tunnel with vacuum to yield vat eons. The filtering procsee must be cob* tttued only as long at a steady raita of water ! exiting from the funnel. ttemu filtration urns could ttouit a aoma lota of 7CU. The wet cake sample <0.10 to 4 grama > la added to a carta vial (including septum end aluminum cap sad immediately soaiad ample wtight u then determined to 9 oeelmai places. Th* sample to then placet m the perkin-Emer bead apace analyear tor equiva lent ) and condmonea for one Hour at 40*C. A mmple or vat esk* u used to determine T (total aoudai. This la inquired for calcu lating the BVCM. 744 Disparaioa retttx starry mtopiee.-- This maunni should nor ho filtered taaple mum ha thoroughly mixed. Ostag a tared vial iiacindinr septum aad aluminum eap) add epproaimecety 4 drape (4J4 to 444 pama) of many or latex umne a medietas dropper. This ihouM he done immediately after mixing Seal the ml ae coon ae pomtMe. Detomunt sample weight accurate to 0.001 frama Total sample wtight must not eicaad 0J0 crams Condition toe vui for one hour onat SPc m the analyser. Determine the T9 the slurry sample (flsetion 144), ?4.e laproeem vanevater sample*-- Using a tarad vial (ineludl&f septum aad aluminum cap) quickly add approximately } ec of water using a medicine dropper heal tbs vui ae eoo * aa possible Determine ample weight accurate to 0001 gram Coo* ditlea toe vial for two hour* at M'C lb tho aaaiyxer. 74 Analysts. 74.1 preparation of gas chromatograph-- Install the chromatographic eaioma and 00a* dttion overnight at 140*C. Do not cobnoct tho eon sad of tho column to tbs dotectar while 144J Mov rate adjustmoan qgjnof flow rates as follows: aubsequent ansiyeas la thaet caset the anaivau time must be edluxted to eliminate the interference Ao automated baekflusb system ca& also be used to solve ttm prob lem. c. B--Flushing--The normal settiBg ta 64 midutee d. WSfbxhMtun tmt. The aormel set* nag is 04 miautm * a. X--Number of analyses par mmple--Th* normal setting le l. 744 Prtperscioa of sample turntable--he- tore placing say sample into turntable, be eertetB that the center section of the s)u- miaum cap has bean removed. The numbered ample bonlea should be placed in the cor* responding &umbered positions id the turn table- insert samples in the following order FoaMene l h t--Old 4000 ppm gtaadarde Set fwndHUmmy. T&see are nsremery only after the aaaiyaar has apt bath used ter ft hoars or Jo&gtr. FOSfOOB paged. Foamoa freshly pee. ppm standard. freshly potiuone--*00 ppm standard, traeblypre- pared. Position 7 temple Ho. 7 (This u the ftrat sample of the day. but la given as 7 to as eon* mtant with the turntable and the istegraser printout.) After all sampiee have been peeitloaed. In* ten the second sot of 40. 000. 4000. aad *000 ppm standards. hatriplet meluding stand* eras must ba conditioned is tha hath of c for 1 hour (sat to oteoad 6 hour*) 14 4 nun chromatograph program-- Whan all aampiaa, including standards. have baas conditioned ax 00* C lor 1 hour, start tho analysis program according ta tha mn&v* taecoran' instructions Thorn taotroctiohs must ba carefully followed whan starting and stopping program to prevsat damage to 144 Determination t total aotlds (Tfl>. For vet cake. slurry, ream eolation, aad sam?ie line from the cylinder must oe purged fistoftocd, ior severs; snr.mrs pr.cr to filling viaie After purging, reduce *.r.e f.cu raw to approximately 600-)000 ce tmr. Piece end er tubmg into vial <oe*r oot:?m> end after one minute slowly remove tubing Place septum in mi as socn u poatible to m:nimice owing air vtin ear.ole After tie stand ard nsi* are edaled. inject I00tti of atstiUad water, 14 Preparation et chromatograph calibra tion curve. Prepare two lo ppm. two mo ppm. two 3000 ppm. aao two eOOO ppm atandarfi rampies Bun the caunrauoa aaapias m exactiv m* came manner as regular aaapies Plot a. the latogratcr area counts ter e*M standard ample vs c, the concentration of vinyl chloride ta eecn standard sample. Draw e line of best at through the points. S. Calculations 9.1 pgflfT factor From tha oaubration curve dtsmbed in lemos 44. above, aelect the vs.u* of C. that corresponds to a. ter oath simple Com pute tne reapanae factor, hr. lor escb sample. ao follows Equation 107-1 4 Roaidual vinyl cniortde monomer con centration. or vinyl chloride monomer cononatreuon. Calculate C,,, 99 foUon-i: -KT:)A,P, /.V.V, C. H,T, \ mxH where: Equation 107-2 C*. C^nerntrafinn of t-int'l chion^e in the *ampje. in ))pm P,Lahoratorr ottno*phrr* pres sure. BID Hg. f| -- Bi*nm temperature, *K. MMolecular weight of VCM 162.5). cylinder to read 40 pag. let regulator on ample by aomimxeiy waighmg ipptwln* ahiwmaugrapb to 14 kg/em*. Wormat flows ataly t toe grams et rnmph ta no aluminum this peeesws should bofll to SO ec/aeiputt- pan bofeee and after picag to a draft Chowx with bubble flow meter wvea (104 to UP* C). flatsplea mutt bs dried b. Burner air eoppiy flat reguinsnr on eyi- te sqflswnt weight. After flat wmghmg iv moar e road M pig. flet rsguiasor m ahramatognph to supply air to Purser ax a piste myrms. 19 is then oalrulstefl so tbs c. Bydrugaa supply got ragulamr on eyi* pm weight. moar te rood 40 psg flat regoiaior oo B rilllrrntiofi ehTtiitngrcph te oupmy appeunmotery ChUbrnuon te be patormod seen mg&t- U*l ee/fltinvte. Opttsnim hydregen flew te mw pwlad wtm Che mscrumest is nasd. yield the mem oeasttivv deceeter retpewee Book day. pntr ta ra'uubg midpim the col without emagubhing the flame. Chaos flow umn eheuid be oowdmanad by roBBiag two with bubble muter nod reened tht* floe 74.14 Ttwipeieture stjwmmn flet twniwamim as fdUows: a. Otm (chromatograph* aotumn). 44* c. of tha previous dsn 4404 ppm sinwdSTits. J Freparauon of dtandsrds. Cauheetma standards are prspfed hy flu- tug the nais wtih the nayl ehlcrwe/nitro* gm staadarda, rapidly anting the septum h. Dmng Una. 140* C. c. ttpsurn bioch. lew c. and matthg with the aluminum eap. Obe s ecainlom ensl Use from the cylinder to the r mdolume of vapor pheea (vtnl volume tern sample volume). m Weight of sample, grams BwOaa eonatam ftt.SdO (oc-mm-moie- dspsm Katvtm ] X * Benry s Law conotaat. For VCM is F^C at 40 e. jr4t X lO-*sr*v Forvcvia 1 et (approximate) waetewwtcr sample at M*c.jcse / io*sjr. T. s Bquuibratic& teaparstor* *X If the following conditions art mat. equa tion 107-3 can ba simplified as follows: 1. T ,x C i7y X> 94.. FT./ws4T0g-fim<4m43H* Xg) t .sTiai volume ec (934) I. flampte fonts ins imi then 04 percent water. C"*"l5 TH MHW |wl 11--.1--1 -- > --* Kr m*T Efluation 107-3 -fclcte mnr.ln. TCM. ITTC tain high boUtt moatrsus whaeh ae& mmm lawrfwwum with ths rawyi ehMnwa sank an c..-^x[^+A.fr5)r.^Jf.n - mr,] Equeoda 107*4 A-18 SPI-0S304 rJ=Total ooHdt. Neva. t m muot b* domralaod to atmplw vttb a npor voiua* to uqui volumo ratio etbr tfcaa 294 to j. Thu roue eao b* obaiBit M Mhtiuu tiu *** vaicht ttmifb !Ti&c oBd*ratiaa to tfc* total ooUdo and teaattf oI tb* PVC. 1mm ofttautfiod oitat BqttftMe 19%4 imwurt eoheoatreoeh baood oa tb* tow aaapl*. To obtarn mult* baood oa dry FVC ooftMftt. Ovtdi by fi, Por a )*ce waotovator iampl* (that is. 294 to l vapor velum* to liquid *eium* ratio), t m if |.o xio*. nia Kvuatwa ltfi4 oaa b* tuapuhod to tho touovwt: C'M*T* [S-^X~V(2.6<X 10-')j SqubUofi 107-5 uliSaadMi<*)tho Ooat Air Act. 48 V*C. isf?o-T ud 1MIf (*).) 10. tMmrmom a. BouaI \riayi Chiondo urmnoair Cotataat of PoiyviByl Cblerldo Pf*w a&d Wot CaM Sampim, B. f. Ooodricb Chemical Co. taaoacd Toot Proeoqur* No 100A-T. B. t. Ooodrteb Toohawal Oaatar, Avoa i^feo. Ohm. January to, ms. b. iirtm. a. Il *Th BeluMfcty or Tibyi CbicrM* te Boirwari ChJondo." AOB-DwMoa c Polymer Choaletry, Poiyaar Pn> priot* tt rat: m, irr< e. Bovaao. A. fc, Th* tXBtwoa Ttayl Chloride ta Mpvtsyi Chloride," ACB-Pm* lot ol Polymer Chomhtry, Polymer Pro* print* if cats aw. irr*. Both* a. a. L. . ewe*. c. a. Team* tok ahd J. U. Whitney, Aaaiyne lor viayi Chkne* ta PVC Powoen by r--a-ijpiir Oa Chrametatraphy,'' to b* ptiumbed. sir4*CBac 114 r tw (U9AC 1414)1 A-19. SPl-05305 APPENDIX B REGIONAL EPA AND INDUSTRIAL CONTACTS SPI-05306 Table B - l. REGIONAL EPA AND INDUSTRIAL CONTACTS o t0o> ciO -M Q. 0> Q_ UJ &- CO) to o 00 re u o <D c o I- > E re o u o .8.2 4- 4-> 4-> o 43-> treo 4-> O ->aJ EC3 OJ O r- 4-J 00 c u "D c re 4-> -O c o CO U. QJ c o U 4- 4uOra-e>; *>2a0-)) O Q. O Q.CX 4-> d) 5 3 -c a; o.4-> to f c re > *a> o sa-) cr c re c o re o to -Q cn >, --aCM>. o "O DaC) "43o-> U c ai to a; ttoo re j+c4 5 oc 3 >, a; o u -o to 3 E o > Q*-lO+jHi-- ai 4s- o 4-> a) X 3 -c GJ Q. 4-> to f- C re > *f- o p a> +> a: ic- c re <_> re o to jQ C7) >i a> o a; j4cJ -o a 3 a to a; to re__ CO +-> 5 c 3 >> _ 0> O u "O E - u to 3 O > Q(/)4- U o 0) 5 3 -C 0) Q.4-* to r- C re >O) *- -Uo> Sa>a: c re *- c o re o to 01-0Q 30>*0>> .4c-> "O cr 43-> ic- to re a> CO -M 3U ->o> E to S<D auOc; to 3 O r- &- Q C/0 4- "O c o zrue uu 3 a. O to Ore r0e> oo o ore *ra"e0> o re "303) o rOue) 4-> CO L. >O- 0) re troe OCO CD <D 4- O) O o o o OC 00 CO CO CO * 1 1 1 CO CVJ VO *m a> *" pH "j* 3 1 1 1 VO CO CO CTs >OtOo. 2a> CD re jE Q. 5 re C cl UJ M on on c c c ooo o jm p < O) cn cn a. UJ <2 <2 <2 B-l SPI-05307 EPA, Durham, NC Robert Laundrie Discuss the scope o f th e VC Review Study. 4o QJ C to -f- O +-> Ctrl Q$) 3 Q. 4-> <D 2 3 -C a> Cl 4-> to c fG > f o 01 a> 4-1 oc c ro f- c o <Q o to 4-> .f- -O CT> QJ O flJ "O -C ec 3 +4 *3 4-1 C 0) to to fO -C CO 4-4 2 3 >> 01 w *3 B F* C S- 0) (J c o u CO 3 O > 4-> S- <U 4- O oo 4- i- o 4-> QJ 3 -c O) Q. 4-> to C IQ > O 0) 0) M i- OL C (Q i- C u to o ^ > 4-> *F- . J3 3 >i 0) O 01*0 jo a s c 4-1*3 4-> &_ - C 0) to to <o jz to 4J 2 3 >> 01 U -3 S i- 0) o c o u tO 3 O > r* 4-> i. OJ 4o cn 4- w o 4-4 01 5 3 -c 01 o<p CO P- c ft5 > V- O 01 01 4-4 S- CL C (Q *- C 0(0 0 to > +4 .f- -O 3> >> 01 O 01 *3 -C CC 3 C 4- -3 4-4 J- C 01 to 01 (/) IQ o to O c 3 >> 01 o U *3 E u tO 3 O > i- +-> &Q 00 4- iO.J 4- o |Q O) 4+-4> c o tcra- S- CO u(Q 4-> c oo 2c>*QoJ> aao; CQ U (Q "30> trs -q 01 C 3 01 "3 O'----- +-> c-- 01 f- Q l/l 3 u WS Q S- CL) Cr 5 01 "O C 0) </) a. (_>!-- 3 1/) </) > O to ro 0) SC a_ JC oaa< 4-1 < u x x o o - 00 *-m >>> un -C UJ o 2Z C **" to to o L. to to --i <_> -0 3 3 > OlC U O LO <u +4 to to O 01 o: 3 f- f- O -C w <D O 3 4-> 201 QC3On 0O(Q1 c IQ 00Iut)1 t<o0 oIQ < 0) oo 4- 3) oC Ol 4-> o CO i o CO 1 Os O CO CO I i r-* m cm GO 1 00 CM 4-> 01 o(Q 0I) CO 1 00 1 CO CM 1 Io r*>.-- o1 p4 Table B -l. Continued 7-31-80 c o to *> IQ 4-> oc IQ >> f" c 4*j (C Q. -c^ w o> CJ> cc oo cn o a01: cOcl o to o> IQ IQ r" o F-- (Q o o MM >> cc oo O) Q0L1 01 CL B-2 4>-1>C<-/3) f- *--o' IQ C 3o O^F- to > < *ro 4-> to (Q C O *O i- o -C 4-1 4-4 -F3e Society o f the P lastics Industry SPI-05308 35 OJ C (O *i-- O 4-> a. a; s- a? 3 Q. 4-5 U <3 4-5 C o o to .c s. >) o to 4J r- f- i- 3 2*- 13 4-J 44 U e c to 3> OJ o u c W i. U3 to to - 4-> i- 0J -r- 4-5 C 3 O to CO o to to r- u OJ EX to o o (J <3 01 C 0J Ol*^ o 4- O 3 to to i. 3 - O' 3 to o. C4J-f- f~ E to 3 c a> E 4-5 o -- -C C OJ c u Jr-U-f- <3 QJ OJ O 0J OJ o CO c a*4J jq > o. to x: s- >> o to 4-> p- 2r- t0 .-- J. 3 to 0-P 4-> a ) c c to a cn O OJ o F- 0s- c s- w U 3 > <tj f- 4-> &- J r- 4-4 C 3 O #> ow cE X to -o o 0)flj OJ c 0) OJ 4- O 3 to to o - cr 3 to OJ C+J'P E to S C 3> E 2o P* C 0) o '--<->m- OJ 0J O OJ 0J o CO C Q.44 -O > cou r-- to X i4-5 -f- > P to r- i- 3 - 3 4D 4-J O 2 C C to 0a> o J o r-- s- c S- i- U "0 to 4-> s- OJ o U 4-)i- 4-4 C 3 O to to to E x o to c OJ ou <0 OJ 4- a3; toOj 5tO E Eo0U) 0J CO "*cO3 2 0c-J --3 oa- . cr 3 to Q. -Cc Oci Ea> cP caji aiOr<-3 +*-->>.J5i :> a. to to cc >> o "D OJ 0J -C 0a. a. ) 0J to to oo ^3 *3 tco > x CD to o> OJ <3 U Haa.: .X <3 a. to oj s. CO OJ 4- 0> OC o o 00 I O fOJ 4-5 4-> a; si Cr-O*I0i0 r-* i O 00i VOi 00 0"O > o u Table B -l. Continued oc >> <o to o <u3 a; -oC oooooc >> <c3 oo <u3 E OJ CoooucJ B-3 e- oo Xu o(. t-o<EC>o "D co 5 SPI-05309 Table B -l. Continued p o llu tio n control techniques c u rre n tly beinq used to c o n tro l VC em issions; discuss p la n t processes. <4o CT 01 c o 44 Cl qj S- 01 Q3. E on >- O on 44 r-- S- 3 r to 44 44 U $ CC c- o u ro cf- 4w4 Q) L- O4^ U "O CD o> r 44 on C o sU o44**i/> on so f- u c 01 X on -o o <J to 0) c QJ *^* c 0*- o 3 to iD s- x> - C" 3 on a. 01 C 44 ro 3 c c e -*j O JC C QJ c o 5 f-- u **-- ro 0) OJ Cl 0J o CO C CL-M XI > Q. f-- on -C X O on 44 i-- S- 3 r rO 4J 44 O 5 C C V) L. cr o c s- a; i- o*<u^ on m 1- 4-3 0J 4-3 C 3 0*- on r-- on o u 44 on on r- u c 0) X oo "C O (J ro 4- 01 c o OJ 3 01 on on o J- "O -r- C~ 3 oo Cl 0) C 44 O tO 3 c r-- -C cr. -> C 01 c u 5 r-- 04 ro 01 o OJ 01 <-) 00 c a+jja> O- p-- on _c s- >> O on 44 r-- i- 3 ro 44 44 U f-- c c on a. o 01 o *- Sro c 4&4- ii_- u "o on QJ 43 C 3 0** on on o o 44 tn on F" pX u on -o c 01 V ro 01 c QJ 01 *- o 4- o 3 on on L. -o *-* C* 3 in c. 01 C 44 *f E ro 3 c cr> e +- O -c C OJ c o S r-- <j - ro 01 CO 01 C o 01 01 0-4 CL 44 ^ > a. -C s_ 44 - ro 5 3 L. C ro f-- - 44 f-- on *1" EX ro 01 <+- `O 01 c ro O U2 01 01 CQ C U > UJ to 44 C X oo S -D c -SCoo: X u c0) j- <_? u <0 a. 0Qc). " s- on " rco 3 JQ ro 44 CO o< "O3 OC o 4- CD 444) 4041 O <0 OJ c o on > o >> rcQo. c O aC:D o CO I so I CO c o to w o Q. w oo .X u o sE -ttCno "Oc oE ro o o CO COI inI CO arO. E oo 1u0 E o2o B-4 a>cro>. g o rao '-iC u o 0o0I o o>I oo --Cs3OQCoC sr* H* rWO oc SPI-05310 t II r- 10 c > t0 -- C 3 s4-> *i- i~ to > O to i-- C 3 Xc 4J r- to O to f-- C 3 xc 4J *f-- v>' 40 P P u e e to r- *C PP u 5 <- c c to *" <0 P P 5 - c c to p- ro 4-> 2 f-- c cr> o 01 O f- c. o 01 O f- a o 01 O *- to On o 0) Purpose o f meeting c c C U "3 to f- +j c C3 .#> Oonl to i- *r- EX Crt3- O) O U CO <J on to -o o oDi aon; to- oOuOcnl "3 *r- CT 3 on a. 0) C 4-> s_ c s- C U "3 to <t> *- 4- c f-- on ou c 3u O * to 01 o01n tEo Ox) c to ^ o> O) O 3 to <J oc 01 TCJ 4->- O' 3 Ct c c c C U *3 O <0 1- *-> c c 3 O -4 r-- to U 4-> to p- f" C 01 s<E-C 0x1 "3 - on*D 01 Ol 3 t/> O' 3 Oi3 to OC on 01 c X P- P" o cc c c 40 f-- 4J c *- c 3 f-- to f" r- o u u E x to 40 01 c 01 C- -o o r- 3 O' 01 c 4-> -r- E (C 3 u 5 -- 0)0)0 xcc0)j *acOf-v) E0> O ^--rco CO C Q. +-> X > CL 0E <c --3 c X o C E 01 c U01 O5l --O QOJ 1r-)U<-0 CO C CL p c > a E0u 01 CO <c 3 C O E 5 01 c f-- X e--oQ.4<aPJi c 01 O' Xa3- i>uu<C E <c 3 C o0 01 CO X 01 c X e--O U01 Q.+J Contact JOOcO O Xoc o oc X O X c 01 cu 01 o to to 2 3 X *3 x <t(/o> o 4-> -co con u ro 3 JO c f-- "cO >01 01 aO.l Place Date o f m e e tin g R egio n /C om p a ny/D ivisio n o CO o CM 00 o CO I i On 0CO0Ii o o COi CO 00 > > > Qc. o or* o o Cl mu u oXl oI u0 (u0 E01 i Xo u "3c "3c o o ooo o o 4o0 c01 u. 10 CO CO tc0o1 being used to c o n tro l VC em issions: d iscu ss plant processes. -03o1 C ou B-5 SPI-05311 Date o f m e e tin g Purpose o f m e e tin g R e g io n /C o m p a n y/D ivisio n Place Contact p-- 1/) i- >> o <0 1-- u 3 fC 4-> 4-> U c c to o*> O o w rc pc- iW 4-> c su 3 u "O to 0) O r CO f-- to o u U 4-> to C to 0) E fC X 0) c to "O V <D o U O 4- o 3 to 4/1 i- T3 O" 3 4/1 o. o; C +J Q *0 3c c E +> 0) c U U " fO 0) QJ o AJ (D U CD c CL 4J -Q > Q. UO) .o o3 o <i-0 -C s- p-- >1 o to to 4-) i-- s. 3 <o 4-> 4-> <J 2 a o c<v c to r-- S- c s- W u "C to <0 p 4-> p* 4-> c s3 O CD p to 4/1 o U 4-> to 40 r u c O) EX l/l "D c u <0 a> c 0) o 4- o 3 to to s. -o CT 3 to C- QJ c M E <0 3 C CT- E O -C C O) c o 2 r-- u *- 0) a* o a; a; t_> CD c W. *-> jC > a. Occ)r >o o X 0 01 S. L3. oo0Qu)). CD o0to01 OO1 0o0in00i1 o u o L. oO CL) O O JZ on B-6 SPI-05312 APPENDIX C CURRENT INDUSTRIAL SOURCES SPI-05313 OPERATING ETHYLENE DICHLORIDE/VINYL CHLORIDE PLANTS CSJ >> to V/) a> u o o > (-5 > o > >o > o o > CJ > o CJ >> ^<1*^ cj >* o > CJ > cj > o > c cj o> o o o o u> o o o o CJ o cj o cj o o u oooooooooo oc co ooo o. UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ UJ trO** CO 0> o 3 CO 4c-> ro c c IQ * * CO c CO tO CO X CO c CO CfO* to c Cr~O CO c Cf*O to CO cco *" to to c 3 CO c ico ,o c <4- u a) o 4-> u 0) (0 f"* to *-- <D 3 3 CO o CO to CO o c o o4-> c(0 o ft--o. >> 3 to r* 3 o <0 * X tO 0) ro h- X 3 o J t-- 0 V) CQ X 0) o3 -J o k J +J o _J 0) 0k 0) h- 0J _J cj > H- 0) a> u o*. 3 0) c 0k 0J u * o 4- o rm o J to t 3 c CO to to a> o 0) -J ^ 3 3 uo o CO -J CO X to CO a; o I-- 3 k * --J u u k CO c o 4-J CO o 4-> u m O) E u CL E CL u CO C- 0) 3 o-J u |R <0 o. 4J 0) u o a> co o 0k o. CO CO o> 0) g #" k 3 OJ Cl c 3 c g > to to 4-> to <u QJ uo a) a. 0> (0 C" 4-> 0) o CO tn 0) +J >) u CO C" o CO 4-> CO to tu Jt a> CO cj cs 2 o Ou o U- CO Ou. CO o E o u to u QJ o 0) QJ o CD c o ai u OJ 4J 4o c to or- c M .X o to CO r-- o o CO oo uo uuu 4-> ur>* CO u to a> CO CO CO o sc <0 a. X Q s <3 u. CO e 0) .5 cj c 0) "O u o CO s u o u o c o cj E <0 t/} "O c O E <0 S E <0 t/> "O Qc g to 5 CO u *E u x o c0 u B w o 5 o o r-" CO u E 0) CJ X o Q fc. o o to CO c* au. CO o. #k o o CO Quj E s E 3> u >> u o o c 4-> UJ a; o oM- o u 4U to 3 "O c CO a. a. u s 0J o r~ 0) 00 uf-- E s o r-- 0J 00 u OJ 4J CO z c CO u r-- 3 > 0) CJ u #ai 3 CO 4- uo CO cn 3E uo Q. a> u c to c o 4-> CO u 0) Q. o cc o >M ^-4 HH HH HH M ^-4 X CO O) aQ:J >>>>>>> >>>> >> >> 0) CD c c-i SPI-05314 OPERATING POLYVINYL CHLORIDE PLANTS Began operation since promulgation o f the re g u la tio n (October, 1976). CO tO 3u o .X k CL *3 -Q O 44 cC oo e o Cce Ooo e o 3N4 *A* ^ tO to to to to to to kk k kkk k X 0) 0) 3 3 3 01 01 X X i> 3 44 3 a. Q. to to a. to pm a. & o. to to to a. 3 to +4 434 3 O -3 3 3 3 *3 3 "3 o. a 4 aa ** a* A cCccC Cec ce c c Cec c c oO ooO ooo ooooooo o o fB pm pm (/) to tO tO to to to to to to to to to to CO to to ccccc cec ccesccc ce 3 3 0) 3 3 3 3 3 X 3 3 3 3 01 01 01 3 3 X a. a. a. a. D .X to in V) to to Q. to a. to a. to 3 44 Cl a. a. a. a. Q. a. Cl a. 3 to to to to to to to to to 44 3 t/> 3 to 3 to 3 (/> 3 to 3 CD 3 to 3 CO 3 to 3 -J 3 to 3 to 3 to 3 to 33 to to 3 to 3 to 3 to to c >- * pmm #o to to CO 3 Z 3 33 2 * 2 > to to u z 1 c3 s. to 3A a a 44 A 3 3 >> to CO 3 Z SB >> 3 >> A to 31 o 2 44 Q. 3 44 > U. W A 3 <i aA pm * to 44 AZ O ck m k 5 3 e e * O 3 A 3 U A-- A 3 o o o 33 3) o Q 14 44 c3 o3 A <$ S c44 k <4- 44 4-1 44 44 3 0) pm* 3 ** oto 3 3 X 3 3) 3 3) 31 to t CL k 44 O 3A c O 3 U CQ fW s3 k kC 3 *" 3) ^a c Oo 44 2 3 3 k o o o3 3 3 3 as & & (J c --_1 Li. Q. Z CO e Ck 3 a to ^a S 3 B to k 3 3 3 ^a CO Li- to ID > 44 3 X > to 44 2 ID k 44 e 3 k 44 ^a 3 Q. CL CL U 3 u 3 > to 3 3 i3 > CO A-- i. J* 3 3 3O 3 G. e s z 3 u o Z 3 u <u O . 3 0) c u 3 3T e C s Sm o 3 t-- 3 E ID -O to to 44 44 k 3 33 TJ fA 911 44 O <s 3 t. CJ 44 3 (/) 3 3 8 3 3O 3 e C k33S k 3 A < 3 ID o o O O T3 o O O 3 O O O e c8 4-> C C 0) 3 to 3 3 > 3 3 3 3 ID 3 u. o 44 e C c cL. <U 44 oOC C c cO <s o 3 3 3 I00 cs ^-1 CD z O. h~ H* C 44 44 to 4. k k o e o OI ID to pC to 3 k Q. Ob 3 44 3 Li. k S3k k 3 Oo A* 3 41 fm o O U. Li. Q. 00 < < 00 o z * C-2 SPI-05315 OPERATING POLYVINYL CHLORIDE PLANTS (C ontinued) Began operation since promulgation o f the re g u la tio n (October, 1976). CO CO au; x ol- o> CL c o ccc c cc oooo tNC CO CO CO CO oo p* CO CO r w0) u 0> J01- &01- Cl CL Cl Cl CO CO CO CO L01. u01 CL a. to CO f* P" -o *3 "3 "3 3 "3 ao. #> * * * cc ccc c c cc c c cc cc o o p-- o p-- o p* o oo oo oo o o oo p- c CO CO to to to CO CO to to CO CO CO CO o CO CO c01 Cl CO c0) c. to c0) CL. CO 0c) Cl to c01 CL to -pX* -x f-- c Ol c. CO c 01 a. CO c01 0c1 c01 CL o. a CO to CO c01 0c1 a. Q. to CO 0c r ) 4-> c01 CL 3 Cl to CO 0c1 a. X01 to 4-1 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 O 3 3 tO 00 CO OO oo 00 CO oa oo OO OO OO OO OO 00 00 OO OO 00 --J cof4J <0 o -oJ X s>-> c z01 o 01 -tJo o>c < p> to ps oa. 1o-- lC pu-- * -o <co 3 z z o * pto* -3Q J4ttC-oo1 c < t 0c1 E301 cr to a. c _J * s-3C CL O3O o * 4>-1> O tO u c ao. o m 4p>---1 o m g -OC r--tC X o X 1-- * atLQ.. L. 01 oA? <-cj * a301i COL C 4ot-o1 00 <-J > to p <C <--J 0c1 E 031 3to" G_ X 1-- 4S->OCL J0011Li- X 1-- * to 30c1 ttoo ato. X *-- * 4>-1> C-3 to tXo H01- < * JC u to C0O1 3> c --oJ <u -C u to C0D1 CD c --oJ cC-3 3U tCo ii-hO- oU lC o LI u X s- o 0) s- T3 L. E o 0u 01 u "3 c. "031 *3 D OO OO C5 CD 01 *-- o 3 p-- o CD 4-1 C -C to "3 -Q c/> 01 Q. S- O S- c .c to &- tO TO o to u o (-3 O 8 3 5c Li. U. 8 1. o CO CO GO 2 <0 s- to 3 lo 4-> O u o o u o Qc g 4-1 to > 3 c i- c c to s- 3oo 4-> *p- co C-3 (_3 C-3 o LU Ll. 01 3) M i- c u 3 c 0C 4o Li- 3 C o O -C oo c 3 h- p* c CC to o-> t/3 p-- c 3 C-3 SPI-05316