UC Irvine

Atmospheric emission of sulfuric acid (H2SO4) and its precursors, sulfate oxides SO2 and SO3, are generated by burning fuels such as coal, oil, and gas for energy production or in the metallurgical industry. While sulfuric acid is important in many industrial applications, in the atmosphere it is one of the culprits for acid rain formation. Steps have been taken to reduce the sulfur content of fuels (for example, the migration to low sulfur fuel in the marine industry), but sulfur emissions are still a significant problem, especially for coal-fired power-plants. In the atmosphere, sulfate constitutes a large fraction of particulate matter air pollutants with aerodynamic diameter <2.5 microns (PM 2.5), which has been found responsible for adverse health effects. Although the use of coal is sharply declining in theUnited States, it is still on the rise in China and in Southeast Asia. In coal-fired power-plants, most of the fuel sulfur content is oxidized to sulfur dioxide, while a small percentage can be further oxidized to sulfur trioxide (SO3) in the boiler and across the NOx reduction catalyst. In the presence of water, sulfur trioxide reacts at low temperatures to form sulfuric acid. The SO3/H2SO4 emissions can be reduced by injecting neutralizing sorbents, entailing an additional operating cost. With continuous monitoring, a closed-loop control for sorbent injection could be developed, reducing sulfate emissions, and improving the operational performance of power plants. Therefore, the continuous monitoring of SO3, H2SO4, and SO2 is a primary need for the coal industry. The goal of this dissertation is to evaluate the feasibility of continuous measurements in flue gas conditions using optical absorption methods, including providing a high temperature spectral database of SO2, SO3, and H2SO4. A further goal is the experimental investigation of the relationship between SO3 and H2SO4, often assumed to be at chemical equilibrium. The equilibrium has been hypothesized and relied on in many measurement techniques, however a direct experimental verification has not been accomplished. Specifically, this work examines the performance of Differential Optical Absorption Spectroscopy (DOAS) and External Cavity Quantum Cascade Laser (EC-QCL) as potential methods for continuous measurement. The results are validated using the Environmental Protection Agency (EPA) Method 8A, the current industry standard for the measurement of sulfur species. The experimental setup includes a unique high-temperature multi-pass optical cell and a flow system able to replicate appropriate levels of flue gas species. The gas temperature and spatial distribution of sulfur species in the measurement cell was also investigated using numerical simulations. The DOAS technique performed well for SO2 measurements in the UV, while SO3 and H2SO4 could not be evaluated with accuracy. The DOAS technique was found unsuitable for continuous sulfur species monitoring in flue gas conditions. The mid-IR ECQCL technique could successfully detect SO2 in flue gas conditions, in the spectral ranges near both 7um and 8um, and for the first time, high-temperature mid-IR spectral libraries for SO3 and H2SO4 were recorded. The equilibrium between SO3 and H2SO4 was verified through direct optical measurements at 3% water vapor concentration within +-12% uncertainty. At 7um, while SO3 measurements are successful, the strong absorbance of SO2 limits measurements only to low-sulfur coals (with SO2 concentrations <500ppm) in the current multi-pass configuration. Sulfuric acid does not absorb at 7um. At 8um, SO2 is a weak absorber, hence high-sulfur coal emission monitoring is feasible. Sulfuric acid is a strong absorber and can be measured even at low levels, while SO3 does not absorb at 8um and needs to be calculated from equilibrium. Unfortunately, however, sulfuric acid reacts with the windows material (BaF2) to form a layer of BaSO4, and at this wavelength, the layer formation causes a significant performance degradation, as BaSO4 absorbs strongly. As the BaSO4 layer thickens, it prevents the laser light from reaching the detector, and this poses a significant engineering challenge and limits the measurement to short time frames (10 to 30 minutes).