EPA mercury regime risks apples-to-oranges results

Pam Boschee

Associate Editor

Mercury made its entrance onto the emissions monitoring stage as directed by the U.S. Environmental Protection Agency (EPA) on Jan. 1, 1999. The fanfare about the public`s right to know failed to mention that standardized testing of coal samples was not part of EPA`s directive. Without standardization, plant-to-plant comparisons of testing results might prove meaningless. Many in the industry think the public`s right to know should be further defined as the public`s right to know accurate, comparable data.

Test methods published by EPA (for example, Solid Waste-846 method 7471A for mercury analysis) and the American Society for Testing and Materials (ASTM methods) are similar to cookbook recipes. The methods detail the type of sample appropriate for testing, preparation of the sample, step-by-step procedures for analysis (including type of instrumentation to be used), and levels of the parameter of interest that can be detected. Using the same test method by all plants assures comparable results because everyone should be following the same procedure.

All coal-fired power plants with a generating capacity of more than 25 MW (about 1,400 plants nationwide) must now report the mercury and chlorine content for every sixth shipment of coal they receive. Facilities with mine-mouth operations must analyze coal samples approximately every 10 days. Seventy-five selected plants (chosen on the basis of coal type and SO2 and particulate matter control equipment) will also be required to perform smokestack testing for quantity and species (elemental or oxidized) of mercury emissions once during the year. This stack test requirement is in addition to the sampling analysis requirement.

EPA expects to decide in two years whether it will regulate mercury, based on the data collected. Beginning in 2000, EPA plans to make utility mercury emissions data available over the Internet as part of its community right-to-know initiatives.

EPA released its new Emissions and Generation Resource Integrated Database (E-GRID) last December to help states develop mechanisms for electricity labeling and verification of green power marketing claims. Integrating 12 different federal databases, E-GRID-developed by EPA`s Acid Rain Division-includes 1996 data for CO2, SO2 and NOx for 4,800 utility plants. It provides information on emissions per unit of electricity, allowing direct comparison of emissions levels from different power sources. E-GRID shows, for individual power plants, companies, states and regions, the amount and percentage of power from different fuels, such as coal, natural gas, nuclear, hydro, solar and wind.

Because existing Energy Information Administration (EIA) policy provides confidential treatment for plant-specific nonutility data such as generation, fuel type, fuel quantity and quality, and emissions, this data is reported only when it could be obtained from other public sources. EIA is expected to publicly release operational data for nonutilities beginning with 1998 data.

Rick Morgan, EPA Acid Rain Division senior energy analyst, said EPA is considering adding more parameters in the future including mercury, nuclear waste, particulates and heated water discharge. For mercury, he emphasized that it is “critical to have information at the plant level. We have to be satisfied that we have good data.”

This all sounds straightforward-after all, coal analyses and stack emissions testing are already familiar to most plants. But mercury monitoring is not familiar, and its analysis presents unique issues for consideration in obtaining “good” data.

Many utilities are exploring their options and have not yet determined if they will gear up to analyze coal samples in-house, or whether they will contract the work with an outside laboratory. Manojit Sukul, American Electric Power`s environmental services engineer, said careful selection of an outside laboratory is critical to ensure accurate results. “The type of analysis depends on the system in place, and the laboratory will have detection limits depending on the method used.” Although the data is initially being collected for the public`s right to know, Sukul said, “it`s hard to second guess EPA.”

Steven Benson, associate director for research, and John Pavlish, senior research manager, with the Center for Air Toxic Metals, Energy and Environment Research Center (EERC) of the University of North Dakota, expressed similar concerns about contracting with laboratories.

Because mercury content of coal is very low (less than 1 part per million), the laboratory and its selected method of testing must be able to detect low levels of mercury. Laboratories and test methods vary in the levels (known as detection limits) they can accurately analyze. The laboratory selected to analyze a plant`s coal samples should be able to demonstrate its ability to achieve low detection limits with accuracy and precision. The accompanying sidebar gives additional guidelines to use in selecting an analytical laboratory.

Chlorine is another parameter now tested for in coal samples. During combustion, chlorine in coal binds to mercury, forming oxidized mercury (converting mercury from its original elemental form). Wet scrubbers capture oxidized mercury; therefore, higher chlorine levels in coal increase the ability of wet scrubbers to capture mercury, decreasing its release into the atmosphere-at least in theory.

However, Pavlish said, “It may be an oversimplification to measure only chlorine. The sodium, calcium or potassium contained in coals may tie up the chlorine, preventing it (chlorine) from being tied up with mercury.” In that case, the mercury would not be oxidized, and the wet scrubbers would not capture it.

Stack emissions testing presents its own set of concerns. According to Benson, there are only five or six laboratories in the United States that have performed these analyses at power plants.

For stack emissions, EPA specified a method, the Ontario Hydro Method. Pavlish said this method evolved many years ago, and modifications have been made to it over the years. He emphasized that a laboratory being contracted to perform the Ontario Hydro Method should be able to demonstrate hands-on experience in the last two years. Simply reading the method and being familiar with it will not be adequate preparation to provide data that can be reported to EPA with confidence.

EPA`s cost estimate of stack emissions testing is $44,500 per test. Benson said the cost will more likely run from $60,000 to $70,000 to “do it right and generate quality data.”

E-GRID can be downloaded from http://www.epa.gov/acidrain/egrid/egrid.htm. The Ontario Hydro Method is available at http://www.epa.gov/ttn/emc/prelim.html. n


Coal sample analysis guidelines provided by Gerald Holmes, analytical chemist, Severn Trent Laboratories, Pensacola, Fla.:

– Use a lab experienced with coal analysis.

– Ask about quality control (QC) tests and documentation.

– Request that the lab run QC tests on the plant`s coal samples to demonstrate their ability to generate reproducible results.

– Request that the lab run a detection limit study on coal to demonstrate their ability to consistently measure very low levels of mercury.

– Send a coal sample with a known amount of mercury (called a “standard”) for analysis. The National Institute of Standards & Technology offers several coal standards. The plant submitting the standard knows its true value, but the laboratory is kept “blind” to its value and must demonstrate its ability to accurately analyze the standard.

– Inquire about the lab`s certification, such as state or A2LA certification. This status indicates some measure of oversight.

– Lab should have written procedures and methods available for review.

– Lab should provide a liaison with the power plant, such as a project manager, to provide informed guidance in testing procedures and explain testing results.

Stack testing guidelines provided by Benson and Pavlish, EERC:

– Use a lab experienced in stack testing.

– Lab must have specific hands-on experience with the Ontario Hydro Method, preferably within the last two years.

– The lab analysts must be able to be trained very quickly in the method, or must team up with another lab that has experience.

Although mercury emissions are not currently regulated, this is not the time for complacency in planning coal-fired plants` procedures for compliance in reporting to EPA. Time and money invested now will deliver payoffs in the form of reliable and valid mercury analysis results, which may ultimately be used as base line data for subsequent EPA comparisons.

Previous articleELP Volume 77 Issue 2
Next articleELP Volume 77 Issue 3

No posts to display