Author: Ana Suarez Negreira

Publisher:

Published: 2013

Total Pages:

ISBN-13:

Mercury emissions from coal-fired power plants represent 32% of the total anthropogenic mercury emissions in the United States (60 tons in 2012, 2000 tons worldwide). In recent years, public concern has increased due to the long-term irreversible effects of mercury on the environment and human health. As a result, the U.S. Environmental Protection Agency (EPA) proposed in December 2011 the Mercury and Air Toxics Standards (MATS); which require U.S. natural gas and coal-fired power plants to install air pollution control devices to prevent 91% of the Hg present in flue gas from being released. Currently, there are several air pollution control devices designed to reduce Hg emissions in power plants and whose working principles depend on the nature of the mercury species. Mercury is present in the flue gas in three forms: elemental (Hg0), oxidized (Hg+2) and particulate (HgP). Oxidized Hg is highly soluble in aqueous solutions, as compared to the insoluble and nonreactive Hg0, thus allowing for the removal of the former by conventional air pollution control devices. As a matter of fact, the promotion of Hg0 oxidation along the path of the flue gas from the boiler to the stack is currently the best approach to remove it by using current emission control technologies. The catalytic oxidation of mercury can be obtained through specific Hg oxidation catalysts such as noble metals or as a co-benefit of existing control technologies such as the Selective Catalyst Reduction (SCR) unit for NOx reduction. The latter option would be particularly attractive due to the associated low economic investment, since 40% of electricity from coal sources is produced in power plants that are already equipped with SCR units. However, little is known about the fate of mercury across the SCR unit, since most of the research work has been devoted to their applicability for NOx reduction. Understanding which are the key factors controlling the oxidation of mercury and developing a detailed mechanism of Hg oxidation across the SCR unit is a primary objective of this dissertation. One of the main achievements of this work has been the integration of an atomic-scale model with bench-scale experiments to identify key factors in Hg oxidation as a co-benefit of the SCR unit. Widely employed materials in commercial SCR catalysts include titania-supported vanadium and tungsten oxides, i.e., V2O5-WO3-TiO2, which were therefore investigated in this study. Theoretical models were used to assess the role of each component, namely, the support (TiO2), the active phase (V2O5) and the promoters (WO3), on the activity of this catalyst towards Hg oxidation. These include both density functional theory and ab-initio thermodynamic calculations; the latter are applied to investigate the effects of temperature and flue gas composition (which is coal dependent) on the reactivity of the catalyst under realistic operating conditions. Active phase, support and structural promoter were incorporated progressively into the analysis, thereby modeling the SCR catalyst with an increased level of complexity. The DFT results show that the active phase, V2O5, alone is not reactive under flue gas conditions and that the presence of the support leads to an increase of its reactivity toward Hg oxidation, presumably due to the higher dispersion of the vanadia phase on the TiO2 surface. Particular focus was given to the interaction of water with the supported system, due the significant concentration of water vapor present in the flue gas (≈ 10%). It is shown that water interacts with the surface in either a molecular or dissociative fashion, depending on the water coverage, which is in turn temperature-dependent. Interestingly, a stabilization effect is observed at low water coverages, as the latter tends to dissociate on the surface, thus yielding a reconstructed surface with attached hydroxyl groups. Moreover, a dehydration process is observed that takes place with increasing temperature and that leads to a water-free surface above 390 K. The analysis of the reactivity of the supported vanadium oxide catalyst was completed by a study of the adsorption energies of gas species that likely play a role in Hg oxidation (i.e., Hg, HgCl, HCl and H2O). Hereby, it was observed that surfaces with high water coverage show higher reactivity towards HgCl (the gas specie with the highest adsorption energy) followed by HCl. The adsorption energies of Hg suggested a negligible interaction with the vanadia dimer. Ab initio thermodynamic calculations were carried out to take into account the effect of temperature and entropy loss on the adsorption energies of these species; based on these results, a mechanism to explain Hg oxidation to HgCl2 was proposed, which involves the adsorption of HCl and HgCl, following a Langmuir-Hinshelwood mechanism. As a final step in the theoretical analysis, the incorporation of WO3 into the model shows that these ternary systems (V2O5-WO3-TiO2) are even more reactive than the binary systems (V2O5-TiO2). First, the effect of the surface coverage was studied by comparing the reactivity of the low- and high-loading binary systems. This analysis indicated enhanced reactivity of the SCR catalyst toward HgCl, HCl and Hg, with increasing loadings of the active phase. The effect of the surface composition on the reactivity of the catalyst was estimated by comparing the reactivity of the binary monolayer systems (i.e., 100% V2O5-TiO2 or 100%WO3-TiO2) against ternary systems (V2O5-WO3-TiO2 with different V2O5/WO3 ratios). This study showed a higher reactivity of the ternary system, with the 75%V2O5-25%WO3-TiO2 system representing the optimal catalyst composition toward Hg oxidation. The theoretical studies were complemented by Hg oxidation experiments carried out in a lab-scale packed-bed reactor with the purpose of benchmarking some of the predictions of the computational work. The effects of flue gas composition, catalyst formulation, temperature and space velocity on the Hg oxidation efficiency of different SCR catalysts were examined under typical flue gas conditions. The effect of the catalyst composition on the activity toward Hg oxidation was analyzed by testing four different SCR catalysts: 4%V2O5-10%WO3-TiO2, 1%V2O5-10%WO3-TiO2, 1%V2O5-TiO2 and 10%WO3-TiO2). It was shown that the binary systems have a lower activity compared to the ternary systems, thus supporting the predictions from first-principles calculations described above. Through the kinetic analysis, parameters such as reaction orders and the apparent activation energy were derived. By using the power law equation, it was found that O2 is zeroth-order and Hg is first-order in terms of the Hg oxidation rate. For the case of HCl, the reaction order could not be estimated using such a simple equation, and a more complex equation is necessary to capture the complexities of the heterogeneous reaction pathway. The activation energy takes a value of about 40 kJ/mol and is in reasonable agreement with data from the literature. It is worth pointing out that the intrinsic difficulty of measuring very low Hg concentration (≈ 5 ppb) results in large uncertainties associated with relevant parameters such as oxidation efficiencies and reaction rates.