Vol. 76 Friday, No. 227 November 25, 2011

Part II

Environmental Protection Agency

40 CFR Part 63 National Emissions Standards for Hazardous Air Pollutants: Mineral Wool Production and Wool Fiberglass Manufacturing; Proposed Rule

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00001 Fmt 4717 Sfmt 4717 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72770 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

ENVIRONMENTAL PROTECTION receives a copy of your comments on or regulations.gov or email. The http:// AGENCY before December 27, 2011. www.regulations.gov Web site is an Public Hearing. If anyone contacts the ‘‘anonymous access’’ system, which 40 CFR Part 63 EPA requesting to speak at a public means the EPA will not know your [EPA–HQ–OAR–2010–1041, EPA–HQ–OAR– hearing by December 5, 2011, a public identity or contact information unless 2010–1042; FRL–9491–9] hearing will be held on December 12, you provide it in the body of your 2011. comment. If you send an email RIN 2060–AQ90 ADDRESSES: Submit your comments, comment directly to the EPA without going through http://www.regulations. National Emissions Standards for identified by Docket ID Numbers EPA– gov, your email address will be Hazardous Air Pollutants: Mineral HQ–OAR–2010–1041 and EPA–HQ– automatically captured and included as Wool Production and Wool Fiberglass OAR–2010–1042, by one of the part of the comment that is placed in the Manufacturing following methods: • http://www.regulations.gov: Follow public docket and made available on the AGENCY: Environmental Protection the on-line instructions for submitting Internet. If you submit an electronic Agency (EPA). comments. comment, the EPA recommends that ACTION: Proposed rule. • Email: [email protected], you include your name and other Attention Docket ID Number EPA–HQ– contact information in the body of your SUMMARY: The EPA is proposing OAR–2010–1041 and EPA–HQ–OAR– comment and with any disk or CD–ROM amendments to the national emissions 2010–1042. you submit. If the EPA cannot read your standards for hazardous air pollutants • Fax: (202) 566–9744, Attention comment due to technical difficulties for Mineral Wool Production and Wool Docket ID Number EPA–HQ–OAR– and cannot contact you for clarification, Fiberglass Manufacturing to address the 2010–1041 or EPA–HQ–OAR–2010– the EPA may not be able to consider results of the residual risk and 1042. your comment. Electronic files should technology review that the EPA is • Mail: U.S. Postal Service, send avoid the use of special characters, any required to conduct by the Clean Air comments to: EPA Docket Center, EPA form of encryption, and be free of any Act. The proposed Mineral Wool West (Air Docket), Attention Docket ID defects or viruses. For additional Production amendments include Number EPA–HQ–OAR–2010–1041 or information about the EPA’s public emissions limits for carbonyl sulfide, EPA–HQ–OAR–2010–1042, U.S. docket, visit the EPA Docket Center hydrogen fluoride and hydrochloric Environmental Protection Agency, homepage at http://www.epa.gov/ acid for cupolas; add combined Mailcode: 2822T, 1200 Pennsylvania epahome/dockets.htm. collection and curing processes as new Ave. NW., Washington, DC 20460. Docket. The EPA has established regulated sources; and include Please include a total of two copies. In dockets for this rulemaking under emissions limits for formaldehyde, addition, please mail a copy of your Docket ID Number EPA–HQ–OAR– phenol and methanol for combined comments on the information collection 2010–1041 (Mineral Wool Production) collection and curing operations. provisions to the Office of Information and EPA–HQ–OAR–2010–1042 (Wool Modifications to the testing and and Regulatory Affairs, Office of Fiberglass Manufacturing). All monitoring and related notification, Management and Budget, Attn: Desk documents in the docket are listed in recordkeeping and reporting Officer for EPA, 725 17th Street NW., the http://www.regulations.gov index. requirements are also proposed. Washington, DC 20503. Although listed in the index, some The proposed amendments for the • Hand Delivery: U.S. Environmental information is not publicly available, Wool Fiberglass Manufacturing source Protection Agency, EPA West (Air e.g., CBI or other information whose category include emissions limits for Docket), Room 3334, 1301 Constitution disclosure is restricted by statute. chromium compounds, hydrogen Ave. NW., Washington, DC 20004, Certain other material, such as fluoride, hydrochloric acid and Attention Docket ID Number EPA–HQ– copyrighted material, is not placed on particulate matter for glass-melting OAR–2010–1041 or EPA–HQ–OAR– the Internet and will be publicly furnaces at major sources; revised 2010–1042. Such deliveries are only available only in hard copy. Publicly emissions limits for formaldehyde, and accepted during the Docket’s normal available docket materials are available the addition of emissions limits for hours of operation, and special either electronically in http://www. phenol and methanol for bonded arrangements should be made for regulations.gov or in hard copy at the product lines at major sources; and deliveries of boxed information. EPA Docket Center, EPA West, Room modifications to testing and monitoring Instructions. Direct your comments on 3334, 1301 Constitution Ave. NW., and related notification, recordkeeping the Mineral Wool RTR to Docket ID Washington, DC. The Public Reading and reporting requirements. Number EPA–HQ–OAR–2010–1041 and Room is open from 8:30 a.m. to 4:30 These proposed rules only apply to direct your comments on the Wool p.m., Monday through Friday, excluding major sources, but we plan to regulate Fiberglass RTR to Docket ID Number legal holidays. The telephone number wool fiberglass area sources in a future EPA–HQ–OAR–2010–1042. The EPA’s for the Public Reading Room is (202) action. policy is that all comments received 566–1744, and the telephone number for We are also proposing to revise will be included in the public docket the EPA Docket Center is (202) 566– provisions addressing periods of without change and may be made 1742. startup, shutdown and malfunction to available on-line at http://www. FOR FURTHER INFORMATION CONTACT: For ensure that the rules are consistent with regulations.gov, including any personal questions about this proposed action, a recent court decision. information provided, unless the contact Ms. Susan Fairchild, Sector DATES: Comments must be received on comment includes information claimed Policies and Programs Division (D243– or before January 24, 2012. Under the to be CBI or other information whose 04), Office of Air Quality Planning and Paperwork Reduction Act, comments on disclosure is restricted by statute. Do Standards, U.S. Environmental the information collection provisions not submit information that you Protection Agency, Research Triangle are best assured of having full effect if consider to be CBI or otherwise Park, North Carolina 27711, telephone the Office of Management and Budget protected through http://www. (919) 541–5167; fax number: (919) 541–

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00002 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72771

3207; and email address: D. What are the proposed decisions and B. How are the impacts for this proposal [email protected]. For specific actions related to electronic reporting? evaluated? information regarding the risk modeling VII. Rationale for the Proposed Actions for C. What are the air quality impacts? methodology, contact Mr. Chris the Mineral Wool Production Source D. What are the water quality and solid Category waste impacts? Sarsony, Health and Environmental A. What data were used for the NESHAP E. What are the secondary impacts? Impacts Division (C539–02), Office of analyses? F. What are the energy impacts? Air Quality Planning and Standards, B. What are the proposed decisions G. What are the cost impacts? U.S. Environmental Protection Agency, regarding surrogacy relationships? H. What are the economic impacts? Research Triangle Park, North Carolina C. What are the proposed decisions I. What are the benefits? 27711; telephone number: (919) 541– regarding certain unregulated emissions J. What demographic groups might benefit 4843; fax number: (919) 541–0840; and sources? the most from this regulation? D. What are the proposed decisions email address: [email protected]. XI. Request for Comments regarding subcategorization? XII. Submitting Data Corrections For information about the applicability E. What are the results from the risk XIII. Statutory and Executive Order Reviews of the NESHAP to a particular entity, assessments performed and the proposed A. Executive Order 12866: Regulatory contact Scott Throwe, Office of decisions for the Mineral Wool Planning and Review and Executive Enforcement and Compliance Production source category? Order 13563: Improving Regulation and Assurance; U.S. EPA Headquarters Ariel F. What are our proposed decisions for the Regulatory Review Rios Building; 1200 Pennsylvania Mineral Wool Production source B. Paperwork Reduction Act category based on risk acceptability and C. Regulatory Flexibility Act Avenue NW. Mail Code: 2227A; ample margin of safety? Washington, DC 20460; telephone D. Unfunded Mandates Reform Act G. What are the results from the technology E. Executive Order 13132: Federalism number: (202) 564–7013; fax number: review and proposed decisions? F. Executive Order 13175: Consultation (202) 564–0050; email address: throwe. VIII. Rationale for the Proposed Actions for and Coordination With Indian Tribal [email protected]. the Wool Fiberglass Manufacturing Governments SUPPLEMENTARY INFORMATION: Source Category G. Executive Order 13045: Protection of Organization of this Document. The A. What data were used for the NESHAP Children From Environmental Health analyses? Risks and Safety Risks information in this preamble is B. What are the proposed decisions organized as follows: H. Executive Order 13211: Actions regarding surrogacy relationships? Concerning Regulations That I. Preamble Acronyms and Abbreviations C. What are the proposed decisions Significantly Affect Energy Supply, II. General Information regarding certain unregulated emissions Distribution, or Use A. Does this action apply to me? sources? I. National Technology Transfer and B. Where can I get a copy of this document D. What are the results from the risk Advancement Act and other related information? assessments and analyses and the J. Executive Order 12898: Federal Actions C. What should I consider as I prepare my proposed decisions for the Wool To Address Environmental Justice in comments for the EPA? Fiberglass Manufacturing Source Minority Populations and Low-Income D. When will a public hearing occur? Category? Populations III. Background Information E. What are our proposed decisions for the A. What are NESHAP? Wool Fiberglass Manufacturing source I. Preamble Acronyms and B. What litigation is related to this category based on risk acceptability and Abbreviations proposed action? ample margin of safety? IV. Mineral Wool and Wool Fiberglass Source F. What are the results from the technology Several acronyms and terms used to Categories review and proposed decisions? describe industrial processes, data A. Overview of the Mineral Wool IX. Summary of Cost, Environmental, and inventories, and risk modeling are Production Source Category and MACT Economic Impacts for the Mineral Wool included in this preamble. While this Standards Source Category may not be an exhaustive list, to ease B. Overview of the Wool Fiberglass A. What are the affected sources in the the reading of this preamble and for Manufacturing Source Category and 1999 Mineral Wool Production source reference purposes, the following terms MACT Rule category? C. What data collection activities were B. How are the impacts for this proposal and acronyms are defined here: conducted to support this action? evaluated? ACGIH American Conference of V. Analyses Performed C. What are the air quality impacts for the Governmental Industrial Hygienists A. How did we estimate risks posed by the Mineral Wool Production source AEGL acute exposure guideline levels source categories? category? AERMOD air dispersion model used by the B. How did we consider the risk results in D. What are the water quality and solid HEM–3 model making decisions for this proposal? waste impacts? ATSDR Agency for Toxic Substances and C. How did we perform the technology E. What are the secondary impacts? Disease Registry review? F. What are the energy impacts? BACT best available control technology D. What other issues are we addressing in G. What are the cost impacts for the BLDS bag leak detection systems this proposal? Mineral Wool Production source BTF beyond the floor E. What analyses were performed for the category? CAA Clean Air Act Mineral Wool Production source H. What are the economic impacts for the CalEPA California EPA category under the Regulatory Flexibility Mineral Wool Production source CA–REL California reference exposure level Act? category? CBI Confidential Business Information VI. Summary of Proposed Decisions and I. What are the benefits for the Mineral CFR Code of Federal Regulations Actions Wool Production source category? CIIT Chemical Industry Institute of A. What are the proposed decisions and J. What demographic groups might benefit Toxicology actions related to the Mineral Wool the most from this regulation? CO carbon monoxide Production NESHAP? X. Summary of Cost, Environmental, and COS Carbonyl sulfide B. What are the proposed decisions and Economic Impacts for the Wool EJ environmental justice actions related to the Wool Fiberglass Fiberglass Manufacturing Source EPA Environmental Protection Agency Manufacturing NESHAP? Category ERPG Emergency Response Planning C. What are the proposed decisions and A. What are the affected sources in the Guidelines actions related to startup, shutdown and Wool Fiberglass Manufacturing source ERT Electronic Reporting Tool malfunction? category? ESP electrostatic precipitators

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00003 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72772 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

FA flame attenuation RACT reasonably available control rather provides a guide for readers GP General Provisions technology regarding the entities likely to be GHG Greenhouse Gases RBLC RACT/BACT/LAER Clearinghouse affected by this proposed action. These RCRA Resource Conservation and Recovery HAP hazardous air pollutants standards, once finalized, will be HCl Hydrogen chloride Conservation HEM Human Exposure Model REL reference exposure level directly applicable to affected sources. HEM–3 Human Exposure Model, Version 3 RFA Regulatory Flexibility Act Federal, state, local, and Tribal HF Hydrogen fluoride RfC reference concentrations government entities are not affected by HI Hazard Index RfD reference dose this proposed action. RS rotary spin HQ Hazard Quotient In 1992 the EPA defined the Mineral IRFA Initial Regulatory Flexibility Analysis RTO regenerative thermal oxidizers Wool Production source category as any IRIS Integrated Risk Information System RTR residual risk and technology review kg/MG kilogram/megawatt SAB Science Advisory Board facility engaged in producing mineral SBA Small Business Administration km kilometer wool fiber from slag or rock. Mineral SBAR Small Business Advocacy Review LAER lowest achievable emissions rate wool is a material used mainly for SCC Source Classification Codes lb/ton pounds per ton SER Small Entity Representatives thermal and acoustical insulation. This lb/yr pounds per year category includes, but is not limited to, SO2 sulfur dioxide MACT maximum achievable control SSM startup, shutdown, and malfunction the following process units: a cupola technology TC Toxicity Characteristics furnace for melting the mineral charge; mg/L milligrams per liter TCLP Toxicity Characteristic Leaching a blow chamber in which air and, in mg/m3 milligrams per cubic meter Procedure MIR maximum individual risk some cases, a binder is drawn over the TLV threshold limit value fibers, forming them to a screen; a NAAQS National Ambient Air Quality TOSHI target organ-specific hazard index Standard tpy tons per year curing oven to bond the fibers; and a NAICS North American Industry TRIM Total Risk Integrated Modeling cooling compartment. Classification System System In 1992 the EPA defined the Wool NaOH sodium hydroxide TTN Technology Transfer Network Fiberglass Manufacturing source NAS National Academy of Sciences UF uncertainty factors category as any facility engaged in NATA National Air Toxics Assessment mg/m3 microgram per cubic meter NESHAP National Emissions Standards for UMRA Unfunded Mandates Reform Act producing wool fiberglass from sand, Hazardous Air Pollutants UPL upper predictive limit feldspar, sodium sulfate, anhydrous NIOSH National Institutes for Occupational URE unit risk estimate borax, boric acid or any other materials. Safety and Health WHO World Health Organization In the wool fiberglass manufacturing NRC National Research Council WWW worldwide web process, molten glass is formed into NTTAA National Technology Transfer and II. General Information fibers that are bonded with an organic Advancement Act resin to create a wool-like material that OAQPS Office of Air Quality Planning and A. Does this action apply to me? is used as thermal or acoustical Standards OMB Office of Management and Budget The regulated industrial source insulation. The category includes, but is PB–HAP hazardous air pollutants known to categories that are the subject of this not limited to the following processes: be persistent and bio-accumulative in the proposed rule are listed in Table 1 of glass-melting furnace, marble forming, environment this preamble. Table 1 of this preamble refining, fiber forming, binder PM particulate matter is not intended to be exhaustive, but application, curing and cooling.

TABLE 1—NESHAP AND INDUSTRIAL SOURCE CATEGORIES AFFECTED BY THIS PROPOSED ACTION

Source category NESHAP NAICS code 1

Mineral Wool Production ...... Mineral Wool Production ...... 327993 Wool Fiberglass Manufacturing ...... Wool Fiberglass Manufacturing ...... 327993 1 North American Industry Classification System.

B. Where can I get a copy of this technology exchange in various areas of that does not contain the information document and other related air pollution control. claimed as CBI must be submitted for information? inclusion in the public docket. If you C. What should I consider as I prepare submit a CD–ROM or disk that does not my comments for the EPA? In addition to being available in the contain CBI, mark the outside of the docket, an electronic copy of this Submitting CBI. Do not submit disk or CD–ROM clearly indicating that proposal will also be available on the information containing CBI to the EPA it does not contain CBI. Information not WWW through the EPA’s TTN. through http://www.regulations.gov or marked as CBI will be included in the Following signature by the EPA email. Clearly mark the part or all of the public docket and the EPA’s electronic Administrator, a copy of this proposed information that you claim to be CBI. public docket without prior notice. action will be posted on the TTN’s For CBI information on a disk or CD– Information marked as CBI will not be policy and guidance page for newly ROM that you mail to the EPA, mark the disclosed except in accordance with proposed or promulgated rules at the outside of the disk or CD–ROM as CBI procedures set forth in 40 CFR part 2. following address: http://www.epa.gov/ and then identify electronically within Send or deliver information identified ttn/atw/rrisk/rtrpg.html. In addition, a the disk or CD–ROM the specific as CBI only to the following address: copy of each rule showing specific information that is claimed as CBI. In Roberto Morales, OAQPS Document changes proposed under this action is addition to one complete version of the Control Officer (C404–02), Office of Air available in their respective dockets. comment that includes information Quality Planning and Standards, U.S. The TTN provides information and claimed as CBI, a copy of the comment Environmental Protection Agency,

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00004 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72773

Research Triangle Park, North Carolina section 112(d)(2)(A)–(E)). The MACT regarding such remaining risk. The EPA 27711, Attention Docket ID Number standards may take the form of design, prepared and submitted this report EPA–HQ–OAR–2010–1041 (Mineral equipment, work practice or operational (Residual Risk Report to Congress, EPA– Wool RTR) or Attention Docket ID standards where the EPA first 453/R–99–001) in March 1999. Congress Number EPA–HQ–OAR–2010–1042 determines either that, (A) a pollutant did not act in response to the report, (Wool Fiberglass RTR). cannot be emitted through a conveyance thereby triggering the EPA’s obligation designed and constructed to emit or D. When will a public hearing occur? under CAA section 112(f)(2) to analyze capture the pollutants, or that any and address residual risk. If a public hearing is held, it will requirement for, or use of, such a Section 112(f)(2) of the CAA requires begin at 10 a.m. on December 12, 2011 conveyance would be inconsistent with us to determine, for source categories and will be held at a location to be law; or (B) the application of subject to certain MACT standards, determined. Persons interested in measurement methodology to a whether those emissions standards presenting oral testimony or inquiring particular class of sources is not provide an ample margin of safety to as to whether a public hearing is to be practicable due to technological and protect public health. If the MACT held should contact Ms. Pamela Garrett, economic limitations (CAA sections standards that apply to a source Office of Air Quality Planning and 112(h)(1)–(2)). category emitting a HAP that is Standards, Sector Policies and Programs The MACT ‘‘floor’’ is the minimum ‘‘classified as a known, probable, or Division, (D243–01), U.S. control level allowed for MACT possible human carcinogen do not Environmental Protection Agency, standards promulgated under CAA reduce lifetime excess cancer risks to Research Triangle Park, North Carolina section 112(d)(3) and may not be based the individual most exposed to 27711; telephone number: (919) 541– on cost considerations. For new sources, emissions from a source in the category 7996; email address: the MACT floor cannot be less stringent or subcategory to less than one-in-one [email protected]. than the emissions control that is million,’’ the EPA must promulgate III. Background Information achieved in practice by the best- residual risk standards for the source controlled similar source. The MACT category (or subcategory) as necessary to A. What are NESHAP? floors for existing sources can be less provide an ample margin of safety to 1. What is the statutory authority for stringent than floors for new sources, protect public health (CAA section NESHAP? but they cannot be less stringent than 112(f)(2)(A)). This requirement is the average emissions limitation procedural. It mandates that the EPA Section 112 of the CAA establishes a achieved by the best-performing 12 two-stage regulatory process to address establish CAA section 112(f) residual percent of existing sources in the risk standards if certain risk thresholds emissions of HAP from stationary category or subcategory (or the best- sources. In the first stage, after the EPA are not satisfied, but does not determine performing 5 sources for categories or the level of those standards (NRDC v. has identified categories of sources subcategories with fewer than 30 emitting one or more of the HAP listed EPA, 529 F. 3d at 1083). The second sources). In developing MACT sentence of CAA section 112(f)(2) sets in CAA section 112(b), CAA section standards, we must also consider 112(d) calls for us to promulgate out the substantive requirements for control options that are more stringent residual risk standards: Protection of NESHAP for those sources. ‘‘Major than the floor. We may establish sources’’ are those that emit or have the public health with an ample margin of standards more stringent than the floor safety based on the EPA’s interpretation potential to emit 10 tpy or more of a based on considerations of the cost of of this standard in effect at the time of single HAP or 25 tpy or more of any achieving the emissions reductions, any the CAA amendments. Id. This refers to combination of HAP. For major sources, non-air quality health and the National Emissions Standards for these technology-based standards must environmental impacts, and energy Hazardous Air Pollutants: Benzene reflect the maximum degree of requirements. emissions reductions of HAP achievable The EPA is then required to review Emissions from Maleic Anhydride (after considering cost, energy these technology-based standards and Plants, Ethylbenzene/Styrene Plants, requirements, and non-air quality health revise them ‘‘as necessary (taking into Benzene Storage Vessels, Benzene and environmental impacts) and are account developments in practices, Equipment Leaks, and Coke By-Product commonly referred to as MACT processes, and control technologies)’’ no Recovery Plants (Benzene NESHAP), (54 standards. Area sources are those that less frequently than every 8 years, under FR 38044, September 14, 1989), emit less than major amounts of HAP. CAA section 112(d)(6). In conducting described in the next paragraph. MACT standards must require the this review, the EPA is not obliged to The EPA may adopt residual risk maximum degree of emissions reduction completely recalculate the prior MACT standards equal to existing MACT through the application of measures, determination, and, in particular, is not standards if the EPA determines that the processes, methods, systems, or obligated to recalculate the MACT existing standards are sufficiently techniques, including, but not limited floors. NRDC v. EPA, 529 F.3d 1077, protective, even if (for example) excess to, measures that (A) reduce the volume 1084 (DC Cir., 2008). cancer risks to a most exposed of or eliminate pollutants through The second stage in standard-setting individual are not reduced to less than process changes, substitution of focuses on reducing any remaining one-in-one million. Id. at 1083 (‘‘If the materials or other modifications; (B) ‘‘residual’’ risk according to CAA EPA determines that the existing enclose systems or processes to section 112(f). This provision requires, technology-based standards provide an eliminate emissions; (C) capture or treat first, that the EPA prepare a Report to ‘ample margin of safety,’ then the pollutants when released from a Congress discussing (among other agency is free to readopt those standards process, stack, storage or fugitive things) methods of calculating the risks during the residual risk rulemaking’’). emissions point; (D) are design, posed (or potentially posed) by sources Section 112(f)(2) of the CAA further equipment, work practice or operational after implementation of the MACT authorizes the EPA to adopt more standards (including requirements for standards, the public health significance stringent standards, if necessary ‘‘to operator training or certification); or (E) of those risks, and the EPA’s prevent, taking into consideration costs, are a combination of the above (CAA recommendations as to legislation energy, safety, and other relevant

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00005 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72774 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

factors, an adverse environmental health from hazardous air pollutants by risk as a metric for determining effect.’’ 1 (1) Protecting the greatest number of acceptability, we acknowledged in the CAA section 112(f)(2) expressly persons possible to an individual 1989 Benzene NESHAP that preserves our use of the two-step lifetime risk level no higher than ‘‘consideration of maximum individual process for developing standards to approximately 1-in-1 million; and (2) risk * * * must take into account the address any residual risk and our limiting to no higher than strengths and weaknesses of this interpretation of ‘‘ample margin of approximately 1-in-10 thousand [i.e., measure of risk.’’ Id. Consequently, the safety’’ developed in the Benzene 100-in-1 million] the estimated risk that presumptive risk level of 100-in-1 NESHAP. The first step in this process a person living near a facility would million (1-in-10 thousand) provides a is the determination of acceptable risk. have if he or she were exposed to the benchmark for judging the acceptability This determination ‘‘considers all health maximum pollutant concentrations for of maximum individual lifetime cancer information, including risk estimation 70 years. risk, but does not constitute a rigid line uncertainty, and includes a presumptive The agency also stated that, ‘‘The EPA for making that determination. limit on MRI [cancer] 2 of approximately also considers incidence (the number of The agency also explained in the 1989 1-in-10 thousand [i.e., 100-in-1 persons estimated to suffer cancer or Benzene NESHAP the following: ‘‘In million]’’ (54 FR 38045). In the second other serious health effects as a result of establishing a presumption for MIR step of the process, the EPA sets the exposure to a pollutant) to be an [maximum individual cancer risk], standard at a level that provides an important measure of the health risk to rather than a rigid line for acceptability, ample margin of safety ‘‘in the exposed population. Incidence the agency intends to weigh it with a consideration of all health information, measures the extent of health risks to series of other health measures and including the number of persons at risk the exposed population as a whole, by factors. These include the overall levels higher than approximately 1-in-1 providing an estimate of the occurrence incidence of cancer or other serious million, as well as other relevant factors, of cancer or other serious health effects health effects within the exposed including costs and economic impacts, in the exposed population.’’ The agency population, the numbers of persons technological feasibility, and other went on to conclude that ‘‘estimated exposed within each individual lifetime factors relevant to each particular incidence would be weighed along with risk range and associated incidence decision’’ (Id.) other health risk information in judging within, typically, a 50- km exposure The terms ‘‘individual most exposed’’, acceptability.’’ As explained more fully radius around facilities, the science ‘‘acceptable level’’, and ‘‘ample margin in our Residual Risk Report to Congress, policy assumptions and estimation of safety’’ are not specifically defined in the EPA does not define ‘‘rigid line[s] of uncertainties associated with the risk the CAA. However, CAA section acceptability,’’ but rather considers measures, weight of the scientific 112(f)(2)(B) preserves the EPA’s broad objectives to be weighed with a evidence for human health effects, other interpretation set out in the Benzene series of other health measures and quantified or unquantified health NESHAP, and the Court in NRDC v. EPA factors (EPA–453/R–99–001, p. ES–11). effects, effects due to co-location of concluded that the EPA’s interpretation The determination of what represents an facilities, and co-emissions of of CAA section 112(f)(2) is a reasonable ‘‘acceptable’’ risk is based on a pollutants.’’ Id. one. See NRDC v. EPA, 529 F.3d at 1083 judgment of ‘‘what risks are acceptable (DC Cir. 2008), which says in the world in which we live’’ In some cases, these health measures ‘‘[S]ubsection 112(f)(2)(B) expressly (Residual Risk Report to Congress, p. and factors taken together may provide incorporates the EPA’s interpretation of 178, quoting the DC Circuit’s en banc a more realistic description of the the CAA from the Benzene standard, Vinyl Chloride decision at 824 F.2d magnitude of risk in the exposed complete with a citation to the Federal 1165) recognizing that our world is not population than that provided by Register.’’ See also, A Legislative History risk-free. maximum individual lifetime cancer of the Clean Air Act Amendments of In the Benzene NESHAP, we stated risk alone. As explained in the Benzene 1990, volume 1, p. 877 (Senate debate that ‘‘the EPA will generally presume NESHAP, ‘‘[e]ven though the risks on Conference Report). We also notified that if the risk to [the maximum judged ‘acceptable’ by the EPA in the Congress in the Residual Risk Report to exposed] individual is no higher than first step of the Vinyl Chloride inquiry Congress that we intended to use the approximately 1-in-10 thousand, that are already low, the second step of the Benzene NESHAP approach in making risk level is considered acceptable.’’ 54 inquiry, determining an ‘ample margin CAA section 112(f) residual risk FR 38045. We discussed the maximum of safety,’ again includes consideration determinations (EPA–453/R–99–001, individual lifetime cancer risk as being of all of the health factors, and whether p. ES–11). ‘‘the estimated risk that a person living to reduce the risks even further.’’ In the In the Benzene NESHAP, we stated as near a plant would have if he or she ample margin of safety decision process, an overall objective: * * * in protecting were exposed to the maximum pollutant the agency again considers all of the public health with an ample margin of concentrations for 70 years.’’ Id. We health risks and other health safety, we strive to provide maximum explained that this measure of risk ‘‘is information considered in the first step. feasible protection against risks to an estimate of the upper bound of risk Beyond that information, additional based on conservative assumptions, factors relating to the appropriate level 1 ‘‘Adverse environmental effect’’ is defined in such as continuous exposure for 24 of control will also be considered, CAA section 112(a)(7) as any significant and hours per day for 70 years.’’ Id. We including costs and economic impacts widespread adverse effect, which may be acknowledge that maximum individual of controls, technological feasibility, reasonably anticipated to wildlife, aquatic life, or lifetime cancer risk ‘‘does not uncertainties and any other relevant natural resources, including adverse impacts on populations of endangered or threatened species or necessarily reflect the true risk, but factors. Considering all of these factors, significant degradation of environmental qualities displays a conservative risk level which the agency will establish the standard at over broad areas. is an upper-bound that is unlikely to be a level that provides an ample margin of 2 Although defined as ‘‘maximum individual exceeded.’’ Id. safety to protect the public health and risk,’’ MIR refers only to cancer risk. MIR, one prevent adverse environmental effects, metric for assessing cancer risk, is the estimated Understanding that there are both risk were an individual exposed to the maximum benefits and limitations to using taking into consideration costs, energy, level of a pollutant for a lifetime. maximum individual lifetime cancer safety, and other relevant factors, as

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00006 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72775

required by CAA section 112(f) (54 FR this benchmark, they become the court 6 that PM is an appropriate 38046). presumptively less acceptable under surrogate for HAP metals, therefore, we CAA section 112, and would be retain that surrogacy relationship in 2. How do we consider the risk results weighed with the other health risk these proposed rules. in making decisions? measures and information in making an In separate litigation, the Court In past residual risk determinations, overall judgment on acceptability. Or, vacated portions of two provisions in the EPA presented a number of human the agency may find, in a particular EPA’s CAA section 112 regulations that health risk metrics associated with case, that a risk that includes an MIR govern emissions of HAP during periods emissions from the category under less than the presumptively acceptable of SSM.7 Specifically, the Court vacated review, including: the MIR; the numbers level is unacceptable in the light of the SSM exemption contained in 40 of persons in various risk ranges; cancer other health risk factors.’’ Similarly, CFR 63.6(f)(1) and 63.6(h)(1) that are incidence; the maximum noncancer HI; with regard to the ample margin of part of regulations commonly referred to and the maximum acute noncancer safety analysis, the Benzene NESHAP as the GP rule. When incorporated into hazard. In estimating risks, the EPA states that: ‘‘the EPA believes the section 112(d) regulations for specific considered source categories under relative weight of the many factors that source categories, these two provisions review that are located near each other can be considered in selecting an ample exempt sources from the requirement to and that affect the same population. The margin of safety can only be determined comply with otherwise applicable EPA provided estimates of the expected for each specific source category. This MACT standards during periods of difference in actual emissions from the occurs mainly because technological SSM. Because both of the Mineral Wool source category under review and and economic factors (along with the and Wool Fiberglass NESHAP relied on emissions allowed pursuant to the health-related factors) vary from source the GP rule for startup and shutdown source category MACT standard. The category to source category.’’ provisions (40 CFR 63.1194 and EPA also discussed and considered risk 63.1386(c)), we are also proposing to estimation uncertainties. The EPA is B. What litigation is related to this proposed action? revise these provisions for both of the providing this same type of information Mineral Wool and Wool Fiberglass in support of these actions. In 2007, the DC Circuit (Court) found source categories. The agency acknowledges that the that the EPA had erred in establishing Recent litigation 8 led to a consent Benzene NESHAP provides flexibility emissions standards for sources of HAP decree under which we must propose regarding what factors the EPA might in the NESHAP for Brick and Structural these amendments no later than October consider in making our determinations Clay Products Manufacturing and Clay 31, 2011; and promulgate no later than and how they might be weighed for each Ceramics Manufacturing, 67 FR 26,690 June 29, 2012. source category. In responding to (May 16, 2003), and consequently comment on our policy under the vacated the rule.3 These errors included IV. Mineral Wool and Wool Fiberglass Benzene NESHAP, the EPA explained incorrectly calculated MACT emission Source Categories that: ‘‘The policy chosen by the limits, instances where EPA failed to set A. Overview of the Mineral Wool Administrator permits consideration of emission limits, and instances where Production Source Category and MACT multiple measures of health risk. Not EPA failed to regulate processes that Standards only can the MIR figure be considered, emitted HAP. We are taking action to but also incidence, the presence of correct errors in both the Mineral Wool The NESHAP (or MACT rule) for the noncancer health effects, and the and Wool Fiberglass NESHAP for HAP Mineral Wool Production source uncertainties of the risk estimates. In that are not regulated. Some pollutants category was promulgated on June 1, this way, the effect on the most exposed were represented in the 1999 MACT 1999 (64 FR 29490), and codified at 40 individuals can be reviewed as well as rules by surrogates; other pollutants CFR part 63, subpart DDD. As the impact on the general public. These were not regulated at all in the rule. In promulgated in 1999, the NESHAP factors can then be weighed in each both these cases, we are establishing applies to affected sources of HAP individual case. This approach complies pollutant-specific emission limits. With emissions at mineral wool production with the Vinyl Chloride mandate that the exception of PM as a surrogate for facilities. As defined in the 1992 EPA the Administrator ascertain an all HAP metals, where surrogacy report, ‘‘Documentation for Developing acceptable level of risk to the public by relationships exist, we are proposing to the Initial Source Category List’’ (EPA– employing [her] expertise to assess remove that surrogacy. We are also 450/3/91/030, July 1992), a ‘‘mineral available data. It also complies with the correcting one unregulated HAP- wool facility’’ is ‘‘any facility engaged in Congressional intent behind the CAA, emitting process in the Mineral Wool producing mineral wool fiber from slag, which did not exclude the use of any NESHAP. rock or other materials, excluding sand particular measure of public health risk In two earlier court decisions 45 the or glass.’’ from the EPA’s consideration with court found EPA had erred in not setting The MACT rule for the Mineral Wool respect to CAA section 112 regulations, MACT standards for every HAP emitted Production source category does not and, thereby, implicitly permits from a source. Therefore, with the apply to facilities that manufacture wool consideration of all measures of health exception of PM as a surrogate for HAP fiberglass from sand, feldspar, sodium risk which the Administrator, in [her] metals, in this action we are proposing sulfate, anhydrous borax, boric acid or judgment, believes are appropriate to emission limits for all HAP emitted other similar materials.9 Although there determining what will ‘protect the from Mineral Wool and Wool Fiberglass. public health.’ ’’ We note that we have established 6 Sierra Club v. EPA, 353 F. 3d 976 (DC Cir. 2004). For example, the level of the MIR is through previous analyses upheld by 7 Sierra Club v. EPA, 551 F. 3d 1019 (DC Cir. only one factor to be weighed in 2008), cert. denied, 130 S. Ct 1735 (2010). determining acceptability of risks. The 3 Sierra Club v. EPA, 479 F. 3d 875 (DC Cir. 8 Consent Decree, Sierra Club v. Jackson (No. 09– Benzene NESHAP explains ‘‘an MIR of March 13, 2007). cv–00152SBA, N.D. Cal., Sept. 27, 2010). 4 Cement Kiln Recycling Coalition v. EPA, 255 9 Wool fiberglass produced from sand, feldspar, approximately 1-in-10 thousand should F.3d 855 (DC Cir. 2001) (per curiam). sodium sulfate, anhydrous borax, boric acid, etc. are ordinarily be the upper end of the range 5 National Lime Ass’n v. EPA, 233 F.3d 625 (DC a part of the wool fiberglass source category, which of acceptability. As risks increase above Cir. 2000). is also addressed in this action.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00007 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72776 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

are some similarities among rock that 11 cupolas and 3 curing ovens are of melt) or a percent reduction standard. may be used for both mineral wool and operated by the facilities in this source PM emissions from existing, new, and wool fiberglass production, the two category. reconstructed cupolas are limited to an industries are distinct. Mineral wool is HAP emission sources at mineral outlet concentration of 0.05 kg/Mg (0.10 used in cases in which fireproofing, wool production facilities include the lb/ton) of melt, 40 CFR 63.1178(a). CO structural strength and sound cupola where the mineral charge is emissions limits from new and attenuation are needed, such as in high melted; a collection chamber, in which reconstructed cupolas are limited to an occupancy commercial and industrial air and a binder are drawn over the outlet concentration of 0.05 kg/Mg (0.10 buildings. Wool fiberglass is used fibers, forming them into a mat against lb/ton) of melt or 99 percent CO primarily for insulation, in residential a screen; and a curing oven that bonds removal, 40 CFR 63.1178(a). and small commercial buildings. Some the fibers (for bonded products). HAP Formaldehyde emissions limits from wool fiberglass facilities also operate a are emitted from the cupolas, curing existing, new, and reconstructed curing ceiling tile or pipe product ovens and collection operations when ovens are limited to an outlet manufacturing line. The manufacturing collection occurs with curing. concentration of 0.03 kg/Mg (0.06 lb/ of ceiling tile is not regulated under the Collection at nonbonded product lines ton) of melt or 80 percent formaldehyde Wool Fiberglass Manufacturing MACT does not emit HAP. COS accounts for removal, 40 CFR 63.1179(a). Standard. the majority of the HAP emissions from Today, there are seven mineral wool these facilities (approximately 224 tpy B. Overview of the Wool Fiberglass facilities that are subject to the MACT and 51 percent of the total HAP Manufacturing Source Category and rule. No new mineral wool facilities emissions by mass). The majority of 1999 MACT Rule have been built in the last 21 years and HAP emissions (approximately 58 The NESHAP (or MACT rule) for the the agency does not anticipate new percent of the total HAP by mass, Wool Fiberglass Manufacturing source mineral wool facilities will be built in including HF and HCl are from the category was promulgated on June 14, the foreseeable future. According to the cupolas. The remainder of the HAP are 1999 (62 FR 31695), and codified at 40 size definition applied to this industry from bonded lines, including phenol, CFR part 63, subpart NNN. As by the U.S. SBA (750 company formaldehyde, and methanol. Although promulgated in 1999, the MACT rule employees or less), 5 of the 7 firms, the majority of HAP are emitted from applies to affected sources of HAP employing 540 employees altogether, the cupola, the emissions (primarily emissions at wool fiberglass are classified as a small business. formaldehyde and phenol) that were manufacturing facilities. Although the Mineral wool is a fibrous, glassy significant in evaluating risk are from source category definition includes all substance consisting of silicate fibers the collection chambers on the bonded manufacturers of wool fiberglass, the typically 4 to 7 micrometers in lines. Formaldehyde and phenol are 1999 MACT rule (40 CFR 63.1381) diameter, made from natural rock (such emitted only from bonded mineral wool defines a ‘‘wool fiberglass as basalt, granite and other rock), blast production lines; these lines include manufacturing facility’’ as ‘‘any facility furnace slag, glass cullet, coke and other emissions from the application of the manufacturing wool fiberglass on a RS similar materials. Products made from binder during collection and curing. manufacturing line producing bonded mineral wool are widely used in The current NESHAP requires control building insulation or on a FA thermal and acoustical insulation and of PM emissions, as a surrogate for HAP manufacturing line producing bonded other products where mineral wool fiber metals, from the cupolas and pipe insulation and bonded heavy- is added to impart structural strength or formaldehyde emissions from the curing density products.’’ The MACT rule for fire resistance. In the mineral wool ovens. Fabric filters are the control the Wool Fiberglass Manufacturing manufacturing process, raw materials devices used by this industry to reduce source category does not apply to (e.g., rock and slag) are melted in a both PM and HAP metal emissions from facilities that manufacture mineral wool cupola using coke as fuel; the molten cupolas. Emissions from collection from rock, slag, and other similar material is then formed into fiber. In the operations are not regulated under the materials. In addition, RS and FA production of mineral wool products current NESHAP, but collection and manufacturing lines that produce that do not require high rigidity, oil is curing ovens are generally controlled nonbonded products (in which no typically applied to suppress dust and using RTOs and fabric filters. phenol-formaldehyde binder is applied) add some strength to the fiber; the fiber The existing MACT rule applies to are not subject to the current standards. is then sized and bagged or baled. This each existing, new and reconstructed Wool fiberglass products are primarily is known as a ‘‘nonbonded’’ product cupola or curing oven in a mineral wool used as thermal and acoustical which is manufactured on a production facility. All mineral wool insulation for buildings, automobiles, ‘‘nonbonded’’ production line. production facilities that are major aircraft, appliances, ductwork and For mineral wool products requiring a sources are subject to the standards. For pipes. Other uses include liquid and air higher structural rigidity, typically a all cupolas, the 1999 MACT rule filtration. Approximately 90 percent of phenol/formaldehyde binder may be specifies a numerical emission limit for the wool fiberglass currently produced applied to the fiber. The binder-laden PM, as a surrogate for metal HAP. For is used for residential and commercial fiber mat is then thermoset in a curing new and reconstructed cupolas, building insulation products. Today, oven and cooled. This is known as a emissions limits are specified for CO, as wool fiberglass is currently ‘‘bonded’’ product which is made on a a surrogate for COS. Emissions limits for manufactured in the United States by 5 ‘‘bonded product’’ line. The major formaldehyde are also specified (as a companies operating 29 facilities across differences between the ‘‘nonbonded’’ surrogate for phenol emissions) for each 16 states. According to the size and ‘‘bonded’’ production lines are the existing, new, and reconstructed curing definition applied to this industry by application of binder during fiber oven. Under the 1999 MACT rule, a the U.S. SBA (750 company employees collection and the use of a curing oven. mineral wool production facility may or less), none of these companies are Four facilities only manufacture elect to comply with a numerical classified as a small business. One new nonbonded products, while the other formaldehyde or CO emission limit wool fiberglass facility was recently three facilities operate both bonded and expressed in mass of emissions per unit built in 2007 and one wool fiberglass nonbonded production lines. A total of of production (kg/MG of melt or lb/ton facility closed in 2010. Because several

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00008 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72777

furnaces have been idled across the recent years.10 Most new binder according to the increasingly exposed industry, current production of wool formulations are now HAP-free. refractory surface area.12 13 14 fiberglass is below production levels According to the information collected In addition, organic HAP from previous years, and several months through a survey by the industry, a few (formaldehyde, phenol, and methanol) of stockpiled products exist at wool pipe insulation products made from may be released from RS forming and fiberglass companies, we do not expect wool fiberglass are still made at two curing processes and FA forming and new wool fiberglass facilities to be built facilities using a phenol-formaldehyde curing processes. The 1999 MACT rule applies to in the near future. based binder. process emissions from each of the Wool fiberglass is manufactured in a After application of the binder and following existing, newly constructed, process that forms thin fibers from formation of the mat, the conveyor and reconstructed sources: Glass- molten glass. Over 90 percent of the carries the newly formed mat through melting furnaces located at a wool wool fiberglass industry produces an oven to cure the thermosetting resin fiberglass manufacturing plant, RS insulation; two plants also operate a and then through a cooling section. manufacturing lines that produce pipe product line and one plant building insulation, and FA operates a ceiling tile line (although the Some products, such as those made on manufacturing lines producing pipe production of ceiling tile is not part of FA manufacturing lines, do not require curing and/or cooling. insulation. The MACT rule also applies this MACT standard). A typical wool to FA manufacturing lines producing Process emissions sources include the fiberglass manufacturing line consists of heavy-density products. the following processes: (1) Heating of furnace where the charge is melted; the The 1999 MACT rule requires control raw materials and/or cullet in a furnace collection process, in which air carrying of PM emissions from the glass-melting to a molten state, (2) preparation of a binder is drawn over the fibers, furnaces and formaldehyde emissions molten glass for fiberization, (3) forming them into a mat; and the curing from the RS and FA lines. Typical formation of fibers into a wool fiberglass oven that bonds the fibers (for bonded control devices to reduce PM and HAP mat or pipe insulation product, (4) products only). emissions from furnaces include both curing the binder-coated fiberglass mat, HAP, including chromium wet and dry ESP and fabric filters. Low (5) cooling the mat (this process is not compounds, are emitted from glass- and high-temperature thermal oxidizers always present), and (6) backing, are used to control phenol, cutting, and packaging. melting furnaces. Glass-melting furnaces are constructed using refractory bricks formaldehyde, and methanol from The primary component of most types or blocks (commonly called curing operations on bonded lines. of wool fiberglass is silica sand, but refractories), that provide thermal The 1999 MACT rule limits PM wool fiberglass also includes varying emissions to an outlet concentration of insulation and corrosion protection. The quantities of feldspar, sodium sulfate, 0.50 lb of PM per ton of glass pulled for refractory bricks re-direct the heat of the anhydrous borax, boric acid, and may be both existing and new furnaces, 40 CFR made entirely of glass cullet, crushed furnace back into the melt. Refractories 63.1382. Emissions of formaldehyde recycled glass. Wool fiberglass are produced to withstand the extreme from RS manufacturing lines are limited manufacturing plants typically operate corrosive thermal conditions of a to an outlet concentration of 1.2 lb/ton one or more manufacturing lines. furnace and may contain a variety of of glass pulled for existing sources and Refined raw materials for the glass batch mineral materials, including chromium, 0.80 lb/ton of glass pulled for new are weighed, mixed, and conveyed to and more specifically chromic oxide.11 sources. Emissions of formaldehyde the glass-melting furnace, which may be In a wool fiberglass glass-melting from FA manufacturing lines producing gas-fired, electric, oxygen-enriched or a furnace, sufficient temperatures are pipe insulation are limited to an outlet combination of gas and electric. reached to drive the transformation of concentration of 6.8 lb/ton of glass Two methods of forming fibers are chromium from the trivalent to the pulled from both existing and new used by the industry, RS and FA. In the hexavalent valence state. Because of the sources, 40 CFR 63.1382. Emissions of RS process, centrifugal force causes corrosive properties of the molten glass formaldehyde from FA manufacturing molten glass to flow through small holes and the fining agents (salts added to the lines producing heavy-density products in the wall of a rapidly rotating top of the molten glass layer which act are limited to an outlet concentration of cylinder. In the FA process, molten to draw the gas bubbles out of the 7.8 lb/ton of glass pulled for new glass flows by gravity from a small molten glass), the refractory of the inner sources; no emission limit is specified for existing FA manufacturing lines furnace, or pot, to form threads that are furnace walls are eroded and fresh producing heavy-density products, 40 then attenuated (stretched to the point refractory is continually exposed along of breaking) with air and/or flame. CFR 63.1382. A surrogate approach, the metal/glass line within the furnace. where PM serves as a surrogate for HAP After the fibers are formed, they are As a result, when the glass-melting metals and formaldehyde serves as a sprayed with a binder to hold the fibers furnace is constructed using refractories surrogate for organic HAP, was used in together. These bonded fibers are then containing high percentages of the 1999 MACT rule to allow for easier collected as a mat on a conveyor. Binder chromium, the emission levels of and less expensive testing and compositions vary with product type. At chromium compounds continuously monitoring requirements. the time of development of the MACT increase over the life of the furnace The industry trade association has standard, wool fiberglass mat was advised us that because the wool typically made using a phenol- formaldehyde resin based binder. 12 Notes of April 14, 2011 telephone discussion According to the trade organization, between Carlos Davis, Environmental Manager, 10 only a few insulation products are Letter from the North American Insulation Certainteed, Kansas City, KS; and Susan Fairchild, currently made using a formaldehyde- Manufacturers Association (NAIMA). June 8, 2011 project lead, USEPA/OAQPS/SPPD. Letter. 13 based binder because new Region 7 Certainteed, Kansas City, KS; meeting 11 Chromium in Refractories. Sept. 2000. Dr. and site visit notes. formaldehyde- and HAP-free binder Mariano Velez, Ceramic Engineering Dept., Univ. 14 Emissions Test Results from Certainteed, formulations have been developed in Missouri-Rolla. Kansas City, KS. 2005 and 2008.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00009 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72778 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

fiberglass industry has voluntarily furnaces, curing ovens, forming, and review in developing the current risk phased out most uses of phenol- collection operations. Emissions test assessment; we will be continuing our formaldehyde based binders, there may data provided by facilities in both efforts to improve our assessments by now be only two wool fiberglass source categories, including the incorporating updates based on the SAB facilities that are subject to the current emission unit and pollutant tested, recommendations as they are developed MACT rule. If this is accurate, 27 of the varied widely by facility. and become available. The risk 29 facilities manufacturing wool The mineral wool industry included assessment consisted of seven primary fiberglass may not be considered major testing for most HAP metals, CO, PM steps, as discussed below. The docket sources due to the phaseout of phenol- and certain organic HAP (formaldehyde, for this rulemaking contains the formaldehyde based binders. We are phenol, methanol and COS). Pollutants following document, which provides soliciting comment on our tested for by the wool fiberglass more information on the risk assessment understanding that there will be no manufacturing source category included inputs and models: Draft Residual Risk major sources in the wool fiberglass most HAP metals, including chromium Assessment for the Mineral Wool insulation source category (other than and hexavalent chromium, PM, Production and Wool Fiberglass pipe insulation products) by the end of formaldehyde, phenol and methanol. Manufacturing Source Categories. the 2012 calendar year. The EPA completed the dataset by 1. Establishing the Nature and C. What data collection activities were assigning emission estimates from tested Magnitude of Actual Emissions and conducted to support this action? processes and their known production Identifying the Emissions Release rates to the similar represented Characteristics In June 2010, the industry conducted processes based on production rates at a voluntary survey among all companies the untested processes. A copy of the For each facility in the Mineral Wool that own and operate mineral wool dataset can be found in the docket to Production and Wool Fiberglass production and wool fiberglass this proposed rule. Manufacturing source categories, we manufacturing facilities. The survey The results of these emission tests developed and compiled an emissions sought test data for PM, CO and HAP were compiled into a database for each profile (including emissions estimates, emissions and information on the source category, which is available in stack parameters, and location data) process equipment, control devices, the docket for this action. based on the information provided by point and fugitive emissions, practices the industry survey, the emissions test used to control point and fugitive V. Analyses Performed data, and various calculations. We used emissions, and other aspects of facility the production rates of tested processes A. How did we estimate risks posed by operations. Facilities were asked to seek to assign emissions to untested but the source categories? and obtain prior EPA approval where similar processes based on known new test data for a subset of processes, The EPA conducted a risk assessment production rates at the untested control devices and operations would be that provided estimates of (1) The MIR processes. The site-specific emissions submitted as representative of an posed by the HAP emissions from the 7 profiles include annual estimates of untested subset of processes, control mineral wool facilities and 29 wool process emissions for the 2010 devices and operations. In addition, fiberglass manufacturing facilities in the timeframe, as well as emissions release facilities were allowed, in lieu of source categories, (2) the distribution of characteristics such as emissions release conducting new testing and with prior cancer risks within the exposed height, temperature, velocity, and EPA approval, to submit existing and populations, (3) the total cancer location coordinates. We are requesting well-documented test data that were incidence, (4) estimates of the maximum comment on the assumptions used to representative of current operations TOSHI for chronic exposures to HAP complete the dataset, including using the recommended test methods in with the potential to cause chronic non- assumptions we made to assign the industry survey. Furthermore, the cancer health effects, (5) worst-case emission rates. EPA requested, and industry agreed, screening estimates of HQ for acute The primary risk assessment is based that a subset of the facilities that were exposures to HAP with the potential to on estimates of the actual emissions thought to be representative of emission cause non-cancer health effects, and (6) (though we also analyzed allowable sources from both the mineral wool and an evaluation of the potential for emissions and the potential risks due to wool fiberglass industries would adverse environmental effects. In June allowable emissions). We received a conduct additional emissions testing for of 2009, the EPA’s SAB conducted a substantial amount of emissions test certain HAP from specific processes. formal peer review of the risk data and other information from the The bases for representativeness assessment methodologies used in its industry survey that enabled us to included design type and size of process review of the document entitled, ‘‘Risk derive estimates of stack emissions of units or equipment; fuel type; operating and Technology Review Assessment certain HAP for all of the facilities in temperatures; control devices; and raw Methodologies.’’ 15 We received the both source categories. The wool material content. Facilities completed final SAB report on this review in May fiberglass industry provided emission and submitted responses to the industry of 2010.16 Where appropriate, we have testing on all known pollutants, survey in the spring of 2011. responded to the key messages from this including total chromium and In summary, the EPA received hexavalent chromium, PM, and other existing emissions test data from all 7 15 U.S. EPA, 2009. Risk and Technology Review metals at furnaces they considered to be mineral wool facilities and 26 of the 29 (RTR) Risk Assessment Methodologies: For Review representative of other furnaces wool fiberglass facilities, with some by the EPA’s Science Advisory Board with Case operated by the company. Where Studies—MACT I Petroleum Refining Sources and facilities submitting data for multiple Portland Cement Manufacturing. EPA–452/R–09– different furnace types were used to years. Mineral wool facilities provided 006. Available at http://www.epa.gov/ttn/atw/rrisk/ melt fiberglass, industry usually tested existing test data on cupolas, curing rtrpg.html. representative furnaces for each furnace ovens, and collection operations. Wool 16 U.S. EPA, 2010. SAB’s Response to EPA’s RTR type. The representative furnaces were Risk Assessment Methodologies. Available at: fiberglass facilities provided existing http://yosemite.epa.gov/sab/sabproduct.nsf/ chosen by industry according to test data on one or more of the following 4AB3966E263D943A8525771F00668381/$File/EPA- production rates and furnace type. For emission sources: Glass-melting SAB-10-007-unsigned.pdf. untested furnaces, industry provided

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00010 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72779

the normal operating rate in terms of standards to provide an operational allowable and that the differences tons of glass produced per hour. We cushion and to accommodate the between actual and allowable emissions estimated emissions at untested variability in manufacturing processes are significant. For the facilities furnaces by using data from the and control device performance. producing bonded product, the representative tested furnaces. To do Facilities’ actual emissions may also be estimated actual emissions were up to this, we used test data from significantly lower than MACT- three times lower than allowable representative furnaces that provided allowable emissions for other reasons emissions. That is, MACT-allowable emissions rates of all tested pollutants such as State requirements, emissions were determined to be three on a pound per hour basis. We applied improvements in performance of control times the actual emissions for all this pound per hour basis to the devices since by the MACT standards, pollutants in the Mineral Wool untested furnaces with the known or reduced production. In this case, we Production category. Therefore, we production rates of those furnaces to are reducing the allowable emissions multiplied the actual stack emissions estimate pounds per hour of pollutants. limits to the levels of actual emissions. from each facility by a factor of 3 to We considered furnace type and For this reason, for the pollutants derive estimates of allowable emissions company when making these emitted, we are using only actual for modeling (whether these emissions assignments. emissions in our risk analysis. were measured by testing or calculated We consider these estimates to be For both the Mineral Wool Production based on representative emission tests). very good because they are based upon and Wool Fiberglass Manufacturing b. Analysis of allowable and actual known emission test methods, have test source categories, we evaluated actual emissions for the Wool Fiberglass reports that verify the results, were and allowable stack emissions. Manufacturing source category. signed as being true and accurate by Appendices 1a and 1b of the Draft The analysis of allowable emissions authorized company representatives, Residual Risk Assessment for the for the Wool Fiberglass Manufacturing and also signed as being accurate by the Mineral Wool Production and Wool source category was largely focused on testing company. In addition, one Fiberglass Manufacturing Source emissions of chromium compounds and testing company was used by the Categories, available in the docket, formaldehyde because these are the only industry to conduct all the emissions further describe the estimates of MACT- pollutants emitted with significant testing using approved EPA methods. allowable emissions and the estimates health risks. To estimate the difference We are requesting comment on our use of risks due to allowable emissions. between the actual and allowable of the available test data to assign a. Actual and allowable emissions for emissions, we averaged the actual emission estimates to untested emission the Mineral Wool Production source formaldehyde emission rates of points. category. The analysis of allowable emissions manufacturing lines provided by 2. Establishing the Relationship for the Mineral Wool Production source facilities and compared those values to Between Actual Emissions and MACT- category was largely focused on the maximum level allowed by the Allowable Emissions Levels formaldehyde emissions, which we existing MACT standard (i.e., 1.2 or 0.8 The emissions data in our data set considered the most important HAP lb/ton of glass pulled for formaldehyde). consists of actual stack emissions and, emitted from this source category based We realize that these estimates of where we did not have actual emissions on our screening level risk assessment allowable emissions are theoretical data, estimates of emissions based on a and the HAP for which we had the most high-end estimates as facilities must subset of operations that were data. However, we also considered maintain average emissions levels at representative of such emission points. allowable emissions for other HAP, some level below the MACT limit to In the EPA’s experience, with most including HAP metals and COS. To ensure compliance with the standard at source categories, we generally have estimate the difference between the all times because of the day-to-day found that ‘‘actual’’ emissions levels are actual and allowable emissions, we variability in emissions. Nevertheless, lower than the emissions levels that a averaged the actual formaldehyde these high-end estimates of allowable facility is allowed to emit under the emission rates of manufacturing lines emissions were adequate for us to MACT standards. The emissions levels provided by facilities and compared estimate the magnitude of allowable allowed to be emitted by the MACT those values to the maximum level emissions and the differences between standards are referred to as the ‘‘MACT- allowed by the existing MACT standard the estimates of actual emissions and allowable’’ emissions levels. This (i.e., 0.06 pounds of formaldehyde per the MACT allowable emissions. Based represents the highest emissions level ton of melt) from all curing ovens. on this analysis, we conclude that that could be emitted by facilities We realize that these estimates of allowable emissions are estimated to be without violating the MACT standards. allowable emissions are theoretical three times higher than actual As we discussed in prior residual risk high-end estimates as facilities must emissions. Therefore, to develop the and technology review rules, assessing maintain average emissions levels at MACT-allowable emissions, the actual the risks at the MACT-allowable level is some level below the MACT limit to stack emissions for formaldehyde, reasonable since these risks reflect the ensure compliance with the standard at phenol and methanol were multiplied maximum level at which sources could all times because of the day-to-day by a factor of 3. The range of differences emit while still complying with the variability in emissions. Nevertheless, between actual and allowable MACT standards. However, we also these high-end estimates of allowable formaldehyde emission levels is explained that it is reasonable to emissions were adequate for us to significant, that is, for some sources consider actual emissions, where such estimate the magnitude of allowable there was little difference between data are available, in both steps of the emissions and the differences between actual and allowable emission levels, risk analysis, in accordance with the the estimates of actual emissions and other times, allowable emissions were Benzene NESHAP (54 FR 38044, the MACT allowable emissions. up to 5 times greater than actual September 14, 1989). Considering actual Based on this analysis, we conclude emissions. MACT-allowable emissions emissions is reasonable because source that all facilities in the mineral wool for chromium compounds were categories typically seek to perform source category are emitting determined to be equal to actual better than required by emissions formaldehyde at levels lower than emissions since there is currently no

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00011 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72780 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

emissions limit for chromium estimated annual average ambient air those currently presented in the compounds. concentration of each of the HAP External Review draft of the assessment emitted by each source for which we and will discuss their strengths and 3. Conducting Dispersion Modeling, have emissions data in the source weaknesses. As recommended by the Determining Inhalation Exposures, and category. The air concentrations at each NAS committee, appropriate sensitivity Estimating Individual and Population and uncertainty analyses will be an Inhalation Risks nearby census block centroid were used as a surrogate for the chronic inhalation integral component of implementing the Both long-term and short-term exposure concentration for all the BBDR model. The draft IRIS assessment inhalation exposure concentrations and people who reside in that census block. will be revised in response to the NAS health risks from each source in both We calculated the MIR for each facility peer review and public comments and the source categories addressed in this as the cancer risk associated with a the final assessment will be posted on proposal were estimated using the HEM continuous lifetime exposure (24 hours the IRIS database. In the interim, we (Community and Sector HEM–3 version per day, 7 days per week, and 52 weeks will present findings using the 1991 2.1 Beta). The HEM–3 performs three per year for a 70-year period) to the IRIS value as a primary estimate, and primary risk assessment activities: (1) maximum concentration at the centroid may also consider other information as Conducting dispersion modeling to of an inhabited census block. Individual the science evolves. As described in the estimate the concentrations of HAP in cancer risks were calculated by risk assessment, the IRIS URE for ¥ ambient air, (2) estimating long-term multiplying the estimated lifetime formaldehyde is 1.3 × 10 5 mg/m3, and short-term inhalation exposures to exposure to the ambient concentration whereas, the CIIT URE for formaldehyde ¥ individuals residing within 50 km of the of each of the HAP (in micrograms per is 5.5 × 10 9 mg/m3. modeled sources, and (3) estimating cubic meter) by its URE, which is an Incremental individual lifetime individual and population-level upper bound estimate of an individual’s cancer risks associated with emissions inhalation risks using the exposure probability of contracting cancer over a from the source category were estimated estimates and quantitative dose- lifetime of exposure to a concentration as the sum of the risks for each of the response information. of 1 microgram of the pollutant per carcinogenic HAP (including those The dispersion model used by HEM– cubic meter of air. For residual risk classified as carcinogenic to humans, 3 is AERMOD, which is one of the assessments, we generally use URE likely to be carcinogenic to humans and EPA’s preferred models for assessing values from the EPA’s Integrated Risk suggestive evidence of carcinogenic 19 pollutant concentrations from industrial Information System (IRIS). For potential ) emitted by the modeled 17 facilities. HEM–3 draws on three data carcinogenic pollutants without an EPA source. Cancer incidence and the libraries to perform the dispersion IRIS value, we look to other reputable distribution of individual cancer risks modeling and to develop the sources of cancer dose-response values, for the population within 50 km of any preliminary risk estimates. The first is a often using CalEPA URE values, where source were also estimated for the library of meteorological data, which is available. We may use dose-response source category as part of these used for dispersion calculations. This values in place of or in addition to other assessments by summing individual library includes 1 year of hourly surface values, if appropriate, in cases where risks. A distance of 50 km is consistent and upper air observations for more new, scientifically credible dose- with both the analysis supporting the than 200 meteorological stations, response values have been developed in 1989 Benzene NESHAP (54 FR 38044) selected to provide coverage of the a manner consistent with the EPA and the limitations of Gaussian dispersion models, including AERMOD. United States and Puerto Rico. A second guidelines and have undergone a peer To assess risk of noncancer health library of United States Census Bureau review process similar to that used by 18 effects from chronic exposures, we census block internal point locations the EPA. and populations provides the basis of summed the HQ for each of the HAP human exposure calculations (Census, With regard to formaldehyde, the EPA that affects a common target organ 2000). In addition, for each census determined in 2004 that the CIIT cancer system to obtain the HI for that target block, the Census library includes the dose-response value for formaldehyde organ system (or TOSHI). The HQ for × ¥9 m 3 elevation and controlling hill height, (5.5 10 per g/m ) was based on chronic exposures is the estimated which are used in dispersion better science than the IRIS cancer dose- chronic exposure divided by the chronic × ¥5 m 3 calculations. A third library of pollutant response value (1.3 10 per g/m ) reference level, which is either the EPA unit risk factors and other health and we switched from using the IRIS RfC, defined as ‘‘an estimate (with benchmarks is used to estimate health value to the CIIT value in risk uncertainty spanning perhaps an order risks. These risk factors and health assessments supporting regulatory of magnitude) of a continuous benchmarks are the latest values actions. Based on subsequent published inhalation exposure to the human recommended by the EPA for HAP and research, however, EPA changed its population (including sensitive other toxic air pollutants. These values determination regarding the CIIT model subgroups) that is likely to be without are available at http://www.epa.gov/ttn/ and in 2010 the EPA returned to using an appreciable risk of deleterious effects atw/toxsource/summary.html and are the 1991 IRIS value. The EPA has been discussed in more detail later in this working on revising the formaldehyde 19 These classifications also coincide with the IRIS assessment and the NAS completed terms ‘‘known carcinogen, probable carcinogen and section. possible carcinogen,’’ respectively, which are the In developing the risk assessment for its review of the EPA’s draft in May of terms advocated in the EPA’s previous Guidelines chronic exposures, we used the 2011. The EPA is reviewing the public for Carcinogen Risk Assessment, published in 1986 comments and the NAS independent (51 FR 33992, September 24, 1986). Summing the risks of these individual compounds to obtain the 17 scientific peer review. The EPA will U.S. EPA. Revision to the Guideline on Air cumulative cancer risks is an approach that was Quality Models: Adoption of a Preferred General follow the NAS Report recommended by the EPA’s SAB in their 2002 peer Purpose (Flat and Complex Terrain) Dispersion recommendations and will present review of EPA’s NATA entitled, NATA—Evaluating Model and Other Revisions (70 FR 68218, results obtained by implementing the the National-scale Air Toxics Assessment 1996 November 9, 2005). biologically based dose-response Data—an SAB Advisory, available at: http:// 18 A census block is generally the smallest yosemite.epa.gov/sab/sabproduct.nsf/ geographic area for which census statistics are (BBDR) model for formaldehyde. The 214C6E915BB04E14852570CA007A682C/$File/ tabulated. EPA will compare these estimates with ecadv02001.pdf.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00012 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72781

during a lifetime,’’ or, in cases where an automatically indicate an adverse health increasing but transient and RfC from the EPA’s IRIS database is not impact. nondisabling odor, taste, and sensory available, the EPA will utilize the AEGL values were derived in irritation or certain asymptomatic, following prioritized sources for our response to recommendations from the nonsensory effects.’’ Similarly, the chronic dose-response values: (1) The NRC. As described in Standing document defines AEGL–2 values as Agency for Toxic Substances and Operating Procedures of the National ‘‘the airborne concentration (expressed Disease Registry Minimum Risk Level, Advisory Committee on Acute Exposure as ppm or mg/m3) of a substance above which is defined as ‘‘an estimate of Guideline Levels for Hazardous which it is predicted that the general daily human exposure to a substance Substances (http://www.epa.gov/ population, including susceptible that is likely to be without an opptintr/aegl/pubs/sop.pdf),20 ‘‘the individuals, could experience appreciable risk of adverse effects (other NRC’s previous name for acute exposure irreversible or other serious, long-lasting than cancer) over a specified duration of levels—community emergency exposure adverse health effects or an impaired exposure’’; (2) the CalEPA Chronic REL, levels—was replaced by the term AEGL ability to escape.’’ which is defined as ‘‘the concentration to reflect the broad application of these ERPG values are derived for use in level at or below which no adverse values to planning, response, and emergency response, as described in the health effects are anticipated for a prevention in the community, the American Industrial Hygiene specified exposure duration’’; and (3), as workplace, transportation, the military, Association’s document entitled, noted above, in cases where and the remediation of Superfund Emergency Response Planning scientifically credible dose-response sites.’’ This document also states that Guidelines (ERPG) Procedures and values have been developed in a manner AEGL values ‘‘represent threshold Responsibilities (http://www.aiha.org/ consistent with the EPA guidelines and exposure limits for the general public 1documents/committees/ have undergone a peer review process and are applicable to emergency ERPSOPs2006.pdf) which states that, similar to that used by the EPA, we may exposures ranging from 10 minutes to 8 ‘‘Emergency Response Planning use those dose-response values in place hours.’’ The document lays out the Guidelines were developed for of or in concert with other values. purpose and objectives of AEGL by emergency planning and are intended as Screening estimates of acute stating (page 21) that ‘‘the primary health based guideline concentrations exposures and risks were also evaluated purpose of the AEGL program and the for single exposures to chemicals.’’ 21 for each of the HAP at the point of National Advisory Committee for Acute The ERPG–1 value is defined as ‘‘the highest off-site exposure for each facility Exposure Guideline Levels for maximum airborne concentration below (i.e., not just the census block Hazardous Substances is to develop which it is believed that nearly all centroids), assuming that a person is guideline levels for once-in-a-lifetime, individuals could be exposed for up to located at this spot at a time when both short-term exposures to airborne 1 hour without experiencing other than the peak (hourly) emission rate and concentrations of acutely toxic, high- mild transient adverse health effects or worst-case dispersion conditions (1991 priority chemicals.’’ In detailing the without perceiving a clearly defined, calendar year data) occur. The acute HQ intended application of AEGL values, objectionable odor.’’ Similarly, the is the estimated acute exposure divided the document states (page 31) that ‘‘[i]t ERPG–2 value is defined as ‘‘the by the acute dose-response value. In is anticipated that the AEGL values will maximum airborne concentration below each case, acute HQ values were be used for regulatory and which it is believed that nearly all calculated using best available, short- nonregulatory purposes by U.S. Federal individuals could be exposed for up to term dose-response values. These acute and state agencies and possibly the 1 hour without experiencing or dose-response values, which are international community in conjunction developing irreversible or other serious described below, include the acute REL, with chemical emergency response, health effects or symptoms which could AEGL and ERPG for 1-hour exposure planning and prevention programs. impair an individual’s ability to take durations. As discussed below, we used More specifically, the AEGL values will protective action.’’ conservative assumptions for emission be used for conducting various risk As can be seen from the definitions rates, meteorology and exposure assessments to aid in the development above, the AEGL and ERPG values location for our acute analysis. of emergency preparedness and include the similarly-defined severity prevention plans, as well as real-time levels 1 and 2. For many chemicals, a As described in the CalEPA’s Air emergency response actions, for severity level 1 value AEGL or ERPG has Toxics Hot Spots Program Risk accidental chemical releases at fixed not been developed; in these instances, Assessment Guidelines, Part I, The facilities and from transport carriers.’’ higher severity level AEGL–2 or ERPG– Determination of Acute Reference The AEGL–1 value is then specifically 2 values are compared to our modeled Exposure Levels for Airborne Toxicants, defined as ‘‘the airborne concentration exposure levels to screen for potential an acute REL value (http:// of a substance above which it is acute concerns. www.oehha.ca.gov/air/pdf/acuterel.pdf) predicted that the general population, Acute REL values for 1-hour exposure is defined as ‘‘the concentration level at including susceptible individuals, could durations are typically lower than their or below which no adverse health experience notable discomfort, irritation corresponding AEGL–1 and ERPG–1 effects are anticipated for a specified or certain asymptomatic nonsensory values. Even though their definitions are exposure duration.’’ Acute REL values effects. However, the effects are not slightly different, AEGL–1 values are are based on the most sensitive, disabling and are transient and often the same as the corresponding relevant, adverse health effect reported reversible upon cessation of exposure.’’ ERPG–1 values, and AEGL–2 values are in the medical and toxicological The document also notes (page 3) that, often equal to ERPG–2 values. literature. Acute REL values are ‘‘Airborne concentrations below AEGL– Maximum HQ values from our acute designed to protect the most sensitive 1 represent exposure levels that can screening risk assessments typically individuals in the population by the produce mild and progressively result when basing them on the acute inclusion of margins of safety. Since margins of safety are incorporated to 20 NAS, 2001. Standing Operating Procedures for 21 ERP Committee Procedures and address data gaps and uncertainties, Developing Acute Exposure Levels for Hazardous Responsibilities. November 1, 2006. American exceeding the acute REL does not Chemicals, page 2. Industrial Hygiene Association.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00013 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72782 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

REL value for a particular pollutant. In production specifics. These data are from the SAB’s peer review of EPA’s cases where our maximum acute HQ used to show compliance with RTR risk assessment methodologies,24 value exceeds 1, we also report the HQ Underwriters Laboratories and other we examine a wider range of available value based on the next highest acute building construction safety standards. acute health metrics than we do for our dose-response value (usually the AEGL– From this specific 10-year data set, the chronic risk assessments. This is in 1 and/or the ERPG–1 value). EPA determined that, on a worst-case response to the acknowledgement that To develop screening estimates of possible basis, formaldehyde could be there are generally more data gaps and acute exposures, we developed emitted at levels no more than three inconsistencies in acute reference estimates of maximum hourly emission times the actual rate. The worst-case values than there are in chronic rates by multiplying the average actual scenario is possible if the binder reference values. By definition, the annual hourly emission rates by a factor contained the maximum amount of acute CA–REL represents a health- to cover routinely variable emissions. resin possible, the resin contained the protective level of exposure, with no We chose the factor based on process maximum amount of formaldehyde risk anticipated below those levels, even knowledge and engineering judgment possible, was sprayed at the maximum for repeated exposures; however, the and with awareness of a Texas study of rate possible, and retained in the health risk from higher-level exposures short-term emissions variability, which product at the minimum level possible. is unknown. Therefore, when a CA–REL showed that most peak emission events, These data were used to in the risk is exceeded and an AEGL–1 or ERPG– in a heavily-industrialized four county assessment to determine the acute 1 level is available (i.e., levels at which area (Harris, Galveston, Chambers and health effects hazard index. For Mineral mild effects are anticipated in the Brazoria Counties, Texas) were less than Wool Production, the plant-specific general public for a single exposure), we twice the annual average hourly acute factors were calculated and ranged have used them as a second comparative emission rate. The highest peak from 1.0 to 1.6. Based on these results, measure. Historically, comparisons of emission event was 74 times the annual and to allow for additional uncertainty the estimated maximum off-site one- average hourly emission rate, and the in emissions, we used an acute factor of hour exposure levels have not been 99th percentile ratio of peak hourly 3.0. The calculation we used to typically made to occupational levels emission rate to the annual average determine this acute factor is available for the purpose of characterizing public 22 hourly emission rate was 9. This in the docket to this rule.23 health risks in RTR assessments. This is analysis is provided in Appendix 4 of In cases where acute HQ values from because occupational ceiling values are the Draft Residual Risk Assessment for the screening step were less than or not generally considered protective for the Mineral Wool Production and Wool equal to 1, acute impacts were deemed the general public since they are Fiberglass Manufacturing Source negligible and no further analysis was designed to protect the worker Categories, which is available in the performed. In cases where an acute HQ population (presumed healthy adults) docket for this action. Considering this from the screening step was greater than for short duration (< 15 minute) analysis, unless specific process 1, additional site-specific data were increases in exposure.25 As a result, for knowledge or data are available to considered to develop a more refined most chemicals, the 15-minute provide an alternate value, to account estimate of the potential for acute occupational ceiling values are set at for more than 99 percent of the peak impacts of concern. The data levels higher than a one-hour AEGL–1, hourly emissions, we apply a refinements employed for these source making comparisons to them irrelevant conservative screening multiplication categories consisted of using the site- unless the AEGL–1 or ERPG–1 levels are factor of 10 to the average annual hourly specific facility layout to distinguish exceeded (U.S. EPA 2009). Such is not emission rate in these acute exposure facility property from an area where the the case when comparing the available screening assessments. The factor of 10 public could be exposed. These acute inhalation health effect reference was used for the Wool Fiberglass refinements are discussed in the draft values for formaldehyde (U.S. EPA Manufacturing source category, but we risk assessment document, which is 2009). determined that a factor of 3 is more available in the docket for each of these appropriate for the Mineral Wool source categories. Ideally, we would The worst-case maximum estimated Production source category (for more prefer to have continuous measurements 1-hour exposure to formaldehyde details see the Acute Effects Factor for over time to see how the emissions vary outside the facility fence line for the Mineral Wool Manufacturing Operations each hour over an entire year. Having a mineral wool source category is 0.47 3 document in the docket for this frequency distribution of hourly mg/m . This estimated worst-case rulemaking). emission rates over a year would allow exposure exceeds the 1-hour REL by a For the mineral wool source category, us to perform a probabilistic analysis to factor of 8 (HQREL = 8) and is below the we used data from the highest estimate potential threshold 1-hour AEGL–1 (HQAEGL–1 = 0.4). This formaldehyde emitting source among exceedances and their frequency of exposure estimate does not exceed the the mineral wool producers. That occurrence. Such an evaluation could AEGL–1, or exceed the workplace company also presented the highest risk include a more complete statistical ceiling level guideline for the due to formaldehyde emissions. This treatment of the key parameters and formaldehyde value developed by the company provided the agency with 10 elements adopted in this screening years of measurements of binder analysis. However, we recognize that 24 The SAB Peer review of RTR Risk Assessment formulation, formaldehyde content in Methodoligies is available at: http:// having this level of data is rare, hence yosemite.epa.gov/sab/sabproduct.nsf/ binders, binder application rates, and our use of the multiplier approach. 4AB3966E263D943A8525771F00668381/$File/EPA- binder retention rates. Because the To better characterize the potential SAB-10-007-unsigned.pdf. industry must manufacture their health risks associated with estimated 25 U.S. EPA. (2009) Chapter 2.9 Chemical Specific product for use in fireproofing, they worst-case acute exposures to HAP, and Reference Values for Formaldehyde in Graphical Arrays of Chemical-Specific Health Effect Reference must keep meticulous records of in response to a key recommendation Values for Inhalation Exposures (Final Report). U.S. Environmental Protection Agency, Washington, DC, 22 See http://www.tceq.state.tx.us/compliance/ 23 Acute Factor Memo. Cindy Hancy and David EPA/600/R–09/061, and available on-line at field_ops/eer/index.html or docket to access the Reeves, RTI; to Susan Fairchild, USEPA/OAQPS/ http://cfpub.epa.gov/ncea/cfm/ source of these data. SPPD; EPA Project Lead. August 30, 2011. recordisplay.cfm?deid=211003.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00014 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72783

NIOSH 26 ‘‘for any 15 minute period in first step, we determined whether any Residual Risk Assessment for the a work day’’ (NIOSH REL-ceiling value facilities emitted any PB–HAP in the Mineral Wool Production and Wool 3 of 0.12 mg/m ; HQNIOSH = 4). The environment. There are 14 PB–HAP Fiberglass Manufacturing Source estimate is at the value developed by the compounds or compound classes Categories, which is available in the ACGIH as ‘‘not to be exceeded at any identified for this screening in the EPA’s docket for this action. time’’ (ACGIH TLV-ceiling value of 0.37 Air Toxics Risk Assessment Library 3 5. Assessing Risks After Control Options mg/m ; HQACGIH = 1). Additionally, the (available at http://www.epa.gov/ttn/ estimated maximum acute exposure fera/risk_atra_vol1.html). They are In addition to assessing baseline exceeds the Air Quality Guideline value cadmium compounds, chlordane, inhalation risks and screening for that was developed by the World Health chlorinated dibenzodioxins and furans, potential multi-pathway risks, where Organization 27 for 30-minute exposures dichlorodiphenyldichloroethylene, appropriate, we also estimated risks 3 (0.1 mg/m ; HQWHO = 5). heptachlor, hexachlorobenzene, considering the potential emission For the wool fiberglass manufacturing hexachlorocyclohexane, lead reductions that would be achieved by source category, the worst-case compounds, mercury compounds, the particular control options under maximum estimated 1-hour exposure to methoxychlor, polychlorinated consideration. In these cases, the formaldehyde outside the facility fence biphenyls, polycyclic organic matter, expected emissions reductions were line is 1.92 mg/m3. This estimated toxaphene and trifluralin. applied to the specific HAP and worst-case exposure exceeds the 1-hour Since three of these PB–HAP (lead, emissions sources in the source category REL by a factor of 30 (HQREL = 30) and cadmium, and mercury compounds) are dataset to develop corresponding the 1-hour AEGL–1 (HQAEGL–1 = 2). This emitted by at least one facility in both estimates of risk reductions. More exposure estimate also exceeds multiple source categories, we proceeded to the information on the risks remaining after workplace ceiling level guidelines for second step of the evaluation. In this controls are in place to meet the formaldehyde, including the value step, we determined whether the emissions limits is available in the Draft developed by the American Conference facility-specific emission rates of each of Residual Risk Assessment for the of Governmental Industrial Hygienists the emitted PB–HAP were large enough Mineral Wool Production and Wool (ACGIH) as ‘‘not to be exceeded at any to create the potential for significant fiberglass Manufacturing Source time’’ (ACGIH TLV-ceiling value of 0.37 non-inhalation human or environmental Categories, which is available in the 3 mg/m ; HQACGIH = 5), and the value risks under reasonable worst-case docket for this action. developed by the National Institutes for conditions. To facilitate this step, we 6. Conducting Facility Wide Risk Occupational Safety and Health developed emission rate thresholds for Assessments (NIOSH) ‘‘for any 15 minute period in each PB–HAP using a hypothetical a work day’’ (NIOSH REL-ceiling value worst-case screening exposure scenario To put the source category risks in 3 of 0.12 mg/m ; HQNIOSH = 16). developed for use in conjunction with context, we also examine the risks from Additionally, the estimated maximum the EPA’s TRIM.FaTE model. The the entire ‘‘facility,’’ where the facility acute exposure exceeds the Air Quality hypothetical screening scenario was includes all HAP-emitting operations Guideline value that was developed by subjected to a sensitivity analysis to within a contiguous area and under the World Health Organization 28 for 30- ensure that its key design parameters common control. In other words, for 3 minute exposures (0.1 mg/m ; HQWHO = were established such that each facility that includes one or more 19). Id. environmental media concentrations sources from one of the source We solicit comment on the use of the were not underestimated (i.e., to categories under review, we examine occupational values described above in minimize the occurrence of false the HAP emissions not only from the the interpretation of these worst-case negatives or results that suggest that source category of interest, but also from acute screening exposure estimates for risks might be acceptable when, in fact, all other emission sources at the facility. both the Mineral Wool Production and actual risks are high) and to also For both source categories, all Wool Fiberglass Manufacturing source minimize the occurrence of false significant HAP sources have been categories. positives for human health endpoints. included in the source category risk analysis and there are no other 4. Conducting Multipathway Exposure We call this application of the significant HAP emissions sources and Risk Modeling TRIM.FaTE model TRIM-Screen. The facility-specific emission rates of each of present. Therefore, we conclude that the The potential for significant human the PB–HAP in each source category facility wide risk is essentially the same health risks due to exposures via routes were compared to the TRIM-Screen as the source category risk for both the other than inhalation (i.e., multi- emission threshold values for each of mineral wool and wool fiberglass source pathway exposures) and the potential the PB–HAP identified in the source categories and that no separate facility for adverse environmental impacts were category datasets to assess the potential wide analysis is necessary. evaluated in a three-step process. In the for significant human health risks or 7. Considering Uncertainties in Risk environmental risks via non-inhalation 26 Assessment National Institutes for Occupational Saffety and pathways. Health (NIOSH). Occupational Safety and Health Uncertainty and the potential for bias Guideline for Formaldehyde; http://www.cdc.gov/ None of the facilities in the Mineral niosh/docs/81-123/pdfs/0293.pdf. Wool Production and Wool Fiberglass are inherent in all risk assessments, 27 WHO (2000). Chapter 5.8 Formaldehyde, in Air Manufacturing source categories including those performed for the Quality Guidelines for Europe, second edition. reported emissions of PB–HAP that source categories addressed in this World Health Organization Regional Publications, were greater than the de minimis proposal. Although uncertainty exists, European Series, No. 91. Copenhagen, Denmark. Available on-line at http://www.euro.who.int/_data/ threshold levels, indicating no potential we believe that our approach, which assets/pdf_file/0005/74732/E71922.pdf. for significant multi-pathway risks from uses conservative tools and 28 WHO (2000). Chapter 5.8 Formaldehyde, In Air these facilities. Therefore, multi- assumptions, ensures that our decisions Quality Guidelinies for Europe, second edition. pathway exposures and environmental are health-protective. A brief discussion World Health Organization Regional Publications, European Series, No. 91. Copenhagen, Denmark. risks were deemed negligible and no of the uncertainties in the emissions Available on-line at http://www.euro.who.int_data/ further analysis was performed. This datasets, dispersion modeling, assets/pdf_file/0005/74732/E71922.pdf. analysis is provided in the Draft inhalation exposure estimates and dose-

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00015 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72784 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

response relationships follows below. A assessment. Specifically, short-term emissions from these sources would not more thorough discussion of these mobility and long-term mobility be affected by uncertainty in the length uncertainties is included in the draft between census blocks in the modeling of time emissions sources operate. risk assessment documentation domain were not considered.29 The The exposure estimates used in these (referenced earlier) available in the assumption of not considering short- or analyses assume chronic exposures to docket for this action. long-term population mobility does not ambient levels of pollutants. Because bias the estimate of the theoretical MIR, most people spend the majority of their a. Uncertainties in the Emissions nor does it affect the estimate of cancer time indoors, actual exposures may not Datasets incidence since the total population be as high, depending on the Although the development of the number remains the same. It does, characteristics of the pollutants MACT datasets involved quality however, affect the shape of the modeled. For many of the HAP, indoor assurance/quality control processes, the distribution of individual risks across levels are roughly equivalent to ambient accuracy of emissions values will vary the affected population, shifting it levels, but for very reactive pollutants or depending on the source of the data, the toward higher estimated individual larger particles, these levels are degree to which data are incomplete or risks at the upper end and reducing the typically lower. This factor has the missing, the degree to which number of people estimated to be at potential to result in an overstatement of assumptions made to complete the lower risks, thereby increasing the 25 to 30 percent of exposures.30 datasets are inaccurate, errors in estimated number of people at specific In addition to the uncertainties estimating emissions values and other risk levels. highlighted above, there are several factors. The emission estimates In addition, the assessment predicted factors specific to the acute exposure considered in this analysis generally are the chronic exposures at the centroid of assessment that should be highlighted. annual totals for certain years that do each populated census block as The accuracy of an acute inhalation not reflect short-term fluctuations surrogates for the exposure exposure assessment depends on the during the course of a year or variations concentrations for all people living in simultaneous occurrence of from year to year. that block. Using the census block independent factors that may vary The estimates of peak hourly emission centroid to predict chronic exposures greatly, such as hourly emissions rates, rates for the acute effects screening tends to over-predict exposures for meteorology, and human activity assessment were based on a people in the census block who live patterns. In this assessment, we assume multiplication factor of 10 applied to further from the facility, and under- that individuals remain for 1 hour at the the average annual hourly emission rate, predict exposures for people in the point of maximum ambient which is intended to account for census block who live closer to the concentration as determined by the co- emission fluctuations due to normal facility. Thus, using the census block occurrence of peak emissions and worst- facility operations. centroid to predict chronic exposures case meteorological conditions. These may lead to a potential understatement b. Uncertainties in Dispersion Modeling assumptions would tend to overestimate or overstatement of the true maximum actual exposures since it is unlikely that While the analysis employed the impact, but is an unbiased estimate of a person would be located at the point EPA’s recommended regulatory average risk and incidence. of maximum exposure during the time dispersion model, AERMOD, we The assessments evaluate the cancer of worst-case impact. recognize that there is uncertainty in inhalation risks associated with ambient concentration estimates continuous pollutant exposures over a d. Uncertainties in Dose-Response associated with any model, including 70-year period, which is the assumed Relationships AERMOD. In circumstances where we lifetime of an individual. In reality, both There are uncertainties inherent in had to choose between various model the length of time that modeled the development of the dose-response options, where possible, model options emissions sources at facilities actually values used in our risk assessments for (e.g., rural/urban, plume depletion, operate (i.e., more or less than 70 years), cancer effects from chronic exposures chemistry) were selected to provide an and the domestic growth or decline of and noncancer effects from both chronic overestimate of ambient air the modeled industry (i.e., the increase and acute exposures. Some concentrations of the HAP rather than or decrease in the number or size of uncertainties may be considered underestimates. However, because of United States facilities), will influence quantitatively, and others generally are practicality and data limitation reasons, the risks posed by a given source expressed in qualitative terms. We note some factors (e.g., meteorology, building category. Depending on the as a preface to this discussion a point on downwash) have the potential in some characteristics of the industry, these dose-response uncertainty that is situations to overestimate or factors will, in most cases, result in an brought out in the EPA 2005 Cancer underestimate ambient impacts. For overestimate both in individual risk Guidelines; namely, that ‘‘the primary example, meteorological data were levels and in the total estimated number goal of the EPA actions is protection of taken from a single year (1991) and of cancer cases. However, in rare cases, human health; accordingly, as an agency facility locations can be a significant where a facility maintains or increases policy, risk assessment procedures, distance from the site where these data its emission levels beyond 70 years, including default options that are used were taken. Despite these uncertainties, residents live beyond 70 years at the in the absence of scientific data to the we believe that the approach considered same location, and the residents spend contrary, should be health protective.’’ in the dispersion modeling analysis for most of their days at that location, then (EPA 2005 Cancer Guidelines, pages 1– off-site locations and census block the risks could potentially be 7). This is the approach followed here centroids should generally yield underestimated. Annual cancer as summarized in the next several overestimates of ambient HAP incidence estimates from exposures to paragraphs. A complete detailed concentrations. discussion of uncertainties and 29 Short-term mobility is movement from one c. Uncertainties in Inhalation Exposure micro-environment to another over the course of hours or days. Long-term mobility is movement 30 U.S. EPA. National-Scale Air Toxics The effects of human mobility on from one residence to another over the course of a Assessment for 1996. (EPA 453/R–01–003; January exposures were not included in the lifetime. 2001; page 85.)

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00016 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72785

variability in dose-response the absence of compound-specific data; accounting for an incomplete database relationships is given in the residual where data are available, UF may also on toxic effects of potential concern. risk documentation, which is available be developed using compound-specific Additional adjustments are often in the docket for this action. information. When data are limited, applied to account for uncertainty in Cancer URE values used in our risk more assumptions are needed and more extrapolation from observations at one assessments are those that have been UF are used. Thus, there may be a exposure duration (e.g., 4 hours) to developed to generally provide an upper greater tendency to overestimate risk in derive an acute reference value at bound estimate of risk. That is, they the sense that further study might another exposure duration (e.g., 1 hour). represent a ‘‘plausible upper limit to the support development of reference Not all acute reference values are true value of a quantity’’ (although this values that are higher (i.e., less potent) developed for the same purpose and is usually not a true statistical because fewer default assumptions are care must be taken when interpreting confidence limit).31 In some needed. However, for some pollutants, it the results of an acute assessment of circumstances, the true risk could be as is possible that risks may be human health effects relative to the low as zero; however, in other underestimated. While collectively reference value or values being circumstances, the risk could also be termed ‘‘uncertainty factor,’’ these exceeded. Where relevant to the greater.32 When developing an upper factors account for a number of different estimated exposures, the lack of short- bound estimate of risk and to provide quantitative considerations when using term dose-response values at different risk values that do not underestimate observed animal (usually rodent) or levels of severity should be factored into risk, health-protective default human toxicity data in the development the risk characterization as potential approaches are generally used. To err on of the RfC. The UF are intended to uncertainties. the side of ensuring adequate health- account for: (1) Variation in Although every effort is made to protection, the EPA typically uses the susceptibility among the members of the identify peer-reviewed reference values upper bound estimates rather than human population (i.e., inter-individual for cancer and noncancer effects for all lower bound or central tendency variability); (2) uncertainty in pollutants emitted by the sources estimates in our risk assessments, an extrapolating from experimental animal included in this assessment, some HAP approach that may have limitations for data to humans (i.e., interspecies continue to have no reference values for other uses (e.g., priority-setting or differences); (3) uncertainty in cancer or chronic noncancer or acute expected benefits analysis). extrapolating from data obtained in a effects. Since exposures to these Chronic noncancer reference (RfC and study with less-than-lifetime exposure pollutants cannot be included in a RfD) values represent chronic exposure (i.e., extrapolating from sub-chronic to quantitative risk estimate, an levels that are intended to be health- chronic exposure); (4) uncertainty in understatement of risk for these protective levels. Specifically, these extrapolating the observed data to pollutants at environmental exposure values provide an estimate (with obtain an estimate of the exposure levels is possible. For a group of uncertainty spanning perhaps an order associated with no adverse effects; and compounds that are either unspeciated, of magnitude) of daily oral exposure (5) uncertainty when the database is or do not have reference values for every (RfD) or of a continuous inhalation incomplete or there are problems with individual compound (e.g., glycol exposure (RfC) to the human population the applicability of available studies. ethers), we conservatively use the most (including sensitive subgroups) that is Many of the UF used to account for protective reference value to estimate likely to be without an appreciable risk variability and uncertainty in the risk from individual compounds in the of deleterious effects during a lifetime. development of acute reference values group of compounds. To derive values that are intended to be are quite similar to those developed for Additionally, chronic reference values ‘‘without appreciable risk,’’ the chronic durations, but they more often for several of the compounds included methodology relies upon an UF use individual UF values that may be in this assessment are currently under approach (U.S. EPA, 1993, 1994) which less than 10. UF are applied based on the EPA IRIS review and revised includes consideration of both chemical-specific or health effect- assessments may determine that these uncertainty and variability. When there specific information (e.g., simple pollutants are more or less potent than are gaps in the available information, irritation effects do not vary appreciably the current value. We may re-evaluate UF are applied to derive reference between human individuals, hence a residual risks for the final rulemaking if values that are intended to protect value of 3 is typically used), or based on these reviews are completed prior to our against appreciable risk of deleterious the purpose for the reference value (see taking final action for these source effects. The UF are commonly default the following paragraph). The UF categories and if dose-response metric values,33 e.g., factors of 10 or 3, used in applied in acute reference value changes enough to indicate that the risk derivation include: (1) Heterogeneity assessment supporting this notice may 31 IRIS glossary (http://www.epa.gov/NCEA/iris/ among humans; (2) uncertainty in significantly understate human health help_gloss.htm). extrapolating from animals to humans; risk. 32 An exception to this is the URE for benzene, (3) uncertainty in lowest observed When we identify acute impacts which is considered to cover a range of values, each adverse effect (exposure) level to no which exceed their relevant end of which is considered to be equally plausible benchmarks, we pursue refining our and which is based on maximum likelihood observed adverse effect (exposure) level estimates. adjustments; and (4) uncertainty in acute screening estimates. For the 33 According to the NRC report, Science and Mineral Wool Production source Judgment in Risk Assessment (NRC, 1994) Therefore, default options are not rules that bind category, we used a refined emissions ‘‘[Default] options are generic approaches, based on the agency; rather, the agency may depart from multiplier of 3 to estimate the peak general scientific knowledge and policy judgment, them in evaluating the risks posed by a specific hourly emission rates from the average that are applied to various elements of the risk substance when it believes this to be appropriate. assessment process when the correct scientific In keeping with EPA’s goal of protecting public rates. For a detailed description of how model is unknown or uncertain.’’ The 1983 NRC health and the environment, default assumptions the refined emissions multiplier was report, Risk Assessment in the Federal Government: are used to ensure that risk to chemicals is not developed for the Mineral Wool Managing the Process, defined default option as underestimate risk). See EPA, 2004, An Production source category see the ‘‘the option chosen on the basis of risk assessment Examination of EPA Rick Assessment Principles policy that appears to be the best choice in the and Practices, EPA/100/B–001 available at: http:// memo on the Acute Effects Factor for absence of data to the contrary’’ (NRC, 1983a, p. 63). www.epa.gov/osa/pdfs/ratf-final.pdf. Mineral Wool Manufacturing

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00017 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72786 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

Operations, which is in the docket for protect the public health and prevent transformation in the vicinity of the this action. For the Wool Fiberglass adverse environmental effects. We sources in these categories. Manufacturing source category, data discuss these methods further in The agency understands the potential were not available to develop a refined Sections V.A., VIII. D and VIII. E of this importance of considering an emissions multiplier; therefore, the preamble. individual’s total exposure to HAP in default emissions multiplier of 10 was In past residual risk actions, the EPA addition to considering exposure to used. has presented and considered a number HAP emissions from the source category of human health risk metrics associated and facility. This is particularly e. Uncertainties in the Multi-Pathway with emissions from the category under important when assessing non-cancer and Environmental Effects Assessment review, including: the MIR; the numbers risks, where pollutant-specific exposure We generally assume that when of persons in various risk ranges; cancer health reference levels (e.g., RfC) are exposure levels are not anticipated to incidence; the maximum non-cancer HI; based on the assumption that thresholds adversely affect human health, they also and the maximum acute non-cancer exist for adverse health effects. For are not anticipated to adversely affect hazard (72 FR 25138, May 3, 2007; 71 example, the agency recognizes that, the environment. For each source FR 42724, July 27, 2006). In our most although exposures attributable to category, we generally rely on the site- recent proposals (75 FR 65068, October emissions from a source category or specific levels of PB–HAP emissions to 21, 2010 and 75 FR 80220, December 21, facility alone may not indicate the determine whether a full assessment of 2010), the EPA also presented and potential for increased risk of adverse the multi-pathway and environmental considered additional measures of non-cancer health effects in a effects is necessary. As discussed above, health information, such as estimates of population, the exposures resulting we conclude that the potential for these the risks associated with the maximum from emissions from the facility in types of impacts is low for these source level of emissions which might be combination with emissions from all of categories. allowed by the current MACT standards the other sources (e.g., other facilities) to (see, e.g., 75 FR 65068, October 21, 2010 which an individual is exposed may be f. Uncertainties in the Facility Wide and 75 FR 80220, December 21, 2010). sufficient to result in increased risk of Risk Assessment The EPA also discussed and considered adverse non-cancer health effects. In Given that the same general analytical risk estimation uncertainties. The EPA May 2010, the EPA SAB advised us approach and the same models were is providing this same type of ‘‘* * * that RTR assessments will be used to generate facility wide risk information in support of the proposed most useful to decision makers and results as were used to generate the actions described in this Federal communities if results are presented in source category risk results, the same Register notice. the broader context of aggregate and types of uncertainties discussed above The agency is considering all cumulative risks, including background for our source category risk assessments available health information to inform concentrations and contributions from apply to the facility wide risk our determinations of risk acceptability other sources in the area.’’ 34 assessments. Because the source and ample margin of safety under CAA Although we are interested in placing category processes are the only section 112(f). The agency source category and facility wide HAP processes at each facility, there is no acknowledges that the Benzene risks in the context of total HAP risks greater uncertainty for facility wide NESHAP provides flexibility regarding from all sources combined in the emissions. what factors the EPA might consider in vicinity of each source, we are making determinations and how these concerned about the uncertainties of B. How did we consider the risk results factors might be weighed for each doing so. At this point, we believe that in making decisions for this proposal? source category. Thus, the level of the such estimates of total HAP risks will Based on our risk assessment we are MIR is only one factor to be weighed in have significantly greater associated proposing that risks due to hexavalent determining acceptability of risks. uncertainties than for the source chromium and formaldehyde are The EPA wishes to point out that category or facility wide estimates, and acceptable, with a maximum individual certain health information has not been hence would compound the uncertainty cancer risk for the source category at 40- considered to date in making residual in any such comparison. This is because in-one million. Emissions testing at the risk determinations. In assessing risks to we have not conducted a detailed facility presenting this risk indicated populations in the vicinity of the technical review of HAP emissions data that 92 percent of the total chromium facilities in each category, we present for source categories and facilities that compounds were hexavalent chromium. risk estimates associated with HAP have not previously undergone an RTR In the second step of the process, the emissions from the source category review or are not currently undergoing EPA sets the standard at a level that alone (source category risk estimates) such review. provides an ample margin of safety. and the risks due to HAP emissions We found from our risk assessment from the entire facility at which the C. How did we perform the technology that risks due to hexavalent chromium covered source category is located review? were acceptable at 40-in-one million. In (facility wide risk estimates). We have For our technology review, we the second step of our risk assessment, not attempted to characterize the risks identified and evaluated the we considered whether any cost- associated with all HAP emissions developments in practices, processes effective measures, technologies or impacting the populations living near and control technologies that have practices are available to reduce risks the sources in these categories. That is, further to an ‘‘ample margin of safety’’. at this time, we do not attempt to 34 The EPA’s response to this and all other key We found two methods whereby quantify those HAP risks that may be recommendations of the SAR’s advisory on RTR risk assessment methodologies (which is available hexavalent chromium emissions can be associated with emissions from other at: http://yosemite.epa.gov/sab/sabproduct.nsf/ reduced at wool fiberglass facilities and facilities that are not included in the 4AB3966E263D943A8525771F00668381/$File/EPA- we are proposing in this action emission source categories in question, including SB-10-007-unsigned.pdf) are outlined in a memo to this rulemaking docket from David Guinnup limits for hexavalent chromium from mobile source emissions, natural source entitled, EPA’s Actions in Response to the Key wool fiberglass facilities that will emissions, persistent environmental Recommendations of the SAB Review of RTR Risk provide an ample margin of safety to pollution, and atmospheric Assessment Methodologies.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00018 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72787

occurred since the 1999 MACT rules refractory products including high In addition, we are proposing other were promulgated. In cases where we chrome refractories and consulted various minor changes with regards to identified such developments, we experts in the refractory manufacturing editorial errors and other revisions to analyzed the technical feasibility of and field. promote the use of plain language. The the estimated impacts (costs, emissions Control technologies, classified as analyses and proposed decisions for reductions, risk reductions, etc.) of RACT, BACT, or LAER apply to these actions are presented in Section VI applying these developments. We then stationary sources depending on of this preamble. decided, based on impacts and whether the sources are existing or new, feasibility, whether it was necessary to and on the size, age and location of the E. What analyses were performed for the propose amendments to the regulation facility. We consulted the EPA’s RBLC Mineral Wool Production source to require any of the identified to identify potential technology category under the Regulatory developments. advances. BACT and LAER (and Flexibility Act? Based on our analyses of the data, sometimes RACT) are determined on a Section 609(b) of the RFA requires a information collected under the case-by-case basis, usually by State or Panel to be convened prior to voluntary industry survey, our general local permitting agencies. The EPA publication of the IRFA that an agency understanding of both of the industries established the RBLC to provide a may be required to prepare under the and other available information on central database of air pollution RFA. The RFA directs the Panel to potential controls for these industries, technology information (including report on the comments of small entity we identified potential developments in technologies required in source-specific representatives and make findings on practices, processes, and control permits) to promote the sharing of the following elements: technologies. information among permitting agencies • A description and estimate of the For the purpose of this exercise, we and to aid in identifying future possible number of small entities to which the considered any of the following to be a control technology options that might proposed rule will apply; ‘‘development’’: apply broadly to numerous sources • • A description of projected Any add-on control technology or within a category or apply only on a reporting, recordkeeping and other other equipment that was not identified source-by-source basis. The RBLC compliance requirements of the and considered during development of contains over 5,000 air pollution control proposed rule, including an estimate of the 1999 MACT rules. permit determinations that can help the classes of small entities that will be • Any improvements in add-on identify appropriate technologies to subject to the requirements and the type control technology or other equipment mitigate many air pollutant emissions of professional skills necessary for (that were identified and considered streams. We searched this database to preparation of the report or record; during development of the 1999 MACT determine whether it contained any • An identification, to the extent rules) that could result in significant practices, processes, or control practicable, of all relevant federal rules additional emissions reduction. technologies for the types of processes • Any work practice or operational which may duplicate, overlap, or covered by the Mineral Wool conflict with the proposed rule; and procedure that was not identified or Production and Wool Fiberglass • Descriptions of any significant considered during development of the Manufacturing MACT rules. 1999 MACT rules. Additionally, we requested alternatives to the proposed rule which • Any process change or pollution information from facilities regarding accomplish the stated objectives of prevention alternative that could be developments in practices, processes, or applicable statutes and which minimize broadly applied to the industry and that control technology. Finally, we any significant economic impact of the was not identified or considered during reviewed other information sources, proposed rule on small entities. This development of the 1999 MACT rules. such as State and local permitting analysis must discuss any significant • Any development in equipment or alternatives such as: agency databases and industry- • technology that could result in supported databases. The establishment of differing increased HAP emissions. compliance or reporting requirements or In addition to reviewing the practices, D. What other issues are we addressing timetables that take into account the processes, and technologies that were in this proposal? resources available to small entities; not considered at the time we developed In addition to the analyses described • The clarification, consolidation, or the 1999 MACT rules, we reviewed a above, we also reviewed other aspects of simplification of compliance and variety of data sources for the mineral the MACT standards for possible reporting requirements under the rule wool and wool fiberglass industries. revision. Based on this review we have for such small entities; Among the data sources we reviewed identified several aspects of the MACT • The use of performance rather than were the NESHAP for various industries standards that we believe need revision. design standards; and that were promulgated after the 1999 This includes proposing revisions to the • An exemption from coverage of the MACT rules. We reviewed the startup, shutdown, and malfunction rule, or any part thereof, for such small regulatory requirements and/or provisions of the MACT rule in order to entities. technical analyses associated with these ensure that they are consistent with a Once completed, the Panel Report regulatory actions to identify any recent court decision in Sierra Club v. presents the results of the analyses practices, processes and control EPA, 551 F. 3d 1019 (DC Cir. 2008). identified in the above list, and is technologies considered in these efforts We are proposing HAP-specific provided to the agency issuing the that could possibly be applied to emission limits for COS, phenol, and proposed rule and is included in the emissions sources in the Mineral Wool methanol in place of surrogacy in the rulemaking record. The agency is to Production and Wool Fiberglass source MACT standards. The proposed rule consider the Panel’s findings when categories, as well as the costs, non-air also would regulate the collection completing the draft of the proposed impacts, and energy implications process as a source of HAP emissions of rule. In light of the Panel Report, and associated with the use of these phenol, methanol and formaldehyde where appropriate, the agency is also to technologies. We reviewed scientific that were not included in the 1999 consider whether changes are needed to and technical literature regarding Mineral Wool MACT standard. the IRFA for the proposed rule or the

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00019 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72788 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

decision on whether an IRFA is emissions sources (i.e., the MACT processes immediately precede curing required. floors), recordkeeping and notification, ovens. Both processes emit HAP when The Panel’s findings and discussion compliance and other proposed they occur on bonded production lines, are based on the information available at decisions and actions related to but of the two processes, only the curing the time the final Panel Report is subcategorization of emissions sources oven was regulated under the 1999 published. The EPA will continue to and the findings of the SBAR Panel. MACT standard. This proposed rule conduct analyses relevant to the 1. Addressing Unregulated Pollutants regulates collection and curing as a proposed rule, and additional and Emissions Sources From Mineral combined process on bonded information may be developed or Wool Production production lines under three obtained during the remainder of the subcategories (one subcategory for each rule development process. In the course of evaluating the 1999 combined process design). The Any options identified by the Panel MACT rule, we identified certain HAP proposed emissions limits were for reducing the rule’s regulatory impact for which we failed to establish calculated using the 99 percent UPL on small entities may require further emission standards in the original method. analysis and/or data collection to ensure MACT (i.e., COS, HF, HCl, phenol, and methanol) and certain unregulated We considered beyond-the-floor that the options are practicable, options for COS, HF, and HCl standards enforceable, environmentally sound and processes (i.e., collection). Some of these HAP (COS, phenol, and methanol) for all cupolas, and for formaldehyde, consistent with the CAA and its phenol and methanol for all combined amendments. The Mineral Wool SBAR were not regulated under the 1999 MACT rule because they were collection and curing operation designs, Panel convened on June 2, 2011, to as required by section 112(d)(2) of the address regulatory flexibility represented by surrogates (i.e., CO and formaldehyde). The EPA did not Act. However, we decided not to alternatives and opportunities for the propose any limits based on the beyond- mineral wool industry. regulate HF and HCl in the 1999 rule although these HAP are emitted from the-floor analyses for COS, HF, HCl, VI. Summary of Proposed Decisions cupolas. The 1999 MACT rule also did formaldehyde, phenol, and methanol for and Actions not regulate any HAP emitted from these sources because of the costs, Pursuant to CAA sections 112(d)(2), collection processes that occur on a potential disadvantages of additional 112(d)(6) and 112(f), we are proposing bonded line even though these controls (including the cost of RTO and to revise the 1999 MACT rules relative processes emit the HAP phenol, unintended SO2 emissions), non-air to mineral wool production and wool formaldehyde, and methanol. According environmental impacts, and adverse fiberglass manufacturing to include the to National Lime v. EPA, 233 F.3d 625, energy implications associated with use standards and requirements 634 (DC Cir. 2000), the EPA has a ‘‘clear of these additional controls. The summarized in this section. More statutory obligation to set emissions beyond-the-floor analyses are presented details of the rationale for these standards for each listed HAP.’’ As a in the technical documentation for this proposed standards and requirements part of the information collected in 2010 action (see MACT Floor Analysis for the are provided in Sections VII and VIII of to support this proposal, we specifically Mineral Wool Production Manufacturing this preamble. In addition, as part of evaluated COS, HF, and HCl from Source Category and the MACT Floor these rationale discussions, we solicit cupolas and formaldehyde, phenol and Analysis for the Wool Fiberglass public comment and data relevant to methanol from collection and curing Manufacturing Source Category), and several issues. The comments we operations. are available in the docket for this receive during the public comment For the Mineral Wool Production action. period will help inform the rule source category, we are proposing In summary, we are proposing the development process as we work toward MACT limits for: (1) COS, HF and HCl following emissions limits for existing, promulgating a final action. for existing, new and reconstructed new, and reconstructed cupolas in the cupolas; and (2) formaldehyde, phenol Mineral Wool Production Source A. What are the proposed decisions and and methanol for existing, new and Category as presented in Table 2. We are actions related to the Mineral Wool reconstructed combined collection and not proposing changes to the PM Production NESHAP? curing operations. The collection emissions limits in the 1999 MACT rule The following sections discuss the process emits HAP when a phenol- for Mineral Wool Production, and for proposed decisions and actions formaldehyde based binder is sprayed this reason they are not included in the regarding unregulated pollutants and during collection. Such collection proposed limits in Table 2 below.

TABLE 2—MINERAL WOOL PRODUCTION PROPOSED EMISSIONS LIMITS FOR EXISTING, NEW, AND RECONSTRUCTED CUPOLAS, POUND OF POLLUTANT PER TON OF MELT

Emission limit (lb/ton of melt) Pollutant New and Existing cupolas reconstructed cupolas

COS ...... 3 .3 0 .017 HF ...... 0.014 0.014 HCl ...... 0.0096 0 .0096

2. Subcategorization among classes, types, and sizes of separate subcategories are established, a sources within a category or subcategory MACT floor is determined separately for Under CAA section 112(d)(1), the EPA in establishing * * *’’ standards. When each subcategory. To determine whether has the discretion to ‘‘* * * distinguish

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00020 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72789

the mineral wool production facilities facilities that operate a bonded line), we and other information collected by the warrant subcategorization for the MACT are proposing to subcategorize agency for development of the NESHAP floor analysis, the EPA reviewed unit combined collection operations and for this source category. Based on this and process designs, operating curing ovens designs into three review, the EPA concluded that there information, and air emissions data subcategories based on what the are significant design and operational compiled in the industry survey data set industry is currently using: Vertical, differences in the collection operations and other information collected by the horizontal and drum. When separate at each of the three facilities that operate agency for development of the NESHAP subcategories are established, a MACT a bonded line in this source category. for this source category. Based on this standard is determined separately for The combined collection and curing review, the EPA concluded that there each subcategory. To determine whether designs consist of three design types: are significant design and operational the mineral wool production facilities Vertical, horizontal and drum. For each differences in the collection operations warrant subcategorization for the MACT existing, new, and reconstructed at each of the three facilities that operate floor analysis, the EPA reviewed unit combined collection and curing a bonded line in this source category. and process designs, operating operation, we are proposing the For the unregulated process that emits information and air emissions data following emissions limits as presented HAP (i.e., collection and curing for compiled in the industry survey data set in Table 3.

TABLE 3—MINERAL WOOL PRODUCTION PROPOSED EMISSIONS LIMITS FOR EXISTING, NEW, AND RECONSTRUCTED COMBINED COLLECTION AND CURING OPERATIONS, POUND OF POLLUTANT PER TON OF MELT

Emission limit Design Pollutant (lb/ton of melt)

Vertical ...... Formaldehyde ...... 0.46 Phenol ...... 0.52 Methanol ...... 0.63 Horizontal ...... Formaldehyde ...... 0 .054 Phenol ...... 0.15 Methanol ...... 0.022 Drum ...... Formaldehyde ...... 0.067 Phenol ...... 0.0023 Methanol ...... 0.00077

3. Reporting and Recordkeeping commenced construction or surrogates in the existing rule and to Notifications reconstruction after the effective date of subcategorize the combined collection this rule must demonstrate compliance We are proposing to revise certain operations and curing oven designs with the requirements of this subpart no reporting and recordkeeping from those facilities operating bonded later than the effective date of the rule requirements of 40 CFR part 63, subpart lines. We also discuss here the findings or upon start-up, whichever is later. of the SBAR panel. DDD. Specifically, we are proposing that We are proposing that compliance facilities maintain records and prepare testing for PM, COS, formaldehyde, a. Surrogacy and submit performance test reports on phenol and methanol be conducted As described in Sections III.B and the frequency described below in using the same test methods as required VII.B of this preamble, the court, in the Compliance Dates and Approaches to by the 1999 MACT rule (i.e., Method 5 Brick MACT decision (Sierra Club v. comply with the proposed emissions for PM and Method 318 for the organic limits for COS, HF, HCl, formaldehyde, EPA, 479 F.3d 875 (DC Cir. March 13, HAP). We are proposing that sources 3 phenol, methanol and the existing PM can use either Test Method 26A or Test 2007)) , found that the EPA has a ‘‘clear limit. Although the PM limits in the Method 320 to determine compliance statutory obligation to set emission existing MACT do not change as a result for HF and HCl. standards for each listed HAP,’’ which of this proposed rule we are proposing We are proposing both an initial does not allow it to ‘‘avoid setting the same reporting, recordkeeping performance test and repeat testing standards for HAP not controlled with requirements for PM as for the other every 5 years or more often if the raw technology.’’ Because we did not pollutants addressed under this materials charged to the cupola change conduct analyses that would support proposed rule. We are also proposing by more than 10 percent of that used for the use of CO as a surrogate for COS, or language that would require the use of the initial performance test. Finally, we formaldehyde for methanol and phenol, electronic reporting for all test methods propose that continuous monitoring of we cannot demonstrate that we that are supported by the ERT. Methods appropriate operating parameters for established emission limits for COS, supported by ERT may be found at control devices (e.g., RTO), cupolas, methanol and phenol in the 1999 MACT http://www.epa.gov/ttn/chief/ert/ curing ovens and/or collection standard. Therefore, the agency is index.html. operations will be required as proposing to add emission limits for both phenol and methanol. Similarly, 4. Compliance Dates and Approaches parametric monitoring. This is to ensure continuous compliance with the PM, the agency is proposing to discontinue We are proposing that facilities that COS, HF, HCl, formaldehyde, phenol the use of CO as a surrogate for COS, commenced construction or and methanol emissions limits. and to set emission limits for COS. The reconstruction on or before November proposed emissions limits for 25, 2011 must demonstrate compliance 5. Other Decisions and Actions formaldehyde, phenol, methanol and with the requirements of this subpart no In addition to the proposed decisions COS are presented in Tables 2 and 3, later than 3 years after the effective date and actions discussed above, we are also above. We are soliciting comment on of this rule. Affected sources that proposing changes to the use of our decisions to discontinue use of

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00021 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72790 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

formaldehyde and CO as surrogates; any 1. Number and Types of Entities the EPA does not have the discretion to person wishing to establish or Affected set the MACT floor emission limits at reestablish surrogacy relationships of Six companies exist in this industry; levels suggested by the SER. The Panel one pollutant for others should provide five of the six companies are small recognizes that EPA has the authority to emissions testing to support their businesses. All small businesses in the review the MACT standard for conclusions. mineral wool production industry completeness, risk, and technology operate under NAICS code 327993. improvements, and that the agency is b. Small Business Advocacy Review currently under court order to conduct Panel 2. Recordkeeping, Reporting and Other the risk and technology review for the For purposes of assessing the impacts Compliance Requirements mineral wool source category and of the proposed rule on small entities, The proposed rule under propose amendments to the standard by the RFA defines small entities as consideration potentially impacts small October 31, 2011, and promulgate the including ‘‘small businesses,’’ ‘‘small businesses by requiring new emission amendments by October 31, 2012. governments,’’ and ‘‘small limits on processes that were not However, whenever opportunities for organizations’’ (5 U.S.C. 601). The regulated under the MACT standard regulatory flexibility arise, and when regulatory revisions being considered by promulgated in 1999, by requiring that regulatory flexibility can work to the EPA for this rulemaking are emission limits for pollutants that were lessen impacts to small businesses, the expected to affect a variety of small not regulated under the MACT, or both Panel recommends that the EPA businesses, but would not affect any processes and pollutants not regulated propose amendments to the mineral small governments or small under the MACT. All companies are wool MACT that offer such regulatory organizations. The RFA references the subject to Title V operating permits flexibility to the maximum extent definition of ‘‘small business’’ found in requirements, and as such will be possible. Specifically, these the Small Business Act, which required to add the newly regulated opportunities arise in the following processes to their operating permits situations: authorizes the SBA to further define • ‘‘small business’’ by regulation. The along with compliance demonstrations Selection of the averaging method SBA definitions of small business by that the processes meet each pollutant in calculating the MACT floor for COS size standards using the NAICS can be emission limit in the rule. Compliance from cupolas and phenol, formaldehyde found at 13 CFR 121.201. For the testing will be required to be conducted and methanol emissions from collection Mineral Wool Production source using EPA methods for each pollutant. and curing processes; and • Subcategorization of regulated category (NAICS code 327993), the SBA Reporting and recordkeeping processes, when appropriate. size standard for a small business is 500 requirements are not expected to change from the MACT, with the exception of The Panel recommends that the EPA employees. Based on this size not require BTF emission limits for the designation, there are currently 5 small additional pollutants and processes included in such reports. mineral wool industry. Such limits are businesses operating with a total likely to have additional cost impacts to number of 540 employees. 3. Related Federal Rules industry. In addition, the EPA did not Under section 609(b) of the RFA, the NAAQS: the most prevalent identify BTF measures for consideration Panel is to report its findings related to technology for reducing COS emissions and has found that the results of the risk these four items: will increase emissions of SO2. Under assessment show acceptable risks from • A description of and, where the current NAAQS, none of the small this source category. feasible, an estimate of the number of entities are in nonattainment areas, so The Panel recommends small entities to which the proposed installation of emissions control subcategorization of collection along the rule will apply; equipment should not subject them to lines described in Section 3 of the Panel • A description of the projected additional permitting requirements Report, specifically, subcategorization reporting, recordkeeping and other under the SO2 NAAQS. However, the for vertical collection and curing, compliance requirements of the EPA cannot make such assurances about horizontal collection and curing, and proposed rule, including an estimate of future NAAQS or future nonattainment drum collection and curing. Based on the classes of small entities which will zones, so there is a risk that future available information, the Panel believes be subject to the requirement and the compliance with this rule could trigger that emission standards based on the type of professional skills necessary for additional emissions control average emission limits across both preparation of the report or record; requirements through the Title V/ collection and curing processes at each of the three subcategories would • Identification, to the extent prevention of significant deterioration minimize the burden on small entities practicable, of all relevant federal rules permit program. while fully complying with the EPA’s which may duplicate, overlap or GHG: Most emissions control obligations under section 112. The conflict with the proposed rule; and strategies identified by the EPA during • the Panel would increase the energy Panel also recommends setting MACT A description of any significant intensity of mineral wool production. limits for new sources equal to MACT alternatives to the planned proposed Although the Panel does not have limits for existing sources. rule which would minimize any specific information about the GHG The Panel recommends that the EPA significant economic impact of the emissions of individual facilities in this allow the maximum amount of time proposed rule on small entities industry, these facilities could be within its discretion (3 years) and work consistent with the stated objectives of subject to GHG permitting as that with state permitting authorities to the authorizing statute. program is phased in under the provide for the additional year The Panel’s most significant findings Tailoring Rule. permitted by the statute. and discussion with respect to each of The Panel recommends that the EPA these items are summarized below. To 4. Regulatory Flexibility Alternatives provide a detailed discussion in the read the full discussion of the Panel The Panel agrees that the EPA does preamble to the proposed rule that findings and recommendations, see not have discretion in a number of areas outlines the manner in which small Section 9 of the Panel Report. that SER commented upon. Specifically, entities may demonstrate compliance

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00022 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72791

with the rule, when finalized, during something called an ‘‘incinerator’’. We emissions standards for each listed start-up and shutdown. The Panel also are also proposing to specify HAP’’. National Lime v. EPA, 233 F. 3d recommends that the EPA propose performance testing frequency for RTOs. 625, 634 (DC Cir. 2000). The EPA allowing an affirmative defense against B. What are the proposed decisions and specifically evaluated HF and HCl, from compliance actions for malfunction actions related to the Wool Fiberglass glass-melting furnaces and events, consistent with other section Manufacturing NESHAP? formaldehyde, phenol and methanol 112 rules recently promulgated. For from RS manufacturing lines and FA more information on the SBAR Panel The following sections discuss the manufacturing lines. review process and findings, see Section decisions proposed by this action with IV.E of this preamble and the Final regard to the following topics: a. Surrogacy Report of the Small Business Advocacy unregulated pollutants and emissions As described in Sections III. B and Review Panel on the EPA’s Planned sources; the risk review; the technology VII.B of this preamble, the Court, in the Proposed Rule Risk and Technology review; our plans regarding area Brick MACT decision, also found that Review (RTR) Amendments to the sources; recordkeeping, reporting and National Emission Standard for notification requirements; compliance the EPA erred when we did not Hazardous Air Pollutants (NESHAP) for requirements; and other proposed establish emission limits for each HAP Mineral Wool Production October 2011 decisions and actions (i.e., changes in emitted from industrial processes in the docket. surrogacy and terminology cleanup). regulated by the MACT standard. We are proposing to replace CO as a c. Technical Corrections to the Rule 1. Addressing Unregulated Pollutants surrogate for COS with COS emissions We are also proposing revisions to and Emissions Sources limits. We are also proposing to certain terms in the existing NESHAP. In the course of evaluating the 1999 discontinue use of formaldehyde as a Specifically, we are proposing to replace MACT rule, we identified certain HAP surrogate for phenol and methanol. We the term ‘‘incinerator’’ with for which we failed to establish are, therefore, proposing to add ‘‘regenerative thermal oxidizer’’ to avoid emission standards in the original emission limits for COS, phenol and confusion with rules promulgated under MACT (i.e., HF, HCl, phenol and methanol. The proposed emissions CAA section 129 and any new methanol). As stated earlier, the EPA limits can be found in Tables 4–6, requirement that may be imposed on has ‘‘clear statutory obligation to set below.

TABLE 4—PROPOSED EMISSIONS LIMITS FOR ROTARY SPIN (RS) MANUFACTURING LINES [Pound of pollutant/ton of melt]

New and Pollutant Existing RS reconstructed lines RS lines

Formaldehyde ...... 0.17 0 .020 Phenol ...... 0.19 0 .0011 Methanol ...... 0.48 0 .00067

TABLE 5—PROPOSED EMISSIONS LIMITS FOR FLAME ATTENUATION (FA) MANUFACTURING LINES [Pound of pollutant/ton of melt]

New and Pollutant Existing FA reconstructed lines FA lines

Formaldehyde ...... 5.6 3.3 Phenol ...... 1.4 0.46 Methanol ...... 0.50 0 .50

TABLE 6—PROPOSED EMISSIONS LIMITS FOR GLASS-MELTING FURNACES [Pound of pollutant/ton of melt]

New and Pollutant Existing reconstructed furnaces furnaces

HF ...... 0.002 0.00078 HCl ...... 0.0015 0 .00078

b. Emission Limits for Unregulated from existing, new, and reconstructed Section 112(d)(3)(B) of the CAA HAPs RS manufacturing lines; and requires that the MACT standards for formaldehyde, phenol and methanol existing sources be at least as stringent For the Wool Fiberglass from existing, new, and reconstructed as the average emissions limitation Manufacturing source category, we are proposing MACT limits for HF and HCl FA manufacturing lines. The proposed achieved by the best performing 12 for glass-melting furnaces; emissions limits can be found in Tables percent of sources (for which the formaldehyde, phenol and methanol 4–6 above. Administrator has or could reasonably

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00023 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72792 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

obtain emissions information) in a tons of glass pulled. As explained in in part from these types of refractories. category with more than 30 sources. The Section VIII of this preamble, we are (Notes of April 14, 2011; Region 7 Wool Fiberglass Manufacturing source proposing these limits as an outcome of Certainteed Notes).12 13 category consists of 29 facilities with our ample margin of safety analysis. It is our understandng that because of approximately 80 glass-melting the corrosive properties of the molten 3. Proposed Decisions Based on the furnaces. Since there are more than 30 glass, fresh refractory is continuously Technology Review for the Wool furnaces, we based the MACT floor limit exposed to the molten glass along the Fiberglass Industry on the average emissions limitation metal/glass contact line in the glass- achieved by the best performing 12 As explained in Sections VI.B and melting furnace process. This increases percent of furnaces. VIII.E of this preamble, we are the surface area of the refractory that is The EPA must exercise its judgment, proposing emissions limits for PM, exposed to the molten glass. As a result, based on an evaluation of the relevant under section 112(d)(6) (see Table 12 of when the glass furnace is constructed factors and available data, to determine Section VIII in this preamble). using high chrome refractories, the the level of emissions control that has Furthermore, as explained in Section emission levels of chromium been achieved by the best performing VIII.F of this preamble, we are compounds continuously increase over sources under variable conditions. It is proposing emissions limits for the life of the furnace (Please refer to recognized in the case law that the EPA chromium compounds under section notes of April 14, 2011, telephone may consider variability in estimating 112(d)(6) of the CAA as part of our discussion between Susan Fairchild and the degree of emissions reduction technology review (see those sections Certainteed). One industry achieved by best-performing sources for details) of 0.006 pounds of total spokesperson estimated that 20,000 lb/ and in setting MACT floors. See chromium per thousand tons of glass yr of refractory are worn away from the Mossville Envt’l Action Now v. EPA, 370 pulled, which is the same limit we are inside walls of one wool fiberglass F.3d 1232, 1241–42 (DC Cir 2004) proposing under Section 112(f)(2) of the furnace and ducted to the control device (holding that the EPA may consider CAA. before venting to the atmosphere.35 emissions variability in estimating In our technology review for this On August 31, 2011, industry performance achieved by best- industry, we discovered and evaluated representatives met with the agency to performing sources and may set the two new technology developments that provide data, in an attempt to improve affect emissions from wool fiberglass floor at a level that a best-performing our understanding of the levels of manufacturing furnaces: furnace control source can expect to meet ‘‘every day chromium content in refractory technologies and high chrome and under all operating conditions’’). products used at wool fiberglass refractories. These are discussed below. More details on how we calculate furnaces and their impacts on MACT floors and how we account for Wool fiberglass furnaces are now equipped with air pollution control chromium compound emissions. In the variability are described in the MACT meeting industry representatives stated Floor Analysis for the Wool Fiberglass devices that achieve emissions of about 0.014 pounds PM per ton of glass the following: Manufacturing Source Category which • The use of chromium in refractories is available in the docket for this produced. This is about 50 times lower than required under the MACT rule (0.5 is important to wool fiberglass proposed action. operations because it extends the useful We considered beyond-the-floor lb PM per ton glass produced). In light of the record and additional data we life of the furnace; options for the HF and HCl standards • Chromium content of furnaces vary for all of the glass-melting furnaces and received on PM emissions, we are proposing revised PM limits under the from 0 to 95 percent; there is no the formaldehyde, phenol and methanol distinction between the types of standards for all RS manufacturing lines technology review of the wool fiberglass source category (as described in Section refractories used at the highest chrome and FA manufacturing lines, as required emitting furnace and the refractories by section 112(d)(2) of the Act. We VIII of this preamble). used to construct other glass furnaces decided not to propose any limits based Glass-melting furnaces are that emit low levels of hexavalent on the beyond-the-floor analyses for any constructed using refractories, which chromium. of these pollutants because of the costs, direct the heat of the furnace back into • The type of furnace used at the high non-air environmental impacts, and the melt. We are aware of a new chromium emitting facility may may be adverse energy implications associated technology that is used to significantly responsible for increased hexavalent with use of these additional controls. extend the life of the wool fiberglass furnace: refractories that are made of chromium emissions. The beyond-the-floor analysis is However, the information from the presented in the technical almost 100 percent chromium compounds and that are used to meeting appears to contradict other documentation for this action (MACT information on the reason for certain Floor Analysis for the Mineral Wool construct entire furnaces or very large parts of furnaces. Based on emission furnaces to have elevated chromium Production Source Category and the emissions. As previously discussed, MACT Floor Analysis for the Wool testing of one furnace, it appears that the levels of chromium compounds that emission test results from the 2010 Fiberglass Manufacturing Source testing and previous statements made to Category). can be emitted when glass-melting furnaces are constructed from high the EPA from owners/operators (Notes 2. Proposed Decisions Based on the Risk chrome refractories can be significant. of April 14, 2011, Certainteed; Region 7 Review This facility operates two furnaces. The Certainteed notes) seem to inply that the Based on the results of our risk total chromium compound emissions at high chromium emissions are due to the assessment and risk review (which are this facility are estimated as 913 lb/yr chromium content of the refractory. described in more detail in Section VIII assuming that both furnaces emit at a Because of this contradictory of this preamble), we are proposing similar rate. This includes 840 pounds information we are requesting emission limits for chromium of hexavalent chromium. Industry 35 Meeting between U.S. EPA, would fiberglass compounds under the authority of information indicates that the furnaces industry representatives and NAIMA (trade section 112(f)(2) of the CAA of 0.006 emitting the highest levels of chromium association). August 31, 2011. At USEPA offices in pounds of total chromium per thousand compounds are constructed in whole or Research Triangle Park, NC.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00024 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72793

additional emissions testing of wool manufacturing furnaces maintain reflect reductions required for the glass fiberglass furnaces (discussed below). records of the refractory brick making furnaces pursuant to CAA We are also soliciting comment on composition from which the furnaces section 112(d)(6) must be met no later whether and how to subcategorize are constructed, including any than one year after the effective date of industry according to furnace type, or rebricking or additional layers of this rule. We believe this time is needed type of refractory. Commenters should refractory that are added to the outside to either enable installation of also provide emissions test data to furnace walls. In addition, owners and replacement bags, or if a facility decides support their assertions regarding the operators are required to keep records of to add a new baghouse in series with an correct manner in which to the occurrence and duration of each existing baghouse, seek bids, select a subcategorize the industry. malfunction or operation of the air vendor, install and test the new As shown in Table 12 of Section VIII pollution control equipment and equipment; prepare and submit the of this preamble, we are proposing monitoring equipment. We are also reports in this proposed rule, if chromium compound emissions limits proposing requirements for the use of finalized. of 0.00006 lb/ton of glass produced. electronic reporting for all test methods Therefore, we are proposing that wool These limits would apply to wool that are supported by the ERT. Methods fiberglass facilities would be required to fiberglass furnaces at major sources. supported by ERT may be found at show compliance with both PM and the However, there are no differences in http://www.epa.gov/ttn/chief/ert/ chromium limits within 1 year of furnaces at major sources and those at index.html. promulgation of this standard. We are area sources. We are concerned about soliciting comments on this aspect of the levels of hexavalent chromium that 5. Compliance Dates and Approaches this proposed action. can be emitted by area sources where With regard to formaldehyde, HCl, Additionally, we propose that furnaces may be constructed or HF, phenol and methanol, we are compliance with the proposed reconstructed using high chrome proposing that facilities that chromium compounds emissions limits refractories. We are announcing today commenced construction or be demonstrated by annual performance our plans to regulate wool fiberglass reconstruction on or before November tests for all glass-melting furnaces area sources in a future action. We have 25, 2011 must demonstrate compliance subject to this rule as described in issued a section 114 information with the requirements of this subpart no Section VI.B.2 of this preamble. We are collection request to the wool fiberglass later than 3 years after the effective date proposing additional annual industry to collect comprehensive of this rule. Affected sources that performance testing no later than 12 information specific to the chrome commenced construction or calendar months following the initial or content of the refractories used to reconstruction after the proposal date of previous performance or compliance construct their glass-melting furnaces this rule must demonstrate compliance test to demonstrate compliance with the and obtain complete chromium with the requirements of this subpart no chromium compounds emissions limit emissions test data. This information later than the effective date of the rule for furnaces. will enable us determine the scope of or upon start-up, whichever is later. We We are proposing both an initial the source category (in terms of the are proposing an initial performance test performance test and repeat testing universe of wool fiberglass producers within 90 days of promulgation of the every 5 years on the RS and FA lines that are area sources and that emit final rule. and each time the binder formulation hexavalent chromium) to be regulated in With regard to total chromium changes by more than 10 percent as the future action. compounds, we are proposing that the compared to the binder formulation We are requesting information requirements under CAA section used in the initial performance test. We specific to wool fiberglass furnaces, 112(f)(2), if finalized, must be are seeking comment on whether the including information on the chromium implemented no later than 90 days after binder formulation variability of 10 content of the refractories used in the effective date of this rule, but the percent as used here is appropriate. furnace construction, process rates and EPA may extend that timeframe for We are proposing that compliance emissions testing. Nevertheless, we are circumstances under which we believe testing for PM, formaldehyde, phenol soliciting comment from the public on the additional time is necessary for and methanol be conducted using the our approach to limit emissions of installation of air pollution control same test methods as required by the chromium compounds as well as other equipment or other measures to reduce 1999 MACT rule (i.e., Method 5 for PM alternatives to reducing emissions of HAP emissions. We are, therefore, and Method 318 for formaldehyde, chromium compounds, especially allowing affected sources up to one year phenol and methanol). We are hexavalent chromium. from the effective date of this rule to proposing Test Method 26A be used to demonstrate compliance with the determine compliance for HF and HCl 4. Reporting, Recordkeeping and chromium emission limits. Consistent and Test Method 0061 be used to ensure Notification Requirements with CAA section 112(f)(4)(B), we are compliance with the chromium We are proposing to revise certain proposing that a one-year compliance compounds emission limit. recordkeeping requirements of 40 CFR period is necessary so that affected We propose that continuous part 63, subpart NNN. Specifically, we facilities have adequate time to install monitoring of temperatures of control are proposing that facilities maintain additional controls and demonstrate devices (e.g., fabric filters, wet and dry records and prepare and submit compliance, including the time ESP, scrubbers) for glass-melting performance test reports to comply with necessary to purchase, install and test furnaces, RS manufacturing lines, and the proposed emissions limits for PM, control equipment. Because these limits FA manufacturing lines will be required chromium compounds, HF, HCl, reflect the reductions from glass making as parametric monitoring to ensure formaldehyde, phenol and methanol. furnaces required under both sections continuous compliance with the PM, Because refractory products can contain 112(d)(6) and 112(f)(2), we believe a chromium compounds, HF, HCl, chromium compounds that can then be one-year compliance timeframe is formaldehyde, phenol and methanol emitted to the ambient air during wool needed for the same reasons provided emissions limits. fiberglass manufacturing, we are above. In addition, we are proposing Because the recent test data for glass- proposing that owners/operators of glass that the PM emissions limit that would melting furnaces show a significant

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00025 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72794 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

portion of the chromium compounds are Consistent with Sierra Club v. EPA, the determined that CAA section 112 does hexavalent chromium, we are requiring EPA is proposing standards in this rule not require that emissions that occur Test Method 0061 be used to ensure that apply at all times. We are also during periods of malfunction be compliance with the chromium proposing several revisions to Table 1 to factored into development of CAA compounds emission limit and as the subparts DDD and NNN of part 63 (the section 112 standards. Under CAA most cost effective method to determine General Provisions Applicability table). section 112, emissions standards for both total chromium and hexavalent For example, we are proposing to new sources must be no less stringent chromium from wool fiberglass furnace eliminate the incorporation of the than the level ‘‘achieved’’ by the best stacks. Sources must report both total General Provisions’ requirement that the controlled similar source and for chromium and hexavalent chromium source develop an SSM plan. We also existing sources generally must be no using this method or all chromium are proposing to eliminate or revise less stringent than the average emissions emissions are assumed to be hexavalent certain recordkeeping and reporting that limitation ‘‘achieved’’ by the best chromium. related to the SSM exemption. The EPA performing 12 percent of sources in the has attempted to ensure that we have category. There is nothing in CAA 6. Other Decisions and Actions not included in the proposed regulatory section 112 that directs the agency to In addition to the proposed decisions language any provisions that are consider malfunctions in determining and actions discussed above, we are also inappropriate, unnecessary, or the level ‘‘achieved’’ by the best proposing surrogacy changes and some redundant in the absence of the SSM performing or best controlled sources general cleanup in terminology to the exemption. We are specifically seeking when setting emissions standards. existing rule. comment on whether there are any such Moreover, while the EPA accounts for a. Surrogacy provisions that we have inadvertently variability in setting emissions As described in Sections III.B and incorporated or overlooked. standards consistent with the CAA VIII.B in this preamble, the Court found section 112 case law, nothing in that that the EPA has a ‘‘clear statutory In proposing the standards in this rule, the EPA has taken into account case law requires the agency to consider obligation to set emission standards for malfunctions as part of that analysis. each listed HAP.’’ Because we did not startup and shutdown periods and, for the reasons explained below, is Section 112 of the CAA uses the concept conduct analyses that would support of ‘‘best controlled’’ and ‘‘best the use of formaldehyde as a surrogate proposing emissions limits for those periods. Information on periods of performing’’ unit in defining the level of for methanol and phenol, we cannot stringency that CAA section 112 currently demonstrate that we startup and shutdown received from the industry survey indicate that emissions performance standards must meet. established emission limits for the HAP Applying the concept of ‘‘best methanol and phenol in the 1999 MACT during these periods are less than emissions during production. Control controlled’’ or ‘‘best performing’’ to a standard. Therefore, we are proposing unit that is malfunctioning presents the emissions limits for phenol and devices such as baghouses for PM and metal HAP particulate control and RTO significant difficulties, as malfunctions methanol, which are presented in are sudden and unexpected events. Tables 4–6, above. for COS control are started up before the b. Technical corrections to the rule. process units, and are operational Further, accounting for malfunctions We are also proposing revisions to during the shutdown phase of a process. would be difficult, if not impossible, certain terms in the existing NESHAP. Therefore, no increase in emissions is given the myriad different types of Specifically, we are proposing to replace expected during these periods. Because malfunctions that can occur across all the term ‘‘incinerator’’ with ‘‘RTO’’ and the processes are ducted to the control sources in the category and given the specify performance test frequency. device before startup and after difficulties associated with predicting or shutdown, and because emissions accounting for the frequency, degree C. What are the proposed decisions and during startup and shutdown are not and duration of various malfunctions actions related to startup, shutdown and more than emissions during production, that might occur. As such, the malfunction? startup and shutdown emissions limits performance of units that are The United States Court of Appeals should be equivalent to the emissions malfunctioning is not ‘‘reasonably’’ for the District of Columbia Circuit limits for production. Production based foreseeable. See, e.g., Sierra Club v. vacated portions of two provisions in emissions limits are expressed in this EPA, 167 F. 3d 658, 662 (DC Cir. 1999) the EPA’s CAA section 112 regulations rule on a pound of pollutant per ton (the EPA typically has wide latitude in governing the emissions of HAP during melt basis. However, during startup and determining the extent of data-gathering periods of SSM. Sierra Club v. EPA, 551 shutdown, there is no melt being necessary to solve a problem. The court F.3d 1019 (DC Cir. 2008), cert. denied, produced. Therefore, separate standards generally defers to the agency’s decision 130 S. Ct. 1735 (U.S. 2010). Specifically, for periods of startup and shutdown to proceed on the basis of imperfect the Court vacated the SSM exemption were developed by translating the scientific information, rather than to contained in 40 CFR 63.6(f)(1) and 40 production-based emissions limits from ‘‘invest the resources to conduct the CFR 63.6(h)(1), that are part of a a pound per ton basis to a pound of perfect study.’’). See also, Weyerhaeuser regulation, commonly referred to as the pollutant per hour basis and are being v. Costle, 590 F.2d 1011, 1058 (DC Cir. ‘‘General Provisions Rule,’’ that the EPA proposed in this rule. Periods of startup, 1978) (‘‘In the nature of things, no promulgated under CAA section 112. normal operations and shutdown are all general limit, individual permit or even When incorporated into CAA section predictable and routine aspects of a any upset provision can anticipate all 112(d) regulations for specific source source’s operations. However, by upset situations. After a certain point, categories, these two provisions exempt contrast, malfunction is defined as a the transgression of regulatory limits sources from the requirement to comply ‘‘sudden, infrequent, and not reasonably caused by ‘uncontrollable acts of third with the otherwise applicable CAA preventable failure of air pollution parties,’ such as strikes, sabotage, section 112(d) emissions standard control and monitoring equipment, operator intoxication or insanity, and a during periods of SSM. process equipment or a process to variety of other eventualities, must be a We are proposing the elimination of operate in a normal or usual manner matter for the administrative exercise of the SSM exemption in this rule. * * *’’ (40 CFR 63.2). The EPA has case-by-case enforcement discretion, not

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00026 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72795

for specification in advance by 40 CFR 63.2 (sudden, infrequent, not acknowledged that in setting standards regulation’’). In addition, the goal of a reasonable preventable and not caused under CAA section 111 ‘‘variant best controlled or best performing by poor maintenance and or careless provisions’’ such as provisions allowing source is to operate in such a way as to operation). For example, to successfully for upsets during startup, shutdown and avoid malfunctions of the source and assert the affirmative defense, the source equipment malfunction ‘‘appear accounting for malfunctions could lead must prove by a preponderance of the necessary to preserve the reasonableness to standards that are significantly less evidence that excess emissions ‘‘[w]ere of the standards as a whole and that the stringent than levels that are achieved caused by a sudden, infrequent, and record does not support the ‘never to be by a well-performing non- unavoidable failure of air pollution exceeded’ standard currently in force.’’ malfunctioning source. The EPA’s control and monitoring equipment, See also, Portland Cement Association approach to malfunctions is consistent process equipment, or a process to v. Ruckelshaus, 486 F.2d 375 (DC Cir. with CAA section 112 and is a operate in a normal or usual manner 1973). Though intervening case law reasonable interpretation of the statute. * * *.’’ The criteria also are designed to such as Sierra Club v. EPA and the CAA In the event that a source fails to ensure that steps are taken to correct the 1977 amendments undermine the comply with the applicable CAA section malfunction, to minimize emissions in relevance of these cases today, they 112(d) standards as a result of a accordance with 40 CFR 63.543(j) and to support the EPA’s view that a system malfunction event, the EPA would prevent future malfunctions. For that incorporates some level of determine an appropriate response example, the source must prove by a flexibility is reasonable. The affirmative based on, among other things, the good preponderance of the evidence that defense simply provides for a defense to faith efforts of the source to minimize ‘‘[r]epairs were made as expeditiously as civil penalties for excess emissions that emissions during malfunction periods, possible when the applicable emissions are proven to be beyond the control of including preventative and corrective limitations were being exceeded * * *’’ the source. By incorporating an actions, as well as root cause analyses and that ‘‘[a]ll possible steps were taken affirmative defense, the EPA has to ascertain and rectify excess to minimize the impact of the excess formalized its approach to upset events. emissions. The EPA would also emissions on ambient air quality, the In a Clean Water Act setting, the Ninth consider whether the source’s failure to environment and human health * * *.’’ Circuit required this type of formalized comply with the CAA section 112(d) In any judicial or administrative approach when regulating ‘‘upsets standard was, in fact, ‘‘sudden, proceeding, the Administrator may beyond the control of the permit infrequent, not reasonably preventable’’ challenge the assertion of the affirmative holder.’’ Marathon Oil Co. v. EPA, 564 and was not instead ‘‘caused in part by defense and, if the respondent has not F.2d 1253, 1272–73 (9th Cir. 1977). But poor maintenance or careless operation’’ met its burden of proving all of the see, Weyerhaeuser Co. v. Costle, 590 40 CFR 63.2 (definition of malfunction). requirements in the affirmative defense, F.2d 1011, 1057–58 (DC Cir. 1978) Finally, the EPA recognizes that even appropriate penalties may be assessed (holding that an informal approach is equipment that is properly designed and in accordance with CAA section 113 adequate). The affirmative defense maintained can sometimes fail and that (see also 40 CFR 22.27). provisions give the EPA the flexibility to such failure can sometimes cause an both ensure that its emission limitations exceedance of the relevant emissions The EPA included an affirmative are ‘‘continuous’’ as required by 42 standard (see, e.g., State Implementation defense in the proposed rule in an U.S.C. 7602(k), and account for Plans: Policy Regarding Excessive attempt to balance a tension, inherent in unplanned upsets and thus support the Emissions During Malfunctions, many types of air regulation, to ensure reasonableness of the standard as a Startup, and Shutdown (Sept. 20, 1999); adequate compliance while whole. Policy on Excess Emissions During simultaneously recognizing that despite Startup, Shutdown, Maintenance and the most diligent of efforts, emission D. What are the proposed decisions and Malfunctions (Feb. 15, 1983)). The EPA limits may be exceeded under actions related to electronic reporting? is, therefore, proposing to add to the circumstances beyond the control of the Records must be maintained in a form final rule an affirmative defense to civil source. The EPA must establish suitable and readily available for penalties for exceedances of emissions emission standards that ‘‘limit the expeditious review, according to limits that are caused by malfunctions. quantity, rate, or concentration of 63.10(b)(1). Electronic recordkeeping See 40 CFR 63.542 (defining emissions of air pollutants on a and reporting is available for many ‘‘affirmative defense’’ to mean, in the continuous basis.’’ 42 U.S.C. records, and is the form considered context of an enforcement proceeding, a 7602(k)(defining ‘‘emission limitation most suitable for expeditious review if response or defense put forward by a and emission standard’’). See generally available. Electronic recordkeeping and defendant, regarding which the Sierra Club v. EPA, 551 F.3d 1019, 1021 reporting is encouraged in this proposal defendant has the burden of proof, and (DC Cir. 2008). Thus, the EPA is and some records and reports are the merits of which are independently required to ensure that section 112 required to be kept in electronic format. and objectively evaluated in a judicial emissions limitations are continuous. Records required to be maintained or administrative proceeding). We also The affirmative defense for malfunction electronically include the output of are proposing other regulatory events meets this requirement by continuous monitors and the output of provisions to specify the elements that ensuring that even where there is a the BLDS. Additionally, standard are necessary to establish this malfunction, the emission limitation is operating procedures for the BLDS and affirmative defense; the source must still enforceable through injunctive fugitive emissions control are required prove by a preponderance of the relief. While ‘‘continuous’’ limitations to be submitted to the Administrator for evidence that it has met all of the on the one hand are required, there is approval in electronic format. elements set forth in 40 CFR 63.552 (40 also case law indicating that in many CFR 22.24). The criteria ensure that the situations it is appropriate for the EPA VII. Rationale for the Proposed Actions affirmative defense is available only to account for the practical realities of for the Mineral Wool Production where the event that causes an technology. For example, in Essex Source Category exceedance of the emissions limit meets Chemical v. Ruckelshaus, 486 F.2d 427, As discussed in Section VI.A of this the narrow definition of malfunction in 433 (DC Cir. 1973), the DC Circuit preamble, we evaluated emissions limits

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00027 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72796 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

for PM, COS, HF, HCl, formaldehyde, resetting the MACT floors in order to D. What are the proposed decisions phenol and methanol at mineral wool correct our own errors. They also agreed regarding subcategorization? production facilities. This section of the that the approach our petitioners The EPA collected information from preamble provides the results of the labeled ‘‘MACT-on-MACT’’ would be the mineral wool companies that RTR, our rationale for the proposed more accurately described as ‘‘MACT- operate bonded lines to better actions and decisions concerning on-Unsupportable-Standards- understand the different equipment changes to the 1999 MACT rule for the Erroneously-Labeled-as-MACT’’ 37. With designs and whether all collection Mineral Wool Production source regard to the evaluation of potential processes are the same, or whether category. MACT limits for HAP metals from this design and manufacturing process A. What data were used for the NESHAP source category, consistent with the differences warranted consideration of analyses? explanation presented in the proposal of subcategories for the collection process. the 1999 MACT rule (NESHAP for This process led to the identification of To perform the technology review and Mineral Wool Production, Proposed three distinct process design residual risk analysis for the Mineral Rule, June 1, 1997, 64 FR 29490) for this subcategories: Vertical, horizontal and Wool NESHAP, we created a source category describing the drum. Because collection processes only comprehensive dataset based on appropriateness of PM as a surrogate for emit HAP if they occur on a bonded existing and new test data provided by line, we are proposing to bundle the 7 mineral wool facilities. As HAP metals, we continue to consider PM as an appropriate surrogate for HAP collection operations and curing ovens described in Section IV.C of this together for each of three subcategories preamble, the voluntary industry survey metals in the proposed amendments to the NESHAP in this action. and propose new emissions limits for requested available information formaldehyde, phenol, and methanol at The agency is proposing emissions regarding process equipment, control combined collection/curing on bonded devices, point and fugitive emissions, limits for phenol and methanol because lines. The following discussion involves practices used to control fugitive the concentration of formaldehyde in a the rationale for subcategorization of emissions, and other aspects of facility specific binder formulation is collection operations into three operations. In addition to the industry independent of phenol and/or subcategories: survey, each owner/operator was asked methanol. The mineral wool industry to submit reports for any recent commented during the small business 1. The Vertical Collection Design emissions tests conducted at their advocacy review that the binder During the production of wool facility and to conduct additional ingredients and formulation can vary fiberglass on a bonded production line emissions tests in 2010 for certain HAP from one mineral wool company to the using a vertical collection design, the from specific processes. Pollutants next, and that the test data from one molten rock/slag mixture is poured from tested for the mineral wool source company is not necessarily relevant for the cupola spout onto a group of category in 2010 included most HAP or representative of another company. stainless steel drums spinning in metals, CO, PM and certain organic HAP In summary, under 112(d)(3) we are opposite directions. The spinning (formaldehyde, phenol, methanol and drums form fine fibers of the mineral carbonyl sulfide). proposing emission limits for COS, HF and HCl from cupolas; and for mixture. High air volume directs the B. What are the proposed decisions formaldehyde, methanol and phenol fibers off the fiberization spinners regarding surrogacy relationships? from bonded lines. toward a fast-moving porous vertical conveyor belt. A strong vacuum is In the 1999 MACT rule, PM serves as C. What are the proposed decisions drawn on the opposite side of the belt the surrogate for metal HAP36 at regarding certain unregulated emissions causing the fibers to lie against the existing and new cupolas, CO serves as sources? vertical belt as it moves upward. At the the surrogate for COS at new cupolas top of the conveyance, the belt travels and formaldehyde serves as the In the course of evaluating the around a curve, the vacuum is released, surrogate for phenol and methanol from Mineral Wool Production source and the fibers are moved onto a second curing ovens. The 1999 MACT standard category, we identified certain HAP for belt that conveys the layer of binder- does not have emissions limits for COS, which we failed to establish emission sprayed mineral wool fibers into the HCl or HF from existing cupolas; limits standards in the original MACT. See curing oven. Because the conveyor belt for phenol or methanol from curing; or National Lime v. EPA, 233 F. 3d 625, is vertical, the air volume drawn emissions limits for any pollutants from 634 (DC Cir. 2000) (the EPA has ‘‘clear through the belt and fiber layer must be collection operations. We are proposing statutory obligation to set emissions very high and the resulting fiber layer HAP-specific emission limits for these standards for each listed HAP’’). that is collected on the belt is thin. In pollutants under CAA section Specifically, we evaluated emissions this design, ‘shot’ (BB-sized black 112(d)(3)in this action. The agency is standards for COS, HF and HCl for granules that are high in iron as a result retaining use of PM as a surrogate for cupolas and formaldehyde, phenol and of using slag from the iron and steel HAP metals. As discussed in Sections methanol for curing ovens and industry) falls out of the fiber layer. The III.B and VII.B. of this preamble, the collection operations at mineral wool vertical design is used to produce a Court found that the EPA must set production facilities, that are not specific type of mineral wool that is low emission limits for each listed HAP specifically regulated in the existing in ‘shot’ and may be used in the (Sierra Club v. EPA, 479 F. 3d 875 (DC 3 1999 MACT standard. We are proposing hydroponic gardening market as well as Cir. March 13, 2007)), and agreed with emissions limits for these pollutants and in a specialized market of insulation the EPA that nothing in the CAA processes pursuant to 112(d)(2) and products in which shot is undesirable. suggests that it is prohibited from 112(d)(3) as discussed in Section V.A of Currently, only one facility operates this type of collection design. 36 this preamble. The HAP metals emitted from mineral wool Formaldehyde, phenol and methanol cupolas include antimony, arsenic, beryllium, cadmium, chromium, cobalt, mercury, manganese, 37 Sierra Club v. EPA, 167 F. 3d 658 (DC Cir. MACT floors for existing, new and nickel, lead and selenium. March 2, 1999). reconstructed sources in this

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00028 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72797

subcategory were based on emissions 3. The Drum Collection Design risk assessment for all HAP emitted test runs for combined curing and In the drum collection design, fibers from the Mineral Wool Production collection operations from this facility. are drawn using a very high volume air source category. We also conducted multipathway screenings for cadmium, 2. The Horizontal Collection Design flow into the center of a rotating drum. The sides of the rotating drum have mercury, and lead. Details of the risk Horizontal collection is similar to small holes that allow the air flow to assessments and additional analyses can vertical collection, but because the exit, but which trap the fibers. The angle be found in the draft residual risk documentation referenced in Section conveyor belt is horizontal it works with of the drum and the use of a vacuum V.A of this preamble, which is available gravitational forces. The layer of mineral and centrifugal force pull the fibers in the docket for this action. The agency wool collected on a horizontal belt is against the inside wall of the drum and considered the available health thinner than that collected on a vertical out the end. The entire drum is enclosed and the air flow may be vented to the information—the MIR; the numbers of belt, and the ‘shot’ is not selectively persons in various risk ranges; cancer removed. The air volume that is drawn hottest part of the cupola exhaust stack or to the existing thermal oxidizer on incidence; the maximum non-cancer HI; through the fiber layer is much lower the curing oven. the maximum acute non-cancer hazard; than in the vertical design, and therefore Currently, only one facility operates the extent of non-cancer risks; the the air stream is conducive to thermal this type of collection. Formaldehyde, potential for adverse environmental oxidation at the hottest part of the phenol, and methanol MACT floors for effects; and the distribution of risks in cupola exhaust stack or the existing existing, new, and reconstructed sources the exposed population (54 FR 38044, thermal oxidizer on the curing oven. in this subcategory were based on September 14, 1989)—in developing the Currently, only one facility operates emissions test runs for combined curing proposed CAA section 112(f)(2) this type of collection design. and collection operations from this standards for the Mineral Wool Formaldehyde, phenol and methanol facility. Production source category. MACT floors for existing, new and E. What are the results from the risk 1. Inhalation Risk Assessment Results reconstructed sources in this assessments performed and the for the Mineral Wool Production Source subcategory were based on emissions proposed decisions for the Mineral Wool Category test runs for combined curing and Production source category? Table 7 of this preamble provides an collection operations from this facility. As described in Section V.A of this overall summary of the results of the preamble, we conducted an inhalation inhalation risk assessment.

TABLE 7—MINERAL WOOL PRODUCTION INHALATION RISK ASSESSMENT RESULTS

Maximum individual cancer risk Maximum chronic non-cancer (in 1 million) 1 Estimated Estimated TOSHI 2 population at annual cancer Maximum screening acute Based on increased risk incidence Based on 3 Based on actual emissions allowable of cancer ≥ (cases per Based on allowable non-cancer HQ level emissions 1-in-1 million year) actual emis- emissions level sions level level

4 ...... 10 1,650 0.0004 0.04 0.1 8 (REL) 0.4 (AEGL–1, ERGP–1). 1 Estimated maximum individual excess lifetime cancer risk due to HAP emissions from the source category. 2 Maximum TOSHI. The highest TOSHI for the Mineral Wool Production source category is for the respiratory system. 3 The maximum HQ acute value of 8 is driven by emissions of formaldehyde. It is also based on a refined emissions multiplier of 3 which was used to estimate the peak hourly emission rates from the average rates. See section V.A. of this preamble for explanation of acute dose-re- sponse values.

The results of the chronic inhalation persons are estimated to have cancer Our screening analysis for worst-case cancer risk assessment indicate that, risks greater than 10-in-1 million, and acute impacts indicates the potential for based on estimates of current actual approximately 1,650 people are only one pollutant, formaldehyde, to emissions, the MIR could be up to 4-in- estimated to have risks greater than 1- exceed an HQ value of 1 at only one 1 million, with formaldehyde primarily in-1 million as a result of emissions facility in this source category, with a driving these risks. The total estimated from 1 facility. When considering the potential maximum HQ up to 8. A cancer incidence from this source risks associated with MACT-allowable refined emissions multiplier of 3 was category based on actual emission levels emissions, the MIR could be up to 10- used to estimate the peak hourly is 0.0004 excess cancer cases per year or in-1 million. The maximum modeled emission rates from the average rates. one case in every 2,500 years, with chronic non-cancer TOSHI value for the Refer to Appendix 7 of the draft residual emissions of formaldehyde and arsenic Mineral Wool Production source risk document in the docket for a compounds contributing 64 percent and category could be up to 0.04 with detailed description of how the refined 33 percent, respectively, to this cancer emissions of formaldehyde dominating emissions multiplier was developed for incidence.38 In addition, we note that no those impacts, indicating no significant the Mineral Wool Production source potential for chronic non-cancer category. The worst-case acute impact 38 We note that the MIR for this source category impacts. would not change if the CIIT URE for formaldehyde estimate occurs at a facility that is had been used in the assessment, although the total located in a rural area with a small cancer incidence would decrease by 52 percent. assessments will use the cancer potency for population. Since the acute modeling The MIR for the source category would remain at formaldehyde that results from that reassessment. 40 due to Cr (VI). There is an ongoing IRIS As a result, the current results may not match those reassessment for formaldehyde, and future RTR risk of future assessments.

VerDate Mar<15>2010 18:47 Nov 23, 2011 Jkt 223001 PO 00000 Frm 00029 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72798 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

scenario is worst-case because of its Based on the inhalation risk additional controls, under section confluence of peak emission rates and assessment, we estimate that the cancer 112(f)(2). Therefore, we are proposing worst-case dispersion conditions, and risks to the individual most exposed that the MACT standards for the mineral since the HQ estimates for could be up to 4-in-1 million due to wool production source category, as formaldehyde based on the AEGL–1 and actual emissions and up to 10-in-1 revised per above, provide an ample ERPG–1 values for this facility are well million due to MACT-allowable margin of safety to protect public health below 1, we are proposing to find that emissions, mainly due to formaldehyde and prevent adverse environmental acute noncancer health impacts of stack emissions. We estimate that the effects. concern are unlikely. incidence of cancer based on actual Nevertheless, we are soliciting With respect to the potential for emissions is 0.0004 excess cancer cases comments and information regarding adverse environmental effects from non per year or one case every 2,500 years, additional control measures and work PB–HAP, we note that that there is a and that about 1,650 people face a practices that may be available and their lack of information about specific cancer risk greater than 1-in-1 million feasibility in further reducing stack adverse environmental effects occurring due to HAP emissions from this source emissions of COS, HF, HCl, at given concentrations of the HAP category. Our assessments also formaldehyde, phenol, and methanol, or emitted by this source category. indicated a low potential for HAP additional monitoring that may be However, given that all chronic non- emissions from these sources to pose warranted to ensure adequate control of cancer HQ values considering actual any significant adverse environmental these emissions. emissions are less than 1 using human effects or human health multi-pathway G. What are the results from the health reference values, we believe that risks or chronic noncancer human technology review and proposed it is unlikely that adverse environmental health risks due to inhalation. While our decisions? effects would occur at the actual HAP acute risk screening ruled out the Based on our technology review, we concentrations estimated in our human possibility of acute impacts of concern believe that the reductions in HAP health risk assessment. for all pollutants except for formaldehyde at one facility, we emissions since promulgation of the 2. Multipathway Risk Assessments and ultimately concluded that the potential 1999 Mineral Wool Production MACT Results for acute impacts of concern at this rule are directly related to There were no exceedances of facility is low. The risk assessment for improvements in two areas: (1) screening emissions rates for the PB this source category was largely based Improvements in fabric filter control HAP emitted by the facilities in the on facility-specific stack test data and technology (e.g., improved bag Mineral Wool Production source emissions estimates, indicating a high materials, replacement of older baghouses) and (2) addition of category, thus we have no concerns degree of confidence in the results. regenerative thermal oxidizers (RTOs) about potential multi-pathway risks Considering all of the above and oxygen injection to control from this source category. information, we are proposing that the current risks due to actual HAP emissions from cupolas. Additional 3. Facility Wide Risk Assessment emissions from this source category are reductions have been achieved due to Results acceptable. the use of low-sulfur raw materials at For all facilities in this source While the estimated chronic risks one facility. The RTOs and lower sulfur category, there are no other significant associated with MACT-allowable raw materials are discussed above (in HAP emissions sources present beyond emissions from this source category are Section VII.C of this preamble) since those included in the source category. slightly higher than risk estimates based these controls and measures are relevant All significant HAP sources have been on actual emission levels, they are still to development of the MACT standards included in the source category risk well below 100 in one million and there for COS and other organic HAPs under analysis. Therefore, we conclude that are no other significant risks. Therefore, Section 112(d)(2) of the CAA, and in the the facility wide risks are essentially the we propose the risks due to allowable beyond the floor analyses (described in same as the source category risks. emissions are also acceptable. Section VII.C of this preamble) that we also do as part of the MACT standard F. What are our proposed decisions for 2. Ample Margin of Safety evaluations under Section 112(d)(2) and the Mineral Wool Production source As explained earlier in Section V of 112(d)(3). category based on risk acceptability and this preamble, the agency again In this section, as part of our ample margin of safety? considers all of the health risks and technology review, we describe other health information considered in developments in development in fabric 1. Risk Acceptability the first step. Beyond that information, filter technologies and the relationship As noted in Section V.A of this we evaluate the cost and feasibility of to PM emissions. preamble, we weigh all health risk available control technologies and other Slight improvements in fabric filter factors in our risk acceptability measures (including the controls, control technology are reflected in the determination, including the MIR; the measures and costs reviewed under the emissions test data collected under the numbers of persons in various risk technology review) that could be industry survey. The emissions limit for ranges; cancer incidence; the maximum applied in this source category to PM under the 1999 MACT rule is a noncancer HI; the maximum acute further reduce the risks due to production-based limit of 0.1 pounds of noncancer hazard; the extent of emissions of HAP identified in our risk PM per ton of melt for new and existing noncancer risks; the potential for assessment. cupolas. Based on our analysis of survey adverse environmental effects; and Based on our research and analyses as responses and test data collected under distribution of risks in the exposed discussed in Section V.C of this the industry survey, this industry population; and risk estimation preamble, we have not identified any primarily uses fabric filters to control uncertainty (54 FR 38044, September feasible control options beyond what we emissions of metal HAP, and sources 14, 1989) in developing the proposed are requiring in our proposed standards affected by the current PM limit are CAA section 112(f)(2) standards for this for emissions sources described above, achieving PM concentrations at control source category. and are therefore not proposing device outlets that are only slightly

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00030 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72799

below the current limit (see Technology tests conducted and to conduct the surrogate for metal HAPs 39 at Review for the Mineral Wool Production additional emissions tests in 2010 for existing and new glass-melting furnaces Manufacturing Source Category). Given certain HAP from specific processes. and formaldehyde serves as the fluctuations in control device Pollutants tested for the wool fiberglass surrogate for phenol and methanol from performance and mineral wool source category in 2010 included most forming and curing at RS manufacturing production fluctuations, we do not HAP metals, PM, and certain organic lines and forming and curing at FA believe that developments in practices, HAP (HF, HCl, formaldehyde, phenol, manufacturing lines. As described in processes, and control technologies and methanol). Sections III.B and VIII.B in this warrant revisions to the PM limit in the As discussed in Section IV.C above, in preamble, the court found that the EPA 1999 MACT rule to reflect HAP metal the emissions testing for the survey, erred when we did not set emission emissions levels achieved in practice. industry requested to conduct emission limits for each HAP emitted by industry Moreover, the RBLC did not identify 40 testing on furnaces they believed were processes in the MACT standards. any practices, processes, or control representative of the other furnaces in Therefore, the agency is proposing HAP- technologies applicable to the emission operation. The EPA and industry agreed specific emissions limits for phenol and sources in this source category that were that the bases for representativeness methanol. not identified and evaluated during the would include a variety of factors such original MACT development. C. What are the proposed decisions In summary, we have not identified as processing the same materials, regarding certain unregulated emissions any additional relevant cost-effective producing the same products and being sources? developments in technologies, practices the same type of furnace. Furnace As discussed earlier in Section VI.B of or processes since promulgation of the construction and refractory composition this preamble, we identified certain MACT rule to further reduce HAP were not factors that were presented by HAP for which we failed to establish emissions. Therefore, we are not industry as having an effect on HAP emission standards in the original 1999 proposing any changes to the MACT emissions, and those factors were not MACT. In the 1999 MACT rule, we used standards in this action as a result of our used as a basis of representativeness for formaldehyde as a surrogate for phenol technology review under Section the resulting data set. During analysis of and methanol, and we did not establish 112(d)(6) for Mineral Wool Production. the test data, the EPA discovered high HAP-specific emission limits for Additional details regarding these emissions of chromium compounds, phenol, methanol, HF and HCl. For this analyses can be found in the following including hexavalent chromium, and action we evaluated emissions technical document for this action that these emissions were mostly from standards for HF, HCl, phenol, and which is available in the docket: certain furnaces constructed of high methanol at wool fiberglass Technology Review for the Mineral Wool chrome refractories. manufacturing facilities, described Production Manufacturing Source The Wool Fiberglass Manufacturing below, that are not specifically regulated Category. source category consists of 29 facilities in the existing 1999 MACT standard. with 80 furnaces, 54 RS manufacturing The EPA is therefore proposing to set VIII. Rationale for the Proposed lines and less than 30 FA manufacturing emissions limits for these HAP Actions for the Wool Fiberglass lines. Since there are more than 30 emissions, under CAA section 112(d)(3) Manufacturing Source Category furnaces and RS lines, we based the in this action. As discussed in Section VI.B of this MACT floor limits on the average D. What are the results from the risk preamble, we evaluated emissions limits emissions limitation achieved by the assessments and analyses and the for PM, chromium compounds, HF, HCl, best performing 12 percent of sources. proposed decisions for the Wool formaldehyde, phenol, and methanol at Therefore, the MACT floor for HF and Fiberglass Manufacturing source wool fiberglass manufacturing facilities. HCl from glass-melting furnaces was category? This section of the preamble provides based on the 10 best performing the results of the RTR, our rationale for furnaces; the 7 best performing RS lines; An inhalation risk assessment was the proposed actions for the Wool and the 5 best performing FA lines. completed for all HAP emitted for the Wool Fiberglass Manufacturing source Fiberglass Manufacturing source The stack test data were used to category, and our proposed decisions category. Details of the risk assessments calculate the MACT floors using the 99 and additional analyses can be found in concerning changes to the 1999 MACT percent UPL for glass-melting furnaces, rule. the residual risk documentation RS manufacturing lines, and FA referenced in Section V.A of this A. What data were used for the NESHAP manufacturing lines from wool preamble. The agency considered the analyses? fiberglass manufacturing plants. The available health information—the MIR; To perform the technology review and UPL analysis is explained in more detail the numbers of persons in various risk residual risk analysis for the Wool in MACT Floor Analysis for the Wool ranges; cancer incidence; the maximum Fiberglass Manufacturing NESHAP, we Fiberglass Manufacturing Source non-cancer HI; the maximum acute non- created a comprehensive dataset based Category, which is available in the cancer hazard; the extent of non-cancer on existing and new test data provided docket for this proposed action. The risks; the potential for adverse by 26 of the 29 wool fiberglass facilities. results from the MACT floor analysis are environmental effects; and distribution As described in Section IV.C of this presented in Section VI.B of this of risks in the exposed population (54 preamble, the voluntary industry survey preamble. FR 38044, September 14, 1989)—in requested available information B. What are the proposed decisions developing the proposed CAA section regarding process equipment, control regarding surrogacy relationships? 112(f)(2) standards for the Wool devices, point and fugitive emissions, practices used to control fugitive A surrogate approach is used to allow 39 The HAP metals emitted from wool fiberglass emissions, and other aspects of facility for easier and less expensive glass-melting furnaces include antimony, arsenic, beryllium, cadmium, chromium, cobalt, mercury, operations. In addition to the ICR measurement and monitoring requirements. In the 1999 MACT rule manganese, nickel, lead, and selenium. survey, each facility was asked to 40 Sierra Club v. EPA, 479 F. 3d 875 (DC Cir. submit reports for any recent emissions for this source category, PM serves as March 13, 2007).

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00031 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72800 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

Fiberglass Manufacturing source 1. Inhalation Risk Assessment Results category. for the Wool Fiberglass Manufacturing Source Category Table 8 of this preamble provides an overall summary of the results of the inhalation risk assessment.

TABLE 8—WOOL FIBERGLASS MANUFACTURING INHALATION RISK ASSESSMENT RESULTS

Maximum individual cancer risk (in 1 mil- Maximum chronic non-cancer lion) 1 Estimated Estimated TOSHI 2 population at annual cancer Maximum screening acute Based on increased risk incidence Based on ≥ Based on non-cancer HQ 3 Based on actual emissions allowable of cancer (cases per actual emis- allowable level emissions 1-in-1 million year) emissions level sions level level

40 ...... 60 849,000 0.05 0.2 0.5 30 (REL) 2 (AEGL–1, ERPG–1). 1 Estimated maximum individual excess lifetime cancer risk due to HAP emissions from the source category. Hexavalent chromium is the pri- mary driver for cancer risk. 2 Maximum TOSHI. The highest TOSHI for the Wool Fiberglass Manufacturing source category is for the respiratory system. 3 The maximum HQ acute value of 30 is driven by emissions of formaldehyde. See section V.A. of this preamble for explanation of acute dose- response values.

The results of the chronic inhalation of 7 facilities due to formaldehyde up to 40-in-1-million based on actual cancer risk assessment indicate that, emissions,41 with one facility in this emissions. The major contributor to this based on estimates of current actual source category indicating the potential cancer risk is hexavalent chromium. emissions, the maximum individual to create a maximum worst-case HQ The greatest amount of risk is from one lifetime cancer risk (MIR) could be up value up to 30. This maximum worst- facility that uses a type of refractory to 40-in-1 million. The major case acute impact corresponds to a brick that is described by the company contributor to this cancer risk is maximum HQ of 2 based on the AEGL– as ‘‘high chrome.’’12 13 ((Notes of April hexavalent chromium that is emitted 1 and ERPG–1 levels for formaldehyde. 14, 2011, Certainteed); (Region 7 from the furnace refractory brick. The Altogether, these results indicate that Certainteed Notes). greatest amount of hexavalent we cannot rule out the potential for Because the use of high chrome chromium emitted from a single source formaldehyde emissions from this refractories extends the life of the is from a facility that currently uses a source category to cause acute impacts furnace from a maximum of 10 years to type of refractory brick that is made of mild concern, such as eye and nose at least 15 years, and the cost of furnace almost entirely of chromium irritation. Repeated exposures to these construction is increased by about 15 compounds. In addition, we note that levels (i.e., at or above the AEGL–1 and percent when it is reconstructed using approximately 12,000 people are ERPG–1) could cause further health high chrome refractories 12 estimated to have cancer risks greater concerns. (Notes of than 10-in-1 million as a result of With respect to the potential for April 14, 2011, Certainteed) we believe formaldehyde and hexavalent adverse environmental effects from non that there is a financial incentive for chromium emissions at 2 facilities, and PB–HAP, we note that that there is a other facilities to switch to this high approximately 849,000 people are lack of information about specific chromium refractory at the time they estimated to have risks greater than 1- adverse environmental effects occurring rebuild their furnaces. For this reason, in-1 million as a result of formaldehyde at given concentrations for the HAP we performed an auxiliary risk and hexavalent chromium emissions emitted by this source category. characterization analysis to assess the from 15 facilities. The maximum However, given that all chronic non- potential maximum individual lifetime estimated chronic non-cancer TOSHI cancer HQ values considering actual cancer risks in the event that the other value for the Wool Fiberglass emissions are less than 1 using human 28 Wool Fiberglass facilities switch to Manufacturing source category is 0.2 health reference values, we believe that the high chromium brick. For the with emissions of formaldehyde it is unlikely that adverse environmental auxiliary risk characterization analysis dominating those impacts, indicating no effects would occur at the actual HAP it was assumed that the hexavalent significant potential for chronic non- concentrations estimated in our human chromium emissions for each facility cancer impacts. health risk assessment. would be the same as that for the facility Based on the acute REL to assess with annual emissions of 420 lbs of possible acute non-cancer effects due to 2. Auxiliary Risk Characterization hexavalent chromium per furnace. Table emissions of formaldehyde, our analysis As indicated in Section VIII.D.1 9 of this preamble provides a summary indicates that the maximum acute HQ above, the MIR for the Wool Fiberglass of the results of this auxiliary inhalation value could exceed a value of 1 at a total Manufacturing source category could be risk assessment.

41 Individual facility acute HQ values for all assessment document that is included in the docket facilities can be found in Appendix 6 of the risk for this proposed rulemaking.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00032 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72801

TABLE 9—WOOL FIBERGLASS MANUFACTURING AUXILIARY INHALATION RISK ASSESSMENT RESULTS

1 Potential maximum individual cancer risk (in 1 million) Estimated Estimated Estimated population at population at population at Estimated an- increased risk increased risk increased risk nual cancer in- ≥ ≥ of cancer ≥ cidence (cases Based on actual emissions level of cancer of cancer 100-in-1 per year) 1-in-1 million 10-in-1 million million

900 ...... 7,300,000 460,000 8,100 0.46 1 Estimated maximum individual excess lifetime cancer risk due to HAP emissions from the source category.

The results of the auxiliary analysis determination, including the MIR; the or environmental impacts, we are indicate that, under this scenario, the numbers of persons in various risk proposing that the current risk levels estimated emissions from 14 facilities ranges; cancer incidence; the maximum due to actual and MACT-allowable could lead to maximum individual noncancer HI; the maximum acute emissions are acceptable. lifetime cancer risks greater than 100-in- noncancer hazard; the extent of 1-million, with the highest emitting noncancer risks; the potential for 2. Ample Margin of Safety Analysis and facility posing a potential maximum adverse environmental effects; and Proposed Decisions individual risk of 900-in-1-million. distribution of risks in the exposed As described above, we are proposing Under this scenario, 8,100 people would population; and risk estimation that the risks associated with the actual be exposed to risks greater than 100-in- uncertainty (54 FR 38044, September and MACT-allowable stack emissions 1-million, 460,000 people would be 14, 1989) in developing the proposed exposed to risks of greater than 10-in-1- CAA section 112(f)(2) standards for this from this source category are acceptable million, and over 7 million people source category. based on the current composition of refractory bricks used by this source would be exposed to cancer risks of Based on the inhalation risk greater than 1-in-1-million. assessment, we estimate that the cancer category. However, as discussed in In summary, the auxiliary risk risks to the individual most exposed Section VIII.D(2) of this preamble, if analysis indicates that if other facilities could be up to as 40-in-1 million due to other wool fiberglass facilities switch to high chromium refractory, actual emissions and up to as 60-in-1 reconstructed their furnaces with high emissions of hexavalent chromium million due to MACT-allowable chromium refractory bricks, the could potentially pose unacceptable emissions, mainly due to formaldehyde maximum individual cancer risks risks to public health due to inhalation and chromium stack emissions. We would be higher and likely result in a exposures resulting from stack estimate that the incidence of cancer finding of unacceptable risks. emissions of hexavalent chromium. based on actual emissions is 0.05 excess According to our 2-step process for 3. Multipathway Risk Assessments and cancer cases per year or one case every assessing risks, after we evaluate Results 20 years, and that about 850,000 people whether risks are ‘‘acceptable’’ we face a cancer risk greater than 1-in-1 evaluate whether cost effective None of the facilities in the Wool million due to the HAP emissions from measures are available to reduce risks Fiberglass Manufacturing source this source category. category reported emissions of PB HAP further, to provide an ‘‘ample margin of Our assessments also indicate a low safety.’’ As stated in Section VIII.F of that were greater than the screening potential for HAP emissions from these emission rates. Therefore, multi- this preamble, both NaOH scrubbers and sources to pose any significant adverse a furnace rebuild are considered cost pathway exposures and environmental environmental effects, human health risks were deemed negligible. effective when hexavalent chromium multi-pathway effects, or chronic levels are high. NaOH scrubbers achieve noncancer human health risks. Our 4. Facility Wide Risk Assessment at least 95 percent reduction in acute risk screening ruled out the Results hexavalent chromium emissions at other possibility of acute impacts of concern For this source category, there are no for all pollutants but one, formaldehyde, industries. Transferring this technology other significant HAP emissions sources at seven facilities, with a maximum to the wool fiberglass industry is present beyond those included in the worst-case HQ estimated to be 30 based reasonable and would reduce source category. All significant HAP on the REL and 2 based on the AEGL– hexavalent chromium to levels that sources have been included in the 1 (or ERPG–1, which is equivalent). would achieve an ample margin of source category risk analysis. Therefore, While this means we cannot rule out the safety. Therefore, we are proposing we conclude that the facility wide risk potential for acute concerns due to emission limits of 0.06 lb of total is essentially the same as the source formaldehyde emissions from these chromium compounds per thousand category risk and that no separate facilities, we note that the use of tons (or 60 lb of total chromium facility wide analysis is necessary. formaldehyde is being phased out in compounds per million tons) of glass E. What are our proposed decisions for this industry, and will be eliminated pulled in this action (as presented in the Wool Fiberglass Manufacturing from all but 2 facilities in the source Table 10) under Section 112(f)(2) of the source category based on risk category. Since the cancer risks due to CAA in this action. We believe this limit acceptability and ample margin of actual and allowable emissions (based would achieve an ample margin of safety? on the current composition of refractory safety to protect public health and bricks used by this source category) are prevent adverse environmental effects. 1. Risk Acceptability well within the acceptable range (i.e., As noted in Section VIII.D of this less than 100-in-1 million) and since we preamble, we weigh all health risk have no additional significant concerns factors in our risk acceptability regarding other potential human health

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00033 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72802 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

TABLE 10—PROPOSED EMISSIONS LIM- on the REL and 2 based on the AEGL– warranted to ensure adequate control of ITS FOR GLASS-MELTING FURNACES 1 or ERPG–1, which is equivalent stack emissions. We specifically request BASED ON RISK REVIEW (formaldehyde). While this means we information on other criteria on which cannot rule out the potential for acute a chromium compounds emission limit Pounds of pollut- concerns due to formaldehyde should be based that would reduce risks Pollutant ant per thousand emissions from these facilities, we note from hexavalent chromium. tons of melt: that the worst-case acute HQs are based on conservative assumptions (e.g., 3. Analysis of the Resulting Risk After Chromium compounds ... 0.06 worst-case meteorology coinciding with the Proposed Requirements Are in Place peak short-term one-hour emissions We conducted an assessment to These emission limits apply to from each emission point, with a person furnaces at major sources in the wool estimate the risks based on a post- located at the point of maximum control scenario reflecting all the fiberglass manufacturing source concentration during that hour). category. However, there are no proposed requirements for the Moreover, the use of formaldehyde is emissions described above (including differences in furnaces at major sources being phased out in this industry, and and area sources. We are concerned the proposed emissions limit for will be eliminated from all but 2 chromium compounds). Details are about the levels of hexavalent facilities in the source category. Since provided in the Draft Residual Risk chromium that can be emitted by area the cancer risks due to actual emissions Assessment for the Mineral Wool sources where furnaces may be are well within the acceptable range Production and Wool Fiberglass constructed using high chrome (i.e., less than 100 in 1 million) and Manufacturing Source Categories, EPA’s refractories. Therefore we plan to collect since we have no additional significant Office of Air Quality Planning and additional information from industry to concerns regarding other potential Standards Office of Air and Radiation, inform regulation of area sources in a human health or environmental September 2011, which is available in future action. impacts, and since we have not the docket to this rule. The emission limits we are proposing identified any additional cost-effective for chromium compounds under controls to further reduce formaldehyde Table 11 of this preamble provides an 112(f)(2) are identical to the chromium emissions, we are proposing that the overall summary of the results of the compounds limits we are proposing MACT rule along with all the proposed post-control inhalation risk assessment. under 112(d)(6), as described in Section amendments described above (including As compared to Table 8, the MIR VIII.F of this preamble. the emissions limits for chromium and decreased from 40 in 1 million to 20 in Our assessments also indicate a low formaldehyde) will provide an ample 1 million, primarily as a result of one potential for HAP emissions from these margin of safety to protect public health facility replacing the high chrome sources to pose any significant adverse and prevent adverse environmental refractory bricks at the facilities that environmental effects, human health effects. currently exceed the proposed multi-pathway effects, or chronic We are soliciting comments and chromium standard. These estimates are noncancer human health risks. Our information regarding additional control based on the dataset compiled using the acute risk screening ruled out the measures, work practices that may be industry’s emissions test data from their possibility of acute impacts of concern available, and their feasibility in further 2010 industry survey responses, which for all pollutants but one, formaldehyde, reducing emissions of formaldehyde, show three furnaces would have to at seven facilities, with a maximum chromium compounds, HCl, and HF, or reduce chromium emissions to meet the worst-case HQ estimated to be 30 based additional monitoring that may be limit in the proposed rule.

TABLE 11—POST CONTROL INHALATION RISK ESTIMATES FOR WOOL FIBERGLASS [Result of chromium control]

Estimated Estimated Maximum Maximum population at annual cancer chronic non- screening Maximum individual cancer risk (in 1 million) based on actual emissions cancer TOSHI 1 increased risk incidence acute level ≥ based on non- of cancer (cases per actual emis- 1 in 1 million year) cancer HQ 3 sions level 2

20 ...... 282,000 0.02 0.2 30

In addition, we estimated that the Essentially non-formaldehyde binders requirements. We further estimate that formaldehyde emissions would be at or are or will be used industry-wide.’’ A there may be two facilities below the MACT standard for all copy of this letter has been placed in the manufacturing pipe insulation or heavy facilities once this rule is fully docket for this action (see NAIMA’s density insulation products that will be implemented and we are not proposing Response for the Fiberglass Industry to major sources of HAP emissions on the that additional control options be EPA’s Formaldehyde and Collection compliance date of these proposed implemented. Questions). Based on this information amendments to subpart NNN. If NAIMA In a letter dated June 8, 2011, the and the information provided by the is correct in that formaldehyde will be industry trade association (NAIMA) industry in their 2010 survey, we phased out by the compliance date of stated that ‘‘NAIMA can provide estimate that 27 of the 29 wool these proposed amendments, we documentation that all major sources fiberglass manufacturing facilities will anticipate that the estimated inhalation have already converted or have have HAP emissions below the 10 and risks due to formaldehyde would further announced plans to convert to non- 25 tpy thresholds and will not be decrease. phenol formaldehyde binders. subject to the major source MACT

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00034 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72803

In summary, we are proposing that PM per ton of glass pulled applicable to followed by EPA Method 0061). In the MACT standard, with the changes all glass melting furnaces. Based on our contrast, the chromium emissions for a we are proposing in this action, will analysis of survey responses and test few furnaces were several orders of provide an ample margin of safety and data collected under the industry magnitude higher than the rest of the prevent adverse environmental effects. survey, this industry primarily uses industry. The facility emitting the fabric filters to control emissions of highest level of hexavalent chromium, F. What are the results from the metal HAP, and the vast majority of at 840 lb/yr, advised us that the reason technology review and proposed sources affected by the current PM limit chromium tested very high was due to decisions? are achieving PM emissions at control the refractory products, high chrome Based on our technology review, we device outlets that are far below the refractories, from which the furnaces are determined that there have been current limit. Id. constructed (Notes of April 14, 2011, advances in emissions control measures Most, if not all, sources reported PM Certainteed) 12. Based on the emissions since the Wool Fiberglass emissions (coming out of the stacks after testing and information on high chrome Manufacturing NESHAP was originally the control devices) that are less than 10 refractories, we believe changes to the promulgated in 1999. Since percent of the current limit, with several 1999 MACT rule are warranted under promulgation, we estimate that sources achieving PM emissions that are CAA section 112(d)(6). industry-wide metal HAP emissions two to three orders of magnitude lower The data indicate that well from process sources have been reduced than the current limit. Based on these performing wool fiberglass furnaces by approximately 76 percent. Due to data, we believe that developments in emit small amounts of chromium industry’s efforts to replace phenol- practices, processes, and control compounds, that is, they emit less than formaldehyde binders, more than 95 technologies warrant revisions to the 0.06 pounds of chromium compounds percent of formaldehyde, phenol, and 1999 MACT rule, under section (Cr) per thousand tons of glass pulled. methanol emissions have been reduced 112(d)(6). Our analysis of emissions However, three facilities currently (or will be by 2012). As a result actual data provided in the survey conducted operate furnaces that emit chromium in PM (metal HAP), formaldehyde, phenol, by industry indicates that stacks excess of this rate. Chromium emissions and methanol emissions from process equipped with a well-performing fabric from these high emitters range from 9 to sources at all wool fiberglass filter or ESP can achieve exhaust PM 840 lb/yr. Furnaces operating below this manufacturing facilities are significantly concentrations of less than 0.014 lb/ton rate generally emit less than 1 pound lower than are allowed under the 1999 of glass pulled. We estimate that all of per year; many of these tested below the MACT rule. the wool fiberglass facilities would be detection level of the test method. The We believe that the reductions in able to comply with this revised limit data indicate that there is a ‘break’ metal HAP emissions since without additional controls. We between the furnaces emitting less than promulgation of the 1999 MACT rule estimate that this would result in small the proposed limit and those emitting are mainly directly related to reductions of metal HAP emissions greater amounts of chromium. Data improvements in two areas: (1) since there will only be a couple of further indicate there are no wool Improvements in fabric filter control facilities subject to the PM limits and fiberglass manufacturers with low glass technology (e.g., improved bag the available data on some of the production rates but high levels of materials, replacement of older furnaces at those facilities indicates they chrome emissions. We are therefore baghouses) and (2) the use of are currently meeting the proposed PM proposing to set a chromium electrostatic precipitators (ESPs). Our emission limit. We do not anticipate compounds emission limit of 0.06 lb of review also indicates that high chrome additional energy use associated with chromium per thousand tons of glass refractories are a new technology used this revised limit. Furthermore, we do pulled as shown in Table 12. in wool fiberglass furnaces that the not anticipate any adverse non-air Under section 112(d)(6), we are available data indicate result in an environmental impacts associated with proposing this emission limit for increase in emissions of chromium the implementation of this revised limit. chromium compounds taking into compounds. The results of our analyses Therefore, we are proposing that account the developments in practices, and our proposed decisions for these reducing the PM limit in the NESHAP processes and technology by the wool areas under CAA section 112(d)(6) are from 0.50 lb of PM per ton of glass fiberglass industry since promulgation presented in the following sections. pulled to 0.014 lb of PM per ton of glass of the 1999 MACT standard. The Based on these data, we believe that pulled (see Table 12) is both feasible emission limits we are proposing for developments in practices, processes, and cost effective. Therefore, we are chromium compounds under 112(d)(6) and control technologies warrant proposing a revised PM limit in the are identical to the chromium revisions to the 1999 NESHAP. NESHAP of 0.014 lb of PM per ton of compounds limits we are proposing Additional details regarding these glass pulled in this action. We have under 112(f)(2), as described in Section analyses can be found in Technology based these statements on information VIII.E of this preamble. Review for the Wool Fiberglass we received from the industry in their We estimate that the 2 remaining Manufacturing Source Category. survey responses; nevertheless, we are major source wool fiberglass facilities The improvements in fabric filter seeking comment on our estimation that would be able to comply with this control technology are reflected in the all wool fiberglass manufacturers can chromium compounds emission limit. emissions test data collected under the meet the PM emission limits without We estimate that if the high chromium industry survey. Two types of PM additional controls. emitting facilities remain major sources, control are used in the wool fiberglass We conducted a review of the these new emission limits would result manufacturing industry: fabric filters available test data for chromium in annual reductions of 1,155 pounds of (baghouses) and electrostatic compounds including hexavalent chromium compounds, specifically precipitators. Electrostatic precipitators chromium emissions from glass hexavalent chromium and there will be (ESP) may be configured as either wet furnaces. We found that for most no reductions at the remaining facilities ESPs or dry ESPs. The emissions limit furnaces, measured emissions were near because data indicate they are currently for PM under the 1999 MACT rule is a or below detection limits of the methods meeting the proposed chromium production-based limit of 0.5 pounds of used for testing (EPA Method 29 emission limit.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00035 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72804 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

Wet scrubbers are not generally in use the cost of the controls to the sales or actions required to meet the proposed in this industry. However, we evaluated revenues of the companies that would emissions limits. The Economic Impact their use to achieve reductions in incur costs to comply with the Analysis considered annual sales and hexavalent chromium for furnaces chromium emission limits. The revenue data from the facilities within emitting chrome above the levels being economic impact on these firms, this source category and their ability to proposed. Sodium hydroxide (NaOH) measured in annual compliance costs as meet the proposed amendments. The scrubbers are in use for furnace a percent of sales or revenues, is less following sections discuss the cost, operations at other industries for than 0.001 percent for each of the environmental, and economic impacts chromium compounds reduction. We affected firms.43 to the Mineral Wool Production source have evaluated the use of NaOH We therefore, we propose that category. (Economic Impact Analysis for scrubbers for the wool fiberglass requiring the 0.06 lb chromium per the Mineral Wool and Wool Fiberglass manufacturing industry and find that thousand tons of melt limit in the RTRs. U.S. EPA. October 2011.) the control technology can be adapted NESHAP is both feasible and cost C. What are the air quality impacts for for use in the wool fiberglass industry effective. We solicit comment on this the Mineral Wool Production source from the chromium electroplating comparison and the use of this value as category? industry and from certain high a reasonable cost to reduce chromium. temperature metallurgical industries.42 The EPA estimated the emissions We do anticipate an additional energy TABLE 12—PROPOSED EMISSIONS LIM- reductions that are expected to result use associated with this revised limit if ITS FOR GLASS-MELTING FURNACES from the proposed amendments to the sources choose to install NaOH BASED ON TECHNOLOGY REVIEW 1999 MACT rule compared to the 2010 scrubbers to remove hexavalent baseline emissions estimates. A detailed chromium from the furnace gases. We Pounds documentation of the analysis can be anticipate the affected sources may Pollutant pollutant per found in: Cost Impacts of the Revised incur disposal costs of hexavalent ton of melt NESHAP for the Mineral Wool Production Manufacturing Source chromium contaminated materials PM ...... 0.14 associated with the implementation of Chromium compounds ...... 0.00006 Category. this emission limit. We anticipate that Emissions of formaldehyde from two sources which currently emit This proposed limit for chromium mineral wool production facilities have chromium at levels slightly higher than compounds (of 0.06 lb per thousand declined over the last 12 years as a the proposed limit will be able to meet tons chromium limit) under CAA result of federal rules, state rules and on it by installing NaOH scrubbers (which Section 112(d)(6) is the same limit being the industry’s own initiative. The selectively remove the hexavalent form proposed under Section 112(f)(2) that current proposal would not reduce of chromium from the exhaust air). This was described earlier in this notice. We formaldehyde, phenol, or methanol cost is about $300 per pound hexavalent believe that these proposed revisions for emissions from their current levels. chromium removed if these companies chromium and PM are cost effective Under the proposed emissions limits for install a NaOH scrubber in series with revisions and reflect the current cupolas, COS, HF, and HCl emissions the existing furnace control. A wool developments in processes and would be reduced by a combined 23 fiberglass facility could also choose to technology by this industry. (i.e., well percent compared to 2010 levels rebuild the glass furnace using performing air pollution control). reported in the industry survey refractories with low chromium responses. We estimated that the COS contents. The cost of that option would IX. Summary of Cost, Environmental, emissions reductions would be 41 tpy be prorated to consider the remaining and Economic Impacts for the Mineral from cupolas. useful life of the existing high Wool Source Category Based on the emissions data available chromium furnace and would cost about Here we discuss the anticipated air, to the EPA, we believe that all facilities $12,000 per pound chromium water, solid waste and energy impacts will be able to comply with the compounds removed. We expect that for in addition to the cost and economic proposed emissions limits for COS, HF, the highest chromium emitting wool impacts to the industry as a result of the HCl, formaldehyde, phenol, and fiberglass furnace emitting 500 lb proposed amendments to the 1999 methanol without additional controls chromium per year, this option would MACT rule. because they can reduce emissions using raw material substitution or be used to meet the proposed limit. We A. What are the affected sources in the base this estimate on two factors: (1) oxygen injection as discussed Mineral Wool Production source previously in Section VII.F of this The furnace is at the end of its useful category? life and is expected to be reconstructed preamble. We anticipate that the 7 mineral wool in 2013 (Notes of April 14, 2011; Region D. What are the water quality and solid 12 13 production facilities currently operating 7 Certainteed Notes) and (2) the waste impacts? NaOH scrubber achieves about 95 in the United States will be affected by We do not anticipate any adverse percent reduction (NaOH Scrubber these proposed amendments. water quality or solid waste impacts Information),42 which is not quite B. How are the impacts for this proposal from the proposed amendments to the enough to meet the proposed chromium evaluated? 1999 MACT rule because the emission limit. The cost of the control requirements proposed would not equipment to wool fiberglass plants is For the proposed amendments to the change the existing requirements that about $225,000 for installation and Mineral Wool Production source impact water quality or solid waste. annual operation and maintenance costs category, the air quality, water quality, solid waste, and energy impacts were of about $5000 per year. We compared E. What are the secondary impacts? determined based on the need for Indirect or secondary air quality 42 NaOH Scrubber Information. Telephone additional control technologies and discussion and emails between vendors, companies, impacts include impacts that will result and EPA. Steffan Johnson, Measurement Policy 43 Economic Impact and Initial Regulatory from the increased electricity usage Group, USEPA/OAQPS/SPPD. Flexibility Analysis. September 2011. associated with the operation of control

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00036 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72805

devices, as well as water quality and memorandum, Cost Impacts of the concentrations of COS in the cupola solid waste impacts (which were just Revised NESHAP for the Mineral Wool exhaust gas to energy that is returned to discussed) that might occur as a result Production Manufacturing Source the cupola. This technology reduces the of these proposed actions. We anticipate Category, includes a complete consumption of coke up to 30 percent that the mineral wool production description of the cost estimate methods and, because of the cost of coke, this facilities will be able to comply with the used for this analysis and is available in technology pays for itself over a period proposed amendments without having the docket. of several years. Emissions of COS are to install additional control technologies We identified several ways in which below 0.04 lb COS per ton melt when such as RTOs. In addition, those mineral wool producers reduce the COS an RTO is installed for energy facilities that switch to low-sulfur raw emissions from cupolas, enabling them reclamation and new source MACT is materials will most likely reduce air to comply with the proposed emission based upon the use of this technology. emissions of SO2. limit of 3.3 lb COS per ton of melt. One facility is expected to incur an These methods include raw material incremental annualized cost of $360,000 F. What are the energy impacts? substitution, oxygen injection, and for low-sulfur raw materials (rock) if Energy impacts in this section are installation of an RTO. We found two they use that option to comply with the those energy requirements associated approaches to raw material substitution: COS requirement for cupolas. That cost with the operation of emission control slag and rock. One mineral wool would be lessened to no more than devices. Potential impacts on the manufacturer purchases low-sulfur slag, $20,000 for installation of oxygen national energy economy from the rule a waste product from a local steel plant. injection, which is another alternative. are discussed in the economic impacts Another plant owns and operates a local We do not anticipate this plant would section. There would be little national quarry from which they obtain rock that install an RTO to comply with the rule. energy demand increase from the does not contain sulfur. The low-sulfur The total industry-wide costs for operation of any of the control options slag or rock is used in the cupola in monitoring for COS, HF, and HCl from analyzed under the proposed NESHAP place of high-sulfur slag. Because sulfur the cupolas is $146,000, while the total amendments. is not added into the cupola with the costs for monitoring for formaldehyde, raw materials, it is not emitted as sulfur phenol, and methanol from the G. What are the cost impacts for the compounds from the stack in the form combined collection and curing Mineral Wool Production source of COS or SO during production. As operations is $42,000. category? 2 shown in their title V permit, another The total annualized costs for the Each facility was evaluated for its plant uses oxygen injection to accelerate proposed rule are estimated at $548,000 ability to meet the proposed emissions the reaction of COS to CO2 and SO2, (2010 dollars). Table 13 provides a limits for PM, COS, HF, and HCl thereby reducing that company’s COS summary of the estimated costs and emissions from cupolas and emissions. emissions reductions associated with formaldehyde, phenol, and methanol However, most mineral wool plants the proposed amendments to the emissions from combined collection have installed regenerative thermal Mineral Wool Production NESHAP operations and curing designs. The oxidizers to convert the high presented in this action.

TABLE 13—ESTIMATED COSTS AND REDUCTIONS FOR THE MINERAL WOOL PRODUCTION PROPOSED STANDARDS IN THIS ACTION

Total HAP Cost effective- Estimated cap- Estimated emissions re- ness in $ per Proposed amendment ital cost ($MM) annual cost ductions (tons ton total HAP ($MM) per year) reduction

COS limit; Low-Sulfur Materials ...... 0 0.360 41 8,780 Additional testing and monitoring ...... 0 0.243 N/A N/A

H. What are the economic impacts for pass through the entire cost of the rule expected to result in approximately 23 the Mineral Wool Production source to consumers, we would expect prices percent reduction in COS; HF, and HCl category? to increase by less than one percent, are not reduced. We have not quantified with no change in output. Conversely, if the monetary benefits associated with We performed an economic impact producers could not pass through any of these reductions. analysis for mineral wool producers the cost by increasing the price, we J. What demographic groups might nationally using the annual compliance would expect output to decline by less benefit the most from this regulation? costs estimated for this proposed than one percent. rule.(Economic Impact and Initial Hence, the overall economic impact of The worst-case nature of our acute Regulatory Flexibility Analysis. October this proposed rule should be low on screening assessment suggests that the 2011).43 The impacts to most producers most of the affected industry and its potential for adverse effects carries a affected by this proposed rule are consumers. For more information, relatively low probability of occurrence. annualized costs of less than one please refer to the Economic Impact The EPA concludes that, based on our percent of their revenues using the most Analysis for this proposed rulemaking analyses, the risks associated with current year available for revenue data. that is available in the public docket. Id. MACT-allowable and actual emissions One producer will experience an (primarily due to formaldehyde I. What are the benefits for the Mineral annualized cost of 6.7 percent of its emissions from stacks) from this source Wool Production source category? revenue, however. Both demand and category are acceptable. Thus, a supply in this sector are inelastic to The proposed Mineral Wool demographic analysis was not price changes. Thus, if producers could Production NESHAP amendments are conducted.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00037 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72806 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

X. Summary of Cost, Environmental, determined based on the need for disposal of high chrome refractories in and Economic Impacts for the Wool additional control technologies and landfills or in other areas that are not Fiberglass Manufacturing Source actions required to meet the proposed designed or permitted to receive Category emissions limits. The Economic Impact hexavalent chromium waste. Water Analysis considered annual sales and quality and solid waste impacts are also A. What are the affected sources in the revenue data from the facilities within Wool Fiberglass Manufacturing source possible from potential reuse of spent this source category and their ability to category? high chrome refractory products. meet the proposed amendments. The Because of their durability, we believe We evaluated the impacts to the following sections discuss the cost, that use of refractory bricks made with affected sources based on all available environmental, and economic impacts high chrome content are becoming information, including two significant to the Wool Fiberglass Manufacturing widespread,44 (Chromium in sources: the 2010 emissions testing and source category. (Economic Impact Refractories),11 as their use can nearly subsequent conversations with NAIMA Analysis for the Mineral Wool and Wool double the life of glass furnaces (Notes and individuals operating industry Fiberglass RTRs. U.S. EPA. October of April 14, 2011, Certainteed; Region 7 facilities. According to the 2010 2011.) Certainteed Notes; August 31, 2011 emissions test data, there are 3 furnaces 12 13 35 C. What are the air quality impacts? Meeting). When glass furnaces at 3 facilities that do not meet this reach the end of their useful life and proposed chromium emission limit. In The EPA estimated the emissions must be rebuilt, the high chrome their responses to the survey conducted reductions that are expected to result refractory brick from demolition of the by the industry, facilities stated the from the proposed amendments to the old furnace is typically discarded, as it tested furnaces were representative of 1999 MACT rule compared to the 2010 typically cannot be used in new furnace the untested furnaces. However, furnace baseline emissions estimates. A detailed construction. As for any industrial construction materials (refractory documentation of the analysis can be waste, the bricks from an old glass composition) were not one of the factors found in: Cost Impacts of the Revised furnace would, when discarded, considered in determining NESHAP for the Wool Fiberglass potentially be subject to the Resource representativeness. Manufacturing Source Category. We Conservation and Recovery Act (RCRA) After the completion of the survey expect reductions of formaldehyde, and its regulations. conducted by industry, we received phenol and methanol, and chromium information that emissions testing for compounds. Additionally, NaOH scrubber solids chromium may not necessarily be Emissions of formaldehyde, PM, and are expected to contain high levels of representative of other furnaces that HAP metals from wool fiberglass hexavalent chromium removed from were not tested. Therefore, we based our manufacturing have declined over the furnace emissions. The proper disposal assessment of the impacts upon the last 12 years as a result of federal rules, procedures for hexavalent chromium- tested furnaces only, and did not state rules and on the industry’s own contaminated waste are provided under include in that assessment untested initiative. The current proposal is RCRA regulations (40 CFR 262.11). furnaces. expected to yield emission reductions E. What are the secondary impacts? Based on this approach, we anticipate for formaldehyde, phenol, and methanol that all 29 wool fiberglass from their current levels. However, the Indirect or secondary air quality manufacturing facilities currently proposed amendments are expected to impacts include impacts that will result operating in the United States will be discourage facilities in the wool from the increased electricity usage affected by these proposed amendments, fiberglass industry from reintroducing associated with the operation of control 2 of the 29 wool fiberglass formaldehyde to their production lines. devices, as well as water quality and manufacturing facilities currently In addition, the proposed chromium solid waste impacts that might occur as operating in the United States will compound emission limit would a result of these proposed actions. We install air pollution controls, and that prevent emissions of chromium estimate the proposed amendments will one facility will reconstruct a furnace to compounds in the future and discourage not result in any significant secondary comply with these proposed the replacement of currently operating impacts from the requirements of the amendments. Additionally, industry has furnaces with those constructed of high Mineral Wool MACT amendments stated that no major wool fiberglass chromium refractory bricks. because facilities can meet the COS residential insulation sources will still Based on the emissions data available limits without installing RTOs. We do exist in this source category by the time to the EPA, we believe that all affected not anticipate significant secondary the proposed rules are promulgated. If facilities will be able to comply with the impacts from the proposed amendments their predictions come to pass, we proposed emissions limits for to the Wool Fiberglass MACT. estimate that two facilities will be formaldehyde, phenol, methanol, HF, F. What are the energy impacts? affected by these proposed amendments; and HCl without additional controls. these are pipe insulation facilities. Additional controls are required for Energy impacts in this section are However, any major sources still in major sources with high-chrome those energy requirements associated operation at the time the amendments refractories. Additionally, as discussed with the operation of emission control are promulgated will be affected by this in Section X.J of this preamble, the EPA devices. Potential impacts on the rule. One new facility was recently has determined that the proposed rule national energy economy from the built, but no facilities are expected to be will not have disproportionately high proposed amendments to the Wool constructed in the foreseeable future. and adverse human health or Fiberglass MACT are expected to be environmental effects on minority or B. How are the impacts for this proposal minimal and will not result in a low-income populations. evaluated? significant increase in national energy demand. For the proposed Wool Fiberglass D. What are the water quality and solid waste impacts? Manufacturing NESHAP amendments, 44 Excel spreadsheet provided by North American the air quality, water quality, solid We anticipate water quality and solid Insulation Manufacturers Association (NAIMA). waste, and energy impacts were waste impacts may result from the Non-CBI NAIMA Response to Cr Emissions 8.11.11.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00038 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72807

G. What are the cost impacts? $60,000 for compliance testing on glass- incurred. However, in the event that the The capital costs for each facility were melting furnaces and $52,000 for three facilities that do not now meet the estimated based on the ability for each compliance testing on the FA chromium compounds limit were to facility to meet the proposed emissions manufacturing line for pipe insulation remain major sources, we estimated the limits for PM, chromium compounds, products. The total annualized costs for annualized control costs as between HF, HCl, formaldehyde, phenol, and the proposed rule are estimated at $100,000 to $300,000 per furnace, methanol. The memorandum, Cost $112,000 (2010 dollars). Table 14 depending on which of two options is Impacts of the Revised NESHAP for the provides a summary of the costs and used. Nine hundred seventy (970) Wool Fiberglass Manufacturing Source emission reductions associated with the pounds of chromium compounds per Category, includes a complete proposed amendments if the three year would be reduced at three major description of the cost estimate methods facilities with high levels of hexavalent sources in the industry, 913 pounds of used for this analysis and is available in chromium install controls or reconstruct this from a single facility. Hexavalent the docket. Under the proposed furnaces to meet the emission limits of chromium is 92% of the total chromium amendments, the majority of wool the proposed rule. Because the industry compounds emitted from wool fiberglass facilities are not expected to is undergoing the phaseout of HAP fiberglass furnaces. Actual facility costs incur any capital costs to comply with binders, no major sources are expected would be determined by the number of the proposed emissions limits. The total to exist by the compliance deadline for furnaces, the associated level of Cr costs estimated for compliance with the this proposed rule, and no costs to emissions, and the major source status amendments proposed in this action are industry beyond testing would be of the facility.

TABLE 14—ESTIMATED COSTS AND REDUCTIONS FOR THE PROPOSED WOOL FIBERGLASS MANUFACTURING STANDARDS IN THIS ACTION

Total HAP emissions Cost effective- Proposed amendment Est. capital Est. annual reductions ness in $ per Number of cost ($MM) cost ($MM) (pounds per pound facilities year)

Change out of refractory brick lining ...... 6.0 0.3 900 333 1 Installation of NaOH scrubber ...... 0.25 0.1 70 1400 2 Additional testing and monitoring for glass-melting fur- naces ...... 0 0.06 N/A N/A ...... Additional testing and monitoring for FA lines for pipe in- sulation products...... 0 0.052 N/A N/A ......

H. What are the economic impacts? Fiberglass RTRs. U.S. EPA. October percentage. See the Risk and 2011.) Technology Review—Analysis of Socio- We performed an economic impact Economic Factors for Populations Living I. What are the benefits? analysis for the wool fiberglass industry Near Wool Fiberglass Facilities in the using the annual compliance costs As stated in section X.C., we expect docket for additional details on the estimated for this proposed rule emissions reductions of PM, phenol, demographic analysis. (Economic Impact and Initial Regulatory formaldehyde, methanol, and The EPA has determined that the Flexibility Analysis for the Proposed chromium compounds. We have not current health risks posed to anyone by Mineral Wool and Wool Fiberglass Risk quantified the monetary benefits emissions from this source category are and Technology Review).43 The impacts associated with these reductions. acceptable. Therefore, the EPA has to producers affected by this proposed J. What demographic groups might determined that the proposed rule will rule are annualized costs of less than 0.1 benefit the most from this regulation? not have disproportionately high and percent of their revenues using the most adverse human health or environmental For the proposed wool fiberglass rule, current year available for revenue data. effects on minority or low-income the EPA has determined that the current populations. With the responsiveness of wool health risks posed to anyone by fiberglass demand and supply at less emissions from this source category are XI. Request for Comments than 1:1 compared to a price change, acceptable. However, there are about We are soliciting comments on all and with the change in product price as 849,000 people nationwide that are aspects of this proposed action. All approximated by the cost to revenue currently subject to health risks which comments received during the comment ratio at less than 0.1 percent, for this are non-negligible (i.e., cancer risks period will be considered. In addition to ratio is the maximum price change that greater than 1-in-1 million) due to general comments on this proposed producers may face, it is expected that emissions from this source category. We action, we are also interested in any wool fiberglass price and output performed an analysis of the additional data that may help to address changes will be less than 0.1 percent. demographic makeup of these 849,000 emissions of chromium compounds Hence, the overall economic impact of people. The demographic distribution of from wool fiberglass manufacturing this proposed rule should be low on the this ‘‘at-risk’’ population is similar to furnaces, such as speciation of the affected industry and its consumers. For the national distribution of different types of chromium compounds more information, please refer to the demographics for all groups except for that may be used in the manufacture of Economic Impact Analysis for this the ‘‘minority’’ group (defined as total refractory bricks, shapes, and castables; proposed rulemaking that is available in population minus the white and the properties of different the public docket. (Economic Impact population), which is 11 percent greater chromium compounds when exposed to Analysis for the Mineral Wool and Wool than its corresponding national temperatures exceeding 1500°C.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00039 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72808 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

Specifically, we are interested in data data or information. Section VII of this identify the data in question, provide we can use to support any of the preamble provides more information on your reason for concern, and provide proposed alternatives and new data that submitting data. any ‘‘improved’’ data that you have, if could support an alternative not XII. Submitting Data Corrections available. When you submit data, we proposed in these actions. We are also request that you provide documentation interested in additional data that may The site-specific emissions profiles of the basis for the revised values to used in the source category risk and help to reduce the uncertainties support your suggested changes. To demographic analyses are available for inherent in the risk assessments and submit comments on the data download on the RTR web page at: downloaded from the RTR Web page, other analyses. We are specifically http://www.epa.gov/ttn/atw/rrisk/ complete the following steps: interested in receiving corrections to the rtrpg.html. The data files include site-specific emissions profiles used for detailed information for each HAP 1. Within this downloaded file, enter risk modeling. Such data should include emissions release point for the facility suggested revisions to the data fields supporting documentation in sufficient included in the source category. appropriate for that information. The detail to allow characterization of the If you believe that the data are not data fields that may be revised include quality and representativeness of the representative or are inaccurate, please the following:

Data element Definition

Control Measure ...... Are control measures in place? (yes or no). Control Measure Comment ...... Select control measure from list provided, and briefly describe the control measure. Delete ...... Indicate here if the facility or record should be deleted. Delete Comment ...... Describes the reason for deletion. Emissions Calculation Method Code For Re- Code description of the method used to derive emissions. For example, CEM, material bal- vised Emissions. ance, stack test, etc. Emissions Process Group ...... Enter the general type of emissions process associated with the specified emissions point. Fugitive Angle ...... Enter release angle (clockwise from true North); orientation of the y-dimension relative to true North, measured positive for clockwise starting at 0 degrees (maximum 89 degrees). Fugitive Length ...... Enter dimension of the source in the east-west (x-) direction, commonly referred to as length (ft). Fugitive Width ...... Enter dimension of the source in the north-south (y-) direction, commonly referred to as width (ft). Malfunction Emissions ...... Enter total annual emissions due to malfunctions (tpy). Malfunction Emissions Max Hourly ...... Enter maximum hourly malfunction emissions here (lb/hr). North American Datum ...... Enter datum for latitude/longitude coordinates (NAD27 or NAD83); if left blank, NAD83 is as- sumed. Process Comment ...... Enter general comments about process sources of emissions. REVISED Address ...... Enter revised physical street address for MACT facility here. REVISED City ...... Enter revised city name here. REVISED County Name ...... Enter revised county name here. REVISED Emissions Release Point Type ...... Enter revised Emissions Release Point Type here. REVISED End Date ...... Enter revised End Date here. REVISED Exit Gas Flow Rate ...... Enter revised Exit Gas Flowrate here (ft 3/sec). REVISED Exit Gas Temperature ...... Enter revised Exit Gas Temperature here (F). REVISED Exit Gas Velocity ...... Enter revised Exit Gas Velocity here (ft/sec). REVISED Facility Category Code ...... Enter revised Facility Category Code here, which indicates whether facility is a major or area source. REVISED Facility Name ...... Enter revised Facility Name here. REVISED Facility Registry Identifier ...... Enter revised Facility Registry Identifier here, which is an ID assigned by the EPA Facility Registry System. REVISED HAP Emissions Performance Level Enter revised HAP Emissions Performance Level here. Code. REVISED Latitude ...... Enter revised Latitude here (decimal degrees). REVISED Longitude ...... Enter revised Longitude here (decimal degrees). REVISED MACT Code ...... Enter revised MACT Code here. REVISED Pollutant Code ...... Enter revised Pollutant Code here. REVISED Routine Emissions ...... Enter revised routine emissions value here (tpy). REVISED SCC Code ...... Enter revised SCC Code here. REVISED Stack Diameter ...... Enter revised Stack Diameter here (ft). REVISED Stack Height ...... Enter revised Stack Height here (ft). REVISED Start Date ...... Enter revised Start Date here. REVISED State ...... Enter revised State here. REVISED Tribal Code ...... Enter revised Tribal Code here. REVISED Zip Code ...... Enter revised Zip Code here. Shutdown Emissions ...... Enter total annual emissions due to shutdown events (tpy). Shutdown Emissions Max Hourly ...... Enter maximum hourly shutdown emissions here (lb/hr). Stack Comment ...... Enter general comments about emissions release points. Startup Emissions ...... Enter total annual emissions due to startup events (tpy). Startup Emissions Max Hourly ...... Enter maximum hourly startup emissions here (lb/hr). Year Closed ...... Enter date facility stopped operations.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00040 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72809

2. Fill in the commenter information Fiberglass Manufacturing. The reported by source operators, only a fields for each suggested revision (i.e., information collection requirements are small number would be expected to commenter name, commenter not enforceable until OMB approves result from a malfunction (based on the organization, commenter email address, them. The information requirements are definition above), and only a subset of commenter phone number, and revision based on notification, recordkeeping, excess emissions caused by comments). and reporting requirements in the malfunctions would result in the source 3. Gather documentation for any NESHAP General Provisions (40 CFR choosing to assert the affirmative suggested emissions revisions (e.g., part 63, subpart A), which are defense. Thus, we believe the number of performance test reports, material mandatory for all operators subject to instances in which source operators balance calculations). national emissions standards. These might be expected to avail themselves of 4. Send the entire downloaded file recordkeeping and reporting the affirmative defense will be with suggested revisions in Microsoft® requirements are specifically authorized extremely small. For this reason, we did Access format and all accompanying by CAA section 114 (42 U.S.C. 7414). not estimate any such occurrences for documentation to Docket ID Number All information submitted to the EPA all sources subject to subparts DDD and EPA–HQ–OAR–2010–1041 for the pursuant to the recordkeeping and NNN over the 3-year period covered by Mineral Wool Production source reporting requirements for which a these ICRs. We expect to gather category and Docket ID number EPA– claim of confidentiality is made is information on such events in the future HQ–OAR–2010–1042 for the Wool safeguarded according to agency and will revise this estimate as better Fiberglass Manufacturing source policies set forth in 40 CFR part 2, information becomes available. category (through one of the methods subpart B. We estimate 7 regulated entities are described in the ADDRESSES section of For this proposed rule, the EPA is currently subject to subpart DDD and this preamble). To expedite review of adding affirmative defense to the will be subject to all proposed the revisions, it would also be helpful estimate of burden in the ICRs. To standards. The annual monitoring, if you submitted a copy of your provide the public with an estimate of reporting, and recordkeeping burden for revisions to the EPA directly at the relative magnitude of the burden this collection (averaged over the first 3 [email protected] in addition to submitting associated with an assertion of the years after the effective date of the them to the docket. affirmative defense position adopted by standards) for these amendments to 5. If you are providing comments on a source, the EPA has provided subpart DDD (Mineral Wool Production) a facility, you need only submit one file administrative adjustments to these is estimated to be $85,348 per year. This for that facility, which should contain ICRs to show what the notification, estimate includes performance tests, all suggested changes for all sources at recordkeeping and reporting notifications, reporting, and that facility. We request that all data requirements associated with the recordkeeping associated with the new revision comments be submitted in the assertion of the affirmative defense requirements for COS, HF, and HCl from form of updated Microsoft® Access might entail. The EPA’s estimate for the cupolas and formaldehyde, phenol, and files, which are provided on the RTR required notification, reports and methanol from combined collection and Web Page at: http://www.epa.gov/ttn/ records for any individual incident curing oven designs. The total burden atw/rrisk/rtrpg.html. totals $3,141 and is based on the time for the Federal government (averaged and effort required of a source to review over the first 3 years after the effective XIII. Statutory and Executive Order relevant data, interview plant date of the standard) is estimated to be Reviews employees, and document the events 22 hours per year at a total labor cost of A. Executive Order 12866: Regulatory surrounding a malfunction that has $970 per year. Burden is defined at 5 Planning and Review and Executive caused an exceedance of an emissions CFR 1320.3(b). We estimate 29 regulated entities are Order 13563: Improving Regulation and limit. The estimate also includes time to currently subject to subpart NNN and Regulatory Review produce and retain the record and reports for submission to the EPA. The only 2 will be subject to all proposed Under Executive Order 12866 (58 FR EPA provides this illustrative estimate standards. The annual monitoring, 51735, October 4, 1993), this action is a of this burden because these costs are reporting, and recordkeeping burden for significant regulatory action because it only incurred if there has been a this collection (averaged over the first 3 raises novel legal and policy issues. violation and a source chooses to take years after the effective date of the Accordingly, the EPA submitted this advantage of the affirmative defense. standards) for these amendments to action to the Office of Management and Given the variety of circumstances subpart NNN (Wool Fiberglass Budget (OMB) for review under under which malfunctions could occur, Manufacturing) is estimated to be Executive Orders 12866 and 13563 (76 as well as differences among sources’ $14,000 per year. This estimate includes FR 3821, January 21, 2011) and any operation and maintenance practices, performance tests, notifications, changes made in response to OMB we cannot reliably predict the severity reporting, and recordkeeping associated recommendations have been and frequency of malfunction-related with the new requirements for PM, documented in the docket for this excess emissions events for a particular chromium compounds, HF, and HCl action. source. It is important to note that the from glass-melting furnaces and EPA has no basis currently for formaldehyde, phenol, and methanol B. Paperwork Reduction Act estimating the number of malfunctions from both RS and FA manufacturing The information collection that would qualify for an affirmative lines. The total burden for the Federal requirements in this rule have been defense. Current historical records government (averaged over the first 3 submitted for approval to the OMB would be an inappropriate basis, as years after the effective date of the under the Paperwork Reduction Act, 44 source owners or operators previously standard) is estimated to be 6.3 hours U.S.C. 3501 et seq. The Information operated their facilities in recognition per year at a total labor cost of $283 per Collection Request (ICR) documents that they were exempt from the year. prepared by the EPA have been assigned requirement to comply with emissions An agency may not conduct or EPA ICR numbers 1799.06 for Mineral standards during malfunctions. Of the sponsor, and a person is not required to Wool Production and 1160.10 for Wool number of excess emissions events respond to, a collection of information

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00041 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72810 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

unless it displays a currently valid OMB category, which has the NAICS code in aggregate, or the private sector in any control number. The OMB control 327993 (i.e., Mineral Wool Production 1 year. The proposed rule imposes no numbers for the EPA’s regulations in 40 and Wool Fiberglass Manufacturing), enforceable duties on any State, local or CFR are listed in 40 CFR part 9. When the SBA small business size standard is Tribal governments or the private sector. these ICRs are approved by OMB, the 500 employees according to the SBA Thus, this proposed rule is not subject agency will publish a technical small business standards definitions. to the requirements of sections 202 or amendment to 40 CFR part 9 in the After considering the economic 205 of the UMRA. Federal Register to display the OMB impacts of this proposed rule on small control numbers for the approved entities in the Mineral Wool Production This proposed rule is also not subject information collection requirements and Wool Fiberglass Manufacturing to the requirements of section 203 of contained in the final rules. source categories, I certify that this UMRA because it contains no regulatory To comment on the agency’s need for action will not have a significant requirements that might significantly or this information, the accuracy of the economic impact on a substantial uniquely affect small governments provided burden estimates, and any number of small entities. Five of the 6 because it contains no requirements that suggested methods for minimizing Mineral Wool Production parent apply to such governments nor does it respondent burden, the EPA has companies affected are considered to be impose obligations upon them. established a public docket for this rule, small entities per the definition which includes this ICR, under Docket provided in this section. However, we E. Executive Order 13132: Federalism ID number EPA–HQ–OAR–2010–1041 estimate that this proposed action will This proposed rule does not have for the Mineral Wool Production source not have a significant economic impact federalism implications. It will not have category and Docket ID number EPA– on those companies. The impact of this substantial direct effects on the States, HQ–OAR–2010–1042 for the Wool proposed action on these companies Fiberglass Manufacturing source will be an annualized compliance cost on the relationship between the national category. Submit any comments related of less than one percent of its revenues. government and the States, or on the to the ICRs to the EPA and the OMB. Only one of the five small parent distribution of power and See the ADDRESSES section at the companies is expected to have an responsibilities among the various beginning of this notice for where to annualized compliance cost of greater levels of government, as specified in submit comments to the EPA. Send than one percent of its revenues. All Executive Order 13132. None of the comments to the OMB at the Office of other affected parent companies are not facilities subject to this action are Information and Regulatory Affairs, small businesses according to the SBA owned or operated by State Office of Management and Budget, 725 small business size standard for the governments, and, because no new 17th Street NW., Washington, DC 20503, affected NAICS code (NAICS 327993). requirements are being promulgated, Attention: Desk Office for the EPA. One Wool Fiberglass Manufacturing nothing in this proposed rule will Since the OMB is required to make a facility is considered to be owned by a supersede State regulations. Thus, decision concerning the ICR between 30 small business, but this facility will not Executive Order 13132 does not apply and 60 days after November 25, 2011, a experience an impact from this to this proposed rule. comment to OMB is best assured of proposed rule. We have determined that In the spirit of Executive Order 13132, having its full effect if the OMB receives the impacts do not constitute a it by December 27, 2011. The final rule significant economic impact on a and consistent with the EPA policy to will respond to any OMB or public substantial number of small entities in promote communications between the comments on the information collection the Wool Fiberglass Manufacturing EPA and State and local governments, requirements contained in this proposal. source category (See: Economic Impact the EPA specifically solicits comment and Small Business Analysis for the on this proposed rule from State and C. Regulatory Flexibility Act proposed Mineral Wool and Wood local officials. The RFA generally requires an agency Fiberglass Production Source Categories F. Executive Order 13175: Consultation to prepare a regulatory flexibility NESHAP). analysis of any rule subject to notice Although this proposed rule will not and Coordination With Indian Tribal and comment rulemaking requirements have a significant economic impact on Governments under the Administrative Procedure Act a substantial number of small entities, This proposed rule does not have or any other statute unless the agency the EPA nonetheless has tried to reduce Tribal implications, as specified in certifies that the rule will not have a the impact of this rule on small entities. Executive Order 13175 (65 FR 67249, significant economic impact on a For more information, please refer to the November 9, 2000). Thus, Executive substantial number of small entities. economic impact and small business Order 13175 does not apply to this Small entities include small businesses, analysis that is in the docket. We action. small organizations, and small continue to be interested in the governmental jurisdictions. potential impacts of the proposed rule The EPA specifically solicits For purposes of assessing the impacts on small entities and welcome additional comment on this proposed of this proposed rule on small entities, comments on issues related to such action from Tribal officials. small entity is defined as: (1) A small impacts. business as defined by the SBA’s G. Executive Order 13045: Protection of regulations at 13 CFR 121.201; (2) a D. Unfunded Mandates Reform Act Children From Environmental Health small governmental jurisdiction that is a This proposed rule does not contain Risks and Safety Risks government of a city, county, town, a Federal mandate under the provisions This proposed rule is not subject to school district or special district with a of Title II of the UMRA of 1995, 2 U.S.C. population of less than 50,000; and (3) 1531–1538 for State, local, or Tribal Executive Order 13045 (62 FR 19885, a small organization that is any not-for- governments or the private sector. The April 23, 1997) because it is not profit enterprise that is independently proposed rule would not result in economically significant as defined in owned and operated and is not expenditures of $100 million or more Executive Order 12866. dominant in its field. For this source for State, local, and Tribal governments,

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00042 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72811

H. Executive Order 13211: Actions Annexes to ASTM D6348–03, Sections The EPA welcomes comments on this Concerning Regulations That A1 through A8 are mandatory; and (2) aspect of the proposed rulemaking and, Significantly Affect Energy Supply, In ASTM D6348–03 Annex A5 (Analyte specifically, invites the public to Distribution, or Use Spiking Technique), the percent R identify potentially-applicable VCS and This action is not a ‘‘significant (percent R) must be determined for each to explain why such standards should energy action’’ as defined under target analyte (Equation A5.5). In order be used in this regulation. Executive Order 13211, ‘‘Actions for the test data to be acceptable for a compound, percent R must be 70 J. Executive Order 12898: Federal Concerning Regulations That percent ≥ R ≤ 130 percent. If the percent Actions To Address Environmental Significantly Affect Energy Supply, R value does not meet the criterion for Justice in Minority Populations and Distribution, or Use’’ (66 FR 28355, May a target compound, the test data is not Low-Income Populations 22, 2001), because it is not likely to have acceptable for that compound and the Executive Order 12898 (59 FR 7629, significant adverse effect on the supply, test must be repeated for that analyte February 16, 1994) establishes federal distribution, or use of energy. This (i.e., the sampling and/or analytical executive policy on EJ. Its main action will not create any new procedure should be adjusted before a provision directs federal agencies, to the requirements and therefore no retest). The percent R value for each greatest extent practicable and additional costs for sources in the compound must be reported in the test permitted by law, to make EJ part of energy supply, distribution, or use report, and all field measurements must their mission by identifying and sectors. be corrected with the calculated percent addressing, as appropriate, I. National Technology Transfer and R value for that compound by using the disproportionately high and adverse Advancement Act following equation: Reported Result = human health or environmental effects (Measured Concentration in the Stack × Section 12(d) of the National of their programs, policies and activities 100)/percent R. Technology Transfer and Advancement on minority populations and low- In addition, ASTM D6784–02 (2008), income populations in the United Act of 1995 (NTTAA), Public Law No. Standard Test Method for Elemental, 104–113 (15 U.S.C. 272 note), directs States. Oxidized, Particle-Bound and Total For the proposed mineral wool rule, the EPA to use VCS in its regulatory Mercury Gas Generated from Coal-Fired activities unless to do so would be the EPA has determined that the rule Stationary Sources (Ontario Hydro will not have disproportionately high inconsistent with applicable law or Method) is acceptable as an alternative otherwise impractical. VCS are and adverse human health or to Method 29 in the subpart NNN rule. environmental effects on minority or technical standards (e.g., materials The search identified four other VCS specifications, test methods, sampling low-income populations, because it that were potentially applicable for the increases the level of environmental procedures, and business practices) that Mineral Wool Production rule in lieu of are developed or adopted by voluntary protection for all affected populations EPA reference methods. However, after without having any disproportionately consensus standards bodies. NTTAA reviewing the available standards, EPA directs the EPA to provide Congress, high and adverse human health or determined that four candidate VCS environmental effects on any through OMB, explanations when the (ASTM D3685/D3685M–98 [2005], ISO agency decides not to use available and population, including any minority or 9096:1992 [2003], CAN/CSA Z223.1– low-income population. applicable VCS. M1977, ANSI/ASME PTC 38 1980 The proposed rule involves technical For the proposed wool fiberglass rule, [1985]) identified for measuring the EPA has determined that the current standards. Therefore, the requirements emissions of pollutants or their of the NTTAA apply to this action. We health risks posed to anyone by surrogates subject to emission standards emissions from this source category are conducted searches for the RTR for the in the rule would not be practical due Mineral Wool Production and Wool acceptable. Therefore, the EPA has to lack of equivalency, documentation, determined that the proposed rule will Fiberglass Manufacturing NESHAP validation data and other important through the Enhanced NSSN Database not have disproportionately high and technical and policy considerations. adverse human health or environmental managed by the American National Under the Wool Fiberglass rule, the effects on minority or low-income Standards Institute (ANSI). We also search identified six other VCS that populations. contacted VCS organizations and were potentially applicable in lieu of accessed and searched their databases. EPA reference methods (EN 13211:2001, List of Subjects in 40 CFR Part 63 Under 40 CFR part 63 subpart DDD, CAN/CSA Z223.26–M1986, ASTM Environmental protection, Air searches were conducted for EPA D3685/D3685M–98 [2005], ISO pollution control, Incorporation by Methods 5, 318, and 320 of 40 CFR Part 9096:1992 [2003], CAN/CSA Z223.1– reference, Mineral wool, Wool 60, Appendix A. Under 40 CFR part 63 M1977, and ANSI/ASME PTC 38 1980 fiberglass, Reporting and recordkeeping subpart NNN, searches were conducted [1985]). However, the EPA determined requirements. for EPA Methods 5, 318, 320, 29, and that these methods would not be 0061 of 40 CFR Part 60, Appendix A. No practical due to lack of equivalency, Dated: November 4, 2011. applicable voluntary consensus documentation, validation data and Lisa P. Jackson, standards were identified for EPA other important technical and policy Administrator. Method 318 and SW–846 Method 0061. considerations. For the reasons stated in the One voluntary consensus standard The VCS searches are documented in preamble, part 63 of title 40, chapter I, ASTM D6348–03 (2010), Determination the Voluntary Consensus Standard of the Code of Federal Regulations is of Gaseous Compounds by Extractive Results for the Risk and Technology proposed to be amended as follows: Direct Interface Fourier Transform Review for the Mineral Wool NESHAP (FTIR) Spectroscopy is acceptable as an and the Voluntary Consensus Standard PART 63—[AMENDED] alternative to Method 320 for both Results for the Risk and Technology subparts DDD and NNN, but with Review for the Wool Fiberglass NESHAP 1. The authority citation for part 63 several conditions: (1) The test plan memorandums as provided in the continues to read as follows: preparation and implementation in the docket. Authority: 42 U.S.C. 7401, et seq.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00043 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72812 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

Subpart DDD—[Amended] 4. Section 63.1180 is amended by (A) Were caused by a sudden, revising paragraphs (a), (b), and (d), and infrequent, and unavoidable failure of 2. Section 63.1178 is amended by adding paragraph (e) to read as follows: air pollution control and monitoring revising paragraph (a)(2) and adding equipment, process equipment, or a paragraphs (a)(3) and (4) to read as § 63.1180 When must I meet these process to operate in a normal or usual standards? follows: manner; and (a) Existing cupolas and combined § 63.1178 For cupolas, what standards (B) Could not have been prevented must I meet? collection/curing operations. (1) Except through careful planning, proper design as noted in paragraph (a)(2) of this or better operation and maintenance * * * * * section, the compliance date for an (a) * * * practices; and owner or operator of an existing plant or (C) Did not stem from any activity or (2) Limit emissions of carbonyl source subject to the provisions of this event that could have been foreseen and sulfide (COS) from each existing, new, subpart is June 2, 2002 or June 3, 2003 avoided, or planned for; and or reconstructed cupola to the if you applied for and received a one- (D) Were not part of a recurring following: year extension under section pattern indicative of inadequate design, (i) 3.3 lb of COS per ton of melt or less 112(i)(b)(3)(B) of the Act. operation, or maintenance. for existing cupolas. (2) The compliance dates for existing (ii) Repairs were made as (ii) 0.017 lb of COS per ton of melt or plants and sources are: expeditiously as possible when the less for new or reconstructed cupolas. (i) [DATE 3 YEARS AFTER applicable emissions limitations were (3) Limit emissions of hydrogen PUBLICATION OF THE FINAL RULE being exceeded. Off-shift and overtime fluoride (HF) from each existing, new, IN THE FEDERAL REGISTER] for labor were used, to the extent or reconstructed cupola to 0.014 lb of cupolas and combined collection/curing practicable to make these repairs; and HF per ton of melt or less. operations subject to emission limits in (iii) The frequency, amount and (4) Limit emissions of hydrogen duration of the excess emissions chloride (HCl) from each existing, new, §§ 63.1178 and 63.1179 which became effective [DATE OF PUBLICATION OF (including any bypass) were minimized or reconstructed cupola to 0.0096 lb of to the maximum extent practicable HCl per ton of melt or less. THE FINAL RULE IN THE FEDERAL REGISTER]. during periods of such emissions; and * * * * * (ii) [DATE OF PUBLICATION OF THE (iv) If the excess emissions resulted 3. Section 63.1179 is amended by FINAL RULE IN THE FEDERAL from a bypass of control equipment or revising the section heading and REGISTER] for the provisions related to a process, then the bypass was paragraphs (a) and (b) introductory text malfunctions and affirmative defense unavoidable to prevent loss of life, to read as follows: provisions of paragraph (e) of this personal injury, or severe property damage; and § 63.1179 For combined collection/curing section and the electronic reporting operations, what standards must I meet? provisions of §§ 63.1192(d) and (v) All possible steps were taken to 63.1193(b)(1) and (g). minimize the impact of the excess (a) You must control emissions from emissions on ambient air quality, the each existing and new combined (b) New and reconstructed cupolas and combined collection/curing environment and human health; and collection/curing operations by limiting (vi) All emissions monitoring and operations. For affected sources that emissions of formaldehyde, phenol, and control systems were kept in operation commenced construction or methanol to the following: if at all possible, consistent with safety reconstruction after November 25, 2011, (1) For combined drum collection/ and good air pollution control practices; you must demonstrate compliance with curing operations: and the requirements of this subpart no later (i) 0.067 lb of formaldehyde per ton of (vii) All of the actions in response to than the effective date of the rule or melt or less, the excess emissions were documented upon start-up. (ii) 0.0023 lb of phenol per ton of melt by properly signed, contemporaneous or less, and * * * * * operating logs; and (iii) 0.00077 lb of methanol per ton of (d) See § 63.1197 for requirements (viii) At all times, the affected source melt or less. during startups and shutdowns. was operated in a manner consistent (2) For combined horizontal (e) Affirmative defense for exceedance with good practices for minimizing collection/curing operations: of emissions limits during malfunction. emissions; and (i) 0.054 lb of formaldehyde per ton of In response to an action to enforce the (ix) A written root cause analysis has melt or less, standards set forth in this subpart, you been prepared, the purpose of which is (ii) 0.15 lb of phenol per ton of melt may assert an affirmative defense to a to determine, correct, and eliminate the or less, and claim for civil penalties for exceedances primary causes of the malfunction and (iii) 0.022 lb of methanol per ton of of such standards that are caused by the excess emissions resulting from the melt or less. malfunction, as defined at § 63.2. malfunction event at issue. The analysis (3) For combined vertical collection/ Appropriate penalties may be assessed, must also specify, using best monitoring curing operations: however, if you fail to meet your burden methods and engineering judgment, the (i) 0.46 lb of formaldehyde per ton of of proving all of the requirements in the amount of excess emissions that were melt or less, affirmative defense. The affirmative the result of the malfunction. (ii) 0.52 lb of phenol per ton of melt defense must not be available for claims (2) Notification. The owner or or less, and for injunctive relief. operator of the affected source (iii) 0.63 lb of methanol per ton of (1) To establish the affirmative experiencing an exceedance of its melt or less. defense in any action to enforce such a emissions limit(s) during a malfunction, (b) You must meet the following limit, you must timely meet the must notify the Administrator by operating limits for each combined notification requirements in § 63.1191 of telephone or facsimile transmission as collection/curing operations this subpart, and must prove by a soon as possible, but no later than two subcategory: preponderance of evidence that: business days after the initial * * * * * (i) The excess emissions: occurrence of the malfunction, s/he

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00044 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72813

wishes to be able to use an affirmative (b) Conduct a performance test as § 63.1189 What test methods do I use? defense to civil penalties for that specified in § 63.1188 of this subpart * * * * * malfunction. The owner or operator while manufacturing the product that (g) Method 318 in appendix A to this seeking to assert an affirmative defense, requires a binder formulation made with part for the concentration of must also submit a written report to the the resin containing the highest free- formaldehyde, phenol, methanol, or Administrator within 45 days of the formaldehyde content specification COS. initial occurrence of the exceedance of range. Show compliance with the * * * * * the standards in this subpart. This formaldehyde, phenol, and methanol (i) Method 26A or 320 in appendix A report must demonstrate that the owner/ emissions limits while the device for to this part for the concentration of HF operator met the requirements set forth measuring the control device operating and HCl. in this paragraph (e) and must include parameter is installed, operational, and 9. Section 63.1190 is amended by all necessary supporting documentation. properly calibrated. Establish the revising paragraph (b) introductory text The owner or operator may seek an average operating parameter based on and the ‘‘MW’’ entry under ‘‘where:’’ extension of this deadline for up to 30 the performance test as specified in and by removing paragraph (c). additional days by submitting a written § 63.1185(a) of this subpart. The revision reads as follows: request to the Administrator before the * * * * * expiration of the 45 day period. Until a (d) Following the performance test, § 63.1190 How do I determine compliance? request for an extension has been monitor and record the free- * * * * * approved by the Administrator, the formaldehyde content of each resin lot (b) Using the results from the owner or operator is subject to the and the formulation of each batch of performance tests, you must use the requirement to submit such report binder used, including the following equation to determine within 45 days of the initial occurrence formaldehyde, phenol, and methanol compliance with the COS, HF, HCl, of the exceedance. content. formaldehyde, phenol, and methanol 5. Section 63.1182 is amended by * * * * * numerical emissions limits: revising the section heading, the * * * * * introductory text, and paragraphs (a) 7. Section 63.1188 is amended by and (b) to read as follows: revising paragraphs (b), (c), (d), (e), and MW = Molecular weight of measured (f) to read as follows: pollutant, g/g-mole: § 63.1182 How do I comply with the § 63.1188 What performance test COS = 60.07, HF = 20.01, HCl = 36.46, carbonyl sulfide, hydrogen fluoride, and Formaldehyde = 30.03, Phenol = 94.11, hydrogen chloride standards for existing, requirements must I meet? new, and reconstructed cupolas? * * * * * Methanol = 32.04. To comply with the COS, HF, and (b) Conduct a performance test, * * * * * HCL standards, you must meet the consisting of three test runs, for each 10. Section 63.1191 is amended by following: cupola and/or combined collection/ revising the introductory text to read as (a) Install, calibrate, maintain, and curing operation subject to this subpart follows: operate a device that continuously at the maximum production rate to § 63.1191 What notifications must I measures the operating temperature in demonstrate compliance with each of submit? the firebox of each thermal incinerator. the applicable emissions limits in You must submit written or electronic For the purposes of this rule, the term §§ 63.1178 and 63.1179 of this subpart. notifications to the Administrator as ‘incinerator’ means ‘regenerative (c) Following the initial performance required by § 63.9(b) through (h) of the thermal oxidizer’ (RTO). or compliance test to be conducted general provisions in subpart A of this (b) Conduct a performance test as within 120 days of the effective date of part. Electronic notifications are specified in § 63.1188 of this subpart this rule, you must conduct a encouraged when possible. These that shows compliance with the COS, performance test to demonstrate notifications include, but are not limited HF, and HCl emissions limits while the compliance with each of the applicable to, the following: device for measuring incinerator emissions limits in §§ 63.1178 and * * * * * (regenerative thermal oxidizer) 63.1179 of this subpart at least once 11. Section 63.1192 is amended by operating temperature is installed, every 5 years and as often as the raw revising paragraph (d) to read as operational, and properly calibrated. material ingredients change by more follows: Establish the average operating than 10 percent of those processed temperature based on the performance during the previous performance test. § 63.1192 What recordkeeping test as specified in § 63.1185(a) of this (d) Measure emissions of PM, COS, requirements must I meet? subpart. HF, and HCl from each existing, new, or * * * * * * * * * * reconstructed cupola. (d) Records must be maintained in a 6. Section 63.1183 is amended by (e) Measure emissions of form suitable and readily available for revising the section heading, the formaldehyde, phenol, and methanol expeditious review, according to § 63.10 introductory text, and paragraphs (b) from each existing, new, or of the General Provisions that are and (d) to read as follows: reconstructed combined collection/ referenced in Table 3 to this subpart. § 63.1183 How do I comply with the curing operation. Electronic recordkeeping is encouraged. formaldehyde, phenol, and methanol (f) Measure emissions at the outlet of * * * * * standards for existing, new, and the control device for PM, COS, HF, 12. Section 63.1193 is amended by reconstructed combined collection/curing HCl, formaldehyde, phenol, or redesignating paragraphs (b) through (f) operations? methanol. as paragraphs (c) through (g), and To comply with the formaldehyde, * * * * * adding a new paragraph (b) and by phenol, and methanol standards, you 8. Section 63.1189 is amended by revising the newly redesignated must meet all of the following: revising paragraph (g) and adding paragraph (g) to read as follows: * * * * * paragraph (i) to read as follows: * * * * *

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00045 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72814 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

(b)(1) As of January 1, 2012, and ‘‘affirmative defense’’ and ‘‘combined (b) The owner or operator must not within 60 days after the date of collection/curing operations’’, and shut down items of equipment that are completing each performance test, as revising the definition for ‘‘incinerator’’ utilized for compliance with this defined in § 63.2, and as required in this to read as follows: subpart. subpart, you must submit performance § 63.1196 What definitions should I be (c) Table 1 to subpart DDD test data, except opacity data, aware of? summarizes the emissions limits during electronically to the EPA’s Central Data startups and shutdowns for existing, Exchange by using the ERT (see http:/ * * * * * Affirmative defense means, in the new, and reconstructed cupolas. /www.epa.gov/ttn/chief/ert/erttool.html/ context of an enforcement proceeding, a ) or other compatible electronic response or defense put forward by a TABLE 1 TO SUBPART DDD—EMIS- spreadsheet. Only data collected using defendant, regarding which the SIONS LIMITS DURING STARTUPS test methods compatible with the ERT defendant has the burden of proof, and are subject to this requirement to be AND SHUTDOWNS FOR EXISTING, the merits of which are independently submitted electronically into the EPA’s NEW, AND RECONSTRUCTED CUPO- and objectively evaluated in a judicial WebFIRE database. LAS or administrative proceeding. * * * * * Combined collection/curing [Pound of pollutant per hour] (g) All reports required by this subpart operations means the combination of Emission limit (lb/hr) not subject to the requirements in fiber collection operations and curing paragraph (b) of this section must be ovens used to make bonded products. Pollutant New and sent to the Administrator at the Incinerator means an enclosed air Existing cupolas reconstructed appropriate address listed in § 63.13. If pollution control device that uses cupolas acceptable to both the Administrator controlled flame combustion to convert and the owner or operator of a source, combustible materials to PM ...... 1.0 1.0 these reports may be submitted on noncombustible gases. For the purposes COS ...... 32 0 .17 HF ...... 0.13 0 .13 electronic media. The Administrator of this rule, the term ‘incinerator’ means HCl ...... 0.092 0.092 retains the right to require submittal of ‘regenerative thermal oxidizer’ (RTO). reports subject to paragraph (b) of this * * * * * section in paper format. 14. Add § 63.1197 to read as follows: (d) Table 2 to subpart DDD 13. Section 63.1196 is amended by summarizes the emissions limits during removing the definitions for ‘‘CO’’ and § 63.1197 Startups and shutdowns. startups and shutdowns for existing, ‘‘formaldehyde’’, adding definitions for (a) The provisions set forth in this new, and reconstructed combined subpart apply at all times. collection/curing operations.

TABLE 2 TO SUBPART DDD—EMISSIONS LIMITS DURING STARTUPS AND SHUTDOWNS FOR EXISTING, NEW, AND RECONSTRUCTED COMBINED COLLECTION/CURING OPERATIONS [Pound of pollutant per hour]

Emission limit Design Pollutant (lb/hr)

Vertical ...... Formaldehyde ...... 4.5 Phenol ...... 5.0 Methanol ...... 6.0 Horizontal ...... Formaldehyde ...... 0 .52 Phenol ...... 1.4 Methanol ...... 0.21 Drum ...... Formaldehyde ...... 0.64 Phenol ...... 0.022 Methanol ...... 0.0074

15. Table 1 to subpart DDD of part 63 DDD of part 63 and revised to read as is redesignated as Table 3 to subpart follows:

TABLE 3 TO SUBPART DDD OF PART 63—APPLICABILITY OF GENERAL PROVISIONS (40 CFR PART 63, SUBPART A) TO SUBPART DDD OF PART 63

Applies to Reference subpart DDD Comment

63.1 ...... Yes. 63.2 ...... Yes. 63.3 ...... Yes. 63.4 ...... Yes. 63.5 ...... Yes. 63.6(a), (b), (c) ...... Yes. 63.6(d) ...... No ...... Section reserved. 63.6(e)(1)(i) ...... No ...... See 63.1180 for general duty requirement. 63.6(e)(1)(ii) ...... No. 63.6(e)(1)(iii) ...... Yes. 63.6(e)(2) ...... No ...... Section reserved.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00046 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72815

TABLE 3 TO SUBPART DDD OF PART 63—APPLICABILITY OF GENERAL PROVISIONS (40 CFR PART 63, SUBPART A) TO SUBPART DDD OF PART 63—Continued

Applies to Reference subpart DDD Comment

63.6(e)(3) ...... No. 63.6(f)(1) ...... No. 63.6(g) ...... Yes. 63.6(h) ...... No ...... No opacity limits in rule. 63.6(i) ...... Yes. 63.6(j) ...... Yes. § 63.7(a)–(d) ...... Yes. § 63.7(e)(1) ...... No ...... See 63.1180. § 63.7(e)(2)–(e)(4) ...... Yes. 63.7(f), (g), (h) ...... Yes. 63.8(a)–(b) ...... Yes. 63.8(c)(1)(i) ...... No ...... See 63.1180 for general duty requirement. 63.8(c)(1)(ii) ...... Yes. 63.8(c)(1)(iii) ...... No. 63.8(c)(2)–(d)(2) ...... Yes. 63.8(d)(3) ...... Yes, except for last sentence. 63.8(e)–(g) ...... Yes. 63.9(a), (b), (c), (e), (g), (h)(1) through (3), (h)(5) and Yes. (6), (i) and (j). 63.9(f) ...... No. 63.9(h)(4) ...... No ...... Reserved. 63.10(a) ...... Yes. 63.10(b)(1) ...... Yes. 63.10(b)(2)(i) ...... No. 63.10(b)(2)(ii) ...... No ...... See 63.1193(c) for recordkeeping of occurrence and duration of mal- functions and recordkeeping of actions taken during malfunction. 63.10(b)(2)(iii) ...... Yes. 63.10(b)(2)(iv)–(b)(2)(v) ...... No. 63.10(b)(2)(vi)–(b)(2)(xiv) ...... Yes. 63.(10)(b)(3) ...... Yes. 63.10(c)(1)–(9) ...... Yes. 63.10(c)(10)–(11) ...... No ...... See 63.1192 for recordkeeping of malfunctions. 63.10(c)(12)–(c)(14) ...... Yes. 63.10(c)(15) ...... No. 63.10(d)(1)–(4) ...... Yes. 63.10(d)(5) ...... No ...... See 63.1193 for reporting of malfunctions. 63.10(e)–((f) ...... Yes. 63.11 ...... No ...... Flares will not be used to comply with the emissions limits. 63.12 to 63.15 ...... Yes.

Subpart NNN—[Amended] of this rule, the term ‘incinerator’ means (D) 0.00006 lb of chromium (Cr) ‘regenerative thermal oxidizer’ (RTO). compounds per ton of glass pulled (60 16. Section 63.1381 is amended by * * * * * lb per million tons glass pulled). adding a definition for ‘‘affirmative 17. Section 63.1382 is amended by (ii) The owner or operator of each new defense’’ and revising the definition for revising paragraphs (a) and (b)(6) to read or reconstructed glass-melting furnace ‘‘incinerator’’. as follows: must not discharge or cause to be discharged into the atmosphere in § 63.1381 Definitions. § 63.1382 Emission standards. excess of: * * * * * (a) Emissions limits. (1) Glass-melting (A) 0.0018 lb of PM per ton of glass Affirmative defense means, in the furnaces. On and after the date the pulled; context of an enforcement proceeding, a initial performance test is completed or (B) 0.00078 lb of HF per ton of glass response or defense put forward by a required to be completed under § 63.7 of pulled; and (C) 0.00078 lb of HCl per ton of glass defendant, regarding which the this part, whichever date is earlier, pulled. defendant has the burden of proof, and (i) The owner or operator of each the merits of which are independently (D) 0.00006 lb of Cr compounds per existing glass-melting furnace must not ton of glass pulled (60 lb per million and objectively evaluated in a judicial discharge or cause to be discharged into or administrative proceeding. tons glass pulled). the atmosphere in excess of: (2) Rotary spin manufacturing lines. * * * * * (A) 0.014 pound (lb) of particulate On and after the date the initial Incinerator means an enclosed air matter (PM) per ton of glass pulled; performance test is completed or pollution control device that uses (B) 0.0020 lb of hydrogen fluoride required to be completed under § 63.7 of controlled flame combustion to convert (HF) per ton of glass pulled; and this part, whichever date is earlier, combustible materials to (C) 0.0015 lb of hydrogen chloride (i) The owner or operator of each noncombustible gases. For the purposes (HCl) per ton of glass pulled. existing rotary spin (RS) manufacturing

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00047 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72816 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

line must not discharge or cause to be manufacturing line, or FA 20. Section 63.1385 is amended by discharged into the atmosphere in manufacturing line must install, revising paragraphs (a)(5) and (6), and excess of: calibrate, maintain, and operate a adding paragraphs (a)(11), and (a)(12). (A) 0.17 lb of formaldehyde per ton of monitoring device that continuously glass pulled; measures an appropriate parameter that § 63.1385 Test methods and procedures. (B) 0.19 lb of phenol per ton of glass is correlated to the emission limit (a) * * * pulled; and performance test. (5) Method 5 and Method 202 (40 CFR (C) 0.48 lb of methanol per ton of * * * * * part 60, appendix A) for the glass pulled. 19. Section 63.1384 is amended by concentration of total PM including (ii) The owner or operator of each new revising paragraph (c) introductory text, condensibles. Each run must consist of or reconstructed RS manufacturing line variables E, C, and MW, and adding a minimum run time of 2 hours and a must not discharge or cause to be paragraphs (d) and (e) to read as follows: minimum sample volume of 60 dry discharged into the atmosphere in standard cubic feet (dscf). The probe excess of: § 63.1384 Performance test requirements. and filter holder heating system may be (A) 0.020 lb of formaldehyde per ton * * * * * set to provide a gas temperature no of glass pulled; (c) To determine compliance with the greater than 177 ± 14°C (350 ± 25°F); (B) 0.0011 lb of phenol per ton of emission limit for formaldehyde, (6) Method 318 (appendix A of this glass pulled; and phenol, or methanol for RS subpart) for the concentration of (C) 0.00067 lb of methanol per ton of manufacturing lines and FA formaldehyde, phenol, and methanol. glass pulled. manufacturing lines, and for chromium Each run must consist of a minimum (3) Flame attenuation manufacturing compounds, HF, or HCl for glass- run time of 2 hours; lines. On and after the date the initial melting furnaces, use the following * * * * * performance test is completed or equation: (11) Method 0061 (appendix A of this required to be completed under § 63.7 of * * * * * subpart) for the concentration of this part, whichever date is earlier, E = Emission rate of formaldehyde, chromium compounds and hexavalent (i) The owner or operator of each phenol, methanol, chromium chromium. Each run must consist of a existing flame attenuation (FA) compounds, HF, or HCl, kg/Mg (lb/ton) minimum run time of 1 hour. manufacturing line that produces heavy- of glass pulled; (12) Method 26A or Method 320 density wool fiberglass and/or pipe C = Measured volume fraction of (appendix A of this subpart) for the insulation must not discharge or cause concentration of HF and HCl. Each run to be discharged into the atmosphere in formaldehyde, phenol, methanol, chromium compounds, HF, or HCl, must consist of a minimum run time of excess of: 1 hour. (A) 5.6 lb of formaldehyde per ton of ppm; glass pulled; MW = Molecular weight of * * * * * (B) 1.4 lb of phenol per ton of glass formaldehyde, 30.03 g/g-mol; molecular 21. Section 63.1386 is amended by pulled; and weight of phenol, 94.11 g/g-mol; revising paragraphs (a)(2) through (4); (C) 0.50 lb of methanol per ton of molecular weight of methanol, 32.04 g/ revising paragraphs (d)(1)(ii) and (iii); glass pulled. g-mol; molecular weight of chromium adding paragraphs (d)(2)(x), (f) and (g). (ii) The owner or operator of each new compounds tested in g/g-mol; molecular The revisions and addition read as or reconstructed FA manufacturing line weight of HF, 20.0064 g/g-mol; follows: molecular weight of HCl, 36.4611 g/g- that produces heavy-density wool § 63.1386 Notification, recordkeeping, and fiberglass and/or pipe insulation must mol. reporting requirements. not discharge or cause to be discharged (d) Following the initial performance (a) * * * into the atmosphere in excess of: or compliance test to be conducted (2) Notification that a source is subject (A) 3.3 lb of formaldehyde per ton of within 90 days of [EFFECTIVE DATE to the standard, where the initial startup glass pulled; OF THE RULE] to demonstrate (B) 0.46 lb of phenol per ton of glass compliance with the chromium is before November 25, 2011. pulled; and compounds emissions limit specified in (3) Notification that a source is subject (C) 0.50 lb of methanol per ton of § 63.1382(a)(1)(i)(D) or (a)(1)(ii)(D), you to the standard, where the source is new glass pulled. must conduct an annual performance or has been reconstructed the initial (b) * * * test for chromium compounds startup is after November 25, 2011, and (6) The owner or operator must emissions from each glass-melting for which an application for approval of operate each control device used to furnace (no later than 12 calendar construction or reconstruction is not control formaldehyde, phenol, and months following the previous required; methanol emissions from forming or compliance test). (4) Notification of intention to curing such that any three-hour block (e) Following the initial performance construct a new affected source or average temperature in the firebox does or compliance test to demonstrate reconstruct an affected source; of the not fall below the average established compliance with the PM, HF, HCl, date construction or reconstruction during the performance test as specified formaldehyde, phenol, and methanol commenced; of the anticipated date of in § 63.1384. emissions limits specified in § 63.1382, startup; of the actual date of startup, where the initial startup of a new or * * * * * you must conduct a performance test to 18. Section 63.1383 is amended by demonstrate compliance with each of reconstructed source occurs after revising paragraph (f) to read as follows: the applicable PM, HF, HCl, November 25, 2011, and for which an formaldehyde, phenol, and methanol application for approval or construction § 63.1383 Monitoring requirements. emissions limits in § 63.1382 of this or reconstruction is required (See * * * * * subpart at least once every 5 years and § 63.9(b)(4) and (5) of this part); (f) The owner or operator who uses a as often as raw material inputs change * * * * * control device to control HAP emissions by more than 10 percent following the (d) * * * from a glass-melting furnace, RS previous test. (1) * * *

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00048 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72817

(ii) The owner or operator may retain air pollution control and monitoring malfunction. The owner or operator records electronically, on a computer or equipment, process equipment, or a seeking to assert an affirmative defense labeled computer disks, or on paper; process to operate in a normal or usual must also submit a written report to the and manner; and Administrator within 45 days of the * * * * * (B) Could not have been prevented initial occurrence of the exceedance of (iii) The owner or operator may report through careful planning, proper design the standards in this subpart. This required information on paper or on a or better operation and maintenance report must demonstrate that the owner/ labeled computer disk using commonly practices; and operator has met the requirements set available and EPA-compatible computer (C) Did not stem from any activity or forth in paragraph (g) of this section and software. Electronic notifications are event that could have been foreseen and must include all necessary supporting encouraged when possible. avoided, or planned for; and documentation. The owner or operator (D) Were not part of a recurring may seek an extension of this deadline * * * * * pattern indicative of inadequate design, (2) * * * for up to 30 additional days by operation, or maintenance. (x) You must report total chromium submitting a written request to the (ii) Repairs were made as and hexavalent chromium emissions Administrator before the expiration of expeditiously as possible when the from glass-melting furnaces using the 45 day period. Until a request for an applicable emissions limitations were Method 0061. extension has been approved by the being exceeded. Off-shift and overtime Administrator, the owner or operator is * * * * * labor were used, to the extent subject to the requirement to submit (f)(1) As of January 1, 2012 and within practicable to make these repairs; and 60 days after the date of completing such report within 45 days of the initial (iii) The frequency, amount and occurrence of the exceedance. each performance test, as defined in duration of the excess emissions § 63.2, and as required in this subpart, * * * * * (including any bypass) were minimized 22. Section 63.1387 is amended by you must submit performance test data, to the maximum extent practicable except opacity data, electronically to the revising paragraphs (a)(1) and (2) to read during periods of such emissions; and as follows: EPA’s Central Data Exchange by using (iv) If the excess emissions resulted the ERT (see http://www.epa.gov/ttn/ from a bypass of control equipment or § 63.1387 Compliance dates. chief/ert/erttool.html/) or other a process, then the bypass was (a) * * * compatible electronic spreadsheet. Only unavoidable to prevent loss of life, (1) Except as noted in paragraph (a)(2) data collected using test methods personal injury, or severe property of this section, the compliance date for compatible with ERT are subject to this damage; and an owner or operator of an existing requirement to be submitted (v) All possible steps were taken to plant or source subject to the provisions electronically into the EPA’s WebFIRE minimize the impact of the excess of this subpart is [DATE OF database. emissions on ambient air quality, the PUBLICATION OF THE FINAL RULE (2) All reports required by this environment and human health; and IN THE FEDERAL REGISTER]. subpart not subject to the requirements (vi) All emissions monitoring and (2) The compliance dates for existing in paragraph (f)(1) of this section must control systems were kept in operation plants and sources are: be sent to the Administrator at the if at all possible, consistent with safety (i) [DATE 1 YEAR AFTER appropriate address listed in § 63.13. If and good air pollution control practices; PUBLICATION OF THE FINAL RULE acceptable to both the Administrator and IN THE FEDERAL REGISTER] for glass- and the owner or operator of a source, (vii) All of the actions in response to melting furnaces, rotary spin these reports may be submitted on the excess emissions were documented manufacturing lines, or flame electronic media. The Administrator by properly signed, contemporaneous attenuation manufacturing lines subject retains the right to require submittal of operating logs; and to emission limits in § 63.1382(a) which reports subject to paragraph (f)(1) of this (viii) At all times, the affected source became effective [DATE OF section in paper format. was operated in a manner consistent PUBLICATION OF THE FINAL RULE (g) Affirmative Defense for with good practices for minimizing IN THE FEDERAL REGISTER]. Exceedance of Emission Limit During emissions; and (ii) [DATE OF PUBLICATION OF THE Malfunction. In response to an action to (ix) A written root cause analysis has FINAL RULE IN THE FEDERAL enforce the standards set forth in this been prepared, the purpose of which is REGISTER] for the provisions related to subpart, you may assert an affirmative to determine, correct, and eliminate the malfunctions and affirmative defense defense to a claim for civil penalties for primary causes of the malfunction and provisions of § 63.1386(g) and the exceedances of such standards that are the excess emissions resulting from the electronic reporting provisions of caused by malfunction, as defined at malfunction event at issue. The analysis § 63.1386(d) and (f). § 63.2. Appropriate penalties may be must also specify, using best monitoring * * * * * assessed, however, if you fail to meet methods and engineering judgment, the 23. Section 63.1388 is revised to read your burden of proving all of the amount of excess emissions that were as follows: requirements in the affirmative defense. the result of the malfunction. The affirmative defense must not be (2) Notification. The owner or § 63.1388 Startups and shutdowns. available for claims for injunctive relief. operator of the affected source (a) The provisions set forth in this (1) To establish the affirmative experiencing an exceedance of its subpart apply at all times. defense in any action to enforce such a emissions limit(s) during a malfunction, (b) The owner or operator must not limit, you must timely meet the must notify the Administrator by shut down items of equipment that are notification requirements in § 63.1386 of telephone or facsimile transmission as required or utilized for compliance with this subpart, and must prove by a soon as possible, but no later than two the provisions of this subpart during preponderance of evidence that: business days after the initial times when emissions are being routed (i) The excess emissions: occurrence of the malfunction, if he/she to such items of equipment, if the (A) Were caused by a sudden, wishes to be able to use an affirmative shutdown would contravene infrequent, and unavoidable failure of defense to civil penalties for that requirements of this subpart applicable

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00049 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 72818 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules

to such items of equipment. This contemporaneous startup or shutdown, (c) Table 1 to subpart NNN paragraph does not apply if the owner of the affected source or a portion summarizes the emissions limits during or operator must shut down the thereof. startups and shutdowns of glass-melting equipment to avoid damage due to a furnaces.

TABLE 1 TO SUBPART NNN—EMISSIONS LIMITS DURING STARTUPS AND SHUTDOWNS OF GLASS-MELTING FURNACES (LB/ HR)

New and Pollutant Existing reconstructed furnaces furnaces

PM ...... 0.25 0.033 Chromium Compounds ...... 0.0019 0.0019 HF ...... 0.036 0.014 HCl ...... 0.026 0.014

(d) Table 1 to subpart NNN startups and shutdowns of rotary spin summarizes the emissions limits during [RS] manufacturing lines.

TABLE 2 TO SUBPART NNN—EMISSIONS LIMITS DURING STARTUPS AND SHUTDOWNS OF ROTARY SPIN (RS) MANUFACTURING LINES (LB/HR)

New and Pollutant Existing RS reconstructed lines RS lines

Formaldehyde ...... 3.1 0.36 Phenol ...... 3.4 0.019 Methanol ...... 8.8 0.012

(e) Table 3 to subpart NNN startups and shutdowns of flame summarizes the emissions limits during attenuation (FA) manufacturing lines.

TABLE 3 TO SUBPART NNN—EMISSIONS LIMITS DURING STARTUPS AND SHUTDOWNS OF FLAME ATTENUATION (FA) MANUFACTURING LINES (LB/HR)

New and Pollutant Existing FA reconstructed lines FA lines

Formaldehyde ...... 100 60 Phenol ...... 25 8 Methanol ...... 9 9

24. Table 1 to Subpart NNN of Part 63 NNN of Part 63 and revised to read as is redesignated as Table 4 to Subpart follows:

TABLE 4 TO SUBPART NNN OF PART 63—GENERAL PROVISIONS APPLICABILITY TO SUBPART NNN

Applies to Reference subpart NNN Comment

63.1 ...... Yes. 63.2 ...... Yes. 63.3 ...... Yes. 63.4 ...... Yes. 63.5 ...... Yes. 63.6(a), (b), (c) ...... Yes. 63.6(d) ...... No ...... Section reserved. 63.6(e)(1)(i) ...... No ...... See 63.1382(b) for general duty requirement. 63.6(e)(1)(ii) ...... No. 63.6(e)(1)(iii) ...... Yes. 63.6(e)(2) ...... No ...... Section reserved. 63.6(e)(3) ...... No. 63.6(f)(1) ...... No. 63.6(g) ...... Yes. 63.6(h) ...... No ...... No opacity limits in rule. 63.6(i) ...... Yes. 63.6(j) ...... Yes.

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00050 Fmt 4701 Sfmt 4702 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2 Federal Register / Vol. 76, No. 227 / Friday, November 25, 2011 / Proposed Rules 72819

TABLE 4 TO SUBPART NNN OF PART 63—GENERAL PROVISIONS APPLICABILITY TO SUBPART NNN—Continued

Applies to Reference subpart NNN Comment

§ 63.7(a)–(d) ...... Yes. § 63.7(e)(1) ...... No ...... See 63.1382(b). § 63.7(e)(2)–(e)(4) ...... Yes. 63.7(f), (g), (h) ...... Yes. 63.8(a)–(b) ...... Yes. 63.8(c)(1)(i) ...... No ...... See 63.1382(b) for general duty requirement. 63.8(c)(1)(ii) ...... Yes. 63.8(c)(1)(iii) ...... No. 63.8(c)(2)–(d)(2) ...... Yes. 63.8(d)(3) ...... Yes, except for last sentence. 63.8(e)–(g) ...... Yes. 63.9(a), (b), (c), (e), (g), (h)(1) through (3), Yes. (h)(5) and (6), (i) and (j). 63.9(f) ...... No. 63.9(h)(4) ...... No ...... Reserved. 63.10 (a) ...... Yes. 63.10 (b)(1) ...... Yes. 63.10(b)(2)(i) ...... No. 63.10(b)(2)(ii) ...... No ...... See 63.1386 for recordkeeping of occurrence and duration of malfunctions and rec- ordkeeping of actions taken during malfunction. 63.10(b)(2)(iii) ...... Yes. 63.10(b)(2)(iv)–(b)(2)(v) ...... No. 63.10(b)(2)(vi)–(b)(2)(xiv) ...... Yes. 63.(10)(b)(3) ...... Yes. 63.10(c)(1)–(9) ...... Yes. 63.10(c)(10)–(11) ...... No ...... See 63.1386 for recordkeeping of malfunctions. 63.10(c)(12)–(c)(14) ...... Yes. 63.10(c)(15) ...... No. 63.10(d)(1)–(4) ...... Yes. 63.10(d)(5) ...... No ...... See 63.1386(c)(2) for reporting of malfunctions. 63.10(e)–((f) ...... Yes. 63.11 ...... No ...... Flares will not be used to comply with the emissions limits. 63.12 to 63.15 ...... Yes.

[FR Doc. 2011–29454 Filed 11–23–11; 8:45 am] BILLING CODE 6560–50–P

VerDate Mar<15>2010 15:25 Nov 23, 2011 Jkt 226001 PO 00000 Frm 00051 Fmt 4701 Sfmt 9990 E:\FR\FM\25NOP2.SGM 25NOP2 wreier-aviles on DSK7SPTVN1PROD with PROPOSALS2