Geoffrey D. Silcox

Fundamentals of Mercury Oxidation in Flue Gas

The objective of this project is to understand the importance of and the contribution of gas-phase and solid-phase coal constituents in the mercury oxidation reactions. The project involves two experimental scales and a modeling effort. The team is comprised of University of Utah, Reaction Engineering International, and University of Connecticut. The objective is to determine the experimental parameters of importance in the homogeneous and heterogeneous oxidation reactions; validate models; and, improve existing models. Parameters to be studies include HCl, NOx, and SO{sub 2} concentrations, ash constituents, and temperature. This report summarizes Year 1 results for the experimental and modeling tasks. Experiments in the drop tube are just beginning and a new, speciated mercury analyzer is up and running. A preliminary assessment has been made for the drop tube experiments using the existing model of gas-phase kinetics.

Analysis of Halogen-Mercury Reactions in Flue Gas

Oxidized mercury species may be formed in combustion systems through gas-phase reactions between elemental mercury and halogens, such as chorine or bromine. This study examines how bromine species affect mercury oxidation in the gas phase and examines the effects of mixtures of bromine and chlorine on extents of oxidation. Experiments were conducted in a bench-scale, laminar flow, methane-fired (300 W), quartz-lined reactor in which gas composition (HCl, HBr, NO{sub x}, SO{sub 2}) and temperature profile were varied. In the experiments, the post-combustion gases were quenched from flame temperatures to about 350 C, and then speciated mercury was measured using a wet conditioning system and continuous emissions monitor (CEM). Supporting kinetic calculations were performed and compared with measured levels of oxidation. A significant portion of this report is devoted to sample conditioning as part of the mercury analysis system. In combustion systems with significant amounts of Br{sub 2} in the flue gas, the impinger solutions used to speciate mercury may be biased and care must be taken in interpreting mercury oxidation results. The stannous chloride solution used in the CEM conditioning system to convert all mercury to total mercury did not provide complete conversion of oxidized mercury to elemental, when bromine was added to the combustion system, resulting in a low bias for the total mercury measurement. The use of a hydroxylamine hydrochloride and sodium hydroxide solution instead of stannous chloride showed a significant improvement in the measurement of total mercury. Bromine was shown to be much more effective in the post-flame, homogeneous oxidation of mercury than chlorine, on an equivalent molar basis. Addition of NO to the flame (up to 400 ppmv) had no impact on mercury oxidation by chlorine or bromine. Addition of SO{sub 2} had no effect on mercury oxidation by chlorine at SO{sub 2} concentrations below about 400 ppmv; some increase in mercury oxidation was observed at SO{sub 2} concentrations of 400 ppmv and higher. In contrast, SO{sub 2} concentrations as low as 50 ppmv significantly reduced mercury oxidation by bromine, this reduction could be due to both gas and liquid phase interactions between SO{sub 2} and oxidized mercury species. The simultaneous presence of chlorine and bromine in the flue gas resulted in a slight increase in mercury oxidation above that obtained with bromine alone, the extent of the observed increase is proportional to the chlorine concentration. The results of this study can be used to understand the relative importance of gas-phase mercury oxidation by bromine and chlorine in combustion systems. Two temperature profiles were tested: a low quench (210 K/s) and a high quench (440 K/s). For chlorine the effects of quench rate were slight and hard to characterize with confidence. Oxidation with bromine proved sensitive to quench rate with significantly more oxidation at the lower rate. The data generated in this program are the first homogeneous laboratory-scale data on bromine-induced oxidation of mercury in a combustion system. Five Hg-Cl and three Hg-Br mechanisms, some published and others under development, were evaluated and compared to the new data. The Hg-halogen mechanisms were combined with submechanisms from Reaction Engineering International for NO{sub x}, SO{sub x}, and hydrocarbons. The homogeneous kinetics under-predicted the levels of mercury oxidation observed in full-scale systems. This shortcoming can be corrected by including heterogeneous kinetics in the model calculations.