This paper presents an overview of research related to mercury control technology for coal-fired power plants and identifies areas requiring additional research and development. It critically reviews measured mercury emissions; the chemistry of mercury transformation and control; progress in the development of promising control technologies: sorbent injection, control in wet scrubbers, and coal cleaning; and projects costs for mercury control. Currently, there is no single best technology that can be broadly applied. Combinations of available control methods may be able to provide up to 90% control for some plants but not others. In August 2000, the National Research Council completed a study that determined that the U.S. Environmental Protection Agency's (EPA) conservative exposure reference dose of 0.1 mug mercury/ kg body weight/day was scientifically justified to protect against harmful neurological effects during fetal development and early childhood. Subsequently, in December 2000, EPA made its regulatory decision that mercury emissions from coal-fired electric generating plants will need to be controlled on a schedule that calls for a proposed rule by December 2003, a final rule by December 2004, and full compliance by the end of 2007. Coal-fired utility boilers are currently the largest single-known source of mercury emissions in the United States. EPA's Information Collection Request (ICR) to coal-burning utilities indicated that there were 75 tons of mercury in the 900 million tons of coal used in U.S. power plants during 1999. Estimates of total mercury emissions from coal-fired plants based on ICR data range from 40 to 52 tons. On average, about 40% of the mercury entering a coal-fired power plant is captured and 60% emitted. Percentage emissions of mercury for individual plants tested under the ICR varied widely depending on coal type and emission control equipment. Western subbituminous coals on average contain only about half as much mercury as Appalachian bituminous coals, but the higher chlorine content of the latter promotes mercury oxidation and results in a higher percentage of mercury capture. Some iron minerals found in coal also catalyze mercury oxidation, whereas calcium and sulfur tend to impede oxidation. Review of ICR data on mercury capture in boilers and existing control devices indicates very little mercury removal within a pulverized coal-fired boiler, and the level of mercury oxidation at the exit of the boiler was increased for higher coal chlorine contents and lower exit temperatures. Mercury removals across cold-side electrostatic precipitators (ESPs) averaged 27%, compared to 4% for hot-side ESPs. Removals for fabric filters (FFs) were higher, averaging 58%, owing to additional gas-solid contact time for oxidation. Both wet and dry flue gas desulfurization (FGD) systems removed 80% to 90% of the gaseous mercury(II), but elemental mercury (Hg-0) was not affected. High mercury removals, averaging 86%, in fluidized-bed combustors with FFs were attributed to mercury capture on high-carbon fly ash. Tests on the two coal-fired integrated gasification combined-cycle plants in the United States suggest that about half of the coal mercury was emitted predominantly in elemental form. ICR tests on selective catalytic reduction and selective noncatalytic reduction used for NO, control were inconclusive, and additional full-scale tests are in progress. The mechanisms responsible for varied levels of mercury oxidation and capture are beginning to be understood. Mercury in coal occurs in association with pyrite and other sulfide minerals and may also be organically bound. Coal mercury is converted to gaseous Hg-0 in the combustion flame and is subsequently partially oxidized (35% to 95%) as the combustion gases cool. Mercury oxidation in boiler systems is kinetically controlled; homogeneous oxidation reactions are promoted by chlorine and atomic chlorine, and heterogeneous oxidation is promoted by fly ash and sorbents. Acid gases critically influence the heterogeneous oxidation of mercury, particularly as it affects capture on sorbents. HCl, NO, and NO, all promote oxidation and capture both individually and in combination. However, the combination of SO2 with NO2 greatly reduces capture of Hg-0 on activated carbon, whereas oxidation continues on the solid surface. Mass transfer of gaseous mercury by diffusion from the bulk gas to the solid surface can also limit heterogeneous oxidation and capture of mercury, but diffusion within a porous sorbent is not believed to be rate-limiting. Reducing the size of the sorbent particles and increasing their dispersion can greatly enhance control where mass transfer is limiting. To achieve 90% control of a mercury concentration of 10 mug/scm in 2-s residence time by activated carbon injection requires a minimum carbon-to-mercury (C/Hg) mass ratio of about 3000:1 for 4-mum particles compared to 18,000:1 for 10-mum particles. Mercury removals in some tests performed to characterize sorbents have been mass transfer-limited by the large particle size of the sorbents used. Mercury sorption capacities between about 200 and 5000 mug Hg/g C have been reported for conditions applying to coal combustion. However, higher measured capacities do not always correlate with higher removal levels in practice because of the effect of other variables. What is important is that several of the activated carbons tested have sufficient capacity to capture mercury at carbon injection rates below a C/Hg mass ratio of 10,000, based on both laboratory and field sorption tests. Since capacity is defined in reference to an assumed sorption equilibrium, the equilibrium capacity of a sorbent determined over a period of hours in the laboratory may have limited relevance to the amount of mercury captured in a few seconds' time of flight or in minutes of contact time on an FF. Laboratory tests that are more representative of the conditions in an actual control device are needed to determine more useful capacity factors. Injection of activated carbon upstream of either an ESP or an FF baghouse is a retrofit control technology that has potential application to 75% of all coal-fired power plants in the United States that are not equipped with FGD scrubbers. Field and pilot-scale tests on activated carbon injection for mercury control have resulted in mercury removals between about 25% and 95% over the range of 2000-15,000 C/Hg mass ratio. The mercury removal data from some tests could be correlated with carbon injection rates by assuming that the removal was mass transfer-limited, whereas in tests on other coals, removals appeared to be controlled by catalytic oxidation and capture on fly ash. Mercury capture on sorbents, therefore, depends on the properties of the coal being burned, and pilot-scale tests on particular coals should be performed before a full-scale sorbent injection system is designed. Development of low-cost, ultrafine sorbents with high effective sorption capacities and rapid reaction kinetics would revolutionize injection technology. Engineering development is also needed to improve sorbent dispersion and to optimize gas-solid contact time. Wet FGD units currently installed on about 25% of the U.S. coal-fired utility boilers remove nearly 90% of the mercury(II) entering but essentially none of the Hg-0. Research to enhance mercury removal in scrubbers focuses on converting Hg-0 to an oxidized form in or ahead of the scrubber using proprietary reagents. Palladium and carbon-based catalysts have shown the most promise for oxidizing Hg-0. Mercury removals from near 0% to about 60% are reported for the physical washing methods of the type that are widely used to remove pyritic sulfur and ash from 77% of all bituminous coal used in the United States. Advanced cleaning methods and hydrothermal treatment offer higher removals, but no coal-cleaning method is likely to reliably meet a 70% or greater removal requirement. Coal cleaning could, however, contribute to overall mercury control under a cap-and-trade form of mercury regulation. Concerns over the release of mercury from coal combustion by-products by leaching or atmospheric reemission will be heightened with the installation of mercury control technologies. Concentrations of mercury in leachates from fly ashes, FGD materials, and activated carbon saturated with mercury are very low and usually below detection limits. Essentially, no mercury emission from these materials into air has been measured at ambient temperature. However, mercury is released from saturated sorbents upon heating above 135 degreesC. Preliminary results on the stability of mercury on fly ash, FGD materials, and saturated carbons are encouraging, but more testing is needed before the concerns are fully resolved. Published by Elsevier Science B.V.