Document MGg7vkabbbxZJkokzg1D2b6Mx

ABD00281961 CATc: SU22SCT ffiOM: 7C: UNITED STATES ENVIRONMENTAL PROTECTION REGION V I-D-40 APR 1983 Vinyl ChlorideNESMP Review S' Larry F/f Kertcher, Chijf Air Compliance Branch Air Management Division James C. Berry, Acting Chief Chemicals and Petroleum Branch (MD-13) Emission Standards and Engineering Division ;< ' . 1 , , . / fhOTFC7!G:i .vr.vr Y A.'T , 10^4 CENTRAL DOCKET SECTION My staff has evaluated the review document (EPA--450/3--82--003) and "Chapter 6" that you sent to Bruce Varner on February 25, 1933. We believe that the need to set vinyl chloride (VC) emission standards for polyvinyl chloride (PVC) fabricators merits further Investigation and possible inclusion in Chapter 6, On Page 6-3, the review document states that the PVC industry has established a 10 ppm VC limit in dried resins. However, on Page 4-58, it states that two bulk resin plants have 150-250 ppm.in the dried resin. The BF Goodrich, Avon Lake, Ohio plant ships latex solution meeting a customer specification of 50 ppm wet basis (IOC ppm dry basis). Thus there could be significant VC emissions from PVC compounders ano fabricators handling bulk and latex resins. The review document is comprehensive and well written. We offer the following comments for your consideration: J Page 2-6. Section 61.69(b)(1) also requires suomir.tal of me informa tion required under Section 61.10. 0 Page 3-12. The dispersion process has no centrifuge, but may include a concentrator. 0 Page 3-14. Steps 6, 7 and 8 are regulated (by resin residual standards 0 Page 3-17. 0.13 cubic feet is a gallon of gas. 3 Page 3-18. Somewhere in the document it would be helpful to list which standards only apply to eouiproent "in vinyl chloride service''. These appear to be Section 61.64(a)(3) and (c), ana Section 61.65(a), (c)(1) through (b)(5), (b)(7) ana (b)(8). 0 Pr.ge 3-32. Reactor emissions are also subject to Sections 61.64(a)(3) and 61.65(a). 0 Page 4-7. A distinction should be maae between seconaary controls that operate in tandem with primary controls, and backup controls that operate only when primary controls are inoperative. *QAM -CV S-TBJ ABD00281962 2 Page 4-8. The residence time for the 54GC (100QF) test should he stated. 0 Page 4-23. Latex reactors have fewer relief valve d'schargps oec3use initiator and monomer are continuously added during pclvme'`)zat on ;see Page 3-22). 0 Page 4-28. At a January 17, 1983, meeting, General fire informed us that alpha methyl styrene (AMS) shortstop temporarily stops polymerization by tying up reaction sites. When it is used up, reaction continues. 3y contrast, isoprene kills the catalyst, preventing any further reaction. 6 Page 4-32. Heat transfer is a function cf coolant flowrate as weP as temperature. Chilled water may be less elective than cooling towe: water if the latter can be supplied to the reactor jacket at a much higne^ flowrate. 0 Page 4-78. Collectively, fugitive missions constitute the largest VC emission source, at a PVC plant meeting the NESHAP. The -esponse time for repairing leaks is a important element of tne plan of action, and a major determinant of fugitive emissions. The vinyl cnloPde stanaaro dees not specify a response time. Region 7 believes a leak car ana should be repaired in 5 days, unless the equipment is in continuous service arc a spare is unavailable. However, the volatile organic guideline document for polyene1" and resin manufacturing specifies a 15-day -esponse time. The proposed benzene fugitive emission NESHAP specifies repair in 15 days, and an initial attempt within 5 days of detection. 0 Page 4-31. There needs to be a discussion of now VC gets emitted to the atmosphere as a result of opening the `edctcc usually an elephant trunk hose is placed in the reactor after it is opened, ana a fan discharges the VC-laden gas to atmosphere through a stack on too of tne polymerization building. This helps keep the VC concentration in the building below OSHA limits. 0 Pages 4-81 and 4-87. The calculation option for reactor opening loss is an alternative test method, not a test waiver. Operating para meters are measured in lieu of emissions. Test waivers are granted in the case of a contained -eactor opening - ^esin -esioual standard as described on Page 4-88. 0 Page 5-5. The relief valve and reactor manual vent valve discharge standards should require the engineering calculation cf tne magnitude, not just a description of the method of calculation. TCs is particularly important if the quantity discharged becomes the oasis for determining tne level of compliance or the action to be taken by the source or tne regulatory agency. Sources have been reluctant to supply the calculation. Frequently, the calculational formula(s) uan only approximate actual conditions. The calculation may use data that the source considers confidential. When furnished, the calculation is occasionally error. A ABD00281963 Page 7-7. The vinyl chloride NESHAP had not been delegated to any states In Region V that contain subject VC sources when the review document was completed In Februa-y 1982. The vinyl chloride NESHAP was delegated to Michigan on March 29, 1982, to Ohio on August 9, 1982, and to Illinois effective Septentoer 18, 1982. 0 Page 7-8. Borden Chemical is under the impression that VC losses from the dryers at suspension and dispersion resin plants meeting the resin residual standards, are equivalent to the 10 ppm standard for exhaust gases. If this is true. South Coast Air Quality Management Dlswlct-Rule 1005.1 does not require that dryer exhaust be vented to the primary control device, and Rule 1005,1 achieves no reduction in VC emissions from this major source beyond that already achieved by NESHAP. However, Rule 1005.1 may significantly reduce fugitive emissions from leaks. If you have any questions on these comments, please contact Mr. Varner at FTS 353-2088. cc: Richard Biondi Compliance Monitor'nq branch SSCO (EN-341)