(14b) Modeling Mercury Chemistry in Coal-Fired Air Preheaters: Effects of Sulfur Trioxide and Surface Area

Mercury emissions from coal-fired power plants are a concern to both the state and federal governments in the U.S.  US EPA had proposed to regulate mercury emissions from coal-fired power plants using the Clean Air Mercury Rule (CAMR).  This two-phase program allowed trading of mercury emissions, similar to the SO₂ trading program already in place.  As of now, all aspects of CAMR have been vacated by a court ruling, and the EPA has committed to proposing draft mercury emissions regulations for coal-fired power plants by March, 2011.  As of October 2009, nineteen states had passed their own mercury emissions regulations.   

Mercury exists as the elemental form (Hg⁰) in the high-temperature regions of coal-fired boilers.  As the flue gas is cooled, a series of complex reactions begin to convert the Hg⁰ to gaseous oxidized forms (Hg²⁺) and particulate-bound mercury (Hg_p). The extent of conversion of Hg⁰ to Hg²⁺ and Hg_p depends on the flue gas composition, the amount and properties of fly ash and the flue gas temperature and quench rate.  The speciation of mercury in coal combustion flue gas affects the performance of activated carbon (a dedicated mercury control technology) and the removal of mercury by wet FGD scrubbers (a ?co-benefit? approach to mercury control).

In order to prepare for the existing and impending mercury emission regulations, utilities must have useful tools for compliance planning.  REI's MerSim^TM mercury simulation tool includes homogeneous and heterogeneous oxidation kinetics, adsorption on fly ash, oxidation across SCRs, and removal and re-emission across wet FGD scrubbers.  This model can be used by utilities, given inputs that are generally available to them.  On the surface, this requirement might seem obvious and inconsequential; however, matching models for the complex chemistry of mercury in practical combustion systems with the information that is generally available to plant engineers is a difficult undertaking. 

We have endeavored to create fundamentally based submodels for mercury behavior, but these models must be based on input data that a plant engineer can reasonably supply.   An integrated model requiring inputs that are not available or difficult to obtain will not be a useful compliance tool.  Therefore, a balance must be struck between the complexity of the submodels and the complexity of the input parameters, in order to provide utilities with a useful and accurate tool.

Detailed homogeneous and heterogeneous kinetic pathways for mercury oxidation are included in the MerSim integrated power plant model.  Overall results of the model have been previously reported.¹  However, details of some of the key submodels have not always been reported in detail.  Innovation and improvement continue in the integrated model.  Recently, a more detailed submodel has been implemented to predict both mercury oxidation and adsorption across regenerative air preheaters.

Using full-scale mercury speciation data assembled from previous DOE- and EPRI-funded programs,¹  one can see that regenerative air preheaters promote formation of oxidized and, to a lesser extent, particulate-bound mercury.  Figure 1 illustrates these points with observed values of oxidized and particulate-bound mercury at air preheater outlets. 

Figure 1.  Observed percentage of oxidized and particulate-bound mercury at air preheater exits from full-scale coal-fired power plants.

Our approach was to use the heterogeneous and homogeneous rates for mercury oxidation and the rate of mercury sorption, which are currently used to predict in-flight mercury transformations in the duct and ESP submodels, to account for deposition of fly ash on the air preheater surfaces.  Furthermore, detailed SO₂ oxidation kinetics (the Leeds mechanism²) were added to the gas-phase kinetic model, to accurately predict the concentration of SO₃ at the air preheater inlet.

Heterogeneous reactions of mercury are assumed to take place on the surface of unburned carbon particles.  Unburned carbon particles from coal combustion are larger than inorganic fly ash particles and have a lower density.² A mean diameter of 50 microns and a density of 0.4 g/cm³ would be reasonable for unburned carbon. Using these mean values for unburned carbon and ash loading, the external surface area of unburned carbon in the duct (in m²/m³) can be calculated. The air preheater surface area is much greater than the external surface area of the unburned carbon that is suspended in the flue gas. Thus, unburned carbon that deposits on the air preheater surface area represents more available surface area.  The available surface area for heterogeneous mercury reaction is thus assumed to be the sum of the in-flight ash and the ash deposited on the air preheater surfaces.

If we can predict accurately the concentration of SO₃ in the air preheater, and the loss of SO₃ in the air preheater, then we can begin to outline a way to predict the effect of SO₃ on oxidation and adsorption of mercury on unburned carbon in the air preheater.  Note that H₂SO₄ and SO₃ will be used interchangeably here.   SO₃ is the oxidized sulfur species created in the flame or across the SCR, but it is converted to H₂SO₄ at lower temperatures.  Since concentrations of SO₃/H₂SO₄ will be expressed in ppmv, the concentrations in the flue gas will be equivalent, no matter which form is thermodynamically stable.

Most power plants do not operate such that the air preheater exit temperature is below the dew point of H₂SO₄.  However, much of the surface within the air preheater is below the dew point, because of contact with the relatively cold air stream.  Thus, there is loss of H₂SO₄ within an air preheater, even when the exit temperature in the flue gas is above the dew point.  H₂SO₄ is condensing on the surfaces inside the air preheater.  It must, therefore, condense on unburned carbon on those surfaces.  The condensation of SO₃ should remove sites for mercury oxidation and adsorption.  Therefore, the reactive surface area in the air preheater available for reaction with mercury was reduced in the model, based on the predicted condensation of SO₃ in the air preheater. 

The results of the calculation of mercury oxidation across the air preheaters in full-scale power plants are shown in Figure 2, in which calculated oxidation is compared against observed oxidation at the air preheater outlet.  Further details of the model and of predictions of mercury emissions from coal-fired power plants will be presented.

Figure 2.  Observed and predicted oxidized mercury at air preheater outlet. References

1. Senior, C., Fry, A., Montgomery, C., Sarofim, A., Wendt, J.  Modeling Tool for Evaluation of Utility Mercury Control Strategies.? Presented at the DOE-EPRI-U.S. EPA -A&WMA Combined Power Plant Air Pollutant Control Symposium - The Mega Symposium, Baltimore, MD, August 28-31, 2006.

2. http://www.chem.leeds.ac.uk/Combustion/Combustion.html.

3. Belbla, V.H.; Martin, C.E.; Baldrey, K.; Lindsey, C.; Altman, R. The Effects of Unburned Carbon Properties on Electrostatic Precipitator Performance Modeling. Presented at the DOE-EPRI-U.S. EPA -A&WMA Combined Power Plant Air Pollutant Control Symposium - The Mega Symposium, Baltimore, MD, August 28-31, 2006.