DATE:	July 24, 2015

SUBJECT:	Technology Review for the Secondary Aluminum Production Source Category - Final Rule

FROM:	Mark Bahner and David Green, RTI, International 

TO:		Rochelle Boyd, U.S. Environmental Protection Agency, OAQPS
Background
Requirements of Section 112(d)(6) of the CAA
Section 112 of the Clean Air Act (CAA) requires the EPA to establish technology-based standards for sources of HAP. These technology-based standards are often referred to as maximum achievable control technology, or MACT, standards. Section 112 also contains provisions requiring the EPA to periodically review these standards. Specifically, paragraph 112(d)(6) states:
      (6) REVIEW AND REVISION.  -  The Administrator shall review, and revise as necessary (taking into account developments in practices, processes, and control technologies), emissions standards promulgated under this section no less often than every 8 years.
      
Description of the Secondary Aluminum Production Source Category and Requirements of the Current NESHAP
The current National Emission Standards for Hazardous Air Pollutants (NESHAP) for the Secondary Aluminum Production source category were promulgated on March 23, 2000 (65 FR 15690) as 40 CFR part 63, subpart RRR. The rule was amended at 67 FR 79808, December 30, 2002; 69 FR 53980, September 3, 2004; 70 FR57513, October 3, 2005; and 70 FR 75320, December 19, 2005. The NESHAP applies to affected sources of HAP emissions at secondary aluminum production facilities. A secondary aluminum production facility is defined as any establishment using clean charge, aluminum scrap or dross from aluminum production, as the raw material and performing one or more of the following processes: scrap shredding, scrap drying/delacquering/decoating, thermal chip drying, furnace operations (i.e., melting, holding, sweating, refining, fluxing, or alloying), recovery of aluminum from dross, in-line fluxing or dross cooling. A secondary aluminum production facility may be independent or co-located with a primary aluminum production facility. For purposes of the NESHAP, aluminum die casting facilities, aluminum foundries and aluminum extrusion facilities are not considered to be secondary aluminum production facilities if the only materials they melt are clean charge, customer returns, or internal scrap, and if they do not operate sweat furnaces, thermal chip dryers, or scrap dryers/delacquering kilns/decoating kilns.
There are an estimated 161 secondary aluminum production facilities that are subject to the NESHAP. Emission limits have been promulgated for particulate matter (PM) as a surrogate for metal HAP, total hydrocarbons (THC) as a surrogate for organic HAP other than dioxins and furans, dioxins and furans (D/F) expressed as toxicity equivalents, and hydrogen chloride (HCl) as a surrogate for HCl, hydrogen fluoride and chlorine. HAP are emitted from the following affected sources: aluminum scrap shredders (regulated for PM), thermal chip dryers (regulated for THC and D/F), scrap dryers/delacquering kilns/decoating kilns (regulated for PM, D/F, HCl and THC), sweat furnaces (regulated for D/F), dross-only furnaces (regulated for PM), rotary dross coolers (regulated for PM), group 1 furnaces (regulated for PM, HCl and D/F), and in-line fluxers (regulated for PM and HCl). Group 2 furnaces and certain in-line fluxers are regulated by work practices. 
Control devices currently in use to reduce emissions from affected sources subject to the NESHAP include fabric filters for control of PM from aluminum scrap shredders; afterburners for control of THC and D/F from thermal chip dryers; afterburners plus lime-injected fabric filters for control of PM, HCl, THC and D/F from scrap dryers/delacquering kilns/decoating kilns; afterburners for control of D/F from sweat furnaces; fabric filters for control of PM from dross-only furnaces and rotary dross coolers; lime-injected fabric filters for control of PM and HCl from in-line fluxers; and lime-injected fabric filters for control of PM, HCl and D/F from group 1 furnaces. All affected sources with add-on controls are also subject to design requirements and operating limits to limit fugitive emissions. 
Developments in Practices, Processes, and Control Technologies
For the purposes of this technology review, a "development" was considered to be a (n): 
add-on control technology that was not identified during the development of the current NESHAP for the source category; 
improvement to an existing add-on control technology that could result in significant additional HAP emissions reductions; 
work practice or operational procedure that was not identified during development of the current NESHAP for the source category; or 
applicable process change or pollution prevention alternative that was not identified and considered during the development of the current NESHAP for the source category. 
       
We investigated developments in practices, processes, and control technologies through a literature review and discussions with industry representatives, and included questions in a Section 114 information collection request (ICR) that was sent to all companies thought to be subject to the NESHAP. The results of these analyses are presented in the following sections.
Literature review and industry contacts 
2.1.1 	Multichamber Delacquering Kiln/Melting Furnace
At least one company supplies multichamber furnaces that combine the functions of a delacquering kiln and a melting furnace. This furnace has the potential to reduce emissions, because all of the emissions from the delacquering of used beverage cans and other coated scrap are swept into another chamber where the delacquered scrap is melted. These emissions are combusted in the melting chamber, reducing energy requirements; destroying THC and D/F; and eliminating the need for an afterburner. A multichamber furnace can therefore be used as a replacement for a delacquering/decoating kiln (with afterburner) plus a group 1 furnace handling only clean charge, or the multichamber furnace can be used as a replacement for group 1 furnace handling other than clean charge.
At least 16 of these furnaces are in operation in Europe, Asia and the Middle East. One furnace of this type is presently operating in the U. S. and is permitted as a group 1 furnace handling other than clean charge.
The emissions test data for the one multichamber furnace operating in the U.S. indicate that the furnace produces D/F emissions that are within the range of emissions test data for other group 1 furnaces handling other than clean charge, and delacquering/decoating kilns. Thus, the multichamber furnace D/F test data are within the range of other equipment using control technology considered by the EPA in the Subpart RRR NESHAP. These data indicate that the multichamber furnace does not produce lower D/F emissions (in toxic equivalents, or TEQs) than any other group 1 furnace handling other than clean charge, or than any delacquering/decoating kiln, as shown in Table 1. Table 1 lists D/F emissions from the one multichamber furnace, operating at Logan Aluminum. Table 1 also lists D/F emissions from three group 1 furnaces handling other than clean (i.e., "dirty") charge, and three delacquering/decoating kilns. The other furnaces and delacquering/decoating kilns in Table 1 were selected because they had low reported D/F emissions, and emission test reports were available to support those reported emissions. The D/F emissions from these other sources are lower than for the multichamber furnace. Therefore, based on available information, we are unable to conclude that the multichamber furnace technology reduces HAP emissions to lower emission levels than technologies that were considered in promulgating the subpart RRR NESHAP and are already used by other facilities in the U.S.
Table 1. Comparison of D/F Emissions from a Multichamber Furnace with Other Technologies
RTI ID
Facility Name
Equipment ID
Equipment Type
Test Date
Test-Condition
Average D/F in ug/Mg TEQ[(1)] 
351
Logan Aluminum
Multichamber Furnace
Grp1 dirty charge
5/13/2008
1-50%Class I/50% Class III
1.061[(2)]
351
Logan Aluminum
Multichamber Furnace
Grp1 dirty charge
5/13/2008
1-100% Class I
0.461
351
Logan Aluminum
Multichamber Furnace
Grp1 dirty charge
5/14/2008
2-80% UBC's/20% Class I
0.496
351
Logan Aluminum
Multichamber Furnace
Grp1 dirty charge
8/12/2008
1-50%Class I/50% Class III
0.26
421
Alcan Rolled Products
DC1M
Grp1 dirty charge
1/29-31/2008
1
0.023
335
Jupiter Aluminum
Furnace 2
Grp1 dirty charge
1/25/2010
1
0.027
211
Alumax Texarkana
EPN 011A
Grp1 dirty charge
4/20/2009
1
0.103
198
Novelis
P9A
Delacquering Kiln
10/01-02/2008
1
0.000419
198
Novelis
P9B
Delacquering Kiln
9/30-10/1/2008
1
0.000173
415
JL French
P30-1
Delacquering Kiln
6/22/2009
1
0.02
    Average of three test runs, except where noted.
 Average of two test runs.

2.2.2	Eddy Current Separators
Sweat furnaces are used in the secondary aluminum industry to separate aluminum from ferrous metals. Certain types of scrap, primarily automotive, are composed of individual pieces of metal which contain both aluminum and iron or steel. Automobile parts such as aluminum engine blocks with cast iron cylinders, aluminum transmission cases with steel gears and inserts and aluminum suspension components with steel inserts are examples of sweat furnace feed. These materials, often containing oil and grease, are separated by melting the aluminum from the ferrous metals with higher melting points. The molten aluminum is tapped to form ingots or "sows," and the ferrous residue is raked from the furnace. Sweat furnace emissions contain D/F, which results from the combustion of the organic contaminants in the scrap. 
Eddy current separators are used to separate a concentrated aluminum fraction from a heterogeneous scrap feed. These units operate at ambient temperature and emit no D/F or other gaseous pollutants. Eddy current separators are used to separate a concentrated aluminum fraction from a heterogeneous scrap feed. These units operate at ambient temperature and emit no D/F or other gaseous pollutants. They are used on the material output from mechanical shredders that shred automobiles and appliances (not on the material from aluminum scrap shredders used in the secondary aluminum industry). These units can potentially decrease the need for sweat furnaces. However, the product of eddy current separators is not an aluminum ingot or sow, as with a sweat furnace. Therefore, the product of eddy current separators must undergo further processing to produce an aluminum ingot or sow, and it is not possible to directly compare eddy current separators with sweat furnaces. 
2.1.3 	Catalytic Filter Bags
Catalytic filtration systems, including catalytic filter bags, are available to reduce D/F emissions. These bags incorporate an expanded polytetrafluoroethylene membrane coated with a precious metal catalyst which promotes the oxidation of D/F. One manufacturer claims that this system is installed in over 100 applications around the world, including at least one secondary aluminum processing plant. To determine the extent that these bags are in use in the secondary aluminum industry in the U. S., the EPA included a question in an information collection request (ICR) sent to all identified secondary aluminum production facilities. The question "Do you use catalytic filters for dioxin control . . .)?" was answered "no" or "not applicable" by 126 of the 159 facilities that responded to the ICR. The remaining facilities did not answer this question. Some, or all, of the blank responses and the "not applicable" responses are attributable to facilities that do not operate fabric filters. Therefore, the EPA has no information about any secondary aluminum facility in the U.S. currently using catalytic filter bags, and no specific secondary aluminum facility in another country that uses catalytic filter bags has been identified.
Catalytic fabric filter bags are potentially problematic for the secondary aluminum production industry because they require a minimum fabric filter inlet temperature of approximately 300 degrees Fahrenheit to produce the catalytic oxidation of D/F. Many fabric filters at secondary aluminum production facilities in the U.S. operate below this temperature, specifically to avoid de novo creation of D/F in the fabric filter, which occurs at temperatures from approximately 300 to 840 degrees Fahrenheit (150 to 450 degrees Celsius),[1] and because large amounts of ambient air flow into hoods are necessary to meet design guidelines of the American Council of Governmental Industrial Hygienists (ACGIH). Therefore, the fact that catalytic filter bags require a temperature above approximately 300 degrees Fahrenheit conflicts with a need to remain below 300 degrees Fahrenheit to avoid de novo synthesis of D/F, and the need to draw in large amounts of ambient air into hoods to meet ACGIH guidelines.
Therefore, the EPA cannot conclude that catalytic fabric filter bags are more effective at reducing D/F emissions at secondary aluminum production facilities than the control technologies considered by the EPA in the 2000 Subpart RRR NESHAP. Catalytic filtration systems are not at present a demonstrated control technology for the Secondary Aluminum Production source category that should be used as the technical basis to establish more stringent emission limits for the source category. 
2.1.4 	Work Practices: Good Combustion Practices
D/F emissions in municipal and medical waste combustors are controlled in part through "good combustion practice (GCP)." For municipal waste combustors, the major technical objectives of GCP were determined to be achieved by monitoring and controlling: (a) the flue gas concentration of CO; (b) steam load (a surrogate for PM carryover) and (c) temperature at the inlet of the PM control device.[1] The first two parameters are not applicable in the operation of secondary aluminum furnaces and kilns, since the fuel in secondary aluminum is not waste, but is instead typically natural gas. [Afterburner operation (for sweat furnaces, chip dryers and delacquering kilns/decoating kilns/scrap dryers) is presently subject to inspection for burners and proper adjustment of combustion air.] The third parameter, temperature at the inlet of the PM control device, must already be monitored in the existing NESHAP. Therefore, GCP, as it relates to D/F formation in secondary aluminum production, is already addressed by the existing NESHAP, and is not a new development in practices, processes, or control technologies for the Secondary Aluminum source category under section 112(d)(6). 
2.1.5 	Other Work Practices
The EPA investigated other work practices such as better scrap inspection and cleaning, and process monitoring. However, no such practices were identified that were not already identified at the time of the original NESHAP. For example, the issue of scrap inspection was investigated extensively in the development of the original NESHAP, and no sampling or analytical procedures were identified then or in our present review that can determine whether scrap of unknown origin is completely free of paints, coatings, or lubricants. 
Responses to ICR
In an attempt to identify developments in emission control technologies in use, an ICR was sent to all identified secondary aluminum production facilities. The following questions were asked:
Please provide details for any alternative control devices (i.e., control devices other than fabric filters, lime-coated fabric filters, or afterburners), monitoring (including particulate matter or HCl continuous emissions monitors), or operating conditions at this facility for equipment regulated under 40 CFR 63, subpart RRR. 
Have you injected activated carbon or other type of sorbent for HAP control (excluding research efforts)? What barriers do you envision to adding carbon injection to fabric filters for HAP control?
Do you have any plans to install any new higher efficiency rated control devices or have any pending applications to add on any new controls? 

2.2.1 	Activated Carbon Injection (ACI) for D/F Control
Three respondents reported using ACI for control of dioxin, and one respondent reported using ACI previously on a unit that has since been shut down. Activated carbon is typically added to control D/F, although in one case it was used with a thermal chip dryer and may also have helped control THC. This technology was known at the time of the development of the current standard and was regarded as a possible control alternative for sources that were unable to demonstrate compliance using lime-injected fabric filters. The EPA also evaluated the broader use of ACI under the technology review. The option considered under the technology review was the same control option evaluated under the ample margin of safety analysis (i.e., an option of lowering the D/F emissions limit from 15 to 10 ug TEQ/Mg for group 1 furnaces processing other than clean charge at all facilities). However, the EPA rejected it as an option under Section 112(d)(6) because the EPA concluded it was not a cost effective option for this source category. The estimated costs and cost effectiveness are described in Section 3.0 of this memorandum. 
2.2.2 	Ammonia Injection for HCl Control
Three respondents reported adding injecting ammonia into furnace exhaust gas as a means of HCl control. Another respondent plans to replace the lime currently used to control HCl emissions with ammonia. We were unable to determine from the ICR responses whether the remaining respondents reporting the use of this technology are reducing lime usage as a result of using ammonia, or adding ammonia in addition to lime currently used. Further, we do not have any HCl test data for the furnaces at these four facilities. The HAP emission reductions achieved from using this technology alone or in addition to lime injection are not known. This technology may be suitable for retrofit to existing fabric filters. The applicability of this technology is unlikely to be influenced by furnace configuration because the ammonia is added between the furnace and the fabric filter. It has the potential of decreasing the amount of fabric filter dust requiring disposal, but would likely result in increased emissions of ammonia (which is not a HAP, but may contribute to nitrogen deposition). We do not have sufficient data to determine whether ammonia injection provides greater control of HCl emissions than lime injection.
DIOXIN/FURAN LIMIT ANALYSIS
The broader use of activated carbon for D/F control was evaluated under the ample margin of safety analysis as a potential approach to reduce multipathway risks from D/F. As mentioned in Section 2.2.1 of this memorandum, this option was also considered under the technology review. The analysis examined lowering the D/F limit for Group 1 furnaces handling other than clean charge from 15 micrograms per megagram (ug/Mg) of charge to 10 ug/Mg of charge. The analysis determined that there are twelve Group 1 furnaces with emissions above 10 ug/Mg. Eleven of the twelve furnaces are presently controlled with fabric filters, and it was assumed that activated carbon injection would be added. One furnace is presently uncontrolled and it was assumed that a direct lime injection fabric filter will be added. In both cases, it was assumed that D/F emissions would be reduced by 85 percent. The results of the analysis indicated an estimated D/F emissions reduction of 2.826 grams toxic equivalency (TEQ), at a cost effectiveness of approximately $765,000 per gram TEQ of D/F.[2]
CONCLUSIONS: Recommended Revisions Based on Developments in Practices, Processes, and Control Technologies
This review identified several developments in practices, processes, or control technologies that have been implemented in this source category since promulgation of the current NESHAP. However, these technologies are not in use by a substantial number of secondary aluminum production facilities in the U. S. and as described below, we did not identify any developments that warranted revision to the NESHAP. 
Specific developments in practices, processes, or control technologies examined included:
Multichamber Delacquering Kiln/Melting Furnace  -  Only one such furnace is currently in use in the U.S. D/F emissions from this furnace are within the range of other group 1 furnaces handling other than clean charge, and delacquering/decoating kilns. Therefore, we are unable to conclude that this technology reduces HAP emissions relative to technologies that were considered by EPA in promulgating the subpart RRR NESHAP and are already used by other facilities in the U.S. 
Eddy Current Separators  -  It is not possible to compare eddy current separators with secondary aluminum production equipment such as sweat furnaces. Sweat furnaces produce aluminum ingots or sows, whereas eddy current separators simply provide a concentrated aluminum fraction from a heterogeneous scrap stream. The concentrated aluminum fraction from an eddy current separator requires further processing to produce an aluminum ingot or sow.
Catalytic Filter Bags  -  The EPA requested all facilities potentially subject to Subpart RRR to identify whether they used catalytic filter bags. None used these bags. Further, there is potential problem in that catalytic filter bags require a minimum temperature of approximately 300 degrees Fahrenheit to destroy D/F, whereas secondary aluminum fabric filters typically operate at temperatures lower than 300 degrees Fahrenheit to avoid de novo synthesis of D/F, and because large volumes of ambient air must be drawn into furnace hoods to promote effective capture. Therefore, we cannot conclude that catalytic filter bags are more effective in reducing D/F emissions at secondary aluminum facilities than control technologies considered by the EPA in the 2000 Subpart RRR NESHAP.
Work Practices: Good Combustion Practices (GCP)  -  GCP as they relate to reducing emissions from municipal waste combustors and medical waste incinerators are not generally applicable to secondary aluminum furnaces, since secondary aluminum furnaces burn natural gas rather than waste. One GCP parameter that is applicable to the secondary aluminum industry is to monitor fabric filter inlet temperature; this must already be monitored in the existing NESHAP. Therefore, GCP, as it relates to D/F formation in secondary aluminum, is already addressed by the existing NESHAP.
Work Practices: Other Work Practices  -  In addition to GCP, we investigated other work practices such as better scrap inspection and monitoring. No sampling or analytical procedures could be identified that could determine whether scrap of unknown origin was completely free of paints, coatings, or lubricants.
Activated Carbon Injection  -  Three facilities reported using activated carbon injection to control D/F. This technology was identified at the time of the development of the current NESHAP, and so is not a new control technology. However, as described previously in this memorandum, we evaluated the potential for broader application of ACI than is currently used in the secondary aluminum production industry as an option under the technology review and under the ample margin of safety analysis as a potential approach to reduce D/F emissions. However, as described previously, the EPA did not propose or promulgate this option under the technology review because the EPA concluded it was not cost effective for this source category.  
Ammonia Injection for HCl Control  -  Four facilities use ammonia injection for HCl control. However, three use ammonia injection in combination with lime injection, so it is not possible to separate the HCl control achieved by lime injection with the HCl control achieved by ammonia. The fourth facility had not yet switched to full ammonia injection. The EPA does not have sufficient data to determine whether ammonia injection provides greater HCl control than lime injection. 
References
    Kilgroe, James D., W. Steve Lanier, and T. Rob Van Alten. "Development of Good Combustion Practice for Municipal Waste Combustors." Presented at 1992 National Waste Processing Conference. Available at: http://www.seas.columbia.edu/earth/wtert/sofos/nawtec/1992-National-Waste-Processing-Conference/1992-National-Waste-Processing-Conference-15.pdf. Accessed July 24, 2015.
   
    Bahner, Mark (RTI, International), to Rochelle Boyd (EPA). "Memorandum: Technical Support Document for the Secondary Aluminum Source Category Final Rule." July 2015. Available in EPA docket for Secondary Aluminum Production Risk and Technology Review. 

