Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A novel fluidized bed reactor for integrated NOx adsorption-reduction with hydrocarbons Yang, Terris Tianxue 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2008_fall_yang_terris.pdf [ 13.3MB ]
Metadata
JSON: 24-1.0058550.json
JSON-LD: 24-1.0058550-ld.json
RDF/XML (Pretty): 24-1.0058550-rdf.xml
RDF/JSON: 24-1.0058550-rdf.json
Turtle: 24-1.0058550-turtle.txt
N-Triples: 24-1.0058550-rdf-ntriples.txt
Original Record: 24-1.0058550-source.json
Full Text
24-1.0058550-fulltext.txt
Citation
24-1.0058550.ris

Full Text

  A NOVEL FLUIDIZED BED REACTOR FOR INTEGRATED NOX ADSORPTION-REDUCTION WITH HYDROCARBONS    by  TERRIS TIANXUE YANG  B.A.Sc., Tsinghua University, 1989 M.A.Sc., Tsinghua University, 1991     A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY   in   The Faculty of Graduate Studies  (Chemical and Biological Engineering)     THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)   September 2008                                                     © Terris Tianxue Yang, 2008  ii  Abstract An integrated NOx adsorption-reduction process has been proposed in this study for the treatment of flue gases under lean-burn conditions by decoupling the adsorption and reduction into two different zones. The hypothesis has then been validated in a novel internal circulating fluidized bed. The adsorption and reaction performance of Fe/ZSM-5 for the selective catalytic reduction (SCR) of NOx with propylene was investigated in a fixed bed reactor. The fine Fe/ZSM-5(Albemarle) catalyst showed reasonable NOx adsorption capacity, and the adsorption performance of the catalyst was closely related to the particle size and other catalyst properties. Fe/ZSM-5 catalyst was sensitive to the reaction temperature and space velocity, and exhibited acceptable activity when O2 concentration was controlled at a low level. Water in the flue gas was found to slightly enhance the reactivity of Fe/ZSM- 5(Albemarle), while the presence of CO2 showed little effect. SO2 severely inhibited the reactivity of Fe/ZSM-5(Albemarle), and the deactivated catalyst could be only partially regenerated. Configurations of the reactor influenced the hydrodynamic performance significantly in a cold model internal circulating fluidized bed (ICFB) reactor. For all configurations investigated, the high gas bypass ratio from the annulus to draft tube (RAD) and low draft tube to annulus gas bypass ratio (RDA) were observed, with the highest RDA associated with the conical distributor which showed the flexible and stable operation over a wide range of gas velocities. Solids circulation rates increased with the increase of gas velocities both in the annulus and the draft tube. Gas bypass was also studied in a hot model ICFB reactor. The Abstract   iii results showed that the orientation of perforated holes on the conical distributor could be adjusted to reduce RAD and/or enhance RDA. Coarse Fe/ZSM-5(PUC) and fine Fe/ZSM-5(Albemarle) catalysts were used in an ICFB and a conventional bubbling fluidized bed to test the NOx reduction performance. Coarse Fe/ZSM-5(PUC) catalyst showed poor catalytic activity, while fine Fe/ZSM- 5(Albemarle) catalyst exhibited promising NOx reduction performance and strong inhibiting ability to the negative impact of excessive O2 in the ICFB reactor, proving that the adsorption-reduction two-zone reactor is effective for the NOx removal from oxygen-rich combustion flue gases.  iv Table of Contents Abstract ...................................................................................................................... ii Table of Contents ..................................................................................................... iv List of Tables ............................................................................................................ ix List of Figures............................................................................................................ x Acronyms ..............................................................................................................xxiii Nomenclature......................................................................................................... xxv Acknowledgements.............................................................................................xxviii 1 Introduction ............................................................................................................ 1 1.1 Formation of NOx.........................................................................................................1 1.2 Methods for the abatement of NOx emissions .............................................................2 1.3 Selective catalytic reduction of NOx with NH3............................................................4 1.4 Research objectives......................................................................................................6 1.5 Thesis layout ................................................................................................................7 2 Literature Review................................................................................................... 9 2.1 Catalysts .......................................................................................................................9 2.1.1 Metal-exchanged zeolites ..................................................................................11 2.1.2 Metal-exchanged metal oxides ..........................................................................14 2.2 Hydrocarbon reducing agents ....................................................................................15 2.3 Reaction mechanisms and kinetics of HC-SCR.........................................................21 2.3.1 Surface redox mechanism .................................................................................21 2.3.2 Oxidation of NO to NO2....................................................................................22 2.3.3 Partial oxidation of hydrocarbons .....................................................................22 2.4 Influence of flue gas compositions on HC-SCR........................................................23 Table of Contents  v 2.4.1 Influence of SO2 and H2O in the flue gas..........................................................23 2.4.2 Influence of excess oxygen in the flue gas........................................................25 2.5 Reactors for SCR of NOx ...........................................................................................28 2.6 NOx storage-reduction (NSR) ....................................................................................30 2.7 Summary ....................................................................................................................31 3 Adsorption and Reaction Kinetics...................................................................... 33 3.1 Experimental setup.....................................................................................................33 3.2 Catalyst preparation ...................................................................................................35 3.2.1 Materials ............................................................................................................36 3.2.2 Preparation of NH4/ZSM-5 and H/ZSM-5 ........................................................37 3.2.3 Preparation of Fe/ZSM-5 by WIE method ........................................................38 3.2.4 Preparation of Fe/ZSM-5 by IMPO method......................................................38 3.3 Performance of Fe/ZSM-5 (PUC) catalyst prepared by IMPO..................................38 3.3.1 Adsorption performance of Fe/ZSM-5(PUC) ...................................................38 3.3.2 Reaction performance of Fe/ZSM-5(PUC) .......................................................42 3.3.2.1 Effect of reaction temperature on catalytic activity.............................43 3.3.2.2 Catalyst deactivation at low temperatures ...........................................49 3.3.2.3 Effect of GHSV ...................................................................................56 3.3.2.4 Effect of O2 and HC concentrations ....................................................58 3.4 Reaction performance of Fe/ZSM-5 (PUC) catalyst prepared by WIE .....................66 3.5 Performance of Fe/ZSM-5(Albemarle) catalyst.........................................................69 3.5.1 Adsorption performance of Fe/ZSM-5(Albemarle) ..........................................70 3.5.2 Effect of O2 concentration on NOx adsorption of Fe/ZSM-5(Albemarle) ........73 3.5.3 Effect of H2O and CO2 on NOx adsorption of Fe/ZSM-5(Albemarle)..............74 3.5.4 Adsorption performance of parent H/ZSM-5(Albemarle) ................................76 3.5.5 Catalytic activity of parent H/ZSM-5(Albemarle) in HC-SCR.........................77 3.5.6 Reaction performance of Fe/ZSM-5(Albemarle) ..............................................79 3.5.6.1 Effect of reaction temperature .............................................................79 3.5.6.2 Effect of GHSV ...................................................................................84 3.5.6.3 Effect of O2 and HC concentrations ....................................................86 3.5.6.4 Effect of CO2 and H2O.........................................................................89 Table of Contents  vi 3.5.6.5 Deactivation of Fe/ZSM-5(Albemarle) catalyst by SO2......................94 3.6 Summary ..................................................................................................................103 4 Hydrodynamic Study of the ICFB Reactor ..................................................... 105 4.1 Experimental setup...................................................................................................106 4.1.1 Estimation of gas bypass ratio.........................................................................111 4.1.2 Estimation of solids circulation rate ................................................................113 4.2 Performance with the flat distributor plate...............................................................114 4.2.1 Gas bypass .......................................................................................................116 4.2.2 Solids circulation rate ......................................................................................120 4.3 Performance with the cylindrical distributor plate...................................................122 4.3.1 Gas bypass .......................................................................................................123 4.3.2 Solids circulation rate ......................................................................................129 4.4 Performance with the conical distributor plate ........................................................131 4.4.1 Gas bypass .......................................................................................................132 4.4.2 Solids circulation rate ......................................................................................136 4.5 Gas bypassing in the hot model ICFB reactor .........................................................140 4.5.1 Fe/ZSM-5(PUC) catalyst.................................................................................141 4.5.2 Fe/ZSM-5(Albemarle) catalyst........................................................................144 4.5.3 Prediction of O2 concentration in the draft tube and annulus..........................146 4.6 Summary ..................................................................................................................148 5 Adsorption and Reduction Performance of the ICFB Reactor ..................... 151 5.1 Experimental setup...................................................................................................151 5.2 Estimation of NOx and HC conversions and NOx adsorption ratio .........................154 5.3 Performance with Fe/ZSM-5(PUC) catalyst............................................................158 5.3.1 Selection of adsorption zone in the ICFB reactor ...........................................158 5.3.2 Effect of catalyst loading in the ICFB reactor.................................................159 5.3.3 Effect of HC:NO molar ratio on NOx conversion ...........................................161 5.3.4 Effect of flue gas O2 content on NOx conversion............................................162 5.3.5 Effect of gas velocities on NOx conversion and adsorption ............................163 5.3.6 Performance of fluidized bed reactor with Fe/ZSM-5(PUC) ..........................165 Table of Contents  vii 5.3.6.1 Effect of inlet NO concentration........................................................165 5.3.6.2 Effect of HC:NO ratio on NOx and HC conversions .........................167 5.3.6.3 Effect of inlet O2 concentration on NOx and HC conversions...........172 5.3.7 Comparison of ICFB reactor and fluidized bed reactor ..................................174 5.4 Performance of Fe/ZSM-5(Albemarle) catalyst.......................................................175 5.4.1 Effect of HC:NO molar ratio on NOx and HC conversions ............................175 5.4.2 Effect of flue gas O2 concentration on NOx conversion and adsorption .........179 5.4.3 Effect of gas velocities on NOx conversion and adsorption ............................182 5.4.4 Performance of the fluidized bed reactor with Fe/ZSM-5(Albemarle) ...........185 5.4.4.1 Effect of HC:NO ratio on NOx and HC conversions .........................185 5.4.4.2 Effect of O2 concentration on NOx and HC conversions...................188 5.4.5 Comparison of reactor types............................................................................190 5.5 Summary ..................................................................................................................191 6 Conclusions and Recommendations for Future Work ................................... 194 6.1 Conclusions..............................................................................................................194 6.1.1 Adsorption and reaction performance of Fe/ZSM-5 catalyst ..........................194 6.1.2 Hydrodynamic study of the ICFB reactor .......................................................195 6.1.3 Adsorption and reduction performance of Fe/ZSM-5 in the ICFB reactor .....196 6.2 Recommendations for future work ..........................................................................197 References .............................................................................................................. 200 Appendices A Summary of Previous Work............................................................................. 219 B Calibration ......................................................................................................... 236 B.1 Gas flow meters.......................................................................................................236 B.2 Peristaltic water pump .............................................................................................239 B.3 Pressure transducers ................................................................................................240 C BET Surface Area ............................................................................................. 243 Table of Contents  viii D Particle Size Distribution.................................................................................. 244 E Adsorption and Reaction Kinetics of Fe/ZSM-5 (Crushed PUC) Catalyst.. 247 E.1 Adsorption performance of Fe/ZSM-5(crushed PUC) ............................................247 E.2 Reaction performance of Fe/ZSM-5(crushed PUC) ................................................251 F HC-SCR Performance of FCC and Fe/FCC Catalyst.................................... 254 G Some Results from the Fixed Bed Experiment............................................... 255 G.1 Adsorption curves for Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle)......................255 G.2 Profiles of NOx and HC conversions and outlet CO concentration from time-on- stream test ...............................................................................................................261 G.3 XPS analysis for Fe/ZSM-5(PUC)..........................................................................269 H Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor......... 273 H.1 Fe/ZSM-5(PUC)......................................................................................................273 H.2 Fe/ZSM-5(Albemarle).............................................................................................277  ix  List of Tables  Table 1.1 Typical compositions of flue gases from thermal power plant stacks ......................2 Table 3.1 Properties of the catalyst supports ..........................................................................37 Table 4.1 Dimensions of the cold model ICFB unit .............................................................108 Table 5.1 Geometric dimensions of the ICFB reactor ..........................................................153 Table 5.2 Properties of catalysts used in the hot model experiments ...................................154 Table A.1 HC-SCR catalysts supported by zeolites..............................................................219 Table A.2 HC-SCR catalysts supported by non-zeolites (Metal oxides)..............................220 Table A.3 Specifications of some typical HC-SCR catalysts ...............................................221 Table A.4 HC-SCR with methane as reducing agent............................................................222 Table A.5 HC-SCR with propane as reducing agent ............................................................224 Table A.6 HC-SCR with propylene as reducing agent .........................................................226 Table A.7 HC-SCR with iso-C4H10 (iso-butane) as reducing agent .....................................229 Table A.8 HC-SCR with other hydrocarbons as reducing agent ..........................................232 Table A.9 HC-SCR with oxygenated hydrocarbons as reducing agent ................................234 Table A.10 NSR with reducing agent ...................................................................................235 Table C.1 BET surface areas for catalysts used in this work................................................243    x  List of Figures  Figure 1.1 NOx control effectiveness for coal-fired tangential boilers .....................................3 Figure 1.2 Schematic of the integrated adsorption-reduction process of NOx .........................7 Figure 2.1 NO conversion to N2 over Cu/ZSM-5 catalysts as a function of Cu/Al ratio at different temperatures..........................................................................................11 Figure 2.2 Dependence of NO conversion on temperature for SCR of NO over various catalysts with alternative reducing agent.............................................................17 Figure 2.3 Temperature dependence of N2 yield over Fe/ZSM-5 with different hydrocarbons .............................................................................................................................18 Figure 2.4 Effect of the nature of the hydrocarbon on NO reduction selectivity on Co/ZSM-5 .............................................................................................................................19 Figure 2.5 NO conversion to N2 on Ag/Al2O3 using alternative reducing agent....................20 Figure 2.6 Effect of H2O on NO conversion to N2 over Fe/ZSM-5 or Cu/ZSM-5.................24 Figure 2.7 Influence of O2 concentration on NO conversion to N2 ........................................27 Figure 3.1 The fixed bed reaction system...............................................................................33 Figure 3.2 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=250oC) ................................................................................39 Figure 3.3 Fitted adsorption isotherms of NOx by Freundlich equation (Catalyst: Fe/ZSM- 5(PUC)) ...............................................................................................................41 Figure 3.4 Relationship between α/ and adsorption temperature (Catalyst: Fe/ZSM-5(PUC)) .............................................................................................................................41 Figure 3.5 Profiles of NOx and HC conversions and outlet CO concentration (Catalyst: Fe/ZSM-5(PUC), T=250oC, [O2]=1%)................................................................44 List of Figures  xi Figure 3.6 Profiles of NOx and HC conversions and outlet CO concentration (Catalyst: Fe/ZSM-5(PUC), T=350oC, [O2]=1%)................................................................45 Figure 3.7 Effect of reaction temperature on NOx conversion (Catalyst: Fe/ZSM-5(PUC)) .47 Figure 3.8 Effect of reaction temperature on HC conversion (Catalyst: Fe/ZSM-5(PUC)) ...47 Figure 3.9 Effect of inlet O2 concentration on the catalytic activity (Catalyst: Fe/ZSM- 5(PUC)) ...............................................................................................................50 Figure 3.10 Comparison of BET surface areas (Catalyst: Fe/ZSM-5 (PUC)) ........................51 Figure 3.11 XPS narrow scan for C 1s (Catalyst: Fe/ZSM-5(PUC)) .....................................52 Figure 3.12 TGA result for the catalyst aged at 275oC...........................................................54 Figure 3.13 TGA result for the catalyst aged at 350oC...........................................................54 Figure 3.14 Recovery of catalyst activity under different regeneration conditions (Catalyst: Fe/ZSM-5(PUC)).................................................................................................55 Figure 3.15 Effect of GHSV on catalytic activity (Catalyst: Fe/ZSM-5(PUC), T=325oC)....56 Figure 3.16 Effect of GHSV on catalytic activity (Catalyst: Fe/ZSM-5(PUC), T=350oC)....57 Figure 3.17 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=275oC, 95% confidence level) ..........................................................59 Figure 3.18 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=300oC)...............................................................................................61 Figure 3.19 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=325oC)...............................................................................................62 Figure 3.20 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=350oC)...............................................................................................63 Figure 3.21 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=375oC)...............................................................................................64 List of Figures  xii Figure 3.22 Effect of temperature on catalytic activity for NOx reduction (Catalyst: Fe/ZSM- 5(PUC, WIE))......................................................................................................66 Figure 3.23 Effect of temperature on NOx and HC conversion (Catalyst: Fe/ZSM-5(PUC, WIE)) ...................................................................................................................68 Figure 3.24 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=275oC).......................................................................70 Figure 3.25 Fitting of adsorption isotherms of NOx by Freundlich equation (Catalyst: Fe/ZSM-5(Albemarle))........................................................................................71 Figure 3.26 Relationship between α/ and adsorption temperature (Catalyst: Fe/ZSM- 5(Albemarle)) ......................................................................................................72 Figure 3.27 Effect of O2 concentration in the flue gas on the adsorption capacity of Fe/ZSM- 5(Albemarle)) catalyst .........................................................................................74 Figure 3.28 Effect of H2O and CO2 in the flue gas on adsorption isotherms (Catalyst: Fe/ZSM-5(Albemarle))........................................................................................75 Figure 3.29 Comparison of adsorption isotherm between Fe/ZSM-5(Albemarle) and parent H/ZSM-5(Albemarle) at T=350oC ......................................................................76 Figure 3.30 Catalytic activity of parent H/ZSM-5(Albemarle) ..............................................78 Figure 3.31 Profiles of NOx and HC conversions and outlet CO concentration (Fe/ZSM- 5(Albemarle), T=275oC, [O2]=1%) .....................................................................80 Figure 3.32 Profiles of NOx and HC conversions and outlet CO concentration (Fe/ZSM- 5(Albemarle), T=350oC, [O2]=1%) .....................................................................81 Figure 3.33 Effect of temperature on the catalytic activity of Fe/ZSM-5(Albemarle) ...........83 Figure 3.34 Effect of GHSV on the catalytic activity of Fe/ZSM-5(Albemarle) ...................85 Figure 3.35 Effect of O2 and HC concentrations (Fe/ZSM-5(Albemarle), T=350oC)............87 List of Figures  xiii Figure 3.36 Effect of O2 and HC concentrations (Fe/ZSM-5(Albemarle), T=375oC)............88 Figure 3.37 Effect of CO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (T=350oC).....90 Figure 3.38 Effect of H2O on the catalytic activity of Fe/ZSM-5(Albemarle) (T=350oC).....91 Figure 3.39 Time-on-stream test on the effect of H2O (Fe/ZSM-5(Albemarle), T=350oC)...92 Figure 3.40 Combined effect of H2O and CO2 on the catalytic activity of Fe/ZSM- 5(Albemarle) (T=350oC) .....................................................................................93 Figure 3.41 Effect of the addition of H2O on outlet CO and CO2 concentrations (Fe/ZSM- 5(Albemarle), T=350oC)......................................................................................94 Figure 3.42 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #1) ................................................................................................................96 Figure 3.43 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #2) ................................................................................................................97 Figure 3.44 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #3) ................................................................................................................98 Figure 3.45 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #4) ................................................................................................................98 Figure 3.46 Comparison of the four consecutive runs with 200ppm SO2 in the flue gas (Catalyst: Fe/ZSM-5(Albemarle)) .......................................................................99 Figure 3.47 Comparison of fresh and regenerated catalysts over four consecutive runs (Catalyst: Fe/ZSM-5(Albemarle)) .....................................................................100 Figure 3.48 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (30ppm SO2)101 Figure 3.49 Comparison of BET surface areas for fresh and deactivated catalyst (Fe/ZSM- 5(Albemarle)) ....................................................................................................102 Figure 4.1 Schematic of the cold model ICFB reactor .........................................................107 List of Figures  xiv Figure 4.2 Schematics of the cold model ICFB system........................................................110 Figure 4.3 Schematic of the mass transfer between draft tube and annulus .........................111 Figure 4.4 Configurations of the flat distributor for the annulus gas flow ...........................115 Figure 4.5 Effect of gas velocities on gas bypass (Flat distributor, HG=10 mm) .................118 Figure 4.6 Effect of gas velocities on gas bypass (Flat distributor, HG=6 mm) ...................119 Figure 4.7 Effect of gas velocities on solids circulation rate (Flat distributor, HG=10 mm) 121 Figure 4.8 Effect of gas velocities on solids circulation rate (Flat distributor, HG=6 mm) ..121 Figure 4.9 Configurations of the cylindrical distributor for the annulus gas flow ...............123 Figure 4.10 Effect of gas velocities on gas bypass (Cylindrical distributor, HG1=10 mm) ..125 Figure 4.11 Effect of gas velocities on gas bypass (Cylindrical distributor, HG1=20 mm) ..126 Figure 4.12 Effect of gas velocities on gas bypass (Cylindrical distributor, HG1=60 mm) ..128 Figure 4.13 Effect of gas velocities on solids circulation rate (Cylindrical distributor, HG1=10mm) .......................................................................................................129 Figure 4.14 Effect of gas velocities on solids circulation rate (Cylindrical distributor, HG1=20mm) .......................................................................................................130 Figure 4.15 Effect of gas velocities on solids circulation rate (Cylindrical distributor, HG1=60mm) .......................................................................................................131 Figure 4.16 Configuration of the conical distributor for the annulus gas flow.....................132 Figure 4.17 Effect of gas velocities on gas bypass (Conical distributor, HG=10 mm) .........133 Figure 4.18 Effect of gas velocities on gas bypass (Conical distributor, HG=15 mm) .........135 Figure 4.19 Effect of gas velocities on solids circulation rate (Conical distributor, HG=10 mm)....................................................................................................................136 Figure 4.20 Effect of gas velocities on solids circulation rate (Conical distributor, HG=15 mm)....................................................................................................................137 List of Figures  xv Figure 4.21 Effect of overall gas velocity on solids circulation rate using conical distributor (HG=10 mm) ......................................................................................................139 Figure 4.22 Effect of overall gas velocity on solids circulation rate using conical distributor (HG=15 mm) ......................................................................................................139 Figure 4.23 Configuration of the conical distributor in the hot model ICFB reactor ...........140 Figure 4.24 Effect of gas velocities on gas bypass (Fe/ZSM-5(PUC), T=350±10oC) .........143 Figure 4.25 Effect of gas velocities on gas bypass (Fe/ZSM-5(Albemarle), T=355±15oC) 145 Figure 5.1 Schematic of hot model ICFB reaction system ...................................................152 Figure 5.2 Effect of gas velocities and HC:NO ratio on NOx conversion using the draft tube as the adsorption zone (ICFB, Fe/ZSM-5(PUC))..............................................159 Figure 5.3 Effect of catalyst loading on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(PUC)) .............................................................................................................160 Figure 5.4 Effect of HC:NO ratio on NOx conversion (ICFB, Fe/ZSM-5(PUC)) ................162 Figure 5.5 Effect of flue gas O2 concentration on NOx conversion (ICFB, Fe/ZSM-5(PUC)) ...........................................................................................................................163 Figure 5.6 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(PUC)) .............................................................................................................164 Figure 5.7 Effect of inlet NO concentration on NOx conversion (Fluidized bed, Fe/ZSM- 5(PUC), [O2] =4%)............................................................................................166 Figure 5.8 Effect of inlet NO concentration on NOx conversion (Fluidized bed, Fe/ZSM- 5(PUC), [O2] =2.5%).........................................................................................166 Figure 5.9 Effect of inlet NO concentration on NOx conversion (Fluidized bed, Fe/ZSM- 5(PUC), [O2] =1%)............................................................................................167 List of Figures  xvi Figure 5.10 Effect of HC:NO on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC), [NO]=300ppm) ..................................................................................................169 Figure 5.11 Effect of HC:NO on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC), [NO]=600ppm) ..................................................................................................170 Figure 5.12 Effect of HC:NO on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC), [NO]=900ppm) ..................................................................................................171 Figure 5.13 Effect of inlet O2 concentration on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC))...............................................................................................173 Figure 5.14 Comparison of NOx conversion between ICFB and fluidized bed reactors (Fe/ZSM-5(PUC)) .............................................................................................174 Figure 5.15 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.6 m/s, [O2]=4%)...............................................................176 Figure 5.16 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.75 m/s, [O2]=4%).............................................................176 Figure 5.17 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.9 m/s, [O2]=4%)...............................................................177 Figure 5.18 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.6 m/s, [O2]=8%)...............................................................177 Figure 5.19 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.75 m/s, [O2]=8%).............................................................178 Figure 5.20 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.9 m/s, [O2]=8%)...............................................................178 Figure 5.21 Effect of flue gas O2 concentration on NOx conversion and adsorption (ICFB, Fe/ZSM-5(Albemarle), UD=0.6 m/s, HC:NO=2) ..............................................180 List of Figures  xvii Figure 5.22 Effect of flue gas O2 concentration on NOx conversion and adsorption (ICFB, Fe/ZSM-5(Albemarle), UD=0.75 m/s, HC:NO=2) ............................................181 Figure 5.23 Effect of flue gas O2 concentration on NOx conversion and adsorption (ICFB, Fe/ZSM-5(Albemarle), UD=0.9 m/s, HC:NO=2) ..............................................181 Figure 5.24 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=4%, HC:NO=1) .................................................................183 Figure 5.25 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=4%, HC:NO=2) .................................................................183 Figure 5.26 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=8%, HC:NO=2) .................................................................184 Figure 5.27 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=12%, HC:NO=2) ...............................................................185 Figure 5.28 Effect of HC:NO ratio on NOx and HC conversions (Fluidized bed, Fe/ZSM- 5(Albemarle), [O2]=1%)....................................................................................186 Figure 5.29 Effect of HC:NO ratio on NOx and HC conversions (Fluidized bed, Fe/ZSM- 5(Albemarle), [O2]=4%)....................................................................................187 Figure 5.30 Effect of HC:NO ratio on NOx and HC conversions (Fluidized bed, Fe/ZSM- 5(Albemarle), [O2]=8%)....................................................................................187 Figure 5.31 Effect of inlet O2 concentration on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(Albemarle), HC:NO=1)...................................................................189 Figure 5.32 Effect of inlet O2 concentration on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(Albemarle), HC:NO=2)...................................................................189 Figure 5.33 Comparison of NOx conversion between ICFB and fluidized bed reactors (Fe/ZSM-5(Albemarle)) ....................................................................................191 List of Figures  xviii Figure B.1 Calibration curve of FL-1473G rotameter (0.6% NO + N2)...............................236 Figure B.2 Calibration curve of FL-1473G rotameter (50% O2 + N2) .................................237 Figure B.3 Calibration curve of FL-1472S rotameter (1.2% Propylene + N2) .....................237 Figure B.4 Calibration curve of FL-1473S rotameter (N2)...................................................238 Figure B.5 Calibration curve of FL-1476G rotameter (N2) ..................................................238 Figure B.6 Calibration curve of peristaltic water pump .......................................................239 Figure B.7 Calibration curve of pressure transducer (Channel #1) ......................................240 Figure B.8 Calibration curve of pressure transducer (Channel #2) ......................................241 Figure B.9 Calibration curve of pressure transducer (Channel #3) ......................................241 Figure B.10 Calibration curve of pressure transducer (Channel #4) ....................................242 Figure B.11 Calibration curve of pressure transducer (Channel #6) ....................................242 Figure D.1 Particle size distribution (Fe/ZSM-5(PUC)).......................................................244 Figure D.2 Particle size distribution (Fe/ZSM-5(crushed PUC)) .........................................245 Figure D.3 Particle size distribution (Fe/ZSM-5(Albemarle)) .............................................245 Figure D.4 Particle size distribution (Fe/FCC(Spent)) .........................................................246 Figure E.1 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(crushed PUC), T=325oC).................................................................248 Figure E.2 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(crushed PUC), T=350oC).................................................................248 Figure E.3 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(crushed PUC), T=375oC).................................................................249 Figure E.4 Fitting of adsorption isotherms of NOx by Freundlich equation (Catalyst: Fe/ZSM-5(crushed PUC)) .................................................................................250 List of Figures  xix Figure E.5 Relationship between /β and adsorption temperature (Catalyst: Fe/ZSM-5 (crushed PUC))..................................................................................................251 Figure E.6 Effect of reaction temperature on catalytic activity (Catalyst: Fe/ZSM-5(crushed PUC)).................................................................................................................252 Figure E.7 Effect of inlet O2 concentration on catalytic activity (Catalyst: Fe/ZSM-5(crushed PUC)).................................................................................................................253 Figure F.1 HC-SCR performance of spent FCC and Fe/FCC ..............................................254 Figure G.1 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=280oC) ..............................................................................255 Figure G.2 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=310oC) ..............................................................................256 Figure G.3 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=340oC) ..............................................................................256 Figure G.4 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=370oC) ..............................................................................257 Figure G.5 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=400oC) ..............................................................................257 Figure G.6 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=300oC).....................................................................258 Figure G.7 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=325oC).....................................................................258 Figure G.8 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC).....................................................................259 List of Figures  xx Figure G.9 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=375oC).....................................................................259 Figure G.10 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC, 10% H2O added).........................................260 Figure G.11 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC, 10% CO2 added) .........................................260 Figure G.12 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC, 10% H2O + 10% CO2 added) .....................261 Figure G.13 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=275oC, [O2]=1%)...........................262 Figure G.14 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=275oC, [O2]=4%)...........................263 Figure G.15 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=300oC, [O2]=1%)...........................264 Figure G.16 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=325oC, [O2]=1%)...........................265 Figure G.17 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(Albemarle), T=300oC, [O2]=1%)..................266 Figure G.18 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(Albemarle), T=325oC, [O2]=1%)..................267 Figure G.19 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(Albemarle), T=375oC, [O2]=1%)..................268 Figure G.20 XPS survey scan for fresh Fe/ZSM-5(PUC) ....................................................269 Figure G.21 XPS survey scan for spent Fe/ZSM-5(PUC)....................................................270 List of Figures  xxi Figure G.22 XPS survey scan for regenerated Fe/ZSM-5(PUC) .........................................270 Figure G.23 Comparison of XPS narrow scan of Al 2p .......................................................271 Figure G.24 Comparison of XPS narrow scan of Si 2p........................................................271 Figure G.25 Comparison of XPS narrow scan of O 1s.........................................................272 Figure G.26 Comparison of XPS narrow scan of Fe 2p 3/2.................................................272 Figure H.1 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.20 m/s, T=350±10oC) ....273 Figure H.2 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.30 m/s, T=350±10oC) ....274 Figure H.3 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.35 m/s, T=350±10oC) ....274 Figure H.4 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.40 m/s, T=350±10oC) ....275 Figure H.5 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.45 m/s, T=350±10oC) ....275 Figure H.6 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.50 m/s, T=350±10oC) ....276 Figure H.7 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.55 m/s, T=350±10oC) ....276 Figure H.8 Effect of gas velocities on RDA (Fe/ZSM-5(Albemarle), T=355±15oC) ............277 Figure H.9 Effect of gas velocities on RAD (Fe/ZSM-5(Albemarle), T=355±15oC) ............278 Figure H.10 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.20 m/s, T=355±15oC).....................................................................................................278 Figure H.11 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.25 m/s, T=355±15oC).....................................................................................................279 Figure H.12 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.30 m/s, T=355±15oC).....................................................................................................279 Figure H.13 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.35 m/s, T=355±15oC).....................................................................................................280 Figure H.14 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.40 m/s, T=355±15oC).....................................................................................................280 List of Figures  xxii Figure H.15 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.45 m/s, T=355±15oC).....................................................................................................281 Figure H.16 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.45 m/s, T=355±15oC).....................................................................................................281 Figure H.17 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.60 m/s, T=355±15oC).....................................................................................................282 Figure H.18 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.75 m/s, T=355±15oC).....................................................................................................282 Figure H.19 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.90 m/s, T=355±15oC).....................................................................................................283  xxiii Acronyms  AOFA  Advanced overfire air CVD   Chemical vapour deposition Fe(AA)3 Iron(III) acetylacetonate GHSV  Gas hourly space velocity, h-1 HC  Hydrocarbon HC-SCR Selective catalytic reduction with hydrocarbons ICFB  Internal circulating fluidized bed I.D.  Inner diameter IMP  Impregnation IMPA   Incipient wetness impregnation in aqueous solution IMPO  Impregnation in organic solution LNB  Low NOx burner MFI  One type of zeolite M.W.  Molecular weight NSCR  Non-selective catalytic reduction NSR  NOx storage-reduction O.D.  Outer diameter PUC  China University of Petroleum (Beijing, China) SCR  Selective catalytic reduction SNCR   Selective non-catalytic reduction SSIE  Solid state ion exchange Std. T  Standard temperature, 25oC Acronyms   xxiv SUB  Sublimation SUZ-4  One kind of zeolite TGA  Thermogravimetric analysis TOS  Time on stream TPD  Temperature programmed desorption VOC  Volatile organic compound WIE  Wet ion exchange XPS  X-ray photoelectron spectroscopy ZSM-5  One kind of MFI         xxv  Nomenclature C0   Equilibrium NOx concentration, ppm CA,O2  O2 concentration in the annulus, % CA0, O2  O2 concentrations at inlet of the annulus, % CCO,out  CO concentration in gas mixture at reactor outlet, ppm CCO2,0  Initial CO2 concentration in flue gas feed, % CCO2,out CO2 concentration in gas mixture at reactor outlet, % CD,O2  O2 concentration in the draft tube, % CD0, O2  O2 concentrations at inlet of the draft tube, % CHC,in   Inlet concentration of propylene, ppm CNOx,0  Initial NOx concentration in flue gas feed, ppm CNOx, ads,in Real initial NOx concentration for adsorption at adsorption zone inlet, ppm CNOx,ads, out NOx concentration at adsorption zone outlet, ppm CNOx,in  Initial NOx concentration in total gas flow , ppm CNOx,out NOx concentration in gas mixture at reactor outlet, ppm Di  Inner diameter of fixed bed reactor, mm dp  Mean particle diameter, mm F  Gas flow rate at ambient temperature, m3/s FA  Volumetric gas flow rate at annulus outlet, m3/s FA,0  Volumetric gas flow rate at annulus inlet, m3/s FAD  Volumetric gas bypass rate from annulus to draft tube, m3/s FD  Volumetric gas flow rate at draft tube outlet, m3/s FD,0  Volumetric gas flow rate at draft tube inlet, m3/s FDA  Volumetric gas bypass rate from draft tube to annulus, m3/s Nomenclature   xxvi FF,0  Flue gas flow rate, m3/s FFR  Calculated gas bypass from adsorption zone to reduction zone, m3/s FHC,0  Flow rate of 40% Propylene + N2, m3/s FR,0  Reductant gas flow rate, m3/s FRF  Calculated gas bypass from reduction zone to adsorption zone, m3/s h  Catalyst packed height in fixed bed reactor, mm HG  Effective gap opening for solids circulation, mm HG1  Gap between draft tube and annulus gas distributor, mm HG2  Gap between draft tube and annulus gas distributor, mm L  Tracer particle downward moving distance along column, m MNO  Molecular weight of NO, g/mol P  Operating pressure, atm qe   Adsorption equilibrium capacity of NOx on catalyst, mg/g R  Universal gas constant, 8.314 J/(mol.K) RAD   Gas bypass ratio from annulus to draft tube, % RDA   Gas bypass ratio from draft tube to annulus, % S  Integrated area from adsorption curve, s SA  Cross-sectional area of annulus, m2 SD   Cross-sectional area of draft tube, m2 t  Adsorption time, s T  Reaction temperature, oC T0  Ambient temperature, oC tm  Time interval for tracer particle moving L, s U  Overall gas velocity in ICFB reactor, m/s Nomenclature  xxvii UA  Gas velocity in annulus, m/s UD  Gas velocity in draft tube, m/s Umf  Minimum fluidizing velocity, m/s Up  Tracer particle downward velocity in annulus, m/s Wcat   Catalyst loading, g Ws  Solids circulation rate based on draft tube cross-sectional area, kg/m2.s xA  Volume fraction of CO2 in gas flow at annulus outlet, - xA,0  Volume fraction of CO2 in gas flow at annulus inlet, - xAD  Volume fraction of CO2 in gas bypass flow from annulus to draft tube, - xD  Volume fraction of CO2 in gas flow at draft tube outlet, - xD,0  Volume fraction of CO2 in gas flow at draft tube inlet, - xDA  Volume fraction of CO2 in gas bypass flow from draft tube to annulus, - XHC  Propylene conversion, % XNOx  NOx conversion, % XNOx, ads  Adsorption ratio of NOx in adsorption zone, % XNOx, BM Benchmark NOx conversion caused by bypassing HC from reduction zone, % α  Coefficient in Freundlich equation, - β  Coefficient in Freundlich equation, -    Bed voidage in annulus, - air  Air density, kg/m3 CO2  CO2 density, kg/m3 ρp   Particle bulk density, kg/m3   xxviii  Acknowledgements I would like to take this opportunity to express my sincere gratitude to my supervisor, Dr. Xiaotao Bi, for his immense guidance and financial assistance during my research. I give my utmost appreciation to my research committee members, Dr. Kevin Smith and Dr. Robert Evans, for their valuable suggestions at the start of this work. Many individuals who contributed to this research would be greatly acknowledged. Mr. Tianzhu Zhang, a former visiting scholar from China, contributed a lot to the setup of the fixed bed reaction system and the design of the cold model unit. Mr. Yuanqing (Larry) Liu, an undergraduate student of class 2008, was of great assistance to the adsorption and reaction kinetics study in the fixed bed reactor during the research on his undergraduate thesis. Mr. Shahin Goodarznia from Dr. Kevin Smith’s research group helped with the measurement of the catalyst BET surface area. Ms. Yonika Wiputri from Dr. Peter Englezos’s research group helped with the measurement of the particle size distribution. I also appreciate the hard work of the technical and administrative staff in the department stores, workshop and offices. Na/ZSM-5(PUC) was purchased from the Petroleum University of China with the help of Professors Weisheng Wei and Xiaojun Bao. H/ZSM-5(Albemarle) was a free sample offered by Dr. Maria M. Ludvig from Albemarle Catalysts Co., USA. This research was financially supported by the Canada Foundation for Innovation (CFI) and the Natural Sciences and Engineering Research Council of Canada (NSERC). I am also grateful to a NSERC postgraduate scholarship (PGS-D). Last but not least, I am sincerely grateful to my wife Jessica, my daughter Rena, my parents and my parents-in-law, for their endless support and encouragement throughout my whole life.   1  Chapter 1 Introduction  Nitrogen oxides (NOx) are emitted primarily from transportation and other industrial sources, and contribute to many problems that threaten the quality of the environment and human health: the formation of acid rain and ground level ozone (smog), and general atmospheric visibility degradation. Man-made nitrogen oxides are mainly classified into two types, depending on their sources: stationary or mobile.  1.1 Formation of NOx NOx form when fuel is burned at high temperatures. N2 and O2 present in the fuel and air can combine chemically to form one or more of seven oxides of N: NO, NO2, NO3, N2O, N2O3, N2O4 and N2O5. For most combustion processes, particularly fossil fuel combustions, both the molecular N in the air and the chemically bound N in the fuel can be oxidized by O2 to form NOx, and the only oxides of nitrogen present in significant concentration in flue gases are NO and NO2 with NO representing 90 to 95% of the total NOx (Wark et al., 1998). The colorless and odorless NO, in the presence of radicals from VOCs or O2 in the air, will be gradually oxidized to NO2 if sufficient time is given. NO2 is a reddish-brown, toxic gas, which is a precursor for photochemical smog formation in the presence of certain hydrocarbons and sunlight, and along with particles in the air, NO2 can often be seen as a reddish-brown layer over many urban areas. NOx are produced in a variety of processes, mainly in the combustion process of fossil fuels, such as engines of vehicles and aircraft, combustion equipment, gas turbines, incinerators, kilns and power plants. NOx also are emitted as by-products from many Chapter 1: Introduction  2 metallurgical processes where nitric acid is used as an oxidant. Typical NOx concentrations in flue gases from power plants are shown in Table 1.1.  Table 1.1 Typical compositions of flue gases from thermal power plant stacks Fuel Natural gas Fuel oil Coal NOx (ppm) 25–160 100–600 150–1000 SOx (ppm) <0.5–20 200–2000 200–2000 CO2 (%) 5–12 12–14 10–15 O2 (%) 3–18 2–5 3–5 H2O (%) 8–19 9–12 7–10 N2 Balance Balance Balance   1.2 Methods for the abatement of NOx emissions Due to the increasingly stringent emission regulations on NOx all over the world, the method for the abatement of NOx emission has gained extensive attention from academic as well as industrial research groups. In industrial combustion processes, NOx control can be accomplished by three primary methods, i.e., pre-combustion, combustion modification and post-combustion control techniques, which involve the use of low nitrogen fuels, modification of the combustion design and operating features of the combustion unit, and flue gas treatment, respectively. As an effective pollution prevention practice, the pre- combustion and combustion modification technologies have been widely used in combustion processes (Wark et al, 1998; de Nevers, 2000). Although NOx abatement at control efficiencies lower than 60% can be achieved by pre-combustion and combustion modification techniques at relatively low cost, expensive downstream treatment (e.g. SCR) is still required to reach control efficiencies higher than 75% in order to meet stringent emission regulations, as shown in Figure 1.1. Chapter 1: Introduction  3    Figure 1.1 NOx control effectiveness for coal-fired tangential boilers (Source: Alternative Control Techniques Document – NOx Emission from Utility Boilers, EPA-453/R-94-023, U.S. EPA OAQPS, 1994).  To date, there are four ways to abate NOx in flue gases: direct catalytic decomposition, non-selective catalytic reduction (NSCR), selective non-catalytic reduction (SNCR) and selective catalytic reduction (SCR). Direct catalytic decomposition of NOx might be the most attractive method for NOx control, but the NOx conversion is not high enough for the practical application (Parvulescu et al., 1998; Liu and Woo, 2006; Shi et al., 2008). The best example of non-selective catalytic reduction (NSCR) of NOx is the three- way catalytic converter which has been commercialized successfully for reducing NOx in the Chapter 1: Introduction  4 exhaust gases from gasoline engines (Parvulescu et al., 1998). However, it could not be applied to lean-burn engines where excess oxygen is present in the exhaust gases (Goralski and Schneider, 2002). Using gas-phase free-radical reactions instead of a catalyst to promote reactions in the selective non-catalytic reduction (SNCR) process, NH3 is injected into the flue gas to reduce NOx via the gas phase homogeneous reaction between NH3 and NOx. In order to achieve a high NOx reduction, the gas mixture must be kept within a relatively narrow and high temperature window of 900~1100oC. If the reaction temperature is lower than 800oC, NOx conversion will be very low and most NH3 remains un-reacted in the flue gas. At a temperature higher than 1200oC, NH3 will be oxidized by O2 to form NO, rather than being oxidized by NO to form N2 (Muzio and Quartucy, 1997; Javed et al., 2007). Selective catalytic reduction, so called SCR, selectively reduces NOx to N2 and H2O by the use of a catalyst and a reducing agent. SCR is the only single method which can achieve more than 75% NOx control efficiency for the flue gas emitted from stationary combustion sources with a high selectivity toward N2. Despite extensive research reported on the selective catalytic reduction (SCR) of NOx with various catalyst /reductant combinations since 1975, only SCR with NH3 or urea as the reducing agent offers a practical approach for NOx abatement, and has been commercialized to date (Forzatti, 2001).  1.3 Selective catalytic reduction of NOx with NH3 SCR with NH3 as the reducing agent has been commercialized in more than 400 industrial and utility plants, mainly in Japan, US and European countries.  In this process, liquid ammonia or aqueous ammonia solution is injected into the flue gas to reduce NOx with V2O5/TiO2 catalyst (WO3 and/or MoO3 as promoters) following reactions in equations 1.1 Chapter 1: Introduction  5 and 1.2, with 75%~90% of NOx conversion being achieved at an NH3-to-NOx molar ratio of 1.2~1.5. OHNONONH 2223 6444          (1.1) OHNONONH 22223 6324      (1.2)  Although NH3-SCR of NOx has been widely implemented, it has the following disadvantages (Ham and Nam, 2002; Zhou et al., 1996):  The handling of large quantities of NH3 is a concern because NH3 is toxic and corrosive;  Un-reacted NH3 may potentially be discharged to the environment (ammonia slippage);  NH3 can react with SO3 and H2O to form ammonium sulphates, causing the fouling of downstream equipment.  Due to the high installation and operating cost and complexity, the NH3-SCR system is only used when the highest level of NOx removal is required, for example, in large power plants located in urban areas where ozone and photochemical smog are serious problems. Taking the drawbacks of using NH3 in SCR process into account, a large number of studies on NOx SCR has been reported in recent years, mainly focusing on the selection and evaluation of NH3-free alternative reducing agents and catalysts, and the mechanism of SCR. Hydrocarbons instead of NH3 are considered the most promising reducing agents for the SCR process. The first study on SCR of NOx with hydrocarbons was reported by Iwamoto et al. (1988). Thereafter, SCR of NOx by hydrocarbons (HC-SCR) has attracted considerable attention as an alternative of ammonia or urea process for the treatment of oxygen-rich flue gases, and numerous studies on various types of catalysts and reducing Chapter 1: Introduction  6 agents have been widely investigated (Parvulescu et al., 1998; Traa et al., 1999; Wichtelova et al., 2003; Liu and Woo, 2006). However, the NOx conversion in HC-SCR process is still much lower than that in NH3-SCR at a given reductant to NOx stoichiometric ratio because of the existence of excess O2 and poisioning components (SOx and H2O) in the flue gas.  1.4 Research objectives As a technology under development, most studies have focused on the development of catalysts, the selection of reducing agents, or the understanding of the reaction mechanisms of the process. In order for such a technology to be used in the industrial processes, engineering aspects still need to be considered to address technical challenges in the reactor design and scale-up for the HC-SCR process. To avoid the negative impact of the poisoning components and excess O2 in the flue gas, we propose a new concept of an integrated NOx adsorption-reduction process, as shown schematically in Figure 1.2, where the NOx adsorption and reduction are carried out in two separate zones of the reactor. The flue gas is passed into the adsorption zone where NOx is adsorbed by the catalyst. The NOx-adsorbed catalyst particles then move into the reaction zone where NOx is reduced by injected hydrocarbons at controlled oxygen concentrations, and at the same time, other adsorbed flue gas contaminants are also stripped from the catalyst. The regenerated catalyst particles are then recirculated back to the adsorption zone to establish a continuous operation. Chapter 1: Introduction  7 Flue gas Reductant gas Adsorption      zone Reduction     zone Catalyst Catalyst  Figure 1.2 Schematic of the integrated adsorption-reduction process of NOx  The main objective of this work was to evaluate the potential application of the integrated adsorption-reduction process for the selective catalytic reduction of NOx with hydrocarbons in a dual-zone fluidized bed reactor, i.e. the internal circulating fluidized bed (ICFB) reactor. To achieve this goal, a fixed bed reactor system was set up first to investigate the adsorption performance and reaction kinetics of selected catalysts at various temperatures and flue gas compositions. After the hydrodynamic studies were performed in a cold model ICFB unit, the adsorption and reaction performance of selected catalysts was conducted in the hot model ICFB reactor, with the results compared with a conventional bubbling fluidized bed reactor.  1.5 Thesis layout A literature review is presented in Chapter 2 to summarize the present status of NOx SCR with hydrocarbons, including catalysts, effects of the flue gas composition and reducing agents, and mechanisms of HC-SCR. The adsorption performance and reaction kinetics of Chapter 1: Introduction  8 HC-SCR over selected catalysts in a fixed bed reaction system are given in Chapter 3. The hydrodynamic study on the ICFB reactor is shown in Chapter 4, including the gas bypass and solids circulation rate. Results of NOx reduction and adsorption over Fe/ZSM-5 catalysts in a hot model ICFB reactor are discussed in Chapter 5. In addition, the results from a conventional fluidized bed reactor are also reported and compared with those from the ICFB reactor. Finally, Chapter 6 summarizes the conclusions of this study and recommendations for the future work.   9  Chapter 2 Literature Review  It was first reported by Iwamoto et al. (1988) that the catalytic activity of Cu/ZSM-5 on the reduction of NOx was remarkably enhanced by the addition of hydrocarbons. Following this discovery, studies on HC-SCR have been expanded to various aspects, including sources of catalyst supports, metal loadings and preparation methods, effects of flue gas compositions and reducing agents, and reaction mechanisms.  2.1 Catalysts Generally, HC-SCR catalysts can be categorized into two groups according to catalyst supports: zeolite (ZSM-5, Mordenite, Y-zeolite, Ferrierite and Beta zeolite) (Iwamoto et al., 1990; Corma et al, 1997; Amiridis et al., 1997; Traa et al., 1999; Chen et al., 2000b; Wichtelova et al., 2003) and non-zeolite (mainly metal oxides, such as Al2O3, SiO2, and ZrO2) (Captain et al., 1998; Garcia-Cortes et al., 2000; Lick et al., 2003a; Liu and Woo, 2006). Based on the supports described above, noble metals (Pt, Pd, Rh, Ir, Ag, Au) (Cho et al., 1995; Parvulescu et al., 1998; Kikuchi et al., 2000; Shi et al., 2002; Sato et al., 2003; Ueda et al., 1997), transition metals (Cu, Co, Fe, Cr, Mo, Ni, Mn, Al, Ti, V, Zn, Zr) (Iwamoto et al., 1990; Witzel et al., 1994; Hall et al., 1998; Wang et al., 2000; Mosqueda- Jimenez et al., 2003; Lucas et al., 2004), alkali and alkaline metals (Na, K, Li, Ba, Ca, Mg, Sr) (Sato et al., 1992), rare earth metals (La, Nd, Ru, Sm) (Bosch and Janssen, 1988) and others (Ga, In) (Tabata et al., 1995) were loaded as active elements in the prepared HC-SCR catalysts.   10  Metals are often loaded onto the supports in four ways:  1. Wet Ion-Exchange (WIE): The support is added to the chemical solution with expected active metal element. The mixture is stirred for a given time period, at a given temperature, with or without inert gas blanket. At the end of this process, the mixture is filtered and washed thoroughly with deionized water, dried and calcined under different conditions. This process may be repeated several times to gain various ion-exchange rates (Wang et al., 2000). 2. Impregnation (IMP): The support is added to the chemical solution (aqueous or organic) with expected metal ion, volume and concentration. The solution is stirred for a given time period, at a given temperature. After that, the mixture is filtered or evaporated, with or without washed thoroughly with deionized water, dried and calcined under different conditions (Chen et al., 1998a; Delahay et al., 2005). 3. Sublimation (SUB) or Chemical Vapour Deposition (CVD): The metal compound is sublimated under given conditions while the support is put into this atmosphere. After a given time period, the sample is removed and washed with deionized water thoroughly, then dried and calcined (Chen et al., 1998b). 4. Solid State Ion Exchange (SSIE): The support and metal compound are physically mixed and grounded. After a given time period, the mixture is washed thoroughly by deionized water, dried and calcined (Berndt et al., 2003).  Different loading methods will give different metal loading ratios. Nevertheless, higher metal loading doesn’t always mean higher reactivity (such as Cu/ZSM-5, as shown in Figure 2.1). The catalyst reactivity is closely related to the preparation process and the reaction Chapter 2: Literature Review  11 conditions. In general, catalysts with various metal loadings described above have all shown catalytic activity for HC-SCR of NOx. 0.4 0.5 0.6 0.7 0.8 0.9 Cu/Al 0 10 20 30 40 50 60 70 80 90 100 N O  c on ve rs io n to  N 2 (% ) 350 oC 450 oC 550 oC  Figure 2.1 NO conversion to N2 over Cu/ZSM-5 catalysts as a function of Cu/Al ratio at different temperatures Reaction conditions: 2000 ppm NO, 2000 ppm C3H6, 2% O2, balanced by He; 1.0 g catalyst and 150 ml/min-1 total flow rate (Seyedeyn-Azad et al., 2001).  2.1.1 Metal-exchanged zeolites Much effort has been devoted to the group of metal-exchanged zeolite. So far, Cu, Fe, Pt, Co, Pd supported by zeolite, mainly as ZSM-5, gained extensive attention. Cu/ZSM-5 might be the most studied catalyst for its higher NO conversion (NO conversion to N2 could reach up to 85% in the absence of H2O and SO2 (Seyedeyn-Azad et al., 2001)) over a wide temperature window (300~550oC) by using various reducing agents. However, deactivation of Cu/zeolite catalyst by H2O and SO2 in the flue gas is still the obstacle in the practical applications (Gilot et al., 1997; Chen et al., 1998a; Li et al., 2004). Many papers also reported Chapter 2: Literature Review  12 that Cu exchanged with other types of zeolite as HC-SCR catalysts could achieve similar results as Cu/ZSM-5, although differences existed in the maximum reactivity and operating temperature windows of these catalysts (Shelf, 1995; Walker, 1995; Corma et al., 1997; Ohtsuka et al., 1997; Park et al., 2000; Cho et al., 2003). As indicated in Figure 2.1, with the increase of Cu/Al ratio, NO conversion to N2 decreased, suggesting that increasing the quantity of implanted Cu2+ did not improve the activity of Cu/ZSM-5. Most recently, Subbiah et al. (2003) reported that a new catalyst Cu/SUZ-4, which was supported by SUZ-4 zeolite, gave a high reactivity (NO conversion of 44~68%, based on 2.3%Cu/SUZ-4) and a wide operating temperature window (350~600oC) even in the presence of H2O and SO2. The reactivity of this catalyst increased as the Cu-exchanged level increased. Fe/ZSM-5 was widely reported in recent years to posses high effectiveness without the production of N2O and good resistance to H2O (Feng and Hall, 1997; Yamada et al., 1998; Kogel et al., 1999; Chen et al., 2000b; Battiston et al., 2003; Sobalik et al., 2002). However, large amount of CO was detected when Fe/ZSM-5 was used as the HC-SCR catalyst (Chen et al., 2000b). According to the literature, Fe loading level on zeolite also played an important role. Although catalysts with low Fe loadings gave some activity for NO reduction (NO conversion to N2 of ~20%), only catalysts with higher Fe loadings prepared by chemical vapour deposition showed satisfactory NO conversion (>60%) (Chen and Sachtler, 1998a; Saaid et al., 2002). Nevertheless, it is very difficult to prepare large amount of catalyst by chemical vapour deposition method from the practical application point of view. More recently, Fe/ZSM-5 catalyst prepared by an impregnation method using a solution of Fe(AA)3 (Iron(III) Acetylacetonate) in toluene was reported to show similar activities for Chapter 2: Literature Review  13 NH3-SCR of NOx as catalyst prepared by chemical vapour deposition method (Delahay et al., 2005). Pt/ZSM-5 is more active at lower temperatures and is less affected by SO2 and water vapour, but its active temperature window is very narrow (200~250oC) and most of NO is reduced to N2O in the presence of oxygen (Le et al., 1994; Xin et al., 1999; Woo et al., 2003). Cobalt-containing zeolite showed good activity when methane was used as the reducing agent (Witzel et al., 1994). Moreover, neither N2O nor CO was detected in the HC- SCR using Co-containing catalyst. However, the effective operating temperature of Co/zeolite is too high (450~550oC) and the catalytic activity is also inhibited by H2O and SO2 (Ohtsuka et al., 1997; Park et al., 2000; Wang et al., 2000). Like Co/ZSM-5, Palladium-containing zeolite can reduce NO with methane. Nishizaka et al. (1994) reported a NO conversion to N2 of ~70% at a temperature around 475oC by using Pd/ZSM-5 in CH4-SCR without water in the model flue gas. Many researchers studied the reduction on Pd-containing zeolites using reducing agents other than methane, e.g. C2H4, C3H6, (CH3)2O and CH3OH (Shin, et al., 1995; Keiski et al., 1996; Adelman and  Sachtler, 1997; Kato et al., 1997; Masuda et al., 1998) . However, very low NO conversions (< 20%) were obtained. Much effort has been devoted to the investigation of HC-SCR on metal-containing zeolite promoted by other metals, such as Co promoted Pd/ZSM-5, Pt promoted Co/ZSM-5, etc (Kagawa et al., 1998; Gutierrez et al., 1998; Bustamante et al., 2002). Ogura et al. (2002) tried the use of Co as the promoter of Pd/ZSM-5, and observed a significant improvement on the durability of catalyst in the presence of water (NO conversion to N2 kept 60% in CH4- SCR, 500oC, 50 hours time-on-stream period). Chapter 2: Literature Review  14 Other metal loadings on zeolite have also received much attention, but no evidence showed that the catalytic performance was good enough even for model flue gases.  2.1.2 Metal-exchanged metal oxides Numerous metal oxides (such as Al2O3, TiO2, ZrO2, V2O5, Fe2O3, CoO, NiO and MnO2 etc.) have been investigated as supports of HC-SCR catalysts, but most of them showed low or no activity to NO reduction (Parvulescu et al., 1998; Liu and Woo, 2006). Among them, only Pt, Cu, Co, Fe, Ag, Pd supported by Al2O3, SO42-/TiO2 and SO42-/ZrO2 exhibited some catalytic activity (Burch et al., 1998; Captain et al., 1998; Lick et al., 2003a). Some researchers found that when promoters were added to metal-exchanged Al2O3, the catalytic activity could have a significant increase (Naito and Tanimoto, 1993; Obuchi et al., 1993). Chin et al. (1999) reported that Pd-Sulphated Zirconia catalyst could tolerate the effect of SO2, and that the inhibition of water in the flue gas was weak (NO conversion dropped from 48% to 38%) and reversible. Efthimiadis et al. (2001) used a fluidized bed reactor loaded with γ-Al2O3, Rh/Al2O3 or 5% Rh/Al2O3-95% FCC for the control of NOx emitted from the regenerator of a FCC pilot-plant unit. They found that γ-Al2O3 with CH3OH exhibited significant reactivity (maximum NO conversion of 97% at 400oC), Rh/Al2O3 also reduced more than 70% of NO to N2 (C3H6 as reducing agent, 300oC). Furthermore, although 5% Rh/Al2O3-95% FCC showed lower NO conversion (40%, 350oC), its catalytic activity was enhanced by the presence of SO2 in the flue gas. This result is notable, although the NO content was low (80 ppm) and no H2O was present in the treated flue gas. Shimizu et al. (2000) reported a high reactivity of Ag/Al2O3 in the SCR with higher hydrocarbons. They also compared the performance of several metal-containing Al2O3 catalysts, and found that higher NO conversion to N2 could be obtained by using n-octane as Chapter 2: Literature Review  15 the reducing agent. Satokawa et al. (2001) further showed the impressive tolerance of Ag/Al2O3 to SO2 in a 50 hours time-on-stream test. Although significant research has been made, the development of catalysts with both very stable supports and highly active catalytic metal ingredients for practical applications is still a challenge to researchers.  2.2 Hydrocarbon reducing agents Hydrocarbons, which have been proven effective as reducing agents in SCR of NOx, include methane, propane, propylene, iso-butane and oxygenated hydrocarbons, such as methanol and acetic acid etc. Among them, taking into account that methane is the main component of natural gas, it is not surprising that methane was the mostly studied reducing agent in the past years (Witzel et al., 1994; Tabata et al., 1995; Mitome et al., 1998; Kameoka et al., 2000; Bustamante et al., 2002; Berndt et al., 2003). Nevertheless, due to its preference to combustion, methane has been shown to be only effective in the SCR of NOx over Co/zeolite, Ga/ZSM-5, In/ZSM-5, Pd/Al2O3 and Pd/TiO2. The catalytic activity of these catalysts decreased significantly when water was added into the model flue gas. For example, the NO conversion to N2 decreased from 75% to 5% over Ga/ZSM-5 and 60% to 10% over In/ZSM- 5 (Tabata et al., 1995) when H2O was added. Ogura et al. (2002) reported that Pd promoted Co/ZSM-5 in the CH4-SCR showed good NO conversion to N2 (60% at 500oC) in the presence of water, similar to the result of Bustamante et al. (2002) (Pd-Co/Mordenite, 60% at 550oC). Berndt et al. (2003) obtained high catalytic activity over Ce-In/ZSM-5. NO conversion to N2 dropped from 100% to 88% when 5% water was added. Chapter 2: Literature Review  16 Due to the non-selective performance of methane, many researchers focused on other hydrocarbons, i.e., propane and propylene. Fe, Cu or Co exchanged zeolite and Co supported by alumina showed acceptable catalytic activity using propane as the reducing agent (Kogel et al., 1999; Park et al., 2000; Lick et al., 2003b; Martinez-Hermandez and Fuentes, 2004). Moreover, Co/Beta-zeolite (Tabata et al., 1996; Ohtsuka et al., 1997), Fe/Beta-zeolite (Chen et al., 2000b) and Ce- In/ZSM-5 (Berndt et al., 2003) exhibited high activity with propane even in the presence of water. This suggests that, by using these catalysts, the SCR of NOx with propane may be a good choice in the practical applications without the presence of SO2. When propylene was used as the reducing agent, Fe or Cu exchanged zeolite gave high NO conversion in the absence of H2O and SO2. At the same reaction conditions, the temperature with the maximum NO conversion over Cu/ZSM-5 was reported to be 100oC higher for propylene (500oC) than for propane (400oC) (Park et al., 2000). The temperature window of Fe/zeolite was found to be relatively low (300~400oC) for both propane and propylene. Figure 2.2 compared the NO conversion to N2 over Cu/ZSM-5 and Co/ZSM-5 using CH4 or C3H6 as the reducing agent under the same reaction condition. Obviously, Cu/ZSM-5 showed significantly different activity for tests with CH4 or C3H6 as the reducing agent, while Co/ZSM-5 kept almost the same activity for both reducing agents. Chapter 2: Literature Review  17 250 300 350 400 450 500 550 600 650 T (oC) 0 10 20 30 40 50 60 70 80 90 100 N O  c on ve rs io n to  N 2 (% ) Cu-ZSM-5 + C3H6 Cu-ZSM-5 + CH4 Co-ZSM-5 + C3H6 Co-ZSM-5 + CH4  Figure 2.2 Dependence of NO conversion on temperature for SCR of NO over various catalysts with alternative reducing agent Reaction conditions: 2000 ppm NO, 2000 ppm CH4 or C3H6, 2% O2, balanced by He; 1.0 g catalyst; total flow rate: 100 ml/min-1. (Seyedeyn-Azad et al., 2001)  Co, Fe, Ni, Mn, Ga and Cu exchanged ZSM-5 gave high NO conversion when iso- butane was used as the reducing agent (Witzel et al., 1994; Chen et al, 2000b; Saaid et al., 2002). Specifically, Co/ZSM-5 was resistant to water in the flue gas and gave very high N2 yield (>94%) (Wang et al., 2000). In the SCR of NOx over Fe/ZSM-5, various types of hydrocarbons were tested as reducing agents by Chen et al. (1998b), and the performance followed the order of i-butane > propane > propylene > methane, as shown in Figure 2.3.  Chapter 2: Literature Review  18 200 250 300 350 400 450 500 T (oC) 0 20 40 60 80 100 N 2 yi el d (% ) CH4 C3H6 C3H8 i-C4H10  Figure 2.3 Temperature dependence of N2 yield over Fe/ZSM-5 with different hydrocarbons Reaction conditions: 0.2 g Fe/ZSM-5, 0.2% NO, 3% O2, 0.2% i-C4H10 (0.8% CH4, 0.27% C3H8 or C3H6), flow rate: 280 ml/min (Chen et al., 1998b)  Witzel and coworkers (1994) examined the NOx reduction performance of methane, propane, iso-butane, n-pentane, 2, 2-dimethylpropane (neopentane), 3, 3-dimethylpentane, 2, 2, 4-trimethylpentane and 3, 3-diethylpentane (neononane) on Co/ZSM-5 catalyst. The maximum NO conversion as a function of the reducing agent followed the order of iso- butane > methane > neopentane > n-pentane > 2, 2, 4-trimethylpentane > propane > 3, 3- dimethylpentane > neononane (see Figure 2.4).  Chapter 2: Literature Review  19  Figure 2.4 Effect of the nature of the hydrocarbon on NO reduction selectivity on Co/ZSM-5 (Witzel et al., 1994)  Some other hydrocarbons, such as C2H6, n-C6H14, n-C4H10, n-C8H18 and n-C10H22, were also examined and found to give similar performance (Shimizu et al., 2000b; Shibata et al., 2002; Sato et al., 2003; Subbiah et al., 2003). Shimizu et al (2000) reported that the SCR activity of 2% Ag/Al2O3 notably increased as the carbon number in the reducing agent increased in the presence of water (Figure 2.5).  Chapter 2: Literature Review  20 500 600 700 800 T (K) 0 20 40 60 80 100 N O  c on ve rs io n to  N 2 (% ) Methane Ethane Propane n-Butane n-Hexane n-Octane  Figure 2.5 NO conversion to N2 on Ag/Al2O3 using alternative reducing agent Reaction conditions: NO/n-alkane/O2 = 1000 ppm/6000 ppm C/10%; 2% H2O, W/F= 0.12 g- s/ml except for CH4-SCR (W/F = 0.9 g-s/ml) (Shimizu et al., 2000)  Oxygenated hydrocarbons were investigated as reducing agents in the SCR (Masters et al., 1999b; Elkaim et al., 2000). The case over Al2O3 or Sulphated-Al2O3 catalyst with methanol as the reducing agent reached a high NO conversion (Efthimiadis et al, 2001; Burch et al., 1998). This implies that Al2O3 may have good tolerance to SO2 under given conditions, but the influence of water needs to be investigated further. Acetic acid was also reported in the SCR of NO over Pd/Mordenite. In the presence of water, the NO conversion reached 94% with a high acetic acid to NO ratio of 22 (Uchida et al., 1995). Chapter 2: Literature Review  21 Although hydrocarbons gained extensive attention in the area of SCR of NOx, the substantial difference between hydrocarbons and ammonia in NOx conversion and H2O and SO2 inhibition shows that there is still a long way to go for the practical application of HC- SCR process.  2.3 Reaction mechanisms and kinetics of HC-SCR The mechanism of HC-SCR has been investigated extensively. Due to the complexity of the HC-SCR process, different mechanisms were proposed for different catalysts, reducing agents and/or reaction conditions. Generally, according to the reaction pathways and rate- limiting steps, reaction mechanisms can be categorized into three groups (Parvulescu et al., 1998; Traa et al., 1999; Sadykov et al., 2003): surface redox, oxidation of NO to NO2, and partial oxidation of hydrocarbons.  2.3.1 Surface redox mechanism Based on microscopic reaction results, Inui et al.(1994) proposed that NO first decomposed to molecular nitrogen and chemisorbed oxygen on reduced surface sites of the catalyst, and the adsorbed oxygen oxidized the catalyst surface. The hydrocarbon then reacted with adsorbed oxygen and reduced the active site (Burch et al., 1996b; Cho et al., 1995). .)(22 2 adsONNO       (2.1) OHCOadsOHC x 2.)(       (2.2) Reaction (2.1) is considered to be the rate-limiting step.  Chapter 2: Literature Review  22 2.3.2 Oxidation of NO to NO2 NO is first oxidized to NO2 and/or surface nitrite-nitrate complexes by oxygen. Then the oxidized compounds are reduced by hydrocarbons to molecular nitrogen, possibly via intermediate organic nitro compounds (Yokoyama et al., 1994; Adelman et al., 1996; Yan et al., 1998; Kikuchi et al., 1996). xNOONO  2       (2.3) OHCONHCNO xx 22      (2.4) As indicated in the literature, the oxidation of NO to NOx (reaction 2.3) is the rate-controlling reaction.  2.3.3 Partial oxidation of hydrocarbons In this mechanism, hydrocarbons are partially oxidized to oxygen- and/or nitrogen-containing organic intermediates (e.g. cyanide or nitro species) by oxygen or NOx. The oxidized hydrocarbon intermediates subsequently react with NO to produce N2, CO and/or CO2, and H2O (Li et al., 1994c; Lobree et al., 1997; Shelf, 1995; Lukyanov et al., 1996; Hayes et al., 1996). * 2 / HCNOOHC x       (2.5) OHCONNOHC xx 22 *      (2.6) The rate-limiting step is reaction (2.5). It should be noted that the mechanism of HC-SCR is still not clearly understood due to the complexity of HC-SCR under different reaction conditions. Therefore, mechanisms suggested above are not universal. For some specific processes, a combination of those Chapter 2: Literature Review  23 mechanisms may work better than any one of them (Meunier et al., 2000). For the same reason, the kinetic expression of the HC-SCR process is still not readily available.  2.4 Influence of flue gas compositions on HC-SCR So far, compositions of the model flue gas in most investigations are far from the real ones. Most researchers reported their results without considering the influence of both water and sulphur dioxide, or either one of them. Hence, although some researchers claimed that they achieved very high NO conversion in the HC-SCR process, the results still need to be verified under practical operating conditions or using real combustion flue gases.  2.4.1 Influence of SO2 and H2O in the flue gas For a HC-SCR system, one of the most important issues for the practical application is the durability of catalyst to the poisoning components, i.e., H2O and SO2. Tabaka et al (1996) reported that Co/Beta-zeolite exhibited stable activity (NO conversion to N2 kept at about 70%) in the SCR with propane for the treatment of model flue gas with 9% H2O and 0.3ppm SO2 in a long term run (4000 hours) at 400oC, while under the same reaction conditions, Co/ZSM-5 was deactivated rapidly. Co/Al2O3 showed stable activity (NO conversion ~45%) in the HC-SCR with propylene when 10% H2O was presented in the model flue gas during 125 hours operation at 550oC (Yan et al., 1997). Berndt et al (2003) revealed that the SCR of NOx with propane over Ce-In/ZSM-5 was enhanced when 5% water was introduced into the system. The NO conversion to N2 increased slightly from 67% to 70%, yet the optimum temperature shifted from 550oC to 600oC. Chen et al. (1998a) reported the performance of Fe/ZSM-5 and Cu/ZSM-5 as a function of time on stream (see Figure 2.6). Although Cu/ZSM-5 showed 60% NO Chapter 2: Literature Review  24 conversion in the absence of water, its activity decreased quickly to 0 when water was injected into the stream, and the deactivation was partially irreversible. On the other hand, Fe/ZSM-5 showed stable activity no matter whether water was present or not.  Figure 2.6 Effect of H2O on NO conversion to N2 over Fe/ZSM-5 or Cu/ZSM-5 (●) Fe/ZSM-5 (temperature=375oC, GHSV=42,000 h-1, iso-C4H10=0.2%, NO=0.2%, O2=3%, H2O=0% for dry; 10% for wet); (■) Cu/ZSM-5 (temperature=400oC, GHSV=120,000 h-1, C3H8= 0.1%, NO=0.1%, O2=2%, H2O: 0% for dry; 10% for wet). (Chen et al., 1998a)  Ag/Al2O3 exhibited some long-term stable reactivity under real lean-burn gas engine conditions (Shimizu et al., 2000b). In a time-on-stream test, in the presence of 4ppm SO2, Ag/Al2O3 deactivated very quickly at temperatures lower than 500oC, but exhibited stable activity at 550oC even after 50 hours (Satokawa et al., 2001). Yan et al. (1997) reported that Co/Al2O3 was durable to the flue gas of 950ppm NO + 1000ppm propylene + 5% O2 + 1.5% H2O + 30ppm SO2 with NO conversion remaining at ~63% over a time-on-stream test of 6.5 hours (T=450oC), except for that the propylene Chapter 2: Literature Review  25 conversion decreased from 78% to 67% over time.  The activities of Ag/Al2O3, Ni/Al2O3, Co/Al2O3 and In/Al2O3 were seriously inhibited by 100ppm SO2 in the flue gas (Meunier et al., 2001), worsened with prolonged exposure time. With the removal of SO2 from the flue gas, the catalytic activities were partially recovered except for Ag/Al2O3. Similarly, Cu/SUZ- 4 (Cho et al., 2003) exhibited some tolerance and stable activity to both H2O and SO2. Decyk et al. (2001) investigated the effect of SO2 on the reactivity of Fe/ZSM-5. Using i-butane as the reducing agent, in the absence of SO2, NO conversion to N2 reached 59.1%. With the feed of 150ppm SO2, NO conversion decreased to 36%, and partially recovered to 43.8% after SO2 was removed from the flue gas. They also found that increasing the concentration of SO2 to 300ppm could further deactivate the catalyst. According to their analysis, the presence of SO2 suppressed the formation of Fe-NOy complexes on the catalyst surface and increased the formation of carbonaceous deposits, thus further deactivated the catalyst. In summary, Fe/ZSM-5, Pt/ZSM-5, Co/Beta-zeolite, Ag/Al2O3 and Ce-In/ZSM-5 exhibited a good hydrothermal resistance. On the other hand, only Co/Al2O3, Ag/Al2O3 and Cu/SUZ-4 showed some tolerance to sulfur dioxide at low concentrations.  2.4.2 Influence of excess oxygen in the flue gas As a key component of the flue gas under lean-burn conditions, the influence of excess oxygen has been widely investigated. One of the main NO reduction mechanisms states that NO is first oxidized to NO2, followed by the reduction by the reducing agent. The presence of O2 is thus essential for the HC-SCR process. This mechanism appears to agree with many experimental findings that the presence of small amount O2 (e.g., <2%) is essential for HC- SCR (Captain et al., 1998; Shi et al., 2002; Lee, et al., 1997). Chapter 2: Literature Review  26 Li et al. (2004) reported that, using Sn/Al2O3 as the catalyst in a TPD experiment, after the adsorption process for the flue gas containing NO and O2, the amount of desorbed species was much higher than that for the gas containing NO only (without O2), which indicates the fact that O2 in the flue gas could enhance the adsorption of NO. As shown in Figure 2.7, Corma et al. (1997) reported that NO conversion to N2 increased from 10% to 70% on Cu/ZSM-5 and 27% to 76% on Cu/Beta as O2 concentration increased from 0% to 2%. Thereafter, NO conversion to N2 decreased significantly to 43% on Cu/ZSM-5 and 42% on Cu/Beta in the presence of 6% O2. For Fe/ZSM-5 catalyst (Chen et al., 1998b), when O2 concentration was 0%, no N2 was produced. As O2 concentration increased, N2 yield increased and reached a maximum of 73% at 2% O2. Further increase of O2 concentration from 2% to 10% made N2 yield drop slightly to 69%. Li and Armor (1994a) studied NO reduction rate with low O2 concentration over Co/Ferrierite with CH4, and found that NO reduction rate increased from 0.12 to 0.58 mmol/h-g catalyst when O2 concentration rose from 0 to 5000ppm (0.5%v/v). Lee (2000) investigated the influence of oxygen on NO reduction using oxygenated hydrocarbons (acetic acid) as the reducing agent. The result revealed that with the increase of oxygen concentration from 2.25% to 9.36% in the model flue gas, NO conversion dropped from 52.2% to 21.8%. Elkaim and Bai (2000) reported high tolerance of V2O5/-Al2O3 catalyst to H2O with acetic acid as the reductant, and the catalyst kept high activity in the O2-free flue gas even in the presence of up to 15% H2O. However, it is sensitive to the concentration of O2 in the flue gas, with the NOx conversion dropped from 100% to 38% when the O2 concentration increased from 0% to 10%. Chapter 2: Literature Review  27 0 2 4 6 8 10 O2 concentration (%) 0 20 40 60 80 100 N O  c on ve rs io n to  N 2 (% ) Cu-Beta - C3H8 Cu-ZSM-5 - C3H8 Fe-ZSM-5 - C4H10 CAT-1 - CH3COOH V2 O5- Al2 O3- CH3COOH Co-FER - CH4  Figure 2.7 Influence of O2 concentration on NO conversion to N2 Reaction conditions: 1. Cu/Beta (●)and Cu/ZSM-5 (■): T=350oC, 850ppm NO, 450ppm C3H8 (Corma et al., 1997); 2. Fe/ZSM-5 (♦): T=350oC, 0.2% NO, 0.2% i-C4H10, 280 ml/min (Chen et al., 1998a); 3. Cat-1(▲): T=320oC, 377ppm NO, 1663ppm acetic acid, GHSV: 8.87s-1 (Lee, 2000); 4. V2O5/-Al2O3 (▼): T=300oC; 500ppm NO, 500-2250ppm acetic acid, GHSV: 5,000h-1 (Elkaim and Bai, 2000); 5. Co/FER (): T=450oC, 1600ppm NO, 1000ppm CH4, GHSV: 30,000h-1 (Li et al., 1994a).  To summarize, O2 influences the selectivity of HC-SCR catalyst. In most cases, NO conversion increased when the oxygen concentration was kept at low levels (0~2%). With the increase of oxygen concentration to beyond 2%, NO conversion decreased significantly. For most combustion processes, O2 concentration in the flue gas is higher than 2% (as listed in Table 1.1). Therefore, O2 will inhibit the activity/selectivity of most HC-SCR catalysts because excess oxygen reacts with hydrocarbons, leading to increased hydrocarbon consumption. In fact, many researchers considered O2 as one of the major factors negatively Chapter 2: Literature Review  28 impacting the performance of HC-SCR under lean-burn conditions. On the other hand, the presence of O2 is essential for the HC-SCR process because the presence of small amount O2 (e.g., <2%) is necessary for NO to be oxidized to NO2, and then adsorbed by the catalyst before it is reduced to N2 in the HC-SCR process.  2.5 Reactors for SCR of NOx The most commonly used reactors in the NH3-SCR process are packed bed reactors with honeycomb monolith, metal plate-type monolith, or coated metal monolith catalyst with parallel channels (Beretta et al., 1998; Forzatti, 2001). Compared to other packed bed reactors, the monolithic reactor has the advantages of low-pressure drop, superior attrition resistance and low tendency to fly ash plugging, and high specific surface area. However, because of the relatively lower gas/solid contacting efficiency, large amounts of catalysts need to be loaded into the reactor. The flux range of the flue gas is also restricted by the volume of catalysts. From the operating point of view, the replacement of catalysts is difficult. For NH3-SCR of NOx, fluidized bed reactors have been used for NOx or simultaneous SOx/NOx removal by various SO2 sorbents and NH3- or urea-SCR catalysts (Snip et al., 1996; Gao et al., 1996; Roh et al., 2003). Compared to conventional monolithic reactors, the catalyst loading is decreased, fly ash can easily pass through and the gas/solid contact efficiency is enhanced in the fluidized bed reactor. Furthermore, the replacement of catalysts does not require the SCR system to be shut down. On the other hand, the high pressure drop in the reactor and attrition of catalyst particles and adsorbent particles may limit the application of fluidized beds in NH3-SCR processes. Chapter 2: Literature Review  29 For the HC-SCR process, since the development of the catalyst and the screening of reducing agents are still the primary impediments for practical applications, most of studies were carried out in fixed bed reactors. Few researchers focused on the selection and application of reactor types for the HC-SCR process. Iwamoto et al. (1998) introduced a new method, intermediate addition of reductant (IAR), for the HC-SCR process. An oxidation catalyst (e.g. Pt/MFI), which oxidized NO to NO2, and a reduction catalyst (e.g. Zn/MFI), which reduced NOx to N2, was combined for the removal of NOx in the presence of excess oxygen. The conversion of NO to N2 at 300oC increased to 54% on combined Pt/MFI and Zn/MFI catalysts from 5% on Zn/MFI or 6% on Pt-MFI using ethane as the reducing agent. Perez-Ramirez et al. (2000) reported a dual-bed catalytic system for the combined deNOx-deN2O process. In this system, NO was removed in the first stage over Pt/AC catalyst with propane. To further reduce the large amounts of N2O produced in the first stage to N2, a second deN2O reactor bed was employed using ex-Co-Rh/Al-HTlc or Fe/ZSM-5 with propane. Efthimiadis et al (2001) reported a fluidized bed HC-SCR process (dimensions not mentioned) using 30g Al2O3, Rh/Al2O3 or 5:95Rh/Al2O3-FCC as the SCR catalyst, and CH3OH (1500ppm) or C3H6 (500ppm) as the reducing agent. The flue gas was the exhaust gas emitted from the regenerator of a FCCU, having following compositions: 80ppm NOx, 4% O2, 7% CO2, 225ppm CO and 235ppm SO2, with a total flow rate of 1000 ml/min treated in the SCR reactor.  Chapter 2: Literature Review  30 2.6 NOx storage-reduction (NSR) Recently, a new NOx removal process, NOx storage-reduction (NSR), gained much attention for the application under O2-rich lean-burn conditions, especially to truck exhaust gas. Although the mechanism of NSR is still not well understood, the common view is that, in the NSR process, NO is first oxidized to NO2 and then trapped during lean operations over a NOx adsorbent. The adsorbent is then regenerated with NOx released with the released NOx being reduced by carbon monoxide, hydrogen or hydrocarbons over the catalyst bed when the engine is run under stoichiometric or rich burn conditions for a short period of time. Similar to HC-SCR catalyst, some catalysts, such as Pt/Al2O3, had been proposed but showed insufficient activity for the NSR process (Olsson et al., 2001; Shinjoh et al., 1998). A new type of catalyst, which combines NOx adsorption and reduction materials, was extensively investigated. Typically, alkali or alkaline earth metals, such as barium and potassium, were chosen as NOx storage components, and noble metals, such as platinum and rhodium, as NOx oxidation and reduction components with the support of some metal oxides (Al2O3, etc.) (Fridell et al., 1999; Epling et al., 2003; Abdulhamid et al., 2004). Among them, Pt-Ba/Al2O3 was the mostly investigated NSR catalyst with CO, H2 or C3H6 as the reducing agent, which can provide high NOx removal efficiency (>90%) in successive sequences of lean and rich modes (Bogner et al., 1995). However, the inhibition of SO2 is still a problem (Hodjati et al., 1998a; Schreier et al., 2005; Hammache et al. 2008). Although NSR gained extensive attention as a potential solution to automobile exhaust gas NOx control because the engine can be relatively easily controlled to be run over the lean and rich cycles, it is difficult to apply NSR for the flue gas emitted from stationary sources because the oxygen concentration cannot be easily adjusted without upsetting the upstream process. Chapter 2: Literature Review  31 Similar to NSR, Elkaim and Bai (2000) proposed a method for removing O2 from the flue gas by adsorption/desorption in order to create an oxygen-free stream for the effective reduction of NOx. According to their patent, NO was first temporarily adsorbed by an adsorbent (activated carbon), then NO in the NO-rich adsorbent was desorbed by heating the adsorbent or by contacting an inert carrier gas (N2) to create an oxygen-free gas stream. NO in the oxygen-free gas stream was subsequently passed to a catalyst bed to have NO reduced by a reducing agent (hydrocarbons). This patent first reported the concept of separating HC- SCR process into two steps: adsorption and reduction.  2.7 Summary HC-SCR has been widely studied in recent years, with focus mainly on the preparation and evaluation of catalysts, effects of flue gas compositions and reducing agents, and reaction mechanisms. ZSM-5 and Al2O3 are commonly used catalyst supports in HC-SCR process, and many metals have been loaded onto catalyst supports as active elements, such as Pt, Ag, Cu, Co and Fe. Among reported catalysts in the literature, Fe/ZSM-5 gained extensive attention due to its stable reactivity and high hydrothermal durability. Although Pt/ZSM-5 and Ag/Al2O3 also exhibited good performance in HC-SCR, they are not economical for the treatment of the flue gas from stationary combustion processes due to the high cost in the preparation of these catalysts using Pt or Ag as active elements. Many hydrocarbons have been proven effective in HC-SCR, such as methane, propane, propylene, and iso-butane. Although methane has been widely investigated as the reducing agent in HC-SCR, it has been found to be effective only in SCR of NOx over Co/zeolites and some other catalysts at temperatures higher than 450oC. Propane, propylene Chapter 2: Literature Review  32 and iso-butane showed different performance in HC-SCR, depending on the type of catalyst and reaction conditions used in the experiment. If Fe/ZSM-5 is chosen as the HC-SCR catalyst, then propylene will be a good candidate as the reducing agent because of its relatively high reaction temperature window which can avoid the catalyst deactivation caused by the formation of a carbonaceous deposit on the catalyst surface at low temperatures (Chen et al., 1999). The poisioning components (SO2 and H2O) and excess O2 in the flue gas showed significant negative impacts on HC-SCR. Fe/ZSM-5, Pt/ZSM-5, Co/Beta-zeolite, Ag/Al2O3 and Ce-In/ZSM-5 exhibited a good hydrothermal resistance. However, only Co/Al2O3, Ag/Al2O3 and Cu/SUZ-4 showed some tolerance to SO2 at low concentrations. As a key component of the flue gas under lean burn conditions, the excess O2 inhibited the activity/selectivity of most catalysts in HC-SCR process. On the other hand, according to one of the reaction mechanisms, a small amount of O2 is necessary for NOx reduction.    33  Chapter 3 Adsorption and Reaction Kinetics  3.1 Experimental setup To investigate the reaction kinetics and adsorption performance of selected catalysts under various temperatures and flue gas compositions, a fixed bed reaction system was constructed as shown schematically in Figure 3.1. The system consists of a tubular reactor, a gas supply and flow rate control unit, a gas preheating unit, a gas heating unit (furnace), a gas analysis unit and a data acquisition unit. A stainless steel tubular reactor with an inner diameter of 13.5mm was used in this experiment, and the catalyst packing height was maintained at 50.8 mm throughout all the tests. PressureTransducer  Heati ng Regul ator O 2 CO 2 NO N 2 HC H 2 O    Wat er Evaporat or     Gas Preheater   Te mp. Controll er Ther mocoupl e Pressure Transducer Vent   Gas Anal yzer Computer Tubul ar React or Furnace Ther mocoupl es      ︵ 3 li nes︶ SO 2  Figure 3.1 The fixed bed reaction system  Chapter 3: Adsorption and Reaction Kinetics  34 The model flue gas used in the experiment was a mixture prepared from several gas cylinders: 50% O2 balanced with N2, 0.6% NO balanced with N2, 10% SO2 balanced with N2, pure CO2 and pure N2 gas cylinders from Praxair Products Inc. The reducing agent used in the experiment was propylene. The gas cylinder containing 1.2% propylene balanced with N2 was also supplied by Praxair. Deionized water was pumped into the system using a LKB Microperpex peristaltic pump, heated and evaporated through a water evaporator, and then pre-mixed with other gases if necessary. The compositions of the effluent gases (NOx, CO, CO2, and SO2) were analyzed by a Horiba PG-250 flue gas analyzer. In the presence of oxygen, NO in the model flue gas will be partially oxidized to NO2. Since the gas analyzer measures NO and NO2 together, all calculations related to NO adsorption and conversion throughout this work were based on NOx (NO + NO2) although only NO was fed into the reaction system. Before the start up of the experiment, the reactor was first heated to the desired temperature with pure N2 gas passing through the catalyst bed. After the reactor temperature was stabilized, the flow of a gas mixture at a preset composition was turned on to start the experiment, with the gas mixture composition determined by the gas analyzer via bypassing the gas from the reactor. In the adsorption experiment, the model flue gas was pre-mixed with pure N2 as the balancing gas according to the preset flow rate and concentrations of NO and O2. The mixed flue gas was then preheated to 150oC by the preheating system, and the composition was measured by the gas analyzer. After stable readings of the analyzer were reached, which corresponded to the inlet model flue gas concentrations, the three way valves were switched to the reactor and the time was recorded as the starting point of the adsorption. During the adsorption process, the gas composition at the reactor’s outlet was continuously monitored Chapter 3: Adsorption and Reaction Kinetics  35 until the NOx concentration became stable, with the final NOx concentration defined as the equilibrium NOx concentration. Thereafter, the NO flow from the gas cylinder was turned off and, at the same time, HC (i.e. propylene) at the same flow rate of NO was added into the O2 + N2 flow to start the catalytic reduction process by reacting propylene with the adsorbed NOx on the catalyst. After 45 minutes, the HC flow was turned off and the gas composition passing through the catalyst bed was adjusted to 10% O2 + balanced N2 at a flow rate of 500 ml/min to further remove the adsorbed HC or other intermediate species produced in the reduction process. This stripping process took 30~45 minutes or even longer to lower the outlet CO concentration to below 5 ppm. Afterward, pure N2 was used to purge the catalyst bed for another 45 minutes to remove all adsorbed COx and O2 from the catalyst. In the reaction kinetics experiment, the flue gas was prepared in the same way as in the adsorption experiment, and the HC flow at the desired HC:NO ratio was injected right after the pre-heater to have it mix with other model flue gases and to minimize the oxidation of the HC outside the reactor. The composition of the gas mixture was first measured via bypassing the gas mixture to the gas analyzer before the gas flow was switched to the reactor for the reaction experiment. After the reaction was finished, the gas flow was switched back to the bypass route to have the inlet gas concentration checked.  3.2 Catalyst preparation In most HC-SCR processes using ZSM-5 as the catalyst support, NH4/ZSM-5 or H/ZSM-5 was first prepared from Na/ZSM-5 via wet ion-exchange method using aqueous NH4NO3 solution, followed by calcination in air at 500oC for 4 hours. As the active element in the SCR process, Fe is often loaded onto the catalyst support in four ways: conventional wet ion-exchange (WIE), incipient wetness impregnation in Chapter 3: Adsorption and Reaction Kinetics  36 aqueous solution (IMPA), impregnation in organic solution (IMPO) and sublimation or chemical vapour deposition (CVD). FeCl2, FeCl3, Fe(NO3)3, Fe2(SO4)3 and iron (III) acetylacetonate (Fe(AA)3) are most common used iron sources. It was reported that catalysts prepared by both WIE and IMPA methods had similar activity for NH3-SCR of NOx, except that the IMPA method was simpler and less costly than WIE method (Long and Yang, 1999; Qi and Yang, 2005). Although Fe/ZSM-5 prepared by CVD method showed high catalytic activity and stability in HC-SCR of NOx (Chen et al, 1998), it is difficult to implement this process for the production of large quantities of catalysts considering the complexity of the preparation process. Fe/ZSM-5 prepared by IMPO method was first used in the study of NH3-SCR process by Delahay et al. (2005). A recent study from Lima et al. (2008) showed that Fe/ZSM-5 prepared by IMPO demonstrated high activity in NH3-SCR and acceptable NO conversion in HC-SCR using n-decane as the reducing agent. Based on all the above aspects, WIE and IMPO were selected for the preparation of the catalysts in this study.  3.2.1 Materials Two types of ZSM-5 were selected as the support of the catalyst in this experiment, i.e., Na/ZSM-5 (PUC) purchased from The China University of Petroleum (Beijing, China) and a free sample of H/ZSM-5 (Albemarle) kindly supplied by Albemarle Corporation (USA). It should be noted that both particles are not pure ZSM-5 but a mixture of pure ZSM-5 and additives which were added in the prilling process. To examine the influence of particle size, the coarse Na/ZSM-5(PUC) was crushed and sieved to obtain a fine Na/ZSM-5 (crushed PUC). A spent FCC catalyst was also selected as the catalyst support. The properties of the catalyst supports are listed in Table 3.1. Chapter 3: Adsorption and Reaction Kinetics  37  Table 3.1 Properties of the catalyst supports Catalyst support Average particle size Apparent bulk density SBET Provider Na/ZSM-5 (PUC) 1042 µm 903 kg/m3 118 m2/g China University of Petroleum H/ZSM-5 (Albemarle) 155 µm 968 kg/m 3 171 m2/g Albemarle Corp.(USA) Na/ZSM-5 (Crushed PUC) 234 µm 764 kg/m 3 192 m2/g Crushed from original Na/ZSM-5 (PUC) Spent FCC 116 µm 922 kg/m3 N/A Chevron Refinery  The following chemicals were used in the preparation of the catalysts:  Ammonium nitrate : NH4NO3, 99.0%, M.W.=80.04 g/mol, Sigma-Aldrich  Iron (III) acetylacetonate (Fe(AA)3): [CH3COCH=C(O-)CH3]3Fe, 97%, M.W.=353.18 g/mol, Sigma-Aldrich  Ferrous chloride: FeCl2 · 4H2O, 102.0% (as FeCl2 · 4H2O), M.W.=198.81 g/mol, Fisher  Toluene: C7H8, >=99.5%, M.W.=92.14 g/mol, Sigma-Aldrich  Deionized water  3.2.2 Preparation of NH4/ZSM-5 and H/ZSM-5 NH4/ZSM-5 was prepared from Na/ZSM-5 by exchanging with NH4NO3 solution. 100g of Na/ZSM-5 was mixed with 100ml of 0.5M NH4NO3 solution at the room temperature, and the slurry was stirred periodically. After 3 hours, the catalyst in the slurry was separated from the NH4NO3 solution, mixed with another batch of fresh 0.5M NH4NO3 solution (100ml). After 3 times of the aqueous ion-exchange process, the catalyst was washed thoroughly with 250ml of deionized water for 5 times, dried in air at 120oC for 12 hours to form NH4/ZSM-5, and then calcined in air at 500oC for 4 hours to form H/ZSM-5. Chapter 3: Adsorption and Reaction Kinetics  38  3.2.3 Preparation of Fe/ZSM-5 by WIE method 20g NH4/ZSM-5 was added to 500ml of 0.5M FeCl2 solution with constant stirring. After 24 hours, the mixture was filtered and then washed five times with deionized water, dried at 120oC for 12 hours and then calcined in air at 500oC for 6 hours.   3.2.4 Preparation of Fe/ZSM-5 by IMPO method 56g H/ZSM-5 was added to a solution containing 20g Fe(AA)3 + 200 ml toluene. The slurry was stirred periodically for 24 hours. Toluene was then evaporated from the slurry and recycled, and the residue after the evaporation was dried in air at 120oC for 12 hours and calcined in air at 500oC for 6 hours.  3.3 Performance of Fe/ZSM-5 (PUC) catalyst prepared by IMPO 3.3.1 Adsorption performance of Fe/ZSM-5(PUC) The adsorption isotherm behaviour of the Fe/ZSM-5(PUC) catalyst prepared by IMPO method was investigated under various temperatures and inlet NO concentrations using a model flue gas with 5% (v/v) O2 balanced with N2 at GHSV=5000 h-1. The catalyst loading in the reactor was 6.58g. The NO concentration of the model flue gas varied from 200 to 1000 ppm with an increment of 200 ppm and the temperature from 250 to 400oC with an increment of 30oC. For each given temperature and gas composition, the adsorption curve was determined experimentally, with the result at T=250oC shown in Figures 3.2. Adsorption curves at other temperatures are shown in Appendix G (Figures G.1 to G.5) Chapter 3: Adsorption and Reaction Kinetics  39  Figure 3.2 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=250oC)  Since there exists a delay in system response time for the gas analyzer, the delay time needs to be subtracted from the recorded adsorption time. Because the response time varied with the operating temperature, the delay time was determined separately for each operating temperature. It is clear from Figure 3.2 that, at a given temperature, with the increase of equilibrium NOx concentration (C0), the adsorption curve moved towards the left, indicating that the catalyst bed was saturated more quickly by NOx at higher C0. The equilibrium adsorption capacity of NOx on the catalyst was calculated by .0 3 0 )273( 10 cat NO e W S TR MFCPq          (3.1) where qe, adsorption equilibrium capacity of NOx on the catalyst, mg/g P, operating pressure, Pa Chapter 3: Adsorption and Reaction Kinetics  40 C0, equilibrium NOx concentration, ppm F, model flue gas flow rate at room temperature, m3/s MNO, molecular weight of NO, 30 g/mol R, universal gas constant, 8.314 J/(mol.K) T0, ambient temperature, oC S, integrated area of the adsorption curve, s Wcat., catalyst loading, g The integrated area, S, of the adsorption curve was calculated by dt C C S t outNOx )1( 0 0 ,        (3.2) where CNOx,out, the outlet concentration of NOx, ppm t, adsorption time, s The Freundlich equation  0Cqe         (3.3) was applied to fit qe as a function of C0. The calculated qe and the curve fitting results using Freundlich equation were shown in Figure 3.3. For a given temperature, the adsorption capacity of the catalyst increased significantly with the increase of the equilibrium NOx concentration. At a given equilibrium NOx concentration, the adsorption capacity of the catalyst decreased with the increase of the adsorption temperature. Using α and β fitted in Figure 3.3, correlations of α and β as a function of temperature were derived (Figure 3.4) by least-square curve fitting with a second-order polynomial function: 27107123.800123.046179.0 TT      (3.4) 251081409.1024.013236.7 TT      (3.5) Chapter 3: Adsorption and Reaction Kinetics  41  Figure 3.3 Fitted adsorption isotherms of NOx by Freundlich equation (Catalyst: Fe/ZSM- 5(PUC))  Figure 3.4 Relationship between α/ and adsorption temperature (Catalyst: Fe/ZSM-5(PUC)) Chapter 3: Adsorption and Reaction Kinetics  42  3.3.2 Reaction performance of Fe/ZSM-5(PUC) To investigate the influence of temperature, gas hourly space velocity (GHSV), HC and O2 concentrations on the SCR performance of the catalyst, the HC-SCR of NOx was conducted in the tubular reactor. The model flue gas consisted of 600ppm NO, 0.5% to 8% (v/v) O2, balanced with N2. The HC:NO molar ratio in the model flue gas ranged from 0.5:1 to 4:1. The time-on-stream test was also performed to evaluate the durability and stability of the catalyst. As a promising catalyst, Fe/ZSM-5 has shown effectiveness in the SCR of N2O by propylene (Yamada et al., 1998). Since N2O can be converted completely to N2 over Fe/ZSM-5 catalyst under experimental conditions similar to this study, and only NO was used in the model flue gas in this study, the concentration of N2O in the flue gas exiting the reactor was expected to exist only at very low level in comparison with NOx. Therefore, N2O in the stream exiting the reactor was not monitored in this study. The NOx conversion was calculated based on measured inlet and outlet NOx concentrations by %100 , ,,  inNO outNOinNO NO x xx x C CC X      (3.6)  Because the gas analyzer used in the experiment can only measure the concentrations of CO and CO2, the conversion of propylene was calculated by carbon balance based on the measured concentrations of CO and CO2 at the reactor’s outlet,  %100 )(3 10(%))( , 4 ,, 2   ppmC CppmC X inHC outCOoutCO HC    (3.7)  Chapter 3: Adsorption and Reaction Kinetics  43 The inlet concentration of propylene (CHC,in) in equation 3.7 was estimated based on the measured flow rate from the certified (1.2% propylene + N2) gas cylinder and the total feed gas flow rate. For example, if the total flow rate of gas feed is 600 ml/min and 60 ml/min is from the (1.2% HC + N2) cylinder, the HC concentration in the feed gas should be 60*1.2%/600=0.12% or 1200ppm. A coefficient of 3 is included in equation 3.7 because one mole of propylene could produce 3 moles of carbon. It should be noted that other intermediate species (or incompletely oxidized products) might be generated from propylene, which had not been monitored and considered in the calculation for HC conversion. Thus, equation 3.7 may underestimate the HC conversion.  3.3.2.1 Effect of reaction temperature on catalytic activity Using 600 ppm NO + 1200 ppm HC + 1% O2 + balanced N2 as the reacting gas mixture, the time-on-stream test at a gas space velocity of 5000 h-1 was conducted on the Fe/ZSM-5 (PUC) catalyst. Typical profiles of NOx and HC conversions and the outlet concentration of CO at 250oC and 350oC are shown in Figures 3.5 and 3.6, respectively. Results at other temperatures are shown in Figures G.13 to G.16 in Appendix G.  Chapter 3: Adsorption and Reaction Kinetics  44  Figure 3.5 Profiles of NOx and HC conversions and outlet CO concentration (Catalyst: Fe/ZSM-5(PUC), T=250oC, [O2]=1%)  Chapter 3: Adsorption and Reaction Kinetics  45  Figure 3.6 Profiles of NOx and HC conversions and outlet CO concentration (Catalyst: Fe/ZSM-5(PUC), T=350oC, [O2]=1%)  Chapter 3: Adsorption and Reaction Kinetics  46  It should be noted that the catalyst was calcined in 20% O2 + N2 at 500oC for 2 hours to fully regenerate the catalyst prior to each test at a given temperature. As shown in Figure 3.5, the NOx conversion increased quickly to a maximum and then decreased steadily with time, indicating a decaying catalytic activity at T=250oC. Meanwhile, HC conversion was low and decreased with time, which is also reflected by the low and decreasing outlet CO concentration. It was found that CO was the main product from HC oxidation and almost no CO2 was detected at this temperature. When the reaction temperature was increased to 350oC, as shown in Figure 3.6, both NOx and HC conversions and outlet CO concentration kept relatively stable within the tested period. Furthermore, the outlet CO concentration and HC conversion increased very quickly with the increase in temperature. At this temperature, the CO to CO2 ratio in the flue gas exiting the reactor was 58%:42%. Although CO2 increased significantly, it was still lower than CO in the exhaust gas stream. This result agreed with those reported in the literature using Fe/ZSM-5 as the deNOx catalyst (Chen et al, 1998). To investigate the effect of the reaction temperature on the catalytic performance of Fe/ZSM-5(PUC), the time-on-stream test was conducted with the temperature ranging from 250 to 350oC with an increment of 25oC, with the results shown in Figures 3.7 and 3.8 for NOx and HC conversions, respectively. Chapter 3: Adsorption and Reaction Kinetics  47  Figure 3.7 Effect of reaction temperature on NOx conversion (Catalyst: Fe/ZSM-5(PUC))   Figure 3.8 Effect of reaction temperature on HC conversion (Catalyst: Fe/ZSM-5(PUC))  Chapter 3: Adsorption and Reaction Kinetics  48 Peak NOx conversions at different reactor temperatures were around 70.5% (250oC), 75% (74.4% for repeated run) (275oC), 70.2% (300oC), 68.2% (325oC) and 65.6% (350oC), respectively, for the current Fe/ZSM-5 (PUC) catalyst, with the highest NOx conversion achieved at 275oC. With the further increase of the temperature beyond 275oC, the peak NOx conversion decreased. However, at lower temperatures (<325oC), the catalyst activity was quite unstable and the conversion decreased very quickly with time. This result is consistent with findings from Chen et al. (1999, 2000b) using iso-butane or n-butane as the reducing agent. According to their study, the deactivation of Fe/ZSM-5 was caused by the formation of a carbonaceous deposit which was not swiftly volatilized at a low temperature (300oC). The durability or stability of the catalyst improved with the increase of reaction temperature. At T=350oC, the catalyst showed stable, yet lowest NOx conversion. This means that to achieve a high NOx conversion, a low temperature is preferred. On the other hand, the stable catalytic activity could be achieved only at high temperatures. Considering the requirement of catalyst durability in practical applications, coupling with the desired high reaction activity of NOx reduction, the optimal reaction temperature should be around 350oC for the Fe/ZSM-5(PUC) catalyst prepared by the IMPO method. Compared to NOx conversion in Figure 3.7, HC conversion always increased with increasing reaction temperature (Figure 3.8). Peak HC conversions were 17.2% (250oC), 38.2% (34.8% for repeated run) (275oC), 40.2% (300oC), 49.6% (325oC) and 71.0% (350oC), respectively. This is because the oxidation reaction generally increased with increasing reactor temperature. It is noted that when the reactor was operated in the right temperature window, the catalytic activity kept stable and the consumption of hydrocarbon by the SCR process also remained at a relatively stable level. On the contrary, when the catalyst lost its activity gradually with time at low temperatures, the amount of hydrocarbon consumed in the Chapter 3: Adsorption and Reaction Kinetics  49 SCR process also decreased, leading to a decrease of HC conversion with time. The peak HC conversion did not always correspond to the peak NOx conversion at each temperature, because HC was catalytically oxidized by both NOx and O2.  3.3.2.2 Catalyst deactivation at low temperatures Further investigation revealed that the catalyst deactivation at temperatures lower than 325oC might be induced by intermediates generated from the incomplete oxidation of hydrocarbons. As shown in Figure 3.9, when O2 concentration in the flue gas increased from 1% to 4%, the peak NOx conversion decreased from 75% to 67.9%, but the catalyst activity remained stable for a much longer time at higher O2 level. One may infer that the intermediate species of the hydrocarbon and carbon deposit produced in the NOx reduction process gradually covered the active sites of the catalyst in the reaction process, and resulted in the catalyst deactivation at low O2 concentrations. At high O2 concentrations, the intermediate species were oxidized by excessive O2 and thus left the catalyst sites active for the adsorption of NOx. Similarly, high temperatures can promote the complete combustion of the HC intermediate species and thus reduce the amount of hydrocarbon intermediates and carbon deposit on active sites. Chapter 3: Adsorption and Reaction Kinetics  50  Figure 3.9 Effect of inlet O2 concentration on the catalytic activity (Catalyst: Fe/ZSM- 5(PUC))  To elucidate the catalyst deactivation, the single point BET surface area of fresh, deactivated and regenerated catalysts under T=275oC was measured and shown in Figure 3.10. The deactivated catalyst was regenerated in the flow of 20% O2 +N2 at T=500oC for 2 hours, followed by N2 flush for 20 minutes. The instrument used for measuring the catalyst surface area was a Micromeritics Flowsorb II 2300 with a gas flow of 30%V N2 + 70%V He. Chapter 3: Adsorption and Reaction Kinetics  51  Figure 3.10 Comparison of BET surface areas (Catalyst: Fe/ZSM-5 (PUC))  The impregnation of Fe led to a slight decrease in the surface area from 118 m2/g (parent Na/ZSM-5) to 110 m2/g (fresh Fe/ZSM-5). The surface area of the deactivated catalyst decreased significantly to 68 m2/g, and was almost completely recovered to 109 m2/g after regeneration. Since only NO, propylene, O2 and N2 were introduced into the reactor, the only explanation for the decrease of the BET surface area of the deactivated catalyst was the deposit of the partially oxidized hydrocarbon intermediates or graphitic carbon. Further evaluation was carried out based on X-ray photoelectron spectroscopy (XPS) on the surface element of fresh, deactivated and regenerated catalysts, with differences identified in the C 1s region as shown in Figure 3.11 (see also Figures G.20 to G.26 in Appendix G for other XPS results). Chapter 3: Adsorption and Reaction Kinetics  52  Figure 3.11 XPS narrow scan for C 1s (Catalyst: Fe/ZSM-5(PUC))  The binding energy corresponding to the peak intensity in the narrow spectra scan for C 1s region of the fresh, deactivated and regenerated catalyst were 285.05, 284.40 and 285.06, respectively. The atomic percentages of the C 1s on the surface of the catalysts were 23.39%, 24.76%, and 22.06%, respectively, for fresh, deactivated and regenerated catalyst. Because of the existence of the adsorbed gaseous carbon, i.e., CO2, the difference among the atomic percentages of the C 1s on the surface of the catalysts could not prove the existence of the graphitic carbon. However, the peak position shift from the higher binding energy for the fresh and regenerated catalysts to the lower binding energy for the deactivated catalyst may be attributed to the deposit of the graphitic carbon on the catalyst surface which covered the active sites of the catalyst, and was considered as the possible reason for the deactivation of the catalyst. Chapter 3: Adsorption and Reaction Kinetics  53 To further investigate the possible reason for the deactivation of the catalyst at low temperatures, a thermogravimetric analysis (TGA) was performed using a Shimazu TGA-50 thermogravimetric analyzer. In the experiment, catalyst samples were aged in the tubular reactor for 6 hours by contacting the flue gas consisting of 600ppm NO + 1200 ppm HC +1% O2 + N2 at 275 and 350oC, respectively. The sample was then placed into the TGA under a N2 atmosphere and heated up from room temperature to 250oC with a ramp of 20oC/min. The temperature was maintained at 250oC for 10 minutes to volatilize adsorbed gas components. Thereafter, the temperature was increased to 500oC with a ramp of 5oC/min and maintained at 500oC for 30 minutes. Weight changes during above period were recorded and are shown in Figure 3.12 for the sample aged at 275oC and Figure 3.13 for the sample aged at 350oC. Both samples showed some weight loss (-2.03% in Figure 3.12 and -1.53% in Figure 3.13) when heated up from room temperature to 250oC, which indicates that both samples contained some adsorbed gas components. Over the period of the temperature rise from 250 to 500oC and stablized at 500oC, the catalyst aged at 275oC lost 1.38% of its total weight while the catalyst aged at 350oC lost only 0.23% of its weight, which suggests that more carbonaceous compounds were deposited onto the catalyst at 275oC than at 350oC, and led to catalyst deactivation. Also, most of deposited carbonaceous compounds on the catalyst aged at 350oC could be removed at relatively lower temperatures than those deposited on the catalyst aged at 275oC, implying that catalyst aged at 350oC could be regenerated at lower temperatures or over a shorter period of time than catalyst aged at 275oC. Chapter 3: Adsorption and Reaction Kinetics  54  Figure 3.12 TGA result for the catalyst aged at 275oC   Figure 3.13 TGA result for the catalyst aged at 350oC Chapter 3: Adsorption and Reaction Kinetics  55 To improve the recovery of the catalytic activity of the catalyst deactivated at low reaction temperatures, several regeneration methods were tested at 275oC with the results shown in Figure 3.14.  Figure 3.14 Recovery of catalyst activity under different regeneration conditions (Catalyst: Fe/ZSM-5(PUC))  At 275oC, the reacting gas mixture was switched off after the catalyst was aged and deactivated, then the catalyst was regenerated by passing 20% O2 + N2 for 40 minutes at the same temperature (i.e. 275oC). It is seen in Figure 3.14 (line A) that the catalytic activity could not be fully recovered after regeneration, with a highest NOx conversion of only 49.7% for the regenerated catalyst. When the same regeneration method was used for the catalyst aged at 350oC (line B in Figure 3.14), the peak NOx conversion increased to 64.9%, which was still much lower than the fresh catalyst (75%). The catalyst activity could be fully recovered only if the catalyst was calcined in air or 20% O2 + N2 at 500oC or higher Chapter 3: Adsorption and Reaction Kinetics  56 temperatures to have the deposited graphitic carbons and other adsorbed intermediate species of HC removed from the catalyst surface completely.  3.3.2.3 Effect of GHSV Effects of gas space velocities on NOx conversion at 325oC and 350oC are shown in Figures 3.15 and 3.16.   Figure 3.15 Effect of GHSV on catalytic activity (Catalyst: Fe/ZSM-5(PUC), T=325oC)  Chapter 3: Adsorption and Reaction Kinetics  57  Figure 3.16 Effect of GHSV on catalytic activity (Catalyst: Fe/ZSM-5(PUC), T=350oC)  It is worth noting that the gas hourly space velocity (GHSV) was calculated based on the gas flow rate at room temperature: hD FGHSV i   2 4 60        (3.8) where GHSV, gas hourly space velocity, h-1 F, gas flow rate at room temperature, ml/min Di, inner diameter of the reactor, 13.5 mm h, catalyst packed height in the reactor, 50.8 mm  At T=325oC (Figure 3.15), the peak NOx conversions were 68.2% (GHSV=5000 h-1), 61.3% (GHSV=7500 h-1), 57.6% (GHSV=10000 h-1), 55.0% (GHSV=15000 h-1) and 50.3% (GHSV=20000 h-1). At GHSV=5000 h-1, the catalyst showed relatively stable activity. Chapter 3: Adsorption and Reaction Kinetics  58 At T=350oC (Figure 3.16), the peak NOx conversions were 65.6% (GHSV=5000 h-1), 59.5% (GHSV=7500 h-1), 55.7% (GHSV=10000 h-1), 52.2% (GHSV=15000 h-1) and 47.1% (GHSV=20000 h-1). The highest NOx conversion was achieved at GHSV=5000 h-1, and the conversion remained stable over the tested time period. For a given GHSV, NOx conversion at T=325oC was always 2~3% higher than at T=350oC. However, at both temperatures, with the increase of GHSV, the peak NOx conversion decreased. At the same time, at a higher GHSV, NOx conversion decreased more quickly with time. This indicates that the space velocity has a significant influence on the catalyst activity, and must be considered as a key operating parameter in the design and operation of the dual-zone reactor.  3.3.2.4 Effect of O2 and HC concentrations Effects of O2 and HC concentrations on the catalytic activity at temperatures from 275oC to 375oC are shown in Figures 3.17 to 3.21. As discussed previously, the catalytic activity was unstable for the reaction temperatures lower than 325oC. Therefore, NOx conversions shown in Figures 3.17 to 3.21 were the peak conversions obtained at each temperature.  Error bars shown in Figure 3.17 are calculated based on the standard deviation method. Four to eight repeated runs were used for the calculation of NOx and HC conversions to obtain the standard deviation, with two times of standard deviations corresponding to 95% of confidence level plotted in this figure. The possible sources of errors include the measurement of gas concentrations by the gas analyzer, rotameter readings of gas flow rates and the reproducibility of the experiment. Throughout all fixed bed experiments, the errors for calculated NOx conversions ranged from 0.37% to 2.86%, with an average around 1.42%. Due to the large number of figures in this work, error bars were added only in a few selected figures. Chapter 3: Adsorption and Reaction Kinetics  59  (a) HC conversion  (b) NOx conversion Figure 3.17 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=275oC, 95% confidence level) Chapter 3: Adsorption and Reaction Kinetics  60  For T=275oC shown in Figure 3.17, the NOx conversion was very low at HC:NO=0.5, ranging from 28.9% (8% O2) to 35.1% (1% O2) while the HC conversion was high from 88.8% (0.5% O2) to 100% (≥4% O2). The increase of O2 concentration had no remarkable impact on both NOx and HC conversions. When the HC:NO ratio was increased to 1, NOx reduction was greatly enhanced. For low O2 concentrations (≤2%), NOx conversion increased to 56.1% (2% O2) and 58.1% (1% O2). On the contrary, the HC conversion dropped significantly to 46.9% (0.5% O2) and 74.9% (2% O2). At high O2 concentrations (≥4%), the increase of HC:NO led to an increase of NOx conversion to 48.9% (4% O2)  and 47.8% (8% O2), but the influence of the change of HC:NO was not as significant as that for O2≤2%. Although the HC conversion also decreased to 84.6% (4% O2) and 96% (8% O2), it was still much higher than that at 0.5% O2. Further increase of HC:NO to 1.5, 2 and 4 showed similar influence on NOx and HC conversions as in the case with HC:NO=1, except for that at HC:NO≥2, the highest NOx conversion occurred at 2% O2 (75% for HC:NO=2 and 77.9% for HC:NO=4). In general, for a given O2 concentration, HC conversion decreased very quickly with the increase of HC:NO ratio. The increase of HC:NO ratio from 0.5 to 1.5 could greatly improve the catalytic activity  for the flue gases containing 0.5% to 8% O2 at T=275oC. For O2≤1%, the increase of HC:NO from 1.5 to 4 had little influence on NOx conversion. For O2≥2%, the NOx conversion improved significantly with the increase of HC:NO from 1.5 to 4. The same trend was observed for T=300, 325, 350 and 375oC, as shown in Figures 3.18 to 3.21, except that the highest NOx conversion was always obtained at 0.5% O2 for HC:NO=1 to 4. Chapter 3: Adsorption and Reaction Kinetics  61  (a) HC conversion  (b) NOx conversion Figure 3.18 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=300oC) Chapter 3: Adsorption and Reaction Kinetics  62  (a) HC conversion  (b) NOx conversion Figure 3.19 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=325oC) Chapter 3: Adsorption and Reaction Kinetics  63  (a) HC conversion  (b) NOx conversion Figure 3.20 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=350oC) Chapter 3: Adsorption and Reaction Kinetics  64  (a) HC conversion  (b) NOx conversion Figure 3.21 Effect of HC:NO ratio and O2 level on catalytic activity (Catalyst: Fe/ZSM- 5(PUC), T=375oC) Chapter 3: Adsorption and Reaction Kinetics  65  For all cases tested, the increase of HC concentration enhanced NOx conversion significantly. NOx conversion remained at a very low level (between 15% to 35%) when HC:NO molar ratio was 0.5 at all temperatures and O2 concentrations. With the increase of HC:NO to 2, NOx conversion increased very quickly, especially for those at low O2 concentrations (≤2%). However, further increase of HC:NO from 2:1 to 4:1 had little influence on NOx conversion at low temperatures (≤300oC) compared to at high temperatures (≥325oC). For HC:NO=0.5, the increase of O2 concentration had less impact over the investigated temperature range. At a given HC:NO ratio, the increase of O2 concentration showed more negative impact on NOx conversion when HC:NO≥1, especially at high temperatures (≥325oC). However, at low temperatures (≤300oC), NOx conversion remained almost independent of HC:NO ratio at O2 concentrations ≤2%.   For a given O2 concentration, the increase of HC:NO ratio led to a significant decrease of HC conversion at all temperatures. In contrast, for a given HC:NO, the HC conversion increased with the increase of O2 concentration in the model flue gas. It is generally anticipated that high HC to NO ratio can lead to a high NOx conversion, at a cost of high hydrocarbon consumption which is not economical for the practical application. From this point of view, a HC to NO molar ratio of 2:1 appears to be a good choice. Oxygen played an important role in the HC-SCR under lean-burn conditions, which agrees well with the results from other researchers (Chen et al, 1998; Lee, 2000; Elkaim and Bai, 2000).  Chapter 3: Adsorption and Reaction Kinetics  66 3.4 Reaction performance of Fe/ZSM-5 (PUC) catalyst prepared by WIE Fe/ZSM-5(PUC) was prepared by the WIE method to evaluate the catalyst performance prepared by different methods. The time-on-stream test was performed at 275oC and 350oC with the result shown in Figure 3.22.  Figure 3.22 Effect of temperature on catalytic activity for NOx reduction (Catalyst: Fe/ZSM- 5(PUC, WIE))   In the experiment, the inlet NOx concentration was first measured by bypassing the model flue gas to the gas analyzer. When switching the model flue gas to the reactor, there was a time delay ranging from 5 to 10 minutes before the real concentration of NOx exiting from the reactor was obtained from the analyzer, which caused the initial increase of NOx conversion in the time-on-stream experiment as shown in Figure 3.22. This part of time-on- stream data was not considered in the analysis of experimental data. Chapter 3: Adsorption and Reaction Kinetics  67 At T=275oC and with a feed gas composition of 600ppm NO + 1200ppm HC + 4%O2 + N2 and GHSV=5000 h-1, the peak NOx conversion for the catalyst prepared by WIE (Fe/ZSM-5(PUC, WIE)) was 68.1%, which was almost the same as the catalyst prepared by IMPO method (Fe/ZSM-5(PUC, IMPO)). However, the NOx conversion decreased quickly to 49.2% in 100 minutes, in comparison to 66.1% for Fe/ZSM-5(PUC, IMPO) within the same time period as shown in Figure 3.9. Previous results showed that Fe/ZSM-5(PUC) prepared by IMPO method exhibited stable activity at 350oC, even for the flue gas containing only 1% O2. Although the peak NOx conversion for Fe/ZSM-5 (PUC, WIE) at 350oC in Figure 3.22 was 57.6%, 11% higher than Fe/ZSM-5(PUC, IMPO), the activity of Fe/ZSM-5(PUC, WIE) was not stable even for the feed gas containing as high as 4% O2. The effect of the reaction temperature on NOx and HC conversions for Fe/ZSM- 5(PUC, WIE) with the feed gas composition of 600ppm NO + 1200ppm HC + 4% O2 + N2 and GHSV=5000 h-1 is shown in Figure 3.23. Again, only the peak NOx conversion and corresponding HC conversion were plotted. The catalyst was calcined in 20% O2 + N2 at 500oC for 2 hours prior to each run in order to recover the catalytic activity. Chapter 3: Adsorption and Reaction Kinetics  68  Figure 3.23 Effect of temperature on NOx and HC conversion (Catalyst: Fe/ZSM-5(PUC, WIE))  The highest peak NOx conversion was 71.3% (T=250oC) within the tested temperature window, and the peak NOx conversion decreased to 47.3% when the reaction temperature was increased to 400oC. The HC conversion increased with the increase in the reaction temperature, but only reached 43% at T=350oC, which was much lower than Fe/ZSM-5(PUC, IMPO). The catalyst was observed to be unstable, even at 400oC. Although the peak NOx conversion was higher than Fe/ZSM-5(PUC, IMPO) at the same reaction condition, the poor stability prevents it from the practical application in the HC-SCR. It was thus concluded that IMPO is a better method than WIE for the preparation of Fe/ZSM-5 catalyst using current Na/ZSM-5(PUC) particles.  Chapter 3: Adsorption and Reaction Kinetics  69 3.5 Performance of Fe/ZSM-5(Albemarle) catalyst Although the Fe/ZSM-5(PUC) catalyst prepared via IMPO method exhibited good performance for the catalytic reduction of NOx, the adsorption performance of Fe/ZSM- 5(PUC) was not satisfactory because the equilibrium adsorption capacity was lower than 0.07 mg/g cat. with a very short breakthrough time of the catalyst bed ranging from 15 to 60 seconds. To increase the adsorption capacity of the catalyst, the parent Na/ZSM-5(PUC) particles were crushed and sieved to obtain fine particles with an average particle size of 234m and BET surface area of 192 m2/g, which were then used to prepare the fine Fe/ZSM- 5(crushed PUC) catalyst by the IMPO method. As discussed in Appendix E, the fine Fe/ZSM-5(crushed PUC) catalyst showed several times the NOx adsorption capacity than the coarse Fe/ZSM-5(PUC) (See Figure E.4). Therefore, it would be desirable to use fine catalyst particles in order to have both a high adsorption capacity for NOx capture in the adsorption zone and an acceptable catalytic activity for NOx reduction in the reduction zone for the proposed dual zone deNOx reactor. Another type of fine particles, a spent Fluid Catalytic Cracking (FCC) catalyst, was also evaluated as a potential catalyst support. Unfortunately, as shown in Appendix F, both spent FCC and Fe/FCC showed very poor catalytic activity in HC-SCR. A newly developed FCC catalyst support with ZSM-5 structure was obtained as a free sample from Albemarle Corporation. Fe/ZSM-5(Albemarle) catalyst was then prepared via the standard IMPO method, and its adsorption and reaction performances were investigated in the tubular reactor.  Chapter 3: Adsorption and Reaction Kinetics  70 3.5.1 Adsorption performance of Fe/ZSM-5(Albemarle) Experiments on the adsorption performance of the Fe/ZSM-5(Albemarle) catalyst were conducted at temperatures from 275oC to 375oC and GHSV=5000h-1 using model flue gases with NO concentrations ranging from 200ppm to 1000ppm and 4% O2. The typical adsorption curves are shown in Figure 3.24, with the remaining curves at other temperatures shown in Figures G.6 to G.9 in Appendix G.   Figure 3.24 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=275oC)   Compared with previously used coarse Fe/ZSM-5(PUC) and fine Fe/ZSM-5(crushed PUC)) catalyst, the breakthrough time for the fine Fe/ZSM-5(Albemarle) catalyst was much longer for all tested conditions, indicating a much higher adsorption capacity of the Fe/ZSM- 5 (Albemarle) catalyst. Furthermore, the equilibrium NOx concentration was not reached at a Chapter 3: Adsorption and Reaction Kinetics  71 time of thousand seconds, which was tens of times of fine Fe/ZSM-5(crushed PUC)) although irregular shapes appeared on the adsorption curves. Based on the adsorption profiles, the equilibrium adsorption capacity of the catalyst was calculated and plotted as a function of the equilibrium NOx concentration in Figure 3.25. Again, the Freundlich equation was used to fit the relationship between qe and C0, and the parameters  and  were fitted with the 2nd order polynomial function with the results shown in Figure 3.26. Equations 3.9 and 3.10 are obtained for  and  as a function of the adsorption temperature (T, K). 251069966.101696.088632.2 TT     (3.9) 251043429.202732.011218.8 TT      (3.10)   Figure 3.25 Fitting of adsorption isotherms of NOx by Freundlich equation (Catalyst: Fe/ZSM-5(Albemarle)) Chapter 3: Adsorption and Reaction Kinetics  72  Figure 3.26 Relationship between α/ and adsorption temperature (Catalyst: Fe/ZSM- 5(Albemarle))  Generally, the coefficient  in the Freundlich equation reflects the intensity of adsorption. =1 indicates one active site adsorbs one molecule of NO. In Figure 3.26, the value of  is around 0.5, meaning that one molecule of NO occupies about 2 active sites indicating a weak adsorption of NO on the Fe active site. Fe/ZSM-5(Albemarle) exhibited excellent adsorption performance as shown in Figure 3.25. As discussed earlier, at T=350oC and C0=0.4 g/m3, the fine Fe/ZSM-5(crushed PUC) showed better adsorption capacity (qe=0.058 mg/g cat.) than the coarse Fe/ZSM-5(PUC) (qe=0.016 mg/g cat.). This difference demonstrates that a small particle size is beneficial to the adsorption of NOx. At the same operating condition, qe for the fine Fe/ZSM-5(Albemarle) catalyst is 0.66 mg/g cat., which is 11.4 and 41.3 times of the capacity of the fine Fe/ZSM- Chapter 3: Adsorption and Reaction Kinetics  73 5(crushed PUC) and the coarse Fe/ZSM-5(PUC), respectively. This result clearly shows that the structure or the manufacturing procedure of the catalyst support could also have a significant effect on the adsorption performance of the catalyst.  3.5.2 Effect of O2 concentration on NOx adsorption of Fe/ZSM-5(Albemarle) According to one of the mechanisms of HC-SCR suggested in the literature (Traa et al., 1999), NO is first oxidized to NO2 by O2 in the flue gas and then reduced by HC at the surface of the catalyst. This mechanism implies that NOx adsorption onto the catalyst surface may be affected by the oxygen concentration. To verify this mechanism, the effect of O2 concentration on NOx adsorption of Fe/ZSM-5(Albemarle) was investigated, with the result shown in Figure 3.27. Without the presence of O2 in the flue gas, the catalyst showed very low equilibrium NOx adsorption capacity (qe). With the increase of O2 concentration from 0 to 4%, qe increased significantly from lower than 0.1 to around 0.7 mg/g cat. Further increase of O2 concentration from 4 to 8% enhanced the adsorption performance, but not as significant as that at low O2 concentrations.  This result clearly demonstrates that the presence of O2 in the flue gas is essential for achieving high NO adsorption ratio using Fe/ZSM-5(Albemarle) catalyst. Based on the experimental result, for Fe/ZSM-5(Albemarle) catalyst, one can postulate that, in the adsorption process, NO is first oxidized to NO2 by the adsorbed O2 on the active site of the catalyst, and NO2 is then captured to the active site. According to the Langmuir-Hinshelwood theory, the following surface reactions can be assumed: * 2 2*2 OO        (3.11) * 2 * NOONO        (3.12) Chapter 3: Adsorption and Reaction Kinetics  74 The increase of O2 concentration in the flue gas increases the surface concentration of O* in reaction (3.11), and therefore more NO can be adsorbed to active sites of the catalyst when the adsorption equilibrium in reaction (3.12) is reached.  Figure 3.27 Effect of O2 concentration in the flue gas on the adsorption capacity of Fe/ZSM- 5(Albemarle)) catalyst  3.5.3 Effect of H2O and CO2 on NOx adsorption of Fe/ZSM-5(Albemarle) Since Fe/ZSM-5(Albemarle) appears to be the most promising potential candidate for the dual zone ICFB reactor, the effect of other major components of the flue gas, such as H2O and CO2, on the adsorption performance of the catalyst was further investigated. The model flue gas containing 4% O2 was supplemented with 10% H2O and/or 10% CO2 and then used in the experiment at T=350oC and GHSV=5000 h-1, with the adsorption curves shown in Figures G.10 to G.12 in Appendix G. Based on the adsorption curve, the equilibrium adsorption capacity qe for the model flue gas with the addition of 10% H2O, 10% CO2 and 10% CO2 + 10% H2O was calculated, Chapter 3: Adsorption and Reaction Kinetics  75 with the result shown in Figure 3.28. For comparison, data from the model flue gas without the addition of H2O and CO2 are also plotted in the same figure.  Figure 3.28 Effect of H2O and CO2 in the flue gas on adsorption isotherms (Catalyst: Fe/ZSM-5(Albemarle))  The addition of 10% H2O to the model flue gas showed no effect on the adsorption capacity of Fe/ZSM-5(Albemarle). However, when compared with the case without H2O and CO2, qe dropped by about 20% when 10% CO2 was added into the flue gas. When 10% H2O was introduced to the flue gas containing 10% CO2, qe increased by about 5% but was still lower than that without the presence of CO2. Clearly, CO2 exhibited a negative impact on the adsorption of NO, which might be caused by the competitive adsorption between CO2 and NO. Although H2O had no effect on the adsorption of NOx in the flue gas without CO2, the addition of H2O to the CO2-containing flue gas could slightly improve the adsorption of NO on Fe/ZSM-5(Albemarle) catalyst. Chapter 3: Adsorption and Reaction Kinetics  76  3.5.4 Adsorption performance of parent H/ZSM-5(Albemarle) The adsorption capacity of H/ZSM-5(Albemarle) was measured and is shown in Figure 3.29. For comparison, the adsorption capacity of Fe/ZSM-5(Albemarle) is also plotted in Figure 3.29.  Figure 3.29 Comparison of adsorption isotherm between Fe/ZSM-5(Albemarle) and parent H/ZSM-5(Albemarle) at T=350oC  The adsorption capacity of the parent H/ZSM-5(Albemarle) was very low compared to Fe/ZSM-5(Albemarle). This means that the high adsorption capacity of Fe/ZSM- 5(Albemarle) was contributed not by the catalyst support, but by the impregnated active Fe site or the combined effect of Fe and the ZSM-5 support. For catalyst prepared by the IMPO method, it is estimated that about 5% (wt.) of Fe was impregnated into Fe/ZSM- 5(Albemarle), which means that there was around 1.0 mmol of Fe per gram of catalyst. The Chapter 3: Adsorption and Reaction Kinetics  77 adsorption capacity of NO on the Fe/ZSM-5 (Albemarle) was measured to be around 1 mg/g cat. or 0.033 mmol/g cat., which is much lower than the loaded Fe content on the catalyst. This may suggest that, in the adsorption process, most of the active sites are occupied by the adsorbed N2 or O2 because of the high N2 and O2 concentrations in the flue gas, while only a small fraction of active sites are taken by the adsorbed NO2.  3.5.5 Catalytic activity of parent H/ZSM-5(Albemarle) in HC-SCR To investigate the HC-SCR capability of H/ZSM-5(Albemarle), the time-on-stream test was conducted with the feed gas consisting of 600ppm NO + 1200ppm HC + 1% O2 + N2 at T=350oC and GHSV=5000 h-1, with the result shown in Figure 3.30. It can be seen that NOx conversion quickly decreased to near zero in 200 minutes. The low HC conversion in Figure 3.30(b), on the other hand, indicates that H/ZSM-5 (Albemarle) does not catalyze the oxidation of hydrocarbons appreciably. Chapter 3: Adsorption and Reaction Kinetics  78  (a) NOx conversion  (b) HC conversion Figure 3.30 Catalytic activity of parent H/ZSM-5(Albemarle) Chapter 3: Adsorption and Reaction Kinetics  79  3.5.6 Reaction performance of Fe/ZSM-5(Albemarle) A series of time-on-stream tests were carried out to investigate the reduction performance of Fe/ZSM-5(Albemarle) catalyst at various temperatures and GHSV. Further study on the effect of O2 and HC concentrations in the model flue gas was also conducted. Moreover, the influence of H2O, CO2 and SO2 on the catalytic activity was investigated preliminarily.  3.5.6.1 Effect of reaction temperature Since the reaction temperature plays an important role in HC-SCR of NOx, it is necessary to seek for the optimum temperature window for the future study. As before, the feed gas mixture of 600 ppm NO + 1200 ppm HC + 1% O2 + N2 was used in the time-on-stream test at GHSV=5000 h-1 while the reaction temperature varied from 275 to 375oC with an increment of 25oC. Typical profiles of NOx and HC conversions and outlet CO concentration are shown in Figure 3.31 (T=275oC) and Figure 3.32 (T=350oC) (see Figures G.17 to G.19 in Appendix G for profiles at other temperatures). At T=275oC, the catalyst showed unstable activity which is similar to Fe/ZSM- 5(PUC). The peak NOx conversion was found to correspond to the time when HC conversion and outlet CO concentration reached their maximum values. The discontinuity in the HC conversion curve in Figure 3.32 is caused by the resolution of the gas analyzer at low CO2 concentrations (<100ppm or 0.01%) while the CO concentration is recorded at the 1ppm level. At all temperatures, large amount of CO was observed in the outlet gas flow, indicating that the production of CO is inevitable if Fe/ZSM-5 is used as the HC-SCR catalyst. Similar trends were observed for the NOx and HC conversions and the outlet CO concentration at all temperatures covered by this study. To avoid the potential release of Chapter 3: Adsorption and Reaction Kinetics  80 unconverted hydrocarbons and produced CO into the environment, a reactor filled with oxidation catalyst can be added to the downstream of the HC-SCR reactor to convert CO and unconverted hydrocarbons to CO2.  Figure 3.31 Profiles of NOx and HC conversions and outlet CO concentration (Fe/ZSM- 5(Albemarle), T=275oC, [O2]=1%) Chapter 3: Adsorption and Reaction Kinetics  81  Figure 3.32 Profiles of NOx and HC conversions and outlet CO concentration (Fe/ZSM- 5(Albemarle), T=350oC, [O2]=1%)    Chapter 3: Adsorption and Reaction Kinetics  82 The NOx and HC conversions at all investigated temperatures are shown in Figure 3.33. Peak NOx conversions were observed to be 65.2% (275oC), 66.3% (300oC), 68.3% (325oC), 71% (350oC) and 61.2% (375oC) with corresponding HC conversions of 24.8% (275oC), 45.3% (300oC), 72.3% (325oC), 100% (350oC) and 100% (375oC). Again, the discontinuous profiles of HC conversion in Figure 3.33(b) were caused by the low resolution of the gas analyzer at low CO2 concentrations. At T≤325oC, NOx and HC conversions increased with time. After the peak was reached, both NOx and HC conversions at T=275oC decreased quickly.  At T=300oC and 325oC, the NOx conversion decreased from the peak values to 38.8% and 54.2%, respectively, and then remained stable afterward. The maximum stable NOx conversion was observed at T=350oC (71%). The catalytic activity remained stable with further increase in temperature up to 375oC although the NOx conversion was lower (61.2%) at this temperature. Chapter 3: Adsorption and Reaction Kinetics  83  (a) NOx conversion  (b) HC conversion Figure 3.33 Effect of temperature on the catalytic activity of Fe/ZSM-5(Albemarle)  Chapter 3: Adsorption and Reaction Kinetics  84 3.5.6.2 Effect of GHSV Fe/ZSM-5(PUC) showed high sensitivity to the variation in GHSV where the catalyst maintained its stable activity with time only at GHSV≤5000 h-1 at 350oC (see Figure 3.16). In contrast, Fe/ZSM-5(Albemarle) exhibited good durability to the change of GHSV. As shown in Figure 3.34, the catalytic activity remained stable with the change in GHSV. However, with the increase of GHSV from 5000 h-1 to 10000 h-1, the NOx conversion decreased from 71% (GHSV=5000 h-1) to 55% (GHSV=7500 h-1) and 45% (GHSV=10000 h-1), respectively. Moreover, the corresponding HC conversion dropped from 100% (GHSV=5000 h-1) to 74% (GHSV=7500 h-1) and 57% (GHSV=10000 h-1). The drop in NOx conversion as well as HC conversion could be induced by the decrease of the resident time in the catalyst bed as the GHSV increased.  Chapter 3: Adsorption and Reaction Kinetics  85  (a) NOx conversion  (b) HC conversion Figure 3.34 Effect of GHSV on the catalytic activity of Fe/ZSM-5(Albemarle)  Chapter 3: Adsorption and Reaction Kinetics  86 3.5.6.3 Effect of O2 and HC concentrations The effect of O2 and HC concentrations in the model flue gas on the catalytic activity of Fe/ZSM-5(Albemarle) was investigated at reactor temperatures of 350oC and 375oC and GHSV=5000 h-1. The inlet NO concentration was kept at 600ppm throughout the experiment. O2 concentration in the feed gas was set as 0.5, 1, 2, 4 or 8% with a HC:NO(V/V) ratio of 0.5, 1, 2 or 4. As shown in Figure 3.35 for T=350oC, at a given O2 concentration, NOx conversion generally increased with increasing HC:NO. At HC:NO=0.5, NOx conversion was in a range of 35%~40% without distinct variation with the change in O2 concentration. HC conversion reached 100% at all O2 concentrations at such a low HC:NO ratio. When HC:NO ratio increased to 1, NOx conversion increased at a given O2 concentration, and HC conversion kept at around 100%, except for 0.5% O2 where HC conversion dropped to 89%. Further increase of HC:NO to 2 greatly enhanced the catalytic activity. However, HC conversion at 0.5% O2 steeply decreased to 58% because of the lack of O2 for the combustion of HC. Over 95% of HC conversion was obtained at HC:NO=2 for cases with 1% or more O2. Although the increase of HC:NO to 4 could improve the catalytic activity further, the HC conversion decreased quickly, especially for cases with O2 concentrations lower than 2%. In all cases, at a given ratio of HC:NO, the maximum NOx conversion was always observed when O2 concentration was around 1%. For 2% or more of O2 concentration, NOx conversion decreased with the increase of O2 concentration. For instance, at HC:NO=2, NOx conversion was in the order of 1% O2 (67.4%) > 0.5% O2 (60.1%) > 2% O2 (55.2%) > 4% O2 (47.2%) > 8% O2 (41.5%). Chapter 3: Adsorption and Reaction Kinetics  87  (a) NOx conversion  (b) HC conversion  Figure 3.35 Effect of O2 and HC concentrations (Fe/ZSM-5(Albemarle), T=350oC)  Chapter 3: Adsorption and Reaction Kinetics  88  (a) NOx conversion  (b) HC conversion Figure 3.36 Effect of O2 and HC concentrations (Fe/ZSM-5(Albemarle), T=375oC)  Chapter 3: Adsorption and Reaction Kinetics  89 Figure 3.36 showed similar trend as in Figure 3.35 on the influence of the HC:NO ratio for T=375oC. The exception is that, at a given HC:NO, NOx conversion always decreased with the increase of O2 concentration. In the range of studied O2 concentrations at HC:NO=2, NOx conversion was in the order of 0.5% O2 (62.2%) > 1% O2 (56.1%) > 2% O2 (49.3%) > 4% O2 (43.8%) > 8% O2 (35.3%). For a given HC:NO ratio and O2 concentration, NOx conversion at T=375oC was always lower while the HC conversion was equal to or higher than at T=350oC.  3.5.6.4 Effect of CO2 and H2O The real flue gas from the combustion process contains not only NOx and O2, but also large amount of CO2 and water vapour (see Table 1.1). To investigate the influence of CO2 on the catalytic activity of Fe/ZSM-5(Albemarle), 5, 10 and 15% CO2 was added into the feed gas of 600ppm NO + 1200ppm HC + O2 + N2 with various O2 concentration of 1, 4 and 8% at T=350oC and GHSV=5000 h-1 with the result shown in Figure 3.37. Compared with the case without the addition of CO2, for feed gas containing 1% O2, CO2 showed minor negative influence with NOx conversion slightly decreased from 69.7% (0% CO2) to 68.1% (5% CO2), 66.9% (10% CO2) and 66.8% (15% CO2). For the feed gas containing 4% and 8% O2, NOx conversion fluctuated at values between 52% (4% O2) and 44% (8% O2) with a difference of around 1% after the addition of 5%, 10% or even 15% CO2. In general, it could be concluded that the presence of CO2 in the feed gas had an insignificant influence on the catalytic activity of Fe/ZSM-5(Albemarle), and the effect could be well neglected.  Chapter 3: Adsorption and Reaction Kinetics  90  Figure 3.37 Effect of CO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (T=350oC)  When 5% H2O was added into the feed gas of 600ppm NO + 1200ppm HC + 1% O2 + N2, as shown in Figure 3.38, the NOx conversion increased from 68.5% (0% H2O) to 71.7% (5% H2O). Further increase in NOx conversion was not observed for the addition of 10% and 15% H2O. For the flue gas containing 4% O2, the NOx conversion without H2O was 47%, and the addition of 5%, 10% and 15% H2O led to the increase of NOx conversion to 51%, 53.3% and 54.2%, respectively. For the flue gas containing 8% O2, NOx conversion increased from 39.1% (0% H2O) to 45.4% (5% H2O), 47% (10% H2O) and 49.1% (15% H2O). It is clear that the presence of H2O in the model flue gas somewhat enhanced the SCR activity of Fe/ZSM-5(Albemarle) catalyst, especially at high O2 concentrations in the flue gas. Chapter 3: Adsorption and Reaction Kinetics  91  Figure 3.38 Effect of H2O on the catalytic activity of Fe/ZSM-5(Albemarle) (T=350oC)  The time-on-stream test was carried out to further investigate the long-term influence of H2O on the SCR activity of Fe/ZSM-5(Albemarle), with the result shown in Figure 3.39. Test A was carried out using the catalyst aged at T=350oC for several runs without regeneration. After test A, the catalyst was regenerated in 10% O2 + N2 at 500oC for 2 hours before test B was performed under the same operating condition. It can be seen that the aged Fe/ZSM-5(Albemarle) still kept a stable activity over a time period of 450 minutes for the feed gas of 600ppm NO + 1200ppm HC + 4% O2 + 10% H2O + N2 at T=350oC and GHSV=5000 h-1. After regeneration, the catalytic activity remained almost the same. Chapter 3: Adsorption and Reaction Kinetics  92  Figure 3.39 Time-on-stream test on the effect of H2O (Fe/ZSM-5(Albemarle), T=350oC)  The combined effect of CO2 and H2O was also investigated for the model flue gas with 10% CO2 and 10% H2O, with the results plotted in Figure 3.40. Clearly, the addition of 10% CO2 had almost no effect on the SCR activity of the catalyst. The presence of 10% H2O showed some improvement on NOx conversion for the flue gas containing various O2 concentrations. Compared with the case with only 10% H2O, further addition of 10% CO2 seemed to have no effect on the catalytic activity when O2 concentration was low (1%). However, a combined positive impact was obtained at 4% O2, with the impact being more significant at 8% O2. Chapter 3: Adsorption and Reaction Kinetics  93  Figure 3.40 Combined effect of H2O and CO2 on the catalytic activity of Fe/ZSM- 5(Albemarle) (T=350oC)  One of the possible mechanisms for the enhanced deNOx performance by the addition of H2O is the contribution of H2 and CO, produced from the water gas shift reaction (equation 3.13) or steam reforming of propylene (equation 3.14) which can take place at low temperatures. 222 HCOCOOH       (3.13) 22 HCOHCOH       (3.14) As shown in Figure 3.41, for the flue gas containing 4% O2, without the presence of H2O, CO and CO2 concentrations at the outlet of the reactor were around 1000ppm and 0.30%, respectively. The addition of 5% H2O led to the decrease of CO concentration to 500ppm, but the increase of CO2 concentration to 0.34%. Further increase of H2O had no significant influence on both CO and CO2 concentrations. The significant drop of the CO Chapter 3: Adsorption and Reaction Kinetics  94 concentration and the increase of the CO2 concentration caused by the addition of H2O reflect the fact that water gas shift reaction (equation 3.13) or the combination of equations 3.13 and 3.14 played a role in the SCR process. It is well known that both CO and H2 are strong reducing agents, which are commonly used for NOx reduction in the three-way catalytic converters. Further controlled experiments, in which both CO and H2 concentrations will be monitored, are needed in the future to elucidate the reaction mechanism on the enhancement of NOx reduction by water.  Figure 3.41 Effect of the addition of H2O on outlet CO and CO2 concentrations (Fe/ZSM- 5(Albemarle), T=350oC)  3.5.6.5 Deactivation of Fe/ZSM-5(Albemarle) catalyst by SO2 Sulphur oxides (SO2 and SO3) are considered as the most poisonous components in the combustion flue gas on SCR process.  Although the content of SO2 in the flue gas could be reduced to a relatively low level (10 to 30ppm) after the wet scrubbing process, this small Chapter 3: Adsorption and Reaction Kinetics  95 amount of SO2 can still seriously inhibit the catalytic activity of most of the catalysts used in the SCR process, damaging the application of HC-SCR. Since Fe/ZSM-5(Albemarle) showed an acceptable activity in the HC-SCR process, the effect of SOx on the catalyst performance is further investigated to examine the durability of the catalyst in the presence of SO2. To monitor the gradual decay of the catalyst activity with time, the time-on-stream test was performed with the result of total four runs shown in Figures 3.42 to 3.45. In each run, four time-on-stream (TOS) tests were conducted as described below: A. The fresh catalyst was first treated with the model flue gas of 600ppm NO +1200ppm HC + 1% O2 + N2 at 350oC and GHSV=5000 h-1 until the stable NOx conversion was reached to establish the baseline. B. Then, SO2 was introduced into the feed gas stream and the outlet NOx and SO2 concentrations were continuously monitored. C. After several hours, SO2 was switched off from the feed gas flow and the SO2 deactivated catalyst was used in another TOS test. When the outlet NOx concentration reached a relatively stable value, the test was terminated. D. The catalyst was then calcined in-situ with 20% O2 + N2 at 500oC for 2 hours in order to regenerate the catalyst. Thereafter, the regenerated catalyst was used in the final TOS test to finish the first cycle or Run #1. E. The TOS result in step D from the regenerated catalyst was then used as the new baseline for the next cycle, and steps B to D were repeated for the 2nd to 4th test cycles, or Runs #2-4.  It could be seen from those figures that, in Run #1 for the baseline without the presence of SO2, the NOx conversion was stable with time. When 200ppm SO2 was introduced into the feed gas flow, the measured SO2 concentration at the outlet of the reactor Chapter 3: Adsorption and Reaction Kinetics  96 showed that SO2 was adsorbed onto the catalyst. As time passed by, the outlet SO2 concentration increased gradually when more catalyst surface was occupied by SO2. Finally, the outlet SO2 reached the inlet concentration, or even higher because some of the adsorbed SO2 might also be washed out by the model flue gas flow. With the addition of 200ppm SO2, the NOx conversion quickly reached a maximum value and then gradually headed down due to the deactivation of the catalyst by SO2 adsorption. Because of the malfunction of the data logging system, the TOS test result of step C in Run #1 was not recorded and thus not plotted in Figure 3.42. The TOS result from the regenerated catalyst without the presence of 200ppm SO2 showed that, the catalytic activity was partially recovered, and the NOx conversion was stable with time but couldn’t reach the level with the original catalyst, which means that the catalyst was permanently poisoned by SO2.  Figure 3.42 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #1) Chapter 3: Adsorption and Reaction Kinetics  97  The result of Run #2 in Figure 3.43 showed the similar trend as in Run #1 for the presence of 200ppm SO2 in the flue gas. When SO2 feed was switched off, the catalytic activity was partially recovered and remained stable with time because the adsorbed SO2 on the catalyst was partially stripped by the SO2-free flue gas flow, which led to part of the active sites being recovered. Immediate calcination of the catalyst could further recover part of the catalytic activity, but the NOx conversion was still much lower than the baseline value. Run #3 (Figure 3.44) and Run #4 (Figure 3.45) revealed the same trend except for that the catalytic activity became lower and lower.   Figure 3.43 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #2) Chapter 3: Adsorption and Reaction Kinetics  98  Figure 3.44 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #3)  Figure 3.45 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (200ppm SO2, Run #4) Chapter 3: Adsorption and Reaction Kinetics  99  To further analyze the results from all four runs, TOS tests with 200ppm SO2 present in the flue gas are compared in Figure 3.46. The highest NOx conversion was obtained in Run #1. After 250 minutes, the NOx conversion dropped from the peak value of 68% to 47%. Over the same period of time, the NOx conversion decreased from 64% to 37%, 56% to 25%, and 36% to 19% for Run #2, #3 and #4, respectively, using regenerated catalyst. The difference of the peak NOx conversion between runs increased, indicating the permanent loss of reactivity after each deactivation and regeneration cycle because more and more active sites on the catalyst were blocked by SO2 or S-containing compounds.  Figure 3.46 Comparison of the four consecutive runs with 200ppm SO2 in the flue gas (Catalyst: Fe/ZSM-5(Albemarle))  Figure 3.47 compares the performance of the fresh and regenerated catalyst on NOx reduction over the four consecutive runs, after repeated exposure to the flue gas containing 200ppm SO2. It can be seen that the permanent damage of SO2 on the catalyst gradually built Chapter 3: Adsorption and Reaction Kinetics  100 up with time by contacting the SO2-containing flue gas with the catalyst. After Run #1, only 7% of permanent loss of the NOx conversion was observed (72% to 65%). Run #2 caused another 17% of loss (65% to 48%). The catalyst lost 11% (48% to 37%) and 12% (37% to 25%) of the reactivity in Run #3 and Run #4, respectively. In other words, the catalytic activity dropped from 72% to 25% in about 27 hours with regeneration between runs because of the addition of 200ppm SO2 into the flue gas.  Figure 3.47 Comparison of fresh and regenerated catalysts over four consecutive runs (Catalyst: Fe/ZSM-5(Albemarle))  Another set of tests was conducted at lower SO2 concentration (30ppm) to simulate flue gases after SO2 removal in the combustion process. Fresh catalyst was used in this test, with the results shown in Figure 3.48. The NOx conversion decreased with time when 30ppm SO2 was added into the flue gas, but at a much slower pace than the one with 200ppm SO2. Over about 11 hours, the NOx conversion dropped from 69% to 50%, in comparison to a drop from 68% to 47% in 4 hours at 200ppm SO2. However, if the catalyst was aged by the 30ppm Chapter 3: Adsorption and Reaction Kinetics  101 SO2 and placed overnight without immediately flushed by SO2-free gas flow or calcined at 500oC, the catalyst was almost totally deactivated, giving only 6.5% of NOx conversion when the reactor was restarted with SO2-free model flue gas. Further calcination of the catalyst could only bring the NOx conversion back to 32%. This permanent activity loss might be caused by the aging of adsorbed SO2 over a long time on the active sites. As time passed, the adsorbed SO2 also reacted with the active Fe, leading to the formation of Fe-S bond and/or FeSOx compounds, which caused the irreversible poisoning to the catalyst.  Figure 3.48 Effect of SO2 on the catalytic activity of Fe/ZSM-5(Albemarle) (30ppm SO2)  Measured BET surface area of the fresh Fe/ZSM-5(Albemarle) catalyst (Figure 3.49) showed that the impregnation of Fe to the parent ZSM-5(Albemarle) led to a slight loss (around 5%) of BET surface area (171 m2/g (parent) to162 m2/g (Fe/ZSM-5(Albemarle))). After the catalyst was aged in the flue gas containing 200ppm SO2 for 27 hours and then regenerated, the BET surface area dropped to 129 m2/g, a 20% loss compared to the fresh Chapter 3: Adsorption and Reaction Kinetics  102 Fe/ZSM-5(Albemarle). For the catalyst aged by 30ppm SO2 for 11 hours and calcined after being placed over night, the BET surface area decreased remarkably by 36% to 103 m2/g. These data proved that SO2 could have serious negative impact on the catalytic activity of Fe/ZSM-5(Albemarle) especially when the catalyst had a long-term exposure to SO2- containing flue gases. The poisoning is irreversible although part of the deactivated sites caused by the adsorbed SO2 could be recovered by immediate calcination. In the proposed dual-zone deNOx reactor, when the SO2-containing flue gas is injected into the adsorption zone and contacts with the catalyst, both SO2 and NOx will be adsorbed onto the catalyst. After the SO2/NOx-adsorbed catalyst moves into the reduction zone, due to the lack of SO2 in the reductant gas, the adsorbed SO2 will be stripped off from the catalyst by the reductant gas flow. Since the contact time between the catalyst and SO2 is very short in the adsorption zone, the negative effect of SO2 to the catalyst could be reduced in the dual-zone deNOx reactor, although such a speculation still needs to be confirmed experimentally in the future.  Figure 3.49 Comparison of BET surface areas for fresh and deactivated catalyst (Fe/ZSM- 5(Albemarle)) Chapter 3: Adsorption and Reaction Kinetics  103  3.6 Summary Fe/ZSM-5 prepared from Na/ZSM-5 (PUC) with an average particle size of 1042 m showed very low NOx adsorption capacity. The adsorption performance was greatly improved when the catalyst was prepared using crushed Na/ZSM-5(PUC) with smaller particle size (234 m). The most promising adsorption performance was offered by Fe/ZSM-5(Albemarle) with the finest particle size (155 m) among tested catalysts. This result proves that the adsorption performance of the catalyst is closely related to the particle size and the structure of the catalyst support. The addition of water vapour to the model flue gas had little effect on the adsorption capacity of Fe/ZSM-5(Albemarle). However, negative impact was observed when CO2 was added into the flue gas, and the further addition of H2O to the CO2-containing flue gas slightly improved the adsorption performance of Fe/ZSM-5(Albemarle). Although Fe/ZSM-5(PUC, WIE) showed the same or higher peak NOx conversion compared with Fe/ZSM-5(PUC, IMPO) at the same reaction condition, its reactivity was not stable with time, even at higher temperatures and excessive O2 concentration. However, Fe/ZSM-5(PUC, IMPO) showed stable HC-SCR performance at T350oC even with very low O2 concentration. From this point of view, IMPO is a preferred method for the preparation of Fe/ZSM-5. The catalytic activity of Fe/ZSM-5 catalyst was sensitive to the reaction temperature and space velocity. The catalyst deactivated very quickly at temperatures lower than 325oC because of the deposit of graphitic carbon on the catalyst surface. Fe/ZSM-5 catalysts using Na/ZSM-5(PUC), Na/ZSM-5(crushed PUC) and H/ZSM-5(Albemarle) as supports exhibited stable reactivity at T350oC and GHSV=5000 h-1. Among them, Fe/ZSM-5(PUC) showed the highest reactivity, but lowest adsorption capacity, at the same reaction condition, which Chapter 3: Adsorption and Reaction Kinetics  104 implies that the reaction activity of the catalyst is not directly proportional to the adsorption capacity in the conventional HC-SCR process. To reach a high NOx conversion, a high HC:NO ratio is needed. Considering both the economic aspect and the reduction efficiency of NOx, HC:NO=2:1 is a reasonable choice. O2 concentration played an important role in the SCR of NOx with propylene as the reducing agent. Fe/ZSM-5(PUC), Fe/ZSM-5(crushed PUC) and Fe/ZSM-5(Albemarle) catalysts exhibited acceptable activity when O2 concentration was controlled at relatively low levels (e.g. ≤1%). Water vapour could slightly enhance the reactivity of Fe/ZSM-5(Albemarle) in HC- SCR, while CO2 showed little effect. SO2 in the flue gas could seriously poison Fe/ZSM- 5(Albemarle) catalyst, especially for the long time contact between SO2-containing flue gas and the catalyst. The damage caused by SO2 was permanent, and only partial catalytic activity could be recovered via calcination at high temperatures.   105  Chapter 4 Hydrodynamic Study of the ICFB Reactor  Based on the literature review, the catalytic activity of most HC-SCR catalysts is severely inhibited by the presence of excess O2 in the flue gas. On the other hand, one of the main NO reduction mechanisms states that NO is first oxidized to NO2, followed by the reduction of NO2 by the reducing agent. That means the presence of O2 is essential for the HC-SCR process. Therefore, in the proposed dual-zone reactor, the presence of oxygen in the reduction zone is essential but must be controlled within an appropriate range in order to keep the O2 concentration at a relatively low level (lower than 2%, according to the results from the fixed bed experiment). On the other hand, gas bypassing from the reduction zone to the adsorption zone will cause an increase in the consumption of the reducing agent in the dual-zone reactor, which should be minimized. In addition, the solids circulation rate will be directly linked to the transfer of the adsorbed NOx from the adsorption zone to the reduction zone and the NOx reduction in the reaction zone of the reactor. Based on the above requirements, the fluidized bed reactor with a draft tube, or so- called internal circulating fluidized bed (ICFB) reactor, could meet the requirement of the dual-zone reactor. The ICFB with different configurations has been studied extensively on its hydrodynamics (Fusey, et al., 1986; Milne et al., 1992; Kim et al., 1997; Ishikura et al., 2003), heat and mass transfer (Freitas and Freire, 2001; Stocker et al., 1990; Kim, et al., 2000), and its applications for particle drying (Ando et al., 2002), coating of tablets (Shelukar et al., 2000), coal combustion and gasification (Yang and Keairns, 1978a, b), and other industrial processes (Yang, 1998). Chapter 4: Hydrodynamic Study of the ICFB Reactor   106 In order to provide design and operating criteria for the hot model ICFB unit, it is essential to investigate the hydrodynamic performance of the ICFB reactor.  4.1 Experimental setup A cold model ICFB reactor was constructed with the configuration shown in Figure 4.1. Gas bypass both from the annulus to the draft tube and the draft tube to the annulus and the solids circulation rate were investigated in this cold model unit, which was made of Plexiglass with dimensions listed in Table 4.1. Chapter 4: Hydrodynamic Study of the ICFB Reactor  107  1 - Draft tube Gas Nozzle  2 - Annulus Gas Inlet  3 - Solids Inlet  4 - Annulus Gas Distribution Chamber  5 - Gas Distributor  6 - Draft Tube  7 - Downcomer (Annulus)  8 - Freeboard  9 - Exhaust Gas Outlet 10 - Solids Drain 11 - Pressure Measurement Solids Flow Draft tube Gas Flow Annulus Gas Flow 11 Draft Tube Gas Feed Annulus Gas Feed 6 2 10 4 P 5 1 4 11 P P 11 7 11 8 Solids Feed 3 P P 11 9  Figure 4.1 Schematic of the cold model ICFB reactor  Chapter 4: Hydrodynamic Study of the ICFB Reactor  108   Table 4.1 Dimensions of the cold model ICFB unit Item Dimension Draft tube diameter, mm 50.8 (I.D.), 63.5 (O.D.) Draft tube diameter, mm (for cylindrical gas distributor only) 38.1 (I.D.), 44.5 (O.D.) Draft tube length, mm 1016.0 Column diameter, mm 101.6 (I.D.), 114.3 (O.D.) Column height, mm 1092.2 Freeboard diameter, mm 254.0 (I.D.), 266.7 (O.D.) Freeboard height, mm 508.0 Annulus gas distributor opening ratio 2.1% Holes on the distributor 52 holes of 1.6mm diameter Gas nozzle diameter, mm 34.9 (I.D.), 38.1 (O.D.)   The geometry of the cold model ICFB unit was developed based on the design criteria described by Yang (1998) in order to achieve high solids circulation rate and retain certain degree of gas bypassing between the core and the annulus regions. The desired solids circulation rate was calculated based on the adsorption performance of the catalyst from the fixed bed experiments, and the reactor diameter was selected with the consideration of the available flow rate of the building air. The bed material used in the experiment was millet with a particle density (ρp=837 kg/m3) and average particle size (dp=1.1 mm) similar to the Fe/ZSM-5(PUC) catalyst used in the fixed bed experiment. The building air was used as the fluidizing gas. Pure CO2 from Praxair was used as the tracer gas to study the gas bypass. The cold model ICFB system is shown schematically in Figure 4.2. In the experiment, the building air was injected into the draft tube via the gas nozzle and the gas flow rate was Chapter 4: Hydrodynamic Study of the ICFB Reactor  109 adjusted to have the draft tube region operated at a pneumatic transport condition to carry particles upward to create a continuous circulation of the particles between the annulus and the draft tube. At the same time, the mixture of the building air and the tracer gas was introduced into the annulus or downcomer region through the gas chamber and the annulus gas distributor. The flow rate was controlled to keep the annulus region being operated at the moving bed or minimum fluidization condition. Flow rates in both the draft tube and the annulus were adjusted to investigate the gas bypass between the two regions and the solids circulation from the annulus to the draft tube. Gas compositions at the inlet of the annulus gas chamber, the inlet of the gas nozzle, the outlet of the draft tube, the outlet of the annulus, and the gas mixture in the freeboard region were measured by a Horiba PG-250 flue gas analyzer. Chapter 4: Hydrodynamic Study of the ICFB Reactor  110 P P C O 2 B ui ld in g ai r IC FB  R ea ct or Bag house Cyclone Exhaust gas to vent Sampling Sampling Vent Gas Analyzer Computer Sampling (Inlet of annulus) (Outlet of draft tube) (Gas mixture) Sampling S am pling (O utlet of annulus) Sampling (Inlet of draft tube) P P B ui ld in g ai r   Figure 4.2 Schematics of the cold model ICFB system   As reported by Yang and Keairns (1983), for a given ICFB reactor, the gas bypass and solids circulation rates are significantly affected by the gap opening between the inlet of the draft tube and the distributor plate. Three types of annulus gas distributors, a flat plate, a cylindrical plate and a conical plate, were tested, together with the effective gap opening between the draft tube and the distributor plate being adjusted to investigate the hydrodynamic performance of the ICFB reactor. All three perforated distributors had the same opening ratio of 2.1% for easy comparison of the experimental results. Chapter 4: Hydrodynamic Study of the ICFB Reactor  111  4.1.1 Estimation of gas bypass ratio The gas bypass between the draft tube and the annulus was evaluated by CO2 tracer method. At given operating conditions, continuous CO2 tracer was injected into the inlet gas flow of the annulus, with CO2 concentrations at the outlet of the draft tube and the annulus, and the inlet of the draft tube and the annulus being analyzed and recorded. Gas bypass ratios are calculated by the mass balance method as illustrated in Figure 4.3. D raft Tube A nnulus FD, xD FA, xA FD,0, xD,0 FA,0, xA,0 FDA, xDA FAD, xAD  Figure 4.3 Schematic of the mass transfer between draft tube and annulus  The total mass balance over the draft tube and the annulus in Figure 4.3 is given by,        AACOAairDDCODair AACOAairDDCODair FxxFxx FxxFxx 22 22 )1()1( )1()1( 0,0,0,0,0,0,          (4.1) where air, air density, kg/m3 CO2, CO2 density, kg/m3 Chapter 4: Hydrodynamic Study of the ICFB Reactor  112 FD,0, volumetric gas flow rate at the inlet of the draft tube, m3/s xD,0, volume fraction of CO2 in the gas flow at the inlet of the draft tube FA,0, volumetric gas flow rate at the inlet of the annulus, m3/s xA,0, volume fraction of CO2 in the gas flow at the inlet of the annulus FD, volumetric gas flow rate at the outlet of the draft tube, m3/s xD, volume fraction of CO2 in the gas flow at the outlet of the draft tube FA, volumetric gas flow rate at the outlet of the annulus, m3/s xA, volume fraction of CO2 in the gas flow at the outlet of the annulus  Since the bed material used in the cold model ICFB unit was millet, which had a very low CO2 adsorpiton capacity, the change of CO2 concentration contributed from particle adsorption was thus neglected in the mass balance equation. The total mass balance on CO2 over the draft tube and the annulus is given by, AACODDCOAACODDCO FxFxFxFx 2222 0,0,0,0,       (4.2) With xA,0, xD,0, xD, and xA being measured, flow rates at the outlet of the annulus (FA) and the outlet of the draft tube (FD) can be calculated by combining equations 4.1 and 4.2: DA ADADDD A xx FxxFxx F   0,0,0,0, )()(     (4.3) AD DADAAA D xx FxxFxx F   0,0,0,0, )()(     (4.4) The total mass balance and the CO2 mass balance over the draft tube are given by equations 4.5 and 4.6 below,        DADACODAairDDCODair ADADCOADairDDCODair FxxFxx FxxFxx 22 22 )1()1( )1()1( 0,0,0,       (4.5) DADACODDCOADADCODDCO FxFxFxFx 2222 0,0,      (4.6) Chapter 4: Hydrodynamic Study of the ICFB Reactor  113 where FDA, volumetric gas bypass rate from draft tube to annulus, m3/s xDA, volume fraction of CO2 in the gas bypass flow from draft tube to annulus FAD, volumetric gas bypass rate from annulus to draft tube, m3/s xAD, volume fraction of CO2 in the gas bypass flow from annulus to draft tube  The volumetric gas bypass rates from the annulus to the draft tube (FAD) and from the draft tube to the annulus (FDA) are derived from equations 4.3 to 4.6 with xDA=xD,0 and xAD=xA,0, D DA DA DDA Fxx xx FF 0,0, 0, 0,        (4.7) D DA DD AD Fxx xx F 0,0, 0,         (4.8)  The percentage gas bypass ratios from the draft tube to the annulus (RDA) and the annulus to the draft tube (RAD) are defined as:    %100 0,  D DA DA F FR       (4.9)    %100 0,  A AD AD F FR       (4.10)  4.1.2 Estimation of solids circulation rate The solids circulation rate was measured by the visual observation method. Coloured millet particles were added into the bed material as tracer particles. For given gas flow rates in the draft tube and the annulus, the motion of a specific tracer particle on the column wall was tracked by marking its moving distance L in a given time interval tm. The tracer particle downward velocity in the annulus was then calculated by Chapter 4: Hydrodynamic Study of the ICFB Reactor  114 m p t LU        (4.11) The solids circulation rate Ws (kg/m2.s) through the unit cross-sectional area of the draft tube was calculated by )1(   pp D A s US SW      (4.12) where SA and SD are cross-sectional areas of the annulus and the draft tube respectively, ρp is the particle density, and  is the bed voidage in the annulus which is assumed to be equal to the packed voidage.  4.2 Performance with the flat distributor plate The flat gas distributor with two configurations used in the experiment is shown schematically in Figure 4.4. For the first configuration (Figure 4.4(a)), the top of the gas nozzle was flush with the surface of the flat distributor, and the gap, called the effective gas opening, between the bottom of the draft tube and the surface of the distributor was fixed at HG1=10 mm. The top of the gas nozzle was flush with the bottom of the draft tube for the second configuration (Figure 4.4(b)) with HG1 remaining at 10 mm, but the effective gap opening for the solids circulation in the second configuration was reduced to HG2=6 mm, which corresponds to the distance between the inner surface of the draft tube and the outer surface of the gas nozzle. Chapter 4: Hydrodynamic Study of the ICFB Reactor  115   D raft Tube A nnulus A nnulus Draft Tube Gas Annulus Gas H G 1 D raft Tube A nnulus A nnulus Draft Tube Gas Annulus Gas H G 1HG2 ( a ) ( b )  Figure 4.4 Configurations of the flat distributor for the annulus gas flow  Chapter 4: Hydrodynamic Study of the ICFB Reactor  116 4.2.1 Gas bypass For the configuration one with an effective gap opening of HG=10 mm, the effect of gas velocities in the annulus (UA) and the draft tube (UD) on the gas bypass from both the annulus to the draft tube and the draft tube to the annulus is shown in Figure 4.5. Within the range of tested gas velocities, the ICFB with a flat distributor plate showed more than 50% of gas bypass from the annulus to the draft tube (RAD) (Figure 4.5(a)), but less than 1.5% from the draft tube to the annulus (RDA) (Figure 4.5(b)). RAD increased with the increase of both UA and UD. Although RDA decreased with the increase of UA at a given UD, there was no clear trend with respect to the effect of UD on the gas bypass at a given UA (see Figure 4.5(b)). One should note that RDA values were very small and could be associated with significant measurement errors. For the second distributor configuration with the top of the gas nozzle flushed with the bottom of the draft tube while keeping HG1=10 mm (see Figure 4.4(b)), it is seen from Figure 4.6(a) that the gas bypass ratio from the annulus to the draft tube was 10% more than that in Figure 4.5(a) at given UA and UD. On the other hand, RDA was further reduced to lower than 0.3%. This can be explained by the fact that the configuration of the distributor used in Figure 4.6 allows gas from the nozzle to directly flow into the draft tube with less chance to escape into the annulus region. Furthermore, this configuration increased the pressure drop around the inlet of the draft tube, and thus more annulus gas was sucked into the draft tube from the annulus region. It was observed that, because of the high gas bypass from the annulus to the draft tube in both cases, insufficient gas was supplied to the annulus. As a result, particles could not be stably circulated if the annulus gas velocity was lower than 0.45 m/s. Chapter 4: Hydrodynamic Study of the ICFB Reactor  117 Error bars plotted in Figure 4.5 were obtained from 6 repeated runs based on a 95% confidence interval. The possible source of errors came from the readings of the rotameters and the measured CO2 concentrations, as well as the reproducibility of the operating conditions. Errors ranged from 1.61% to 4.52%, with an average around 3.35% for calculated gas bypass ratios from the annulus to the draft tube, and from 0.08% to 0.69%, with an average around 0.49% for calculated gas bypass ratios from the draft tube to the annulus. Chapter 4: Hydrodynamic Study of the ICFB Reactor  118  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.5 Effect of gas velocities on gas bypass (Flat distributor, HG=10 mm) Chapter 4: Hydrodynamic Study of the ICFB Reactor  119  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.6 Effect of gas velocities on gas bypass (Flat distributor, HG=6 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  120 4.2.2 Solids circulation rate The influence of gas velocities on the solids circulation rate (Ws) is shown in Figures 4.7 and 4.8 for distributor configurations with HG=10 mm and 6 mm, respectively. In both cases, Ws increased with increasing annulus gas velocity (UA) at a given UD. Meanwhile, the same tendency was observed for the influence of UD at a given UA. However, UA clearly showed more influence on Ws than UD.  At given UA and UD, although the gas bypass from the annulus to the draft tube (RAD) was higher at HG=6 mm than HG=10 mm, Ws at HG=10 mm (Figure 4.7) was always higher than at HG=6 mm in Figure 4.8. In addition, the influence of UA and UD on Ws in Figure 4.7 where HG=10 mm was more significant than HG=6 mm shown in Figure 4.8. This is because the small effective gap opening (HG) obstructed the solids circulation, while the increase in gas bypassing was counteracted by the smaller HG in the second configuration. Error bars plotted in Figure 4.7 were based on 4 repeated runs for each point with 90% confidence interval. The possible source of errors came from the readings and the timing of the traveling distance of the coloured particles, the change of bed voidage and the reproducibility of the experimental conditions. Errors for calculated solids circulation rates ranged from 0.79 to 8.05 kg/m2.s, with an average around 4.55 kg/m2.s. Chapter 4: Hydrodynamic Study of the ICFB Reactor  121  Figure 4.7 Effect of gas velocities on solids circulation rate (Flat distributor, HG=10 mm)  Figure 4.8 Effect of gas velocities on solids circulation rate (Flat distributor, HG=6 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  122 4.3 Performance with the cylindrical distributor plate The flat distributor plate showed very high gas bypassing from the annulus to the draft tube which is undesirable for the dual zone reactor if the annulus is operated as the adsorption zone for the HC-SCR of NOx. To decrease gas bypassing from the annulus to the draft tube, another type of gas distributor for the annulus, namely cylindrical distributor plate, was designed and built, as shown in Figure 4.9. The draft tube was also changed to a size of   38.1 mm I.D. with the wall thickness of 3.2 mm in order to keep the same opening ratio of the distributor (2.1%). In the experiment, the top of the gas nozzle was always flush with the bottom of the cylindrical distributor, and three configurations were tested in the experiment, with the distance (HG1) between the bottom of the draft tube and the top of the gas nozzle set at 10, 20 and 60 mm, respectively. Note that, for HG1=10 mm and 20 mm, as shown in Figure 4.9(a), the effective gap openings were the same, i.e., the gap between the outer surface of the draft tube and the inner surface of the cylinder (HG2=9.5 mm). For HG1=60 mm, since the bottom of the draft tube was 9 mm higher than the surface of the annulus gas distributor, the effective gap opening was thus changed to HG2=13 mm (see Figure 4.9(b)). Chapter 4: Hydrodynamic Study of the ICFB Reactor  123 A nnulus A nnulus D raft Tube Draft Tube Gas Annulus Gas H G 1 A nnulus A nnulus D raft Tube Draft Tube Gas Annulus Gas H G 1H G2 ( a ) ( b ) HG2  Figure 4.9 Configurations of the cylindrical distributor for the annulus gas flow  4.3.1 Gas bypass Gas bypass ratios at HG1=10 mm are shown in Figure 4.10. When the annulus gas velocity was low (UA=0.37 m/s), the draft tube gas velocity (UD) showed less influence on the gas bypass ratio from the annulus to the draft tube (RAD) (Figure 4.10(a)). However, the increase Chapter 4: Hydrodynamic Study of the ICFB Reactor  124 of UD decreased the gas bypass from the draft tube to the annulus (RDA) effectively at UA=0.37 m/s (Figure 4.10(b)). Further increase of UA to 0.45 m/s resulted in an increase of RAD by 5%~15%, while RDA dropped to lower than 1.5% in most cases for the UD range of 3.31 to 4.55 m/s. RAD increased by another 15% with RDA remaining at around 1% when UA increased from 0.45 to 0.52 m/s. Although RAD increased with the increase of both UA (>0.37 m/s) and UD, when compared with the flat distributor (Figure 4.5(a)), RAD for the flat distributor was always 20~30% higher than for the cylindrical distributor at given UA and UD. The change of the annulus distributor showed no significant influence on RDA, which always remained at a very low level in both cases. It is not surprising to see the significant decrease of RAD in Figure 4.10(a) in comparison to Figure 4.5(a) in view of the fact that the use of a smaller size draft tube for the cylindrical distributor led to an increase of the annulus cross-sectional area from 4940 to 6556 mm2, and a corresponding decrease of the draft tube cross-sectional area from 2027 to 1140 mm2. When HG1 was set to 20 mm, the influence of gas velocities on gas bypass is shown in Figure 4.11. RAD was also observed to increase with the increase in both UA and UD, although the influence of UD was not as significant as UA. It appears that the increase of HG1 from 10 to 20 mm had no effect on the gas bypass from both directions, which could be attributed to the fact that the effective gap opening remained the same for both cases. RDA remained lower than 2.5% within the range of gas velocities tested. Chapter 4: Hydrodynamic Study of the ICFB Reactor  125  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.10 Effect of gas velocities on gas bypass (Cylindrical distributor, HG1=10 mm) Chapter 4: Hydrodynamic Study of the ICFB Reactor  126  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.11 Effect of gas velocities on gas bypass (Cylindrical distributor, HG1=20 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  127 The increase in both UD and UA increased RAD when the draft tube was lifted further to HG1= 60 mm, as shown in Figure 4.12. At UA=0.45 m/s, RAD was even higher than that at UA=0.52 m/s in Figure 4.11, but still much lower than that in Figure 4.5 where the flat distributor was used. This indicates that the increase of the effective gap opening enhanced gas bypassing from the annulus to the draft tube, but had no significant effect on gas bypassing from the draft tube to the annulus for this type of distributor. Chapter 4: Hydrodynamic Study of the ICFB Reactor  128  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.12 Effect of gas velocities on gas bypass (Cylindrical distributor, HG1=60 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  129 4.3.2 Solids circulation rate Solids circulation rates (Ws) for the cylindrical distributor with HG1=10 mm are shown in Figure 4.13. For a given UA, Ws increased with the increase of UD. For a given UD, lower Ws was observed at UA=0.37 m/s. Further increase of UA beyond 0.45 m/s showed less influence on Ws indicating that Ws was mainly affected by UD. This trend is different from the flat distributor with HG=10 mm where Ws was more sensitive to UA than to UD (see Figure 4.7). Moreover, the cylindrical distributor exhibited much higher Ws than the flat one. For example, at UA=0.52 m/s and UD=3.26 m/s, Ws was 105 kg/m2.s and 49 kg/m2.s for the cylindrical and the flat distributor, respectively.  Figure 4.13 Effect of gas velocities on solids circulation rate (Cylindrical distributor, HG1=10mm)  With the gap between the bottom of the draft tube and the top of the gas nozzle increased from 10 to 20 mm, both UA and UD showed minor effect on Ws at UA0.45 m/s and Chapter 4: Hydrodynamic Study of the ICFB Reactor  130 UD3.72 m/s, although the trend that Ws increased with the increase of both UA and UD was still observed. The change of HG1 from 10 to 20 mm showed no influence on Ws, which could be explained by the fact that the effective gap opening remained the same for both cases.  Figure 4.14 Effect of gas velocities on solids circulation rate (Cylindrical distributor, HG1=20mm)  As shown in Figure 4.15, UA dominated the change of Ws while UD showed less effect when the draft tube was lifted to HG1=60 mm, which is different from HG1=10 and 20 mm but consistent with the trend from the flat distributor experiment. Considering that the bottom of the draft tube was 9 mm higher than the surface of the annulus gas distributor, which was almost identical to that for the flat distributor (Figure 4.5), it is not surprising that the same trend was observed. The higher Ws for the cylindrical distributor is associated with its bigger gap opening (HG2=13 mm). Chapter 4: Hydrodynamic Study of the ICFB Reactor  131  Figure 4.15 Effect of gas velocities on solids circulation rate (Cylindrical distributor, HG1=60mm)  The cylindrical distributor exhibited lower gas bypassing from the annulus to the draft tube and relatively higher solids circulation rate than the flat distributor, at given UA and UD. However, as observed in the experiment, constrained by the shape of the distributor, the particles could not be circulated at low draft tube gas velocities, which limited the possible operation window of the gas flow rate.  4.4 Performance with the conical distributor plate As one of the most commonly used gas distributors for ICFB reactors, the conical distributor was also tested in this study. The perforated holes on the distributor were aligned in perpendicular to the surface of the distributor and the effective gap opening for the solids Chapter 4: Hydrodynamic Study of the ICFB Reactor  132 circulation was defined as HG, as shown in Figure 4.16. In the experiment, two configurations with HG=10 and 15 mm were investigated, separately, with the top of the gas nozzle always flush with the bottom of the conical plate. A nnulus A nnulus D raft Tube Draft Tube Gas Annulus Gas H G  Figure 4.16 Configuration of the conical distributor for the annulus gas flow  4.4.1 Gas bypass As seen in Figure 4.17(a), the increase of UD increased the gas bypass ratio from the annulus to the draft tube, but not as significantly as the flat and cylindrical distributors. For a given UD, UA showed less influence on RAD, with RAD ranging from 20 to 50% over the tested range of gas velocities. Relatively high values of RDA were observed using the conical distributor, ranging from 5~10%, as shown in Figure 4.17(b). RDA remained almost constant with the variation of UA at a given UD, but decreased with increasing UD at a given UA. Chapter 4: Hydrodynamic Study of the ICFB Reactor  133  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.17 Effect of gas velocities on gas bypass (Conical distributor, HG=10 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  134 When the effective gap opening for the solids circulation was set to 15 mm, rather than 10 mm, the increase in UA enhanced gas bypass from the annulus to the draft tube, as shown in Figure 4.18(a). The influence of UD on the gas bypass showed a trend opposite to HG=1.0 cm. At UD=1.63 m/s, RAD was very high at 70%~75%. As UD increased to 2.79 m/s, RAD dropped to around 40%. This indicates that the gas bypass from the annulus to the draft tube was reduced with the increase in UD. According to the configuration of the conical plate, at HG=10 mm, all perforated holes were blocked by the outer surface of the draft tube and the annulus gas could not flow directly into the draft tube but was forced into the annulus region, and the gas could only be dragged from the annulus to the draft tube by the circulating particles. As a result, UD dominated RAD. At HG=15 mm, over 50% of the perforated holes on the conical plate were directly exposed to the inlet of the draft tube. Therefore, the gas flow through these annulus distributor holes could partially flow upward into the draft tube, and thus enhanced gas bypassing from the annulus to the draft tube. With the increase of UD, the increased gas flow to the draft tube pushed part of the annulus gas back to the annulus area and thus decreased RAD.  As a result of the coupling impact from both UA and UD, RDA increased by around 5% compared to that at HG=10 mm, as shown in Figure 4.18(b). Chapter 4: Hydrodynamic Study of the ICFB Reactor  135  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.18 Effect of gas velocities on gas bypass (Conical distributor, HG=15 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  136 4.4.2 Solids circulation rate For HG=10 mm, Ws increased with increasing the annulus gas velocity (UA) at a given draft tube gas velocity (UD), as shown in Figure 4.19. When UA reached a certain level, i.e. >0.90 m/s, further increase in UA had only little influence on Ws. Meanwhile, the same tendency was observed for the influence of UD at a given UA. However, the increase of UA clearly showed more influence on Ws than that of UD. For HG=15 mm, as shown in Figure 4.20, a similar trend was observed except with a relatively higher Ws than that in Figure 4.19 at given UA and UD. Moreover, both UA and UD exhibited significant influence on Ws.  Figure 4.19 Effect of gas velocities on solids circulation rate (Conical distributor, HG=10 mm)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  137  Figure 4.20 Effect of gas velocities on solids circulation rate (Conical distributor, HG=15 mm)  Since the solids circulation rate was mainly influenced by both UA and UD for a given configuration of the conical plate, and there existed uncertainties on the actual gas velocities inside the draft tube and the annulus because of the gas bypass between the annulus and the draft tube, an overall gas velocity is introduced here to evaluate its impact on Ws. The overall gas velocity U is defined as the total gas flow rate divided by the total cross-sectional area of the draft tube and the annulus DA DA SS FF U   0,0,      (4.13) where FA,0 and FD,0 are gas flow rates at the inlet of the annulus and the draft tube, and SA and SD are the cross-sectional areas of the annulus and the draft tube, respectively. Chapter 4: Hydrodynamic Study of the ICFB Reactor  138 As shown in Figure 4.21, for HG=10 mm, Ws increased with the increase of U, and could be well fitted by a 2nd order polynomial equation with most of the data points falling into an error range of ±20%. According to the fixed bed experimental data for Fe/ZSM-5(PUC) catalyst, to keep a stable NOx conversion for a long term operation in the hot model ICFB reactor, the reaction temperature must be remained at 350oC or higher and the gas space velocity in the reduction zone should be lower than 5000h-1 (Std. T, i.e., 25oC). Based on the results from the cold model unit which has the same configuration as the hot model ICFB reactor, the gas velocity in the draft tube (UD) must be lower than 1.41 m/s (Std. T) if the draft tube is to be used as the reduction zone. If UD is set to 1.0 m/s and UA=0.4 m/s (Std. T) which are equivalent to UD=2.09 m/s and UA=0.84 m/s at T=350oC in the hot model ICFB reactor, the particle circulation rate will be around 50 kg/m2.s based on the annulus cross-sectional area, according to the cold model experimental result shown in Figure 4.21. For a model flue gas containing 600 ppm (0.75 g/m3) NO at a gas velocity of UA=0.4 m/s (Std. T), the amount of NO input to the annulus will be 0.3 g/m2.s, while the adsorption capacity (i.e., the amount of NO at the equilibrium adsorption) of the catalyst at 350oC will be 1.34 g/m2.s according to the adsorption performance data for Fe/ZSM-5(PUC) catalyst in the fixed bed reactor, which is higher than the amount of input NO. This means that the input NO in the annulus region of the ICFB reactor could be completely adsorbed by the catalyst. For HG=15 mm, in Figure 4.22, although the shape of the fitted curve switched from minor concave to convex style in comparison with Figure 4.21, the data can also be well correlated by the overall gas velocity, with an error range of ±10%. Chapter 4: Hydrodynamic Study of the ICFB Reactor  139  Figure 4.21 Effect of overall gas velocity on solids circulation rate using conical distributor (HG=10 mm)  Figure 4.22 Effect of overall gas velocity on solids circulation rate using conical distributor (HG=15 mm) Chapter 4: Hydrodynamic Study of the ICFB Reactor  140  The same effect of the overall gas velocity on the solids circulation rate was also observed for the ICFB with a cylindrical distributor plate which is not shown here. However, for the flat distributor, the data showed a great scattering with poor correlation between U and Ws.  4.5 Gas bypassing in the hot model ICFB reactor In the cold model experiment, the conical distributor plate exhibited high solids circulation rate, high gas bypass from the draft tube to the annulus and relatively low gas bypassing from the annulus to the draft tube compared to the flat and cylindrical distributor plates. Moreover, the conical distributor plate showed flexible and stable operation within wide ranges of gas velocities for both UA and UD. As a result, the conical distributor plate was selected as the annulus gas distributor for the hot model ICFB reactor. To further reduce gas bypassing from the annulus to the draft tube, the perforated holes were aligned in parallel to the draft tube or the column wall in the hot model unit, as shown in Figure 4.23. A nnulus A nnulus D raft Tube Draft Tube Gas Annulus Gas H G  Figure 4.23 Configuration of the conical distributor in the hot model ICFB reactor Chapter 4: Hydrodynamic Study of the ICFB Reactor  141  In the hot model experiments, gas bypassing was studied using O2 as the tracer gas. Using the same method shown in Figure 4.3 and equations 4.1 to 4.10, RAD and RDA could be evaluated based on O2 balance instead of CO2 balance for the hot model test. Two kinds of catalysts were investigated in the hot model experiment, Fe/ZSM- 5(PUC) with an average particle diameter of 1042 µm and Fe/ZSM-5(Albemarle) with an average particle diameter of 155 µm. The effective gap opening (HG) for the solids circulation was set to be 10 mm for all experiments.  4.5.1 Fe/ZSM-5(PUC) catalyst Effects of gas velocities on gas bypass for Fe/ZSM-5(PUC) in the hot model ICFB reactor are shown in Figure 4.24. In the range of velocities investigated, the gas bypass from the draft tube to the annulus (RDA) was always higher than that from the annulus to the draft tube (RAD) for Fe/ZSM-5(PUC) catalyst at given UA and UD (see Figures H.1 to H.7 in Appendix H). For cases with UA0.3 m/s, with the increase of UD from 0.45 to 1.30 m/s, RAD increased from 3~7% to a peak value of 17~18% at UD=0.9~1.1 m/s, then decreased to 10~13% (Figure 4.24(a)) with further increase in UD, while RDA increased monotonically from 18% to 40% (Figure 4.24(b)) with the increase in UD. For UA=0.20 m/s, as UD increased from 0.45 to 0.59 m/s, RDA increased from 11% to 14%, while RAD increased from 4% to 6%. Compared to the effect of UD, it seemed that UA had less influence on both RAD and RDA. Furthermore, for a given UD, there was no clear relationship between UA and RAD because of the scattering of data. Based on those results, if the annulus is selected as the adsorption zone and the draft tube as the reduction zone, the gas bypass from the annulus to the draft tube could supply Chapter 4: Hydrodynamic Study of the ICFB Reactor  142 sufficient O2 to keep the O2 concentration in the draft tube at a desired level. To minimize RDA but also achieve a reasonably high solids circulation rate, UD should be controlled to be in the range of 0.45~0.6m/s. In conclusion, the change of the opening direction of perforated holes on the annulus gas distributor has been demonstrated to be successful in reducing the gas bypass from the annulus to the draft tube. At the same time, RDA was enhanced, which is beneficial for the SCR of NOx in the adsorption zone. Chapter 4: Hydrodynamic Study of the ICFB Reactor  143  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.24 Effect of gas velocities on gas bypass (Fe/ZSM-5(PUC), T=350±10oC)  Chapter 4: Hydrodynamic Study of the ICFB Reactor  144  4.5.2 Fe/ZSM-5(Albemarle) catalyst Effects of gas velocities on gas bypass using Fe/ZSM-5(Albemarle) catalyst are shown in Figure 4.25.  Unlike Fe/ZSM-5(PUC), the change of both UA and UD had almost no effect on the gas bypass from the annulus to the draft tube (RAD), which fluctuated in the range of 5~9% as shown in Figure 4.25(a) (see also Figures H.9 to H.19 in Appendix H). This feature could be useful for controlling the O2 concentration to a desired constant level when the draft tube region is used as the reduction zone. For all cases in Figure 4.25(b), with the increase of UD at a given UA, the gas bypass ratio from the draft tube to the annulus (RDA) was higher than using Fe/ZSM-5(PUC), but following a similar trend. As UA increased, RDA decreased (Figure H.8 in Appendix H). At UD=0.45 m/s, the decrease of RDA was not significant, remaining at a level around 14% (see Figure H.16 in Appendix H). With UD increased from 0.60 to 0.90 m/s, RDA decreased quickly with the increase in UA (see Figures H.17 to H.19 in Appendix H). In the gas velocity range of the current experiment, a higher UA and a lower UD are preferred in order to control RDA to a relatively low level. The reactor would not work properly if UD<0.40 m/s and UA>0.45 m/s.  Chapter 4: Hydrodynamic Study of the ICFB Reactor  145  (a) Annulus to draft tube  (b) Draft tube to annulus Figure 4.25 Effect of gas velocities on gas bypass (Fe/ZSM-5(Albemarle), T=355±15oC)   Chapter 4: Hydrodynamic Study of the ICFB Reactor  146 4.5.3 Prediction of O2 concentration in the draft tube and annulus For a given UA and UD, gas bypass ratios from the annulus to the draft tube (RAD) and from the draft tube to the annulus (RDA) can be obtained from Figures 4.24 and 4.25, and the volumetric gas bypass rates can then be calculated by ADDDAD RSUF       (4.13) DAAADA RSUF       (4.14) where FAD, volumetric gas bypass rate from annulus to draft tube, m3/s FDA, volumetric gas bypass rate from draft tube to annulus, m3/s SA, cross-sectional area of the annulus, m2 SD, cross-sectional area of the draft tube, m2  If O2 concentrations at inlets of both the annulus (CA0, O2) and the draft tube (CD0, O2) are measured, O2 concentrations in the annulus and the draft tube can be predicted by: DAADAA ODDAOAADAA OA FFSU CFCFSU C   2,02,02, )(     (4.15) ADDADD OAADODDADD OD FFSU CFCFSU C   2,02,02, )(     (4.16) where  CA,O2, O2 concentration in the annulus, % CD,O2, O2 concentration in the draft tube, %   It should be noted that certain amount of O2 in the flue gas could be adsorbed onto the catalyst surface and carried into the draft tube region. However, this part of O2 was negligibly Chapter 4: Hydrodynamic Study of the ICFB Reactor  147 small compared to the amount of O2 bypassing in the bulk gas phase from the annulus to the draft tube, and thus was not considered in deriving eqs. (4.15) and (4.16).  For Fe/ZSM-5(PUC) catalyst, if the draft tube is used as the adsorption zone, i.e., CA0,O2=0, at UA=0.4 m/s and UD=0.6 m/s, with CD0, O2 increasing from 4 to 8 and 12%, O2 concentration in the annulus or reduction zone  increases from 0.8 to 1.5 and 2.3%, respectively. However, if UD increases to 1.2 m/s while UA remaining at 0.4 m/s, the increase of CD0, O2 from 4 to 8 and 12% leads to an increase of O2 concentration in the annulus from 1.2 to 2.5 and 3.7%. On the other hand, if the annulus is selected as the adsorption zone, at UA=0.4 m/s, with UD in the range of 0.6 to 1.2 m/s and the flue gas O2 concentration from 4 to 12%, O2 concentration in the draft tube will vary from 0.7 to 2.2%, which is much lower than that with the draft tube serving as the adsorption zone. This result shows that, to treat flue gases with high O2 cocentrations in the ICFB with Fe/ZSM-5 (PUC) catalyst, the annulus should be selected as the adsorption zone in order to control O2 to a relatively low level in the reduction zone.  For Fe/ZSM-5(Albemarle) catalyst, at a low UD, both the annulus and the draft tube can be used as the adsorption zone. For example, at UA=0.3 m/s and UD =0.5 m/s, the increase of O2 concentration from 4 to 12% in the flue gas in the adsorption zone increases O2 concentrations in the reduction zone from 0.4 to 1.2% when the daft tube is used as the reduction zone, and from 0.6 to 1.7% when the annulus is used as the reduction zone. However, at UA=0.3 m/s and UD=0.9 m/s, O2 concentration increases from 0.5 to 1.4% in the reduction zone if the daft tube is used as the reduction zone due to the relatively constant RAD, and from 1.3 to 4% if the annulus serves as the reduction zone because of the increased gas byass from the draft tube to the annulus at high draft tube gas velocities. This implies that Chapter 4: Hydrodynamic Study of the ICFB Reactor  148 only the annulus can be used as the adsorption zone if the ICFB reactor is to be operated at a high UD for the treatment of flue gases with high O2 concentrations.  Considering the operation feature of the ICFB reactor where the annulus is operated at a moving bed or at the minimum fluidization while the draft tube is at pneumatic transport, the annulus should be used as the adsorption zone to provide a relatively long contact or adsorption time between the flue gas and the catalyst to achieve high adsorption ratio.  4.6 Summary The hydrodynamic performance of an ICFB cold model unit was investigated using three types of annulus gas distributors, i.e., flat, cylindrical and conical distributor plates. For each distributor, the influence of the effective gap opening on the gas bypass and the solids circulation was studied at various gas velocities in the annulus and the draft tube. For the flat distributor, the gas bypass ratio from the annulus to the draft tube (RAD) was over 50% and the gas bypass ratio from the draft tube to the annulus (RDA) was very low (0~1.5%) in the tested range of gas velocities. The increase of both UA and UD could enhance the gas bypass from the annulus to the draft tube. The solids circulation rate (Ws) increased with the increase of both UA and UD, although UA showed more influence on Ws than UD. Because of the high gas bypass from the annulus to the draft tube, particles could not be stably circulated if the annulus gas velocity was lower than 0.45 m/s. For the cylindrical distributor, the same trend as the flat distributor was observed with respect to the influence of UA and UD on RAD, RDA and Ws, except that the use of cylindrical distributor could greatly decrease RAD and enhance the solids circulation when compared with the flat distributor. Chapter 4: Hydrodynamic Study of the ICFB Reactor  149 As observed in the experiment, both distributors showed a relatively narrow operating window of gas velocities. The conical distributor showed a flexible and stable operation in a wide range of velocities for both UA and UD. The increase of UD could enhance RAD while UA showed less effect on RAD. RDA ranged from 5% to 15% in the test, which is much higher than that using the other two distributors. RAD and RDA were sensitive to the change of the effective gap opening. The increase of both UA and UD could increase Ws, but UA showed more influence on Ws than UD. The solids circulation rate was found to be well correlated with the overall gas velocity for the ICFB with conical and cylindrical gas distributors. In the hot model ICFB reactor, the gas bypass ratio of Fe/ZSM-5(PUC) from the annulus to the draft tube (RAD) reached a maximum value when the draft tube gas velocity (UD) was equal to 0.9~1.0 m/s in this study. Meanwhile, RDA continuously increased with increasing UD. In the range of velocities investigated, the gas bypass ratio from the draft tube to the annulus (RDA) was always higher than that from the annulus to the draft tube (RAD) for Fe/ZSM-5(PUC) catalyst. UA had less influence on both RAD and RDA than UD. For Fe/ZSM-5(Albemarle) in the hot model ICFB reactor, at a given UA, the gas bypass ratio from the draft tube to the annulus (RDA) increased with increasing UD. At given UA and UD, the ICFB using Fe/ZSM-5(Albemarle) possessed higher RDA but lower RAD than that using Fe/ZSM-5(PUC). The increase of UA could decrease RDA, and the increase of UD or UA had almost no effect on the gas bypass ratio from the annulus to the draft tube (RAD). It was found in the cold model experiment that all three types of distributors showed very low gas bypassing from the draft tube to the annulus, but high gas bypassing in the reverse direction. The data from the hot model experiment showed that the change in the direction of perforated holes on the conical distributor reduced the gas bypass from the Chapter 4: Hydrodynamic Study of the ICFB Reactor  150 annulus to the draft tube, but enhanced the gas bypass from the draft tube to the annulus. Thus, the change of the reactor configurations could have a significant impact on the operating characteristics of the ICFB reactor. Further detailed investigation is needed in order to fully understand the hydrodynamics of the ICFB reactor.    151  Chapter 5 Adsorption and Reduction Performance of the ICFB Reactor  According to the reaction kinetic test results, Fe/ZSM-5 catalyst exhibited promising performance on the selective catalytic reduction of NOx with propylene as the reductant. O2 exhibited a significant negative impact on NOx reduction when O2-rich flue gas was used in the experiment. To achieve a high NOx conversion, O2 concentration in the flue gas must be controlled to be lower than 2%, which is not commonly encountered for most combustion flue gases. To overcome the negative impact of excessive O2 in the flue gas, the integrated adsorption-reduction process has been proposed in this study. Based on the result obtained from the hydrodynamic experiment, a hot model internal circulating fluidized bed (ICFB) reactor with a similar configuration as the cold model unit presented in Chapter 4 (Figure 4.1) was built and tested.  5.1 Experimental setup In the proposed configuration, the flue gas is passed into the adsorption zone (annulus) where NOx is adsorbed by the catalyst particles. The NOx-rich catalyst particles then move downward and into the reduction zone (draft tube) where NOx is reduced by injected hydrocarbons. Meanwhile, the catalyst particles are depleted of the adsorbed NOx in the draft tube. The NOx-depleted catalyst particles are then recirculated back to the adsorption zone to keep a continuous operation. By adjusting the gas flow rates in the annulus and the draft tube, the bypassing of the flue gas from the adsorption zone to the reduction zone can be controlled Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  152 to have the O2 concentration in the reduction zone maintained at desired levels. The hot model ICFB reaction system is shown schematically in Figure 5.1. T T T T 40%HC+N2 N2 N2 20%NO+N2 Building air Mass Flowmeter Heater Mass Flowmeter Heater IC FB  R ea ct or Bag house Cyclone Exhaust gas to vent Sampling Sampling Vent Gas Analyzer Computer Sampling (Flue gas inlet) (Outlet of the drafttube) (Gas mixture) Sampling S am pling (O utlet of annulus)   Figure 5.1 Schematic of hot model ICFB reaction system  In the experiment, the ICFB reactor was first preheated by passing the preheated building air through both the annulus and the draft tube. After the desired reaction temperature was reached in the reactor, the gas bypass between the annulus and the draft tube was first measured based on the O2 mass balance method over the annulus and the draft tube by adjusting the flow rate of either the flue gas or the reductant gas, as illustrated in Chapter 4. After the gas bypass measurement, NO from the gas cylinder was blended with preheated building air and pure N2 to prepare the simulated model flue gas at desired NO and O2 concentrations. At the same time, propylene was injected into the preheated N2 to prepare the reductant gas mixture. The model flue gas was injected into the adsorption zone (annulus) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  153 through the conical distributor plate and NOx was adsorbed by the catalyst in the annulus. The reductant gas mixture was injected into the reduction zone (draft tube) via a gas nozzle and NOx adsorbed on the catalyst particles flowing from the annulus into the draft tube was reduced by the hydrocarbon in the reductant gas. The adsorption and reduction performance was evaluated by measuring the gas composition at the inlet and outlet of the annulus, the outlet of the draft tube and the gas mixture at the exit of the reactor. The geometric dimensions of the ICFB reactor are given in Table 5.1. Table 5.1 Geometric dimensions of the ICFB reactor Item Dimension Draft tube diameter, mm 54.8 (I.D.), 60.3 (O.D.) Draft tube length, mm 1016.0 Column diameter, mm 108.2 (I.D.), 114.3 (O.D.) Column height, mm 1092.2 Freeboard height, mm 1016.0 Freeboard diameter, mm 260.4 (I.D.), 273.1 (O.D.) Annulus distributor opening ratio 1.62%, 52 holes of 1.6mm diameter Gas nozzle diameter, mm 34.9 (I.D.), 38.1 (O.D.) Gas nozzle distributor opening ratio 9.19%, 61 holes of 1.6mm diameter Fluidized bed distributor opening ratio 3.25%, 151 holes of 1.6mm diameter Effective gap opening for the solids circulation, mm 10.0   Two types of Fe/ZSM-5 catalysts were selected and tested in this experiment, i.e., Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle). Both catalysts were prepared following the same procedure of the impregnation in organic solution (IMPO) method, as described in Chapter 3. The properties of the catalysts were given in Table 5.2.  Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  154 Table 5.2 Properties of catalysts used in the hot model experiments Catalyst Average particle size Bulk density SBET Umf  (T=350oC) Fe/ZSM-5 (PUC) 1042 µm 926 kg/m3 110 m2/g 0.39 m/s Fe/ZSM-5 (Albemarle) 155 µm 979 kg/m3 162 m2/g 0.01 m/s  The model flue gas used in the experiment was a mixture prepared from a gas cylinder containing 20% NO balanced with N2 and a liquid N2 Dewar, with both supplied from Praxair Products Inc. Building air was used as the source of O2. NOx concentration in the model flue gas was controlled to be 300~900 ppm with O2 concentration ranging from 4 to 12%. The reducing agent used in the experiment was propylene. The gas cylinder containing 40% propylene balanced with N2 was supplied by Praxair Products Inc. The reducing agent stream consisted of propylene + N2, with propylene-to-NOx molar flow ratio varied from 1 to 4. It should be noted that NO and HC were in different gas streams in the ICFB reactor, i.e., the flue gas stream in the adsorption zone and the reductant gas stream in the reduction zone. To investigate the effect of gas velocities and other factors on the performance of the catalyst without changing HC:NO ratio, the ratio of HC:NO was defined as the molar flow rate of propylene to the reduction zone divided by the molar flow rate of NO to the adsorption zone.  5.2 Estimation of NOx and HC conversions and NOx adsorption ratio The overall NOx conversion (XNOx) was calculated by equations 5.1 and 5.2 based on the initial concentration of NOx in the total gas flow of the annulus and the draft tube, and the concentration of NOx in the gas mixture at the exit of the reactor. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  155 )( 0,0, 0,0, , RF FNOx inNOx FF FC C        (5.1) %100 , ,,  inNOx outNOxinNOx NOx C CC X     (5.2) where CNOx,0, initial NOx concentration in the flue gas feed, ppm CNOx,in, initial NOx concentration in the total gas feed into the annulus and the draft tube, ppm        CNOx,out, NOx concentration in the gas mixture at the exit of the reactor, ppm        FF,0, flue gas flow rate, m3/s        FR,0, reductant gas flow rate, m3/s        XNOx, overall NOx conversion, %  The conversion of propylene was calculated based on the measured concentrations of CO and CO2 in the gas mixture at the exit of the reactor by assuming that all converted propylene goes to CO and CO2, as shown in equation 5.3: %100)( %)40(3 ) (%) (%)(10)( 0,0, 0, 0,0, 0,0, , 4 , 2 2      RF HC RF FCO outCOoutCO HC FFF FF FC CppmC X       (5.3) where CCO2,0, initial CO2 concentration in the flue gas feed, % CCO,out, CO concentration in the gas mixture at the outlet of the reactor, ppm CCO2,out, CO2 concentration in the gas mixture at the outlet of the reactor, % FF,0, flue gas flow rate, m3/s FR,0, reductant gas flow rate, m3/s Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  156 FHC,0, flow rate of 40% Propylene + N2 gas mixture from the cylinder, m3/s XHC, overall HC conversion, %  Since the building air contained a certain amount of CO2, the measured initial concentration of CO2 in the flue gas feed must be converted to the concentration based on the total flow rate of the flue gas and the reductant gas (FF, 0 + FR, 0) and then subtracted from the measured CO2 concentration at the exit of the reactor.  The initial concentration of propylene was estimated based on the flow rate of certified 40% propylene + N2 added to the reductant gas feed. Again, a coefficient of 3 was introduced in equation 5.3 to reflect the fact that one mole of propylene contains 3 moles of carbon. Because the overall mass balance on carbon was used to calculate HC conversion in eq. (5.3), the carbon balance could not be carried out to check the experiment reliability. Because pure N2 was used as the carrier gas for the reducing agent while NO concentration was much lower than N2 in the total gas flow, it is inaccurate to use total nitrogen balance to check the experiment reliabiltiy. To evaluate the adsorption performance of the catalyst in the ICFB reactor, NOx adsorption ratio in the annulus was calculated and compared with the overall NOx conversion. Because the gas bypass from both the annulus to the draft tube and the draft tube to the annulus existed, the real NOx concentration and the gas flow rate inside the adsorption zone must be corrected by the gas bypass data, and the NOx reduction caused by the hydrocarbon in the bypassing gas from the reduction zone to the adsorption zone should also be deducted from the initial NOx concentration. The adsorption ratio of NOx (XNOx, ads) in the adsorption zone was calculated by %100 ,, ,,,, ,   realadsNOx outadsNOxrealadsNOx adsNOx C CC X    (5.4) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  157 where CNOx,ads, out was the NOx concentration at the outlet of the adsorption zone, and the real initial NOx concentration, CNOx, ads,in, in the adsorption zone was calculated by )1( )( , 0, 0,0, ,, BMNOx RFFRF FRFNOx inadsNOx XFFF FFC C     (5.5)   where CNOx,0, initial NOx concentration in the flue gas feed, ppm FF,0, flue gas flow rate, m3/s FFR, calculated gas bypass from the adsorption zone to the reduction zone, m3/s FRF, calculated gas bypass from the reduction zone to the adsorption zone, m3/s XNOx, BM, benchmark NOx conversion caused by the bypassing HC from the reduction zone, % CNOx, ads,,in, real initial NOx concentration used for the adsorption in the adsorption zone, ppm  The benchmark NOx conversion used in equation 5.5 was determined individually for each case based on the fixed bed experimental results and the real concentrations of O2 and HC at the inlet of the adsorption zone with the effect of the gas bypass being considered. The fluidized bed experiment was also conducted and compared to the results from the ICFB reactor. The draft tube in the ICFB reactor was removed and a new flat gas distributor plate was installed to construct the fluidized bed reactor system. The model flue gas and the reductant gas were mixed in the gas chamber below the distributor before entered the catalyst bed for tests conducted in the fluidized bed. Effects of gas velocities, NO, O2 and HC concentrations on the NOx conversion were investigated.  Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  158 5.3 Performance with Fe/ZSM-5(PUC) catalyst 5.3.1 Selection of adsorption zone in the ICFB reactor Either the annulus or the draft tube could be used as the adsorption zone, while the other one as the reduction zone. The adsorption and reduction zones were switched between the annulus and the draft tube using Fe/ZSM-5(PUC) catalyst to investigate the effect of the zone switch on NOx conversion in the ICFB reactor.  The influence of gas velocities and the HC:NO ratio on NOx conversion using the draft tube as the adsorption zone is shown in Figure 5.2. For a given HC:NO molar ratio of 1 or 2 and a reductant gas velocity (UA), the increase in the flue gas velocity (UD) led to the decrease in NOx conversion. This could be attributed to the fact that the increase of UD decreased the contact time between the flue gas and catalyst particles in the adsorption zone, leading to a decrease in the amount of NOx adsorbed. Although NOx conversions in all runs were generally lower than 15%, the data clearly showed that increasing the HC:NO ratio and the reductant gas velocity (UA) improved the reactor performance. Such a low NOx conversion for this kind of configuration confirms that the annulus should be used as the adsorption zone and the draft tube as the reduction zone in order to achieve a high NOx conversion in the current ICFB reactor design. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  159  Figure 5.2 Effect of gas velocities and HC:NO ratio on NOx conversion using the draft tube as the adsorption zone (ICFB, Fe/ZSM-5(PUC))  The experimental results presented below were thus all obtained with the annulus as the adsorption zone.  5.3.2 Effect of catalyst loading in the ICFB reactor The overall NOx conversion and adsorption ratio in the annulus are given in Figure 5.3 for the ICFB reactor with Fe/ZSM-5(PUC) with a catalyst loading of 1.1, 2.2 and 3.3 kg, respectively. The model flue gas was 600 ppm NO + 4% O2 + N2, with the reductant gas containing propylene + N2 at HC:NO=2. The experiment was carried out with a fixed reductant gas velocity of UD=0.6 m/s, while the flue gas velocity in the annulus changed from 0.27 to 0.55 m/s. It clearly demonstrated that the increase in the catalyst loading increased both NOx conversion and NOx adsorption ratio. Both NOx conversion and adsorption ratio Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  160 decreased slightly with the increase of UA in all cases, and the NOx conversion and adsorption ratio appear to be matched each other very closely, indicating that all NOx adsorbed in the annulus zone was completely converted to N2 in the draft tube zone no matter how much the catalyst loading was. Although increasing the catalyst loading will be beneficial to NOx adsorption as well as NOx conversion, the maximum catalyst loading for both Fe/ZSM-5(PUC) and Fe/ZSM- 5(Albemarle) was set to 3.3 kg in all experiments carried out in this study.  Figure 5.3 Effect of catalyst loading on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(PUC))  Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  161 5.3.3 Effect of HC:NO molar ratio on NOx conversion The influence of HC:NO molar ratio on NOx conversion is shown in Figure 5.4. For a given HC:NO ratio, NOx conversion decreased with increasing UA because the decrease in the contact time between the flue gas and the catalyst in the annulus reduced the adsorption rate. On the contrary, NOx conversion increased with increasing UD, likely due to the increased reaction rate in the draft tube and catalyst circulation rate. The NOx conversion increased when HC:NO ratio increased from 1 to 2, which agrees with the results from the fixed bed experiment. The difference in NOx conversion between HC:NO=1 and 2 at a given UA decreased with increasing UD. For example, at UA=0.4m/s, the difference in NOx conversion was 6%, 5% and 2% for UD=0.6, 0.9 and 1.2 m/s, respectively, which indicates that the benefit of a larger HC:NO ratio for a higher NOx conversion diminished at high UD. At low UD, the reactor performance was limited by the reaction in the draft tube region. Therefore, increasing the HC:NO ratio improved the reactor performance significantly. At a high UD, the reactor performance was controlled by the adsorption in the annulus zone, insensitive to the change of HC:NO ratio, due to the enhanced mass transfer of hydrocarbons from the reductant gas flow to the catalyst surface. UA showed less influence on NOx conversion than UD because NOx conversion was controlled mainly by the draft tube gas velocity. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  162  Figure 5.4 Effect of HC:NO ratio on NOx conversion (ICFB, Fe/ZSM-5(PUC))   5.3.4 Effect of flue gas O2 content on NOx conversion As shown in Figures 5.5, for a given UD, increasing UA caused a steady decrease in NOx conversion for tests with 4% and 8% O2, and only a slight decrease at 12% O2. A highest NOx conversion of 31% was observed at low UA at 4% O2. With the increase of O2 concentration from 4% to 12% in the model flue gas, the maximum NOx conversion decreased significantly from 31% to ~20%. For a given O2 concentration and UA, the increase of UD from 0.6 to 0.75 m/s showed little influence on NOx conversion. This result shows that the flue gas O2 concentration still imposed a significant impact on NOx conversion in the ICFB reactor when coarse PUC catalyst was used. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  163  Figure 5.5 Effect of flue gas O2 concentration on NOx conversion (ICFB, Fe/ZSM-5(PUC))  5.3.5 Effect of gas velocities on NOx conversion and adsorption The influence of gas velocities on the adsorption and conversion of NOx for the model flue gas with 600 ppm NO + 4% O2 + N2 and HC:NO=2 is shown in Figure 5.6. When the draft tube gas velocity was low (UD=0.6 m/s) which created low solids circulation rate, the increase of UA had little influence on the NOx adsorption ratio. This is likely because that the shortened gas residence time in the annulus was compensated by the increased solids circulation rate, as UA increased. Taking the experimental error into account, one may say that all adsorbed NOx was reduced when UA<0.4 m/s for UD=0.6 m/s and UA<0.3 m/s for UD=0.9 m/s. For UD≥0.9 m/s with a high solids circulation rate, the adsorption ratio of NOx decreased with the increase of UA because the solids circulation rate was less sensitive to UA at a high UD, and the reduction rate was always lower than the adsorption ratio, which suggests that part of the adsorbed NOx was not reduced in the reduction zone, probably due Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  164 to the shortened gas and catalyst residence time in the draft tube zone at a high UD. In general, if the reactor was operated at steady state, NOx adsorbed in the annulus should always be equal to NOx converted in the draft tube. If NOx conversion is lower than the adsorption ratio, less NOx should be adsorbed when the catalyst returns to the adsorption zone. Hence, another possible reason for NOx conversion lower than the adsorption ratio is that NOx is weakly adsorbed by Fe/ZSM-5(PUC) catalyst, and the un-reacted NOx is desorbed before the catalyst particles leave the reduction zone. As a result, NOx conversion in the draft tube is lower than the NOx adsorption ratio in the annulus. For each given UA, with an increase in UD, both the conversion and the adsorption ratio of NOx increased, although the influence of UD on NOx adsorption ratio was more pronounced at low UA. To achieve a high NOx conversion in the ICFB reactor with Fe/ZSM-5(PUC) catalyst, the annulus gas velocity should be kept below 0.3 m/s.  Figure 5.6 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(PUC)) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  165   5.3.6 Performance of fluidized bed reactor with Fe/ZSM-5(PUC) 5.3.6.1 Effect of inlet NO concentration Figures 5.7 to 5.9 show that the inlet NO concentration in the model flue gas had some influence on NOx conversion. At O2 concentrations of 4% (Figure 5.7) and 2.5% (Figure 5.8), NOx conversion increased monotonically with increasing NO concentration from 300 to 900 ppm at a given fluidizing gas velocity. This may be explained by the fact that the increase in NOx concentration enhanced the NOx adsorption capacity of the catalyst and also the reaction rate. For the flue gas with 1% O2 (Figure 5.9), a 10~15% increase in NOx conversion was observed when the inlet NO concentration increased from 300 to 600 ppm. The NOx conversion slightly decreased as the flue gas NO concentration increased from 600 to 900 ppm. As discussed earlier, NO must be oxidized to NO2 before it is reduced by the HC. At an O2 concentration of 1% and a NO concentration of 900 ppm, the HC concentration in the gas flow was 1800 ppm at a HC:NO ratio of 2. If all HC was completely converted to CO2, it could have reduced the flue gas O2 level from 1% to 0.44%. Therefore, there may not be enough O2 for the oxidization of NO to NO2, resulting in a NOx conversion slightly lower than the case with 600 ppm of NO in the flue gas. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  166  Figure 5.7 Effect of inlet NO concentration on NOx conversion (Fluidized bed, Fe/ZSM- 5(PUC), [O2] =4%)   Figure 5.8 Effect of inlet NO concentration on NOx conversion (Fluidized bed, Fe/ZSM- 5(PUC), [O2] =2.5%) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  167   Figure 5.9 Effect of inlet NO concentration on NOx conversion (Fluidized bed, Fe/ZSM- 5(PUC), [O2] =1%)  5.3.6.2 Effect of HC:NO ratio on NOx and HC conversions It is anticipated that an increase in HC:NO will enhance the NOx conversion at given O2 and NO concentrations in the fluidized bed. For the flue gas with 300 and 600 ppm NO, as shown in Figure 5.10 and 5.11, at 1% flue gas O2, the increase of HC:NO ratio from 1 to 2 increased NOx conversion significantly. At O2 concentrations of 2.5% and 4%, NOx conversion also increased by increasing HC:NO ratio, although the effect was not as significant as at 1% O2. At low gas velocities, HC conversion decreased notably with the increase in HC:NO ratio at all tested O2 concentrations. When a high gas velocity was used, the effect of HC:NO ratio on HC conversion was insignificant. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  168 For an inlet NO concentration of 900 ppm, as shown in Figure 5.12, the increase of HC:NO led to significant increase in NOx conversion and a slight decrease in HC conversion at both 2.5% and 4% O2. For the flue gas with 1% O2, the effect of HC:NO on NOx conversion was insignificant at 300 and 600 ppm NO, and the HC conversion dropped to a relatively low level at HC:NO=2. Again, this is likely due to the insufficient supply of O2. In fact, as observed in the experiment, the outlet O2 concentration approached zero when the flue gas of 900 ppm NO + 1% O2 was used with a HC:NO ratio of 2. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  169  (a) NOx conversion  (b) HC conversion Figure 5.10 Effect of HC:NO on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC), [NO]=300ppm) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  170  (a) NOx conversion  (b) HC conversion Figure 5.11 Effect of HC:NO on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC), [NO]=600ppm) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  171  (a) NOx conversion  (b) HC conversion Figure 5.12 Effect of HC:NO on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC), [NO]=900ppm)  Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  172 5.3.6.3 Effect of inlet O2 concentration on NOx and HC conversions Fe/ZSM-5(PUC) showed high reactivity on NOx reduction in the fluidized bed for flue gas with low O2 concentration, as shown in Figure 5.13. NOx conversion reached 60% at O2=1% and U=0.27 m/s. Increase of O2 concentration led to the decrease of NOx conversion at a given gas velocity. It is worthwhile to point out that the performance of Fe/ZSM-5(PUC) was negatively influenced by the gas velocity due to the shortened residence time of gas reactants. The catalytic activity was more sensitive to the gas velocity at lower O2 concentrations than at higher O2 concentrations. The change in O2 concentration showed less influence on HC conversion than on NOx conversion, with the highest HC conversion consistently lower than 50%. However, the increase of the gas velocity decreased HC conversion significantly under current experimental conditions, likely due to the reduced gas residence time, which appears to be consistent with the result from the fixed bed reactor. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  173  (a) NOx conversion  (b) HC conversion Figure 5.13 Effect of inlet O2 concentration on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(PUC))  Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  174 5.3.7 Comparison of ICFB reactor and fluidized bed reactor Figure 5.14 compares the NOx conversion in the fluidized bed reactor and the ICFB reactor. The results with UD=0.60 and 1.20 m/s for the ICFB reactor were selected as the lower and upper bounds, because the NOx conversion reached the lowest value at UD=0.6 m/s and the highest at UD=1.20 m/s. Although the ICFB reactor showed higher NOx conversion (20~25% higher) than the fluidized bed at the same O2 concentration (4%), the NOx conversion in the ICFB reactor at high O2 concentrations was still much lower than the fluidized bed with 1% O2 only. This suggests that the ICFB reactor with Fe/ZSM-5(PUC) catalyst could improve the HC-SCR performance, but cannot completely eliminate the O2 effect, as reflected by the significant negative impact of flue gas O2 level on the overall NOx conversion in the ICFB reactor.   Figure 5.14 Comparison of NOx conversion between ICFB and fluidized bed reactors (Fe/ZSM-5(PUC)) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  175  5.4 Performance of Fe/ZSM-5(Albemarle) catalyst 5.4.1 Effect of HC:NO molar ratio on NOx and HC conversions For the flue gas containing 4% O2 and a draft tube gas velocity (UD) of 0.6 m/s, the influence of HC:NO ratio on NOx reduction was shown in Figure 5.15. It is very clear that as the HC:NO molar ratio increased from 1 to 4, NOx conversion increased notably, especially when HC:NO increased from 2 to 4. At HC:NO=1, the injected HC was completely converted. When HC:NO reached 2, HC was almost completely converted to CO or CO2 at a low UA. However, as UA increased to 0.42 m/s, HC conversion decreased quickly from 100% to 70%. The lowered HC conversion appeared to be directly related to the lowered NOx conversion. Further increase of HC:NO from 2 to 4 led to even lower HC conversion. As UD increased from 0.6 to 0.9 m/s, as shown in Figures 5.15, 5.16 and 5.17, NOx conversion increased, while HC conversion decreased. The increase in UD increased the solids circulation rate, thus increasing the NOx adsorption ratio in the annulus, leading to increased NOx conversion. On the other hand, high UD decreased the HC residence time in the draft tube, and thus lowered HC conversion. At UD=0.9 m/s (Figure 5.17), HC conversion was lower than 50% for HC:NO=4, and the influence of HC:NO on NOx conversion was not as significant as for the case with UD=0.60 and 0.75 m/s. It is observed that, at a given UA, high NOx conversions could be reached by either increasing the HC:NO molar ratio or increasing the gas velocity in the draft tube (UD). In other words, a high UD can be used to achieve the same NOx conversion at a low HC:NO ratio. The same conclusion could be drawn from Figures 5.18 to 5.20 for flue gases with 8% O2.  It is also noticed that NOx conversion became lower and HC conversion became higher as the flue gas O2 concentration increased from 4 to 8%. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  176  Figure 5.15 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.6 m/s, [O2]=4%)   Figure 5.16 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.75 m/s, [O2]=4%) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  177  Figure 5.17 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.9 m/s, [O2]=4%)   Figure 5.18 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.6 m/s, [O2]=8%) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  178  Figure 5.19 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.75 m/s, [O2]=8%)   Figure 5.20 Effect of HC:NO ratio on NOx and HC conversions (ICFB, Fe/ZSM- 5(Albemarle), UD=0.9 m/s, [O2]=8%) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  179 5.4.2 Effect of flue gas O2 concentration on NOx conversion and adsorption The influence of flue gas O2 concentration at the reactor inlet on NOx conversion and adsorption with HC:NO=2 and UD=0.6 m/s is shown in Figure 5.21. At a given UA, NOx adsorption ratio decreased when O2 concentration increased from 4% to 8%, and the further increase of O2 concentration from 8% to 12% showed little influence on NOx adsorption. Correspondingly, NOx conversion decreased with the increase in O2 concentration. For a given HC:NO molar ratio, as the O2 concentration in the flue gas increased, the O2 concentration in the draft tube increased proportionally due to gas bypassing from the annulus to the draft tube. Therefore, more hydrocarbons were oxidized to CO2 in the draft tube. As a result, more active sites on the catalyst were occupied by adsorbed CO2 when the regenerated catalyst circulated back to the annulus region, which led to a decrease in the NOx adsorption ratio. At a given O2 concentration, NOx adsorption ratio was always lower than NOx conversion, except for 4% and 8% O2 at UA<0.25 m/s. When UD increased from 0.75 to 0.9 m/s, as shown in Figures 5.22 to 5.23, NOx conversion was always higher than NOx adsorption ratio under a given operating condition. The highest NOx conversion and adsorption ratio were always reached at 4% O2, and the increase of O2 from 8% to 12% showed only marginal impact on both NOx conversion and adsorption. On the other hand, NOx adsorption ratio is seen to increase with increasing UD, likely resulting from increased solids circulation rate. In all cases, the difference of NOx conversion between various flue gas O2 contents decreased with increasing UD, indicating that the increase of O2 concentration exerted less negative impact on NOx conversion at a higher UD. In other words, higher UD is preferred to maintain a high NOx conversion by enhancing NOx adsorption ratio when a flue gas with a high O2 concentration is to be treated. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  180 The fact that NOx conversion was generally higher than NOx adsorption ratio in most cases in Figures 5.21 to 5.23 indicated that the overall performance of the ICFB with Fe/ZSM-5(Albemarle) was limited by the relatively poor adsorption performance of the catalyst. Doping of other NOx adsorbent, such as barium oxide, to the current Fe/ZSM- 5(Albemarle) catalyst can probably further improve the adsorption performance and thus increase the NOx conversion.  Figure 5.21 Effect of flue gas O2 concentration on NOx conversion and adsorption (ICFB, Fe/ZSM-5(Albemarle), UD=0.6 m/s, HC:NO=2) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  181  Figure 5.22 Effect of flue gas O2 concentration on NOx conversion and adsorption (ICFB, Fe/ZSM-5(Albemarle), UD=0.75 m/s, HC:NO=2)  Figure 5.23 Effect of flue gas O2 concentration on NOx conversion and adsorption (ICFB, Fe/ZSM-5(Albemarle), UD=0.9 m/s, HC:NO=2) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  182  5.4.3 Effect of gas velocities on NOx conversion and adsorption The influence of gas velocities on NOx conversion and adsorption is shown in Figure 5.24 with HC:NO=1, and Figure 5.25 with HC:NO=2 at a flue gas inlet O2 concentration of 4%. It is seen that both UA and UD significantly influenced NOx conversion and adsorption. In Figure 5.24, NOx conversion increased monotonically with increasing UD, while NOx adsorption ratio first decreased as UD increased from 0.6 to 0.75 m/s, and then increased as UD further increased from 0.75 to 0.9 m/s. For HC:NO=2 (Figure 5.25), UD showed more influence on NOx conversion than on NOx adsorption, although there is a lack of clear trend on the effect of UD on NOx adsorption. For both cases, NOx conversion and adsorption ratio generally decreased with increasing UA, and the influence of UD on NOx adsorption ratio was more significant at a low UA than at a high UA. For UD=0.6 m/s (Figure 5.24) or UD=0.45 m/s (Figure 5.25), NOx conversion was approximately equal to the NOx adsorption ratio, meaning that the adsorbed NOx was completely converted by the catalyst in the draft tube. At high values of UD, NOx conversion was higher than the NOx adsorption ratio, and their difference increased as UD increased from 0.75 to 0.9 m/s in Figure 5.24 or from 0.6 to 1.05 m/s in Figure 5.25. One of possible explanations is that, at low draft tube gas velocities (UD), when catalyst particles moved up the draft tube and left from the top outlet, they immediately dropped down to the annulus region without ejecting to the upper freeboard region. At a high UD, particles leaving the draft tube might continue their upward movement, forming a “fountain” region, as in a spouted bed, which delayed the catalyst particles returning to the annulus region. As a result, NOx escaping from the annulus adsorption region and the draft tube reaction region was reduced in the freeboard region due to the high solids holdup in the “fountain” region, creating the difference between the measured total NOx conversion and the Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  183 adsorption ratio. The increase in gas velocity in the annulus region (UA) decreased the contact time between the flue gas and the catalyst, leading to a decrease in the NOx adsorption ratio, which, as a limiting step, reduced the overall NOx conversion.  Figure 5.24 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=4%, HC:NO=1)  Figure 5.25 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=4%, HC:NO=2) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  184  The influence of gas velocities on NOx conversion and adsorption for 8% and 12% flue gas O2 concentration is shown in Figures 5.26 and 5.27, respectively, both with HC:NO=2. UA showed similar influence on NOx conversion and adsorption ratio as at 4% O2. However, NOx adsorption ratio showed a steady increase with increasing UD. This further confirms the postulate that the high solids circulation rate at a high draft tube gas velocity, UD, promotes NOx adsorption in the annulus region because of lower NOx concentration on the catalyst surface. Again, for the same reason described earlier, NOx conversion was always higher than the adsorption ratio, especially at a high UD. For O2=12%, the annulus gas velocity UA had little influence on both NOx conversion and adsorption ratio.  Figure 5.26 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=8%, HC:NO=2)  Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  185  Figure 5.27 Effect of gas velocities on NOx conversion and adsorption (ICFB, Fe/ZSM- 5(Albemarle), [O2]=12%, HC:NO=2)  5.4.4 Performance of the fluidized bed reactor with Fe/ZSM-5(Albemarle) 5.4.4.1 Effect of HC:NO ratio on NOx and HC conversions Figures 5.28 to 5.30 show that the gas velocity in the fluidized bed with Fe/ZSM- 5(Albemarle) catalyst had less influence on NOx conversion than with Fe/ZSM-5(PUC). Increasing HC:NO ratio had a positive impact on NOx conversion but negative impact on HC conversion in all cases, especially at low gas velocities. For O2=1% (Figure 5.28), NOx conversion increased by ~10%, while HC conversion decreased by ~5% as the HC:NO ratio increased from 1 to 2. For O2=4% (Figure 5.29), when HC:NO increased from 1 to 2, NOx conversion increased to ~40% from less than 30%, while HC conversion decreased. Further increase of HC:NO from 2 to 4 gave 10% further increase of NOx conversion and about a 5% Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  186 drop in HC conversion. A similar trend was observed when HC:NO increased from 1 to 2 at 8% O2 (Figure 5.30). As discussed previously for Fe/ZSM-5(PUC) catalyst, NOx and HC conversions were significantly influenced by varying the gas velocity in the fluidized bed. Figures 5.28 to 5.30 show that NOx conversion was enhanced with increasing gas velocity in the fluidized bed with Fe/ZSM-5(Albemarle) particles under most experimental conditions, contrary to the Fe/ZSM-5(PUC) catalyst where NOx conversion decreased as the gas velocity increased. One possible explanation is that the interphase mass transfer between the bubble phase and the dense phase in the fluidized bed with fine Fe/ZSM-5(Albemarle) catalyst improved significantly due to increased solids holdup in the freeboard region as the gas velocity increased. As a result, NOx conversion increased. Although HC conversion decreased with increasing gas velocity for Fe/ZSM-5 (Albemarle), the influence of the gas velocity on HC conversion was not as significant as for the Fe/ZSM-5(PUC) catalyst.  Figure 5.28 Effect of HC:NO ratio on NOx and HC conversions (Fluidized bed, Fe/ZSM- 5(Albemarle), [O2]=1%) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  187  Figure 5.29 Effect of HC:NO ratio on NOx and HC conversions (Fluidized bed, Fe/ZSM- 5(Albemarle), [O2]=4%)  Figure 5.30 Effect of HC:NO ratio on NOx and HC conversions (Fluidized bed, Fe/ZSM- 5(Albemarle), [O2]=8%) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  188  5.4.4.2 Effect of O2 concentration on NOx and HC conversions Figures 5.31 and 5.32 show that increasing the flue gas O2 concentration had significant negative impact on NOx conversion in the fluidized bed using Fe/ZSM-5(Albemarle) catalyst, where the increase of O2 concentration could greatly reduce NOx conversion with a slightly increased HC conversion. In Figure 5.31, at 1% O2 and HC:NO=1, NOx conversion reached around 51%, which was the same as obtained in the fixed bed experiment, but the HC conversion was much lower than in the fixed bed experiment. However, at O2=4% and 8% and HC:NO=1, both NOx and HC conversions in the fluidized bed were much lower than in the fixed bed. This means that the SCR performance of the Fe/ZSM-5(Albemarle) catalyst in the fluidized bed became worse than in the fixed bed when the reactor was operated at high O2 concentrations with a low reductant-to-NOx feed ratio. This is likely related to the existence of interphase mass transfer between the bubble and the dense phase in the fluidized bed and the reduced mean gas-solids contact time due to the gas bypass through bubbles. The increase in the gas velocity slightly enhanced NOx reduction but decreased HC conversion at a given O2 concentration. This could be explained by the improved interphase mass transfer, on one hand, and shortened contact time, on the other hand, as the gas velcoity increased in the fluidized bed. As HC:NO ratio increased from 1 to 2, as shown in Figure 5.32, the same trend was observed for the influence of O2 concentration on NOx and HC conversions as found in Figure 5.31. Although NOx conversion was improved by 10% compared to HC:NO=1 at the same O2 concentration, it was still 6-10% lower (for NOx conversion) and 45~50% lower (for HC conversion) than that in the fixed bed. The increase of the gas velocity had no significant influence on NOx conversion, but decreased the HC conversion at a given O2 concentration. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  189  Figure 5.31 Effect of inlet O2 concentration on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(Albemarle), HC:NO=1)  Figure 5.32 Effect of inlet O2 concentration on NOx and HC conversions (Fluidized bed, Fe/ZSM-5(Albemarle), HC:NO=2) Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  190  5.4.5 Comparison of reactor types Figure 5.33 compares NOx conversion in the fluidized bed reactor and the ICFB reactor. The results from the fluidized bed experiment with 1% and 4% O2 were plotted. For the ICFB reactor, the conditions with UD=0.45 and 1.05 m/s and O2=4% were selected because they represented the lowest (UD=0.45 m/s) and highest (UD=1.05 m/s) NOx conversion. NOx conversions at UD=0.9 m/s with O2=8% and 12%, are also plotted. When the ICFB reactor was operated at UD=0.45 m/s which had a low solids circulation rate, the ICFB showed similar performance as the fluidized bed reactor with a NOx conversion of ~40% for the flue gas with 4% O2, which means that the ICFB reactor had no advantage if operated at a low UD. When UD increased to 1.05 m/s, ICFB showed much better performance than the fluidized bed, with the NOx conversion close to, or even higher than achieved in the fluidized bed with a flue gas of only 1% O2. At UD=0.9 m/s, NOx conversion for 8% and 12% O2 in the ICFB reactor was 10% higher than that achieved in the fluidized bed for the flue gas with 4% O2 content when U (or UA) was lower than 0.4 m/s. This result clearly demonstrates the advantage of the ICFB reactor over the fluidized bed reactor for the treatment of flue gases containing excessive O2, and proves that the dual-zone ICFB reactor can almost completely eliminate the negative impact imposed by the high flue gas O2 concentration. Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  191  Figure 5.33 Comparison of NOx conversion between ICFB and fluidized bed reactors (Fe/ZSM-5(Albemarle))  5.5 Summary Experiments on the catalytic reduction of NOx with propylene in the presence of excessive O2 were carried out over Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle) catalysts in the fluidized bed and ICFB reactor, respectively. To pursue a better performance of the ICFB reactor for the reduction of NOx, the annulus must be used as the adsorption zone and the draft tube as the reduction zone. Increasing the catalyst load will be of benefit to NOx adsorption ratio as well as NOx conversion in ICFB reactor. Increasing the ratio of HC:NO improved the reduction performance of both Fe/ZSM- 5(PUC) and Fe/ZSM-5(Albemarle) catalysts in both the fluidized bed and ICFB reactor. For both catalysts, at a given ratio of HC:NO, NOx conversion decreased with increasing UA in Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  192 the ICFB reactor. On the contrary, NOx conversion increased with an increase of UD. Using a higher UD, the catalyst operated under a lower HC:NO ratio could reach the similar catalytic activity as that using a lower UD but a higher HC:NO ratio. NOx conversion and adsorption ratio decreased with increasing UA for both catalysts. NOx conversion increased with increasing UD because more particles were entrained into the gas mixture of the flue gas and the reductant gas, and formed a “fountain” in the freeboard region at higher UD. In most cases, the NOx conversion was lower than adsorption ratio for Fe/ZSM-5(PUC) which means that the catalytic activity was too low and could not match the adsorption performance. In contrast, the NOx conversion was higher than the adsorption ratio for Fe/ZSM-5(Albemarle) meaning that the adsorption performance should be improved in order to fully utilize the catalytic activity of the catalyst. Inlet NO concentration of the model flue gas had some impacts on NOx conversion over Fe/ZSM-5(PUC) in the fluidized bed, because the change of NOx concentration influences the ability of NOx to be adsorbed by the catalyst in a fixed period of contact time (or gas velocity). Lower O2 concentration could greatly enhance the catalytic activity of both Fe/ZSM- 5(PUC) and Fe/ZSM-5(Albemarle) in the fluidized bed, which agrees well with the findings in the fixed bed experiment from this study and other research groups. The performance of Fe/ZSM-5(PUC) was not as good as in the fixed bed experiment. In contrast, Fe/ZSM-5(Albemarle) showed promising reduction performance and strong inhibiting ability on the negative impact of excessive O2 in the ICFB reactor. Using Fe/ZSM- 5(Albemarle), at the same gas velocity in the annulus of the ICFB reactor and the fluidized bed reactor, and HC:NO=2, the NOx conversion in the ICFB reactor with the flue gas Chapter 5: Adsorption and Reduction Performance of the ICFB Reactor  193 containing 4% O2 was higher than in the fluidized bed with the flue gas containing 1% O2 if UD=1.05 m/s was used in the ICFB reactor.   194  Chapter 6 Conclusions and Recommendations for Future Work  6.1 Conclusions The objectives of this research were to identify the possible application of the integrated adsorption-reduction process for the selective catalytic reduction of NOx with hydrocarbons (HC-SCR) over selected catalysts in an internal circulating fluidized bed. The adsorption and reaction performances of the selected catalysts were investigated in a fixed bed reaction system. Hydrodynamic studies were carried out in a cold model ICFB unit. The concept of the integrated adsorption-reduction process of NOx was implemented by means of a hot model ICFB reactor, and the reactor performance was tested under various operating conditions. The following conclusions can be drawn.  6.1.1 Adsorption and reaction performance of Fe/ZSM-5 catalyst The adsorption performance of Fe/ZSM-5 catalysts tested in this study is closely related to the particle size and the structure of the catalyst support. The fine Fe/ZSM-5(Albemarle) catalyst exhibited highest NOx adsorption capacity. The presence of H2O in the model flue gas flow had almost no effect on the adsorption capacity of Fe/ZSM-5(Albemarle). In contrast, a negative impact was observed when CO2 was added into the flue gas, and the addition of H2O into a CO2-containing flue gas slightly improved the adsorption performance of the Fe/ZSM-5(Albemarle) catalyst. The catalytic activity of Fe/ZSM-5 was sensitive to the reaction temperature and space velocity. The catalyst deactivated very quickly at temperatures lower than 325oC Chapter 6: Conclusions and Recommendations for Future Work  195 because of the deposit of graphitic carbon on the catalyst surface. Among tested catalysts, Fe/ZSM-5 prepared by IMPO method exhibited stable activity at T350oC and GHSV=5000 h-1, and the reaction activity of the catalyst is not directly proportional to the adsorption capacity in the conventional HC-SCR process. A HC:NO molar ratio of 2 is recommended with the consideration of both the economic aspect and the reduction efficiency of NOx. O2 concentration played an important role in the SCR of NOx with propylene as the reducing agent.  Fe/ZSM-5 exhibited acceptable activity when O2 concentration was controlled at relatively low level (≤1%). H2O could slightly enhance the activity of Fe/ZSM- 5(Albemarle) in HC-SCR, while CO2 showed negligible effect. The catalyst could be permanently deactivated by SO2 in HC-SCR, and only partial catalytic activity could be recovered via calcination at a high temperature.  6.1.2 Hydrodynamic study of the ICFB reactor In the cold model experiment, both the flat and cylindrical distributors showed relatively narrow window of operating gas velocities, while the conical distributor exhibited flexible and stable operation in a wide range of velocities for both UA and UD. All three types of distributors showed very low gas bypass from the draft tube to the annulus but high gas bypass at the reverse direction. The increase of UD could enhance RAD while UA showed less effect on RAD. RAD and RDA were sensitive to the change of the effective gap opening between the draft tube and the annulus gas distributor. The solids circulation rate was determined by the combined effect of UA and UD for the conical and cylindrical gas distributors. In the ICFB hot model experiment, RDA increased with the increase of UD, and RDA was always higher than RAD for both Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle) catalysts. Chapter 6: Conclusions and Recommendations for Future Work  196 At given UA and UD, the fine Fe/ZSM-5(Albemarle) gained higher RDA but lower RAD than the coarse Fe/ZSM-5(PUC). The gas bypass ratio of Fe/ZSM-5(PUC) from the annulus to draft tube (RAD) reached a maximum value when the draft tube gas velocity (UD) was 0.9~1.0 m/s in this study. The difference in the gas bypass characteristics between the cold model and the hot model unit further shows that the reactor configuration could have a significant impact on the performance of the ICFB reactor.  6.1.3 Adsorption and reduction performance of Fe/ZSM-5 in the ICFB reactor For both Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle) catalysts, NOx conversion and adsorption ratio decreased with the increase of UA. NOx conversion increased with the increase of UD. At a high UD, the catalyst operated at low HC:NO molar ratio could reach similar catalytic reactivity as that using a lower UD but a higher HC:NO ratio. Increasing the HC:NO molar ratio improved the reduction performance in both the fluidized bed and the ICFB reactor. To obtain better catalytic activity, lower O2 concentration is preferred in the fluidized bed. The performance of Fe/ZSM-5(PUC) in the ICFB reactor was not as good as exhibited in the fixed bed experiment. In contrast, Fe/ZSM-5(Albemarle) showed promising reduction performance and great ability to inhibit the negative impact of excessive O2 in the ICFB reactor. Using Fe/ZSM-5(Albemarle) catalyst, at the same gas velocity in the annulus of the ICFB reactor and the fluidized bed reactor, the NOx conversion in the ICFB reactor with the flue gas containing 4% O2 was higher than that in the fluidized bed with the flue gas containing 1% O2 if UD=1.05 m/s was used in the ICFB reactor. However, the adsorption Chapter 6: Conclusions and Recommendations for Future Work  197 performance should be improved in order to further improve the catalytic activity of Fe/ZSM-5(Albemarle) catalyst. Overall, this study shows that such an ICFB reactor exhibited the ability to overcome the negative impact of excessive O2 in the flue gas using Fe/ZSM-5(Albemarle) as the deNOx catalyst.  6.2 Recommendations for future work Due to the complexity of HC-SCR process and the variation of the combustion flue gases, further study should be undertaken in the following aspects: 1. Although Fe/ZSM-5(Albemarle) catalyst showed acceptable catalytic activity in HC- SCR, its adsorption capacity is still not satisfactory according to the result from the hot model ICFB experiment. The impregnation of other metals, such as Ba, could improve the adsorption performance of Fe/ZSM-5(Albemarle). In addition, the catalytic activity of Fe/ZSM-5(Albemarle) was severely inhibited by SO2. Therefore, new catalysts with good tolerance to SO2 and hydrothermal stability need to be developed. 2. Other hydrocarbons, such as propane and i-butane, should be studied as reducing agents to compare their NOx reduction performance with propylene. 3. More detailed investigation on the effect of H2O, SO2, CO2 and the Fe loading on the catalyst performance should be conducted in the fixed bed experiment. 4. An in-situ gas chromatograph (GC) is needed for the measurement of the hydrocarbon, CO2 and N2O concentrations in order to evaluate the HC conversion more precisely and to monitor the conversion of NOx to N2O. Chapter 6: Conclusions and Recommendations for Future Work  198 5. To minimize the experimental error, rotameters used in the fixed bed experiment should be replaced by mass flow meters for the better control of gas flow rates. 6. The solids circulation rate should be measured in the cold model unit using the annulus gas distributor with the same configuration as in the hot model unit and Fe/ZSM-5(Albemarle) or similar bed materials. The attrition of the catalyst should be examined as well. 7. Effect of H2O, SO2 and CO2 on the reduction performance of Fe/ZSM-5(Albemarle) catalyst should be investigated in the hot model ICFB reactor. 8. The time-on-stream experiment should be performed using the real flue gases from a natural gas burner or a gasoline or diesel engine to evaluate the performance of the ICFB reactor. 9. As discussed previously, the ICFB reactor performance was limited by the catalyst adsorption capacity, because the overall NOx conversion increased with decreasing gas velocity in the annulus. To promote the adsorption performance of the catalyst, a configuration with larger annulus cross-sectional area should be considered to decrease the annulus gas velocity (or increase the residence time) while keeping the annulus gas flux at a specific amount. For the same reason, the effect of the catalyst loading in the reactor should be investigated as well. 10. The gas bypass from the annulus to draft tube should be more flexible in order to control the O2 concentration in the draft tube to a desired level. In light of this, other types of gas distributors and/or the relative position between the draft tube and gas nozzle should be investigated, and small diameter gas nozzles should be used to investigate the gas bypass from the draft tube to annulus, in order to minimize the bypass of the reductant gas to the annulus. Chapter 6: Conclusions and Recommendations for Future Work  199 11. Temperature programmed desorption (TPD), in-situ electron paramagnetic resonance (EPR), in-situ Fourier transform infrared spectroscopy (FT-IR) or other methods should be performed to elucidate the nature of NOx adsorption and reduction over Fe/ZSM-5(Albemarle) catalyst. 12. Based on the result from the cold model experiment and the reaction kinetics from fixed bed tests, a mathematical model of the ICFB reactor needs to be developed for the prediction of gas and solids flow structure in the ICFB reactor and the reactor performance for NOx reduction.   200 References Abdulhamid, H.; Fridell, E.,; Skoglundh, M. (2004). Influence of the type of reducing agent (H2, CO, C3H6 and C3H8) on the reduction of stored NOx in a Pt/BaO/Al2O3 model catalyst. Topics Catal., 30/31, 161-168.  Adams, K.M.; Cavataio, J.V.; Hammerle, R.H. (1996). Lean NOx catalysis for diesel passenger cars: Investigating effects of sulfur dioxide and space velocity. Appl.Catal.B: Environ., 10, 157-181.  Adelman, B. J.; Beutel, T.; Lei, G.-D.;  Sachtler, W. M. H. (1996). Mechanistic cause of hydrocarbon specificity over Cu/ZSM-5 and Co/ZSM-5 catalysts in the selective catalytic reduction of NOx. J. Catal., 158, 327-335.  Adelman, B.J.; Sachtler, W.M.H. (1997). The effect of zeolitic protons on NOx reduction over Pd/ZSM-5 catalysts. Appl. Catal., B: Environ., 14, 1-11.  Amiridis, M.D.; Roberts, K.L.; Pereira, C.J. (1997). The selective catalytic reduction of NO by propylene over Pt supported on dealuminated Y zeolite. Appl. Catal., B: Environ., 14, 203-209.  Amiridis, M.D.; Zhang, T.; Farrauto, R.J. (1996). Selective catalytic reduction of nitric oxide by hydrocarbons. Appl. Catal. B: Environ., 10, 203-227.  Ando, S.; Maki, T.; Nakagawa, Y.; Namiki, N.; Emi, H.; Otani, Y. (2002). Analysis of the drying process of seed particles in a spouted bed with a draft tube. Advanced Powder Technol., 103(1), 73-91.  Angelidis, T. N.; Christoforou, S.; Bongiovanni, A.; Kruse, N. (2002). On the promotion by SO2 of the SCR process over Ag/Al2O3: influence of SO2 concentration with C3H6 versus C3H8 as reductant.  Appl. Catal., B: Environ., 39(3), 197-204.  Armor, J.N. (1997). Catalytic solutions to reduce pollutants. Catal. Today, 38, 163-167.  Armor, J.N.; Li, Y. (1995). The effect of SO2 on the catalytic performance of Co-ZSM-5 and Co-ferrierite for the selective reduction of NO by CH4 in the presence of O2. Appl. Catal., B: Environ., 5, L257-L270.  Baerns, M.; Rottlander, C.; Andorf, R.; Plog , C.; Krutzsch B. (1996). Selective NO reduction by propane and propene over a Pt/ZSM-5 catalyst: A transient study of the reaction mechanism. Appl. Catal., B: Environ., 11, 49-63.  Battiston, A.A.; Bitter, J.H.; Koningsberger, D.C. (2003). Reactivity of binuclear Fe complexes in over-exchanged Fe/ZSM5, studied by in situ XAFS spectroscopy 2. Selective catalytic reduction of NO with isobutane  J. Catal., 218, 163-177.  References  201 Beretta, A.; Orsenigo, C.; Ferlazzo, N.; Tronconi, E.; Forzatti, P.; Berti, F. (1998). Analysis of the performance of plate-type monolithic catalysts for selective catalytic reduction DeNOx applications. Ind. Eng. Chem. Res., 37, 2623-2633.  Berndt, H.; Schutze, F.H.; Richter, M.; Sowade, T.; Grunert, W. (2003). Selective catalytic reduction of NO under lean conditions by methane and propane over indium/cerium- promoted zeolites. Appl. Catal. B: Environ., 40, 51-67.  Bogner, W.; Krämer, M.; Krutzsch, B.; Pischinger, S.; Voigtländer, D.; Wenninger, G.; Wirbeleit, F.; Brogan, M.S.; Brisley R.J.; Webster, D.E. (1995). Removal of nitrogen oxides from the exhaust of a lean-tune gasoline engine. Appl. Catal. B: Environ., 7, 153-171.  Boix, A.; Miro, E. E.; Lombardo, E. A.; Banares, M. A.; Mariscal, R.; Fierro, J. L. G. (2003). The nature of cobalt species in Co and Pt-Co/ZSM5 used for the SCR of NOx with CH4. J. Catal.,  217(1), 186-194.  Bosch, H.; Janssen, F. (1988).  Catalytic reduction of nitrogen oxides. A review on the fundamentals and technology. Catal. Today, 2(4), 369-521.  Brilhac, J.F.; Sultana, A.; Gilot, P.; Martens, J.A. (2002). Adsorption and pressure swing desorption of NOx in Na-Y zeolite: Experiments and Modeling. Environ. Sci. Tech., 36, 1136-1140.  Burch, R.; Halpin, E.; Sullivan, J.A. (1998). A comparison of the selective catalytic reduction of NOx over Al2O3 and sulphated Al2O3 using CH3OH and C3H8 as reductants. Appl. Catal., B: Environ., 17, 115-129.  Burch, R.; Ottery, D. (1996a). Selective catalytic reduction of NOx by hydrocarbons on Pt/Al2O3 catalysts at low temperatures without the formation of N2O. Appl. Catal., B: Environ., 9, L19-L24.  Burch, R.; Scire, S. (1994). Selective catalytic reduction of nitric oxide with ethane and methane on some metal exchanged ZSM-5 zeolites. Appl. Catal. B: Environ., 3,295-318.  Burch R.; Watling, T.C. (1996b). Adsorbate-assisted NO decomposition in NO reduction by C3H6 over Pt/Al2O3 catalysts under lean-burn conditions. Catal. Lett., 37, 51-55.  Bustamante, F.; Cordoba, F.; Yates, M.; Correa, C. (2002). The promotion of cobalt mordenite by palladium for the lean CH4-SCR of NOx in moist streams. Appl. Catal., A: General, 234, 127-136.  Campa, M.C.; De Rossi, S.; Ferraris, G.; Indovina, V. (1996) Catalytic activity of Co-ZSM- 5 for the abatement of NOx with methane in the presence of oxygen. Appl. Catal., B: Environ., 8(3), 315-331.  Cant, N.W.; Liu, I.O.Y. (2000). The mechanism of the selective reduction of nitrogen oxides by hydrocarbons on zeolite catalysts. Catal. Today, 63, 133-146.  References  202 Capek, L.; Vradman, L.; Sazama, P.; Herskowitz, M.; Wichterlova, B.; Zukerman, R.; Brosius, R.; Martens, J.A. (2007). Kinetic experiments and modeling of NO oxidation and SCR of NOx with decane over Cu- and  Fe -MFI catalysts. Appl. Catal., B: Environ., 70(1-4), 53-57.  Captain, D. K.; Amiridis, M. D. (1999). In situ FTIR studies of the selective catalytic reduction of NO by C3H6 over Pt/Al2O3.  J.  Catal., 184(2), 377-389.  Captain, D. K.; Roberts, K.L.; Amiridis, M. D. (1998). The selective catalytic reduction of nitric oxide by propylene over Pt/SiO2. Catal. Today, 42, 93-100.  Centi, G.; Generali, P.; dall’Olio, L.; Perathoner, S. (2000). Removal of N2O from industrial gaseous streams by selective adsorption over metal-exchanged zeolites. Ind. Eng. Chem. Res., 39, 131-137  Chen, H.Y.; Sachtler, W. M. H. (1998a).  Activity and durability of Fe/ZSM-5 catalysts for lean burn NOx reduction in the presence of water vapor. Catal. Today, 42(1-2), 73-83.  Chen, H.Y.; Voskoboinikov, T.; Sachtler, W. M. H. (1998b). Reduction of NOx over Fe/ZSM-5 catalysts: Adsorption complexes and their reactivity toward hydrocarbons. J. Catal.,  180(2),  171-183.  Chen, H.Y.; Voskoboinikov, T.; Sachtler, W. M. H. (1999). Reduction of NOx over Fe/ZSM-5 catalysts: Mechanistic causes of activity differences between alkanes. Catal. Today, 54, 483-494.  Chen, H.Y.; Wang, X.; Sachtler, W. M. H. (2000a). Reduction of NOx over zeolite MFI supported iron catalysts: Nature of active sites. Phys. Chem. Chem. Phys., 2(13), 3083- 3090.  Chen, H.Y.; Wang, X.; Sachtler, W.M.H. (2000b). Reduction of NOx over various Fe/zeolite catalysts. Appl. Catal. A: General, 194/195, 159-168  Chin, Y.-H.; Pisanu, A.; Serventi, L.; Alvarez, W. E.; Resasco, D. E. (1999). NO reduction by CH4 in the presence of excess O2 over Pd/sulfated zirconia catalysts.  Catal. Today, 54(4), 419-429.  Cho, B.K. (1997). Activity normalization for lean-NOx catalysts. Appl. Catal. B: Environ., 12, 147-160.  Cho, B. K.; Blint, R. J.; Subbiah, A. (2003). Hydrothermally stable catalyst for improved lean NOx reduction, Eur. Pat. Appl., EP 1304165, 12 pp.  Cho, B.K.; Yie, J.E.; Rahmoeller, K.M. (1995). Autonomous kinetic oscillations during NO + C2H4 + O2 reaction over Pt-ZSM-5 under highly oxidizing conditions. J. Catal., 157, 14-24.  References  203 Corma, A.; Fornes V.; Palomares E. (1997). Selective catalytic reduction of NOx on Cu- beta zeolites. Appl. Catal., B: Environ., 11, 233-242.  Cowan, A.D.; Dumpelmann, R.; Cant, N.W. (1995). The rate-determining step in the selective reduction of nitric-oxide by methane over a Co-ZSM5 catalyst in the presence of oxygen. J. Catal., 151, 356-363.  Decyk, P.; Kim, D.K.; Woo, S.I. (2001). Poisoning effect of SO2 on NO reduction by i- butane over Fe/ZSM-5 prepared by sublimation method. J. Catal., 203 (2), 369–374.  Delahay, G.; Ensuque, E.; Coq, B.; Figueras, F. (1998). Selective catalytic reduction of nitric oxide by n-decane on Cu/sulfated-zirconia catalysts in oxygen rich atmosphere: Effect of sulfur and copper contents. J.  Catal., 175, 7-15.  Delahay, G.; Guzman-Vargas, A.; Coq, B. (2007).   Deactivation of a Fe-ZSM-5 catalyst during the selective catalytic reduction of NO by n-decane: An operando DRIFT study. Appl. Catal., B: Environ., 70(1-4), 45-52.  Delahay, G.; Valade, D.; Guzman-Vargas, A.; Coq, B.. (2005). Selective catalytic reduction of nitric oxide with ammonia on Fe-ZSM-5 catalysts prepared by different methods. Appl. Catal., B: Environ., 55, 149-155.  De Lucas, A.; Valverde, J.L.; Dorado, F.; Romero, A.; Asencio, I. (2005). Influence of the ion exchanged metal (Cu, Co, Ni and Mn) on the selective catalytic reduction of NOx over mordenite and ZSM-5. J. Mol. Catal. A: Chem., 225, 47-58.  de Nevers, N. (2000). Air pollution control engineering. 2nd ed. McGraw-Hill Higher Education. 439-470.  Desai, A. J.; Kovalchuk, V. I.; Lombardo, E. A.; D'Itri, J. L. (1999). Co/ZSM-5: why this catalyst selectively reduces NOx with methane.  J.  Catal. , 184(2), 396-405.  Efthimiadis, E.A.; Lappas, A.A.; Iatrides, D.K.; Vasalos, I.A. (2001). Selective catalytic reduction of NOx by hydrocarbons/oxygenates. Application for the control of NOx from the regenerator of a fluid catalytic cracking pilot-plant unit. Ind. Eng. Chem. Res., 40, 515-521.  Efthimiadis, E. A.; Lionta, G. D.; Christoforou, S.C.; Vasalos, I.A. (1998). The effect of CH4, H2O and SO2 on the NO reduction with C3H6. Catal. Today, 40, 15-26.  Elkaim, D.; Bai, D. (2000). Selective catalytic reduction of nitrogen oxides. Canadian Patent Appl., CA 2,249,708, 46 pages.  Epling, W. S.; Campbell, G. C.; Parks, J. E. (2003). The effects of CO2 and H2O on the NOx destruction performance of a model NOx storage/reduction catalyst. Catal. Lett., 90, 45-56.  References  204 Feng, X.; Hall, W.K. (1996). On the unusual stability of overexchanged Fe-ZSM-5. Catal. Lett., 41, 45-46.  Feng, X.; Hall, W. K. (1997a). FeZSM-5: A durable SCR catalyst for NOx removal from combustion streams. J. Catal., 166, 368-376.  Feng, X.; Hall, W.K. (1997b). Limitations on the formation of oxygen-bridged divalent cations in FeZSM-5. Catal. Lett., 46, 11-16.  Forzatti, P. (2001). Present status and perspectives in De-NOx SCR catalysis. Appl. Catal., A: General, 222, 221-236.  Freitas, L.A.P; Freire, J.T. (2001). Gas-to-particle heat transfer in the draft tube of a spouted bed. Drying Tech., 19(6), 1065-1082.  Fridell, E.; Skoglundh, M.; Westerberg, B.; Johansson, S.; Smedler, G. (1999). NOx storage in barium-containing catalysts. J. Catal., 183, 196-209.  Fritz, A.; Pitchon, V. (1997). The current state of research on automotive lean NOx catalysis. Appl. Catal. B: Environ., 13, 1-25.  Fusey, I.; Lim, C.J.; Grace, J.R. (1986). Fast fluidization in a concentric circulating bed. in Circulating Fluidized Bed Technology II (P. Basu ed.) Pergamon Press, Oxford, pp.409- 416.  Gao, S.; Nakagawa, N.; Kato, K.; Inomata, M.; Tsuchiya, F. (1996). Simultaneous SO2/NOx removal by a powder-particle fluidized bed. Catal. Today, 29, 165-169.  Gao, Z.X.; Sun, Q.; Sachtler, W.M.H. (2001). Adsorption complexes of O2 on Fe/MFI and their role in the catalytic reduction of NOx. Appl. Catal., B: Environ., 33, 9-23.  Garcia-Cortes, J.M.; Perez-Ramirez, J.; Illan-Gomez, M.J.; Kapteijn, F.; Moulijn, J.A.; Salinas-Martinez de Lecea, C. (2001). Comparative study of Pt-based catalysts on different supports in the low-temperature de-NOx-SCR with propene. Appl. Catal., B: Environ., 30(3,4), 399-408.  Garcia-Cortes, J. M.; Perez-Ramirez, J.; Illan-Gomez, M. J.; Kapteijn, F.; Moulijn, J. A.; De Lecea, C. S.M. (2000). Effect of the support in de-NOx HC-SCR over transition metal catalysts.  React. Kinet. Catal. Lett., 70(2), 199-206.  Garcia-Cortes, J. M.; Perez-Ramirez, J.; Rouzaud, J.N.; Vaccaro, A.R.; Illan-Gomez, M. J.; De Lecea, C. S.M. (2003). On the structure sensitivity of denox HC-SCR over Pt-beta catalysts. J. Catal., 218, 111-122.  Gerlach, T.; Illgen, U.; Bartoszek, M.; Baerns, M. (1999). Reduction of nitrogen dioxide by propene over acidic mordenites: influence of acid site concentration, formation of by- products and mechanism. Appl. Catal., B: Environ., 22, 269-278.  References  205 Gilot P.; Guyon, M.; Stanmore, B.R. (1997). A review of NOx reduction on zeolitic catalysts under diesel exhaust conditions. Fuel, 76(6), 507-515.  Gomez, S. A.; Campero, A.; Martinez-Hernandez, A. ; Fuentes, G. A. (2000). Changes in Cu2+ environment upon wet deactivation of Cu-ZSM-5 deNOx catalysts. Appl. Catal. A: General, 197, 157-164.  Goralski, C.T.; Schneider, W.F. (2002). Analysis of the thermodynamic feasibility of NOx decomposition catalysis to meet next generation vehicle NOx emissions standards. Appl. Catal., B: Environ., 37(4), 263–277.  Goursot, A.; Coq, B.; Fajula, F. (2003). Toward a molecular description of heterogeneous catalysis: transition metal ions in zeolites. J. Catal., 216, 324-332.  Gutierrez, L.; Boix, A.; Petunchi, J.O. (1998). Promoting effect of Pt on Co-zeolites upon the SCR of NOx. J. Catal., 179(1), 179-191.  Hall, W.K.; Feng, X.; Dumesic, J.; Watwe, R. (1998). Problems in preparation of Fe/ZSM- 5 catalysts. Catal. Lett., 52, 13-19.  Ham, S.W.; Nam, I.S. (2002). Selective Catalytic Reduction of Nitrogen Oxides by Ammonia. Catalysis (The Royal Society of Chemistry), Vol.16, 236-271.  Hamada, H.; Kintaichi, Y.; Inaba, M.; Haneda, M.;Sato, K.; Fujimoto, T.; Kanai, S. (1997). Catalytic performance of silver ion-exchanged saponite for the selective reduction of nitrogen monoxide in the presence of excess oxygen. Appl. Catal., B: Environ., 13, 27-33.  Hammache, S.; Evans, L.R.; Coker, E.N.; Miller, J.E. (2008). Impact of copper on the performance and sulfur tolerance of barium-based NOx storage-reduction catalysts. Appl. Catal., B: Environ., 78(3-4), 315-323.  Heinisch, R.; Jahn, M.; Langohr, J.; Pohl, M. (2001). Modeling of a fixed bed reactor and kinetic investigations for the catalytic reduction of nitrogen oxides with methane. Chem. Eng. Technol., 24(4), 381-386.  Henderson M.A.; Szanyi J.; Peden C.H.F. (2003). Conversion of N2O to N2 on TiO2(110). Catal. Today, 85, 251-266.  Hodjati, S.; Bernhardt, P.; Petit, C.; Pitchon, V.; Kiennemann, A. (1998a). Removal of NOx: Part I. Sorption/desorption processes on barium aluminate. Appl. Catal., B: Environ., 19, 209-219.  Hodjati, S.; Bernhardt, P.; Petit, C.; Pitchon, V.; Kiennemann, A. (1998b). Removal of NOx: Part II. Species formed during the sorption/desorption processes on barium aluminates. Appl. Catal., B: Environ., 19, 221-232.  Hums E. (1998). Is advanced SCR technology at a standstill? A provocation for the academic community and catalyst manufacturers. Catal. Today, 42, 25-35. References  206  Hayes, N.W.; Joyer, R.W.; Shpiro, E.S. (1996). Infrared spectroscopy studies of the mechanism of the selective reduction of NOx over Cu-ZSM-5 catalysts. Appl. Catal., B: Environ., 8, 343-363.  Inui, T.; Iwamoto, S.; Kojo, S.; Shimizu, S.; Hirabayashi, T. (1994). Removal of nitric oxide on metallosilicate catalysts. Catal. Today, 22, 41-57.  Inui, T.; Iwamoto, S.; Kon, S.; Sakimon, T.; Kagawa, K. (1997). Evidently advantageous features of metallosilicates as the catalysts for elimination of NO in the exhaust gases containing a large excess of O2 and H2O. Catal. Today, 38, 169-174.  Ishikura, T.; Nagashima, H.; Ide, M. (2003). Hydrodynamics of a spouted bed with a porous draft tube containing a small amount of finer particles. Powder Technol., 131, 56- 65.  Iwamoto, M.; Hernandez, A. M.; Zengyo, T. (1997). Oxidation of NO to NO2 on a Pt-MFI zeolite and subsequent reduction of NOx by C2H4 on an In-MFI zeolite: a novel de-NOx strategy in excess oxygen. Chem. Comm., 37-38.  Iwamoto, M.; Takeda, H. (1996). Pulse study on reactivity of ethane adsorbed on Cu-MFI with nitrogen oxides and oxygen. Catal. Today, 27, 71-78.  Iwamoto, M.; Yahiro, H.; Tanda, K. (1988). Successful design of catalysts. Inui, T., ed., Elsevier, Amsterdam, p. 219.  Iwamoto, M.; Yahiro, H.; Yuu, Y.; Shundo, S.; Mizuno, N. (1990). Selective reduction of NO by lower hydrocarbons in the presence of O2 and SO2 over copper ion-exchanged zeolites. Shokubai, 32(6), 430-433.  Iwamoto M.; Zengyo, T.; Hernandez, A.M.; Araki, H. (1998). Intermediate addition of reductant between an oxidation and a reduction catalyst for highly selective reduction of NO in excess oxygen. Appl. Catal. B: Environ., 17, 259-266.  James, D.; Fourre, E.; Ishii, M.; Bowker, M. (2003). Catalytic decomposition / regeneration of Pt/Ba(NO3)2 catalysts: NOx storage and reduction.  Appl. Catal., B: Environ., 45(2), 147-159.  Javed, M.T.; Irfan, N.; Gibbs, B.M. (2007). Review: Control of combustion-generated nitrogen oxides by selective non-catalytic reduction. J. Environ. Manag., 83, 251–289.  Jen, H.W. (1998). Study of nitric oxide reduction over silver/alumina catalysts under lean conditions: Effects of reaction conditions and support. Catal. Today, 42, 37-44.  Jeong, S.M.; Kim, S.D. (2000). Removal of NOx and SO2 by CuO/γ-Al2O3 sorbent/catalyst in a fluidized-bed reactor. Ind. Eng. Chem. Res., 39, 1911-1916.  References  207 Jeong, S.K.; Park, T.S.; Hong, S.C. (2001). Simultaneous SOx/NOx removal in a fluidized bed reactor using natural manganese ore. J. Chem. Technol. Biotechnol., 76, 1080-1084.  Joyner, R.; Stockenhuber, M. (1997). Unusual structure and stability of iron-oxygen nano- clusters in Fe-ZSM-5 catalysts. Catal. Lett., 45, 15-19.  Joyner, R.; Stockenhuber, M. (1999). Preparation, Characterization, and performance of Fe-ZSM-5 catalysts. J. Phys. Chem. B, 103, 5963-5976.  Kagawa, K.; Ichikawa Y.; Iwamoto S.; Inui T. (1998). Broadening of the effective temperature range for NO removal on Co-Pd-modified H-ZSM-5 catalyst by suppression of CH4 combustion. Catal. Lett., 52, 145-149.  Kameoka, S.; Suzuki, T.; Yuzaki, K.; Takeda, T.; Tanaka, S.; Ito, S.; Miyadera, T.; Kunimori, K. (2000). Selective catalytic reduction of N2O with methane in the presence of excess oxygen over Fe-BEA zeolite. Chem. Comm., 745-746.  Kato, H.; Yokoyama C.; Misono, M. (1997). Rate-determining step of NO-CH4-O2 reaction catalyzed by Pd/H-ZSM-5. Catal. Lett., 47, 189-191.  Kato, K.; Takarada, Y.; Sagara, H.; Inomata, M. (1994). Simultaneous desulfurization and denitration of combustion flue gases with fluidized-bed catalysts. Jpn. Kokai Tokkyo Koho, JP 06170165, 6 pp.  Kato, T.; Shinoda, K.; Masuda, K.; Tsujimura, K. (1998). Activity enhancement of Ag/mordenite catalysts by addition of palladium for the removal of nitrogen oxides from diesel engine exhaust gas. Appl. Catal., B: Environ., 15, 29-35.  Kaucky, D.; Vondrova, A.; Dedecek, J.; Wichterlova, B. (2000). Activity of Co ion sites in ZSM-5, ferrierite, and mordenite in selective catalytic reduction of NO with methane. J. Catal., 194(2), 318-329.  Kawai, Y. (1992). Catalyst for treatment of exhaust gases.  Eur. Pat. Appl.  EP 499,286, 16 pp.  Keiski, R.L.; Raeisaenen, H.; Haerkoenen, M.; Maunula, T.; Niemistoe, P. (1996). NOx abatement in lean exhaust gas conditions over metal substrated zeolite catalysts. Catal. Today, 27, 85-90.  Kikuchi, E.; Ogura, M. (2000). Palladium species in Pd/H-ZSM-5 zeolite catalysts for CH4-SCR.   Res. Chem. Int., 26(1), 55-60.  Kikuchi, E.; Ogura, M.; Terasaki, I.; Goto Y. (1996). Selective reduction of nitric oxide with methane on gallium and indium containing H-ZSM-5 catalysts: Formation of active sites by solid-state ion exchange. J. Catal., 161, 465-470.  Kim, Y.J.; Lee, J.M.; Kim, S.D. (2000). Modeling of coal gasification in an internally circulating fluidized bed reactor with draught tube. Fuel, 79, 69-77. References  208  Kim, Y.T.; Song, B.H.; Kim, S.D. (1997). Entrainment of solids in an internally circulating fluidized bed with draft tube. Chem. Eng. J., 66, 105-110.  Kogel, M.; Monnig, R.; Schwieger, W.; Tissler, A.; Turek, T. (1999). Simultaneous catalytic removal of NO and N2O using Fe-MFI. J. Catal., 182, 470-478.  Kogel, M.; Sandoval, V.H.; Schwieger, W.; Tissler, A.; Turek, T. (1998). Simultaneous catalytic reduction of NO and N2O using Fe-MFI prepared by solid-state ion exchange. Catal. Lett., 51, 23-25.  Laurent F.; Pope, C.J.; Mahzoul, H.; Delfosse, L.; Gilot, P. (2003). Modeling of NOx adsorption over NOx adsorbers. Chem. Eng. Sci., 58, 1793-1803.  Lee, C.W.; Chong, P.J.; Lee, Y.C.; Chin, C.S.; Kevan, L. (1997). Wetness method preparation of catalysts for selective catalytic reduction of NO by propane. Catal. Lett., 48, 129-133.  Lee, L. (2000). Oxygen effect on the selective catalytic reduction of NOx using hydrocarbon reductants. Undergraduate thesis, UBC.  Lee, Y.W.; Choi, D.K.; Park, J.W. (2002). Characteristics of NOx adsorption and surface chemistry on impregnated activated carbon. Sep. Sci. Technol., 37(4), 937-956.  Li, J.; Hao, J.; Fu, L.; Liu, Z.; Cui, X. (2004). The activity and characterization of sol-gel Sn/Al2O3 catalyst for selective catalytic reduction of NOx in the presence of oxygen. Catal. Today, 90 (3–4), 215–221.  Li, L.; Chen, J.; Zhang, S.; Guan, N.; Richter, M.; Eckelt, R.,; Fricke, R. (2004). Study on metal-MFI/cordierite as promising catalysts for selective catalytic reduction of nitric oxide by propane in excess oxygen. J. Catal., 228 (1), 12–22.  Li, Y.; Armor, J.N. (1992a). Catalytic reduction of nitrogen oxides using methane in the presence of oxygen. US Patent 5,149,512, 10 pp.  Li, Y.; Armor, J.N. (1992b). Catalytic decomposition of nitrous oxide on metal exchanged zeolites.  Appl. Catal., B: Environ., 1, L21-L29.  Li, Y.; Armor, J.N. (1993a). Selective catalytic reduction of NOx with methane over metal exchange zeolites. Appl. Catal., B: Environ., 2, 239-256.  Li, Y.; Armor, J.N. (1993b). Metal exchanged ferrierites as catalysts for the selective reduction of NOx with methane. Appl. Catal., B: Environ., 3, L1-L11.  Li, Y.; Armor, J.N. (1994a). Selective reduction of NOx by methane on Co-Ferrierites: I. Reaction and kinetic Studies. J. Catal., 150, 376-387.  References  209 Li, Y.; Armor, J.N. (1994b). Selective catalytic reduction of NO with methane on gallium catalysts. J. Catal., 145, 1-9.  Li, Y.; Slager, T.L.; Armor, J.N. (1994c). Selective reduction of NOx by methane on Co- Ferrierites: II. Catalyst characterization . J. Catal., 150, 388-399.  Lick, I. D.; Carrascull, A.; Ponzi, M.; Ponzi, E.N. (2003a). The catalytic activity of Co/ZrO2 for NO reduction with propane in O2 presence. Catal. Lett., 89(3-4), 179-184.  Lick, I. D.; Ponzi, E. N. (2003b). NO reduction with C3H8 on Co/Al2O3 catalyst. Effect of O2 concentration.  React. Kinet.  Catal. Lett., 79(1), 197-203.  Lietti, L.; Forzatti, P.; Nova I.; Tronconi, E. (2001). NOx storage reduction over Pt---Ba/γ- Al2O3 catalyst. J. Catal., 204, 175-191.  Lima, E.; Guzmam-Vargas, A.; Mendez-Vivar J.,; Pfeiffer, P.; Fraissard, J. (2008). Fe-ZSM- 5 catalyst: Preparation in organic media, Fe-particle morphology and NOx reduction activity. Catal. Lett., 120, 244-251.  Liotta, L. F.; Pantaleo, G.; Macaluso, A.; Di Carlo, G.; Deganello, G. (2003). CoOx catalysts supported on alumina and alumina-baria: influence of the support on the cobalt species and their activity in NO reduction by C3H6 in lean conditions. Appl. Catal., A: General, 245(1), 167-177.  Liu, Z.; Woo, S. (2006). Recent advances in catalytic deNOx science and technology. Catal. Rev., 48(1), 43-89.  Lobree, L.J.; Aylor, A.W.; Reimer, J.A.; Bell, A.T. (1999). NO reduction by CH4 in the presence of O2 over Pd-H-ZSM-5. J. Catal., 181, 189-204.  Lobree, L.J.; Aylor, A.W.; Reimer, J.A.; Bell, A.T. (1997). Role of cyanide species in the reduction of NO by CH4 over Co-ZSM-5. J. Catal., 169, 188-193.  Lombardo, E.A.; Sill, G.A.; Ditri, J.L.; Hall, W.K. (1998). The possible role of nitromethane in the SCR of NOx with CH4 over M-ZSM5 (M=Co, H, Fe, Cu). J. Catal., 173, 440-449.  Long, R.Q.; Yang, R.T. (1998). Pt/MCM-41 catalyst for selective catalytic reduction of nitric oxide with hydrocarbons in the presence of excess oxygen. Catal. Lett., 52, 91-96.  Long, R. Q.; Yang, R. T. (1999). Catalytic performance of Fe–ZSM-5 catalysts for selective catalytic reduction of nitric oxide by ammonia. J. Catal., 188, 332–339.  Lucas, A.D.; Valverde, J.L.; Dorado, F.; Romero, A.; Asencio, I. (2004). Influence of the ion exchanged metal (Cu, Co, Ni and Mn) on the selective catalytic reduction of NOx over mordenite and ZSM-5. J. Mole. Catal. A, 225(1), 47–58.  References  210 Lukyanov, D.B.; Lombardo, E. A.; Sill, G. A.; Ditri, J. L.; Hall, W.K. (1996). Selective catalytic reduction (SCR) of NO with methane over CoZSM-5 and HZSM-5 zeolites: On the role of free radicals and competitive oxidation reactions. J. Catal., 163, 447-456.  Lukyanov, D.B.; Sill, G.; Ditri, J.L.; Hall, W.K. (1995). Comparison of catalyzed and homogeneous reactions of hydrocarbons for selective catalytic reduction (SCR) of NOx. J. Catal., 153, 265-274.  Mahzoul, H.; Brilhac, J.F.; Gilot, P. (1999). Experimental and mechanistic study of NOx adsorption over NOx trap catalysts. Appl. Catal., B: Environ., 20, 47-55.  Machida, M.; Kurogi, D.; Kijima, T. (2003). NOx storage and reduction characteristics of Pd/MnOx-CeO2 at low temperature. Catal. Today, 84(3-4), 201-207.  Machida, M.; Murakami, H.; Kijima, T. (1998). Temperature-swing sorption/desorption cycles of nitric oxide through intercalation by double layered cuprate. Appl. Catal., B: Environ., 17, 195-203.  Machida, M.; Yoshii, A.; Kijima, T. (2000). Temperature swing adsorption of NOx over ZrO2-based oxides. Int. J. Inorg. Mater., 2, 413-417.  Machida, M.; Masuda, N.; Kijima, T. (1999). Sorption-desorption properties of nitric oxide over La2-xSr1+xCu2O6-δ (0≤x≤2). J. Mater. Chem., 9, 1369-1374.  Martinez-Hernandez, A.; Fuentes, G.A. (2005). Redistribution of cobalt species in Co- ZSM5 during operation under wet conditions in the reduction of NOx by propane. Appl. Catal., B: Environ., 57 (3), 167–174.  Masters, S. G.; Chadwick, D. (1999a). Effect of SO2 on selective catalytic reduction of NO by CH3OCH3 over γ-alumina in excess oxygen. Catal. Lett., 61(1,2), 65-69.  Masters, S. G.; Chadwick, D. (1999b). Selective reduction of nitric oxide by methanol and dimethyl ether over promoted alumina catalysts in excess oxygen. Appl. Catal., B: Environ., 23(4),  235-246.  Matsumoto, H.; Tanabe, S. (1995). Reduction mechanism of nitrogen monoxide with propane over palladium-Y zeolite catalyst. J. Phys. Chem., 99, 6951-6956.  Matsumoto, S. (1996). DeNOx catalyst for automotive lean-burn engine. Catal. Today, 29, 43-45.  Matsumoto, S.; Ikeda, Y.; Suzuki, H.; Ogai, M.; Miyoshi, N. (2000). NOx storage-reduction catalyst for automotive exhaust with improved tolerance against sulfur poisoning. Appl. Catal., B: Environ., 25, 115-124.  Marturano, P.; Kogelbauer A.; Prins, R. (1999). Different methods for preparation of Fe- ZSM5 as catalyst for the selective catalytic reduction of nitrogen oxides. Porous Materials in Environmentally Friendly Processes , Studies in Surf. Sci. Catal., 125, 619-625. References  211  Masuda, K.; Shinoda, K.; Kato T.; Tsujimura, K. (1998). Activity enhancement of Ag/mordenite catalysts by addition of palladium for the removal of nitrogen oxides from diesel engine exhaust gas. Appl. Catal., B: Environ., 15, 29-35.  Meunier, F.C.; Ukropec, R.; Stapleton, C.; Ross, J.R.H. (2001). Effect of the silver loading and some other experimental parameters on the selective reduction of NO with C3H6 over Al2O3 and ZrO2-based catalysts. Appl. Catal., B: Environ., 30(1-2), 163-172.  Meunier, F. C.; Zuzaniuk, V.; Breen, J. P.; Olsson, M.; Ross, J. R. H. (2000). Mechanistic differences in the selective reduction of NO by propene over cobalt- and silver- promoted alumina catalysts: kinetic and in situ DRIFTS study. Catal. Today, 59, 287- 304.  Mihaylov, M.; Hadjiivanov, K.; Panayotov, D. (2004). FTIR Mechanistic studies on the selective catalytic reduction of NOx with methane over Ni-containing zeolites: Comparison between NiY and Ni-ZSM-5. Appl. Catal., B: Environ., 51(1),33-42.  Milne, B.J.; Berruti, F.; Behie, L.A.; de Brujin, T.J.W. (1992). The internally circulating fluidized bed (ICFB): a novel solution to gas bypassing in spouted beds. Can. J. Chem. Eng., 70, 910-915.  Mitome, J.; Karakas, G.; Bryan, K.A.; Ozkan, U.S. (1998). Effect of H2O and SO2 on the activity of Pd/TiO2 catalysts in catalytic reduction of NO with methane in the presence of oxygen. Catal. Today, 42, 3-11.  Mongkolsiri, N.; Praserthdam, P.; Silveston, P. L.; Hudgins, R. R. (2000). Transient study of the effect of residual cations in Cu/ZSM-5 for SCR of NO by hydrocarbon.  Chem. Eng. Sci., 55(12), 2249-2256.  Mosqueda-Jimenez, B.I.; Jentys, A.; Seshan, K.; Lercher, J.A. (2003). Structure-activity relations for Ni-containing zeolites during NO reduction II. Influence of acid sites. J. Catal., 218(2), 348-353.  Muslehiddinoglu J.; Vannice M.A. (2003). Adsorption of NO on promoted Ag/α-Al2O3 catalysts. J. Catal., 217, 442-456.  Muzio, L.J.; Quartucy, G.C. (1997). Implementing NOx control: Research to application. Progr. Energ. Combust. Sci., 23, 233–266.  Naito, S.; Tanimoto, M. (1993). The role of oxygen in the reduction of nitric oxide with propene over rhodium/alumina in an oxidizing atmosphere. Chem. Lett., 11, 1935-1938.  Nelli, C.H.; Rochelle, G.T. (1996). Nitrogen dioxide reaction with alkaline solids. Ind. Eng. Chem. Res., 35, 999-1005.  References  212 Nikolopoulos, A. A.; Stergioula, E. S.; Efthimiadis, E. A.; Vasalos, I. A. (1999). Selective catalytic reduction of NO by propene in excess oxygen on Pt- and Rh-supported alumina catalysts.  Catal. Today, 54(4), 439-450.  Nishizaka, Y.; Misono, M. (1994). Essential role of acidity in the catalytic reduction of nitrogen monoxide by methane in the presence of oxygen over palladium-loaded zeolites. Chem. Lett., 2237-2240.  Nova, I.; Castoldi, L.; Lietti, L.; Tronconi, E.; Forzatti, P. (2002). On the dynamic behavior of “NOx-storage/reduction” Pt-Ba/Al2O3 catalyst. Catal. Today, 75, 431-437.  Obuchi, A.; Ohi, A.; Nakamura, M.; Ogata, A.; Mizuno, K.; Ohuchi, H. (1993). Performance of platinum-group metal catalysts for the selective reduction of nitrogen oxides by hydrocarbons.  Appl. Catal., B: Environ., 2(1), 71-80.  Ogura, M.; Kage, S.; Hayashi, M.; Matsukata, M.; Kikuchi, E. (2000). Remarkable enhancement in durability of Pd/H-ZSM-5 zeolite catalysts for CH4-SCR.  Appl. Catal., B: Environ., 27(4), L213-L216.  Ogura, M.; Kage, S.; Shimojo, T.; Oba, J.; Hayashi, M.; Matsukata, M.; Kikuchi, E. (2002). Co cation effects on activity and stability of isolated Pd(II) cations in zeolite matrices for selective catalytic reduction of nitric oxide with methane. J. Catal., 211, 75-84.  Ohman, L. O.; Ganemi, B.; Bjornbom, E.; Rahkamaa, K.; Keiski, R. L.; Paul, J. (2002). Catalyst preparation through ion-exchange of zeolite Cu-, Ni-, Pd-, Cu/Ni- and Cu/Pd- ZSM-5. Mater. Chem. Phys. 73(2-3), 263-267.  Ohtsuka, H.; Tabata, T.; Okada, O.; Sabatino, L.M.F.; Bellussi, G. (1997). A study on selective reduction of NOx by propane on Co-Beta. Catal. Lett., 44, 265-270.  Olsson L.; Persson, H.; Fridell, E.; Skoglundh, M.; Andersson, B. (2001). A kinetic study of NO oxidation and NOx storage on Pt/Al2O3 and Pt/BaO/Al2O3. J. Phys. Chem. B, 105, 6895-6906.  Pandey, R.N. ; Ratnani, K.; Varma, R. L.; Elkaim, D.; Pandey, R.N (1997). Selective catalytic reduction of nitrogen oxides. US Patent 5,612,010.  Park, P.W.; Ragle, C.S.; Boyer, C.L.; Balmer, M.L.; Engelhard, M.; McCready, D. (2002). In2O3/Al2O3 Catalysts for NOx reduction in lean condition. J. Catal., 20, 97-105.  Park, S.K.; Park, Y.K.; Park, S.E.; Kevan, L. (2000). Comparison of selective catalytic reduction of NO with C3H6 and C3H8 over Cu(II)-ZSM-5 and Co(II)-ZSM-5. Phys. Chem. Chem. Phys., 2(23), 5500-5509.  Parvulescu, V.I.; Grange, P.; Delmon, B. (1998). Catalytic removal of NO. Catal. Today, 46, 233-316.  References  213 Perez-Ramirez, J.; Garcia-Cortes, J. M.; Kapteijn, F.; Illan-Gomez, M. J.; Ribera, A.; Salinas-Martinez de Lecea, C.; Moulijn, J. A. (2000). Dual-bed catalytic system for NOx- N2O removal: a practical application for lean-burn deNOx HC-SCR.  Appl. Catal., B: Environ., 25(2, 3), 191-203.  Perez-Ramirez, J.; Garcia-Cortes, J. M.; Kapteijn, F.; Mul, G.; Moulijn, J. A.; Salinas- Martinez de Lecea, C. (2001). Characterization and performance of Pt-USY in the selective catalytic reduction of NOx with hydrocarbons under lean-burn conditions. Appl. Catal., B: Environ., 29(4), 285-298.  Pieterse, J. A. Z.; van den Brink, R. W.; Booneveld, S.; de Bruijn, F. A. (2002). Durability of ZSM5-supported Co-Pd catalysts in the reduction of NOx with methane.  Appl. Catal., B: Environ., 39(2), 167-179.  Qi, G.; Yang, R.T. (2005). Ultra-active Fe/ZSM-5 catalyst for selective catalytic reduction of nitric oxide with ammonia. Appl. Catal., B: Environ., 60, 13–22.  Ramachandran, B.; Herman, R. G.; Choi, S.; Stenger, H. G.; Lyman, C. E.; Sale, J. W. (2000). Testing zeolite SCR catalysts under protocol conditions for NOx abatement from stationary emission sources.  Catal. Today, 55(3), 281-290.  Roh, S. A.; Jung, S. H.; Jeong, S. M.; Kim, S. D. (2003). Selective catalytic reduction by urea in a fluidized-bed reactor. J. Chem. Tech. Biotech, 78(10), 1104-1109.  Saaid, I.M.; Mohamed, A. R.; Bhatia, S. (2002). Comparative study of Cu-ZSM-5 and Fe- ZSM-5 in the SCR of NOx with i-C4H10. React. Kinet. Catal. Lett. 75(2), 359-365.  Sadykov, V.A.; Lunin, V.V.; Matyshak, V.A.; Paukshtis, E.A.; Rozovskii, A. Y.; Bulgakov, N.N.; Ross J.R.H. (2003). The reaction mechanism of selective catalytic reduction of nitrogen oxides by hydrocarbons in excess oxygen: Intermediates, their Reactivity, and routes of transformation. Kinet. Catal., 44, 379-400.  Sato, K.; Yoshinari, T.; Kintaichi, Y.; Haneda, M.; Hamada, H. (2003). Rh-post-doped Ag/Al2O3 as a highly active catalyst for the selective reduction of NO with decane. Catal. Comm., 4(7), 315-319.  Sato, S.; Hirabayashi, H.; Yahiro, H.; Mizuno, N.; Iwamoto, M. (1992). Iron ion-exchanged zeolite: The most active catalyst at 473K for selective reduction of nitrogen monoxide by ethene in oxidizing atmosphere.  Catal. Lett., 12(1-3), 193-200.  Satokawa, S.; Yamaseki, K.; Uchida, H. (2001). Influence of low concentration of SO2 for selective reduction of NO by C3H8 in lean-exhaust conditions on the activity of Ag/Al2O3 catalyst. Appl. Catal., B: Environ., 34(4), 299-306.  Satsuma A.; Shimizu, K. (2003). In situ FT/IR study of selective catalytic reduction of NO over alumina-based catalysts. Prog. Energy Combust. Sci., 29, 71-84.  References  214 Schmitz, P.J.; Baird, R.J. (2002). NO and NO2 adsorption on barium oxide: Model study of the trapping stage of NOx conversion via lean NOx traps. J. Phys. Chem. B, 106, 4172- 4180.  Schreier, E.; Eckelt, R.; Richter, M.; Fricke, R. (2005). Investigation on sulphation and regeneration of a commercial monolithic NOx-storage/reduction catalyst. Catal. Comm., 6(6), 409-414.  Schwidder, M.; Santhosh Kumar, M.; Bentrup, U.; Perez-Ramirez, J.; Brueckner, A.; Gruenert, W. (2008). The role of Bronsted acidity in the SCR of NO over Fe-MFI catalysts. Micro. Meso. Mater., 111(1-3), 124-133.  Sedlmair, Ch.; Seshan, K.; Jentys, A.; Lercher,J.A. (2003). Elementary steps of NOx and surface reaction on a commercial storage-reduction catalyst. J. Catal., 214, 308-316.  Seyedeyn-Azad, F.; Zhang, D.K. (2001).  Selective catalytic reduction of nitric oxide over Cu and Co ion-exchanged ZSM-5 zeolite: the effect of SiO2/Al2O3 ratio and cation loading. Catal. Today, 68(1-3), 161-171.  Shelef, M. (1995). Selective catalytic reduction of NOx with N-free reductants. Chem. Rev. 95, 209-225.  Shelukar, S.; Ho, J.; Zega, J.; Roland, E.; Yeh, N.; Quiram, D.; Nole, A.; Katdare, A.; Reynolds, S. (2000). Identification and Characterization of Factors Controlling Tables Coating Uniformity in a Wurster Coating Process. Powder Technol., 110, 29-37.  Shen, S.C.; Kawi, S. (2003). Selective catalytic reduction of NO by hydrocarbons in the presence of excess oxygen using Pt/MCM-41 catalysts.  Appl. Catal., B: Environ., 45(1), 63-76.  Shi, C.; Cheng, M.; Qu, Z.; Yang X.; Bao, X. (2002). On the selectively catalytic reduction of NOx with methane over Ag-ZSM-5 catalysts. Appl.Catal., B: Environ., 36, 173-182.  Shi, Y.; Pan, H.; Li, Z.; Zhang, Y.; Li, W. (2008).  Low-temperature decomposition of NOx over Fe-Mn/H-beta catalysts in the presence of oxygen. Catal. Comm., 9(6), 1356- 1359.  Shibata, J.; Shimizu, K.; Satsuma, A.; Hattori, T. (2002). Influence of hydrocarbon structure on selective catalytic reduction of NO by hydrocarbons over Cu-Al2O3.   Appl. Catal., B: Environ., 37(3), 197-204.  Shimizu, K.; Kawabata, H.; Maeshima, H.; Satsuma, A.; Hattori, T. (2001a). Intermediates in the Selective Reduction of NO by Propene over Cu-Al2O3 Catalysts: Transient in-Situ FTIR Study. J. Phys. Chem. B, 104(13),  2885-2893.  Shimizu, K.; Satsuma, A.; Hattori, T. (2001b). Metal oxide for selective reduction of NOx by hydrocarbons: Toward molecular basis for catalyst design. Catal. Surv. from Japan, 4(2), 115-123. References  215  Shimizu, K.; Satsuma, A.; Hattori, T. (2000). Catalytic performance of Ag–Al2O3 catalyst for the selective catalytic reduction of NO by higher hydrocarbons. Appl. Catal., B: Environ., 25, 239-247.  Shin,H.K.; Hirabayashi, H.; Yahiro, H.; Watanabe, M.; Iwamoto, M. (1995). Selective catalytic reduction of NO by ethene in excess oxygen over platinum ion-exchanged MFI zeolites. Catal. Today, 26, 13-21.  Shinjoh, H.; Takahashi, N.; Yokota, K.; Sugiura, M. (1998). Effect of periodic operation over Pt catalysts in simulated oxidizing exhaust gas. Appl. Catal., B: Environ., 15, 189- 201.  Snip, O.C.; Woods, M.; Korbee, R.; Schouten, J.C.; van den Bleek, C.M. (1996). Regenerative removal of SO2 and NOx for a 150 MWe power plant in a interconnected fluidized bed facility. Chem. Eng. Sci., 51(10), 2021-2029.  Sobalik, Z.; Vondrova , A.; Tvaruzkova, Z.; Wichterlova, B. (2002). Analysis of the structural parameters controlling the temperature window of the process of SCR-NOx by low paraffins over metal-exchanged zeolites. Catal. Today, 75, 347-351.  Stakheev, A.Y.; Lee, C.W.; Park, S.J.; Chong, P.J. (1996). Selective catalytic reduction of NO with propane over CoZSM-5 containing alkaline earth cations. Appl. Catal., B: Environ., 9, 65-76.  Stocker, R.K.; Eng, J.H.; Behie, L.A. (1990). Hydrodynamics and thermal modelling of a high temperature spouted bed reactor with a draft tube. Can. J. Chem. Eng., 68, 302– 311.  Subbiah, A.; Cho, B. K.; Blint, R. J.; Gujar, A.; Price, G. L.; Yie, J. E. (2003). NOx reduction over metal-ion exchanged novel zeolite under lean conditions: activity and hydrothermal stability. Appl. Catal., B: Environ., 42(2), 155-178.  Sullivan, J.A.; Keane, O. (2007). A combination of NOx trapping materials and urea- SCR  [selective catalytic reduction] catalysts for use in the removal of NOx from mobile diesel engines. Appl. Catal., B: Environ., 70(1-4), 205-214.  Tabata, T.; Kokitsu, M.; Ohtsuka, H.; Okada, O.; Sabatino, L. M. F.; Bellussi, G. (1996). Study on catalysts of selective reduction of NOx using hydrocarbons for natural gas engines.  Catal. Today, 27(1-2), 91-98.  Tabata, T.; Kokitsu, M.; Okada, O. (1994). Study on patent literature of catalysts for a new NOx removal process. Catal. Today, 22, 147-169.  Tabata, T.; Kokitsu, M.; Okada, O. (1995). Relationship between methane adsorption and selective catalytic reduction of nitrogen oxide by methane on gallium and indium ion- exchanged ZSM-5. Appl. Catal., B: Environ., 6, 225-236.  References  216 Takahashi, N.; Shinjoh, H.; Iijima, T.; Suzuki, T.; Yamazaki, K.; Yokota, K.; Suzuki, H.; Miyoshi, N.; Matsumoto, S.; Tanizawa, T.; Tanaka, T.; Tateishi, S.; Kasahara, K. (1996). The new concept 3-way catalyst for automotive lean-burn engine: NOx storage and reduction catalyst. Catal. Today, 27, 63-69.  Tonetto, G. M.; Damiani, D. E. (2003). Performance of Pd-Mo/γ-Al2O3 catalysts for the selective reduction of NO by methane. J. Mol. Catal. A: Chem., 202(1-2), 289-303.  Toubeli, A.; Efthimiadis, E. A.; Vasalos, I. A. (2000). NO reduction by C3H6 in excess oxygen over fresh and sulfated Pt- and Rh-based catalysts. Catal. Lett., 69(3,4), 157-164.  Traa, Y.; Burger, B.; Weitkamp, J. (1999). Zeolite-based materials for the selective catalytic reduction of NOx with hydrocarbons. Micro. Meso. Mater. 30, 3-41.  Uchida, H.; Yamaseki, K.; Takahashi, I. (1995). Reduction of NOx from exhaust gases of lean-burn gas engines. Envir. Catal., 331-334  Ueda, A.; Oshima, T.; Haruta, M. (1997). Reduction of nitrogen monoxide with propene in the presence of oxygen and moisture over gold supported on metal oxides. Appl. Catal., B: Environ., 12 (2–3), 81–93.  van de Grift, C.J.G.; Woldhuis, A.F.; Maaskant, O.L. (1996). The Shell DENOX system for low temperature NOx removal. Catal. Today, 27, 23-27.  Vernoux, P.; Leinekugel-Le-Cocq, A.-Y.; Gaillard, F. (2003). Effect of the addition of Na to Pt/Al2O3 catalysts for the reduction of NO by C3H8 and C3H6 under lean-burn conditions. J. Catal., 219(1), 247-257.  Walker, A.P. (1995). Mechanistic studies of the selective reduction of NOx over Cu/ZSM- 5 and related systems. Catal. Today, 26, 107-128.  Wang, A.; Liang, D.; Xu, C.; Sun, X.; Zhang, T. (2001a). Catalytic reduction of NO over in situ synthesized Ir/ZSM-5 monoliths. Appl. Catal., B: Environ., 32(3), 205-212.  Wang, C.; Wang, X.; Yu, S.; Xu, Y. (2005). Acetylene: A new reducing agent used in SCR  of NO on zeolite catalysts in an oxygen-rich atmosphere. React. Kinet. Catal. Lett., 86(1), 59-66.  Wang, X.; Chen, H.Y.; Sachtler, W. M. H. (2001b). Mechanism of the selective reduction of NOx over Co/MFI: Comparison with Fe/MFI. J. Catal., 197(2), 281-291.  Wang, X.; Chen, H.Y.; Sachtler W.M.H. (2000). Catalytic reduction of NOx by hydrocarbons over Co/ZSM-5 catalysts prepared with different methods. Appl. Catal., B: Environ., 26, L227-L239.  Wark, K.; Warner, C.F.; Davis, W.T. (1998). Air pollution: Its origin and control. 3rd ed. Addison Wesley Longman, Inc., 430-438.  References  217 Wichtelova, B.; Sobalik, Z.; Dedecek, J. (2003). Redox catalysis over metallo-zeolites: Contribution to environmental catalysis. Appl. Catal., B: Environ., 41, 97-114.  Williams, J.L. (2001). Monolith structures, materials, properties and uses. Catal. Today, 69, 3-9.  Witzel, F.; Sill, G.A.; Hall, W.K. (1994). Reaction studies of the selective reduction of NO by various hydrocarbons. J. Catal. 149, 229-237.  Woo, S.I.; Kim, D.K.; Park, Y.K.; Kim, M.R.; Decyk, P. (2003). Wide temperature window in the catalytic activity of novel Pt/ZSM-5 prepared by a sublimation method for selective catalytic reduction of NO. Catal. Lett., 85(1-2), 69-72.  Xin, M.; Hwang, I. C.; Kim, D. H.; Cho, S. I.; Woo, S. I. (1999). The effect of the preparation conditions of Pt/ZSM-5 upon its activity and selectivity for the reduction of nitric oxide. Appl. Catal., B: Environ., 21(3), 183-190.  Xu, G.; Wang, B.; Suzuki, H.; Kato, K. (1999). Performance with respect to flue gas composition of a combined desulfurization/denitration process using powder-particle fluidized bed. Chinese J. Chem. Eng., 7(4), 295-306.  Xu, G.; Luo, G.; Akamatsu, H.; Kato, K. (2000). An adaptive sorbent for the combined desulfurization/denitration process using a powder-particle fluidized bed. Ind. Eng. Chem. Res., 39, 2190-2198.  Yamada, K.; Pophal, C.; Segawa, K. (1998). Selective catalytic reduction of N2O by C3H6 over Fe-ZSM-5. Micro. Meso. Mater. 21, 549-555.  Yan, J.; Kung, M.C.; Sachtler,W.M.H.; Kung, H.H. (1997). Co/Al2O3 lean NOx  reduction catalyst. J. Catal., 172(1), 178–186.  Yan, J.Y.; Lei, G.D.; Sachtler, W.M.H.; Kung, H.H. (1996). Deactivation of Cu/ZSM-5 catalysts for lean NOx reduction: Characterization of changes of Cu state and zeolite support. J. Catal., 161(1), 43-54.  Yan, J.Y.; Kung, H.H.; Sachtler, W.M.H.; Kung, M.C. (1998). Synergistic effect in lean NOx reduction by CH4 over Co/Al2O3 and H-zeolite catalysts. J. Catal., 175, 294-301.  Yang, W.C. (1998). Engineering and Application of Recirculating and Jetting Fluidized Beds. Fluidization, Solids Handling, and Processing: Industrial Applications, P236-330. Noyes Publications, Westwood, New Jersey.  Yang, W.C.; Keairns, D.L. (1978a). Design and operating parameters for a fluidized bed agglomerating combustor/gasifier. Fluidization, Cambridge: Cambridge University Press, 208-214.  Yang, W.C.; Keairns, D.L. (1978b). Design of recirculating fluidized beds for commercial applications. AIChE Symposium Series, 176(74), 218-228 References  218  Yang, W.C.; Keairns, D.L. (1983). Studies on the solid circulation rate and gas bypassing in spouted fluid-bed with a draft tube. Can. J. Chem. Eng., 61, 349-355  Yentekakis, I.V.; Tellou, V.; Botzolaki, G.; Rapakousios, I.A. (2005). A comparative study of the C3H6-NO-O2, C3H6-O2 and NO-O2 reactions in excess oxygen over Na- modified Pt/Al2O3 catalysts. Appl. Catal., B: Environ., 56 (3), 229–239.  Yogo K.; Kikuchi, E. (1994). Reaction mechanism of selective reduction of nitric oxide by methane on Ga- and In-ZSM-5 catalysts. Study Surf. Sci. Catal., 84, 1547-1554.  Yokoyama, C.; Misono, M. (1994a). Selective reduction of nitrogen monoxide by propene over cerium-doped zeolites. Catal. Today, 22, 59-72.  Yokoyama, C.; Misono, M. (1994b). Catalytic reduction of nitrogen oxides by propene in the presence of oxygen over cerium ion-exchanged zeolite. II. Mechanistic study of roles of oxygen and doped metals. J. Catal., 150, 9-17.  Zelenka, P.; Cartellieri, W.; Herzog, P. (1996). Worldwide diesel emission standards: current experiences and future needs. Appl. Catal. B: Environ., 10(1996), 3-28.  Zhou, X.; Zhang, T.; Xu, Z.; Lin, L. (1996). Selective catalytic reduction of nitrogen monoxide with methane over impregnated In/HZSM-5 in the presence of excess oxygen. Catal. Lett., 40, 35-38  Zuzaniuk, V.; Meunier, F.C.; Ross, J. R.H. (2001). Differences in the reactivity of organo- nitro and nitrito compounds over Al2O3-based catalysts active for the selective reduction of NOx. J. Catal., 202, 340-353.     219  Appendix A Summary of Previous Work   Table A.1 HC-SCR catalysts supported by zeolites Catalyst Active site Support Authors Iwamoto et al. (1990) Zelenka et al. (1996) Saaid et al. (2002) Ohman et al. (2002) Garcia-Cortes et al. (2000) Cu/ZSM-5 Cu2+ Na/ZSM-5 Park et al. (2000) Cu/Beta Cu2+ Beta-Zeolite Corma et al. (1997) Na/ZSM-5 Hall et al. (1998) Na/ZSM-5 Garcia-Cortes et al. (2000) Na/ZSM-5 Chen et al. (2000b) Na/ZSM-5 Joyner et al. (1999) H/ZSM-5 Lee et al. (1997) H/ZSM-5 Joyner et al. (1999) Na/ZSM-5 Lombardo et al. (1998) Fe/ZSM-5 Fe2+ Na/ZSM-5 Gao et al. (2001) Fe/Beta NH4/Beta-Zeolite Fe/FER Ferrierite Fe/MOR Mordenite Fe/Y Fe2+ Y-Zeolite Chen et al. (2000b) Pt-Fe/ZSM-5 Fe2+, Pt2+ H/ZSM-5 Kogel et al. (1998) Wang et al. (2000) Park et al. (2000) Co/ZSM-5 Co2+ H/ZSM-5 Kaucky et al. (2000) Co/Beta Co2+ Beta zeolite Ohtsuka et al (1997) Co/FER Co2+ Ferrierite Li et al. (1994a) Xin et al (1999) Pt/ZSM-5 Pt2+ Na/ZSM-5 Cho et al. (1995) Pt/Y Pt2+ Y-Zeolite Amiridis et al (1997) Pt/Beta Pt2+ Beta-Zeolite Garcia-Cortes et al.(2003) Pd/ZSM-5 Pd2+ H/ZSM-5 Kikuchi et al. (2000) Ogura et al. (2002) Co-Pd/ZSM-5 Co2+, Pd2+ H/ZSM-5 Bustamante et al. (2002)   Appendix A: Summary of Previous Work  220    Catalyst Active site Support Authors Ga/ZSM-5 Ga3+ In/ZSM-5 In3+ H/ZSM-5 Kikuchi et al. (1996) Ce-In/ZSM-5 Ce3+, In3+ H/ZSM-5 Berndt et al. (2003) Ag/ZSM-5 Ag+ H/ZSM-5 Shi et al. (2002) Cu/SUZ-4 Cu2+ K/SUZ-4 Cho et al. (2003) Ni/ZSM-5 Ni2+ Na/ZSM-5 Ni/MCM-22 Ni2+ Na/MCM-22 Ni/MOR Ni2+ Na/Mordenite Mosqueda-Jimenez et al.(2003) Ni/Y Ni2+ Na/Y Mihaylov et al. (2004)       Table A.2 HC-SCR catalysts supported by non-zeolites (Metal oxides) Catalyst Active site Support Authors Cu/Al2O3 Cu2+ γ-Al2O3 Garcia-Cortes et al. (2000) Pt/Al2O3 Pt2+ γ-Al2O3 Burch et al. (1998) Co/Al2O3 Co2+ γ-Al2O3 Garcia-Cortes et al. (2000) γ-Al2O3 Jen (1998) Shimizu et al. (2000) Ag/Al2O3 Ag+ Boehmite Satokawa et al. (2001) Rh-Ag/Al2O3 Rh3+, Ag+ γ-Al2O3 Sato et al (2003) Co/ZrO2 Co2+ ZrO2 Lick et al. (2003a) Pt/SiO2 Pt2+ SiO2 Captain et al. (1998) Co/Al2O3 Co2+ γ-Al2O3 Liotta et al. (2003) Pd/Al2O3 Pd2+ γ-Al2O3 Pd-Mo/Al2O3 Pd2+, Mo3+ γ-Al2O3 Tonetto et al. (2003) Au/Al2O3 Au3+ γ-Al2O3 Ueda et al. (1997)  Appendix A: Summary of Previous Work  221   Table A.3 Specifications of some typical HC-SCR catalysts Catalyst Metal Source Metal Loading (wt%) Loading Method Specifications Authors 2.3 WIE Si/Al=14.0, Co/Al=0.44 5.0 WIE+IMP Si/Al=14.0, Co/Al=1.01 5.0 SSI Si/Al=14.0, Co/Al=0.98 Co/ZSM-5 CoCl2 5.6 SUB Si/Al=14.0, Co/Al=1.13 Wang et al. (2000) FeCl3 - SUB Si/Al=23 Fe/Al=1 Chen et al. (1998b) FeSO4·7H20  IMP Si/Al=14.2, Na/Al=0.49, Fe/Al=0.29 FeC2O4·2H2O  IMP Si/Al=14.2, Na/Al=0.42 Fe/Al=0.52 Chen et al. (1998a) Fe/ZSM-5 FeSO4 2.9 WIE Kameoka et al. (2001) Rh/Al2O3 RhCl3·3H2O 1.86 IMP SBET=182m2/g Efthimiadis et al. (2001) Cu/Beta Cu(CH3COO)2·H2O 9.2(CuO) WIE Si/Al=11.9, Cu/Al=0.97 Corma et al. (1997) Pt/Y Pt(NH3)4(OH)2 1.2 IMP Amiridis et al. (1997) Co- Pd/ZSM-5 Pt(NH3)4Cl2 Co(CH3COO)2 Pd:0.4 Co:3.3 WIE+IMP Si/Al=35.0, Pd/Al=0.03, Co/Al=0.45 Ogura et al. (2002) Ga/ZSM-5 Ga(NO3)3 1.20 WIE In/ZSM-5 In(NO3)3 1.40 WIE Tabata et al. (1995) Ce- In/ZSM-5 In(NO3)3,Ce(OH)3, Ce(NO3)3 Ce:8 In:13 WIE+ SSIE SBET=260m 2/g Berndt et al. (2003) Ag/Al2O3 AgNO3 2 IMP SBET=226m2/g Jen (1998) Ag/Al2O3 AgNO3 2 IMP SBET=189m2/g Satokawa et al. (2001) Rh- Ag/Al2O3 RhCl3·3H20 AgNO3 Rh:0.05 Ag:4 IMP SBET=161m 2/g Sato et al. (2003) Pd/ZSM-5 Pd(NH3)4Cl2 4 WIE  Kikuchi et al. (2000) Cu/SUZ-4 Cu(CH3COO)2 1.9-5.5 WIE  Subbiah et al. (2003) Ni/ZSM-5 Ni(NO3)2.6H2O 2.0 WIE Si/Al=16.4, Ni/Al=0.4, Na/Al=0.3 Mosqueda-Jimenez et al. (2003)     222  Table A.4 HC-SCR with methane as reducing agent  Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) CH4 (ppm) CH4 /NO GHSV (h-1) T (oC) NO Conversion to N2 (%) CH4 Conversion (%) Authors Fe/Beta 950(N2O) 10   500 0.5 60,000 350 100 44 Fe/ZSM-5 950(N2O) 10   500 0.5 60,000 400 100 52 Kameoka et al. (2000) Pd-Co/ ZSM-5 100 10 10  2000 20 100ml/min (0.1g Cat.) 500 60  Ogura et al. (2002) 0  500 75 40 Ga-ZSM-5 1000 10 9  1000 1 15,000 600 5 10 0  400 60 43 In/ZSM-5 1000 10 9  1000 1 15,000 500 10 7 Tabata et al. (1995) 0 0  100 12 0  96 9 1300 5 8 2000 1.5 30,000 500 40 72 Pd- Co/Mordenite 1000 6 8  2700 2.7 30,000 550 60 90 Bustamante et al. (2002) 0  550 100 - Ce-In/ ZSM-5 1000 10 5  2000 2 24,000 600 88 - Berndt et al. (2003) Co/ZSM-5 450 78 41 Ni/ZSM-5 510 78 24 Mn/ZSM-5 550 65 52 Ga/ZSM-5 550 30 8 H/ZSM-5 2000 10   8000 4 75ml/min (0.25g Cat.) 550 60 20 Witzel et al. (1994) Pd/Al2O3    500 86 100 Pd-Mo/Al2O3 735.4    183.9 0.25 40,000feed mol/hmol pd 450 100 100 Tonetto et al. (2003)  Appendix A: Sum m ary of Previous W ork  223    Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) CH4 (ppm) CH4 /NO GHSV (h-1) T (oC) NO Conversion to N2 (%) CH4 Conversion (%) Authors Pt-Co/ZSM-5 1000 2   3000 3 30,000 500 91  Gutierrez et al. (1998) Cu/ZSM-5   350 10 70 Co/ZSM-5 2000 2   2000 1 100ml/min (1.0g Cat.) 450 50 81 Seyedeyn-Azad et al. (2001) Ag/Al2O3 1000 10 2  6000 6 0.9 gs/ml 550 42 Shimizu et al (2000) Fe/MFI 1000+1000 (N2O) 4   1000 1 30,000 450 5 (N2O:100) 90 Kogel et al. (1999) Ni/Y 1300 (NOx) 1.5   1000 0.77 30,000 500 3 - Ni/ZSM-5 1300 (NOx) 1.5   1000 0.77 30,000 500 28 - Mihaylov et al. (2004) 0  98 - Pd/ZSM-5 100(NO2) 10 10  2000 20 100ml/min (0.1g Cat.) 450 22 Kikuchi et al. (2000) 0 98 Pd/TiO2 1780 0.2  178 2.13% 12 62ml/min (37.5mg Cat.) 500 10 Mitome et al. (1998)  Appendix A: Sum m ary of Previous W ork  224  Table A.5 HC-SCR with propane as reducing agent  Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) C3H8 (ppm) C3H8/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) C3H8 Conversion (%) Authors Cu/ZSM-5 1500 3   4500 3 20,000 400 96 - Co/ZSM-5 1500 3   4500 3 20,000 550 75 - Park et al. (2000) Co/Beta 500 10 9  1000 2 15,000 450 87 99.7 Cu/Beta 750 2.4   470 0.63  350 80 - Ohtsuka et al. (1997) γ-Al2O3 500 5   1000 2 200ml/min 550 80 12 (CO), 62(CO2) Sulphated-Al2O3 500 5   1000 2 200ml/min 550 10 Burch et al. (1998) 0.8 0  15,000 500 60 43 2.5 0   500 42 65 0  500 28 42 Co/ZrO2 1500 2.5 8 2000 1.3 30,000 550 15 70 Lick et al. (2003a) 0  550 67 - Ce-In/ZSM-5 1000 10 5  2000 2 24,000 600 70 - Berndt et al. (2003) Co/ZSM-5 2000 10   2700 1.35 75ml/min(0.25g Cat.) 400 62 34 Witzel et al. (1994) 0   692 527 77 3 0.61   527 20 5 0.8   527 34 6 2   527 53 38 Co/Al2O3 1500 8 2000 1.3 8600 527 60 73 Lick et al. (2003b) Cu/Al2O3 1000 6.7   6000 6 18,000 400 13 8 Shibata t al.(2002)  Appendix A: Sum m ary of Previous W ork  225   Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) C3H8 (ppm) C3H8/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) C3H8 Conversion (%) Authors  0 21(N2 Yield), 22 (NO2 Yield) 100 Ag/Al2O3 1000 10  50 1000 1 500ml/min (0.66g Cat.) 480 6(N2 Yield), 5 (NO2 Yield) 58 Angelidis et al. (2002) After 6h in steam Co/Beta 150 10 9 0.3 500 3.3 15,000 400 ~70 Tabata et al. (1996) During 4000h on stream Cu/Beta    350 72 Cu/ZSM-5 750 2.4   470 0.63  350 68  Corma et al. (1997) 10  325 67(Yield) 58 (CO yield), 24 (CO2 yield) Fe/Beta 0 280ml/min (0.2g Cat.) 325 69(Yield) 62(CO yield), 26(CO2 yield) Fe/ZSM-5 10  325 44(Yield) 34 (CO yield), 16(CO2 yield)  2000  3  0 2700  2.7 280ml/min (0.2g Cat.) 325 59(Yield) 37(CO yield), 20(CO2 yield) Chen et al. (2000b) Ag/Al2O3 1000 10 2  6000 6 0.12 gs/ml 450 100  Shimizu et al (2000) 0 527 70 58.2 4 450 0 - 4 500 ~0 - Ag/Al2O3 91 9.1 9.1 4 303 3.33 132,000 550 55 - Satokawa et al (2001) Fe/MFI 1000+1000 (N2O) 4   1000 1 7,500 275 70 100(N2O, >450oC) 87 Koget et al., 1999 0  61 53 Co/ZSM-5 1250 4 20  3250 2.6 150,000 500 30 10 Martinez- Hermandez and Fuentes (2005)  Appendix A: Sum m ary of Previous W ork  226  Table A.6 HC-SCR with propylene as reducing agent  Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) C3H6 (ppm) C3H6/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) C3H6 Conversion (%) Authors Cu/ZSM-5 1500 3   4500 3 20,000 500 98 100 Park et al. (2000) Co/ZSM-5 1500 3   4500 3 20,000 460 31 61(450oC) Park et al. (2000) Fe/ZSM-5 950 (N2O) 10   500 0.5 60,000 370 100 - Kameoka et al. (2001) Rh/Al2O3  80 4  235 500 6.25  300 70 90 Efthimiadis et al. (2001) Pt/Y 1850 1 10 20 300+100 (C3H8) 0.22  300 24 - Amiridis et al. (1997) 0 0 500 62 82 Ag/Al2O3 1000 10 9 18 950+520 (C3H8) 1.47 500ml/ min (0.2g Cat.) 550 28 57 Jen (1998) Cu/ZSM-5 405 53 95 Co/ZSM-5 1000 5   1000 1 30,000 530 37 90 Inui et al. (1997) Pt/Beta 1000 5   1500 1.5 15,000 218 91(N2 Selectivity: 32%) 1 Garcia-Cortes et al. (2003) 0 24(N2 Yield), 42 (NO2 Yield) 100, 59(C3H8) 25 50 (N2 Yield) 100, 80(C3H8) 50 38(N2 Yield) 50, 20(C3H8) 100 500+500 (C3H8) 35(N2 Yield) 23, 0(C3H8) 0 1000 21(N2 Yield), 30 (NO2 Yield) 100 Ag/Al2O3 1000 10 50 1 500ml/ min(0.6 6g Cat.) 480 47(N2 Yield), 3 (NO2 Yield) 58 Angelidis et al. (2002)  During  6h on stream  Appendix A: Sum m ary of Previous W ork  227   Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) C3H6 (ppm) C3H6/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) C3H6 Conversion (%) Authors Cu/ZSM-5   350 85 100 Co/ZSM-5 2000 2   2000 1 100ml/ min(1.0 g Cat.) 450 50 100 Seyedeyn-Azad et al. (2001) Pt/Al2O3  0  300 45 15 Toubeli et al. (2000)  0  250 48 93  Sulfated Pt/Al2O3 1000 5  300 1000 1  250 42 78 0    35 0 0.5    70 11 Pt/SiO2 1000 1 1000 1  267 13 33 Captain et al. (1998) 0 0 62 100 Pt/Al2O3 2000 5 5 0 2000 1 285 55 100 0 0 53  Pt/Al2O3 2000 5 0 200 2000 1 1000ml/ min (4g Cat.) 267 47 Efthimiadis et al. (1998) 1.5 30 1000 1.05 350ml/ min (1.25g Cat.) 450 63 (0h) 62 (after 6.5h) 78 (0h) 67 (after 6.5h) 10 0 1000 1.05 500ml/ min (0.5g Cat.) 450 45 80 Co/Al2O3 950 5 1.7 0 1000 1.05 100ml/ min (0.5g Cat.) 450 76 54 (to CO2) 24 (to CO) Yan et al. (1997)  Appendix A: Sum m ary of Previous W ork  228       Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) C3H6 (ppm) C3H6/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) C3H6 Conversion (%) Authors Ag/Al2O3 1000 5   1000 1 25,000 450 100 100 Meunier et al. (2001) Fe/MFI 1000 + 1000(N 2O) 4   1000 1 30,000 350 30 100 (N2O, >450oC) 95 Kogel et al., 1999 Au/ Al2O3 940 4.7 1.8  940 0.5 100ml/ min (0.3g Cat.) 450 70 - Ueda et al. (1997) Pt/ Al2O3 1000 5   1000 1 70mg Cat. (Contac t time: 4s) 325 75 100 Yentekakis et al. (2005) Cu/MOR 1000 5   1000 1 15,000 375 65.1 - Cu/ZSM-5 1000 5   1000 1 15,000 375 78.9 - Co/MOR 1000 5   1000 1 15,000 375 81.4 - Co/ZSM-5 1000 5   1000 1 15,000 375 66.8 - Ni/MOR 1000 5   1000 1 15,000 450 79.8 - Ni/ZSM-5 1000 5   1000 1 15,000 425 76.6 - Mn/MOR 1000 5   1000 1 15,000 400 73.6 - Mn/ZSM-5 1000 5   1000 1 15,000 425 65.2 - De Lucas et al. (2005)   Appendix A: Sum m ary of Previous W ork  229   Table A.7 HC-SCR with iso-C4H10 (iso-butane) as reducing agent  Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) iso- C4H10 (ppm) iso- C4H10 /NO GHSV (h-1) T (oC) NO Conversion to N2 (%) iso-C4H10 Conversion (%) Authors 0 450 88 (N2 yield) 21(CO yield) 46(CO2 yield) Co/ZSM-5 (WIE) 2000 3 10 0 2000 1 42,000  98(N2 yield) 31(CO yield) 43(CO2 yield) 0 450 80(N2 yield) 3(CO yield) 89(CO2 yield) Co/ZSM-5 (IMP) 2000 3 10 0 2000 1 42,000 450 72(N2 yield) 4(CO yield) 73(CO2 yield) 0 400 99(N2 yield) 4(CO yield) 87(CO2 yield) Co/ZSM-5 (SUB) 2000 3 10 0 2000 1 42,000 400 92(N2 yield) 4(CO yield) 85(CO2 yield) 0 400 56(N2 yield) -(CO yield) 59(CO2 yield) Co/ZSM-5 (SSI) 2000 3 10 0 2000 1 42,000 400 64(N2 yield) 2(CO yield) 51(CO2 yield) Wang et al (2000) Fe/ZSM-5 (WIE) 2000 3 20 0 2000 1 42,000 450 60 70 Hall et al (1998) Fe/ZSM-5 (SUB) 2000 3 10 0 2000 1 42,000 350 76.6 55 (to CO) 28(To CO2) Fe/ZSM-5 (WIE) 2000 3 0 0 2000 1 42,000 350 55 44 Chen et al (1998a)    Appendix A: Sum m ary of Previous W ork  230  Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) iso- C4H10 (ppm) iso- C4H10 /NO GHSV (h-1) T (oC) NO Conversion to N2 (%) iso-C4H10 Conversion (%) Authors Cu/ZSM-5 (WIE) 350 100 Cu/ZSM-5 (IMP) 350 100 Fe/ZSM-5 (WIE) 400 42 Fe/ZSM-5 (IMP) 350 42 Fe/ZSM-5 (SUB) 400 60 Fe/ZSM-5 (SSI) 1000 3 0 0 1500 1.5 100ml/mi n (0.18g Cat.) 400 38 Saaid et al (2002) 0 76 45 (CO yield), 51 (CO2 yield) Fe/MFI 10 350 77 53 (CO yield), 34 (CO2 yield) 0 67 37 (CO yield), 80 (CO2 yield) Fe/BEA 10 375 62 41 (CO yield), 57 (CO2 yield) 0 18 0 (CO yield), 12 (CO2 yield) Fe/FER 10 350 15 3 (CO yield), 4 (CO2 yield) 0 12 0 (CO yield), 14 (CO2 yield) Fe/MOR 10 400 9 0 (CO yield), 5 (CO2 yield) 0 24 6 (CO yield) Fe/Y 2000 3 10 0 2000 1 280ml/mi n (0.2g Cat.) 450 10 4 (CO yield) Chen et al. (2000b)  Appendix A: Sum m ary of Previous W ork  231   Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) iso- C4H10 (ppm) iso- C4H10 /NO GHSV (h-1) T (oC) NO Conversion to N2 (%) iso-C4H10 Conversion (%) Authors Co/ZSM-5 400 82 100 Ni/ZSM-5 400 90 70 Mn/ZSM-5 400 87 87 Ga/ZSM-5 400 87 50 La/ZSM-5 500 62 100 Cu/ZSM-5 2000 10 0 0 2000 1 75ml/min (0.25g Cat.) 350 95 98 Witzel et al (1994) 2000 3  0 2000 1 42,000 350 59.1 40 (to CO) 31.9 (to CO2) 2000 3  150(*) 2000 1 42,000 350 36 21.5 (to CO) 13.3 (to CO2) 2000 3  0 (after *) 2000 1 42,000 350 43.8 27 (to CO) 20.6 (to CO2) 2000 3  300 (#) 2000 1 42,000 350 24.7 15 (to CO) 9.1 (to CO2) Fe/ZSM-5 2000 3  0 (after #) 2000 1 42,000 350 36.5 20.6 (to CO) 17 (to CO2) Decyk et al. (2001)   Appendix A: Sum m ary of Previous W ork  232   Table A.8 HC-SCR with other hydrocarbons as reducing agent Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) HC (ppm) HC/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) HC Conversion (%) Authors Fe/ZSM-5 420 73 3.9%Cu- 3.5%Sn/ZrO2 1000 9   n-C10H22, 300 0.3 70,000 380 43 Delahay et al. (1998) Ag/Al2O3 0 300 28 52 Ag-Rh/Al2O3 0 300 46 60 0 300 54 53 1 350 74  Rh-Ag/Al2O3 20 350 18 0 350 42 1 350 30  Cu/ZSM-5 1000 10 10 20 n-C10H22, 333 0.333 0.05gs/ ml 350 28 Sato et al. (2003)   C2H6, 6000(C) 500 3 16   C6H14, 6000(C) 375 25 58 Cu/Al2O3 1000 6.7   C8H18, 6000(C) 6 18,000 350 32 55 Shibata t al. (2002)  n-butane, 6000(C) 450 100  C2H6, 6000(C) 550 73  n-C6H14, 6000 (C) 325 80 Ag/Al2O3 1000 10 2  n-C8H18, 6000(C) 6 0.12 gs/ml 350 100 Ag/Al2O3 327-377 100 Co/Al2O3 450 70 Ni/Al2O3 450 62 Ga/Al2O3 450 58 Sn/Al2O3 450 68 Al2O3 1000 10 2  n-C8H18, 6000 (C) 6 0.12 gs/ml 450 60 Shimizu et al. (2000)  Appendix A: Sum m ary of Previous W ork  233     Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) HC (ppm) HC/NO GHSV (h-1) T (oC) NO Conversion to N2 (%) HC Conversion (%) Authors Co/ZSM-5 1000 10   n-C8H18, 1000 1 30,000 510 58 Inui et al. (1997) Pt/Al2O3 500 5   n-C8H18, 1000 2  210 60 100 Burch et al. (1998) 1.3%Fe/SUZ-4 0 0 500 5(NOx), 72(NO) 55 5.9%Ag/SUZ-4 0 0 550 8(NOx), 80(NO) 40 0.1%Co/SUZ-4 230 7 0 0 C2H4, 1200 5.2 550 20(NOx), 76(NO) 72 0 0 450 68(NOx), 65(NO) 100 2.5 0 475 42(NOx), 40(NO) 0 15 450 46(NOx), 42(NO) 2.3%Cu/SUZ-4 230 7 2.5 15 C2H4, 1200 5.2 500 43(NOx), 42(NO) 0 0 525 55(NOx), 85(NO) 90 2.5 0 525 44(NOx), 91(NO) 83 Aged 2.3%Cu/SUZ-4 230 7 2.5 15 C2H4, 1200 5.2 550 40(NOx), 68(NO) 88 2.3%Cu/ZSM-5  350 63(NO) Aged 2.3%Cu/ZSM- 5 230 7 0 0 C2H4, 1200 5.2  450 24(NO) Subbiah et al. (2003) Fe/ZSM-5 (CVD) 53 100 Fe/ZSM-5 (IMPO) 53 100 Fe/ZSM-5 (IMPA) 1000 9 0 0 n-C10H22, 300 0.3 35,000 400 40 100 Lima et al (2008)  Appendix A: Sum m ary of Previous W ork  234    Table A.9 HC-SCR with oxygenated hydrocarbons as reducing agent  Flue Gas Composition Catalyst NO (ppm) O2 (%) H2O (%) SO2 (ppm) O.H. (ppm) O.H./ NO GHSV (h-1) T (oC) NO Conversion (%) O.H. Conversion (%) Authors γ-Al2O3 80 4  235 CH3OH, 1500 18.75  400 80 60 Efthimiadis et al. (2001) γ-Al2O3 500 5   CH3OH, 3000 6 12,000 350 100 16(CO2), 40(CO), 25(DME) Sulphated-Al2O3 500 5   CH3OH, 3000 6 12,000 350- 400 90 Burch et al. (1998) γ-Al2O3   485 68 82 5%MoO3- Al2O3   420 72 25 6%V2O5- Al2O3 1700 3  CH3OH, 3400 2 25,500 475 27 48 Masters et al. (1999b)  Ethanol, 500 5.5 350 25 100  Acetaldehyde, 500 5.5 350 30 100  Acetic acid, 500 5.5 350 56  Acetic acid, 1000 11 350 78 Pd/Mordenite 91 9.1 9.1  Acetic acid, 2000 22 44,000 350 94 Uchida et al. (1995) 2.25   52.2 96.1 4.8   28.4 99.3  377 9.36 Acetic acid, 1663 4.4 31,932 320 20.4 96.1 Lee (2000) 0 0  Acetic Acid, 500 1 100 1 0  Acetic Acid, 2250 4.5 81 2 0  Acetic Acid, 2250 4.5 60 10 0  Acetic Acid, 2250 4.5 38 0 5  99.4 0 10  99.7 V2O5-γ-Al2O3 500 0 15 Acetic Acid, 500 1 5000 300 97 Elkaim et al., 2000  Appendix A: Sum m ary of Previous W ork  235  Table A.10 NSR with reducing agent Exhaust Gas Composition Catalyst Sorption Cycle Reduction Cycle GHSV (h-1) T (oC) Sorption Efficiency (%) Reduction Efficiency (%) Overall NO Conversion (%) Authors 175 80.2 90.4 72.6 200 91.0 94.2 85.7 225 95.6 96.1 91.9 250 85.1 96.0 81.7 278 88.9 96.1 85.4 303 91.6 95.3 87.2 330 92.9 92.7 86.1 250ppm NO, 150ppm CO, 8%O2, N2 0ppm NO, 150ppm CO, 0%O2, 1500ppm H2, N2 25,000 360 92.3 88.0 81.2 175 54.5 85.3 46.5 200 69.6 90.4 62.9 225 77.9 93.6 72.9 250 63.0 93.1 58.6 278 66.3 92.4 61.2 303 66.2 90.4 59.9 330 65.3 87.3 57.0 Pt-Ba/Al2O3 250ppm NO, 150ppm CO, 8%O2, 8%H2O, 8%CO2, N2 0ppm NO, 150ppm CO, 0%O2, 8%H2O, 8%CO2,1500ppm H2, N2 25,000 360 61.9 82.1 50.8 Epling et al. (2003) Pt/Al2O3 1200pppm NO, 800ppm C3H6, 4%O2, He 1200ppm NO, 800ppm C3H6, 0.3%O2, He   25-36% Pt-Ba/Al2O3 1200pppm NO, 800ppm C3H6, 4%O2, He 800ppm C3H6, 0.3%O2, He 95,000 300-500   44-55% Shinjoh et al. (1998) Pt-Ba/Al2O3 1000 ppm NO, 3%O2, He 2000ppm H2, He 60s storage period: NO adsorbed: 0.74x10-4, NO desorbed:0.91x10-4mol/g Cat, 350oC 120s storage period: NO adsorbed: 1.5x10-4, NO desorbed:1.88x10-4mol/g Cat, 350oC Lietti et al. (2001) Pt-Rh/Al2O3 No NOx storage could be observed, N2O was formed BaO-Al2O3 No NOx storage could be observed, no reaction took place Pt-Rh/BaO- Al2O3 300-1100ppm NO, 900 ppm C3H6 8% O2, 400oC 300-1100ppm NO, 900ppm C3H6, 400oC NOx storage could be observed, N2O was formed Fridell et al. (1999)  Appendix A: Sum m ary of Previous W ork  236  Appendix B Calibration  B.1 Gas flow meters Calibration curves for gas flow meters used in the fixed bed reaction system are shown in Figures G.1 to G.5.   Figure B.1 Calibration curve of FL-1473G rotameter (0.6% NO + N2)  Appendix B: Calibration  237  Figure B.2 Calibration curve of FL-1473G rotameter (50% O2 + N2)  Figure B.3 Calibration curve of FL-1472S rotameter (1.2% Propylene + N2) Appendix B: Calibration  238  Figure B.4 Calibration curve of FL-1473S rotameter (N2)   Figure B.5 Calibration curve of FL-1476G rotameter (N2)  Appendix B: Calibration  239 B.2 Peristaltic water pump In the fixed bed reaction system, liquid water was pumped into the system by a LKB Microperpex peristaltic pump. The pump was calibrated with the relationship between the pump reading and the water flow rate shown in Figure B.6.   Figure B.6 Calibration curve of peristaltic water pump Appendix B: Calibration  240  B.3 Pressure transducers Pressure transducers used in the cold model and hot model ICFB systems were calibrated using a U-tube manometer. Relationships between the measured voltage and pressure are shown in Figures B.7 to B.11.   Figure B.7 Calibration curve of pressure transducer (Channel #1)  Appendix B: Calibration  241  Figure B.8 Calibration curve of pressure transducer (Channel #2)   Figure B.9 Calibration curve of pressure transducer (Channel #3) Appendix B: Calibration  242  Figure B.10 Calibration curve of pressure transducer (Channel #4)    Figure B.11 Calibration curve of pressure transducer (Channel #6)  243  Appendix C BET Surface Area   Table C.1 BET surface areas for catalysts used in this work Sample Sample weight (g) Measured surface area (m2) SBET (m2/g) Average SBET (m2/g) 0.19 21.98 115.68 0.20 23.96  119.80 Na/ZSM-5(PUC) 0.20 23.45 117.25 118 0.18 20.03 111.28 Fresh Fe/ZSM-5 (PUC)  0.17 18.62 109.53 110 0.19 20.85 109.74 Regenerated Fe/ZSM-5 (PUC) 0.19 20.60 108.42 109 0.21 14.63 69.67 Spent Fe/ZSM-5 (PUC) 0.23 14.01 65.26 67 Na/ZSM-5 (Crushed PUC) 0.11 21.12 192 192 Fresh Fe/ZSM-5 (Crushed PUC) 0.10 18.42 184.2 184 H/ZSM-5 (Albemarle) 0.11 18.77 170.64 171 Fresh Fe/ZSM-5 (Albemarle) 0.11 17.79 161.73 162 Fe/ZSM-5 (Albemarle) aged by 30ppm SO2 (Calcined after placed overnight) 0.11 11.36 103.27 103 Fe/ZSM-5 (Albemarle) aged by 200ppm SO2 (Calcined immediately) 0.11 14.19 129 129    244  Appendix D Particle Size Distribution  The particle size distribution for the four catalysts used in this study was measured by a Mastersizer 2000 particle size analyzer using the wet dispersion method. Three runs were carried out for each catalyst, with the results shown in Figures D.1 to D.4.  Figure D.1 Particle size distribution (Fe/ZSM-5(PUC))  Appendix D: Particle Size Distribution  245  Figure D.2 Particle size distribution (Fe/ZSM-5(crushed PUC))    Figure D.3 Particle size distribution (Fe/ZSM-5(Albemarle))     Appendix D: Particle Size Distribution  246       Figure D.4 Particle size distribution (Fe/FCC(Spent))    247  Appendix E Adsorption and Reaction Kinetics of Fe/ZSM-5 (Crushed PUC) Catalyst  To investigate the effect of the particle size on the adsorption and reaction performances of the Fe/ZSM-5 (PUC) catalyst prepared via IMPO method, the parent Na/ZSM-5 (PUC) particles were crushed and sieved to obtain fine particles for the preparation of the Fe/ZSM- 5(crushed PUC) catalyst by the IMPO method. The adsorption and reaction performances were investigated and compared with the coarse Fe/ZSM-5(PUC) catalyst thereafter.  E.1 Adsorption performance of Fe/ZSM-5(crushed PUC) The typical adsorption curves for Fe/ZSM-5(crushed PUC) catalyst using the model flue gas containing 4% O2 and GHSV=5000 h-1 at 325oC, 350oC and 375oC are shown in Figures E.1 to E.3 respectively.  Appendix E: Adsorption and Reaction Kinetics of Fe/ZSM-5(Crushed PUC) Catalyst  248  Figure E.1 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(crushed PUC), T=325oC)   Figure E.2 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(crushed PUC), T=350oC)   Appendix E: Adsorption and Reaction Kinetics of Fe/ZSM-5(Crushed PUC) Catalyst  249  Figure E.3 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(crushed PUC), T=375oC)  It can be seen from Figure E.1 that the adsorption curve had an irregular shape, which was different from the one for the coarse Fe/ZSM-5(PUC) catalyst (see Figure 3.2) where the outlet NOx concentration reached the equilibrium concentration very quickly after the breakthrough of the catalyst bed. At a given temperature, the breakthrough point was reached more quickly at higher inlet NOx. As the temperature increased, the breakthrough time decreased because more NOx was adsorbed at low temperatures (Figures E.2 and E.3). Although the breakthrough time did not increase much (60-115s (T=325oC), 55-90s (T=350oC) and 38-55s (T=375oC)) when compared with Fe/ZSM-5(PUC),  the equilibrium NOx concentration was not reached until 100-200 seconds later after the breakthrough of the catalyst bed. The equilibrium adsorption capacity (qe) of the catalyst was calculated using equations 3.1 and 3.2. The relationship between qe and equilibrium NOx concentration (C0) Appendix E: Adsorption and Reaction Kinetics of Fe/ZSM-5(Crushed PUC) Catalyst  250 was well fitted by the Freundlich equation (equation 3.3), as shown in Figure E.4. The obtained Freundlich coefficients  and  as a function of the adsorption temperature were also fitted by 2nd order polynomial functions with the results given in Figure E.5. It is clearly demonstrated in Figure E.4 that the adsorption capacity of the fine Fe/ZSM-5(crushed PUC) catalyst was much higher than the coarse Fe/ZSM-5(PUC). For example, at T=350oC and C0=0.4 g/m3, qe for the fine Fe/ZSM-5(crushed PUC) is 0.058 mg/g cat., which is 3.6 times of the coarse Fe/ZSM-5(PUC) (0.016 mg/g cat.). Note that the only remarkable difference in the adsorption experiment of the two catalysts was the particle size, with the average particle size of coarse Fe/ZSM-5(PUC) as 1042m, 4.5 times of the fine Fe/ZSM-5(crushed PUC) (234m). It can thus be concluded that the adsorption performance of Fe/ZSM-5 catalyst prepared from Na/ZSM-5 (PUC) is closely related to the particle size, and a fine catalyst is preferred in order to achieve a high adsorption capacity.  Figure E.4 Fitting of adsorption isotherms of NOx by Freundlich equation (Catalyst: Fe/ZSM-5(crushed PUC)) Appendix E: Adsorption and Reaction Kinetics of Fe/ZSM-5(Crushed PUC) Catalyst  251  Figure E.5 Relationship between /β and adsorption temperature (Catalyst: Fe/ZSM-5 (crushed PUC))  E.2 Reaction performance of Fe/ZSM-5(crushed PUC) The time-on-stream experiment was conducted to evaluate the reaction performance of the selective catalytic reduction of NOx over the fine Fe/ZSM-5(crushed PUC) catalyst at various temperatures, with the results shown in Figure E.6. The feed gas contained 600ppm NO, 1200ppm HC and 1% O2, balanced with N2, and the GHSV=5000 h-1. Appendix E: Adsorption and Reaction Kinetics of Fe/ZSM-5(Crushed PUC) Catalyst  252  Figure E.6 Effect of reaction temperature on catalytic activity (Catalyst: Fe/ZSM-5(crushed PUC))   At T=275oC, a peak NOx conversion of 67.9% was obtained, but the catalytic activity was unstable with time. With the temperature increased to 325oC, the NOx conversion became stable over a longer period of time with a peak NOx conversion of 61.6% when compared with T=275oC. The stable yet even lower catalytic activity (55.6% for T=350oC and 51.9% for T=375oC) was obtained when T350oC. The influence of reaction temperature on the catalytic activity of both the fine Fe/ZSM-5(crushed PUC) and the coarse Fe/ZSM- 5(PUC) is similar with the peak NOx conversion decreased with the increase in temperature and the catalytic activity remained stable with time at T350oC. The effect of  O2 concentration was also investigated using model flue gas containing 1% and 4% O2 at T=350oC and GHSV=5000 h-1, with the result shown in Figure E.7. With the inlet O2 concentration increased from 1% to 4%, the NOx conversion dropped from Appendix E: Adsorption and Reaction Kinetics of Fe/ZSM-5(Crushed PUC) Catalyst  253 55.6% to 42.5%, reflecting the negative impact of higher O2 concentration on the performance of HC-SCR reactions.  Figure E.7 Effect of inlet O2 concentration on catalytic activity (Catalyst: Fe/ZSM-5(crushed PUC))  Although the fine Fe/ZSM-5(crushed PUC) exhibited higher NOx adsorption capacity, for unknown reason, however, its catalytic activity was observed to be lower than the coarse Fe/ZSM-5(PUC) at the same operating conditions. The activity of the catalyst in the HC-SCR process is thus not always directly correlated with the adsorption capability of the catalyst.  254  Appendix F HC-SCR Performance of FCC and Fe/FCC Catalyst  A spent Fluid Catalytic Cracking (FCC) catalyst was tested for its possible use as the HC- SCR catalyst. The spent FCC was first calcined in air at 500oC for 4 hours, and then impregnated with Fe using the IMPO method. The SCR test results for both spent FCC and Fe/FCC are shown in Figure F.1. Unfortunately, both spent FCC and Fe/FCC showed very poor catalytic activity. In the temperature range of 300 to 375oC, only 5-10% of the NOx conversion was achieved.   Figure F.1 HC-SCR performance of spent FCC and Fe/FCC  255  Appendix G Some Results from the Fixed Bed Experiment  G.1 Adsorption curves for Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle) Adsorption curves for the coarse Fe/ZSM-5(PUC) and fine Fe/ZSM-5(Albemarle) catalyst are shown in Figures G.1 to G.5 and G.6 to G.9, respectively, with the model flue gas containing NO + O2 + N2 at various temperatures. Figure G.10 shows adsorption curves for Fe/ZSM-5(Albemarle) catalyst with the model flue gas containing NO + 4% O2 + 10% H2O + N2. The effect of the addition of 10% CO2 into the model flue gas of NO + 4% O2 + N2 is shown in Figure G.11. The combined effect of 10% H2O and 10% CO2 on the adsorption performance of Fe/ZSM-5(Albemarle) is shown in Figure G.12.  Figure G.1 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=280oC) Appendix G: Some Results from the Fixed Bed Experiment  256  Figure G.2 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=310oC)   Figure G.3 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=340oC) Appendix G: Some Results from the Fixed Bed Experiment  257  Figure G.4 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=370oC)  Figure G.5 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(PUC), T=400oC) Appendix G: Some Results from the Fixed Bed Experiment  258  Figure G.6 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=300oC)   Figure G.7 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=325oC) Appendix G: Some Results from the Fixed Bed Experiment  259  Figure G.8 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC)  Figure G.9 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=375oC) Appendix G: Some Results from the Fixed Bed Experiment  260  Figure G.10 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC, 10% H2O added)  Figure G.11 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC, 10% CO2 added) Appendix G: Some Results from the Fixed Bed Experiment  261  Figure G.12 Adsorption curves based on NOx concentrations at the reactor outlet (Catalyst: Fe/ZSM-5(Albemarle), T=350oC, 10% H2O + 10% CO2 added)   G.2 Profiles of NOx and HC conversions and outlet CO concentration from time-on-stream test Profiles of NOx and HC conversions and the outlet CO concentration in the time-on-stream test over Fe/ZSM-5(PUC) and Fe/ZSM-5(Albemarle) catalysts at various temperatures are shown in Figures G.13 to G.16 and G.17 to G.19, respectively.  Appendix G: Some Results from the Fixed Bed Experiment  262  Figure G.13 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=275oC, [O2]=1%)  Appendix G: Some Results from the Fixed Bed Experiment  263  Figure G.14 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=275oC, [O2]=4%)  Appendix G: Some Results from the Fixed Bed Experiment  264  Figure G.15 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=300oC, [O2]=1%) Appendix G: Some Results from the Fixed Bed Experiment  265   Figure G.16 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(PUC), T=325oC, [O2]=1%)  Appendix G: Some Results from the Fixed Bed Experiment  266   Figure G.17 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(Albemarle), T=300oC, [O2]=1%)  Appendix G: Some Results from the Fixed Bed Experiment  267  Figure G.18 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(Albemarle), T=325oC, [O2]=1%)  Appendix G: Some Results from the Fixed Bed Experiment  268   Figure G.19 Profiles of NOx and HC conversions and outlet CO concentration in time-on- stream test (Catalyst: Fe/ZSM-5(Albemarle), T=375oC, [O2]=1%)   Appendix G: Some Results from the Fixed Bed Experiment  269 G.3 XPS analysis for Fe/ZSM-5(PUC) XPS survey scan spectra for the fresh, spent and regenerated Fe/ZSM-5(PUC) catalysts are shown in Figures G.20 to G.22, respectively. Comparisons of XPS narrow scan spectra of Al 2p, Si 2p, O 1s and Fe 2p 3/2 for the fresh, spent and regenerated Fe/ZSM-5(PUC) catalysts are shown in Figures G.23 to G.26, respectively.  Figure G.20 XPS survey scan for fresh Fe/ZSM-5(PUC)  Appendix G: Some Results from the Fixed Bed Experiment  270  Figure G.21 XPS survey scan for spent Fe/ZSM-5(PUC)   Figure G.22 XPS survey scan for regenerated Fe/ZSM-5(PUC)   Appendix G: Some Results from the Fixed Bed Experiment  271   Figure G.23 Comparison of XPS narrow scan of Al 2p    Figure G.24 Comparison of XPS narrow scan of Si 2p  Appendix G: Some Results from the Fixed Bed Experiment  272  Figure G.25 Comparison of XPS narrow scan of O 1s    Figure G.26 Comparison of XPS narrow scan of Fe 2p 3/2   273  Appendix H Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  H.1 Fe/ZSM-5(PUC) Gas bypass ratios from the annulus to draft tube (RAD) and the draft tube to annulus (RDA) using the coarse Fe/ZSM-5(PUC) catalyst in the hot model ICFB reactor at a given annulus gas velocity (UA) are shown in Figures H.1 to H.7.   Figure H.1 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.20 m/s, T=350±10oC)   Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  274  Figure H.2 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.30 m/s, T=350±10oC)  Figure H.3 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.35 m/s, T=350±10oC) Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  275  Figure H.4 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.40 m/s, T=350±10oC)   Figure H.5 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.45 m/s, T=350±10oC) Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  276  Figure H.6 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.50 m/s, T=350±10oC)  Figure H.7 Effect of UD on gas bypass (Fe/ZSM-5(PUC), UA=0.55 m/s, T=350±10oC)  Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  277  H.2 Fe/ZSM-5(Albemarle) The gas bypass ratio (RDA or RAD) as a function of the annulus gas velocity (UA) for the fine Fe/ZSM-5(Albemarle) catalyst in the hot model ICFB reactor is shown in Figures H.8 and H.9, respectively. At a given UA, the influence of the draft tube gas velocity (UD) on gas bypass ratio is shown in Figures H.10 to H.15. At a given UD, the influence of the annulus gas velocity (UA) on gas bypass ratio is shown in Figures H.16 to H.19.  Figure H.8 Effect of gas velocities on RDA (Fe/ZSM-5(Albemarle), T=355±15oC)  Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  278  Figure H.9 Effect of gas velocities on RAD (Fe/ZSM-5(Albemarle), T=355±15oC)  Figure H.10 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.20 m/s, T=355±15oC)  Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  279  Figure H.11 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.25 m/s, T=355±15oC)  Figure H.12 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.30 m/s, T=355±15oC) Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  280  Figure H.13 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.35 m/s, T=355±15oC)  Figure H.14 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.40 m/s, T=355±15oC) Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  281  Figure H.15 Effect of UD on gas bypass (Fe/ZSM-5(Albemarle), UA=0.45 m/s, T=355±15oC)  Figure H.16 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.45 m/s, T=355±15oC) Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  282  Figure H.17 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.60 m/s, T=355±15oC)  Figure H.18 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.75 m/s, T=355±15oC)  Appendix H: Effects of Gas Velocities on Gas Bypass in Hot Model ICFB Reactor  283   Figure H.19 Effect of UA on gas bypass (Fe/ZSM-5(Albemarle), UD=0.90 m/s, T=355±15oC) 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0058550/manifest

Comment

Related Items