Coalition for Responsible Waste Incineration

Ms. Beth Antley
USEPA, Region 4
980 College Station Road
Athens, GA 30605

Dear Ms. Antley:

The Coalition for Responsible Waste Incineration (CRWI) is pleased to submit comments on the Guidance on Collection of Emissions Data to Support Site-Specific Risk Assessments at Hazardous Waste Combustion Facilities (EPA530-D-98-002, August 1998). CRWI represents eleven companies with either captive or commercial hazardous waste incineration interests. These companies account for a significant portion of the U.S. capacity for hazardous waste incineration. In addition, CRWI is advised by a number of academic members with research interests in hazardous waste incineration. Since its inception, CRWI has encouraged its members to reduce the generation of hazardous waste. However, for certain hazardous waste streams, CRWI believes that incineration is a safe and effective method of treatment, reducing both the volume and toxicity of the waste treated. CRWI seeks to help its member companies both to improve their incineration operations and to provide lawmakers and regulators helpful data and comments.

CRWI believes that conducting a trial burn is one of the tools used to confirm that a hazardous waste combustor meets current standards. CRWI also believes that realistic risk assessments can be used as a tool to reassure employees, the regulating agencies, and the public that their hazardous waste operation is safe and protective. However, CRWI does not believe that risk assessments should be used to set individual permit limits that are more stringent than the current standards except under unusual conditions. These standards have already been shown to be protective of human health and the environment. The only circumstance where CRWI can envision using risk assessments to further restrict permit conditions is where a realistic risk assessment shows an unacceptable risk and an omnibus determination can clearly be documented. CRWI does not believe that risk assessments should be routinely used as permitting tools unless the risk assessment protocols are subjected to the rigors of the rulemaking process. To that end, CRWI is offering a series of comments on the draft trial burn guidance that we believe will make it a more practical tool. These are addressed in the following paragraphs and in Attachment A.

However, before discussing our comments, CRWI believes that a 30 day comment period was not sufficient to properly review the trial burn guidance document. In addition, this document often refers to the Human Health Risk Assessment Protocol for Hazardous Waste Combustion Facilities. While the two documents are closely related and should be reviewed together, the risk assessment document is more than 2000 pages. Thirty days is simply too short a time frame to review and comment on the trial burn guidance document itself, much less to develop comments on both documents. CRWI will continue to examine these documents, and as issues arise, will bring them to the attention of the Agency.

Permitting issues

CRWI is concerned with the additional permit restrictions resulting from the incorporation of a risk burn. Hazardous waste combustion units are required to conduct trial burn tests in order to demonstrate compliance with the current requirements of RCRA. Operating limits (feed limits, process parameter limits, continuous emission monitoring, etc.) are established based on the trial burn using existing guidance for permit writers. The existing guidance was developed to ensure compliance with the existing regulations under worst case conditions. The Agency has determined that these regulations are protective of human health and the environment. Recent guidance, including several documents under comment, has been developed by EPA in order to evaluate the theoretical risk from these facilities when operating in compliance with these regulations. The guideline for acceptable risk has been established at 1 X 10^-5. Therefore, CRWI believes that additional operating limits for risk control (i.e. those described in the draft trial burn guidance document) should only be required where emissions are shown to present an unacceptable risk from a realistic risk assessment. However, the draft guidance is structured such that operating limits (e.g., Section 4.2 and 6.3) would always be established to control emissions of certain compounds (D/F, non-dioxin PICs, metals, etc.) at levels demonstrated by a risk burn. This would impose a number of unnecessary restrictions in several combustor operations that already meet current standards as demonstrated by their SRE and DRE tests. In the event that a realistic risk assessment showed unacceptable risks, then revised permit limits or other mechanisms to reduce risk are appropriate.

Until the Hazardous Waste Combustor MACT standards are finalized and operators are required to demonstrate compliance with these standard, D/F emissions should only be required to be at levels of acceptable risk. Additional operating limits on batch feeding, kiln temperature, and THC are not necessary. Therefore, CRWI believes that additional operating conditions should only be required in the event that a realistic risk assessment shows an unacceptable risk.

Metals

The draft trial burn guidance document does not appear to allow upward extrapolation of metals from what has been demonstrated during either the SRE test or the risk burn (if required). In the May 2, 1997, Revised Technical Standards for Hazardous Waste Combustion Facilities--Notice of Data Availability and Request for Comments, EPA proposed to allow upward metals extrapolation for determining metals feed rates. This allowed facilities to avoid the expense of metals spiking and reduced the overall emissions of spiked metals to the environment. Then and now, CRWI supports the idea of metals extrapolation. Our comments submitted on June 17, 1997, stated:

Although most facilities normally burn low amounts of metals, these facilities need the flexibility to burn occasional waste streams that contain higher levels of regulated metals. Historically, facilities have been forced to feed high amounts of these metals during testing in order to receive the higher feed limits in their permits. It is likely that the actual stack test emitted many times more emissions than the normal day-to-day operation. This is not to say that all spiking should be discontinued. However, the use of extrapolation would greatly reduce the amount of metals unnecessarily released to the environment. In addition, working with less metal is obviously much safer for the personnel trying to prepare for this burn.

CRWI suggests that the trial burn guidance should recognize that upward extrapolation is appropriate and allow this tool to be used to minimize the amount of metal spiking needed in either the SRE or risk burns. However, should a facility want higher metal feed limits than would be allowed by scaling up, the facility and the permit writer should be allowed to set metal spiking at appropriate levels during the trial burn.

CRWI is also concerned about the continued expansion of the number of metals for which testing is required. The regulations for hazardous waste incinerators (Part 264, Subpart O) do not contain any requirements for metal feed rate limits. The BIF regulations (Part 266.106) require testing for 10 metals. Two additional metals were added to the list through a memorandum from Shiva Garg to the ten EPA regions. Now, this guidance document adds an additional six metals to the list of compounds of potential concern. The requirement to test for eight of these metals has been developed through either memorandum or guidance. In addition, the application of the 10 BIF metals to hazardous waste incinerators has been accomplished through omnibus authority, not formal rulemaking. In fact, when the currently proposed hazardous waste combustor MACT rule becomes final, hazardous waste incinerators will be regulated for only six metals. If the Agency believed that the other metals should have been controlled, CRWI suspects that they would have been included in this regulation. The continued addition of metals and other testing requirements through guidance rather than the formal rulemaking process is very disturbing and should be stopped. If the Agency wants to add compounds for testing where permit limits are to be established, this should be done in a formal rulemaking process rather than through memorandums, guidance, or omnibus authority. If EPA has sufficient information to show that any compound should be regulated, this must to be accomplished via the formal rulemaking process. If such information is lacking, EPA should initiate research projects to determine if a compound should be included. EPA should not use omnibus authority to force facilities to run research projects for the Agency to determine which compounds are important.

In addition to these concerns, one of the metals (cobalt, page 12) suggested in this guidance to be included has no toxicological data for either human or ecological risk. Including this compound makes little sense because the data will simply be discarded when running the risk assessment due to the lack of toxicological data. Several other compounds have toxicological data for ecological risk only. A protocol for ecological risk has not yet been developed. The generation of this data seems to have little use in risk assessments and CRWI suggests that these requirements be removed from the guidance document until EPA can show that the data will be used and, if so, how it will be used.

Particle Size Distribution

CRWI remains concerned about requiring the measurement of particle size distribution from incinerator stacks. While this guidance suggests that this be included as part of the data necessary to run the risk assessment, the guidance does not give any clear direction on how this parameter can be measured. In fact, in Section 2.4 the guidance states that a "widely accepted method for determining particle-size distribution is not yet available..." A number of facilities have attempted to measure particle size distribution with a limited amount of success. A scanning electron microscope has been tried at least two facilities (DuPont Wilmington Experimental Station and Chemical Waste Management Port Arthur) with little or no success. A cascades impactor was used at 3M's facility in Cottage Grove and at WTI in East Liverpool with limited success. In addition, there are stack gas temperature and moisture limitations on the use of a cascade impactor. In certain high moisture conditions, it is not possible to prevent condensation on the filter papers in the collection device. If these filter papers become wet, it becomes nearly impossible to remove the filters intact, without leaving portions on the metal screen. Failure to remove the entire filter will result in incorrect measurement of particulate catch. Similar problems with removal of the filter paper occur when the gas temperature exceeds 500° F. However, in this case, parts of the filter melt to the screen and complete removal becomes impossible. While in theory, the cascade impactor should work, there are several limitations to this method that will prevent universal application to hazardous waste combustors since approximately 60% of these combustors have wet stacks.

In comments on an earlier draft of this guidance, CRWI presented data that showed that the majority of particles have diameters less than two microns and suggested that sensitivity analysis be used to determine if the air dispersion models are sensitive to this parameter. A sensitivity analysis of the ISCST3 model was run and the report published in May 1997 (Model Parameter Sensitivity Analysis, The Air Group, May 23, 1997). While some questions were answered by this sensitivity analysis, others were not. This analysis shows that the model is not sensitive to the number of particle size categories as long as the same range of sizes is included. Stated another way, it did not matter whether there were 3 or 9 particle size categories as long as the range of categories did not change from the base case (0 to 15 micron diameters). When the analysis was run with 80% of the particles having a diameter less than 3.6 microns, there was no significant difference between this case and the base case for the particle and particle bound phase concentrations. There was a 50% reduction from the base case in the both dry deposition rates. The wet deposition rates were more complicated in that the smaller particles had less wet deposition up to 9 km from the source and a maximum of 30% greater wet deposition about 45 km from the source. However, this was only for the particle phase portion of the deposition. In addition, the report states that "small particles in the range of 1 micrometer (micron) have very low terminal velocities and effectively are suspended in air."

Given the information from the sensitivity analysis and the difficulty of developing accurate particle size distribution data with existing methods, CRWI strongly recommends that the default particle size distribution be changed to reflect the fact that best measurements to date indicate that 80% of the particles are less than 2 microns. This will more accurately reflect the state of the art for coarse particle removal and will create more accurate deposition estimates. CRWI sees no reason to measure a parameter where there is no known accurate method and available sensitivity analyses show the parameter not to be of major importance. Measuring particle size distribution will become even more inconsequential when facilities install equipment to meet the new MACT PM standard. At that point in time, it is likely that virtually all particles emitted from all facilities will be below 1 micron.

Again, thank you for the opportunity to comment on this draft trial burn guidance document. If you need additional information, please contact me.

Sincerely yours,

Melvin E. Keener
Executive Director

ATTACHMENT A - Specific Comments

1. Page 18 - "a facility may need to consider collection of SSRA emissions data during a separate risk burn if actual wastes are not burned during DRE and SRE tests." This is usually not such a black and white decision. Some alteration of waste feeds is required for several reasons. These include spiking with POHC for DRE testing, spiking with metals to establish metal feed rate limits, and meeting specifications for ash, chlorine, and heat content. Therefore, it is important that the operator and the permit writer have the ability to show that the trial burn waste meets the definition of "actual waste" that has been modified to acceptable levels.
2. Figure 3-1 (page 21) - CRWI sees no reason for the right hand side of this figure. There is no reason to impose additional permit conditions if the facility passes the DRE/SRE test and passes a realistic risk assessment.
3. Page 29 - The requirements to "assure combustion quality" are a part of the forthcoming HWC MACT rule. Under this rule, facilities have three years after promulgation date to come into compliance with these new regulations. Requiring a facility to meet these conditions prior to the promulgation date plus three years is not appropriate. In addition, the batch feed restrictions do not match what will be in the HWC MACT rule. CRWI suggests that the guidance document be modified to match the final rule.
4. Page 31 - Permit parameters in Table 4-1 are only required when the facility is required to meet the HWC MACT requirements. Early compliance with these standards is not appropriate and often not possible since compliance may require additional air pollution control equipment.
5. Page 31 - The entire discussion of transient conditions is somewhat disturbing because CRWI is only aware of a few instances where transient conditions occurred during trial burns and actual emissions were measured. A preliminary examination of this data indicates that there is little impact of transient conditions on emissions. We are currently gathering these data and will forward them when available. In addition, most transient conditions last a short period of time (few minutes) where the sample gathering normally takes at least an hour. With this testing protocol, it would be very difficult to determine whether short-term transient conditions contributed to increased emissions. This can only be determined in a research mode with instantaneous measurements of the pollutant of concern. It can not be determined as a part of the trial burn testing. In addition, it would be difficult, if not impossible, to create all the possible transient conditions during the trial burn process. CRWI suggests that the currently available data indicate that transient conditions do not result in a significant increase in emissions and, as such, should not be addressed during trial burn or risk assessment.
6. Page 32 - The batch feed controls listed do not correspond with the controls to be required in the soon-to-be-promulgated HWC MACT rule. In this rule, one-hour rolling average CO limit of 100 ppm is considered to be protective. This guidance should be changed to match the new rule.
7. Page 36 (Section 4.2) - This is one of a number of places where the guidance document reflects the soon-to-be-promulgated HWC MACT rule. The guidance document should make it clear that meeting MACT prior to the promulgation date plus three years is not expected and, in some cases, not possible.
8. Table 4-1 - The 1-hr RHA maximum PCC temperature is the basis for controlling metals but makes no sense in evaluating D/F precursors formed when feeding containerized waste. CO control is adequate; temperature spike control represents an inappropriate regulatory limit.
9. Page 51 - It should be noted that attempts to collect data and complete a mass balance for total organics have yielded questionable results and should be used only in the discussion of uncertainty. Quantitative adjustment of risk is not technically correct or appropriate.
10. Page 63 and 64 - Selenium has generally behaved as a semi-volatile metal. It should not be placed in the same category as mercury. In addition, the groupings of metals in Figure 6-1 do not match with the categories of metals soon to be regulated under the HWC MACT rule. This guidance must be consistent with that rule.
11. Page 70 - It is not clear how Method 5i procedures would apply to a cascade impactor. Has EPA redesigned the cascade impactor filter holder or is the document referring to recovery and weighing procedures for the filters? If so, what are the procedures?

Return to Comments

Return to Main Page