Volume 4, Number 11
A Gossman Consulting, Inc. Publication
November 1998

Review and comment of EPA530-D-98-002Guidance on Collection of Emissions Data to Support Site-SpecificRisk Assessments at Hazardous Waste Combustion Facilities

Jim Woodford and Dave Constans

The EPA has written several Trial Burn Guidance documents. Few of them have made it down the long road to become the "official" guidance document. Some have been superseded while still in early draft forms, some were technically obsolete before being issued. This is another such effort which has proven to be information rich and technically superior to any other trial burn related guidance of which GCI is aware. EPA is to be congratulated on this effort to date. EPA had requested public comment on this document with the comment period closing on November 16, 1998. The following are GCI's comments on this important document.

General comments
This document entitled, Guidance on Collection of Emissions Data to Support Site-specific Risk Assessments at Hazardous Waste Combustion Facilities (PEER REVIEW DRAFT) cites numerous EPA documents including Components 1-5 of the voluminous Hazardous Waste Combustion Unit Permitting Manual out of EPA Region 6. A simple cursory review of Component 1 entitled, "How to Review a Trial Burn Plan" reveals that Component 1 is actually a review of the U.S. EPA Region 6 Generic Trial Burn Plan which has been written "for a new incinerator train that consists of a rotary kiln incinerator, a secondary combustion chamber (SCC), a boiler for energy recovery, and a wet air pollution control system (APCS)." This clearly does not have direct bearing on cement kilns.

Specific comments

The next to last paragraph in this section states that "EPA Region 6 has also developed additional guidance for reviewing trial burn plans, RCRA Part B permit applications, and documents such as quality assurance project plans." While document EPA530-D-98-002 is now going through public comment, it should be noted that the Region 6 guidance documents (approximately 2,300 pages worth) have not been made available for public comment and should absolutely not carry the weight of a publicly reviewed and commented upon document.

Review of RCRA Trial Burn Provisions
The second paragraph on page 5 incorrectly refers to 12 BIF metals. There are only ten BIF metals. Two metals, selenium and nickel, were specifically excluded from the BIF regulations. The reasons for this were given in the BIF Preamble, FR 56, page 7171, footnote 54. The EPA has not provided any information to refute this reasoning.

Human Health and Ecological Concern for Specific Metals
While there may be some level of general concern of the metals discussed in this section, no evidence has been presented that they are a concern from combustion sources. As pointed out in previous comments, adding nickel is in direct conflict with footnote #54 in the BIF preamble, as noted above, which discounts the possibility of a combustion device emitting carcinogenic Nickel compounds. Plus, the discussions about Nickel on page 13 are of refinery dust from reducing conditions which in no way relate or compare to the oxidizing conditions found in HWCs.

Defining average or normal operating conditions
1. Can the facility provide sufficient information to define normal feed and operating conditions?

In answer to this question, this section cites "Hazardous Waste Combustion Unit Permitting Manual. Component 1 - How to Review a Trial Burn Plan." Center for Combustion Science and Engineering, Multi Media Planning Division, EPA Region 6. December (as EPA 1997f). The manual states, "that data collection for SSRAs under normal operating conditions will only be considered if the facility burns wastes that have little temporal variation in chemical and physical properties, at nearly constant rates, and under operating conditions that do not fluctuate widely."

Unfortunately, in making this citation EPA has "made official" a document that has not been placed into the public record and subjected to public comment. This document came to GCI's attention during our involvement with Texas Industries' permit hearing where it was repeatedly referenced, even though it had been written after the TXI trial burn had been completed. The following is a review of how the above referenced document is being applied to this Trial Burn Guidance.

Review of Component 1 has only turned up part of this cited statement and that was in Attachment A which is a generic trial burn "for a new incinerator train that consists of a rotary kiln incinerator, a secondary combustion chamber (SCC), a boiler for energy recovery, and a wet air pollution control system (APCS)." Widely fluctuating operating conditions may be a concern for incinerators, but the overall stability of the cement kiln manufacturing system/process requires "operating conditions that do not fluctuate widely."

Clearly, in the end it is up to the trial burn plan reviewer/permit writer to make a decision. Stating that waste input must be "at nearly constant rates" and calling for "operating conditions that do not fluctuate widely" unnecessarily promotes and invites a very conservative interpretation. Temporal variations in chemical and physical parameters of hazardous waste fuels in cement kilns have no significant impact on emissions.

There is a related statement elsewhere in Component 1 ( Operating Conditions) which points out that "it is not anticipated that commercial hazardous waste TSDFs will be eligible" ["to conduct a risk burn at normal operating conditions"] which in turn references Component 3 entitled, How to Review a Part B Permit Application. However, review of Component 3 did not reveal clarification of this issue. Under Section 5.1.5, Waste in Incinerators, there is a statement that actually appears contradictory, "Risk burn test typically requires worst-case waste streams operating under normal conditions in order to measure PIC formation." No similar or even related comment is found under Section 5.1.8, Waste Boilers and Industrial Furnaces.

Since the guidance cites Component 3, it is also relevant to point out that Section 6.5.3, Data Submitted in Lieu of Trial Burn, attempts to narrow BIF regulatory language almost to the point of exclusion by stating that "To be classified as similar, the unit should use the same source of hazardous waste feed, have identical dimensions, and have similar operating controls and comparable(CEMS). All maintenance should also be identical" This criteria is impossible to meet, but more importantly, is not part of existing regulation. It also seems quite curious that guidance involving trial burns and risk assessments would not point out that the Director has a regulatory obligation to review data submitted in lieu of a trial burn. Then, in accordance with §270.22(a)(6) "The Director shall (emphasis added) approve a permit application without a trial burn if he finds that hazardous wastes are sufficiently similar, the devices are sufficiently similar, and the data from the compliance tests, trial burns or operational burns are adequate to specify (under 266.102 of this chapter) operating conditions that will ensure conformance with 266.102(c) of this chapter."

Component 3, Section 6.9, Reviewing Information Regarding Boiler and Industrial Furnace Units, incorrectly ignores the preamble to SW-846 by stating that "The description of hazardous constituents must be based on an analysis conducted according to techniques specified in 'Test Methods for Evaluating /Solid Waste, Physical/Chemical Methods,' U.S. EPA Publications SW-846, or equivalent methods." The preamble to SW-846 states quite clearly that "except for those situations where the RCRA regulations specify use of a particular method, it is appropriate for the chemist to use judgment, tempered by experience, in selecting an appropriate set of methods from SW-846 or the scientific literature (emphasis added) for preparing and analyzing a given sample." The preamble also contains provisions for a chemist to modify SW-846 methods. In addition, there are no SW-846 methods for a number of relevant parameters, such as heat content.

GCI supports the basic example logic presented on Figure 3-1, entitled Site Specific Risk Assessment Data Collection Flow Chart. The left side of the chart addresses normal operating condition testing and concludes that "long-term average limits" should be established. The right side of the chart addresses worst-case testing and concludes that "no long-term average permit limits [are] needed."

Other Factors
The point is made that bromine affects dioxins but what type of affect?

Operating Parameters Associated with D/F Production
Discussion under Secondary operating parameters suggests that permit writers have the authority to limit the organics in raw material input to cement kilns. EPA has no authority to require such a limitation under RCRA since raw materials are not a hazardous waste and resulting emissions would not be coming from hazardous waste. Neither does EPA have authority to require such a limitation under the Clean Air Act, according to the Conference Report filed regarding the Clean Air Act amendments (House Conference Report No. 101-952), whereby the conference committee wrote on page 339: "For categories and subcategories of sources of hazardous air pollutants engaged in mining, extraction, beneficiation, and processing of nonferrous ores, concentrates, minerals, metals, and related in-process material, the Administrator shall not consider the substitution of, or other changes in, metal or mineral bearing raw materials processed or derived form such feedstocks, or materials in setting emission standard, work practice standards, operating standard or other prohibitions or requirements of limitations under this section for such categories or subcategories."

Consequently, it seems quite clear that EPA does not have the authority to regulate raw materials under either RCRA or the Clean Air Act.

(As an aside, Greg Rigo and James J. Cudahy have written a paper for the A&WMA Journal entitled National Annual Dioxin Emissions Estimate for Hazardous Waste Incinerators wherein it is stated "Estimate ITEQ dioxin emissions from HWIs are most likely 15 g/yr rather than the 79 g/yr the EPA estimated as part of their HWC rule development effort." See A&WMA's Journal web site at http://www.awma.org/journal/)

Discussion under tertiary operating parameters also suggests that there is some possibility that dioxin precursors could be limited even though they "have not routinely been correlated with D/F emissions during full scale testing." It is inappropriate to limit parameters based upon speculation.

D/F Emissions from Cement Kilns
GCI agrees with the comment in line six of the first paragraph whereby it is stated that, "because of the chemical composition of the raw materials, carbon monoxide and total hydrocarbon concentrations may not always serve as indicators of good combustion." GCI congratulates EPA for agreeing with an issue studied in detail by GCI and others and concluded in a presentation published in conjunction with the 1993 A&WMA BIF Conference. The paper was entitled, Suitability of Hydrocarbon and Carbon Monoxide Measurement as Combustion Indicators in Cement Kilns.

The second paragraph states that "data presented by Harris and others (1994) and Lanier and others (1996) demonstrate that D/F emissions from cement kilns increase exponentially with increases in inlet temperatures to the dry APCD while within the D/F formation window (400 to 750o F.)" GCI, Inc. does not deny that temperature plays some sort of role with respect to dioxin emissions but it has previously been brought to EPA attention in HWC comments that data from Volume III of the Technical Support Document for the HWC MACT Standards contains 19 data points related to D/F emissions in relation to APCD temperature which do not support the exponential increase claim here in this guidance document. That the exponential increase claim is included in this guidance document is even more incredible given that Volume III is cited in this document as EPA 1996c. [See this GCI Tech Note]

In the final sentence of the second paragraph in this section it is stated that "maximum APCD inlet temperatures...may coincide with the SRE test...." GCI would point out that SREs require maximum APCD inlet temperatures rather than "may coincide."

Table 4-2 mentions 10-minute averages but 10-minute averages are not in accordance with any applicable regulation at this time. A regulatory basis for this guidance is needed. If omnibus authority is the basis then EPA must provide data which demonstrates this need.

GCI requests that the specific reference to the DRE failure attributed to poor atomizer design be designated. In addition, GCI is unaware of any DRE failures in cement kilns related to BIF testing.

On the last page of Section 4.5, the second to last sentence in the top paragraph, states that "Carbon monoxide may not always be a good indicator of organic emissions from cement kilns." Given that EPA is now in agreement with a widely accepted issue within the cement industry, it now seems appropriate to change the regulations to match this conclusion.

5.1 Organic Emissions from HWIs and Boilers
EPA suggests that certain operating conditions may be required for worst case PIC emissions from Hazardous Waste Combustors. While this may be true for some devices, it is not true for cement kilns. This has been demonstrated in a number of tests conducted since 1992 comparing organic PIC emissions measured under worst case HWF and coal only conditions. There were no significant differences in emissions. EPA has acknowledged that PICs in cement kilns arise from organics in the kiln feed operating conditions, including the use of HWF does not impact those emissions. Data supporting this can be found in the following papers.

An Evaluation of a Cement Kiln's Emissions While Under Worst Case Operating Conditions (Dec., 1997, Rock Products Conference) Available here

A Comparison of Normal and Worst Case Cement Plant Emissions by Jim Woodford, David Gossman, Norris Johnson. This paper was presented at the 1996 A&WMA BIF Conference.

Organic Emissions from Cement Kilns and LWAKs The discussion on page 56 suggests "an alternate ranking scheme ...to reflect the most prevalent waste codes and toxic, persistent, and bioaccumulative constituents in the blended fuel." The use of waste codes to indicate waste constituents cannot and will not work. The derived from rule and tendency for generators to use multiple and large numbers of waste codes makes this suggestion nonsensical.

Metal Emissions
The first sentence in this section is misleading. There is a definite difference in the character of the metals emissions from cement kilns versus other types of HWC's. This is documented in a paper presented at the March 1993 A&WMA BIF Conference entitled Comparison of Metal Emissions from Cement Kilns Utilizing Hazardous Waste Fuels with Commercial Waste Incinerators. This paper demonstrates that incinerators consistently emitted higher levels of toxic metals per dry standard cubic meter of stack gas. This paper is available for viewing here.

The next to last line in the first paragraph states that "Nickel and selenium were later regulated through omnibus authority of RCRA (EPA 1992a)." GCI is not familiar with the Shiva Garg memo cited as EPA 1992a. In addition, facilities faced with trial burn plan negotiations will sometimes agree to parameter testing that may not make sense but preserves good faith efforts in getting the trial burn accomplished. Now that this logic appears in print in this guidance document (also at the bottom of page 4), it seems appropriate to prove the need for omnibus authority for selenium and nickel. What data demonstrates that there is a risk to human health thus justifying omnibus authority? Development of RACs is not sufficient justification. Please demonstrate the need for omnibus authority in regulating nickel and selenium.

GCI wishes to support EPA in their recognition of allowing "sufficient operating flexibility" for a regulated facility as noted in the first sentence of the fourth paragraph of Section 6.0.

The second to last paragraph on page 58 refers to nickel as being "considered as a potential carcinogen." However, GCI would point out that this is only in its reduced form.

Metal Speciation
There is metal speciation data for chromium & nickel based upon leachability and solubility in CKD. GCI questions why that data is not included or at least referenced in this guidance document.

Section 6.2 (Second to last paragraph) EPA correctly notes that cement kilns do not exhibit the same volatility profile for metals as other HWC systems. There is considerable literature on this issue available, including the following:

EPA should reconsider the statement made in the first sentence of section 6.0 in light of this data. Guidance more specifically tailored to cement kiln data is needed to avoid confusion over this issue.

Volatile Metals (Mercury and Selenium)
The first bullet item on page 64 states that "Permit limits should be established for total maximum mercury feedrate (including hazardous waste, raw materials, and fossil fuels)." All cement kilns already monitor and limit metals input to the kiln from all feedstreams as a total metals input limit. However current regulations do not specifically limit metals content of the raw material or the fossil fuels fed to the kiln. GCI supports this approach as consistent with the BIF regulations but wishes to insure the clarity of this guidance by strongly reiterating that EPA does not have the authority to impose specific permit limits on raw materials or fossil fuels, under RCRA or the Clean Air Act, concerning metals, for the same reason stated previously in Section 4.2 comments.

There are a number of issues in the next to last paragraph on Section 6.1.1 on page 65. In the third sentence, it is stated that additional mercury removal can be achieved through "treatment with carbon injection or carbon beds." Neither of these techniques have been nor are currently practiced in cement kilns. This statement needs to be clarified as speculative or if testing has been performed then it should be cited. The very next sentence claims that mercury can "partition to the clinker product." This is highly unlikely given the extreme heat of the clinker and the volatility of mercury. GCI is unaware of such partitioning. Please cite data for this statement. The final sentence of the paragraph notes that SREs range "from zero to 90 percent," citing the April 19, 1996 proposed HWC rule. GCI is aware of at least one test situation in Texas where SREs exceeded 90% and this data is available in the EPA database. This statement needs to be modified.

Semivolatile Metals
The bulleted items listed here may be appropriate for incinerators but they are not relevant for cement kilns. It would appear that the cement kiln literature has been largely ignored in this regard. The first bullet in particular, the increase of metal volatility by chlorine, has been demonstrated not to be the case. See citations in comments at the beginning of this section.

Operating Conditions and Parameters for Metal
The first bullet on page 67 lists "maximum combustion temperatures" as a traditional operating parameter related to high temperature testing of BIF metals. This is certainly a correct statement. However, GCI would like to take this opportunity to point out that the proposed HWC rule (EPA 1996a) does not include a maximum combustion chamber temperature limit for metals [FR 61 17429; Part Five Implementation, II. Selection of Proposed Monitoring Requirements, C. Compliance Monitoring Requirements, 4. Semivolatile Metals (SVM) AND Low Volatile Metals (LVM) & §63.1210 @ FR 61 17520] and we support the elimination of such a limit. GCI is aware that supporting logic was previously submitted with comments on the proposed HWC MACT.

Measuring Particle-Size Distribution

It is clear that EPA desires regulated facilities to determine a particle size distribution which will then be factored into the risk assessment, however, GCI questions how this can be justified when EPA can't seem to demonstrate a valid method?

Permit Limits for Key Operating Parameters
Short term limits "(such as instantaneous or 10-minute rolling averages)" are suggested as possibly "appropriate in situations where short-term perturbations outside of a certain operating range could result in high emissions rates that cannot be offset by lower emission rates during periods of normal operations." GCI is unaware of any situation regarding a cement kiln where this is even a possibility. This appears to be speculation that is not based upon reality as no specific instances are cited as examples.

The repetition of this short term limit concept, as proposed in the April 19, 1996 HWC MACT rule does not reflect the numerous negative public comments on this issue.

Application of Trial Burn Data in the SSRA
GCI wishes to express strong support for the "reasonable maximum exposure" (RME) discussion in the final paragraph of this section on page 77.

Appendix A Trial Burn Conditions and Permit Limits for an Example HWC Facility
It is clear from just reading the very first sentence of this Appendix that the title needs to be changed since the example is for a "hazardous waste incinerator (HWI)" facility which has little or no bearing on cement plants. This example should not muddy the water for cement plant permit writers.

Appendix B Sampling and Analysis
GCI points out that the phrase "highly variable waste" is used throughout this appendix but no definition is offered. Liquid waste fuel burning cement kilns burn a blended waste fuel that is consistent and should not be considered "highly variable" as appears to be suggested as always the case at commercial facilities.

B.6 Chlorobenzenes/chlorophenols
The "higher" of the two results (due to dual analysis of constituents, SVOC and CB/CP) should not be used as stated in the next to last sentence of this paragraph. It is the average that should be used. Anything different is questionable science.

B.7 Polycyclic Aromatic Hydrocarbons
The "higher" of the two results (due to dual analysis of constituents, SVOC and PAHs) should not be used as stated in the third sentence of the first paragraph. It is the average that should be used. Anything different is questionable science.

B.10 Aldehydes/ketones
GCI wishes to express general support of this section. Good work!

B.12 Total Organics
The recommendations in this section are surprising given that EPA funded a project by EER that essentially found no new risks. Why would this be required for each trial burn if no new risks were revealed?

Gravimetric Mass Fraction (boiling points > 300o C)
Discussion on page B-40 readily admits that the "GRAV fraction may include organic and/or inorganic mass not directly attributable to organic incinerator emissions. These artifacts may be composed of inorganic salts, super-fine particulate ... or some other unknown (Lamiuex and Ryan 1998)."

Such is virtually certain to be the case for cement kilns. It is known that cement kilns emit salts, super-fine particulates and organic materials from the kiln raw feed. Due to the variability of the sources of the raw feed materials (be it different locations within the quarry for the limestone, clay or shale, or perhaps even different quarries for each of the raw feed ingredients) the GRAV mass values may be grossly different from test to test, irrespective of whether or not hazardous waste fuel was utilized.

This analysis is highly speculative. Indeed this is acknowledged in comments made in the document "The TO measurement is an estimate." "...the estimated mass is calculated based on detector response to specific compounds that are presumed (emphasis added) to be representatives of the specific function." ... "measurements are strongly believed to be" ... "for the purpose of indicating uncertainty...." There are more such carefully phrased statements that lead one to believe that the analysis is highly subjective and open to the interpretation of the analyst.

The document continues "The GRAV portion is extremely difficult to analyze. However, it would be expected (emphasis added) to contain high molecular weight organics of C17 or greater, D/Fs, many PAHs and high molecular weight organic

acids and salts." Clearly EPA would have a prejudicial expectation if higher than previously experienced Gravimetric Mass Fractions were determined for a cement kiln. This expectation would be difficult to refute due to the GRAV portion being "extremely difficult to analyze" particularly in the cement kiln process matrix which would include inorganic salts, super-fine particulates and a variety of unknowns attributable to the raw feed.

Clearly, until the analytical portion of the GRAV testing can be refined to provide specific speciation of the organic constituents, the use of the GRAV test on cement kiln stack emissions should be set aside.

B.16 Hydrogen Chloride and Chlorine
GCI points out here that none of these methods have ever been validated on cement kilns. In addition, while the correction for NH4+ is recommended, other cations should also be tested for quality control purposes. Gossman Consulting Inc. had made this recommendation in early Method 26 discussions when EPA stated that this method should not be used on cement kilns because of interferences.

B.17 Process Samples
No justification is provided for requiring proximate analysis.