Studies of the Fate of Sulfur Trioxide in Coal-Fired Utility Boilers Based on Modified Selected Condensation Methods YAN CAO,* ,† HONGCANG ZHOU, †,‡ WU JIANG, †,§ CHIEN-WEI CHEN, † AND WEI-PING PAN † Institute for Combustion Science and Environmental Technology (ICSET), Western Kentucky University (WKU), 2413 Nashville Road, Bowling Green, Kentucky 42101, School of Environmental Science and Engineering, Nanjing University of Information Science & Technology, Nanjing 210044, P.R. China, and School of Energy and Environmental Engineering, Shanghai University of Electric Power, Shanghai 200090, P. R. China Received December 3, 2009. Revised manuscript received March 26, 2010. Accepted March 29, 2010. The formation of sulfur trioxide (SO 3 ) in coal-ﬁred utility boilers can have negative effects on boiler performance and operation, such as fouling and corrosion of equipment, efﬁciency loss in the air preheater (APH), increase in stack opacity, and the formation of PM 2.5 . Sulfur trioxide can also compete with mercury when bonding with injected activated carbons. Tests in a lab-scale reactor conﬁrmed there are major interferences between ﬂy ash and SO 3 during SO 3 sampling. A modiﬁed SO 3 procedure to maximize the elimination of measurement biases, based on the inertial-ﬁlter-sampling and the selective-condensation-collecting of SO 3 , was applied in SO 3 tests in three full-scale utility boilers. For the two units burning bituminous coal, SO 3 levels starting at 20 to 25 ppmv at the inlet to the selective catalytic reduction (SCR), increased slightly across the SCR, owing to catalytic conversion of SO 2 to SO 3, and then declined in other air pollutant control device (APCD) modules downstream to approximately 5 ppmv and 15 ppmv at the two sites, respectively. In the unit burning sub-bituminous coal, the much lower initial concentration of SO 3 estimated to be approximately 1.5 ppmv at the inlet to the SCR was reduced to about 0.8 ppmv across the SCR and to about 0.3 ppmv at the exit of the wet ﬂue gas desulfurization (WFGD). The SO 3 removal efﬁciency across the WFGD scrubbers at the three sites was generally 35% or less. Reductions in SO 3 across either the APH or the dry electrostatic precipitator (ESP) in units burning high-sulfur bituminous coal were attributed to operating temperatures being below the dew point of SO 3 . 1. Introduction During combustion of sulfur-containing fuels, the majority of sulfur in these fuels is converted into sulfur dioxide (SO 2 ), with approximately 0.5-1.5% of the sulfur converted into sulfur trioxide (SO 3 ). Coals containing sulfur between 0.5-5 wt % can produce SO 3 concentrations of approximately ten parts per million by volume (ppmv) (1, 2). This is enough to create serious issues in coal-ﬁred utility boilers (1-4). The SO 3 can increase opacity in the stack by forming sulfuric acid droplets and sulfate aerosols, which are ﬁne particulate (PM 2.5 ) sources. Generally, the stack opacity can be seen as the color blue when the SO 3 concentration in the ﬂue gas is greater than 5 ppmv. Sulfur trioxide can also react with ammonia or chloride in the ﬂue gas to form submicrometer aerosols that create a white, opaque plume. The potential corrosion of utility equipment (e.g., down- stream ductwork, air preheater (APH), dry electrostatic precipitator (ESP), baghouse, and induced draft fan (I.D. fan)) is another important issue for boiler performance and operability. To avoid corrosion, the outlet temperature of the APH should be above the dew point of SO 3 , which will decrease the electricity-generation efﬁciency of the boiler (the electricity generation efﬁciency of the boiler is decreased by 1% when the APH outlet temperature is increased by 20 °C(3)). Recently, additional applications of SCR for NO x emission control and mercury oxidation have begun to focus on the SO 3 issue since SCR catalysts may be responsible for additional SO 2 conversion to SO 3 in the ﬂue gas (5, 6). The SCR temperature should be low enough to eliminate this potential SO 3 conversion. This would mean losing part of the capability of the SCR catalyst on NO x removal efﬁciency. On the other hand, the more slipped NH 3 that leaves the SCR, the greater the likelihood that it will react with SO 3 to form ammonia bisulfate (ABS). It is a very sticky and heavy material, which could plug up downstream at the APH and baghouse. Activated-carbon injection (ACI) is currently the prevailing technology for mercury-emission control (7). Available SO 3 in the ﬂue gas will deteriorate the adsorbent performance on mercury adsorption because of its competi- tive adsorption on the active sites of mercury adsorbents (8-11). Thus, the many and varied issues caused by SO 3 in and out of the utility boiler have produced the need for the effective control of SO 3 in these boilers (12-16). The potential value of effective SO 3 control can exceed $2-3 million per year for a 500 MW unit (3). However, the beneﬁts are likely to be deducted by the additional costs for the application of SO 3 abatement technologies, which depend on SO 3 con- centrations in the ﬂue gas. Thus, accurate, repeatable, and veriﬁable SO 3 sampling and analysis methods are very important. However, this is not the case for currently available SO 3 sampling methods because of the occurrence of potential interference from ﬂy ash. This paper presents a detailed investigation of quality-control issues related to SO 3 mea- surements in coal-ﬁred utility boilers with conventional air pollution control devices (APCDs) in three coal-ﬁred power plants. 2. Experimental Section SO 3 Sampling Methods. There are three standard SO 3 sample-collection methods (EPA Method 17, Method 8A, and ASTM 3226-73T; details on these methods have been presented in the Supporting Information). These methods are based on either chemical absorption (EPA Method 17) using 80% isopropyl alcohol (IPA) as a trapping solution or the selective condensation of SO 3 from gas streams laden with SO 2 by controlling condensation temperatures (Method 8A and ASTM 3226-73T). Currently, the selective condensa- tion method accurately separates SO 3 and SO 2 in coal-ﬁred ﬂue gas. However, this is not the case for EPA Method 17 * Corresponding author e-mail: yan.cao@wku.edu. † Western Kentucky University. ‡ Nanjing University of Information Science & Technology. § Shanghai University of Electric Power. Environ. Sci. Technol. 2010, 44, 3429–3434 10.1021/es903661b  2010 American Chemical Society VOL. 44, NO. 9, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3429 Published on Web 04/09/2010