                 SUMMARY OF PUBLIC COMMENTS AND RESPONSES ON 
                           SECONDARY ALUMINUM NESHAP
                               December 14, 1999



      National emission standards for hazardous air pollutants (NESHAP) were proposed for the secondary aluminum production source category on February 11, 1999 (64 FR 6946).  The public comment period ended on April 12, 1999.  Copies of the comment letters have been places in project docket A-92-61.  Dockets are on file at EPA Headquarters in Washington, DC, and at the Office of Air Quality Planning and Standards (OAQPS) in Durham, NC.  A list of commenters, their affiliations, and the EPA docket number assigned to their correspondence are given in Table 1.
      For the purpose of presentation, the comments have been categorized and summarized according to the following subject areas:
      1.  Applicability
      2.  Definitions
      3.  Emission Standards and Operating Requirements
      4.  Monitoring and Compliance Requirements
      5.  Impacts


 TABLE 1.  LIST OF COMMENTERS ON PROPOSED SECONDARY ALUMINUM PRODUCTION NESHAP


Docket Item Number

                           Commenter and affiliation

IV-D-1

Kevin J. Bald
Corporate Environmental Quality
Reynolds Metals Company
P.O. Box 27003 
Richmond, Virginia 23261-7003

IV-D-2

Kenneth D. Woodrow
Counsel for the Aluminum Association, Inc.
Baker & Hostetler, Counselors at Law
Washington Square, Suite 1100
1050 Connecticut Avenue, NW
Washington, DC 20036-5304

IV-D-3

Drek Newton
Naval Facilities Engineering Service Center
Code ESC426
1100 23[rd] Avenue
Port Huneme, CA 93043-4370

IV-D-4

Michael Palazzolo
Alcoa, Incorporated
201 Isabella Street
Pittsburgh, PA 15212-5858

IV-D-5                                                                            

Dana K. Mount
Director, Division of Environmental Engineering
North Dakota Department of Health
Environmental Health Section
P.O. Box 5520
Bismark, ND 58506-5520

IV-D-6

Mike Ling
Director, Dioxin Project
Cambridge Health Project
41 Lee Street, #12
Cambridge, MA 02139

IV-D-7

Jancey Pettit, President
United Group, Inc.
4008 NW 14[th] St.
Topeka, Kansas 66618

IV-D-8

Kevin J. Bald
Corporate Environmental Quality
Reynolds Metals Company
P.O. Box 27003
Richmond, Virginia 23261-7003

IV-D-9

Gail M. Graban
Manager, Environmental Affairs
Century Aluminum of West Virginia, Inc.
P.O. Box 98
Ravenswood, WV 26164

IV-D-10

Andrew Stein, CEO
TST, Inc.
11601 Etiwanda Ave.
Fontana, CA 92325

IV-D-11

Samuel H. Bruntz
Staff Environmental Engineer
Commonwealth Aluminum
1372 State Road 1957
P.O. Box 480
Lewisport, Kentucky 42351-0480

IV-D-12

Gary G. Walker and Daniel L. Twarog
North American Die Casting Association
9701 West Higgins Road, Suite 880
Rosemont, Illinois 60018-4721

IV-D-13

Gerry Smarte
Compliance Manager
Vista Metals Corp
13425 Whittram Avenue
Fontana, CA 92335-2999

IV-D-14

Carl Johnson
New York State Department of Environmental                       Conservation
Office of Air & Waste Management
Room 608
50 Wolf Road
Albany, NY 12233-1014

IV-D-15

David A. Gallogly
Werner Company
93 Werner Road
Greenville, PA 16125

IV-D-16

Donn W. Sanford
Aluminum Extruders Council
1000 N. Rand Road, Suite 214
Wauconda, IL 60084

IV-D-17

Michael A. Palazzolo
EHS Services North America for 
Alcoa Corporate Center
201 Isabella Street
Pittsburgh, PA 15212-5858

IV-D-18

Kimberly Bowden
Sr. Environmental Engineer
Delphi Automotive Systems
485 West Milwaukee Avenue
Detroit, Michigan 48202

IV-D-19

Robert D. Fletcher, Chief
California Environmental Protection Agency
Air Resources Board
2020 L Street
P.O. Box 2815
Sacramento, CA 95812

IV-D-20

Peter J. Maciejewski
Regulatory/Legislative Interface Group
General Motors Corporation
485 West Milwaukee Avenue
Detroit, MI 48202

IV-D-21

Debra J. Jezouit
Counsel to the American Foundrymen's Society
Baker and Botts
1299 Pennsylvania Avenue, NW
Washington, DC 20004-2400

IV-D-22

Steven R. DuBois
Alcan Aluminum Corporation 
Manager Environmental Affairs

IV-D-23

Robert P. Strieter
Director, Environmental Affairs
The Aluminum Association
900 19[th] Street, NW
Washington, DC 20006

IV-D-24

Kelle R. Vigeland
Environmental Engineer
Spokane County Air Pollution Control Authority
West 1101 College, Suite 403
Spokane, WA 99201

IV-D-25

Edward J. Campobenedetto
Institute of Clean Air Companies
1660 L Street NW, Suite 1100
Washington, DC 20036-5603

IV-D-26

Michael A. Coffman
Hayes Lemmerz
38481 Huron River Dr.
Romalus, MI 48174

IV-D-27

Tom Tyler
Associate Counsel
Institute of Scrap Recycling Industries, Inc.
1325 G Street NW, Suite 1000
Washington, DC 20005-3104

IV-D-28

Terry D. Snell,
The William L Bonnell Company, Inc.
25 Bonnell Street
P.O. Box 428
Newnan, Georgia 30263

IV-D-29

Judith Ann Van Houten
LAN Associates
66 Cuna Street
St. Augustine, FL 32084-3619
for VAW of America Inc.

IV-D-30

Kevin M. Butt
Toyota Motor Manufacturing 
North America, Inc.
25 Atlantic Avenue
Erlanger, KY 41018

IV-D-31

Wayne E. Wooster
Goldendale Aluminum Company
85 John Day Dam Road
Goldendale, WA 98620-4689

IV-D-32

J.W. Vinzant
Kaiser Aluminum
1201 Airline Highway 
Baton Rouge, LA 70805

IV-D-33

Steven K. Curreri
IMCO Recycling, Inc.
Central Tower at Williams Square
5215 North O'Conner Blvd., Suite 940
Irving, TX 75039

IV-D-34

Gerry Smarte
Vista Metals Corp
Aluminum Alloys
13425 Whittram Avenue
Fontana, CA 92335-2999

IV-D-35 

Joseph S. Viland
Wabash Alloys, L.L.C
4525 W Old 24
P.O. Box 466
Wabash, IN 46992-0466
	

1.  Applicability

      1.1  Comment:  Several commenters (IV-D-9, IV-D-17,IV-D-23, IV-D-31, IV-D-32) wanted to exempt unvented in-line flux boxes from testing and monitoring requirements and suggested regulating them via work practices based on the following statements:
      1)  emissions do not have the potential to exceed the emission limit because small amounts (< 0.2 lbs/ton) of chlorine gas flux are used;
  	2)  there is no acceptable method for sampling their fugitive emissions, so exclusion from testing and monitoring would improve the SAPU concept and substantially reduce costs;
      3)  they are a pollution prevention design that operate within allowable OSHA limits, and should be considered representative of the MACT floor when properly installed.
      Response:  Unvented in-line fluxers are capable of using and emitting chlorine and HCl in excess of the HCl emission standard for in-line fluxers, 0.04 lb/ton.  One manufacturer of unvented in-line fluxers specifies a flux rate of 0.92 pounds chlorine per ton aluminum (Docket item IV-J-1).  The Agency has no reason to believe that fluxing at 0.2 to 0.9 lb/ton in an "unvented" in-line fluxer will meet the MACT floor level of emissions.  Owner/operators can meet the emission limit by capturing and venting emissions to add-on controls or limiting the chlorine flux input to the fluxer.  Limiting chlorine flux input to levels below the emission limit and monitoring flux addition is a work practice that would avoid the need for testing to determine the emission rate.  If testing is necessary, testing costs may be reduced through like-for-like testing allowed in the final rule, i.e., with multiple uncontrolled flux boxes of same design and same operating practice, only one need be tested to demonstrate compliance.
      The commenter's claim that such units cannot be tested is not valid.  One unvented flux box at a facility that will be subject to this rule has been tested since proposal, and the results reported to the Agency (see Docket Item IV-D-40).  This particular unit was tested by measuring emissions at the point where fluxed metal exits the flux box.  Another method of testing is to construct a temporary enclosure around the fluxer for the short duration of performance tests to capture fugitive emissions for measurement purposes (see Docket item IV-A-1).  Following the performance tests, flux usage must be monitored and the flux box operating procedures must be maintained to ensure continuous compliance with the SAPU HCl standard.
      
      With regard to "unvented" fluxers being a pollution prevention design that should be considered a MACT floor, commenters have referred to perceived lower emissions that presumably are achieved by lower and more efficient use of fluxing agents.  The MACT floor technology for control of in-line flux boxes upon which the emission limit is based is a lime-injected fabric filter; this technology can achieve an emission limit of  0.04 lb HCl/ton aluminum.  No data were provided by the commenters to demonstrate equal or lower emissions from "unvented" fluxers over the full range of input flux as compared to vented fluxers with the floor technology.

      1.2  Comment: Two commenters (IV-D-31, IV-D-32) state that they support the SAPU concept.
      Response: The EPA acknowledges this support for the proposed SAPU concept.

      1.3  Comment:  Several commenters (IV-D-5, IV-D-7, IV-D-8, IV-D-9, IV-D-11, IV-D-16, IV-D-15, IV-D-17, IV-D-23, IV-D-32, IV-D-33) opposed regulation of area manufacturing sources of D/F emissions, such as extrusion, die casting, and foundry facilities.  Another commenter asserted that the EPA assumed area and major source D/F emitting processes emit at about equal rates per ton of feed, but data available to EPA for side-charge and roll top melters processing clean charge show those furnaces are not significant sources of D/F as compared to furnaces charging dirty scrap.  This commenter also contended the EPA assumption that 55 percent of all delacquering furnaces are located at area sources was the basis for regulating area sources.  In comments on the Integrated Urban Air Toxics strategy the commenter claimed there were inappropriate assumptions and errors in the inventories for 112(c)(6) and 112(k).
      Response:	 The EPA acknowledges the error regarding delacquering furnaces in the inventory for the Integrated Urban Air Toxics strategy. The EPA recognizes that emissions of D/F from affected sources in secondary aluminum production facilities are site-specific and depend on the type of materials (scrap) fed to the process, flux type, flux rate, and flux practices among other variables.  For both major and area sources the materials fed to the furnace and combustion processes contain varying amounts of oil (hydrocarbons) and coatings (hydrocarbons and chlorides). These compounds found in scrap containing oils and coatings, as well as some fluxes, are D/F precursors.  Processes, located at facilities that are area sources that use the same feed and flux materials as are used at major sources, will emit D/F at levels equal to the same processes at major source facilities.
      The EPA is not claiming that the total D/F emissions from affected sources located at facilities that are area sources are equal to the total D/F emissions from facilities that are major sources.  However, there were also other commenters who mentioned large numbers of sweat furnaces in their states whose emissions were not counted, suggesting there are additional D/F emissions beyond those estimated in the national impacts at proposal.  The EPA has developed an estimate of D/F emissions from sweat furnaces located at facilities that are area sources.  That estimate is now included in the national impact calculations.


      1.4  Comment: Numerous commenters (IV-D-5, IV-D-7, IV-D-8, IV-D-9, IV-D-11, IV-D-12, IV-D-14, IV-D-15, IV-D-16, IV-D-17, IV-D-18, IV-D-19, IV-D-20, IV-D-21, IV-D-23, IV-D-26, IV-D-30, IV-D-32, IV-D-33) representing aluminum extruders, aluminum die casters, and aluminum foundries stated that their facilities should not be regulated because they differ fundamentally from large secondary aluminum production facilities and in emission potential, particularly D/F emissions.  The commenters raised the following issues:
      1) Extruders encompass a broad spectrum of facilities and appear to fall within the broad definition of secondary aluminum production facilities, which range from relatively small facilities owned by large companies to facilities owned by independent business people, many of which the commenter claimed are small businesses.
      2) Some extruders, die casters, and foundries use no purchased scrap but do use internally generated scrap, while other facilities use small amounts of "clean" purchased scrap. Some are concerned that regulation may interfere with the effort to recycle at the plant, while others who purchase scrap see the regulation as creating a disincentive to recycle from outside the plant. 
      3)  Impurities in scrap are a principal source of D/F precursors.  The commenters stated that extruders, die casters, and foundries cannot be large contributors to D/F emissions because they use or process only small amounts of higher quality scrap and do limited fluxing. One commenter argued that EPA should exclude extruders as small contributors to D/F emissions as in Alabama Power vs. Costle, 636 F.2d 323 (D.C. Cir. 1980) using the de minimis exception articulated in that case. 
      4)  Previous EPA publications support the distinction between die casters and secondary aluminum production facilities:
      a) the Documentation for Developing the Initial Source Category List defines secondary aluminum production as facilities that smelt, and not including die casters;
      b) an EPA New Source Review guidance memo (Memo from Thomas Curran, December 1998) states "die casting facilities typically need not be considered secondary metal production plants" (the commenters argued that this memo acknowledges that die casters could engage in in-house recycling of castings and not be considered a secondary aluminum production facility);
      c)  the Secondary Brass and Bronze NSPS distinguishes between facilities that reclaim brass and bronze and those that create a finished product.  
      5) Applying the D/F standard to affected sources located at facilities that are area sources will subject facilities such as extruders, die casters, and foundries to the burden of Title V permitting and MACT monitoring and reporting.  One of these commenters stated that no environmental benefit will be gained from regulating area source aluminum production facilities since they already meet the emission limitation.
      
      Response: 1)  The EPA agrees that based on the definition of secondary aluminum production facility and current operations of some facilities that are extruders, those extruders would be subject to this rule.  Numerous comments on the applicability section and definitions in the proposed rule were received and after consideration of those comments, the EPA has revised those sections of the final rule.  As part of the revisions, the EPA has concluded that aluminum extruding, aluminum die casting, and aluminum foundry facilities that process no materials other than a) materials generated within the facility, and b) "clean charge" (defined in the rule), and that do not operate a thermal chip dryer, sweat furnace, or scrap dryer/delacquering kiln/decoating kiln are not secondary aluminum production facilities and are exempted from the requirements of the rule.  Based on comments and information received in response to the proposal and subsequent meetings with the sources, the Agency believes that most small businesses will not fall under the definition of secondary aluminum production facility.  Those aluminum extruding, die casting, and foundry facilities that do purchase or otherwise obtain materials other than "clean charge" and/or operate a thermal chip dryer, sweat furnace, or scrap dryer/delacquering kiln/decoating kiln are secondary aluminum production facilities and are subject to this rule.  
      The commenter's reference to some small facilities being owned by large companies is consistent with the EPA's knowledge that large companies in the secondary aluminum production industry engage in extruding operations.  The commenter also claimed that some extruders are owned by independent businesses, many of which are small, however no specific quantitative data were provided to assist the Agency in assessing potential impacts.  The Agency has no information indicating that large numbers of small business aluminum extruders will be affected by this rule.
      2)  With regard to the regulation discouraging recycling within the plant, the final rule does not prevent facilities that are area sources from using internally-generated scrap as charge to their group 1 furnaces.  Regarding purchased scrap, some extruders, die casters, and foundries use only small amounts of purchased scrap in their operations, while other information provided to the EPA since proposal indicates that some of these facilities use more than half scrap (purchased, returns, and internally generated) as feed/charge in their operations.  This range of purchased scrap is typical of other secondary aluminum processors from which data were obtained to develop the NESHAP. (See Docket Items IV-E-1 and IV-E-2). 
      
      The issue with purchased scrap is the level of contamination with D/F emission precursors.  The EPA worked with industry representatives during the regulatory development phase to establish definitions and specifications for purchased scrap that would yield lower HAP emissions.  Data collected indicated that the percentage of oil and coatings in scrap (hydrocarbon and chloride content) varies over a large range.  No concurrence was achieved on the levels of scrap oil and coatings content that would reliably limit the processing of D/F precursors from affected sources, nor was concurrence achieved on a way to measure these levels of oil and coatings.  Further, a similar discussion with industry representatives failed to reach a consensus on how to define limited reactive fluxing, the other important aspect of  D/F emission potential.  The EPA has concluded that facilities in which aluminum scrap is processed, whether purchased or otherwise acquired from outside the facility, fall within the secondary aluminum production source category. 
      3)  These commenters assert that these facilities are not large contributors to D/F emissions because they purchase only small amounts of scrap or "clean" scrap, thus limiting the availability of D/F precursors in the affected sources.  However three factors, i) the total quantity of scrap fed to processes,  ii) the percentage of oil and coatings contamination of the scrap, and iii) the flux rate, are significant variables that affect generation of D/F precursors.  As mentioned above, some facilites use significant amounts of purchased scrap.
       With regard to the comment citing Alabama Power vs. Costle, 636 F.2d 323 (D.C. Cir. 1980) and requesting de minimis exemption for extruders, the Clean Air Act does not provide for de minimis exemptions.
      4)  Documentation for the Source Category Listing states that the secondary aluminum production source category includes "any facility engaged in the cleaning, melting, refining, alloying, and pouring of aluminum recovered from scrap, foundry returns, and dross, to form aluminum products such as alloy ingots, billets, notched bars, shot, hot metals, and hardeners."  The documentation also states that the category includes pretreatment processes which include drying, burning, and sweating, among others.  Although there can be differences in operations and products between secondary aluminum production facilities and those facilities that are primarily die casting, foundry, and extrusion facilities, for the purposes of this NESHAP the Agency considers the die casting, foundry, and extrusion facilities that use aluminum scrap and other coated/painted aluminum bearing materials obtained from outside their facilities to be engaging in secondary aluminum production operations.
      
      The EPA New Source Review guidance memo (Memo from Thomas Curran, December 1998) (Docket item IV-B-3) has, in addition to the commenter's quote, an extensive discussion of the fact that some facilities whose primary activity is die casting also perform secondary metals production from post-consumer scrap or unspecified aluminum scrap.  This type of facility was identified in the memo as a "nested" secondary aluminum support facility.  Such facilities also use processing equipment that is defined as an affected  source under this rule.  It is the acquisition of aluminum-bearing materials from outside the facility that are not "clean charge," and the presence of affected sources that subject the facility to this rule.  The difference in products is not the determining factor.  
      The final rule clarifies that aluminum die casting, aluminum foundry, and aluminum extrusion facilities that process only clean charge (as distinct from scrap) are not secondary aluminum production facilities (regardless of the remelting of internally generated scrap) provided they do not operate thermal chip dryers, scrap dryer/delacquering kilns/decoating kilns, or sweat furnaces.  Aluminum die casting, extruding, and foundry facilities that process aluminum scrap, etc., (i.e., materials that are not clean charge) from outside the facility in the furnaces are secondary aluminum production facilities and subject to the final rule.
      5) Regarding the burden of Title V permitting, monitoring, and reporting for area sources, the final rule has been changed to provide the discretion to individual permitting authorities to defer Title V permitting requirements for area source facilities, if those facilities are not subject to title V permitting requirements under other regulatory actions.  At present the permitting authority may defer the Title V permit requirement for these area sources until December 9, 2004.  (In a separate action, the Agency proposed final amendments on August 18, 1999 to extend title V operating permit deferrals for area sources in five source categories (64 FR 45116).) 
      A further change that will reduce the burden for area sources is that they will only be required to conduct an initial performance test to demonstrate compliance.  The requirement to repeat the performance test every 5 years has been eliminated for area sources.
      Facilities that use add-on controls will be required to monitor parameters in accordance with their approved site-specific OM&M plan.  Facilities that are area sources which use purchased scrap, but meet the D/F emission limit without add-on controls, i.e. use work practices, will also be required to monitor in accordance with their site-specific OM&M plan.  Their monitoring provisions are likely to include a scrap inspection or a scrap certification program to demonstrate they are not exceeding the scrap quantity and oil and coatings contamination levels, and flux rate established during the initial performance test of their group 1 furnace(s).
      The environmental benefit of controlling D/F from these affected sources is reduction of emissions of an environmentally persistent HAP.  The benefits of monitoring for those sources who meet the limit without add-on controls is continuing evidence that the operating practices used during the compliance tests are maintained and emissions remain at a level below the limit. 


      1.5  Comment:	Commenters (IV-D-8, IV-D-9, IV-D-23, IV-D-31, IV-D-32) desired to allow new or reconstructed units into the SAPU and encourage EPA to do it with a discount applied to the new unit's allowed emissions.  The commenters stated that:
      1)  It will allow sources to take advantage of the more efficient fluxing achievable in new flux boxes, in particular the "unvented" flux boxes.
      2)  It promotes pollution prevention and is consistent with common sense initiatives and project XL innovations, that allow plant-wide applicability limits.
      3)  It is not standard avoidance, but a more effective way of complying.
      Response: To allow new or reconstructed units into a SAPU consisting of existing units would involve combining control requirements for new and existing sources for which there is no legal construct under the Clean Air Act.  Therefore, EPA will not allow new units to become part of a SAPU comprised of existing units.  In order for new units to have the same benefits available to existing emission units, the Agency has revised the rule to allow for a SAPU that is composed entirely of simultaneously constructed new sources and/or simultaneously reconstructed sources, in addition to SAPU for the existing emission units.

      1.6  Comment:	In comments on combining and treating emissions from existing sources with those from new sources in a single control system:
      1) One commenter (IV-D-8) asked to group an existing or new furnace with a new in-line fluxer as a separate affected source (outside the SAPU).  The combined unit would have the same limits as the furnace by itself for PM, HCl, and D/F (i.e. no emission increment for the new in-line fluxer).  The industry claimed a significant improvement in fluxing efficiency with much lower emissions are associated with moving fluxing from the furnace to in-line fluxers.
      2) Another commenter (IV-D-23) requested that the rule be expanded to affirm that new emission units may be ducted to existing control systems, if capacity is available or can be expanded to accommodate the new source.
      Response:  The problem with combining a new affected source with an existing affected source is that the new source is required to meet the specified emission limits, but once combined, the new source emissions are not measurable separately from the emissions from the existing source.  As noted in the response to the previous comment, there is no legal construct under the Clean Air Act that permits combining control requirements for existing and new sources, therefore the combination of an existing furnace and new in-line fluxer is not permitted.
      The revisions to the final rule do provide for the establishment of a SAPU composed entirely of simultaneously constructed new emission units.  This will allow the combination of a new furnace and new in-line fluxer as a SAPU, but not allow combining a new furnace or in-line fluxer with an existing SAPU.
      
      1.7  Comment:  One commenter (IV-D-7), a manufacturer of sweat furnaces,  expressed concern about economic impacts on small aluminum reclamation operators.  This commenter estimated that there are at least several hundred sweat furnaces manufactured by them being currently used nationally with capacities considerably less than the model sweat furnace used in EPA's analysis of impacts (5,000 tons/year).  All of their furnaces are equipped with integral afterburners.  This commenter also submitted an afterburner performance test report showing 97.8 percent removal of PM by the afterburner and claimed, but did not have measurements, that D/F removal should be similar. The commenter stated that the preamble did not show D/F results upstream of the afterburners or what destruction efficiency was achieved.
      Another commenter (IV-D-5) attached a brochure from a manufacturer claiming to have distributed over 2000 small sweat furnaces.  This commenter states that the proposal underestimated the number of these sources.  The commenter believes that testing and control costs will eliminate small businesses from the market, and suggested that regulations for area sources be withdrawn until small business, health and environmental impacts have been reassessed.  One commenter (IV-D-14) suggested a technology-based standard for area source sweat furnaces with no testing required. 
      Response: The EPA has no test data to support a comparison between PM and D/F removal efficiencies.  The D/F emission limit in the proposed and final rules has been proven to be achievable with MACT floor technology.
      Based on the information contained in these comments, the EPA requested additional information and data from sweat furnace manufacturers to further assess impacts of regulating D/F emissions from the furnaces.  The large number of units reported to be manufactured suggested large numbers of these affected sources are currently in operation.  The EPA's further investigation found that although one manufacturer who commented only sells sweat furnaces with integral afterburners for emission control, that is not the case for all domestic manufacturers.
      
      Due to the large number of these sources and the types of scrap materials processed, their D/F emission potential is significant both individually and in the aggregate.  Recognizing this, the EPA considered additional regulatory strategies for sweat furnaces and performed an economic analysis to examine the impacts of those strategies.  The conclusion from this analysis is that the cost of measuring D/F emissions from sweat furnaces through a performance test is significant in comparison to the cost of the furnace and afterburner.  Based on this analysis the EPA has revised the rule to add an alternative means of compliance.  Owner/operators electing to install and operate an afterburner meeting the design criteria of operating temperature of at least 1600 F and a 2 second residence time will not have to conduct performance tests.  The final rule retains the numerical standard so that owner/operators with control equipment that does not meet the design criteria have the option to test to show that the D/F emissions are below the limit.  These revisions to the proposed rule, combined with many anticipated State permitting authority decisions to exercise their discretion to defer the requirement for title V permits, will significantly reduce the burden for both large and small businesses operating sweat furnaces.  The economic impact analysis conducted for this regulation reports minimal economic impacts to owners and operators of sweat furnaces.

      1.8  Comment: One commenter (IV-D-32) states that the applicability language for area sources "and is an area source of D/F" could be misinterpreted, and should be replaced by "and is an area source operating certain emission units or affected sources with D/F limits in this rule."
      Response:  The preamble makes clear that area source status is determined on a facility-wide basis.  Facilities subject to the rule which are area sources, are subject to regulation of only those affected sources for which a D/F limit is established, without regard to the actual or potential mass of D/F emitted.

      1.9  Comment:  One commenter (IV-D-3) states that the rule is confusing as to its coverage of group 2 furnaces and recommends that the rule be revised to clarify that group 2 furnaces are affected, except those in SIC 336 (facilities that use only clean aluminum and aluminum foundries).
      Response:  The applicability section of the rule has been clarified to indicate that aluminum extruding, aluminum die casting, and aluminum foundry facilities that process no materials other than materials generated within the facility and clean charge and that do not operate a thermal chip dryer, scrap dryer/ delacquering kiln/decoating kiln, or sweat furnace are not subject to the rule.  These facilities, without regard to SIC codes, are exempt even if they use reactive fluxes and/or process internally generated scrap containing oil and/or coatings.  Group 2 furnaces located at facilities that are area sources are not affected sources under this rule.

      1.10  Comment: One commenter (IV-D-14) states that they would like applicability clarified and that it is much more straightforward to simply state up-front that the rule applies to all chip dryers, scrap dryers/delacquering/decoating kilns, sweat furnaces etc., regardless of the pollutants that are emitted from the source.  The commenter states that at present there is a question of whether a test showing zero D/F emissions would exempt a facility. 
      
      Response:  The rule has been clarified to indicate that all operators of sweat furnaces and all operators of thermal chip dryers and scrap dryers/delacquering kilns/decoating kilns at secondary aluminum production facilities, are subject to the rule.  If these affected sources are located at a facility that is an area source, they are subject to those provisions pertaining to D/F testing to show compliance with the D/F emission limit (except sweat furnaces complying with the equipment standard), operating, monitoring, reporting and recordkeeping.  A test showing zero D/F emissions would not exempt an affected source because parameters would have to be established and monitored to ensure continuous compliance with the standard.
 
      1.11  Comment:  Two commenters (IV-D-8, IV-D-17) state that research and development facilities are not exempted from the rule but should be; they are to be regulated under a separate NESHAP. 
      Response:  The final rule has been clarified to exempt R & D facilities that do not produce salable products.

      1.12  Comment: Several commenters asserted that EPA did not collect enough information or data from die casting and foundry facilities in the regulatory development process to justify establishment of standards that apply to their facilities.
      1) One commenter stated that EPA did not follow proper procedures during development of the proposed regulation because  neither the commenter nor any of its member companies actively participated in the development.  The commenter claimed that EPA could not have enough information about die casting to justify inclusion under this regulation because only two die casters were included in the ICR process.  The commenter contended that none of EPA's determinations are valid including MACT floor emission levels, quantification of emissions, and economic impacts.  Further, the preamble request for comments from the die casting industry was inappropriate because the preamble stated die casters are excluded.  The commenter does not possess information, estimates, data, and detailed analysis on HAPs and is unable to obtain such information within the time constraints [of the public comment period]. (IV-D-12)    
      
      2) Another commenter stated that EPA included aluminum foundries in the proposal with little if any evidence that aluminum foundries generate sufficient quantities of HAPs to require regulation.  They argued that insufficient foundry data were collected for proper consideration.  They acknowledged that a few foundries may have received the Information Collection Request (ICR) and two site visits were made to foundries, but these were incidental to the overall information gathering process.  The EPA improperly assumed foundry processes are similar to other facilities being regulated without regard to differences in output such as aluminum products from foundries, and the fact that smelters reclaim aluminum from aluminum scrap.  They claimed that without exclusion of foundries from the final rule, the action would surely be held to be arbitrary and capricious as is evidenced by a Supreme Court case, Motor Vehicles Mfrs. Assn v. State Farm Mut. [463 U.S.29,43(1983)]; regulating them without proper factual research and documentation is an arbitrary and capricious action. (AFS) 
      3) Other commenters stated that D/F emissions are the issue with potential to draw die casters and foundries into the NESHAP applicability.  They stated that EPA does not have any emissions data from die casters and foundries that show they emit D/F.  Since EPA did not test any source at die casters, the area source finding cannot be imposed on die casting operations.  The conclusion that area and major sources...emit D/F at about equal rates per ton of feed is not supported by data.  The processes and materials fed to die casting operations are fundamentally different.  There should be a way for specific facilities to show they are not sources of D/F; the fact that clean charge group 1 furnaces have no D/F limit implies this. (IV-D-18, IV-D-20)
      4)    Two commenters state that EPA's presumption that because D/F emissions from a chip dryer (or group 1 furnace) were found in a non-die casting facility that means chip dryers at die casters will also be sources of D/F is not valid. (IV-D-12, IV-D-26) 
      5)  One commenter (referring to facilities that are area sources) understood from previous discussions with EPA that the rule would only apply to those units having D/F emissions, however as proposed, the rule would require record keeping, labeling, and other requirements to continually demonstrate group 2 furnace status if the facility has a group 1 furnace. (IV-D-12)
      Response: 1) As the commenter acknowledged, at least two die casting facilities were included in the initial ICR wherein the EPA collected information and data regarding facilities potentially subject to the secondary aluminum MACT standards.  In addition, during the early phase of the regulatory development activities there were discussions between the EPA and North American Die Caster's Association (NADCA) staff regarding the nature of operations in die casting facilities.  Die casting facility operations were described as involving the use of "clean" feed materials and low levels of fluxing.  Based on this description, the EPA was not certain that such facilities should be regulated under this rulemaking.  Further, at the time of these discussions, the area source categories of section 112(c)(6) pollutants had not been identified.
      As the regulatory process progressed, the 112(c)(6) area source listing was published and included secondary aluminum smelting for D/F emissions, and more data were collected on secondary aluminum operations that generate D/F.  Also, meetings and discussions continued with die casting and foundry representatives and these revealed that there are major sources as well as area sources of HAP emissions within these sectors.  These sources perform operations identical to those of other secondary aluminum processors.
      
      Although the end products of such die casting (also foundry and extruding) facilities may be different than those of the facilities that are primarily SAPF, the affected source/emission unit operations at some of the die caster, foundry, and extruding facilities do not differ significantly from those at major source SAPF.  Those that process purchased scrap contaminated with oil, paint, and other coatings have similar potential to emit D/F because the precursors are present in the feed, and process operating conditions are similar. 
      On the basis of the information submitted to EPA during the public comment period and changes made to the proposed rule that narrow the applicability to facilities that are area sources, primarily aluminum extruders, die casters, and foundries, the EPA believes the number of those facilities subject to the rule to be small.
      2) This commenter also acknowledged that their industry segment was included in the ICR and that two plant visits were made to facilities that are aluminum foundries.  Similar to the rationale above, those aluminum foundries that purchase scrap and process it in operations similar to those found at major source SAPF have the potential to emit D/F due to the presence of D/F precursors (oil, paint, and coatings) on the charge material and similar process operating conditions.  
      In regard to the question of whether these facilities emit sufficient quantities of D/F to be regulated, under section 112(c)(6) EPA is required to subject not less than 90 percent of the aggregate emissions of D/F to standards (not necessarily control) under section 112(d)(2) or (d)(4).  Those facilities using similar processes and feed materials at aluminum foundries (also die casters and extruders) are expected to have similar emissions to SAPF and therefore should be subject to such standards.
      3)  The EPA's statement that secondary aluminum processes located at facilities that are area and major sources emit D/F at equal levels on a lb/ton basis is a reasonable conclusion for those that process like amounts of purchased oily, painted, or coated scrap and/or conduct reactive fluxing in melting furnaces; or process these materials in thermal chip dryers, sweat furnaces, or scrap dryer/delacquering kilns/decoating kilns.
      4)  Chip dryers process oil-laden (and sometimes painted) chips from machining operations and it is reasonable to assume that they have D/F emissions based on test data reported in the proposal.
      5)  The applicability section in the final rule excludes group 2 furnaces from the affected source list at facilities that are area sources.  An owner/operator of a group 2 furnace at a facility that is an area source is not required to do any monitoring, recordkeeping, or reporting related to that group 2 furnace.   


 	1.13  Comment:  One commenter representing an extruder stated that it is unfair to automatically include all area sources irrespective of size, purchased vs. in-house scrap, types of lubricants, and types of fluxes.  Three commenters want a clarification that only D/F producing units regulated for D/F within an area source are subject to the rule and no other HAPs from that source must be controlled.  One commenter requests that area sources be exempted or the comment period be reopened. (IV-D-15, IV-D-16, IV-D-29)
      Response: Facilities that are area sources are subject only to those provisions relating to D/F emission standards for the affected sources, and the testing, monitoring, recordkeeping, and reporting requirements for the same.  The rule does not apply to die casters, foundries, and extruders, that purchase or otherwise obtain only clean charge (and that do not purchase scrap), or that use only material generated from within the facility, and do not operate a thermal chip dryer, sweat furnace, or scrap dryer/delacquering kiln/decoating kiln.

2.  Definitions

2.1  Comment:  Numerous comments were received on the definition of "clean charge."
      1)  One commenter (IV-D-23) stated that the definition should include as clean charge, outside runaround that is contractually ensured to be clean. 
      2)  Other commenters (IV-D-8, IV-D-31, IV-D-32) stated that they support inclusion of "noncoated runaround" scrap in the definition, which may have small amounts of lubricant, and that some runaround is returned from customers.  These commenters stated that the rule should allow external, preconsumer, non-coated runaround scrap in group 2 furnaces. 
       3)  Several commenters (IV-D-15, IV-D-16, IV-D-30) requested that EPA define noncoated runaround scrap or redefine clean charge to allow scrap covered with lubricants or substances low in materials that could generate D/F.  Many die casters use scrap generated on-site (miscast material, defective parts, and cutoffs of excess aluminum) that may have inorganic agents (clay or talc) or die release agents (heavy waxes or high molecular weight oils) that do not generate D/F when burned.
      4)  Commenters (IV-D-16, IV-D-29) representing extruders also wanted to revise the definition to include purchased scrap low in materials that contribute to D/F generation.
      5)  Other commenters (IV-D-12, IV-D-18, IV-D-20) noted the proposed definition of clean charge allows only pure aluminum (pure aluminum is an incorrect term) that cannot be cast in a die casting machine.  They stated that the definition of clean charge also restricts the use of chips that have not been processed in a chip dryer.  This is a disincentive for exploration of new technology (presses, centrifuges, and washers) alternatives to chip drying.  Some facilities that do dry chips do not heat to 343C because it may oxidize the metal. The temperature to which chips must be heated to qualify as clean charge is arbitrary and was not considered with any input from foundries and die casters.
      
      6)  Two commenters (IV-D-23, IV-D-31) recommend that the definition include billets and pig.
      7)  One commenter (IV-D-8) suggests including alloyed aluminum in the definition, along with pure aluminum.
      Response:  As a result of the many comments received, EPA has reviewed and reconsidered the definition of clean charge.  
      1, 2, 3, and 4)  The definition of clean charge at proposal erroneously included non-coated runaround scrap which commenters wanted clarified to include runaround from outside the facility (i.e., external, relatively "clean," preconsumer, non-coated runaround).  The commenters acknowledged that the runaround may have "small" amounts of lubricant and coatings.  Lubricants, oils, and coatings are D/F precursors.  As explained in a previous response, the EPA worked with industry representatives during the regulatory development phase to establish definitions and specifications for purchased scrap that would yield consistently lower HAP emissions when charged to furnaces.  Data collected indicated that the percentage of oil and coatings in scrap (hydrocarbon and chloride content) varies over a large range.  No concurrence was achieved on the levels of scrap oil and coatings content, or a universal method of measuring the scrap content of oils/coatings, that would reliably limit the processing of D/F precursors from affected sources.  Group 2 and those group 1 furnaces that are "clean charge" furnaces have no D/F emission limit.  It is not consistent with the concept of clean charge furnaces to allow oil- and lubricant-bearing scrap purchased or otherwise obtained from outside the facility to be charged as clean charge.  For this reason, the Agency has clarified that the definition of clean charge includes internally generated runaround.  Internal runaround is defined in the final rule as scrap material generated on-site by aluminum extruding, rolling, scalping, forging, forming/stamping, cutting, and trimming operations that do not contain paint or solid coatings.  Aluminum chips generated by turning, boring, milling, and similar machining operations that have not been dried at 343 C (650 F) or higher, or by an equivalent non-thermal drying process, are not considered internal runaround.  Clean charge also does not include "runaround" scrap that is purchased or otherwise obtained from outside the facility.
      
      Secondary aluminum production facilities may use painted and/or purchased runaround in group 2 furnaces by drying or delacquering it to meet the definition of clean charge, so as to eliminate the possibility of D/F formation in the furnace.  Owner/operators may also charge painted and/or purchased runaround scrap to uncontrolled group 1 furnaces in a SAPU, provided they achieve an initial compliance demonstration and operate according to an OM&M plan approved by the permitting authority.  For group 1 furnaces operated without add-on controls, the plan would likely include a site-specific scrap inspection or certification program of some type to indicate the contamination level and to define the percentage of scrap in the total furnace charge.
      As noted in a response to a previous comment, facilities that are primarily aluminum die casters, foundries, and extruders that process only on-site materials or clean charge, and that do not operate a thermal chip dryer, scrap dryer/delacquering kiln/decoating kiln, or sweat furnace are not secondary aluminum production facilities and are not subject to this rule.
      5)  With regard to the use of the term "pure aluminum" in the definition in the proposed rule, the definition has been revised in the final rule to eliminate the word "pure" as a modifier of aluminum and instead describe it as oil- and lubricant-free uncoated/unpainted aluminum.
      With respect to the issue of chip drying and the potential for oxidation of the aluminum, the final rule does not contain a minimum temperature requirement for thermal chip drying to make the chips "clean charge."  With regard to other chip processing methods, the Agency is not precluding new technologies such as presses, centrifuges, and washers that may be capable of producing chips with no oily residue thus qualifying those processed chips as clean charge.
      6)  The EPA agrees that billets and pig are clean aluminum charge.  The definition of "clean charge" has been revised to include billets and pig.
      7)  As noted above the definition of clean charge has been revised to eliminate the word "pure" from the phrase "pure aluminum".  Because all aluminum is alloyed to some extent, EPA declined to add alloyed aluminum to the definition.

      2.2  Comment: For the definition of "afterburner," one commenter (IV-D-8) prefers the term "oxidizer" to "incinerator."  They state that "oxidizer" typically refers to a device used for the destruction of VOC whereas "incinerator" refers to a device used for the destruction of hazardous and infectious waste.
      Response:  EPA uses "incinerator" and "afterburner" to refer to devices that use high temperatures to oxidize organic vapors without distinction as to whether the organic vapor is a HAP or not.  In the proposed rule, both "afterburner" and "incinerator" were used.  EPA clarified the final rule by using "afterburner" throughout and revised the definition of "afterburner" to state that afterburners are also known as incinerators or thermal oxidizers.

      2.3 Comment: One commenter stated that the definition of "chip dryer" is a problem. The commenter questioned if coolants are included by virtue of being oil/water mixtures.  Chips may be coated by other chemicals that are not in the definition.  Die casters use dryers to remove water, machining fluids, and coolants from chips. (IV-D-12)
      
      Response: The definition in the proposed rule stated the chip dryer is a device that uses heat to remove water, oil, or oil/water mixtures from unpainted/uncoated aluminum chips.  Coolants and machining fluids that contain oil or oil/water mixtures are included in this definition.  The intent of this part of the definition is to include substances that are D/F emission precursors.  The words "unpainted/uncoated" preclude chips with paint, lacquer, and other coatings from this definition.  The latter painted/coated chips are covered in the definition of "scrap dryer/delacquering /decoating kiln".  Because there are other methods of drying chips, for example by centrifuging chips, the term "chip dryer" is modified in the final rule to "thermal chip dryer" to indicate a chip dryer that uses heat to clean aluminum chips.

      2.4 Comment:  Two commenters (IV-D-14, IV-D-24) requested that "D/F" (dioxin/furan) be defined.
      Response: The definition of D/F has been added to the promulgated regulation.  Dioxins and furans (D/F) means tetra-, penta-, hexa-, hepta-, and octa-chlorinated dibenzo dioxins and furans.  The emission limits for affected sources and emission units are expressed in terms of D/F TEQ.  TEQ was defined at proposal and remains in the promulgated rules.

      2.5 Comment:  One commenter (IV-D-23) states that the definition of "dross-only furnace" should be revised to include adding salts to feedstock materials because salt is common and was included in testing parameters to establish emission limit.
      Response: The EPA agrees that salt is commonly added along with dross and has revised the definition of "feed/charge" to clarify that for dross-only furnaces, the feed/charge includes  the addition of solid flux.

      2.6  Comment: Several comments were received on the definition of "feed/charge weight."
      1)Two commenters (IV-D-8, IV-D-23) recommend deleting solid fluxes from the definition because they are not part of charge materials per se; solid fluxes were not included as charge material in MACT emission testing. 
      2)  One commenter (IV-D-31) recommends monitoring on the basis on metal cast rather than charged.
      3) One commenter (IV-D-24) recommends that EPA either define "feed", or else use the term "feed/charge" throughout.    
      Response: 1)  The definition of feed/charge as proposed included solid flux as part of feed/charge.  However, the EPA intended the measurement of feed/charge to provide a uniform basis of aluminum-bearing materials processed in the affected sources/emission units from which to calculate allowable emissions and determine compliance.  Solid flux usage is a relatively small weight percentage in comparison to the aluminum-bearing feed materials processed.  Solid flux has been deleted from the definition of feed/charge in the final rule (except in cases relating to dross-only furnaces).  
      
      2)  The rule has been revised to allow the option of compliance demonstrations on a production weight basis as opposed to feed/charge basis.  If an owner/operator elects to comply on a production weight basis, monitoring must also be performed on a production weight basis.  It is essential to have a consistent basis for establishing the emission limit for the SAPU and monitoring compliance.
      3)  The commenter correctly points out that "feed" was sometimes used in the proposed rule when what was intended was "feed/charge" weight.  The rule has been revised to be consistent in the use of the term feed/charge.

      2.7 Comment: Two commenters addressed the topic of cover fluxes.
      1) One commenter stated that EPA uses the term cover flux in the definition  of "fluxing." It is contrary to the requirement for sidewell furnaces to do all their fluxing in the sidewell. (IV-D-18)
      2) One commenter (IV-D-8) wants to use cover fluxes in group 2 furnaces.
      Response: 1)  A definition of cover flux has been added to the final rule and the term "cover flux" has been removed from the definition of reactive fluxing.  The operating requirements section of the final rule has also been modified so as to allow the use of cover flux in the hearth of a sidewell furnace where the hearth emissions are not vented to a control device.
      2) The definition of cover flux states that it is used in both group 1 and group 2 furnaces. 
  
      2.8 	Comment:  One commenter stated that the definitions of group 1 and group 2 furnaces should be revised to clarify that certain heat producing units holding or applying heat to aluminum are not intended to be included in these definitions.  The language melts, holds, or processes aluminum scrap, or processes clean charge could be interpreted to apply to ingot preheaters (used prior to placement of ingots in a furnace) or homogenizing ovens which apply further heat after material has been melted, cooled, and cast.  It does not appear that EPA intended to regulate such units and the issue could be resolved by modifying the definitions.  The commenter suggested for group 1, "melts aluminum scrap or holds or processes such scrap in molten form" as the alternate language.  For group 2 furnaces, "melts only clean charge or holds or processes only such clean charge in molten form" as the alternate language. (IV-D-28)
      Response: The commenter is correct, the EPA did not intend to regulate ingot preheaters or homogenizing ovens.  Neither of these sources are identified as affected sources under this rule. 


      2.9  Comment: One commenter stated that group 1 furnaces should be individually discussed in the preamble.  The commenter asserted that the definition of group 1 furnaces is overly broad and should have some de minimis level for reactive fluxing and scrap contamination below which the furnace is exempt from the group 1 definition.  Instead of being totally exempt from regulation, EPA could consider such furnaces part of group 2 and therefore not subject to the extensive reporting requirements of group 1 furnaces.  The commenter argued that EPA should consider a third category of furnace allowing a minimal level of reactive fluxing and semi-clean charge, these furnaces would not have to apply controls with low actual benefit to the environment. (IV-D-18)
      Response:  In developing this regulation the EPA thoroughly reviewed the various combinations of melting, holding, charging, and fluxing practices.  This review included the wide variations in contamination level in feed/charge materials and types and chemical characteristics of fluxes.  Extensive discussions were held with representatives of the industry, and their insights into the effects of these variables were solicited.  The EPA considered establishing more than two groups of furnaces, but concluded there was no basis nor data to support, or upon which to establish, further furnace categories or floors.  No additional data that would change this situation have been provided to the Agency.  The final regulation retains only two categories of furnaces, group 1 and group 2. 

      2.10 Comment: One commenter (IV-D-11) supports the definition of "group 2 furnace."
      Response: The EPA acknowledges this supporting comment.  
      
      2.11 Comment: Two commenters requested that the definition of group 2 furnaces be revised to allow reactive fluxing up to 1% by weight of the metal produced and charges with less than 5% by weight of metal produced of painted and/or coated scrap, otherwise costly controls would be required for die caster furnaces.  (IV-D-18, IV-D-20) 
      Response:  The response to comment 2.1 above clearly presents the rationale for not allowing painted and/or coated scrap as feed/charge material and reactive fluxing for group 2 furnaces.  These practices generate D/F emission precursors.  
If the aluminum die casting facilities owned or operated by the commenters are not major sources, they may not be subject to the final rule depending on the source of their feed/charge materials, fluxing practice, and other processes operated at the site (See comment response 1.4 and the final rule). 

      2.12	Comment:	Referring to in-line fluxers two commenters (IV-D-8, IV-D-23) wanted the EPA to remove the word "typically" from the sentence referring to located in a transfer line from a furnace, to avoid including pretreatment units not tested in development of the MACT floor technology emission limits.
      Response:  The EPA agrees with the commenter.  The definition has been revised to remove the word "typically" (to exclude crucible fluxing with aluminum fluoride).
      
      2.13 Comment:  One commenter (IV-D-13) stated that the definition of in-line fluxers describes different units than those used by the commenter.  The commenter's units are large semi-stationary structures using small amounts of reactive flux. South Coast Air Quality Management District (SCAQMD) requires no controls.  Due to variations in these units and potential to emit, EPA needs to differentiate between different types of in-line fluxers.
      Response: As noted above, if the commenter is conducting crucible fluxing, this operation is not considered to be in-line fluxing.

      2.14  Comment: One commenter (IV-D-17) states that the definition of "lime injection" should be clarified to provide consistency throughout the rule; the definition defines it as "continuous" but nowhere in the rule is lime injection required to be continuous.  
      Response:  The MACT floor technology upon which the emission standards are based for several affected sources and emission units includes continuous lime injection.  Owner/operators are free to meet the emission limits anyway they can, including use of intermittent lime addition systems.  However, in the final rule, lime addition requirements have been specified only for continuous lime injection systems.  Lime addition requirements for intermittent lime addition have been eliminated because it is not the MACT floor control technology.  Provisions for obtaining approval for intermittent lime addition and establishing operating requirements have been added to the rule.

      2.15  Comment: Several commenters (IV-D-8, IV-D-23, IV-D-31) recommend adding the word "routinely" to transfer molten aluminum to or from another furnace, which would allow molten aluminum transfers to and from melter/holders.  This is necessary during alloy changes and during startup, shutdown, and malfunction  events.
      Response: The EPA established the melter/holder type of group 1 furnace for furnaces that do both melting and holding in the same furnace, exclusively.  An owner/operator who finds it necessary to infrequently use a melter/holder for transferring molten metal to or from another furnace as a result of a malfunction, startup, or shutdown event can cover that situation by including it in the facility's Startup, Shutdown, and Malfunction plan which is required to be incorporated by reference into the facility's Title V permit.

      2.16  Comment: One commeter (IV-D-24) states that "Melter/holder" is not defined and asked if this is the same as "melting/holding furnace."
      Response:  The term melter/holder, where it was used in the proposal package was intended to mean melting/holding furnace.


      2.17  Comment: One commenter (IV-D-23) wants to add that "produce significant gaseous" HAP emissions to the definition of reactive fluxing.  Another commenter (IV-D-8) states that the present definition discourages the use of low emitting cover fluxes.  Another commenter states that the definition of reactive flux that includes the words "results in a HAP emission" is overly broad; it should be defined to allow a de minimis level.  Any measurable emission of HAP would cause a flux to be termed reactive whether the HAP came from the flux or not. (IV-D-20) 
      Response:  With regard to HAP emissions from furnaces and in-line fluxers, the EPA's analysis of the industry identified the existence of a MACT floor technology that reduces HAP emissions from fluxing operations.  The Clean Air Act does not provide for de minimis exemptions when floor controls exist.  The EPA has clarified the definition of reactive flux to exclude cover flux.  The final rule includes a separate definition of cover flux and provides that cover fluxes can be used in group 1 and 2 furnaces for prevention of oxidation.  Use of cover flux is not considered to be reactive fluxing.

      2.18 Comment: Several commenters (IV-D-8, IV-D-9, IV-D-22, IV-D-23, IV-D-28, IV-D-32, IV-D-33, IV-D-35) propose a long list of chloride/fluoride salts (lithium, sodium, potassium, calcium, and zinc chlorides, and lithium, sodium, potassium, magnesium, calcium, and aluminum fluorides) that would be added to non-reactive list (presumably in group 2 furnace definition) based on IV-D-8 thermodynamic calculations and lab data submitted.  Another commenter (IV-D-11) disagrees that any flux salt be defined as reactive if any level of HAP emissions is detected; sodium and potassium fluoride salts should be considered nonreactive.  Results of emissions tests using fluoride salt flux  are summarized by the commenter.
      Response: Reactive fluxing was defined in the proposed rule as the use of any gas, liquid, or solid flux that results in a HAP emission.  Only argon and nitrogen were specifically defined as not reactive and not producing any HAP.  The question at issue is whether chloride and fluoride salt compounds in flux yield HAP emissions (e.g. HCl) when used in emission unit processes.  Some commenters submitted theoretical calculations based on reactions of pure salts, and the results of bench-scale tests to support their assertion that the compounds, when used as fluxes would not yield HAP emissions and, therefore, should not be identified as reactive fluxes.
      
      The EPA reviewed the calculations and bench-scale results submitted and agrees with the theoretical trends estimated by the commenters for pure salt compounds (see Docket item IV-B-2).  The results suggest HCl will not be formed.  However, the fluxing agents used in actual practice contain proprietary ingredients and are not typically pure salt compounds.  Furthermore, the furnaces in which these fluxing agents are used contain residues from previous furnace cycles that are available to react with current furnace cycle ingredients.  The EPA does not accept the bench-scale test results and theoretical calculations submitted by the commenters as sufficient evidence that HAP emissions are not produced in the actual furnace environment.  The Agency is not precluding the possibility of a demonstration, with full-scale emission tests at the furnace, by which an owner/operator can show that an element or compounds other than nitrogen and argon are not "reactive" fluxes.
      The emission test data submitted by one commenter (see Docket item IV-D-11) involved tests with potassium aluminum fluoride flux in a melting furnace processing non-coated runaround scrap and delacquered metal feed/charge (a group 1 furnace as defined in the final rule).  The test results by EPA Method 26A indicated 0.0067 lb HF/ton of feed for three runs (non-detectable in two runs and detectable in one run).  In another case, sodium silicofluoride salt was added to a holding furnace prior to skimming.  Measured emissions using a measurement device similar to that used in EPA Method 13A indicated 0.00057 lb HF/ton of feed.
      According to the definitions in the final rule, the melting furnace case would be a group 1 furnace subject to PM, HCl, and D/F emission standards and performance testing.  The presence of measured HF emissions, although relatively low, by definition makes the latter flux not a "non-reactive" flux and puts the furnace in the group 1 category.

      2.19 Comment: Several commenters (IV-D-22, IV-D-23, IV-D-28, IV-D-35) recommend that EPA establish a procedure by which the owner can petition for exclusion of a particular fluxing agent from the definition of reactive flux based on laboratory studies, sound science, and acceptance by the permitting agency.  Another commenter (IV-D-33) suggests that EPA develop a lab test for non-reactivity that facilities can use in support of a delisting.
      Response: The EPA has reviewed the laboratory study  suggested by one commenter, and supported by others, to provide a means of classifying the fluxes as non-reactive.  As noted in the previous comment, measurements performed at bench-scale on pure salt compounds are not sufficient evidence to place the flux on a "non-reactive" list.  The Agency would accept emission test data obtained with EPA reference methods and conducted on full-scale furnaces showing no HAP emissions as evidence a flux is non-reactive.

      2.20 Comment:  Two commenters (IV-D-11, IV-D-23) support the proposed definition of reconstruction.
      Response: The EPA acknowledges this support for the definition.


      2.21 Comment: One commenter (IV-D-11) states that furnace rebuilds with no changes in design and capacities should also be excluded from reconstruction as routine maintenance. Other commenters (IV-D-8, IV-D-31) suggest that routine burner replacements are not reconstructions, and that replacements of in-line fluxers with like replacement units are not reconstructions.
      Response:   The proposed and final rule definitions clearly state that refractory replacement is routine maintenance and not a reconstruction.  If a furnace rebuild encompasses more than refractory replacement, the determination for reconstruction is based on the fixed capital cost of the new components in comparison to the cost of a comparable new affected source.  The affected source in the case of group 1 furnaces is the SAPU as opposed to the individual furnace.  The EPA considered the comment regarding excluding furnace rebuilds with no changes in design and capacity from reconstruction, but decided it was not necessary to modify the definition of reconstruction in this regard.  The emission standards are identical for new and existing sources in any case. 
      The proposed and final rules state that replacement of an in-line fluxer with a repaired unit is not a reconstruction.  The definition in the final rule was modified to make clear that replacement of in-line fluxer components (e.g., rotors/shafts, burner tubes, refractory, warped steel) is considered to be routine maintenance.  Routine burner replacements are subject to the same capital cost criterion as furnace rebuilds and seem unlikely to cause an affected source to become "reconstructed."    
      2.22 Comment:  One commenter (IV-D-22) requests a clarification of the rule by adding a definition of "combustion zone" and modifying the residence time definition so as to allow them to demonstrate their design retains scrap dryer/delacquering kiln/decoating kiln gases at or above 1,400 F for the required time.
      Response: In the definition of residence time, combustion zone refers to the reaction chamber of the afterburner in which the waste gas stream is exposed to the direct combustion flame and in which combustion of the pollutants takes place; and in which a temperature at or above the design temperature is achieved.  (Typical design practice is for this chamber to have a length to diameter ratio of 2.0 to 3.0 with average velocities of 10 to 50 feet per second according to the Air Pollution Engineering Manual [Air and Waste Management Association, 1992]).  The EPA considered the commenter's request, but decided not to modify the definition.

      2.23 Comment: One commenter (IV-D-10) stated that the definition for scrap shredders is too broad and should be narrowed to exclude "bale breakers" used for safety reasons, enabling furnace operators to see the contents of the bale.  The commenter believes facilities will take them out-of-service if forced to control the minimal emissions.
      
      Another commenter stated that the definition of scrap shredders should be modified because they shred all types of scrap, not just aluminum.  It should be changed to "secondary aluminum sizing device" or "aluminum charge preparation."  EPA should also clarify that those shredders that process de minimis or incidental amounts will not be regulated (low percentages of aluminum-containing scrap in comparison to total weight processed). (Institute of Scrap Recyclers)
      Response: The definition in the final rule has been modified to clarify that an "aluminum scrap shredder" is the affected source as opposed to a scrap shredder.  The EPA also agrees with the commenter that bale breakers are not intended to be regulated.  The definition has been revised to state that a bale breaker is not an aluminum scrap shredder.

      2.24	Comment: One commenter (IV-D-23) wants to add the words "subject to a D/F limit" to the end of the last sentence, referring to the definition of a group 1 furnace in the proposed rule.
      Response: Group 1 furnaces are emission units that are part of a SAPU and subject to emission standards for PM, HCl, and D/F.  The definition as recommended by the commenter would not be correct.  This change has not be made in the final rule.

      2.25 Comment: According to one commenter the definitions for secondary aluminum production facility, group 1 and 2 furnaces, reactive fluxing, clean charge, and chip dryer are all poorly constructed and give the impression that die casting facilities can or should be regulated.  De minimis levels of reactive fluxing should be allowed in group 2 furnaces. (IV-D-12)
      Another commenter stated that the definitions must be modified to distinguish among source types.  For example, distinguish between facilities making new products from scrap (using processed scrap as raw material) from those only making processed scrap from unprocessed scrap.  Post-consumer scrap must be removed from the list of materials, the processing of which subjects the facility to the rule.  It can be misunderstood to apply to all sorts of scrap materials rather than just aluminum recyclables.  The commenter also stated that more specific terminology should be used in the rule.  In addition, the definition of secondary aluminum processing facility should be modified to avoid regulating scrap recyclers, except those that are part of a larger aluminum production operation. (Institute of Scrap Recyclers)
      
      Response:  The EPA received many comments explaining the differences between die casters, extruders, and foundries, as opposed to secondary aluminum production operations.  Upon consideration of these comments, the EPA has revised the definition of secondary aluminum production facilities as it may or may not apply to die casters, extruders, and foundries in the final rule.  The definition has been changed such that facilities that process no materials other than 1) materials generated within the facility, or 2) clean charge, purchased or otherwise obtained from outside the facility, and that do not operate sweat furnaces, thermal chip dryers, or scrap dryer/delacquering kilns/decoating kilns are not secondary aluminum production facilities.
      Foundries, extruders, and die casters that use external scrap are, however, considered to be secondary aluminum production facilities under the final rule, and are subject to all appropriate major or area source provisions of the rule.  The Agency considered that an important determining factor regarding who is part of the secondary aluminum production source category is the recovery of aluminum from scrap.  The Agency believes that owner/operators can reliably identify clean charge as opposed to scrap and thereby reliably limit the processing of D/F precursors.  The request of commenters to allow the use of "clean purchased scrap" in a facility not subject to the rule requires generally accepted, verifiable, defined limits to distinguish scrap types from a continuum of contamination that cannot be converted into acceptable vs. unacceptable levels of D/F emissions.  Therefore the processing of any purchased scrap in an affected source subjects the facility to the rule.

      2.26  Comment: One commenter (IV-D-32) states that they support the use of the "TEQ" definition of dioxin.
      Response: The EPA acknowledges support of the TEQ definition.

3.  Emission Standards and Operating Requirements

      3.1  Emission Standards

      3.1.1  Comment:  One commenter (IV-D-25) urged EPA to review the application of fluoride and chlorine fluxes in the secondary aluminum industry and to verify the appropriateness of HCl as a surrogate.
      Response:  Emission limits for HCl were proposed because test data indicate that HCl is emitted when chlorine and reactive chloride fluxes are used, and the technology representing the MACT floor for HCl removal, which was determined to be lime injected fabric filters, also achieves MACT floor level removal of chlorine and HF.  Although some fluoride fluxes are used by the industry, differences in flux properties, cost relative to chlorine/chloride fluxes, and occupational health considerations related to in-plant particulate levels limit the amounts used, thus limiting the potential for HF emissions.


      3.1.2  Comment:  One commenter (IV-D-25) stated that emission limits do not reflect limits achievable using currently available technology and that neither the limits nor the selected MACT accurately reflect MACT.  The commenter stated that the MACT floor emission levels violate section 112 of CAA in that they are not based on the best-controlled sources for new sources and are not at least as stringent as the best performing 12 percent for existing sources.  According to the commenter, EPA should consider, but did not, emission limits more stringent than the floor.  In a related comment, another commenter (IV-D-6) disagreed with the dioxin emission standards and stated that they are unsupported by emissions data.  According to the commenter, the method of developing the limit is inconsistent with the CAA and fails to recognize the law of averages and, in the case of SAPUs, is illegal because it permits individual group 1 furnaces to emit dioxin at levels in excess of MACT floor.
      Response:  The commenters argued that EPA did not properly consider the available emissions data in establishing the MACT floor emissions limits.  In the case, Sierra Club v. EPA (March 2, 1999), the DC Circuit held that because MACT standards must be achievable in practice, EPA must assure that the standards are achievable "under most adverse circumstances which can reasonably be expected to recur" (assuming proper design and operation of control technology).  Slip op. p. 13.  The case further holds that EPA can reasonably interpret the MACT floor methodology language so long as the Agency's methodology in a particular rule allows it to "make a reasonable estimate of the performance of the top 12 percent of units", slip op. p. 7; that evaluating how a given MACT technology performs is a permissible means of estimating this performance, id. at 13; and that new source standards need not be based on performance of a single source, id.  The court's decisions gives EPA latitude in determining the MACT floor and the MACT floor emission limits.  EPA determined the MACT floor based on information available for each affected source and emission unit.  At proposal the EPA selected emission limits at the floor level of control and the commenters provided no additional emissions data for any pollutant for EPA to consider.  The emission standards are based on the emission levels achieved through the application of MACT floor technologies and account for variation in the process and in the air pollution control device effectiveness.

       3.1.3  Comment:  One commenter (IV-D-32) stated that they generally support the emission limits proposed.
      Response:  The EPA acknowledges support for the proposes emission limits.

      3.1.4  Comment:  One commenter (IV-D-23) supports the emission classes of sources as proposed.  Another commenter (IV-D-8) stated that they approve of a clean furnace source category and what they call a clean in-line fluxer source category.  Other commenters (IV-D-8, IV-D-32) support the decision not to regulate D/F emissions from group 1 furnaces using clean charge
      
      Response:  The proposed rule does not refer to a "clean in-line fluxer" affected source.  Emission limits have not been proposed for group 2 furnaces, or for in-line fluxers that operate with no reactive flux because these affected sources do not emit HAP.  

      3.1.5  Comment:  One commenter (IV-D-23) supports EPA on the selection of pollutants and surrogates.
      Response:  The EPA acknowledges support for selection of pollutants and surrogates.

      3.1.6  Comment:  One commenter (IV-D-23) supports EPA's selection of MACT floor technologies except for lime-injected FF for Group 1 furnaces, which the commenter believes should be lime-conditioned FF.  However, the emission limits based on lime injection are not opposed by the commenter; the preamble should state these emission limits are based on above-the-floor technology.  The commenter supports the work practice standards for Group 2 furnaces and supports the decision to not require above-the-floor technologies for other sources.
      Response:  EPA acknowleges the support for the proposed standards, however the commenter is mistaken with regard to the floor technology for group 1 furnaces.  Lime injected fabric filters are in use at the best performing 12 percent of group 1 furnaces and thus represent the MACT floor.

      3.1.7  Comment:  One commenter (IV-D-27) stated that several of the identified MACT floor technologies are poorly known, or potentially unworkable.  One scrap recycling firm claimed that a lime injection system installed on a fabric filter was unable to achieve removal percentages comparable to those specified in the rule.  One firm has an existing FF that would be very expensive to modify for lime injection.
      Response:  All of the technologies determined to represent MACT are in place and operating on the best performing affected sources, and have been demonstrated through emission test data to be capable of achieving the emission limits.  Sources emitting HAP in excess of the standards must install or upgrade control devices or implement pollution prevention practices to meet the standard.

       3.1.8  Comment:  One commenter (IV-D-25) stated that EPA selected system design criteria  that are inappropriate and do not necessarily reflect the performance of add-on controls, and that the criteria are inconsistent from one application to the other within the source category.  For example, for fabric filters, air-to-cloth ratio is not a good indicator of performance, and the specified ratio is high for applications using polyester and woven fiberglass. According to the commenter, EPA has used an inventory approach rather than assess currently available materials providing enhanced performance.
      
      Response:  EPA determined the MACT floor technologies based on information available on control technologies presently in use in the secondary aluminum industry.  The add-on control design criteria discussed in the proposal preamble for the various affected sources and emission units are representative of MACT floor technology systems.  Air-to-cloth ratio and fabric type are two parameters that affect performance, but not the only parameters.  Add-on controls with design parameters other than those cited may also achieve emission levels that meet the standards.  The commenter provided no additional information to establish a different floor nor did they provide any performance test data to support a more stringent floor for new or existing sources.

      3.1.9  Comment:  One commenter (IV-D-25) stated that HCl emission limits are far above what is actually achievable based on emissions data presented.  The commenter further stated that emission limits should not be specified in terms of feed to the process since a properly designed system is capable of limiting total outlet emission regardless of feed.  According to the commenter, a properly designed baghouse with continuous lime injection is capable of <15 ppm HCl; this should be the MACT limit.
      Response:  EPA does not have sufficient data to conclude that a concentration standard of 15 ppm can be consistently achieved under all operating conditions expected within the industry, and the commenter did not provide any data to support their position.  Performance data from lime injected fabric filter tests indicate that these control devices can achieve 90 percent HCl removal rates.  Under some conditions, a 15 ppm standard may not be achievable with MACT floor level control equipment due to variations in processes and uncontrolled HCl levels, as explained in the proposal preamble.
    	
      3.1.10 Comment:  One commenter (IV-D-25) is concerned that chlorine is being emitted and is not controlled by required systems; wet scrubbing is necessary.	
      Response:  Based on recent test reports (see Docket item IV-D-36) approximately 98 percent of the chlorine emitted is in the form of HCl.  Chlorine gas may not be as effectively removed as HCl, however, under the final rule, it will be subject to the MACT floor level of control.  Regarding the commenter's suggestion that wet scrubbing should be required, EPA did not identify wet scrubbing as MACT for either existing sources or new sources.

      3.1.11 Comment:  One commenter (IV-D-19) stated that they have chip dryers and scrap dryer/delacquering kilns/decoating kilns without afterburners in California and that EPA's testing program may not be adequate to represent the population of sources.  They stated that EPA did not consider increases in NOx (presumably from afterburners) in ozone nonattainment areas.
      
      Another commenter stated that the inventory of sources is inaccurate.  The commenter is aware of at least 6 chip dryers at aluminum die casting facilities that were not considered in establishing the NESHAP; this calls into question the technical evaluation and economic impact analysis. (IV-D-20)
      Response: The EPA worked with all affected entities, sent information collection requests to facilities in the affected industries, and met with the affected industry representatives and State agencies to obtain information.  The Clean Air Act, Section 112(d)(3) states emission standards are to be promulgated considering ". . . the average emission limitation achieved . . . (for which the Administrator has emissions information) . . ."  The MACT floor technology was determined based on the information available to EPA at the time of proposal and no information was received following proposal that would justify a revision in this determination.  The numerical emission limit which can be achieved with MACT floor technology has been determined from performance tests conducted on affected sources equipped with MACT floor technology.
      Taking additional NOx emissions into consideration, the afterburner equipped affected sources are still considered to represent the best controlled sources which make up the MACT floor.  However, owner/operators are not constrained to the use afterburners to meet the standards.

  	3.1.12 Comment:  One commenter (IV-D-19) stated that their PM standard for non-ferrous metal melting is 99 percent control across the control device.  They suggested allowing this as an alternative PM standard.
      Response:  The PM emission standard is based on the use of fabric filters which commonly achieve 99.9 percent, or greater, removal at the floor.  The commenter's suggested alternative limit is less stringent than the floor technology and therefore is not an acceptable alternative.

      3.1.13 Comment:  One commenter (IV-D-24) suggested consideration of carbon injection for D/F control; a Municipal Solid Waste (MSW) facility in their county uses this technology to limit emissions to one tenth the emission standard.
      Response:  No one in the secondary aluminum industry uses activated carbon injection, therefore it was not considered for the MACT floor.  Furthermore, EPA has no data upon which to evaluate carbon injection as a beyond the floor control option for the secondary aluminum industry.  Although activated carbon injection is not cost effective at this time, industry can use this, or any technology they choose, to comply with the D/F emission limit.  They are not constrained to the use of an afterburner or a lime-injected fabric filter.


      3.1.14 Comment:  Commenters stated that the test program carried out by EPA did not investigate process factors that could cause D/F emissions in determining which activities to regulate.  The EPA D/F data for chip dryers did not consider process parameters such as the presence of cutting fluids on the chips.  Afterburners not followed by a fabric filter would have a different temperature profile than those with a fabric filter and therefore have a different D/F emission potential.  One commenter cited the June 1994 EPA report on Estimating Exposure to Dioxin-Like Compounds and effects of radiant gas cooling on formation of D/F.  These factors could cause chip dryer D/F emissions to be different.  Also, EPA tested only one chip dryer to establish the emission limit. (IV-D-12, IV-D-20)
      Another commenter stated that D/F emissions data shown in the preamble Figure 11 show variation in emissions at a single facility.  This suggests key parameters affecting emissions were not identified; EPA should have adjusted the test program to identify the causes of wide variability.  The cause might be a work practice or operational parameter that should be prohibited for high results rather than regulating an entire class. (IV-D-18)
      Response:  The floor determination and emission limit were based on the best data available to the EPA.  The variation in D/F emissions represents typical chip drying emissions profile due to the different levels of contamination of the chips.  EPA recognized the potential for variability in emissions in establishing the D/F emission limit.  The commenter provided no additional data upon which to reanalyze the emission limit.

      3.1.15 Comment:  Different types of in-line fluxers are not recognized in the rule, specifically semi-permanently sealed in-line fluxers that are not vented to a fabric filter are significantly different in operation than those described in the rule. (IV-D-27)
      Response:   The commenter did not provide any information upon which EPA could base a decision to establish a separate class of in-line fluxers.  "Sealed flux boxes" appear to be the same as other in-line fluxers and are not really sealed (See Comment 1.1 and response).  The EPA believes that an in-line fluxer ducted to a lime injected fabric filter represents MACT floor technology for this emission unit.  In-line fluxers are not required to install a control device if the owner or operator can demonstrate initial compliance by means of a performance test.  Also, no emission limits are established for in-line fluxers that use only nonreactive flux materials.  These units are required to keep records of the flux materials used and submit semiannual certification that only nonreactive flux materials are used.  Limiting chlorine flux input to in-line fluxers is a work practice that also can be used.  Sources willing to monitor flux addition and assume that all reactive flux used is emitted may demonstrate compliance for a SAPU on this basis, and avoid the need for testing. 

      3.1.16 Comment:  Several comments were received regarding the format of the proposed emission standards.  
      
      1)  Two commenters (IV-D-8, IV-D-23) support EPA's choice of format for emission limits based on lbs/ton; also, they do not oppose the concentration based limits for shredders.
      2)  Another commenter stated that tons produced would be a better choice than charge weight as a basis for limits due to the burden of charge weighing. (IV-D-27)
      3)  One commenter (IV-D-25) stated that limits should be in gr/dscf instead of lb/ton of feed since performance is defined by outlet emissions.
      4)  Two commenters (IV-D-8, IV-D-23) support the use of D/F emissions measured on the basis of TEQ. 
      Response:  1) The EPA acknowledges support for the lb/ton emission limit format and concentration limit for aluminum scrap shredders.
      2)  Sources that have difficulty determining accurate charge weights may instead measure aluminum production weights, recognizing that, as a result of refining, the production weight will be lower than the charge weight and will result in a somewhat more stringent emission limit.
      3)  The EPA believes that, generally, production based standards are preferable to concentration standards (except for aluminum scrap shredders, where production rates are difficult to reliably determine), because they encourage pollution prevention and the achievement of emission limits in the most cost effective way.  For sources within a SAPU, production based standards permit more straightforward calculation of allowable emissions.
      4) The EPA acknowledges support for D/F emissions measurements on the basis of D/F TEQ.

      3.1.17 Comment:  Two commenters (IV-D-22, IV-D-23) support the method used by EPA to select the limits except for D/F from delacquering kilns. Other commenters (IV-D-23, IV-D-33) stated that on the basis of their assessment of MACT floor performance, the short residence time delacquering kiln D/F standard should be increased from 0.25 to 0.35 μg/Mg (they supplied data from a facility not included in the EPA assessment in a table).  Another commenter (IV-D-22) supports a limit of 1.0 μg/Mg instead of the proposed 0.25 μg/Mg.
      Response:  The numerical emission limit for D/F was based on the D/F emissions rate that could be consistently achieved  (taking into account process variability and control system variability) by an affected source equipped with the MACT floor level of control.  The EPA rejected the additional data supplied by commenters because of problems in the measurements of D/F emissions.  Furthermore, the first test performed at that source met the D/F limit.  The limit of 1.0 μg/Mg recommended by one commenter is based on a source controlled only by an afterburner. Because the MACT floor technology for delacquerers is an afterburner and a lime-injected fabric filter, the EPA rejects this commenter's suggestion.


      3.1.18 Comment:  One commenter (IV-D-6) stated that the proposed dioxin emission standard for new scrap dryers and delacquering kilns does not include a work practice requirement for afterburner operation at 1400 F minimum and a minimum residence time of 1 second; these limits were established as MACT.
      Response: The proposed rule contained two alternate sets of emission standards for these affected sources.  To use one of the sets of numerical standards, the afterburner must have an afterburner an operating temperature of at least 1400 F and a minimum design residence time of 1 second.  The second set of numerical standards does not specify the use of an afterburner with a particular temperature and design residence time, however, the Agency believes that an afterburner that achieves the more stringent numerical standards will necessarily meet or exceed the 1400 F operating temperature and 1 second residence time.  No matter which set of standards the owner/operator chooses for compliance, the operating provisions of the final rule require that the afterburner temperature be maintained at the performance tested level and monitored to show continuous compliance. 

      3.1.19 Comment:  One commenter (IV-D-7) questioned the use of a single test as the basis of the emission limit for sweat furnaces and stated that it seems arbitrary and that the preamble did not present any uncontrolled emissions data.
      Response:  The EPA acknowledges that the standard was based upon data from one test.  The numerical emission limit for D/F was based on the D/F emissions rate that could be consistently achieved  (taking into account process variability and control system variability) by an affected source equipped with the MACT floor control technology.  The EPA attempted to get uncontrolled emissions data in one test by turning off the afterburner, but self-sustaining combustion occurred at 1400F (which is typical of afterburner operation) and emission levels were within the range of those with the burner on.

      3.1.20 Comment:  One commenter (IV-D-32) stated that although based on limited data, they support the emission standards for rotary dross coolers.  Other commenters (IV-D-23, IV-D-8) stated that they do not oppose the concentration-based limits for rotary dross coolers.
      Response:  The EPA acknowledges support for the emission standard and its format for rotary dross coolers.

      3.1.21 Comment:  Several comments were received on the emission standards for group 1 furnaces. 
      1) One commenter (IV-D-23) stated that test data in the Culp Aluminum report and that in the preamble differ.  They stated that all new bags were placed in the FF before the test so the data represent optimum performance.
      
      2) One commenter (IV-D-6) disagreed with dioxin limit for new and existing group 1 furnaces as not representative of MACT.  They stated that at expected stack temperatures, dioxin will be condensed on PM.  According to this commenter, a lax particulate HAP standard (8 times higher than than SCAQMD BACT for group 1 furnaces) will increase total dioxin emissions.
      3) One commenter (IV-D-6) disagreed with allowing the charging of PVC and neoprene coated scrap, dioxin precursors, to uncontrolled sidewell furnaces.
      4) One commenter (IV-D-23) supported the option of 0.4 lb HCl/ton  or 90 percent reduction.
      5) According to several commenters (IV-D-11, IV-D-22, IV-D-23, IV-D-33, IV-D-35), EPA needs to clarify the regulation so that hearth emissions from group 1 sidewell furnaces are not included unless fluxing is performed in the hearth.
      6) Other commenters (IV-D-18, IV-D-20) stated that reactive fluxing should be allowed in the hearth of sidewell furnaces without requiring control of hearth emissions, otherwise metal oxides build up in the hearth and the efficiency of the furnace is reduced.  More metal is lost when dross is skimmed and the cost of operations will increase.  Work practices and pollution prevention could be used to minimize emissions from these operations.
      7) One commenter (IV-D-24) stated that a facility in their district is operating with 10 melters feeding 8 holders.  All fluxing is done in the holders.  They interpret the rule to mean that the holders are not subject to D/F limits, but they have found that D/F from the holders is about the same as from the melters.  They suggest subjecting clean charge furnaces that flux to the D/F standard based on their data.
      Response:  1)The preamble data were preliminary.  The final data were slightly different, but were lower for HCl emissions and thus provide even stronger support for an emission limit at least as stringent as proposed.  Also, sources are free to replace filter bags prior to initial and periodic performance tests, and are expected to replace bags between performance tests (and conduct other maintenance operations) as necessary for proper operation and to maintain control device performance.
      2) Some fraction of the D/F present in the particulate matter control device will be adsorbed on the particulate matter (PM).  The EPA has no data upon which to estimate this fraction and none was provided by the commenter. The D/F emissions are controlled by dry lime injection and maintenance of the temperature at the inlet to the FF.  More than 99.99 percent of the PM is expected to be controlled by a FF representative of the MACT floor level of particulate matter control.  The PM emission limit is based on the emission rate that could be consistently achieved (taking into account process variability and control system variability) by an affected source equipped with the MACT floor control technology.
      
      3) As described by the commenter, the furnace would be subject to the emission limits for a group 1 furnace.  The final rule permits emission units within a SAPU to operate without control devices (although most sidewell furnaces are controlled) if the complete SAPU can demonstrate compliance with the emission standards through a performance test.  This performance test must be conducted at the highest load or capacity reasonably expected to occur.  Sources must establish initial compliance during a test with scrap containing the highest level of coatings and oils that the owner/operator plans to charge to the furnace (emission unit).  The EPA expects that furnaces charged with large amounts of oily and coated scrap will be controlled.  Uncontrolled emission units must operate within the limits of an approved site-specific operating and monitoring plan to ensure continuous compliance with the SAPU emission limit. 
      4) The EPA acknowledges support for HCl emission limits.
      5) Although the EPA agrees with the commenters in that hearth emissions are not subject to control unless reactive fluxing takes place in the hearth, the EPA did not revise the rule because the operating requirements make it clear that control is necessary only if reactive fluxing is performed in the hearth.
      6) Sources using reactive flux in the hearth of sidewell furnaces are required to control emissions emissions of D/F, HCl, and PM. The final rule definitions make clear that solid salts added as cover fluxes to prevent oxidation are not considered reactive fluxes.
      7) This commentor did not provide any data to support their contention.  The EPA already had data on the referenced facility and had used it for proposal in determining that holders with clean charge and reactive fluxing do not require emission limits for D/F.  In addition, the facility in question has already installed the floor controls (a continuous alkaline injection system with a fabric filter) to control HCl and PM emissions.  This is the same floor technology for furnace D/F emissions. 

      3.1.22 Comment:  The following comments were received on the melting/holding furnace emission standards:
      1) One commenter (IV-D-32) stated that although based on limited data, they support the emission standards for melting/holding furnaces.
      2) Two commenters (IV-D-8, IV-D-23) supported the method used by EPA to select the limits except for HCl for melter/holders.  They argued that the proposed melter/holder limit will restrict operational flexibility in producing high purity alloys in a melter/holder as compared to separate furnaces.  They also claimed they have provided EPA data clearly supporting at least 0.6 lb/ton and request a limit of 0.6 lb/ton. Other commenters (IV-D-31, IV-D-33) suggested increasing the melter/holder HCl limit to 0.6 lb/ton.
      3) One commenter (IV-D-31) suggested a separate emission limit section in the rule for melter/holders.
      
      4) One commenter (IV-D-24) is confused about section 1505(i)(1) of the proposed rule.  They suggested clarifying language, "Except for a melter/holder processing only clean charge,...".
      Response:  1) The EPA acknowledges support for the emission standards.
      2) The EPA has determined that the emissions data support the MACT floor level of HCl from group 1 furnaces, i.e., 0.40 lb/ton, or 90 percent removal of HCl across a lime-injected fabric filter.  One of the data sets supporting this emission limit was obtained during tests on a furnace with heavy fluxing occurring.  None of the commenters have provided data indicating that these levels could not be achieved with a well operated, properly maintained, MACT floor level control device.  Also, a source that is unable to meet the 0.40 lb/ton emission limit for an individual group 1 furnace without a control device can comply by meeting the 3-day, 24-hour rolling average emission limit for a SAPU.
      3) The final rule contains a separate paragraph to address the PM emission limit for melter/holder emission units.  The HCl limit for melter/holders is not addressed separately from other group 1 furnaces because the limit is the same.  The calculation of PM and HCl emission limits for SAPUs that include one or more melter/holders is described in the SAPU emission limit section of the rule.
      4) As proposed, EPA intended section 1505(i)(1) (i) -(iii) and 1505 (i)(2) to apply to group 1 furnaces, except for melter/holders processing clean charge at facilities that are major sources; 1505(i)(1)(ii) to apply to group 1 furnaces, except for melter/holders processing clean charge at facilities that are area sources, and 1505(i)(3) to apply to group 1 melter/holders processing clean charge.  This section has been rewritten to clarify EPA's intent. 

      3.1.23 Comment:  Although based on limited data, one commenter (IV-D-32) supported the emission standards for flux boxes.  One commenter (IV-D-8) stated that the PM and HCl limits for in-line fluxers are too tight, which will discourage the use of new in-line fluxers and cause sources to flux in the furnace to get the higher SAPU limit (and not install an APCD for the new in-line fluxer).  This commenter thinks that the determination of the lime baghouse as the MACT floor for in-line fluxers is technically accurate but will discourage the use of in-line fluxers and result in increased emissions. The commenter wanted no PM limit on in-line fluxers and stated that EPA has not made the case that an emission limit is necessary to limit HAP metals.  The commenter stated that the requirement has a bad cost effectiveness. The same commenter wanted no limit, only work practices, for unvented in-line fluxers.  The commenter stated that the standards for in-line fluxers did not account for capture efficiency and should not apply to fugitive sources.
      
      Response: The EPA has determined that the MACT floor technology for new in-line fluxers is a lime-injected FF.  Emission limits have been promulgated based on the level of emissions that can be obtained on a consistent basis when MACT floor control technology is applied.  Sources may demonstrate compliance with the HCl and PM standards without the installation of a control device through a performance test.  The promulgated standards are based on the MACT floor.  New unvented in-line fluxers must achieve the MACT floor level of emissions, regardless of whether those emissions are discharged through a stack or are fugitive.  Sources may not avoid compliance merely by eliminating capture equipment and converting stack emissions to fugitive emissions.   

      3.1.24 Comment:  The following comments were received on SAPUs:
      1) Several commenters (IV-D-8, IV-D11, IV-D-23, IV-D-32) supported the establishment of the SAPU which affords some compliance flexibility; the equations to calculate limits are also supported.
      2) One commenter (IV-D-31) wanted individual compliance provisions for emission units as an alternative to SAPU compliance.
      3) One commenter (IV-D-17) stated that the 3-day, 24-hour rolling average for SAPU is problematic as a result of reduced operating schedule on weekends and suggested revising the requirement to a 7-day, 24-hour rolling average.  Other commenters (IV-D-11, IV-D-31) recommended a monthly basis similar to monthly basis for determining compliance for primary aluminum reduction plants.  Another commenter (IV-D-32) supported the emission standard based on a 3-day, 24-hour rolling average, which provides flexibility for facilities to use the most efficient techniques to control overall emissions.
      4)  One commenter requested monthly compliance  demonstrations, rather than compliance by cycle or each 24 hours.  This commenter notes that some processes have cycles greater than 24 hours. (IV-D-8)
      5) Two commenters (IV-D-17, IV-D-32) stated that the cross reference to required information in section 63.1505(l)(3)(i) of the proposed rule is incorrect; should be (l)(2) rather than (m)(2).
      6) One commenter (IV-D-24) stated that in section 1505(a) of the proposed rule (emission standards), they are confused about the reason for excepting SAPUs.
      Response:  1) The EPA acknowledges support for the SAPU provisions.	
      2) The preamble to the final rule includes language which explicitly states that a source that demonstrates that all existing emission units within a SAPU are in compliance with the limits for group 1 furnaces and in-line fluxers, has established that its SAPU is in compliance with the SAPU emission limits.
      
      3) The EPA considered other averaging times and concluded that the 3-day rolling average is an appropriate averaging time.  Based on test data, individual emission units have the capability of complying on a batch-by-batch basis.  Because of the different cycle times, and overlapping cycles, a three day, 24 hour rolling average provides a meaningful way of representing the emissions contributions from the assemblage of emission units within the SAPU.
      4)  The EPA recognizes that some processes may have cycles longer than 24 hours, but this is not typical.  The establishment of compliance demonstrations based on one cycle or 24 hours is related to the period of time over which the data were collected during performance tests.  To establish a longer period for compliance requires development of longer-term performance information not currently available to EPA.  Emission limits based on monthly compliance periods would necessarily be lower than those limits established for one cycle or 24-hour periods based on the averaging time involved. 
      5) The EPA agrees with the commenters and has made the correction.
      6) In the proposed rule the emission standards for SAPUs were written differently than the other standards because the SAPU could comply on a 3-day, 24-hour rolling average basis.  The exception is not needed; the paragraph has been revised in the final rule.

      3.1.25 Comment:  One commenter (IV-D-11) supports no emission limits for group 2 furnaces.
      Response:  The EPA acknowledges support for no emission limits for group 2 furnaces.

      3.1.26 Comment:  Two commenters stated that the selection of facilities tested did not represent the entire regulated community because no die casting and foundry operations were tested.  They claimed that EPA did not solicit sufficient information from die casters and foundries through the ICR process.  They stated that only one set of tests per operation is insufficient to establish standards, the test data obtained were severely limited and statistically weak, that operational parameters and conditions were not presented, nor a determination of representativeness of the industry.  They also questioned how  one set of data can be used to develop emission limits for chip dryers. (IV-D-18, IV-D-20) 
      Another commenter claimed that EPA has no basis to assume that die caster chip dryers emit D/F at levels requiring control. (IV-D-26)   
      Two other commenters stated that emission standards and operating requirements were developed without reference to die casting, foundry facilities, and scrap recyclers and that this acted to bias the data.  (IV-D-20, IV-D-27)
      
      Response:  Many meetings were held during the regulatory development process with relevant industry associations, thus providing them with the opportunity to participate in the development process, and some die casters were included in the ICR activity.  In discussions with die casting industry staff early in the project, representatives of the industry stated that they charged only clean materials to their processes, thus leading EPA to conclude at that time that their portion of the industry would be unlikely to be subject to the rule.  Subsequently the Agency obtained additional information that indicated some aluminum die casting, foundry, and extrusion facilities do operate sources using purchased scrap.  These sources are similar to those operated in facilities whose primary purpose is secondary aluminum recovery.  
      Much of the test data on which the emission standards are based were collected in a test program developed, and at sites selected through discussions with industry staff as representative of the affected sources and with MACT floor control systems in place.  The test program covered the full range of furnace operations including operations that are representative of those facilities at die casting and foundry facilities.  Therefore the data are representative of emissions from die casting and foundry facilities which perform the same process operations and which use the floor technologies/work practices and/or pollution prevention measures.  These data were considered in establishing the numerical emission limits which could be consistently achieved through application of MACT floor technology.  Also, at proposal, the Agency requested any additional data from die casting and foundry facilities relevant to HAP emissions from affected sources that might be small businesses and which processed offsite scrap as well as additional information on major sources.  Commenters provided no additional test data which would indicate a decreased potential for D/F emissions from die caster chip dryers, relative to chip dryers operated at other types of facilities, nor were D/F emissions data provided for potentially affected sources at other die caster and foundry facilities.  No data provided indications that small business were impacted.  The response to Comment 1.4 in the applicability section of this document explains why D/F emissions are expected from the foundry and die casting facilities that use purchased scrap.  
      
      
      3.1.27 Comment:  Two commenters stated that EPA's testing program was insufficient to justify setting numerical limits because only one chip dryer was tested.  EPA chose not to set limits for other nonferrous metal industries because of insufficient data.  The commenter cited the Background Information Document for the Secondary Lead NESHAP (60 FR 121, June 23, 1995).  They argued that it cannot be sufficient for EPA to set a D/F limit on chip dryers using data from tests of a single emission point if six tests at three lead facilities is not sufficient data and gives no technical basis to regulate.  The inventory reported in the April 1998 draft of "The Inventory of Sources of Dioxin in the United States" assigned low confidence to emissions estimated for secondary aluminum operations, but medium confidence for emissions estimated for secondary lead.  (IV-D-18, IV-D-20)
      Response:  As noted in the response to the previous comment, the sites for testing were selected jointly by EPA and industry through discussions and meetings with industry staff.  The chip dryer site selected was equipped with the MACT floor technology and was representative of the affected sources, chip dryers charging oily chips.  The emission limits are based on the level of pollutant control which EPA believes to be consistently achievable by affected sources employing the MACT floor level of control, taking into consideration affected source variability and control device variability. 

      3.1.28 Comment:  One commenter (IV-D-13) recommended that EPA provide for alternative means of compliance.  The commenter stated that the local agency requires 99 percent PM removal for baghouses, instead of the standard's mass per unit of production.  Methods of compliance should allow for differences in production rates, type of scrap melted, amount of reactive flux actually used, baghouse inlet loading, and actual measured emissions.
      One commenter supported alternate requirements for area sources.  They stated that limits and control requirements need to be reasonably applied.  Permitting agencies should have the discretion to classify sources as insignificant or fugitive within clear guidelines.  NESHAP rules anticipated the possibility of giving states authority to exclude area sources from controls that should not apply, or to establish alternate limits and practices.  Some equipment may meet source definitions, but does not emit HAPs, or there are no visible or fugitive emissions.  The commenter stated that these sources could be reviewed and classified insignificant by the state permitting authority. (IV-D-27)
      Response:  In regard to the first part of the comment, EPA notes that fabric filters are routinely capable of removing greater than 99 percent of PM and that a mass per unit of production was selected for the very reason that it does allow for the variations cited by the commenter.  It is also permissible for States to have different requirements as long as they are not less stringent.
      
      In regard to the second part of the comment, since EPA has established limits for facilities that are area sources, States as delegated authorities must enforce the MACT limits or more stringent limits.  The Clean Air Act does not provide for de minimis emission thresholds when a MACT floor exists.  Control devices are not required for affected sources which can demonstrate through performance tests that they are in compliance with the numerical emission limits equivalent to the application of MACT floor technology.  In the absence of performance tests and monitoring, EPA has no means of establishing that sources are in continuous compliance with the emission limit representing the MACT floor level of control. 

      3.1.29 Comment:  One commenter (IV-D-9) disagreed with EPA's assumption that pollution prevention will be limited to type of scrap charged and the type and amount of fluxing agent used and argued that MACT floor equivalence is achievable through furnace design, production controls, and revised operating practice.  According to the commenter, the rule does not adequately support innovative design and alternative operating approaches as means to limiting emissions as opposed to reliance on control systems.
      Response:  The EPA agrees with the commenter that furnace design, production controls, and revised operating practices, in addition to type of scrap charged and fluxing practices, can all be used to meet the MACT floor emission limits.  The proposed rule, as well as the final rule, allows the use of any approach to achieve compliance.  Sources may demonstrate compliance, by  performance tests, with the standard by any means, including pollution prevention or alternate control devices. Sources must monitor appropriate parameters to demonstrate compliance between tests.

      3.1.30 Comment:  One commenter (IV-D-11) is concerned that the rule does not address situations where commonly ducted sources have different emission limit formats, e.g., lbs/ton expressed on an instantaneous basis vs 3-day, 24-hour rolling average lbs/ton basis.
      Response:  Sources may be vented to a common control device if the same pollutants are being controlled and the standards are on a lb/ton basis.  For other existing sources, such as delacquering kilns and furnaces ducted to a common FF, a case-by-case compliance demonstration is required with the permitting authority providing approval of the test plan before the tests are conducted.

      3.1.31 Comment:  One commenter (IV-D-17) stated that the equations in 63.1505(k) and Table 1 should be clarified to note that "feed rate" in "total tons processed" should be used for batch processes.
      Response:  In order for the calculation to yield the emission limit in the proper units, i.e., mass of emissions per unit of charge, the feed rate must always be expressed in mass per unit of time.  If total tons processed is used, it must be over a specified period.

      3.2 Operating Requirements

      3.2.1  Comment:  Several commenters (IV-D-8, IV-D-22, IV-D-23, IV-D-32) did not want an exceedance of an operating parameter to be a violation of an operating requirement.  According to the commenters:
      
      1) The rule is not clear as to what constitutes a violation of the operating requirements.
      2) Operating parameters are only indicators of process and control performance, not a direct measure of excess emissions.
      3) An exceedance should not to be a violation until six exceedances occur in a 6-month period.
      4) No more than one violation should be counted per 24 hour period for any one parameter.
      5) The rule is not clear on whether a failure to take corrective action in response to an exceedance is a violation of the standard.
      6) A failure to initiate corrective action within 1 hour should constitute a violation.
      7) The rule should specify that if corrective action is begun within 1 hour and completed in accordance with the startup, shutdown, malfunction (SSM) plan, no violation has occurred.
      Response:  1)The language in the final rule has been revised to make clear that a deviation of an operating parameter is a violation of the operating requirements.  Each owner/operator is required to define the compliance parameters to be monitored in their OM&M plan.  Then, during the initial performance tests they are required to monitor and establish the value or range of the parameters.  These values must be reported in the results of the test and notification of compliance status to the permitting authority and must be approved by the permitting authority.  During subsequent operations, if the monitored parameters exceed the values or fall outside the range determined during the initial performance test, it is a violation of the operating requirements of the standard (unless it is the result of a malfunction to which the facility responds to in accordance with the SSM plan).
      2) The owner/operator may use CEMs as a direct measure of the emissions rather than using operating parameters if such CEMs can be demonstrated to the satisfaction of the permitting agency to reliably measure emissions.
      3) The EPA has no basis for allowing six deviations before considering the facility to be in violation.  The owner/operator has ample opportunity to establish a range for the operating parameters and must thereafter operate within that range.
      4) Any deviation of an operating parameter limit is a violation of the operating standard, regardless of when it occurs, unless it is the result of a malfunction to which the owner or operator responds in accordance with the SSM plan.
      5) and 6)  The rule requires corrective action as a result of an operating parameter deviation or bag leak detector alarm.  Corrective action must be conducted in accordance with the OM&M plan.  Failure to take corrective action and to complete corrective action as expeditiously as practicable, is a violation of the operating requirements.
      
      7) A deviation that is the result of a malfunction, to which the facility responds in accordance to its SSM plan, is excluded as a violation.

      3.2.2  Comment:  Related to the corrective action requirements after deviations from process or add-on air pollution control equipment operating parameter values or ranges, one commenter suggested replacing "deviation" with "excursion" and "deviates" with "experiences an excursion from".  According to the commenter this will avoid confusion with Title V program requirements to report deviations from permit requirements.
(IV-D-28)
      Response: Sources that must obtain a Title V permit will have the process and add-on air pollution control equipment operating parameter values and ranges incorporated in their Title V permit. As a result there is no opportunity for confusion.  Unless a deviation results from a malfunction, a deviation under the regulation is also a deviation under the Title V permit.

      3.2.3  Comment:  Several commenters (IV-D-17, IV-D-18, IV-D-20, IV-D-22,IV-D-23) disagreed with the requirement that capture and collection systems meet the criteria established by the American Conference of Governmental Industrial Hygienists (ACGIH) for hooding and ventilation systems. The commenters claimed EPA has not shown that MACT floor facilities' hooding and ventilation systems met ACGIH criteria so that the requirement is arbitrary; EPA should show that the facilities met the ACGIH criteria. Several commenters (IV-D-1, IV-D-23, IV-D-31) stated that because EPA has no data to support the requirement for ACGIH criteria for capture and collection equipment for existing sources, they recommended the requirement only apply to new sources.  Other commenters stated that although currently protecting work space air quality, most existing systems would not meet ACGIH criteria meaning significant expenditures to upgrade those systems.  EPA did not likely account for these costs in their economic analysis; they agree with the commenters who stated that the requirement should be limited to new or modified sources. (IV-D-18, IV-D-20)
      Response:  For affected sources and emission units that require an air pollution control device, a capture and control system meeting ACGIH criteria is necessary for occupational safety, and to meet the emission standards.  The emission standards are based on systems that effectively capture and contain emissions at the source (minimizing fugitives) and convey them to the control device for removal.  In addition, a capture and control system meeting ACGIH criteria with good hooding design will result in a lower volume of exhaust air to be treated, and in many cases, a smaller, lower-cost control device.  EPA considers an ACGIH capture and collection system to be part of MACT floor technology for affected sources with add-on controls.


      3.2.4  Comment:  One commenter (IV-D-32) supported not counting false alarms of the bag leak detection system in the alarm time.  Another commenter (IV-D-25) stated that the monitoring and reporting requirements are reasonable in order to confirm compliance, with exception of bag leak detectors.  The commenter stated that a facility should not be penalized for rapid response to an alarm and recommends that the actual time be counted and delete the 1-hour minimum alarm time. 
      Response:  The rule has been clarified so that false alarms are not counted.  Alarms that occur during conditions covered by the facility's startup, shutdown, and malfunction (SSM) plan also are not counted.  A one hour minimum has been retained in the final rule to encourage proactive fabric filter maintenance.  This provision does not penalize the owner/operator if action is started in accordance with the facility's OM&M plan within one hour.

      3.2.5  Comment:  Several commenters (IV-D-1, IV-D-9, IV-D-10, IV-D-17, IV-D-18, IV-D-20, IV-D-22, IV-D-23, IV-D-27, IV-D-32) did not want the labeling requirement.  They argued that:
      1) The inspectors can get this information from the OM&M plan in the office before entering the plant.
      2) The labels will be hard to maintain in a plant environment.
      3) It creates opportunity for violation with no commensurate benefit and increases/duplicates regulatory paperwork.
      4) The labeling requirement generates safety concerns.
      Response:  The EPA believes that labeling requirements are necessary for enforcement and operating purposes, and should be retained due to the complexity of the industry and the numerous possible facility configurations (and emission units that could be combined within a SAPU).  Labeling will help prevent operators from charging the wrong materials or improperly operating the units and will help inspectors in identifying units and determining if the units are being properly operated.  However, EPA understands industry's concerns over the implementation of the labeling requirements and has revised the rule to require labeling only at those affected sources and emission units that can be operated in more than one mode and/or which are physically very similar, including group 1 furnaces with and without add-on controls, group 2 furnaces, scrap dryer/delacquering kilns/decoating kilns, and in-line fluxers.  In addition, the rule requires that labels contain only the identification of the unit and the applicable operational standards.  These revisions respond to industry's concerns regarding increased regulatory paperwork with no commensurate benefits while maintaining enforceability of the standards, since both operators and inspectors will clearly know the operating standards/requirements of each emission unit.
     

      3.2.6  Comment:  Several commenters (IV-D-11, IV-D-17, IV-D-23, IV-D-31, IV-D-32,) disagreed with requirement to maintain the same flux injection schedule as used in performance test. One commenter (IV-D-32) stated that they should be given flexibility to develop schedule procedures during performance tests subject to approval by the permitting authority.  According to one commenter (IV-D-11), the requirement to maintain the same flux injection schedule as used in performance test, which would be done under worst case conditions, would result in an increase in HCl emissions and cause other negative environmental impacts.  Another commenter (IV-D-17) stated that this requirement will cause increased HCl emissions for uncontrolled group 1 furnaces and will restrict work practices to minimize chlorine use.  The commenter suggested a separate provision to maintain the same flux injection schedule for baghouses with semi-continuous lime feed systems.
      One commenter (IV-D-31) wanted flux monitoring on a monthly basis and the schedule requirement eliminated.  The rule could be interpreted to preclude a system with computerized monitoring of furnace operations/controls with correlated emissions and online continuous emissions calculations.
      Response:  Owners or operators are required to conduct performance tests under the highest load or capacity reasonably expected to occur. This is represented by the maximum reactive flux rate.  The final rule provides that sources may flux (on a lb per ton of feed/charge basis), up to the limit established during a successful performance test and does not require maintaining the same schedule.  The rule also does not require owners or operators to use more flux than necessary to produce a saleable product.  These requirements will not lead to increased HCl emissions.
      Reactive flux monitoring on a monthly basis is not acceptable in that it is inconsistent with the emission standards based on 3-day 24-hour rolling averages and the established monitoring parameter values or ranges derived during the performance test.  Monitoring over a period consistent with the basis of the emission standards provides the necessary evidence of continuous compliance.
      The issue of flux injection rate and schedule is related to lime injection practice for the FF control systems.  The final rule provides operating requirements for the floor technology, continuous lime injection systems with lime-injected fabric filters.  Owner/operators who want to use intermittent lime feed systems (as opposed to continuous injection) must show compliance with the emission limits and must apply to the permitting authority for approval of an alternative lime addition monitoring procedure.  The owner/operator must provide information as necessary to show that the applicable emission limits will be achieved on a continuous basis.
      
      The rule does not preclude the use of computerized systems that correlate controls and operating practices with emissions and calculate emissions on a continuous basis once this approach is approved by the permitting authority and incorporated into the site-specific OM&M plan.

      3.2.7  Comment:  Several commenters  (IV-D-11, IV-D-17, IV-D-22,IV-D-23) disagreed with the +25 F associated with the inlet temperature limit for FF established during the initial performance test.  According to two commenters (IV-D-17, IV-D-23) the operating temperature of these FF will vary more than 25 F due to changes in ambient temperature.  This creates an unnecessary risk of violation and provides no environmental benefit. Another commenter (IV-D-11) stated that instead of the temperature requirement, electrochemical HCl sensors for automatic lime feed adjustment and other automatic systems should be considered to allow greater operating flexibility.  One commenter (IV-D-17) stated that in-line fluxers are not regulated for dioxin emissions and therefore do not need a temperature limit.
      Response:  The final rule has been changed to eliminate this requirement for fabric filters used only to control in-line fluxers since these units operate at temperatures that are close to ambient temperature.  For other affected sources and emission units, or fluxers ducted to a device cocontrolling other emission units, the +25 F limit is retained.  Operators commonly use  dilution air or water sprays as required to maintain the FF inlet temperature within the range.  Also, performance tests could be conducted at worst case conditions.  For example, performance tests could be conducted so that the inlet temperature is much higher than the normal operating inlet temperature (450 F vs 380 F, for example) thus providing a larger operating range.  Dioxin formation is strongly influenced by the temperature at the FF inlet, and temperature control is the means of preventing D/F formation (and enhancing HCl removal) in the fabric filter.  Temperature is also a parameter which is monitored to ensure continuous compliance between periodic performance tests.  This is because it is an indicator of control device performance for D/F and HCl emissions.

      3.2.8  Comment:  According to one commenter (IV-D-32), owners and operators could demonstrate compliance with the HCl emission limit by monitoring total chlorine input and showing it to be less than the emission limit.
      Response:  The EPA agrees, for in-line fluxers and group 1 furnaces processing only clean charge, operators may demonstrate compliance with the HCl emission standards (in lieu of performance tests) by demonstrating that reactive flux injection is limited to a rate which would not exceed the SAPU standard if emitted in its entirety.  However, the operators must also show the affected sources or emission units are meeting the PM limit as well. 

      3.2.9 Comment:  One commenter (IV-D-24) stated that there are two sections 1506(g)(2) in the proposed rule.
      
      Response:  The commenter is mistaken.

      3.2.10 Comment:  One commenter (IV-D-31) wanted a SAPU wide bubble on flux and lime monitoring.
      Response:  Flux and lime monitoring is required to ensure that emissions do not increase above the level measured during a successful performance test.  A SAPU wide bubble would, for example, allow all of the flux to be added to one furnace and all of the lime to a FF controlling a different furnace.  This would not ensure continuous compliance with the standard, and would permit affected sources and emission units to operate with less than the MACT floor level of control. 

      3.2.11 Comment:  Several commenters (IV-D-17, IV-D-22, IV-D-23) agreed with the selection of temperature as an operating parameter for FF used on group 1 furnaces and delacquering/drying kilns with FF, but disagreed with requirement to maintain the inlet temperature within +25 F of the performance test value. Two commenters (IV-D-17, IV-D-23) claimed most FFs operate in range of 300-350 F to protect bags and want to allow 50 F above that determined during performance test instead of 25 F.  They stated that EPA has not shown that HCl adsorption is a strong function of temperature at 350 F. One commenter (IV-D-22) recommended that the temperature simply be limited to 400 F.  One commenter (IV-D-11) further disagreed with the requirement to maintain temperatures for delacquering kiln afterburners and fabric filters relative to the established average temperatures.
      One commenter (IV-D-6) stated that the 25 F operating tolerance is arbitrary and there is no evidence that this will not result in dioxin emissions in excess of standard. They stated that EPA should require worst case performance testing and not permit sources to exceed the control device inlet temperature of the performance test.
      Another commenter (IV-D-17) stated that the language in section 63.1506(g) of the proposed rule to monitor the inlet temperature of fabric filter needs to be rewritten to be more clear. 
      Response:  There is evidence to suggest that D/F formation is strongly influenced by temperature in the FF.  There is no assurance that an affected source that meets the standard at one temperature will continue to meet the standard at a higher temperature.  The rule allows a 25 F tolerance in recognition of the potential time lag in temperature sensing and control system response.  Operators are free to conduct performance tests under worst case conditions, e.g., 350 F and then operate 300 F to achieve a 50 F window, or to upgrade temperature control systems to maintain the FF inlet temperature within the 25 F tolerance under all conditions.


      3.2.12 Comment:  One commenter (IV-D-10) stated that the requirement for continuous lime injection for FF will generate excessive quantities of potentially hazardous waste to be accommodated by declining landfill capacity.  They stated that lime precoating on finer mesh polyester bags may be just as effective and that continuous feeding does not ensure continuous injection because feeders become clogged.  Another commenter (IV-D-13) is concerned about the requirement to add lime to baghouse continuously since their current practice is to inject lime at the beginning of 24 hour melting cycle; at the end of the cycle, bags are cleaned and fresh lime is added.  They stated that the pressure drop across filters is monitored and when the pressure reaches a predetermined limit, melting is halted or bags cleaned and new lime is added. The commenter is uncertain as to how to add lime continuously and questioned the need to do so because only a finite amount of lime can be applied to bags and the rest would be wasted.
      Response:  Continuous lime injection is the MACT floor technology upon which the emission limits are based.  The operating requirements in the final rule are tailored to continuous lime injection systems.  Operators may use batch injection or another technology with the approval of the permitting authority, and if they can demonstrate compliance with the emission limits.  Operators may conduct performance tests under intermittent lime addition conditions and maintain intermittent lime addition at performance test conditions when approved in advance by the permitting authority.  The rule includes requirements for determining if continuous lime feeders are clogged and corrective action in the event that they are not operating properly. 
 
      3.2.13 Comment:  One commenter (IV-D-17) stated that the cross reference to emission standards for group 1 furnaces without add-on controls in 63.1506(n)(3) is incorrect; it should be 63.1505(i)(3) rather than (2). One commenter (IV-D-24) pointed out an error in 1506(n)(3); the reference should be to 1505(i)(1) and (3).
      Response:  These references have been corrected in the final rule.

      3.2.14 Comment:  One commenter (IV-D-9) recommended allowing furnace redesign as a compliance alternative for group 1 furnaces.
      Response:  Operators may use any means they choose to comply with the emission limits.  If a successful performance test is conducted without the use of a control device, the source is not required to install a control device.

      3.2.15 Comment:  One commenter (IV-D-31) wanted site-specific monitoring for uncontrolled emission units.
      
      Response:  A source may develop (with the approval of the permitting authority) alternate monitoring procedures for emission units within a SAPU or affected sources without add-on controls. 

      3.2.16 Comment:  Several comments were received on the feed/charge weight requirements.
      1) One commenter (IV-D-17) stated that the requirement for feed/charge weight determinations (63.1506(d)) should be revised to be consistent with the Part 70 operating permit program and recommended language to that effect.
      2) One commenter (IV-D-32) supported the flexibility provided by feed/charge weight options but stated that the facility should also be allowed to substitute production weight for feed weight.  This would cause the denominator for emission rate calculation to be smaller and therefore give a conservative measure of compliance.
      3) One commenter  (IV-D-31) wanted facility-wide charge weight monitoring. 
      4) One commenter (IV-D-31) wanted charge weight monitoring frequency to be site-specific.
      Response:  1) Accurate determination of charge weight is required to ensure compliance with production-based standards.  Sources that are unable to meet the charge weight measurement tolerances established in the rule may apply to the permitting authority for approval of alternate monitoring procedures.
      2) Operators are permitted to substitute production rate for charge weight and the final rule states this.
      3) Operators demonstrate compliance for SAPUs based on  performance tests.  Charge weights must be determined on an emission unit basis to ensure that the calculations for SAPU wide emissions are correct.  In addition, flux rate monitoring is based on individual emission unit production rates.  Control devices must be tested under the highest load reasonably anticipated to occur, and if production is shifted from emission unit to emission unit, performance test data would not be valid, therefore this is not allowed.
      4) For batch processes, charge weight must be monitored for each furnace cycle to properly calculate SAPU compliance, and to determine that individual emission units are being operated within established parameters.

      3.2.17 Comment:  Two commenters (IV-D-23, IV-D-32) stated that air breaks in ductwork are needed to achieve cooling of gases upstream of FFs to protect FFs. They recommended deleting proposed section 63.1506(c)(2).
      Response:  The final rule has been clarified to permit air breaks in ductwork for gas cooling.


      3.2.18 Comment:  One commenter (IV-D-17) stated that the wording in Table 1 of section 63.1506 of the proposed rule requiring corrective action when "any" visible emission occurs, effectively prohibits any visible emissions. The commenter recommended corrective action when a 6-minute average opacity reading of greater than or equal to 5 percent is determined via a visible emissions evaluation. 
      Response:  The commenter correctly interpreted the rule in that EPA requires corrective action when a visible emission occurs.  This is required in instances when a continuous opacity monitor is not installed and the operating requirement is periodic visual inspection for emissions.  The EPA and industry generally concurred that no visible emissions occur when the control technology is operated properly.

4.  Monitoring And Compliance Requirements 

      4.1	Monitoring Requirements

      4.1.1 Comment:  Several comments were received that requested more flexibility in the monitoring requirements aimed at reducing the burden to the industry:
      1)  One commenter stated that the operating and monitoring requirements of §§63.1506 and 63.1510 are too prescriptive and not consistent with preamble statements regarding flexibility. (IV-D-32)
      2)  Several commenters stated that EPA should allow alternative site-specific monitoring and operating plans to improve feasibility and cost effectiveness. (IV-D-9, IV-D-31, IV-D-33, IV-D-35)
      3)  Another commenter stated that separate provisions should be included in each of §§63.1506 and 63.1510 allowing facilities to develop alternative procedures approvable by the applicable permitting agency. (IV-D-32)
      4)  Two commenters claimed the provisions will result in burdensome, labor-intensive requirements without commensurate benefit to the environment. (IV-D-22, IV-D-35)
      5)  Another commenter with a rolling mill facility claimed their plant is operating at demonstrated low emission levels and seeks monitoring plan flexibility to allow their facility to continue in its present mode.  Referring to this plant, another commenter stated that the plant has developed a correlation between opacity and PM which has been used for over a year, in accordance with a regulatory order.  This monitoring has been approved by EPA and the local agency and is federally enforceable. (IV-D-32)
      Response:  The final rule has been written to incorporate more flexibility in the monitoring requirements:
      
      1), 2), and 3) The final rule includes explicit provisions for obtaining approval to use alternative monitoring procedures and lists the types of information needed in the application.  It includes data or information to justify the request such as technical or economic infeasibility, a description of the proposed alternative monitoring requirements including operating parameters and how the limit for SAPUs (if SAPUs are included in the application) will be calculated, and information as to how the alternative monitoring requirements would provide equivalent or better assurance of compliance with the standards.
      In addition, in response to the numerous comments received regarding the proposed monitoring and operating provisions, the final rule has been written to provide more flexibility to individual facilities in developing their OM&M plans and for approval of site-specific monitoring and operating alternatives, within EPA guidelines, by the permitting authority.  Additional comment responses below discuss some specific changes made in the final rule.
      4) The monitoring requirements are necessary to demonstrate continuous compliance and, as such, are environmentally beneficial.  Most, if not all, of the monitoring data collection or logging can be computerized and, therefore, will not be labor intensive.
      5) Specifically, the final rule allows the owner/operator of a plant to apply to the Administrator for alternative monitoring, if necessary, or document their current procedures in the facility OM&M plan.  The OM&M plan is submitted to the permitting authority for review and approval.  The final rule gives more flexibility, for example, through guidance for scrap inspections (used in operating limits and monitoring) that is less prescriptive and more options for lime injection monitoring.

      4.1.2  Comment:  Several commenters stated that the monitoring frequencies and data quality objectives are too restrictive and specific for application across a diverse industry and bear no relevance to the emission standards or ensuring proper operation of emission controls. (IV-D-9, IV-D-22, IV-D-31, IV-D-35,)  Another commenter agreed with the selection of the monitoring parameters in the proposed rule, but stated that the monitoring intervals are too frequent. (IV-D-8)
      Response:  Monitoring frequency requirements are related to the need for evidence of continuous compliance, and frequent readings are essential to provide the demonstration.  However, the final rule changes the frequency of recording monitored parameter values from that proposed.  For example, the frequency of recording fluxing rates has been reduced by requiring readings only during periods when flux additions are occurring.  Additional options included for monitoring free-flowing lime change those monitoring and frequency requirements and increase the monitoring options.  Furthermore, the provisions for site-specific OM&M plans approved by permitting authorities allow opportunity for adjustment of monitoring, within EPA guidelines, to fit site-specific conditions.  Comments dealing with data-quality objectives for specific monitored parameters are addressed in more detail below.


      4.1.3  Comment:  Several commenters argued that the requirements for accuracy of 1 percent when applied to feed/charge weight and flux injection rates are overly stringent and burdensome and create an unnecessary increment for a violation. (IV-D-1, IV-D-9, IV-D-10, IV-D-11, IV-D-22, IV-D-23, IV-D-27, IV-D-32)
      Response:  The EPA has retained the 1 percent accuracy requirement in the final rule.  However, the EPA recognizes there may be situations in which 1 percent accuracy for feed/charge weight and chlorine flux injection rate is not workable.  An example of this may be operating at a very low flux injection rate.  The final rule has been written to allow the permitting authority to approve alternative accuracy requirements for monitoring equipment, on a site-specific basis, in situations where the 1 percent accuracy requirement is not workable and where the owner/operator provides data/information to substantiate that emission standards will be achieved on a continuous basis.

      4.1.4  Comment:  In comments on accuracy of performance test measurements and feed/charge weight measurements:
      1)  One commenter stated that the EPA reference methods are not better than 10 percent repeatable, so the requirement for 1 percent accuracy in charge weight is arbitrary and unnecessarily burdensome. (IV-D-32)
      2)  Another commenter requested less stringency in the accuracy requirement for the sources whose emissions are well under the emission limit, noting that the expected accuracy of Methods 26A and 5 is 10 percent.  This commenter suggested that the charge weight monitoring be restricted to only those sources having to comply with a lb/ton emission limit. (IV-D-8)
      3)  An additional commenter stated that an aggregate accuracy of 5 percent is more representative of reproducible floor practice. (IV-D-9)
      4)  Another commenter wanted the weight monitoring not to be required for each emission unit, but allowed to be aggregated across emission units. (IV-D-31)
      Response:  The EPA considered the measurement accuracy issue raised by the commenters and addresses their points as follows:
      1), 2), and 3)  Related to test method accuracy, the EPA notes that the variability in the test methods, process, and control equipment is incorporated into the testing results upon which the emission limits are based.  The limits have been established to accommodate that variability.  Given that the emission limits are on a lb-of-emission/ton-of-feed (or charge) basis, it is also in the owner/operators best interest to make an accurate weight determination because inaccurate measurements could cause them to be out of compliance.  As noted in the previous response, the final rule provides additional flexibility with regard to feed/charge measurement in situations where the 1 percent accuracy is not workable. 
      
      4)  Weight monitoring is required because the emission limits are based on lb/ton of feed/charge or product.  Under the site-specific OM&M plans, individual emission units of the same type may have different allowable emission rates based on the presence of add-on control devices, fluxing practice, and feed/charge practices.  The only way to determine compliance is to monitor weights for individual emission units. 
      
      4.1.5  Comment:  One commenter stated that inspection of lime feed systems once per 8-hour shift and more frequently when found to be plugged may be difficult, arguing that visual inspection at silo and bin tops is dangerous.  The commenter suggested alternate language that reduces the required checks from every 4 hours for 3 days, if plugged, to checks for only 2 consecutive 4-hour periods following restoration to free flow.  Another commenter also disagreed with the requirement to inspect every 4 hours for 3 days, even if the problem is corrected earlier.
      Response:  Based on the comments received the final rule has been written to provide other options to demonstrate free-flowing lime.  In addition to the option to perform visual checks to verify free-flowing lime, the owner/operator may use devices such as load cells to demonstrate this via weight changes in lime feed bins, use pressure sensors in pneumatic conveying systems to distinguish low or "no flow" conditions, continuously monitor lime feed rate, use an HCl monitoring device at the fabric filter outlet, or another method subject to approval by the permitting authority.

      4.1.6  Comment:  One commenter requested that lime feeder inspection requirements and corrective action requirements demonstrate compliance and that discovery and correction of a blockage or feeder setting drift not be an automatic violation.  The commenter suggested that the rule be rewritten to require corrective action when necessary and not to make blockage or feeder setting drift a violation.
      Response:  As noted in the response to the previous comment, the final rule provides additional options for monitoring the lime system to maintain free-flowing lime.  One of those options, the HCl monitor, provides a direct indication of continued effective operation of the control system which is the desired goal of any monitoring option selected.  Other options that detect lime feeder blockages are not direct and immediate performance indicators, so the time until remedied is a critical variable.  For this reason, EPA requires maintenance of free flowing lime in the feed hopper or silo at all times.  Blockages that occur as a result of equipment breakage or failure would potentially fall under the malfunction provision, and if determined to be a malfunction, would be covered by the SSM plan and would not be a violation, if corrected in accordance with the SSM plan.  However, continued and frequent blockages indicate a system design and operating problem rather than a malfunction.


      4.1.7  Comment:  Several commenters objected to the proposed regulatory requirements for scrap inspection programs.  They stated that the requirements are too onerous, expensive, complex and overly prescriptive, and further, some provisions are not technically feasible or cannot be reasonably met.  Three of the commenters suggested that the broadly stated scrap inspection requirements provided in the preamble to the proposed rule could be acceptable, and that approval of site-specific plans by the permitting authority would be a more acceptable requirement.  Two commenters also stated that the scrap should not have to be inspected if the necessary control systems are in place.  According to these commenters, inspection is only needed for control by work practices or pollution prevention.  They stated that the EPA needs to be clearer as to which sources are covered; the preamble says all furnaces and the rule says uncontrolled group 1 furnaces.
      Response:  The scrap inspection program requirements apply only to those facilities that elect to use such a program as a monitoring technique to ensure the oil and coatings content of scrap charged to a group 1 furnace stays below levels established during the performance tests.  Such a program could apply to facilities that have only uncontrolled group 1 furnaces, or facilities that have both add-on controlled and uncontrolled group 1 furnaces.
      As a result of the numerous comments received regarding the scrap inspection program elements, the EPA has modified the proposed rule.  The detailed requirements contained in the proposed rule have been deleted and the general scrap inspection guidelines provided in the proposal preamble have been adopted.  This change will provide more flexibility to owner/operators to tailor the program to specific conditions for their facility.  The scrap inspection program, if selected by the facility, will become part of the site-specific OM&M plan.  The specific inspection program elements, which must be consistent with guidance in the rule, will be approvable by the permitting authority as part of the site-specific OM&M plan and will be enforceable under the facility's permit.

      4.1.8 Comment:  One commenter stated that firms would likely need to hire additional staff to process paperwork, especially the monitoring data recorded every 15 minutes.  The commenter argued that these requirements have no corresponding benefit to the environment; requiring a report to document no exceedances is unnecessary when the absence of a report of exceedances should presume compliance. (Institute of Scrap Recyclers)
      
      Response:  The facilities' burden for processing and reporting results from  monitoring data were included in the impacts estimates reported with the proposal package for this rule.  Recording data every 15 minutes would require few if any additional staff.  Data logging equipment and software automate this process.  Such systems may be set up to sound warning devices when monitored levels approach or exceed the established operating limits.  The requirements to process and report results periodically ensures that some staff are assigned the responsibility to identify problems and take corrective actions when necessary, and that a corporate official is obligated to report the results of these efforts.  The benefit to the environment is the continuous prevention of HAP emissions in excess of established emission limits.

      4.1.9  Comment:  Two commenters wanted to establish parameters for control devices other than baghouses, lime baghouses and afterburners.  Wet scrubbers used in one facility illustrate such a need.  Such alternative monitoring plans could be handled on a site-specific basis subject to permit authority review. (IV-D-31, IV-D-8) Another commenter suggested provisions for and language permitting the use of alternate control devices with site-specific monitoring plans. (IV-D-28)
      Response:	The General Provisions provide for the use of alternatives to monitoring requirements or procedures upon approval by the Administrator (40 CFR 63.8(b)).  Monitoring parameters for scrubbers or other alternative control devices can be established under this provision.  Consistent with the General Provisions, the final rule has been modified to specify some procedural requirements for applying and obtaining approval of alternative monitoring programs. 

      4.1.10  Comment:  Two commenters requested that the requirement to calibrate lime injection rate, flux injection rate, and other monitoring instrumentation quarterly should be changed to "in accordance with manufacturer's specifications" throughout the rule.  The EPA requirement may conflict with the manufacturer's recommendation. (IV-D-32, IV-D-8)
      Response:  The final rule has been revised in the appropriate sections to require monitoring equipment calibration in accordance with the manufacturers' recommended schedule or if no schedule is specified at least once every 6-months.  It was not EPA's intention to have the calibration occur less frequently than that recommended by the manufacturers.

      4.1.11  Comment:  One commenter wants to restrict application of the Operation, Maintenance, and Monitoring (OM&M) plan to the control device, not the affected source, and the commenter also wants to exempt uncontrolled sources from the plan. (IV-D-8)
      
      Response:  Although the add-on air pollution control device is the means by which emission reductions are ensured for those affected sources or secondary aluminum processing unit- emission units vented to such a device, process operating conditions for the attached sources also affect the level of emissions achieved.  It is for this reason that EPA is requiring the establishment of operating limits for process parameters such as flux injection rate.  For uncontrolled emission units that are part of a secondary aluminum processing unit, the level of emissions is directly related to the type and amount of flux used as well as the level of contamination in the scrap fed, in the case of melting furnaces.  It is essential that the OM&M plan document the work practices and pollution prevention measures that will be used to maintain compliance and that these practices and measures then become an enforceable part of the permit.

      4.1.12  Comment:  According to two commenters, by the time the rule is finalized most affected sources will have Part 70 or 71 permits.  These two commenters wanted a requirement that the OM&M plans be available for review not that they be required to submit them as part of the Part 70 or 71 process.  The costs would be burdensome and were not included in the economic analysis. (IV-D-18, IV-D-20)
      Response:  The proposed rule required the OM&M plan to be submitted to the applicable permitting authority as part of the application for a Part 70 or 71 permit.  Based on the comments received and a further review of the proposed requirements the EPA is clarifying the requirements for the OM&M plan.  As the commenters mention, some facilities will already have their permits, so it is not feasible for them to submit the OM&M plan as part of the application process.  Also it was not EPA's intention to have the OM&M plan itself incorporated into the permit, such that it would be necessary to reopen the permit if changes to the OM&M plan are desired.  The permit can incorporate the OM&M plan by reference upon the plan approval by the permitting authority. The OM&M plan would be external to the permit so that a permit revision is not necessary when an OM&M plan change is desired.  For those plants already having permits when the final rule is promulgated, the OM&M plan will be incorporated by reference.

      4.1.13  Comment:  One commenter suggested the requirement for verification (of accuracy in feed measurements) every three months should worded to be either once per quarter, or else within three months of the last verification.  This commenter also suggested changing the requirement to verify weight calibration every three months to a requirement to check it and recalibrate if the error is found to be greater than 1%. (IV-D-20)
      Response:	The EPA has considered this, and other comments related to measurement accuracy for feed/charge.  The final rule has been changed to require verification of the calibration at the frequency recommended by the manufacturer, or if no calibration frequency is specified, at least once every six months.

      4.1.14  Comment:  One commenter wanted solid flux excluded from charge weight. (IV-D-8)
      
      Response:  Most of the test data that were used as the basis for selection of emission limits for the affected sources and emission units were computed on the basis of exclusion of solid fluxes from the feed/charge weight. Therefore, this definition is modified in the final rule to exclude solid fluxes, except for dross-only furnaces.  The EPA has determined that it is appropriate to continue to include solid fluxes as part of the feed/charge for dross-only furnaces.  

      4.1.15  Comment:  One commenter supported the monitoring options provided for fabric filters. (IV-D-11)  Another commenter wanted Method 9 or daily visual inspection added to the monitoring options available for fabric filters (Method 9 is an option for scrap shredders, but not other affected sources or emission units). (IV-D-23)
      Response:	 The EPA proposed the daily use of EPA Method 9 as an alternative means of monitoring compliance for fabric filters and lime-injected fabric filters serving scrap shredders due to the intermittent nature of shredder operations.  The final rule retains the requirement for all affected sources or emission units with fabric filters or lime-injected fabric filters (other than scrap shredders) to install, operate, and maintain a bag leak detector system, or install and operate a continuous opacity monitoring system to provide evidence of continuing compliance. 

      4.1.16  Comment:  One commenter suggested alternative wording for calibration of the bag leak detection system at 63.1510(f)(1)(ix) of the proposed rule to clarify the intent. (IV-D-17)  Another commenter supported the option to use bag leak detectors that operate on principles other than triboelectric.  This commenter recommended that recording milliamp current be allowed as an alternative to output voltage.(IV-D-11)  One commenter requested clarification that non-triboelectric bag leak detectors can be used.  Also the commenter questioned whether the EPA guidance is intended to apply only to triboelectric detectors.  The commenter stated that a clarification is needed as to whether record keeping and response tests specified in the guidance are required by the rule and whether recommended and suggested procedures are required by the rule.  Details of what is intended to be mandatory and enforceable should be provided. (IV-D-28)
      Response:  The guidance referenced in the proposed and final rule is intended to apply only to triboelectric detectors.  The record keeping and response tests specified in the guidance are required by the rule.  The recommended procedures and suggested procedures are also required by the rule.  Some flexibility is provided in the guidance, so an owner/operator can select an option and follow that option.
      However, bag leak detection systems based on triboelectric principle are not the only type that can be used.  For other types of bag leak detection systems, the manufacturer's recommendations should be followed.
      
      With respect to the requirement to monitor output voltage from the system that was in the proposed rule at 63.1510(f)(1)(v), the intent was to monitor the electrical signal from the triboelectric sensor that indicates relative PM loadings.  If the bag leak detection system installed uses current or some other type signal instead of voltage to provide that relative PM loading, then current or the other signal should be monitored instead of voltage.
 
      4.1.17  Comment:  Two commenters stated that PS-2 was established for SO2 and NOx monitoring systems, not temperature monitoring and PS-2 is therefore not applicable.  One commenter gave six cites for requirements that do not apply or work, stating that existing equipment is not suitable for temperature traversing because it is not equipped with ports for that purpose.  The other commenter also stated that adding sampling ports would decrease structural integrity of the system and potentially expose workers to high temperatures. The latter commenter also requested EPA to modify the temperature monitoring device provisions that call for use of a NIST reference method (63.1510(h)(2)(v) of the proposed rule) to clarify that what is being required is the use of a calibration standard, not a reference test method such as those in Appendix A, part 60. (IV-D-23, IV-D-17)  Another commenter stated that the relative accuracy requirements are unnecessary for temperature monitoring. (IV-D-11)
      Response:	 The EPA agrees that Performance Specification-2 is not appropriate for application to temperature measurement. Based on the other comments received, the requirements for the temperature monitoring device for affected sources using an afterburner or fabric filter have been revised.  The final rule requires that the recorder response range include zero and 1.5 times the average temperatures established as the monitoring parameter and that the calibration method must be a National Institute of Standards and Technology calibrated reference thermocouple-potentiometer system or alternate reference, subject to approval by the Administrator.  

      4.1.18  Comment:  One commenter stated that the proposed rule included two sections 1510(g)(2). (IV-D-20)
      Response:	 This assertion is incorrect, there was only one section 1510(g)(2) in the proposed rule.

      4.1.19  Comment:  Two commenters supported options for monitoring injection weight, and maintaining and recording injection rate. (IV-D-23, IV-D-11)
      Response:	The EPA acknowledges these comments in support of the rule provisions and notes that the final rule also includes a provision for obtaining approval of alternate monitoring procedures.


      4.1.20  Comment:  One commenter stated that inspection of lime feed systems once per 8 hour shift and more frequently when found to be plugged may be difficult, arguing that visual inspection at silo and bin tops is dangerous.  The commenter suggested alternate language that reduces the required checks every 4 hours for 3 days if plugged to checks for only 2 consecutive 4-hour periods when restored to free flow.  Another commenter also disagreed with the requirement to inspect every 4 hours for 3 days, even if the problem is corrected earlier (IV-D-23, IV-D-11)
      Response:  Based on the comments received, the final rule has been modified to provide other options to demonstrate free-flowing lime.  In addition to the option to perform visual checks to verify free-flowing lime, the owner/operator may use devices such as load cells to demonstrate this via weight changes in lime feed bins, use pressure sensors in pneumatic conveying systems to distinguish low or "no flow" conditions, continuously monitor lime feed rate, use an HCl monitoring device at the fabric filter outlet, or another method subject to approval by the permitting authority.  
      
      4.1.21  Comment:  One commenter requested that lime feeder inspection requirements and corrective action requirements demonstrate compliance and that discovery and correction of a blockage or feeder setting drift not be an automatic violation.  The commenter suggested that the rule be rewritten to require corrective action when necessary, not to make blockage or feeder setting drift a violation.
(IV-D-28)
      Response: As noted in the response to the previous comment, the final rule provides additional options for monitoring the lime system to maintain free-flowing lime.  One of those options, the HCl monitor, provides a direct indication of continued effective operation of the control system which is the desired goal of any monitoring option selected.  Other options that detect lime feeder blockages are not direct and immediate performance indicators, so the time until remedied is a critical variable.  For this reason EPA requires maintenance of free flowing lime in the feed hopper or silo at all times.  Blockages that occur as a result of equipment breakage or failure would potentially fall under the malfunction provision, and if determined to be a malfunction, would be covered by the Startup, Shutdown, and Malfunction (SSM) plan and not a violation, if corrected in accordance with the SSM plan.  However, continued and frequent blockage indicates a system design or operating problem rather than a malfunction.
      
      
      4.1.22  Comment:  Several commenters suggested alternatives to the manual inspection of lime feed hoppers.  One commenter states that a facility has installed devices that continuously monitor lime feed rate. (IV-D-17) Another commenter states that manual inspection should be changed to allow computerized weight loss monitoring and demonstration of reliability of alternatives as part of development of the OM&M plan. (IV-D-32)   A third commenter recommends that lime feed pneumatic conveyor blower pressure be monitored to determine if lime is free flowing and that a sensor be installed that activates an alarm if no lime feed is detected. (IV-D-11)
      Response:	 Based on the comments received, the final rule has been modified to provide other options to demonstrate free-flowing lime.  In addition to the option to perform visual checks to verify free-flowing lime, the owner/operator may use devices such as load cells to demonstrate this via weight changes in lime feed bins, use pressure sensors in pneumatic conveying systems to distinguish low or "no flow" conditions, continuously monitor lime feed rate, use an HCl monitoring device at the fabric filter outlet, or another method subject to approval by the permitting authority.  

      4.1.23  Comment:  One commenter agreed with need to monitor chlorine usage on in-line fluxers and group 1 furnaces without add-on controls and to maintain rate at or below the average measured during the performance test. (IV-D-11)
      Response:  The EPA acknowledges support for those provisions of the rule.		

      4.1.24  Comment:  Several commenters claimed continuous flux measurements are not needed since much of the time during furnace cycles there is no fluxing.  The requirement for continuous measurement and recording 15-minute blocks is an undue burden when applied across entire industry.  Having data in 15-minute blocks does not necessarily yield accurate results.  Many facilities only flux 10 to 30 minutes out of a 5-7 hour cycle and so would be recording zero values for 97% of the operating time.  Likewise solid flux additions and other fluxing may occur in batches.  The commenters stated that the requirement to continuously measure and record is unreasonable and provides no environmental benefit. (IV-D-23, IV-D-22, IV-D-17, IV-D-11, IV-D-32) Another commenter stated that they will have to measure chlorine before dilution to meet this requirement, and thus pipe pure chlorine, rather than a 12% mixture, through the plant to each furnace which they view as a safety problem. (IV-D-8)
      Response:	 The EPA has reviewed the flux injection rate monitoring requirements and, based on comments received, modified the requirements in the final rule.  As noted by the commenters many furnaces do inject flux for only short time periods within a furnace cycle and others add solid flux in batches.  The monitoring provisions have been modified to require recording flux injection rate in 15-minute blocks only during periods when fluxing is occurring.
      
      The EPA's intent is to achieve high accuracy in the measurement of reactive flux usage. The latter commenter mixes flux gases at a central location prior to distribution through the plant to eliminate the presence of pure chlorine gas in work areas.  The weight of reactive flux component in the diluted mixture must be measured to 1 percent accuracy, and the quantity of mixture delivered to each emission unit must also be measured to yield an accurate quantitative measurement of reactive component used per emission unit during the monitoring period.  The final rule does not preclude the use of central mixing stations for reactive flux gases provided that the amount of reactive flux component delivered to each emission unit is determined within 1 percent accuracy.  As described in the response to Comment 4.1.3,  the final rule does provide for the permitting authority to approve alternative monitoring accuracy requirements on a site-specific basis when 1 percent may not be workable.

      4.1.25  Comment:  Several commenters suggested alternatives to the flux monitoring requirements.  One commenter suggested measuring flux cylinder weights before and after casting operations for monitoring for in-line fluxers.(IV-D-23)  Two commenters stated the regulation should provide an option for a site-specific monitoring scheme better suited to each operation. (IV-D-22, IV-D-8)   Another commenter urged EPA to provide monitoring flexibility in the form of alternative monitoring techniques. (IV-D-32)   One commenter suggested an HCl monitor in the waste gas stream to automatically control the baghouse lime injection system should be allowed in lieu of monitoring chlorine flux input. (IV-D-11)   EPA should allow proper operation of pollution control systems to replace continuous weighing of flux salts. (IV-D-11) Other commenters requested SAPU bubble monitoring of flux (no monitoring for individual emission units) (IV-D-1, IV-D-32).  One commenter also wanted to drop the fluxing schedule requirements and substitute a monthly mass input to the SAPU. (IV-D-8)
      Response:	 The EPA recognizes the wide variety of fluxing practices among facilities and within facilities for different affected sources or emission units.  The purpose of requiring monitoring of fluxing rates and flux consumption is to provide continuous verification that the practices (as monitored by selected parameter values) do not exceed those practices in use during the performance tests (that could cause emissions to exceed the levels achieved in the compliance demonstration).  The alternate monitoring practice, used by one commenter, of operating a HCl monitor to control the amount of lime injected to the fabric filter control system is an acceptable alternative to flux injection rate monitoring for controlled sources.  This is not an option for uncontrolled affected sources or emission units.  The final rule has been modified to include the option for a HCl monitoring system that controls the lime injection feed rate.
      
      The use of bubble monitoring of flux is not a satisfactory alternative for SAPU that include both controlled and uncontrolled emission units because it is necessary to distinguish between consumption in controlled and uncontrolled units.  Even for controlled units that are served by different control systems, it is necessary to monitor rates and consumption of flux in relation to rates and consumption of lime so as to monitor the performance of the separate control systems.  For systems in which all emission units are connected to the same control system, bubbled flux monitoring across all units is acceptable provided HCl concentrations are monitored and demonstrate that the allowable HCl emission rate is not exceeded.  	Extending the flux monitoring period to a monthly basis permits daily or cycle-to-cycle variations in the relative usage of lime and flux that will not necessarily coincide.  This will lead to increased emissions as compared to those measured during performance tests and thus is not acceptable.  
      The final rule has been modified to allow bubbled flux monitoring for all emission units connected to the same HCl emissions control system if monitoring control system outlet waste gas HCl concentrations to demonstrate that the allowable HCl emission rate is not exceeded.  The final rule retains the requirements for monitoring over the operating cycle or period of the performance test.

      4.1.26  Comment:  One commenter disagreed with the record keeping and compliance certification requirements, stating they are unnecessary.  In-line refining of molten aluminum for product quality purposes without use of reactive flux should not be subject to rule. (IV-D-17)  Another commenter stated that the requirement to identify fluxing agents on a cycle-by-cycle basis is unnecessary since in-line fluxers not using reactive fluxes use the same flux every day. The certification requirement of 63.1510(m)(2) of the proposed rule is sufficient, (m)(1) should be deleted. (IV-D-32) A third commenter wanted a monitoring exemption for in-line fluxers that do not use reactive flux. (IV-D-31)
      Response:  The EPA agrees that in-line fluxers that do not use reactive flux can be subject to less rigorous record keeping and compliance certification requirements than those using reactive fluxes.  The final rule has been revised to require that owner/operators of in-line fluxers that don't use reactive fluxes keep a record of the flux materials used (does not have to be on a cycle-by-cycle basis) and certify every six months that only those materials were processed in the fluxer.


      4.1.27  Comment:  One commenter suggested adding the word sidewell between "Group 1" and "furnace" in 63.1510(n) of the proposed rule. (IV-D-32)   Another commenter asked for clarifications of which requirements in 1510(n), only apply to sidewell furnaces.   In 1510(n)(1), the latter commenter suggested that the operator record whether or not the level of molten metal was above the top of the arch and adding a requirement for corrective action for the situation if the level is not. (IV-D-20)
      Response:	 At present all the provisions of 63.1510(n) deal with sidewell furnaces, so the commenter's suggested addition to clarify this is appropriate.  The final rule has been changed to clarify this point.  The rule as proposed does have a corrective action requirement in 63.1506(q), however, it may not be clear that the operating requirement for metal level in sidewell furnaces is covered.  The final rule states that maintenance of the metal level above the top of the arch is an operating requirement.  Failure to maintain that level is a violation.
      
      4.1.28  Comment:  One commenter interpreted 63.1510(o) of the proposed rule to allow the use of COMs as approved alternatives to scrap inspection.  The requirement in 63.1510(o)(ii) for a relationship between parameters and emissions should allow for the use of minimum or maximum values below which or above which compliance is assured.  For example, if flux usage is less than the emission limit that is sufficient to demonstrate compliance. (IV-D-32)  Another commenter suggested that one worst case scenario (rather than multiple parametric tests) as required in 63.1510(o)(1)(ii) (as well as in 63.1505(l) for SAPU's) is adequate to establish operating limits for SAPUs. (IV-D-8)
      Response:  The commenter's interpretation of 63.1510(o) relative to Group 1 furnaces without add-on control devices is not correct.  A COM cannot be used as an alternative to scrap inspection.  COMs do not provide an indication of D/F, HCl, or Cl2 emissions.
      Testing of maximum or worst case flux addition values to establish operating limits for SAPU is acceptable.  As suggested by the commenter, if flux usage is monitored and it is less than the emission limit for the affected source or emission unit, this would demonstrate compliance.

      4.1.29  Comment:  One commenter wanted the deadline for submittal of the site-specific monitoring plan moved to six months before the compliance date, as opposed to six months before the performance test.(IV-D-31)
      Response:	 After consideration the EPA agrees that is appropriate to change the deadline for submittal of the site-specific monitoring plan to at least six months prior to the compliance date.  That change has been made in the final rule. 
 

      4.1.30  Comment:  Several commenters objected to the proposed regulatory requirements for scrap inspection programs:  	1) Several stated that the requirements are too onerous, expensive, complex and overly prescriptive, and further, as provided in some examples, some provisions are not technically feasible or not reasonably met. (IV-D-23, IV-D-22 IV-D-17, IV-D-32, IV-D-18, IV-D-20)
      2) One commenter stated that the procedures for inspector training, certification and program validation and verification will place large burden on one of their facilities to maintain on-site inspection program.  The commenter wanted flexibility of alternate inspector training and a less stringent inspection program with review and approval by the permit authority. (IV-D-11)
      3) Three of the commenters suggested that the broadly stated scrap inspection requirements provided in the preamble could be acceptable and that approval of site-specific plans by the permitting authority would be a more acceptable requirement.
      4) Two commenters also stated that scrap should not have to be inspected if necessary controls are in place.  Inspection is only needed for control by work practices or pollution prevention.  They stated that EPA needs to be clearer as to which sources are covered, the preamble says all furnaces and the rule says uncontrolled group 1 furnaces. (IV-D-18, IV-D-20)
      Response:	The requirement for a scrap inspection program is only intended to apply to those facilities that have group 1 furnaces without add-on control devices (including those that are part of a SAPU).  For those facilities that have both add-on controlled and uncontrolled group 1 furnaces, the scrap inspection requirements would apply only to the materials charged to the uncontrolled units.
      As a result of the numerous comments received regarding the scrap inspection program elements, the EPA has modified the final rule to delete the detailed requirements and adopt the general scrap inspection guidance as presented in the proposal preamble.  The scrap inspection program, if selected by the facility, will remain part of the site-specific monitoring plan.  The specific inspection program elements (which must be consistent with guidance in the regulation) will be approvable by the permitting authority as part of the site-specific monitoring plan and will be enforceable under the permit.    

      4.1.31  Comment:  One commenter stated that the rule should clarify that it is the responsibility of the source to demonstrate that scrap inspection program meets intent of rule, including use of worst case emission testing, verification sampling of scrap, and enforceable contractual arrangements with offsite entities.  The commenter cited other inspection programs as precedents. (IV-D-9)
      Response:	See the response above. 

      4.1.32  Comment: One commenter supported an alternative scrap contamination determination by calculation procedure. (IV-D-11) Another commenter supported simplified scrap inspection procedures for furnaces dedicated to one type of scrap. (IV-D-8)
      
      Response: In addition to the changes to scrap inspection procedures described in the previous responses, the final rule contains a calculation method for scrap contamination monitoring for group 1 furnaces without add-on control devices that may be appropriate for a furnace dedicated to one type of scrap.
      
      4.1.33  Comment: In 63.1510(p) of the proposed rule, one commenter wants the term "acceptable scrap" to be defined in the scrap inspection program (as scrap that will not cause excess emissions, for example). (IV-D-20)
      Response:	 As noted in the responses to previous comments, the EPA has substantially revised the requirements for scrap inspection programs.  The EPA has clarified its intent in the final rule by deleting the term "acceptable scrap." which appeared in several places.  The intent in using the term was to describe scrap that meets site-specific acceptance criteria under an approved scrap inspection program and, that when used, will not cause the affected source to exceed the emission limits.

      4.1.34  Comment:  Section 63.1510(q)(2) of the proposed rule dealing with scrap contamination level by calculation, limits charges to having the same composition as in the performance test which a commenter claimed is impossible.  The commenter claimed they should be allowed to use a reasonable worst case in the performance test and stay below that oil and coating content for ongoing compliance. (IV-D-32)
            Response:  The language in the final rule has been modified to explain that "same composition" means oil and coatings content equal to or less than that of the scrap used during the successful performance test.

      4.1.35  Comment:  One commenter stated that the scrap inspection requirements do not provide for the use of a limited amount of coated scrap that could be charged under the emission limit provisions for a SAPU.  That scrap for an individual furnace might have the potential for emissions that exceed an individual furnace limit, but be acceptable when the unit is part of the SAPU. (IV-D-11)
      Response:  Under the scrap inspection requirements and site-specific monitoring plan requirements for a SAPU, the testing program can be used to establish an emission rate for the furnace operating conditions associated with the consumption of a limited amount of coated scrap.  When an emission unit is operated in such a manner, the emission rate associated with that condition would be the rate used to compute the 24-hour emissions for that unit as part of the SAPU emissions calculations for each pollutant.  However, the furnace charge cannot be more contaminated than what was used in the performance test for that furnace.


      4.1.36  Comment:  One commenter states that the dates and provisions of the Site-Specific Test plan (SST), Secondary Aluminum Processing Unit (SAPU) emission plan, Operation, Maintenance and Monitoring plan (OM&M)  Startup, Shutdown, and Malfunction (SSM) plan are too complex and not consistent with each other.  The commenter requests that EPA combine the SAPU emission plan and OM&M into one plan, and fix the dates references. (IV-D-23)  Another commenter requests the requirements be reviewed and simplified to eliminate duplication and reduce the burden to industry. (IV-D-22)  A third commenter suggests a single compliance plan instead of a site specific test plan, a notification of compliance, and an OM&M plan. (IV-D-8)  
      Response:	In response to numerous comments related to the submission dates for each of the plans and some apparent confusion regarding the purpose of each, the final rule has been revised to clarify the submission dates and reduce the complexity of the requirements.  All the dates by which the various plans must be submitted are now related to the compliance date which is defined in the final rule as not later than 3 years after the promulgation date for existing units, or upon startup for new units.
      The Startup, Shutdown, and Malfunction (SSM) plan requirements are provided in the General Provisions (40CFR 63.6(e)(3) and have not been changed.  The plan must be developed by the owner/operator by the source's compliance date for the relevant standard.
      The Operating, Maintenance, and Monitoring (OM&M) plan, which includes the site-specific monitoring plan for uncontrolled units, if any are present at the facility, must be submitted no later than 6 months prior to the compliance date or the date by which the facility intends to comply, whichever is sooner.  Those facilities with a SAPU are also required to submit the site-specific SAPU emission plan showing the calculation of the SAPU emission limits, explaining how SAPU emission limits will be achieved, presenting the test plan for performance tests, and other required elements no later than 6 months prior to the compliance date or the date by which the facility intends to comply, whichever is sooner.

  	4.1.37  Comment:  One commenter is concerned with the stipulations of 63.1505(k) that require complying with all requirements of the SAPU emission plan plus all the requirements for each individual emission unit in the SAPU.  The commenter wants to negotiate with the permitting agency for collection of needed data for the site-specific monitoring plan and the SAPU emissions plan to demonstrate compliance.  The commenter does not want to collect monitoring data for each and every emission unit within the SAPU, especially for multiple units connected to a common control device.  The commenter suggests alternate language for 63.1505(k)(5) that requires compliance with the requirements of an approved site-specific SAPU emission plan and, if applicable, a site-specific monitoring plan (under the OM&M plan) without having to do it for each individual emission unit. (IV-D-32)

      Response:	 To simplify the compliance process, the requirement for a separate site-specific emissions plan contained in the proposed rule at 63.1505(k) has been removed from the final rule.  The elements of data and information required for submittal to the permitting authority in this plan under the proposed rule have been incorporated into the OM&M plan in the final rule as described at 63.1510(s).  Across the industry affected by these rules there will be significant variations between facilities in terms of combinations of controlled and uncontrolled emission units and control systems.  In view of this fact, the commenter's request to negotiate with the permitting agency for collection of data for the site-specific monitoring plan (and previous emission plan data) under the OM&M plan is reasonable and allowed under the final rule.  The Agency has revised the final rule and provided some additional guidance to assist the owner/operator and permitting authorities in this matter.
      For controlled and uncontrolled emission units it is necessary to monitor the individual units with respect to charge weight and flux consumption.  The charge weight is essential to the calculation of allowable emissions for a SAPU, and the flux usage must be maintained at or below the values at which the individual units were tested when operating at the defined worst case conditions during the compliance test.  The need for these data is present even if emission units are connected to a common control device.  Just as for a system that controls an individual affected source or emission unit, the lime injection in a system that provides cocontrol must be coordinated with emission unit operations to maintain sufficient lime availability to remove HCl, Cl2, and HF emissions.  Solid reactive flux usage must also be considered in regard to maintaining adequate lime availability.

      4.1.38  Comment:  One commenter wanted clarification that SSM plans are not part of the operating permit. They requested a 30 day deadline for plan approval by EPA, with unreviewed plans considered approved by default. (IV-D-8)
      Response:	The General Provisions (40CFR 63.6(e)(3) direct that the SSM plan be incorporated by reference into the source's Title V permit.  With regard to the issue of approval within 30 days, owner/operators of potentially affected existing sources will have up to 3 years from the promulgation date during which they can prepare and submit the SSM plan for approval. They need to work with the permitting authority to develop a workable plan sufficiently far in advance that approval within 30 days is not necessary.  Unreviewed plans cannot be considered approved by default.

      4.1.39  Comment:  Two commenters stated that the SSM plan should apply only to units with control devices as is the case in the Printing and Publishing MACT. (IV-D-23, IV-D-8)
      
      Response:	 While the add-on air pollution control devices play the primary role in emission control, variation in the process operations generating the waste gas stream and pollutants affect the emissions as well.  Further, these standards anticipate the presence of emission units using pollution prevention approaches as part of secondary aluminum processing units in many facilities.  These emission units and the associated work practices and pollution prevention measures are the means by which emissions are limited in the SAPU.  For these reasons, and consistent with the General Provisions, all affected sources and emission units must be included in the SSM plan.

      4.1.40  Comment:  One commenter suggested that EPA allow facilities to establish alternative site-specific OM&M plans subject to permit authority review and approval.  These alternative plans would include monitoring provisions different from those in the proposed rule, and reflect changes needed to deal with difficulties specific facilities would encounter in adopting the EPA proposed monitoring provisions. (IV-D-11)
      Response:	 All OM&M plans submitted will be site-specific by definition.  As noted in comment responses above, changes have been made in the final rule to provide more flexibility in monitoring the quantity and accuracy of feed/charge weight, flux usage, lime addition, and scrap inspection.  If an owner/operator wants to use alternative monitoring schemes different than those described as acceptable in the final rule, the procedure to gain approval is to submit a request and obtain approval of the alternative scheme from the Administrator.  The final rule provides the requirements for obtaining approval in 63.1510(w).
      
      4.1.41  Comment:  Two commenters did not want the OM&M and SAPU emissions plan incorporated in Title V permit because it will make all operating, maintenance, and monitoring procedures enforceable permit conditions.  The commenters stated that it will also be too difficult to make minor changes to their operations without amending the permit which could cause delays for weeks or months. (IV-D-23, IV-D-8)  Another commenter also disagreed with the requirement to have the OM&M plan be part of the Part 70 and 71 permit.  They stated that the site-specific monitoring plan should be developed for and approved by the applicable permitting authority. (IV-D-17)
      
      Response:  The proposed rule required the OM&M plan to be submitted to the applicable permitting authority as part of the application for a Part 70 or 71 permit.  Based on the comments received and a further review of the proposed requirements, the EPA is clarifying the requirements for the OM&M plan.  It was not EPA's intention to have the OM&M plan itself be incorporated into the permit, such that it would be necessary to reopen the permit if changes to the OM&M plan are desired.  The OM&M plan is to be incorporated into the permit by reference so that it is not necessary to revise the permit when a change in the OM&M plan is desired.  For those facilities that already have a Title V permit when the final rule is promulgated, the OM&M plan will be incorporated by reference.  The OM&M plan requirements become enforceable under the permit upon approval by the permitting authority.
      If an owner/operator wants to revise the ranges or limits established for the operating standards, they must:
      (1)  Submit a request to and obtain approval from the permitting authority to conduct a performance test to revise the operating requirement ranges or limits.
      (2)  After receiving approval from the permitting authority, conduct a performance test to demonstrate that compliance with the emissions standards can be achieved at the revised operating conditions.
      (3)  Submit the performance test results and the revised OM&M plan to the permitting authority for approval.
      (4)  Pending Administrator approval of the revised OM&M plan, the owner/operator must comply with the provisions of the approved plan.
      If the owner/operator is revising an aspect of the plan that does not require an additional performance test, for example, maintenance procedures, they must submit only a final plan to the permitting authority for approval.  Pending the Administrator's approval of the changes, they must comply with the provisions of their approved plan.
      
      4.1.42  Comment:  Two commenters agreed with EPA and support no monitoring or performance testing provisions for group 2 furnaces. (IV-D-23, IV-D-32)  One of the commenters urged EPA to include external preconsumer runaround scrap in the definition of clean charge.  The commenter claimed this material, often returned from their customers who do not have melting capability, is as clean and predictable as internal runaround. (IV-D-32)
      Response:  The EPA acknowledges the support for the decision not to require monitoring and performance testing for group 2 furnaces.  For a discussion of EPA's consideration of a revision to the definition of "clean charge," see the response to comment 2.1.

      4.1.43  Comment:  Identification of all charge materials in 63.1510(r)(1) of the proposed rule is unnecessary according to two commenters.  Some group 2 melters are used in similar service day after day, and cycle after cycle.  Only monitoring sufficient to supply the certification in 63.1510(r)2 of the proposed rule should be required. (IV-D-32, IV-D-8)
      
      Response:  The EPA agrees that in such situations it is not necessary to keep cycle-by-cycle records for Group 2 furnaces.  The owner/operator must record all the materials charged to the furnace that make it a Group 2 furnace.  Then, the six-month certification of compliance must contain the information specified in 63.1516(b)(2)(v) of the proposed rule and certify that only the materials they recorded as charged to the furnace were used during that 6-month period.

      4.1.44  Comment:  Another commenter opposed the requirement to record charge and flux materials, unless those materials deviate from the OM&M plan.(IV-D-31)
      Response:  Records of charge and flux materials used in the SAPU are necessary to demonstrate continuous compliance with the emission and operating limits.  Such monitoring observations that show deviations from the ranges or limits established during the initial compliance or performance test are violations of the standard unless caused by events covered by the SSM plan.  

      4.1.45  Comment:  In 63.1510(s)(2) of the proposed rule referring to calculation of 3-day, 24-hour rolling average emissions, one commenter stated the words "most recent" should be added when referring to the performance test because tests are conducted every 5 years. (IV-D-20)
      Response:	The commenter was referring to using the results from the most recent performance test in the calculations to monitor compliance.  However, the emission factors established for each affected source or emission unit that is part of a SAPU can only be changed by revising the site-specific OM&M plan (the portion required by 63.1510(s)) per the procedures described in a comment response above under Title V permit comments.  The most recent performance test may or may not be the basis for the currently approved site-specific OM&M plan incorporated in the permit.  Only if the most recent performance test is the basis for the approved OM&M plan can those results be used to calculate the 3-day, 24-hour rolling average emissions under 63.1510(t).

      4.1.46  Comment:  Another commenter wants the instructions for calculating and recording the average emissions based on up to a 30-day average and a 30-day SAPU compliance period. (IV-D-31)
      Response:  Data used to establish the emission limits and operating requirements were based on source tests conducted over relatively short durations.  The test periods required for initial compliance determinations are based on full process operating cycles or 24-hour periods, whichever is less.  For this reason, the final rule retains the requirement to demonstrate compliance on a 3-day 24 hour rolling average basis.  Sufficient data to establish lower 30-day emission limits and monitoring parameters are not available.


      4.1.47  Comment:  Several commenters wanted an alternative to, or believe that EPA did not intend to require, multiple testing over different furnace operating (range of charge/flux) conditions.  One commenter wanted to test under worst case scenario rather than across range of variables (uncontrolled units) as required in 63.1510(o) and 63.1505(l) of the proposed rule.  The commenter gave examples of situations where it would be appropriate and estimates of associated cost savings. The commenter suggested language allowing a permitting authority to approve a test plan if judged to contain sufficient provisions to ensure compliance. (IV-D-23)   Another commenter wanted an alternative to correlations between measured emissions and parameters to be monitored.  For example a minimum or maximum parameter value under which emissions may be less than the limit, i.e., flux usage is less than the limit therefore the limit will not be exceeded. (IV-D-32) Several other commenters stated that a defined worst case condition test will often be adequate to set upper limits on flux use and charge material.  Such an approach should be allowed if approved by the permitting authority. (IV-D-17, IV-D-22, IV-D-11)
      Response:	The purpose of requiring tests over a range of different furnace operating conditions is to show that the selected monitoring parameters are valid indicators of relative emission levels over the range of furnace operating conditions for uncontrolled furnaces.  Testing at maximum or worst case charge/flux addition values to establish operating limits for emission units in a SAPU is acceptable provided that the owner/operator can show the tested cases produce maximum emissions for all regulated pollutants.  However, the measured value of emissions for that maximum case is the level of emissions that must be used in the 24-hour average emission calculations for that unit within a SAPU.  Also, to be in compliance after the compliance date, the flux rate used in the each emission unit must be equal to or less than the amount used during the maximum or worst case test, and the charge materials must be of equal or lower oils and coatings content than the charge materials used during the maximum or worst case test.
      There are some cases in which it may not be possible to establish that a particular test represents a maximum or worst case scenario.  For example, it may not possible to show that a particular combination of oil and coating scrap contamination is worse than another combination without actually performing a test under both conditions. 
      As suggested by the commenter, if flux usage is monitored and it is less than the emission limit for the affected source or emission unit, this would demonstrate compliance with the HCl limit.

      4.1.48  Comment:  A commenter believes that the language of the proposed rule requires all operators of secondary aluminum processing units to develop a correlation between emissions and monitored parameters, and that it is not clear that this correlation is only required for uncontrolled Group 1 furnaces.  The commenter requests explicit language that parameter/emission correlations are necessary only when used to demonstrate compliance with the standard. (IV-D-28)
      
      Response:  The proposed rule at 63.1510(o)(1)(ii) required that the owner/operator of a SAPU show a relationship between emissions and monitoring parameters over the entire range of charge and flux materials processed in the furnace (emission unit).  Also, the proposed rule at 63.1505(l)(2)(vii) required that the site-specific secondary aluminum processing unit emissions plan include a correlation of measured emissions with the selected process or operating parameter to be monitored, as part of the application for an operating permit.  The final rule has been revised to eliminate the requirement for such a correlation.  The Agency has concluded that it will be sufficient to identify maximum or worst case charge and flux conditions under which tests are performed to establish the monitoring parameter values or ranges.  As noted above in a previous comment response, in some situations it may be necessary to perform tests at several conditions to establish which of the conditions are representative of maximum or worst case conditions for an emission unit. 	
      
      4.1.49  Comment:  One commenter supported a scaled down monitoring program with reduced monitoring requirements for those units operating at half (for example) the federal emission limitation to provide an incentive for applying controls and operating practices that exceed the MACT floor. (IV-D-22)
      Response:  Measurements that demonstrate operation at low emission rates are short-term indicators of performance.  Process operations vary with time and cannot be relied upon to remain constant.  For this reason the EPA has retained the basic monitoring requirements for all units regardless of the emission levels achieved in relation to the emission limits.

      4.1.50  Comment:  One commenter requested clarification of what is meant by "relative accuracy" in Section 63.1510 of the proposed rule.  When used in conjunction with continuous monitoring systems there are usually set protocols giving procedures and equations to determine relative accuracy. (IV-D-20)
      Response:  This comment refers to requirements in the proposed rule for temperature monitoring devices at fabric filter inlets.  That section has been revised in the final rule to delete the relative accuracy requirement.  The language has been revised to only require that 1) the recorder response range must include zero and 1.5 times the average temperature established according to the requirements in §63.1512(n) for affected sources or emission units using fabric filters for D/F or HCl control, and 2) the reference method must be a National Institute of Standards and Technology calibrated reference thermocouple-potentiometer system or alternate reference, subject to approval by the Administrator.

4.2  Performance Testing and Compliance


      4.2.1  Comment:  Commenters wanted the compliance date to be 3 years after promulgation rather than "on or after the date of the initial performance test."  (IV-D-8, IV-D-22, IV-D-23) They argued that:
      1) Carrying out performance tests prior to the end of 3 years is essential to completing the monumental job; but they do not like having to comply "on and after the date of the initial performance test" which could be the emission test program for the SAPU.
      2) The submittal deadlines for the OM&M plan, the SAPU emission plan, and the site-specific test plan are inconsistent with each other; they wanted EPA to remove all the interim compliance requirements to give the necessary flexibility to evaluate and agree with the permitting agency on compliance requirements before the 3-year deadline.
      Response:  A facility must be in compliance on and after the date of the initial performance test.  The date of that initial performance test, for existing sources, may be up to 3 years after the promulgation date of the standard.  For existing SAPUs, the initial performance test is considered to be the date of approval of the OM&M plan by the permitting authority. 
      In response to the comments regarding the inconsistent plan requirements and dates for submittal, the EPA has revised and clarified those requirements.  The final rule requires the owner/operator of a SAPU to perform tests that will define the operating modes of the controlled and uncontrolled emission units within the SAPU, and to define which parameters to monitor to demonstrate continuous compliance.  These same tests can be used to measure the emission rates from the affected sources and emission units for performance test purposes.  A site-specific test plan for this program must be submitted to the permitting authority for review and approval before the tests are conducted.  The plan must identify the parameters to be monitored during the tests, the test methods to be used, the units to be tested, and planned operating modes for each unit during the tests.  After the test plan has been approved by the permitting authority, the owner/operator is required to notify the Administrator of the test dates.
      
      The results from this test program, including the emission rates measured, values of parameters monitored, monitoring parameters selected by the owner/operator for compliance demonstration, and values of the parameters to be used as operating limits must be submitted to the permitting authority for review and approval.  As a result of the review, the permitting authority may request changes to selected monitoring parameters or values of the parameters used for compliance demonstration if it is determined the parameters or values do not provide an adequate means of demonstrating continuous compliance.  When all of these elements are approved by the permitting authority, the owner/operator prepares an OM&M plan using the approved monitoring scheme and submits the OM&M plan to the permitting authority for approval.  The compliance date is the approval date of the OM&M plan.  The approved OM&M plan will be included by reference in the operating permit.
      The latest date for an existing facility to achieve compliance is 3 years from the date the standard is promulgated.  The OM&M plan must be submitted to the permitting authority for approval no later than 6 months before the planned compliance date.  Given these conditions and lead times for preparing plans and conducting tests, it is clear that owner/operators must act expeditiously to develop test plans and execute the test programs. 
      Facilities that choose to comply by demonstrating that each emission unit in the SAPU meets the emission limit for that unit, and by monitoring the parameters as designated in the rule for each emission unit and control device, are also required to develop a test plan and notify the permitting authority of the test date(s).

      4.2.2  Comment:  Several commenters wanted EPA to allow testing one representative unit from a group of similar sources, that is one unit to represent similar furnaces or in-line fluxers, instead of having to test every emission unit.  One of the commenters stated that this practice should be allowed for either controlled or uncontrolled units.  Several commenters claim this approach is widely used under existing State permits and has been used by EPA in other NESHAPs.  Commenters claimed that it would significantly reduce costs, provide flexibility, and provide more cost-effective test programs. (IV-D-9, IV-D-17, IV-D-22, IV-D-23, IV-D-32, IV-D-33, IV-D-35)
      Response:  Based on the comments received, the EPA is modifying the testing requirements to allow representative or similar uncontrolled emission units that use like charge and flux materials to be tested, instead of requiring each unit to be tested.  Testing of representative or similar units may be used provided approval is obtained from the applicable permitting authority.  The representative unit selected for testing must be subject to the same work practices and be of the same design as those emission units it is representing for test purposes.  The representative unit must be tested under worst case conditions.  It is up to the owner/operator to define the worst case scenario(s) for review and approval by the permitting authority.  At least one of each different style unit must be tested.  Each add-on control device controlling emissions from an affected source or emission unit must be tested.


      4.2.3  Comment:  One commenter suggested exempting clean charge furnaces and in-line fluxers from the HCl performance test if chlorine usage is less that emission limit. (IV-D-8) Another commenter requested an alternative to performance testing to show that chlorine usage is less than the emission limit and specifically requested this as an alternative for uncontrolled in-line fluxers. (IV-D-31).  Two other commenters stated that Part II of Pharmaceutical Product NESHAP permitted owners to use calculational methodologies to determine emission rates for individual process vents.  (IV-D-22, IV-D-35)
      Response:	 The EPA agrees that it is reasonable to provide alternate procedures for owner/operators to demonstrate that HCl emissions are less than the emission limit for clean charge furnaces and uncontrolled in-line fluxers.  If the flux (weight equivalent as HCl) usage rate is less than the emission limit for an affected source this is a sufficient demonstration that the  source is in compliance with the standard.  The weight equivalent HCl must be determined for both gaseous and solid reactive fluxes used during the furnace cycle or performance test period. 
            
      4.2.4  Comment:  Submitting the site-specific test plan in accordance with general provision 63.7 would require the plan going to the "Administrator."  The commenter stated that they believe that it's EPA's intent for the plan to go to the applicable regulatory authority. (IV-D-17) Another commenter stated that the proposed requirement to submit performance test plans for approval goes beyond the General Provisions.  The commenter argued that this requirement will greatly increase the burden to permitting agencies and may interfere with schedules needed to complete tests before compliance deadline. (IV-D-32)
      Response:	 The general provisions state that before conducting a performance test the owner/operator of the affected source must develop and, if requested by the Administrator, submit the plan for approval.  Due to the complexity of testing for demonstration of SAPU compliance (facilities with combinations of controlled and uncontrolled emission units, multiple emission units controlled through common control devices, etc.) the Agency has decided that the test plans must be submitted for approval.  The General Provisions at 63.2 define Administrator to mean the Administrator of the U.S. Environmental Protection Agency, or his or her authorized representative (e.g., the State that has delegated authority to implement the provisions of this part).  In the case of authorized states, the plan would go to the permitting authority for approval.
      Facilities having Group 1 furnaces without add-on control devices must develop test data to support their OM&M plan.  Specifically they are required to develop a site-specific monitoring plan as part of the OM&M plan based in part on test data.  If facilities wish to use test data collected during development of the SAPU OM&M plan later as a compliance demonstration, a test plan should be prepared for approval prior to the OM&M plan testing activity.  The benefit to owner/operators in submitting the plan for approval is that they may avoid the necessity of conducting two separate sets of tests, one for OM&M development and one for compliance demonstration. 
            
            
      4.2.5  Comment:  One commenter stated that the requirement to operate all affected sources connected to a common stack during tests of that common stack is not feasible and not necessary.  The commenter claimed that even at the highest production rate, some facilities may not operate all affected sources simultaneously.  The commenter supports the EPA decision to allow a single test to characterize emissions from multiple sources ducted to a common stack.  Further, the commenter requested that the regulatory paragraph should be expanded to affirm that new emissions units may be ducted to existing control systems, if capacity is available or can be expanded to accommodate the new source. (IV-D-32)
      Response:	 The EPA understands that it may not be feasible to operate all emission units connected to a common stack simultaneously during a performance test.  The intent has been clarified in the final rule by stating that the performance test must be conducted at a maximum production rate scenario, i.e. the number of operating emission units or affected sources connected to the common control device or stack must represent a maximum or highest emission rate condition for the commonly controlled units or sources.
      For the reasons stated in the response to comment 1.6 in this document new emission units may not be cocontrolled with existing units.
      
      4.2.6  Comment:  One commenter opposed the mandatory retesting every 5 years and claimed that alternative evidence ensuring compliance should suffice.  That is, if the required controls are in place, the floor limits are literally guaranteed according to the preamble of the proposed rule. (IV-D-10)
      Response:	 In general, the presence of required controls is not sufficient to ensure compliance.  Proper operation and maintenance of the control systems are also necessary elements to ensure compliance.  Over a period of years changes may also occur in process operations that affect emissions and control device performance.  For these reasons the requirement to retest every 5 years is being retained in the final rule.

      4.2.7  Comment:  One commenter stated that it is not clear what is meant in 1511(e) of the proposed rule by "at the time of permit renewal."  If the source test results are meant to be used in the renewal application, the testing needs to happen far in advance.  Also the commenter observed that the permit renewal dates may be more than 5 years apart in some cases. (IV-D-20)
      Response:	 It is the EPA's intent that each new or existing affected source or emission unit be retested every 5th year after the initial performance test.  Given that this may not coincide with permit renewal, the final rule has been modified by deleting that phrase from the retesting requirement.
      
      4.2.8  Comment:  One commenter supported the choice of performance test methods designated by EPA. (IV-D-23)
      
      Response:	 The EPA acknowledges support of the test method selection.
      
      4.2.9  Comment:  One commenter requested additional language in the regulation stating that if the format of the emission limit (units, lb/ton) is the same for all units (units ducted to a common control device), compliance is demonstrated if the combined emission limit for all units is not exceeded. (IV-D-23) 
      Response:	 For units that are all part of the same SAPU no additional language is necessary, emission limits are all written in the same format (e.g. lbs/ton).  Compliance is demonstrated if the combined emission limit for all units is not exceeded.  For other existing sources, such as delacquering kilns and furnaces going to a common fabric filter, a case-by-case compliance demonstration is required with the permitting authority providing approval of the test plan before the tests are conducted.

      4.2.10  Comment:  Section 63.1511(f)(3) of the proposed rule, regarding establishment of operating and monitoring parameter values, that existing data may be used if, among other conditions, no design or work practice changes have been made since the time the previous data were reported in order to make the previous data acceptable for compliance demonstration.  Two commenters observed that not all work practice or design changes will affect emissions or the operating parameters.  The commenters requested a clarification that only design and work practices affecting emissions may not have changed since the previous report in order for the data to be an acceptable demonstration. (IV-D-17, IV-D-23)
      Response:	 The proposed provision required the owner/operator to certify that no changes in work practice or design have been made since the previous data were obtained.  It also stated that the conditions must be met to the satisfaction of the permitting authority.  If changes have been made and the owner/operator believes the changes have no bearing on emissions, the owner/operator can include an explanation for review and consideration by the permitting authority.  The permitting authority would determine whether to accept the previous data after consideration of whether all the necessary conditions have been met.  The final rule retains the provision as proposed in 63.1511(g).

      4.2.11  Comment:  One commenter stated that sources should be able to use test results obtained prior to the compliance deadline as a demonstration of compliance. (IV-D-28) 
      
      Response:  Test results obtained with the methods required for performance testing in the final rule, in coordination with development of the required operating and monitoring parameter values or ranges as required in the final rule, may be used as a demonstration of compliance at the compliance deadline, provided that all the conditions of 63.1511(g) cited in the previous comment have been met.  Submittal of all the necessary supporting information and data would be part of the OM&M plan preparation, review, and approval process with the permitting authority.  

      4.2.12  Comment:  One commenter requested that different affected sources connected to a common control device be permitted the flexibility to allow consistent monitoring and compliance periods, for example a delacquering kiln and SAPU furnaces. (IV-D-23)
      Response:	 Although the example provided by the commenter is a case in which the emission limit formats are the same for different affected sources, the pollutants regulated, and the monitoring and compliance periods will be different.  The difference in the monitoring and compliance period for different affected sources can be significant.  For example SAPU compliance is determined on the basis of 3-day, 24-hour rolling average emissions and delacquering kilns are based on the average of three 3-hour test runs.  Furthermore, there is no legal construct under the CAA that provides for emissions from two different affected sources to be averaged.  The final rule requires that compliance for different affected sources must be determined separately. 

      4.2.13  Comment:  One commenter suggested improved wording for 1511(b)(4) of the proposed rule as follows: "Where multiple affected sources or emission units are exhausted through a common stack, pollutant sampling for each run must be conducted over a period of time during which all affected sources or emission units complete at least one entire process operating cycle or for 24 hours, whichever was shorter. (IV-D-20)
      Response:  The EPA has adopted the suggested wording change in the final rule.

      4.2.14  Comment:  One commenter stated that the wording in Section 63.1512(j) of the proposed rule presumes there are always control devices present on emission units within the SAPU because testing is required at the outlet to the control device.  However, some emission units may not be controlled, so the requirements will need to be reworded. (IV-D-20)
      Response:	The commenter correctly noted that a wording change is necessary to include testing for uncontrolled emission units.  The final rule has been modified to incorporate the language changes at 63.1512(e).
      
      
      4.2.15  Comment:  One commenter stated that testing, especially for area sources, is needlessly burdensome and expensive.  According to the commenter, the EPA should provide an option to use reliable emission factors.  The owner should not be required to test at highest production level if they can show it would never or could never be operated at such.  There should be something like synthetic minor status available.  The commenter claimed that having to submit results of "any" test will discourage owners from exploring alternative operating scenarios; owners should have the choice to test actual emissions and control HAPs based on feed material certification instead of controlling PM. (IV-D-27)
      Response:  Area sources mentioned by the commenter are only  being regulated for D/F emissions from the processes for which D/F emission limits have been established.  Therefore area sources are not required to control PM under this rule.  The EPA recognizes that testing of sources to determine D/F emissions may be relatively costly in comparison to the cost of some process equipment subject to the rule, and amount of material processed in the affected source, i.e. sweat furnaces.  There are not reliable emission factors available for estimation of D/F emissions for regulatory purposes.
      As a result of comments received, and after further consideration, the EPA has determined to revise the standard for sweat furnaces.  The final rule provides that a sweat furnace equipped with an afterburner with 2 second residence time and with a design temperature of at least 1600 F is not required to conduct a performance test to demonstrate compliance with the D/F emission limit.  The D/F emission limits proposed are being retained in the final rule.  Owner/operators of sweat furnaces located at facilities which are area sources can elect to demonstrate compliance with the D/F emission limits for those existing afterburners that may not meet the equipment specifications through a one-time test. 

4.3  Notifications, Reports, and Records

      4.3.1  Comment:  One commenter noted that the compliance certification requirement in Section 63.1516(c) of the proposed rule limits the information required in annual compliance certifications.  The commenter suggested making the language consistent with 40 CFR 70.6(5)(iii) which is included in their comment.  The suggested language is much broader than that of the proposed rule. (IV-D-20)
      Response:	The EPA did not intend to limit the information that could be required in annual compliance certifications by the permitting authority.  The wording has been changed in the final rule to state that compliance certification must "at least" contain the items listed.

      4.3.2  Comment:  One commenter requested the requirement to develop a Startup, Shutdown, and Malfunction (SSM) plan to apply only to controlled sources.  The same commenter wanted to eliminate the requirement to certify the use of only "non-HAP containing, non-HAP generating flux gases" from the reporting provision of Section 63.1517(b)(2)(vi) of the proposed rule for in-line fluxer using no reactive flux, and certify only "non-reactive flux gases" are used.  (IV-D-31)
      
      Response:	While the add-on air pollution control devices play the primary role in emission control, variation in the process operations generating the waste gas stream and pollutants affect the emissions as well.  Further, these standards anticipate the presence of emission units using pollution prevention approaches rather than add-on controls as part of secondary aluminum processing units in many facilities.  These  emission units and the associated work practices and pollution prevention measures are the means by which emissions are limited in the SAPU.  For these reasons, and consistent with the General Provisions, all affected sources and emission units must be included in the SSM plan.
      The certification requirement for "in-line fluxer using no reactive flux" has been retained in the final rule as it was proposed.  This provision is necessary for consistency with the requirements for group 2 furnaces.  Both emission units have no emission limits and are therefore subject to this more stringent certification statement. 
 
      4.3.3  Comment:  One commenter stated that the requirement for the quality improvement plan (QIP) in 63.1515(b)(11) of the proposed rule should be deleted because this is the only place in the rule it is mentioned. (IV-D-17)  Another commenter asked for a clarification stating that if the QIP in 1515(b)(11) of the proposed rule refers to the QIP of 40 CFR 64, EPA should clearly state that. (IV-D-20)
      Response:  This reference to the Quality Improvement Plan was inadvertently left in the proposed rule when published; it was not intended to be a requirement and has been removed.

4.4  Applicability of General Provisions

      4.4.1  Comment:  One commenter provided a table with comments supporting or opposing, and noting no or neutral comments, on the applicability of each of the General Provisions requirements shown in Appendix A to Subpart RRR. (IV-D-23)
      Response:  The EPA has reviewed applicability of the general provisions and, except as noted in Appendix A to the final rule, all of the general provisions apply to this subpart.

5.	Impacts

      5.1  Comment:  Several commenters (IV-D-12, IV-D-14, IV-D-15, IV-D-16, IV-D-18, IV-D-19, IV-D-20, IV-D-21, IV-D-23, IV-D-27, IV-D-30, IV-D-32) disagreed with the results of EPA's regulatory impact analysis and believed that EPA underestimated the cost of the rule.  The commenters identified the following as deficiencies in the impact analysis:  
      
      1)  The EPA underestimated the number of area sources that would be impacted as a result of the area source D/F standard.  In particular, owners or operators of sweat furnaces, die casting facilities, foundries, and extruders were identified as potentially affected area sources that were either excluded or not adequately accounted for in the analysis.  Furthermore, the commenters claimed that the proposed monitoring, recordkeeping, and reporting, and title V permit requirements would impose a significant burden on area sources.
      2)  The EPA understated the number of small businesses that would be affected by the rule and, as a result, EPA's analysis of impacts on small entities was not adequate.  According to several of the commenters, the small business impacts analysis underestimated small business impacts because it did not accurately account for sweat furnaces, die casting facilities, foundries, and extrusion facilities, many of which are small businesses and would be subject to the rule.  The commenters also claimed that the proposed monitoring, recordkeeping, reporting, and title V permit requirements would impose significant burdens on these small businesses.  They argued that the rule would have a significant impact on a substantial number of small entities and that EPA must, therefore, perform a regulatory flexibility analysis as required by the Regulatory Flexibility Act.
      3)  Commenters took issue with the methods and assumptions used by EPA to estimate the costs and economic impacts of the rule, including failure to adequately account for the large number of affected area sources, title V permitting costs for area sources, and underestimating performance test costs due to the assumption of shared stacks.  As a result, the commenters state that EPA's costs and economic impact estimates are too low.  They argue that the annualized cost of the rule exceeds $100 million and is, therefore, a significant regulatory action under Executive Order 12866.
      Response:  Based on the numerous comments received regarding the regulatory impact analysis, the EPA has reviewed, revised, updated, and refined the analysis to address commenters' points:
      
      1)  With regard to commenters' first point, for the proposed rule, the EPA used the information available on area sources of D/F emissions and requested additional information on the number of area sources, levels of emissions from these sources, the level of control currently employed, and the number of area sources that are also small businesses.  In response to the comments on the proposed rule, and using the information provided by commenters on sweat furnaces, die casting facilities and foundries, EPA has reassessed the cost of the rule on area sources (see Docket item IV-B-4).  In addition, EPA has clarified and, in some cases, revised the proposed rule to address commenter concerns that the proposed rule will be overly burdensome for area sources.  These changes include clarifications or revisions in the applicability of the rule, the performance testing requirements, the scrap inspection program, and giving the State permitting authorities the discretion to defer the requirements for a title V permit for area sources.  On the basis of the information submitted to EPA during the public comment period and changes made to the proposed rule that narrow the applicability to facilities that are area sources (primarily aluminum extruders, die casters, and foundries), the EPA believes the number of those facilities subject to the rule to be small.  
      2)  Regarding the commenters' second point, after reviewing the comments on the small business impacts of the proposed rule and using the information on sweat furnaces, die casting facilities, and foundries provided by commenters, EPA has refined its small business impacts analysis (see Docket item IV-B-6).  The analysis shows that the final rule will not have a significant impact on a substantial number of small businesses; therefore, no regulatory flexibility analysis is required.  The small business impact analysis shows that the impact to small businesses operating sweat furnaces, and to small firms in the aluminum die casting and aluminum foundry industries is minimal.
      3)  Regarding the commenters' third point, EPA considered the comments objecting to the costing methods and assumptions it used to estimate the impacts of the proposed rule.  The EPA has reexamined its cost estimating procedures and believes that overall it has overstated the cost of the proposed rule.  However, in view of the changes in the proposed rule, and to incorporate revisions in the estimated number of affected area sources, EPA has updated its estimate of the cost of the rule (see Docket item IV-B-4 and IV-B-5).  The revised cost of the rule is below the $100 million per year threshold, therefore, the rule is not a significant regulatory action as defined under Executive Order 12866.

      5.2  Comment: One commenter claimed that under the proposed rule one of their facilities would be unable to use their existing monitoring procedures, which were approved by EPA and the State of Washington, e.g., data from existing COMs under existing SIP procedures, single upstream chlorine feed monitoring, and an existing weigh feeder for trona injection (more accurate and reliable than proposed EPA visual inspection).  The cost of the facility's existing monitoring systems exceeds those monitoring costs estimated by EPA for similar model plants.  This commenter also claims that EPA substantially underestimated costs for installing and operating monitoring systems by comparing costs for their facility to EPA Model Plant 6. (IV-D-32)
      
      Response: In considering the comment, EPA notes that the monitoring and operating requirements of the rule have been revised to be less prescriptive and provide greater flexibility for affected sources. The final rule has been revised to provide for approval of site-specific monitoring and operating alternatives, within EPA guidelines, by the permitting authority.  However, should an owner/operator find that the provisions in the final rule for getting approval of site-specific plans do not cover a particular situation or fail to meet their needs, the General Provisions provide for the use of alternatives to monitoring requirements or procedures upon approval by the Administrator.  EPA believes that these revisions to the rule provide industry with the flexibility necessary to comply with the monitoring requirements efficiently and at a reasonable cost.
      
      5.3  Comment: One commenter stated that because they never have been required to measure D/F emissions, they may have to spend $100,000 per plant site just to establish what their emissions are and that the additional cost of controls is unknown until they establish their baseline emissions.  In addition, record keeping requirements will require additional staffing just to take readings.  The commenter concludes that the rule will have major economic impacts and potential to shut down a number of operations.  Another commenter stated that they surveyed 10 testing firms and found the median cost of D/F testing to be $12,000 per source based on tests on medical/infectious waste incinerators and concluded that the average sweat furnace testing costs could easily be 25% of the capital cost of a furnace ($45,000), yielding a significant impact. (IV-D-13, IV-D-7)
      Response: In response to the first commenter, EPA has revised the rule so that it no longer requires testing of each piece of equipment where sources are uncontrolled and are similar from the standpoint of design and operation.  In addition, because area sources only perform an initial compliance test, they will incur lower costs for testing than major sources which are required to retest every 5 years.
      Regarding the concerns of the second commenter, a manufacturer of sweat furnaces, EPA notes that the cost used by EPA to assess impacts is almost double the cost cited by the commenter.  Furthermore, where the sweat furnace afterburner meets the final rule's design specifications, the owner/operator is not required to perform emissions testing, but must only monitor the afterburner operating temperature.
 
      5.4  Comment:  One commenter stated that the rule should provide incentives to identify innovative approaches to emission reductions, and if not, it should allow innovative approaches with the least net economic effect. (IV-D-9)
      Response: The rule does provide for the use of pollution prevention and innovative approaches to reduce emissions by allowing owners/operators to develop site-specific plans for monitoring and operations, grouping in-line fluxers and group 1 furnaces to allow SAPU compliance demonstrations to be conducted in the most cost-effective way, and requiring site specific work practices to prevent or limit the use of materials in furnace operations that result in HAP emissions.


      5.5  Comment:  Benefits of recycling should be considered in regulating this category.  Recycling offsets environmental harms associated with production and use of virgin materials.  Because recycling itself is a pollution prevention activity the Agency might consider providing a recycling credit that reduces or modifies limits for affected facilities, such as an offset determination that contributes toward meeting emission control levels. (IV-D-27)
      Response: EPA agrees that recycling is environmentally beneficial in that it saves resources.  But because the secondary aluminum production industry also releases hazardous air pollutants listed in the Act, and the industry is listed as a source category for regulation under the Act, EPA is required to develop technology-based (MACT) standards for the industry.

      5.6  Comment:  One commenter took issue with certain health-related documents contained in the docket as well as length of the preamble's discussion of the health effects of the pollutants.(IV-D-23)
      Response: The commenter's criticisms were not directed at the rule but at certain docket items and the discussion of health effects in the preamble.  Although consideration of health risks do not enter into the analysis of regulations that are set at the floor level of control as is the case with this regulation, EPA believes that the information presented in the preamble and the supporting docket material is needed to fully inform the public of the health risks associated with the hazardous air pollutants covered by this rule.
      
      5.7  Comment:  EPA presented model results based on four model plants for 384 meteorological sites, the worst risk from which was in Hawaii.  The commenter stated that because this site has unusual air dispersion characteristics, and EPA did not consider activity patterns and indoor/outdoor exposure gradients, that the  potential impacts are overly conservative. (IV-D-23)
        The same commenter stated that the potential health impacts from HCl exposure are exaggerated due to large safety factors applied in developing the RfC from the Lowest Observable Adverse Effect Level (LOAEL).  The commenter provided a review of the derivation of the IRIS RfC for HCl in a report by a consultant.  In addition the consultant addressed other conservative assumptions related to the exposure assessment that in their opinion causes EPA's exposure estimates to be overstated.
      Response:  The comments address methods EPA used in assessing exposure to hazardous air pollutants and the associated health risks.  As noted above, the consideration of health risks do not enter into the analysis of regulations that are set at the MACT floor and the rule for secondary aluminum production facilities is set at the MACT floor.


      5.8  Comment:  One commenter feels that enforcement of the NESHAP will be unequal because some state agencies (e.g., South Coast Air Quality Management District) are better staffed and trained to implement new laws.  The commenter is concerned that facilities in Southern California will be handicapped by the standards in comparison to facilities in other locales. (IV-D-10)
      Response: EPA understands that implementation and enforcement of any rule may vary among agencies depending on several factors but believes that the differences will be minimal.  In addition, nothing in the CAA precludes adoption of more stringent standards by any State.
      
      5.9  Comment:  One commenter noted that there are many differences between the proposed NESHAP and the control measures required by the state regulatory agency.  This commenter stated the most effective way to integrate the federal and state requirements is to revise the state permits to incorporate the relevant federal requirements.  The proposed amendments to 40 CFR Part 63, Subpart E, Approval of State Programs and Delegation of Federal Authorities, allow the use of district operating permits or Title V permits for making equivalency determinations, but the amendments place a limit on the number of individual sources for which permits may be used in determining equivalency between federal and state programs.  The commenter opposes limits on the number of individual sources that can use permits to determine equivalency. (IV-D-19)
      Response: The commenter's concern is in reference to State permit programs and is not relevant to this action.
