Back to Library

Gossman Consulting, Inc.

GCI TECH NOTES©


Volume 3, Number 10                 A Gossman Consulting, Inc. Publication                    October, 1997

Comments on the 5/2/97 Draft Guidance on Structuring RCRA Trial Burns for Collection of Risk Assessment Data

Jim Woodford

This issue of GCI Tech Notes was substantively taken from Gossman Consulting, Inc.'s comments on this draft guidance document as submitted to the USEPA. Some modifications have been made for clarification purposes for those of our readers who do not have access to the draft guidance document.

USEPA guidance can be very useful. This is particularly true when that guidance factors experience into previously finalized regulations. Recognizing that guidance does not and cannot legally take the place of final regulations, well written guidance can still be very helpful. In many ways this Guidance on Structuring RCRA Trial Burns for Collection of Risk Assessment Data is very well written. However, this guidance repeatedly attempts to get industry to perform data collection testing for issues about which EPA is curious, rather than just providing guidance concerning existing regulations. Granted, Administrator Browner did point out that "EPA will use its omnibus authority on a case-by-case basis as necessary to protect human health and the environment to include appropriate conditions in permits being issued" (5-18-93) but omnibus authority does not authorize the requirement for facilities to do unnecessary testing just so EPA can collect curiosity data. In addition, Administrator Browner's pledge to "conduct further analysis on the nature and toxicity of combustion emissions and their contribution to overall risk" does not direct permit writers to force facilities to perform unnecessary testing in order to satisfy EPA data curiosity.

The purpose of this Guidance on Structuring RCRA Trial Burns for Collection of Risk Assessment Data document is stated, in the abstract as being, "to establish guidelines for generating emissions data for incorporation into multi-pathway, site-specific risk assessments." The need for this was purportedly initiated/established by the May 18, 1993 U.S. Environmental Protection Agency Administrator Browner announcement of a draft Hazardous Waste Minimization and Combustion Strategy which "was finalized in November, 1994." The combustion strategy announcement called for full risk assessments, an appropriate PCDD/PCDF emission standard and more stringent control for metals without providing scientific justification for such actions. And while the BIF rules are mentioned in the May 18, 1993 combustion strategy announcement, nowhere in the announcement does it state that the BIF rules are not protective of human health and the environment. In addition, the November 18, 1994 "Final EPA Strategy for Hazardous Waste Minimization and Combustion" provides no scientific justification either but rather cites concerns raised in a series of public meetings.

With that in mind, it is of great interest to note that the Agency was quite confident that human health and the environment were protected by the BIF rule which went into effect approximately two years earlier. "The Agency firmly believes that the 0.08 gr/dscf PM standard, when used as a supplement to the risk-based metal controls provided by today's rule, provides protection of human health and the environment." (emphasis added) [FR 56 7145] The Agency went on to state later in the BIF preamble that, "Hypothetical risk assessments have shown that a 99.99 percent DRE standard for POHCs is protective of risks posed by emissions of organic constituents in the waste in virtually every scenario of which the Agency is aware." (emphasis added) [FR 56 7146] and cited the Engineering Science Background Information Document for the Development of Regulations to Control the Burning of Hazardous Waste in Boilers and Industrial Furnaces, Volume III as support for their statement. And while EPA recognized in the BIF preamble, "that the contribution of indirect pathways may be significant", the Agency also pointed out "that other conservative procedures, such as apportioning 75% of exposures to either indirect pathways or other emissions sources (that can contribute to background levels) in the calculation of RACs, will help offset the contribution of indirect pathways." And then finally, "Use of the MEI in the Screening Limits procedure comprises yet another conservative element in the risk assessment process which would offset direct estimation of indirect pathway exposure. Therefore, the Agency has not modified the risk assessment process to address indirect pathways." [FR 56 7169] While the draft Hazardous Waste Minimization and Combustion Strategy announced "full risk assessments including those for indirect exposure", it is clear that the BIF rules "offset the contribution of indirect pathways" exposures through "other conservative procedures."

One would certainly have to conclude, by the aforementioned Agency statements, that the BIF rules were in fact quite protective of human health and the environment. These rules took several years to develop, going through an extensive rule making process, including public comment, before they went into effect on August 21, 1991. Less than two years later, without going through any kind of a rule making process of any kind, and without any formal scientific input process, EPA Administrator Browner announced her so called combustion strategy. It would appear that the 1992 election somehow negated seven years of detailed rule-making which culminated in the BIF rules. It is worth repeating that nowhere in the combustion strategy announcement nor in the final strategy is it stated that the BIF rules are not protective of human health and the environment but rather that the rules should just be more stringent based upon public concern and political reasons rather than upon scientific reasons.

It is interesting and then disappointing that Attachment 1 is divided into only two sections, one for hazardous waste incinerators and one for BIFs. While this may fit the regulatory distinction, a cement kiln is a very different beast from a boiler or other industrial furnaces. Separate guidance is needed. It is also disappointing that not one of the examples in this guidance document specifically addressed cement kilns. This unfortunately continues to indicate that EPA just doesn't understand cement kilns. Given that the overall document tends to lump cement kiln requirements with other combustion device requirements, except that the guidance does note that organics can come from cement kiln raw materials, the following comments are provided.

Page 1 - I. Introduction

It would seem to be highly irregular to cite documents that are clearly stamped "REVIEW DRAFT (Do Not Cite or Quote)" as is the case with both the draft Mercury Study Report to Congress and the draft Dioxin Reassessment. Both of these documents are being reviewed to be finalized and in the case of the Dioxin Reassessment, USEPA's own Science Advisory Board has very clearly stated that "scientific findings portrayed in the draft document's conclusions is not balanced vis-a-vis the possible risks posed by exposure to dioxin, with a tendency to overstate the possibility of danger." In addition, the SAB pointed out that "important uncertainties associated with the Agency's conclusions are not fully identified and are not subjected to feasible analyses. Finally, the characterization of non-cancer risk is not performed in a manner that can facilitate meaningful analysis of the incremental benefits of risk management alternatives." To mention these documents as studies that have been conducted is certainly valid but to cite them as studies that "indicate that there can be significant risks from indirect exposure pathways" perpetuates the problems and technical inaccuracies of these documents as identified by credible reviewers and is highly inappropriate.

The third sentence in the first paragraph greatly exacerbates the problems of the draft report "findings", widely publicized across the country upon initial release of the report, "findings" which have since been seriously questioned as noted above. While a reading of the draft report may result in a belief by some that, "In many cases, risks from indirect exposure may constitute the majority of the risk from a hazardous waste combustor", SAB review conclusions clearly bring this statement into serious question and it certainly does not belong in this guidance document. It directly reflects on the credibility of the document and the Agency. And last but not least, surely it is a violation of internal Agency policy.

Page 2, I. Introduction, C. Relationship to Existing Guidance

This section states that "This document replaces the June 1994 'Guidance on Trial Burns'", yet it is cited as a reference source in Attachment 1A. Consequently, a document that has been "replaced" is used as a "current" reference document. This needs to be rectified.

Page 3, 1. Metals Data

Where or what is the justification for the statement that, "It may also be necessary to obtain data on previously unregulated metals (Ni, Se, Al, Zn, Mn, Cu, etc.)"? Given that the BIF rules took seven years of development, logic would dictate that there was probably no valid reason to regulate additional metals. In fact, according to footnote 54 (FR 56 7171) "Nickel is not controlled because the two nickel compounds suspected at this time of being potential human carcinogens are not likely to be emitted from combustion devices, given their highly oxidizing conditions." The guidance needs to justify this statement with scientific data or drop the suggestion.

Page 3, 2. Dioxins/Furans and Other Organic Data

Dioxins/Furans is often used to denote emission concerns however it is more technically accurate to reference the specifically targeted subset of PCDDs/PCDFs. While this may not be essential throughout the document, it is appropriate at least the first time mentioned in a document.

Page 4, C. Development of Trial Burn Plans and Risk Assessment Protocols

The final sentence in the first paragraph of this section "The trial burn should not be conducted until the protocol has been approved by the regulatory agency" presents a potential conflict. At first reading the sentence may make sense, given the intent of this guidance document. However, what if a facility has a trial burn plan that has already been approved? While this sentence may be sound advice, it should be specifically directed at the Agency rather than leaving it open for interpretation that somehow the regulated facility is at fault. In addition, what if a facility has already conducted their trial burn? There needs to be some clarification here.

Page 5, D. Alternative Data Sources

This section of the guidance, albeit brief, is to be applauded for including "data in lieu of" alternate sources such as "similar units burning similar wastes." It is clear that previous data from the exact same unit would exceed the "similar units" specification. A slight text modification is recommended here. Not only can alternate sources be used "if the permitting authority deems them to be appropriate" but "the Director shall approve a permit without a trial burn if he finds that the hazardous wastes are sufficiently similar, the devices are sufficiently similar, the operating conditions are sufficiently similar, and the data from other compliance tests, trial burns, or operation burns are adequate to specify (under 266.102) operating conditions that will ensure conformance" (emphasis added) This is a regulatory requirement and this brief section should be modified to reflect this directive.

Page 5, III. Trial Burn Objectives

It should be noted that the requirement to "demonstrate that emissions from hazardous waste combustion and ancillary operations produce risks which are less than or equal to the target cancer risk (1x10-5) and hazard index (0.25), as outlined in the May 5, 1994 Implementation Guidance Memorandum [site]" in the second bullet refers to a guidance memorandum and not final rule requirements.

Page 6, Trial Burn Objectives, A. Normal Versus Worst-Case Operation

In the very first paragraph of this section it is stated that "The term 'worst-case' is generally meant to mean 'maximum allowable'." This is a gross over-simplification and misleading. In addition, it is not consistent with section IV, B language which discusses maximizing or minimizing rather than "maximum allowable". It is recommended that this inaccurate generalization be modified to reflect maximum or minimum or it should be struck from the document.

Page 7, General Policy Statement

It seems highly inappropriate to include any kind of a policy statement in a guidance document. Surely policy should undergo the rule promulgation and public comment process.

Page 8, worst case and normal conditions

Section 1. For testing performed under worst-case conditions, the guidance suggests that all three runs must be in compliance. Three runs was initially implemented back in the early days of compliance testing in order to have three points to average since one run does not a valid test make. This is also evident in the BIF rule. 40CFR 266.102 requires only that the runs be "valid" and then the limits are established by taking the average of those test runs. [266.102(e)(2)(ii)(B) & 266.102(e)(6)(i)(B)(2)] In addition, what if your worst case is also your normal case? There is data available which indicates that in cement kilns, "normal" test days are very similar, if not statistically the same as "worst-case" days.

In addition, the "general policy statement" appears to, in effect, force a facility to use "worst-case" over "normal" conditions. It is as if USEPA wishes only to pay lip service to "normal" operating conditions when the facility is actually being forced into a "worst-case" risk assessment. Frankly, this does a disservice to the Agency as it comes across as if the Agency doesn't really know how to deal with "normal" conditions so it forces the facility into "worst-case" conditions.

Page 9, Trial Burn Objectives, C. Specific Emissions/Performance Standard Trial Burn Objectives, 2. Dioxin/Furan (D/F) Emissions

This section states that "Within this critical temperature window, D/F levels have been generally observed to increase by an order of magnitude for every 125o F increase in dry air pollution control device temperature." The structuring of this statement implies that a very strong correlation between PCDD/PCDF emissions and the temperature of the APCD exists. The data does not merit such a statement.

This statement references the data in Volume III of the Technical Support Document for HWC MACT Standards. TSD Volume III contains 53 tests of incinerators of which nine have APCD temperatures at or above 450F. These are listed in Table 1 below in increasing temperatures rather than increasing average PCDD/PCDF emission rate as was done in the document.

Table 1: Reported APCD Temperatures Versus Reported Average PCDD/PCDF TEQ Values for Incinerators as Listed in Volume III of the Technical Support Document

Temperature (F) PCDD/PCDF TEQ ng/dscm (avg.)

450

18.36

450

20.10

457

8.50

460

1.31

460

2.13

460

2.37

500

4.51

500

8.02

First of all, this is too narrow a temperature range to draw such a conclusion. Secondly, the trend actually appears to be the reverse from that stated.

In addition to this incinerator data, there are 19PCDD/PCDF values listed for temperatures between 450 and 750F for cement kilns. These are listed in Table 2 below. Six additional values were also listed for temperatures below 450F as well.Table 2: 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/PCDF TEQ ng/dscm (avg) Temperature of APCD (EF) PCDD/PCDF TEQ ng/dscm (avg)
383 5.06 492 0.05
400 0.79 493 3.82
400 0.59 497 0.06
433 1.02 498 1.02
436 1.76 499 0.62
443 1.12 500 5.6
455 0.09 527 7.54
459 0.37 593 19.73
460 1 608 10.97
462 0.34 641 49.86
470 0.2 718 32.42
474 5.79 741 49.46
490 5.18

Clearly, there is a general correlation between APCD temperature and PCDD/PCDF emission rates. However, "order of magnitude" higher concentrations were measured at lower temperatures than the two lowest values which occur at 492 and 497F. It is clear that the statement that there is a "one order of magnitude increase for every 125F increase in temperature" for the temperature range 450 to 750F is at the very best highly simplistic and definitely misleading, even with the modifier of "generally" added to the statement. A more logical conclusion based on this data is that between 400 and 500F the PCDD/PCDF rates are primarily impacted by some feature other than temperature, and that between 500 and 750F, temperature may become a more significant factor.

It must be remembered that the tests from which this data was extracted represent a number of different kilns. Also, the majority of this data was taken from BIF COC tests that were specifically designed to produce a "worst case" operating envelope - that is, maximum feed rates, maximum temperatures and maximum flue gas velocities as well as other operating conditions that could adversely impact PCDD/PCDF emissions. Such variability in the devices tested and the selection of the operating parameters greatly increase the difficulty of making generalized statements regarding increasing PCDD/PCDF emissions versus APCD temperature.

By removing some of this variability, specifically by limiting the data to one cement kiln, one can

see an entirely different temperature and emission relationship. Source number 319 executed a number of PCDD/PCDF tests. The values for these tests are listed in Table 3 by increasing temperature.

Table 3: Reported APCD Temperatures Versus Reported Average PCDD/PCDF TEQ Values for Cement Kiln Source 319 as Listed in Volume III of the Technical Support Document

Condition Temperature of APCD (F) PCDD/PCDF TEQ Ng/dscm (avg.)
C9 426 0.16
C5 443 1.12
B1 462 0.34
C7 474 5.79
C52 497 0.06
C53 499 1.11
C54 518 0.61
C6 527 7.54
C50 562 0.95
C51 568 0.13
C2 593 19.71

Clearly, the operator of this kiln is manipulating the PCDD/PCDF emission rate by controlling some operating parameter(s) other than APCD temperature. If one kiln can demonstrate such a scatter gram of PCDD/PCDF emissions versus increasing temperature, it is impossible to accept the statement in question. The bottom line is that such simplistic statements can be very problematic for the regulated community. The problem is exacerbated even more when the statement is not supported by the data. It also undermines the reliability of a guidance document.

In addition, this over simplification of PCDD/PCDF increase with temperature is supported from a draft regulation as cited in footnote #2 on page 10. Draft regulations should not be used as support for EPA guidance documents.

Page 11, III. Trial Burn Objectives, C. Specific Emissions/Performance Standard Trial Burn Objectives, 4. Mercury and Other Metals Emission Limits Here, it is stated that "the associated monthly, quarterly, or annual permit limitations would not, by themselves, be sufficient to prevent short-term, extremely high emission spikes. (emphasis added) Potential health concerns associated with these spikes are not necessarily addressed as part of the chronic risk assessments, nor would they be prevented by permit limits based on chronic risk assessments." This text is not consistent with the cited paper, Establishing Averaging Times for Metal Emissions from Hazardous Waste Incinerators. Consequently, it is misleading at best and if intentionally miscited, then it is also intellectually dishonest.

The following quote is taken from the Executive Summary of that cited paper, "According to...multiple of the trial burn emission rate necessary to reach acute exposure concentrations at the point of highest off-site concentration with the potential for adverse health effects" ... "emissions would have to be 100,000 to 100-million times higher than measured emission rates...." While emissions that were 100,000 times higher would certainly qualify as an "extremely high emission spike", it is also quite clear that it is impossible. In addition, the paper USEPA cites goes on to state that "long averaging times, on the order of months rather than days or hours, are recommended for metal emission standards for facilities combusting hazardous waste." Therefore, the paper that USEPA cites recommends the very "associated monthly, quarterly, or annual permit limitation" upon which EPA has chosen to cast doubt. In summary, the statements included in the draft document are not consistent with the published report/paper that is cited supposedly supporting the statements. In fact, the paper recommends almost the exact opposite. Reading just the Executive Summary of the paper cited makes it clear that the paper recommends long averaging periods and offers information casting strong doubt on the effects of any metals spikes that might occur during day-to-day operation. GCI strongly recommends that this section be modified to more accurately reflect the cited paper rather than promote unsubstantiated opinions.

Page 11 & 12, Example #1, 2 & 3

None of these examples involve cement kilns.

Page 14, 6. Products of Incomplete Combustion (PICs)

In this section the "conceptual framework for PIC testing" is introduced as the "reasonable worst-case approach," since USEPA does not yet know the difference between normal and worst-case PIC emissions. The "reasonable worst-case approach" is later defined on page 18 in the third complete paragraph. Because "specific correlations between individual operating/waste feed conditions and the type and quantity of PIC emissions have not been identified at this time" the Agency provides general guidelines for test conditions which include: 1) will result in the highest CO/THC emissions (worst-case combustion): 2) includes reasonable worst-case "real" waste feed streams; 3) represents minimum combustion temperature; and 4) represents reasonable worst-case conditions for remaining parameters listedin Attachment 1. It is of interest to note that "real" wastes are never defined. One is left to assume that this means an actual waste stream or representative waste stream received at the facility. There is some indication of what a "real" waste is expected to be on page 19 (IV.C.1) and in footnote 2/ of Attachment 1.

It is worth noting here that general guideline "1) will result in the highest CO/THC emissions (worst-case combustion):" once again demonstrating that EPA clings to an erroneous assumption, one that has been debunked in court. As noted in the Page 16, Section IV response, found subsequent to this response, following demonstration that CO and THC do not track together, the court found that "neither CO nor THC emissions are an accurate indicator of the combustion efficiency of wet kilns."

The guidance later discusses "'reasonable worst-case' methodology" on page 24. The guidance states that "the wastes and/or chemicals to be burned should be ranked based on the constituents in the various wastes that could significantly affect the risk assessment if trace amounts of those constituents went through the combustion process undestroyed." This guidance is provided because "every waste generally cannot be represented during the trial burn." No doubt, this guidance is provided to help make things easier, but consider the following:

On page 14, the guidance introduces "reasonable worst-case" approach (emphasis added) of which the guidance does not provide additional information until page 18.

On page 15, the guidance directs that PIC emissions testing should be conducted under "reasonable worst-case" conditions. (emphasis added)

On page 18, the guidance provides more information on the "reasonable worst-case" approach, but introduces reasonable worst-case "real" wastes, the quotation marks now being placed around the word, real, rather than the words, reasonable worst-case. This new quotation mark arrangement continues into page 19, but eventually refers to worst-case "real wastes" whereby the word, reasonable, has been dropped and the word, wastes, is now in quotation marks with the word, real.

By page 24, the guidance is discussing "reasonable worst-case" methodology (emphasis added).

While all of this may be clear to the authors of the guidance document, this commentor suggests that this entire scenario needs to be significantly reworked/reworded. Plus, whatever became of the term "representative" wastes used throughout currently existing and proposed regulations?

Page 13, last paragraph before 5. HCl and Cl2, second sentence

The statement that "these limits are required by the RCRA regulations anyway" is true in a general sense. However, there is no requirement for determination of the average feed rate "based on the last three years of operation" as suggested in the second paragraph of Example #3 on page 12.

Page 15, IV. Trial Burn Waste Feed and Operating Conditions, C. Selection of Waste Feeds for the Trial Burn, 1. Selection of Reasonable Worst-case "Real" Wastes, page 21

The top paragraph lists metals which are "thought to promote" PCDD/PCDF production reactions. A test study on an Ash Grove Cement, Foreman Arkansas cement kiln dated May of 1995 reports that "spiking the waste with copper had no impact on PCDD/PCDF formations."

Page 16, Section IV. Trial Burn Waste Feed and Operating Conditions, B. Introduction to Attachment 1, page 18

In the third complete paragraph, the following is stated: "1) will result in the highest CO/THC emissions (worst-case combustion)." It is clear that EPA continues to promote HC/CO as reliable combustion indicators even though this approach has been discredited. In CKRC versus USEPA, a US District Court 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) "allegedly demonstrated that emissions of THC are quantifiable and actually decrease when hazardous waste fuels are burned (in comparison to a non-hazardous waste burning baseline)." However, the Court held that re-analyzed data from this same testing "reveals that the inhomogeneity of the organic material in the raw materials precludes the establishment of a reliable baseline, and thus the baselines are not a valid indicator of good combustion."

Page 20 & 21, general dioxin theories

"Data from municipal waste combustion systems, medical waste incinerators, and cement kilns indicate that the majority of dioxins and furans emitted from these facilities are generally created in the relatively low temperature regions provided by particulate control devices." This is not true for cement kilns. Data has been submitted to the Agency which demonstrates that a certain level of PCDD/PCDFs are re-suspended from the processing of raw materials such as rock quarried from the ground. Comments not included here but recently provided in response to the 5-2-97 HWC NODA, from Continental Cement Company, LLC provides data demonstrating that dioxins are not always or entirely formed in the air pollution control device but rather may be input with the raw materials.

Page 21, top paragraph

Discussion about metals "which are thought to promote" PCDD/PCDF formation "include copper, iron, zinc, nickel, and aluminum" is only speculation. In fact, in a report entitled Emissions Testing of Ash Grove Cement Company Foreman, AR Waste-derived Fuel Facility Cement Kiln No. 3, obtained through FOIA, it was found that "spiking the waste with copper had no impact on PCDD/PCDF formations." The guidance document itself makes a case against including these metals on page 23 when it refers to related research as "inconclusive." In addition, iron and aluminum are major ingredients in the manufacture of cement clinker and there is no correlation between significant PCDD/PCDF emissions data from cement kilns and Fe/Al concentrations.

Page 22, a. DRE Testing, third paragraph

No data has ever been cited concerning the alleged impact of "batch feeds with high volatility/instantaneous oxygen demand" on DREs as stated in item 3. In fact, all of the data to date indicates that cement kilns that use any type of "batch feed" of hazardous waste have in fact passed on all aspects of DREs and PICs. Not one shred of data has been published since implementation of the BIF regulations which would indicate otherwise.

Page 23, b. Dioxin/Furan Testing, last sentence in top paragraph on page 23

The directive that "for industrial furnaces, the level of organics in raw material feeds should be as high as levels experienced during normal operations" implies that raw materials are being placed in the same category as hazardous waste. This statement is also very frustrating. First EPA arbitrarily picked 20 ppm THC as an alleged emission level indicating "good combustion." Then cement kilns that chose to burn hazardous waste were forced to find expensive substitutes for their raw materials in order to reduce their THC emissions resulting from volatilized raw material hydrocarbons to meet the 20 ppm THC limit even though it was recognized that these emissions were not combustion related. Now EPA guidance is "requiring" that hydrocarbons in the raw materials be monitored as part of a trial burn test. Raw materials are not hazardous wastes and EPA has no authority to treat them as such.

Page 23, second paragraph, last sentence

The statement that there is a need to insure that "potential dioxin inhibitors (i.e., sulfur, calcium, sodium and ammonia) are not biased high during the trial burn" is laughable for cement kilns. Recognize that the major component of cement clinker is calcium. This is a meaningless statement for cement kilns.

Page 25, top partial paragraph

This text seems to imply that a permit writer should determine chemical similarity, structure, halogens present, etc. of chemically similar POHCs. What about the incinerability index which has been used in every cement kiln DRE test burn on record? That is what the incinerability index was designed to do. Why recreate the wheel? The incinerability index needs to be referenced here as the primary source of this type of activity.

Page 27, 2. Variability of Batch-charged Waste Feed

The phenomenon discussed here may be true for some combustion devices but it has no bearing on cement kilns. There have been a number of overall analyses of cement kiln data generated since the testing portion of BIF went into effect in 1992 and every single DRE test of cement kilns that feed containers has demonstrated destruction efficiencies that exceed regulatory limits. This requirement is an unnecessary burden, to add to the many other testing burdens, for cement kilns that burn hazardous waste.

Page 27, 3. Feed Streams with a High Content of Chlorides/Halogens

It is worth noting here that there is no data to support the alleged impact of HCl/Cl2 on metals emissions. There is over four years worth of testing from cement kilns and analysis of this data has demonstrated that there is no direct effect between HCl/Cl2 and metals emissions. A paper presented at the A&WMA BIF Conference in March of 1995, The Effect of Process Differences on System Removal Efficiencies (SREs) and the Fate of Metals in Cement Kilns, indicates that there is no direct relationship.

Page 28, 4.a. Precursors

The third sentence in this paragraph seems to summarize good logic for not worrying about PCDD/PCDF precursors, "It is inconclusive at this time whether existing data shows a correlation between dioxin/furan precursors in the waste or fuel feeds and dioxin/furan emissions." This seems like a pretty clear case of "we don't know, so please do our research for us." Or worse yet, "Keep collecting data until we find a shred of evidence that we can hang a regulation on to."

Page 30, top paragraph

Where is the data to support the following statement? "High levels of chlorine in the hazardous waste feed coupled with high levels of organics in the industrial furnace raw materials could increase the levels of dioxins/furans and other PICs." (emphasis added) Extensive data analysis has demonstrated that this is not the case. Why keep perpetuating a concept that no longer has a sound scientific basis. The dioxin workshop held in November 1996 presented paper after paper of first-hand test data which destroyed the idea of chlorine input having any kind of a direct effect on PCDD/PCDF emissions. This has been particularly demonstrated in the widely publicized Greg Rigo ASME paper(1). In this paper, the authors concluded that "The failure to find simultaneous increases in most cases and finding inverse relationships in a few indicates that any effect chlorine has on PCDD/F emissions is smaller than the variability of other causative factors. Whatever effect chlorine has on PCDD/F emissions in commercial scale systems is masked by the effect of APCS (air pollution control systems), temperature, ash chemistry, combustion conditions, measurement imprecision, and localized flow stratification."

Page 30, 6. Maximum ash feed rate

While Attachment 1 addresses the issue of ash not being limited with regard to cement kilns and the regulation is also quite clear, that should be stressed here in this section as well, in order to avoid confusion.

Page 31, middle paragraph

This paragraph significantly varies from established and/or allowable regulatory practice. Take the first sentence, "Under current EPA policy, final metals feed rate limits will be established directly from the levels fed during the trial burn." What about "data in lieu of" from certification of compliance or recertification of compliance testing? And where did this "EPA policy" come from? It is in direct conflict with the regulations. Perhaps this was inadvertently overstated.

Now let's look at the second sentence, "System Removal Efficiencies (SREs) generally may not be used to extrapolate from the trial burn data to establish metals feed limits." This statement is incorrect. Extrapolation protocol is provided in the BIF Technical Implementation Document (Section 10.5 Extrapolation/Interpolation of Metals Emissions Data). In addition, a letter from Michael Shapiro, Director of the Office of Solid Waste, points out that "Generally, SRE decreases at lower feedrates." It seems clear that using such lower SRE values to calculate metals feedrate limits using EPA's conservative models would produce metals feedrate limits that are more conservative, that is more protective of human health and the environment. What is even more curious about this comment disallowing extrapolation is that the current NODA allows extrapolation. Section II. Compliance Requirements, C. Metals Extrapolation and Interpolation Considerations, which very clearly discusses "use [of] a statistical extrapolation methodology." [FR 62 24238]

And last but not least, if extrapolation were not allowed then metals spiking would be essentially required, especially if facilities are forced into worst case operating conditions as this trial burn guidance document would seem to indicate. The currently proposed Revised Standards for Hazardous Waste Combustors describes "the common process of spiking metals during compliance testing to ensure an adequate operating envelope [as] expensive, potentially dangerous to the testing crew that must handle the toxic metals, and causes higher than normal emission rates during compliance testing." (FR 61 17429) Why would this latest guidance take a different approach from what is currently proposed as regulation?

Page 32, D. Selection of Operating Conditions for the Trial Burn, 1. Minimum combustion temperature and residence time

Minimum combustion chamber temperature may apply to hazardous waste incinerators but it is not particularly applicable to cement kilns. Temperatures hot enough to produce cement clinker (2600F) are inherent in the cement manufacturing process/system and residence time in the burn zone is several seconds. Heat and residence time sufficient to destroy hazardous waste would be present in a cement kiln for several minutes even if you shut the kiln down on the spot.

Page 33, top paragraph

Here is another case of pure speculation. There is no cement kiln test data that even remotely indicates "inadequate mixing in the combustion chamber." This might be relevant for a hazardous waste incinerator. However, discussion of a "need to establish a minimum production rate" clearly puts the onus on cement kilns since hazardous waste incinerators produce nothing. Where is this coming from? What data supports your contention that "inadequate mixing is normally indicated by increased CO/HC emission rates"? This is pure speculation without one shred of data to support such a contention concerning cement kilns. This does not apply to cement kilns and should be noted as such in the guidance.

Page 33, 2. Amount and Distribution of Combustion Air - All Combustors: Minimum Oxygen Limit or Improved Air-to-Fuel Ratio

The guidance "encouragement" to provide controls which would maintain fuel/air ratios and thereby minimum excess oxygen levels is undoubtedly well intentioned. It may be warranted for certain types of incinerators, but this guidance, by not differentiating cement kiln specific issues, again illustrates EPA's apparent lack of understanding of the cement manufacturing process. The cement manufacturing process, in order to manufacture quality clinker, must maintain an oxidizing environment in the combustion zone. In order to accomplish this objective, cement kilns have for decades monitored and controlled excess oxygen levels at the kiln exit. Because of the oxidizing process inherent in the mineral processing in a cement kiln, as well as the fuel combustion oxidation and extremely high temperatures, the establishment of limits is not so straight-forward as EPA seems to suggest in this guidance. Optimum excess oxygen levels are not only different kiln to kiln, but can vary for a specific kiln over time based on a wide variety of factors. Given the benefits to product quality and fuel efficiency, had this been a straight-forward parameter for setting automated limits, the industry would have done so years ago. Instead, most plants continue to rely on trained, experienced operators to control kiln excess oxygen and numerous other control parameters for optimum combustion efficiency. EPA, in this guidance, should recognize this critical difference and the inherent controls built into cement kiln technology.

Page 34, 3. Amount and Distribution of Combustion Air - batch feed systems

Once again, container feeding to cement kilns has been included in numerous tests since 1992. This data has been analyzed, re-analyzed, probably over-analyzed and in no case did any kiln, container feeding or otherwise, fail a DRE test. While this may apply to hazardous waste incinerators, it does not apply to cement kilns. This is an unnecessary testing and subsequent ongoing operational burden for hazardous waste burning cement kilns.

Page 35, 5. Maximum Combustion Temperature

Extremely high temperatures are always present in a cement kiln. Because of these extremely high temperatures, necessary to produce cement clinker, almost all the metals vaporize whether they are considered volatile metals or not. It is the cement clinker chemistry and counter-current materials/gas flow that affects where the metals in the cement manufacturing process will end up. Maximum combustion temperature is an unnecessary control parameter in a cement kiln. It would seem more logical to set an exit temperature limit but the data does not support such a control limit either.

Page 36, top of the page

It is worth noting here that data presented at the 1996 Air & Waste Management Association BIF conference indicated that SREs were very similar even at APCD maximum inlet temperature vs. normal inlet temperature. The paper, entitled A Comparison of Normal and Worst Case Cement Plant Emissions, compares metals emissions from a worst case day and a normal day and the emissions are very similar.

Page 40, top paragraph

Excellent point! The guidance presented in this paragraph (V. Application of Trial Burn Data in the Risk Assessment, A. Emission levels to be evaluated in the high-end risk assessment, second full paragraph - excluding the four bullets) varies from existing regulation but is highly logical. Permit emission limits should be based upon the highest allowable rate as determined by the risk assessment. This makes complete and total sense. Unfortunately, the BIF regulation requires that the emission limits be based upon the trial burn. This is in definite conflict, but in this case the BIF regulations need to be modified.

The EPA has chosen the value of 0.02 ng/dscm TEQ PCDD/PCDF as a maximum emission level that is still protective of the environment. There are a variety of methods for calculation of these TEQ values. As an example, the EPA used a different set of toxic equivalency factors (TEFs) in a 1981 document than in the 1989 version accepted today. Also, Ontario (Canada) and New York State as well as the USFDA, the Swiss, and Britain uses a different method of TEQ calculation. It is possible that the EPA will change the TEF values in the future based on continuing research. Consequently, this Trial Burn Guidance document should clearly state which version of the TEF and method of calculation of the TEQs should be utilized. Indeed, this is nearly as important as specifying the emissions test method. It is currently assumed that the EPA approved method of calculating the TEQ values is that detailed in 40CFR266 Appendix IX, section 4. If such is the case, they should be confirmed in the guidance documents. This is especially important as this document will be used by permit writers to establish permit limits and to document adequate risk assessments for other RCRA regulated facilities that may not be subject to 40CFR266.

Page 43, C. Detection limits for use in the Risk Assessment, final one line paragraph

"If the permit writer is setting an emission limit on the compound of concern, then they must use the full detection limit in order to ensure enforceability." Using the full detection limit, in setting an emission limit on a compound of concern, is the only way to ensure enforceability. This is an imminently logical point and it is good to see it in some EPA guidance.

Pages 43 - 45, Trial Burn Sampling and Analysis, A. Process Sampling/Analysis

There is no justification for the inclusion of so many metals for which to quantify input rates for cement kilns, and this should be noted. The ten BIF metals are required by the BIF regulations but the other metals included here (Ni, Se, Cu, Fe, Al, and Zn) are not necessary as noted previously in these comments. This would also hold true for so many of the other parameters listed in this section. What is the justification for total organic carbon? Or carbon, hydrogen & nitrogen? These have absolutely no bearing on anything in the cement kiln. The one thing that should have been mentioned but did not get mentioned was a mass balance for metals and chlorine. A mass balance for metals and chlorine, for the purpose of determining the regulatorily required fate of these elements, is far more telling/revealing than requiring obtuse parameters that have no bearing on the process.

The compounds listed at the top of page 45 are a good reference list. But it should be noted that phthalates are frequently found in background samples. USEPA should provide some logical guidance on how to address commonly found background compounds that are listed here. Perhaps some sort of subtraction of background levels. USEPA should address this in the guidance.

Page 45, 2. Metals

Review of PCDD/PCDF research data indicates that the metals listed here have no bearing on PCDD/PCDF emissions from cement kilns. All of these metals are present in fairly high concentrations in the cement kiln process anyway as previously addressed in these comments.

Page 45, 3. HCl/Cl2

This guidance section is ever so brief in stating that "risk assessment data will need to include stack sampling and analysis for HCl/Cl2 compounds" and then that the "sampling train (Method 0050) may be utilized for both HCl/Cl2 and particulate emissions determinations." The guidance fails to point out that the proposed HWC regulations point out that "However, the test method used to determine HCl and Cl2 emissions (BIF methods 0050, 051, and 9057, commonly referred to as Method 26A) may not be able to distinguish between HCl and Cl2 in all situations." (FR 61 17376) Similar comments have been provided in response to the 5-2-97 HWC NODA. In addition to EPA's own admittance that 0050 will not provide reliable data there are a number of other related items which have pertinent bearing on lack of reliability of Method 0050.

A meeting occurred in late 1989 or early 1990, related to municipal waste combustor (MWC) regulation, between a cement company representative, their 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."

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 suggest 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 showing higher HCl results than 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 & Dithrich - 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 some 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.

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 cations 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 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 who point out that the data is not valid. It seems all too clear that EPA reliance upon this method can only produce invalid data. The suspect nature of any data produced through use of Method 0050 should be discussed in this guidance. Ideally, the method should not be recommended.

Page 46, 5. Total Organics

While the guidance provided here may be valid in some sense, since the goal is to determine the total organics, why not simply determine total hydrocarbons (THC) with a flame ionization detector (FID)? These systems are already present on hazardous waste burning cement kilns and are readily available from stack testing companies.

It is confusing as to exactly what is desired here. In item 5 there is talk of total organics yet in item 6 there is discussion about CEMs and hydrocarbons. This would appear to be somewhat redundant. Please clarify this in the guidance. It would appear that item 6 could be substituted for item 5, and item 5 could be completely eliminated.

The final comment on this issue again addresses redundancy. Cement kilns have repeatedly characterized their hydrocarbons. Much data exists. Both from normal testing and research oriented data. It seems rather vague and obtuse to continue to require that hydrocarbons repeatedly be characterized. What is the objective here?

1. 1 Rigo, G.H.; Chandler, J.A.; Lanier, S.W. (1995) The relationship between chlorine and dioxin emissions from waste combustors. In: Solid Waste Management: Thermal Treatment & Waste-to-Energy Technologies, Kilgroe, J., ed. Proceedings of an international specialty conference sponsored by the Air and Waste Management Association, Washington, D.C. April 18-21, 1995, pp 433-438.