GCI TECH NOTES ©
This month's GCI Tech Notes will start an intermittent series of comments on selected aspects of the proposed EPA Hazardous Waste Combustor Rule. Contributions were made by the entire GCI staff.
The EPA has determined that cement kilns burning hazardous wastes should be regulated as major sources under 112 (c)(6) primarily because of HCl and Cl2 emission rates (FR17365). Also, the EPA has determined that cement kilns are area sources specifically due to PCDD/PCDF emissions and mercury emissions (FR17366), In making these determin-ations the EPA has ignored several facts and has also not followed the requirement of the CAA in making these determinations as subsequent applicability to MACT regulation.
HCl and Cl2 Emission Values
EPA's implication in the HWC preamble (FR17365), that cement kilns burning hazardous waste are major sources due to HCl and Cl2 emissions, is based on highly questionable data. The EPA has acknowledged that the sampling/ analysis test for determining HCl and Cl2 in stack emissions in cement kilns is an inappropriate method for that emission source. This is documented in a letter from Fosten Curtis, Test Support Section, Emission Measurement Branch USEPA to Walter H. Stevenson, Regulation Develop-ment Section Standards Development Branch, ESO (MD-13), dated June 1990. In addition, the test data used by the EPA to calculate the emission values were based primarily on data obtained during BIF Certification of Compliance (COC) tests. As such, this test data is representative of stack emissions from kilns operating at maximum chlorine input values, maximum flue gas temperature and flow rate and a number of other operating parameters associated with the "worst case" operating conditions man-dated by the BIF regulation. Consequently, the data set used by the EPA to make the implied determination that 80+% of the cement kilns burning hazardous waste should be "major sources" due to HCl and Cl2 emissions is highly suspect.
There can be no question that the EPA is aware of the inappropriateness of Method 0050 for the determination of HCl emissions from cement kilns. Nor could the EPA not be aware of the high chlorine feedrate and extreme operating conditions inherent in COC Tests. (FR17390) Consequently, the preamble has been careful not to make the determination that cement kilns burning hazardous waste are a major source solely on this questionable HCl and Cl2 data, merely implying that these devices may be major emitters of HCl and Cl2. GCI challenges such an implication and did so in our comments.
PCDD/PCDF and Mercury Emissions from Cement Kilns
EPA has made the determination that cement kilns are area sources subject to regulation under 112(c)(6) due to their emissions of mercury and PCDD/PCDF. However, the EPA must have "list(ed) categories and subcategories of sources assuring that sources accounting for not less than 90 per centum of the aggregate emissions of each such pollutant are subject to standards under subsection (d)(2) or (d)(4) of this section" 112(c)(6). EPA admits that this list has not been prepared. That is, the EPA has not prepared the list specifying cement kilns burning hazardous waste as a category or subcategory subject to the MACT nor has the EPA documented that such categories or subcategories constitute 90% of the aggregate emissions of these pollutant though the EPA has implied that such would be the case based on data they currently possess.
Here again, the database used by the EPA relies heavily on emissions data from COC Tests, tests that resulted in higher than normal emissions. In some cases PCDD/PCDF emission rates may be one order of magnitude or more higher when the kiln is operated at the extreme COC Test conditions versus normal operating feed rates and conditions. Similarly, mercury emissions will be higher at higher kiln exit temperatures and maximum feed rates.
By not following the procedures specified in the CAA, that is, by not publishing a "list of categories and subcategories of sources assuring that sources accounting for less than 90 per centum of the aggregate emissions of each such pollutant" the industry has been denied the opportunity to comment on the appropriateness of the data used for this determination.
With respect to mercury emissions, the EPA has assumed that hazardous waste fuel is the primary mercury input into the kiln (FR17393). This is not the case for many cement plants. For one plant, all of the mercury sample analysis for the input feed streams were below the level of detection for the analytical method. Consequently, the raw feed slurry alone could account for 91% of the mercury input. Similar results have been seen in other 1992 COC Test data. Indeed it is EPA's intent as stated in the preamble (FR17393), "Thus, the MACT floor level is based on hazardous waste fuel control". It is ludicrous to expect a cement kiln to use hazardous waste fuel mercury feedrate to control mercury input when as much as 90% of the mercury input is from the raw feed with coal likely contributing an additional 5%, especially when all of the analytical values are "non-detect".
This is another instance of the EPA performing an inadequate and uncritical analysis of available data. The EPA acknowledges that mercury in the raw feed and coal can "...contribute to cement kiln mercury feedrates and emission..." (FR17393), but discounts as "anomalous" data from one kiln that indicates "...substantially higher mercury feedrates in the raw material..." (FR17393). A cursory perusal of the metals mass balance submitted with several COC Test reports would quickly indicate that the raw feed and coal are the major sources of mercury input into many kilns and that this "anomalous" data may very well indeed be accurate. With such being the case, it is highly unlikely that a proper accounting of mercury emissions sources in aggregate would place cement kilns burning hazardous wastes as part of the "90 per centum". Nor may MACT be applied when the pollutant is the result of contaminants present in the raw material feeds.
With respect to PCDD/PCDF emissions, the EPA has used data, as noted above, from COC Tests. Tests that were specifically designed to produce high emission rates for metals, hydrogen chloride, chlorine and particulates, condi-tions that also produce high PCDD/PCDF emissions. Again, by not complying with 112(c)(6), the EPA has denied industry the opportunity to comment on the appropriateness of the data to be used in this determination. In addition, the EPA has used the "Dioxin Reassessment Document", a flawed and highly criticized document (See GCI Tech Notes, Volume 3, No. 2) as a reference in "an example of how today's proposal constitutes an action to assure that sources accounting for at least 90% of emissions of 112(c)(6) pollutants are subject to MACT standards". Clearly, the EPA has purposely ignored provisions of 112(c)(6) and then used selected, highly questionable data to justify their actions. They further ignored the large body of technical comments provided by the public on the "Dioxin Reassessment Document." Furthermore, GCI believes that EPA may be in violation of its own policies and possible federal law by referring to a document clearly marked "do not cite or quote."
Additionally, for some kilns the kiln raw feed may be a, or the, major source of PCDD/PCDF. This is pointed out in a paper presented at the Incinerator Conference in Savannah Georgia in May, 1996. (Ref. David Constans, Sources of PCDDs/PCDFs in Cement Kiln Emissions)
One final point should be made: 112(c)(6) specifically lists the compounds 2,3,7,8-tetrachlorodibenzofuran and 2,3,7,8-tetrachlorodibenzo-p-dioxin. It does not speak of dioxin like compounds, it does not speak of toxic equivalent values (TEQs). Logic dictates that Congress was being specific in carefully listing these named compounds. This will make a significant difference to cement kiln operators in that although these devices emit dioxin like compounds, they are primarily of the low toxicity hepta and octa congeners. If, indeed Congress was being this specific, the EPA may have exceeded their authority in setting emission limits as "0.2ng/dscm TEQ", and in evaluating anything but the specific congeners listed in the law.