EPA’s Mercury Information Collection Request

Ralph Roberson

RMB Consulting & Research, Inc., 5104 Bur Oak Circle, Raleigh, North Carolina 27612

ABSTRACT

On April 9, 1998 EPA published a notice in the Federal Register stating that the Agency intends to obtain additional information on mercury emissions from coal-fired power plants under authority of Section 114 of the Clean Air Act (CAA). To accomplish its intent, EPA must prepare, submit, and have approved an Information Collection Request (ICR) by the Office of Management and Budget (OMB). EPA held a public meeting in Washington on May 21, 1998 and accepted comments from the public on its draft ICR through June 8, 1998.

Under EPA's draft ICR, all coal-fired power plants would be required to periodically sample and analyze the coal burned for mercury, and perhaps some additional elements such as chloride. Specifically, EPA would require . . . "weekly as-fired coal analyses from each distinct coal storage pile, including silos, etc., in use at the facility." EPA would also require each affected facility to measure and record the amount of coal burned in each week and identify the source (e.g., State, seam. etc.) of the coal.

The draft ICR also included provisions for obtaining speciated mercury stack sampling data. According to the draft ICR, EPA would require quarterly stack sampling data for about 30 coal-fired utility boilers. Sampling would be required at the inlet and outlet to the control devices. In identifying the 30 boilers, EPA suggests it would place each coal-fired utility boiler into one of nine bins. The nine bins are defined by intersection of three coal ranks: (1) bituminous, (2) subbituminous, and (3) lignite with three categories of flue gas desulfurization (FGD) systems: (1) wet, (2) dry, and (3) no FGD system.

The purpose of this paper is to review some of the comments submitted to EPA by the utility industry and to describe the ICR process as it evolved from EPA’s April 9^th draft through the ICR that was approved by OMB on or about November 16, 1998. Finally, this paper summarizes the requirements contained in the ICR that was approved by OMB.

BACKGROUND AND RATIONALE

On April 9, 1998, EPA published a notice in the Federal Register¹ stating the Agency planned to submit the following information collection request (ICR) to the Office of Management and Budget (OMB): Electric Utility Steam Generating Unit Mercury Information Collection Request; EPA ICR No. 1858.01. Before formally submitting the request to OMB for approval, EPA established a 60-day period for submitting written comments and held a public meeting in Washington, D.C. on May 21, 1998.

In its Final Report to Congress², EPA stated that mercury is the hazardous air pollutant (HAP) emission of greatest potential concern from cold-fired electric utility boilers and that additional research and monitoring are merited. EPA also listed a number of research needs related to mercury emissions. These included obtaining additional data on the mercury content of various types of coal burned by the electric utility industry and additional data on mercury emissions to the atmosphere, especially with respect to the various form or specie of the emitted mercury.

EPA’S DRAFT MERCURY ICR

To address issues related to coal characteristics, EPA’s draft ICR required the owner/operators of all coal-fired electric utility boilers to conduct periodic sampling and analysis of all coals burned at each facility. This would be accomplished by "obtaining weekly as-fired coal analyses from each distinct coal storage pile, including silos, etc., in use at the facility."³ In the draft ICR, EPA would require each facility to measure and record the amount of coal burned each week and identify the source of the coal (e.g., State, seam, etc.). Each collected coal sample would be analyzed for mercury and chlorine. EPA also initially requested an ultimate and proximate analysis of each sample.

To address the issue of mercury speciation, the draft ICR required quarterly, triplicate simultaneous before and after control device stack sampling to be performed on an EPA-designated set of boilers. In the April 1998 Federal Register notice, EPA stated that it considered 30 units to be a statistically representative sample of the coal-fired utility boiler population. EPA divided the coal-fired boiler population into nine categories as defined by the intersection of three coal ranks (bituminous, subbituminous, and lignite) with three scrubber scenarios (wet, dry, and none). [EPA realized that one of its categories probably had no members (i.e., bituminous coal with a dry scrubber).] In specifying the boilers to be tested, EPA stated it would use a proportional allocation methodology, with a provision for selecting at least two members from each of the remaining eight categories.

Initial Industry Response

In general, the electric utility industry was quite surprised and somewhat blindsided by EPA’s April 9^th Federal Register mercury ICR notice. The utility industry, especially through its associations like the Electric Power Research Institute (EPRI) and the Utility Air Regulatory Group (UARG), had cooperated with EPA both before and during the Agency’s preparation of the Utility HAPs Report to Congress. The industry was miffed to learn, via a Federal Register notice, that EPA believes it needs thousands and thousands of additional coal samples analyzed for mercury.

During the 60-day comment period for the draft ICR, the utility industry prepared and submitted written comments to EPA, participated in the public meeting in Washington, DC, and met with EPA staff in Research Triangle Park, NC. Upon review of the April 9^th Federal Register notice, EPRI immediately provided EPA with a 1995 EPRI report titled, Characterizing the Trace Element Content of Utility Coals. This report provides the results of a successful collaboration between coal companies and the utility industry, and summarizes the trace element content of over 1,500 privately collected coal samples. Through a hurriedly conducted telephone survey, EPRI was able to determine that between 5,000 and 7,500 additional coal samples had been analyzed for mercury. EPRI offered to obtain as many of these data as possible, conduct a quality assurance review of the data, and provide the results to EPA in a timely fashion. EPA did not respond.

The stack-testing component of EPA’s draft mercury ICR was equally perplexing. First, utility industry researchers were well ahead of EPA in understanding the environmental significance of mercury speciation and the importance of measuring speciated mercury emissions. In fact, EPRI and the Department of Energy’s Federal Energy Technology Center (FETC) collaborated on developing and evaluating a method for determining speciated mercury emissions for coal-fired utility boilers.⁴

In the face-to-face meeting with EPA staff, as well as in the written comments, the utility industry showed EPA how much mercury speciation data had been collected since establishing a valid sampling method (i.e., Ontario Hydro method). Industry representatives also explained to EPA how the mercury speciation data would be used to develop a predictive model so that speciated mercury emissions could be estimated for all coal-fired utility boilers. Lastly, industry reasoned that since its program was already underway, EPA could probably get results sooner by allowing the EPRI/DOE work to be completed than by requiring the stack sampling specified in the draft ICR.

EPA’s Initial Response to Comments

According to EPA, comments were received on the draft ICR from over 130 entities including utility companies, trade associations, cities, State agencies, State associations, environmental/public interest groups, private citizens, and even members of Congress. Unfortunately, EPA did not appear overly persuaded by the voluminous comments. Clearly, EPA was influenced by comments from some of the environmental groups that advocated increasing the number of boilers for stack testing from about 30 to 10 percent of the coal-fired boiler population. Of course, the environmental groups failed to provide any technical support or justification for the 10 percent position.

EPA’S SUBMITTAL TO OMB

On September 22, 1998, EPA published a notice in the Federal Register⁵ stating the Agency is submitting the following information collection request (ICR) to the Office of Management and Budget (OMB) for review and approval: Electric Utility Steam Generating Unit Mercury Information Collection Request; EPA ICR No. 1858.01. The Federal Register notice also announced the beginning of a 30-day public comment period. EPA’s mercury ICR consists of three parts. Part I is for general facility information and is to be completed once for each plant. Part II deals with coal sampling and analysis; Part III is a requirement to conduct speciated mercury stack testing for a selected group of coal-fired utility boilers.

Part I – General Facility Information

Part I is a questionnaire that asks for general plant information (e.g., owner, address, name, title, and phone number of appropriate plant contact) and specific information on all boilers that burn coal. For each coal-fired unit, EPA requires a table to be completed that will provide the boiler identification, MWe capacity, furnace type, type of NO_x control technology, type of SO₂ control technology, and type of particulate matter (PM) control technology. Part I must be completed and returned to EPA within 30 days of receipt.

Part II – Coal Sampling and Analyses

Part II would require that coal samples, which are collected for contract verification purposes, be analyzed. Specifically, EPA stated "[t]he owner/operator would also be required to document and provide the amount and type of each coal received for each unit and identify the source of each coal (i.e., State, county, seam, etc.). Each owner/operator would be required to provide the proximate and ultimate analyses of each coal obtained for contract verification purposes and provide analyses for mercury and chlorine content."⁶ The frequency of sampling and analysis would depend on the mode of transportation of coal to the plant.

Shipment by Rail Car, Barge, or Ship. Plants receiving coal by rail car, barge, or ship would be required to provide the results of the analysis of a composited coal sample that represents each particular coal shipment that is collected to determine compliance with individual coal contracts. Also, all owner/operators would be required to provide certain basic information (i.e., quantity, coal rank, and coal origin) for each shipment.

Shipment by Truck. Plants receiving coal by truck would be required to provide the results of the analysis of a composited coal sample that represents each particular coal shipment that is collected for normal contract verification purposes. EPA would anticipate that the analyses would be conducted on samples composited from individual truck samples, which are taken to verify performance with individual supplier contracts. Also, all owner/operators would be required to provide certain basic information (i.e., quantity, coal rank, and coal origin) for each shipment.

Shipment by Conveyor. Plants receiving coal by conveyor, slurry pipeline, etc. would be required to provide the results of the analysis of a sample that represents a coal increment no greater than 10,000 tons. The analysis would also include those parameters that are used for contract verification purposes. Finally, owner/operators are required to report certain basic information (i.e., quantity, coal rank, and coal origin) for each 10,000-ton increment.

Part III – Speciated Mercury Emissions Testing

In the documentation submitted to OMB for review and approval, EPA states that 138 boilers need to be sampled for speciated mercury emissions. The boilers were to be selected from 15 categories. Nine of the boiler categories arise from the intersection of three coal ranks (bituminous, subbituminous, and lignite) with three scrubber types (wet, dry, and none). Three categories are for fluidized bed combustion units that burn bituminous, subbituminous, or lignite coals. The final three categories are for coal gasification units that burn bituminous, subbituminous, or lignite coals.

For each unit selected for testing, EPA’s supporting documentation indicated that simultaneous before and after control device stack testing must be performed. Also, the documentation stated that three individual sampling runs must be performed using a paired sampling train at each location. The method to be used for mercury testing is titled "Standard Test Method for Elemental, Oxidized, Particle-Bound, and Total Mercury in Flue Gas from Coal-Fired Stationary Sources." This method is often referred to as the "Ontario Hydro Method."

INDUSTRY COMMENTS TO EPA AND OMB

Comments submitted by individual utility companies as well as many of the representing associations (e.g., EPRI and UARG) were similar to comments submitted on EPA’s draft ICR, which was issued in April 1998. That is, the utility industry generally believes that there exist more than enough data to adequately and accurately estimate annual mercury emissions for coal-fired power plants. Annual mercury emission estimates compiled by groups as diverse as industry consultants, EPA, and environmental organizations are always in close agreement and center around 50 tons per year. In one analysis, RMB calculated that the ICR, as submitted to OMB, would result in the analysis of 274,000 coal samples; the cost would be on the order of $50 million. We also observed that ultimate analysis is not routinely performed on coal samples collected for contract verification purposes; moreover, we know of no relationship between mercury speciation in flue gas and the various components of an ultimate analysis (e.g., carbon, hydrogen, oxygen, etc.).

Industry was also critical of the speciation testing, as described in the ICR documents submitted to OMB. While the utility industry has consistently acknowledged the need for additional mercury speciation data, the test program developed by EPA is unnecessarily expensive and not guaranteed to yield the necessary information. First, EPA misapplied a statistical formula to arrive at the 138 boilers to be sampled. Perhaps more importantly, EPA has neither developed nor put forth any mathematical framework with which to analyze or extrapolate the speciation data. Industry also observed that EPA’s sampling design probably does not reflect some of the important parameters that may control or at least significantly affect flue gas mercury speciation. Lastly, EPA’s ICR fails to incorporate or even acknowledge any of the mercury speciation data collected by EPRI and/or DOE.

THE MERCURY ICR APPROVED BY OMB

On or about November 17, 1998, EPA announced that OMB had approved an ICR on mercury emissions from coal-fired electric utility plants. Part I (General Facility Information) of EPA’s approved ICR is basically the same as discussed by EPA in the September 22^nd Federal Register notice and as previously summarized in this paper. Part II (Coal Sampling and Analysis) of EPA’s approved ICR is somewhat revised from that described in the September 22^nd Federal Register notice.

One significant change to Part II is that each coal-fired facility is, at least initially, required to analyze every sixth shipment for mercury and chlorine. However, each facility must obtain a minimum of three analyses of mercury and chlorine per month. Since mine-mouth plants do not receive coal shipments, per se, they are required to conduct the mercury and chlorine analysis once every 10 days in order to have the minimum of three analyses per month. At the end of the first quarter, EPA requires that each owner/operator conduct an evaluation to determine whether a 90 percent confidence interval about the average mercury in coal concentration is within ± 10 percent of the average concentration. If the statistical evaluation achieves the ± 10 percent criterion, the owner/operator will continue to analyze each sixth sample for mercury and chlorine. If the statistical test is failed, the owner/operator must begin to analyze each third sample for mercury and chlorine.

It should be pointed out that in EPA’s final Supporting Statement,⁷ the Agency states if data become available before reporting begins on January 1, 1999 indicating that the tolerance is either too stringent (or too lenient), EPA may adjust the ± 10 percent criterion to a more reasonable value. Another important revision to the final ICR as compared to the version submitted to OMB for approval includes deleting the requirement to conduct an ultimate analysis on each ICR sample.

Part III (Speciated Mercury Emissions Testing) of the approved ICR is somewhat different from the Part III version submitted to OMB. First, instead of having 15 categories or bins of coal-fired utility boilers, EPA expanded the number of categories to 45. The 45 categories are listed in Table 1. Basically, EPA has created three new subcategories for each of its original 15 categories. The new subcategories focus on particulate control technology and include (1) cold-side electrostatic precipitators (ESPs), (2) hot-side ESPs, and (3) any other particulate control technology. Then, EPA states that a maximum of three boilers will be selected from each of the 45 categories yielding a maximum total of 135 units to be tested. Also, in the final Supporting Statement,⁸ EPA acknowledges that DOE and EPRI have already collected speciated mercury emission data from several of the Agency’s boilers categories (i.e., various coal ranks, wet scrubbers, and ESP configurations) and these categories need not be retested pursuant to this ICR. In addition, EPA acknowledges that some of its 45 boilers categories have "zero" population, further reducing the number of boilers to be tested to about 75 – instead of 135.

Part I information is to be submitted to EPA by January 4, 1999. EPA intends to use the Part I submittals to place all of the coal-fired units into one of its 45 boiler categories. EPA will then choose the units that will be required to conduct mercury speciation testing. "Lottery" winners will be notified by means of a second Section 114 letter, which probably will be mailed in late February or early March 1999. Recipients of the second Section 114 letter will be required to prepare and submit a site-specific test plan to EPA for review and approval by April 15, 1999. Mercury speciation testing must be completed on or before May 31, 2000.

COMMENTARY

Implementing EPA’s mercury ICR will not bankrupt the electric utility industry. However, EPA’s ICR will result in the expenditure of considerable resources without commensurate benefits. At the end of the day, annual mercury emissions from coal-fired power plants will still be estimated to be 50 tons per year, ± 5-10 percent.

With respect to speciated mercury emissions testing, what will result from implementing the ICR is far from clear. That is, EPA’s framework for determining which units will be required to conduct mercury speciation testing is not well conceived. A number of variables that probably play a critical role in determining the form (specie) of the mercury in the flue gas are not addressed in EPA’s experimental design. We can only hope that EPA’s mercury ICR will create enough mercury speciation data from the appropriate facilities to permit the development of robust predictive models so that reasonable speciation profiles can be estimated for each and every coal-fired power plant in the U.S. Such emission profiles are important to improving our understanding of mercury control options as well as to better estimating the environmental and health benefits of any mercury emission reduction strategy.

REFERENCES

1. Federal Register, Vol. 63, 17406, April 9, 1998.

2. U.S. EPA. "Study of Hazardous Air Pollutant Emissions from Electric Utility Steam Generating Units – Final Report to Congress," Office of Air Quality Planning and Standards, Research Triangle Park, NC, EPA-453/R-98-004, February 1998.

3. Federal Register, Vol. 63, 17407, col. 3, April 9, 1998.

4. EPRI. "Evaluation of Flue Gas Mercury Speciation Methods," EPRI TR-108988, Electric Power Research Institute, Palo Alto, CA, December 1997.

5. Federal Register, Vol. 63, 50567, September 22, 1998.

6. U.S. EPA. "Standard Form 83-I Supporting Statement for OMB Review of EPA ICR No. 1858.01: Information Collection Request for Electric Utility Steam Generating Unit Mercury Emissions Information Collection Request," Office of Air Quality Planning and Standards, Research Triangle Park, NC, p. 13, September 3, 1998.

7. U.S. EPA. "Standard Form 83-I Supporting Statement for OMB Review of EPA ICR No. 1858.01: Information Collection Request for Electric Utility Steam Generating Unit Mercury Emissions Information Collection Request," Office of Air Quality Planning and Standards, Research Triangle Park, NC, p. 25, November 16, 1998.

8. Ibid, p. 28.

TABLE 1. EPA’S MERCURY ICR BOILER CATEGORIES