The characteristics of the submicrometer mode of ash and mercury speciation on combustion of coal in an oxygen-carbon dioxide and air (conventional) system were compared. The experiments were conducted at different O 2 :CO 2 and O 2 :N 2 :CO 2 mixing ratios. On replacing the nitrogen in air with carbon dioxide, the total mass of the particles in the submicrometer mode (<0.5 µm) of ash was smaller. Correspondingly, the geometric mean size (d pg ) of the submicrometer aerosol was smaller by approximately 28%. When the O 2 :CO 2 ratio was increased from 1:4 to 4:4, the geometric mean size of the submicrometer mode increased from 29 to 54 nm because of a faster vaporization rate as a result of a higher coal particle temperature. An increase in the geometric mean size was observed on increasing the N 2 :CO 2 ratio at a fixed O 2 concentration because of the same reasons. The shape of the primary particles was spherical in all the tests, indicating that a vapor to particle transformation pathway was prevalent. The ratios of elemental to oxidized mercury (approximately 4:1) were similar for O 2 -CO 2 and air combustion.

Addition of halogens or halides has been reported to promote mercury removal in coal-fired power plants. In this study, bench-and pilot-scale experiments were conducted using potassium iodide (KI) for capture and removal of Hg in air and coal combustion exhaust. Two bench-scale reactor systems were used: (1) a packedbed reactor (PBR) packed with granular or powder KI and (2) an aerosol flow reactor (AFR) with injection of KI particles. It was found that a higher temperature, a higher concentration of KI, and a longer gas residence time resulted in a higher Hg removal efficiency. A 100% Hg removal was achieved in the PBR above 300 °C using 0.5 g of powder KI and in the AFR above 500 °C with a KI/Hg molar ratio of 600 at a 5.8 s residence time. The low KI injection ratio relative to Hg indicated that KI is highly effective for Hg removal in air. Formation of I 2 vapor by the oxidation of KI by O 2 at high temperatures, which then reacts with Hg to produce HgI 2 , was identified as the pathway for removal. The pilot-scale experiments were conducted in a 160 kW pulverized coal combustor. KI was introduced in two ways: as a powder mixed with coal and by spraying KI solution droplets into the flue gas. In both cases the Hg removal efficiency increased with an increase in the feed rate of KI. Mixing KI powder with coal was found to be more effective than spraying KI into the flue gas, very likely due to the higher temperature, longer residence time of KI, and the formation of a secondary reactive sorbent. The Hg removal by KI was less efficient in the pilot-scale tests than in the bench-scale tests probably due to certain flue gas components reacting with KI or I 2 . Hg speciation measurements in both benchand pilot-scale experiments indicated no oxidized mercury in the gas phase upon introduction of KI, indicating that the oxidation product HgI 2 was captured in the particulate phase. This is very beneficial in coal-fired power plants equipped with electrostatic precipitators where particulate-bound Hg can be efficiently removed.