GCI TECH NOTES ©
A Revised Technical Standards for Hazardous Waste Combustion Facilities notice of data availability and request for comments was published May 2, 1997. This NODA is part of the on-going EPA efforts to finalize the standards for hazardous waste combustors proposed April 19, 1996. Two previous NODAs were published in August, 1996 and January 7, 1997. In part, the May 2, 1997 NODA was published to inform the public of EPA's re-evaluation of MACT standard setting approaches based on some new data and voluminous public comments. This Tech Notes presents GCI's comments on this NODA.
by
Jim Woodford and David Gossman
Part Three: Implementation, I. Compliance Date Considerations (62FR 24235 & 24236).
The Agency states they have "become persuaded by commentors concerns regarding the ability of HWC sources in particular to comply with the proposed standards by the compliance date" because, in part, "sources will have to modify their RCRA permits." It is a simple fact that on May 18, 1993 USEPA Administrator Carol Browner announced her "New Hazardous Waste Reduction and Combustion Strategy" in which she directed "EPA's regional offices to immediately give their highest priority over the next 18 months to bringing existing facilities under rigorous permit controls." Eighteen months occurred in November 1994, the same time that the draft combustion strategy was finalized. Here it is June of 1997, over forty-eight months [4 years] after that initial announcement and only one cement kiln BIF has a final RCRA Part B permit. How can anyone in the Agency seriously believe that the time table proposed in this latest NODA is even remotely realistic? Let's see, one location out of approximately 20 locations in four years, given the same progress for the other 19 that would require at least 75 years. Granted, 75 years is an absurd exaggeration but given that only one location has been permitted since 1993, the current proposed time table is every bit as absurd.
Part Two: Revisions to proposed standards using the revised emissions database and data analysis method., F. Re-evaluation of proposed MACT standards for cement kilns, 2. Dioxins and furans (D/F) [62FR 24226]
EPA continues to rely on limited data from selected cement plants to extrapolate to the rest of the industry. EPA's similar past actions in this regard have met with unmitigated disaster. A good example of this is EPA's use of HC/CO as an indication of combustion conditions where hydrocarbons in the raw materials may completely mask any HC/CO related to the combustion process. Because of the key role that raw materials play in the operations and emissions from every cement plant, no two plants are the same. The control scheme that works at one plant will not necessarily work at another, and visa versa. As an example, Continental Cement Company, LLC has effectively reduced PCDD/PCDF emissions two orders of magnitude by changing raw materials with no change in APCD inlet temperatures. Given the lack of EPA authority to require changes in raw materials, the choice by EPA to ignore this data may be understandable but nevertheless scientifically regrettable. Tests at Continental have further demonstrated little, if any, effect by using activated carbon injection and temperature reducing water sprays. The agency has again chosen to ignore this data, referencing only that data which supports its misguided theories. Attempts to extrapolate across an industry where inherent differences will forever exclude such an approach are the role of "junk science." EPA has no technical basis for using such an approach given its knowledge of the industry and past failings at taking limited data from selected cement plants and then attempting to extrapolate the data across the industry.
Temperature Control
As part of this NODA, EPA suggests that temperature control at the APCD inlet at <400F is both a method for controlling PCDD/PCDF emissions and is "readily achievable." Gossman Consulting, Inc. (GCI) challenges both of these conclusions. First, EPA has provided no data supporting the need to drop from the previous proposal of 418 to 400F. Indeed, the agency's own statement that surface catalyzed formation occurs in the range of 450 to 750F would only support 450F as a limit. Attempts of one cement plant, Continental Cement, to use temperature control as a means of controlling PCDD/PCDF emissions have not been successful. A test that included a reduction of 40F from 520 to 480F at the APCD inlet did not have a statistically significant impact on PCDD/PCDF emissions. GCI further challenges EPA's assertion that because 20% of the data in its database was gathered with APCD inlet temperatures at <400F that such a temperature is "readily achievable" at all cement kilns all the time. The 350 to 400F APCD inlet temperature range in which EPA is suggesting all plants operate is, in fact, a very narrow window and not readily achievable by many plants. Differences in raw materials and plant design invalidate this extrapolation. While GCI acknowledges the overall trend in PCDD/PCDF data as it relates to APCD inlet temperature, this data cannot be scientifically extrapolated to prove that this control technique will work at all plants. This may be because of the PCDD/PCDF contribution from raw materials. Technical accuracy and honesty dictates that EPA must acknowledge this both in words and substance as it redrafts its proposal.
Data from Volume III of the Technical Support Document for HWC MACT Standards contains 19 PCDD/PCDF values listed for temperatures between 450 and 750F for cement kilns. These are listed in Table 1 below. Six additional values were also listed for temperatures below 450F.
While not conclusive proof beyond a shadow of a doubt, the 5.06 TEQ at <400F certainly casts doubt on control-ling APCD inlet temperatures to below 400F as an effective means of control.
Table 1: Reported APCD Temperatures Versus Reported Average PCDD/PCDF TEQ Values for Cement Kilns as Listed in Volume III of the Technical Support Document
| Temperature of APCD (EF) | PCDD/DF TEQ ng/dscm (avg) |
| 383 | 5.06 |
| 400 | .79 |
| 400 | .59 |
| 433 | 1.02 |
| 436 | 1.76 |
| 443 | 1.12 |
| 455 | 0.09 |
| 459 | 0.37 |
| 460 | 1.0 |
| 462 | 0.34 |
| 470 | 0.2 |
| 474 | 5.79 |
| 490 | 5.18 |
| 492 | 0.05 |
| 493 | 3.82 |
| 497 | 0.06 |
| 498 | 1.02 |
| 499 | 0.62 |
| 500 | 5.60 |
| 527 | 7.54 |
| 593 | 19.73 |
| 608 | 10.97 |
| 641 | 49.86 |
| 718 | 32.42 |
| 741 | 49.46 |
The 0.4 TEQ ng/dscm @ 7% O2 Standard
Continuing to use a limited database, EPA also suggests that the <400F limit can achieve a 0.4 TEQ ng/dscm @ 7% O2 standard. Further research reveals that EPA "discarded" data that did not meet its criteria on grounds that cannot be considered scientifically valid. It is difficult to argue with the specific "reasons" given for discarding the data. EPA only describes the conditions under which the data was gathered and, in fact, provide no "reason" whatsoever. We believe this is arbitrary and capricious on the part of EPA and may be in legal error. Finally, we must again point out that extrapolation across the industry is both unwarranted and has previously proven to be a technically flawed approach by the agency. This was the basis for comments on subcategorization. We will discuss, in a following section, evidence that PCDD/PCDFs in cement plant raw materials does impact emissions. Because of this, and the contradicting data in EPA's own database, EPA's conclusion that temperature control to 400F will always achieve PCDD/PCDF emissions below 0.4 TEQ ng/dscm @ 7% O2 is not technically correct, is not justified, and is in error.
Activated Carbon Injection
It is disheartening and frustrating that EPA continues to propose that activated carbon injection (ACI) can be used to achieve a BIF level of 0.2 TEQ ng/dscm @ 7% O2. EPA should be well aware that tests performed at Continental in 1994 using ACI and with an EPA contracted observer present conclusively demonstrated that ACI did not work on the Continental cement kiln. During this test Continental injected 100 lbs/hr (207.8 mg/dscm) of AC into the duct work between the kiln exit and the ESP. Samples of gas collected, both prior to the ESP and in the stack, clearly show that PCDD/PCDF emissions increase across the ESP while AC was being injected by approximately a factor of two. This is similar to results obtained while not utilizing ACI. More significant to note, however, was that total PCDD/PCDF concentration in the CKD increased by two orders of magnitude. This suggests that the ACI actually increased PCDD/PCDF generation rates in the system, an effect that has reportedly been observed when using ACI under other circumstances. While ACI may reduce stack emissions, EPA would be hard pressed to suggest that increases of emissions to another media are a legitimate control technology. EPA has not responded to previously identified problems of ACI use related to CKD recycling, reuse, and disposal. It would again appear that EPA has chosen a predetermined path and is apparently ignoring data or information that might interfere. The similarity between these circumstances and EPA's use of CO/HC as combustion indicators, despite conflicting data from the Continental kiln prior to the BIF regulations, is striking and disheartening.
Raw Materials as a Source of PCDD/PCDFs
Continental has over the past few years and in conjunction with stack testing for PCDD/PCDFs, also monitored PCDD/PCDF concentration in the raw materials fed to Continental's kiln. The results have been enlightening as to the source of PCDD/PCDF emissions because the raw materials contain a considerable quantity of naturally occurring PCDD. This information has been provided to EPA, yet it has obviously been ignored. EPA's repeated failure to recognize and accept valid data submittals strongly suggests politically motivated reasons for why EPA seems to prefer to continue to view PCDD/PCDFs as "man-made" pollutants, yet facts are facts. EPA cannot continue to deny the data demonstrating PCDD/PCDFs in raw materials. Other plants have observed the same pheno-menon. A GCI paper, which published some of these results, lends support to these comments. This paper can be found here. Since it is clear that under the MACT rules EPA does not have the authority to regulate a source of raw materials or the pollutants they contain, the agency must make an allowance in the rule where data demonstrates that raw materials are part or potentially all of the source of PCDD/PCDF emissions from a plant. It is further noted that despite probably 50+ years of "high" PCDD/PCDF emissions, Continental sampling and analysis of soil, water, flora, and fauna in the area around their plant reveals levels which are not elevated above normal background. We believe that there may be a decomposition mechanism that prevents these emissions from impacting the environment. One possibility is the photo-decomposition of PCDD/PCDFs on the alkaline particulate emissions associated with cement production. It appears to us that EPA and industry is expending a large amount of time and money to "solve" a problem that does not exist. Surely such data warrants further serious scientific attention by the EPA despite any political implications. Failure to do so may be legal error.
Where are PCDD/PCDFs Formed in the Cement Kiln process?
Along with the testing performed at Continental, testing has been performed for PCDD/PCDFs both prior to and after the APCD. This data indicates that EPA's assumption that PCDD/PCDFs are formed in the APCD system is not correct for all cement kilns. The data clearly shows that from half to all of the total PCDD/PCDF emissions were present prior to gases from the kiln entering the APCD. It is quite possible that these results are related to the naturally occurring PCDD/PCDFs found in raw materials which are simply boiled off and carried with the gas flow up the stack to be sampled. Partial destruction would account for the varying congener concentrations seen in the duplicate stack emissions tests. EPA cannot continue to ignore quality data that conflicts with its pet theories. The re-proposal of the MACT standard must consider the full range of data and information available without excluding information that does not fit EPA's predetermined goals. Clearly, the fact that PCDD/PCDF have been found in the gas stream prior to the APCD significantly impacts EPA conjectures regarding potential control technologies as well as mechanisms of formation, and lends support of the effects of raw material born PCDD/PCDF contaminants which occur naturally in some raw materials.
PCDD/PCDF Sampling and Testing
During the considerable testing of PCDD/PCDF emissions at Continental, there are a number of things about the data gathered that cause concern about the accuracy and quality of PCDD/PCDF emission data for cement kilns. There has been a high degree of variability in the data despite nearly identical operating conditions (as much as one full order of magnitude). During the recently completed trial burn, one day's variability over three runs was a factor of three despite very tightly controlled operating conditions. There have been a few instances at Continental of enormously high data - so high that the data was clearly in error and was not utilized to attempt to draw further conclusions. This has occurred despite Continental's use of highly qualified and technically experienced stack testing crews and laboratories. Based on these observations as well as data we understand CKRC has gathered and analyzed, GCI must conclude that EPA has not properly allowed for potential errors inherent in EPA's testing methods in the establishment of the MACT standard . Based on available data, GCI also suggests that EPA needs to research the potential for PCDD/PCDF formation inside the Method 23 sampling train given the wide, varied and perhaps heretofore unrecognized precursor and reactive sites that may exist in the sampling train when applied to cement plant emissions.
Conclusion
Gossman Consulting, Inc. believes that EPA continues to "cherry-pick" PCDD/PCDF emission data from cement kilns to establish predetermined MACT limits and potential control technologies. Unfortunately, this is not unlike similar criticisms directed towards the EPA, by its own Science Advisory Board, where they drew the conclusion that "research data does not support the report conclusions" in EPA's dioxin reassessment. Continental Cement, LLC has performed extensive testing of their kiln system that clearly demonstrates the major errors in EPA assumptions and conclusions. EPA's reconsideration of industry data for preparation of this NODA constitutes nothing less than the clearest form of "junk-science" imaginable. We strongly urge the agency to step back away from their predetermined conclusions. Return to the law and quality science. Continued failure to analyze the available data in a scientifically sound manner is an embarrassment to both the agency and the industry. We need sound rules and limits based on scientific principles designed to protect human health and the environment, not the politically motivated, predetermined approach used to date. Gossman Consulting, Inc. continues to support sound regulation based on congressional intent and scientific principals. Regulations developed to satisfy political agendas does not best serve human health and the environment.
Part Two: Revisions to proposed standards using the revised emissions database and data analysis method., F. Re-evaluation of proposed MACT standards for cement kilns, 7. Hydrochloric Acid and Chlorine (HCl/Cl2) [62FR 24230]
In the proposed hazardous waste combustor (HWC) regulations (FR 61 17358 - 17536), USEPA states that "Section 112 of the CAA requires that the EPA promulgate regulations requiring the control of hazardous air pollutants emissions associated with categories or subcategories of major and area sources." EPA then goes on to point out that cement kilns "meeting the major source criteria do so for HCl and Cl2 emissions...." Interestingly enough, EPA then goes on to say that "Notwithstanding the fact that some hazardous waste-burning cement kilns may not meet the definition of major source, the Agency is proposing to subject all HWCs to regulation under MACT, as major sources, under the authority of §112(c)(6)." (Part Three: A & C, FR 61 17365)
In Part Four: I. B. 3, EPA begins to identify the weakness in their efforts to pull cement kilns under MACT by using unreliable HCl data by stating, "However, the test method used to determine HCl and Cl2 emissions (BIF methods 0050, 0051, and 9057, commonly referred to as 'Method 26A') may not be able to distinguish between HCl and Cl2 in all situations." (FR 61 17376) EPA then points out in footnote #37 that even though "owners and operators of cement kilns have argued that this method provides measurements that are biased high because metallic salts penetrate the filter and chloride is incorrectly reported as HCl. EPA has considered this concern and continues to believe that metallic salts do not significantly bias the results." It is significant to note that while EPA makes an attempt to debunk industry claims, they offer no data to support their position.
This NODA response will further demonstrate that EPA has ignored their own findings. A meeting occurred in late 1989 or early 1990, related to municipal waste combustor (MWC) regulation, between a cement company representative, the company's consultant, USEPA Test Support Section, Emission Measurement Branch representatives and someone from the Atmospheric Research and Exposure Assessment Laboratory (AREAL) who also happened to be the individual who supervised the Method 26 development work. This meeting resulted in the generation of a June, 1990 USEPA memorandum to Walter Stevenson, Regulation Development Section, Standards Development Branch. That memo was also copied to five other individuals. The memo states, "It appears to us in the Emission Measurement Branch, as well as to [AREAL], that Method 26 is not suitable for testing cement kilns because no HCl should be present theoretically in the emissions, and the potential exists for anomalous results from interferences." A research paper, which discusses the cement plant data presented at this meeting, can be found here.
While no specific publication date is available, a USEPA document entitled Summary of Comments and Responses for Methods 23 and 26 came out about one year after the aforementioned meeting. The introduction of these Method 23 & 26 comments indicates that publication was after the March 5, 1990 Method 23 & 26 comment period (FR 54 52190). The question response beginning on the page numbered 22 and ending on the page numbered 23, states in part, "Filter penetration by more volatile particulate chloride salts, such as ammonium chloride (NH4Cl), may occur when testing sources such as cement kilns. However, the method [26] was not developed for use at cement kilns, and these sources will be granted a waiver from the hydrochloric acid (HCl) testing requirements." A further comment on the page numbered 23 is responded to on the page numbered 24 by stating, "The method [26] was not developed for use at cement kilns: therefore the quality of data obtained by sampling this matrix with Method 26 is unknown. Cement kilns that qualify as MWCs will be granted a waiver from the HCl testing requirements."
A situation that further compounds the issue of the formation of metallic salts vs HCl is recent data indicating that HCl may actually form in the sampling train. Consequently, results of stack sampling that may appear to be pulling HCl into the sampling train from the stack gases, are possibly reflecting the formation of HCl in the sampling train, thus producing HCl positive results which are not the result of HCl in the stack gases. There is also evidence that some of the more recent FTIR data is also flawed because of this same mechanism. The data suggests that HCl "emissions" increase the longer the sampling train is operated, pointing to a mechanism for the production of HCl from gases in the train interacting with particles on the filter. A study presented at the 1996 AWMA BIF Conference, on GFCIR analysis of HCl showed higher HCl results than Method 26. Method 26 was likely producing more HCl in the sampling train by virtue of the higher train temperatures (Hot-wet Instrumental Hydrogen Chloride Emissions Quantification using GFCIR - Method Validation and Comparison by Powell & - page 125 of proceedings).
In addition, there is cement kiln manufacturing system logic to consider here. Kurt Peray discusses the alkaline environment of the cement kiln manufacturing system in Chapter nine of The Rotary Cement Kiln. EPA has even provided supporting information concerning cement kilns in a number of documents including The Combustion Emissions Technical Resource Document (CETRED). Given that cement clinker production requires a manufacturing system that essentially makes lime in a highly alkaline environment, it doesn't even make sense that hydrochloric acid would get produced in such an environment. There may be an appearance that Method 26 results indicate HCl is being produced in cement kiln stack gases, however, analysis for cations at several cement kiln compliance tests since 1992 clearly indicate the overwhelming presence of significant quantities of ammonium salts rather than HCl. And even though analysis of the sampling trains reveals the presence of HCl, there is increasing potential that this analysis simply reflects formation of HCl in the sampling train after the stack gases have been sampled.
There is also a procedural point to ponder. The cement industry commented on the BIF rule that Method 26 was not valid. But this was not a major issue at the time since the BIF limits were comfortably well above any erroneous results that might be obtained with Method 26 and the cement industry chose not to vigorously challenge the issue. But now EPA chooses to use this challenged Method 26 data to pull cement kilns in under MACT. The fact that data from Method 26 exists for so many cement kilns does not automatically make the data valid. In fact, evidence presented in this NODA response clearly indicates there are serious problems with Method 26 for cement kilns. This in turn would undermine the appropriateness of using that data for a major source determination under the CAA.
Consequently in conclusion, the highly alkaline environment of the cement clinker production system undermines the chemical likelihood that HCl would even be formed in the first place. In addition, sample train analysis of anions clearly demonstrates the presence of salts rather than HCl. And the HCl that is detected increasingly appears to reflect the formation of HCl in the sampling train rather than HCl from the stack gases. It is obvious by reviewing the history of Method 26, as laid out in this document, not only was Method 26 not developed for cement kilns and recognized as technically flawed as early as 1990, but by EPA's own admission, "The method [26] was not developed for use at cement kilns: therefore the quality of data obtained by sampling this matrix with Method 26 is unknown."
EPA has proposed an entire set of regulations based upon the legal requirement that "Section 112 of the CAA requires that the EPA promulgate regulations requiring the control of hazardous air pollutants emissions associated with categories or subcategories of major and area sources" and then the flawed technical logic that cement kilns "meeting the major source criteria do so for HCl and Cl2 emissions...." It seems clear that the supposed HCl data used to pull cement kilns under MACT is not valid HCl data at all and removes the basis for EPA's cement kiln MACT standard. It can't be valid since EPA has not validated Method 26 for cement kilns and it is crystal clear from review of historical information on Method 26 that there are technically competent people within the USEPA that point out that the data is not valid. It seems all too clear that EPA is proposing MACT standards for cement kilns based upon invalid data.
Part Three: Implementation, II. Compliance Requirements, A. Compliance with CO and/or HC Emission Standards.[62FR 24237]
The Agency states that "Even though the vast majority of the data indicate that HC will be low when CO levels are low, a requirement to confirm this relationship on a site-specific basis may be warranted." The Agency continues to perpetuate CO and HC as indicators of good combustion. This statement also implies that CO and HC track together. Both of these issues were addressed as part of a court challenge. A US District Court (Horsehead Resource Development Company, Inc. vs. Carol M. Browner, Administrator USEPA, 16 F. 3rd 1246, 305 U.S. App. D.C. 35) held that "neither CO nor THC emissions are an accurate indicator of the combustion efficiency of wet kilns...." In addition, the results of test runs from a single wet process cement kiln located at Hannibal, Missouri (56 Fed. Reg. 7157 col. 1, 7163 col. 2 & Id: Emissions Testing of a Wet Cement Kiln at Hannibal, Missouri, BBSP-S0147I) were re-analyzed after the results were referenced in the BIF rules (FR 56 7163, footnote #42 and FR 56 7180, footnote #71) where it was discovered that the highest CO value occurred at the exact same time as the lowest HC value. This is a clear indication that HC & CO do not track together. Much other data has also been submitted to the EPA, however, this particular set of data happened to be data that EPA cited in a finalized rule. Data that disproves their long time HC/CO contentions.
Part Three: Implementation, II. Compliance Requirements, B. Startup, Shutdown, and Malfunction Plans [62FR 24237 & 24238]
In the next to last paragraph of this section, the Agency states that "if a source is temporarily not burning waste and a malfunction occurs that is followed by an exceedance of an applicable standard, the source will not be in violation as long as it is complying with the procedures outlined in the malfunction plan." The fact of the matter is that the source would not be out of compliance for any parameter if they were not burning hazardous waste at the time, with or without a malfunction plan. What kind of an example is this?
And since when is a standard exceedance following a waste feed cutoff a violation? When the cement kiln is not burning hazardous waste, there is no reason for any of the hazardous waste combustor limitations to apply. Very simply put, if the source is not burning hazardous waste then the hazardous waste parameters do not apply. The Agency says as much in footnote #49 at the bottom of this same page. This is some pretty twisted logic proposed by the Agency.
Part Three: Implementation, II. Compliance Requirements, C. Metals Extrapolation and Interpolation Consideration.[62FR 24238]
Gossman Consulting, Inc. supports Agency efforts to allow metals extrapolation similar to the guidelines provided long ago in the BIF Technical Implementation Document. It is worth noting however, just as has been previously commented upon many times, there is no need to include beryllium in the regulated metals of concern.
Part Three: Implementation, II. Compliance Requirements, D. Consideration of Site-specific Variances for Cement Kilns and LWAKs. [62FR 24238 & 24239]
Gossman Consulting, Inc. strongly disagrees with the Agency contention that congressional intent has no bearing on a final rule/law. There are no doubt court cases which bear this out. Documents such as the Federalists Papers, the pre-congressional equivalent to congressional intent, are often referred to in order to understand the intent of rules and laws such as the Constitution. To state that congressional intent has no bearing on the finalized rule/law is really quite naive.
Gossman Consulting, Inc. supports the variance procedures proposed as items (1) & (2) in the first paragraph of this section. GCI also is providing comments on the following issues:
Part Three: Implementation, III. DRE Testing Considerations. [62FR 24240]
Gossman Consulting, Inc. has been involved in examining the issue of HC and CO in cement kilns since 1992. GCI has been involved in discussions with EPA representatives and have provided research papers on the subject. One of those papers was mentioned along with comments on HC and CO earlier in this NODA. The GCI comments also notes a well known court case which found that HC and CO are not good combustion indicators across the board. Yet EPA continues to ignore all the data and cling to their misguided contention that not only are HC and CO good combustion indicators but that they also track together. This is hogwash. Data cited by EPA in the BIF rule and subsequent data made available to the Agency demonstrates that this is not an iron clad law of combustion science as EPA would have us believe. And now the Agency wants to throw DRE into this equation. These efforts are absolutely wrongheaded. Why not have HC and CO be indicators of metals and PM compliance as well? There may be other ways to ensure DRE related compliance but HC and CO are definitely not a consistently reliable approach.
A final note on this issue. Four nines could probably be achieved in any cement kiln at just about any level of HC & CO. In fact, the Agency makes a similar statement at the end of the second paragraph of section A. Options for Ensuring Compliance with a DRE Standard. Where is the logic here?