Journal of Catalysis 181, 233–243 (1999) Article ID jcat.1998.2305, available online at http://www.idealibrary.com on Oxidation of SO2 over Supported Metal Oxide Catalysts Joseph P. Dunn,1 Harvey G. Stenger, Jr., and Israel E. Wachs2 Zettlemoyer Center for Surface Studies, Department of Chemical Engineering, Lehigh University, Bethlehem, Pennsylvania 18015 Received June 19, 1998; revised October 10, 1998; accepted October 19, 1998 INTRODUCTION A systematic catalytic investigation of the sulfur dioxide oxida- tion reactivity of several binary (MxOy/TiO2) and ternary (V2O5/ Sulfur dioxide (SO2) is formed from both the oxidation of MxOy/TiO2) supported metal oxide catalysts was conducted. sulfur contained in fossil fuels and the industrial processes Raman spectroscopy characterization of the supported metal oxide that treat and produce sulfur-containing compounds. The catalysts revealed that the metal oxide components were essentially catalytic oxidation of sulfur dioxide appears in numerous 100% dispersed as surface metal oxide species. Isolated fourfold co- industrial processes and has a significant environmental im- ordinated metal oxide surface species are present for most oxides pact because of the associated sulfur oxide (SOx) emissions. tested at low coverages, whereas at surface coverages approaching Approximately two-thirds of the 50 billion pounds of sulfur monolayer polymerized surface metal oxide species with sixfold co- oxides released annually in the United States are emitted ordination are present for some of the oxides. The sulfur dioxide from coal fired power plants. Industrial fuel combustion and oxidation turnover frequencies (SO2 molecules converted per sur- face redox site per second) of the binary catalysts were all within industrial processes (primarily sulfuric acid manufacture, an order of magnitude (V2O5/TiO2 >Fe2O3/TiO2 >Re2O7/TiO2 petroleum refining, and smelting of nonferrous metals) ac- CrO3/TiO2 Nb2O5/TiO2 >MoO3/TiO2 WO3/TiO2). An excep- count for the remainder of the emissions (1). tion was the K2O/TiO2 catalyst system, which is inactive for sulfur The oxidation of sulfur dioxide to sulfur trioxide is un- dioxide oxidation under the chosen reaction conditions. With the desirable during the selective catalytic reduction (SCR) of exception of K2O, all of the surface metal oxide species present in nitrogen oxides (NOx) found in the flue gas of power plants. the ternary catalysts (i.e., oxides of V,Fe, Re, Cr,Nb, Mo, and W) can SCR removes NOx in the flue gas by reacting the nitrogen undergo redox cycles and oxidize sulfur dioxide to sulfur trioxide. oxides with ammonia and oxygen to form nitrogen and wa- The turnover frequency for SO2 oxidation over all of these catalysts ter at approximately 370C over titania supported vanadia is approximately the same at both low and high surface coverages, catalysts, e.g., V O /WO –MoO /TiO (2). Under typical despite structural differences in the surface metal oxide overlayers. 2 5 3 3 2 This indicates that the mechanism of sulfur dioxide oxidation is not SCR design and operating conditions, NOx reduction ef- sensitive to the coordination of the surface metal oxide species. A ficiency is directly proportional to the NH3 :NOx ratio up comparison of the activities of the ternary catalysts with the corre- to NOx reduction levels of about 80%. Ammonia readily sponding binary catalysts suggests that the surface vanadium oxide combines with sulfur trioxide at temperatures below 250 C and the additive surface oxide redox sites act independently without to form ammonium sulfates, which can block the catalyst’s synergistic interactions: the sum of the individual activities of the pores and foul downstream heat exchangers. This problem binary catalysts quantitatively correspond to the activity of the cor- is so serious that industrial specifications for SCR processes responding ternary catalyst. The V2O5/K2O/TiO2 catalyst showed include upper limits for the outlet concentration of sulfur a dramatic reduction in catalytic activity in comparison to the un- trioxide corresponding to approximately 1 to 2% sulfur promoted V2O5/TiO2 catalyst. The ability of potassium oxide to sig- dioxide conversion. Commercial SCR catalysts tradition- nificantly retard the redox potential of the surface vanadia species ally have low vanadia loadings due to vanadia’s propensity is primarily responsible for the lower catalytic reactivity. c 1999 to catalyze sulfur dioxide oxidation. Additives that can ef- Academic Press Key Words: sulfur dioxide; sulfur trioxide; oxidation; metal oxide; ficiently promote the SCR reaction without significantly in- SCR; Raman; catalyst. creasing the rate of SO2 oxidation would allow lower SCR operating temperatures without the worry of ammonium sulfate production and deposition. In an effort to develop a catalyst for the selective catalytic reduction of nitric oxide with ammonia with minimal sulfur dioxide oxidation activity, Morikawa et al. (3) tested several 1 Current address: BOC Gases Technology, 100 Mountain Avenue, modifications of vanadia/titania catalysts for sulfur dioxide Murray Hill, NJ 07974. 2 To whom correspondence should be addressed. Fax: (610) 758-6555. oxidation and SCR activity (e.g., V2O5/Mx Oy/TiO2, where E-mail: [email protected]. Mx Oy = GeO2, ZnO, MoO3,orWO3). Catalysts containing 233 0021-9517/99 $30.00 Copyright c 1999 by Academic Press All rights of reproduction in any form reserved. 234 DUNN, STENGER, AND WACHS tungsten oxide or molybdenum trioxide showed an increase TABLE 1 in the activity of sulfur dioxide oxidation, while catalysts Composition of Catalysts Studied containing germanium oxide or zinc oxide showed a dras- tic decrease in sulfur dioxide oxidation activity. Similarly, VVMM Sazonova et al. (4) tested the performance of V2O5/TiO2 (atoms/ (surface (atoms/ (surface catalysts doped by tungsten oxide and niobium oxide for Catalyst nm2) coverage) nm2) coverage) sulfur dioxide oxidation and selective catalytic reduction TiO2 ———— of nitric oxide by ammonia. In contrast to the results of 1% V2O5/TiO2 1.3 0.17 — — Morikawa et al., the data suggest that tungsten oxide sub- 1% Fe2O3/TiO2 — — 0.6 0.2 stantially decreases the oxidation of sulfur dioxide to sulfur 1% Re2O7/TiO2 — — 0.4 0.2 trioxide. Niobium doped catalysts also exhibited a slight 1% CrO3/TiO2 — — 1.0 0.2 1% Nb O /TiO — — 0.8 0.1 decrease in oxidation activity. Neither of these studies pro- 2 5 2 1% MoO3/TiO2 — — 0.7 0.1 vided molecular structural information on the surface ar- 1% WO3/TiO2 — — 0.5 0.1 rangement of the catalysts and, consequently, a funda- 1% K2O/TiO2 — — 2.4 1 mental explanation of the observed results could not be 5% V2O5/TiO2 6.5 0.83 — — provided. 5% Fe2O3/TiO2 — — 3.0 0.8 5% Re O /TiO — — 2.0 0.8 The objective of this investigation is to combine 2 7 2 5% CrO3/TiO2 — — 5.3 0.8 molecular structural information on several binary (e.g., 5% Nb2O5/TiO2 — — 4.1 0.7 V2O5/TiO2,Re2O7/TiO2, CrO3/TiO2,Fe2O3/TiO2,Nb2O5/ 5% MoO3/TiO2 — — 3.3 0.7 TiO2, MoO3/TiO2,WO3/TiO2, and K2O/TiO2) and ternary 7% WO3/TiO2 — — 3.3 0.8 1% V2O5/5% Fe2O3/TiO2 1.3 0.17 3.0 0.8 (e.g., V2O5/Re2O7/TiO2,V2O5/CrO3/TiO2,V2O5/Fe2O3/ 1% V2O5/5% Re2O7/TiO2 1.3 0.17 2.0 0.8 TiO2,V2O5/Nb2O5/TiO2,V2O5/MoO3/TiO2,V2O5/WO3/ 1% V2O5/5% CrO3/TiO2 1.3 0.17 5.3 0.8 TiO2, and V2O5/K2O/TiO2) supported metal oxide cata- 1% V2O5/5% Nb2O5/TiO2 1.3 0.17 4.1 0.7 lysts with the corresponding sulfur dioxide oxidation reac- 1% V2O5/5% MoO3/TiO2 1.3 0.17 3.3 0.7 tivities in order to assist in the molecular engineering of 1% V2O5/7% WO3/TiO2 1.3 0.17 3.3 0.8 SCR catalysts with low sulfur dioxide oxidation activities. 1% V2O5/1% K2O/TiO2 1.3 0.17 2.4 1 This study does not investigate the abilities of these cata- lysts to efficiently promote the SCR of NO with NH , since x 3 Catalysts were prepared on a TiO support (Degussa a recent comprehensive study by Amiridis et al. (5) has ad- 2 P-25, 55 m2/g), which was washed with distilled water and dressed this issue. isopropanol (Fisher-certified ACS, 99.9% pure), dried at 120C for 4 h, ramped at 5C/min from 120 to 450C, METHODS and calcined at 450C for 2 h prior to impregnation. The V O /TiO catalysts were prepared by incipient wetness Preparation of Catalysts 2 5 2 impregnation. Vanadium triisopropoxide was used as the The oxidation catalysts used in this research program vanadium precursor. The air and moisture sensitive nature were supported metal oxide catalysts possessing two- of the vanadium alkoxide precursor required the prepara- dimensional metal oxide overlayers on a high surface area tion of vanadia catalysts to be performed under a nitro- titanium dioxide support. The two-dimensional metal oxide gen environment using nonaqueous solutions. Solutions of overlayers are confirmed by Raman spectroscopy and pre- known amounts of vanadium triisopropoxide (Alfa) and sented under Results. As shown in Table 1, the kinetic stud- isopropanol, corresponding to incipient wetness impregna- ies employed both low surface coverage (0.1 to 0.2 mono- tion volume and the final amount of vanadia required, were layer) and high surface coverage (0.7 to 0.9 monolayer) of prepared in a glove box and dried at room temperature for supported vanadium oxide, rhenium oxide, chromium ox- 16 h. The impregnated samples were subsequently heated ide, iron oxide, niobium oxide, molybdenum oxide, tung- at 120C (2 h) and 300C (2 h) in flowing nitrogen (Linde, sten oxide, and potassium oxide catalysts. The fractional 99.995% pure). The final calcination was performed in oxy- surface coverages given in Table 1 are calculated based gen at 450Cfor2h.A5C/min temperature ramp was used on the weight percent of the specific supported metal ox- between each stage of heating.