fs095-01 10/29/01 11:37 AM Page 1

Mercury in U.S. Coal—Abundance, Distribution, and Modes of Occurrence

Introduction In February 1998, The U.S. Environ- mental Protection Agency (EPA, 1998a,b) issued a report citing mercury emissions Atmospheric from electric utilities as the largest Hg Anthropogenic Hg remaining anthropogenic source of mer- Global deposition cury released to the air. EPA officials Natural Hg estimated that about 50 tons of elemental mercury are emitted each year from U.S. terrestrial marine coal-burning powerplants, with lesser Local and amounts coming from oil- and gas-burn- regional Reemitted Reemitted ing units. According to EPA estimates, deposition anthropogenic anthropogenic emissions from coal-fired utilities and natural Hg and natural Hg account for 13 to 26 percent of the total (natural plus anthropogenic) airborne emissions of mercury in the United States. On December 14, 2000, the EPA Methylated announced that it will require a reduction Hg in mercury emissions from coal-fired powerplants, with regulations proposed by 2003 and final rules for implementa- tion completed by 2004 (EPA, 2000). Figure 1. Simplified geochemical cycle of mercury (Hg).

Environmental Significance of uptake via the food chain. Cases of mer- mercury to the environment. Additional Mercury cury poisoning have been documented in options include selective mining of coal The mercury (Hg) directly emitted people who eat contaminated fish for pro- (avoiding parts of a coal bed that are from powerplants generally is not consid- longed periods, both in the United States higher in mercury content), coal washing ered harmful; however, in the natural and abroad. Pregnant women and subsis- (to reduce the amount of mercury in the environment, mercury can go through a tence fishermen are particularly vulnera- coal delivered to the powerplants), switch- series of chemical transformations that ble. Because high levels of mercury have ing from coal to natural gas, and postcom- convert elemental mercury to a highly been detected in fish, many U.S. States bustion removal of mercury from the toxic form that is concentrated in fish and have issued advisories that restrict fishing. powerplant stack emissions. Information birds (fig. 1). The most toxic form of Reduction in mercury emissions from on the abundance, distribution, and forms mercury is methylmercury, an organic U.S. coal-fired powerplants may help of mercury in coal may be helpful in form created by a complex bacterial con- minimize or avoid health problems caused selecting the most efficient and cost-effec- version of inorganic mercury. Methyla- by exposure to excess mercury. There are tive options for mercury reduction. tion rates (creation of methylmercury) in several ways in which this reduction can ecosystems are a function of mercury be accomplished. One option to reduce Abundance and Distribution of availability, bacterial population, nutrient the quantity of mercury in the atmosphere Mercury in Coal load, acidity and oxidizing conditions, is to use high-rank coals. Generally, mois- The U.S. Geological Survey (USGS) sediment load, and sedimentation rates ture in coal decreases and calorific value has compiled a nationwide coal informa- (National Research Council, 1978). (thermal energy) increases as coal rank tion database over the last 25 years. A Methylmercury enters the food chain, (degree of maturation) increases. There- subset of the data, called COALQUAL particularly in aquatic organisms, and fore, powerplants that burn high-rank coal (Bragg and others, 1998) contains analy- bioaccumulates. Bioaccumulation is the in their boilers require less coal for a ses of over 7,000 coal samples that have enrichment of a substance in an organism given thermal output. Thus, for coals hav- been collected or calculated to represent and includes bioconcentration from envi- ing similar mercury concentrations, the the entire thickness of a coal bed in the ronmental concentrations and additional higher rank coals will contribute less ground.

U.S. Department of the Interior USGS Fact Sheet FS–095–01 U.S. Geological Survey September 2001 fs095-01 10/29/01 11:37 AM Page 2

Figure 2 is a histogram of the mercu- 25 ry values in the COALQUAL database for conterminous U.S. coal. Statistics for all analyses indicate a mean of 0.17 part 20 per million (ppm), with a median and standard deviation of 0.11 ppm and 0.17, respectively. About 80 percent of the 15 mercury concentrations in the database are less than 0.25 ppm. The maximum mercury database value for coal in the 10 ground is 1.8 ppm, after deleting one higher value as a statistical outlier. Table 1 shows the median and mean 5 values for mercury concentrations (in IN PERCENT FREQUENCY, ppm) and calorific values (British thermal units per pound (Btu/lb)), as well as the 0 number of analyses, for selected coal- 0.003 0.253 0.503 0.753 1.003 1.253 1.503 1.753 producing regions in the United States, using the COALQUAL database. The MERCURY, IN PARTS PER MILLION mercury data in table 1 have been calcu- lated back to an as-received basis, Figure 2. Histogram of mercury concentrations (remnant moisture, whole coal basis) for conter- approximately the mercury concentration minous U.S. coal from the COALQUAL database. of the coal in the ground. Northern Appalachian area coal has percent, respectively). About 14 percent Powder River Basin are slightly higher in the highest mean and median values for of U.S. production comes from coal in mercury levels than the subbituminous mercury, with coal from the southern the Interior areas. coals of the San Juan River Basin. Appalachian area having the second high- On the basis of the information est value and coal from the central shown in figure 3, the Gulf Coast lignite Modes of Occurrence and Reduction Appalachian area slightly lower. Coal may have the highest potential for mercu- of Mercury from these three areas has extremely high ry emissions, and the Green River coal The COALQUAL data set does not calorific values. Coal from the Uinta from western Wyoming may have the take into account the potentially substantial region has the lowest mean and median lowest mercury emissions on an equal- reduction of mercury by physical coal mercury values of all indicated areas. energy basis. Of the two major bitumi- cleaning, because the analyses represent Some western U.S. coals are low in mer- nous coal-producing regions, samples coal as it exists in the ground. The modes cury but are also low in calorific value, from the Appalachian region contain of occurrence of an element in coal can because they are low in rank. higher mercury levels than those from affect the way the element behaves during The concentration of mercury can the Eastern Interior. Samples from the coal cleaning, combustion, and leaching. also be presented on an equal-energy basis (input load) in pounds (lb) per tril- Table 1. Median and mean values for mercury concentrations (in parts per million (ppm)) and 12 lion (10 ) Btu to provide a convenient calorific values (in British thermal units per pound (Btu/lb) on an as-received, whole coal basis unit of comparison between coal from for selected coal-producing regions in the United States. different areas (fig. 3). This is a simple [No. = number of analyses] calculation, dividing as-received mercury ppm values by Btu/lb and expressing the Mercury Calorific value Coal-producing (ppm) (Btu/lb) value on a 1012 Btu basis. The data from region COALQUAL used in this analysis yield a Median Mean No. Median Mean No. mean U.S. input load of 14 lb Hg/1012 Btu (with a median of 9.7 and a standard Appalachian, northern 0.19 0.24 1,613 12,570 12,440 1,506 Appalachian, central .10 .15 1,747 13,360 13,210 1,648 deviation of 15). The calculated input Appalachian, southern .18 .21 975 12,850 12,760 969 loads from individual samples were used Eastern Interior .07 .10 289 11,510 11,450 255 to calculate a mean value for each of the Fort Union .08 .10 300 6,280 6,360 277 selected coal-producing regions listed in Green River .06 .09 388 9,940 9,560 264 table 1. Mean mercury input loads were Gulf Coast .13 .16 141 6,440 6,470 110 Pennsylvania divided into arbitrary 5-unit intervals and Anthracite .10 .10 51 12,860 12,520 39 are color-coded in figure 3. According to Powder River .06 .08 612 8,050 8,090 489 the Energy Information Administration Raton Mesa .05 .09 40 12,500 12,300 34 (EIA, 2001), U.S. coal production, which San Juan River .04 .08 192 9,340 9,610 173 can be roughly correlated with usage, is Uinta .04 .07 253 11,280 10,810 226 Western Interior .14 .18 286 11,320 11,420 261 similar between coal regions east and Wind River .08 .15 42 9,580 9,560 42 west of the Mississippi River (38 and 48 fs095-01 10/29/01 11:37 AM Page 3

Figure 3. Mercury input loadings (in pounds of mercury per 1012 British thermal units (lb Hg/1012 Btu)) of in-ground coal for selected U.S. coal-producing regions.

Thus, the element’s mode of occurrence niques. The results of these studies indi- coal-producing region. On an equal-ener- has an important influence on its environ- cate that, on the average, 37 percent of the gy basis, Gulf Coast coal samples have mental and technological impacts. Because mercury is removed by coal cleaning the highest input load values (27.0 lb of the low concentrations (commonly less (Toole-O’Neil and others, 1999). The Hg/1012 Btu), and the Green River region than 0.2 ppm) of mercury and its volatility, information that the USGS is generating samples have the lowest values (6.5 lb it is particularly difficult to determine the on mercury distribution and modes of Hg/1012 Btu). modes of mercury occurrence in coal. occurrence is also relevant to mercury The COALQUAL database is an USGS research indicates that much of the reduction by fuel switching, selective extremely valuable source of information mercury in coal is associated with pyrite, mining, and chemical coal cleaning. Flue for raw or in-ground trace-element con- which generally forms after the coal is gas controls on mercury (sorbent injection centrations in U.S. coals and, if adjusted compacted (fig. 4). Other forms of mercu- and hydrothermal treatment technologies) for the effect of coal cleaning in appropri- ry that have been reported in coal are are also being evaluated by research ate coals, can provide a first estimate of organically bound, elemental, and in sul- organizations as possible economic solu- as-shipped mercury concentration in coal fide and selenide minerals (fig. 5). tions for mercury reduction. where data are not available. Physical coal The U.S. Geological Survey is col- cleaning is a viable method of reducing laborating on research to determine if the Summary mercury that enters the combustion sys- modes of occurrence of mercury in coal The concentration of mercury in coal tem and, therefore, reducing mercury that influence the formation of mercury samples from the U.S. Geological Sur- enters the atmosphere. The mean mercury species during the combustion process vey’s COALQUAL database averages concentration of eastern U.S. coals may and thus the likelihood of mercury capture 0.17 ppm for in-ground coal in the con- be less than reported, if the impact of from the gas. The USGS has also collabo- terminous United States. Mean values physical coal cleaning is considered. rated with industry on research to assess range from 0.07 ppm for coal samples —By Susan J. Tewalt, the removability of mercury from coal by from the Uinta region to 0.24 ppm for Linda J. Bragg, and conventional physical coal-cleaning tech- samples from the northern Appalachian Robert B. Finkelman fs095-01 10/29/01 11:37 AM Page 4

MERCURY 100 80 60 40 20

PERCENT LEACHED 0

Wyee

Wyodak-1 Pittsburgh-2 Illinois #6-1Illinois #6-2Illinois #6-3 Pittsburgh-1 Pittsburgh-3Pittsburgh-4 Warrior Basin Upper Freeport Elkhorn/Hazard Roseud-McKay Upper Kittanning Gascoigne Wood #2

CH34 COONH HCI HF HNO3 Ammonium Hydrochloric Hydrofluoric Nitric acid acetate acid acid

Figure 4. Selective leaching results for 15 coal samples (12 from the United States) (Palmer and others, 1998). The yellow bars indicate the proportion of mercury leached by nitric acid. This mercury is believed to be associated with the sulfide minerals, such as pyrite. Direct analysis of pyrite grains by a laser ablation mass analyzer indicated mercury concentrations consistent with selective leaching data. The green bars indicate the mercury leached by hydrochloric acid; much of this mercury may have come from oxidized pyrite. Arrows indicate minimum values.

National Research Council, 1978, An ——— 2000, EPA to regulate mercury assessment of mercury in the envi- and other air toxics emissions from ronment: National Academy of Sci- coal- and oil-fired powerplants: EPA ences, 192 p. fact sheet available online at http: Palmer, C.A., Finkelman, R.B., //www.epa.gov/ttncaaa1/t3/fact_sheets Mroczkowski, S.J., Willett, J.C., Tay- /fs_util.pdf. (Accessed April 2001.) lor, K.C., and Bullock, Jr., J.H., 1998 [abs.], Arsenic and mercury modes of occurrence in coal—Implications for For more information on the following reducing emissions of these elements subjects, please contact: from coal-fired powerplants, in Pro- ceedings of the Air Quality Confer- Mercury input loads Figure 5. Scanning electron photomicrograph ence, McLean, Va.: Energy and Envi- Susan J. Tewalt of a polished block of lignite from California. ronmental Research Center, Universi- (E-mail: [email protected]) The minute (less than 1 micrometer) bright ty of North Dakota, unpaginated. spots are rare grains of mercury selenide. Toole-O’Neil, B., Tewalt, S.J., Finkel- COALQUAL database man, R.B., and Akers, D.J., 1999, Linda J. Bragg References Cited Mercury concentration in coal— (E-mail: [email protected]) Bragg, L.J., Oman, J.K., Tewalt, S.J., Unraveling the puzzle: Fuel, v. 78, Oman, C.L., Rega, N.H., Washing- p. 47Ð54. Mercury modes of occurrence ton, P.M., and Finkelman, R.B., U.S. Environmental Protection Agency, Robert B. Finkelman 1998, The U.S. Geological Survey 1998a, Mercury emissions and elec- (E-mail: [email protected]) Coal Quality (COALQUAL) Data- tric utilities: EPA fact sheet available base—Version 2.0: U.S. Geological online at http://www.epa.gov/ttncaaa1/ U.S. Geological Survey Survey Open-File Report 97Ð134, t3/fact_sheets/hg17th.html. National Center, MS 956 one CD-ROM. (Accessed April 2001.) Reston, VA 20192 Energy Information Adminstration, 2001, ——— 1998b, Mercury study report to Weekly coal production: EIA web Congress: EPG white paper available site at http://www.eia.doe.gov/cneaf/ online at http://www.epa.gov/oar/ coal/weekly/weekly_html/wcppage. merwhite.html. (Accessed April htm. (Accessed July 18, 2001.) 2001.)