March 26, Dear Sir or Madam:

Transcription

1 March 26, 2012 EPA Docket Center, EPA West (Air Docket) Attn: Docket ID No. EPA-HQ-OAR U.S. Environmental Protection Agency Mail Code 2822T 1200 Pennsylvania Avenue, N.W. Washington, D.C Re: National Emission Standards for Hazardous Air Pollutant Emissions: Hard and Decorative Chromium Electroplating and Chromium Anodizing Tanks Supplemental Proposed Rule Docket No. EPA-HQ- OAR Dear Sir or Madam: Enclosed please find comments submitted on behalf of the National Association for Surface Finishing (NASF) in response to EPA s February 8, 2012 supplemental proposal for National Emission Standards for Hazardous Air Pollutant Emissions: Hard and Decorative Chromium Electroplating and Chromium Anodizing Tanks. 77 Fed. Reg (2012). If you have any questions, would like additional information, or would like to discuss these comments, please contact me by telephone at or by at jhannapel@thepolicygroup.com. Respectfully submitted, Jeffery S. Hannapel The Policy Group On Behalf of NASF

2 March 26, 2012 Comments on the Supplemental Proposal for National Emission Standards for Hazardous Air Pollutant Emissions: Hard and Decorative Chromium Electroplating and Chromium Anodizing Tanks Docket No. EPA-HQ-OAR These comments are submitted on behalf of the National Association for Surface Finishing (NASF) in response to the United States Environmental Protection Agency (EPA) February 8, 2012 supplemental proposal for National Emission Standards for Hazardous Air Pollutant Emissions: Hard and Decorative Chromium Electroplating and Chromium Anodizing Tanks. 77 Fed. Reg (2012). NASF members with hard and decorative chromium electroplating and chromium anodizing processes will be impacted by this proposed rule. These comments summarize some of the NASF s concerns regarding the proposed rulemaking and NASF requests that EPA issue the final regulation with revisions that are consistent with the comments provided below. EXECUTIVE SUMMARY In summary, these comments raise the following important issues and concerns: EPA Has Unjustifiably Reversed Its Earlier Position on the Rule In this supplemental proposal, EPA concludes that the residual risks associated with the chromium electroplating and anodizing industry are lower that the agency 2

3 estimated in the October 2010 proposal. EPA, however, inexplicably proposes to impose more stringent surface tension levels and emission limits based on technology and additional controls that it now claims can reduce risks at reasonable costs. This about-face makes no sense, particularly in light of the further information that NASF is providing today. The Industry Has Already Reduced Emissions by More Than 99.7 Percent As context, over the past 15 years, the surface finishing industry has dramatically reduced chromium emissions. Since 1995, the industry has reduced nationwide emissions of chromium from 173 tons to at least EPA s conservative current estimate of 1,140 pounds a reduction of 99.7 percent. Actual Industry Data Show Much Lower Emissions The actual emissions associated with the industry are significantly lower than EPA emissions estimates. NASF s newly collected data and our analysis indicates that a more accurate estimate of current emissions is approximately 200 pounds a year, or 80 percent lower that EPA s estimate and a 99.9 percent reduction since Such reductions in emissions and the associated risks represent an environmental policy success story that should be acknowledged and celebrated. Instead, EPA is proposing more stringent new limits on the industry. Further Emission Reductions are Unnecessary and Newly Added Costs Are Unreasonable Because actual emissions are so much lower than EPA s conservative estimate, there is no justification for imposing the stringent limits included in this proposed rule. The risks posed by emissions from chromium electroplating and anodizing operations are miniscule and have been deemed acceptable by EPA, further reduction of emissions is unnecessary and the costs of attempting to achieve additional reductions are unreasonable. EPA Data Underpinning the Rule Contain Major Errors and Outdated Information In addition, there are significant quality problems with the data underpinning this rule. While EPA now has more data that it had for the October 2010 proposal, most of this data is still outdated, incomplete, inaccurate, and unreliable. In addition, there are numerous inconsistencies in the data. As NASF has continued to review the data, in the limited time available, it continues to identify errors and outdated information. The number and scale of the problems completely undercut the validity of the proposed rule. The Majority of Facilities in EPA s Database are Closed, Do Not Process Chromium or Have Lower Emissions vs. EPA s Modeled Data In order to check the reliability of EPA s data, NASF reached out to its members and requested information on approximately 300 of the highest emissions facilities identified by EPA. Over a third of those facilities have responded to date and these responses indicate that actual emissions are 94 percent lower than EPA s estimates. Of the facilities that responded or for which new information was collected, 35 percent were found to be closed, 38 percent no longer, or never did, 3

4 process chromium, 25 percent provided data that showed emissions lower than EPA s estimate, and only 2 percent actually confirmed EPA s emissions estimate. EPA s Statistical Analysis is Problematic The problems do not end with the data. The statistical analyses that EPA conducted with the data it had available also suffer from reliability problems. In addition, EPA s emissions and dispersion modeling for hexavalent chromium is systematically biased high compared to actual monitoring data collected by EPA, leading to greatly over-estimating ambient concentrations (and risks) at those sites where higher concentrations are modeled. EPA Has Committed Procedural Errors and Has Not Allowed Industry to Complete Its Analysis and Provide Meaningful Comment EPA has committed procedural errors in this rulemaking. Despite the fact that NASF raised significant concerns about the data and other aspects of the proposed rule, EPA denied the NASFs request for an extension of the comment period. This has left NASF and others with insufficient time to fully and meaningfully comment on this proposed rule. In addition, EPA should have, but did not, conduct a Small Business Regulatory Enforcement Fairness Act (SBREFA) panel review of this rule pursuant to its obligations under the Regulatory Flexibility Act. EPA Reversed the Sequence of Proper Analytical Steps under the Clean Air Act EPA s regulatory approach in this rulemaking reversed the proper analysis under the Clean Air Act. Had EPA first evaluated the existing standards under 112(f), it should have concluded that those existing standards were appropriately protective, because it reached precisely that conclusion in October 2010, and EPA in this proposal concludes that the risk is even lower than what it believed then. There was no need to conduct a section 112(d)(6) review. Though, in any event, the record does not support a lower standard based on such a review section 112(d)(6), as EPA did not identify practices, processes, or control technologies that warrant stricter standards. EPA Has Little or No Data to Show that Non-PFOS Fume Suppressants are Equivalent to PFOS Fume Suppressants Finally, the technology that EPA is proposing to require in these rules, non-pfos fume suppressants, is still a technology in transition. The new technology is not a simple drop-in replacement for PFOS fume suppressants and has not been proven effective in many chromium applications. Even more troubling, EPA has provided no data in the record that non-pfos fume suppressants can achieve the stricter proposed new surface tension levels. EPA merely assumes that non-pfos fume suppressants are equivalent in performance to PFOS fume suppressants without presenting any scientific proof or supporting data. For all of these reasons, and as discussed in more detail below, EPA needs to fundamentally reconsider these proposed regulations. NASF welcomes the agency s 4

5 approach proposed in the October, 2010, proposal, but believes that the stricter regulations included in this proposal are arbitrary and capricious and unsupported by the record. I. SUMMARY OF THE SURFACE FINISHING INDUSTRY The NASF has nearly 2,000 members that include metal finishing companies, metal finishing suppliers, and individual and professional members. The NASF represents the business, management, technical and educational programs as well as the regulatory and legislative advocacy interests of the surface finishing industry to promote the advancement of the North American surface finishing industry globally. The surface finishing industry plays a vital role in the lives of consumers and in the nation s economic future. The industry s role in corrosion protection alone provides an estimated $200 billion annual economic benefit to the nation, including significant corrosions protection for military equipment that provides national defense. Surface finishing ensures that the products people use every day last longer, work better, and look better. Metal finishing operations are performed in two ways: 1) as a "captive" operation or department of a manufacturing company; and 2) on a job-shop basis where the work is performed under contract for the owner of the product or material that is to be finished. Although many manufacturers continue to operate metal finishing departments, the increasing trend is to subcontract this work to independent firms. This trend is a result of the high operating costs and a realization that metal finishing is both a regulatory and process specialty. Over 80 percent of the job-shops in business employ fewer than 75 people, while nearly 40 percent employ fewer than 20 people. Most job-shop surface finishing firms are family-owned businesses, located in urban areas, with a large percentage of minority employees. 5

6 II. RULEMAKING HISTORY The Chromium Electroplating NESHAP regulates emissions of chromium compounds from three source categories: hard chromium electroplating, decorative chromium electroplating and chromium anodizing. The NESHAP was first promulgated on January 25, 1995 (60 Fed. Reg. 4963) and applies to both major and area sources, although all sources subject to the rule are area sources with respect to chromium emissions. The NESHAP was amended on July 19, 2004 (69 Fed. Reg ) to revise some control technology, monitoring, and implementation issues. On October 21, 2010 (75 Fed. Reg ) in accordance with the Clean Air Act (CAA) requirement that standards be reviewed every eight years, EPA proposed a rule based on its initial residual risk and technology review of the Chromium Electroplating NESHAP. In that proposal EPA indicated that the residual risk due to HAP emissions from the source categories were acceptable, and no additional controls were necessary to provide an ample margin of safety to protect public health because [EPA] had not identified additional controls that would reduce risks at reasonable costs. 77 Fed. Reg. at EPA did, however, propose to: 1) incorporate new housekeeping practices; 2) phase out the use of perfluorooctyl sulfonates (PFOS) fume suppressants; 3) revise the startup, shutdown, and malfunction (SSM) provisions; 4) revise monitoring and testing requirements; and 5) make several technical corrections to the standard. Despite the fact that EPA did not propose to revise the NESHAP pursuant to 112(f)(2) of the CAA, it requested comments on potential cancer risks posed by emissions from these industry source categories. After reviewing comments on the proposed rule (presumably comments on the potential cancer risks posed by emissions from chromium electroplating and anodizing operations), EPA decided not to finalize the proposal and issued the February 8, 2012 supplemental proposal that is the subject of these comments. In this supplemental proposal, EPA concludes that the residual risks are lower that it estimated in the October 2010 proposal. EPA, however, inexplicably, proposes to impose 6

7 more stringent surface tension levels and emission limits based on technology and additional controls that it now claims can reduce risks at reasonable costs. Based on the comments below, NASF will demonstrate that the proposed revisions to the NESHAP in this supplemental proposal are inappropriate, unwarranted and unsupported by the record. III. SMALL BUSINESS IMPACTS As part of the October 2010 proposed rule, NASF identified potential impacts on the facilities subject to this NESHAP, nearly all of which are small businesses. EPA did not conduct a Small Business Regulatory Enforcement Fairness Act (SBREFA) panel pursuant to its obligations under the Regulatory Flexibility Act, in part, because it was not proposing any new surface tension levels, emission limits, or additional controls. In preparation for this supplemental proposal, and the new revisions to surface tension levels and emission limits, NASF urged EPA to conduct a SBREFA panel to review EPA s data and to assess the potential impacts of the proposed changes on small business. Such a valuable effort would have greatly informed EPA s decision making process for this supplemental rulemaking. In response to the NASF s request, EPA began the administrative process for convening a SBREFA panel. Unfortunately, at the eleventh hour EPA reversed course, concluding that the supplemental proposal would not have a significant impact on a substantial number of small entities, and declined to convene a SBREFA panel. EPA s justification for the decision was also partly due to time constraints and the Agency s failure to adhere to its own rulemaking schedule and deadlines. The Small Business Administration (SBA), Office of Advocacy, along with NASF, opposed EPA s decision not to convene a SBREFA panel and continues to express opposition in its comments that were recently submitted to EPA on this supplemental proposal. 7

8 IV. DATA QUALITY Many of the concerns regarding this proposal center on the poor quality of the data used by EPA in this rulemaking. Much of the data is more than 15 years old, and most of it was generated prior to amendments to the NESHAP and changes to OSHA s workplace exposure standard for hexavalent chromium in 2006 that caused many facilities to improve their performance significantly in capturing and controlling hexavalent chromium in the workplace. EPA has claimed that it has improved its database through its outreach to 28 state and regulatory agencies last year. While EPA now has more data that it had for the October 2010 proposal, much of this data is still outdated, incomplete, inaccurate, and unreliable. In addition, numerous inconsistencies in the data are apparent in the administrative record for this rulemaking. Some of the flaws in the data will be discussed below in more detail. Pursuant to the Data Quality Act (DQA), EPA is obligated to use the best data to make regulatory decisions. The DQA requires that information used by EPA should be useful to the intended users and should be presented in accurate, clear, complete, and unbiased manner. As a substantive matter, the information must be accurate, reliable and unbiased. The quality of the data is particularly important when its use will have a clear and substantial impact on important public policies such as this proposed rule. Finally, regulatory agencies must also be transparent about the data, methods, models, assumptions and statistical procedures such that the analysis can be replicated by a qualified person. Using these criteria, the information used as the basis for this proposed rule has serious data quality deficiencies. Given that EPA recognized that its data and technical support for this rule was inadequate, the Agency should have conducted a more formal information collection request (ICR) effort that also include the regulated industry before proceeding with this rulemaking. In fact in recent risk and technology NESHAP reviews, EPA has imposed new standards and additional controls only in those cases where it had a substantial 8

9 amount of recent and reliable data or had recently conducted an ICR. In this NESHAP, EPA has not only relied upon limited and flawed data, but it has also chosen to stretch the data beyond its logical and reasonable limits. Imposing new regulatory standards based on unreliable data creates an unnecessary and unwarranted burden on an industry that serves several critical supply chains (such as military, aerospace and automotive) and is experiencing significant economic challenges. V. COMMENT DEADLINE EXTENSION REQUEST For the reasons listed below, NASF had requested an extension of the comment period so that it would have a meaningful opportunity to comment on the supplemental proposal. 1) The February, 2012 supplemental proposal is substantially different from the October 2010 proposed rule. Specifically, EPA has now proposed more stringent emission limits and surface tension levels for chromium electroplating and anodizing tanks. 2) To support the supplemental proposal, EPA conducted a data collection effort with over 28 state and local regulatory agencies regarding chromium electroplating and anodizing operations. Based on our initial review of these new data, NASF identified several concerns about the data, including: data from facilities that have been closed for several years, data from facilities that no longer process hexavalent chromium, data from facilities that predated the implementation of the 1995 chromium electroplating NESHAP, and data that are not representative of current operations at facilities. 3) This supplemental proposed rule relies on an extensive set of data to support several technical claims regarding issues such as revised surface tension levels, revised emission limits, current industry emissions control practices, new fume 9

10 suppressant technology, residual risk analysis and cost impacts. Accordingly, NASF needed additional to review the data in the record, particularly the new data added for the supplemental proposal, to have the opportunity to comment effectively on the proposed rule. 4) It is critical to ensure that the data in the record are accurate and reflective of the practices and operational status of chromium electroplating and anodizing facilities. Additional time to review and evaluate the data would assist EPA s effort to make an informed decisions based on credible data regarding available control technologies and practices in the surface finishing industry. NASF recognizes that EPA is facing a court-ordered deadline for the issuance of the final rule, but EPA has apparently elected not to request an extension of that deadline to give commenters adequate time to review and comment on the record and EPA adequate time to address the comments. An extension of the comment period would have facilitated EPA efforts by improving the quality of the technical basis for the rule. Specifically, additional time to review and evaluate the data in the record would have provided EPA with more effective feedback on the technical support for the proposed revisions, the measurable environmental benefits that may result from the rule, and the potential impacts on the surface finishing industry. Unfortunately, EPA denied the NASFs request for an extension of the comment period. These comments provide significant information for EPA s consideration, but NASF did not have enough time to review and evaluate all the data and analysis provided by EPA in the record. NASF will continue to work with EPA officials to identify issues with the data, analysis, and methodologies used and ignored by EPA in this rulemaking. EPA has agreed to consider the information on the potential impacts these issues may have on the technical basis and conclusions reached by EPA in this rule, provided that NASF submits within a reasonable time period after the close of the comment period. 10

11 VI. THE CHROMIUM ELECTROPLATING NESHAP HAS BEEN EFFECTIVE IN CONTROLLING CHROMIUM EMISSIONS When EPA promulgated the Chromium Electroplating NESHAP in 1995, the Agency estimated that the total chromium emissions from the industry were 173 tons per year. Over the past 15 years, the surface finishing industry has implemented engineering controls and management practices to reduce chromium emissions significantly. Since 1995 the industry has reduced nationwide emissions of chromium from 173 tons to EPA s conservative current estimate of 1,140 pounds a reduction of 99.7 percent. As discussed below, EPA continues to over-estimate chromium emissions. NASF s analysis indicates that a more accurate estimate of current emissions is approximately 200 pounds a year, or 80 percent lower that EPA s estimate. So if corrected, the results would be an even more impressive 99.9 percent reduction. As a result of these 11

12 reductions, the risks posed by emissions from chromium electroplating and anodizing operations are miniscule and the costs of attempting to achieve additional reductions are unreasonable. Such reductions in emissions and the associated risks represent an environmental policy success story that should be acknowledged and celebrated. Rather than imposing new regulatory burdens and costs on the chromium electroplating and anodizing operations with little if any environmental gains, EPA should work with the industry to continue the success through the implementation of work practices and the elimination of PFOS-based fume suppressants. VII. EPA S PROPOSED REVISIONS TO CHROMIUM ELECTROPLATING NESHAP ARE NOT CONSISTENT WITH THE CLEAN AIR ACT EPA s proposal to significantly reduce emissions standards for the plating industry sector is a significant and unexplained change from initial proposal in 2010 to retain the existing emission standards for this sector, and is inconsistent with the record and contrary to the Clean Air Act ( CAA ). In October 2010, EPA evaluated the chromium plating sector and determined that no new emission standards were warranted because of the low risk posed by emissions from the sector. 75 Fed. Reg In its revised proposal, EPA has now concluded that the risks due to the estimates of actual emissions are considerably lower than those presented in the October 21, 2010 proposal FR notice for hard chromium and decorative chromium plants and about the same for chrome anodizing. 77 Fed. Reg. at As discussed in detail below, even these new findings still significantly overstate current chromium emissions, and thus residual risk. Despite these new findings that the risks posed are considerably lower or about the same for the industry sector, and there being no evidence of new technological developments to control emissions, the Agency is now proposing to significantly ratchet down the emission limits that it previously found adequate not even two years ago. This makes no sense. 12

13 A. EPA Reversed the Proper Order of Standards Review under the CAA EPA s regulatory approach in this rulemaking reversed the proper analysis under the CAA. EPA first evaluated whether there were changes to the standards warranted under CAA section 112(d)(6) and, after proposing to set new standards under section 112(d)(6), then moved on to determine whether those newly proposed standards should be further revised under CAA section 112(f). However, EPA should have first analyzed risk under section 112(f) before determining whether to adjust the standards under section 112(d)(6). The history of CAA Section 112 makes clear that the technology-driven section 112(d)(3) standard-setting process was put in place to accelerate stalled progress towards imposition of health-based emissions standards. See Sierra Club v. EPA, 353 F.3d 976, (D.C. Cir. 2004). The CAA originally focused on a risk-based approach to setting emission standards for hazardous air pollutants. Id. But, finding that approach was not proving successful, Congress passed the 1990 CAA amendments and established a twophase approach to setting standards for hazardous air pollutants the first phase is to set technology based standards, followed by a second phase, intended to be initiated within eight years after technology-based standards were set, where the agency determines whether residual risks remain that warrant more stringent standards. Id. If, as is the case here, the initial technology-driven section 112(d)(3) standard-setting process results in a standard that is appropriately protective of human health, as judged by the residual risk review under section 112(f), there is no reason for further technologybased analysis under 112(d)(6). Put differently, EPA should have conducted the section 112(f) review of the existing standard first to determine if there was a residual risk of concern. The Agency should only have moved on to the section 112(d)(6) review if stricter standards were warranted based on the results of the section 112(f) review. EPA has previously explained this exact logical approach: 13

14 If EPA were to determine that standards adopted under section 112(f) (or section 112(d) standards evaluated pursuant to section 112(f)) provide an ample margin of safety to protect public health and prevent adverse environmental effects, one can reasonably question whether further reviews of technological capability are necessary (section 112(d)(6)). 69 Fed. Reg Here, however, EPA reversed the steps. It first proposed unnecessarily strict standards based on its section 112(d)(6) technology review which, as is discussed further below, were not warranted because EPA did not base the proposed lower emission standards on any new technology that has become available since the promulgation of the original rule. EPA then moved on to (not surprisingly) find those overly-strict standards met the risk requirements of section 112(f). Had EPA first evaluated the existing standards under 112(f), it should have concluded that those existing standards were appropriately protective, because it reached precisely that conclusion in October 2010, and EPA in this proposal concludes that the risk is even lower than what it believed then. Further, as discussed in detail elsewhere in these comments, the data demonstrate that the risk is even significantly lower than what EPA suggests now. There was no need to conduct a section 112(d)(6) review, (though, in any event, the record does not support a lower standard based on such a review). B. EPA Does Not Have the Authority under CAA Section 112(d)(6) to Establish the Strict Emission Standards Included in the Proposal EPA is proposing to require stricter standards based on its conclusion that the industry sector can achieve them at a relatively low cost. As the Agency explains: [EPA] made the decision to consider more stringent emissions limits than the limit in the current NESHAP primarily because the revised data set indicated that most facilities were operating well below the current emissions limit. This 14

15 indicated that more stringent emissions limits could be implemented without significant economic burden to the industry. 77 Fed. Reg This is not, however, an appropriate basis for revising emissions standards under section 112(d)(6). That section requires that any changes in the standards be driven by changes in practices, processes, [or] control technologies. CAA section 112(d)(6). The section does not allow EPA to change standards simply because portions of the industry are operating below existing standards or because compliance with new limits may not be cost prohibitive. EPA recently described its standard practice for conducting technology reviews as follows: Based on specific knowledge of each source category, we began by identifying known developments in practices, processes and control technologies. For the purpose of this exercise, we considered any of the following to be a development : any add-on control technology or other equipment that was not identified and considered during MACT development; any improvements in add-on control technology or other equipment (that was identified and considered during MACT development) that could result in significant additional emission reduction; any work practice or operational procedure that was not identified and considered during MACT development; and any process change or pollution prevention alternative that could be broadly applied that was not identified and considered during MACT development. 77 Fed. Reg. 1268, 1283 (Jan. 9, 2012); see also 77 Fed. Reg. 8576, 8593 (Feb. 14, 2012); 75 Fed. Reg , (October 21, 2010). EPA s consistent past practice and statements confirm that EPA should only set new standards under section 112(d)(6) if the Agency determines that new technology is 15

16 available. See e.g., 77 Fed. Reg. 1268, 1288 (Jan. 9, 2012) (deciding not to change the standards, as the agency identified no advancements in practices, processes, and control technologies ). 77 Fed. Reg. 8576, 8596 (declining to base new restrictions on new technology where it is not clear that this technology would reduce HAP emissions relative to technologies that were considered by the EPA in promulgating the subpart RRR NESHAP and are already used by other facilities ). In the October 2010 rulemaking, EPA identified emission elimination devices, high efficiency particulate air (HEPA) filters, wetting agent fume suppressants (WAFS), and housekeeping practices and determined that none of those technologies warranted setting lower emission standards. 75 Fed. Reg EPA has not, in this rulemaking, identified any additional practices, processes, [or] control technologies that were not identified and considered for this sector during the development of the original MACT or the 2010 rulemaking and that warrant stricter standards. Instead, EPA s technology analysis stops when the Agency concludes that facilities are achieving better emissions results than the current standard. See e.g., 77 Fed. Reg Once EPA reaches that conclusion, the Agency turns to creating hypothetical options for combining existing technologies to achieve those reduced emission results. Id. Rather than identifying a new technology that might be driving the emission results, EPA uses the emission results to drive the identification of possible combinations of existing technologies. Fundamentally, the agency sets the new standards based on cost, rather than technology, considerations. See e.g., 77 Fed. Reg. at That is not the appropriate method for conducting a technology review under 112(d)(6). Thus, while there have been continual incremental changes in the industry sector that may be resulting in reduced emissions, EPA has not identified any practices, processes, or control technologies that exist now that were not in use when the MACT standards were originally set. Further, EPA has not even analyzed whether, much less determined that, particular practices, processes, [or] control technologies are more widespread now than when the original standards were set. 16

17 Even assuming that EPA s data is correct (which it is not, as demonstrated below), EPA has simply demonstrated that some facilities are below the existing standard. This does not warrant the establishment of revised standards under section 112(d)(6). EPA should expect that some facilities will decide to reduce emissions below the existing standard in order to ensure a compliance buffer. EPA should not set the precedent that an industry that operates with a compliance buffer will be subject to ratcheting down of the standards. That would create the perverse incentive of encouraging industry sectors not to reduce their emissions at all below the existing MACT standards. Finally, EPA s authority under CAA section 112(d)(6) is limited by the terms of the statute to only making necessary revisions to existing standards. The term necessary in section 112(d)(6) sets a very high standard for when changes may be made and EPA has offered no basis for concluding that its proposed changes are necessary. C. Section 112(d)(6) Does Not Require EPA to Engage in a Section 112(d)(3) Benchmarking Exercise NASF agrees with EPA s conclusion that it would be inappropriate for the Agency to import the strict benchmarking required by section 112(d)(3) into the periodic section 112(d)(6) reviews. Doing so would create a one way ratchet for MACT standards that would be in violation of the clear statutory mandate. The D.C. Circuit has already rejected the argument that section 112(d)(6) requires EPA to completely recalculate the maximum achievable control technology in other words, to start from scratch. NRDC v. EPA, 529 F.3d 1077, 1084 (D.C. Cir. 2008). The setting of a MACT floor under section 112(d) is a one-time requirement. Section 112(d)(3) explains the process and requirements for promulgation of a standard and carefully lays out how a MACT standard is set in the first instance. For example, the standards for existing sources shall not be less stringent than the average emission limitation achieved by the best performing 12 percent of the existing sources. CAA Section 112(d)(3). Section 112(d)(6), however, speaks only of revision of standards 17

18 that have already been promulgated and establishes its own parameters for such revisions they must be necessary and must tak[e] into account developments in practices, processes, and control technologies. 1 D. EPA Correctly Decided on the Framework to Analyze Whether to Adjust the Standards under Section 112(f) NASF also agrees with EPA s conclusion that maximum individual risks as high as 100 in one million can be presumptively considered safe or acceptable. In order to reach that conclusion, EPA correctly relied on the 1987 Vinyl Chloride decision (NRDC, Inc. v. EPA, 824 F.2d 1146 (D.C. Cir. 1987)), and the 1989 Benzene Rule (54 Fed. Reg. 38,044 (Sept. 14, 1989)), codified by statute under section 112(f)(2)(B). This approach to residual risk MACT rulemakings has been upheld by the D.C. Circuit, and EPA has been consistently applying it. NRDC v. EPA, 529 F3d (D.C. Cir. 2008); 77 Fed. Reg. at VIII. EPA HAS GREATLY OVER-ESTIMATED CHROMIUM EMISSIONS One of the most important components in EPA s analysis to support the proposed regulations is the Agency s estimates of current emissions of hexavalent chromium from each of the more than 1,300 electroplating and anodizing facilities outside of California. Facilities in California are already subject to emission limitations more stringent than what EPA has proposed for the rest of the nation, and virtually all of the costs and benefits of the proposed rule will result from the proposed tighter regulations for non- California facilities. EPA s estimates of current emissions from the non-california facilities provide the starting point for most of EPA s analyses supporting the proposed 1 In other places in the CAA, Congress has expressly provided EPA authority to revise standards in the same manner as the initial promulgation with respect to other standards in the Act. See, e.g., 42 U.S.C. 7409(b)(1) ( Such primary standards may be revised in the same manner as promulgated. ); 7409(b)(2) ( Such secondary standards may be revised in the same manner as promulgated. ). Congress did not do so in CAA section

19 rule: the risk analysis, the environmental justice analysis, the estimates of emissions reductions and risk reductions, the cost analysis, the cost-effectiveness analysis, etc. EPA has estimated current actual emissions for the non-california facilities in two steps. EPA first estimated annual emissions for each of 170 non-california facilities for which the Agency believed adequate site-specific information was available, and EPA then extrapolated from these results to estimate emissions for the remaining 1,149 non- California facilities about which virtually no site-specific information was available. How EPA Estimated Actual Annual Emissions for the 1,319 Non-CA Facilities # of Non-CA Facilities EPA claims to have all the facility-specific data needed to estimate emissions 113 EPA claims to have some of the needed data (+ need some default assumptions) 57 EPA has none of needed site-specific data (each facility s estimate is extrapolated based on estimates for the 170 facilities above) * Source: EPA. Simulation of Actual and Allowable Emissions for Cr Electroplating Facilities. Table 1 Significant flaws, misleading information and faulty assumptions plague both steps in EPA s process, thereby producing inaccurate and unreliable emissions estimates. On the whole, EPA s various errors result in emissions estimates that are far too high for most facilities and for the industry as a whole. 1,149 1,319 We reviewed several aspects of EPA s emissions estimates and report the results of our review in these comments. We reviewed EPA s documents describing the Agency s procedures for generating the emissions estimates and the resulting estimates for the 1,319 non- California facilities. The primary documents that we reviewed include: 1) Simulation of Actual and Allowable Emissions for Chromium Electroplating Facilities (January 5, 2012); 2) Profile of Chromium Electroplating Processes 19

20 and Emissions (January 6, 2012); and 3) Information on Chromium Electroplating Facilities Collected from State and Local Agencies (December 19, In the following pages we provide several comments based on this review of EPA s documentation. We reviewed the underlying data and calculation procedures that EPA used in estimating actual annual emissions for several of the 170 non-california facilities for which the Agency developed site-specific emissions estimates. In effect, we spot-checked some of the information and methods that EPA used in a few of the instances where the Agency believed that available data were sufficient to allow a reasonably confident estimate of a facility s current emissions. We would like to have had the time to do much more of this spot-checking. We conducted a major effort to assess and either verify or correct EPA s emissions estimates for each of the 300 non-california facilities that EPA has estimated as posing MEI risks of 1 x10-6 or more. We focused this review specifically on higher risk facilities because that is also what EPA had done in last year s request to State and local agencies. EPA had requested additional and presumably better information from State and local agencies on these higher risk facilities; we thought that we should evaluate EPA s resulting emissions estimates for exactly these facilities for which the Agency had received presumably better and more accurate information. At present, we have obtained responses regarding 180 of these 300 non-california facilities. We would have obtained more responses and could also perhaps have evaluated EPA s estimates for some of the lower risk facilities had there been an extension of the comment deadline. We report some of the results of these three reviews in the following pages. 20

21 A. Spot-Checking Some of EPA s Data and Procedures in Using Site-Specific Information to Estimate a Facility s Actual Annual Emissions EPA uses site-specific information to estimate actual annual emissions for 170 non- California facilities. To estimate actual annual emissions for each facility, EPA uses one of three approaches: For a little less than half of the facilities, EPA estimates annual emissions by multiplying the facility s estimated hourly emission rate by the number of hours per year that the facility operates. In some instances EPA multiplies three quantities: 1) chromium concentration in the stack gas, 2) stack gas flow rate, and 3) the annual number of operating hours (hours during the year that the plating tanks at the facility are energized and producing potential emissions). In other instances EPA multiplies only two quantities: 1) the facility s estimated hourly emission rate, and 2) the facility s annual number of operating hours. Again for a little less than half of the facilities, EPA instead has somehow obtained a direct estimate of the facility s actual emissions over the course of a year. Typically a State agency has at some time in the past generated an estimate of the facility s annual emissions, and EPA simply accepts that as an estimate of the facility s current annual emissions. For about 20 facilities, EPA has site-specific data on a facility s operating hours and/or stack gas flow rate, but no information on emissions concentration or hourly emissions rate. For these facilities, EPA has estimated the plant s emission concentration by assuming that the plant emits exactly at the applicable concentration limit allowed by the existing NESHAP. This approach for purportedly estimating actual emissions instead estimates the facility s allowable emissions. All facilities for which EPA uses this approach should be dropped from the list of plants for which EPA claims to have estimated actual 21

22 annual emissions based on site-specific data. For these plants, EPA has estimated only allowable annual emissions, and has done so without having site-specific data on perhaps the most important quantity -- emissions concentration -- needed for estimating emissions. (See Table 5 of EPA s Profile of Chromium Electroplating Processes document and Table 1 of EPA s Simulation of Actual and Allowable Emissions document for a list of the 170 non-ca facilities for which EPA claims to have used site-specific data in estimating actual annual emissions and identification of the method used for each of these facilities.) We wanted to spot check the data and procedures that EPA used in estimating sitespecific actual emissions for several dozen facilities where EPA used the first approach and several dozen where EPA used the second approach. (The third approach is clearly inappropriate on its face, and we do not need to review the estimates for facilities where it has been used.) We were not able to complete this much spot checking during the comment period. We have completed the following reviews: For the first approach, we chose to review the data EPA had acquired and the conclusions the Agency drew regarding operating hours; the only one of the quantities needed for emissions estimation that is required for all of the emissions estimates that EPA developed via the first approach. We decided to review EPA s data on operating hours for two of the four industry subcategories -- anodizing and decorative chrome. We selected these two subcategories to limit our workload; EPA claims to have acquired site-specific data on operating hours for 21 anodizing or decorative chrome facilities, whereas the Agency has acquired such data for more than 30 hard chrome facilities. 22

23 For the first approach, we also reviewed for one facility all of EPA s data and calculations from start to finish leading to an estimate of actual annual emissions for that facility. We randomly selected Chrome Crankshaft Company of Illinois as this one facility. For the second approach, we randomly selected for review three facilities for which EPA claims to have obtained from State agencies a direct estimate of the facility s actual emissions over the course of a year. 1. Operating Hours Data EPA claims to have reasonably accurate facility-specific estimates of hours of operation for more than fifty facilities. We believe that it would be difficult for EPA to secure accurate information regarding annual operating hours for a facility without requesting this information directly from facilities through a survey which EPA has not done in this rulemaking. We wondered how EPA had obtained the data that the Agency says it has, whether the data appears accurate, and how EPA has used the data. We checked EPA s data on operating hours specifically for the anodizing and decorative chrome facilities for which EPA claims to have used information on the facility s operating hours in estimating a facility s actual emissions (see Table 1 of EPA s Simulation of Actual and Allowable Emissions document). There are 21 facilities in this Table for which EPA claims to have developed site-specific information on a facility s operating hours from data provided by State and local agencies. We conclude that EPA s data on operating hours appears to be inaccurate/wrong for 16 among these 21 facilities: for 8 facilities, the information shown in the table is imputed rather than actual (i.e., not obtained for the specific facility in question), as denoted by shaded (imputed) cells in the table; 23

24 for 2 more facilities, the information shown in the table appears to be incorrect (different from that shown in Table 19 of EPA s Profile document -- the supposed source document for the operating hours data in Table 1) 2 ; and for 6 more facilities, the information shown in the table is clearly incorrect because it doesn t incorporate the assumption that the facility energizes its electroplating tanks for only 60 percent of the hours that the facility operates (decorative chrome). The result is that for only 5 of the 21 facilities is the operating hours data really as advertised by EPA. For 16 of the 21 facilities, the operating hours data that EPA uses in estimating emissions is either not specific to the facility that is being represented or not developed appropriately and accurately from the information provided by State and local agencies. The net result of these errors is to overstate these facilities operating hours, which would in turn result in over-estimating these facilities actual emissions, which in turn would result in over-estimating emissions for the remaining 1149 non-ca facilities comprising the remainder of the industry. 2. Review of EPA s Several Data Elements Used to Estimate Annual Actual Emissions for Chrome Crankshaft Co. of Illinois EPA estimated actual annual emissions for this small hard chrome plating facility in Chicago, IL by combining information on stack flow, operating hours and emissions concentration. In Table 1, EPA shows this facility as having stack flow of 44,360 dscfm (purportedly actual data for the facility, since this cell in the Table is not shaded), operating hours/yr. of 2,080 (identified as EPA as imputed data rather than specific to the facility since this cell is shaded; presumably 2,080 operating hours is some sort of 2 For Co-Operative Plating Co., the number shown for operating hours in Table 1 appears to be the number for flow rate which is incorrectly transcribed into the operating hours column. For Delphi Corporation, the number shown for operating hours in Table 1 is not shown in the source document, Table 19. In fact, Delphi Corporation does not appear at all in Table 19, and there is no indication where EPA might have obtained the estimate of 5,880 operating hours/yr. for Delphi Corporation that appears in Table 1. 24

25 average for small hard chrome facilities), and average emissions concentration of mg/dscm (purportedly actual data for the facility, since this cell is not shaded). In fact, EPA received from Illinois a copy of a stack test report from July, 2001 for this facility. The docket reflects no other information provided by the State or otherwise available regarding this facility. A first question is whether a stack test in 2001 accurately reflects a facility s current emissions. Many facilities have changed their plating processes and equipment during the past 11 years, and many have improved their capture and/or control equipment for chromic acid mists, particularly due to OSHA s tightened worker exposure standard for hexavalent chromium. The test report provided by Illinois does not indicate the facility s stack flow. In fact, the figure of 44,360 that EPA characterizes as actual facility data is instead imputed; this figure represents the facility s 4 stacks each having the imputed industry-wide average estimated stack flow of 11,090 dscfm. This cell should be shaded in EPA s table. We wonder how many additional instances of supposed facility-specific data are in fact imputed. The test report does not indicate the facility s actual operating hours. EPA characterizes the figure shown of 2,080 as imputed rather than specific to the facility. EPA nowhere indicates how this particular imputed figure was derived. We suspect that the 2,080 hour estimate represents some assumption that some class of facilities operates 40 hours per week for 52 weeks per year = 2,080 operating hours. If this in fact is the derivation of this estimate, the estimate should be, but is not, further scaled down by 20% to reflect EPA s assumption that rectifiers are energized and potential emissions are generated for only 80% of the hours that a hard chrome facility operates (see page 5 of EPA s Profile document). EPA s estimate for Chrome Crankshaft of 2,080 operating hours per year appears to be both imputed instead of site-specific and incorrect as an imputation. 25

26 We can replicate from the 2001 test report how EPA may have calculated that this facility had an average emissions concentration of as the simple average across the test results for the facility s four stacks. Instead, the more appropriate concentration would be the flow-weighted average across the four stacks. More importantly, though, the test report provides a statement to the effect that under the test conditions the tanks were run at maximum rectifier settings (maximum current) and thus generated a maximum emissions concentration (emissions are a linear function of current applied). The tanks are typically run at less than maximum amperage. The concentrations observed during these stack tests thus likely exceed those that would prevail on average over all the hours that the tank rectifiers are operated hours during the year, and it is this lower average emissions concentration that should be used in estimating this plant s actual annual emissions. This is likely an issue in all of the many instances in which EPA has used stack test emissions concentration information in estimating a plant s actual annual emissions. Most stack tests are run as stress tests to see whether emission limitations are met under worst case operating conditions. Most facilities will emit at average levels significantly below their emission concentrations observed during stack tests. More reasonable assumptions should be used to estimate actual emissions from real facilities. In sum, for Chrome Crankshaft Co. of Illinois, none of the figures that EPA provides for the three data elements used in estimating this facility s actual annual emissions -- stack flow, emissions concentration and operating hours -- accurately represents current conditions at the facility. Each of the errors that EPA makes in estimating these quantities serves to over-estimate this facility s current actual annual emissions. 26

27 3. Direct Estimates of Actual Annual Facility Emissions Supposedly Obtained from State/Local Agencies EPA also claims to have received direct estimates of facility emissions from state and local regulatory agencies. Based on a spot check for three such facilities, the rulemaking docket does not include any such information. A discussion of these examples is provided below. a. Jobsite, Inc. of Grand Junction, CO (hard chrome, large) In Table 1, EPA shows this facility as having an emission rate of 2.7 lbs./yr., with this estimate having been provided directly by the state instead of being calculated by EPA using information on concentration, flow rate and operating hours. We searched the information provided by Colorado to EPA and searched the docket more generally, but in neither could we find any information with respect to Jobsite, Inc. There appears to be no information in the record to support EPA s estimate of 2.7 lbs./yr. emissions for this facility. b. Dover Industrial Chrome, Inc. of Chicago, IL (hard chrome, large) In Table 1, EPA shows this facility as having an emission rate of 4.76 lbs./yr., with this estimate having been provided directly by the state instead of being calculated by EPA using information on concentration, flow rate and operating hours. We found the spreadsheet provided by Kevin Mattison of Illinois EPA to the U.S. EPA with information about numerous electroplating sources of hexavalent chromium emissions in Illinois, including Dover Industrial Chrome, Inc.. Nowhere in the spreadsheet is there any mention of the emission rate or concentration or operating hours or stack gas flow rate for Dover Industrial Chrome. The data provided in the spreadsheet for this facility includes only: a) 27

28 information to the effect that the facility has six hard chrome plating tanks; and b) information to the effect that the permittee claims the use of fume suppressant to lower the chrome plating solution surface tension to dynes/cm reduces chromic acid vapor emissions by 97 to 99 percent. There appears to be no information in the record to support EPA s estimate of 2.7 lbs./yr. emissions for this facility. Perhaps EPA has somehow imputed an estimate of annual emissions for this facility by assuming some typical average uncontrolled rate of emissions for six hard chrome plating tanks and then reducing this figure by 97 to 99 percent. If so, the Agency should not claim this to be a facility-specific estimate of annual emissions provided by a State or local agency. c. Starrett Company - Evans Rule Division of Charleston, SC (decorative chrome) In Table 1, EPA shows this facility as having an emission rate of 4.00 lbs./yr., with this estimate having been provided directly by the state instead of being calculated by EPA using information on concentration, flow rate and operating hours. We looked for information provided by State or local agencies in South Carolina to EPA in the Agency s document on Information on Chromium Electroplating Facilities Collected from State and Local Agencies. We found that South Carolina had neither been asked to provide information, nor had they provided any. We searched the docket more generally, but could not find any information with respect to Starrett Company - Evans Rule Division. 28

29 There appears to be no information in the record to support EPA s estimate of 4.0 lbs./yr. emissions for this facility. In sum, for none of the three facilities that we investigated did we find that EPA had accurately used facility-specific information from State or local agencies to develop the Agency s estimates of annual emissions for these facilities. For only 170 of the non-ca facilities that will be affected by the proposed rule does EPA claim to have site-specific information with which to reasonably estimate the facility s annual emissions. Based on our spot-checking, including 21 estimates of operating hours and 4 estimates of facility emissions, EPA is likely incorrect in claiming that the Agency has accurate, representative and documented site-specific data with which to estimate these emissions. Accurate emission estimates are key to analyzing risks, risk reduction, environmental justice issues, costs, benefits, and economic impacts. Despite last year s data collection effort from State and local agencies, EPA has very little of the needed information. EPA also appears to misuse much of the limited information the Agency has most of the information that EPA has on this industry is at least five years old and misses the important impacts of the following trends in the industry: compliance with OSHA s tighter PEL, shift from use of hexavalent chromium to trivalent to avoid toxicity issues, and the continued shrinking of U.S. manufacturing and the recent recession. The quality of the data relied upon by EPA has serious deficiencies. The major reason why EPA has insufficient information is the Agency s failure to conduct a recent Information Collection Request or industry survey. For most of the recent residual risk and technology reviews where EPA has sought to significantly tighten the NESHAP, the 29

30 Agency has appeared to find it advisable to develop a good information basis for the regulation via a recent ICR or joint survey with industry. EPA s failure to collect the data necessary to support this rulemaking, and nonetheless proposing to impose stricter standards even after it concluded that emissions were lower than previously estimated, is an arbitrary and capricious exercise of the Agency s authority. B. EPA s Extrapolated Emissions Estimates For the Remaining 1,149 Non- CA Facilities Are Erroneous EPA uses the imputed, inaccurate facility-specific data discussed above to calculate emissions estimates for the remaining 1,149 facilities for which the Agency has virtually no site-specific information. In addition, for many of these facilities, EPA has no information on their size or on the plating processes that they conduct. Due to confusing and sometimes conflicting documentation in the record, it is difficult for us to understand and then comment on how EPA performed this second extrapolation step in the Agency s emissions analysis. A description of some of this confusion is summarized below. In the January 6, 2912 Data Profile document (page 4), EPA indicated the following. [A]verage emissions for the non-california plants were not used in any of the analyses, other than as a basis of comparison to the simulated annual emissions for the non-california plants. The average emissions for the California plants were used as default values in the risk assessment for those plants for which actual emissions were unknown or for which the process type was unknown. In the latter case, it was assumed that any unknown plant was a hard chromium electroplating facility. This appears to suggest for the current analysis that EPA is still using a procedure like that for the October 2010 proposal, in which emissions for facilities of unknown size and process are estimated using conservative (i.e., potentially too high relative to actual conditions) default values, in this case, average emissions from large hard chromium facilities. 30

31 However, in the Simulation of Actual and Allowable Emissions document (page 6), EPA indicated the following. Because emissions were to be simulated according to process type, it was necessary to assign a process type to each of these "unknown" plants. a distinction was made between chromium electroplating plants located in California and plants located elsewhere The reason for this is that plants in California are subject to emission limits that are more stringent than the emission limits specified in the Chromium Electroplating NESHAP, and annual emissions from chromium electroplating plants located in California are significantly lower than emissions from non- California plants. Therefore, emissions from California plants are not representative of emissions from non-california plants. Based on the total numbers of plants, plant types were randomly assigned to each of the unknown plants, while ensuring that the total numbers of each type of plant nationwide were preserved. This appears to suggest that EPA followed an entirely different procedure for the February 2012 supplemental proposal. The non-california unknown plants are randomly assigned to different plant types in proportions reflecting the national totals. Emissions for the unknown non-california plants are estimated in some manner that reflects emissions from the relatively few non-california plants for which EPA has developed emissions estimates. Emissions from California plants are not used in any way in estimating emissions from unknown non-california plants. NASF supports EPA s decision to exclude California facilities from the analysis because they are not representative of chromium electroplating and anodizing facilities nationwide. It is not clear which of these two different approaches described above did EPA use to calculate emission estimates for the facilities for which the Agency does not have any emissions information, does not know the size of the facility, or does not know which chromium processes are conducted at the facility, if any. Based on information collected by NASF regarding emissions estimates, it appears that EPA may have categorized unknown facilities as hard chromium, rather than randomly assigning process types to unknown facilities. A more detailed discussion on this topic is provided below. 31

32 1. Monte Carlo Analysis EPA recognized that the Agency needed to find some way to assign emissions estimates to the large number of facilities for which it has little or no facility-specific information. EPA claims to have used a Monte Carlo analysis to assign these emissions estimates. As best we can tell, EPA appears to have used Monte Carlo procedures only to estimate what the n emission values would be if there were n plants in a subcategory (subcategory meaning hard chrome/large, hard chrome/small, etc.). For all the 1,149 plants for which EPA had neither a facility-specific emission estimate nor full information on process and size, EPA then, we believe: 1) Assigned each facility randomly to a subcategory, with the number of facilities assigned to each subcategory being chosen so as to yield the appropriate industry-wide total # of plants for each subcategory; and 2) Assigned the n emission values to the plants assigned as being within a subcategory by using an unusual approach involving alphabetical ordering -- EPA assigned emissions estimates ranging from low to high according to the alphabetical order among facilities. EPA s major failing with regard to Monte Carlo was using it for only a less important uncertainty in the analysis, and not using it on the major uncertainty which is what subcategory an unknown plant really is in. The alphabetical approach by which emissions estimates were ultimately assigned is odd, obviously inaccurate, and inexplicable. For each subcategory of facilities, EPA s estimated emissions increase with facility names in alphabetical order. The last facility in the alphabet in each subcategory was assigned emissions by EPA ranging from 5 to 761 times the emissions of the first facility in the alphabet. Table VIII-1 below summarizes EPA s emissions estimates for the 1,149 facilities for which it does not have any facility-specific data. 32

33 TABLE VIII-1: EPA s Emissions Estimates for Facilities for Which EPA Has No Site-Specific Data Facility Name Process Type Simulated As Estimated Ratio of Emissions: Emissions (lbs/yr) High vs. Low Cr anodizing Cr anodizing Advanced Coating Techniques to to Valley Plating, Inc Decorative Cr Decorative Cr A&A Bumper Plating to to Xtreme Metal Finishing Hard Cr, unknown Hard Cr, large \AH\ Metal Fabricators 1.82 to to Youngstown Hard Chrome Plating Hard Cr, unknown Hard Cr, small 3-M to to Worldcolor Richmond Unknown process Cr anodizing A-1 Plating Co., Inc to to Yellowstone Hydraulics Unknown process Decorative Cr Acadia Polishing and Plating to to Xerox Joseph C Wilson Center Unknown process Hard Cr, large AAA Hydraulic Cylinder Services to to Wearmaster Unknown process Hard Cr, small ABC Custom Chrome to to Worthington Steel Co While this unusual alphabetical process does not necessarily bias the overall emissions estimates high or low, the pattern is nonsensical and without justification. Facilities at the beginning of the alphabet are assigned such low emissions estimates that they have 33

34 very little impact on any of the parameters that EPA must analyze such as baseline risks, risk reduction, costs and cost effectiveness. Conversely, facilities at the end of the alphabet would have very high emissions estimates and a disproportionately large impact on the parameters that EPA must analyze. EPA could have achieved equivalent irrelevant results by simply throwing darts at a board. Using models to assign values to unknown variables is a necessary process for many rulemakings. Nonetheless, even the most elegant and sophisticated models must generate relevant and useful information to predict a reasonable representation of a regulated industry. EPA s methodology for estimating emissions for the non-california facilities demonstrates the Agency s extreme lack of knowledge about the specific chromium electroplating and anodizing operations that the Agency seeks to regulate and provides no useful information or any reasonable simulation or prediction of real-world emissions from each of these facilities. The results from this process cannot be used reliably to impose additional regulatory controls on chromium electroplating and anodizing operations. EPA could have produced more reliable results if it had focused on other more relevant variables such as facility size and chromium processes for the Monte Carlo analysis. 2. NASF Data Collection Effort to Verify Facility Emissions Estimates In addition to critiquing EPA s flawed methodology for estimating emissions, NASF has collected data from facilities to evaluate and perhaps verify the emissions estimates assigned to facilities. The process and results from this effort are summarized below. EPA has summarized its emissions estimates and risk calculations for 1,455 facilities in its data base. Specifically, EPA provided facility name, address, process type, annual emissions estimate, and a maximum individual risk value for each facility. NASF reached out to its members and requested information on approximately 300 of the highest emissions facilities identified by EPA, most of which are hard chromium facilities. 34

35 As of the date of these comments, NASF has received information on130 facilities. The information was provided by facilities directly, persons knowledgeable about the facility, phone conversations, research of public records, and site visits to confirm the information. The results indicated that among the 130 facilities responding: 35 percent were closed, 38 percent no longer, or never did, use hexavalent chromium, 25 percent provided data that showed emissions lower than EPA s estimate, and 2 percent confirmed EPA s emissions estimate. These responses represent a 94 percent reduction from EPA s estimates from an estimated total of 325 pounds to only 19 pounds for these 130 facilities. The results of the NASF data collection are provided in Attachment 1 and summarized in the diagram below. a. Facility Closures The results from NASF s data collection effort demonstrate two critical trends in the chromium electroplating and anodizing industry sectors: facility closures and a switch 35

36 away from hexavalent chromium processes. First, the surface finishing industry has suffered through the decline of U.S. manufacturing over the past two decades. As manufacturing operations closed and left the U.S., the market for domestic plating services significantly declined, particularly through the recent recession of In addition, rising costs of raw materials (e.g., metals prices), increased operating costs, higher energy prices, more intense global competition for manufacturing to relocate abroad, and mounting regulatory burdens have forced a large number of plating operations to close In its comments on the October 2010 proposed rule, NASF estimated facility closures in the range of 20 to 30 percent. The 35 percent closure rate from this data collection effort may be conservative because the focus was on the larger, higher emissions facilities. Closure rates among smaller, lower emissions facilities (including many decorative chromium and chromium anodizing facilities) could be even higher due to the lower economies of scale for these facilities. Nonetheless, a closure rate of 35 percent would be a reasonable estimate across the all of the chromium facilities in EPA s data base. b. Facilities That No Longer Use Hexavalent Chromium The second trend from the results of the NASF data collection effort was 38 percent of the facilities that responded indicated that they no longer use hexavalent chromium in their processes. This trend can be explained in part because surface finishing companies are facing a growing number of global and domestic regulatory pressures to cease chromium electroplating and anodizing processes, including the European Union (EU) End of Life Vehicle (ELV) directive, the EU Restrictions on Hazardous Substances (RoHS) directive, and the EU Waste Electrical and Electronic Equipment (WEEE) directive. Because of the direct bans and restrictions on the use of hexavalent chromium as well as market de-selection pressures stemming from the restrictions, many facilities have chosen to abandon chromium electroplating and anodizing processes. 36

37 The U.S. Department of Defense (DOD) has also initiated efforts to eliminate the use of hexavalent chromium in the defense supply chain. Because many chromium electroplating and anodizing processes have been defined as mission critical in military applications, DOD has not yet banned the use of hexavalent chromium for military corrosion protection and surface technology. Nonetheless, pressures continue to mount to find alternatives to chromium electroplating and anodizing processes for defense applications. Finally, the Occupational Safety and Health Administration (OSHA) promulgated a more stringent workplace exposure standard for hexavalent chromium in February In response to this new standard, many metal finishing operations implemented additional controls and practices to minimize the release of hexavalent chromium into the workplace and the environment. The additional workplace exposure requirements also prompted many facilities to eliminate potential hazards to employees by switching to processes that did not include hexavalent chromium. Like the results on facility closures, the trend toward no longer using hexavalent chromium is universal and can be reasonably expected to occur throughout the industry. In fact, the trend of no longer using hexavalent chromium for decorative chromium may be even greater than the rate for hard chromium electroplating because more alternatives are available for decorative chromium than are available for hard chromium. Consequently, a 75 to 80 percent reduction in EPA s estimated emissions would be reasonable based on the rates for facility closures and switch from hexavalent chromium processes from the NASF data collection effort and the overall trends in the industry on these two criteria. c. Facilities That Never Had Hard Chromium Processes One of the other more common responses from facilities in the NASF data collection effort was that the facility never had hard chromium processes despite being identified by EPA as a hard chromium facility. In addition to this common response, the distribution 37

38 of hard chromium facilities in EPA s estimated emissions analysis was significantly higher than in other distributions prepared by EPA in the record for this rule. For example, in the preambles to the October 2010 proposal and the February 2012 supplemental proposal, the number of hard chromium facilities was very similar to the number of decorative chromium facilities. Yet, in the estimated emissions document, the number of hard chromium facilities is 2.5 times more than the number of decorative chromium facilities. See, Table VIII-2 below. TABLE VIII-2: Distribution of Facility Process Categories October 2010 Rule Preamble February 2012 Rule Preamble Estimated Emissions Model Hard Decorative Anodizing The responses from facilities that indicated that they never had hard chromium processes together with the disproportionately high number of hard chromium facilities in EPA s emissions estimates suggest that EPA may have identified most of the facilities in the unknown categories as hard chromium. Given that the highest emissions are expected from hard chromium processes, EPA s apparent assumptions for unknown facilities would tend to over-estimate emissions significantly. d. Revised Emissions Estimates Provided by Facilities As discussed above, facility closure and switch from chromium processes accounted for approximately 75 percent of the 94 percent reduction in EPA s estimated emissions. The remaining 25 percent of the reductions resulted from revised data provided by facilities on annual emissions. EPA officials have asked NASF if it could find information on the lower emissions facilities to determine if those emissions estimates are lower than the actual emissions. NASF requested an extension of the comment period, in part, to verify 38

39 the emissions estimates for the lower emissions facilities. EPA, however, denied the request for the extension, so NASF did not have the time to collect the data in response to EPA s question. NASF will continue to work with EPA to provide additional information on this issue, information that EPA indicated in its letter of March 22, 2012 it would accept and take into account in this rulemaking. If in assigning emission estimates alphabetically to the 1,147 extrapolated non-ca facilities EPA has both: 1. Performed the Monte Carlo analysis correctly so as to accurately estimate the n emissions estimates for the n facilities in each subcategory; and 2. Assigned an appropriate number of unknown size or process facilities to each subcategory so as to match the expected actual numbers of facilities in each subcategory. Then the too-high emissions estimates that we find for the higher-emitting, higher-risk facilities that we have investigated could in part be counterbalanced by emissions estimates that we would find as being too low if we were also to investigate the loweremitting, lower risk facilities. The reduction in emissions resulting from the revised data provided by the facilities that we have investigated could be counterbalanced by an increase in emissions from the lower end of the distribution. This would still leave, however, the very significant 73 percent reduction resulting from facility closures and switch from hexavalent chromium process that would still be expected throughout the entire distribution of facilities. If the EPA emissions estimates are biased high because more unknown facilities were incorrectly identified as hard chromium facilities, then the reductions at the high end of the emissions distribution would be greater than the increases at the low end of the emissions distribution. Such a result is more likely than the purely random, normal distribution based on the disproportionately high number of hard chromium facilities in 39

40 EPA s emissions estimates and the responses from facilities that indicated that they never had hard chromium processes. Accordingly, it would be reasonable to conclude that EPA s nationwide annual emissions estimate of 1,140 pounds should be reduced, conservatively, by 80 percent (less than the total 94 percent reduction that we have found for the higher emissions facilities that we have investigated, but more than the 73 percent reduction from closures and switch from hexavalent chromium), resulting in a revised current emissions estimate of 228 pounds, a massive decrease from EPA s estimate of 1,140 pounds and approximately 4.5 times the emissions decrease EPA hopes to achieve by imposing a more stringent standard. In summary, EPA has proposed a more stringent standard to reduce 208 pounds of emissions. The analysis discussed in detail above demonstrates that emissions nationwide are over 900 pounds lower than EPA had estimated that is 4.5 times greater than the emissions and risk reduction that EPA had sought with these new standards. Based on these substantial reductions in emissions for chromium electroplating and anodizing facilities and the associated risks, it is not clear why EPA would continue to pursue this proposed regulation. IX. EMISSIONS REDUCTION ESTIMATES In the preamble to the supplemental proposal, EPA estimates that the revisions to the NESHAP will result in a reduction of 208 pounds of emissions annually from its estimate of 1,140 pounds of current emissions. 77 Fed. Reg. at A summary of EPA s emissions reduction estimate is provided below in Table IX-1. 40

41 TABLE IX-1: EPA s Emissions Reduction Estimate (Pounds) Emissions Before Emissions After Emissions Reduction Large Hard Small Hard Decorative Anodizing TOTAL 1, As discussed in detail above, EPA has grossly exaggerated the emissions estimates. Even assuming for purposes of this following analysis that EPA s emissions estimates are correct, a closer examination of EPA s estimated emissions reductions indicates that EPA has over-estimated the emissions reductions expected from this rule. First, EPA claims that all decorative and anodizing facilities already meet the proposed emissions limits, so the only reductions from these two categories would result from the lowering of the surface tension levels. Lowering the surface tension levels of 35 dynes/cm to 33 dynes/cm is a 5.7 percent reduction, and 45 dynes/ cm to 40 dynes/cm is an 11.1 percent reduction. As discussed in detail below in Section XIII, reduction of surface tension levels may not reduce emissions at all, particularly with respect to non-pfos fume suppressants. Furthermore, even though the reduction in surface tension does not equate to a linear reduction in emissions (if any reduction in emissions), for purposes of this analysis assume a linear relationship between the two. A 5.7 percent reduction of 280 pounds for decorative chromium equals 16 pounds and an 11.1 percent reduction equals 31 pounds. Using a midpoint of these reductions would yield an emissions reduction of 24 pounds for decorative chromium. 41

42 Similarly, a 5.7 percent reduction of 66 pounds for anodizing equals 4 pounds and an 11.1 percent reduction equals 7 pounds. Using a midpoint of these reductions would yield an emissions reduction of 6 pounds for chromium anodizing. The calculation for hard chromium is more complicated because both reductions from lower surface tension levels and lower emission limits should be considered. First, in Table 4 of the August 16, 2010 document entitled, Nationwide Costs for Conversion to Non-PFOS Fume Suppressants, EPA estimated that approximately 20 percent of small hard chromium facilities (i.e., 111 facilities) and 10 percent of medium hard chromium facilities (i.e., 16 facilities) use fume suppressants. Again for simplicity, assume that these facilities are complying with the NESHAP by surface tension levels and the reduction is linear. In addition, because not all hard chromium facilities are reducing emissions through lower surface tension levels or lower emission limits, we shall use EPA s estimate of average annual emissions of 2.62 pounds for large hard chromium facilities and 0.56 pounds for small hard chromium facilities that are cited in the preamble to the supplemental proposal. 77 Fed. Reg. at small hard chromium facilities with average emissions of 0.56 pounds would be 62 pounds. A 5.7 percent reduction of 62 pounds for small hard chromium equals 4 pounds and an 11.1 percent reduction equals 7 pounds. Using a midpoint of these reductions would yield an emissions reduction of 6 pounds for small hard chromium. 16 large hard chromium facilities with average emissions of 2.62 pounds would be 42 pounds. A 5.7 percent reduction of 42 pounds for large hard chromium equals 2 pounds and an 11.1 percent reduction equals 5 pounds. Using a midpoint of these reductions would yield an emissions reduction of 4 pounds for large hard chromium. Instead of lower surface tension levels, many hard chromium facilities would choose to comply with lower emission limits: 25 percent lower for large hard chromium and 50 percent lower for small hard chromium. EPA estimates that 41 large hard chromium and 85 mall hard chromium facilities will need to lower emissions to meet the new proposed 42

43 emission limits. Accordingly 41 large hard chromium facilities with average emissions of 2.62 pounds are 107 pounds. A 25 percent reduction of 107 pounds is 27 pounds. Similarly, 85 small hard chromium facilities with average emissions of 0.56 pounds are 48 pounds. A 50 percent reduction of 48 pounds is 24 pounds. Based on this analysis (which includes assumptions that are all more favorable to EPA with a higher emissions reduction estimate), EPA s revised emissions reduction should be at most 91 pounds instead of 208. A summary of this emissions reduction analysis is provided below in Table IX-2. TABLE IX-2: Revised EPA Emissions Reduction Estimate Process Emissions Reductions (Pounds) Large Hard (Surface Tension) 4 Large Hard (Emission Limits) 27 Small Hard (Surface Tension) 6 Small Hard (Emission Limits) 24 Decorative 24 Anodizing 6 TOTAL 91 Keep in mind that the revised emissions reduction estimate above was still based on EPA s baseline emissions estimate of 1,140 pounds. Based on the NASF data collection effort and analysis, it is reasonable to reduce EPA s baseline emissions estimate by 80 percent. If the baseline emissions estimate are reduced by 80 percent (i.e., 228 pounds), then the revised emissions reduction estimate (i.e., 91 pounds) should also be reduced by 80 percent, because the emissions reduction are a function of the baseline emissions estimates. Accordingly, the final emissions reduction estimate would be 80 percent of the revised emissions reduction referenced above or 18 pounds. A summary of the emissions reduction estimate is summarized below in Table IX-3. Even if EPA is correct about its original estimate of an emissions reduction of 208 pounds, with 80 percent lower baseline emissions, the emissions reduction estimate would be only 40 pounds. 43

44 TABLE IX-3: Final Emissions Reduction Estimate Baseline Emissions Emissions Reduction EPA s Original Estimate 1,140 pounds 208 pounds EPA s Revised Estimate 1,140 pounds 91 pounds 80% Reduction of EPA s Estimate 228 pounds 18 pounds X. REDUCED RISKS A. EPA Significantly Over-Estimated the Risks that Will Result from Facilities Actual Emissions In the October 2010 proposed rule, EPA concluded that the risks posed by chromium electroplating and anodizing were acceptable, yet the Agency continued to express concerns regarding potential cancer risks. Initially, EPA estimated that 14 million individuals were exposed to a cancer risk of greater than one in a million. Subsequently, EPA has significantly reduced its estimate from 14 million individuals to 360,000 individuals and now to 180,000 individuals potentially exposed to a cancer risk of one in a million. See Figure X-1 below. 44

45 FIGURE X-1: Individuals Potential Exposed to Cancer Risk from Chromium Electroplating and Anodizing Emissions Based on the reasons discussed below, even this risk level is overstated and, few, if any, individuals may be exposed to a potential cancer risk of one in a million from chromium electroplating and anodizing emissions. Because EPA concluded in October 2010 that stricter standards were not necessary even when it believed that 14 million individuals were exposed to a potential cancer risk of one in a million, there is no rational basis for imposing such controls when EPA itself has adjusted its assessment of risk downward by two orders of magnitude, and the record demonstrates that the residual risk is even lower than that. EPA s risk analysis is primarily a function of the baseline emissions estimates. Based on EPA s current emissions estimates, the residual risk is very low. A summary of EPA s distribution of maximum individual risks for facilities is provided below in Table X-1. 45

46 TABLE X-1: EPA s Facility Risk Distribution (Without Corrections to EPA s Emissions Estimates) Hard Chrome Actual Emissions Estimates MIR Number of Facilities % of Total % % % % % < % Decorative Actual Emissions Estimates MIR Number of Facilities % of Total % % % < % Anodizing Actual Emissions Estimates MIR Number of Facilities % of Total % % % < % Based on this information, which for the reasons discussed in detail above significantly overstates emissions, nearly 80 percent of the facilities have a MIR cancer risk of less than one in a million. Most of the remaining facilities have a MIR cancer risk at or near one in a million. Adjusting the overstated emissions estimates downward as discussed above would substantially reduce the risks such that few, if any, facilities would pose a MIR cancer risk equal to or greater than one in a million. Few, if any, individuals would, therefore, be exposed to a MIR risk of one in a million or more. 46

47 B. Risks Based on Allowable Emissions In the supplemental proposed rule, EPA identified maximum individual cancer risks (MIR). As referenced above, the highest MIR based on actual emissions was 22 for hard chromium, 9 for decorative, and 5 for anodizing. EPA also identified MIRs based on allowable emissions with 50 reported as the highest for hard chromium, 70 for decorative, and 60 for anodizing. 77 Fed. Reg. at Despite EPA s statement that it did not include California facilities as part in its analysis for estimating allowable emissions (77 Fed. Reg. at 6634), the facilities with the highest MIRs for allowable emissions for decorative and anodizing were nearly all in California. Allowable emissions in California should be significantly lower due to stringent state air emission limits for hexavalent chromium. As a result, EPA has artificially inflated its risk estimates based on allowable emissions for decorative and anodizing operations, and in direct contradiction to its claims of not using data from facilities in California. C. Critique of EPA s Risk Modeling Procedures We believe that EPA greatly overestimates the MIR risks posed by electroplating and anodizing facilities. The primary reason, as we have discussed, is that EPA greatly overestimates current actual hexavalent chromium emissions from these facilities. Based on our review of EPA s emission estimates for the roughly 300 highest risk facilities, we judge that current actual emissions from the industry may be roughly 80% lower (only 1/5 as much as) what EPA estimates. Since the risk a facility poses is a linear function of the facility s emissions, this emissions correction would reduce EPA s estimated risks by a similar 80% or so. (We recognize that the reduction in risks across the entire industry would depend on whether the reduction in emissions is concentrated among higher or among lower risk facilities.) In addition to believing that EPA badly over-estimates risks because the Agency badly over-estimates emissions, NASF believes that inappropriate choices in dispersion 47

48 modeling also contributes to EPA over-estimated the risks. A discussion of EPA s dispersion modeling procedures is provided below. 1. A Previous EPA Analysis that Models Long-Term Average Ambient Concentrations of Hexavalent Chromium Using Techniques Identical to the Agency s Current Electroplating Analysis Shows Modeled Concentrations at Monitoring Sites That Are Far Higher than Actual Monitored Concentrations at These Sites EPA s process for estimating hexavalent chromium emissions and using dispersion modeling to predict resulting annual average concentrations at receptor sites involves fundamental and systematic errors. Across those higher-risk sites at which hexavalent chromium concentrations have been both modeled and monitored, EPA s modeled or predicted concentrations appear to be some 3 to 5 or more times too high relative to actual monitoring results. It is not clear exactly where in EPA s sequence of modeling steps the most significant errors occur, but significant problems likely exist in EPA s dispersion modeling as well as in EPA s emission estimates. In 2007, EPA completed an analysis of information on hexavalent chromium as part of the Urban Air Toxics Monitoring Program (UATMP). 3 This 2007 study still today remains EPA s only substantial national analysis addressing ambient concentrations of hexavalent chromium. In the study, EPA compared the concentrations of hexavalent chromium that had been modeled in the Agency s National Air Toxics Assessment (NATA) risk analysis against the actual UATMP monitored concentrations at the particular sites where concentrations had been both modeled and monitored. The results for higher risk sites (those where modeled cancer risks from hexavalent chromium exceeded 0.5 x 10-6 ) provided strong evidence of excessive conservatism in EPA s risk modeling procedures for hexavalent chromium: EPA s modeling predicted ambient concentrations and risks that were more than four times higher on average than monitored concentrations and resulting projected risks at the same sites. The dispersion and risk modeling procedures used for hexavalent chromium in the NATA risk analysis for 3 U.S. EPA Urban Air Toxics Monitoring Program (UATMP) -- Hexavalent Chromium. Final Report. February,

49 hexavalent chromium are virtually identical to those the Agency has used for the current supplemental proposed chromium electroplating NESHAP. We believe this suggests that EPA s dispersion modeling procedures used in estimating risks for the proposed NESHAP may also be contributing to overestimating risks, in addition to the problems we have discussed involving overestimating emissions. We will elaborate. In EPA s UATMP study on hexavalent chromium, the Agency compiled the results of special ambient air monitoring for this HAP that was conducted at 22 UATMP sites and one Hurricane Katrina site during The 23 monitoring sites represented a wide variety of conditions, ranging from highly urban (e.g., Detroit, Chicago and Seattle), near-downwind of an urban area (Atlanta), on-site heavy industrial (Sloss Industries outside Birmingham, AL, with a metals plant and coke ovens), residential (Oregon), and rural/background (Wisconsin, Alabama). Each site was intended to be monitored for hexavalent chromium every 6 or 12 days during the entire year, but this was not achieved at all sites. Monitoring began late or ended early in the year at some sites, and some monitoring dates were missed for varying reasons. In total, more than 1,200 hexavalent chromium samples (including some colocated samples) were collected across the 23 sites. EPA calculated a daily average concentration at each site as the average concentration of all detects. EPA also calculated an annual average concentration at each site where monitoring began no later than February and ended no earlier than November, and nondetects were averaged into the annual average calculations at one-half the method detection limit (MDL). For each site where it was calculated, the annual average was, thus, always less than or equal to the daily average due to the inclusion of non-detects in the annual average calculations at one-half the MDL. In addition to compiling daily and annual average hexavalent chromium concentrations for the monitoring sites, in the report EPA also compared the concentrations of hexavalent chromium that were modeled in the Agency s National Air Toxics 49

50 Assessment (NATA) risk analysis against the actual UATMP monitored concentrations at these sites. EPA compared the annual average modeled concentration of hexavalent chromium at the centroid of the Census Tract within which each UATMP monitor was located against the calculated annual average monitored concentration at the corresponding site. Although EPA did not characterize the results in this manner, the results were striking -- across all the higher-modeled-risk UATMP sites, EPA s predicted ambient concentrations of hexavalent chromium were 3 to 5 times higher than the actual monitored concentrations. Specifically, across the UATMP sites for which projected lifetime cancer risks from hexavalent chromium were 1 x 10-6 or more, EPA s modeled annual average ambient concentrations exceeded the actual monitored annual average concentrations by a factor of more than 5. The degree to which EPA s modeling overestimated actual monitored hexavalent chromium concentrations declines as one looks at UATMP sites posing progressively lower-risks. Across all UATMP sites for which modeled risks exceeded 0.5 x 10-6, modeled concentrations exceeded monitored concentrations by an average factor of about 4. Across the UATMP sites for which modeled risks exceeded 0.3 x 10-6, modeled concentrations exceeded monitored concentrations by an average factor of about 3.5. Across the UATMP sites for which modeled risks were less than 0.3 x 10-6, modeled concentrations were typically slightly less than monitored concentrations. These results are shown in the Table X-2 below. 50

51 TABLE X-2: EPA Modeling for Hexavalent Chromium Over-Estimates Actual Ambient Concentrations 2005 Hex Chrome UATMP Monitoring Site Location Monitored Annual Average* (ng/m 3 ) Census Tract ID NATA-Modeled Concentration (ng/m 3 ) NATA-Modeled Cancer Risk (x 10-6 ) Ratio: Modeled vs. Actual LTA Cr(VI) Concentration Seattle, WA St. Louis, MO Detroit, MI Providence, RI North Birmingham, AL Birmingham, AL (Sloss Industries) Sydney, FL (rural, nr Tampa) East Thomas, AL (Birmingham) Northbrook, IL (Chicago) Bountiful, UT (Ogden) Providence, AL (rural) Boston, MA Decatur, GA (downwind Atlanta) Washington, DC Gulfport, MS * Annual averages shown include 6 sites where EPA calculated only daily averages. Across all hexavalent chromium UATMP sites where EPA calculated an annual average, this annual avg. = 0.67 daily avg. For the 6 sites where EPA didn't calculate an annual avg., we estimated annual avg. as 0.67 x the site's daily average Average Ratio: Modeled vs. Actual Cr(VI) Concentration 7 Sites w/modeled risk 1 x Sites w/modeled risk between 0.5 and 1.0 x Sites w/modeled risk between 0.3 and 0.5 x Sites w/modeled risk 1 x Sites w/modeled risk > 0.5 x Sites w/modeled risk > 0.3 x These results suggest that EPA s emissions and dispersion modeling for hexavalent chromium is systematically biased high compared to actual monitoring data collected by EPA, leading to greatly over-estimating ambient concentrations (and risks) at those sites where higher concentrations are modeled. 51

52 EPA s 2005 NATA risk modeling (analyzed above) and EPA s risk modeling in this rulemaking for maximum individual risks (MIR) employ nearly identical procedures. EPA s MIR emissions and dispersion modeling for hexavalent chromium for the chromium electroplating risk and technology review is, therefore, very likely to be biased high for higher-risk sites in a similar manner as is the NATA modeling. EPA should, in the risk evaluation performed in this rulemaking, not rely on modeling that replicates the errors of the NATA modeling and should explain how the Agency s risk assessment has avoided these errors. The modeling procedures used for both the NATA and the current risk analysis are nearly identical. Both analyses start with NEI or NEI-like estimates of HAP emissions and emission point characteristics (including identical default assumptions regarding stack height, exit velocity, etc. when information is not available) for the relevant source categories. They then use HEM for dispersion and risk modeling, including: 1) accessing the same meteorology data; 2) running the model in a mode that does not account for reduction/transformation or for wet/dry deposition of hexavalent chromium. Finally risks are estimated for HAPs based on the IRIS toxicity values for each HAP and based on EPA s standard assumptions about exposure constants, body weight, and 70-year lifetime. Some or all of the following factors likely account for EPA s badly over-estimating hexavalent chromium concentrations when the agency models MIR and population risks at receptor sites around chromium electroplating/anodizing facilities: EPA has overestimated hexavalent chromium emissions by perhaps a factor of five, as discussed in detail above. EPA assumed that 97 to 100 percent of every facility s chromium emissions is hexavalent, despite evidence to the contrary (e.g., Shin and Pak (2000) found fairly rapid reduction of hexavalent chromium in electroplating mists to trivalent:hexavalent chromium equals 100 percent of total chromium immediately 52

53 after mist generation, but declines to 85 percent after one hour, and to 77 percent after three hours). EPA s default assumptions regarding stack height and diameter, exit velocity and temperature are all nearly worst case, and may be inaccurate. EPA has failed to run HEM/AERMOD in a mode that accounts for rapid reduction/reactivity of hexavalent chromium to trivalent. A study for CARB found an average half-life of 14.4 hours for hexavalent chromium under common atmospheric conditions (Grohse, et al ). EPA has also not run HEM/AERMOD in a mode that accounts for wet and dry deposition of hexavalent chromium and plume depletion. (All hexavalent chromium emitted from electroplating/anodizing facilities is in particulate/aerosol form and subject to deposition.) EPA chose to run HEM/AERMOD in rural rather than urban mode, thereby simulating less mixing and less dispersion (thus higher concentrations and higher risks, particularly MIR) near emissions sources EPA assumed a 30 meter default distance from facility stack to property fence line (where MIR risk modeling may begin). This is too small for some facilities. 2. EPA Projects Ambient Concentrations of Hexavalent Chromium at MIR Sites from Electroplating/Anodizing Emissions Alone That Exceed Any Recently Monitored Ambient Levels of Hexavalent Chromium Anywhere in the U.S. (Reflecting Emissions From All Sources, Not Just Electroplating/Anodizing) It is Difficult to Believe That Ambient Concentrations from Electroplating/Anodizing Emissions Could Really be This High. It seems highly improbable that the concentrations of hexavalent chromium at MIR sites due to electroplating emissions alone could really be higher than any observed actual ambient concentrations of hexavalent chromium, which reflect emissions from all source types, not just electroplating. After all, surface finishing accounts for less than 1 percent of total national emissions of hexavalent chromium and chromic acid, as estimated in the 2005 NEI (see Attachment 2). 53

54 EPA cannot plausibly model ambient concentrations from a single source type that accounts for less than 1 percent of total emissions as being higher than the highest actual ambient concentrations that have been monitored, when the actual ambient concentrations presumably reflect emissions from all source types. According to EPA s modeling of electroplating facility emissions and dispersion, the estimated annual average concentrations of hexavalent chromium from electroplating sources at MIR sites are summarized below in Table X-3. TABLE X-3: EPA s Modeled Hexavalent Chromium Concentrations from MIR Electroplating Facilities 1.9 ng/m 3 is the annual average hexavalent chromium concentration at the MIR site for the highest-risk electroplating facility. (MIR of > 22 x 10-6 ) 1.5 ng/m 3 is the annual average hexavalent chromium concentration at the MIR site for the secondhighest-risk electroplating facility. (MIR of 18 x 10-6 ) ng/m 3 or more are the annual average hexavalent chromium concentrations at the MIR sites for the 14 highest-risk electroplating facility. (According to EPA, 14 facilities pose MIR exceeding 10 x 10-6 ). At roughly 1 to 2 ng/m 3, these annual average concentrations of hexavalent chromium that EPA models at MIR sites from electroplating facility emissions alone exceed the highest annual average concentrations of hexavalent chromium actually found at monitoring sites -- even though the actual monitored levels reflect hexavalent chromium emissions from all sources, with all other sources responsible for at least 100 times the quantity of hexavalent chromium emissions as electroplating facilities. The following summarizes the highest recent long-term average monitored levels of hexavalent chromium that we have found. 54

55 Data from EPA. In the 2005 Urban Air Toxics Monitoring Program, special report on monitoring results for hexavalent chromium (2007). 4 There are 22 monitoring sites, some were near the centers of heavily populated cities (e.g., Chicago, Seattle), while others were in more rural areas (e.g., Chesterfield, SC and Hazard, KY). The aim was to sample each site every 6 or 12 days during the year (for various reasons some of the sites did not meet this goal). The highest annual average level of hexavalent chromium monitored at any site was 0.06 ng/m 3 (resulting in lifetime excess cancer risk at that location of 0.7 x 10-6 ). The total average annual concentration across all sites appears to be about 0.03 ng/m 3 (corresponding to lifetime excess cancer risk of 0.35 x 10-6 ). Data from California. Available data for 2009 and 2010 from all California monitoring sites with data on hexavalent chromium from more than 15 sites each year. 5 The annual average level of hexavalent chromium across all sites was ng/m 3. The highest quarterly average level of hexavalent chromium at any site was 0.19 ng/m 3 (124 quarterly averages, across two years and more than 15 sites). This concentration corresponds to a lifetime excess cancer risk of about 2 x Data from Texas. Data was collected during from four monitors near heavily industrialized sites (not electroplating). 6 The range of annual average hexavalent chromium concentrations at three monitors was 0.1 to 0.2 ng/m 3, while the range at the final site was <0.001 to 0.4 ng/m 3. Based on its review of EPA, California and Texas data on ambient levels of total chromium and hexavalent chromium, the Texas CEQ concluded that: Overall, the data reviewed on speciated chromium indicate that hexavalent chromium measured in ambient air makes up less than 10% of the total chromium, and the USEPA assumption (34% of total atmospheric chromium is hexavalent) is very conservative. Monitored data, while variable, is fairly consistent 4 U.S. EPA Urban Air Toxics Monitoring Program (UATMP) -- Hexavalent Chromium. Final Report. February, Available at: 6 Texas Commission on Environmental Quality. Development Support Document: Chromium -- All Compounds Except Hexavalent Chromium. Final, October 8,

56 across data-sets. In other words, forms of chromium other than hexavalent typically make up more than 90% of measured ambient chromium levels. 7 The highest annual average ambient Cr(VI) concentration observed at any site (and reflecting emissions from all source categories) among data from EPA, California and Texas was 0.4 ng/m 3, much less than the concentrations of 1-2 ng/m 3, modeled by EPA at MIR sites due to electroplating emissions alone. In conclusion, EPA s modeling results are not plausible. XI. COST EFFECTIVENESS In the October 2010 proposed rule, EPA concluded that no additional controls were needed because the Agency had not identified controls that would reduce risk at reasonable costs. 75 Fed. Reg. at Specifically, EPA identified the cost of control options, including HEPA filter retrofit costs in Table A.7 of the preamble to the October 2010 proposed rule. 75 Fed. Reg. at For large hard chromium facilities, EPA estimated the annualized costs were $18.4 million per year with an expected emissions reduction of one ton (i.e., 2,000 pounds). Although EPA reported the cost effectiveness in Table A.7 as $36.3 million per ton (or approximately $18,000 per pound), the cost effectiveness should have been $9,000 per pound ($18.4 million per year divided by 2,000 pounds per year). Based on this assessment, EPA concluded that this technology option would not reduce risks at reasonable costs. 75 Fed. Reg. at In the preamble to the February 2012 supplemental proposed rule, for large hard chromium facilities EPA estimated the annualized costs for the control options (including primarily the use of fume suppressants with only a small portion of add-on controls and HEPA filter retrofit) were $2.196 million with an expected emissions reduction of 121 pounds. 77 Fed. Reg. at EPA reported the cost effectiveness in Table 2 as approximately $18,000 per pound. Id. Based on its assessment now, EPA has somehow concluded that this technology option would reduce risk at reasonable costs. 7 Ibid., p

57 It is curious EPA concluded in October 2010 that control options with cost effectiveness of $9,000 per pound would not reduce risks at reasonable costs, but now has reached the conclusion that less burdensome controls based primarily on the use of fume suppressants with a cost effectiveness of $18,000 per pound are reasonable. Consequently, proposed controls of $18,000 per pound cannot be cost effective, particularly because EPA has already concluded that more burdensome controls identified in October 2010 with a cost effectiveness of $9,000 per pound would not reduce risk at reasonable costs. As discussed above, EPA s estimated emissions reduction is off by at least an order of magnitude, and should be only 18 pounds instead of 208. Adjusting further for this revision would yield a cost effectiveness of approximately $180,000 per pound. Furthermore, this revised cost effectiveness does not yet take into account any changes to EPA s cost estimates. As discussed below NASF contends that EPA has under-estimated cost and the cost effectiveness would need to be adjusted further based on those revised costs. Even if EPA does not accept all of the revisions claimed by NASF, the overwhelming evidence suggests that EPA s cost effectiveness, as expressed in this supplemental proposal, is inaccurate. A conservative estimate of cost effectiveness of $90,000 per pound (approximately one half of $180,000 per pound) is reasonable based on the information provided above. Consequently, if $9,000 per pound was not cost effective in October 2010, then $90,000 per pound cannot be cost effective in March XII. EPA HAS UNDER-ESTIMATED COMPLIANCE COSTS FOR THIS PROPOSED RULE Based on an initial review of EPA s compliance cost estimates for this proposal, it appears that EPA has under-estimated the costs. Due to time constraints and the limited time to provide comments on this proposal, NASF has not been able to provide a detailed analysis of EPA s cost estimates, but NASF plans to provide EPA with additional 57

58 information on the topic consistent with EPA s letter of March 22, 2012 indicating that the Agency would accept such information and take it into account in this rulemaking. Provided below is a brief discussion on EPA s cost analysis. A. EPA Has Under-Estimated the Number of Facilities that Do Not Meet the Proposed Emissions Limits In the January 24, 2012 document entitled, Procedures for Determining Control Costs and Cost Effectiveness for Chromium Electroplating Supplemental Proposal, EPA calculated the number of facilities that would need to reduce emissions or make adjustments to meet the proposed limits, and therefore, incur costs to meet the new proposed emissions limits. EPA assumes that only those facilities exceeding 85 percent of the proposed new emissions limits would need to reduce emissions or make adjustments to meet the new proposed emission limits. For the reasons discussed below, this narrow 15 percent compliance margin should be much broader (e.g., 50 percent), and thus, include more facilities that would have to incur costs to meet the new proposed emissions limits. EPA provided no data in the record to support the 85 percent value for estimating compliance with the new proposed emissions limits. EPA s assumption fails to take into account the variability over time regarding a facility s emissions concentration. A review of typical stack tests demonstrates a significantly larger range of variability among sampling results for a facility. Based on this typical variability, it is reasonable to conclude facilities that exceed 50 percent of the proposed limit would need to reduce emission or make adjustments to meet the new proposed emissions limits. EPA should set a compliance margin of 50 percent based on typical variability in stack test results or provide some evidence to support this narrow compliance margin. The net effect of a more realistic compliance margin would be more facilities that would have to reduce emissions or make adjustments to ensure compliance with the new proposed emission limits and that would have to incur additional costs to meet the new 58

59 proposed emission limits. EPA should, therefore, amend its cost analysis to include these facilities that will have to reduce emissions or make adjustments to ensure compliance with the new proposed limits, which would significantly increase the compliance costs for this proposal. B. Add-On Controls or Retrofits of Existing Controls Are Needed to Meet the New Proposed Emissions Limits In the Control Costs document, EPA assumes that 90 percent of hard chromium facilities will meet the new proposed emissions limits by using fume suppressants (either begin using fume suppressants or using more fume suppressants). As discussed in detail below, EPA has not provided any evidence that the use of non-pfos fume suppressants will lower emissions to meet the new proposed emissions limits. Consequently, contrary to EPA s assumptions, all facilities would need to use add-on controls or retrofit existing controls to meet the new proposed emissions limits. This makes sense because add-on controls are a critical component for meeting the current NESHAP standards (particularly for hard chromium facilities), so they should also be a critical component for meeting the new proposed emissions limits. EPA needs to include the additional costs of add-on controls and retrofits in its cost analysis for all facilities that need to reduce emissions or make adjustments to meet the new proposed emissions limits. When these costs are included, the analysis will demonstrate that EPA technology options are not cost effective (as discussed in detail above in these comments). EPA has already concluded in the October 2010 rule that such add-on controls and retrofits are not cost effective. A summary of EPA s cost estimates and cost effectiveness results for controls are included in Attachment 3. C. EPA Has Under-Estimated the Cost of Non-PFOS Fume Suppressants In its August 16, 2010 document entitled, Nationwide Costs for Conversion to Non- PFOS Fume Suppressants, EPA calculated that the costs of non-pfos fume 59

60 suppressants will be 15 percent more than the costs of PFOS fume suppressants. EPA then uses this 15 percent cost increase in its cost analysis for the proposal. If for purposes of this analysis you assume that non-pfos fume suppressants can be used to reduce emissions (which as discussed in detail below they cannot), EPA has under-estimated the costs associated with using non-pfos fume suppressants. The Conversion to Non-PFOS document includes cost quotes from several chemical suppliers of fume suppressants. Before reaching any final conclusions on the costs of non-pfos fume suppressants compared to PFOS fume suppressants, EPA must confirm that it is evaluating comparable products. For example, fume suppressants are available in a number of different formulations that contain non-pfos and PFOS in various concentrations. A fume suppressant with a lower concentration of non-pfos may appear more cost-comparable to a fume suppressant with a higher concentration of PFOS. Furthermore, EPA has not included all of the additional costs associated with the use of non-pfos fume suppressants. Non-PFOS fume suppressants may require more frequent additions to the plating bath because they are not as stable or persistent in the chemically aggressive conditions of the chromium plating bath. Facilities must also monitor surface tensions levels more frequently when using non-pfos and use more labor associated with demonstrating and maintaining compliance with the chromium electroplating NESHAP. All of this translates into additional costs that EPA has not considered. Finally, several facilities have reported to NASF that their costs for converting to non- PFOS fume suppressant may be more than 30 percent higher than using PFOS fume suppressants. One facility estimated that its annual costs for fume suppressants would increase by approximately $100,000 with the switch to non-pfos fume suppressants. EPA needs to reassess its claims regarding the use on non-pfos fume suppressants. First, EPA has not provided any evidence in the record that non-pfos fume suppressant can effectively reduce emissions to meet the new proposed emissions limits. Second, if non-pfos fume suppressants can be used, they are significantly more expensive to use 60

61 than EPA claims. As a result, non-pfos fume suppressants are not a technologically or economically feasible option for ensuring compliance with the new proposed emissions limits. D. EPA s Costs Associated with Stack Tests May Not Be Accurate EPA s costs associated with stack tests appear to be under-estimated. First, EPA estimates in the Control Costs document that only 181 facilities would need stack tests pursuant to the proposed rule. This estimate appears to be low because EPA has estimated that there are approximately hard chromium facilities. According to EPA most of these facilities are relying on some type of control equipment to meet current emissions limits. To demonstrate compliance with the new proposed limits, whether it can be achieved with new controls or the use of fume suppressants with existing controls, many more facilities would appear to need a new stack test. EPA s estimate of only 181 facilities needing new stack test only makes sense if EPA accepts the claim supported by NASF s data collection effort that nearly 75 percent of chromium electroplating and anodizing facilities are closed or no longer use hexavalent chromium in their processes. Otherwise, EPA has under-estimated the costs associated with stack testing for this proposed rule. Another factor that may impact the costs associated with stack tests is how EPA has calculated the annualized costs. EPA assumed that a stack test is good for 15 years, so the costs were amortized over 15 years at an interest rate of 7 percent. Capital recovery factors (CRF) for converting a capital cost into an annual cost are summarized below in Table XII-1. Based on this information, the cost of stack tests would be 30 percent higher for a period of 10 years and 130 percent higher for a period of 5 years. If EPA requires more frequent stack tests, then the costs associated with stack tests would be significantly higher. 61

62 TABLE XII-1: Capital Recovery Factors (CRF) for Converting Capital Costs into Annual Costs Years Interest Rate CRF 15 7% % % E. With Increased Costs, Technology Options Are Less Cost Effective NASF believes that EPA has significantly under-estimated the costs of this rule as discussed above and plans to provide more detailed cost analysis information. To the extent that EPA has under-estimated compliance costs for this proposed rule, the proposed technology options discussed below would be even less cost effective. For example, if the costs double or triple, then the proposed options would be less effective by a multiple of two or three. XIII. TECHNOLOGY OPTIONS A. Proposed New Surface Tension Levels The chromium electroplating NESHAP allows facilities to demonstrate compliance with the standard by maintaining plating baths below the allowable surface tension levels or meeting a concentration based emission limit. In this supplemental proposal, EPA has proposed to lower the surface tension level from 35 dynes/cm to 33 dynes/cm as measured by a tensiometer and from 45 dynes/cm to 40 dynes/cm as measured by a stalagmometer. With no changes in the current practices, this proposed change appears to be feasible, but unnecessary because it yields little, if any, environmental benefit or risk reductions. While most facilities should be able to meet the new levels with PFOS-based fume 62

63 suppressants, several facilities have expressed concerns that the lower surface tension levels will provide no environmental benefit or risk reduction and will eliminate a muchneeded margin of compliance. To maintain the same margin of compliance, facilities may have to add more fume suppressants. The added costs to ensure continued compliance with a sufficient buffer does not appear to be warranted given the minimal environmental benefits or risk reductions, if any that are expected from this change. In fact, EPA has not provided any documentation or data in the record to demonstrate that the proposed changes to the surface tension levels will provide any measurable benefits (i.e., emissions reductions). The biggest challenge in meeting the revised surface tension levels stems from the phaseout of PFOS-based fume suppressants and switch to non-pfos fume suppressants. PFOS is a persistent and bio-accumulative substance that is believed to pose potential risks to human health and the environment. Because of its persistent nature, PFOS is a very effective fume suppressant in the chemically aggressive conditions of chromium electroplating and anodizing baths. PFOS and other long-chain perfluorinated compounds (PFCs) are being phased out by EPA and by other regulatory agencies globally because of the environmental impacts that may result from the use of PFOS. As part of this rulemaking EPA has proposed to ban the use of PFOS fume suppressants and replace them with non-pfos fume suppressants. This action is consistent with other actions within EPA and by states. For example, EPA has developed an Action Plan pursuant to the Toxic Substances Control Act (TSCA) to ban the use of long-chain PFCs, including PFOS. EPA has already taken action under a TSCA significant new use rule (SNUR) to ban the use of PFOS in most application, with the exception of chromium electroplating. EPA also has developed a Voluntary Stewardship Program with chemical companies to eliminate the use of PFOA by Finally, the State of Minnesota has already banned the use of PFOS fume suppressants for chromium electroplating and anodizing processes. In summary, chromium electroplating and anodizing operations that rely on PFOS fume suppressants to comply with the chromium electroplating NESHAP 63

64 will likely no longer be able to use them, whether EPA decides to include the phase-out as part of this rulemaking or not. The switch to non-pfos fume suppressants is a technology in transition. Facilities that have switched to non-pfos fume suppressants have achieved moderate success in meeting the current surface tension levels, but many challenges and problems persist. The limited experience in Minnesota with non-pfos demonstrates that facilities can meet the current surface tension levels generally, but have seen scattered excursions above the current levels. The switch to non-pfos fume suppressants alone diminishes a facility s margin of compliance in meeting the current surface tension levels. An example of data from the use on non-pfos fume suppressants from a facility in Minnesota is provided in Attachment 4. This facility (which closed December 31, 2011) used non-pfos fume suppressants and maintained surface tension levels that averaged just over 30 dynes/cm. The facility did, however, have a few data points that exceeded the proposed surface tension level of 33 dynes/cm, but none that were above the current standard of 35 dynes/cm. These data demonstrate that the current surface tension levels can met, with little or no margin of safety for compliance. The proposed new surface tension level would be problematic on a consistent basis with virtually no margin of safety for compliance. Nonetheless, problems in transitioning to non-pfos fume suppressants continue. A facility in Texas switched from PFOS to non-pfos fume suppressants. After the initial additions and adjustments, the facility was able to maintain surface tension levels below the current surface tension levels for approximately one year. At that point, the surface tension levels began to rise, so more non-pfos fume suppressant was added to maintain the surface tension levels. Ultimately, the facility experienced chemical failure with the non-pfo and the surface tension levels spiked. The facility s production was disrupted for several days until PFOS fume suppressants were added to the bath to restore surface tension levels. A summary of this case study is provided in Attachment 5. 64

65 The examples above demonstrate that non-pfos fume suppressants are still a technology in transition. The new technology is not a simple drop-in replacement for PFOS fume suppressants. Non-PFOS fume suppressants have not been proven effective in many chromium applications, particularly with military, aerospace and automotive specifications. Where non-pfos has shown promise in lowering surface tension levels, it requires more frequent additions, more frequent monitoring, and more labor to maintain surface tension levels compared to the use of PFOS fume suppressants. In addition, non- PFOS products are generally more expensive than PFOS products. While the challenges of using non-pfos fume suppressants to meet the current surface tension levels are illustrated above, EPA has proposed to exacerbate this challenge by also lowering the surface tension levels in the face of this technology in transition. Nonetheless, EPA has provided no data in the record that non-pfos fume suppressants can achieve the proposed new surface tension levels. EPA merely assumes that non- PFOS fume suppressants are equivalent in performance to PFOS fume suppressants without presenting any scientific proof or supporting data. Given the nature of the new non-pfos technology in transition, the many challenges facing chromium electroplating and anodizing operations in using the new technology to meet the current surface tension levels, and the lack of any data in the record to demonstrate that non-pfos fume suppressants can consistently achieve the proposed surface tension levels for chromium electroplating and anodizing applications, EPA should forego the proposed revisions to the surface tension levels. With minimal, if any, environmental benefits or risk reductions expected from this proposed change, the burdens clearly outweigh any perceived benefits. B. Proposed New Emission Limits EPA has proposed to lower emission limits for all existing and new chromium electroplating and anodizing operations: 30 percent lower for existing decorative and anodizing facilities, 65

66 25 percent lower for existing large hard chromium facilities, 50 percent lower for existing small hard chromium facilities, 40 percent lower for new decorative and anodizing facilities, and 60 percent lower for new hard chromium facilities. EPA has not based these proposed reduction on record evidence that new technology has been introduced that can be linked to achieving these new limits (i.e., CAA 112 (d)(6)) nor, as discussed in detail above, is there ongoing residual risk associated with chromium emissions from these source categories that justifies the stricter standards (i.e., CAA 112(f)(2)). Therefore, there is neither a legal nor factual basis for the proposed changes. Nonetheless, NASF will address each of these proposed new emissions limits in turn below. 1. Existing Decorative and Anodizing Emissions Limits EPA reviewed data from 17 decorative chromium facilities and one anodizing facility, and concluded that all decorative and anodizing facilities already comply with the new proposed emissions limits. 77 Fed. Reg. at Despite this claim, EPA acknowledged that 8 decorative facilities may need to make adjustments and achieve reductions to meet the new limits. 77 Fed. Reg. at EPA dismissed this data, however, by claiming that these facilities would choose to comply with the new NESHAP with the surface tension levels rather than the new emissions limits. Id. As noted above, EPA also misunderstands the applicable legal standard, which does not allow EPA to continue to lower emissions standards based simply on evidence that facilities have been successful in achieving superior environmental performance. In addition, EPA admitted that it did not perform any detailed analysis for anodizing facilities. Rather, EPA concluded that anodizing processes are similar enough to decorative processes so the proposed limits would also be appropriate. Such a conclusion ignores the fact the fume suppressants often perform differently in decorative chromium and chromium anodizing plating baths. 66

67 Due to time constraints from the limited comment period, NASF did not have any opportunity to review thoroughly EPA s data to support the proposed emission limits for decorative and anodizing processes and to evaluate its claims that all of the operations would comply with the proposed emission limits. The limited data and EPA s superficial analysis regarding the proposed emissions limits for decorative and anodizing facilities appear to provide weak scientific and technical basis to support or justify the proposed limits. 2. Existing Hard Chromium Emissions Limits EPA reviewed data from 38 large hard chromium facilities and 56 small hard chromium facilities. EPA concluded that 41 large hard chromium and 85 small hard chromium facilities would need to reduce emissions or make adjustments to meet the proposed emission limits. 77 Fed. Reg. at 6638 & Because many of the hard chromium facilities are using PFOS fume suppressants in addition to control equipment to reduce emissions, more hard chromium facilities will need to reduce emissions or make adjustments to meet the proposed emissions limits when they switch to non-pfos fume suppressants, which are not as effective as PFOS in reducing emissions. With respect to the hard chromium facilities that would need to reduce emissions or make adjustments to meet the proposed emission limits, EPA assumes that most facilities would achieve these extra reductions with the addition of fume suppressants. 77 Fed. Reg. at 6638 & EPA did acknowledge that some facilities would need to install add-on control devices or retrofit their existing controls to meet the limit. EPA has, however, failed to incorporate the cost of new control devices and retro fit equipment appropriately in the cost analysis of the supplemental proposal. Further, the use of more fume suppressants that have equivalent or lesser effectiveness is not a new technology under CAA 112(d)(6). In addition, as discussed above, EPA has already determined in the October 2010 proposal that such add-on controls and retrofits were not cost effective, and there is no evidence in the record to support a change in EPA s position. 67

68 3. New Source Emissions Limits The proposed new source emissions limits could be problematic for facilities, particularly hard chromium facilities. EPA claims that the new source emissions limit can be achieved for decorative and anodizing by maintaining the new surface tension levels, yet offers no proof to support this claim. The new source emissions limit for hard chromium facilities may be beyond the performance capabilities of existing control technology for hard chromium operations. EPA has not provided any data in the record to demonstrate that the proposed limit is feasible for new hard chromium facilities. EPA continues to assume that the proposed emissions limits can be achieved, and NASF remains concerned that the stringent new source emissions limits will discourage new production and job creation for chromium electroplating and anodizing operations. C. Technology Options for Meeting New Emissions Limits For those facilities that need to reduce emissions or make adjustments to meet the proposed emissions limits, EPA claims that they can begin using fume suppressants or use more fume suppressants if they are already using them. This assumption is the key to EPA s proposed technology options. Unfortunately for EPA, the assumption is faulty on numerous levels. Some facilities cannot use fume suppressants because of operational considerations such as the size of tanks, configuration of tanks, and the mechanics of the plating process. In addition, many facilities do not use fume suppressants due to problems that may arise with product quality associated with the use of fume suppressants. EPA has continued to acknowledge this limitation on the use of fume suppressants (see, January 18, 2012 document entitled, Draft Development of Revised Surface Tension Limits for Chromium Electroplating and Anodizing Tanks Controlled with Wetting Agent Fume Suppressants ), but appears to ignore it as it insists that fume suppressants, which are not a new technology, are the answer for all facilities to meet the proposed limits. 68

69 Originally, the chromium electroplating NESHAP did not allow the use of fume suppressants to demonstrate compliance for hard chromium operations. After the 2004 amendments, hard chromium facilities could use fume suppressants to demonstrate compliance (accordingly, for purposes of this rulemaking, fume suppressants are not new ). In addition, following the promulgation of OSHA s revised workplace exposure standard for hexavalent chromium, many hard chromium facilities started using fume suppressants to reduce workplace exposures even further to ensure compliance with the new OSHA standard. As a result, it is reasonable to conclude facilities that can use fume suppressants beneficially in their operations are currently using them. Those facilities that are not using them have operational, product quality or cost considerations for not using them. Accordingly, EPA s oversimplified solution to start using fume suppressants or using more fume suppressants is misguided and without merit, and is not supported by evidence in the record. D. EPA Lacks Technical Basis to Conclude the Use of Fume Suppressants Will Lower Emissions to Meet the Proposed Emissions Limits As discussed above, facilities are using primarily PFOS-based fume suppressants. These PFOS fume suppressants are effective tools for lowering surface tension levels and reducing emissions. PFOS fume suppressants are being phased out, so facilities must now rely only on the use of non-pfos fume suppressants. EPA claims that the use of any fume suppressant that lowers surface tension levels will result in a reduction of emissions. EPA has not supported this claim with any credible data in the record. EPA has limited data on the use of fume suppressants to reduce emissions. These data are summarized in its January 18, 2012 document entitled, Draft Development of Revised Surface Tension Limits for Chromium Electroplating and Anodizing Tanks Controlled with Wetting Agent Fume Suppressants. The document included data from 69

70 15 facilities: 9 decorative and 6 hard chromium. EPA attempts to correlate the surface tension levels observed with the emissions limits measured with the very limited and flawed data for these facilities. The value of this data and observations is questionable for a number of reasons. First, for decorative chromium, one facility is responsible for 63 of the 110 data points. More than half of the surface tension levels measured exceeded the new proposed surface tension levels so they offer no support for EPA s proposal. In addition, the one facility from which the majority of the data were used has closed and is no longer operating, and was located in California, and thus used other control equipment in addition to fume suppressants. With respect to hard chromium facilities, all of the data were from the 1990s, shortly after the promulgation of the original chromium electroplating NESHAP. Such outdated data hardly represents new technology worthy of consideration for a NESHAP technology review standard. In addition, nearly 75 percent of the data points for hard chromium facilities were from a single facility that used fume suppressants in conjunction with a packed bed scrubber. The facility no longer demonstrates compliance with the NESHAP standards through the emissions limits, but now uses only fume suppressants to maintain surface tension levels to demonstrate compliance. In summary, the data that EPA references to support its claim that fume suppressants effectively reduce emissions to meet the proposed limits is flawed and scientifically lacking. Even if the data could be construed to demonstrate a correlation between the use of fume suppressants and the reduction of emissions, EPA provides no scientific evidence that fume suppressants can be used to achieve the proposed emissions limits. While the data relied upon by EPA provide a vague correlation for EPA s claim that the use of fume suppressants in conjunction with control equipment may be able to lower emissions, there is no quantitative data that show how much emissions reduction could be expected from the use of a specific amount of fume suppressants. Perhaps the most 70

71 serious criticism of EPA s data is that none of it includes the use of non-pfos data. While non-pfos fume suppressants may lower surface tension levels for some applications, there is no evidence presented by EPA that the use of non-pfos fume suppressants will achieve the proposed emissions limits. As discussed above, non-pfos fume suppressants are a new technology in transition. Non-PFOS fume suppressants do not perform in the same manner as PFOS fume suppressants, particularly with respect to reducing emissions. In the case study of the Texas facility that is provided in Attachment 3the facility observed a significant decline in performance controlling emissions with the use of non-pfos fume suppressants. Specifically, as measured by the permissible exposure limit for employees, the facility had a ten-fold increase in emissions (although it maintained compliance with the applicable OSHA standard) when it switched from a PFOS fume suppressant to a non- PFOS fume suppressant. During that same period, the surface tension levels were maintained at the same levels achieved with PFOS (at least for a year). This suggests that the non-pfos fume suppressants are not as effective in reducing chromium emissions as PFOS fume suppressants. The results from this case study are consistent with the anecdotal information that NASF has received regarding benchmarking trials with non- PFOS fume suppressants within the industry. EPA cannot claim, in the absence of any credible data in the record, that non-pfos fume suppressants can reduce emissions as effectively as PFOS fume suppressants. Consequently, non-pfos fume suppressants are not feasible technology options because they cannot reduce risks, regardless of cost. EPA has not provided any data to demonstrate that non-pfos fume suppressants can be used to achieve the proposed emissions limits. Accordingly, the only effective methods to reduce emission to meet the new proposed emission limits are add-on devices and retrofits of existing controls. Yet, EPA has already concluded that these technology options are not cost effective. Add-on controls and retrofits are not, therefore, feasible because they cannot reduce risks at a reasonable cost. 71

72 XIV. ENFORCEMENT AND IMPLEMENTATION CONCERNS FOR PROPOSED NEW NESHAP The proposed revisions to the chromium electroplating NESHAP present some enforcement and implementation concerns for regulated facilities. Pursuant to the current standard, facilities demonstrate compliance with the NESHAP through emission limits by conducting a performance test of the concentration of chromium in the stack gases. Ongoing compliance with the standard is demonstrating by maintain the pressure drop across the control equipment within a specific range. EPA has proposed new emissions limits that it claims can be achieved by facilities using control equipment with the addition of fume suppressants. EPA does not specific how much fume suppressants are needed to achieve the necessary emissions reductions to meet the proposed standard. Under the new NESHAP with the new emissions limits, facilities would presumably demonstrate compliance with the new emission limits by conducting a performance test with the addition of fume suppressants. Because the non- PFOS fume suppressants will be consumed over time, if EPA is correct in its assumption that additions of non-pfos fume suppressants reduce emissions (which NASF believes is not true), then facilities would have to continue adding fume suppressants to meet the new emissions limits. As part of the supplemental proposal, EPA has not identified any new procedures for demonstrating ongoing compliance with the NESHAP emissions limits. If facilities can continue to demonstrate compliance by maintaining the pressure drop across the control equipment within the specific range, then the use of fume suppressants to achieve compliance would be irrelevant. If EPA requires facilities that use fume suppressants with control equipment to achieve the new proposed emissions limits to measure surface tension, then the emissions limits would be irrelevant. 72

73 The conundrum described above demonstrates that EPA proposed new emission limits and the proposed technology option to achieve them (i.e., the use of fume suppressants) are not feasible. Furthermore, the new standards present an implementation and enforcement nightmare for facilities and regulatory officials to measure compliance with the NESHAP as proposed by EPA. Before proceeding with this rulemaking, EPA must address this enforcement and implementation issues and provide a reasonable and workable solution. XV. ENVIRONMENTAL JUSTICE CONCERNS ARE ADDRESSED EPA has claimed that the emissions from chromium electroplating and anodizing operations may pose a disproportionate risk for environmental justice communities. As discussed above, EPA has significantly over-estimated the emissions and risk associated with chromium electroplating and anodizing operations, and thus the potential risks to individuals are extremely low, so the impacts on all communities are minimal, and do not, therefore, pose a specific threat to environmental justice communities. XVI. THE PROPOSED RULE DOES NOT MEET EPA S CRITERIA TO IMPOSE ADDITIONAL CONTROLS AS PART OF NESHAP RISK AND TECHNOLOGY REVIEW EPA s decision to impose additional controls for the chromium electroplating NESHAP is not consistent with its decisions made with regard to other NESHAP risk and technology reviews. Specifically, EPA appears to use different criteria to justify the proposed changes to the chromium electroplating NESHAP. Provided below is a brief discussion on these criteria. A. Risk Criteria EPA has proposed to impose new controls for NESHAPs when it is justified by either risk or technology. With respect to risks, EPA has imposed new controls when: 73

74 unacceptable MIR cancer risks from inhalation are greater than 100 in a million; a large population is exposed to a cancer risk from inhalation greater than one in a million (76,000 individuals potential exposed to such risk was clearly acceptable in one case, and 849,000 individuals potentially exposed was acceptable in another); unacceptable MIR chronic non-cancer risks from inhalation have a target organspecific hazard index (TOSHI) well above one, and there is reasonable confidence that the modeled risks are real; and significant acute inhalation risks and/or multi-pathway risks exist. With respect to these criteria, EPA concluded generally that the risks associated with this proposed rule are acceptable. Specifically, EPA has not identified any MIR cancer risks greater than 100 in a million, and very few greater than 10 in a million. Based on the reasonable revisions discussed above in these comments, the risks may all be below one in a million. The number of individuals exposed to a potential cancer risk of one in a million is 180,000 individuals based on EPA s estimate in the supplemental proposal, but this number should decrease significantly below the criteria used by EPA. Similarly, the TOSHI for the rule is well below one, and EPA has not identified any significant acute risks or any HAPs that may present multi-pathway risk concerns. Consequently, EPA cannot justify imposing any new controls for chromium electroplating and anodizing operations based on these risk criteria. B. Margin of Safety and Technology Criteria If the risks are deemed acceptable, then EPA may impose additional controls based on the following ample margin of safety and technology criteria. The source category emits one or more HAPs that were not regulated in the original NESHAP. 74

75 There has been significant improvement in performance of one or more emission control technologies that were the basis of the original NESHAP. There is one or more new control technology identified since original NESHAP that is both cost-effective and can significantly reduce HAP emissions. Control has been found cost-effective if it is less than $30,000 per ton of HAPs reduced, generally, or less than $1,400 per pound of hexavalent chromium or mercury reduced. If co-benefits from the control of criteria pollutants greatly exceed costs. With respect to these criteria, EPA has no basis to justify new controls for chromium electroplating and anodizing operations because: EPA has not identified any additional HAPs since the promulgation of the original NESHAP EPA has not documented any significant improvement in performance for fume suppressants, mesh pads, scrubbers EPA has not identified any new cost-effective control technologies, and concluded that the proposed control technology options have a cost effectiveness of approximately $15,000 per pound of chromium emissions reduced. A discussed, these technology options are significantly less cost effective that EPA claims. EPA has also not identified any co-benefits from the control of criteria pollutant emissions, in addition to chromium emissions. C. Quality of Data for Decision Making For the most part, EPA has also been concerned to have a strong information basis before imposing new controls. For example, a recent ICR, good information on number of sources, good emissions information, and good data on performance of control technologies have been criteria for EPA to justify additional controls for a NESHAP based on a risk and technology review. As discussed above in these comments, EPA's 75

76 information basis for this proposed rule is limited and unreliable because: 1) the data and other information are outdated, 2) EPA has poor data on number, size and source category of the facilities impacted, 3) EPA has very little facility-specific data for emissions estimates (less than 1 percent of facilities), 4) EPA has no data on performance of non-pfos fume suppressants, 5) EPA modeling systematically over-estimates emissions, ambient concentrations, and the associated risks. Based on all of the criteria discussed above, EPA does not have any justification that is consistent with the criteria that have been used for imposing additional control for other NESHAP risk and technology reviews. XVII. CONCLUSION The existing Chromium Electroplating NESHAP has been very effective in controlling chromium emissions and reducing risks to potentially exposed populations. Consistent with these comments, EPA has systematically over-estimated the emissions from chromium electroplating and anodizing operations by at least 80 percent. The NASF has conducted a data collection effort to verify these current emissions. The risks associated with the emissions from chromium electroplating and anodizing operations are even lower that EPA has estimated and would impact few, if any, potentially exposed individuals. The proposed technology options identified by EPA cannot be used to meet the new proposed emissions limits. EPA has failed to provide any evidence in the record to support it claims that the use of non-pfos fume suppressants can achieve the new proposed emissions limits. Furthermore, EPA has not identified any new technology that can reduce risks at reasonable costs. Finally, EPA s approach to the proposed rule is inconsistent with the requirements of the Clean Air Act. Based on the substantial reductions in emissions for chromium electroplating and anodizing facilities and the associated risks, as well as the many procedural flaws in this rulemaking, EPA should not continue to pursue this proposed regulation further. 76

77 On behalf of the National Association for Surface Finishing (NASF), we appreciate the opportunity to submit these comments on EPA s February 8, 2012 supplemental proposal for National Emission Standards for Hazardous Air Pollutant Emissions: Hard and Decorative Chromium Electroplating and Chromium Anodizing Tanks. If you have any questions, would like additional information, or would like to discuss these comments, please contact Christian Richter (crichter@thepolicygroup.com) or Jeff Hannapel (jhannapel@thepolicygroup.com) of The Policy Group on behalf of the NASF. 77

78 ATTACHMENTS 78

79 Attachment 1 NASF Emissions Estimates Corrections 79

80 Summary of Hexavalent Chromium Emissions Company Location Process Action Rotary Die Inc. Advanced Graphics Tech EPA Emission s Estimate (lbs/yr) Correct ed Emissio ns (lbs/yr) Percenta ge Differenc e Addison, IL Hard % Grapevine, TX Hard % Reason for Correction Revised data provided by facility Revised data provided by facility AH Metal Fabricators Moore, OK Hard % Facility is closed Alloy Chrome Inc Anoplate Corp Schiller Park, IL Syracuse, NY Hard % Anodizin g % BIC Corp Milford, CT Hard % Facility confirmed emission estimate Facility has no emissions from anodizing line Facility no longer processes chromium, nor any plating Birken Mfg. Co. Bloomfield, CT Hard % Facility is closed Boeing Commercial Airplane Group Wichita, KS Hard % Facility scheduled to close in

81 Borden St Louis, MO Hard % Facility is closed Briggs & Stratton Capital Electroplati ng C & G Aircraft CMP Anodizing Corp Co- Operative Plating The Crown Group Danbury Metal Finishing Popular Bluff, MO Hard % Houston, TX Hard % Ferris, TX Hard % Elk Grove, IL Hard % St. Paul, MN Portland, TN Decorati ve % Hard % Danbury, CT Hard % Facility has not processed hard chrome since 2000 Facility decommision ed chrome plating tanks in 1998, moved out of city and is now only a machine shop Facility closed in 2008 Facility confirmed emission estimate Facility switched to trivalent chromium Facility no longer does hard chrome, only phosphating Revised data provided by facility, facility is very small and buys only small amounts of chromium 81

82 Delstar Metal Finishing DOE KC Allied Signal Dolloff Industries, Inc Engine Component s Fanning Electric Plating Feldkircher Wire Fabricating Co. Houston, TX Hard % Kansas City, MO Cleveland, OH San Antonio, TX Sherman, TX Hartsville, TN Hard % Hard % Hard % Hard % Hard % Fusion, Inc. Houston, TX Hard % GE Engine Services Griep, Inc, Tri-City Plating Dallas, TX Hard % Beloit, WI Hard % Facility has never processed hard chrome Facility does not process hard chrome Facility is closed Facility no longer processes chromium, only silicon nickel Revised data provided by facility Facility is closed Facility decommissio ned hard chrome plating in January 1995 Facility's hard chrome plating operation is closed Facility is closed Guild Platers, Inc. Long Island City, NY Hard % Facility is closed 82

83 Gull Industries (Cochran St.) Gull Industries (Cochran St.) Houston, TX Houston, TX Decorati ve Decorati ve % % Revised data provided by facility Revised data provided by facility Hard Chrome Evansville, IN Hard % Facility is closed HBM Co Lansing, IL Hard % Hiawatha Metalcraft Hi-Tech Chrome Plating & Polishing Co Hi-Tech Metal Finishing H & M Plating Co. Minneapolis, MN North Windham, CT Hard % Hard % Denton, TX Hard % Houston, TX Hard % Revised data provided by facility Facility never processed hard chrome and had only processed chromic acid anodizing, but not since 2008 Facility is very small with only two employees and processes only very small amounts of chromium Facility no longer processes hard chrome Revised data provided by facility 83

84 Hoffman Plating Company Houston Mfg & Speciality Hytek Finishes Cleveland, OH Hard % Houston, TX Hard % Kent, WA Hard % Facility is closed Facility decomissione d hard chrome tank 8 years ago Revised data provided by facility Indianhead Plating Chippewa Falls, WI Anodizin g % No anodizing process Industrial Chrome Plate Industrial Hard Chrome San Antonio, TX Hard % Newark, NJ Hard % Facility moved location and installed hooded tank system with zero emissions Revised data provided by facility Kilburn Plating Co Adamsville, TN Hard % Facility is closed Lane Plating Dallas, TX Hard % Revised data provided by facility Lincoln Plating Macs Plating Works Lincoln, NE Hard % Tulsa, OK Hard % Revised data provided by facility Facility is closed 84

85 Marine Corps Logistics Base Masterson Industries, Inc. Mendoza Meatal Coating Metal Finishing Corp. Metalplate & Products Metokite Corp Minnesota Rubber Mirror Industries National Chromium Co, Inc NEO Industries Albany, GA Hard % Round Rock, TX Andrews, TX Grand Prairie, TX Milwaukee, WI Lebanon, TN St. Louis Park, MN Hard % Hard % Hard % Hard % Hard % Hard % Houston, TX Hard % Woonsocke t, RI Hard % Decatur, AL Hard % Facility no longer does chrome plating Revised data provided by facility Facility is closed Facility changed name over 4 years ago and has never processed hard chrome Facility is closed (since 2009) Facility no longer does hard chrome, only phosphating Facility is closed and has not processed hard chrome for more than 8 years Revised data provided by facility Facility is closed Revised data provided by facility 85

86 NEO Industries NEO Industries NEO Industries Portage, IN Hard % Beaver Dam, KY Weirton, WV Hard % Hard % Revised data provided by facility Revised data provided by facility Revised data provided by facility Nuclear Plating Inc Cleveland, OH Hard % Facility is closed Nussmeier Engraving Evansille, IN Hard % Facility is closed Oshkosh Plating Serrvices, Inc. Perfection Plating Inc. Pitney Bowes Plastic Plate Inc. Poly-Metal Finishing Inc Oshkosh, WI Nashville, TN Stamford, CT Grand Rapids, MI Springfield, MA Hard % Hard % Hard % Hard % Hard % Faciity no longer processing hard chrome, only zinc and chromating Very small facility, revised data Facility no longer processes hard chrome, nor performs any plating Facility only processes decorative chrome, but is closing for refurbishing Facility no longer processes hard chrome, only 86

87 anodizing and chromating Precision Engineering Precision Finishing Inc. Precision Hard Chrome Precision Plating Corp Premier Turbines Progressive Plating Co Quebecor Printing Seattle Hard % Sellersville, PA Hard % Canton, OH Hard % Vernon, CT Hard % Neosho, MO Indianapolis, IN Mt. Morris, IL Hard % Hard % Hard % Facility is closed (August 2004) Facility has never processed hard chrome and has not performed any plating since 1965 Facility was closed when United Hard Chrome bought the company Facility is very small, a oneman operation and only buys minimal amounts of chromium each year Facility no longer processes hard chrome Facility no longer processes hard chrome, only zinc plating Facility no longer processes hard chrome 87

88 Remmingto n Products Co. LLC Renfrow Hard Chrome Co. Rhimco Industries Rolls-Royce Corporation Brideport, CT Hard % Lubbock, TX Hard % Mansfield, TX Indianapolis, IN Hard % Hard % Facility is closed Facility has only one tank with no stack and no emissions from the facility Facility does not have hard chrome plating and has never had hard chrome Facility no longer processes hard chrome Roll Service, Inc Glenwood, IL Hard % Facility is closed Royal Plating & Polishing Co. R & R Machine Pittsfield, MA China Springs, TX Hard % Hard % Facility does not process hard chrome, only a very small amount of decorative Revised data provided by facility, facility only has one hard chrome tank with KCH CMP system and estimate represents worst case scenario 88

89 Sandy's Bumpers Mart Syracuse, NY Hard % Facility is very small and does not process hard chrome, only decorative chromium Saueressig Jackson, MO Hard % Facility is closed Schumache r Co. Inc. Seidel Inc Sikorsky Aircraft SK Williams Co S & K Industries Houston, TX Hard % Waterbury, CT Stratford, CT Wauwatosa, WI Hard % Hard % Hard % Tulsa, OK Hard % Revised data provided by facility Facility no longer processes any chromium Facility do not process any chromium, nor conduct any plating operations Facility is closed (since 2006) Facility is closed Snap-On Tools Johnson City, TN Hard % Facility is closed Southern Aluminum Finishing Atlanta, GA Hard % Facility has never processed hard chrome, does anodizing, but not chromic acid 89

90 anodizing Southern Graphics Southern Hard Chrome Plating South Shore Plating Company Southwest Plating Company Speciality Plating Springfield Electroplati ng SPX Service Solutions St. Louis, MO Spartanburg, SC Hard % Hard % Quincy, MA Hard % Bridgeview, IL Mt. Vernon, WA Springfield, VT Owatonna, MN Hard % Hard % Hard % Hard % SR of TN Ripley, TN Hard % Facility no longer processes hard chrome Facility is closed (since 2005) Facility does not process hard chrome, only a very small amount of decorative Facility is closed Facility does not process hard chrome, only decorative chromium Facility is very small and processes only small amounts of chromium, primarily EN Facility does not process any hard chrome Facility no longer does hard chrome, only decorative 90

91 chromium Standard Plating Co. Standard Aero Standard Plating Inc. Stanley Tools Div State Plating LLC Superior Metal Finishing Superior Plating Inc. Supreme Bumpers Inc Surface Finishing Technologie s Milwaukee, WI San Antonio, TX West Springfield, MA Hard % Hard % Hard % New Britain hard % Elwood, IN Hard % Tualatin, OR Hard % Minneapolis, MN Hard % Toledo, OH Hard % Elmira, NY Hard % Facility is closed (since 2007) Revised data provided by facility Facility is closed Facility is closed and no longer proceses any chromium Facility is closed (since 2007) Facility has never processed hard chrome and stopped processing decorative chromium in late 1980's Facility closed in December 2011 Facility is closed EPA was using data from 1980 s before MACT controls were 91

92 Surftech Finishes Surgichrom e Inc. Tech Industries Tenneco Automotive Tenth Street Industries Texas Hydraulics BDA Precise Hard Chrome TMT Industries Inc. Kent, WA Hard % Clackamas, OR Woonsocke t, RI Hard % Hard % Cozad, NE Hard % Plano, TX Hard % Waco, TX Hard % Rochelle, IL Hard % installed, revised data was submitted to state in 1997 Facility is closed (June 2003) Faclity is closed -- burned downed Facility is closed, burned to the ground Facility scheduled to close July 2012 Facility is closed Facility confirmed emission estimate Facility is closed Torrington Co Torrington, CT Hard % Facility is closed Triple Plate Chrome Spokanne, WA Hard % Facility does not process hard chrome and only decorative chromium 92

93 Truelove & MaClean Turbine Chrome Services TWA Tyco Valves & Controls U.S. Polestar Marine Engineering Warterbury, CT Hard % Houston, TX Hard % Kansas City, MO Harlingen, TX Hard % Hard % Houston, TX Hard % Facility no longer processes any chromium Facility is closed American Airlines bought bankrupt TWA in 2001 and operated facility until it closed in 2010 Facility no longer processes hard chrome, but has decorative chrome that is exclusively trivalent (since 2007) Facility no longer at the location Utility Platers, Inc. Kingston, NY Hard % Facility is closed Val-Kro Inc. Van der Houst, USA Corp North Tonawanda, NY Hard % Terrell, TX Hard % Facility is no longer processing chromium Facility has been closed for 5 years 93

94 Vision-Ease lens Vought Aircraft Warsaw Plating Works Weatherfor d Aerospace Wearmaste r Wesco Valve & Mfg. Co Wilson Electroplati ng Wilson Manufactur ing Woodhill Plating Co. Ramsey, MN Hard % Dallas, TX Hard % Warsaw, IN Hard % Weatherfor d, TX Kennedale, TX Hard % Hard % Marshall, TX Hard % Binghamton, NY Green Park, MO Cleveland, OH Hard % Hard % Decorati ve % Facility has small 19 gallon tank and revised data provided by the facility Facility now owned by Triumph and has no hard chrome plating Facility is closed Facility is no longer processing chromium Facility has never processed hard chrome, now HVOF Facility decomissione d hard chrome tank over 10 years ago Facility is no longer processing chromium Facility no longer does hard chrome Facility only processes trivalent chrmium 94

95 TOTALS % ( lbs.) 95

96 Attachment 2 NEI Sources of Hexavalent Chromium Emissions 96

97 Sources of Hexavalent Chromium and Chromic Acid Emissions, from 2005 NEI Category Name Pollutant SOURCE Emissions (tpy) Waste Disposal Chromium (VI) POINT Indus Process - Metals Chromium (VI) POINT Indus Process - NEC Chromium (VI) NONPOINT Fuel Comb - Commercial/Institutional Chromium (VI) POINT Indus Process - Pulp & Paper Chromium (VI) POINT Construction Chromium (VI) POINT Non-Road Equipment - Gasoline Chromium (VI) NONROAD Fuel Comb - Commercial/Institutional Chromium (VI) NONPOINT Planes, Trains, & Ships Chromium (VI) NONROAD On-Road Vehicles - Diesel Chromium (VI) ONROAD On-Road Vehicles - Gasoline Chromium (VI) ONROAD Solvent - NEC Chromium (VI) POINT E-06 Indus Process - Petroleum Refineries Chromium (VI) POINT Fuel Comb - Industrial Boilers, ICEs Chromium (VI) NONPOINT Fuel Comb - Industrial Boilers, ICEs Chromium (VI) POINT Gas Stations Chromium (VI) NONPOINT 0 Indus Process - NEC Chromium (VI) POINT Fuel Comb - Residential Fireplaces Chromium (VI) NONPOINT 0 Indus Process - Oil & Gas Production Chromium (VI) POINT Graphic Arts Chromium (VI) POINT Surface Finishing Surface Coating - Industrial Chromium (VI) POINT Indus Process - Chemical Manuf Chromium (VI) POINT Fuel Comb - Residential Fossil Chromium (VI) NONPOINT Indus Process - Storage & Transfer Chromium (VI) POINT Non-Road Equipment - Diesel Chromium (VI) NONROAD Waste Disposal - Open Burning Chromium (VI) POINT Degreasing Chromium (VI) POINT Fuel Comb - Electric Utility Chromium (VI) POINT Indus Process - Cement Manuf Chromium (VI) POINT Fuel Comb - Electric Utility Chromium (VI) NONPOINT Gas Stations Chromium (VI) POINT E-06 Bulk Gasoline Terminals Chromium (VI) POINT E-08 Total for Cr (VI) % Indus Process - Pulp & Paper Chromic Acid (VI) POINT Graphic Arts Chromic Acid (VI) POINT Solvent - NEC Chromic Acid (VI) POINT Indus Process - Storage & Transfer Chromic Acid (VI) POINT Degreasing Chromic Acid (VI) POINT Fuel Comb - Commercial/Institutional Chromic Acid (VI) POINT Surface Coating - Industrial Chromic Acid (VI) POINT Fuel Comb - Industrial Boilers, ICEs Chromic Acid (VI) POINT Indus Process - Chemical Manuf Chromic Acid (VI) POINT Indus Process - NEC Chromic Acid (VI) POINT Fuel Comb - Electric Utility Chromic Acid (VI) POINT Total for Chromic Acid (VI) % 97

98 Attachment 3 EPA s Estimated Costs Associated with Add-On Controls and Retrofits 98

99 Chromium Electroplating -- Estimated Cost Effectiveness for HEPA Retrofit No. of plants Net nationwide reductions, Annualized tons/yr cost, $/yr Electropl ating Type Hard Chromium Cost effectiveness, $/ton NESHAP reductio ns, tons/yr NESHAP annualized cost, $/ton NESHAP cost effectiven ess, $/ton Small $17,640,000 $73,400, $4,020,000 $210,000 Medium $12,953,000 $27,200, $3,990,000 $90,000 Large $13,375,000 $18,400, $4,330,000 $50,000 Total $43,968, $12,340,000 Overall $30,500,000 $80,000 Decorative Chromium (a) Small $2,355,000 $117,700, ($680,000) ($270,000) Medium $1,467,000 $89,000, ($170,000) ($90,000) Large $2,740,000 $44,200, $200,000 $40,000 Total $6,562, ($650,000) Overall $66,600,000 ($70,000) Chromium Anodizing (a) Small $718,000 $213,100, ($730,000) ($920,000) Large $2,457,000 $148,100, ($530,000) ($180,000) Total $3,175, ($1,260,000) Overall $159,100,000 ($330,000) Total for All Electroplating Types Total 1, $53,705,000 Overall $34,400, $10,430,000 $70,000 (a) NESHAP cost effectiveness is negative due to assumption that the use of WAFS would decrease operating costs (net negative cost). 99

100 Attachment 4 Data on the Use of Non-PFOS Fume Suppressants from Minnesota Facility 100

101 101

102 Attachment 5 Gull Industries Case Study on the Use of Non-PFOS Fume Suppressants 102

103 3233 GANO STREET HOUSTON, TX TEL (713) FAX (713) Incorporated March 20, 2012 TO: Jeff Hannapel, The Policy Group President J. KELLY MOWRY 3233 GANO STREET HOUSTON, TX TEL. (713) CELL (713) FAX. (713) FM: RE: J. Kelly Mowry PFOS Alternative Test/Fail-Gull Industries, Inc Dear Jeff; Per our recent conversation, I ve put together the following for your review. I ll make this a quick and to the point report and keep my supplier somewhat protected but if you need more details, I ll gladly provide them. On July 23, 2010, well in advance of EPA s deadline to eliminate PFOS from the market place, we began experimenting with a non-pfos fume suppressant. The test was conducted on a production Armoloy bath at our facility which is a proprietary 54 ounce per gallon chrome 6, TDC bath with a redundant MAPCO mesh pad mist elimination system and existing PFOS fume suppressant in place. As the test began, no anomalies were noted other than in increase in the amount of non-pfos agent needed to maintain our surface tension as well as the timing of the adds and the testing needed. We continued the test with what seemed like successful results through 9/13/2010 when we tested for PFOS in our effluent and found that number to be 11, 500ppt and used that as our base line. On 5/11/11 we changed to a slightly different formula that had a little more foaming agent in it and after seeing little difference in performance and an increase in the amount of product needed, changed back to the original, non-pfos formula on 6/29/11. During this period, we also conducted our regularly scheduled PEL testing and found that while we were still within OSHA limits, our measured PEL had risen almost an order of magnitude and while we could not confirm it was a result of the non-pfos change, it was the only factor in our operation that had been altered. We repeated the test, insured all other aspects were perfect and found similar results. At our 6 month point, we again sampled for PFOS in our effluent and found that it was all but gone we were PFOS free! Unfortunately on September 19 th, 2011, the non-pfos product seemed to fall apart chemically. Our surface tension rose rapidly and uncontrollably. We began smelling chromic smells in the plant and within minutes, shut our entire operation down. Not only did we smell chrome, the product seemed to become suspended and airborne coming down onto flat surfaces in the plant. After two days of zero production due to the threat of risk to our staff, the non-pfos product was replaced and we resumed running with a PFOS surface tension product. We quickly regained our surface tension numbers as well as our clean air work environment. Once in place and stable, our non-pfos additive required approximately twice the amount of product to maintain less than our desired surface tension (we could never achieve the same low surface tension numbers). It required us more than double the amount of time to make adds (as they were needed more frequently) and with that, it 103