UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fluidized bed membrane reactor for steam reforming of higher hydrocarbons Rakib, Mohammad Abdur 2010

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

Item Metadata

Download

Media
24-ubc_2010_fall_rakib_mohammad.pdf [ 4.24MB ]
Metadata
JSON: 24-1.0058773.json
JSON-LD: 24-1.0058773-ld.json
RDF/XML (Pretty): 24-1.0058773-rdf.xml
RDF/JSON: 24-1.0058773-rdf.json
Turtle: 24-1.0058773-turtle.txt
N-Triples: 24-1.0058773-rdf-ntriples.txt
Original Record: 24-1.0058773-source.json
Full Text
24-1.0058773-fulltext.txt
Citation
24-1.0058773.ris

Full Text

  FLUIDIZED BED MEMBRANE REACTOR FOR STEAM REFORMING OF HIGHER HYDROCARBONS  by  Mohammad Abdur Rakib   B.Sc., Aligarh Muslim University, India (1996) M.Sc., King Fahd University of Petroleum & Minerals, Saudi Arabia (2001)      A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY   in   The Faculty of Graduate Studies   (Chemical & Biological Engineering)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   July 2010    © Mohammad Abdur Rakib, 2010   ii ABSTRACT  With growing demand for hydrogen in the industrial and energy sectors, research on novel hydrogen production processes is gaining importance. Fluctuations in price and availability of different hydrocarbons emphasize the need to diversify feedstock options beyond natural gas, the major source for hydrogen. Traditional steam reformers for making hydrogen from hydrocarbons suffer from low catalyst effectiveness factors, poor heat transfer and limited hydrogen yield due to thermodynamic equilibrium constraints. A fluidized bed membrane reactor (FBMR) was designed, fabricated, installed with close attention to safety and operated with methane, propane and heptane as feedstocks at average bed temperatures up to 550°C and pressures up to 800 kPa. When operated without membranes, near-equilibrium conditions were achieved inside the reactor with fluidized catalyst due to the fast reforming reactions. Installation hydrogen permselective Pd77Ag23 membrane panels inside the reactor to extract pure hydrogen shifted the reaction towards complete conversion of the hydrocarbons, including methane, the key intermediate when propane and heptane were the feed hydrocarbons.  Reforming of higher hydrocarbons was found to be limited by the reversibility of the steam reforming of this methane. To assess the performance due to hydrogen in situ withdrawal, experiments were conducted with one and six membrane panels along the reactor. The results demonstrated that the FBMR could produce pure hydrogen from higher hydrocarbon feedstocks at moderate operating temperatures of 475-550°C. A two-phase fluidized bed reactor model was developed, with gas assumed to be in plug flow in both the bubble and dense phases. Diffusional mass transfer, as well as bulk convective flow between the phases, was incorporated to account for concentrations changing due to reactions predominantly in the dense phase, and due to increased molar flow due to reaction. Membranes withdraw hydrogen from both the dense and bubble phases. These studies show that an FBMR can provide compact reactor system with favourable hydrogen yield, and high purity. The model predicted feedstock flexibility capabilities achieved by the experiments, with the higher hydrocarbon feedstock rapidly producing methane and the non-permeate mixture approaching chemical equilibrium.   iii TABLE OF CONTENTS   ABSTRACT .................................................................................................................................. ii TABLE OF CONTENTS ............................................................................................................ iii LIST OF TABLES..................................................................................................................... viii LIST OF FIGURES..................................................................................................................... ix NOMENCLATURE .................................................................................................................. xiv ACKNOWLEDGEMENTS .................................................................................................... xviii DEDICATION ........................................................................................................................... xix CO-AUTHORSHIP STATEMENT.......................................................................................... xx CHAPTER 1.      INTRODUCTION........................................................................................... 1 1.1  Thesis Overview ............................................................................................................. 1 1.1.1   Research objectives............................................................................................... 1 1.1.2   Thesis outline......................................................................................................... 2 1.2   Hydrogen Demand ......................................................................................................... 3 1.2.1   Climate change and the hydrogen economy....................................................... 3 1.2.2   Industrial uses of hydrogen ................................................................................. 4 1.3  Manufacture of Hydrogen............................................................................................. 5 1.3.1   Processes for hydrogen from hydrocarbons ...................................................... 6 1.3.2   Steam reforming for hydrogen production ........................................................ 8 1.3.3   Steam reforming of higher hydrocarbons .......................................................... 9 1.3.4   Industrial hydrogen producing units using steam reforming process ........... 10 1.4  Steam Reforming Catalysts ......................................................................................... 13 1.4.1   Common steam reforming catalysts.................................................................. 13 1.4.2   Coke formation and catalyst deactivation ........................................................ 13 1.4.3   Promotion of steam reforming catalysts........................................................... 14 1.5   FBMR for Steam Reforming of Hydrocarbons .......................................................... 15 1.5.1 Limitations of fixed bed steam reformer .............................................................. 15 1.5.2 Hydrogen removal using permselective membranes........................................... 15 1.5.3 Fluidization of catalysts.......................................................................................... 18 1.5.4 Fluidized bed membrane reformer ....................................................................... 18 1.6   FBMR for Steam Reforming of Higher Hydrocarbons ............................................. 18 1.6.1 Previous studies....................................................................................................... 18 1.6.2 Current research..................................................................................................... 19 1.7   References .................................................................................................................... 31  iv CHAPTER 2.      PILOT SCALE EXPERIMENTAL SETUP FOR HYDROGEN PRODUCTION FROM HIGHER HYDROCARBONS: SAFETY CONSIDERATIONS AND IMPLEMENTATION ...................................................................................................... 44 2.1 Introduction ................................................................................................................. 44 2.2 Pilot Plant Layout ........................................................................................................ 45 2.2.1 Feeding section ........................................................................................................ 45 2.2.2 Fluidized Bed Membrane Reactor (FBMR)......................................................... 47 2.2.3 Hydrogen permeation section ................................................................................ 53 2.2.4 Reformer gas withdrawal section.......................................................................... 53 2.2.5 Gas sampling ........................................................................................................... 53 2.3 Objectives of Experimental Setup ............................................................................... 53 2.4 Toxicological and Safety Information of Materials Encountered............................. 54 2.5 Safety Considerations during FBMR Operation ........................................................ 55 2.5.1 Temperature control .............................................................................................. 55 2.5.2 Pressure control ...................................................................................................... 56 2.5.3 Prevention of backflow........................................................................................... 56 2.5.4 Hazardous gas leakage ........................................................................................... 57 2.5.5 Air ingress into hydrogen permeate section ......................................................... 57 2.5.6 Gas sampling ........................................................................................................... 58 2.5.7 Trips and emergency shutdown ............................................................................ 58 2.5.8 Catalyst handling .................................................................................................... 59 2.5.9 Insulation ................................................................................................................. 59 2.5.10 Electrical safety................................................................................................... 59 2.5.11 Safety apparel ..................................................................................................... 59 2.5.12 Safe working habits ............................................................................................ 60 2.5.13 Process control for FBMR operation ................................................................ 60 2.6   Conclusions .................................................................................................................. 60 2.7   References .................................................................................................................... 81 CHAPTER 3.      STEAM REFORMING OF PROPANE IN A FLUIDIZED BED MEMBRANE REACTOR FOR HYDROGEN PRODUCTION........................................... 84 3.1  Introduction ................................................................................................................. 84 3.1.1  Background ............................................................................................................. 84 3.1.2  Equilibrium compositions in steam reforming of propane................................. 86 3.1.3  Steam reforming of propane: Industrial practice................................................ 87 3.1.4  Fluidized bed membrane reactor (FBMR)........................................................... 87 3.2 Catalysts for Steam Reforming of Propane ................................................................ 88 3.2.1  Catalyst selection .................................................................................................... 88 3.2.2  Micro-reactor testing of RK-212 catalyst particles ............................................. 89 3.3  Experimental Set-up and Procedure ........................................................................... 90 3.3.1 Selective hydrogen removal ................................................................................... 90 3.3.2  Heat supply to the reactor...................................................................................... 91 3.3.3  Experiments ............................................................................................................ 91 3.4 Experimental Results................................................................................................... 93  v 3.5 Discussion .................................................................................................................... 97 3.6 Conclusions .................................................................................................................. 98 3.7 References .................................................................................................................. 118 CHAPTER 4.      STEAM REFORMING OF HEPTANE IN A FLUIDIZED BED MEMBRANE REACTOR ....................................................................................................... 123 4.1  Introduction ............................................................................................................... 123 4.1.1 Background ........................................................................................................... 123 4.1.2 Catalyst issues in steam reforming of higher hydrocarbons ............................ 124 4.1.3 Naphtha steam reforming: Industrial practice .................................................. 125 4.1.4 Thermodynamics of n-heptane steam reforming............................................... 127 4.2 FBMR for Steam Reforming of Heptane ................................................................. 127 4.2.1 FBMR experimental set-up.................................................................................. 127 4.2.2 Experimental plan and performance characterization ..................................... 128 4.3 Results and Discussion .............................................................................................. 129 4.3.1 FBMR experiments............................................................................................... 129 4.3.2 Influence of key operating parameters ............................................................... 130 4.3.3 Hydrogen purity ................................................................................................... 134 4.3.4 Discussion .............................................................................................................. 134 4.4 Conclusions ................................................................................................................ 135 4.5 References .................................................................................................................. 158 CHAPTER 5.      MODELING OF A FLUIDIZED BED MEMBRANE REACTOR FOR HYDROGEN PRODUCTION BY STEAM REFORMING OF HYDROCARBONS ...... 162 5.1 Introduction ............................................................................................................... 162 5.1.1 Hydrogen from higher hydrocarbons feedstock ................................................ 162 5.1.2 Fluidized Bed Membrane Reactors (FBMR) ..................................................... 163 5.2 Description ................................................................................................................. 164 5.2.1 Reactions and rate equations ............................................................................... 164 5.2.2 Model simplifications ........................................................................................... 164 5.2.3 Fluidized bed hydrodynamic model.................................................................... 165 5.3 Model Predictions versus Experimental Results ...................................................... 170 5.3.1 Experimental data for comparison with model predictions ............................. 170 5.3.2 Membrane effectiveness factor ............................................................................ 171 5.3.3 Test results with no membrane panels................................................................ 171 5.3.4 Test results with one membrane panel present.................................................. 172 5.3.5 Test results with six membrane panels ............................................................... 173 5.4 Discussion of Results ................................................................................................. 173 5.4.1 Comparison between model and experimental data ......................................... 173 5.4.2 Membrane permeation effectiveness factor ....................................................... 174 5.4.3 Two-phase fluidization model.............................................................................. 175 5.4.4 FBMR performance.............................................................................................. 176 5.5 Conclusions ................................................................................................................ 176 5.6  References .................................................................................................................. 199  vi CHAPTER 6.      CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK .................................................................................................................................................... 203 6.1 Conclusions ................................................................................................................ 203 6.2 Limitations of FBMR Steam Reforming .................................................................. 204 6.3 Recommendations for Future Work ......................................................................... 205 6.4 Specific Recommendations for Reactor Built in the Current Study........................ 207 6.5 References .................................................................................................................. 208 APPENDIX A.      KINETIC SIMULATION OF A COMPACT REACTOR SYSTEM FOR HYDROGEN PRODUCTION BY STEAM REFORMING OF HIGHER HYDROCARBONS.................................................................................................................. 209 A.1 Introduction ............................................................................................................... 209 A.2 Irreversibility of Steam Reforming of Higher Hydrocarbons .................................. 211 A.3 Kinetic Modeling of a Fluidized Bed Membrane Reactor ....................................... 212 A.3.1 Model assumptions ............................................................................................... 212 A.3.2 Model equations for reactor side......................................................................... 213 A.3.3   Model equations for separation side ................................................................... 213 A.3.4 Interphase mass exchange coefficient ................................................................. 214 A.4 Results and Discussion .............................................................................................. 214 A.5 Conclusions ................................................................................................................ 216 A.A Appendices ................................................................................................................. 217 A.A.1 Kinetic expressions for reactions in reformer................................................ 217 A.A.2 Hydrodynamic equations for the 2-phase model ........................................... 218 A.6 References .................................................................................................................. 229 APPENDIX B.      FBMR OPERATING MANUAL............................................................. 232 B.1 Introduction ............................................................................................................... 232 B.2 Steam Reforming Experiments: Reactor Start-up.................................................... 232 B.3 During Steam Reforming Experiments .................................................................... 236 B.4 Keep-Warm Mode of Operation ................................................................................ 236 B.5 Normal Shutdown ...................................................................................................... 237 B.6 Emergency Shutdown ................................................................................................ 238 APPENDIX C.      FBMR ASSEMBLY DRAWINGS........................................................... 239 APPENDIX D.      CATALYST EVALUATION UNIT........................................................ 247 APPENDIX E.      HYDRODYNAMIC BEHAVIOUR IN A PLEXIGLAS COLUMN AND THE FBMR ............................................................................................................................... 251 E.1   Fluidizability of the Catalyst Particles ...................................................................... 251 E.2   Superficial Gas Velocities in the FBMR................................................................... 252  vii E.3   Future Work with Cold Model .................................................................................. 253 E.4 References .................................................................................................................. 262 APPENDIX F.      MODEL SENSITIVITY ANALYSIS ...................................................... 263 F.1   Model Sensitivity to Reaction Rate Constants .......................................................... 263 F.2   Model Sensitivity to Interphase Mass Transfer ........................................................ 263 F.3  Conclusions ................................................................................................................ 264 F.4 References .................................................................................................................. 271 APPENDIX G.      EXPERIMENTAL DATA TABULATION (FBMR) ............................ 272  viii LIST OF TABLES  Table 1.1: World fertilizer consumption (Calendar year basis)........................................... 20  Table 2.1: Gases and liquids in FBMR steam reforming process ........................................ 61 Table 2.2:  Heaters distribution for the FBMR....................................................................... 61 Table 2.3: Micro-GC column information for product gas analysis.................................... 61 Table 2.4: Controlled parameters for FBMR steam reforming process ............................. 62 Table 2.5: Performance parameters for FBMR steam reforming process.......................... 63 Table 2.6: Flammability and safety information of some species encountered .................. 63 Table 2.7: Nodes in P&ID and HAZOP worksheet ............................................................... 64 Table 2.8: Hazards and Operability Study Worksheet ......................................................... 65 Table 2.9: Cause & Effect matrix for actions by the PLC .................................................... 68  Table 3.1: Density of liquid/ liquefied hydrocarbons at ambient pressure ....................... 100 Table 3.2: Composition of RK-212 (catalyst provided by Haldor Topsoe A/S)................ 100 Table 3.3: Steady-state reactor measurements .................................................................... 101 Table 3.4: Location of sampling ports, thermocouples and pure hydrogen withdrawal, and height intervals of active membrane surface.................................................................. 102 Table 3.5: Experimental runs for steam reforming of propane ......................................... 103  Table 4.1: Steady-state reactor measurements .................................................................... 137 Table 4.2: Location of sampling ports, thermocouples and pure hydrogen withdrawal, and height intervals of active membrane surface.................................................................. 138 Table 4.3: Micro-GC column information for product gas analysis.................................. 139 Table 4.4: Experimental runs for steam reforming of n-heptane ...................................... 140  Table 5.1: Reactor physical details........................................................................................ 178 Table 5.2: Reaction rate equations........................................................................................ 179 Table 5.3: Kinetic parameters ............................................................................................... 180 Table 5.4: Experimental conditions for runs where data are compared with model predictions ......................................................................................................................... 181 Table 5.5: FBMR performance with variations in permeation effectiveness factor ........ 181  Table A.1: Reactor geometry and base simulation parameters .......................................... 219  ix LIST OF FIGURES   Figure 1.1: Hydrogen demand and sources ............................................................................. 21 Figure 1.2: Natural range of H2/CO ratio for natural gas...................................................... 21 Figure 1.3: Global distribution of oil and gas reserves ........................................................... 22 Figure 1.4: Natural gas price fluctuation ................................................................................. 23 Figure 1.5: Crude oil price fluctuation..................................................................................... 23 Figure 1.6: Typical flow-chart configuration of a pre-reformer............................................ 24 Figure 1.7: Pre-reformer temperature profiles for different feeds........................................ 24 Figure 1.8: Reforming section for production of ammonia synthesis gas............................. 25 Figure 1.9: Reforming for the production of methanol synthesis gas ................................... 25 Figure 1.10: Relative rates of carbon formation on nickel catalysts ................................... 26 Figure 1.11: Equilibrium methane conversion in steam reforming of methane as a function of temperature and pressure (steam-to-carbon ratio = 3) ............................... 26 Figure 1.12: Enhancement of methane conversion with in-situ hydrogen removal. Steam- to-carbon ratio = 3, reactor pressure = 1000 kPa ............................................................ 27 Figure 1.13: Hydrogen permeabilities of selected metals ..................................................... 27 Figure 1.14: Pressure-composition isotherms of Pd-H system............................................. 28 Figure 1.15: Equilibrium compositions for heptane steam reforming................................ 29 Figure 1.16: Circulating fluidized bed membrane reformer configuration for steam reforming of heptane ..............................................................................................................  .............................................................................................................................. 30  Figure 2.1: FBMR pressure vessel supported on a mobile stand……………………………71 Figure 2.2 (a): P&ID of pilot plant layout (Supplementary gas feeding section)………… 72 Figure 2.2 (b): P&ID of pilot plant layout (Steam and hydrocarbon feeding section)…... 73 Figure 2.2 (c): P&ID of pilot plant layout (FBMR)………………………………………... 74 Figure 2.2 (d): P&ID of pilot plant layout (Gas sampling and Permeate sections)……….75 Figure 2.3 (a): Strength of SA-240 grade 304H plate material as a function of temperature……………………………………………………………………………….. 76 Figure 2.3 (b): MAWP rating of the FBMR pressure vessel………………………………. 77 Figure 2.4 (a): Instrumentation ports on a lateral flange cover, also showing a membrane panel installed……………………………………………………………………………...78 Figure 2.4 (b): Membrane panel dimensions……………………………………………….. 79 Figure 2.5:  Pressure transducers arrangement……………………………………………... 80  Figure 3.1: Equilibrium compositions (dry gas) in propane steam reforming for steam-to- carbon molar ratio = 5.0: (a) P = 400 kPa; (b) P = 1000 kPa………………………….104 Figure 3.2: Schematic of micro-reactor set-up to study steam reforming of propane……105 Figure 3.3: Propane conversion for steam reforming in micro-reactor, T = 525°C, H2O = 30 g/h……………………………………………………………………………... 106 Figure 3.4: The FBMR pressure vessel, showing dimensions of membrane panel, and arrangement of ports on each lateral flange cover where membrane panels are installed…………………………………………………………………………………... 107 Figure 3.5: Schematic of experimental setup to study steam reforming of propane in an FBMR ……………………………………………………………………………………. 108  x Figure 3.6: Experimental yields and temperature for propane steam reforming without membrane panels at average reactor temperature of 500°C and steam-to-carbon ratio molar ratio of 6.0. Total reactor feed = 0.673 mols/min…………………………109 Figure 3.7: Parity plot of experimental hydrogen yield without membrane panels against local equilibrium values………………………………………………………………… 110 Figure 3.8: Experimental yields and temperature for propane steam reforming at average reactor temperature of 485°C and steam-to-carbon molar ratio 5.0. One membrane panel installed, spanning from 0.95 to 1.16 m above distributor. Total reactor feed = 0.717 mols/min……………………………………………………………………………111 Figure 3.9: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feed = 0.614 mols/min…………... 112 Figure 3.10: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, permeate pressure 25 kPa, and steam-to- carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.410, 0.614, and 0.819 mols/min………………………………………………………………. 113 Figure 3.11: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, and permeate pressure 25 kPa. Six membrane panels installed. Total reactor feed = 0.614 mols/min………….. 114 Figure 3.12: Experimental yields and temperature for propane steam reforming at pressure of 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.635, 0.614, and 0.595 mols/min for 475, 500, and  525°C respectively…………………………………………………... 115 Figure 3.13: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed………………. 116 Figure 3.14: Parity plot of experimental yields against equilibrium values at local temperatures  if there was no hydrogen removal: (a) Hydrogen yield (b) Methane yield………………………………………………………………………………………. 117  Figure 4.1: Key components in a modern steam reforming plant for hydrogen from higher hydrocarbon feedstock………………………………………………………………….. 141 Figure 4.2: Dry gas equilibrium composition for steam-to-carbon molar ratio of 5.0: (a) P = 400 kPa; (b) P = 800 kPa. No membranes present……………………………. 142 Figure 4.3: Dry gas composition for reactor pressure of 400 kPa: (a) Steam-to-carbon molar   ratio = 4.0; (b) Steam-to-carbon molar ratio = 6.0. No membranes present.. 143 Figure 4.4: Drawing of FBMR pressure vessel supported on mobile stand……………… 144 Figure 4.5: (a) Dimensions of membrane panel. (b) Ports arranged on each side-opening cover where membrane panels are installed…………………………………………... 145 Figure 4.6: Schematic of experimental set-up to study steam reforming of n-heptane….. 146 Figure 4.7: Experimental yields and temperature for heptane steam reforming without active membrane panels at reactor pressure of 470 kPa and steam-to-carbon ratio molar ratio of 5.0. Total reactor feed = 0.673 and 0.766 mols/min at 520 and 450°C respectively………………………………………………………………………………. 147 Figure 4.8: Experimental yields and temperature for heptane steam reforming without active membrane panels at average reactor temperature of 520°C and steam-to- carbon ratio molar ratio of 5.0. Total reactor feed = 0.673 mols/min………………...148  xi Figure 4.9: Experimental yields and temperature for heptane steam reforming without active membrane panels at average reactor temperature of 520°C and reactor pressure of 725 kPa. Total reactor feed = 0.673 mols/min……………………………. 149 Figure 4.10: Parity plot of experimental yields without active membrane panels against local equilibrium values: (a) Hydrogen yield (b) Methane yield……………………... 150 Figure 4.11: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 480°C and steam-to-carbon molar ratio 5.0. One membrane panel installed, spanning from 0.95 to 1.16 m above distributor. Total reactor feed = 0.717   mols/min…………………………………………………………. 151 Figure 4.12: Experimental yields and temperature for heptane steam reforming at pressure of 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.635, 0.614, and 0.595 mols/min for 475, 500, and  525°C respectively…………………………………………………... 152 Figure 4.13: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 500°C, permeate pressure 25 kPa, and steam-to- carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.410, 0.614, and 0.819 mols/min    for P = 400, 600, and 800 kPa respectively…………….. 153 Figure 4.14: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 475°C, pressure 600 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feed = 0.635 mols/min…….. 154 Figure 4.15: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, and permeate pressure 25 kPa. Six membrane panels installed. Total reactor feed = 0.614 mols/min………….. 155 Figure 4.16: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed……………… 156 Figure 4.17: Parity plot of experimental yields against equilibrium values at local temperatures  if there was no hydrogen removal: (a) Hydrogen yield (b) Methane yield………………………………………………………………………………………. 157  Figure 5.1: Schematic of reactor geometry…………………………………………………. 182 Figure 5.2: Schematic of the FBMR kinetic model………………………………………… 183 Figure 5.3: A cake of catalyst formed around the ROG filter…………………………….. 184 Figure 5.4: Particulate coating formed on the membranes during FBMR operation: (a) A fresh membrane before installation (b) the membrane surface covered by the coating (c) a view of other side of the same membrane showing a clean shining membrane foil exposed after tapping off a part of the coating………………………………………... 186 Figure 5.5: FBMR performance for experiment Heptane 1.b…………………………….. 187 Figure 5.6: FBMR performance for experiment Propane 1.b…………………………….. 188 Figure 5.7: FBMR performance for experiment Heptane 4.b…………………………….. 189 Figure 5.8: FBMR performance for experiment Propane 2.b…………………………….. 190 Figure 5.9: FBMR performance for experiment Methane 2.c…………………………….. 191 Figure 5.10: FBMR performance for experiment Heptane 5.b………………………….. 192 Figure 5.11: FBMR performance for experiment Propane 3.c…………………………... 193 Figure 5.12: FBMR performance for experiment Methane 1.c………………………….. 194 Figure 5.13: Parity plot for permeate hydrogen yields…………………………………… 195 Figure 5.14: Parity plot for methane yields……………………………………………….. 196 Figure 5.15: Parity plot for carbon oxides yields…………………………………………. 197  xii Figure 5.16: Effect of membrane permeation effectiveness factor (Propane Experiment 3.c)………………………………………………………………………………………... 198  Figure A.1:  Equilibrium compositions in n-heptane steam reforming at varying temperatures and pressures…………………………………………………………….. 220 Figure A.2:  Schematic diagram of the kinetic model……………………………………..221 Figure A.3: Predicted species concentrations in the two phases at 650°C, 10 bars: (a) Dense phase (b) Bubble Phase……………………………………………………….222 Figure A.4:  Predicted methane and hydrogen yields at 650°C, 10 bars: (a) Methane (b) Hydrogen…………………………………………………………………………….. 223 Figure A.5: Predicted heptane conversions at 650°C, 10 bars…………………………... 224 Figure A.6: Predicted effect of S/C ratios on yields at 650°C, 10 bars: (a) Permeate hydrogen   (b) Retentate hydrogen…………………………………………………….. 225 Figure A.7:  Reaction zones in FBMR system for higher hydrocarbons: Pre-reforming, reforming and purification in a single unit……………………………………………. 226 Figure A.8: Dependence of hydrogen yields on membrane thickness at 650°C, 10 bars: (a) Permeate hydrogen (b) Retentate hydrogen…………………………………..…... 227 Figure A.9: Dependence of hydrogen yields on specific membrane area at 650°C, 10 bars: (a) Permeate hydrogen (b) Retentate hydrogen……………………………... 228  Figure C.1: FBMR assembly: Shell weldment.................................................................... 240 Figure C.2: FBMR pressure vessel assembly...................................................................... 241 Figure C.3: Typical rectangular cover for side opening.................................................... 242 Figure C.4: Assembly of rectangular cover and membrane panel ................................... 243 Figure C.5: Assembly of inlet head, showing feed distributor.......................................... 244 Figure C.6: General arrangement of FBMR on reactor stand ......................................... 245 Figure C.7:  Location of band heaters (denoted in brown) mounted on the FBMR........ 246  Figure D.1: Catalyst evaluation unit: Micro-reactor feeding system, Part I…………... 248 Figure D.2: Catalyst evaluation unit: Micro-reactor feeding system, Part II………….. 249 Figure D.3: Catalyst evaluation unit: Micro-reactor and gas analysis…………………. 250  Figure E.1: Plexiglas column for hydrodynamic studies................................................... 254 Figure E.2: Plexiglas column dimensions ........................................................................... 255 Figure E.3: Catalyst bed being lifted by the gas ................................................................ 256 Figure E.4: Fresh catalyst particles..................................................................................... 257 Figure E.5: Used catalyst particles ...................................................................................... 258 Figure E.6: Gas superficial velocities for experiment Propane 1.a .................................. 259 Figure E.7: Gas superficial velocities for experiment Propane 2.d .................................. 259 Figure E.8: Gas superficial velocities for experiment Propane 3.c .................................. 260 Figure E.9: Gas superficial velocities for experiment Heptane 1.a .................................. 260 Figure E.10: Gas superficial velocities for experiment Heptane 4.d .................................. 261 Figure E.11: Gas superficial velocities for experiment Heptane 5.b .................................. 261    xiii Figure F.1: FBMR performance with variations of reaction rate constants for experiment Propane 3.c ........................................................................................................................ 265 Figure F.2:  Methane concentrations with variations of reaction rate constants for experiment Propane 3.c.................................................................................................... 266 Figure F.3:  Hydrogen concentrations with variations of reaction rate constants for experiment Propane 3.c.................................................................................................... 267 Figure F.4:  FBMR performance with variations of interphase mass transfer for experiment Propane 3.c.................................................................................................... 268 Figure F.5: Methane concentrations with variations of interphase mass transfer for experiment Propane 3.c.................................................................................................... 269 Figure F.6:  Hydrogen concentrations with variations of interphase mass transfer for experiment Propane 3.c.................................................................................................... 270   xiv NOMENCLATURE  ba   Specific surface area of gas bubbles (m2/m3) A  Cross-sectional area of bed (m2) PA    Membrane permeation area (m) ' PA   Membrane permeation area per unit length of membrane (m 2/m) Ar Archimedes number (-) ATR Autothermal reforming biC ,  Molar concentration of species i per unit volume of bubble phase (mol/m 3) diC ,  Molar concentration of species i per unit volume of dense phase (mol/m3) CFD Computational Fluid Dynamics db   Bubble diameter (m) dbm   Maximum bubble diameter (m) dp   Mean particle diameter (m) Die  Effective molecular diffusivity of component i (m2/s) Dij   Binary diffusivity of component i in j, with i ≠j (m2/s) 2H E    Activation energy for permeation (J/mol) biF ,  Molar flow rate of species i in bubble phase (mol/s) diF ,  Molar flow rate of species i in dense phase (mol/s) fbiF ,  Molar flow rate of species i in freeboard (mol/s) g  Acceleration due to gravity (m/s2) h Vertical coordinate measured from distributor (m) fbh  Vertical coordinate measured from dense catalyst bed surface (m) ∆H    Heat of reaction (kJ/mol) iqk  Interphase mass transfer component for species i (m/s) NC Number of components Nor Number of orifices in distributor NR Number of reactions O2R Oxidative reforming PH2,M  Hydrogen partial pressure on permeate side PH2,R  Hydrogen partial pressure on reactor side P  FBMR pressure monitored in freeboard (kPa)  xv Pm  Permeate side pressure (kPa) iP  Partial pressure of species i (bar) ' iP  Partial pressure of species i (kPa) bHP ,2   Partial pressure of hydrogen in bubble phase (atm) dHP ,2   Partial pressure of hydrogen in dense phase (atm) pHP ,2   Partial pressure of hydrogen on permeate side (atm) 0MP   Pre-exponential factor for permeation (mole/(m.min.atm0.5)) POX  Partial oxidation dbQ →  Volumetric balancing cross-flow from bubble to dense phase (m3/s) bdQ →  Volumetric balancing cross-flow from dense to bubble phase (m3/s) reqdQ ,  Flow requirement for dense phase to prevent de-fluidization (m3/s) bmiQ ,   Permeation rate (molar) of species i per unit length from reactor side to permeate side for bubble phase (mol/(m.s)) dmiQ ,  Permeation rate (molar) of species i per unit length from reactor side to permeate side for dense phase (mol/(m.s)) QH2  Hydrogen diffusion flux through membrane R  Universal gas constant (J/mol/K) Rj  Rate of jth reaction (mol/kg catalyst/s) SCR  Steam-to-carbon molar ratio SMR  Steam methane reforming T  Temperature (K) Tav Bed average temperature (based on temperatures at the six membrane panels or dummies) (°C) U  Superficial gas velocity (m/s) yi Mole fraction of component i in gas mixture (-) Y Correction coefficient for modified two-phase theory (-)  Greek Letters 2H δ   Thickness of hydrogen selective membranes (m) mfε   Bed voidage at minimum fluidization (-) bε  Volume fraction of catalyst bed occupied by bubble phase (-)  xvi εd  Volume fraction of catalyst bed occupied by dense phase (-) bφ  Volume fraction of catalyst bed occupied by solid particles in bubble phase (-) dφ  Volume fraction of catalyst bed occupied by solid particles in dense phase (-) fbφ  Volume fraction of freeboard occupied by solid particles (-) ijγ    Stoichiometric coefficient of component i in jth reaction gρ   Gas density (kg/m3) pρ   Density of catalyst particle (kg/m3) gµ   Gas viscosity (kg/m/s)  Subscripts b   Bubble phase d  Dense phase i  Species i in  Quantity at reactor inlet j  Reaction j mf  Minimum fluidization condition  Abbreviations AAHH  High limit of toxic and combustible gas detectors BCSA  British Columbia Safety Authority C  Closed ESD  Emergency shut down FAHH  High limit flow alarm FALL  Low limit flow alarm FBMR Fluidized bed membrane reactor FFDAHH High limit steam-to-carbon ratio alarm FFDALL Low limit steam-to-carbon ratio alarm FI  Mass flow meter FICV  Mass flow controller FO  Fully open HAZOP Hazard analysis and operability HT-BA  Band heaters  xvii HT-CA  Cable heaters HT-ST  Strip heaters HTS  High temperature shift HX  Heater HYSYS Commercial process simulation software LNG  Liquefied natural gas LPG  Liquefied petroleum gas LTS  Low temperature shift MTS  Medium temperature shift MAWP  Maximum allowable working pressure MSDS  Material Safety Data Sheet O  Open P&ID  Process and instrumentation diagram PAHH  High limit pressure alarm PALL  Low limit pressure alarm PI  Pressure indicator (dial) PLC  Programmable logic controller PLOT  Porous layer open tube gas chromatography column PPU  Porous polymer-U PSV  Relief valve PT  Pressure transducer (gauge or absolute) PTD  Differential pressure transducer ROG  Reformer off-gas SP  Stop ST  Start TAHH  High limit temperature alarm TALL  Low limit temperature alarm TIC  Temperature indicator and controller TLV  Threshold limit value TT  Thermocouple TWA  Time weighted average V  Valve (either needle valve or on/off valve) VCK  Check valve XV  Solenoid valve   xviii ACKNOWLEDGEMENTS  All Praise and Glory is to the Almighty who has guided me through this achievement, and for the most precious and green earth he has gifted us all.  I would like to thank all my supervisors, who have guided me throughout this challenging project. Working under Dr. Grace has been one of the most pleasant experiences, not only for my PhD research, but also in my all-round professional development. Continuous follow-up by Dr. Grace and Dr. Lim, while allowing a great degree of independence for a challenging experimental project, developed in me a sense of confidence for tackling future projects. Discussions with Dr. Elnashaie inspired me to look at Chemical Reaction Engineering from a unique way of interpretation.  I express gratitude to Dr. Xiaotao Bi, Dr. Tom Troczynski, and Dr. Tony Boyd for serving as supervisory committee members. A course with Dr. Bi also helped me significantly in understanding fluidization, one of the major topics of my research.  I am grateful to the Natural Sciences and Engineering Research Council (NSERC) and the Canada Foundation for Innovation (CFI) for providing funding for this project. I would like also to acknowledge financial support from NSERC and for recognizing my professional achievements in the form of a 2-year doctoral fellowship.  I would like to thank Alan Keelan, Bahman Ghiasi, Zaid Ahmad, Negar Zakipour, Dr. Hengzhi Chen, and Dr. Hongbo Jiang for helping me in different phases of my experimental program. Technical discussions with fellow graduate students Nabeel Aboghander, and Alexandre Vigneault, helped me in my research analyses.  Assistance of the researchers at Membrane Reactor Technologies, Dr. Tony Boyd, Dr. Anwu Li, and Ali Gulamhusein is gratefully acknowledged. They shared their experiences with FBMR, allowed me to use their instruments, and helped me carry out safety audits for my experimental set-up, a necessity for this kind of research.  I am also grateful for timely input from Darren Johnston from Varian Inc. for sample analysis with the micro-GC, and to Jeff Gomach of Haldor Topsoe for providing valuable literature on industrial practice. I would like to thank the department of Chemical and Biological Engineering for the excellent facilities, and to its staff for providing me help and support whenever I needed.  A number of friends Promod Patil, Praveen Linga, Tumuluru Jayashankar, Sujay Sarkar, Saifur Rahaman, Tabrez Siddiqui, Sharif Zaman, Venkata Tayi, Nagu Daraboina and Mohammad Masnadi - made my PhD studies an enjoyable experience.  I must acknowledge the patience and support of my parents, my wife and my son during this period. I hope their wait is now over!!!!  xix  DEDICATION   Dedicated to my Parents, and To my wife, and to little Aziz    “Earth provides enough to satisfy every man's need, but not every man's greed”  - ‘Mahatma’ Mohandas Karamchand Gandhi  xx CO-AUTHORSHIP STATEMENT This is a manuscript-based thesis. Drs. John Grace, C. Jim Lim, and Said Elnashaie are my PhD research supervisors, and are co-authors in the manuscripts. During the fabrication and installation of the FBMR, setting up the infrastructure, and ensuring safety procedures in operation, input was received from Tony Boyd and Ali Gulamhusein from Membrane Reactor Technologies Limited, Mark Epp from Jenmar Concepts Inc., and Alan Keelan, then an undergraduate student at the University of British Columbia, who are co-authors in Chapter 2 of this thesis. Bahman Ghiasi conducted catalyst activity and stability tests in a catalyst evaluation unit as a part of his Master of Engineering research project, and is a co-author for Chapter 4. Dr. Yasemin Bolkan was involved during the first year of my research, and is a co-author of the paper included in Appendix A. My role in the research leading to these manuscripts involved: 1. Planning, conceptualization, design, oversight of fabrication, and commissioning of all research equipments relating to the FBMR. 2. Planning, building, commissioning of the Catalyst Evaluation Unit. 3. Planning, designing, oversight of fabrication of a Plexiglas column for cold hydrodynamic studies. 4. Designing and conducting experiments, data collection, modelling and analysis of the FBMR experiments. 5. Conducting experiments and guiding Mr. Bahman Ghiasi for experiments with the Catalyst Evaluation Unit for testing activity and stability of steam reforming catalysts. 6. Preparing manuscripts, coordinating with co-authors in revising the drafts, corresponding with the publishers for publication of the papers. The use of these manuscripts is under the permission of all the co-authors.  1 CHAPTER 1.      INTRODUCTION*   1.1  Thesis Overview 1.1.1   Research objectives Growing demand for hydrogen as an energy carrier and a widely used-industrial commodity has intensified research on hydrogen production. Since steam reforming was introduced to industry in the 1930’s, it has become the most attractive method for making hydrogen. Over the years, it has seen significant improvements, leading to the development of less costly and more efficient methods of producing synthesis gas and/or hydrogen1,2.  The fluidized bed membrane reactor (FBMR) concept was pioneered at the University of British Columbia (Canada) by a group of researchers3-5, and commercialized by Membrane Reactor Technologies Ltd.6 Since then, the FBMR concept has been studied worldwide for hydrogen production and various other applications. The current research explores feedstock diversification for hydrogen in a fluidized bed membrane reactor, underlining the need for a flexible reformer system to be able to adapt to fluctuating feedstock availability and prices. Specific objectives include: 1. Modeling an FBMR for sizing a proof-of-concept reactor. 2. Reactor fabrication, installation, and commissioning with proper safety procedures. 3. Experimentation with the FBMR for different hydrocarbons - heptane, propane and methane - representing different categories of the most widely used feedstocks for steam reforming. Heptane is a surrogate for naphtha, which is a liquid under ambient conditions, and a feedstock for hydrogen/ syngas production in many parts of the world. Propane is a key component of LPG, which is gaseous under ambient conditions, but can be easily liquefied at relatively low pressures and abundantly available from refinery operations. Methane is the main component in natural gas, the most widely used steam reforming feedstock worldwide. 4. Model verification with requisite improvements and elucidation of the physical phenomena inside the FBMR.  * This chapter presents the background and motivation for the present research. It starts with a statement of the main research objectives, providing an outline of the thesis write-up. Subsequently, a background is offered regarding the demand for hydrogen, methods of hydrogen production, and research trends in the steam reforming process. It ends with a note about the background work done for sizing of the reactor (FBMR) in the form of a modeling work, which has been placed in Appendix A.  2 1.1.2   Thesis outline Chapter 1 details the growing demand of hydrogen in various applications, and the techniques for hydrogen production from hydrocarbons. It then describes the steps involved in traditional steam reforming as practiced by industry. This is followed by outlining the key limitations of traditional steam reformers, and considerations in introducing higher hydrocarbon feedstocks. The final section introduces the FBMR concept for higher hydrocarbons as the research focus for the rest of the thesis. An FBMR was modeled, and designed using a two-phase fluidization model, for steam reforming of heptane. A published paper7 with model details and predicted performance is included as Appendix A. The experiments in this project were conducted at temperatures up to 600°C, at elevated pressures (up to 10 bars), and with hazardous substances (methane, propane, heptane, carbon monoxide, hydrogen, and nickel catalyst powder), requiring close attention to safety issues. Chapter 2 describes the pilot plant layout details, installation, and safety considerations and implementation. Chapter 3 describes steam reforming of heptane in the FBMR. Experiments were conducted without membranes, with one membrane and with six membranes to assess the effect of membrane area on permeation of hydrogen. Reactor performance was evaluated with variation of temperature, reactor pressure, permeate pressure, steam-to-carbon molar ratio, and superficial velocity. Chapter 4 describes experiments and parametric studies along similar lines for steam reforming of propane in the FBMR. Chapter 5 presents an FBMR simulation model, with the dense fluidized bed described by a two-phase model, with removal of hydrogen in-situ from both phases. Both phases were treated as in plug flow, with mass exchange between the two phases due to concentration difference of species and maintenance of minimum fluidized conditions in the dense phase. An explanation was provided to the physical phenomena occurring inside the FBMR and the possible reasons for model prediction deviations from experimental data. Chapter 6 summarizes the conclusions of the thesis with important findings of this research, and proposes recommendations for future research.   3 1.2   Hydrogen Demand 1.2.1   Climate change and the hydrogen economy Recently there have been many reports projecting an alarming increase in average global temperature8. The effects could be catastrophic, ranging from rise in sea levels9, submerging community habitats and already shrinking agricultural lands10, vanishing of glaciers, unpredictable climatic patterns with severe droughts11 or hurricanes, and even diseases12. Global warming has been attributed mainly to the release of greenhouse gases (especially CO2) due to wide-spread dependence on fossil fuels8,13 (which are depleting) as a source of energy. But, at the same time, energy consumption is constantly on the rise due to increase in world population, increasing industrialization, and improved average standards of living. To maintain the balance of demand and supply, new sources of energy need to be investigated and developed, while decreasing greenhouse gas emissions. For many, the solution lies in a gradual transition to a hydrogen economy14,15, where the main carrier of energy, hydrogen, can  be utilized in all parts of this economy. Having the highest gravimetric energy density, and the only product of combustion being water, hydrogen has been projected as an environmentally benign energy carrier. This, however, is not universally accepted, with several severe criticisms about enthusiastic projections of a hydrogen economy16- 18, the main contrary arguments against being: 1. Hydrogen is not a clean fuel as is usually claimed to be, since it does not occur naturally, and needs to be derived mainly from fossil fuels, this process releasing large amounts of CO2. 2. Critics describe hydrogen as the most dangerous of all fuels known to man, both in terms of usage, as well as storage. 3. Transportation of hydrogen either by pipelines or shipping in liquefied form is not energy- efficient, is much more costly than for other fossil fuels, and is subject to leakage. There is, however, little disagreement to the fact that fossil fuels, especially oil and gas, which currently form the backbone of major economies, are fast depleting. So, alternative sources of energy must be explored, the focus being on carbon-free fuels. While feasibility of a full-fledged hydrogen economy is debatable, hydrogen is already being incorporated into the energy matrix, and its application in the energy domain is continuously expanding19,20. Hydrogen production from alternative sources like biomass or bio-oil, or by utilizing solar energy to generate hydrogen from water is being explored. While these techniques are expected to take  4 some time to produce hydrogen on large scales, hydrogen from fossil fuels will continue to be important during the transition to a hydrogen economy. 1.2.2   Industrial uses of hydrogen Hydrogen is one of the most widely used commodities in industry, as a key intermediate for many chemicals and fertilizers. Hydrogen in refineries A refinery can be envisioned as a system of processes separating crude oil into products of varying hydrogen contents, and then manipulating the hydrogen distribution among the products to maximize the yield of mid-range products21. Thus, in terms of net hydrogen usage, some processes can be classified as hydrogen sinks which consume hydrogen, while some others are hydrogen sources. The main consumers of hydrogen in refineries are hydrocracking and hydrotreating, sometimes referred to together as hydroprocessing. The main process producing hydrogen as a by-product is catalytic reforming, which produces aromatic compounds by cyclization and dehydrogenation processes, to increase the octane number of naphtha22. Hydrogen could also be available by recovery from hydrogen-rich off-gases. With aromatics being increasingly unwanted in reformulated gasoline23 due to stricter environmental regulations, and with the increasing hydrogen demand for treating increasingly heavy and sour crude oils, refineries are turning from being net producers to net consumers of hydrogen. Using hydrogen pinch analysis techniques, refinery hydrogen management plans have been established to enable an optimum use of available hydrogen “sources” and “sinks”22,24,25. In many cases, however, as an alternative, it may be more cost-effective in the long run to build new hydrogen plants to meet long-term hydrogen requirements26. Canada has huge resources of non-conventional oil in the form of bitumen in oil sands. Large amounts of hydrogen are required for the upgrading of bitumen and heavy oil, since these are deficient in hydrogen, and the hydrogen demand in this sector is on the rise27,28. Hydrogen in fertilizer industries The world population is growing, increasing the demand for food crops, whereas the available agricultural lands are constrained by growing urbanization, especially in the developing world. Global demand for crop nutrients is steadily on the rise. This is likely to be augmented further by the surging interest in production of bio-fuels, especially U.S. corn, Brazilian sugarcane, and palm oils in Malaysia and Indonesia for bio-diesel production29. Among the major crop nutrients, phosphorus and potassium reserves are usually more or less sufficient in soils, while this is not  5 the case for nitrogen. Table 1 gives a projection of increasing demand of various nutrients, including nitrogen. More than 99% of the world’s nitrogenous fertilizers production is based on ammonia as a raw material30, the basic raw material for which is hydrogen. The global ammonia capacity is projected to increase by about 20% from 181 million metric tons NH3 in 2008 to 218 million metric tones in 2013, with a third of this from revamping activities, and the rest to be provided by about 55 new units projected to become on-stream worldwide31. Hydrogen in methanol manufacture Methanol is one of the most widely used commodities in the petrochemical industries, the main applications being production of formaldehyde, dimethyl ether, acetic acid, MTBE and synthetic gasoline. Methanol could also have a significant share as a fuel for fuel cell vehicles. The most widely practiced method of producing methanol requires hydrogen as a raw material. About 90% of methanol is produced globally from natural gas. As of 1995, the worldwide methanol production capacity was 28 million tons/year32, while in 2004, the capacity was 33 million tons33, and the demand is expected to rise. Other industrial uses of hydrogen Hydrogen is also used in manufacture of aldehydes, in Fischer-Tropsch synthesis for producing liquid hydrocarbons from synthesis gas, as a reducing agent in the metallurgical industry34,35, for hydrogenation of unsaturated edible oils in the food industry, etc.36. Synthesis gas, produced as an intermediate stream by the major hydrogen producing technologies, has a wide range of applications in the synthesis of chemicals, as described by Wender37. Thus there is a huge and growing demand for hydrogen in both the industrial and energy sectors. Figure 1.1 (a)38 shows the global hydrogen consumption pattern, with the main consumers being the fertilizer industry, refineries and methanol production. Currently hydrogen is severely limited as an energy carrier, the only major application being as a rocket fuel.  1.3  Manufacture of Hydrogen Figure 1.1 (b)39 indicates the main sources of hydrogen on a global basis. Feedstocks can range from sources with no carbon content, e.g. water, to sources with high carbon content, e.g. coal. In fact, carbon can also be looked at as a hydrogen carrier40, and usually CO2 is released while hydrogen is recovered from the carbon skeleton.  Electrolysis of water can be a viable and a renewable path of hydrogen production if the energy used for electrolyzing aqueous solutions is derived from renewable sources like  6 hydroelectricity, wind or solar energy. Hydrogen plants working on the electrolysis technique are located where there is cheap and vast hydro-electric power. The largest water electrolysis plants, with capacities around 30,000 Nm3/h, are located in Norway and the Aswan dam in Egypt41.  Coal gasification is one of the oldest techniques for syngas and/or hydrogen production42. Coal-based fertilizer plants have been phased out over the years with cheaper production costs being achieved by newer technologies based on liquid hydrocarbons and natural gas. However, this process is mainly practiced in places where there are huge deposits of coal, but no oil resources, like South Africa43, where coal is the main feedstock for the country’s unique synfuels and petrochemicals industry44. Based on the relative abundance of coal compared to other fossil fuels, whose prices fluctuate and increase in unpredictable manners, while coal remains relatively inexpensive45, coal-based hydrogen and synthesis gas can be economically competitive in future46. As depicted in Figure 1.1(b), currently the main contenders among the feedstocks for hydrogen manufacture are natural gas and oil (the bulk contribution being in the form of naphtha, and to a smaller extent as fuel oil), all accounting for more than three quarters of the hydrogen produced. 1.3.1   Processes for hydrogen from hydrocarbons This section outlines the methods for hydrogen production using hydrocarbon feedstocks. All these techniques except hydrocarbon decomposition produce syngas or synthesis gas, which is a mixture of hydrogen, carbon monoxide and carbon dioxide in various proportions. Steam Reforming Steam reforming of natural gas is the most widely practiced means of hydrogen production23,47. The major reactions are: CH4 + H2O  '  CO + 3 H2    ∆H°298 =  206 kJ/mol    (1.1) CO + H2O  '  CO2 + H2     ∆H°298 = - 41 kJ/mol   (1.2) CH4 + 2H2O  '  CO2 + 4H2    ∆H°298 = 165 kJ/mol   (1.3) For higher hydrocarbons47-49, 2 22n m mC H nH O nCO n H⎛ ⎞+ → + +⎜ ⎟⎝ ⎠    ∆H 0 298 = 1108 kJ/mol  (for n = 7) (1.4) Once H2 and CO are available by steam reforming of higher hydrocarbons, a reverse steam reforming reaction (reverse of Equation 1.1) produce CH4, and thereafter the process proceeds as  7 simple steam reforming of methane. The steam reforming technique is discussed in more detail below. CO2 (Dry) Reforming CH4 + CO2  '  2CO + 2H2    ∆H°298 = 247 kJ/mol   (1.5) CnHm + nCO2  '  2nCO + 2 m H2        (1.6) CO2 or dry reforming has the advantage of utilizing greenhouse gases as a feed, and is therefore environmentally attractive. CO2 reforming of methane produces a syngas with a H2/CO ratio lower than that from steam reforming, and it is more suitable for specific processes like the Fischer Tropsch synthesis47,50-52. It has also been studied for other hydrocarbons like propane53,54 and heptane55,56. However, a major disadvantage is carbon deposition51,52, implying the need for coke-resistant catalysts, or process variants like combination with steam reforming or a reactor- regenerator combination. The state-of-the-art for dry reforming of hydrocarbons has been reviewed by Wang et al57. Hydrocarbon Decomposition CH4  '  C + 2H2     ∆H°298 = 75 kJ/mol   (1.7) CnHm →  nC + 2 m H2          (1.8) The biggest attraction of this process is that no greenhouse gases are produced by hydrocarbon decomposition. In addition, pure H2 can be produced directly without separation of the H2 from other components in the product gas stream58,59. Special operating conditions for decomposition have also enabled production of nanocarbons which might be attractive as catalyst supports, or in the semi-conductor industry60. However, since this process is afflicted by catalyst deactivation, a coke-resistant catalyst is required, in addition to continuous regeneration of the catalysts. In a review paper, Muradov and Veziroglu61 proposed that the hydrocarbon decomposition process could be important during transition to a hydrogen economy. Partial Oxidation Partial oxidation is exothermic and therefore does not need external heat transfer (such as firing in a furnace). The reaction can be written as: CH4 + 2 1 O2 '  CO + 2H2     ∆H°298 =  - 36 kJ/mol   (1.9) CnHm + 2 n O2 '  nCO + 2 m H2        (1.10)  8 Partial oxidation can be conducted with or without catalysts. Catalytic partial oxidation consumes less oxygen than the non-catalytic process. However, the feedstock choice is limited from natural gas to naphtha47,62. Non-catalytic partial oxidation processes are characterized by their ability to operate with feedstocks ranging from natural gas to heavy fuel oil, regardless of their sulfur content41. Hence the process can operate with various feedstocks42. Also referred to as thermal partial oxidation or gasification, this process is sometimes carried out with steam added to moderate operating temperatures and suppress carbon formation63. Severe operating conditions, like pressures as high as 70 atm and temperatures of 1200 to 1600°C, are used64. Autothermal Reforming This process consists of combining steam reforming and partial or total oxidation of the hydrocarbon. Part of the hydrocarbon undergoes combustion, thus providing energy for the highly endothermic steam reforming reactions65-70. Optimal control of the operating parameters can make the overall reaction thermally neutral. In autothermal reforming, air is usually the oxygen source, and nitrogen must be separated from the syngas product mixture, or oxygen from the air (usually cryogenically) before being fed to the syngas reactor. This requirement is critical to avoid nitrogen build-up in the process loop. Usually upstream nitrogen separation from air is more favorable than costly downstream purification, and an oxygen separation plant would be necessary71,72. Since the cryogenic oxygen plant is an expensive section in a reforming process layout, autothermal reforming is economically attractive only for large-scale production73. Studies are being conducted for production of syngas using oxygen-selective ceramic membranes which can introduce oxygen in a distributed fashion along the reactor length, thus avoiding the need for separation of the nitrogen74-77. To be industrially viable, lower cost, high selectivity to oxygen permeation, and a high permeation flux need to be achieved. The appropriate choice of technology depends on several factors like the nature of downstream applications and product distribution, but in general is dominated by economic considerations. When the goal of the whole exercise is production of hydrogen, steam reforming of hydrocarbons is clearly the preferable choice. 1.3.2   Steam reforming for hydrogen production Steam reforming of hydrocarbons is the most widely used process for hydrogen production 23,47,78. Its greatest advantage is that hydrogen is extracted not only from the hydrocarbon, but  9 from steam as well, thereby giving maximum H2 produced per mole of a certain hydrocarbon. Excess steam in the reaction mixture suppresses the coking reactions, the extent of which depends on the temperature and the type of hydrocarbon. The H2/CO ranges79 shown in  Figure 1.2 for different processes using natural gas as feedstock indicate why steam reforming is preferred for producing hydrogen. Compared to liquid hydrocarbon feedstocks like naphtha or diesel, natural gas has several operational advantages. The tendency of catalyst deactivation due to carbon formation increases with the average carbon number of the feedstock. Thus, for similar operating conditions of temperature and pressure, a lower steam-to-carbon ratio compared to those required by liquid hydrocarbons can be applied. Natural gas feedstocks tend to be better also in terms of energy efficiency and lower reformer volume requirement. In addition, natural gas needs less desulfurization prior to feeding due to its generally lower sulfur-content. Currently synthesis gas as well as pure hydrogen is produced from natural gas as the major feedstock. Methane is the major component of natural gas and has the highest hydrogen density per mole of carbon, among all hydrocarbons. Natural gas is widely available worldwide. The overall economics, starting from a generally favorable feedstock pricing to a cheaper cost of hydrogen production, makes it the major feedstock for steam reforming. 1.3.3   Steam reforming of higher hydrocarbons Countries have varying degrees of availability of natural gas and oil, as depicted in Figure 1.3. In places where natural gas in not available, it may be imported via cross-country pipelines, or the natural gas is compressed and transported in liquefied form. LNG terminals and gasification facilities need to be installed in many instances to gain access to this preferred feedstock. Depending on proximity to sources and the dynamics of feedstock prices, oil-based feedstocks like naphtha may become competitive in some areas80. Even when natural gas is liquefied by compression, its volumetric hydrogen density remains lower than for liquid hydrocarbons, although the H/C ratio of methane is high81. Therefore, an easily deliverable and safely storable hydrogen source, such as gasoline and diesel, is preferred for mobile applications82,83. On-board hydrogen generation systems prefer liquid hydrocarbon feedstocks, such as gasoline, kerosene and diesel oil, which have higher energy density and a wider distribution network, compared to methanol84. Methanol, proposed by some as a feedstock for hydrogen, may not be used widely due to its toxicity and miscibility in water, and due to overall energy efficiency, since hydrogen itself is a major feedstock for methanol  10 production. In addition to on-board hydrogen generation, for distributed hydrogen generation systems as in hydrogen re-fuelling stations, liquid hydrocarbons have a potential advantage over natural gas, with the ability to utilize existing gasoline/ diesel distribution systems. Due to fluctuations in supply and market demand, different refinery products may be either scarce or in surplus. Many refineries benefit from flexibility in feedstocks, taking advantage of the surplus of various hydrocarbons in the refinery. With proper desulfurization, it has been possible to convert light gas oils and diesel fuel into syngas with no higher hydrocarbons in the product gas23,85. Feedstock flexibility is therefore an important consideration in refinery hydrogen management as the hydrogen demand soars86,87. Refinery off-gases, which are high in hydrogen content, constitute a possible  substitute for the primary feedstock to the hydrogen plant. Traditionally, these were flared or used as a fuel for firing reactors88. Fertilizer plants for ammonia production have often been designed to accept variable feeds, e.g. 100% naphtha, 100% natural gas or intermediate mixtures89,90. Higher hydrocarbons are generally more reactive than methane, with aromatics showing the lowest reactivity, approaching that of methane.  With huge deposits of unconventional oil reserves in the form of oilsands being developed, hydrogen demand for processing them is on the rise in Canada27,91. At the same time, liquid hydrocarbon feedstocks are likely to be available for meeting hydrogen requirement inside refineries or for local hydrogen producing facilities. Recently, there have been great fluctuations in the price of natural gas92 as well as crude oil93, as seen from Figures 1.4 and 1.5. As a result, hydrogen producers favor flexibility of feedstock choices for their reforming units. Hence there is a need to do further research on hydrogen and/or syngas production from a wider range of hydrocarbons like LPG, naphtha, diesel and kerosene. 1.3.4   Industrial hydrogen producing units using steam reforming process This project investigates a compact reactor configuration for steam reforming of higher hydrocarbons to produce pure hydrogen. In order to understand the merits of the proposed reactor configuration, it is necessary to understand the layout of a traditional hydrogen producing unit. Basic information is outlined in this section, while more detailed information is available elsewhere2,94-97. Desulfurizer: The hydrocarbon feedstocks first need to be desulfurized, since sulfur poisoning can rapidly and irreversibly deactivate catalysts96,98. In the desulfurizer, the sulfur compounds  11 are first converted to H2S by hydrodesulfurization using a Co-Mo catalyst for low sulfur content98 or a Ni-Mo catalyst for a high sulfur feedstock like naphtha95, and then reactively scrubbed with an adsorbent like ZnO. R-SH + H2 → RH + H2S         (1.11) H2S + ZnO → ZnS + H2O         (1.12) Steam Reformer: The sweetened feedstock is then fed to a steam reformer, which consists of hundreds of catalyst-filled vertical tubes, operating in a fixed bed mode, housed inside a furnace. The furnace can be top-fired, side-fired, or of terrace-walled design, depending on the scale of the plant. The side-fired design offers the maximum flexibility, and allows more severe operating conditions65, since relatively high average heat flux can be maintained through the tubes without exceeding critical limits 72. External heating provides the heat required by the highly endothermic process. In another version, also practiced widely in industry, the reformer is operated autothermally by partially combusting part of the hydrocarbon feed simultaneously72,95. The operating temperature and steam-to-carbon molar ratio in the steam reformer depends on the nature of the hydrocarbon feed. The catalysts employ nickel supported on ceramic supports. Major reactions occurring in the steam reformer: Higher hydrocarbons steam reforming CnHm + nH2O → nCO + (n + 2 m )H2   ∆H0298 = 1175 kJ/mol (for n = 7) (1.4) Methane steam reforming (reverse reaction is methanation) CH4 + H2O  '  CO + 3 H2    ∆H°298 = 206 kJ/mol    (1.1) Water gas shift CO + H2O  '  CO2 + H2     ∆H°298 = - 41 kJ/mol   (1.2) Methane overall steam reforming CH4 + 2H2O  '  CO2 + 4H2    ∆H°298 = 165 kJ/mol   (1.3) Reactions in the steam reformer are net endothermic, and thus, the hydrocarbon conversion and hydrogen yield are favored at higher temperatures. However, as discussed below, high operating temperatures can lead to catalyst deactivation by carbon formation as well as by sintering. Carbon formation is usually worse for heavier feedstocks94. Thus, old-generation steam reformers were designed for specific ranges of feedstock since naphtha steam reformers are operated at lower temperatures and higher steam-to-carbon ratios, compared to natural gas feedstocks.  12 Pre-reformer: To add flexibility in feedstock for steam reformers, most new reformer systems are equipped with pre-reformer units. A typical pre-reformer installation99 is shown in Figure 1.6. A pre-reformer operates almost adiabatically, due to a combination of endothermic steam reforming of higher hydrocarbons (equation 1.1), followed by an exothermic methanation (reverse of equation 1.2). Typical temperature profiles for pre-reformers with different feedstocks are shown in Figure 1.787. The effluent of the pre-reformer is fed to the reformer as a methane-rich gas. With higher hydrocarbons completely converted in the pre-reformer23,85, the reformer can be operated under the same conditions as for natural gas reformers with reduced risk of catalyst deactivation. Secondary reformer: The steam reformer is referred to as a primary steam reformer when it is followed by a secondary reformer100. Controlled amounts of air or oxygen are fed when the purpose of the steam reforming process is to generate hydrogen for production of ammonia  or methanol, as shown in Figures 1.8 and 1.9 65. Water gas shift section: The reforming section is followed by a shift reaction section. Traditionally, a two-step shift process is employed (see reaction (1.2)). With no net change in molar flow due to this reaction, the equilibrium conversion of the shift reaction is independent of pressure, but, being exothermic, is favored at low temperatures. However, to avoid the low kinetic rates at low temperatures, the shift reaction is carried out first in a high-temperature shift (HTS) reactor, followed by a low-temperature shift (LTS) reactor to increase hydrogen yield. The HTS reactor is loaded with an iron-chromium catalyst and operates at ~320-350°C95. The exit gas is cooled down, and fed to the LTS reactor operating at ~200-250°C with a copper-zinc- alumina catalyst. The shift reactor section improves the hydrogen yield and reduces the CO concentration. However recent developments have seen a single shift reactor, called a medium temperature shift (MTS) reactor or a HTS reactor with efficient removal of impurities (mainly CO) using efficient Pressure Swing Adsorption (PSA). Purification section: Downstream of the shift section is the gas purification section, which conventionally included CO2 removal, followed by a methanation unit to remove residual CO and CO2. In modern hydrogen plants, the CO2 removal and methanation units are replaced by a PSA system to produce high-purity hydrogen (>99.999% purity)101.     13 1.4  Steam Reforming Catalysts 1.4.1   Common steam reforming catalysts Attempts to use non-metallic catalysts for steam reforming of hydrocarbons have not had commercial success because of low activity47,102. Metals active for steam reforming include Ni, Co, Pt, Ir, Ru and Rh47,94,103. The specific activities of metals supported on alumina or magnesia have been found to be Rh, Ru > Ni, Pd, Pt > Re > Co103. A common problem with steam reforming catalysts is deactivation, due to various mechanisms such as catalyst poisoning (e.g. due to sulfur in the process stream), coking, or sintering. Ni catalysts present major coking problems because of the formation, diffusion, and dissolution of carbon in the metal. Neither Ru nor Rh dissolves carbon to the same extent and, as a result, carbon formation is less in these systems104. Pd is the only noble metal that forms carbon, probably because of carbide formation. In addition, in steam reforming, Pd is considerably more active per unit mass than nickel. However, Ni is much cheaper and sufficiently active to be widely used in steam reforming catalysts on an industrial scale 105. These Ni-based commercial catalysts are supported on refractory materials like Al2O3 and SiO296,106,107. 1.4.2   Coke formation and catalyst deactivation In industrial fixed bed steam reformers, operational ills can sometimes lead to excessive rates of coking. The consequences can be catastrophic. Based on the morphology of the carbon deposits, carbon formation on Ni-based catalysts are of three types: encapsulating carbon, pyrolytic carbon, and whisker carbon94,108,109. Only the former two types lead to deactivation of the catalyst. The type of carbon formed depends on several factors such as: (a) Operating temperature, (b) Steam-to-carbon ratio, and (c) Feedstock. Pyrolytic carbon: Pyrolytic carbon is formed generally by exposure of hydrocarbons to high temperatures108,109, and occurs due to thermal cracking of the hydrocarbon, and subsequent deposition of C precursors on the catalyst. This might lead to encapsulation of the catalyst particle itself, leading to deactivation and an increase in bed pressure drop. Encapsulating carbon: Metal catalysts may also form carbonaceous gum-type encapsulating carbon deposits on the metal particle resulting in deactivation. These may form during reforming of heavy hydrocarbons with high proportions of aromatic compounds. The deposits consist of thin CHx films, or layers of graphite covering the nickel particles, resulting in loss of activity108,109. Conditions favouring formation of this type of carbon include: low operating  14 temperature (<500°C), low steam-to-carbon ratio, low hydrogen-to-hydrocarbon ratio, and high aromatic content in the feed109. Whisker Carbon: Nickel catalysts may form whisker carbon in a process where the hydrocarbon or carbon monoxide dissociates into carbon atoms on one side of the metal crystal and a carbon fiber (whisker) nucleates from the opposite side110-113. Thus, the carbon deposits consist of numerous carbon filaments, many with a metal particle at the top. Hence the carbon whisker can grow without deactivation of the active site and progresses without blocking processes responsible for their growth114. Heavy coking occurs, for example, if the reactor is operated at temperatures higher than designed for, or at steam-carbon ratios lower than a critical value109. These conditions can lead to large quantities of pyrolytic carbon and carbon filaments. Filamentous carbon plugs catalyst pores and bed voids, pulverizes catalyst pellets, and can bring about process shutdown, all within a few hours, by adversely increasing the pressure drop108,109. Hence whisker carbon has been described as the most destructive form of carbon in steam reforming over nickel catalysts. When carbon forms at the inner perimeter of reforming tubes, this can seriously impair the process since external heat transfer is very important for the endothermic steam reforming reactions. Other than these process difficulties, coking itself might lead to serious loss of active sites of the catalyst, and the catalyst must ultimately be replaced. Figure 1.10 shows the rate of carbon formation on nickel catalysts from different hydrocarbons based on TGA measurements94. The aromatic content is considered to be critical. It has also been found that for a given hydrocarbon, at a given steam-to-carbon ratio, there is a temperature “window”, below which gum-type encapsulating carbon formation occurs, while high temperature and whisker carbon formation can occur at the other end85. Olefins content is also critical for steam reforming catalysts, with the rate of carbon formation being several orders of magnitudes higher than for other hydrocarbons. Though not usually present in feedstocks, olefins may form in overheated sections of reformer tubes62. 1.4.3   Promotion of steam reforming catalysts Depending on the hydrocarbon feedstock, steam reformers are operated at different temperatures and steam-to-carbon ratios to minimize catalyst deactivation by carbon formation. Research is also being conducted on catalyst formulation, doping with different promoters to minimize carbon formation by continuously gasifying carbon deposits. Since the acidic nature of the Al2O3 support can contribute to carbon formation by promoting hydrocarbon cracking, one approach is  15 to decrease the acidity, e.g. by adding K2O96,115, MgO73,116,117 or CaO116. The mechanism of inhibition of carbon formation has been interpreted differently118. The specific activity of the catalyst for steam reforming is reduced as a result of adding promoters94,119. There may also be a slow loss of K2O over time due to its high volatility96. In many industrial steam reforming catalysts, the Al2O3 support is replaced by MgAl2O4, thereby preventing the formation of NiAl2O4, which can render the catalyst inactive120. One also avoids MgO which can form Mg(OH)2, and being bigger in volume than MgO, cause rupture of the catalyst pellet. MgO has a tendency to hydrate in steam at temperatures below 500°C98.  1.5   FBMR for Steam Reforming of Hydrocarbons 1.5.1 Limitations of fixed bed steam reformer The steam reforming process has evolved over decades resulting in less costly and more efficient plants due to better materials for reformer tubes, improved catalysts and closer control of carbon formation limits121. However, fixed bed reformers are limited by intra-particle diffusional and equilibrium limitations, in addition to high radial temperature and concentration gradients. These gradients become worse as catalyst deactivates due to reduced energy absorption in deactivated zones. To overcome these limitations, we focus on: (1) hydrogen removal using permselective membranes, and (2) fluidization of the catalysts. 1.5.2 Hydrogen removal using permselective membranes Figure 1.11  shows the thermodynamic limits for steam reforming of methane at a steam-to- carbon ratio of 3 with varying temperature and pressure122. Removal of the main products can drive the reaction towards completion, in accordance with Le Chatelier’s principle. In-situ removal of CO2 from the steam reformer is an attractive method achieving dual targets towards a green economy; enhancing hydrogen yield, as well as concentrating CO2 for sequestration. Alternatively, CO2 recovered from steam reforming may be recycled back to the reformer for applications requiring lower H2/CO ratios79. Any required H2/CO ratio within the natural range79 as depicted in Figure 1.2,  can be achieved by manipulating the CO2 recycle. On the other hand, recycling CO2-lean gas to the reformer can significantly enhance the hydrogen yield123,124. More information about sorption-enhanced steam reforming can be found elsewhere 125-127.  16 Enhancement of hydrogen yield can also be achieved by selectively removing hydrogen. In fact, yield enhancement is more sensitive to hydrogen removal than CO2 removal 123. Selective hydrogen removal using membranes enhances hydrocarbon conversion by favorably shifting the equilibrium conversion, while also purifying the hydrogen product128-130. Equivalent conversions can be achieved at much lower temperatures with membranes as for much higher temperatures without membranes66, as seen in Figure 1.12. Among the various options available for hydrogen selective membranes, Pd or Pd-alloy membranes theoretically offer infinite selectivity, but are currently expensive and suffer from challenges of structural integrity like hydrogen embrittlement and defects131-134. Microporous and dense ceramic membranes show great promise due to their comparatively low cost, and ability to withstand higher temperatures, but permselectivity and permeation flux remain issues to be overcome. Reviews of membranes for hydrogen separation are available132,135-138. Reviews of reaction systems using membranes for selective removal of hydrogen have also been published134,139-143. 1.5.2.1 Dense metallic membranes for hydrogen separation Figure 1.13 shows the hydrogen permeability of some materials132. As seen, metals like niobium, vanadium and tantalum have relatively high hydrogen permeability, but they are difficult to use for hydrogen separation due to formation of oxide layers132. Palladium is widely accepted as the most practical material for hydrogen separation membranes143, with infinite selectivity and excellent resistance to oxidation and corrosion. 1.5.2.2 Critical temperature for Pd membranes Below a critical temperature of 298°C, the α and β (sometimes referred to as α’) phases of palladium hydrides co-exist138. Depending on the composition of the hydride, there can be a transition between the two phases. Figure 1.14 shows hydrogen pressure-composition isotherms for the Pd-H system144,145. Difference in lattice constants between the two phases causes severe lattice strains during phase transition. After a few cycles of α ' β transitions, the palladium membrane becomes brittle142, and eventually develops micro-cracks. Thus, exposure to hydrogen at conditions where the β phase can form must be avoided, and the reactor must be thoroughly purged with inert gas to desorb hydrogen prior to cooling from high temperature146. Alloying with other materials like Ag, Cu or Ru decreases the transition temperature significantly. Alloying also improves the  17 hydrogen flux through the membranes. For example, a Pd77Ag23 alloy leads to a 70% gain in permeability compared to pure Pd142. 1.5.2.3 Permeation flux The mechanism of hydrogen permeation through Pd membranes can be described as follows: 1. Dissociative adsorption on the membrane surface. 2. Solution of atomic hydrogen in the Pd metal, and diffusion across the membrane in a solid solution to the permeate side. 3. Associative desorption from the permeate side. When diffusion is the rate-limiting step, the hydrogen permeation rate across the membranes follows Sieverts’ equation: )exp 22 2 2 2 0 ,MH,RH H H M PH PP(RT E δ PAQ −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −=        (1.13) 1.5.2.4 Membrane reactor for steam reforming of higher hydrocarbons There has been recent interest in extending steam reforming to higher hydrocarbons for hydrogen production. Steam reforming of higher hydrocarbons has been found to be practically irreversible and complete at industrial steam reforming temperature49,85. It is limited by the thermodynamic equilibrium, as for methane steam reforming7, see Figure 1.15. Methane is formed due to methanation reactions as noted above. In principle, selective removal of hydrogen could also enhance the overall hydrocarbon conversion and hydrogen yield. Few experimental data are available for steam reforming of higher hydrocarbons in membrane reactors. Chen et al147,148 studied steam reforming of a synthetic fuel prepared by mixing liquid hydrocarbons (C4 to C8) with an average formula of C6.43H14.84 with in-situ 4.5 µm thick Pd membrane supported on a porous α-alumina tube. The feasibility of a one-step process for production of hydrogen was demonstrated, with a permeate hydrogen product purity of 99.5% reported for a fixed bed membrane reactor operating at 450 - 550°C and 200 - 900 kPa. Damle149 used 25 µm thick planar Pd-Ag membrane supported on a porous stainless steel frame for hydrogen removal from a small packed bed steam reformer with feedstocks of methanol, butane and Clearite®, a commercially available sulfur-free grade of kerosene. With the latter as feed and operation at 620°C, 5.84 bars and a steam-to-carbon ratio of 3.5, a hydrogen yield higher than equilibrium prediction was obtained, reasonably matching simulation predictions150.    18 1.5.3 Fluidization of catalysts Steam reforming reactions for hydrocarbons are rapid, making catalysts highly under-utilized due to high intra-particle diffusional resistances in traditional fixed bed reactors. The large catalyst pellets result in effectiveness factors of only 10-2 to 10-3 in industrial operating conditions151,152. Elnashaie et al153 proposed a fluidized bed reactor with fine catalyst powders suspended in the reaction environment. This enables the reactor to operate with a smaller pressure drop than in a packed bed. Effectiveness factors can also be greatly improved to almost unity in this manner. However, since fluidized bed steam reforming is not practiced on large scales, being instead limited to pilot-scale demonstration units. Fluidizable steam reforming catalysts are also not available commercially. They can be prepared by crushing commercially available catalyst pellets to the required size distribution154, or prepared in-house155,156. Materials strength is of significant concern for such catalysts which suffer from attrition. 1.5.4 Fluidized bed membrane reformer Combining catalyst fluidization with a membrane-assisted reforming, Adris et al.3,4,157 proposed the fluidized bed membrane reactor (FBMR). Over the years, this concept has been extended to other reaction systems158-161. Grace et al162 summed up the potential advantages of hydrogen production in a fluidized bed membrane reactor: a) higher yields by reducing thermodynamic equilibrium limitations, b) process intensification by combining three vessels into one, c) reduced temperatures of operation, d) countering the adverse effects of pressure on equilibrium conversion, e) virtually eliminating intra-catalyst diffusional limitations, f) high productivity per unit volume of reformer, and g) flexibility in using alternative feedstocks.  1.6   FBMR for Steam Reforming of Higher Hydrocarbons 1.6.1 Previous studies Chen163 and Chen et al.48,49,164,165 envisioned and modeled an FBMR for steam reforming of higher hydrocarbons based on heptane as the simulated feedstock, and a circulating fluidized bed configuration, as shown in Figure 1.16. This configuration allowed incorporation of a  19 regenerator in the loop, where carbon deposited on the catalyst particles could be burned, with the energy generated due to the combustion carried by the hot catalyst particles back to the reformer. By carefully choosing the process parameters, it was shown that an autothermal mode of operation could be achieved. A second autothermal mode was proposed by distributed feeding of oxygen through oxygen-selective membranes immersed in the reformer. However, this configuration has some practical challenges: a) Hydrogen permeation flux available with currently commercially available membrane panels remains low. The flux is strongly dependent on the membrane layer thickness. With the current thickness of membranes, a very high height-to-diameter ratio would be required for a reformer operating in the fast fluidized bed regime166. b) Given the high particle velocities in fast fluidized beds, membranes are likely to be subjected to a more erosive environment than in bubbling fluidized beds. c) Nickel catalysts would be oxidized in the regenerator while burning off the carbon deposits. 1.6.2 Current research Rakib et al7 investigated a bubbling fluidized bed mode of operation for hydrogen production from heptane, as a surrogate for naphtha. A two-phase bubbling bed model was written, with hydrogen withdrawn selectively through palladium membranes. A practical upper limit temperature of 650°C and a pressure of 10 bars were the base cases for the simulation. It was predicted that the FBMR could be considered as two overlapping zones, a short zone above the distributor where the heptane is fully consumed, and an extended zone where the steam reforming of methane and the water gas shift reaction occur, while pure hydrogen is continuously withdrawn. The FBMR was predicted to provide a compact reactor system for hydrogen production from higher hydrocarbons, by combining the pre-reformer, reformer and hydrogen purification in a single unit. Simulation results7 also showed that thinner membranes could minimize the residual methane and hydrogen in the reformer, and maximize the pure hydrogen yield. The membrane packing factor, expressed as the membrane surface area per unit volume of the reactor, was found to be important, demonstrating that the permeation flux of the currently commercially available membrane panels is a limiting factor. A higher membrane packing factor could significantly reduce the volume of the reactor, but hydrodynamic factors provide practical constraints on the number and spacing of membrane surfaces167.   20 Table 1.1: World fertilizer consumption (Calendar year basis)  Million tons 2008 2009 (estimated) 2013 (forecast) Nitrogen, N 99.3 101.0 110.4 Phosphorus, P2O5 35.9 37.2 43.9 Potassium, K2O 24.8 25.0 31.0 Total 160.0 163.2 185.3   21  (a)  Global hydrogen consumption sectors (Adapted from Kruse et al.38)   (b)  Global sources of hydrogen (Adapted from Ewan et al.39)  Figure 1.1: Hydrogen demand and sources    Figure 1.2: Natural range of H2/CO ratio for natural gas79.     22  Figure 1.3: Global distribution of oil and gas reserves  23    Figure 1.4: Natural gas price fluctuation 92     Figure 1.5: Crude oil price fluctuation 93   24    Figure 1.6: Typical flow-chart configuration of a pre-reformer 99      Figure 1.7: Pre-reformer temperature profiles for different feeds 87  25    Figure 1.8: Reforming section for production of ammonia synthesis gas 65    Figure 1.9: Reforming for the production of methanol synthesis gas 65   26   Figure 1.10: Relative rates of carbon formation on nickel catalysts 94    Figure 1.11: Equilibrium methane conversion in steam reforming of methane as a function of temperature and pressure (steam-to-carbon ratio = 3) 122   27    Figure 1.12: Enhancement of methane conversion with in-situ hydrogen removal. Steam-to- carbon ratio = 3, reactor pressure = 1000 kPa 66    Figure 1.13: Hydrogen permeabilities of selected metals 132    28   Figure 1.14: Pressure-composition isotherms of Pd-H system 144, 145   29    Figure 1.15: Equilibrium compositions for heptane steam reforming    30    Figure 1.16: Circulating fluidized bed membrane reformer configuration for steam reforming of heptane 49,164   31  1.7   References 1. Adris, A. M.; Pruden, B. B.; Lim, C. J.; Grace, J. R., On the reported attempts to radically improve the performance of the steam methane reforming reactor. Canadian Journal of Chemical Engineering 1996, 74, (2), 177-186. 2. Ferreira-Aparicio, P.; Benito, M. J.; Sanz, J. L., New trends in reforming technologies: From hydrogen industrial plants to multifuel microreformers. Catalysis Reviews-Science and Engineering 2005, 47, (4), 491-588. 3. Adris, A. M.; Lim, C. J.; Grace, J. R., The fluidized bed membrane reactor system: A pilot scale experimental study. Chemical Engineering Science 1994, 49, (24B), 5833-5843. 4. Adris, A. M.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. Fluidized bed reaction system for steam/hydrocarbon gas reforming to produce hydrogen. US Patent US 5326550, 1994. 5. Adris, A. M.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. Fluidized bed reaction system for steam/hydrocarbon gas reforming to produce hydrogen. Canadian Patent CA 2081170, 2002. 6. Adris, A. M.; Boyd, T.; Brereton, C.; Grace, J.; Lim, C. J.; Wolfs, W. In Production of pure hydrogen by the fluidized bed membrane reactor, 14th World Hydrogen Energy Conference, Montreal, June 2002; Montreal, June 2002. 7. Rakib, M. A.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J.; Bolkan, Y. G., Kinetic simulation of a compact reactor system for hydrogen production by steam reforming of higher hydrocarbons. The Canadian Journal of Chemical Engineering 2008, 86, (3), 403-412. 8. IPCC (2007) Climate Change 2007: Synthesis Report, Geneva, Switzerland. 9. Miller, L.; Douglas, B. C., On the rate and causes of twentieth century sea-level rise. Philosophical Transactions of the Royal Society a-Mathematical Physical and Engineering Sciences 2006, 364, (1841), 805-820. 10. Nicholls, R. J.; P.Leatherman, S., The implication of accelerated sea level rise for developing countries: A discussion. J. Coastal Res. (Special Issue) 1995, 14. 11. Arnell, N. W., Climate change and global water resources. Global Environmental Change-Human and Policy Dimensions 1999, 9, S31-S49. 12. Patz, J. A.; Epstein, P. R.; Burke, T. A.; Balbus, J. M., Global climate change and emerging infectious diseases. Journal of the American Medical Association 1996, 275, (3), 217- 223. 13. How to combat global warming: A report prepared for the CC8 Conference, Oslo, . The Bellona Foundation: June 5-6, 2008.  32  14. Johnston, B.; Mayo, M. C.; Khare, A., Hydrogen: the energy source for the 21st century. Technovation 2005, 25, (6), 569-585. 15. Midilli, A.; Ay, M.; Dincer, I.; Rosen, M. A., On hydrogen and hydrogen energy strategies I: current status and needs. Renewable & Sustainable Energy Reviews 2005, 9, (3), 255-271. 16. Bossel, U., Does a hydrogen economy make sense? Proceedings of the IEEE 2006, 94, (10), 1826-1837. 17. Shinnar, R., The hydrogen economy, fuel cells, and electric cars. Technology in Society 2003, 25, 455-476. 18. Shinnar, R., The mirage of the H2 economy. Clean Techn Environ Policy 2004, 6, 223- 226. 19. Forsberg, C. W., The hydrogen economy is coming - The question is where? Chemical Engineering Progress 2005, 101, (12), 20-22. 20. Sherif, S. A.; Barbir, F.; Veziroglu, T. N., Towards a hydrogen economy. The Electricity Journal 2005, 18, (6), 62-76. 21. Towler, G. P.; Mann, R.; Serriere, A. J. L.; Gabaude, C. M. D., Refinery hydrogen management: Cost analysis of chemically-integrated facilities. Industrial & Engineering Chemistry Research 1996, 35, (7), 2378-2388. 22. Hallale, N.; Liu, F., Refinery hydrogen management for clean fuels production. Advances in Environmental Research 2001, 6, (1), 81-98. 23. Rostrup-Nielsen, J. R.; Rostrup-Nielsen, T., Large-scale hydrogen production. CatTech 2002, 6, (4), 150-159. 24. Alves, J. J.; Towler, G. P., Analysis of refinery hydrogen distribution systems. Industrial & Engineering Chemistry Research 2002, 41, (23), 5759-5769. 25. Hallale, N., Boosting refinery profitability through hydrogen management technology. Chemical Engineer 2001, 725, 40-41. 26. Boyce, C. A.; Crews, M. A.; Ritter, R., Time for a new hydrogen plant? Hydrocarbon Engineering February 2004. 27. Forsberg, C. W., Future hydrogen markets for large-scale hydrogen production systems. International Journal of Hydrogen Energy 2007, 32, (4), 431-439. 28. Soderbergh, B.; Robelius, F.; Aleklett, K., A crash programme scenario for the Canadian oil sands industry. Energy Policy 2007, 35, (3), 1931-1947.  33  29. Emerging markets. Canadian Fertilizer Products Forum, Agriculture and Agri-Food Canada 2008. 30. Mineral fertilizer production and the environment. Part 1. The fertilizer industry's manufacturing processes and environmental issues. Technical Report No. 26 Part 1, United Nations Industrial Development Organization, in collaboration with International Fertilizer Industry Association 1998. 31. Heffer, P.; Prudhomme, M. In Fertilizer Outlook 2009-2013, International Fertilizer Industry Association, 77th IFA Annual Conference, Shanghai, China, 25-27 May, 25-27 May, 2009; Shanghai, China, 25-27 May, 2009. 32. Ogden, J. M.; Steinbugler, M. M.; Kreutz, T. G., A comparison of hydrogen, methanol and gasoline as fuels for fuel cell vehicles: implications for vehicle design and infrastructure development. Journal of Power Sources 1999, 79, (2), 143-168. 33. Masih, A. M. M.; Albinali, K.; deMello, L., Price dynamics of natural gas and the regional methanol markets. Energy Policy 2010, 38, 1372-1378. 34. de Lima, L. C.; Duarte, J. B. F.; Veziroglu, T. N., A proposal of an alternative route for the reduction of iron ore in the eastern Amazonia. International Journal of Hydrogen Energy 2004, 29, (6), 659-661. 35. Eliezer, D.; Eliaz, N.; Senkov, O. N.; Froes, F. H., Positive effects of hydrogen in metals. Materials Science and Engineering A: Structural Materials Properties Microstructure and Processing 2000, 280, (1), 220-224. 36. Ramachandran, R.; Menon, R. K., An overview of industrial uses of hydrogen. International Journal of Hydrogen Energy 1998, 23, (7), 593-598. 37. Wender, I., Reactions of synthesis gas. Fuel Processing Technology 1996, 48, (3), 189- 297. 38. Kruse, B.; Grinna, S.; Buch, C., Hydrogen - Status and Possibilities. The Bellona Foundation: 2002. 39. Ewan, B. C. R.; Allen, R. W. K., A figure of merit assessment of the routes to hydrogen. International Journal of Hydrogen Energy 2005, 30, (8), 809-819. 40. deJong, K. P.; Vanwechem, H. M. H., Carbon: Hydrogen carrier or disappearing skeleton. International Journal of Hydrogen Energy 1995, 20, (6), 493-499. 41. Scholz, W. H., Processes for industrial production of hydrogen and associated environmental effects. Gas Separation & Purification 1993, 7, (3), 131-139  34  42. Balthasar, W.; Hambleton, D. J., Industrial scale production of hydrogen from natural gas, naphtha and coal. Advances in Hydrogen Energy 1978, 2, 1007-1028. 43. Dry, M. E., The production of hydrocarbons from coal. Endeavour 1984, 8, (1). 44. van Dyk, J. C.; Keyser, M. J.; Coertzen, A., Syngas production from South African coal sources using Sasol-Lurgi gasifiers. International Journal of Coal Geology 2006, 65, (3-4), 243- 253. 45. Ogden, J. M.; Williams, R. H.; Larson, E. D., Toward a hydrogen-based transportation system. http://www.princeton.edu/pei/energy/publications/texts/Ogden_01- Toward_a_hydrogen_based_transportation_system.pdf (2001) Accessed 20 Feb 2010. 46. Shoko, E.; McLellan, B.; Dicks, A. L.; da Costa, J. C. D., Hydrogen from coal: Production and utilisation technologies. International Journal of Coal Geology 2006, 65, (3-4), 213-222. 47. Rostrup-Nielsen, J. R.; Sehested, J.; Norskov, J. K., Hydrogen and synthesis gas by steam and CO2 reforming. Advances in Catalysis 2002, 47, 65-139. 48. Chen, Z.; Elnashaie, S. S. E. H., Autothermal CFB membrane reformer for hydrogen production from heptane. Chemical Engineering Research & Design 2005, 83, (A7), 893-899. 49. Chen, Z. X.; Yan, Y. B.; Elnashaie, S. S. E. H., Modeling and optimization of a novel membrane reformer for higher hydrocarbons. AIChE Journal 2003, 49, (5), 1250-1265. 50. El-Solh, T. Dry reforming of methane in a fast fluidized bed reactor: Catalysis & kinetics. University of Western Ontario, 2002. 51. Mleczko, L.; Malcus, S.; Wurzel, T., Catalytic reformer-combustor: A novel reactor concept for synthesis gas production. Industrial & Engineering Chemistry Research 1997, 36, (11), 4459-4465. 52. Tomishige, K., Syngas production from methane reforming with CO2/H2O and O2 over NiO-MgO solid solution catalyst in fluidized bed reactors. Catalysis Today 2004, 89, 405-418. 53. Jensen, M. B.; Raberg, L. B.; Sjastad, A. O.; Olsbye, U., Mechanistic study of the dry reforming of propane to synthesis gas over a Ni/Mg(Al)O catalyst. Catalysis Today 2009, 145, (1-2), 114-120. 54. Sutton, D.; Parle, S. M.; Ross, J. R. H., The CO2 reforming of the hydrocarbons present in a model gas stream over selected catalysts. Fuel Processing Technology 2002, 75, (1), 45-53. 55. Puolakka, K. J.; Juutilainen, S.; Krause, A. O. I., Combined CO2 reforming and partial oxidation of n-heptane on noble metal zirconia catalysts. Catalysis Today 2006, 115, (1-4), 217- 221.  35  56. Puolakka, K. J.; Krause, A. O. I., Combined CO2 reforming and partial oxidation of fuel compounds. Catalysis Letters 2007, 116, (3-4), 87-93. 57. Wang, S. B.; Lu, G. Q. M.; Millar, G. J., Carbon dioxide reforming of methane to produce synthesis gas over metal-supported catalysts: State of the art. Energy & Fuels 1996, 10, (4), 896-904. 58. Muradov, N.; Smith, F.; T-Raissi, A., Catalytic activity of carbons for methane decomposition reaction. Catalysis Today 2005, 102, 225-233. 59. Otsuka, K.; Ogihara, H.; Takenaka, S., Decomposition of methane over Ni catalysts supported on carbon fibers formed from different hydrocarbons Carbon 2003, 41, 223-233. 60. Abbas, H. F.; Daud, W. M. A. W., Hydrogen production by methane decomposition: A review. International Journal of Hydrogen Energy 2010, 35, 1160-1190. 61. Muradov, N. Z.; Veziroglu, T. N., "Green" path from fossil-based to hydrogen economy: An overview of carbon-neutral technologies. International Journal of Hydrogen Energy 2008, 33, (23), 6804-6839. 62. Rostrup-Nielsen, J. R.; Dybkjaer, I.; Christiansen, L. J., Steam reforming: Opportunities and limits of the technology. Kluwer Academic Publishers: Ontario, Canada, Aug 25 - Sep 5, 1991. 63. Simbeck, D. R.; Chang, E., Hydrogen supply: Cost estimate for hydrogen pathways - Scoping analysis. 2002. 64. Reyes, S. C.; Sinfelt, J. H.; Feeley, J. S., Evolution of processes for synthesis gas production: Recent developments in an old technology. Industrial & Engineering Chemistry Research 2003, 42, (8), 1588-1597. 65. Dybkjaer, I., Tubular reforming and autothermal reforming of natural gas - an overview of available processes. Fuel Processing Technology 1995, 42, (2-3), 85-107. 66. Boyd, T.; Grace, J.; Lim, C. J.; Adris, A. E. M., Hydrogen from an internally circulating fluidized bed membrane reactor. International Journal of Chemical Reactor Engineering 2005, 3, A58. 67. Rakib, M. A.; Alhumaizi, K. I., Modeling of a fluidized bed membrane reactor for the steam reforming of methane: Advantages of oxygen addition for favorable hydrogen production. Energy & Fuels 2005, 19, (5), 2129-2139.  36  68. Dogan, M.; Posarac, D.; Grace, J.; Adris, A. M.; Lim, C. J., Modeling of autothermal steam methane reforming in a fluidized bed membrane reactor. International Journal of Chemical Reactor Engineering 2003, 1, A2. 69. Patil, C. S.; Annaland, M. V.; Kuipers, J. A. M., Design of a novel autothermal membrane-assisted fluidized-bed reactor for the production of ultrapure hydrogen from methane. Industrial & Engineering Chemistry Research 2005, 44, (25), 9502-9512. 70. Tiemersma, T. P.; Patil, C. S.; Annaland, M. V. S.; Kuipers, J. A. M., Modelling of packed bed membrane reactors for autothermal production of ultrapure hydrogen. Chemical Engineering Science 2006, 61, (5), 1602-1616. 71. Aasberg-Petersen, K.; Christensen, T. S.; Nielsen, C. S.; Dybkjaer, I., Recent developments in autothermal reforming and pre-reforming for synthesis gas production in GTL applications. Fuel Processing Technology 2003, 83, (1-3), 253-261. 72. Aasberg-Petersen, K.; Hansen, J. H. B.; Christensen, T. S.; Dybkjaer, I.; Christensen, P. S.; Nielsen, C. S.; Madsen, S. E. L. W.; Rostrup-Nielsen, J. R., Technologies for large-scale gas conversion. Applied Catalysis A: General 2001, 221, (1-2), 379-387. 73. Rostrup-Nielsen, J. R., New aspects of syngas production and use. Catalysis Today 2000, 63, (2-4), 159-164. 74. Hamakawa, S.; Sato, K.; Inoue, T.; Nishioka, M.; Kobayashi, K.; Mizukami, F., Design of one-component ceramic membrane-reactor for natural gas conversion. Catalysis Today 2006, 117, (1-3), 297-303. 75. Jin, W.; Gu, X.; Li, S.; Huang, P.; Xu, N.; Shi, J., Experimental and simulation study on a catalyst packed tubular dense membrane reactor for partial oxidation of methane to syngas. Chemical Engineering Science 2000, 55, (14), 2617-2625. 76. Ritchie, J. T.; Richardson, J. T.; Luss, D., Ceramic membrane reactor for synthesis gas production. Aiche Journal 2001, 47, (9), 2092-2101. 77. Ikeguchi, M.; Mimura, T.; Sekine, Y.; Kikuchi, E.; Matsukata, M., Reaction and oxygen permeation studies in Sm0.4Ba0.6Fe0.8Co0.2O3-δ membrane reactor for partial oxidation of methane to syngas. Applied Catalysis A: General 2005, 290, (1-2), 212-220. 78. Rostrup-Nielsen, J. R., Syngas in perspective. Catalysis Today 2002, 71, (3-4), 243-247. 79. Crews, M. A.; Shumake, B. G., Hydrogen production and supply: Meeting refiners’ growing needs. In Practical Advances in Petroleum Processing, Hsu, C. S.; Robinson, P. R., Eds. Springer: New York, 2006; Vol. 1.  37  80. Raina, J. L., Outlook on feedstocks for petrochemical and fertilizer industry. Chemical Engineering World 1998, XXXIII, (12), 33-38. 81. Kaila, R. K.; Krause, A. O. I., Reforming of higher hydrocarbons. Studies in Surface Science and Catalysis 2004, 147, 247-252. 82. Melo, F.; Morlane´s, N., Naphtha steam reforming for hydrogen production. Catalysis Today 2005, 107-108, 458-466. 83. Springmann, S.; Bohnet, M.; Docter, A.; Lamm, A.; Eigenberger, G., Cold start simulations of a gasoline based fuel processor for mobile fuel cell applications. Journal of Power Sources 2004, 128, (1), 13-24. 84. Zhu, W.; Han, W.; Xiong, G.; Yang, W., Mixed reforming of heptane to syngas in the Ba0.5Sr0.5Co0.8Fe0.2O3 membrane reactor. Catalysis Today 2005, 104, 149-153. 85. Christensen, T. S., Adiabatic prereforming of hydrocarbons - An important step in syngas production. Applied Catalysis A: General 1996, 138, (2), 285-309. 86. Shumake, G.; Abudiab, T., Operation hydrogen. Hydrocarbon Engineering February 2006. 87. Broadhurst, P. V.; Cotton, B. J., Taking feedstock. Hydrocarbon Engineering March 2005. 88. Shumake, G.; Small, J., Mixing things up in hydrogen plants. Hydrocarbon Engineering November 2006. 89. Nielsen, S. E.; Dybkjaer, I. In Use of adiabatic prereforming in ammonia plants, AIChE Ammonia Safety Symposium, Boston, Massachusetts, September 1996; Boston, Massachusetts, September 1996. 90. Christensen, P. V., Revamping ammonia plants to follow feedstock situation and market demands. In FAI Symposium on "Improving Productivity of Ammonia and Urea Plants", New Delhi, April 19-20, 2001. 91. Forsberg, C. W., Hydrogen markets: Implications for hydrogen production technologies. In 2005 AIChE Spring Meeting, Atlanta, Georgia, April 10-14, 2005. 92. Natural Gas Year-In-Review 2008. U.S. Energy Information Administration April 2009. 93. Short Term Energy Outlook 2009. U.S. Energy Information Administration June 2009. 94. Rostrup-Nielsen, J. R., Catalytic steam reforming. In Advances in Catalysis Science and Technology, Anderson, J. R.; Boudart, M., Eds. Springer-Verlag: Berlin, 1984; Vol. 5, pp 1-117. 95. Appl, M., Ammonia: Principles and industrial practice. Wiley-VCH: 1999. 96. Ridler, D. E.; Twigg, M. V., Catalyst Handbook. Wolfe publishing ltd.: London, 1989.  38  97. Rostrup-Nielsen, J. R.; Christiansen, L. J., Tubular steam reforming. In Chemical Reaction and Reactor Design, Tominaga, H.; Tamaki, M., Eds. John Wiley: 1997. 98. Bartholomew, C. H.; Farrauto, R. J., Fundamentals of industrial catalytic processes. 2 ed.; John Wiley: 2005. 99. Rostrup-Nielsen, J. R.; Christensen, T. S.; Dybkjaer, I., Steam reforming of liquid hydrocarbons. In Studies in Surface Science and Catalysis, Rao, T. S. R. P.; Muralidhar, G., Eds. 1998; Vol. 113. 100. Yu, Y. H., Simulation of secondary reformer in industrial ammonia plant. Chemical Engineering & Technology 2002, 25, (3), 307-314. 101. Sircar, S.; Golden, T. C., Purification of hydrogen by pressure swing adsorption. Separation Science and Technology 2000, 35, (5), 667-687. 102. Rostrup-Nielsen, J. R.; Dybkjaer, I.; Christiansen, L. J., Steam reforming: Opportunities and limits of the technology. In NATO Advanced Study Institute on "Chemical Reactor Technology for Environmentally Safe Reactors and Products", deLasa, H. I.; Dogu, G.; Ravella, A., Eds. Kluwer Academic Publishers: Ontario, Canada, 1992; pp 249-282. 103. Rostrup-Nielsen, J. R., Activity of nickel catalysts for steam reforming of hydrocarbons. Journal of Catalysis 1973, 31, (2), 173-199. 104. Trimm, D. L.; O¨nsan, Z. I., Onboard fuel conversion for hydrogen-fuel-cell-driven vehicles. Catalysis Reviews-Science and Engineering 2001, 43, (1 & 2), 31-84. 105. Trimm, D. L., Coke formation and minimization during steam reforming reactions. Catalysis Today 1997, 37, 233-238. 106. Wang, L. S.; Murata, K.; Inaba, M., Control of the product ratio of CO2/(CO+CO2) and inhibition of catalyst deactivation for steam reforming of gasoline to produce hydrogen. Applied Catalysis B-Environmental 2004, 48, (4), 243-248. 107. Rase, H. F., Synthesis gas and its products. In Handbook of Commercial Catalysts: Heterogeneous Catalysts, CRC Press: 2000; pp 403-482. 108. Sehested, J., Four challenges for nickel steam-reforming catalysts. Catalysis Today 2006, 111, (1-2), 103-110. 109. Bartholomew, C. H., Mechanisms of catalyst deactivation. Applied Catalysis A: General 2001, 212, (1-2), 17-60. 110. Rostrup-Nielsen, J. R.; Sehested, J., Whisker carbon revisited. Studies in Surface Sciences and Catalysis 2001, 139, 1-13.  39  111. Goralski, J.; Grams, J.; Paryjczak, T.; Rzeznicka, I., Investigation of the coke deposit on Ni-Al2O3 and Co-Al2O3 catalysts. Carbon 2002, 40, 2025-2028. 112. Zhang, T. J.; Amiridis, M. D., Hydrogen production via the direct cracking of methane over silica-supported nickel catalysts. Applied Catalysis A: General 1998, 167, (2), 161-172. 113. Snoeck, J. W.; Froment, G. F.; Fowles, M., Filamentous carbon formation and gasification: Thermodynamics, driving force, nucleation, and steady-state growth. Journal of Catalysis 1997, 169, (1), 240-249. 114. Tavares, M. T.; Alstrup, I.; Bernardo, C. A.; RostrupNielsen, J. R., Carbon formation and CO methanation on silica-supported nickel and nickel-copper catalysts in CO + H2 mixtures. Journal of Catalysis 1996, 158, (2), 402-410. 115. Golebiowski, A.; Stolecki, K.; Prokop, U.; Kusmierowska, A.; Borowiecki, T.; Denis, A.; Sikorska, C., Influence of potassium on the properties of steam reforming catalysts. Reaction Kinetics and Catalysis Letters 2004, 82, (1), 179-189. 116. Lisboa, J. D.; Santos, D. C. R. M.; Passos, F. B.; Noronha, F. B., Influence of the addition of promoters to steam reforming catalysts. Catalysis Today 2005, 101, (1), 15-21. 117. Mehr, J. Y.; Jozani, K. J.; Pour, A. N.; Zamani, Y., Influence of MgO in the CO2-steam reforming of methane to syngas by NiO/MgO/α-Al2O3 catalyst. Reaction Kinetics and Catalysis Letters 2002, 75, (2), 267-273. 118. Bengaard, H. S.; Norskov, J. K.; Sehested, J.; Clausen, B. S.; Nielsen, L. P.; Molenbroek, A. M.; Rostrup-Nielsen, J. R., Steam reforming and graphite formation on Ni catalysts. Journal of Catalysis 2002, 209, (2), 365-384. 119. Rostrup-Nielsen, J. R.; Christensen, T. S.; Dybkjaer, I., Steam reforming of liquid hydrocarbons. Studies in Surface Sciences and Catalysis 1998, 113, 81-95. 120. Sehested, J.; Gelten, J. A. P.; Helveg, S., Sintering of nickel catalysts: Effects of time, atmosphere, temperature, nickel-carrier interactions, and dopants. Applied Catalysis A: General 2006, 309, (2), 237-246. 121. Rostrup-Nielsen, J., Steam reforming of hydrocarbons. A historical perspective. Natural Gas Conversion Vii 2004, 147, 121-126. 122. Boyd, D. A. Internally circulating fluidized bed membrane reactor for high-purity hydrogen production. University of British Columbia, Vancouver, 2007. 123. Chen, Z. X.; Po, F.; Grace, J. R.; Lim, C. J.; Elnashaie, S.; Mahecha-Botero, A.; Rakib, M.; Shirasaki, Y.; Yasuda, I., Sorbent-enhanced/membrane-assisted steam-methane reforming. Chemical Engineering Science 2008, 63, (1), 170-182.  40  124. Po, F.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J., Hydrogen production from ex situ “Breaking” of the steam reforming thermodynamic equilibrium barrier. In 11th Asian Pacific Confederation of Chemical Engineering Congress, Kuala Lumpur, Malaysia, August 27th-30th, 2006. 125. Barelli, L.; Bidini, G.; Gallorini, F.; Servili, S., Hydrogen production through sorption- enhanced steam methane reforming and membrane technology: A review. Energy 2008, 33, (4), 554-570. 126. Lee, K. B.; Beaver, M. G.; Caram, H. S.; Sircar, S., Novel thermal-swing sorption- enhanced reaction process concept for hydrogen production by low-temperature steam-methane reforming. Industrial & Engineering Chemistry Research 2007, 46, (14), 5003-5014. 127. Johnsen, K.; Ryu, H. J.; Grace, J. R.; Lim, C. J., Sorption-enhanced steam reforming of methane in a fluidized bed reactor with dolomite as CO2-acceptor. Chemical Engineering Science 2006, 61, (4), 1195-1202. 128. Barbieri, G.; Violante, V.; DiMaio, F. P.; Criscuoli, A.; Drioli, E., Methane steam reforming analysis in a palladium-based catalytic membrane reactor. Industrial & Engineering Chemistry Research 1997, 36, (8), 3369-3374. 129. Shu, J.; Grandjean, B. P. A.; Kaliaguine, S., Methane steam reforming in asymmetric Pd- Ag and Pd-Ag/porous SS membrane reactors. Applied Catalysis A: General 1994, 119, (2), 305- 325. 130. Barbieri, G.; Marigliano, G.; Perri, G.; Drioli, E., Conversion-temperature diagram for a palladium membrane reactor. Analysis of an endothermic reaction: Methane steam reforming. Industrial & Engineering Chemistry Research 2001, 40, (9), 2017-2026. 131. Shu, J.; Bongondo, B. E. W.; Grandjean, B. P. A.; Kaliaguine, S., Morphological changes of Pd-Ag membranes upon hydrogen permeation. Journal of Materials Science Letters 1997, 16, (4), 294-297. 132. Adhikari, S.; Fernando, S., Hydrogen membrane separation techniques. Industrial & Engineering Chemistry Research 2006, 45, (3), 875-881. 133. Gao, H. Y.; Lin, Y. S.; Li, Y. D.; Zhang, B. Q., Chemical stability and its improvement of palladium-based metallic membranes. Industrial & Engineering Chemistry Research 2004, 43, (22), 6920-6930. 134. Saracco, G.; Versteeg, G. F.; Vanswaaij, W. P. M., Current hurdles to the success of high-temperature membrane reactors. Journal of Membrane Science 1994, 95, (2), 105-123.  41  135. Ockwig, N. W.; Nenoff, T. M., Membranes for hydrogen separation. Chemical Reviews 2007, 107, (10), 4078-4110. 136. Lu, G. Q.; da Costa, J. C. D.; Duke, M.; Giessler, S.; Socolow, R.; Williams, R. H.; Kreutz, T., Inorganic membranes for hydrogen production and purification: A critical review and perspective. Journal of Colloid and Interface Science 2007, 314, (2), 589-603. 137. Phair, J. W.; Donelson, R., Developments and design of novel (non-palladium-based) metal membranes for hydrogen separation. Industrial & Engineering Chemistry Research 2006, 45, (16), 5657-5674. 138. Uemiya, S., State-of-the-art of supported metal membranes for gas separation. Separation and Purification Methods 1999, 28, (1), 51-85. 139. Zaman, J.; Chakma, A., Inorganic Membrane Reactors. Journal of Membrane Science 1994, 92, (1), 1-28. 140. Basile, A., Hydrogen production using Pd-based membrane reactors for fuel cells. Topics in Catalysis 2008, 51, (1-4), 107-122. 141. Lukyanov, B. N.; Andreev, D. V.; Parmon, V. N., Catalytic reactors with hydrogen membrane separation. Chemical Engineering Journal 2009, 154, (1-3), 258-266. 142. Shu, J.; Grandjean, B. P. A.; Vanneste, A.; Kaliaguine, S., Catalytic palladium-based membrane reactors: A review. Canadian Journal of Chemical Engineering 1991, 69, (5), 1036- 1060. 143. Uemiya, S., Brief review of steam reforming using a metal membrane reactor. Topics in Catalysis 2004, 29, (1-2), 79-84. 144. Lewis, F. A., Hydrogen in palladium and palladium alloys. International Journal of Hydrogen Energy 1996, 21, (6), 461-464. 145. Bruning, H.; Sieverts, A., Z. Phys. Chem. A. 1933, 163, 409. 146. Paglieri, S. N.; Way, J. D., Innovations in palladium membrane research. Separation and Purification Methods 2002, 31, (1), 1-169. 147. Chen, Y. Z.; Wang, Y. Z.; Xu, H. Y.; Xiong, G. X., Integrated one-step PEMFC-grade hydrogen production from liquid hydrocarbons using Pd membrane reactor. Industrial & Engineering Chemistry Research 2007, 46, (17), 5510-5515. 148. Chen, Y. Z.; Xu, H. Y.; Wang, Y. Z.; Xiong, G. X., Hydrogen production from the steam reforming of liquid hydrocarbons in membrane reactor. Catalysis Today 2006, 118, (1-2), 136- 143.  42  149. Damle, A. S., Hydrogen production by reforming of liquid hydrocarbons in a membrane reactor for portable power generation - Experimental studies. Journal of Power Sources 2009, 186, (1), 167-177. 150. Damle, A. S., Hydrogen production by reforming of liquid hydrocarbons in a membrane reactor for portable power generation - Model simulations. Journal of Power Sources 2008, 180, (1), 516-529. 151. Elnashaie, S. S. E. H.; Elshishini, S. S., Modelling, Simulation and Optimization of Industrial Fixed Bed Catalytic Reactors. Gordon and Breach Science Publishers: London 1993. 152. Soliman, M. A.; Elnashaie, S. S. E. H.; Alubaid, A. S.; Adris, A., Simulation of Steam Reformers for Methane. Chemical Engineering Science 1988, 43, (8), 1801-1806. 153. Elnashaie, S. S. E. H.; Adris, A. M., A fluidized bed steam reformer for methane. In Fluidization VI, Grace, J. R.; Shemilt, L. W.; Bergougnou, M. A., Eds. Engineering Foundation: New York, 1989. 154. Adris, A. M. A fluidized bed membrane reactor for methane steam reforming: Experimental verification and model validation. University of British Columbia, Vancouver, 1994. 155. Magrini-Bair, K.; Czernik, S.; French, R.; Parent, Y.; Ritland, M.; Chornet, E., Fluidizable catalysts for producing hydrogen by steam reforming biomass pyrolysis liquids. Proceedings of the 2002 U.S. DOE Hydrogen Program Review (NREL/CP-610-32405), 1-11. 156. El Solh, T.; Jarosch, K.; de Lasa, H. I., Fluidizable catalyst for methane reforming. Applied Catalysis A: General 2001, 210, (1-2), 315-324. 157. Adris, A. M.; Elnashaie, S. S. E. H.; Hughes, R., A fluidized-bed membrane reactor for the steam reforming of methane. Canadian Journal of Chemical Engineering 1991, 69, (5), 1061-1070. 158. Deshmukh, S. A. R. K.; Laverman, J. A.; Annaland, M. V.; Kuipers, J. A. M., Development of a membrane-assisted fluidized bed reactor. 2. Experimental demonstration and modeling for the partial oxidation of methanol. Industrial & Engineering Chemistry Research 2005, 44, (16), 5966-5976. 159. Alonso, M.; Lorences, M. J.; Pina, M. P.; Patience, G. S., Butane partial oxidation in an externally fluidized bed-membrane reactor. Catalysis Today 2001, 67, (1-3), 151-157. 160. Ahchieva, D.; Peglow, M.; Heinrich, S.; Morl, L.; Wolff, T.; Klose, F., Oxidative dehydrogenation of ethane in a fluidized bed membrane reactor. Applied Catalysis a-General 2005, 296, (2), 176-185.  43  161. Gimeno, M. P.; Z.T.Wu; Soler, J.; Herguido, J.; Li, K.; Menéndez, M., Combination of a two-zone fluidized bed reactor with a Pd hollow fibre membrane for catalytic alkane dehydrogenation. Chemical Engineering Journal 2009, 155, 298 - 303. 162. Grace, J.; Elnashaie, S. S. E. H.; Lim, C. J., Hydrogen production in fluidized beds with in-situ membranes. International Journal of Chemical Reactor Engineering 2005, 3, A41. 163. Chen, Z. A novel circulating fluidized bed membrane reformer for efficient pure hydrogen production for fuel cells from higher hydrocarbons. Auburn University, Auburn, 2004. 164. Chen, Z. X.; Yan, Y. B.; Elnashaie, S. S. E. H., Hydrogen production and carbon formation during the steam reforming of heptane in a novel circulating fluidized bed membrane reformer. Industrial & Engineering Chemistry Research 2004, 43, (6), 1323-1333. 165. Chen, Z. X.; Elnashaie, S. S. E. H., Bifurcation behavior and efficient pure hydrogen production for fuel cells using a novel autothermic membrane circulating fluidized-bed (CFB) reformer: Sequential debottlenecking and the contribution of John Grace. Industrial & Engineering Chemistry Research 2004, 43, (18), 5449-5459. 166. Mahecha-Botero, A.; Chen, Z. X.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J.; Rakib, M.; Yasuda, I.; Shirasaki, Y., Comparison of fluidized bed flow regimes for steam methane reforming in membrane reactors: A simulation study. Chemical Engineering Science 2009, 64, (16), 3598-3613. 167. Adris, A. E. M.; Grace, J. R., Characteristics of fluidized-bed membrane reactors: Scale- up and practical issues. Industrial & Engineering Chemistry Research 1997, 36, (11), 4549- 4556.       44  CHAPTER 2.      PILOT SCALE EXPERIMENTAL SETUP FOR HYDROGEN PRODUCTION FROM HIGHER HYDROCARBONS: SAFETY CONSIDERATIONS AND IMPLEMENTATION†   2.1 Introduction Growing demand for hydrogen and synthesis gas in the process sector, and projected utilization of pure hydrogen for fuel cells make research on steam reforming very important, the main focus being overcoming the several drawbacks associated with traditional fixed bed steam reformers. When higher hydrocarbon feedstocks are used for hydrogen generation, the main reactions can be written1 as: 22mn H2 mnnCOOnHHC ⎟⎠ ⎞⎜⎝ ⎛ ++→+   (∆Ho298 = 1108 kJ/mol   for n = 7) (2.1) The following reactions are also important: CH4 + H2O  '  CO + 3 H2    ∆H°298 =  206 kJ/mol    (2.2) CO + H2O  '  CO2 + H2     ∆H°298 = - 41 kJ/mol   (2.3) CH4 + 2H2O  '  CO2 + 4H2    ∆H°298 = 165 kJ/mol   (2.4) Although methane is not present in the feed, it immediately starts to appear in the system due to the methanation reactions (reverse of reactions (2.2) and (2.4)), once H2, CO and CO2 appear in the system by reactions (2.1) and (2.3). Hence the process proceeds in an identical manner to steam reforming of methane2,3. The methane yield decreases with increasing temperature due to the endothermicity of the steam reforming reaction of methane.  As a result, H2 yield continues to increase.  If this H2 can be selectively removed from the system as in steam methane reforming4,5, the CH4 yield will decrease further, due to forward equilibrium shift of reactions (2.2), (2.3) and (2.4). Thus,   † A version of this chapter has been submitted for publication: Rakib, M.A., Grace, J.R., Lim, C.J., Elnashaie, S.S.E.H., Epp, M., Gulamhusein, A., Boyd, T., and Keelan, A., Pilot Scale Experimental Setup for Hydrogen Production from Higher Hydrocarbons: Safety Considerations and Implementation (2010).  45  (a) Hydrogen generation by steam reforming of higher hydrocarbon feedstock is constrained by the equilibrium of steam reforming of methane, once CO and H2 are produced. Continual removal of hydrogen from the reaction stream by hydrogen permselective membranes can minimize this limitation, and maximize the production of hydrogen. (b) To overcome very low catalyst effectiveness factors due to diffusional limitations in traditional fixed bed reformers, fine catalyst powders can be used in the fluidization mode. Based on these two concepts, a fluidized bed membrane reactor was modeled6, built, commissioned and operated using model compounds for proof-of-concept. This paper briefly describes the process flow layout and the safety considerations that were implemented to ensure personnel safety and control leading to efficient experimentation.  2.2 Pilot Plant Layout The fluidized bed membrane reactor (FBMR), which is at the heart of the process, is shown in Figure 2.1. Figures 2.2(a) through (d) show the P&ID for the pilot plant layout, segmented into nodes for a HAZOP study. Based on this, the pilot plant layout can be visualized to consist of: (a) The feeding section covered in Nodes 1 through 6 in the P&ID in the HAZOP worksheet, (b) The main reactor, the FBMR is covered as Node 7 in the P&ID, (c) The pure hydrogen product section is covered as Node 10 in the P&ID, (d) The reformer off-gas section is shown in Node 8, and (e) Sample gas analysis is included in Node 9. 2.2.1 Feeding section For outdoor gaseous hydrogen systems, the minimum clearance distance from installations of flammable and combustible liquids above ground (0 to 3785 liters) is 3.1 m 7,8.  Alternatively, a minimum of one hour fire-rating capable barrier needs to be installed between these two storage sections to limit direct propagation of flames should a fire start in either of the sections. The heptane tank storage capacity is 28 liters, and the propane tank capacity (FX size) is 108.5 liters. In our case, a two-hour-fire-rating capable barrier was installed between the liquid hydrocarbons section (propane and heptane) and the hydrogen cylinders storage section. Hydrocarbon storage and feeding section: The hydrogen storage facilities are shown as part of the P&ID in Figure 2.2(b). The reactor can be fed with three categories of hydrocarbons: C1 represented by either natural gas or pure methane, C3 represented by propane, or C7 represented  46  by heptane. The liquid hydrocarbon feeding system is shown as Node 4, while the natural gas or methane feeding system appears as Node 3. Sulphur compounds are removed from the natural gas by passing through a bed of desulfurizer sorbent Sulfusorb-8, a CuO-impregnated activated charcoal product from Calgon Carbon Corporation. When pure methane (Grade 2.0) is fed as a substitute for mains natural gas, the desulfurizer is bypassed. A helium (Grade 4.5) size T cylinder with a delivery pressure set to 15 barg, is connected to the top of the propane (Grade 2.0) cylinder (FX size), so that it creates a helium “pad” over the liquid propane. A gas regulator regulates the helium pressures up to 17 barg. With this helium padding system, the head pressure in the propane tank is maintained at a constant value. For heptane supply, a 203 mm ID x 864 mm deep vertically positioned cylindrical tank is used to store heptane. Under the NFPA 30 9 classification, heptane is classified as a Class IB liquid, and has a fire rating of 3 as per NFPA 704 10. When the FBMR is fed with heptane, the vapor headspace is pressurized with helium to provide a pressure head to push the heptane into the feed system. Proper care also needs to be taken to purge the vapor headspace in the heptane tank. Residual air in the headspace may create an explosive mixture with heptane vapor, and an explosion could be triggered by electrostatic charge buildup. The whole liquid hydrocarbon feeding system is therefore fully grounded. Since the FBMR is fed with one hydrocarbon at a time, a common liquid mass flow controller is used to meter either propane or heptane into the feed system. A separate gas mass flow controller is used for natural gas or methane. The propane tank as well as the heptane tank has its own pressure rating. To avoid injury by accidentally over pressurizing either tank, the pressurizing helium line is fitted with a pressure relief valve set at a maximum pressure of 15 barg. A 3-way selector valve is installed to select either a propane or heptane feed. Steam feeding system: Water is accurately metered and pumped through a vaporizer, as shown in Figure 2.2(b). The liquid water pumping and metering system is indicated in Node 5, whereas the feed mixing and vaporizer section is included under Node 6. A low-level indicator switch is installed in the water tank to safeguard against dry-out of the tank as water shortage would lead to heavy deactivation of the catalyst so that the hydrocarbon feed would then need to be immediately stopped. Control logic described in detail below accomplishes this by closing the corresponding hydrocarbon solenoid valve. Metered natural gas or pure methane is pre-mixed with the pumped water before passing through the vaporizer. Metered propane or heptane is  47  introduced as a liquid feed into superheated steam after the vaporizer, with the feed line being completely heat-traced to the entrance into the FBMR. The vaporizer needs to evaporate and superheat water at temperatures as high as 600°C, while satisfying the material of construction temperature constraints. This is achieved by creating control logic which turns off the vaporizer heaters if the vaporizer skin temperature should reach its limit. Nitrogen supply system: The nitrogen supply system is indicated as Node 1 in the P&ID section in Figure 2.2(a). Nitrogen from the bank of cylinders at the storage rack is split into two streams: one is set at a maximum pressure of 10 barg for purging the reactor or leak testing of the reactor and other items at the pressure boundary, whereas the other is a low pressure one to purge the permeate line. Hydrogen supply system: The Ni-based catalysts may experience oxidation of the active phase to NiO with time during storage. Hence hydrogen is used to reduce the catalysts to the active Ni phase, prior to introducing the hydrocarbon feeds. The hydrogen supply system is indicated as Node 2 in the P&ID section in Figure 2.2(a). Hydrogen can also be used to test the permeability of the installed membranes as a function of time and operating conditions. The nitrogen and hydrogen manifolds are each fitted with relief valves. In addition, the pressure sensor upstream of the vaporizer inlet, as well as the one before the FBMR feed inlet, is coupled with control logic to shut off all the solenoid valves if the upper pressure limits should be breached. Hydrogen, methane, and nitrogen are procured from Praxair Inc. Storage, securing and handling of such compressed gases are governed by NFPA 5511. Table 2.1 lists the consumable gases and liquids used, together with their purity specifications. 2.2.2 Fluidized Bed Membrane Reactor (FBMR) The FBMR, the heart of the process, is shown in Figure 2.1, and depicted schematically as Node 7 in the P&ID. Based on preliminary modeling6, the FBMR was conceptualized as a vertically positioned cylindrical pressure shell, with vertical slits alternating in sides along the height of the shell, to accommodate six membrane panels. These panels are arranged vertically one above the other, each passing through the centerline of the reactor shell dividing the cross-section into two communicating sections. The mechanical design was executed by Jenmar Concepts Inc. to conform to the Pressure Vessel code (ASME Section VIII, Division 1). The design was certified by British Columbia Safety Authority (BCSA). Fabrication was executed by Axton Inc., which specializes in fabrication of industrial pressure vessels, tanks, and heat exchangers.  48  FBMR material of construction: SS304 has outstanding weldability, where all standard welding techniques can be used, although machinability of SS304 is lower than for most carbon steels. SS316 has virtually the same mechanical, physical and fabrication characteristics as 304 with better corrosion resistance, particularly to pitting corrosion in warm chloride environments, and has excellent corrosion resistance in a wide range of media. Low carbon “L” grades are used where high temperature exposure occurs, including welding of medium or heavy sections. The low carbon assists in delaying or preventing grain boundary carbide precipitation (often referred to as sensitization) which can result in intergranular corrosion in corrosive service environments. Stainless steel with higher carbon content (>0.04% C) as in the “H” grades, increases the strength, particularly at temperatures above ~500 oC. Long-term creep strength is also higher. Among the materials compatible enough for the design conditions (621°C and 10.3 barg), SS316H and SS304H would be suitable choices. However, because SS316H was not available, SS304H was chosen as the material of construction, selected for its strength and corrosion resistance at high temperature. Although SS304H has excellent strength and corrosion resistance for the design temperatures, operation at these high temperatures will cause precipitation of inter-granular carbides causing a reduction in the ability of the material to resist corrosion. However in applications where no liquids (condensation or steam) are present this is usually not an issue.  During operation, the process stream contains superheated steam.  On shutdown, the pressure vessel is purged to minimize water condensation. Hydrogen can seriously affect the properties of materials by several mechanisms: (a) Hydrogen embrittlement (HE) occurs by ingress of hydrogen into metallic materials, seriously affecting the ductility and load-bearing capacity, sometimes followed by catastrophic brittle fracture at stresses below the yield stress. Hydrogen embrittlement is seen with carbon and low alloy steels, ferritic and martenstitic stainless steels, and duplex stainless steels. Although it is normally not a problem with austenitic stainless steels12-14, metastable austenitic stainless steels (e.g. 304 or 316 type) may experience hydrogen embrittlement whereas stable austenitic stainless steels like 310S are not affected by hydrogen environment15,16. However, embrittlement effects are generally confined to near-ambient temperatures, with a maximum effect at room temperatures13. In our case, hydrogen is not introduced into the FBMR below 400°C for catalyst reduction, and hydrocarbons are only fed to the FBMR for hydrogen generation above 475°C. Hence, under normal experimental conditions, hydrogen embrittlement is not expected to be important for the FBMR. This is also ensured by shutting off the hydrogen delivery solenoid  49  valve if the FBMR temperature drops below 300°C, the primary objective of which is to prevent swelling of the Pd-Ag membranes. (b) Hydrogen attack, also known as high temperature hydrogen attack (HTHA), occurs above ~220°C  as a consequence of hydrogen ingress into steel14. Dissolved hydrogen attacks iron carbide (Fe3C), generating methane gas. Since the methane molecule is too large to diffuse within the solid alloy, the gas stays trapped along the grain boundaries. Gas build-up can tear the grain boundaries apart, ultimately leading to severe cracks17,18. However, austenitic stainless steels are generally unaffected by hydrogen attack, and hence it is not a concern for the FBMR14. Carbonyl corrosion is a concern is industrial systems encountering carbon monoxide, e.g. in manufacture of methanol. However for the FBMR operation, the probability of this is minimal, since: i) At operating temperatures of interest (<600°C), CO production is normally very low (about 2% of dry gas composition). ii) Carbonyl corrosion commences only at pressures above 100 bars, and is a serious issue only between temperatures of 150°C to 200°C19-21. The operating temperatures of this vessel fall into the creep range for the materials used. For this reason, it is required that the entire pressure vessel assembly be uniformly insulated to minimize temperature gradients to avoid thermal stresses and thermal fatigue from temperature cycling.  In addition, heating of the pressure vessel at start-up is to be strictly controlled, with a maximum heat up rate of 5°C/min.  A similar gradual cool down rate on shutdown is also implemented. From the allowable stress characteristics for SS304H as a function of temperature the reactor pressure rating would drop dramatically if its temperature were to exceed 600°C, as shown in Figure 2.3(a). This is reflected in the MAWP (Maximum Allowable Working Pressure) rating of the pressure vessel for different operating temperatures. As the pressure vessel design calculations are based on the yield and allowable stresses under a particular combination of temperature and pressure, the safest or ideal operating conditions would be to work within the maximum ranges of both the temperature and MAWP. However, if the test conditions demand otherwise, the MAWP can be flexible depending on the operating temperature, as depicted in Figure 2.3(b). Being a non-standard flange, the six lateral rectangular flanges were the main challenge in the mechanical design of the pressure vessel. Forged bars of SA182 SS304H were used for the lateral flanges and SA240 SS304H for the blind flange covers. Forged bars are preferred over  50  plates to overcome the tendency to delaminate under stress. Rectangular slots 241.3 mm x 22.2 mm are provided for insertion of the rectangular membrane panels. These panels are supported onto the blind rectangular cover flanges. Reactor shell weld-on ports: Seven weld-on ports were provided on the reactor shell along the height, to allow future additional feed ports, or pressure/ sample gas taps. These holes could also allow withdrawal of catalyst samples at different heights during down periods. Near the bottom, and above the distributor, an additional port, angled downwards at 45° to the vertical, was provided which could function as a catalyst recirculation port if the reformer were to be operated in the future under fast fluidization conditions. Lateral flange weld-on ports: Fittings are installed for one thermocouple to measure bed temperature, one pressure tap, two sample gas outlets, and one tap for removal of hydrogen permeated through the palladium surface into the membrane panels. This is shown in Figure 2.4(a). Membrane panels: Pd membranes are infinitely selective to permeation of hydrogen due to the unique solution-diffusion mechanism of permeation22,23. Diffusion depends on the difference of the square roots of partial pressures on the two sides according to Sieverts’ law when hydrogen diffusion is the rate determining step24. Pd membranes are susceptible to hydrogen embrittlement due to phase transition at temperatures around 300°C, resulting in expansion of the metal lattice5,25-27. Pd is often alloyed with other metals like Ag, Cu and Ru to improve mechanical stability, resistance to hydrogen embrittlement and hydrogen permeation flux. A Pd77Ag23 alloy, for example, leads to a 70% gain in hydrogen flux compared with pure Pd5,28. Packing too many vertical surfaces with small gaps between adjacent ones in a fluidized bed decreases the quality of fluidization, e.g., may cause defluidization due to solids bridging in the gaps between the surfaces29.  Thus, although a bundle of tubular membranes could provide the maximum membrane area per unit volume of reactor, in practice this cannot be utilized in fluidized beds. Moreover, in terms of fabrication, numerous small diameter tubes connected to a header or manifolding poses challenges of sealing. In addition, membrane panels produced by bonding membrane foils onto the porous support provides a more robust performance for operating temperatures about 500°C or higher in a fluidized bed environment, compared to membrane panels prepared by coating methods. Planar membrane panels also provide better and easier sealing compared to tubular panels. So, flat planar membrane panels were used for withdrawing hydrogen from the reaction environment. Double-sided Pd-Ag membrane panels are inserted through vertical slits on the wall of the FBMR reactor. These panels, shown in Figure 2.4(b)  51  were manufactured and supplied by Membrane Reactor Technologies30. The overall dimensions of the panels are 231.8 mm x 73.0 mm x 6.35 mm thick. Accounting for weld space and bonding space, the active area of each membrane is 206.4 mm x 50.8 mm on each side of the membrane panel. The flux of hydrogen through the membranes can be increased by reducing the hydrogen partial pressure on the permeate side. This can be achieved either by using sweep gas or by evacuating the permeate side. In our experiments, a vacuum pump downstream of the hydrogen permeate manifold was employed. FBMR bed cross-section: From reactor modeling prior to sizing, it was clear that the membrane permeation area per unit volume of reactor needs to be maximized in order to optimize the pure hydrogen yield. In practice, an upper limit to this is set by the tendency of the catalyst particles to form an immobile bridge in the gap between the membrane panel and the reactor wall if this gap is too small. To avoid defluidization in this gap and thereby avoid gas channeling, the width of this gap should be greater than about 30dp (where dp is the mean particle diameter)29. Given this constraint, the need to maximize permeation area per unit volume indicates that for the rectangular flat membrane panels, the reactor cross-sectional area should be rectangular. The circular cross-section of a 73.6 mm ID SS304H pipe was converted to rectangular by using reinforced ceramic cement. The ceramic blockers were intermittently grooved laterally on the surface to accommodate rope gaskets which prevent vertical channeling of gas. Distributor: The gas distributor is doughnut-shaped, with six equally-spaced holes drilled on the inner side. These holes point radially inwards and downwards at an angle of 45° to the vertical, to reduce back-sifting of catalyst particles into the windbox, located inside and at the bottom of the FBMR. The FBMR feed line is welded onto the bottom of the distributor housing, passing through the bottom head cover via sealed fittings. This design allows spent catalyst particles from the FBMR to be discharged by unfastening a cap through a simple catalyst drain in the bottom head cover, without requiring complete disassembly of the bottom head. A similar concept was used by Wang et al.31, involving a non-conventional manifold distributor with holes though the wall around the vessel. End flanges and Rupture disk: ASME SA182F SS304H weld-neck flanges are welded onto the reactor shell ends, with blind covers fabricated from the same material. The inlet (bottom) cover is fitted with a catalyst drain, a feed gas inlet, and four heater tubes. Fittings at the outlet (top) allow catalyst filling and gas to leave the reactor through a sintered metal filter. The catalyst filling port is also connected to a rupture disc to protect the vessel from damage due to over-pressurization.  A rupture disc from Fike Canada was installed to avoid over pressurization  52  of the reactor. This rupture disc, which also holds an ASME Code, Section VIII, Div 1, UD/CRN certification, is a 25.4 mm AXIUS style reverse bulged, non-fragmenting rupture disc, made from SS316 seal with a seat ring on the vent side. The stamped pressure rating for the FBMR pressure vessel is 10.3 barg at 621°C. Reformer off-gas filter: A sintered metal filter from Mott Corporation, with the porous part consisting of a 152.4 mm long x 12.7 mm OD tube and a wall thickness 2 mm, captured the catalyst fines from the reformer off-gas. The porous sintered metal filter is SS316L, with a media grade of 40. This filter was inside the reactor, with the off-gas leaving through a 6.35 mm line. Probe filters: The probe filters each contained a 6.35 mm sintered Hastelloy-C276 disc, 3.18 mm thick, spot-welded onto a 6.35 mm SS316 tube. Hastelloy-C276 is ideal up to temperatures of ~540°C under reducing conditions. These filter lines protect the gas sample analyzer, as well as the pressure sensor, against damage due to fine particles. The pressure sensor filter to be as thin as possible to minimize damping of pressure fluctuations. Reformer heaters: 1524 mm long, 3.18 mm diameter narrow cable heaters, sealed by weld-on Conax fittings, provide the primary heat needed to bring the reactor up to ~600°C, and then to maintain the desired temperature. Depending on the space available, 88.9 mm ID band heaters, either 76.2 or 152.4 mm long, are installed in the semicircular spaces opposite each lateral flange. One strip heater per lateral flange is deployed on either of the vertical sides of these flanges. Due to their mounting mode, the cable heaters are termed internal heaters, whereas the band and strip heaters are called external heaters. When operated at maximum capacity, the total power supplied is 7.65 kW for the external heaters and 3.60 kW for the internal heaters. This power is drawn from a 220VAC supply. The location of the heaters, mounting mode and maximum power output of the heaters are listed in Table 2.2. During the operation, roughly 30 to 40% of the full power rating was required to maintain the FBMR at the operating temperature. To minimize thermal stresses, the output power of the heaters is adjusted such that the heating rate of all portions of the FBMR material does not exceed 5°C/min. Catalysts: A proprietary RK-212 catalyst from Haldor Topsoe A/S is employed. It is available as 7-holed black tableted pellets, in the pre-reduced form, with the size and shape optimized for a fixed bed catalyst loading with the required material strength and low pressure drop. In order to use them in a fluidized bed mode, the pellets are carefully crushed and sieved to different size cuts. The FBMR catalyst load consisted of an equal weight mixture of +150 µm -180 µm and +180 µm -212 µm size cuts giving a Sauter mean particle diameter of 179 µm. The catalyst  53  loading was just sufficient to immerse all the membrane panels, facilitating temperature uniformity and thereby preventing membrane failure and leakage. FBMR insulation: Proper insulation is required to minimize heat losses. A ceramic thermal insulator jacket of minimum thickness of 50 mm is wrapped around the outside of the vessel. The entire pressure vessel assembly is uniformly insulated to minimize temperature gradients to avoid thermal stresses and thermal fatigue from temperature cycling. 2.2.3 Hydrogen permeation section The hydrogen permeation section appears on the P&ID in Figure 2.2(d), referred to as Node 10. The flow of pure hydrogen permeating through each of the membranes is determined by hydrogen mass flowmeters, FMA1818 from Omega Instruments. To facilitate pure hydrogen permeation, thereby enhancing conversion of the hydrocarbons, the permeate streams are extracted by a powerful spark-proof hydrogen vacuum pump. Given the wide flammability range of hydrogen, the permeate section must be adequately purged with an inert gas like nitrogen before operating the vacuum pump, and the oxygen content is monitored during operation. 2.2.4 Reformer gas withdrawal section The reformer off-gas is cooled using a condenser; the condensed water is caught in a condensate trap, and the off-gas throttled through a pressure control valve before being vented. The reactor pressure is controlled by this pressure control valve. This section is denoted Node 8 in Figure 2.2(c). 2.2.5 Gas sampling To monitor reactor performance, gas samples were vented from specific locations in the reactor through a sample selection valve to a Varian CP-4900 micro-GC. The gas composition for steam reforming products from methane, propane and heptane is analyzed by a combination of GC columns as listed in Table 2.3, using micro-machined thermal conductivity detectors (TCD).  2.3 Objectives of Experimental Setup The experimental setup was designed and installed for a Proof-of-Concept study to investigate whether higher hydrocarbon feedstocks can be steam reformed at temperatures lower than 600°C, while minimizing catalyst deactivation and achieving high yield of pure hydrogen by means of hydrogen permselective Pd-alloy membranes. Variables include reformer temperature,  54  reformer pressure, pressure of the pure hydrogen permeate stream, steam-to-carbon molar ratio of the feed, and superficial gas velocity. Hydrocarbons studied are n-heptane (a model component for naphtha), propane (a key constituent of LPG), and methane (the major component of natural gas). Other subjects of special interest include membrane fouling, and catalyst deactivation. Table 2.4 lists the parameters which are varied, their ranges and how this is accomplished. Table 2.5 indicates the quantities which are directly monitored or calculated to evaluate the FBMR performance.  2.4 Toxicological and Safety Information of Materials Encountered Table 2.6 shows that special care has to be taken in handling pressurized systems containing substances like hydrogen, carbon monoxide, methane, propane or heptane. The laboratory is equipped with mono-ammonium phosphate based dry powder fire extinguishers, one at each of the three laboratory entry points, and two adjacent to the experimental setup. A strong exhaust ventilation system at the roof of the laboratory prevents accumulation of gases such as hydrogen. Toxic substances: (a)  Carbon monoxide: Carbon monoxide is an asphyxiant and a highly toxic gas that combines with the hemoglobin of the blood, forming carboxyhaemoglobin, decreasing the delivery of oxygen to the tissues. (b) Nickel dust: Occupational exposure to nickel aerosols due to inhalation can result in development of asthma specific to nickel. Inhalation of nickel dust can cause chronic ailments as well as carcinogenic effects to the respiratory system. “Evidence for the carcinogenicity of nickel metal and other compounds is relatively weak or inconclusive, but insoluble dusts of nickel oxides, and soluble aerosols of nickel sulfate, nitrate, and chloride, have been implicated as potential carcinogens.”32. Nickel carbonyl formation due to the interaction of nickel and CO Ni + 4CO → Ni(CO)4          (2.5) is unlikely to occur at the operating conditions of the reformer, but must be considered, given the toxicity of the material, not only due to its CO content which itself is toxic, but also due to nickel which can be released into the body33,34. Flammable substances: The flammability characteristics of species encountered in the operation of the system is given in Table 2.6. Although the auto-ignition temperature for hydrogen is much higher than for propane  55  or heptane, the minimum energy for ignition is much less, and it is hence more readily ignitable35. (a) Hydrogen burns with an invisible flame and may also form a fireball. It has a wide flammability range. (b) Carbon monoxide also has a wide flammability range, but fire hazard caused by CO is less probable for this installation, since the dry gas composition of the synthesis gas mixture produced is expected to be about 1%. However, as mentioned above, CO build-up must be monitored due to its highly toxic nature. (c) Hydrocarbons: Methane, propane and heptane have narrower flammability ranges than hydrogen, but a number of other factors are also important: (i)  The LEL of propane and heptane are considerably lower than for hydrogen and methane; (ii)  The diffusivity of hydrogen is much greater than for propane, methane or heptane; and (iii) Propane and heptane vapor are significantly denser than air.. However, for a confined space with inadequate ventilation, hydrogen and methane would form combustible mixtures more rapidly than propane and heptane. In such a case, hydrogen is expected to form combustible mixtures more rapidly than methane since hydrogen has higher buoyant velocity and slightly lower flammable limit35. Thus, a fire hazard is likely to be in the orders of hydrogen, methane, propane and heptane, and to persist in the reverse order35.  2.5 Safety Considerations during FBMR Operation The laboratory setup comes under the scope of NFPA 45 36. It is adequately equipped with portable fire extinguishers. There are three exit doors installed to swing in the direction of exit.  2.5.1 Temperature control Proper temperature control is needed to: (a) Safeguard the pressurized FBMR and the feed and product lines against mechanical failure. (b) Minimize catalyst deactivation by enhanced carbon formation or sintering at temperatures exceeding the recommended range. (c) Prevent membrane failure causing leakage at higher temperatures. (d) Avoid the presence of hydrogen in the vicinity of the Pd-Ag membranes below the recommended temperature of 300°C to prevent membrane swelling5,25,27.  56  (e) Avoid feeding steam to the FBMR at temperatures below 350°C. (f) Avoid use of the hydrogen vacuum pump above recommended temperatures. A cause and effect matrix of the control logic is used to specify suitable upper and lower temperatures for important locations in the process layout, thereby cutting electrical power to relevant heaters or pumps. Thermocouples were inserted close to the center of the membrane panels, through the lateral rectangular flange covers, and also at the level of the ROG filter. In addition, thermocouples were added below each external heater to limit the skin temperatures, which are likely to be at local maxima directly beneath the heaters, by cutting off the electrical power if the limit is breached. 2.5.2 Pressure control Reliable pressure control is required to safeguard the FBMR vessel and its components against mechanical failure. For example, excessive pressure in the permeate line could cause the membrane foils to detach from their supports. A rupture disc on the FBMR vessel and pressure relief valves in each of the feed lines (hydrogen, nitrogen, individual hydrocarbons, water pump discharge outlet and permeate line) were installed. Indirect mechanisms were also incorporated in the form of pressure sensors with input to the PLC logic capable to shut solenoid valves on the feed lines or open the pneumatically-controlled FBMR pressure control valve. Under normal operating conditions, the pressure control valve is designed to ensure proper pressure regulation. The absolute pressure sensor in the freeboard, and differential pressure transducers on the lateral rectangular flanges indicate the quality of fluidization inside the bed37. The differential pressure transducer outputs were recorded only, in addition to online graphical visualization so that corrective measures could be taken. The absolute pressure transducers data were input to the PLC so that the FBMR pressure can be regulated. This is necessary, for example, to avoid channeling, where the bed would operate as a packed bed, with poor gas-solid contacting. The arrangement of pressure transducers connected directly to the FBMR is shown in Figure 2.5. 2.5.3 Prevention of backflow Flow reversals could conceivably occur due to unexpected process parameters deviation, or when there is a sudden blockage of the FBMR off-gas filter due to catalyst cake formation. Such flow reversals are prevented by check valves at appropriate locations, as outlined in the HAZOP worksheet.  57  2.5.4 Hazardous gas leakage Leakage of dangerous gas such as carbon monoxide, hydrogen, methane, propane or heptane into the workspace and/or around the high temperature equipments must be prevented. Flammable mixtures of air and hydrogen can auto-ignite if in contact with hot surfaces above 500°C 7, and may be accompanied by explosion or toxic gas poisoning. The interfaces between the flanges and covers for the lateral, top and bottom flanges were sealed using SS316-reinforced graphoil gaskets. The flange nut and bolts were tightened using a torque-wrench, taking care not to exceed the specified design torques. A CO gas detector, TS400 and a combustible gas detector, S4000C, both from General Monitors, were installed close to the reactor. Output signals from these detectors are continuously logged during operation. If the specified safe limits should be exceeded at any moment, an emergency shutdown would be triggered. A hydrogen sensor (MSA Orion multigas detector) was also deployed periodically to identify any leaks. In addition, a CO monitor was located near the top of the FBMR near the reactor exit, and another adjacent to the sampling system to the micro-GC. 2.5.5 Air ingress into hydrogen permeate section Hydrogen has a wide explosive range in mixtures with air or pure oxygen. It is essential that the hydrogen suction system be leak-proof. Leak testing by pressurization was performed for each of the six permeate lines individually and for the entire permeate manifold system by determining the pressure holding capability. The permeate lines, manifold system, and vacuum pump were also subjected to vacuum, and left overnight to check the vacuum loss rate due to air ingress. As per the Canadian Hydrogen Installation Code7 under Article 7.7 for compressor requirement, if hydrogen comes from a sub-atmospheric pressure source, the oxygen content of the hydrogen needs to be continuously monitored, and the compressor needs to be shut down should the oxygen content exceed 1% by volume. This will ensure that the oxygen concentration is too low to sustain combustion should the concentration of hydrogen be in the flammable range. In our case, since the feed to the hydrogen vacuum pump comes from membrane panels under vacuum, an oxygen sensor from Teledyne Analytical Instruments was used to monitor the oxygen content of the permeate hydrogen product. A purging line was installed upstream of the hydrogen pump to remove any oxygen buildup.   58  2.5.6 Gas sampling Gas samples could contain high moisture content (e.g. 30-50%), depending on the operating conditions and hydrocarbon conversion. High moisture content could damage some of the columns in the micro-GC. Hence precautions were implemented to avoid this: (a) Installation of a condenser which could be emptied periodically. (b) Installation of a genie filter which does not allow liquid water to pass through it, but allows passage of every other component, including uncondensed steam and condensed heptane. (c) Installation of a micro-gasifier before sample injection, which allows controlled heated pressure reduction of sample gases originating in the high pressure FBMR. (c) Maintaining the sample injection and column temperatures ≥110°C to prevent condensation after sample injection. (d) Adequate back flushing time settings for each of the GC columns to limit steam flow. (e) The sample gas bypass line is connected to the venting line to prevent hazardous gas from reaching the workspace. 2.5.7 Trips and emergency shutdown Some operating conditions deviations could have hazardous consequences. For example, excessive temperatures or pressure limitations of the FBMR pressure vessel could cause the vessel to fail mechanically, leading to explosion and toxic gas release. The system was designed so that pressures exceeding specified values trigger shut-off of all feed streams and opening of the pressure control valve completely. Excessive temperatures cause shutdown of the corresponding heater. In both cases, user input is required to restore the system to regular operation. Some situations lead to immediate emergency shutdown. Toxic or combustible gas release in the surrounding atmosphere indicates a leak which could endanger safety. Similarly low instrument air means that proper regulation of the FBMR pressure could be disrupted. In such situations, the priority is to check the system integrity, and an emergency shutdown is thus actuated by the Cause & Effect matrix of the PLC. As an additional safeguard, an Emergency Shutdown button is within reach of the operators. Pressing this button immediately cuts electrical power to the instruments and opens the reactor pressure control valve.    59  2.5.8 Catalyst handling The commercial RK-212 catalyst pellets (pre-reduced nickel supported on alumina) were crushed and sieved to obtain a narrow size distribution of Sauter mean diameter 179 µm. When stored under atmospheric conditions, the catalyst particles contain roughly 15 to 20% by weight Ni or NiO. NiO is toxicologically classified as a potential carcinogen. Thus all the relevant catalyst handling steps of crushing, sieving, and loading into the FBMR require special care to avoid exposure to personnel. 2.5.9 Insulation Proper insulation is required to minimize heat losses, and to ensure personnel safety by avoiding exposure to hot surfaces. A refractory ceramic fiber product, Cerablanket, supplied by Thermal Ceramics Inc., insulates the reactor. The MSDS for this material does not indicate any respiratory disease attributed to exposure to this material, while many health agencies like International Agency for Research on Cancer (IARC), Canadian Environmental Protection Agency (CEPA), and American Conference of Governmental Industrial Hygienists (ACGIH) classify it as possible human carcinogen38. A particulate face mask was therefore used while handling and installing the insulation. While heating up the system for the first time with new insulation, significant amounts of volatiles are emitted which could be harmful to health, and also is a potential fire hazard. Hence, powerful ventilation and gradual temperature increase were implemented during initial reactor operation. 2.5.10 Electrical safety All power distribution systems, heaters and other electrical appliances were properly grounded and checked for ground faults. Similar precautions were taken to eliminate electrostatic charge buildup in the combustible substances storage area. 2.5.11 Safety apparel Safety apparel required for regular operation of the reactor system include: (a) Regular lab coats (b) Latex gloves (c) Goggles (d) Full-toed shoes Additional precautions are required while working with particulates, e.g. while handling the catalyst and insulation:  60  (e) Full head-to-toe clothing. (f) Particulate masks. In addition, appropriate gloves were required to prevent frostbite when dealing with propane line commissioning. 2.5.12 Safe working habits The following additional measures promote safety: (a) Clearly written and distinctly visible “Emergency Shutdown Procedure” (ESD) instructions. (b) No food or drinks in the laboratory work area. (c) Clear visible locations of emergency showers, eyewash stations, first aid and firefighting equipment. (d) Material safety data sheets (MSDS) are conspicuously located, and referred to before working with any new substance. (e) Written working plans, not only to ensure well-organized and efficient experimentation, but also to think through all procedures and steps which could affect safety. (f) Prohibition of working alone when the FBMR is operating. 2.5.13 Process control for FBMR operation The FBMR process control system was configured by Membrane Reactor Technologies in their control (Delta V) system. Safety issues discussed in previous sections were taken into account during the Hazard and Operability studies as shown in Tables 2.7 and 2.8, leading to the Cause and Effect matrix shown in Table 2.9. To ensure safe operation, a high level HMI ladder logic is used to specify PLC actions based on the matrix.  2.6   Conclusions A novel fluidized bed membrane reactor system has been designed, built and commissioned at the University of British Columbia for a Proof-of-Concept study for steam reforming of higher hydrocarbons to produce pure hydrogen. Operation at high temperature and pressure, combined with toxic and combustible substances, required close attention to safety issues. Detail of the experimental setup and safety measures are summarized. A process control plan was developed based on a cause and effect matrix with respect to the PLC.  61  Table 2.1: Gases and liquids in FBMR steam reforming process  Substance Chemical formula Molecular weight Supplier Purity/ Grade Methane CH4 16 Praxair 2.0 Propane C3H8 44 Praxair 2.5 Heptane C7H16 100 Sigma-Aldrich 99.5% Nitrogen N2 28 Praxair Industrial Hydrogen H2 2 Praxair Industrial Water / Steam H2O 18 UBC Distilled water  Table 2.2:  Heaters distribution for the FBMR  Heater ID Description Height (mm) Maximum power output (W) HT-CA-901 0 – 1524  900 HT-CA-902 0 – 1524 900 HT-CA-903 0 – 1524 900 HT-CA-904 4 cable heaters, one at each corner of the rectangular channel of the FBMR 0 – 1524 900  HT-BA-901 143 – 448 900 HT-BA-902 473 – 625 450 HT-BA-903 651 – 956 900 HT-BA-904 981 – 1133 450 HT-BA-905 1159 – 1464 900 HT-BA-906 1413 – 1718 900 HT-BA-907 1800 – 1952 450 HT-BA-908 1800 – 1952 450 HT-BA-909 2000 – 2152 450 HT-BA-910 The band heaters are mounted on the semi- circular spaces opposite to each lateral flange. 2050 – 2202 450  HT-ST-901 143 – 448 225 HT-ST-902 397 – 702 225 HT-ST-903 651 – 956 225 HT-ST-904 905 – 1210 225 HT-ST-905 1159 – 1464 225 HT-ST-906 The strip heaters are vertically mounted on the lateral flanges. 1413 – 1718 225 *Height is measured from distributor level Table 2.3: Micro-GC column information for product gas analysis  Channel  Column Description Carrier Gas Gases Analyzed Detection limits 1 10 m molsieve 5A with pre-column backflush Argon He, H2, O2, N2, CH4 and CO 10 – 100 ppm 2 10 m PPU with pre-column backflush Helium CO2, C2H4, C2H6, C2H2, H2S and COS 10 – 100 ppm 3 8 m Silica PLOT with pre-column backflush Helium C3 and C4 isomers 10 – 100 ppm 4 8 m CP-Sil 5 with no pre-column Helium C5 to C12 components 1 – 10 ppm   62   Table 2.4: Controlled parameters for FBMR steam reforming process  Parameter Range of parameters Monitoring instrument and variation mechanism Bed temperature 475 – 550°C Feed temperature, bed temperatures, and reformer off-gas temperature close to gas filter located in the freeboard are measured by K-type thermocouples. Temperatures are varied by changing the power supplied to the internal and external heaters. Reactor pressure 300 – 900 kPa Figure 5 includes a schematic of the differential and absolute pressure probe arrangements. This could be varied by a pneumatically-controlled Pressure Control Valve downstream of the FBMR in the off-gas vent line. Permeate pressure 25 – 101 kPa An absolute pressure transducer was located downstream of the hydrogen flow meters, and upstream of the hydrogen vacuum pump. This could be varied by changing the speed of the hydrogen vacuum pump. Steam-to-carbon molar feed ratio 4 – 6 Monitored based on flow rates of the steam and hydrocarbon feeds. This could be varied by changing the mass flow controller set points. Feed superficial velocity 6 – 9 cm/s Feed superficial velocity is calculated based on feed conditions and flow rates.   63  Table 2.5: Performance parameters for FBMR steam reforming process  Parameter Monitoring Instrument or Basis Bed gas composition Online injection to MicroGC model no. CP-4900 from Varian Inc. Exit gas composition Online injection to MicroGC model no. CP-4900 from Varian Inc. Permeate product flow rate Hydrogen flowmeters from Omega Instruments, model no. FMA-1818 Permeate product purity Syringe injection to MicroGC model no. CP-4900 from Varian Inc. Hydrocarbon conversion Carbon balance based on known feed flow rates, and measured sample gas compositions. Hydrogen yield Moles of pure hydrogen produced per mole of hydrocarbon fed. * Methane yield Moles of methane produced per mole of higher hydrocarbon fed. *  * Normally product yields would be defined based on the amount of a feed species consumed. However, for the higher hydrocarbons, the conversion is virtually 100%, so yields are defined based on the hydrocarbon fed.  Table 2.6: Flammability and safety information of some species encountered 35,36   Hydrogen Carbon Monoxide Methane Propane Heptane Auto-ignition temperature (°C) 560 620 595 470 222 Flame temperature in air (°C) 2210 2468 1950 1980 2000 Flash point Flammable Gas Flammable Gas Flammable Gas -104°C -1°C Flammability limits (vol% in air) 4 - 75  12.5 - 74.2  5 - 15  2.1 - 9.5  1.1 - 7 Minimum energy for ignition in air (MJ) 0.02 < 0.3 0.29 0.305 0.24 NFPA fire rating 4 4 4 4 3 Diffusion coefficient in air (cm2/s), 25°C 0.61  0.16 0.10 0.05 Gas/vapor density (Air = 1) 0.0696 0.968 0.55 1.55 3.5 TLV Simple Asphyxiant 25 ppm TWA Simple Asphyxiant 2500 ppm TWA 500 ppm TWA   64  Table 2.7: Nodes in P&ID and HAZOP worksheet  Nodes Description 1 - 6 Deals with feed delivery section. Consequences focus on causes that restrict the delivery, i.e. leaks, empty cylinders etc  7 - 8 Directly to the FBMR; mainly consequences of temperature and pressure deviation 9 - 10 Relate to the products sampling and analysis. Consequences focus on causes resulting from the unit operation itself.   65  Table 2.8: Hazards and Operability Study Worksheet  Hazards and Operability Study Process: Fluidized Bed Membrane Reactor for Steam Reforming of Higher Hydrocarbons Node and Description Parameter/Deviation Cause(s) Consequences Safeguards Pressure / Less Empty cylinder Purging unavailable Low pressure alarm for PT-101 Pressure / More Pressure regulator  failure Exceed system design pressure Relief valve PSV-010. High pressure trips PAHH-1001 and PAHH-1200 1 Nitrogen feed line from facility. Considers effects on FBMR purging and  startup. Flow / Reverse Higher downstream  pressure Mixing of streams: Poor reactor performance Check valves VCK-101, VCK- 107, VCK-1101 Pressure / Less Empty cylinder No safety issue but no catalyst reduction: Poor reactor performance Periodically check cylinder pressure Pressure / More Pressure regulator  failure Exceed system design pressure Relief valve PSV-301. High pressure trips PAHH-1001 and PAHH-1200 2 Hydrogen feed line  from facility. Considers effect on   start-up Flow / Reverse Higher downstream  pressure Mixing of streams: Poor reactor performance Check valves VCK-301, VCK-107 Pressure / Less Empty cylinder No safety issue but no reaction. Catalyst re-oxidation: Poor reactor performance Check valves VCK-101, VCK- 107, VCK-1101 Pressure / More Pressure regulator  failure Exceed system design pressure Relief valve PSV-403. High pressure trips PAHH-1001 and PAHH-1200 Flow / Less System fluctuation Catalyst re-oxidation; Poor reactor  performance Low steam to carbon mole ratio alarm FFDALL-401 Flow / More System fluctuation Catalyst deactivation High steam to carbon mole  ratio alarm FFDAHH-401 Flow / Reverse Higher downstream  pressure Mixing of streams: Poor reactor performance Check valves: VCK-405 3 Natural gas feed lines  from facility. Includes  desulfurizer. Composition/ Fluctuation Spent desulfurization sorbent Reforming catalyst poisoning Sample port V-413 for NG sulfur content analysis. Replace sorbent when saturated  66   Table 2.8:   Hazards and Operability Study Worksheet (….continued)  Node and Description Parameter/Deviation Cause(s) Consequences Safeguards Pressure / Less Empty helium cylinder Poor reactor performance (no reactor feed); Catalyst re-oxidation Check valves VCK-101, VCK- 107, VCK-1101 Pressure / More Helium pressure regulator (V-1201) failure Exceed propane tank/ helium tank design pressure Relief valve PSV-403 Level / More Over-filling of tanks Level / Less Empty tanks No safety issue but no reaction. Catalyst re-oxidation: Poor reactor performance  MI-001 scale for net propane in tank; LS-1203 and LS-1204 high / low heptane tank level alarms Flow / Less System fluctuation Poor reactor performance Low steam-carbon ratio alarms FFDALL-1204 for propane, FFDALL-1205 for heptane Flow / More System fluctuation Catalyst deactivation High steam-carbon ratio alarms FFDAHH-1204 for propane FFDAHH-1205 for heptane 4 Liquid hydrocarbon feed line.  Includes liquid propane and heptane and helium pressurization lines. Flow / Reverse Higher downstream pressure Mixing of streams: Poor reactor performance Check valve VCK-1202 Pressure / Less Pump malfunction Low water flow/  catalyst deactivation Alarms for low steam-carbon ratios for each hydrocarbon Pressure / More Pump malfunction Exceed system design  pressure Relief valve PSV-501. High pressure trips PAHH-1001 and PAHH-1200 Level / Less Empty tank No steam in feed. Coking in preheater Catalyst deactivation. Pump damage. Water tank low level alarm LS- 1202 Flow / Less MFC malfunction Catalyst deactivation Alarms for low steam-carbon ratios for each hydrocarbon Flow / More MFC malfunction Poor reactor performance Alarms for high steam-carbon ratios for each hydrocarbon 5 Water feed line. Flow / Reverse Higher downstream pressure Hydrocarbons enter water line Check valve VCK-501  67  Table 2.8:   Hazards and Operability Study Worksheet (….continued)  Node and Description Parameter/Deviation Cause(s) Consequences Safeguards Temperature / More Heaters controller malfunction High feed temperature, high preheater tube skin temperature High level alarms TAHH-1000 and TAHH-1001 6 Preheater System (HX-1001 to HX- 1051) Temperature / Less Heaters failure/ high water flow rate Poor reactor performance Low level alarm TALL-1001. Shuts off water pump to prevent liquid water from entering FBMR FBMR Pressure / More Clogged ROG filter, or feed flow fluctuations Exceeds design pressure at operating temperature. Physical injury High pressure alarms PAHH-1001 and PAHH-600 will shut off all feeds and open PCV-600 fully. Rupture disc E-RD-001 also protects against overpressurization. FBMR Temperature / More Heater Temperature controller malfunction Exceeds design temperature at operating pressure. Physical injury High temperature alarms for skin temperatures below any band and strip heater will turn off all external heaters 7 FBMR:  Includes FBMR vessel, rupture disc, catalyst filling line FBMR Temperature / Less Heaters failure/ high water flow rate Poor reactor performance Low level alarms TALL-601 to TALL-608. Shuts off water pump to prevent liquid water from entering FBMR 8 ROG line to vent Temperature / More High ROG temperature Damage to pressure control valve (PR-600) High temperature alarm TAHH- 640 Pressure / Less Blockage of sample line filters Inability to sample gases. No safety issues Clean filters on next reactor shutdown. 9 Sample lines. Pressure / More  No safety issues; Pressure won't be higher than FBMR Pressure / Less High suction from hydrogen pump No safety issues Pressure / More Membrane leak Poor reactor performance. High pressure alarm PAHH-703 Temperature / More Membrane leak Damage to vacuum pump,  Poor reactor performance High temperature alarm TAHH- 702 Temperature / Less   No safety issues Low temperature alarm TALL-702 Flow / More Membrane leak Poor reactor performance. Taken care of by high pressure alarm PAHH-703 Flow / Less   No safety issues 10 Permeate lines. Considers all 6 permeate lines and venting operation Composition/Fluctuation Membrane leak Poor reactor performance.  Occasional sampling  of permeates  68  Table 2.9: Cause & Effect matrix for actions by the PLC   1  2  3  5  6  7  8  9 1 1  1 2  1 3  1 4   T A G  #   H S - 5 0 1  P C V - 6 0 1  H S - 7 0 1  X V - 1 1 9  X V - 4 1 9  X V - 3 1 5  X V - 1 2 0 4  X V - 2 0 5  H X - 6 0 0  H X - 9 0 0  H X - 1 0 0 1  H X - 1 0 5 0   P & I D                          Legends: C = Closed O = Open FO = Fully Open SP = Stop ST= Start  Notes:  (1)  FALL-401 / 1204 effects not triggered until FI-401 / 1204 has initially surpassed FALL-401 / 1204 value E F F E C T  ( O U T P U T S )  S E R V I C E  D E S C R I P T I O N  W a t e r  P u m p   R e a c t o r  P r e s s u r e  C o n t r o l  V a l v e  H y d r o g e n  V a c u u m  P u m p  P r o c e s s  N 2  S o l e n o i d  P r o c e s s  N G  S o l e n o i d  S t a r t u p  H y d r o g e n  S o l e n o i d  P r o p a n e  /  H e p t a n e  S o l e n o i d  W a t e r  S o l e n o i d  R e a c t o r  I n t e r n a l  H e a t e r s  R e a c t o r  E x t e r n a l  H e a t e r s  P r o c e s s  F e e d  S u p e r  H e a t e r  F B M R  F e e d  H e a t  R o p e   CAUSE (INPUTS)  TAG # P&ID Sheet SERVICE DESCRIPTION             COMMENTS 1 ESD BUTTON    SP FO SP O C C C C SP SP SP SP ESD 2 AAHH-603 3 Toxic Gas Monitor SP FO SP O C C C C SP SP SP SP Triggers ESD 3 AAHH-605 3 Combustible Gas Monitor SP FO SP O C C C C SP SP SP SP Triggers ESD 4 FALL-1204 2 Low Propane / Heptane Flow       O     C            Note 1 below 5 FALL-401 2 Low Process NG Flow       O C                Note 1 below 6 FFDAHH-500 2 Steam to Carbon Mole Ratio SP             C 7 FFDALL-500 2 Steam to Carbon Mole Ratio       O C   C 8 HS-703 - off 4 Scroll Pump cooling fan off     SP  69  Table 2.9:   Cause & Effect Matrix for actions by the PLC (…. Continued)   H S - 5 0 1  P C V - 6 0 1  H S - 7 0 1  X V - 1 1 9  X V - 4 1 9  X V - 3 1 5  X V - 1 2 0 4  X V - 2 0 5  H X - 6 0 0  H X - 9 0 0  H X - 1 0 0 1  H X - 1 0 5 0   9 PAHH-1001 2 Feed line pressure SP FO SP C C C C C SP SP SP SP High P Trip 10 PAHH-1200 2 Vaporizer upstream Pressure SP FO SP C C C C C SP SP SP SP High P Trip 11 PAHH-501 2 High Water Pump Discharge SP             C 12 PAHH-600 3 Freeboard Absolute Pressure SP FO SP C C C C C SP SP SP SP High P Trip 13 PAHH-701 4 Permeate Header Pressure     SP 14 PAHH-703 4 H2 Vacuum Pump Dishcarge     SP 15 PALL-005 1 Low Instrument Air SP FO SP O C C C C SP SP SP SP Triggers ESD 16 PALL-701 4 Permeate Header Pressure     SP 17 PAHH-912 3 FBMR pressure above distribuor SP FO SP C C C C C SP SP SP SP High P Trip 18 TAHH-1000 2 High Vaporizer Skin Temperature                     SP 19 TAHH-1001 2 High Vaporizer Product Temperature                     SP SP 20 TAHH-601 3 FBMR Temperature (Flange 1)                 SP SP SP SP 21 TAHH-602 3 FBMR Temperature (Flange 2)                 SP SP SP SP 22 TAHH-603 3 FBMR Temperature (Flange 3)                 SP SP SP SP 23 TAHH-604 3 FBMR Temperature (Flange 4)                 SP SP SP SP 24 TAHH-605 3 FBMR Temperature (Flange 5)                 SP SP SP SP 25 TAHH-606 3 FBMR Temperature (Flange 6)                 SP SP SP SP 26 TAHH607 3 FBMR Temperature (Freeboard)                 SP SP SP SP 27 TAHH-608 3 FBMR Temperature (Above Distributor)                 SP SP SP SP Grouped as common high FBMR temp 28 TAHH-640 3 High ROG Temperature                 SP SP SP SP 29 TAHH-702 4 H2 Vacuum Pump Discharge     SP  70  Table 2.9:   Cause & Effect Matrix for actions by the PLC (…. Continued)   H S - 5 0 1  P C V - 6 0 1  H S - 7 0 1  X V - 1 1 9  X V - 4 1 9  X V - 3 1 5  X V - 1 2 0 4  X V - 2 0 5  H X - 6 0 0  H X - 9 0 0  H X - 1 0 0 1  H X - 1 0 5 0   34 TAHH-915 3 HT-BA901 Temperature                   SP 35 TAHH-916 3 HT-BA902 Temperature                   SP 36 TAHH-917 3 HT-BA903 Temperature                   SP 37 TAHH-918 3 HT-BA904 Temperature                   SP 38 TAHH-919 3 HT-BA905 Temperature                   SP 39 TAHH-920 3 HT-BA906 Temperature                   SP 40 TAHH-921 3 HT-BA907 Temperature                   SP 41 TAHH-922 3 HT-BA908 Temperature                   SP 42 TAHH-923 3 HT-BA-909 Temperature                   SP 43 TAHH-924 3 HT-BA-910 Temperature                   SP 44 TAHH-925 3 HT-ST901 Temperature                   SP 45 TAHH-926 3 HT-ST902 Temperature                   SP 46 TAHH-927 3 HT-ST903 Temperature                   SP 47 TAHH-928 3 HT-ST904 Temperature                   SP 48 TAHH-929 3 HT-ST905 Temperature                   SP 49 TAHH-930 3 HT-ST906 Temperature                   SP Grouped as common high external heaters temperature 50 TAHH-939 3 High Rupture Disc Temperature SP FO SP O C C C C SP SP SP SP ESD 51 TALL-1001 2 Low Vaporizer Product Temperature SP             C 52 TALL-601 3 FBMR Temperature (Flange 1) SP             C 53 TALL-602 3 FBMR Temperature (Flange 2) SP             C 54 TALL-603 3 FBMR Temperature (Flange 3) SP             C 55 TALL-604 3 FBMR Temperature (Flange 4) SP             C 56 TALL-605 3 FBMR Temperature (Flange 5) SP             C 57 TALL-606 3 FBMR Temperature (Flange 6) SP             C 58 TALL-607 3 FBMR Temperature (Freeboard) SP             C 59 TALL-608 3 FBMR Temperature (Above Distributor) SP             C Grouped as common low FBMR Temp  71    Figure 2.1: FBMR pressure vessel supported on a mobile stand    72  PR.301 CG-H2-001 XV.119 S FO PR.101 CG-N2-001 PSV-101 FICV-301 XV-315 S FC VCK.301 Sheet No. 4 FICV-101 L Sheet No. 2 VCK.107 Sheet No. 2 N2 to Reactor H2 to Reactor V.101 N2 Cylinder H2 Cylinder Purge N2 (for Permeate Line) PSV.301 PT 101 VCK.101 V.115 101 FI V-009 PR-002 V-015V-010 Vent Vent I I FCV 101 301 FIFCV 301 PI 002 N 2 fo r p ur gi ng  H 2 lin e V-109 V-309 V-311 Flammables Vent Sheet No. 1 Node 1 Node 2   Figure 2.2 (a): P&ID of pilot plant layout (Supplementary gas feeding section)   73   Reformer Feed W at er P-501 ST/SP V-523 CG-HE-001 V-1201b PSV-1201 Vent XV-1204 FC S 501 HS VCK-501 E-HP-001 E-PR-001 V-1202 V-1203 V-1204 N at ur al  G as 00 1MI 1001 TT 1000 TTH X- 10 01 Sheet No. 3 1204 FIC FICV-1204 H L 502 FIC H L FICV-502 Sheet No. 1 CG-NG-001 VCK-405 PR.401 PSV.403 Vent FC LLK.403 V.409 DS.401 Desulfurizer V.411 Hot Water NG Supply Sheet No. 1 H2 to Reactor PI 405PI 403 VCK-1202 VCK-1203 401 FIC FICV-401 XV-419 S V.420 V-415 E-WTR-001 PSV-501 501 PT I HHH L V-529 I 401 FFDI From FIC-502 H L HH I HH LLI 1001 TIC LL I 1204 FFDI V-1201 V-564 V.401 L 1201 PI 1202 PI 1203 PI 1203 LS 1204 LS V- 42 1 V- 42 3 1202 LS 1202 LSL 1204 LSL 1203 LSH I I HH LL H L H L 1200 PI HH HH H X- 10 50 1050 TIC 1001 PI XV-501 FC S I V-1205 VCK-1204 N2 to Reactor PI 501 PD-501 Drain Sheet No. 2 Node 3 Node 4 Node 5 Node 6   Figure 2.2 (b): P&ID of pilot plant layout (Steam and hydrocarbon feeding section)   74  Permeate 2 Permeate 4 Sheet No. 4 Reformer Feed 915 TT 602 TT 919 PTD Sample 2a Sample 2b Sheet No. 4 604 TT 916 PTD Sample 4a Sample 4b Permeate 6 Sheet No. 4 606 TT 917 PTD Sample 6a Sample 6b Permeate 1 Sheet No. 4 Permeate 3 Sheet No. 4 Permeate 5 Sheet No.4 Rupture Disc ROG Product 939 TTH ROG thru Filter 610 PT GC GC GC Sample 1a Sample 1b Sample 3a Sample 3b Sample 5a Sample 5b GC GC GC GC GC GC GC GC GC GCSample ROG E-RD-001 902 DP 608 TT 912 PTA V-045 Sheet No. 2 640 TT L I HH E-CHRP-001 PSV.600 To Vent E-WTR-002 Water Trap To Vent PR.600 600 PIC Instrument Air 901 DP FBMR 600 TIC AT 603 AT 605 I HH I HH HH Toxic Gas Sensor Combustibles Gas Sensor Gas Alarm System I HH HH I H L L H H H (Calculated) H I HH LL 600 EC 600 HS H ST/SP I HH 900 EC 900 HS ST/SP V-1211 Drain V-1210 HT-ST-901 HT-ST-902 HT-ST-903 HT-ST-905 HT-ST-905 HT-ST-906 HT-BA-901 HT-BA-902 HT-BA-903 HT-BA-904 HT-BA-905 HT-BA-906 HT-BA-908 HT-BA-907 HT-BA-909HT-BA-910 HT-CA-903 HT-CA-901 HT-CA-904 HT-CA-902 917 TT 919 TT 921 TT 916 TT 918 TT 920 TT 924 TT 922 TT 923 TT 930 TT 928 TT 926 TT 929 TT 927 TT 925 TT 914 TT 912 TT 913 TT 911 TT 605 TT 913 PTD 603 TT 915 PTD 601 TT 918 PTD 1001 PT 607 TT 914 PTD 600 PT Sheet No. 3 Node 7 Node 8 Vent   Figure 2.2 (c): P&ID of pilot plant layout (FBMR)   75   PSV-701 P-701 To Vent V-717 V-718 V-720 N2 Purge for H2 Line FT-H2-1101 Permeate 5 Permeate 6 FT-H2-1102 Permeate 3 Permeate 4 FT-H2-1103 Permeate 1 Permeate 2 Sheet No. 3 Sheet No. 3 Sheet No. 3 Sheet No. 3 Sheet No. 3 Sheet No. 3 To Vent Reformer Sample 1a Reformer Sample 1b Reformer Sample 2a Reformer Sample 2b Reformer Sample 3a Reformer Sample 3b Reformer Sample 4a Reformer Sample 4b Reformer Sample 5a Reformer Sample 5b Reformer Sample 6a Reformer Sample 6b V-1152 V-1153 V-1154 V-1155 V-1156 V-1157 V-1158 V-1159 V-1160 V-1161 V-1162 V-1163 Micro-GC Reformer Sample ROG V-1145 FI-701 V-1191 V-1192 V-1193 V-1194 V-1195 V-1199 V-1197 V-1198 V-1196 I HH LL H L 703 PT I HH LL I HH H 701 PIC ST/SP 701 PT H L Sheet No. 1 VCK.1101 M-703 M ST/SP 702 TT H I HH I Sample Condenser Cooling Water VCK-1102 701 HS 703 HS Sheet No. 4 Node 9 Node 10   Figure 2.2 (d): P&ID of pilot plant layout (Gas sampling and Permeate sections)     76  0 10 20 30 40 50 60 70 80 90 300 350 400 450 500 550 600 650 700 750 Temperature (°C) A llo w ab le  S tr es s (N /m m 2 )  Figure 2.3 (a): Strength of SA-240 grade 304H plate material as a function of temperature (Ref: ASME - International Boiler and Pressure Vessel Code – 2007 Edition)  77  300 350 400 450 500 550 600 650 10 11 12 13 14 15 16 Maximum Allowable Working Pressure (barg) O pe ra tin g Te m pe ra tu re  (° C )   Figure 2.3 (b): MAWP rating of the FBMR pressure vessel (provided by Jenmar Concepts Inc.)   78  Membrane panel Supporting flange coverGas sample Gas sample Pressure port Permeate line Thermocouple   Figure 2.4 (a): Instrumentation ports on a lateral flange cover, also showing a membrane panel installed  79  6. 35  m m  O D  S S tu be 231.78 mm 206.38 mm 73 .0 3 m m 6.35 mm Pd77Ag23 Membrane Foil Membrane Panel Frame 50 .8 0 m m   Figure 2.4 (b): Membrane panel dimensions  80   Feed Reformer Off Gas Opening 6 Above Distributor Opening 4 Opening 2 Freeboard PTD-919 PTD-916 PTD-917 PTD-914 Opening 1 Opening 3 Opening 5 PTD-913 PTD-915 Above Distributor PTD-918 PT-600 PT-1001 PT-610   Figure 2.5:2 Pressure transducers arrangement   81  2.7   References 1. Rostrup-Nielsen, J. R.; Rostrup-Nielsen, T., Large-scale hydrogen production. CatTech 2002, 6, 150-159. 2. Chen, Z.; Yan, Y.; Elnashaie, S. S. E. H., Modeling and optimization of a novel membrane reformer for higher hydrocarbons. AIChE Journal 2003, 49, 1250-1265. 3. Christensen, T. S., Adiabatic prereforming of hydrocarbons - An important step in syngas production. Applied Catalysis a-General 1996, 138, (2), 285-309. 4. Shirasaki, Y.; Tsuneki, T.; Ota, Y.; Yasuda, I.; Tachibana, S.; Nakajima, H.; Kobayashi, K., Development of membrane reformer system for highly efficient hydrogen production from natural gas. International Journal of Hydrogen Energy 2009, 34, (10), 4482-4487. 5. Shu, J.; Grandjean, B. P. A.; Vanneste, A.; Kaliaguine, S., Catalytic palladium-based membrane reactors - a review. Canadian Journal of Chemical Engineering 1991, 69, (5), 1036- 1060. 6. Rakib, M. A.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J.; Bolkan, Y. G., Kinetic simulation of a compact reactor system for hydrogen production by steam reforming of higher hydrocarbons. Canadian Journal of Chemical Engineering 2008, 86, (3), 403-412. 7. Canadian hydrogen installation code CAN/BNQ 1784-000/2007. National standard of Canada 2007. 8. Rivkin, C. H., NFPA’s hydrogen technologies code project. Process Safety Progress 2008, 27, (4), 328-335. 9. NFPA 30. Flammable and combustible liquids code. National Fire Protection Association, Inc.: Quincy, MA, 2008 Edition. 10. NFPA 704. Standard system for the identification of the hazards of materials for emergency response. National Fire Protection Association, Inc.: Quincy, MA, 2007 Edition. 11. NFPA 55. Compressed gases and cryogenic fluids code National Fire Protection Association, Inc.: Quincy, MA, 2010 Edition. 12. Herms, E.; Olive, J. M.; Puiggali, M., Hydrogen embrittlement of 316L type stainless steel. Materials Science and Engineering a-Structural Materials Properties Microstructure and Processing 1999, 272, (2), 279-283. 13. Robinson, S. L.; Somerday, B. P.; Moody, N. R., Hydrogen embrittlement of stainless steel. Proceedings of 11th International Conference on Fracture (Turin, Italy, March 20–25, 2005), 6.  82  14. Albright's Chemical Engineering Handbook. CRC Press, Taylor & Francis group: Boca Raton, 2009. 15. Han, G.; He, J.; Fukuyama, S.; Yokogawa, K., Effect of strain-induced martensite on hydrogen environment embrittlement of sensitized austenitic stainless steels at low temperatures. Acta Materialia 1998, 46, (13), 4559-4570. 16. Tsay, L. W.; Liua, Y. F.; Huang, R. T.; Kuo, R. C., The effect of sensitization on the hydrogen-enhanced fatigue crack growth of two austenitic stainless steels. Corrosion Science 2008, 50, (5), 1360-1367. 17. Sharp, W. R.; Mostert, R. J., Practical experience in the early detection and assessment of vessels with HTHA degradation. 4th Middle East NDT Conference and Exhibition (Bahrain, December 2007), 8. 18. RP 941 - Steels for hydrogen service at elevated temperatures and pressures in petroleum refineries and petrochemical plants. 7 ed.; American Petroleum Institute: Washington D.C., 2008. 19. Tsinman, A. I.; Gru, B. A., Carbonyl corrosion of steels in methanol synthesis. Chemical and Petroleum Engineering 1977, 13, (4), 331-332. 20. Savitskaya, O. P., Investigation of carbonyl corrosion in metals. Chemical and Petroleum Engineering 1965, 1, (8 ), 604-607. 21. Horst, J. M. A. v. d., Gas coil collapse due to iron carbonyl. Materialwissenschaft Und Werkstofftechnik 1975, 6, 55-58. 22. Paglieri, S. N.; Way, J. D., Innovations in palladium membrane research. Separation and Purification Methods 2002, 31, (1), 1-169. 23. Uemiya, S., State-of-the-art of supported metal membranes for gas separation. Separation and Purification Methods 1999, 28, (1), 51-85. 24. Sieverts, A.; Zapf, G., The solubility of deuterium and hydrogen in solid palladium. Zeitschrift für Physikalische Chemie 1935, 174, 359--364. 25. Okazaki, J.; Tanaka, D. A. P.; Tanco, M. A. L.; Wakui, Y.; Mizukami, F.; Suzuki, T. M., Hydrogen permeability study of the thin Pd–Ag alloy membranes in the temperature range across the α–β phase transition. Journal of Membrane Science 2006, 282, (1-2), 370-374. 26. Roa, F.; Block, M. J.; Way, J. D., The influence of alloy composition on the H2 flux of composite Pd-Cu membranes. Desalination 2002, 147, (1-3), 411-416.  83  27. Sjardin, M.; Damen, K. J.; Faaij, A. P. C., Techno-economic prospects of small-scale membrane reactors in a future hydrogen-fuelled transportation sector. Energy 2006, 31, (14), 2523-2555. 28. Adhikari, S.; Fernando, S., Hydrogen membrane separation techniques. Industrial & Engineering Chemistry Research 2006, 45, (3), 875-881. 29. Adris, A. E. M.; Grace, J. R., Characteristics of fluidized-bed membrane reactors: Scale- up and practical issues. Industrial & Engineering Chemistry Research 1997, 36, (11), 4549- 4556. 30. Li, A.; Grace, J. R.; Lim, C. J., Preparation of thin Pd-based composite membrane on planar metallic substrate - Part 1: Pre-treatment of porous stainless steel substrate. Journal of Membrane Science 2007, 298, (1-2), 175-181. 31. Wang, A. L. T.; Stubington, J. F.; Xu, J., Hydrodynamic performance of a novel design of pressurized fluidized bed combustor. Journal of Energy Resources Technology 2006, 28, 111- 117. 32. http://www.npi.gov.au/database/substance-info/profiles/62.html (Accessed 31 October 2009). 33. Brynestad, J., Iron and nickel carbonyl formation: Literature survey in steel pipes and its prevention. Oak Ridge National Laboratory: Oak Ridge, Tennessee, 1976. 34. http://en.wikipedia.org/wiki/Nickel_tetracarbonyl (Accessed 31 October 2009). 35. Hord, J., Is hydrogen a safe fuel? International Journal of Hydrogen Energy 1978, 3, (2), 157-176. 36. NFPA 45. Standard on fire protection for laboratories using chemicals. National Fire Protection Association, Inc.: Quincy, MA, 2004 Edition. 37. Bi, H. T., A critical review of the complex pressure fluctuation phenomenon in gas-solids fluidized beds. Chemical Engineering Science 2007, 62, (13), 3473-3493. 38. . Material Safety Data Sheet, Cerablanket, Thermal Ceramics Inc.      84   CHAPTER 3.      STEAM REFORMING OF PROPANE IN A FLUIDIZED BED MEMBRANE REACTOR FOR HYDROGEN PRODUCTION∗   3.1  Introduction 3.1.1  Background Hydrogen is an important feedstock in several industries, especially for making ammonia and in petrochemical and petroleum refining processes1-3. Stringent environmental regulations require increasing quantities of hydrogen for hydro-treating processes in oil refineries, especially as available crudes become heavier. The demand is likely to increase sharply in the future due to projected hydrogen demand by the automobile sector4,5. Many uses of hydrogen like fuel cell processes also put special demand on the purity of the hydrogen. Table 3.1 shows hydrogen content in liquid methane (for conditions similar to LNG storage), liquid propane (conditions similar to commercial LPG tanks) and n-heptane, a liquid under normal ambient conditions. While the hydrogen content is highest for methane, the volumetric hydrogen density is most favorable for higher hydrocarbons which are liquids at or near ambient conditions. Currently, the most favored feedstock for hydrogen production is natural gas due to its availability and advantageous price. In addition, compared with higher hydrocarbon feedstocks like LPG or naphtha, the challenges from the feedstock sweetening process and catalyst deactivation are much easier to handle with natural gas as feedstock.  However, higher hydrocarbons are preferred in many places, depending on local availability and local prices relative to natural gas. This is especially important in oil refineries where demand for hydrocarbon feedstock and products vary over time, so that the industry gains from flexibility of feedstock choice. Fuel cell applications for on-board hydrogen generation or distributed  ∗ A version of this chapter has been accepted for publication: Rakib, M.A., Grace, J.R., Lim, C.J., Elnashaie, S.S.E.H., and Ghiasi, B, Steam Reforming of Propane in a Fluidized Bed Membrane Reactor for Hydrogen Production, International Journal of Hydrogen Energy (2010)   85  hydrogen filling stations also demand that the feedstock have a high volumetric hydrogen density, preferably at atmospheric or near-ambient pressures. LPG or Liquefied Petroleum Gas can be liquefied under relatively low pressures and is an abundant feedstock from refinery operations. The advantage of LPG relative to heavier hydrocarbon feedstocks like naphtha or diesel as a source of hydrogen is that it is cleaner and contains a higher weight percent of hydrogen. Also, depending on seasonal demand, refinery operations often result in a surplus of different feedstocks. For example, LPG demand soars in winter due to increased heating requirements, whereas it is usually in surplus throughout the summer. LPG is a mixture of hydrocarbons, predominantly propane and n-butane, with its composition depending on the source, recovery processes, and season. It can be a mixture of either predominantly butane, or predominantly propane, with propylene and butylenes also present in small amounts. The most common LPG is predominantly propane. In the US and Canada, LPG is at least 95% propane. This paper deals with an experimental study of a novel steam reformer and its operation with propane as the feedstock. There are various methods for generating hydrogen from propane: Propane steam reforming C3H8 + 3H2O → 3CO + 7H2    ∆H°298 = 499 kJ/mol   (3.1) Propane CO2 reforming C3H8 + 3CO2 • 6CO + 4H2    ∆H°298 = 622 kJ/mol   (3.2) Propane partial oxidation C3H8 + 3/2 O2 → 3CO+4H2     ∆H°298 = −227 kJ/mol   (3.3) Propane decomposition C3H8 → 3C + 4H2     ∆H°298 = 105 kJ/mol   (3.4) Among these, CO2 reforming of propane6,7 is relatively slow, and hence un-economical compared to steam reforming8. The decomposition pathway9-11 is attractive since the hydrogen produced is free of CO, while also ensuring that no additional greenhouse gases like CO2 or CH4 are produced. However, this reaction is challenging from a catalyst stability point of view. Partial oxidation12 is preferred when a carbon-monoxide-rich syngas is desired or if inexpensive oxygen is available. Steam reforming is the most economical pathway in terms of hydrogen yield13-16, since hydrogen is produced from steam as well as propane. Autothermal reforming has also been employed as a combination of partial oxidation and steam reforming17-21. Steam  86  reforming has been studied in our research. Several other reactions take place following the main steam reforming process: Methanation and Methane steam reforming CO + 3 H2 • CH4 + H2O    ∆H°298 = - 206 kJ/mol   (3.5) Water gas shift CO + H2O • CO2 + H2     ∆H°298 = - 41 kJ/mol   (3.6) Methane overall steam reforming CH4 + 2H2O • CO2 + 4H2    ∆H°298 = 165 kJ/mol   (3.7) Summing equation (3.1) and 3 times equation (3.6) leads to Propane overall steam reforming C3H8 + 6H2O = 3CO2 + 10H2    ∆H°298 = 499 kJ/mol   (3.8) Since industrial operations always use excess steam to minimize catalyst deactivation, the maximum yield of hydrogen per mole of propane fed can be 10. The following carbon formation processes can also take place as unwanted side reactions. C3H8 → 3C + 4H2          (3.4) CH4 • C + 2H2          (3.9) 2CO • C + CO2          (3.10) CO + H2 • H2O + C          (3.11) CO2 + 2H2 • 2H2O + C         (3.12) 3.1.2  Equilibrium compositions in steam reforming of propane Figure 3.1 shows the dry gas equilibrium compositions from HYSYS process simulation software, version Aspen HYSYS 7.1, for temperatures from 450 to 800°C and two pressures. Propane is seen to be fully converted at all temperatures in the range considered, indicating that steam reforming of propane is almost irreversible. No intermediate hydrocarbons were formed, except traces of ethane (~0.01%). Methane appears as an intermediate component by reverse steam reforming, and the overall conversion of hydrocarbons is limited by the steam reforming of methane. Since steam reforming of methane is highly endothermic, the methane yield decreases while the hydrogen yield increases, as temperature is increased. Since the overall reactions result in a net increase in the number of moles, increasing pressure causes a drop in the hydrogen yield, while increasing methane yield.   87  3.1.3  Steam reforming of propane: Industrial practice Refineries are turning from net producers of hydrogen to be net consumers due to increasing demand of hydrotreating operations3,22. With stricter governmental regulations of industrial emissions and growing need for additional hydrogen, refiners are sending off-gases, instead of flaring, and natural gases with varying contents of propane, in addition to LPG, to reformers for hydrogen production. If used in distributed production facilities for hydrogen, e.g. for hydrogen filling stations, propane or LPG may be fed to hydrogen production units. In refineries, the feedstock is fed first to a pre-reformer, operated at relatively low temperatures of 450 to 550°C23. All hydrocarbons heavier than methane are completely converted to C1 components (methane or carbon oxides), producing a methane-rich gas which is introduced to the steam reformer, which operates at higher temperatures of ~850 to 950°C. Using a pre-reformer means that the higher temperature steam reformer does not see variations in feedstock composition, and catalyst deactivation is minimized. The steam reformer is followed by a shift reactor section, followed by a pressure swing adsorption to produce hydrogen of 98 to 99.999% purity24. 3.1.4  Fluidized bed membrane reactor (FBMR) Fixed bed membrane reactors have been extensively studied for steam reforming of natural gas or methane25-30, and to a limited extent for LPG31 and liquid hydrocarbons32,33. Achieving very high hydrogen yield in reforming propane is prevented by the equilibrium of the steam reforming of methane and water-gas-shift reaction. From Le Chatelier’s principle, hydrogen yield can be maximized by selectively removing product hydrogen by perm-selective membrane panels. Another major drawback of traditional steam reformers arises from large intra-particle diffusional resistances. The effectiveness factors of the catalysts can be improved greatly from ~0.01 to 0.001 for industrial steam reforming catalyst pellets34 to almost unity by reducing the catalyst particle size. Pressure drop limitations then dictate that the bed be operated in a fluidized mode for fine catalyst powders. Combining these concepts, a fluidized bed membrane reactor (FBMR) was developed at the University of British Columbia35 for steam reforming of natural gas. The FBMR concept for hydrogen production from natural gas has been studied extensively35-38. This concept is extended to steam reforming of propane in this study. Rakib et al.39 utilized a two-phase model for sizing an FBMR for our proof-of-concept experiments. Details of the layout planning, safety considerations and installation are described elsewhere40. The same reactor was also used to reform heptane41.   88  3.2 Catalysts for Steam Reforming of Propane 3.2.1  Catalyst selection Steam reforming of hydrocarbons can be catalyzed by several transition metals. The specific activities of metals supported on alumina or magnesia have been found  to be42,43 in rank order Rh, Ru > Ni, Pd, Pt > Re > Co. Catalyst selection from this list is predominantly an economic consideration. However, catalyst activity and stability are also important considerations. In terms of activity as well as stability, both ruthenium and rhodium are more effective catalysts than nickel44, on which carbon formation appears to occur via a different mechanism. In addition to its lower activity, more coking arises with nickel because of formation, diffusion and dissolution of carbon in the metal, whereas dissolution of carbon in ruthenium and rhodium is considerably less. Despite their advantages, the cost and availability of Rh and Ru are such that these catalysts are not used widely in industrial applications. The most widely used catalysts for large scale industrial reformers are Al2O3-supported nickel. Especially for higher hydrocarbon feedstocks, these catalysts are frequently modified by promoters such as earth alkaline metals like Mg and Ca to improve their stability and selectivity, by reducing the acidity of the support, thereby suppressing cracking and polymerization reactions. Resistance to coke formation on Ni-based steam reforming catalysts can be significantly increased by adding K2O, MgO or CaO45-48. Commercial steam reforming catalysts are usually designed for operation at 850-900°C and above. However, FBMR operation typically does not exceed 600°C, so the catalyst must be active at this lower temperature. RK-212 naphtha steam reforming catalyst from Haldor Topsoe A/S was chosen based on the fact that it is used industrially for steam reforming of naphtha or lighter hydrocarbons at temperatures of 650°C or higher. The composition of the RK-212 catalyst is summarized in Table 3.2. It is available as 7-holed black pellets in pre-reduced form, with pellet size and shape optimized for fixed bed operation, adequate crushing strength and low pressure drop. In order to use the catalyst in a fluidized mode, the pellets were carefully crushed and sieved to narrow size cuts. The catalyst loaded to the FBMR was an equal weight mixture of +150 µm -180 µm and +180 µm -212 µm, size cuts, giving a mean particle diameter of 179 µm. This particle size provided good hydrodynamics (as observed in a Plexiglas cold model replica of the reactor), and was small enough to give favorable effectiveness factors, but large enough to minimize entrainment. The minimum fluidization velocity estimated at ambient conditions was 0.026 m/s. For shapes of the particles, see Appendix E.  89  3.2.2  Micro-reactor testing of RK-212 catalyst particles Compared to methane, steam reforming of higher hydrocarbon feedstocks, including propane, has a higher carbon formation propensity. Even though the minimum steam-carbon ratio is 2.2, ratios of 4 to 6 are common industrially23,49-51. A catalyst evaluation unit (CEU), shown schematically in Figure 3.2, was installed to identify the favorable operating conditions for the FBMR. The feed materials were water, hydrogen, and propane, accurately metered using mass flow controllers from Brooks Inc. Water is vaporized by passing through a steam generator, and mixes downstream with propane and hydrogen, before being fed to the vertical stainless steel microreactor of internal diameter 6.9 mm and length 457 mm. The actual catalyst bed height was about 50 mm, and the screen size cut +106 µm –125 µm. The product gases from the micro-reactor passed through a condenser to remove liquid water before venting or analysis. A pressure regulator was installed downstream of the condenser to maintain the required reactor pressure. Part of the product gas was sent to a Shimadzu GC-14B gas chromatograph for analysis. Catalyst activity was monitored by following the propane conversion, based on a carbon mass balance. The total catalyst mass was fixed at 1 g. Figure 3.3 shows stability plots of the catalyst for an operating temperature of 525°C, steam-to-carbon ratios of 4 to 6, and a hydrogen- to-carbon feed molar ratio 1, with the feed rate of steam fixed at 30 g/h, and the propane flow rates adjusted accordingly. Thus one of the factors to be borne in mind while interpreting the results for the conversion of propane is that the feed gas superficial velocities differed from case to case. As expected, the conversion improved with higher steam-to-carbon ratio and higher temperature. For the time spans of operation, it is seen that the propane conversions were quite stable. Although no drop in conversion can be observed from these plots, TEM analysis showed growth of filamentous carbon, with a nickel crystallite at the tip, whereas no encapsulating carbon formation could be seen from EDX analysis. TEM pictures for spent catalysts subjected to different operating conditions indicate that higher steam-carbon ratios decrease the rate of filamentous carbon formation. A lower steam-to-carbon ratio of 3.5 led to heavy blockage of the catalyst bed due to large amounts of filamentous carbon.  The micro-reactor data were used to identify operating conditions where the FBMR could be operated for long durations without significant catalyst deactivation. A base steam-to-carbon ratio of 5.0, with 4.0 as the minimum, was used for the FBMR.   90  3.3  Experimental Set-up and Procedure Figure 3.4(a) shows a schematic of the FBMR vessel, of height 2.33 m. The main section above the distributor is 1.87 m in height, with rectangular cross-sections of 1.82 x 10-3 m2 and 2.30 x 10-3 m2, with and without membrane panels installed respectively. Above is an expanded circular cross section of 73.7 mm diameter. 3.3.1 Selective hydrogen removal Palladium membranes are infinitely selective to permeation of hydrogen due to the unique solution-diffusion mechanism of permeation 52,53. Diffusion depends on the difference of square roots of partial pressures on the two sides according to Sieverts’ law when hydrogen diffusion is the rate determining step54. )exp 22 2 2 2 0 ,MH,RH H H M PH PP(RT E δ PAQ −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −=        (3.13) Pd membranes are susceptible to hydrogen embrittlement due to phase transition at temperatures around 300°C, resulting in expansion of the metal lattice28,55-57. Pd is often alloyed with other metals like Ag, Cu and Ru to improve their mechanical stability, resistance to hydrogen embrittlement and H2 permeation. A Pd77Ag23 alloy, for example, leads to a 70% gain in hydrogen flux compared with pure Pd28,58. Six double-sided Pd-Ag membrane panels were inserted through vertical slits on the wall of the FBMR from alternating sides along the reactor height. These panels were arranged vertically one above the other, passing through the centerline of the reactor shell, resulting in two communicating sections at all levels. They were manufactured and supplied by Membrane Reactor Technologies59, and are shown in Figure 3.4(b). The overall dimensions of the membrane panels were 231.8 x 73.0 mm x 6.35 mm. Accounting for the weld space for the stainless steel frame and bonding space for the Pd membrane on the substrate, the active membrane foil cross-sections were 206.4 x 50.8 mm with a thickness of 0.025 mm, the thinnest membranes where pinhole-free surface could be guaranteed at the time of purchase. These foils were bonded onto each side of the membrane panel. Each side opening supporting a membrane panel also corresponds to several ports: (i) one centrally-located for pure permeate hydrogen withdrawal, (ii) one for a thermocouple close to the permeate port, and (iii) two reactor gas sampling ports vertically equidistant from each end on either side of the permeate port. The arrangement of these ports and a schematic of a membrane assembly are shown in Figure 3.4(c).  91   Decreasing the hydrogen partial pressure on the permeate side, e.g. by providing a suitable sweep gas, can improve the hydrogen flux according to Sieverts’ law (Equation 3.13). Alternatively, suction can be provided. In our experiments, a vacuum pump downstream of the permeate hydrogen manifold was employed to increase the driving force for permeation. 3.3.2  Heat supply to the reactor Electrical band heaters were installed on the semi-circular areas opposite the side-opening flanges holding the membrane panels (henceforth referred to as lateral flanges), and on the reactor shell in the freeboard region. Electrically heated tubular heaters were inserted through the bottom flange cover, extending from just above the distributor to the top of the highest membrane panel. Heat losses occur most notably from the various flanges. The band heaters, strapped onto the reactor wall, provided additional localized heating. This array of heaters led to irregular temperature distributions in the bed, as detailed below. 3.3.3  Experiments 3.3.3.1 Operation Figure 3.5 shows a schematic of the experimental setup. Propane was pressurized above its vapor pressure using compressed helium, and its flow was metered by a Bronkhorst mass flow controller. Distilled water was pumped and metered by a Brooks mass flow controller, and fed to the vaporizer. Propane was mixed with the superheated steam downstream of the vaporizer. This superheated stream then fluidized the catalyst particles in the reactor where the generated hydrogen was separated from the reaction gas mixture by the membrane surfaces. Sufficient RK-212 catalyst particles were loaded into the reactor to completely immerse all six membrane (or dummy) panels, leaving 30 mm of catalyst above the upper edge of the topmost panel. The catalyst was reduced by introducing a nitrogen flow diluted with 1/3 parts (by volume) of hydrogen overnight at a temperature of 525°C, taking special care that hydrogen was only introduced when the bed temperatures exceeded 350°C. To initiate the experiments, superheated steam at temperatures above 500°C was introduced with the hydrogen, gradually decreasing the nitrogen flow. To avoid catalyst oxidation, the steam-to-hydrogen molar ratio was maintained below 6. Propane was introduced after ensuring a steady flow of steam, and the hydrogen feed rate was then gradually decreased to 0. The hydrogen vacuum pump was then adjusted to maintain the required partial pressure of hydrogen on the permeate side. When the desired operating conditions of temperature, flow rates, reactor pressure and permeate pressure were achieved, sample gas compositions were continuously monitored from one of the gas  92  sampling ports. Steady state was deemed to have been achieved when the dry gas hydrogen composition oscillated by less than ±1% about its mean value. 3.3.3.2 Sample gas analysis Once steady state was reached, sample gases were analyzed using a Varian CP-4900 micro-GC from different sampling points along the reactor height, repeating the first sampling after all other samples had been analyzed. The GC columns were able to quantify the gas concentrations as follows: Channel 1:  10 m Molsieve 5A with pre-column back-flush.  With Argon carrier gas, the Molsieve is capable of analyzing He, H2, O2, N2, CH4 and CO.  The argon allows for increased sensitivity and a linear range of He and H2.  The detection limits range was from 10 to 100 ppm with the Molsieve. Channel 2:  10 m PPU with pre-column back-flush.  To optimize sensitivity and analysis time, helium was the carrier gas.  With the back-flush enabled, the PPU was able to analyze CO2, C2H4, C2H6, C2H2, H2S and COS. Detection limits were in the 10 to 100 ppm range with the PPU. Channel 3:  8 m Silica PLOT with pre-column back-flush.  The carrier was again helium. With the back-flush enabled, the Silica PLOT was capable of analyzing C3 and C4 isomers.  The Silica PLOT was chosen because it allows separation of alkenes in the presence of water vapour. Detection limits were again in the 10 to 100 ppm range. 3.3.3.3 Experimental plan The FBMR was operated in three combinations of active or dummy membrane panels. (a) Six stainless steel dummies: These experiments was carried out to assess the catalyst stability under the planned operating conditions and to determine the baseline performance without any membranes, similar to the performance of a pre-reformer in traditional steam reforming process, where all the higher hydrocarbons would be fully converted to a gas mixture, limited by equilibrium of the methane steam reforming reactions. The dummy panels were stainless steel plates of dimensions 231.8 x 73.0 mm x 6.35 mm, i.e. the same dimensions as for the active membrane panels, so that the reactor internal geometry was identical for all experimental runs. (b) One active membrane panel: Initially, only one membrane was installed before conducting experiments with the full set of membranes. This single membrane panel was installed at the fifth opening from the bottom. (c) Six active membrane panels: The final and major set of experiments was conducted to utilize the full hydrogen extraction capacity of the FBMR.  93  In the current study, average bed temperatures ranged from 475 to 525°C at a steam-to- carbon ratio of 5.0. For most runs, the feed rates were adjusted to give similar gas superficial velocities based on the average bed temperature and the feed molar flow. Due to heat losses and the inability to locate a heater too near the bottom of the reactor, and the low superficial gas velocities (resulting in limited axial dispersion of solids and hence of heat), the temperatures there were much lower than the average bed temperature so that the superficial velocities were also lower there. Due to the increase in molar flow due to reaction, the gas superficial velocity increased along the reactor when no hydrogen was withdrawn via the membrane panels. In all cases examined, the bed operated in the bubbling fluidization flow regime. In most experiments, two or three samples were analyzed at each position. The error bars plotted here correspond to the standard deviations (±1σ) for each gas sampling position. In many cases these error bars are not visible due to the scale of plotting and to stable conditions for these runs. Table 3.3 shows the steady state reactor measurements. The locations of the probes are listed in Table 3.4. Table 3.4 also gives the spans and heights of the active membrane areas. The temperatures, pressures and flow rates are time-average values over the duration of experiment for each set of steady state operating conditions, and are provided in Table 3.5. To compare the experimental values with the equilibrium values corresponding to local conditions, local temperatures at sample withdrawal locations were determined by interpolation of the recorded bed temperatures. Hydrogen permeate purities were also measured for each membrane panel from time to time. Product hydrogen purity was > 99.99% at the beginning of the experiments, while for two membranes, this purity decreased to > 99.96% towards the end of the experiments.  3.4 Experimental Results The FBMR performance is analyzed by the extent of pure hydrogen production and the degree of approach towards complete conversion of the hydrocarbons: Pure hydrogen yield = stream feedin propane of flowmolar membranes  viaextractedhydrogenpure of flowar mol    (3.14) Retentate hydrogen yield = stream feedin propane of flowmolar stream retentatein hydrogen of flowmolar    (3.15) Total hydrogen yield = Pure hydrogen yield + Retentate hydrogen yield   (3.16)  94  The amount of carbon oxides generated is indicative of the conversion of propane and intermediate methane; hence we also calculated: Carbon oxides yield = ( ) stream feedin  propane of flowmolar  x 3 stream retentatein  CO of flowmolar   CO of flowmolar 2+    (3.17) Methane yield = stream feedin  propane of flowmolar  x 3 stream retentatein  methane of flowmolar     (3.18) In all of our propane steam reforming experiments, the propane conversion exceeded 99% from the lowermost sampling point, and hence it is not plotted here. Intermediate hydrocarbons like propylene and ethylene were not detected, whereas traces of ethane (less than 0.01% by volume in all cases) were detected, but are not considered in the performance calculations given their low levels. Figure 3.6 depicts the performance of propane steam reforming with no in-situ hydrogen removal, i.e. with dummy stainless plates inserted through the six lateral openings, corresponding to experiments 1.a, 1.b and 1.c in Table 3.5. There is considerable axial temperature variation due to the low superficial gas velocity and limited coverage of the heater sections. As expected from thermodynamics, the carbon oxides and hydrogen yields decrease and methane yield increases with increasing reactor pressure. This could also have been affected somewhat by the fact that for these runs, the total molar feed rate was unchanged, meaning that higher pressure led to higher residence time. The parity plot in Figure 3.7 shows that most experimental hydrogen yields were very close to the local equilibrium values. The hydrogen yields follow the same trend as the local bed temperatures, clearly indicating that the reactor behavior was controlled by local equilibrium conditions when no hydrogen was withdrawn from the reactor. Figure 3.8 corresponds to experiments 2.a to 2.d where only one active membrane panel was installed at the 5th level from the bottom (corresponding to the shaded region), with dummy panels in the other five lateral openings. Pure hydrogen was drawn from the middle location of this opening (1.31 m above the distributor), whereas reactor gas samples were drawn from two different locations (1.235 m and 1.387 m respectively above the distributor) just below and above a permeate hydrogen port. For calculation purposes, the pure hydrogen drawn from this was allocated in equal proportions to the levels of the gas sample ports. All four sets of experiments reported in Figure 3.8 had identical feed molar flow rates, so that the feed superficial velocities decreased with increasing pressure. Thus experiments 2.a and 2.b had lower residence times than 2.c and 2.d. Temperature again varied significantly in the axial  95  direction. Two levels of permeate pressure were studied for each reactor pressure. Experiments 2.a and 2.c had similar performances. This can be attributed to a higher driving force for hydrogen permeation, as well as a higher residence time for permeation for 2.c compared to 2.a, counteracted by a negative impact on thermodynamic equilibrium. Similarly, experiments 2.b and 2.d showed similar performance for the same permeate pressures of 30 kPa, but this pair gave better performance than 2.a and 2.c where no vacuum was applied downstream of the membranes (permeate pressure = 101 kPa). Note also that the difference for these two pairs only became prominent after reaching the 5th lateral flange where the single membrane panel was installed. Higher pressure also increases the gas species concentrations, accelerating the reaction rates, an effect which cannot be confirmed here due to the already-fast kinetics of the steam reforming reactions. From Figure 3.9 onwards, all plots depict results for experiments conducted with active membrane panels installed at all six lateral flanges. Compared to Figure 3.8, the performance is affected by membrane permeation starting from the lowest lateral flange, instead of the 5th flange. The shaded regions again correspond to height intervals where there were membrane surfaces. Figure 3.9 corresponds to experiments 3.a, 3.b and 3.c, in which only the permeate side pressure was varied, other parameters remaining the same. It shows significant improvements in hydrogen yield and a drop in methane yield as the permeate pressure was reduced. This confirms the effect of the driving force as a key parameter for higher pure hydrogen yield and a significant equilibrium shift. Figure 3.10 depicts the influence of reactor pressure with the average reactor temperature maintained at 500°C for reactor pressures of 400, 600 and 800 kPa, the permeate side pressures being 25 kPa for all three cases. The flow rates were adjusted to give similar superficial gas velocities. Higher reactor pressure led to poorer hydrogen yield. This can be understood in the light of a higher pressure increasing the hydrogen permeation driving force, but the performance being negatively affected by thermodynamics which means that higher pressures may or may not always translate into higher FBMR performance. Figure 3.11 shows the effect of varying the steam-to-carbon ratio for an average bed temperature of 500°C and a reactor pressure of 600 kPa. Total feed molar flow rates were identical in these three runs. From a thermodynamics point of view, higher steam partial pressure positively affects the reactor performance, leading to a higher yield of hydrogen. A higher hydrogen partial pressure in the reactor results in more pure hydrogen production as seen in  96  Figure 3.11, followed by higher yields of carbon oxides and total hydrogen, and lower yields of methane as the steam-to-carbon ratio increases. Excessive steam could actually decrease the performance by diluting the hydrogen, leading to less permeation. Higher reactor temperature greatly enhances the reaction rates. For equilibrium-limited endothermic reactions, higher temperatures also shift the reaction in the forward direction. Higher temperature also improves the permeability of hydrogen through the membranes. To show the influence of temperature, Figure 3.12 compares results from experiments 6.a and 6.b, with average bed temperatures of 475 and 525°C respectively, to those from experiment 3.c, where the average bed temperature was 500°C, all other operating conditions being the same, with the reactor feed rates adjusted to give the same superficial velocities. Higher average temperatures were not investigated to protect the membranes. One data point corresponding to 525°C is missing for the off-gas sample because the reservoir ran out of propane at that point of time. However, the more important data for samples from the dense catalyst bed for this temperature could be collected adequately as reported. Note from Figure 3.12 that the total hydrogen yield was as high as 9.26, close to the maximum possible value of 10. Consistently higher carbon oxides yield and dwindling methane yield are found as temperature increased. The effect of total molar feed rate, and hence superficial gas velocity inside the reactor was also investigated. Experiment 7 had a 33% higher molar feed rate than experiment 3.c, all other conditions being essentially the same. Results are shown in Figure 3.13. One of the important effects of higher gas velocity was an improvement in temperature uniformity, probably due to more axial solids mixing. A slight increase in pure hydrogen yield with decreasing total molar feed rate is evident, but there was little effect on any of the yields. In Figure 3.14, the effects of hydrogen removal on the methane and total hydrogen yields are compared in a parity plot of the experimental yields versus the equilibrium values if no hydrogen was removed. The experimental data correspond to values at 1.64 m above the distributor, i.e. the top of the sixth (uppermost) membrane panel, with the equilibrium values based on the local temperatures. Without hydrogen removal, the experimental hydrogen yield is somewhat lower than the equilibrium values. With six membrane panels, evacuating the permeate side to 50 kPa boosted the hydrogen yield above the equilibrium value, while evacuating it further to 25 kPa gave further improvement. The methane slip also decreased with membranes present and increased vacuum on the permeate side. Decreasing the partial pressure of hydrogen on the permeate side, and increasing the active membrane surface area  97  (corresponding to the number of membrane panels here) are clearly the most important means of improving the hydrogen and methane yields from the FBMR.  3.5 Discussion Steam reforming of propane can be viewed as producing CO and H2 by Equation 3.1, followed by production of methane by Equation 3.5, with the propane completely consumed within a very short distance above the distributor, and the water gas shift (Equation 3.6), and methane steam reforming (Equations 3.5 and 3.7) reactions taking place in the bulk of the FBMR. The latter equilibrium-limited reactions are driven towards producing more hydrogen, as hydrogen is progressively and selectively removed from the system. The results demonstrate that by steam reforming propane in an FBMR, a pre-reformer is not needed since high hydrogen yields can be attained in the reformer itself due to shifting the performance well beyond the normal equilibrium, even at the relatively low temperatures typical of pre-reformer operations. A moderate operating temperature of ~500°C can reach conversions otherwise only achievable at >750°C, and minimize catalyst deactivation problems for higher hydrocarbons. Operation at such moderate temperatures also assists in saving energy and in avoiding the expensive containment alloys for high-temperature operation required in conventional steam reformers. A separate downstream purifier for extracting pure hydrogen is also not required. Flexibility in feedstock is possible, since the FBMR was demonstrated here to work well with propane as feedstock, while it has also been shown to work with natural gas35 and heptane41. However, heavier feedstocks require higher steam-to-carbon ratios. The hydrodynamic implications resulting from this may need to be investigated. Membrane separation of pure hydrogen from the reactor gas mixture improves the reactor performance significantly, resulting in higher hydrogen yield. Performance can be improved either by increasing the reactor pressure or by decreasing the permeate side hydrogen partial pressure. However, increasing the reactor pressure causes a decrease in equilibrium hydrogen yield in the FBMR, and hence does not necessarily translate into higher hydrogen partial pressure. Decreasing the hydrogen partial pressure on the permeate side, on the other hand, is effective and relatively straightforward, achievable either by a vacuum pump or by using a sweep gas like steam.  98  At the relatively lower steam reforming temperatures (450 to 550°C) of this study, the CO content of the reformer outlet is very low, since the water gas shift reaction is favoured at low temperatures. This was observed with the dry gas content of CO mostly less than 1.5%, much lower than in traditional steam reformers. So, in addition to high hydrogen yield, the product gas very-low CO content is an additional advantage of FBMR, where the retentate dry gas composition was typically 45-50% CO2, 30-40% H2, 7-15% CH4, and 0-2% CO, depending on the operating conditions. Parametric studies were carried out to characterize the reactor performance. For an average bed temperature of 500°C, reactor pressure of 800 kPa, permeate pressure of 25 kPa, steam-to-carbon molar ratio of 5.0, a total feed rate of 0.819 mols/min, and a total membrane permeation area of 0.126 m2, the FBMR produced 0.435 Nm3/h of pure hydrogen. The maximum hydrogen yield in the experiments, for a low feed rate (0.595 mols/min) and a higher average bed temperature (525°C), was 7.71 moles of pure hydrogen (and 9.26 moles of total hydrogen) per mole of propane fed, the theoretically possible maximum being 10. High-purity hydrogen (~99.99% hydrogen) was produced from each of the membrane panels. Theoretically, Pd-based membranes have infinite selectivity for hydrogen. However, in these experiments, the hydrogen permeate was not 100% pure, probably due to structural defects or faults, commonly called pinholes, in the membranes.  3.6 Conclusions Steam reforming of propane was studied in a fluidized bed membrane reactor for production of pure hydrogen. Experiments with membrane dummies, instead of active membranes, indicate that for the operating conditions studied, the local gas compositions in the reactor closely approach equilibrium corresponding to the local temperatures and pressures. Continuous removal of pure hydrogen selectively through Pd-Ag membranes shifts the equilibrium towards production of more hydrogen. Experiments with one membrane panel and with six membrane panels show that the performance of an FBMR improves relative to the no-membrane case, but the extent of improvement is limited by the membrane permeation capacity. Higher reactor side pressure do not always improve the hydrogen yield, because higher driving forces of permeation are offset by lower equilibrium conversions. Since there is a continuous shift in the equilibrium towards more hydrogen production when there is continuous pure hydrogen removal, an FBMR for steam reforming of propane can combine the functions of a pre-reformer, reformer, shift  99  converter and purification section, into a single unit. High-purity hydrogen (~99.99% hydrogen) was produced from each membrane panel. Since high hydrogen yields can be obtained at relatively low temperatures like 475 to 550°C, catalyst deactivation by sintering and high temperature carbon formation can be minimized. Propane was fully consumed at the beginning of the FBMR, which predominantly acts as a methane steam reformer. Since the bulk of the FBMR does not see the higher hydrocarbon, the FBMR can be operated flexibly with a variety of feedstocks.    100  Table 3.1: Density of liquid/ liquefied hydrocarbons at ambient pressure  Hydrocarbon Conditions Liquid density Molar density Hydrogen content Hydrogen proportion   (kg/m 3) (kmol/m3) (kg H2/ m3 hydrocarbon) (kg H2/ kg hydrocarbon) Methane Ambient pressure, -162.5°C 424.9 26.48 106.7 0.25 Methane 11.86 bar, -120.5°C 352.9 21.99 88.66 0.25 Propane 8.62 bar, 20°C 500.3 11.35 91.53 0.18 n-Butane 2.07 bar, 20°C 578.8 9.96 100.4 0.17 n-Heptane Ambient pressure, 20°C 682.4 6.81 109.8 0.16  Table 3.2: Composition of RK-212 (catalyst provided by Haldor Topsoe A/S)  % (w/w) Component 12-15   Nickel   Ni 0-3 Nickel monoxide NiO 25-30 Magnesium oxide MgO 60-65 Aluminium oxide Al2O3 1-2 Potassium oxide K2O 1-4 Calcium oxide CaO MgO bound as magnesium aluminate spinel (MgAl2O4) CaO bound as calcium aluminate spinel (CaAl4O7)       101  Table 3.3: Steady-state reactor measurements  Performance variables Device and Location Bed temperatures One thermocouple just above distributor, one thermocouple       close to the center of each membrane panel, one for the       freeboard, and one just before reformer exit. Gas compositions Two sampling ports for each of the lateral flanges       supporting a membrane panel, one for the ROG. These       thirteen sample gases are analyzed online by a Varian       micro-GC CP-4900 using sample selection valves. Permeate hydrogen Flow rate of permeate hydrogen from each membrane panel       using mass flow meters. Purity of permeate hydrogen Checking hydrogen purity in permeate product from each       membrane panel by the micro-GC. Pressures Absolute pressures in the feed line, freeboard, and at the       distributor level in bed. Differential pressure between       alternate levels of side flanges (i.e. the pairs 1-3, 3-5, 2-       4, and 4-6) and between the distributor and freeboard.   102  Table 3.4: Location of sampling ports, thermocouples and pure hydrogen withdrawal, and height intervals of active membrane surface   Description (Side opening counted from bottom) Location above distributor holes (m) Height interval covered by active membrane Thermocouple (Bottom) 0.01 - Thermocouple (Side opening 1) 0.32 - Thermocouple ( Side opening 2) 0.52 - Thermocouple ( Side opening 3) 0.78 - Thermocouple ( Side opening 4) 1.08 - Thermocouple ( Side opening 5) 1.29 - Thermocouple ( Side opening 6) 1.59 - Thermocouple (Freeboard) 2.33 - Gas samples ( Side opening 1) 0.22, 0.37 - Gas samples ( Side opening 2) 0.47, 0.63 - Gas samples ( Side opening 3) 0.73, 0.88 - Gas samples ( Side opening 4) 0.98, 1.13 - Gas samples ( Side opening 5) 1.24, 1.39 - Gas samples ( Side opening 6) 1.49, 1.64 - Pure hydrogen ( Side opening 1) 0.30 0.19 – 0.40 Pure hydrogen ( Side opening 2) 0.55 0.45 – 0.65 Pure hydrogen ( Side opening 3) 0.80 0.70 – 0.91 Pure hydrogen ( Side opening 4) 1.06 0.95 – 1.16 Pure hydrogen ( Side opening 5) 1.31 1.21 – 1.41 Pure hydrogen ( Side opening 6) 1.57 1.46 – 1.67  103  Table 3.5: Experimental runs for steam reforming of propane  Expt No. Active Membranes (Location) Total feed rate Tav  P  Pm SCR   (mols/min) (°C) (kPa) (kPa)  1.a 0.673 500 400 NA 6.0 1.b 0.673 500 600 NA 6.0 1.c None 0.673 500 700 NA 6.0  2.a 0.717 485 515 101 5.0 2.b 0.717 485 515 30 5.0 2.c 0.717 485 700 101 5.0 2.d 1 (#5) 0.717 485 700 30 5.0  3.a 0.614 500 600 101 5.0 3.b 0.614 500 600 50 5.0 3.c 6 (#1 to #6) 0.614 500 600 25 5.0  4.a 0.410 500 400 25 5.0 4.b 6 (#1 to #6) 0.819 500 800 25 5.0  5.a 0.614 500 600 25 4.0 5.b 6 (#1 to #6) 0.614 500 600 25 6.0  6.a 0.635 475 600 25 5.0 6.b 6 (#1 to #6) 0.595 525 600 25 5.0  7 6 (#1 to #6) 0.819 500 600 25 5.0     104   (a)     P  = 400 kPa, SCR  = 5.0 0 25 50 75 450 500 550 600 650 700 750 800 Temperature (oC) M ol e % C3H8 CH4 H2 CO CO2     (b)     P  = 1000 kPa, SCR  = 5.0 0 25 50 75 450 500 550 600 650 700 750 800 Temperature (oC) M ol e % C3H8 CH4 H2 CO CO2   Figure 3.1: Equilibrium compositions (dry gas) in propane steam reforming for steam-to- carbon molar ratio = 5.0: (a) P = 400 kPa; (b) P = 1000 kPa   105  Cooling Water To Vent H2 + CH4/C3H8 Steam Generator Reactor Condensed water To GC for sampling Hydrogen Propane Nitrogen Water Tank Propane MFC Hydrogen MFC Water MFC Heater Temperature Catalyst Bed Temperature Condenser Feed Pressure Exit Pressure Back Pressure Regulator  Figure 3.2: Schematic of micro-reactor set-up to study steam reforming of propane   106   0 20 40 60 80 100 0 10 20 30 40 50 60 70 80 Time on-stream (h) Pr op an e co nv er si on  (% )    SCR = 4    SCR = 5    SCR = 6   Figure 3.3: Propane conversion for steam reforming in micro-reactor, T = 525°C, H2O = 30 g/h  107  6. 35  m m  O D  S S tu be 231.8 mm 206.4 mm 73 .0  m m 6.35 mm Pd77Ag23 Membrane Foil Membrane Panel Frame 50 .8  m m Membrane panel Supporting flange coverGas sample Gas sample Pressure port Permeate line Thermocouple (a) (b) (c)   Figure 3.4: The FBMR pressure vessel, showing dimensions of membrane panel, and arrangement of ports on each lateral flange cover where membrane panels are installed  108  Liquid Water Helium Propane Tank Propane MFC Liquid Propane Va po riz er Water Tank Water Pump Water MFC Permeate 2 Reformer Feed Sample 2a Sample 2b Permeate 4 Sample 4a Sample 4b Permeate 6 Sample 6a Sample 6b Permeate 1 Permeate 3 Permeate 5 Reformer Off Gas Sample 1a Sample 1b Sample 3a Sample 3b Sample 5a Sample 5b Sample ROG FBMR H2 MFM H2 MFM H2 MFM H2 MFM H2 MFM H2 MFM H2 MFM H2 Vacuum Pump    Figure 3.5: Schematic of experimental setup to study steam reforming of propane in an FBMR   109  Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 P = 400 kPa P = 600 kPa P = 700 kPa To ta l H yd ro ge n  Y ie ld 3 5 7 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.3 0.4 0.5 0.6   Figure 3.6: Experimental yields and temperature for propane steam reforming without membrane panels at average reactor temperature of 500°C and steam-to-carbon ratio molar ratio of 6.0. Total reactor feed = 0.673 mols/min   110  0 2 4 6 8 10 0 2 4 6 8 10 Equilibrium H2 Yield Ex pe rim en ta l H 2 Y ie ld   P = 400 kPa   P = 600 kPa   P = 700 kPa   Figure 3.7: Parity plot of experimental hydrogen yield without membrane panels against local equilibrium values  111  Lo ca l Te m p.  (o C ) 460 490 520 550 580 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 P = 515 kPa, Pm = 101 kPa P = 515 kPa, Pm = 30 kPa P = 700 kPa, Pm = 101 kPa P = 700 kPa, Pm = 30 kPa To ta l H yd ro ge n  Y ie ld 1 3 5 Pu re  H yd ro ge n  Y ie ld 0.0 0.5 1.0 1.5 2.0 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.0 0.2 0.4 0.6   Figure 3.8: Experimental yields and temperature for propane steam reforming at average reactor temperature of 485°C and steam-to-carbon molar ratio 5.0. One membrane panel installed, spanning from 0.95 to 1.16 m above distributor. Total reactor feed = 0.717 mols/min  112  Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 Pm = 101 kPa Pm = 50 kPa Pm = 25 kPa To ta l H yd ro ge n  Y ie ld 3 5 7 9 Pu re  H yd ro ge n  Y ie ld 1 3 5 7 9 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 3.9: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feed = 0.614 mols/min  113  Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 P = 400 kPa P = 600 kPa P = 800 kPa To ta l H yd ro ge n  Y ie ld 3 5 7 9 Pu re  H yd ro ge n  Y ie ld 1 3 5 7 9 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75  Figure 3.10: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.410, 0.614, and 0.819 mols/min  114  Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 SCR = 4 SCR = 5 SCR = 6 To ta l H yd ro ge n  Y ie ld 3 5 7 9 Pu re  H yd ro ge n  Y ie ld 1 3 5 7 9 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 3.11: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, and permeate pressure 25 kPa. Six membrane panels installed. Total reactor feed = 0.614 mols/min  115  Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 Tav = 475 oC Tav = 500 oC Tav = 525 oC To ta l H yd ro ge n  Y ie ld 3 5 7 9 Pu re  H yd ro ge n  Y ie ld 1 3 5 7 9 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 3.12: Experimental yields and temperature for propane steam reforming at pressure of 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.635, 0.614, and 0.595 mols/min for 475, 500, and 525°C respectively  116  Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 Total Feed = 0.614 mols/min Total Feed = 0.819 mols/min To ta l H yd ro ge n  Y ie ld 3 5 7 9 Pu re  H yd ro ge n  Y ie ld 1 3 5 7 9 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 3.13: Experimental yields and temperature for propane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed  117  0 2 4 6 8 10 0 2 4 6 8 10 Equilibrium H2 Yield Ex pe rim en ta l H 2 Y ie ld   Without Membrane Panels   1 Membrane Panel   6 Membrane Panels P m  = 50 kPa P m  = 101 kPaP m  = 101 kPa P m  = 30 kPa P m  = 25 kPa (a)  0.0 0.2 0.4 0.6 0.8 0.0 0.2 0.4 0.6 0.8 Equilibrium CH4 Yield Ex pe rim en ta l C H 4 Y ie ld   Without Membrane Panels   1 Membrane Panel   6 Membrane Panels P m  = 50 kPa P m  = 101 kPa P m  = 25 kPa P m  = 101 kPa P m  = 30 kPa (b)   Figure 3.14: Parity plot of experimental yields against equilibrium values at local temperatures if there was no hydrogen removal: (a) Hydrogen yield (b) Methane yield    118  3.7 References 1. Fonseca, A.; Sá, V.; Bento, H.; Tavares, M. L. C.; Pinto, G.; Gomes, L. A. C. N., Hydrogen distribution network optimization: a refinery case study. Journal of Cleaner Production 2008, 16, (16), 1755-1763. 2. Ramachandran, R.; Menon, R. K., An overview of industrial uses of hydrogen. International Journal of Hydrogen Energy 1998, 23, (7), 593-598. 3. Rostrup-Nielsen, J. R.; Rostrup-Nielsen, T., Large-scale hydrogen production. CatTech 2002, 6, 150-159. 4. Johnston, B.; Mayo, M. C.; Khare, A., Hydrogen: the energy source for the 21st century. Technovation 2005, 25, (6), 569-585. 5. Ball, M.; Wietschel, M., The future of hydrogen - opportunities and challenges. International Journal of Hydrogen Energy 2009, 34, (2), 615-627. 6. Jensen, M. B.; Raberg, L. B.; Sjastad, A. O.; Olsbye, U., Mechanistic study of the dry reforming of propane to synthesis gas over a Ni/Mg(Al)O catalyst. Catalysis Today 2009, 145, (1-2), 114-120. 7. Olafsen, A.; Daniel, C.; Schuurman, Y.; Raberg, L. B.; Olsbye, U.; Mirodatos, C., Light alkanes CO2 reforming to synthesis gas over Ni based catalysts. Catalysis Today 2006, 115, (1- 4), 179-185. 8. Whittington, B. I.; Jiang, C. J.; Trimm, D. L., Vehicle exhaust catalysis .1. The relative importance of catalytic-oxidation, steam reforming and water-gas shift reactions. Catalysis Today 1995, 26, (1), 41-45. 9. Takenaka, S.; Kawashima, K.; Matsune, H.; Kishida, M., Production of CO-free hydrogen through the decomposition of LPG and kerosene over Ni-based catalysts. Applied Catalysis A: General 2007, 321, (2), 165-174. 10. Solovyev, E. A.; Kuvshinov, D. G.; Ermakov, D. Y.; Kuvshinov, G. G., Production of hydrogen and nanofibrous carbon by selective catalytic decomposition of propane. International Journal of Hydrogen Energy 2009, 34, (3), 1310-1323. 11. Muradov, N. Z., How to produce hydrogen from fossil-fuels without CO2 emission. International Journal of Hydrogen Energy 1993, 18, (3), 211-215. 12. Minutillo, M., On-board fuel processor modelling for hydrogen-enriched gasoline fuelled engine. International Journal of Hydrogen Energy 2005, 30, (13-14), 1483-1490.  119  13. Dokupil, M.; Spitta, C.; Mathiak, J.; Beckhaus, P.; Heinzel, A., Compact propane fuel processor for auxiliary power unit application. Journal of Power Sources 2006, 157, (2), 906- 913. 14. Karakaya, M.; Avci, A. K., Simulation of on-board fuel conversion in catalytic microchannel reactor-heat exchanger systems Topics in Catalysis 2009. 15. Resini, C.; Delgado, M. C. H.; Arrighi, L.; Alemany, L. J.; Marazza, R.; Busca, G., Propene versus propane steam reforming for hydrogen production over Pd-based and Ni-based catalysts. Catalysis Communications 2005, 6, (7), 441-445. 16. Suzuki, T.; Iwanami, H.; Iwamoto, O.; Kitahara, T., Pre-reforming of liquefied petroleum gas on supported ruthenium catalyst. International Journal of Hydrogen Energy 2001, 26, (9), 935-940. 17. Ma, L.; Trimm, D. L.; Jiang, C., The design and testing of an autothermal reactor for the conversion of light hydrocarbons to hydrogen .1. The kinetics of the catalytic oxidation of light hydrocarbons. Applied Catalysis A: General 1996, 138, (2), 275-283. 18. Avci, A. K.; Onsan, Z. I.; Trimm, D. L., On-board hydrogen generation for fuel cell- powered vehicles: the use of methanol and propane. Topics in Catalysis 2003, 22, (3-4), 359- 367. 19. Gokaliler, F.; Caglayan, B. S.; Onsan, Z. I.; Aksoylu, A. E., Hydrogen production by autothermal reforming of LPG for PEM fuel cell applications. International Journal of Hydrogen Energy 2008, 33, (4), 1383-1391. 20. Wang, W. Y.; Turn, S. Q.; Keffer, V.; Douette, A., Study of process data in autothermal reforming of LPG using multivariate data analysis. Chemical Engineering Journal 2007, 129, (1- 3), 11-19. 21. Resini, C.; Arrighi, L.; Delgado, M. C. H.; Vargas, M. A. L.; Alemany, L. J.; Riani, P.; Berardinelli, S.; Marazza, R.; Busca, G., Production of hydrogen by steam reforming of C3 organics over Pd-Cu/γ-Al2O3 catalyst. International Journal of Hydrogen Energy 2006, 31, (1), 13-19. 22. Shumake, G.; Abudiab, T., Operation hydrogen. Hydrocarbon Engineering February 2006. 23. Christensen, T. S., Adiabatic prereforming of hydrocarbons - An important step in syngas production. Applied Catalysis A: General 1996, 138, (2), 285-309. 24. Sircar, S.; Golden, T. C., Purification of hydrogen by pressure swing adsorption. Separation Science and Technology 2000, 35, (5), 667-687.  120  25. Gallucci, F.; Paturzo, L.; Fama, A.; Basile, A., Experimental study of the methane steam reforming reaction in a dense Pd/Ag membrane reactor. Industrial & Engineering Chemistry Research 2004, 43, (4), 928-933. 26. Gallucci, F.; Comite, A.; Capannelli, G.; Basile, A., Steam reforming of methane in a membrane reactor: An industrial case study. Industrial & Engineering Chemistry Research 2006, 45, (9), 2994-3000. 27. Barbieri, G.; Violante, V.; DiMaio, F. P.; Criscuoli, A.; Drioli, E., Methane steam reforming analysis in a palladium-based catalytic membrane reactor. Industrial & Engineering Chemistry Research 1997, 36, (8), 3369-3374. 28. Shu, J.; Grandjean, B. P. A.; Vanneste, A.; Kaliaguine, S., Catalytic palladium-based membrane Reactors - A review. Canadian Journal of Chemical Engineering 1991, 69, (5), 1036- 1060. 29. Matsumura, Y.; Tong, J. H., Methane steam reforming in hydrogen-permeable membrane reactor for pure hydrogen production. Topics in Catalysis 2008, 51, (1-4), 123-132. 30. Shirasaki, Y.; Tsuneki, T.; Ota, Y.; Yasuda, I.; Tachibana, S.; Nakajima, H.; Kobayashi, K., Development of membrane reformer system for highly efficient hydrogen production from natural gas. International Journal of Hydrogen Energy 2009, 34, (10), 4482-4487. 31. Loffler, D. G.; Taylor, K.; Mason, D., A light hydrocarbon fuel processor producing high-purity hydrogen. Journal of Power Sources 2003, 117, (1-2), 84-91. 32. Damle, A. S., Hydrogen production by reforming of liquid hydrocarbons in a membrane reactor for portable power generation - Experimental studies. Journal of Power Sources 2009, 186, (1), 167-177. 33. Damle, A. S., Hydrogen production by reforming of liquid hydrocarbons in a membrane reactor for portable power generation - Model simulations. Journal of Power Sources 2008, 180, (1), 516-529. 34. Elnashaie, S. S. E. H.; Adris, A. M.; Alubaid, A. S.; Soliman, M. A., On the non- monotonic behavior of methane-steam reforming kinetics. Chemical Engineering Science 1994, 49, (7), 1079-1079. 35. Adris, A. M.; Lim, C. J.; Grace, J. R., The fluidized bed membrane reactor system: A pilot scale experimental study. Chemical Engineering Science 1994, 49, (24B), 5833-5843. 36. Boyd, T.; Grace, J.; Lim, C. J.; Adris, A. E. M., Hydrogen from an internally circulating fluidized bed membrane reactor. International Journal of Chemical Reactor Engineering 2005, 3, A58.  121  37. Chen, Z.; Po, F.; Grace, J. R.; Lim, C. J.; Elnashaie, S.; Mahecha-Botero, A.; Rakib, M.; Shirasaki, Y.; Yasuda, I., Sorbent-enhanced/membrane-assisted steam-methane reforming. Chemical Engineering Science 2008, 63, (1), 170-182. 38. Patil, C. S.; Annaland, M. V. S.; Kuipers, J. A. M., Fluidised bed membrane reactor for ultrapure hydrogen production via methane steam reforming: Experimental demonstration and model validation. Chemical Engineering Science 2007, 62, (11), 2989-3007. 39. Rakib, M. A.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J.; Bolkan, Y. G., Kinetic simulation of a compact reactor system for hydrogen production by steam reforming of higher hydrocarbons. Canadian Journal of Chemical Engineering 2008, 86, (3), 403-412. 40. Rakib, M. A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H.; Epp, M.; Gulamhusein, A.; Boyd, T.; Keelan, A., Pilot scale experimental setup for hydrogen production from higher hydrocarbons: Safety considerations and implementation. Journal of Chemical Health and Safety Under review. 41. Rakib, M. A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H., Steam reforming of heptane in a fluidized bed membrane reactor. Journal of Power Sources 2010 (In Press). 42. Rostrup-Nielsen, J. R., Activity of nickel catalysts for steam reforming of hydrocarbons. Journal of Catalysis 1973, 31, (2), 173-199. 43. Rostrup-Nielsen, J. R., Catalytic steam reforming. In Catalysis, Science and Technology, Andersen, J. R.; Boudart, M., Eds. Springer-Verlag: Berlin 1984; pp 1-117 44. Trimm, D. L., Coke formation and minimisation during steam reforming reactions. Catalysis Today 1997, 37, (3), 233-238. 45. Golebiowski, A.; Stolecki, K.; Prokop, U.; Kusmierowska, A.; Borowiecki, T.; Denis, A.; Sikorska, C., Influence of potassium on the properties of steam reforming catalysts. Reaction Kinetics and Catalysis Letters 2004, 82, (1), 179-189. 46. Lisboa, J. D.; Santos, D. C. R. M.; Passos, F. B.; Noronha, F. B., Influence of the addition of promoters to steam reforming catalysts. Catalysis Today 2005, 101, (1), 15-21. 47. Jeong, H.; Kim, K. I.; Kim, D.; Song, I. K., Effect of promoters in the methane reforming with carbon dioxide to synthesis gas over Ni/HY catalysts. Journal of Molecular Catalysis A: Chemical 2006, 246, (1-2), 43-48. 48. Borowiecki, T.; Chmiel, B.; Grzegorczyk, W.; Nazimek, D.; Pielach, M.; Barcicki, J., Influence of metal-oxides of alkaline-earths on activity of nickel catalysts in the reaction of hydrocarbon steam reforming process. Przemysl Chemiczny 1981, 60, (9-10), 456-459.  122  49. Gunardson, H., Industrial gases in petrochemical processing Marcel Dekker: New York, 1998. 50. Haldor Topsoe A/S Manual for Topsoe naphtha reforming catalysts, RK-212/ RK-202/ RK-201. 51. Gomach, J., Personal communication, Sales & Technical Service Engineer, Syngas Catalyst & Technology, Haldor Topsoe Inc. 2007. 52. Uemiya, S., State-of-the-art of supported metal membranes for gas separation. Separation and Purification Methods 1999, 28, (1), 51-85. 53. Paglieri, S. N.; Way, J. D., Innovations in palladium membrane research. Separation and Purification Methods 2002, 31, (1), 1-169. 54. Sieverts, A.; Zapf, G., The solubility of deuterium and hydrogen in solid palladium. Zeitschrift für Physikalische Chemie 1935, 174, 359-364. 55. Sjardin, M.; Damen, K. J.; Faaij, A. P. C., Techno-economic prospects of small-scale membrane reactors in a future hydrogen-fuelled transportation sector. Energy 2006, 31, (14), 2523-2555. 56. Okazaki, J.; Tanaka, D. A. P.; Tanco, M. A. L.; Wakui, Y.; Mizukami, F.; Suzuki, T. M., Hydrogen permeability study of the thin Pd–Ag alloy membranes in the temperature range across the α–β phase transition. Journal of Membrane Science 2006, 282, (1-2), 370-374. 57. Roa, F.; Block, M. J.; Way, J. D., The influence of alloy composition on the H2 flux of composite Pd-Cu membranes. Desalination 2002, 147, (1-3), 411-416. 58. Adhikari, S.; Fernando, S., Hydrogen membrane separation techniques. Industrial & Engineering Chemistry Research 2006, 45, (3), 875-881. 59. Li, A.; Grace, J. R.; Lim, C. J., Preparation of thin Pd-based composite membrane on planar metallic substrate - Part 1: Pre-treatment of porous stainless steel substrate. Journal of Membrane Science 2007, 298, (1-2), 175-181.       123  CHAPTER 4.      STEAM REFORMING OF HEPTANE IN A FLUIDIZED BED MEMBRANE REACTOR∗   4.1  Introduction 4.1.1 Background Hydrogen demand is increasing in the petrochemicals and petroleum processing  sectors1-4 and for other industrial applications. It may  also  increase significantly in the energy and transportation sectors5-8. Being a carbon-free fuel, hydrogen  can assist in mitigating global warming due to greenhouse gas emissions if CO2 emissions can be minimized during hydrogen production9. About 48% of industrial hydrogen is produced from natural gas as feedstock10, largely due to the widespread availability of natural gas, as well as having the highest hydrogen-to- carbon ratio. However, for onboard hydrogen generation for mobile applications, liquid hydrocarbons like gasoline, naphtha, kerosene or diesel are advantageous  feedstocks11,12, safely storable under ambient conditions, and with much higher volumetric energy density than natural gas13. Liquid  feedstocks like naphtha are often used for hydrogen production when natural gas is not available, accounting for about 30% of  hydrogen production10,14. In refineries, feedstock versatility for steam reformers would be a great advantage due to fluctuating demand and supply of different feedstocks15. Naphtha is the most common liquid hydrocarbon feedstock for hydrogen production. For steam reforming, low aromatic-content naphtha (LAN) is preferred. Recently, naphtha prices have been unstable due to fluctuations in oil prices. For places with access to both naphtha and natural gas, naphtha tends to be an unprofitable feedstock for hydrogen production during peaks, while being preferred during slumps. Many steam reforming facilities worldwide, especially in India and China, have installed pre-reformer units upstream of natural gas steam reformers to facilitate feedstock flexibility.  ∗ A version of this chapter has been published: Rakib, M.A., Grace, J.R., Lim, C.J., and Elnashaie, S.S.E.H., Steam Reforming of Heptane in a Fluidized Bed Membrane Reactor, Journal of Power Sources (2010) 195, 5749-5760    124  For steam reforming of higher hydrocarbons, the major reactions can be written: Higher hydrocarbons steam reforming 22mn H2 mnnCOOnHHC ⎟⎠ ⎞⎜⎝ ⎛ ++→+   ∆H0298 = 1108 kJ/mol for n=7 (4.1) Methanation and methane steam reforming CO + 3 H2  '  CH4 + H2O    ∆H°298 = - 206 kJ/mol   (4.2) Water gas shift CO + H2O  '  CO2 + H2     ∆H°298 = - 41 kJ/mol   (4.3) Methane overall steam reforming CH4 + 2H2O  '  CO2 + 4H2    ∆H°298 = 165 kJ/mol   (4.4) Summing equation (4.1) and n times equation (4.3) leads to 222mn H2 m2nnCO    O2nHHC ⎟⎠ ⎞⎜⎝ ⎛ ++=+ For n = 7 (i.e. n-heptane): C7H16 + 14H2O = 7CO2 + 22H2        (4.5) Since, under industrial operating conditions, excess steam is always used to minimize catalyst deactivation, the maximum hydrogen yield is 22 moles per mole of heptane fed. 4.1.2 Catalyst issues in steam reforming of higher hydrocarbons Commercial catalysts for steam reforming of hydrocarbons are generally based on Ni, dispersed on a refractory support, due to its high activity and low cost. Other possible candidates include Co, Pt, Pd, Ru and Rh, the order of specific activities of metals supported on alumina or magnesia being Rh, Ru >Ni, Pd, Pt >Re >Co 16. Ni catalysts present major coking problems because of the formation, diffusion and dissolution of carbon. Higher hydrocarbons show a greater tendency to form carbon on Ni than methane. Therefore, special catalyst formulations containing alkali or rare earths, or based on an active magnesia support, are required17. For higher hydrocarbons, there is a potential for various forms of carbon formation18-23. A common technique to reduce carbon formation is to employ a higher steam-to-carbon ratio than required stoichiometrically, the excess increasing with the number of carbons in the hydrocarbon chain. For example, in industrial naphtha steam reforming, steam-to-carbon ratios of 4 to 6 are common24-26 compared with ~3 for natural gas. However, a high steam-to-carbon ratio decreases the thermal efficiency of the process, and also leads to a larger reformer due to  125  the higher volumetric gas flow rates. On the other hand, in addition to resulting in higher rates of carbon formation, lower steam-to-carbon ratios also lead to higher methane leaving the reformer, which must then be compensated by maintaining a higher exit temperature. Intensive research on catalyst design is being carried out to decrease this ratio27. 4.1.3 Naphtha steam reforming: Industrial practice 4.1.3.1 Conventional naphtha steam reforming Since steam reforming of methane is endothermic and equilibrium-limited, industrial natural gas steam reformers operate at temperatures >850°C to achieve high conversions. However, the same operating conditions cannot be applied to higher hydrocarbon feedstocks like naphtha because such high temperatures would cause rapid catalyst deactivation due to carbon formation and shorter reformer tube life. A conventional naphtha steam reformer uses catalysts promoted with alkali compounds to suppress carbon formation28. In many cases, two catalysts are provided, with the entrance of the reformer loaded with a more robust catalyst to handle heavier feeds. A high steam-to-carbon ratio, usually >4.0, is used to suppress catalyst deactivation29,30. A lower average operating temperature is employed, with typical inlet and outlet temperatures of 485 and 850°C, respectively. Commercially available naphtha steam reforming catalysts have nickel loadings from 15% to ~25%, most again promoted by K2O. 4.1.3.2 Steam reforming with pre-reformer A modern hydrogen plant accepting naphtha feedstock starts with an additional unit, the pre- reformer, after feed desulfurization. Pre-reforming of the desulfurized hydrocarbon feedstock makes the gas feed to the primary reformer practically free of higher hydrocarbons, which are converted directly to C1 components with no intermediate hydrocarbon products. Thus, while the pre-reformer operates with specially designed pre-reformer catalysts at temperatures from 450 to 550°C23,29, the methane-rich gas from the pre-reformer can be heated to  >650°C before entering the reformer operating at exit temperatures of ~950°C29. Industrial pre-reformer catalysts are typically highly Ni-loaded, ~25-30% (by weight) for pre-reforming of lighter hydrocarbons up to LPG, and >50% for the naphtha range. The catalysts are characterized by high resistance to sulphur-poisoning and coke formation. At the practiced pre-reforming temperatures, undesired reactions like pyrolysis, steam cracking of higher hydrocarbons, and polymerization of alkenes are minimal. All forms of carbon formation can be avoided by properly choosing the temperature window for steam reforming23,28. The higher hydrocarbon steam reforming reactions are practically irreversible, and thus the hydrogen yields are limited by the equilibrium of the  126  methane steam reforming reactions. Downstream of the pre-reformer, the steam reformer therefore tends to operate at typical methane steam reforming operating conditions, and utilize regular methane steam reforming catalysts. The first naphtha steam reformer dates back to 1962 at ICI, with an operating pressure of 15 bars28,31,32. Some naphtha steam reformers have been operated at low temperatures to produce a methane-rich substitute natural gas. A Topsoe naphtha steam reformer was introduced in 1965, and a pre-reformer was first installed by Topsoe in 198631. Figure 4.1 shows the block diagram of a modern higher hydrocarbon steam reforming set-up incorporating a pre-reformer. Pre- reforming catalysts have high nickel loadings, typically in excess of 25% by weight and some as high as 55%. In the process for making hydrogen, the synthesis gas mixture leaving the steam reformer has few downstream units to purify the hydrogen. Traditionally, the shift conversion reaction following the reformer used to be conducted in two stages: a high-temperature shift (HTS) converter followed by a low-temperature shift (LTS) converter. With more recent steam reforming plants operating at low steam-to-carbon ratios, these reactors are replaced by a single medium-temperature shift (MTS) converter. A CO2 removal section and a methanator (to remove CO traces) may follow the shift conversion. Recent developments also have CO2 removal and methanation units replaced by pressure swing adsorption (PSA) to produce hydrogen of purity up  to 99.999%33. 4.1.3.3 Fluidized bed membrane reformer (FBMR)   Fine catalyst particles ideal for fluidization increase the catalyst  effectiveness factor from as low as 0.01-0.001 in fixed bed reformers to almost unity34,35. Better thermal uniformity in a fluidized bed can prevent hotspots. Selective removal of hydrogen from the reaction environment via permselective Pd alloy membranes drives the equilibrium methane steam reforming and water gas shift conversions forward, thereby significantly enhancing the hydrogen yield36-39. The fluidized bed and membrane reactor concepts developed at the University of British Columbia40,41, has been  commercialized by Membrane Reactor Technologies42. Rakib et al.43 provided a FBMR model for steam reforming of heptane, and predicted that an FBMR for higher hydrocarbons can result in a compact reformer system, combining pre-reforming, reforming and hydrogen purification in a single unit. This paper focuses on the technical feasibility of such a reformer unit, with n-heptane as a surrogate for naphtha, as in some previous studies23,43-47.   127  4.1.4 Thermodynamics of n-heptane steam reforming A HYSYS steady state simulator, version Aspen HYSYS 7.1, was first used to examine the thermodynamics of heptane steam reforming for operating conditions spanning the experimental conditions. Figure 4.2 shows the dry gas compositions at a steam-to-carbon molar ratio of 5 for pressures of 400 and 800 kPa. Figure 4.3 shows dry gas compositions at a pressure of 400 kPa and steam-to-carbon molar ratios of 4 and 6. It is seen that heptane is fully consumed, indicating that heptane reforming is essentially irreversible for temperatures from 400 to 800°C. Irreversibility of steam reforming is a general feature for higher hydrocarbons having different degrees of reactivity28. Industrial steam reforming of light gas oils and diesel fuels produces syngas with no traces of higher hydrocarbons in the product1. Equilibrium predictions also show the absence of intermediate hydrocarbons other than methane, except for a trace of ethane (~0.1% typically). Hydrogen production increases as temperature is increased, decreasing the equilibrium content of methane. This is because the steam reforming of methane is endothermic. Also, since the reactions involve a net increase in molar flow, Le Chatelier’s principle requires that increasing pressure decreases the hydrogen production, as is evident from Figure 4.2. Higher steam partial pressure has a positive effect on hydrogen production, as seen in Figure 4.3.  4.2 FBMR for Steam Reforming of Heptane 4.2.1 FBMR experimental set-up An FBMR pressure vessel, shown in Figure 4.4, was fabricated to allow experiments up to 10 barg and 621°C. A commercial naphtha steam reforming catalyst, RK-212 from Haldor Topsoe A/S, was crushed and sieved to a Sauter mean particle diameter of 179 µm. Pd membranes are infinitely selective to hydrogen permeation  due to the unique solution-diffusion mechanism of permeation48,49. Hydrogen diffusion flux depends on the difference between the square roots of partial pressures on the two sides according to Sieverts’ law, with diffusion as the rate- determining step50: )exp 22 2 2 2 0 ,MH,RH H H M PH PP(RT E δ PAQ −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −=        (4.6) Pd is often alloyed with other metals like Ag, Cu and Ru to improve mechanical stability, resistance to hydrogen embrittlement and hydrogen permeation flux. In our study, double-sided  128  membrane panels, manufactured by Membrane Reactor Technologies51 with a 25 µm thick Pd77Ag23 alloy foil layer, were inserted through six alternately arranged vertical slots  on the wall of the  reactor.   These panels, shown schematically in Figure 4.5(a), are 231.8 mm x 73.0 mm x 6.35 mm thick. Accounting for welding and bonding space, the active area of each membrane is 206.4 mm x 50.8 mm on each side of the membrane panels to withdraw hydrogen along the reactor height. High-purity hydrogen, metered by mass flow meters, FMA-1818 from Omega Instruments, passed through the membrane panels to a spark-proof hydrogen vacuum pump. In some experiments, stainless steel dummies of the same dimensions as the active membrane panels were installed, as explained below. Figure 4.5(b) shows a membrane panel installed onto a supporting side flange cover. Figure 4.6 depicts the experimental set-up. Before starting the experiments, the catalyst was reduced overnight at about 500°C. The required steady flow rate of steam was established before feeding heptane.  The vapour head-space in the heptane storage tank was pressurized by helium, pushing the heptane through a liquid heptane Bronkhorst mass flow controller. Distilled water was pumped, metered by a Brooks mass flow controller, and flowed through an electrically-heated vaporizer. Heptane was mixed with the steam downstream of the vaporizer. The heptane/steam mixture was fed to the FBMR through a doughnut-shaped gas distributor, located inside and at the bottom of the FBMR, with six equally-spaced holes drilled on the inner side. This allowed spent catalysts to be discharged through a catalyst drain in the bottom head cover, without completely disassembling the bottom head. The steam-to-carbon ratio in the feed was maintained by adjusting the mass flow rates of water and heptane. 4.2.2 Experimental plan and performance characterization Table 4.1 summarizes the steady reactor measurements made to characterize the reactor performance. Table 4.2 lists the location of the monitoring probes, and the location and height intervals covered by the membrane panels. The reactor performance was characterized by measuring the pure hydrogen produced and the gas compositions at different locations. The composition of the gas samples was analyzed by a Varian micro-GC CP-4900 (see Table 4.3). Table 4.4 gives key details of the experiments on the steam reforming of heptane. A steam-to- carbon ratio (SCR) of 5 was used for all experiments, except when SCR itself was a parameter. While most of these experiments maintained similar feed superficial velocities for parametric studies, some provided similar molar feeds. These experiments were conducted in three phases: Sets 1 to 3 were carried out with six membrane dummies, set 4 with five dummies and one active  129  membrane panel (at the 5th side opening from the bottom), and sets 5 to 9 with six active membranes installed. The fluidized bed reactor without membranes is comparable to a pre- reformer without removal of hydrogen. Experiments with one and six membrane panels help to elucidate the effect of hydrogen removal on the reactions inside the reactor. Steam reforming of higher hydrocarbons is very rapid, and the conversion of the higher hydrocarbons is irreversible, limited only by equilibrium of methane steam reforming. Thus, conversion of the higher hydrocarbon fed becomes irrelevant, being essentially 100% from near the entrance of industrial setups23, and also for an FBMR with heptane feed43,52. Intermediate hydrocarbons were not detected, except for traces of ethane (less than 0.01% by volume in all cases). However, ethane was not considered in the performance calculations given their low levels. Since the objective is to produce pure hydrogen, pure hydrogen yield is the most relevant performance metric. To compare the reformer with and without membranes, the total hydrogen yield, including both permeated pure and retentate hydrogen, is calculated and plotted.  The yield of carbon oxides, especially carbon dioxide, is an equivalent measure to describe the conversion of the hydrocarbons, including the intermediate. Carbon dioxide is a co-product from reactions (4.3) to (4.5). Based on the dry composition of gas samples withdrawn from the FBMR at different heights, local yields of retentate hydrogen, carbon oxides and methane are calculated: Pure hydrogen yield = rate feed heptanemolar membranes  viaextractedhydrogenpureof flowmolar   (4.7) Retentate hydrogen yield = rate feedheptanemolar stream retentatein hydrogen of flowmolar    (4.8) Total hydrogen yield = pure hydrogen yield + retentate hydrogen yield   (4.9) Carbon oxides yield = rate feed heptanemolar  x 7 stream retentatein  )CO of flowmolar   CO of flow(molar 2+        (4.10) Methane yield = rate feed heptanemolar  x 7 stream retentatein  methane of flowmolar              (4.11)  4.3 Results and Discussion 4.3.1 FBMR experiments In most experiments, two or three samples were analyzed at each location. Error bars, corresponding to the standard deviations (±σ) for each sample gas location, are plotted below with some data points shifted very slightly sideways to allow clear display.  130  For each membrane panel there is one thermocouple close to the hydrogen removal port. An average bed temperature was calculated based on the temperatures recorded at all six membrane levels. For each parametric study, the time-average bed temperature was kept constant, except where the average bed temperature was itself the study parameter. Gas samples were withdrawn from two levels for each side opening. A cubic spline function was used to estimate the temperatures corresponding to these sampling port levels.  For each parametric study, fitted temperature profiles are plotted with profiles of the carbon oxides yield, methane yield, and hydrogen yield. The heptane conversion exceeded 99% at the lowest sampling point, and was 100% for all samples above that. Hence, heptane conversion is not plotted here. For cases with one or more membranes present, the pure hydrogen and total hydrogen yields are plotted. The retentate hydrogen can be estimated from the difference between these two values. 4.3.2 Influence of key operating parameters Figure 4.7 depicts the performance of heptane steam reforming with no in-situ hydrogen removal, representing experiments 1.a and 1.b in Table 4.4. Higher temperature is seen to favour the steam reforming of methane. This is also accompanied by higher hydrogen and carbon oxides yield by favouring reaction 4.2 in the backward, and 4.3 and 4.4 in the forward, direction. Carbon dioxide was the major carbon oxide, with carbon monoxide only ~1% of the dry gas. Figure 4.8 examines the influence of reactor pressure by comparing results for experiments 2 and 1.a with identical total molar feed rates and average temperature. The experimental hydrogen yield was higher at the lower pressure of 460 kPa, as expected from thermodynamics. Correspondingly, the yield of carbon oxides was found to be higher, and of methane lower, for 460 kPa than for 725 kPa. This indicates that the experiments were thermodynamically, rather than kinetically, controlled. Figure 4.9 plots information from experiments 2, 3.a, and 3.b to show the effect of varying the steam-to-carbon molar ratio (SCR), with the same total molar feed rates. In the range of operation of these experiments, increasing steam partial pressure positively affected the conversion of the intermediate component methane, resulting in a lower methane yield. This also gave higher yields of hydrogen and carbon oxides. Higher SCR also probably enhanced gasification of any deposited carbon, thereby reducing catalyst deactivation. However, for the maximum possible hydrogen yield (see equation 4.5), an SCR of 2 is required. Thus, a higher SCR is likely to decrease the energy efficiency of the process due to the energy required to raise excess steam.  131  Figure 4.10 plots the experimental hydrogen yield against the thermodynamic equilibrium values computed corresponding to the local temperatures for experiment sets 1 to 3. The experimental data closely follow the equilibrium values, indicating that the reactor without membranes is controlled by thermodynamic equilibrium. Figure 4.11 corresponds to experiments 4.a through 4.d, where only one active membrane panel was installed with the active membrane length spanning from 1.21 to 1.41 m above the distributor. The shaded band in this figure denotes the zone where pure hydrogen is removed by the membrane. For structural similarity among all experiments, dimensionally identical stainless steel dummy plates were installed in the other five openings. Two reactor pressures were studied, with and without suction on the membrane permeation side for each level. The total molar feed rate was the same for these four runs, with identical average reactor temperatures, so that experiments 4.a and 4.b had lower residence times than 4.c and 4.d. It is seen that experiments 4.a and 4.c had similar performance. This is due to the higher driving force and higher residence time available for hydrogen permeation for 4.c, compared to 4.a, counteracted by a negative impact of the thermodynamic equilibrium for the higher reactor pressure of 4.c. This also applies to similar performance exhibited by 4.b and 4.d for the permeate side operated under vacuum (35 and 26 kPa respectively). The two runs with evacuated permeate (4.b and 4.d) showed better performances than without vacuum (4.a and 4.c). Note that the difference between these two pairs of runs became prominent after reaching the 5th flange where the single membrane panel was installed. Figures 4.12 to 4.16 correspond to experiment sets 5 to 9, each conducted with six active membrane panels along the reactor. The shaded bands in these figures represent intervals where pure hydrogen was withdrawn by membrane panels. Figure 4.12 presents the effect of reactor temperature, with increments of 25°C in the average reactor temperature. The most important reactions (reactions 4.1 to 4.4 as listed) are endothermic on an overall basis, with only the water gas shift reaction (equation 4.3) exothermic. In addition to the effect on equilibrium, an increase in membrane temperature increases hydrogen permeation (equation 4.6), shifting the reversible reactions in the forward direction. This is reflected in the higher yield of permeate hydrogen, contributing to the greater total hydrogen yield as the average reactor temperature increased. The methane yield decreased due to higher consumption of methane (equation 4.4). These trends are reflected in increased yield of carbon oxides.  132  Figure 4.13 portrays the effect of the reactor pressure (400, 600 and 800 kPa), with the average bed temperature maintained at 500°C. To keep the superficial gas velocities similar for all three pressures, the feed total molar flow rates were adjusted. The permeate side pressure was 25 kPa for all three cases, set by modulating the speed of the hydrogen vacuum pump. The total hydrogen yield decreased significantly when the pressure increased from 400 to 600 kPa, but a further increase from 600 to 800 kPa affected the hydrogen yield only marginally. Increased pressures negatively affect the equilibrium of the system, while also causing more hydrogen permeation flux due to increased pressure difference between the reactor and permeate sides. The thermodynamic effect is dominant at lower reactor pressures, but not at higher reactor pressures. This substantiates the fact that the fast kinetics of the steam reforming reactions makes the system reach local equilibrium rapidly so that the performance is limited by the membrane permeation capacity. Figure 4.14 investigates the effect of the permeate side pressure with the reactor pressure and average bed temperature fixed at 600 kPa and 475°C respectively. The feed flow rates were the same for runs 5.a, 7.a and 7.b. Little hydrogen permeated through the membranes when the permeate side was at ambient pressure (vacuum pump not operated). The hydrogen permeation rate jumped significantly when the permeate side was evacuated to 50  or 25 kPa, reflected in increases in total hydrogen yield and carbon oxides yield, and a decreasing methane yield, with greater removal of hydrogen from the reactor. In these experiments, the feed steam-to-carbon molar ratio was 5.0, while stoichiometrically only 2 is required (equation 4.5). As a result, the bulk of the reactor gas stream consists of steam, and a higher reactor pressure does not necessarily translate to higher hydrogen partial pressure inside the reactor. When the permeate side pressure was atmospheric, the local partial pressure of hydrogen on the reactor side was estimated to be between 60 and 90 kPa, depending on the local conditions, with an average of 76 kPa. Thus, there was no driving force to promote hydrogen permeation through the membranes, and no hydrogen permeation was recorded. The average local hydrogen partial pressures were estimated to have been 67 kPa for Pm = 50 kPa, and 59 kPa for Pm = 25 kPa. Accordingly, hydrogen then permeated through the membranes, due to the positive driving force. Figure 4.15 shows the effect of the steam-to-carbon molar ratio (SCR). As for the experiments with no hydrogen removal, higher steam partial pressure positively influenced the hydrogen yield. A similar effect is also seen with the six membranes installed. More methane was consumed, reflected in the dwindling methane yield with increasing SCR.  133  Figure 4.16 investigates the effect of superficial velocity.  Gas superficial velocities increase as a result of the increasing molar flow provided by the steam reforming reactions, but decrease when hydrogen is removed from the system through the membranes.  They are also affected by local temperature and pressure. Hence, the influence is described in terms of the feed molar flow rates, instead of the superficial velocity. Other operating conditions like average bed temperature, reactor pressure, permeate pressure and SCR were maintained constant for the two cases (5.b and 9) compared. Performance profiles are seen to differ near the entrance of the reactor, suggesting different hydrodynamic behaviour near the entrance. Beyond the entrance region, the performance shows only marginal differences, indicating that the overall reactor performance was dominated by the reaction equilibria. However, it is interesting to note that with an increase in the feed flow rate, the actual permeate hydrogen withdrawn also increased significantly (as the yields were almost the same at a 33% higher heptane feed). This was probably due to differences in temperature profile even though the average bed temperature was very nearly the same. It may also be due to weakening of any lateral concentration gradient, likely to be caused by hydrogen depletion near the membrane wall, at higher superficial gas velocities. In Figure 4.17, experimental hydrogen and methane yields are plotted against the corresponding equilibrium values at local temperatures without hydrogen removal. The experimental data were obtained 1.64 m above the distributor, i.e. at the top of the sixth membrane panel. With no hydrogen removal corresponding to experiment sets 1 to 3, hydrogen yield was close to, but less than the equilibrium value, whereas methane slip was more than predicted by equilibrium. For the permeate side operating  at ambient pressure, the performance did not improve much relative to cases without membranes, regardless of whether only one membrane or all six were installed. As expected, there was a significant improvement in the hydrogen and methane yields with six membranes compared with one, demonstrating that the reactor performance was dominated by the available membrane permeation area, as well as by the permeate side pressure. The carbon oxides and methane yields generally follow the temperature profile along the length of the reactor, since the gas composition in the reactor is governed by the local thermodynamic equilibrium. This has been observed for most of the sampling points along the reactor length. For each membrane interval, the effect of hydrogen permeation was apparent, with a higher carbon oxides yield and a lower methane yield at the downstream location than at the upstream one. However, a discontinuity was often (e.g. Figure 4.15) observed for the  134  methane and carbon oxides yields just beyond the second membrane panel. The molar flow rate of gas was found to vary along the reactor height, probably as a result of the uneven temperature profile, which can significantly affect the reaction rate as well as the hydrogen permeation rate. The discontinuity in the methane and carbon oxides yields may have been due to hydrodynamic effects above the second membrane. A similar smaller discontinuity appears above the fourth membrane panel as well. 4.3.3 Hydrogen purity Hydrogen purities were monitored separately for each membrane panel after each day of experiments.  In most cases, the permeate stream was ~99.99% hydrogen. However, for the fourth and sixth membranes, the purity decreased to >99.95% towards the end of the series of experiments. The pure hydrogen production rate depended on the operating conditions and feed flow rates. The highest production rate was 0.39 Nm3/h in experiment 9. 4.3.4 Discussion The experimental results show that an FBMR for heptane reforming can be operated at the industrial operating temperatures of naphtha pre-reformers, while achieving hydrogen yields comparable to a second stage steam reformer, which operates at temperatures as high as 850°C. This is because of the continuous shift of equilibrium limitation as hydrogen is progressively removed. In terms of total hydrogen yield, the FBMR gives the combined performance of a pre- reformer and a reformer. In addition, separate hydrogen purification is not needed, since pure hydrogen is available as a membrane permeate stream. Thus the FBMR combines the function of a pre-reformer, reformer, shift converter, and hydrogen purification section. However, some hydrogen is also lost in the off-gas retentate stream. Since the FBMR operating temperature is moderate, ~550°C, catalyst deactivation is minimized, both in terms of carbon formation and sintering. Moderate temperature operation also avoids expensive alloys for high-temperature tubing used in conventional industrial steam reformers. Heptane conversion exceeded 99% at the lowermost sampling point, and was complete (100%) above that. Except at the very bottom, the FBMR reaction zone sees practically no higher hydrocarbon during steam reforming of heptane. Similar behaviour was observed for steam reforming of propane53. Thus the FBMR is flexible in feedstock, similar to what is achieved by addition of a pre-reformer prior to a conventional steam reformer. However, higher hydrocarbon feedstocks require high steam-to-carbon ratios, which can affect the pressure drop  135  in the steam reformer due to variations in volumetric flow rate. Fluidized beds operate with little or no variation of bed pressure drop, although variations of superficial gas velocity may change the hydrodynamic behaviour. The FBMR process has been widely studied in the past for steam reforming of natural gas. In that case, operation at 550°C is sufficient to achieve high conversion, equivalent to that at temperatures above 800°C without membranes41. Temperatures >550°C, although not essential, could improve the hydrogen yield further by enhancing the equilibrium conversion, as well as the hydrogen permeation. The practical temperature limitation arises from the structural integrity of the membranes, which could develop pinholes or cracks. For steam reforming of liquid hydrocarbons like naphtha or its surrogate heptane, as employed in this study, the upper temperature limit is likely to be similar to that for a naphtha pre-reformer. This study used a model component to emulate steam reforming of naphtha. However, the olefinic components (which must be less than 1% by volume32) of naphtha can cause low temperature catalyst deactivation. To study the feasibility of the FBMR for naphtha steam reforming, the effects of naphthenes and aromatics must also be considered. Nevertheless, the current study provides valuable background information for higher hydrocarbon feedstocks like naphtha, gasoline, kerosene and diesel fuel.  4.4 Conclusions Steam reforming of heptane was studied in a fluidized bed membrane reactor, providing insight into the feasibility of FBMR application for hydrogen production from liquid hydrocarbon feedstocks. Experiments were conducted without and with hydrogen removal. The composition of the reactor gas samples without membranes closely followed the equilibrium values at local temperatures and pressures.  The reactor without membranes was equivalent to a pre-reformer for naphtha steam reforming. Effects of hydrogen removal were studied with one and six membrane panels installed. With hydrogen removal through selective membranes, the FBMR provides a compact reformer system, combining the pre-reformer, reformer, shift conversion and hydrogen purification steps into a single unit. The FBMR system is appropriate for steam reforming of higher hydrocarbons, since the temperature limitations of the Pd-Ag membranes closely match the usual pre-reformer temperatures to avoid catalyst deactivation by coking. The FBMR can also accept different hydrocarbon feedstocks. Hydrogen purities as high as 99.99% were achieved from individual membrane panels. The reactor was tested under different  136  operating conditions and flow rates for parametric studies. A pure hydrogen production rate of 0.39 Nm3/h was achieved at an average bed temperature of 500°C, reactor pressure of 600 kPa, permeate pressure of 25 kPa, steam-to carbon molar ratio of 5, total feed rate of 0.819 mols/min, and a total membrane permeation area of 0.126 m2. The maximum hydrogen yield was 14.7 moles of pure hydrogen (and 18.5 moles of total hydrogen) per mole of heptane fed, compared with the theoretical maximum of 22.     137  Table 4.1: Steady-state reactor measurements  Quantity Device and Location FBMR temperatures One thermocouple just above distributor. One thermocouple close to center of each membrane panel. One         thermocouple for freeboard just upstream of reformer exit. Gas composition Two sampling ports for each of the six lateral flanges supporting a         membrane panel. One sampling line for ROG. Gas sampled from these sampling points are analyzed online by a         Varian micro-GC CP-4900 using sample selection valves. Permeate hydrogen production Flow rates of permeate hydrogen from each membrane panel are         measured using FMA-1818 mass flow meters from Omega         Instruments. Purity of permeate hydrogen Hydrogen purity in permeate product from each membrane panel are         analysed by the micro-GC. Pressures Absolute pressures in the feed line, freeboard, and at the distributor         level in bed are determined using PX-309 absolute pressure         transducers from Omega Instruments. Differential pressure between alternate levels of side flanges (i.e. pairs         1-3, 3-5, 2-4, and 4-6) and between distributor and freeboard are         measured using PX-2300 differential pressure transducers from         Omega Instruments.   138  Table 4.2: Location of sampling ports, thermocouples and pure hydrogen withdrawal, and height intervals of active membrane surface   Description (Side opening counted from bottom) Location above distributor holes (m) Height interval covered by active membrane Thermocouple (Bottom) 0.01 - Thermocouple (Side opening 1) 0.32 - Thermocouple ( Side opening 2) 0.52 - Thermocouple ( Side opening 3) 0.78 - Thermocouple ( Side opening 4) 1.08 - Thermocouple ( Side opening 5) 1.29 - Thermocouple ( Side opening 6) 1.59 - Thermocouple (Freeboard) 2.33 - Gas samples ( Side opening 1) 0.22, 0.37 - Gas samples ( Side opening 2) 0.47, 0.63 - Gas samples ( Side opening 3) 0.73, 0.88 - Gas samples ( Side opening 4) 0.98, 1.13 - Gas samples ( Side opening 5) 1.24, 1.39 - Gas samples ( Side opening 6) 1.49, 1.64 - Pure hydrogen ( Side opening 1) 0.30 0.19 – 0.40 Pure hydrogen ( Side opening 2) 0.55 0.45 – 0.65 Pure hydrogen ( Side opening 3) 0.80 0.70 – 0.91 Pure hydrogen ( Side opening 4) 1.06 0.95 – 1.16 Pure hydrogen ( Side opening 5) 1.31 1.21 – 1.41 Pure hydrogen ( Side opening 6) 1.57 1.46 – 1.67  139  Table 4.3: Micro-GC column information for product gas analysis  Channel  Column Description Carrier Gas Gases Analyzed Detection limits 1 10 m molsieve 5A with pre-column backflush Argon He, H2, O2, N2, CH4 and CO 10 – 100 ppm 2 10 m PPU with pre-column backflush Helium CO2, C2H4, C2H6, C2H2, H2S and COS 10 – 100 ppm 3 8 m Silica PLOT with pre-column backflush Helium C3 and C4 isomers 10 – 100 ppm 4 8 m CP-Sil 5 with no pre-column Helium C5 to C12 components 1 – 10 ppm    140  Table 4.4: Experimental runs for steam reforming of n-heptane  Expt No. Active Membranes (Location) Total feed rate Tav  P  Pm SCR   (mols/min) (°C) (kPa) (kPa)  1.a 0.673 520 460 NA 5.0 1.b None 0.766 450 460 NA 5.0  2 None 0.673 520 725 NA 5.0  3.a 0.673 520 725 NA 4.0 3.b None 0.673 520 725 NA 6.0  4.a 0.717 480 585 101 5.0 4.b 0.717 480 585 35 5.0 4.c 0.717 480 720 101 5.0 4.d 1 (#5) 0.717 480 720 26 5.0  5.a 0.635 475 600 25 5.0 5.b 0.614 500 600 25 5.0 5.c 6 (#1 to #6) 0.595 525 600 25 5.0  6.a 0.410 500 400 25 5.0 6.b 6 (#1 to #6) 0.819 500 800 25 5.0  7.a 0.635 475 600 101 5.0 7.b 6 (#1 to #6) 0.635 475 600 50 5.0  8.a 0.614 500 600 25 4.0 8.b 6 (#1 to #6) 0.614 500 600 25 6.0  9 6 (#1 to #6) 0.819 500 600 25 5.0    141  Hydrocarbon Feed Fuel Steam Sulfur Removal Pre-Reformer Tubular Reformer Flue Gas Channel MT Shift Converter H2 Product PSA Reformer Offgas Figure 4.1: Key components in a modern steam reforming plant for hydrogen from higher hydrocarbon feedstock.  (Adapted from Rostrup-Nielsen and Rostrup-Nielsen1)  142  0 10 20 30 40 50 60 70 80 400 450 500 550 600 650 700 750 800 Temperature (°C) M ol e % Heptane Methane CO2 CO H2 (a)           SCR  = 5.0, P = 400 kPa   0 10 20 30 40 50 60 70 80 400 450 500 550 600 650 700 750 800 Temperature (°C) M ol e % Heptane Methane CO2 CO H2 (b)           SCR  = 5.0, P = 800 kPa  Figure 4.2: Dry gas equilibrium composition for steam-to-carbon molar ratio of 5.0: (a) P = 400 kPa; (b) P = 800 kPa. No membranes present  143  0 10 20 30 40 50 60 70 80 400 450 500 550 600 650 700 750 800 Temperature (°C) M ol e % Heptane Methane CO2 CO H2 (a)           SCR  = 4.0, P = 400 kPa   0 10 20 30 40 50 60 70 80 400 450 500 550 600 650 700 750 800 Temperature (°C) M ol e % Heptane Methane CO2 CO H2 (b)           SCR  = 6.0, P = 400 kPa  Figure 4.3: Dry gas composition for reactor pressure of 400 kPa: (a) Steam-to-carbon molar ratio = 4.0; (b) Steam-to-carbon molar ratio = 6.0. No membranes present  144     Figure 4.4: Drawing of FBMR pressure vessel supported on mobile stand  145   6. 35  m m  O D  S S tu be 231.8 mm 206.4 mm 73 .0  m m 6.35 mm Pd77Ag23 Membrane Foil Membrane Panel Frame 50 .8  m m  (a)  Membrane panel Supporting flange coverGas sample Gas sample Pressure port Permeate line Thermocouple (b) Figure 4.5: (a) Dimensions of membrane panel. (b) Ports arranged on each side-opening cover where membrane panels are installed  146   Liquid Water Helium Heptane Tank Heptane MFC Liquid Heptane Va po riz er Water Tank Water Pump Water MFC Permeate 2 Reformer Feed Sample 2a Sample 2b Permeate 4 Sample 4a Sample 4b Permeate 6 Sample 6a Sample 6b Permeate 1 Permeate 3 Permeate 5 Reformer Off Gas Sample 1a Sample 1b Sample 3a Sample 3b Sample 5a Sample 5b Sample ROG FBMR H2 MFM H2 MFM H2 MFM H2 MFM H2 MFM H2 MFM H2 MFM H2 Vacuum Pump    Figure 4.6: Schematic of experimental set-up to study steam reforming of n-heptane  147   C ar bo n O xi de s Y ie ld 0.3 0.5 0.7 H yd ro ge n Y ie ld 3 7 11 Lo ca l Te m p.  (o C ) 400 450 500 550 600 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.30 0.45 0.60 Tav = 520 oC Tav = 450 oC   Figure 4.7: Experimental yields and temperature for heptane steam reforming without active membrane panels at reactor pressure of 470 kPa and steam-to-carbon ratio molar ratio of 5.0. Total reactor feed = 0.673 and 0.766 mols/min at 520 and 450°C respectively  148   C ar bo n O xi de s Y ie ld 0.3 0.5 0.7 H yd ro ge n Y ie ld 3 7 11 Lo ca l Te m p.  (o C ) 450 500 550 600 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.30 0.45 0.60 P = 460 kPa P = 724 kPa   Figure 4.8: Experimental yields and temperature for heptane steam reforming without active membrane panels at average reactor temperature of 520°C and steam-to-carbon ratio molar ratio of 5.0. Total reactor feed = 0.673 mols/min  149   C ar bo n O xi de s Y ie ld 0.3 0.5 0.7 H yd ro ge n Y ie ld 3 7 11 Lo ca l Te m p.  (o C ) 450 500 550 600 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.30 0.45 0.60 SCR = 4 SCR = 5 SCR = 6    Figure 4.9: Experimental yields and temperature for heptane steam reforming without active membrane panels at average reactor temperature of 520°C and reactor pressure of 725 kPa. Total reactor feed = 0.673 mols/min  150  0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 Equilibrium H2 Yield Ex pe rim en ta l H 2 Y ie ld Expt 1.a Expt 1.b Expt 3.a Expt 2.b Expt 3.c (a)  0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Equilibrium CH4 Yield Ex pe rim en ta l C H 4 Y ie ld Expt 1.a Expt 1.b Expt 3.a Expt 2.b Expt 3.c (b)   Figure 4.10: Parity plot of experimental yields without active membrane panels against local equilibrium values: (a) Hydrogen yield (b) Methane yield  151  Lo ca l Te m p.  (o C ) 480 530 580 C ar bo n O xi de s  Y ie ld 0.35 0.45 0.55 P = 585 kPa, Pm = 101 kPa P = 585 kPa, Pm = 35 kPa P = 720 kPa, Pm = 101 kPa P = 720 kPa, Pm = 26 kPa To ta l H yd ro ge n  Y ie ld 4 6 8 10 Pu re  H yd ro ge n  Y ie ld 0.0 1.5 3.0 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.3 0.4 0.5 0.6   Figure 4.11: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 480°C and steam-to-carbon molar ratio 5.0. One membrane panel installed, spanning from 0.95 to 1.16 m above distributor. Total reactor feed = 0.717 mols/min  152   Lo ca l Te m p.  (o C ) 450 500 550 600 650 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 Tav = 475 oC Tav = 500 oC Tav = 525 oC To ta l H yd ro ge n  Y ie ld 7 12 17 Pu re  H yd ro ge n  Y ie ld 1 5 9 13 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 4.12: Experimental yields and temperature for heptane steam reforming at pressure of 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.635, 0.614, and 0.595 mols/min for 475, 500, and 525°C respectively  153   Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 P = 400 kPa P = 600 kPa P = 800 kPa To ta l H yd ro ge n  Y ie ld 7 12 17 Pu re  H yd ro ge n  Y ie ld 1 4 7 10 13 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 4.13: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 500°C, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feeds = 0.410, 0.614, and 0.819 mols/min for P = 400, 600, and 800 kPa respectively  154   Lo ca l Te m p.  (o C ) 460 490 520 550 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 Pm = 101 kPa Pm = 50 kPa Pm = 25 kPa To ta l H yd ro ge n  Y ie ld 5 9 13 Pu re  H yd ro ge n  Y ie ld 1 3 5 7 9 11 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75  Figure 4.14: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 475°C, pressure 600 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed. Total reactor feed = 0.635 mols/min  155   Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 SCR = 4 SCR = 5 SCR = 6 To ta l H yd ro ge n  Y ie ld 7 12 17 Pu re  H yd ro ge n  Y ie ld 1 5 9 13 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 4.15: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, and permeate pressure 25 kPa. Six membrane panels installed. Total reactor feed = 0.614 mols/min  156   Lo ca l Te m p.  (o C ) 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 Total Feed = 0.614 mols/min Total Feed = 0.819 mols/min To ta l H yd ro ge n  Y ie ld 7 12 17 Pu re  H yd ro ge n  Y ie ld 1 5 9 13 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  Y ie ld 0.00 0.25 0.50 0.75   Figure 4.16: Experimental yields and temperature for heptane steam reforming at average reactor temperature of 500°C, pressure 600 kPa, permeate pressure 25 kPa, and steam-to-carbon molar ratio 5.0. Six membrane panels installed  157  0 4 8 12 16 20 0 4 8 12 16 20 Equilibrium H2 Yield Ex pe rim en ta l H 2 Y ie ld   Without Membrane Panels   1 Membrane Panel   6 Membrane Panels (a) P m  = 25 kPa P m  = 101 kPa P m  = 50 kPa P m  = 26 kPa P m  = 35 kPa  0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 Equilibrium CH4 Yield Ex pe rim en ta l C H 4 Y ie ld   Without Membrane Panels   1 Membrane Panel   6 Membrane Panels (b) P m  = 101 kPa P m  = 35 kPa P m  = 50 kPa P m  = 26 kPa P m  = 25 kPa   Figure 4.17: Parity plot of experimental yields against equilibrium values at local temperatures if there was no hydrogen removal: (a) Hydrogen yield (b) Methane yield   158  4.5 References 1. Rostrup-Nielsen, J. R.; Rostrup-Nielsen, T., Large-scale hydrogen production. CatTech 2002, 6, (4), 150-159. 2. Hallale, N.; Liu, F., Refinery hydrogen management for clean fuels production. Advances in Environmental Research 2001, 6, (1), 81-98. 3. Fonseca, A.; Sá, V.; Bento, H.; Tavares, M. L. C.; Pinto, G.; Gomes, L. A. C. N., Hydrogen distribution network optimization: a refinery case study Journal of Cleaner Production 2008, 16, (16), 1755-1763. 4. Basini, L., Issues in H2 and synthesis gas technologies for refinery, GTL and small and distributed industrial needs. Catalysis Today 2005, 106, (1-4), 34-40. 5. Johnston, B.; Mayo, M. C.; Khare, A., Hydrogen: The energy source for the 21st century. Technovation 2005, 25, (6), 569-585. 6. Ball, M.; Wietschel, M., The future of hydrogen - opportunities and challenges. International Journal of Hydrogen Energy 2009, 34, (2), 615-627. 7. Winter, C. J., Electricity, hydrogen - competitors, partners? International Journal of Hydrogen Energy 2005, 30, (13-14), 1371-1374. 8. Winter, C. J., Into the hydrogen energy economy - milestones. International Journal of Hydrogen Energy 2005, 30, (7), 681-685. 9. Winter, C. J., Hydrogen energy - Abundant, efficient, clean: A debate over the energy- system-of-change. International Journal of Hydrogen Energy 2009, 34, (14), S1-S52. 10. Ewan, B. C. R.; Allen, R. W. K., A figure of merit assessment of the routes to hydrogen. International Journal of Hydrogen Energy 2005, 30, (8), 809-819. 11. Trimm, D. L.; Adesina, A. A.; Praharso; Cant, N. W., The conversion of gasoline to hydrogen for on-board vehicle applications. Catalysis Today 2004, 93-95, 17-22. 12. Avci, A. K.; Onsan, Z. I.; Trimm, D. L., On-board fuel conversion for hydrogen fuel cells: comparison of different fuels by computer simulations. Applied Catalysis A: General 2001, 216, (1-2), 243-256. 13. Kaila, R. K.; Krause, A. O. I., Reforming of higher hydrocarbons. Studies in Surface Science and Catalysis 2004, 147, 247-252 14.http://www.airproducts.com/Products/MerchantGases/HydrogenEnergy/FrequentlyAskedQue stions.htm. Date accessed: 10 November 2009.  159  15. Shumake, G.; Small, J., Mixing things up in hydrogen plants. Hydrocarbon Engineering 2006, 11, 49-52. 16. Rostrup-Nielsen, J. R., Activity of nickel catalysts for steam reforming of hydrocarbons. Journal of Catalysis 1973, 31, (2), 173-199. 17. Rostrup-Nielsen, J. R., Industrial relevance of coking. Catalysis Today 1997, 37, (3), 225-232. 18. Rostrup-Nielsen, J. R.; Hojlund-Nielsen, P. E., Catalyst deactivation in synthesis gas production, and important syntheses. In Deactivation and poisoning of catalysts, Oudar, J.; Wise, H., Eds. Marcel Dekker Inc.: New York, 1985; pp 259-323. 19. Bartholomew, C. H., Mechanisms of catalyst deactivation. Applied Catalysis A:General 2001, 212, (1-2), 17-60. 20. Sehested, J., Four challenges for nickel steam-reforming catalysts. Catalysis Today 2006, 111, (1-2), 103-110. 21. RostrupNielsen, J. R., Innovation and the catalytic process industry - The science and the challenge. Chemical Engineering Science 1995, 50, (24), 4061-4071. 22. Rostrup-Nielsen, J. R.; Sehested, J., Whisker carbon revisited. Studies in Surface Science and Catalysis 2001, 139, 1-12. 23. Christensen, T. S., Adiabatic prereforming of hydrocarbons - An important step in syngas production. Applied Catalysis A: General 1996, 138, (2), 285-309. 24. Gunardson, H., Industrial gases in petrochemical processing Marcel Dekker: New York, 1998. 25. Gomach, J., Personal communication, Sales & technical service engineer, Syngas catalyst & technology, Haldor Topsoe, Inc. In 2007. 26. Haldor Topsoe A/S Manual for Topsoe naphtha reforming catalysts, RK-212/ RK-202/ RK-201. 27. Melo, F.; Morlanes, N., Naphtha steam reforming for hydrogen production. Catalysis Today 2005, 107-08, 458-466. 28. Rostrup-Nielsen, Catalytic steam reforming. In Catalysis Science and Technology, Andersen, J. R.; Boudart, M., Eds. Springer-Verlag: 1984; pp 1-117 29. Rostrup-Nielsen, J. R.; Christensen, T. S.; Dybkjaer, I., Steam reforming of liquid hydrocarbons. Recent Advances in Basic and Applied Aspects of Industrial Catalysis 1998, 113, 81-95. 30. Topsoe technology for hydrogen production, ; Haldor Topsoe A/S.  160  31. Rostrup-Nielsen, J. R., Syngas in perspective. Catalysis Today 2002, 71, (3-4), 243-247. 32. Ridler, D. E.; Twigg, M. V., Steam reforming. In Catalyst Handbook, Twigg, M. V., Ed. Wolfe Publishing Ltd.: London, U.K., 1989. 33. Sircar, S.; Golden, T. C., Purification of hydrogen by pressure swing adsorption. Separation Science and Technology 2000, 35, (5), 667-687. 34. Elnashaie, S. S. E. H.; Adris, A. M., A fluidized bed steam reformer for methane. In Fluidization VI, Grace, J. R.; Shemilt, L. W.; Bergougnou, M. A., Eds. Engineering Foundation: 1989; pp 319-326. 35. Soliman, M. A.; Elnashaie, S. S. E. H.; Al-Ubaid, A. S.; Adris, A., Simulation of steam reformers for methane. Chemical Engineering Science 1988, 43, 1801-1806. 36. Brunetti, A.; G.Barbieri; E.Driolia, Upgrading of a syngas mixture for pure hydrogen production in a Pd-Ag membrane reactor. Chemical Engineering Science 2009, 64, 3448--3454. 37. Matsumura, Y.; Tong, J. H., Methane steam reforming in hydrogen-permeable membrane reactor for pure hydrogen production. Topics in Catalysis 2008, 51, (1-4), 123-132. 38. Damle, A. S., Hydrogen production by reforming of liquid hydrocarbons in a membrane reactor for portable power generation - Experimental studies. Journal of Power Sources 2009, 186, (1), 167-177. 39. Shirasaki, Y.; Tsuneki, T.; Ota, Y.; Yasuda, I.; Tachibana, S.; Nakajima, H.; Kobayashi, K., Development of membrane reformer system for highly efficient hydrogen production from natural gas. International Journal of Hydrogen Energy 2009, 34, (10), 4482-4487. 40. Adris, A. M.; Lim, C. J.; Grace, J. R., The fluidized bed membrane reactor system: A pilot scale experimental study. Chemical Engineering Science 1994, 49, (24B), 5833-5843. 41. Boyd, T.; Grace, J.; Lim, C. J.; Adris, A. E. M., Hydrogen from an internally circulating fluidized bed membrane reactor. International Journal of Chemical Reactor Engineering 2005, 3, A58. 42. Deshmukh, S. A. R. K.; Heinrich, S.; Morl, L.; Annaland, M. V. S.; Kuipers, J. A. M., Membrane assisted fluidized bed reactors: Potentials and hurdles. Chemical Engineering Science 2007, 62, (1-2), 416-436. 43. Rakib, M. A.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J.; Bolkan, Y. G., Kinetic simulation of a compact reactor system for hydrogen production by steam reforming of higher hydrocarbons. Canadian Journal of Chemical Engineering 2008, 86, (3), 403-412. 44. Chen, Z. A novel circulating fluidized bed membrane reformer for efficient pure hydrogen production for fuel cells from higher hydrocarbons. Auburn University, Auburn, 2004.  161  45. Chen, Z.; Yan, Y.; Elnashaie, S. S. E. H., Modeling and optimization of a novel membrane reformer for higher hydrocarbons. AIChE Journal 2003, 49, 1250-1265. 46. Darwish, N. A.; Hilal, N.; Versteeg, G.; Heesink, B., Feasibility of the direct generation of hydrogen for fuel-cell-powered vehicles by on-board steam reforming of naphtha. Fuel 2004, 83, (4-5), 409-417. 47. Tottrup, P. B., Evaluation of intrinsic steam reforming kinetic parameters from rate measurements on full particle size. Applied Catalysis 1982, 4, (4), 377-389. 48. Paglieri, S. N.; Way, J. D., Innovations in palladium membrane research. Separation and Purification Methods 2002, 31, (1), 1-169. 49. Uemiya, S., State-of-the-art of supported metal membranes for gas separation. Separation and Purification Methods 1999, 28, (1), 51-85. 50. Sieverts, A.; Zapf, G., The solubility of deuterium and hydrogen in solid palladium. Zeitschrift für Physikalische Chemie 1935, 174, 359-364. 51. Li, A.; Grace, J. R.; Lim, C. J., Preparation of thin Pd-based composite membrane on planar metallic substrate - Part II. Preparation of membranes by electroless plating and characterization. Journal of Membrane Science 2007, 306, (1-2), 159-165. 52. Chen, Z.; Yan, Y. B.; Elnashaie, S. S. E. H., Modeling and optimization of a novel membrane reformer for higher hydrocarbons. AIChE Journal 2003, 49, (5), 1250-1265. 53. Rakib, M. A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H.; Ghiasi, B., Steam reforming of propane in a fluidized bed membrane reactor for hydrogen production. International Journal of Hydrogen Energy 2010 (In Press).    162  CHAPTER 5.      MODELING OF A FLUIDIZED BED MEMBRANE REACTOR FOR HYDROGEN PRODUCTION BY STEAM REFORMING OF HYDROCARBONS∗   5.1 Introduction 5.1.1 Hydrogen from higher hydrocarbons feedstock Industrial use of hydrogen is projected to increase due to rising fertilizer demand1 and increased hydrotreating requirements of various feedstocks in refineries2 as the available crude becomes heavier and increasingly sour. Hydrogen is also often foreseen as a major energy carrier, whose implementation could help to mitigate global warming due to greenhouse gas emissions from direct use of fossil fuels3. To make hydrogen readily available for automobiles, a distributed network of small-scale to medium-scale hydrogen production units is needed4. Steam reforming of hydrocarbons is the major global pathway for hydrogen production2,5,6. The feedstock consumptions for hydrogen production are 48% natural gas and 30% oil/ naphtha7. Compared to higher hydrocarbons like naphtha, natural gas is favored, mainly due to its widespread availability and lower cost. However, higher hydrocarbon feedstocks are used in places where natural gas is not available. Refineries also prefer flexible feedstock options to take advantage of seasonal surplus products or off-gases rich in higher hydrocarbons. For distributed small and medium scale hydrogen generation units, liquid hydrocarbon feedstocks may be advantageous due to (a) higher volumetric hydrogen density, (b) existing infrastructure of propane/ gasoline/ naphtha fuels, and (c) ease of storage and transportation at or near ambient conditions, compared to natural gas. Traditionally, higher hydrocarbon feedstocks were mainly naphtha, operated at feed temperatures of about 450 to 550°C, and a product temperature of ~750 to 850°C, in order to minimize catalyst deactivation by carbon formation, a problem which becomes more serious as the carbon number of the reforming feedstock increases.  ∗ A version of this chapter has been submitted for publication: Rakib, M.A., Grace, J.R., Lim, C.J., and Elnashaie, S.S.E.H., Modeling of a Fluidized Bed Membrane Reactor for Hydrogen Production by Steam Reforming of Hydrocarbons (2010).    163  Recent versions of industrial naphtha-based steam reforming systems are installed with a pre-reformer, operated at relatively lower temperatures of 450 to 550°C, followed by a reformer at about 800 to 900°C. An important advantage of the pre-reformer is flexibility with relatively low impacts on the reformer operation, since the higher hydrocarbons are reformed in the pre- reformer to methane-rich gas feed for the reformer. 5.1.2 Fluidized Bed Membrane Reactors (FBMR) Steam reforming reactions feature fast kinetics. High intra-particle diffusional limitations lead to very low effectiveness factors. Fine catalyst powders are therefore useful to reduce internal mass transfer limitations. These fine catalysts can be deployed in fluidized bed reactors, which also help to minimize heat transfer limitations for the highly endothermic reactions like steam reforming. Permselective Pd membranes enhance the hydrogen yield by shifting the equilibrium of the reaction. Fluidized bed membrane reactors have been studied experimentally for the production of pure hydrogen by steam reforming of methane or natural gas8-11. Other processes studied experimentally in an FBMR, mostly in the bubbling flow regime, include oxidative dehydrogenation of ethane to ethylene12, partial oxidation of butane to maleic anhydride13, and partial oxidation of methanol to formaldehyde14. FBMRs have also been modelled for hydrogen production and other processes. A comprehensive review has been provided by Deshmukh et al.15.  Fluidized bed membrane reactors for steam reforming of heptane in circulating fast fluidized bed mode of operation were modeled by Chen et al.16-19. Catalyst deactivation by carbon formation could be transformed into an advantage if the catalysts were regenerated by combustion in a separate regenerator, and the regenerated hot catalysts were recycled to the riser reformer, thereby enabling autothermal operation. One-dimensional two-phase models are the most widely used to represent bubbling fluidized beds. A two-phase model originally proposed by Toomey and Johnstone20 considers the dense fluidized bed comprised of two pseudo-phases: a bubble phase, containing very few particles, and a dense phase which contains most of the solids. The flow of gas required to maintain minimum fluidization velocity goes to the dense phase, while gas flow in excess of this amount appears as bubbles. FBMR models based on this approach have treated the two phases differently, with the bubble phase treated as a plug flow in most cases. Rakib et al.21 wrote a one dimensional two-phase model for heptane steam reforming in an FBMR to size and predict its performance. Subsequently, the FBMR system was built,  164  installed and operated safely22. This paper deals with a modeling approach in order to understand the various phenomena taking place during steam reforming of higher hydrocarbons in an FBMR.  5.2 Description Membrane panels or dummies were immersed vertically along the reactor height, the main section of which has a rectangular cross-section. Figure 5.1 gives a schematic of the reactor geometry. Table 5.1 provides key reactor physical details and membrane permeation parameters. A detailed description of the experimental setup, operating conditions and experimental results are provided elsewhere22-24. Figure 5.2 shows a schematic of the model considerations in this work. 5.2.1 Reactions and rate equations The FBMR has been studied for steam reforming of heptane as a model component for naphtha, and propane as a key component of LPG. Heptane steam reforming C7H16 + 7H2O → 7CO + 15H2   ∆H°298 = 1108 kJ/mol   (5.1) Propane steam reforming C3H8 + 3H2O → 3CO + 7H2    ∆H°298 = 499 kJ/mol   (5.2) With the appearance of CO and H2 as products, the following reactions also occur: Methanation (reverse of methane steam reforming) CO + 3H2  '  CH4 + H2O    ∆H°298 = - 206 kJ/mol   (5.3) Water gas shift CO + H2O  '  CO2 + H2     ∆H°298 = - 41 kJ/mol   (5.4) Methane overall steam reforming CH4 + 2H2O  '  CO2 + 4H2    ∆H°298 = 165 kJ/mol   (5.5) Side reactions involving carbon formation which deactivates the catalyst are not considered in this model. The kinetic rate equations governing these reactions are listed in Tables 5.2 and 5.3. 5.2.2 Model simplifications (1) Steady state conditions. (2) Ideal gas law. (3) Solids (catalyst) temperature identical to the local gas temperature.  165  (4) Intra-particle gas diffusional limitations are ignored: Catalyst particles of Sauter mean diameter 179 µm were used in the FBMR experiments, leading to effectiveness factors close to unity8,25,26. (5) No energy balance is included in the model. Instead, the experimentally recorded temperatures were used to estimate axial temperature profiles, which were then imposed on the reactor in the model. The FBMR was heated by: (a) Four internal cable heaters above the feed distributor, (b) Six semi-circular external band heaters mounted opposite to each lateral flange opening, (c) Two external circular band heaters in the freeboard zone, and (d) Six external strip heaters mounted vertically on each lateral flange. The temperature at any position along the height of the reactor depended on the location and power output of each of the heaters, and non-uniform heat loss from different sections of the FBMR pressure vessel. Figure C.7 shows the arrangement of the external band heaters. The limited axial solids mixing due to relatively low superficial gas velocities (in the range of 0.06 to 0.12 m/s, see Appendix E) and slug flow was the probable reason for the temperature not being as uniform as would normally be expected in fluidized beds. (6) Permselective membrane, with infinite selectivity for hydrogen permeation: The poorest purity obtained in our experiments was 99.95% hydrogen. (7) A two-phase model is adopted, with the dense catalyst bed treated as two pseudo-phase compartments in parallel. Given the high aspect ratio of the bed, each phase is treated as a plug flow reactor with exchange between the two compartments. The membranes withdraw hydrogen from both phases. (8) Catalyst deactivation is neglected: A base case steam-to-carbon molar ratio of 5.0 was used for the experimental runs. As recommended by the supplier (Haldor Topsoe A/S) of the RK- 212 catalyst, this is adequate to minimize catalyst deactivation by carbon formation. 5.2.3 Fluidized bed hydrodynamic model Since the bubble size and local gas superficial velocities vary axially due to variations of temperature, pressure, total molar flow, and diffusive and convective mass transport, as described below, the bed expansion is calculated iteratively. Researchers have often found that the two-phase theory overestimates the volume of gas passing through the bed as “visible” bubbles27,28. The gas flow rate in the bubbles is often multiplied by a factor Y (≤ 1.0) to allow for deviations observed experimentally from the two- phase theory, due to greater flow through the dense phase and/or increased flow through the  166  bubbles. Y has typically been found to be in the range 0.8 < Y < 1.0 for Group A powders and 0.6 < Y < 0.8 for Group B powders29,30. Gas split between the two phases at reactor inlet: The entrance region close to the distributor is also treated by means of the two-phase model in this work, although separate treatment of this zone, referred to as the grid zone, is also common31,32. Thus, the feed gas splits and distributes into the two phases at the entrance as: ( ) inmfinb UUAYQ |.., −=          (5.6) inbinind QUAQ ,, . −=           (5.7) with AP TRF U in in n i ini in . . 1 , ⎟⎠ ⎞⎜⎝ ⎛ = ∑ =          (5.8) Hydrodynamic equations: The minimum fluidization velocity is calculated to correspond to the local conditions based on the Wen & Yu correlation33, with the constants as suggested by Grace34: Umf  = ( )[ ] 2.270408.02.27 212 −+⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ Ar d gp g ρ µ       (5.9) The fraction of bed occupied by bubbles is given by: ( ) b mf b U UUY −= 0.ε           (5.10) where, the bubble velocity is given by: ( ) ( ) 210 711.0. bmfb gdUUYU +−=         (5.11) In this work, Y = 1 is assumed. Almost all particles are in the dense phase, with only a very small fraction in the bubbles. The solids content in the bubbles is commonly neglected, but their contribution to reaction in a fluidized bed reactor may be important for fast reaction kinetics. The bubbles may contain about 0.1-1.0% solids by volume30,35. For this study, the solids volume fraction is taken35 as: bb εφ 001.0=              (5.12) Assuming the dense phase voidage to be constant and equal to εmf, the volume fraction of solids in the dense phase is given by: ( )( )mfbd εεφ −−= 11           (5.13) The bubble size is estimated from the semi-empirical equation of Darton et al.36:  167  ( ) 2.08.04.0 454.0 −⎟⎠⎞⎜⎝⎛ +−= gNAhUUd ormfb        (5.14) with a maximum bubble size calculated 37 as: ( )[ ] 4.0064.1 mfbm UUAd −=          (5.15) The bubble surface area per unit volume of the bubble is then approximated by: b b d a 6=             (5.16) Interphase diffusional mass transfer: The bubble phase contains very few catalyst particles, so that very little reaction usually takes place there. Most of the reaction takes place in the dense phase which is rich in catalyst. This can cause significant concentration differences between the phases, leading to an interphase diffusion mass transfer between the two phases. The interphase mass exchange coefficient is estimated by the correlation of Sit & Grace38. For the ith component: 5.04 3 ⎥⎦ ⎤⎢⎣ ⎡+= b bmfiemf iq d UDU k π ε          (5.17) where Die is the effective diffusivity of component i in the gas mixture, calculated based on the average composition of the bubble and the dense phases. Based on the correlation of Wilke39: ( ) ji D y D y n i ij i ie i ≠⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛=− ∑ =1 1         (5.18) where Dij is the binary diffusivity of component i in j. Interphase bulk mass convection: Due to gas exchange between the two phases by diffusion due to a concentration gradient and simultaneous removal of one component (hydrogen) from both phases by permeation through membranes, flow rates in the dense phase could vary, becoming greater or less than required for minimum fluidization. If less, this could cause de-fluidization. We assume, however, that fluidization is maintained by bulk convection between the bubbles and dense phase so that the dense phase interstitial velocity always satisfies the minimum requirement. In modeling studies, this has been treated by introducing an inter-phase bulk convection term25,40,41, equal in amount to the excess or deficit of the flow required to fluidize the dense phase. With U and Umf calculated at a given height, the flow requirement in the dense phase at that position can be written as: ( )bmfreqd AUQ ε−= 1,          (5.19)  168  At a given height, the composition of the bulk inter-phase convective flow matches that of the source phase42. The volumetric exchange terms can be written as: when reqd N i di QFP TR C , 1 , . >∑ =  ,  reqd N i dibd QFP TRQ C , 1 , . −= ∑ = → 0=→dbQ        (5.20) when reqd N i di QFP TR C , 1 , . ≤∑ =  ,  ∑ = → −= CN i direqddb FP TRQQ 1 ,, . 0=→bdQ        (5.21) As a reasonable, though unproven, means of balancing the phase flows, any increase of flow of a species by bulk convection to a receiving phase is accounted for by a corresponding decrease in the species molar flow rate from the source phase. Hydrogen removal from the catalyst bed: Hydrogen permeation through Pd-based membranes occurs via a solution-diffusion mechanism 43. When the diffusion of atomic hydrogen through the solution is the rate-limiting step, the hydrogen flux follows Sieverts’ law43, so that: ( ),pH,bHH H M bHm PPRT E δ PAQ P 22 2 2 2 exp.. 0', −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −= α       (5.22) ( ),pH,dHH H M dmH PPRT E δ PAQ P 22 2 2 2 exp.. 0', −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −= α       (5.23) The permeation parameters, provided by the supplier of the membrane panels, based on permeation testing with helium/ hydrogen mixtures, are included in Table 5.1. The membranes are assumed to be impermeable to species other than hydrogen. The permeation effectiveness factor, α, accounts for blockage of the membranes by dust or foulant. It is treated below as an adjustable parameter to fit the simulated permeation hydrogen yields to the experimental results, with 1 ≥ α ≥ 0. Mole balance equations for dense catalyst bed: For each phase, Mole balance for ith species in bubble phase:   169  (5.24) where i = C7H16, C3H8, CH4, H2O, CO, CO2, and H2, and bimQ , = 0 for i  ≠ H2 Mole balance for ith species in dense phase:  (5.25) where i = C7H16, C3H8, CH4, H2O, CO, CO2, and H2, and dimQ , = 0 for i  ≠ H2 Model equations for membrane permeate side: The differential mole balance equation for the permeate hydrogen is: ( ) dHmbbHmbpH QQdh dF ,, ,2 22 1 εε −+=        (5.26) Freeboard Calculations: Catalysts particles are entrained above the expanded dense bed surface into the freeboard after being ejected by bursting bubbles. The flux of entrained particles decreases with height in the freeboard, depending on the gas velocity, as well as the gas and particle properties. In order to minimize catalyst loss due to entrainment of fines, a sintered metal filter (12.7 mm OD, 152 mm long, media grade 40) was located at the exit of the FBMR, attached to the top flange cover. A layer of catalyst fines could accumulate onto the filter medium surface, and in one case, as shown in Figure 5.3, a very loosely-bound cake of catalyst fines, retained on this filter, was retrieved intact. In most cases, however, no filter cake was found when the top cover of the FBMR was opened after operating the equipment. A thermocouple, installed close to the exit as seen in Figure 5.3, gives an indication of the temperature of the filter cake. In order to make reasonable predictions of overall conversion and exit product distributions, it was found essential to account for catalytic reaction in the freeboard, due to the catalyst dispersion there. An amount of catalyst equivalent to 0.8 mm of static bed depth was assumed to be distributed uniformly in the freeboard region. This assumption was made on the basis of least squares error minimization with respect to the experimental yields of methane, CO2, and H2 in the reformer off-gas (ROG). The freeboard was then modeled as a single-phase dilute catalyst suspension. The inlet species flow rate to the freeboard region is estimated as the sum of species flow rates from the dense and bubble phase at the dense bed surface:  170  at 0=fbh , dibifbi FFF ,,, +=          (5.27) Mole balance for ith species in freeboard: ∑ = = NR j bjijpfb fb fbi RA dh dF 1 , , γρφ          (5.28) 5.3 Model Predictions versus Experimental Results 5.3.1 Experimental data for comparison with model predictions An in-house code was written using Matlab, version 7.3 (2006), to solve the model equations. The differential equations were solved using a built-in variable order stiff ordinary differential equations solver, ode15s. A non-negativity criterion for the components molar flow rates was imposed on the solution to impart stability to the solutions. Relative and absolute tolerance values of 1 x 10-8 were used for the solver. In our earlier experimental work, heptane23, propane24 and methane were steam reformed in an FBMR. The experiments with each hydrocarbon were conducted in three different combinations of dimensionally identical dummies and active membrane panels: (i) Six dummy panels, (ii) One active membrane panel installed in the fifth lateral opening from the bottom, the other five being dummies, and (iii) Six active membrane panels. Experimental results are tabulated in Appendix G. The experimental details of the runs used to compare the experimental profiles with the model predictions are listed in Table 5.4. Experimental data were collected only after steady state operation was achieved. After fixing the operating conditions with respect to FBMR pressure, permeate side pressure, temperature profile, and feed flow rates, the sample gas concentration was monitored by the micro-GC. Steady state was assumed to be attained when the gas composition was seen to oscillate with absolute deviations less than 1%. Simulated dry molar gas compositions are compared with experimental gas compositions analyzed by the micro-GC after condensing the moisture from the sample gas streams. The following quantities are calculated to assess the reactor performance: Pure hydrogen yield = stream feedin nhydrocarbo of flowmolar membranes  viaextractedhydrogenpure of flowar mol    (5.29) Retentate hydrogen yield = stream feedin n hydrocarbo of flowmolar stream retentatein hydrogen of flowmolar    (5.30) Total hydrogen yield = Pure hydrogen yield + Retentate hydrogen yield   (5.31)  171  Carbon oxides yield = ( ) stream feedin  n)hydrocarbo(in carbon  of flowMolar stream retentatein  CO of flowmolar   CO of flowmolar 2+    (5.32) Methane yield = stream feedin  n)hydrocarbo(in carbon  of flowMolar stream retentatein  methane of flowmolar    (5.33) 5.3.2 Membrane effectiveness factor A genetic algorithm-based optimization routine was used to estimate the membrane permeation effectiveness factor, α, assumed to be the same for all 6 membrane panels. There is a probability that individual permeation effectiveness factors may differ from one membrane to another, e.g. due to thickness variations in the foil or hydrodynamic changes with height which could lead to variations in the thickness of any coating accumulated on the membrane foils (see Figure 5.4). The least squares fitted value of α = 0.248 was used for the current model. 5.3.3 Test results with no membrane panels Figure 5.5 shows the results for experiment 1.b where heptane was the feedstock23 with no membranes present. Figure 5.6 depicts simulations for experimental conditions for experiment 1.b with propane as the feedstock24.  The conditions are similar to those of a pre-reformer used to reform higher hydrocarbons. For both feedstocks, the higher hydrocarbon is consumed almost completely, and methane appears in the reformer from right near the bottom, due to the methanation reaction, Eq. (5.3) above.  Both propane and heptane are almost fully consumed right near the bottom, within 220 mm of the bottom, and hence their conversion profiles are not plotted here. Figures 5.5 and 5.6 also indicate the corresponding yields of methane, carbon oxides and local hydrogen yield. These results indicate that, except for a very small zone near the entrance, the reactor behaves like a methane steam reformer. Model predictions of the local hydrogen yield show the effects of competing phenomena among higher hydrocarbon steam reforming, methanation, and methane steam reforming at the entrance of the FBMR. Since the by-products of steam reforming are CO and CO2, the local yields of carbon oxides are indicators of conversion of the higher hydrocarbons as well as methane, the predominant intermediate. Figures 5.5 and 5.6 also show the simulated dry gas compositions for methane and hydrogen in the bubble and dense phases, and the experimental dry gas compositions. As seen, the predicted dry mole fractions in the dense phase are slightly higher than in the bubble phase for reaction products (e.g. hydrogen) and lower for reaction consumables (e.g. methane). Note that also that where there is a drop in temperature, there is a corresponding drop in yields of carbon oxides and  172  hydrogen, with an increase in methane yield due to reverse reaction, i.e. methanation. In such sections, the methane mole fraction is higher in the catalyst-rich dense phase. However, the gas compositions of both phases are very similar, indicating that the inter-phase mass transfer resistance is relatively unimportant for the operating conditions of the reactor. The simulated gas compositions for both phases are very close to the experimental composition profile in the reactor. The model predictions closely match the experimental data, confirming its applicability for cases without hydrogen removal. The reversal in product distribution (yields of methane, hydrogen and carbon oxides) also indicates that without any hydrogen removal, the FBMR performance is overwhelmingly dictated by thermodynamic equilibrium. 5.3.4 Test results with one membrane panel present Experiments with one membrane panel installed were simulated, as depicted in Figure 5.7, showing results for heptane experiment 4.b23, in Figure 5.8 for propane experiment 2.b24, and Figure 5.9 for experiment 2.c where methane was the feedstock. The shaded part indicates the span of the lone membrane panel installed in the 5th lateral opening from the bottom. As depicted in these figures, the simulated yields for methane with higher hydrocarbons, and conversion of methane (with methane as feedstock), show changes in slope corresponding to the start and end of the lone membrane panel, indicating faster consumption of methane in the interval corresponding to the membrane panel. Since hydrocarbon consumption produces carbon oxides, the carbon oxides yield also show a corresponding increase in slope. Thus, the one- membrane case clearly shows an equilibrium shift due to hydrogen removal. These three figures also show the local hydrogen yields. At the bottom of the membrane panel, there is an increase in the total hydrogen yield due to removal of pure hydrogen, shown as permeate hydrogen yield. The retentate hydrogen yield, which is the difference between these two yields, also exhibited a drop in the span of the membrane panel, due to hydrogen removal. Dry gas mole fractions for methane and hydrogen in the two phases, shown in these three figures, also indicate the effect of hydrogen removal. Compared to the immediately preceding or succeeding sections, in the interval occupied by the lone membrane panel, the difference in composition between the two phases increased for methane, and decreased for hydrogen. Membranes remove hydrogen from the dense phase as well as the bubble phase. Hydrogen is mostly produced in the dense phase where the vast majority of catalyst particles reside. However, the dense phase also occupies most of the volume, and hence covers most of the  173  membrane area. Hence the drop in hydrogen concentration is greater in the dense phase, even resulting in some crossing of the two profiles. 5.3.5 Test results with six membrane panels With all six membrane panels installed, the full capacity of the membrane permeation flux in the reactor was available. Figures 5.10, 5.11 and 5.12 depict the reactor performance compared with experimental results from heptane23, propane24, and methane. Simulated axial profiles of methane yield (with higher hydrocarbons feed), methane conversion (with methane feed) and carbon oxides all clearly show the effects of hydrogen withdrawal. Reductions in slopes of these profiles are seen in the short sections between adjacent membrane panels, where there was no hydrogen removal. Similar behavior could be seen from the total hydrogen yield. The dry mole fractions of methane in the dense phase are seen to decrease more quickly than for the bubble phase, except in sections where there was a drop in temperature.  5.4 Discussion of Results 5.4.1 Comparison between model and experimental data Parity plots comparing model predictions with experimental values reported for steam reforming of heptane23, propane24, and methane appear in Figures 5.13 to 5.15. Model estimates at a height of 1.64 m above the distributor are plotted against experimental data at the same location, which is the closest to the top (1.67 m) of the highest membrane panel. Figure 5.13 compares permeate hydrogen yields predicted by the model against experimental values for all three hydrocarbon feedstocks. The maximum hydrogen yields per mole of hydrocarbon fed with steam in excess are 22, 10 and 4 respectively for heptane, propane and methane. In order to allow comparison on the same plot, the permeate hydrogen yields have been normalized so that the permeate hydrogen yields have been divided by 22, 10 and 4. Six membrane panels extract much more hydrogen than a single membrane panel. However, for the heptane 7.a and propane 3.a experiments, with the FBMR pressure at 600 kPa, and an ambient permeate side pressure, little permeate was produced due to the fact that the driving force for permeation is provided by the difference of square roots of hydrogen partial pressures on both sides, rather than the total pressures. This was predicted very closely by the model.  Figure 5.14 compares the methane yields from higher hydrocarbons estimated by the model with the experimental values. Regardless of whether the feed was heptane or propane, the  174  FBMR acts predominantly as a methane steam reformer, with the higher hydrocarbons fully consumed near the distributor. Removal of hydrogen in the permeate stream caused the equilibrium limited reactions to be shifted accordingly, by consuming more methane. Thus more membrane surface area leads to lower methane yields, as depicted in Figure 5.14.  Faster removal of hydrogen as permeate leads to higher consumption of methane, thereby yielding more carbon oxides via reactions (5.3), (5.4), and (5.5). Thus carbon oxides yield is an indirect measure of the conversion of any hydrocarbon in the process stream, including the higher hydrocarbon and the intermediate methane. In Figure 5.15, these have been plotted for the three hydrocarbon feeds for the three membrane configurations, i.e. without membranes, and with 1 and 6 membrane panels. As expected, more membrane area led to more carbon oxides.  In general, there is good agreement between model predictions and the experimental data. The model is therefore helpful in understanding the various phenomena taking place in the FBMR. Some deviation near the bottom of the reactor may be due a to more non-uniform temperature distribution in this region. In this model, the energy balance equation was not considered since the distribution of heaters dominates the temperature profile. However, in the entrance region, where highly endothermic steam reforming of higher hydrocarbons (Eqs.(5.1) and (5.2)) followed by exothermic methanation (Eq. (5.3)) are very important, some interesting heat effects may be occurring. This suggests that it would have been useful to have recorded temperature at more locations along the height of the FBMR, especially near the bottom. 5.4.2 Membrane permeation effectiveness factor As noted above, the model used a membrane permeation effectiveness factor, α = 0.248, to account for the rate of production of pure hydrogen extracted via the membranes. Various reasons can be postulated for the loss in membrane effectiveness: (a) A cake persistently formed on the membrane surface during the fluidized bed operation, as shown in Figure 5.4. XRD and EDX analysis of this cake indicated that the source was mainly catalyst dust cemented (probably in the presence of steam) with traces of Pd (possibly from abrasion of membranes). Hydrogen must pass through this cake, before permeating selectively through the membrane foil. The remaining gas mixture components, predominantly steam, carbon dioxide and methane can form a diffusion layer between the cake and the membrane foil. Thus, fresh hydrogen produced from the steam reforming reactions faces two diffusional resistances: the cake, and an almost stagnant layer of gas mixture, before it can adsorb on the membrane surface.  175  (b) The permeation equation was developed from experimental data based on hydrogen permeation from pure hydrogen streams and hydrogen-helium mixtures. Some previous research44-46 indicates that steam, CO or CO2 can competitively adsorb on the membrane foil, thereby reducing the rate of permeation of hydrogen through the membrane, with steam adsorption being the most prominent. (c) Permeate side porous substrate resistance, and resistances from valves and fittings in the permeate line could also mean that the recorded permeation side pressure was actually slightly lower than actual. To understand the influence of the membrane effectiveness factor, α was varied, for one of the experimental runs, heptane experiment 5.b. The simulated profiles for α = 0.15, α = 0.248 (the fitted value), and α = 0.35 are shown in Figure 5.16. The same reactor temperature profile was implemented. Table 5.5 shows the key performance parameters at the top of the 6th membrane panel (1.67 m above distributor). As seen, the FBMR performance is heavily dependent on the permeability of the membranes. It also shows that a much smaller reactor length would be sufficient for higher membrane permeabilities for otherwise similar conditions. 5.4.3 Two-phase fluidization model Phenomena captured in our model include maintenance of minimum fluidization conditions in the dense phase, change in the number of moles due to gas-solid catalytic reaction, mass transfer between the dense and bubble phases, removal of hydrogen from both phases by membranes, and interactions among all these phenomena. The experiments without membranes assisted in determining the effects of hydrogen removal by membranes. For these experiments without membranes, using heptane or propane, the close agreement among experimental, simulated, and local equilibrium values indicates that (i) mass transfer between the dense and bubble phases is reasonably fast; (ii) the steam reforming kinetics are also relatively fast; and (iii) as a result of these two factors, the FBMR performance is governed closely by local equilibrium conditions. As indicated by the simulation results, the two-phase fluidization model promotes near- equality of compositions due in part to mass transfer and convection between the phases. The two-phase fluidization model simulates the experimental performance of the FBMR well. A sensitivity analysis of the model to the reaction rate constants and to the interphase mass transfer rate as shown in Appendix F indicates that, for this particular process, the model is sensitive to accurately characterizing the chemical equilibrium and hydrogen permeation, but relatively  176  insensitive to predicting the chemical kinetics, interphase mass transfer and hydrodynamics with precision. 5.4.4 FBMR performance The FBMR performance is predominantly determined by local equilibrium. Any withdrawal of hydrogen therefore significantly influences the performance. This makes proper estimation of the membrane permeation very important. As outlined above, a permeation effectiveness factor was necessary to account for the decrease in hydrogen permeation relative to that predicted on the basis of permeation experiments in a permeation rig without particles. The FBMR was operated with three different hydrocarbons, heptane, a model component for naphtha, propane, a key component of LPG, and methane, the major component in natural gas. Under the operating conditions of the experiments, satisfactory hydrogen yields were obtained for all three feedstocks. Since the higher hydrocarbons were fully consumed near the entrance of the reactor, the bulk of the reactor does not see the higher hydrocarbon, and an equilibrium-governed methane-rich gas composition occurs in the reactor. Removal of hydrogen steers the methane steam reforming and water gas shift reactions to produce more hydrogen, thereby enhancing the total hydrogen yield. Membrane-assisted reforming at relatively low temperatures of 500°C can achieve hydrogen yields comparable to a reformer operating at >750°C, and is thus compatible with higher hydrocarbons steam reforming, with minimal catalyst deactivation due to carbon formation or sintering. Steam reforming of higher hydrocarbons starts with a low-temperature pre-reformer to produce a methane-rich feed gas for the steam reformer operated at high temperatures to achieve a desired methane conversion. This is followed by a shift reaction system, and finally pressure swing adsorption to generate pure hydrogen. The FBMR combines the functions of the pre- reformer, reformer, shift converter and purification system into a single unit due to in-situ removal of pure hydrogen.  5.5 Conclusions A fluidized bed membrane reactor for steam reforming of hydrocarbons was modeled by a two- phase fluidization model. With no membrane panels installed, the model closely predicted the reformer performance, which was dominated by equilibrium. Membrane panels immersed in the bed extracted pure hydrogen, enhancing conversion of the hydrocarbons including the key  177  intermediate methane by favourably shifting the equilibrium. The model gave good predictions of the reactor behaviour, aided by a single fitted parameter, a membrane permeation effectiveness factor.  Development of durable membranes with higher hydrogen permeation flux would make the FBMR smaller for similar production capacities. Our FBMR was operated with three different hydrocarbon feeds, and the higher hydrocarbons were consumed close to the entrance of the reactor. Irrespective of the feedstock, the bulk of the FBMR operates as a methane steam reformer. Thus, an FBMR can be operated as a flexible reactor for hydrogen production. Compared to a traditional steam reformer, the FBMR offers a compact one-step reactor for producing hydrogen from higher hydrocarbons. Since membrane-assisted reforming enables high hydrogen yields at temperatures below 575°C, chances of catalyst deactivation are also minimized.  178  Table 5.1: Reactor physical details  Quantity Value Description Nor 6 Number of orifices in the distributor housing 1.88 x 10-3 m2 Main section of FBMR where membrane panels or dummies are not present 2.31 x 10-3 m2 Main section of FBMR where membrane panels or dummies are present A 4.26 x 10-3 m2 Expanded section above the main section hstatic 1.7 m Static bed height Lreactor 2.32 m Total height of FBMR including the main rectangular section and the expanded circular section dp 179 µm Mean size of catalyst particle ρp 2600 kg/m3 Particle density of catalysts powder 2H δ  25 µm Thickness of Pd-Ag membranes 0MP  0.00207 mole/(m.min.atm0.5) Pre-exponential factor for membrane permeation equation 2H E  9180 J/mol Activation energy for membrane permeation equation    179  Table 5.2: Reaction rate equations  Reaction number Rate equation and kinetic parameters Reference (5.1) 2 1 1 2 2 2 2167 167 1 ⎥⎥⎦ ⎤ ⎢⎢⎣ ⎡ ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+ = H OH b OH HHC a HC P P K P PP K Pk r Tottrup47 (5.2) 86.0' 53.0'93.0' 2 2 2 283 1 H OHHC P PPk r θ+= −  Ma48 (5.3) 2 5.2 3 5.0 33 2 242 DEN P PP K PP kr H OHCHHCO ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ − = (5.4) 2 4 2 44 2 2 DEN K P P PP kr CO H OHCO ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ − = (5.5) 2 43 5.0 5.3 2 55 22 2 24 DEN KK PP P PP kr HCO H OHCH ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ − = (5.3), (5.4), and (5.5) 2 22 4422 1 H OHOH CHCHHHCOCO P PK PKPKPKDEN ++++= Xu & Froment49   180  Table 5.3: Kinetic parameters  Reaction number Rate parameters Units Reference RTek 67800 5 1 .108 − −×=  ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ .h.barg mol catalyst  2.25=aK  ( )1bar −  (5.1) 077.0=bK  (-) Tottrup47 RTek 31063.189 14 2 .101428.2 ×− ×= ⎟⎠ ⎞⎜⎝ ⎛ Ni.h.mkPa mol 20.4 (5.2) θ = 1 ( )86.0kPa − Ma48 RTek 3101.240 15 3 .1049.9 ×− ×=  ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ kgcat.h kmol.bar0.5 RTek 31013.67 6 4 .1039.4 ×− ×=  ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ rkgcat.h.ba kmol RTek 3109.243 15 5 .1029.2 ×− ×=  ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ kgcat.h kmol.bar0.5 RT CO eK 31065.70 5.1023.8 × −×=  ( )1bar − RT CH eK 3 4 1028.38 4.1065.6 × −×=  ( )1bar − RT OH eK 3 2 1068.88 5.1077.1 ×− ×=  ( )1bar − RT H eK 3 2 109.82 9.1012.6 × −×=  (-) ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ +×− = 114.301083.26 3 3 TeK ( )2bar (5.3), (5.4), and (5.5) ⎟⎠ ⎞⎜⎝ ⎛ −= 036.4 4400 4 TeK (-) Xu & Froment49   181  Table 5.4: Experimental conditions for runs where data are compared with model predictions  Expt No. Active Membranes (Location) Total feed rate Tav  P  Pm SCR   (mols/min) (°C) (kPa) (kPa)  Heptane1.b None 0.766 450 460 NA 5.0  Propane1.b None 0.673 500 600 NA 6.0  Heptane4.b 1 (#5) 0.717 480 585 35 5.0  Propane2.b 1 (#5) 0.717 485 515 30 5.0  Heptane5.b 6 (#1 to #6) 0.614 500 600 25 5.0  Propane3.c 6 (#1 to #6) 0.614 500 600 25 5.0  Methane1.a 6 (#1 to #6) 0.819 500 800 50 5.0  Methane1.b 6 (#1 to #6) 0.614 500 600 50 5.0  Methane1.c 6 (#1 to #6) 0.614 500 600 25 5.0  Methane2.a 1 (#5) 0.695 500 500 101 5.0  Methane2.b 1 (#5) 0.695 500 750 101 5.0  Methane2.c 1 (#5) 0.695 500 500 30 5.0  Methane2.d 1 (#5) 0.695 500 75 30 5.0   Table 5.5: FBMR performance with variations in permeation effectiveness factor  α Methane Yield Carbon Oxides Yield Permeate H2 Yield Retentate H2 Yield Total H2 Yield 0.15 0.233 0.769 4.378 2.773 7.151 0.2484 0.112 0.890 6.561 2.041 8.602 0.35 0.031 0.972 8.161 1.356 9.517   182  Sample 2a Sample 2b Sample 4a Sample 4b Sample 6a Sample 6b Sample 5a Sample 5b Sample 3b Sample 3a Sample 1b Sample 1a H2 Permeate H2 Permeate H2 Permeate H2 Permeate H2 Permeate H2 Permeate ROG FBMR Feed (Hydrocarbon + Steam)  Figure 5.1: Schematic of reactor geometry    183  ReactionsReactions Interphase Mass Transfer Bulk Convection M em br an es Fr ee bo ar d Bubble Phase Dense Phase FBMR Feed Reformer Off-Gas Pe rm ea tio n Permeation Hydrogen   Figure 5.2: Schematic of the FBMR kinetic model   184    Figure 5.3: A cake of catalyst formed around the ROG filter   185    (a)    (b)   186    (c)  Figure 5.4: Particulate coating formed on the membranes during FBMR operation: (a) A fresh membrane before installation (b) the membrane surface covered by the coating (c) a view of other side of the same membrane showing a clean shining membrane foil exposed after tapping off a part of the coating  187  Lo ca l Te m p.  (o C ) 400 450 500 C ar bo n O xi de s  Y ie ld 0.3 0.5 (Model) (Expt) H yd ro ge n  Y ie ld s 0 4 8 (Model) Total H2 (Expt) M et ha ne  Y ie ld 0.1 0.3 0.5 0.7 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment   Figure 5.5: FBMR performance for experiment Heptane1.b  188  Lo ca l Te m p.  (o C ) 350 400 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 (Model) (Expt) H yd ro ge n  Y ie ld s 0 4 (Model) Total H2 (Expt) M et ha ne  Y ie ld 0.1 0.3 0.5 0.7 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 65 80 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 65 80 Bubble phase Dense phase Freeboard Experiment   Figure 5.6: FBMR performance for experiment Propane1.b   189  Lo ca l Te m p.  (o C ) 350 400 450 500 550 C ar bo n O xi de s  Y ie ld 0.3 0.5 (Model) (Expt) H yd ro ge n  Y ie ld s 0 4 8 12 16 20 Total H2 (Model) Permeate H2 (Model) Total H2 (Expt) Permeate H2 (Expt) M et ha ne  Y ie ld 0.1 0.3 0.5 0.7 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 65 80 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment   Figure 5.7: FBMR performance for experiment Heptane 4.b  190  Lo ca l Te m p.  (o C ) 420 470 520 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 (Model) (Expt) H yd ro ge n  Y ie ld s 0 4 8 (Model) (Model) Total H2 (Expt) Permeate H2 (Expt) M et ha ne  Y ie ld 0.1 0.3 0.5 0.7 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 65 80 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment   Figure 5.8: FBMR performance for experiment Propane 2.b  191  Lo ca l Te m p.  (o C ) 420 470 520 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 (Model) (Expt) H yd ro ge n  Y ie ld s 0 1 2 3 4 (Model) (Model) Total H2 (Expt) Permeate H2 (Expt) M et ha ne  C on ve rs io n 0.1 0.3 0.5 0.7 0.9 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 65 80 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 65 Bubble phase Dense phase Freeboard Experiment   Figure 5.9: FBMR performance for experiment Methane 2.c  192  Lo ca l Te m p.  (o C ) 350 400 450 500 550 600 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 (Model) (Expt) H yd ro ge n  Y ie ld s 0 4 8 12 16 20 Total H2 (Model) Permeate H2 (Model) Total H2 (Expt) Permeate H2 (Expt) M et ha ne  Y ie ld 0.1 0.3 0.5 0.7 0.9 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment   Figure 5.10: FBMR performance for experiment Heptane 5.b  193  Lo ca l Te m p.  (o C ) 420 470 520 570 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 (Model) (Expt) H yd ro ge n  Y ie ld s 0 4 8 (Model) (Model) Total H2 (Expt) Permeate H2 (Expt) M et ha ne  Y ie ld 0.1 0.3 0.5 0.7 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 65 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment   Figure 5.11: FBMR performance for experiment Propane 3.c  194  Lo ca l Te m p.  (o C ) 420 470 520 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 (Model) (Expt) H yd ro ge n  Y ie ld s 0 1 2 3 4 (Model) (Model) Total H2 (Expt) Permeate H2 (Expt) M et ha ne  C on ve rs io n 0.1 0.3 0.5 0.7 0.9 (Model) (Expt) M et ha ne  D ry  F ra ct io n (% ) 5 20 35 50 65 80 Bubble phase Dense phase Freeboard Experiment Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n  D ry  F ra ct io n (% ) 5 20 35 50 Bubble phase Dense phase Freeboard Experiment   Figure 5.12: FBMR performance for experiment Methane 1.c  195  Permeate H2 Yield (Experimental) 0.0 0.2 0.4 0.6 0.8 Pe rm ea te  H 2 Y ie ld  (M od el ) 0.0 0.2 0.4 0.6 0.8 1 Membrane Panel Heptane 6 Membrane Panels Heptane 1 Membrane Panel Propane 6 Membrane Panels Propane 1 Membrane Panel Methane 6 Membrane Panels Methane   Figure 5.13: Parity plot for permeate hydrogen yields   196  CH4 Yield (Experimental) 0.0 0.2 0.4 0.6 0.8 C H 4 Y ie ld  (M od el ) 0.0 0.2 0.4 0.6 0.8 No Membrane Panels Heptane 1 Membrane Panel Heptane 6 Membrane Panels Heptane No Membrane Panels Propane 1 Membrane Panel Propane 6 Membrane Panels Propane   Figure 5.14: Parity plot for methane yields    197  Carbon Oxides Yield (Experimental) 0.0 0.2 0.4 0.6 0.8 1.0 C ar bo n O xi de s Y ie ld  (M od el ) 0.0 0.2 0.4 0.6 0.8 1.0 6 Membrane Panels Heptane 6 Membrane Panels Propane 6 Membrane Panels Methane 1 Membrane Panel Heptane 1 Membrane Panel Propane 1 Membrane Panel Methane No Membrane Panels Heptane No Membrane Panels Propane   Figure 5.15: Parity plot for carbon oxides yields  198  R et en ta te  H 2 Y ie ld 0.0 1.5 3.0 α = 0.15 α = 0.2484 α = 0.35 Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 To ta l H 2 Y ie ld 0 2 4 6 8 10 α = 0.15 α = 0.2484 α = 0.35 Pe rm ea te  H 2 Y ie ld 1 3 5 7 α = 0.15 α = 0.2484 α = 0.35 M et ha ne   Y ie ld 0.1 0.3 0.5 0.7 α = 0.15 α = 0.2484 α = 0.35 Lo ca l Te m p.  (o C ) 420 470 520 570 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 α = 0.15 α = 0.2484 α = 0.35   Figure 5.16: Effect of membrane permeation effectiveness factor (Propane Experiment 3.c)  199  5.6  References 1. Heffer, P.; Prud’homme, M., Medium-Term Outlook for Global Fertilizer Demand, Supply and Trade (2008 – 2012): Summary Report. International Fertilizer Industry Association (IFA): Paris (France), 2008. 2. Rostrup-Nielsen, J. R.; Rostrup-Nielsen, T., Large-scale hydrogen production. CatTech 2002, 6, (4), 150-159. 3. Winter, C. J., Hydrogen energy - Abundant, efficient, clean: A debate over the energy- system-of-change. International Journal of Hydrogen Energy 2009, 34, (14), S1-S52. 4. Ferreira-Aparicio, P.; Benito, M. J.; Sanz, J. L., New trends in reforming technologies: from hydrogen industrial plants to multifuel microreformers. Catalysis Reviews-Science and Engineering 2005, 47, (4), 491-588. 5. Rostrup-Nielsen, J., Steam reforming of hydrocarbons. A historical perspective. Natural Gas Conversion Vii 2004, 147, 121-126. 6. Rostrup-Nielsen, J. R., Syngas in perspective. Catalysis Today 2002, 71, (3-4), 243-247. 7. Ewan, B. C. R.; Allen, R. W. K., A figure of merit assessment of the routes to hydrogen. International Journal of Hydrogen Energy 2005, 30, (8), 809-819. 8. Adris, A. M.; Lim, C. J.; Grace, J. R., The fluidized bed membrane reactor system: A pilot scale experimental study. Chemical Engineering Science 1994, 49, (24B), 5833-5843. 9. Patil, C. S.; Annaland, M. V. S.; Kuipers, J. A. M., Fluidised bed membrane reactor for ultrapure hydrogen production via methane steam reforming: Experimental demonstration and model validation. Chemical Engineering Science 2007, 62, (11), 2989-3007. 10. Boyd, T.; Grace, J.; Lim, C. J.; Adris, A. E. M., Hydrogen from an internally circulating fluidized bed membrane reactor. International Journal of Chemical Reactor Engineering 2005, 3, A58. 11. Roy, S.; Pruden, B. B.; Adris, A. M.; Grace, J. R.; Lim, C. J., Fluidized-bed steam methane reforming with oxygen input. Chemical Engineering Science 1999, 54, (13-14), 2095- 2102. 12. Ahchieva, D.; Peglow, M.; Heinrich, S.; Morl, L.; Wolff, T.; Klose, F., Oxidative dehydrogenation of ethane in a fluidized bed membrane reactor. Applied Catalysis A: General 2005, 296, (2), 176-185. 13. Alonso, M.; Lorences, M. J.; Pina, M. P.; Patience, G. S., Butane partial oxidation in an externally fluidized bed-membrane reactor. Catalysis Today 2001, 67, (1-3), 151-157.  200  14. Deshmukh, S. A. R. K.; Laverman, J. A.; Annaland, M. V.; Kuipers, J. A. M., Development of a membrane-assisted fluidized bed reactor. 2. Experimental demonstration and modeling for the partial oxidation of methanol. Industrial & Engineering Chemistry Research 2005, 44, (16), 5966-5976. 15. Deshmukh, S. A. R. K.; Heinrich, S.; Morl, L.; Annaland, M. V. S.; Kuipers, J. A. M., Membrane assisted fluidized bed reactors: Potentials and hurdles. Chemical Engineering Science 2007, 62, (1-2), 416-436. 16. Chen, Z.; Yan, Y.; Elnashaie, S. S. E. H., Hydrogen production and carbon formation during the steam reformer of heptane in a novel circulating fluidized bed membrane reformer. Industrial & Engineering Chemistry Research 2004, 43, (6), 1323-1333. 17. Chen, Z.; Yan, Y.; Elnashaie, S. S. E. H., Modeling and optimization of a novel membrane reformer for higher hydrocarbons. AIChE Journal 2003, 49, 1250-1265. 18. Chen, Z.; Elnashaie, S. S. E. H., Optimization of reforming parameter and configuration for hydrogen production. AIChE Journal 2005, 51, (5), 1467-1481. 19. Chen, Z.; Elnashaie, S. S. E. H., Autothermal CFB membrane reformer for hydrogen production from heptane. Chemical Engineering Research & Design 2005, 83, (A7), 893-899. 20. Toomey, R. D.; Johnstone, H. F., Gaseous fluidization of solid particles. Chemical Engineering Progress 1952, 48, (5), 220-226. 21. Rakib, M. A.; Grace, J. R.; Elnashaie, S. S. E. H.; Lim, C. J.; Bolkan, Y. G., Kinetic simulation of a compact reactor system for hydrogen production by steam reforming of higher hydrocarbons. Canadian Journal of Chemical Engineering 2008, 86, (3), 403-412. 22. Rakib, M. A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H.; Epp, M.; Gulamhusein, A.; Boyd, T.; Keelan, A., Pilot scale experimental setup for hydrogen production from higher hydrocarbons: Safety considerations and implementation. Journal of Chemical Health and Safety Under Review. 23. Rakib, M. A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H., Steam reforming of heptane in a fluidized bed membrane reactor. Journal of Power Sources Accepted for publication (2010). 24. Rakib, M. A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H.; Ghiasi, B., Steam reforming of propane in a fluidized bed membrane reactor for hydrogen production. International Journal of Hydrogen Energy Accepted for publication (2010).  201  25. Adris, A. M.; Lim, C. J.; Grace, J. R., The fluidized-bed membrane reactor for steam methane reforming: Model verification and parametric study. Chemical Engineering Science 1997, 52, (10), 1609-1622. 26. Bartholomew, C. H.; Farrauto, R. J., Fundamentals of Industrial Catalytic Processes 2ed.; Wiley-AIChE: 2005. 27. Grace, J. R.; Clift, R., 2-Phase Theory of Fluidization. Chemical Engineering Science 1974, 29, (2), 327-334. 28. Valenzuela, J. A.; Glicksman, L. R., Gas-Flow Distribution in a Bubbling Fluidized-Bed. Powder Technology 1985, 44, (2), 103-113. 29. Hilligardt, K.; Werther, J., Local bubble gas hold-up and expansion of gas/solid fluidized beds. German Chemical Engineering 1986, 9, 215-221. 30. Kunii, D.; Levenspiel, O., Fluidization Engineering. 2 ed.; Butterworth-Heinemann: Boston, 1991. 31. Sit, S. P.; Grace, J. R. In Interphase mass transfer during bubble formation in fluidized beds, Fluidization V, Elsionore, Denmark, 1986; Ostergaard, K.; Soorensen, A., Eds. Engineering Foundation, New York: Elsionore, Denmark, 1986; pp 39-46. 32. Grace, J. R.; DeLasa, H. I., Reaction near the grid in fluidized beds. AIChE Journal 1978, 24, 364-366. 33. Wen, C. Y.; Yu, Y. H., A generalized method for predicting minimum fluidization velocity. AIChE Journal 1966, 12, (3), 610-612. 34. Grace, J. R., Fluidized-bed hydrodynamics. In Handbook of Multiphase Systems, Hetsroni, G., Ed. Hemisphere: Washington, 1982; pp 8-5 - 8-64. 35. Kunii, D.; Levenspi.O, Bubbling bed model for kinetic processes in fluidized beds - Gas- solid mass and heat transfer and catalytic reactions. Industrial & Engineering Chemistry Process Design and Development 1968, 7, (4), 481-492. 36. Darton, R. C.; Lanauze, R. D.; Davidson, J. F.; Harrison, D., Bubble growth due to coalescence in fluidized-beds. Transactions of the Institution of Chemical Engineers 1977, 55, (4), 274-280. 37. Mori, S.; Wen, C. Y., Estimation of bubble diameter in gaseous fluidized beds. AIChE Journal 1975, 21, (1), 109-115. 38. Sit, S. P.; Grace, J. R., Effect of bubble interaction on interphase mass transfer in gas fluidized beds. Chemical Engineering Science 1981, 36, (2), 327-335.  202  39. Wilke, C. R., Diffusion properties of multicomponent gases. Chemical Engineering Progress 1950, 46, (2), 95-104. 40. Gallucci, F.; Annaland, M. V. S.; Kuipers, J. A. M. In Comparison of packed bed and fluidized bed membrane reactors for methane reforming, 6th International Symposium on Multiphase Flow, Heat Mass Transfer and Energy Conversion, Xi'an, 2009; Xi'an, 2009. 41. Abba, I. A.; Grace, J. R.; Bi, H. T., Variable-gas-density fluidized bed reactor model for catalytic processes. Chemical Engineering Science 2002, 57, (22-23), 4797-4807. 42. Mahecha-Botero, A.; Grace, J. R.; Lim, C. J.; Elnashaie, S. S. E. H.; Boyd, T.; Gulamhusein, A., Pure hydrogen generation in a fluidized bed membrane reactor: Application of the generalized comprehensive reactor model. Chemical Engineering Science 2009, 64, (17), 3826-3846. 43. Sieverts, A.; Zapf, G., The solubility of deuterium and hydrogen in solid palladium. Zeitschrift für Physikalische Chemie 1935, 174, 359--364. 44. Gielens, F. C.; Knibbeler, R. J. J.; Duysinx, P. F. J.; Tong, H. D.; Vorstman, M. A. G.; Keurentjes, J. T. F., Influence of steam and carbon dioxide on the hydrogen flux through thin Pd/Ag and Pd membranes. Journal of Membrane Science 2006, 279, (1-2), 176-185. 45. Hou, K.; Hughes, R., The effect of external mass transfer, competitive adsorption and coking on hydrogen permeation through thin Pd/Ag membranes. Journal of Membrane Science 2002, 206, (1-2), 119-130. 46. Kulprathipanja, A.; Alptekin, G. O.; Falconer, J. L.; Way, J. D.,. Effects of water gas shift gases on Pd−Cu alloy membrane surface morphology and separation properties. Industrial & Engineering Chemistry Research, 2004, 43, 4188-4198. 47. Tottrup, P. B., Evaluation of intrinsic steam reforming kinetic parameters from rate measurements on full particle size. Applied Catalysis 1982, 4, (4), 377-389. 48. Ma, L. Hydrogen production from steam reforming of light hydrocarbons in an autothermic system. University of New South Wales, Sydney, 1995. 49. Xu, J. G.; Froment, G. F., Methane steam reforming, methanation and water-gas shift .1. Intrinsic kinetics. AIChE Journal 1989, 35, (1), 88-96.     203  CHAPTER 6.      CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK   6.1 Conclusions The demand for hydrogen is projected to increase in the energy sector, as well as for industrial processes. While natural gas is the most widely used feedstock for steam reforming, other hydrocarbon feedstocks may be desirable alternatives in refinery operations and in syngas production in locations where natural gas is not available or where the alternative feedstocks are in over-supply. This research deals with pure hydrogen production from higher hydrocarbons in a novel fluidized bed membrane reactor (FBMR).  In this project, an FBMR was designed, fabricated and installed with safety requirements adequately addressed. It was operated with methane, propane and heptane, representing different categories of hydrocarbon feedstock. Average bed temperatures up to of 550°C, and reactor pressures up to 800 kPa were studied. The following conclusions can be drawn from the experimental results: (a)  The FBMR produced pure hydrogen while enhancing overall hydrocarbon conversion. As in several previous studies involving steam reforming of methane or natural gas in an FBMR, our experiments showed that FBMR operation achieves higher overall hydrogen yields for propane and heptane than predicted by equilibrium conditions, while simultaneously producing a pure hydrogen stream. These results were achieved at moderate temperatures (< 550°C). (b) FBMR gives a compact reactor configuration for hydrocarbons steam reforming. For higher hydrocarbon feedstocks, an industrial hydrogen production setup includes several units like pre-reformer, reformer, shift conversion section and purification section. The FBMR operation demonstrated that pure hydrogen can be produced in a single reactor, combining the functions of all these units. Thus a compact reactor configuration has been demonstrated, which could be useful for small to medium scale distributed hydrogen production at fuelling stations. (c) FBMR is a suitable configuration for higher hydrocarbons. FBMR operation can achieve high hydrogen yields and lower methane yields at the usual operating temperature of the pre-reformer, assisted by selective removal of hydrogen through  204  membranes, thereby continuously shifting the equilibrium limitation towards complete conversion of methane, the intermediate hydrocarbon. (d) FBMR allows flexibility of feedstock. The FBMR system can adapt to feedstock variations, with the bulk of the reactor behaving as a methane steam reformer and little influence of the actual feed hydrocarbon.  In addition to the major conclusions above, this research also achieved the following: (a)  The safety issues involving a laboratory scale reactor setup have been analyzed in detail. This can be useful for other small scale FBMR reactors. A Cause & Effect matrix has been developed to take care of any process upsets. (b) In order to understand the phenomena occurring in the FBMR, a kinetic model of the reactor has been written, based on the two-phase model of fluidization, with the dense catalyst bed treated as two phases in parallel, each with plug flow of gas and with exchange between the two phases. Hydrogen is withdrawn from both phases in proportion to the volume fractions they occupy in the bed. Interphase diffusional mass transfer occurs due to the concentration difference of the various components in the two phases, and bulk convectional cross-flow is assumed to ensure that the gas flow corresponding to minimum fluidization conditions in the dense phase is maintained. The model provides a reasonably good fit between predicted and experimental yields of total hydrogen, carbon oxides and methane, for all three hydrocarbons tested, i.e. methane, propane and heptane.  6.2 Limitations of FBMR Steam Reforming (a) Hydrogen at lower pressure A major limitation for the process is that hydrogen is available at a very low pressure compared to hydrogen line pressure from a traditional steam reformer. The low pressure hydrogen then has to be compressed to meet downstream process requirements. (b) Membrane cost Palladium or its alloys are very expensive, augmenting the capital investment in the process. For the process to be economically attractive, the membranes must be thinner, as well as durable over extended periods of operation.    205  6.3 Recommendations for Future Work As a follow-up of a proof-of-concept research presented in this dissertation, some directions are proposed: (a) Energy balance equation in the reactor model Due to the difficulty of accurate estimation of the heat lost from the FBMR through the insulation, the energy balance equation was not considered in the predictive reactor model, as described in Chapter 5. With the electrical energy input known, the heat lost can be estimated by heating the FBMR to the usual operating temperatures with no reactants being fed. Thereby, a complete predictive model can be developed with the incorporation of the energy balance equation. (b) Further sophistications of the model An axial dispersion model for the two phase model would be able to capture the phenomenon of gas back-mixing which occurs in fluidized beds. Such a model would likely offer better predictions than a plug flow model. The model used in this work is one-dimensional model. A two- or three-dimensional model can capture the local recirculation and back mixing phenomena due to the presence of membrane panels and intermittent widening of the cross-sectional area between the panels. It may also be able to analyze any lateral concentration gradient due to hydrogen withdrawal via the vertical membrane panels. A computational fluid dynamics (CFD) model may be a useful tool to describe the complex phenomena and geometry. CFD calculations should be able to predict the interphase balancing mass transfer when flow rate in the dense phase differs from that under the minimum fluidization conditions1. (c)   Real hydrocarbon feedstock Model hydrocarbon compounds were used in this study, so that the main underlying concept of steam reforming of higher hydrocarbons could be proved in a step-by-step approach. However, real hydrocarbon feedstocks like naphtha, kerosene or diesel contain naphthenic, aromatic or olefin components, which may promote carbon formation, causing catalyst deactivation. Steam reforming of such real feedstocks is therefore required to study any effect on the membranes, e.g. possible fouling due to coke deposition. (d) CO2 capture With the ills of climate change being clearly visible, control of greenhouse gas emissions is becoming increasingly more important. Hydrogen extraction from fossil fuels by steam  206  reforming produces large quantities of CO2 as a byproduct. This CO2 is usually released to the atmosphere. In-situ CO2 capture in a steam reforming process can enhance hydrogen yield by promoting equilibrium shift, in a manner similar to withdrawal of hydrogen through membranes. This has been extensively studied for methane steam reforming. In principle, this could also be extended to steam reforming of higher hydrocarbons. (e) Lowering of steam-to-carbon molar ratio Keeping in view the main scope of the research presented in this thesis as a proof-of-concept, the base steam-to-carbon molar ratio was maintained as 5.0 for these experiments. A lower steam-to- carbon ratio could improve the energy efficiency of the process, and also decrease the reactor volume due to decreased volumetric flow. With high hydrogen yields achieved at temperatures of 550°C or lower, a lower steam-to-carbon ratio may be able to achieve deactivation-free operation. This needs to be investigated experimentally. (f) Autothermal reforming A major limitation of industrial steam reformers is heat supply to the highly endothermic steam reforming reactions. Hundreds of catalyst filled tubes need to be housed in a furnace to decrease the radial non-uniformity of temperature. Autothermal reforming introduces a controlled amount of oxygen which consumes some of the hydrocarbon, supplying the heat requirement. Fluidized bed operation reduces this radial non-uniformity of temperature. In addition, in-situ supply of heat due to autothermal reforming eliminates the heat transfer barrier. Autothermal reforming of methane or natural gas could also be extended to steam reforming of higher hydrocarbons. (g) Other configurations: Ex-situ membranes One of the main challenges for this process to be commercially implemented is the durability of the membranes against abrasion in a particulate environment and with temperature cycling during start-ups and shut-downs. This could be avoided by placing the membranes downstream in a separate vessel, and recycling the retentate partially to the FBMR for further reaction. (h) Catalyst improvement: Fluidizable catalyst development Fluidizable catalysts prepared by crushing commercial catalyst pellets are liable to further breakage in an FBMR due to particle-particle and particle-wall collisions. This can affect the fluidization characteristics, as well as leading to loss of fines by entrainment. Attrition-resistant fluidizable catalysts need to be developed. (i) Long-term durability tests Resistance of the membranes against development of pinholes for continuous operation over long periods of time needs to be established. Effect of the olefins and aromatic content in the  207  feed may cause membrane fouling by coke deposition, thereby decreasing the hydrogen recovery. (j) Scale-Up: Hydrogen filling stations as a first target As Deshmukh et al.2 observed, FBMRs show greater promise of commercialization than fixed bed membrane reactors for steam reforming. Further studies on scale-up for this newly developed process are required for higher hydrocarbon feedstocks.  6.4 Specific Recommendations for Reactor Built in the Current Study (a)  A better temperature control scheme should be implemented, with an individual temperature controller for each heater to provide more uniform temperature profiles, and better safeguards against overheating some zones while leaving some zones significantly colder. Additional rope heaters could be placed in locations where heat losses are greater (e.g. at the flanges). (b) Two additional ports are available per rectangular flange on the side openings. Utilizing these to record bed temperatures could improve the reactor performance monitoring. (c)  One or two additional off-gas filters should be installed in parallel to the existing one, with proper bypass capabilities. Fine catalyst cakes build-up on the filters could then be dislodged during operation with periodic reverse injection of inert gas, while the other filters are still available for off-gas venting. Filters need to be cleaned in rotation, so that the catalyst particles can drop back onto the bed, and also to ensure that these lines are not blocked. (d) This set-up could also be used to investigate other reactions, e.g. propane dehydrogenation or the water-gas shift reaction.   208  6.5 References 1. Li, T. W.; Mahecha-Botero, A.; Grace, J. R., Computational fluid dynamic investigation of change of volumetric flow in fluidized bed reactors. Industrial & Engineering Chemistry Research 2010 (In Press). 2. Deshmukh, S. A. R. K.; Heinrich, S.; Morl, L.; Annaland, M. V. S.; Kuipers, J. A. M., Membrane assisted fluidized bed reactors: Potentials and hurdles. Chemical Engineering Science 2007, 62, (1-2), 416-436.   209 APPENDIX A.      KINETIC SIMULATION OF A COMPACT REACTOR SYSTEM FOR HYDROGEN PRODUCTION BY STEAM REFORMING OF HIGHER HYDROCARBONS*   A.1 Introduction Hydrogen is frequently discussed as a future energy carrier. Key applications are as a carbon- free fuel, and as a fuel for hydrogen fuel cells for automotive and other applications. Hydrogen has been used effectively in a number of internal combustion engine vehicles mixed with natural gas (Hythane)1. Hydrogen can also be combined electrochemically with oxygen without combustion to produce direct-current electricity in fuel cells, and is used in a growing number of fuel cell vehicles. As a feedstock in chemical processes, the demand for hydrogen is increasing, both for the petrochemical industries and for petroleum refining processes. Synthesis gas, a mixture of hydrogen, carbon monoxide and carbon dioxide in various proportions, is used by Fisher Tropsch catalytic technology to produce a wide range of chemicals from methanol up to diesel. Steam-reforming-based hydrogen plants are installed in refineries to meet the fast-rising demand-supply gap in their daily operations2. Hydrogen is used in the metallurgical industry to create a reducing atmosphere in metal extraction3, and in annealing of steel. It is also used in the electronics industry to manufacture semiconductor devices, and in the food industries for hydrogenation of fats and oils3,4. Thus the demand of hydrogen is projected to increase, and this has motivated research into improving methods of hydrogen production, separation, purification, storage and transportation. Many of the hydrogen uses put special demand on the purity of the hydrogen from these reformers. Steam reforming remains the leading pathway of hydrogen from hydrocarbon sources, especially natural gas2,5.  The greatest advantage of the steam reforming pathway is that hydrogen is extracted not only from a hydrocarbon, but from steam as well, thereby enhancing H2 production, giving the maximum H2 production per mole of hydrocarbon. The presence of  * A version of this Appendix has been published: Rakib, M.A., Grace, J.R., Elnashaie, S.S.E.H., Lim, C.J., and Bolkan Y.G. Kinetic Simulation of a Compact Reactor System for Hydrogen Production by Steam Reforming of Higher Hydrocarbons, Canadian Journal of Chemical Engineering (2008) 86, 403-412.  210 excess steam in the reaction mixture suppresses coking reactions, the extent of which depends largely on the reaction temperature and the type of hydrocarbon. Currently methane is the major feedstock for production of synthesis gas, as well as pure hydrogen. However, compared to liquid hydrocarbons, the volumetric hydrogen density remains low even after natural gas is compressed to liquid for transportation, although the H/C ratio of methane is high6. Therefore, an easily deliverable and safely storable hydrogen source, such as gasoline and diesel, is preferred for mobile applications7. On-board hydrogen generation systems prefer liquid hydrocarbon feedstocks, such as gasoline, kerosene and diesel oil, which have a higher energy density and a wider distribution network, compared to methanol8. In addition, many refineries benefit from flexibility in feedstocks, taking advantage of the surplus of various hydrocarbons in the refinery. Traditional steam reforming plants have a fixed bed steam reformer. For naphtha steam reforming, the desulfurized hydrocarbon is fed to a pre-reformer, which is operated adiabatically, where the higher hydrocarbons are directly converted to methane, giving a methane-rich gas feed for the reformer9. In the primary reformer there are hundreds of externally fired catalyst-packed tubes, in which steam reforming of methane takes place. The fixed bed reformer is followed by the shift reactors (HTS and LTS reactors) section for further reaction of carbon monoxide with steam to enhance hydrogen yield. The gas purification system consists of a CO2 removal unit, a Methanator, and finally a Pressure Swing Adsorption unit to produce pure hydrogen. Steam reforming is limited by diffusional resistances inside the catalyst pellet, resulting in very low effectiveness factors, of the order of 10-2 to 10-3  10-12.  In addition, with external firing needed for the highly endothermic reactions, formation of hot spots can lead to problems related to temperature control. Pressure drop limitations block attempts to improve the effectiveness factor by using smaller diameter particles. Adris et al.13 and Elnashaie et al.10 proposed a novel Fluidized Bed Steam Reformer, with the heat supplied through immersed heat transfer tubes. Heat transfer limitations of the fixed bed reactor are also minimized in the fluidized bed because of better mixing characteristics. The other major limitation for the steam reforming reactions is thermodynamic equilibrium. Removal of the main products can drive the reaction towards completion, following Le Chatelier’s principle. Permselective membranes of Pd or Pd-based alloys can remove H2, thus serving dual objectives: enhancing the hydrocarbon conversion by favourably shifting the equilibrium conversion, and producing a stream of pure H2 as permeate14-16.  211  This study deals with modeling a fluidized bed membrane reactor for steam reforming of higher hydrocarbons, carried out to size an experimental reformer setup. Typically, naphtha consists predominantly of saturated hydrocarbons (>90% by volume), the balance being made mainly of aromatics, and some unsaturated hydrocarbons17. n-Heptane is treated in the current simulations as a model compound for steam reforming of naphtha, as also earlier assumed by Chen17,18, Tøttrup19, Christensen9, and Darwish et al.20.  Others have assumed it to be a model component for gasoline6,8,21.  A hydrocarbon feed mixture composed of n-heptane and n-hexane (in a weight ratio of C7/C6 = 2) was taken as a synthetic feed for steam reforming of naphtha by Melo et al.7.  A.2 Irreversibility of Steam Reforming of Higher Hydrocarbons Equilibrium calculations, in Figure A.1 show that the steam reforming of heptane is practically irreversible, indicated by its complete consumption at the representative conditions of reaction. The temperature was varied from 400 to 800°C at four different pressures from 1 to 20 bars, and equilibrium compositions were predicted using a Gibbs Reactor in HYSYS simulation software. The feed composition, consisting of n-heptane, steam and H2, used for the equilibrium predictions are listed in Table A.1, and is the same as employed for the base simulation conditions in the kinetic model. For higher hydrocarbons, the reaction can be written as2,17,18,22, 2 22n m mC H nH O nCO n H⎛ ⎞+ → + +⎜ ⎟⎝ ⎠  .........   r1 ∆H 0 298 = 1108 kJ/mol   (for n=7) (A.1) Once H2 and CO are available by steam reforming of higher hydrocarbons, a reverse steam reforming reaction (reverse of Equation A.2) produces CH4 (methanation reaction),  and thereafter the process proceeds as simple steam reforming of methane2,17,18,22. 224 3HCOOHCH +⇔+   .........   r2  ∆H0298 = 206.1 kJ/mol (A.2) 222 HCOOHCO +⇔+   .........    r3  ∆H0298 = -41.1 kJ/mol  (A.3) 2224 42 HCOOHCH +⇔+  ..........  r4  ∆H0298 = 165 kJ/mol  (A.4) Although methane is not present in the feed, it immediately starts to appear in the system due to the methanation reactions (reverse of reactions (2) and (4)), once H2, CO and CO2 appear in the system by reactions (1) and (3). The methane yield decreases with increasing temperature due to the endothermicity of the steam reforming reaction of methane.  As a result, H2 yield  212 continues to increase.  If this H2 is selectively removed from the system, CH4 yield will decrease further, due to forward equilibrium shift of reaction (2). The irreversibility for steam reforming applies to all higher hydrocarbons with different degrees of reactivity. The higher hydrocarbons are generally more reactive than methane, with aromatics showing the lowest reactivity, approaching that of methane23. Industrially, with proper desulfurization, it has been possible to convert light gas oils and diesel fuel into syngas with no trace of higher hydrocarbons in the product gas2. Pilot scale experiments on adiabatic prereforming of natural gas, which also contained higher hydrocarbons in the range C2-C7, showed that the concentration of all higher hydrocarbons decreased continuously through the bed and that no intermediate compounds were observed9.  A.3 Kinetic Modeling of a Fluidized Bed Membrane Reactor A two-phase model of a fluidized bed membrane reactor (FBMR) was prepared to assist with the sizing of an experimental reactor. The bubbling bed regime of operation has been adopted for the simulations for this paper since the experimental reformer will be focused mainly on this regime. Pd-based membrane panels supplied by Membrane Reactor Technologies Limited, a Vancouver based company, will be used in the reactor immersed in the fluidized bed of the catalyst. A distributor design has been adopted in the experimental design which minimizes any effect of jetting. The geometry and reaction base conditions are tabulated in Table A.1. Simulations were performed for a 1 m membrane length. Figure A.2 shows a schematic of the model developed. Double-sided membrane panels are inserted through vertical slits along the height of the reformer shell. The membrane panels pass through the centerline of the reformer, dividing the cross-section into two communicating sections. Thus, the membranes will be in contact with the bubble and dense phases nearly proportionally to the fractions they occupy in the fluidized bed. A.3.1 Model assumptions 1. Steady-state reactor conditions. 2. Isothermal bed. The experimental reactor setup will be externally heated to overcome the high endothermicity of the reaction in addition to allowing isothermal operation. 3. Only the lower dense catalyst bed is simulated; the lean freeboard regime is not treated in this paper. 4. The lower dense catalyst bed is treated as two parallel phases made up of a dense phase and a bubble phase.  213 5. Plug flow behaviour is assumed for the dense phase as well as the bubble phase. The high aspect ratio of the FBMR simulated justifies this assumption. 6. Catalyst diffusion resistance is taken to be negligible. Very fine catalyst particles with a mean particle size of 100 µm will be used for the experiments. 7. Catalyst deactivation is neglected in this paper. 8. Any jetting just above the distributor is neglected. A.3.2 Model equations for reactor side Mole Balance for ith Species in the Bubble Phase. ( ) ∑ = −+−= 4 1j ibbjbijpbibidbbiq ib QRACCAak dL dF εγρφε      (A.5) i = CH4, H2O, CO, CO2, H2, and C7H16 Mole Balance for ith Species in the Dense Phase. ( ) ∑ = −+−= 4 1j iddjdijpdidibbbiq id QRACCAak dL dF εγρφε      (A.6)     i = CH4, H2O, CO, CO2, H2, and C7H16 Subscripts b and d refer to the bubble and dense phases, respectively; γij is the stoichiometric coefficient of component i in the jth reaction (negative for species consumed and positive for products); Qib and Qid are the permeation rates per unit length from the reactor side to the permeation side for the bubble phase and the dense phases, respectively, for species i. A.3.3   Model equations for separation side The differential mole balance equation for the permeate hydrogen is written as: dHdbHb pH QQ dL dF ,, ,2 22 εε +=         (A.7) The hydrogen permeation rate from each phase is calculated from Sieverts’ law: )exp 22 2 2 2 0' , ,pH,bH H H M bH PP(RT E δ PAQ P −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −=       (A.8) )exp 22 2 2 2 0' , ,pH,dH H H M dH PP(RT E δ PAQ P −⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ −=       (A.9) The membranes are assumed to be impermeable to all other species. where PM0 = Pre-exponential factor for permeation = 0.00207 mole/(m.min.atm0.5) and EH2 = Activation energy for permeation = 9180 J/mol  214 A.3.4 Interphase mass exchange coefficient The interphase mass exchange coefficient is calculated based on the correlation by Sit and Grace24. For the ith component: 2 1 4 3 ⎥⎦ ⎤⎢⎣ ⎡+= b bmfiemf iq d UDU k π ε          (A.10) where Die is the effective diffusivity of component i in the gas mixture and is calculated based on the average composition of the bubble and the dense phases, using the correlation25: ( ) ji D y D y n i ij i ie i ≠⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛=− ∑ =1 1         (A.11) where Dij is the binary diffusivity of components i and j.  A.4 Results and Discussion Figure A.3 shows the predicted species concentrations for the 2 phases for operation at 650°C (close to the current maximum temperature of palladium membrane) and 10 bars absolute pressure.  As can be seen, although the reaction occurs predominantly in the dense phase, and there is almost no reaction in the bubble phase, the species concentrations in the two phases are almost identical. This is attributable to the relatively fast mass transfer between the two phases at the temperature of the reformer. Figure A.4 shows that as hydrogen is withdrawn from the reaction mixture, the methane yield decreases, enhancing the hydrogen production. Thus, while on the one hand pure hydrogen is produced due to membrane separation, on the other hand, overall hydrogen yield increases, which is a measure of the reactor performance in this case. Retentate hydrogen yield, which represents the hydrogen left inside the reactor, goes on decreasing as more and more hydrogen permeates through the membranes. Figure A.5 shows that heptane conversion is completed within a few centimetres after the entrance, especially for higher steam-to-carbon ratios. The rest of the reactor then proceeds as in steam reforming of methane. As seen from Figure A.6, with increasing steam-to-carbon ratio, the hydrogen permeate yield is predicted to be enhanced, correspondingly increasing the overall hydrogen yield. Based on these observations, as shown in Figure A.7, the fluidized bed membrane reactor (FBMR), can be considered to be composed of two overlapping zones: Zone 1, a short zone,  215 where steam reforming of heptane is completed, and Zone 2, for steam reforming of methane. Thus, in this bi-functional reaction and separation set-up, a separate pre-reformer is not needed, since with hydrogen permeation, the reaction can proceed towards completion in the same unit. In view of the pure hydrogen permeation, PSA units are also not required. The main challenge for the competitiveness of this technology lies with membrane issues, in particular in assuring pin-hole-free high-flux perm-selective membranes. Figure A.8 shows the effects of decreasing the membrane thickness for a reformer operating at 650°C and 10 bars. Thinner membranes minimize the residual methane and hydrogen in the reformer, and maximize the pure (permeate) hydrogen yield. Figure A.9 shows the increase of hydrogen permeate yield with increasing specific membrane surface area for a reformer operating at 650°C and 10 bars. Steam reforming reactions being very rapid, and hydrogen permeation being slow, an important parameter is the membrane packing factor, ‘a’, defined as the membrane surface area per unit volume of reactor.  As this factor is increased, the reformer performance as measured in terms of pure hydrogen yield, is significantly enhanced, and a significantly smaller reformer can be used.  Thus this multifunctional reactor is predicted to be able to combine the units from a pre-reformer, reformer and hydrogen purifier into a single unit. The sequence of events can be considered to be: i. Steam reforming of higher hydrocarbon, depicted in Figure A.5. ii. Methanation, indicated in Figure A.4a when the peak is attained for the methane yield. iii. Steam reforming of methane, depicted in Figure A.4a, when the methane conversion becomes zero, thus completing the full conversion of the hydrocarbons. iv. Hydrogen permeation until the hydrogen partial pressure in the retentate equalizes with that in the permeate stream, evident from Figure A.4b. v. In parallel with step (iv), net interphase mass transfer between the bubble and dense phases is also completed. When this sequence of events is complete, the species concentrations in the two phases do not change any further, and the concentration profiles remain flat thereafter, as in Figure A.3. The reformer heights corresponding to this sequence of events depend on the operating parameters including reformer pressure, membrane permeate side pressure, reformer temperature, steam-to- carbon ratio in the feed, and superficial velocity.   216 A.5 Conclusions n-Heptane was used as a model component for higher hydrocarbons, close to the naphtha cut. In- situ permselective membranes should be able to produce ultra-pure hydrogen as required by some sectors like the fuel cell industry. Higher conversion of methane (produced by the methanation reaction) allows the reformer to be operated at much lower temperature to achieve the same hydrogen yield as for much higher temperatures without membranes.  A FBMR system for higher hydrocarbons can result in a compact reformer system combining the units from a pre- reformer, reformer and hydrogen purification into a one single unit. However, for the system to be economically viable and competitive, major challenges remain for the membranes.  Desirable membrane features are: • High flux. • High selectivity to hydrogen. • Low cost. • Longevity. • Higher membrane packing, while maintaining a minimum separation requirement in a fluidized space to prevent solids bridging and gas bypassing. Challenges specific to higher hydrocarbons include catalyst deactivation and possible membrane fouling. These have not been considered in this paper, but will be key factors to be examined in the experimental work. The model considers the bubbling bed mode of operation, as this will be the main operating regime in the forthcoming experiments. However, many industrial fluidized bed reactors are operated in the turbulent regime in view of the higher throughput and advantageous features26. The transition from bubble to turbulent flow happens earlier for powders with smaller mean particle diameter and wider particle size distributions27. The experimental work will include determination of this transition at the temperature and pressure of the reformer, and investigate how it affects the hydrogen yield.     217 A.A Appendices A.A.1 Kinetic expressions for reactions in reformer • Steam Reforming of Higher Hydrocarbons:    (Tottrup19) 22 2 HmnnCOOnHHC mn ⎟⎠ ⎞⎜⎝ ⎛ ++→+  r1 For Heptane, n = 7, mol kJH o 1108298 =∆ 2 1 1 2 2 2 2167 167 1 ⎥⎥⎦ ⎤ ⎢⎢⎣ ⎡ ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛+ = H OH b OH HHC a HC P P K P PP K Pk r ⎟⎠ ⎞⎜⎝ ⎛−×= − RT k 67800exp108 51     barhrg mol catalyst ..  2.25=aK       bar-1 077.0=bK       [ - ] • Steam Reforming of Methane:      (Xu and Froment28) CH4 + H2O ⇔ CO + 3H2   r2   mol kJH o 1.206298 =∆ CO + H2O ⇔ CO2 + H2   r3   mol kJH o 1.41298 −=∆ CH4 + 2H2O ⇔ CO2 + 4H2   r4   mol kJH o 0.165298 =∆ 2 1 5.0 5.2 22 2 2 24 DEN K PP P PP kr HCO H OHCH ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ − = 2 2 2 33 2 2 DEN K P P PP kr CO H OHCO ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ − = 2 21 5.0 5.3 2 44 22 2 24 DEN KK PP P PP kr HCO H OHCH ⎟⎟⎠ ⎞ ⎜⎜⎝ ⎛ − = 2 22 4422 1 H OHOH CHCHHHCOCO P PK PKPKPKDEN ++++= The equation parameters are available in Xu and Froment 28.  218 A.A.2 Hydrodynamic equations for the 2-phase model Bubble Size Distribution: ( ) Dhbbmbmb edddd /3.00 −−−= ( )[ ] 4.0064.1 mfbm UUAd −= ( ) ⎥⎦ ⎤⎢⎣ ⎡ −= or mf b N UUA g d 02.00 38.1    (Mori and Wen29) Fraction of Bed Occupied by Bubbles: b mf b U UU −= 0ε Bubble Rise Velocity:   ( ) 2/10 711.0 bmfb gdUUU +−= (Davidson and Harrison30) Minimum Fluidization Velocity: ( )[ ] 7.330408.07.33Re 2/12 −+= Armf (Wen and Yu31,32) Volume Fraction of Solids:  ( )( )mfbd εεφ −−= 11    bb εφ 001.0=      219 Table A.1: Reactor geometry and base simulation parameters  Reformer empty cross-sectional area 2.0 x 10-3 m2 Specific membrane area 64 m2/m3 of reactor volume Total membrane length (along height of reformer) 1 m  Reformer Catalyst type Ni-Al2O3 Catalyst particle mean diameter 100 µm Catalyst particle density 2270 kg/m3  Catalyst Steam:Carbon ratio in feed 3 n-Heptane mole fraction in feed 0.0454 Steam mole fraction in feed 0.9538 H2 mole fraction in feed 0.0008 Feed temperature 650°C Feed pressure 10 bars abs Membrane permeate side pressure 0.3 bars abs Reactor inlet gas superficial velocity 0.23 m/s   Process operating conditions   220 Equilibrium Mole Fractions, P = 20 bars 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 400 450 500 550 600 650 700 750 800 Temperature, oC M ol e Fr ac tio ns H2O H2 CH4 CO2 CO C7H16 Equilibrium Mole Fractions, P = 10 bars 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 400 450 500 550 600 650 700 750 800 Temperature, oC M ol e Fr ac tio ns H2O H2 CH4 CO2 CO C7H16 Equilibrium Mole Fractions, P = 5 bars 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 400 450 500 550 600 650 700 750 800 Temperature, oC M ol e Fr ac tio ns H2O H2 CH4 CO2 CO C7H16 Equilibrium Mole Fractions, P = 1 bar 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 400 450 500 550 600 650 700 750 800 Temperature, oC M ol e Fr ac tio ns H2O H2 CH4 CO2 CO C7H16   Figure A.1:  Equilibrium compositions in n-heptane steam reforming at varying temperatures and pressures  221    Figure A.2:  Schematic diagram of the kinetic model  222 0 20 40 60 80 100 120 140 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Height above Distributor, m Sp ec ie s C on ce nt ra tio n,  m ol s/m 3 H2O CO2 H2 CH4 COC7H16  (a) 0 20 40 60 80 100 120 140 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Height above Distributor, m Sp ec ie s C on ce nt ra tio n,  m ol s/ m3 H2O CO2 H2 CH4 COC7H16   (b)  Figure A.3: Predicted species concentrations in the two phases at 650°C, 10 bars: (a) Dense phase (b) Bubble Phase   223 0 0.5 1 1.5 2 2.5 3 0 0.2 0.4 0.6 0.8 1 Height above Distributor, m CH 4 Y ie ld Mols CH4 Yield/Mol C7H16 Fed  (a) 0 5 10 15 20 25 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Height above Distributor, m H 2 Y ie ld s Permeate H2 Retentate H2 Total H2   (b)  Figure A.4:  Predicted methane and hydrogen yields at 650°C, 10 bars: (a) Methane (b) Hydrogen   224 0 0.2 0.4 0.6 0.8 1 1.2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Height above Distributor, m C7 H 16  C on ve rs io n SCR = 1 SCR = 3 SCR = 6   Figure A.5: Predicted heptane conversions at 650°C, 10 bars  225 0 5 10 15 20 25 0 0.2 0.4 0.6 0.8 1 Height above Distributor, m H 2 P er m ea te  Y ie ld SCR=1 SCR=3 SCR=6  (a) 0 2 4 6 8 10 12 14 16 0 0.2 0.4 0.6 0.8 1 Height above Distributor, m H 2 R et en ta te  Y ie ld SCR=1 SCR=3 SCR=6   (b)  Figure A.6: Predicted effect of S/C ratios on yields at 650°C, 10 bars: (a) Permeate hydrogen (b) Retentate hydrogen   226 OHCHHCO 2423 +⇔+ 222 HCOOHCO +⇔+ 2224 4HCOOHCH +⇔+ 2 22n m mC H nH O nCO n H⎛ ⎞+ → + +⎜ ⎟⎝ ⎠ 224 3HCOOHCH +⇔+ Zone 2 Zone 1 222167 22714 HCOOHHC +→+ 22167 1577 HCOOHHC +→+ Steam + Hydrocarbon (CnHm) H2 H2 Reformer Off-Gas H2 H2 St ea m  R ef or m in g of  C nH m St ea m  R ef or m in g  o f C H 4 H2 Zo ne  2 Zo ne  1   Figure A.7:  Reaction zones in FBMR system for higher hydrocarbons: Pre-reforming, reforming and purification in a single unit  227 0 5 10 15 20 25 0 0.2 0.4 0.6 0.8 1 Height above distributor, m Pu re  H 2 Y ie ld δH2 = 5µm δH2 = 20µm δH2 = 35µm  (a) 0 1 2 3 4 5 6 7 8 9 10 0 0.2 0.4 0.6 0.8 1 Height above distributor, m R et en ta te  H 2 Y ie ld δH2 = 35µmδH2 = 20µmδH2 = 5µm   (b)  Figure A.8: Dependence of hydrogen yields on membrane thickness at 650°C, 10 bars: (a) Permeate hydrogen (b) Retentate hydrogen       228    0 5 10 15 20 25 0 0.2 0.4 0.6 0.8 1 Height above distributor, m Pu re  H 2 Y ie ld a=80m-1 a=64m-1 a=50m-1 a=100m-1 a=190m-1  (a) 0 2 4 6 8 10 12 14 16 18 0 0.2 0.4 0.6 0.8 1 Height above distributor, m R et en ta te  H 2 Y ie ld a=50m-1 a=64m-1 a=80m-1 a=100m-1 a=190m-1   (b)  Figure A.9: Dependence of hydrogen yields on specific membrane area at 650°C, 10 bars: (a) Permeate hydrogen (b) Retentate hydrogen      229 A.6 References 1. Johnston, B.; Mayo, M. C.; Khare, A., Hydrogen: The energy source for the 21st century. Technovation 2005, 25, 569-585. 2. Rostrup-Nielsen, J. R.; Rostrup-Nielsen, T., Large-scale hydrogen production. CatTech 2002, 6, 150-159. 3. Eliezer, D.; Eliaz, N.; Senkov, O. N.; Froes, F. H., Positive effects of hydrogen in metals. Materials Science and Engineering 2000, A280, 220-224. 4. Ramachandran, R.; Menon, R. K., An overview of industrial uses of hydrogen. Int. J. Hydrogen Energy 1998, 23, 593-598. 5. Rostrup-Nielsen, J. R.; Sehested, J.; Norskov, J. K., Hydrogen and synthesis gas by steam and CO2 reforming Advances in Catalysis 2002, 47, 65-139. 6. Kaila, R. K.; Krause, A. O. I., Reforming of higher hydrocarbons. Studies in Surface Science and Catalysis 2004, 147, 247-252. 7. Melo, F.; Morlanes, N., Naphtha steam reforming for hydrogen production. Catal. Today 2005, 107-108, 458-466. 8. Zhu, W.; Han, W.; Xiong, G.; Yang, W., Mixed reforming of heptane to syngas in the Ba0.5Sr0.5Co0.8Fe0.2O3 membrane reactor. Catal. Today 2005, 104, 149-153. 9. Christensen, T. S., Adiabatic prereforming of hydrocarbons - Important step in syngas production. Applied Catalysis A: General 1996, 138, 285-309. 10. Elnashaie, S. S. E. H.; Adris, A. M. In A fluidized bed steam reformer for methane, Fluidization VI, 1989; Grace, J. R.; Shemilt, L. W.; Bergougnou, M. A., Eds. Engineering Foundation: 1989; pp 319-326. 11. Elnashaie, S. S. E. H.; Elshishini, S. S., Modelling, simulation and optimization of industrial fixed bed catalytic reactors. Gordon and Breach Science Publishers: London, 1993. 12. Soliman, M. A.; Elnashaie, S. S. E. H.; Al-Ubaid, A. S.; Adris, A., Simulation of steam reformers for methane. Chem. Eng. Sci. 1988, 43, 1801-1806. 13. Adris, A. M. A fluidized bed steam reformer for methane. University of Salford, UK, Salford, 1989. 14. Adris, A. M.; Elnashaie, S. S. E. H.; Hughes, R., A fluidized bed membrane reactor for the steam reforming of methane. Can. J. Chem. Eng. 1991, 69, 1061-1070.  230 15. Adris, A. M.; Lim, C. J.; Grace, J. R., The fluidized bed membrane reactor for steam methane reforming: Model verification and parametric study. Chem. Eng. Sci. 1997, 52, 1609- 1622. 16. Grace, J.; Elnashaie, S. S. E. H.; Lim, C. J., Hydrogen production in fluidized beds with in-situ membranes. Int. J. Chem. React. Eng. 2005, 3, A41. 17. Chen, Z. A novel circulating fluidized bed membrane reformer for efficient pure hydrogen production for fuel cells from higher hydrocarbons. Auburn University, 2004. 18. Chen, Z.; Yan, Y.; Elnashaie, S. S. E. H., Modeling and optimization of a novel membrane reformer for higher hydrocarbons. AIChE J 2003, 49, 1250-1265. 19. Tøttrup, P. B., Evaluation of intrinsic steam reforming kinetic parameters from rate measurements on full particle size. Applied Catalysis A: General 1982, 4, 377-389. 20. Darwish, N. A.; Hilal, N.; Versteeg, G.; Heesink, B., Feasibility of the direct generation of hydrogen for fuel-cell-powered vehicles by on-board steam reforming of naphtha. Fuel 2004, 83, 409-417. 21. Puolakka, K. J.; Krause, A. O. I., CO2 reforming of n-Heptane on a Ni/Al2O3 catalyst. Studies in Surface Science and Catalysis 2004, 153, 329-332. 22. Chin, S. Y.; Chin, Y. H.; Amiridis, M. D., Hydrogen production via the catalytic cracking of ethane over Ni/SiO2 catalysts. App. Catalysis A: General 2006, 300, 8-13. 23. Rostrup-Nielsen, Catalytic steam reforming. Catalysis Science and Technology, ed. J.R. Andersen and M. Boudart 1984, Springer-Verlag, 1-117 24. Sit, S. P.; Grace, J. R., Effect of bubble interaction on interphase mass transfer in gas- fluidized beds. Chem. Eng. Sci. 1981, 36, 327-335. 25. Wilke, C. R.; Lee, C. Y., Estimation of diffusion coefficients for gases and vapors. Ind. Eng. Chem 1955, 47, 1253. 26. Bi, H. T.; Ellis, N.; Abba, I. A.; Grace, J. R., A state-of-the-art review of gas-solid turbulent fluidization. Chem. Eng. Sci. 2000, 55, 4789-4825. 27. Sun, G.; Grace, J. R., Effect of particle size distribution in different fluidization regimes. AIChE J 1992, 38, 716-722. 28. Xu, J.; Froment, G. F., Methane steam reforming, methanation and water-gas shift: I. intrinsic kinetics. AIChE J 1989, 35, 88-96. 29. Mori, S.; Wen, C. Y., Estimation of bubble diameter in gaseous fluidized beds. AIChE J 1975, 21, 109-115.  231 30. Davidson, J. F.; Harrison, D., Fluidized Particles. Cambridge University Press: Cambridge 1963. 31. Wen, C. Y.; Yu, Y. H., A generalized method for predicting the minimum fluidization velocity. AIChE J. 1966, 12, 610-612. 32. Wen, C. Y.; Yu, Y. H., Mechanics of fluidization. Chemical Engineering Progress Symposium Series 1966, 62, 100-111.  232 APPENDIX B.      FBMR OPERATING MANUAL3   B.1 Introduction This document outlines the start-up strategy, experimentation, emergency and normal shutdown processes to ensure safe operation. It is assumed at this stage that the reactor system has already been tested for the basic safety requirements, and that commissioning has been completed successfully. However, routine checks must be done before every start-up.  B.2 Steam Reforming Experiments: Reactor Start-up The MAWP of the reactor pressure vessel is 1020.5 kPag at 621°C. The maximum heating rate for any part of the pressure vessel is to be 5°C/min. The temperature is increased using electrical heaters and decreased gradually to allow for compression/expansion of parts. Keeping in view the personnel safety issues, temperature tolerance of the membranes, and catalyst stability, it was decided to use a maximum temperature of 575°C, and a maximum pressure of 1000 kPa for any combination of operating conditions. Step 1: Prior Preparations (Inventory Check) (1.a)  Catalysts • Install membrane dummies or membrane panels as applicable. Fill the reactor with RK- 212 catalyst powder to be able to just submerge the topmost membrane/dummy. The reactor must be leak tested by pressurization every time the reactor is started. Put on the insulation jackets. (1.b)  Desulfurizer sorbent • When operated with natural gas, replace sorbent in desulfurizer every 48 hours of operation. (1.c)  Water • Fill water tank. • Prime the water pump by opening V-515.2, and allowing water to flow out until no air bubbles are detected at the outlet.  3 Input was received from Ali Gulamhusein (Membrane Reactor Technologies Limited) while preparing this document.  233 • Close V-515.2 and cap line to prevent leakage. (1.d)  Gases & liquid hydrocarbons • Check gas pressures (nitrogen, hydrogen, and natural gas or helium cylinder for liquid hydrocarbon headspace). Cylinders must be changed/ refilled if pressure close to 550 psi. • Open valve on nitrogen cylinder. Set PI-101.2 to 150 psig. • Open valve on hydrogen cylinder. Set PI-301.2 to 150 psig. • If performing natural gas steam reforming, start natural gas compressor, and set it on AUTO mode, which starts itself to refill cylinder and stops when a preset pressure value is reached in the cylinder. • If performing natural gas steam reforming, start warm water flow for irrigating PR-401.2. • If performing liquid hydrocarbons steam reforming, set PI-1201.2 to 250 psig. • Make sure that there is enough liquid hydrocarbons storage in the respective tanks. Step 2: Prior Preparations (System Purging) (2.a)  Purge the hydrogen feed line with nitrogen ƒ Close V-421.2 to prevent N2 purge to reactor ƒ Open XV-315.1 (H2 solenoid) ƒ Fully open V-115.1 and partially open the mass flow controller FICV-301.1 (~3 slpm). ƒ Partially open the mass flow controller FICV-101.1 (~3 slpm) ƒ Gradually open nitrogen supply via V-101.1. ƒ Allow pressure to build-up. ƒ Gradually release N2 via V-311.1 ƒ Close V-311.1 and allow pressure to rebuild, then gradually release via V-311.1 ƒ Repeat 2 or 3 times ƒ Close V-101.1, FICV-301.1, V-115.1, and XV-315.1  (2.b)  Purge the reactor fully with nitrogen This is done automatically, while fluidizing the reactor during process heaters start-up, in Step 3. (2.c)  Purge the permeate section with nitrogen If membrane panels are installed, purge the permeate lines including the hydrogen pump. ƒ Ensure V-015.1 is closed ƒ Open V-010.1  234 ƒ Gradually open V-015.1, avoid sudden spikes in flow as this will hamper control of PCV-600.3 (uses same N2 supply) ƒ Allow purge to continue for 15 minutes. Also, repeat this, with pump bypass valve V- 718.4 open. Step 3: Heating up the FBMR and Catalyst Reduction (3.a)  Fluidizing with nitrogen • Close V-101.1. • Open V-421.2 and the solenoid valve XV-119.1. • Set nitrogen flow rate on FICV-101.1. Recommended nitrogen flow rate = 6 slpm, for good fluidization behaviour. (Also monitor via the differential pressure transducers along the FBMR height). (3.b)  Turning on the heaters • Prepare the connections for the external heaters: Turn on the main power supply, turn on the voltage transformer. Next, turn on the heaters switch on the power distribution box on the FBMR rig. Reset Emergency Stop button, if engaged. • On the HMI program, start the External Heaters, and the Internal Heaters. • Heat up the reactor, with only nitrogen flowing. ƒ Specify 30% of full-scale heating rate for the internal as well as the external heaters. Adjust the settings to ensure heating rate of 5°C/min to 500°C. The skin temperature of the reactor vessel at any point should not exceed 600°C. • Gradually ramp up vaporizer heaters to maintain a maximum feed line temperature of 450°C (TT-1001.2), while not exceeding 500°C skin temperature (TT-1000.2) (3.c)  Introducing hydrogen • Reduce the catalyst overnight (12 hours) with hydrogen-nitrogen mix. ƒ Hydrogen should not be introduced at temperatures below 350°C, especially when membranes are in use. ƒ Open hydrogen solenoid valve XV-315.1. ƒ Set nitrogen and hydrogen flow rates on FICV-101.1 and FICV-301.1 respectively. Both of these are controlled by a stand-alone Brooks read-out box. For hydrogen- nitrogen mix, recommended flow rates are hydrogen and nitrogen are 1.5 slpm and 4.5 slpm respectively.   235 Step 4: Introducing Steam and Hydrocarbon Feeds The hydrocarbons are to be sequentially used as desulfurized natural gas, propane, and finally heptane. This stage of steam reforming experiments is similar for all three types of hydrocarbons, and is described only for natural gas here. (4.a)  Steam introduction • Start cooling water to ROG vent condenser line, and to gas sampling condenser line. • After overnight catalyst reduction, the reactor is ready for steam introduction. Ramp up or down the FBMR temperature to the initial desired operating temperature. • Pressurize the FBMR to the initial desired operating pressure, by setting the PCV-600.3. • Set the nitrogen flow rate to be 3 slpm (FICV-101.1); however, stop the nitrogen solenoid valve XV-101.1. Set the hydrogen flow rate to 6 slpm. • Open the water solenoid valve XV-501.2. • Start water pump, and gradually increase the water flow to the required value, and correspondingly decrease the hydrogen flow rate so as to give a superficial velocity of 6 cm/s. At no time (while there is no hydrocarbon feed), should the steam-hydrogen ratio exceed 6. The recommended steam-hydrogen ratio would be 4. • After steam flow stabilizes (indicated by a stable temperature for the ROG exit temperature TT-1002.3, introduce the hydrocarbon feed. (4.b)  Natural gas introduction • Before initiating NG flow, ensure that cooling water to the NG regulator is on. • After period of stability, establish NG flow: ƒ Open XV-419.2 (NG solenoid) ƒ Set set-point on FICV-401.2 to required set-point. • Monitor reactor temperatures. Increase in heater duty may be required due to reforming action • Shut off hydrogen upon GC confirmation of reaction (presence of CO2 detected) by closing solenoid XV-315.1.  Now the reformer is fully operational!!!    236 B.3 During Steam Reforming Experiments The permeate gas flow rates, temperatures, and bed pressures are logged into the computer automatically, and are also recorded manually every 10 minutes. Once the hydrocarbon has been introduced, one of the sample gas lines is monitored for gas composition. When the gas composition variation is within absolute deviations of ~ ±1%, and verified with another gas sampling line, steady state is assumed to have reached. Gas samples are analyzed at various locations, repeated as required.  B.4 Keep-Warm Mode of Operation (1.a)  Transition from experiments mode to keep-warm mode  • Set Internal Heaters to 20% • Set External Heaters to 20% • Adjust above two power input rates, so as to avoid temperature overshoot, and triggering ESD, which may happen feeds are turned off. • Release FBMR pressure by slowly opening PCV-600. • Gradually ramp down the hydrocarbon and water flows and ramp up nitrogen flows, maintaining the superficial velocity at 6 cm/s. The SCR should never fall below 4. • Stop NG flow by closing solenoid valve XV-419.2. • Stop water flow by closing solenoid valve XV-501.2. • Monitor reactor temperatures, further decrease in heater duty may be required due to no reforming action. Maintain temperatures to 500°C, and fully open the PCV-600. • Maintain 10% hydrogen and 90% nitrogen flow rate to maintain a superficial velocity of 5 cm/s. (1.b)  Transition from keep-warm mode to experiments mode  • Continue with Step 4 as described in Section B • Resume experiments when operating conditions are reached  237 B.5 Normal Shutdown (1.a)  Depressurize FBMR, shut off feeds and introduce purge gas • Reduce temperature to 450°C • Open shunt valves for the differential pressure transducers • Release FBMR pressure by slowly opening PCV-600. • Gradually ramp down hydrocarbon and water flows and ramp up nitrogen flows, maintaining the superficial velocity at 6 cm/s. The SCR should never fall below 4. • Stop hydrocarbon flow, and start hydrogen flow. A steam/hydrogen molar ratio of 4 should be maintained to keep the catalyst reduced. • Continue steaming for 10 minutes. • Stop steam and hydrogen. Close water (XV-501.2) and hydrogen (XV-315.1) solenoid valves. • Close steam feed isolation valve. Close manual and solenoid valve on feed delivery from water tanks. • Maintain nitrogen flow rate to maintain a superficial velocity of 5 cm/s. (1.b)  Reduce temperature • Shut down hydrogen pump • Turn off the internal heaters • Reduce power input to external heaters to allow FBMR cooling at the rate of 5°C/min • Continue nitrogen purging till the temperature falls below 100°C. • Shut down all heaters completely. • It is important that all steam be purged from the vaporizer and reactor. Maintain nitrogen flow for at least 2 hours after steam shutoff. Shut off nitrogen flow. (1.c)  If shutting down for a lengthy period • If shutting down for a lengthy period, isolate helium pad from the liquid hydrocarbon tanks and vent pressure. • Isolate all gas cylinders and feed valve. • Empty the water tank. • Reduce and stop process nitrogen via the HMI controller. Close process nitrogen supply valve and isolate cylinders. • Close purge nitrogen supply valve and isolate cylinders.  238 B.6 Emergency Shutdown • Push the STOP button on the PLC Panel door, and confirm the shutdown from the Control Computer adjacent to the Control Panel OR Click the HMI Emergency Shutdown button on the Control Computer and confirm the action. • Press the red Emergency Shutdown button on the FBMR heaters power distribution box, and turn off the external heaters switch. • Close all hydrogen bottles in flammables storage area outside east exit of the building. • Close natural gas compressor located outside north east corner of the building. Close both natural gas tanks in flammables storage area outside east exit of the building. • Shut off natural gas feed (V-420.2) and higher hydrocarbon feed (V-1206.2) adjacent to small window (east wall) where all gas/liquid lines enter the building. • Shut off propane valve (V-1202.2) and heptane valve (V-1203.2) in liquid flammables storage area outside east exit of the building. • Contact concerned persons as per emergency contact list in sheet above. • Pressing the red Emergency Shutdown button on the PLC panel does the following: ƒ shuts off the feed flow solenoid valves ƒ opens nitrogen solenoid for purge ƒ stops the water pump, and hydrogen pump ƒ turns off all heaters ƒ opens PCV-600.3   239 APPENDIX C.      FBMR ASSEMBLY DRAWINGS   This Appendix shows some of the representative and important FBMR assembly drawings. The full set of fabrication drawings and design calculations are available from M. Rakib and J. Grace as electronic files in a folder Rakib_FBMR_Assembly_Drawings.  240   Figure C.1: FBMR assembly: Shell weldment  241   Figure C.2: FBMR pressure vessel assembly  242   Figure C.3: Typical rectangular cover for side opening  243   Figure C.4: Assembly of rectangular cover and membrane panel  244   Figure C.5: Assembly of inlet head, showing feed distributor  245   Figure C.6: General arrangement of FBMR on reactor stand   246   Figure C.7:  Location of band heaters (denoted in brown) mounted on the FBMR  247 APPENDIX D.      CATALYST EVALUATION UNIT   This Appendix shows the main process flow diagrams for the catalyst evaluation unit (micro- reactor set-up). These were prepared in collaboration with Alexandre Vigneault, a fellow graduate student.  248  Figure D.1: Catalyst evaluation unit: Micro-reactor feeding system, Part I   249   Figure D.2: Catalyst evaluation unit: Micro-reactor feeding system, Part II  250  Figure D.3: Catalyst evaluation unit: Micro-reactor and gas analysis   251 APPENDIX E.      HYDRODYNAMIC BEHAVIOUR IN A PLEXIGLAS COLUMN AND THE FBMR   E.1   Fluidizability of the Catalyst Particles A commercial naphtha steam reforming catalyst RK-212 manufactured by Haldor Topsoe A/S was used for the experiments. The catalyst, available in the form of 7-holed cylindrical catalyst pellets, was crushed and sieved to obtain the desired size ranges of the particles.  During the commissioning stage, the reactor was loaded with particles of the size cut +90 µm -125 µm. In another trial, the FBMR was loaded with a mixture of equal fractions of particles of size cuts +63 µm -90 µm, +90 µm -125µm, and +125µm -150µm. The gas superficial velocities was varied up to 0.18 m/s, using dry nitrogen gas (Industrial grade nitrogen, supplied by Praxair Inc.). In all these cases, the pressure transducers along the height of the FBMR (as shown in Figure 2.5) indicated very poor quality of fluidization, with different degrees of bed activity indicated along the FBMR height, some locations even indicating no fluidization at all. A Plexiglas column, shown in Figure E.1, was constructed for cold hydrodynamic studies to understand, by better visualization, the hydrodynamics in the FBMR unit. The dimensions of this cold model, provided in Figure E.2, are similar to, but not identical to the FBMR vessel due to constraints of materials availability. Bed activity trends monitored by pressure transducers for the cold column were similar to those for the FBMR, the reason attributed to channelling in the long vertical column equipped with the membrane panels. The particles exhibited behaviour similar to Group C particles, with occasional lifting of the whole bed with a horizontal gap translating upwards, as seen in Figure E.3. On the contrary, spent FCC particles of mean size 100 µm exhibited smooth fluidization. Crushed RK-212 particles of size cuts +150 µm -180 µm, and +180 µm -212 µm were tested separately in the cold column, and good fluidization was achieved in both cases. However, in both cases for the RK-212 particles, initially there were small vertical bridges of un-fluidized zones, which disappeared in about two hours. Alumina-supported nickel and/or precious metal catalysts were used in previous research1, with mean particle sizes in the range of 108 µm or lower. However, those catalysts were prepared by a catalyst manufacturer by impregnating the active material on high-quality  252 alumina powder supplied to them. As such these particles with higher sphericity or on-purpose fluidizable catalyst particles, e.g. FCC particles, would demonstrate easier and better fluidizability than those of irregular shape prepared by crushing commercial catalyst pellets . Figure E.4 shows enlarged views of fresh catalyst particles (after crushing and sieving); Figure E.5 shows enlarged views of catalyst particles unloaded at the end of a series of experiments. These show that the freshly loaded particles were very irregular and jagged, whereas the used catalyst is much more rounded and regular, the sphericity going from ~0.3 to ~0.8. While particles smaller than 125 µm remained stagnant in general, initial mobilization of bigger particles in the cold model column was still found to be difficult, which could have been due to entanglement of the jagged surfaces of the fresh particles, aggravated by Van der Waals forces as well as by moisture absorbed while stored after crushing. However, once mobilized, the hydrodynamic activity improved in general, possibly due to drying of the catalyst particles while fluidizing with dry nitrogen gas, as well as rounding of the particles during operation.  E.2   Superficial Gas Velocities in the FBMR The cold column also gave an understanding of the bubble behaviour in the range of the superficial gas velocities encountered in the FBMR. Working at ambient conditions of temperature and pressure, superficial velocities higher than ~ 0,08 m/s led to big bubbles or slugs in the upper half of the bed. This indicates that in the high-temperature tests, the FBMR operated in the bubbling bed flow regime in some cases, and formed slugs in others, especially with the superficial velocities varying widely along height as shown below. However, caution is needed as behaviour could differ at elevated temperatures and pressures in the FBMR operation. Figures E.6 to E.11 plot the superficial gas velocities for representative cases of steam reforming experiments with no membranes, one membrane panel, and six membrane panels installed for propane and heptane. The temperature profiles for the experiments are also shown with each plot. Four factors caused the changes in superficial velocity: (1) Intermittent abrupt variations of the superficial gas velocity occur due to changes in cross- sectional area in the spaces between adjacent membrane panels (or dummies), compared to the spans covered by these panels. (2) In general the superficial gas velocity follows the trend of the temperature variations.  253 (3) The steam reforming reactions led to a net increase in the molar flow rate in the FBMR. This caused steep increases in the superficial velocity near the FBMR entrance, where the higher hydrocarbon (propane or heptane) conversion is completed. Colder sections in the FBMR can also encounter methanation reactions leading to decrease in superficial velocity, while a temperature increase can increase the molar flow rate due to methane steam reforming, as well as increasing the molar volumetric flow. (4) Volumetric flow, and hence superficial velocity, is also affected by the hydrogen removal via membranes. Especially in regions with an increasing temperature, a decrease in the gas superficial velocity can be seen in the spans occupied by active membrane panels.  E.3   Future Work with Cold Model Experiments with the cold Plexiglas replica of the FBMR were conducted to understand the difficulties of fluidizing the RK-212 catalyst particles, and to select a suitable catalyst particle size for good fluidizability. The column can be used in future to study the degree of gas back- mixing in the fluidized bed of such a high aspect ratio. It could also be used to study the hydrodynamics in a column with an unusual fluidized bed geometry, as well as to provide guidance on reactor modelling.  254   Figure E.1: Plexiglas column for hydrodynamic studies  255  27 6.875 66.875 .875 27 4.5 6.0 3.5 3.0 9.125 2.8 2 2 .373 3.0 1.375 6.0 0.252.0 3.0 4.0 3/8" Holes 0.03" Holes 3.3 2.7 S ec tio n 1 S ec tio n 2 S ec tio n 3 Section 1 Baffles Distributor Plate Section 2   Figure E.2: Plexiglas column dimensions  256    Figure E.3: Catalyst bed being lifted by the gas  257       Figure E.4: Fresh catalyst particles  258        Figure E.5: Used catalyst particles   259 Lo ca l Te m p.  (o C ) 420 470 520 570 Height above Distributor (m) 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Su pe rfi ci al  G as V el oc ity  (m /s ) 0.08 0.10 0.12 0.14 0.16   Figure E.6: Gas superficial velocities for experiment Propane 1.a  Lo ca l Te m p.  (o C ) 420 470 520 570 Height above Distributor (m) 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Su pe rf ic ia l G as V el oc ity  (m /s ) 0.06 0.08   Figure E.7: Gas superficial velocities for experiment Propane 2.d  260 Lo ca l Te m p.  (o C ) 420 470 520 570 Height above Distributor (m) 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Su pe rf ic ia l G as V el oc ity  (m /s ) 0.06 0.08   Figure E.8: Gas superficial velocities for experiment Propane 3.c Lo ca l Te m p.  (o C ) 420 470 520 570 Height above Distributor (m) 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Su pe rf ic ia l G as V el oc ity  (m /s ) 0.08 0.10 0.12 0.14   Figure E.9: Gas superficial velocities for experiment Heptane 1.a  261 Lo ca l Te m p.  (o C ) 420 470 520 Height above Distributor (m) 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Su pe rf ic ia l G as V el oc ity  (m /s ) 0.06 0.08   Figure E.10: Gas superficial velocities for experiment Heptane 4.d Lo ca l Te m p.  (o C ) 420 470 520 570 Height above Distributor (m) 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 Su pe rf ic ia l G as V el oc ity  (m /s ) 0.06 0.08   Figure E.11: Gas superficial velocities for experiment Heptane 5.b    262 E.4 References  1. Boyd, T.; Grace, J.; Lim, C. J.; Adris, A. E. M., Hydrogen from an internally circulating fluidized bed membrane reactor. International Journal of Chemical Reactor Engineering 2005, 3, A58.  263 APPENDIX F.      MODEL SENSITIVITY ANALYSIS   This appendix examines the sensitivity of the reactor model described in Chapter 5 in order to understand the relative importance of the various resistances to reaction inside the FBMR, as well as the effect of uncertainties in estimating different parameters in the model. Experiment Propane 3.c is taken as a basis for estimating the effects.  F.1   Model Sensitivity to Reaction Rate Constants The kinetic rate constants for all reactions involved in the FBMR were first varied by a factor of 10 upwards and downwards compared with those based on earlier studies in the literature. Figure F.1 shows the reactor performance with these major variations in rate constants. Some variations in the performance can be seen near the entrance of the reactor, affected mainly by the propane steam reforming kinetics. However, in general, it can be seen that in the bulk of the bed, there was very little difference in the local yields of methane, carbon oxides or hydrogen for these variations in the reaction rate constants.  Figures F.2 and F.3 show the methane and hydrogen concentrations in the dense and bubble phases. The model considers a very lean concentration of catalyst particles in the bubble phase (Equation 5.12). While reduced catalyst rate constants (multiplication by a factor of 0.1) do not appreciably change the deviation between the dense and bubble phase concentrations, multiplying by a factor of 10 lowers the concentration difference between these two phases for methane and hydrogen.  F.2   Model Sensitivity to Interphase Mass Transfer The rates of diffusional and convective mass transfer of components between the bubble and dense phases were varied by a factor of 10 lower and higher compared with those obtained from the Sit and Grace (1981) correlation.  Figure F.4 shows the reactor performance with variations of the interphase mass transfer. Propane fed to the distributor also goes to the bubble phase, and higher mass transfer (factor of 10) transfers the propane almost immediately to the dense phase. However, slower mass transfer (factor of 0.1) retains more propane in the bubbles, thereby delaying the overall conversion of propane, which can only occur where there are catalyst particles, i.e. in the dense phase. Since  264 methane is an intermediate component, it appear more slowly in the reactor, and its overall conversion is also delayed compared to the actual mass transfer rate case. This is explained from the methane concentrations in the bubble and dense phases as seen from Figure F.5, with a cross- over occurring between the two concentration curves.  While the tenfold changes in interphase mass transfer are discernible, the effects are not very significant.  Hence, interphase mass transfer, while not a negligible factor, plays a secondary role with respect to the overall reaction model.  Since the hydrodynamics of the bed mostly enter the model through the interphase mass transfer, one may also conclude that accurate portrayal of the bed hydrodynamics is of secondary importance for this particular process and the operating conditions investigated.  F.3  Conclusions The model sensitivity studies with respect to reaction rate, show that in general for this FBMR, the kinetics are fast enough at all temperatures tested for the role of chemical kinetics to be insignificant in determining the FBMR performance. Interphase diffusional mass transfer rate is shown to be somewhat more significant in affecting the reactor performance, but to nevertheless still play a secondary role.  Given these findings, it is evident from these studies that the FBMR performance is primarily controlled by chemical equilibrium and by hydrogen permeation through the membranes.  Hence the model is sensitive to accurately characterizing the chemical equilibrium and hydrogen permeation, but relatively insensitive to predicting the chemical kinetics, interphase mass transfer and hydrodynamics with precision.    265 Pr op an e C on ve rs io n 0.00 0.25 0.50 0.75 1.00 0.1 x  Kinetic rate constants 1 x  Kinetic rate constants 10 x  Kinetic rate constants Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 To ta l H 2 Y ie ld 0 2 4 6 8 10 0.1 x  Kinetic rate constants 1 x  Kinetic rate constants 10 x  Kinetic rate constants Pe rm ea te  H 2 Y ie ld 1 3 5 7 0.1 x  Kinetic rate constants 1 x  Kinetic rate constants 10 x  Kinetic rate constants M et ha ne   Y ie ld 0.1 0.3 0.5 0.7 0.1 x  Kinetic rate constants 1 x  Kinetic rate constants 10 x  Kinetic rate constants Lo ca l Te m p.  (o C ) 420 470 520 570 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 0.1 x  Kinetic rate constants 1 x  Kinetic rate constants 10 x  Kinetic rate constants  Figure F.1: FBMR performance with variations of reaction rate constants for experiment Propane 3.c  266 M et ha ne  C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 Bubble phase Dense phase Freeboard M et ha ne  C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 Bubble phase Dense phase Freeboard Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 Bubble phase Dense phase Freeboard Lo ca l Te m p.  (o C ) 420 470 520 570 1 x Kinetic rate constants 0.1 x Kinetic rate constants 10 x Kinetic rate constants   Figure F.2:  Methane concentrations with variations of reaction rate constants for experiment Propane 3.c  267 H yd ro ge n C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 14 16 Bubble phase Dense phase Freeboard H yd ro ge n C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 14 16 Bubble phase Dense phase Freeboard Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 14 16 Bubble phase Dense phase Freeboard Lo ca l Te m p.  (o C ) 420 470 520 570 0.1 x Kinetic rate constants 1 x Kinetic rate constants 10 x Kinetic rate constants   Figure F.3:  Hydrogen concentrations with variations of reaction rate constants for experiment Propane 3.c    268 Pr op an e C on ve rs io n 0.00 0.25 0.50 0.75 1.00 0.1 x  Mass Transfer 1 x  Mass Transfer 10 x  Mass Transfer Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 To ta l H 2 Y ie ld 0 2 4 6 8 10 0.1 x  Mass Transfer 1 x  Mass Transfer 10 x  Mass Transfer Pe rm ea te  H 2 Y ie ld 1 3 5 7 0.1 x  Mass Transfer 1 x  Mass Transfer 10 x  Mass Transfer M et ha ne   Y ie ld 0.1 0.3 0.5 0.7 0.1 x  Mass Transfer 1 x  Mass Transfer 10 x  Mass Transfer Lo ca l Te m p.  (o C ) 420 470 520 570 C ar bo n O xi de s  Y ie ld 0.3 0.5 0.7 0.9 0.1 x  Mass Transfer 1 x  Mass Transfer 10 x  Mass Transfer  Figure F.4:  FBMR performance with variations of interphase mass transfer for experiment Propane 3.c  269 M et ha ne  C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 Bubble phase Dense phase Freeboard M et ha ne  C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 Bubble phase Dense phase Freeboard Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 M et ha ne  C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 Bubble phase Dense phase Freeboard Lo ca l Te m p.  (o C ) 420 470 520 570 1 x Mass Transfer 0.1 x Mass Transfer 10 x Mass Transfer   Figure F.5: Methane concentrations with variations of interphase mass transfer for experiment Propane 3.c   270 H yd ro ge n C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 14 16 Bubble phase Dense phase Freeboard H yd ro ge n C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 14 16 Bubble phase Dense phase Freeboard Height above Distributor (m) 0.0 0.5 1.0 1.5 2.0 2.5 H yd ro ge n C on ce nt ra tio n (m ol s/ m 3 ) 0 2 4 6 8 10 12 14 16 Bubble phase Dense phase Freeboard Lo ca l Te m p.  (o C ) 420 470 520 570 0.1 x Mass Transfer 1 x Mass Transfer 10 x Mass Transfer   Figure F.6:  Hydrogen concentrations with variations of interphase mass transfer for experiment Propane 3.c   271 F.4 References 1. Sit, S. P.; Grace, J. R., Effect of bubble interaction on interphase mass transfer in gas- fluidized beds. Chemical Engineering Science 1981, 36, 327-335.  272  APPENDIX G.      EXPERIMENTAL DATA TABULATION (FBMR)    FBMR Temperature (°C) 0.01 m 400 0.32 m 487 0.52 m 471 0.78 m 502 1.08 m 497 1.29 m 514 1.59 m 525 Bed Average 499 2.26 m 423  Time Averaged Conditions P (kPa abs) 803 Pm (kPa abs) 52 Feed CH4 (slpm) 3.059  Feed Water (g/h) 737.4 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 1.169 Permeate 2 0.832 Permeate 3 0.610 Permeate 4 0.350 Permeate 5 0.943 Permeate 6 1.403  Experiment: Methane 1.a Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 37.39 52.15 0 10.46 0.219 37.10 52.63 0 10.28 0.371 40.91 35.32 0.53 23.24 0.371 41.27 34.62 0.55 23.56 0.473 29.63 34.81 0.27 35.28 0.473 29.05 35.19 0.27 35.49 0.625 29.01 25.32 0.51 45.16 0.727 45.42 37.16 0.40 17.01 0.727 44.63 36.66 0.43 18.28 0.879 45.20 28.16 0.90 25.73 0.981 41.63 30.47 0.70 27.20 0.981 41.25 30.38 0.71 27.66 1.133 41.43 30.45 0.70 27.42 1.235 47.27 26.29 1.02 25.43 1.387 48.15 22.65 1.36 27.84 1.489 43.50 24.59 1.10 30.81 1.489 43.27 24.33 1.15 31.25 1.489 43.28 24.03 1.14 31.54 1.489 43.02 23.19 1.17 32.62 1.489 42.20 22.49 1.33 33.99 1.641 40.20 26.02 1.05 32.74 2.327 35.45 29.91 0.58 34.05 2.327 35.27 29.57 0.59 34.57   273  FBMR Temperature (°C) 0.01 m 402 0.32 m 486 0.52 m 470 0.78 m 494 1.08 m 499 1.29 m 515 1.59 m 538 Bed Average 500 2.26 m 440  Time Averaged Conditions P (kPa abs) 610 Pm (kPa abs) 52 Feed CH4 (slpm) 2.29  Feed Water (g/h) 553 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 0.793 Permeate 2 0.643 Permeate 3 0.434 Permeate 4 0.245 Permeate 5 0.695 Permeate 6 1.051  Experiment: Methane 1.b Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 40.29 47.78 0 11.93 0.219 41.29 46.25 0 12.46 0.371 41.45 32.91 0.47 25.17 0.371 41.71 33.23 0.49 24.57 0.473 32.22 34.63 0.30 32.86 0.473 32.25 34.26 0.31 33.18 0.625 32.70 26.66 0.63 40.01 0.727 47.49 34.48 0.48 17.55 0.727 46.27 35.56 0.47 17.71 0.727 47.15 34.73 0.48 17.65 0.879 44.60 29.87 0.79 24.74 0.981 43.96 26.22 0.91 28.91 0.981 44.70 27.53 0.88 26.88 1.133 42.60 21.54 1.20 34.67 1.235 46.09 27.78 0.87 25.26 1.387 47.24 24.04 1.09 27.63 1.387 46.26 25.35 1.06 27.33 1.489 45.14 19.63 1.56 33.67 1.489 45.50 19.18 1.56 33.75 1.489 45.22 19.54 1.51 33.73 1.489 45.74 20.03 1.52 32.71 1.489 46.03 20.16 1.51 32.30 1.489 46.14 20.02 1.55 32.30 1.489 45.85 20.66 1.52 31.97 1.489 45.52 20.67 1.50 32.31 1.489 45.59 19.82 1.49 33.10 1.489 45.35 19.65 1.46 33.54 1.489 45.13 20.54 1.43 32.89 1.489 45.87 19.49 1.55 33.09 1.641 44.00 23.63 1.31 31.06 2.327 40.08 25.08 0.88 33.96 2.327 40.70 24.93 0.86 33.51  274  FBMR Temperature (°C) 0.01 m 405 0.32 m 482 0.52 m 463 0.78 m 490 1.08 m 496 1.29 m 515 1.59 m 545 Bed Average 498 2.26 m 450  Time Averaged Conditions P (kPa abs) 608 Pm (kPa abs) 25 Feed CH4 (slpm) 2.29  Feed Water (g/h) 553 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 1.331 Permeate 2 0.991 Permeate 3 0.697 Permeate 4 0.377 Permeate 5 0.894 Permeate 6 1.409  Experiment: Methane 1.c Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 38.00 49.58 0.00 12.41 0.219 37.29 50.63 0.00 12.08 0.371 36.69 31.66 0.47 31.18 0.371 36.34 32.10 0.47 31.09 0.473 29.38 28.04 0.37 42.21 0.473 29.02 29.24 0.35 41.40 0.625 28.45 16.42 0.67 54.45 0.727 44.98 35.06 0.45 19.50 0.727 43.18 36.25 0.42 20.14 0.879 42.66 27.78 0.82 28.74 0.981 41.49 23.76 0.96 33.79 0.981 41.21 24.38 0.99 33.42 0.981 41.94 24.72 0.97 32.37 1.133 39.04 17.61 1.25 42.11 1.235 42.45 24.71 0.96 31.88 1.387 41.81 20.65 1.19 36.34 1.387 41.36 20.39 1.22 37.04 1.489 40.47 13.10 1.82 44.61 1.489 39.53 12.85 1.74 45.87 1.489 39.68 13.65 1.75 44.91 1.489 39.86 14.13 1.70 44.30 1.641 35.97 18.84 1.40 43.79 2.327 35.57 17.56 1.11 45.75 2.327 35.09 18.11 1.08 45.72   275  FBMR Temperature (°C) 0.01 m 402 0.32 m 442 0.52 m 457 0.78 m 537 1.08 m 512 1.29 m 523 1.59 m 524 Bed Average 499 2.26 m 457  Time Averaged Conditions P (kPa abs) 498 Pm (kPa abs) 101 Feed CH4 (slpm) 2.596  Feed Water (g/h) 625 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 0.181 Permeate 6 NA  Experiment: Methane 2.a Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 29.79 63.03 0.00 7.18 0.219 30.46 62.29 0.00 7.25 0.219 30.85 61.65 0.00 7.49 0.371 38.36 52.25 0.00 9.39 0.371 38.69 51.81 0.00 9.50 0.473 47.34 40.51 0.00 12.15 0.473 46.81 41.78 0.00 11.41 0.727 57.33 27.42 0.83 14.42 0.727 57.52 27.11 0.84 14.54 0.981 59.88 23.74 1.05 15.34 0.981 58.84 25.87 1.01 14.28 1.235 57.87 27.00 0.84 14.29 1.235 57.49 27.31 0.77 14.43 1.235 56.98 28.18 0.79 14.05 1.235 56.80 28.47 0.75 13.99 1.235 57.16 27.89 0.77 14.18 1.235 56.80 28.25 0.76 14.19 1.235 56.70 28.38 0.75 14.18 1.235 56.88 27.83 0.79 14.50 1.235 56.97 28.03 0.77 14.24 1.387 54.97 29.93 0.72 14.37 1.387 55.38 29.20 0.72 14.70 1.489 59.82 23.87 1.13 15.18 1.489 59.77 23.64 1.14 15.44 1.641 57.77 26.13 0.94 15.16 1.641 57.68 26.43 0.92 14.96 1.641 54.87 29.99 0.73 14.41 2.327 56.25 28.01 0.80 14.95 2.327 56.57 27.61 0.79 15.03 2.327 55.82 28.63 0.83 14.72   276  FBMR Temperature (°C) 0.01 m 407 0.32 m 446 0.52 m 461 0.78 m 541 1.08 m 506 1.29 m 520 1.59 m 528 Bed Average 500 2.26 m 460  Time Averaged Conditions P (kPa abs) 747 Pm (kPa abs) 102 Feed CH4 (slpm) 2.596  Feed Water (g/h) 625 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 0.723 Permeate 6 NA  Experiment: Methane 2.b Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 34.08 57.91 0.00 8.01 0.371 36.98 54.59 0.00 8.43 0.473 43.65 45.86 0.00 10.49 0.727 54.43 31.26 0.68 13.63 0.727 54.53 31.16 0.69 13.61 0.981 57.20 27.86 0.92 14.03 0.981 57.24 27.69 0.90 14.17 1.235 53.24 32.19 0.63 13.95 1.235 53.09 32.68 0.60 13.64 1.235 53.23 32.33 0.60 13.84 1.387 49.45 34.47 0.67 15.40 1.387 48.63 34.67 0.68 16.02 1.387 48.15 35.98 0.70 15.17 1.387 49.87 32.83 0.68 16.62 1.387 50.40 32.51 0.66 16.42 1.387 50.43 32.49 0.66 16.42 1.387 50.22 32.54 0.66 16.58 1.387 49.55 33.29 0.66 16.51 1.387 50.05 32.98 0.64 16.34 1.387 49.84 33.20 0.64 16.31 1.489 55.39 26.90 1.04 16.68 1.489 55.30 26.91 1.03 16.76 1.641 52.63 30.48 0.82 16.07 2.327 49.46 34.30 0.62 15.62 2.327 49.57 34.22 0.59 15.62   277  FBMR Temperature (°C) 0.01 m 407 0.32 m 446 0.52 m 461 0.78 m 541 1.08 m 515 1.29 m 521 1.59 m 525 Bed Average 501 2.26 m 458  Time Averaged Conditions P (kPa abs) 495 Pm (kPa abs) 29 Feed CH4 (slpm) 2.596  Feed Water (g/h) 625 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 1.645 Permeate 6 NA  Experiment: Methane 2.c Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 34.26 57.05 0.00 8.69 0.219 34.24 57.41 0.00 8.35 0.371 39.22 51.69 0.00 9.08 0.371 39.00 51.86 0.00 9.14 0.473 47.73 40.58 0.00 11.70 0.473 47.47 40.88 0.00 11.64 0.727 57.79 26.52 0.87 14.82 0.727 57.69 26.68 0.88 14.75 0.981 60.66 23.22 1.10 15.03 0.981 59.65 24.34 1.06 14.94 1.235 55.51 27.53 0.79 16.17 1.235 55.58 27.00 0.82 16.60 1.235 55.00 27.94 0.81 16.25 1.387 48.89 26.52 1.04 23.56 1.387 48.66 25.85 1.00 24.49 1.387 48.64 25.24 0.98 25.15 1.387 48.75 25.54 0.99 24.72 1.387 48.22 25.05 1.02 25.72 1.387 48.40 24.46 1.00 26.14 1.387 49.25 25.77 0.96 24.03 1.489 56.11 20.79 1.46 21.64 1.489 56.95 21.06 1.39 20.60 1.641 54.23 23.35 1.21 21.21 1.641 54.46 23.71 1.16 20.68 2.327 52.85 25.58 0.95 20.62 2.327 53.02 25.30 0.96 20.72 2.327 52.62 26.15 0.95 20.29   278  FBMR Temperature (°C) 0.01 m 406 0.32 m 449 0.52 m 468 0.78 m 542 1.08 m 507 1.29 m 518 1.59 m 530 Bed Average 502 2.26 m 457  Time Averaged Conditions P (kPa abs) 756 Pm (kPa abs) 29 Feed CH4 (slpm) 2.596  Feed Water (g/h) 625 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 NA Permeate 6 NA  Experiment: Methane 2.d Sample Gas Dry Composition Height (m) H2 CH4 CO CO2 0.219 33.64 58.35 0.00 8.01 0.371 37.75 53.23 0.00 9.02 0.473 43.61 45.91 0.00 10.49 0.727 54.63 31.09 0.73 13.56 0.981 57.62 27.28 0.96 14.15 1.235 51.90 32.76 0.64 14.70 1.235 51.99 32.85 0.63 14.53 1.387 44.72 30.16 0.80 24.33 1.387 44.48 29.96 0.80 24.77 1.387 44.65 29.86 0.78 24.72 1.387 43.77 30.88 0.82 24.53 1.387 44.29 29.97 0.79 24.94 1.387 44.65 30.12 0.82 24.41 1.489 52.22 24.90 1.30 21.58 1.489 52.66 24.50 1.26 21.58 1.641 48.85 27.03 1.14 22.98 1.641 49.23 28.13 1.03 21.62 2.327 47.63 30.34 0.77 21.27 2.327 47.50 30.42 0.75 21.32   279  FBMR Temperature (°C) 0.01 m 418 0.32 m 443 0.52 m 467 0.78 m 483 1.08 m 508 1.29 m 537 1.59 m 554 Bed Average 499 2.26 m 442  Time Averaged Conditions P (kPa abs) 410 Pm (kPa abs) NA Feed C3H8 (g/h) 93.6  Feed Water (g/h) 688.9 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 NA Permeate 6 NA  Experiment: Propane 1.a Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 50.10 26.54 0.37 20.63 2.74 0.219 53.71 21.79 0.25 22.01 2.49 0.371 52.59 23.94 0.37 20.93 1.55 0.371 53.50 23.36 0.38 20.73 1.02 0.473 47.82 32.49 0.39 19.68 0.00 0.473 51.54 26.83 0.27 21.04 0.00 0.473 49.90 29.47 0.37 20.03 0.00 0.473 49.35 30.48 0.39 19.52 0.00 0.473 50.87 28.95 0.39 19.41 0.00 0.727 53.08 25.39 0.27 21.01 0.00 0.727 53.59 26.22 0.29 20.19 0.00 0.981 57.77 20.89 0.37 21.34 0.00 0.981 58.88 19.66 0.29 21.46 0.00 1.235 62.29 16.28 0.50 21.43 0.00 1.235 61.96 16.79 0.37 21.25 0.00 1.489 67.43 11.18 0.38 21.38 0.00 1.489 66.69 11.24 0.28 22.06 0.00 1.641 63.49 14.29 0.29 21.94 0.00 1.641 64.58 13.54 0.80 21.88 0.00 2.327 59.62 18.83 0.50 21.55 0.00 2.327 59.58 18.12 0.37 22.30 0.00 2.327 59.28 18.59 0.28 22.13 0.00 2.327 59.45 18.94 0.45 21.61 0.00   280  FBMR Temperature (°C) 0.01 m 403 0.32 m 439 0.52 m 469 0.78 m 485 1.08 m 507 1.29 m 546 1.59 m 560 Bed Average 501 2.26 m 462  Time Averaged Conditions P (kPa abs) 598 Pm (kPa abs) NA Feed C3H8 (g/h) 93.6  Feed Water (g/h) 688.9 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 NA Permeate 6 NA  Experiment: Propane 1.b Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 51.96 25.94 0.37 22.10 0.00 0.219 48.02 30.18 0.25 21.78 0.02 0.371 47.51 31.16 0.37 21.27 0.06 0.371 48.57 30.74 0.38 20.69 0.00 0.473 45.39 34.66 0.39 19.85 0.00 0.473 45.32 34.51 0.27 20.10 0.00 0.473 45.49 34.64 0.37 19.87 0.00 0.473 45.74 33.94 0.39 20.29 0.00 0.473 48.93 30.99 0.39 20.03 0.00 0.727 48.71 30.36 0.27 20.84 0.00 0.727 49.69 29.35 0.29 20.54 0.00 0.981 54.90 23.32 0.37 21.69 0.00 0.981 55.51 23.03 0.29 21.34 0.00 1.235 58.31 19.88 0.50 21.66 0.00 1.235 59.03 19.11 0.37 21.80 0.00 1.489 63.90 14.46 0.38 21.54 0.00 1.489 64.24 13.95 0.28 21.69 0.00 1.641 61.32 17.25 0.37 21.37 0.00 1.641 61.71 16.31 0.80 21.90 0.00 2.327 55.02 23.49 0.50 21.46 0.00 2.327 55.42 22.96 0.37 21.60 0.00 2.327 55.56 23.11 0.28 21.27 0.00 2.327 55.22 23.24 0.45 21.51 0.00   281  FBMR Temperature (°C) 0.01 m 386 0.32 m 436 0.52 m 464 0.78 m 484 1.08 m 505 1.29 m 548 1.59 m 562 Bed Average 500 2.26 m 467  Time Averaged Conditions P (kPa abs) 703 Pm (kPa abs) NA Feed C3H8 (g/h) 93.6  Feed Water (g/h) 688.9 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 NA Permeate 6 NA  Experiment: Propane 1.c Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 50.75 27.59 0.37 21.41 0.25 0.219 59.72 16.62 0.25 22.58 1.08 0.371 49.40 29.55 0.37 20.82 0.22 0.371 51.59 27.08 0.38 21.18 0.16 0.473 46.67 32.85 0.39 20.37 0.00 0.473 47.04 32.07 0.27 20.72 0.00 0.473 44.05 36.25 0.37 19.44 0.00 0.473 43.97 35.62 0.39 20.16 0.00 0.727 45.12 35.43 0.39 19.46 0.00 0.727 49.21 29.48 0.27 21.22 0.00 0.981 53.38 25.46 0.29 20.89 0.00 0.981 56.80 21.33 0.37 21.63 0.00 1.235 58.13 20.30 0.29 21.30 0.00 1.235 59.04 19.89 0.50 20.94 0.00 1.489 63.75 14.15 0.37 21.84 0.00 1.489 64.81 13.07 0.38 21.89 0.00 1.641 60.71 16.92 0.28 22.16 0.00 1.641 62.55 16.16 0.37 21.20 0.00 2.327 56.43 22.07 0.80 21.41 0.00 2.327 56.30 21.81 0.50 21.80 0.00 2.327 56.95 21.16 0.37 21.67 0.00 2.327 56.56 23.32 0.28 20.12 0.00   282  FBMR Temperature (°C) 0.01 m 410 0.32 m 453 0.52 m 462 0.78 m 490 1.08 m 492 1.29 m 502 1.59 m 504 Bed Average 484 2.26 m 443  Time Averaged Conditions P (kPa abs) 518 Pm (kPa abs) 101 Feed C3H8 (g/h) 118.3  Feed Water (g/h) 725.8 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 0.015 Permeate 6 NA  Experiment: Propane 2.a Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 46.74 27.46 0.14 18.67 6.99 0.219 42.30 31.21 0.12 17.98 8.38 0.219 39.66 33.56 0.12 17.54 9.12 0.219 42.12 30.68 0.14 18.34 8.72 0.371 50.82 13.15 0.22 18.96 3.40 0.371 50.68 12.49 0.24 18.32 2.10 0.371 50.95 12.84 0.21 18.52 2.10 0.625 50.26 19.46 0.49 22.14 0.00 0.625 48.89 21.52 0.46 22.07 0.00 0.625 47.44 22.93 0.49 21.45 0.00 0.727 46.59 31.74 1.04 20.63 0.00 0.727 53.83 23.33 0.91 21.92 0.00 0.727 49.30 28.03 0.97 21.70 0.00 0.981 47.20 32.63 0.84 19.12 0.00 0.981 44.16 36.49 0.86 18.23 0.00 0.981 51.97 26.89 0.80 20.33 0.00 1.133 56.14 22.29 0.82 20.75 0.00 1.133 44.44 35.13 1.03 19.26 0.00 1.133 54.44 23.93 0.88 20.68 0.00 1.235 54.90 23.62 0.93 20.54 0.00 1.235 52.62 26.03 1.01 20.34 0.00 1.235 53.43 24.70 1.01 20.86 0.00 1.235 52.37 25.92 0.99 20.72 0.00 1.235 52.46 26.00 0.99 20.49 0.00 1.235 48.34 30.70 1.06 19.90 0.00 1.235 55.02 22.91 1.05 21.02 0.00 1.235 51.73 26.36 1.11 20.80 0.00 1.387 54.18 23.74 1.11 20.96 0.00 1.387 51.79 26.73 1.30 20.18 0.00 1.387 53.29 24.64 1.28 20.79 0.00 1.387 52.87 25.27 1.22 20.64 0.00 1.489 47.22 33.08 0.99 18.71 0.00 1.489 53.96 24.35 0.97 20.71 0.00 1.489 52.80 26.07 0.93 20.20 0.00 1.641 49.40 28.76 0.84 21.00 0.00 1.641 49.24 29.28 0.89 20.59 0.00 1.641 49.30 28.83 0.87 21.00 0.00 2.327 45.50 35.03 0.77 18.70 0.00 2.327 50.59 28.32 0.73 20.35 0.00 2.327 50.16 27.86 0.80 21.18 0.00   283  FBMR Temperature (°C) 0.01 m 414 0.32 m 453 0.52 m 464 0.78 m 494 1.08 m 496 1.29 m 505 1.59 m 509 Bed Average 487 2.26 m 446  Time Averaged Conditions P (kPa abs) 515 Pm (kPa abs) 29 Feed C3H8 (g/h) 118.3  Feed Water (g/h) 725.8 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 1.377 Permeate 6 NA  Experiment: Propane 2.b Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 41.78 31.76 0.11 19.67 6.68 0.219 41.06 32.50 0.12 20.46 5.85 0.219 39.01 34.41 0.10 19.52 6.96 0.473 38.40 36.61 0.15 21.35 3.49 0.473 39.01 36.34 0.15 22.22 2.28 0.473 37.73 38.10 0.15 21.51 2.51 0.727 52.57 27.52 0.75 18.90 0.26 0.727 46.60 31.53 0.92 20.95 0.00 0.727 48.93 28.03 0.94 22.09 0.00 0.727 46.16 32.24 0.89 20.71 0.00 1.133 53.87 24.96 0.82 20.26 0.09 1.133 56.14 21.50 0.83 21.52 0.00 1.133 51.80 26.13 0.91 21.17 0.00 1.235 52.66 23.59 1.10 22.64 0.00 1.235 52.47 23.62 1.12 22.79 0.00 1.235 51.73 25.13 1.13 22.01 0.00 1.235 52.23 24.50 1.14 22.13 0.00 1.387 49.78 20.32 1.58 28.32 0.00 1.387 49.00 21.16 1.57 28.27 0.00 1.387 46.83 24.46 1.63 27.09 0.00 1.387 51.04 17.05 1.54 30.37 0.00 1.387 53.25 16.18 1.33 29.24 0.00 1.387 52.06 19.97 1.33 26.63 0.00 1.489 49.41 25.20 1.29 24.10 0.00 1.489 49.73 25.44 1.26 23.56 0.00 1.489 52.71 20.47 1.28 25.54 0.00 1.641 46.87 27.83 0.94 24.36 0.00 1.641 47.04 27.12 0.93 24.90 0.00 1.641 46.95 27.29 0.97 24.79 0.00 2.327 53.34 14.88 1.37 30.41 0.00 2.327 52.52 17.24 1.33 28.91 0.01 2.327 51.88 19.21 1.34 27.57 0.00   284  FBMR Temperature (°C) 0.01 m 414 0.32 m 448 0.52 m 459 0.78 m 494 1.08 m 494 1.29 m 505 1.59 m 509 Bed Average 485 2.26 m 456  Time Averaged Conditions P (kPa abs) 706 Pm (kPa abs) 101 Feed C3H8 (g/h) 118.3  Feed Water (g/h) 725.8 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 0.273 Permeate 6 NA  Experiment: Propane 2.c Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 36.35 35.82 0.00 17.17 6.66 0.219 39.83 33.41 0.00 19.14 7.62 0.219 36.34 33.79 0.12 17.73 5.50 0.371 47.20 18.67 0.21 19.72 2.10 0.371 48.54 15.42 0.21 17.85 1.00 0.371 45.35 17.90 0.21 17.38 0.56 0.625 46.60 26.77 0.29 21.27 0.00 0.625 44.57 28.01 0.29 20.33 0.00 0.625 44.30 28.89 0.27 21.19 0.00 0.727 46.68 32.35 0.70 20.24 0.00 0.727 49.42 29.44 0.68 20.41 0.00 0.727 48.77 29.04 0.73 21.41 0.00 0.981 45.47 34.09 0.86 19.51 0.00 0.981 49.36 30.37 0.80 19.38 0.00 0.981 44.60 35.37 0.81 19.10 0.00 1.133 48.86 30.20 0.76 20.16 0.00 1.133 49.21 27.77 0.82 22.18 0.01 1.133 47.95 32.05 0.81 19.18 0.00 1.235 54.02 25.01 0.84 20.11 0.00 1.235 50.15 28.52 0.86 20.42 0.00 1.235 47.91 31.27 0.93 19.84 0.00 1.235 51.46 26.74 0.94 20.84 0.00 1.235 52.17 26.64 0.92 20.26 0.00 1.235 48.11 30.73 0.98 20.17 0.00 1.235 52.72 25.57 0.91 20.75 0.00 1.235 50.12 28.28 0.91 20.63 0.00 1.235 49.75 30.04 0.89 19.30 0.00 1.387 53.02 24.15 1.00 21.71 0.00 1.387 50.45 28.07 1.01 20.38 0.00 1.387 49.08 27.09 1.12 22.62 0.00 1.489 46.98 32.11 0.97 19.88 0.00 1.489 51.30 27.04 0.93 20.67 0.00 1.489 49.17 28.34 0.96 21.37 0.00 1.641 46.21 31.05 0.75 21.96 0.00 1.641 45.58 31.99 0.75 21.64 0.00 1.641 47.48 31.21 0.69 20.59 0.00 2.327 47.45 29.79 0.62 22.01 0.00 2.327 45.34 33.06 0.61 20.98 0.00 2.327 45.96 32.21 0.58 21.23 0.00   285  FBMR Temperature (°C) 0.01 m 415 0.32 m 449 0.52 m 460 0.78 m 496 1.08 m 496 1.29 m 505 1.59 m 510 Bed Average 486 2.26 m 455  Time Averaged Conditions P (kPa abs) 698 Pm (kPa abs) 28 Feed C3H8 (g/h) 118.3  Feed Water (g/h) 725.8 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 NA Permeate 2 NA Permeate 3 NA Permeate 4 NA Permeate 5 1.600 Permeate 6 NA  Experiment: Propane 2.d Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 40.60 32.55 0.00 18.75 8.09 0.219 39.18 33.88 0.00 19.17 7.76 0.219 38.36 34.59 0.00 18.99 8.06 0.473 36.50 40.05 0.12 20.40 0.05 0.473 37.62 39.05 0.00 20.89 0.02 0.473 36.58 42.40 0.00 21.01 0.01 0.727 51.30 26.74 0.70 21.24 0.00 0.727 50.90 27.30 0.70 20.98 0.00 0.727 51.28 26.51 0.71 21.34 0.00 1.133 49.60 28.75 0.77 20.86 0.00 1.133 48.30 30.74 0.74 20.22 0.00 1.133 49.44 29.31 0.77 20.49 0.00 1.235 49.26 28.29 0.83 21.43 0.18 1.235 49.76 27.22 0.91 22.03 0.00 1.235 48.46 29.67 0.91 20.87 0.00 1.387 48.81 22.48 1.26 27.40 0.00 1.387 43.74 26.89 1.31 27.98 0.00 1.387 46.10 25.15 1.25 27.37 0.00 1.387 43.73 26.20 1.33 28.64 0.00 1.387 49.69 18.51 1.23 30.51 0.00 1.387 50.50 18.00 1.27 30.11 0.00 1.489 48.67 26.13 0.98 24.13 0.00 1.489 48.21 26.54 1.04 24.11 0.00 1.489 46.30 29.56 1.04 23.00 0.00 1.641 47.88 24.47 0.83 26.73 0.00 1.641 46.56 25.40 0.85 27.17 0.00 1.641 46.76 26.56 0.83 25.80 0.00 2.327 43.35 28.26 0.79 27.59 0.01 2.327 42.49 30.71 0.79 26.00 0.01 2.327 41.99 30.86 0.75 26.39 0.01  286  FBMR Temperature (°C) 0.01 m 438 0.32 m 483 0.52 m 472 0.78 m 515 1.08 m 494 1.29 m 506 1.59 m 531 Bed Average 500 2.26 m 480  Time Averaged Conditions P (kPa abs) 610 Pm (kPa abs) 101 Feed C3H8 (g/h) 101.4  Feed Water (g/h) 622.1 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 0.036 Permeate 2 0.026 Permeate 3 0.088 Permeate 4 0.019 Permeate 5 0.110 Permeate 6 0.237  Experiment: Propane 3.a Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 42.19 37.42 0.42 19.93 0.04 0.219 39.05 42.58 0.35 18.02 0.00 0.219 42.87 36.32 0.36 20.43 0.01 0.371 50.14 28.80 0.58 20.47 0.00 0.371 44.42 35.84 0.63 19.10 0.00 0.371 50.18 27.84 0.65 21.32 0.00 0.473 36.29 43.94 0.30 19.38 0.07 0.473 43.14 36.65 0.29 19.84 0.02 0.625 42.24 37.35 0.42 19.94 0.05 0.625 39.20 42.45 0.42 17.92 0.01 0.625 42.13 37.42 0.44 19.97 0.04 0.727 46.91 31.79 0.76 20.53 0.02 0.727 45.33 34.34 0.79 19.53 0.02 0.727 45.40 33.85 0.80 19.92 0.03 0.879 43.99 34.53 0.68 20.76 0.05 0.879 45.02 33.25 0.68 20.98 0.06 0.981 43.99 35.71 0.64 19.64 0.01 0.981 45.71 33.00 0.64 20.63 0.01 0.981 43.18 36.94 0.64 19.23 0.01 1.133 47.39 31.05 0.76 20.80 0.00 1.133 47.21 31.35 0.74 20.70 0.00 1.133 46.31 32.15 0.75 20.79 0.00 1.133 46.96 31.64 0.75 20.65 0.00 1.235 47.21 30.65 0.71 21.40 0.03 1.235 47.23 31.17 0.69 20.89 0.02 1.235 46.61 31.66 0.71 21.01 0.01 1.387 45.70 33.28 0.79 20.21 0.02 1.387 47.42 30.50 0.79 21.27 0.01 1.489 49.99 27.40 1.10 21.46 0.05 1.489 50.17 26.94 1.11 21.71 0.08 1.489 50.51 27.05 1.08 21.36 0.00 1.489 50.55 27.23 1.12 21.02 0.07 1.489 50.62 26.69 1.10 21.55 0.05 1.489 49.43 27.42 1.17 21.94 0.04 1.489 49.75 26.56 1.19 22.44 0.06 1.641 49.92 27.24 1.15 21.68 0.00 1.641 48.86 28.34 1.18 21.62 0.00 2.327 49.08 27.49 1.12 22.32 0.00 2.327 49.90 26.37 1.13 22.59 0.00 2.327 49.25 26.68 1.13 22.94 0.00   287  FBMR Temperature (°C) 0.01 m 443 0.32 m 475 0.52 m 464 0.78 m 520 1.08 m 502 1.29 m 513 1.59 m 539 Bed Average 502 2.26 m 484  Time Averaged Conditions P (kPa abs) 609 Pm (kPa abs) 48 Feed C3H8 (g/h) 101.4  Feed Water (g/h) 622.1 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 0.993 Permeate 2 0.942 Permeate 3 0.459 Permeate 4 0.238 Permeate 5 0.585 Permeate 6 0.928  Experiment: Propane 3.b Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 38.98 39.01 0.36 21.66 0.00 0.219 40.30 37.29 0.34 22.05 0.01 0.219 40.12 38.11 0.35 21.41 0.01 0.219 41.33 35.29 0.36 23.01 0.02 0.371 47.82 24.68 0.65 25.55 0.05 0.371 40.32 34.10 0.66 23.42 0.01 0.371 49.87 22.13 0.61 25.63 0.04 0.473 33.25 41.99 0.31 24.00 0.06 0.473 37.04 39.03 0.32 23.60 0.01 0.473 38.66 36.98 0.38 23.98 0.00 0.625 39.89 32.80 0.51 26.78 0.01 0.625 39.21 32.49 0.56 27.71 0.02 0.625 35.79 38.62 0.57 25.01 0.01 0.727 42.49 23.83 1.09 32.54 0.05 0.727 39.04 28.53 1.12 31.30 0.01 0.879 38.56 23.21 1.04 37.17 0.02 0.879 37.54 23.13 1.11 38.17 0.05 0.981 39.86 24.24 1.07 34.83 0.00 0.981 40.12 22.09 1.07 36.72 0.01 1.133 37.78 18.51 1.37 42.34 0.00 1.133 37.44 16.74 1.42 44.40 0.00 1.235 40.89 24.47 1.18 33.44 0.01 1.235 41.39 22.15 1.17 35.29 0.01 1.387 39.33 21.67 1.36 37.64 0.01 1.387 39.59 20.92 1.28 38.21 0.00 1.489 41.66 15.29 1.78 41.26 0.01 1.489 40.24 16.80 1.83 41.11 0.02 1.489 41.04 15.83 1.82 41.27 0.03 1.489 40.43 16.95 1.70 40.88 0.03 1.489 41.20 16.48 1.70 40.58 0.04 1.641 39.53 20.32 1.58 38.57 0.00 1.641 38.57 21.26 1.63 38.55 0.00 2.327 42.03 17.33 1.66 38.98 0.00 2.327 41.01 18.13 1.65 39.20 0.00 2.327 41.50 17.56 1.65 39.28 0.00   288  FBMR Temperature (°C) 0.01 m 443 0.32 m 469 0.52 m 451 0.78 m 517 1.08 m 499 1.29 m 510 1.59 m 540 Bed Average 498 2.26 m 486  Time Averaged Conditions P (kPa abs) 605 Pm (kPa abs) 25 Feed C3H8 (g/h) 101.4  Feed Water (g/h) 622.1 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 1.511 Permeate 2 1.280 Permeate 3 0.633 Permeate 4 0.321 Permeate 5 0.763 Permeate 6 1.108  Experiment: Propane 3.c Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 36.21 39.69 0.33 23.75 0.02 0.219 35.95 40.47 0.35 23.22 0.01 0.219 39.70 36.43 0.32 23.54 0.01 0.371 40.19 29.18 0.45 27.32 0.05 0.371 39.70 29.56 0.48 27.43 0.04 0.371 36.31 34.36 0.43 26.17 0.01 0.473 32.89 37.30 0.30 29.48 0.03 0.473 32.45 38.58 0.31 28.65 0.02 0.473 33.99 36.07 0.32 29.60 0.01 0.625 37.04 30.92 0.49 31.48 0.07 0.625 33.78 35.47 0.50 30.23 0.03 0.625 33.96 36.01 0.48 29.53 0.02 0.727 36.68 23.53 1.19 38.54 0.01 0.727 37.01 21.03 1.22 40.64 0.02 0.879 34.57 15.50 1.13 48.75 0.05 0.879 32.77 18.05 1.19 47.98 0.02 0.981 35.43 18.76 1.12 44.67 0.02 0.981 34.02 21.08 1.19 43.71 0.00 1.133 31.07 12.30 1.33 55.30 0.00 1.133 30.96 12.80 1.41 54.83 0.00 1.235 37.43 19.76 1.25 41.54 0.02 1.235 37.20 20.40 1.23 41.10 0.03 1.387 35.99 16.06 1.33 46.59 0.03 1.387 34.90 16.56 1.36 47.15 0.02 1.489 36.33 11.63 1.90 50.13 0.01 1.489 35.97 10.88 1.86 51.28 0.01 1.489 35.39 11.23 1.81 51.55 0.01 1.489 36.22 13.96 1.76 48.02 0.04 1.489 36.39 13.36 1.90 48.33 0.02 1.641 31.52 16.05 1.51 50.92 0.00 1.641 31.50 17.33 1.58 49.59 0.00 2.327 36.54 12.00 1.73 49.72 0.00 2.327 36.77 12.60 1.78 48.84 0.00 2.327 35.94 13.14 1.82 49.11 0.00   289  FBMR Temperature (°C) 0.01 m 402 0.32 m 480 0.52 m 460 0.78 m 515 1.08 m 490 1.29 m 499 1.59 m 544 Bed Average 498 2.26 m 458  Time Averaged Conditions P (kPa abs) 406 Pm (kPa abs) 25 Feed C3H8 (g/h) 67.6  Feed Water (g/h) 414.8 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 1.034 Permeate 2 1.097 Permeate 3 0.435 Permeate 4 0.266 Permeate 5 0.488 Permeate 6 0.689  Experiment: Propane 4.a Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 40.91 22.94 0.69 35.36 0.09 0.219 38.99 25.85 0.69 34.45 0.02 0.219 39.08 25.19 0.70 35.04 0.00 0.371 41.14 20.60 0.83 36.31 0.22 0.371 36.38 17.62 0.85 44.85 0.08 0.371 36.69 18.32 0.86 43.75 0.05 0.473 41.48 31.27 0.38 26.53 0.05 0.473 40.51 32.41 0.39 26.28 0.12 0.473 41.74 30.59 0.38 26.74 0.03 0.625 38.83 26.90 0.64 33.48 0.00 0.625 38.85 27.35 0.66 33.05 0.00 0.727 40.62 16.06 1.19 42.13 0.00 0.727 40.15 16.87 1.27 41.71 0.00 0.727 39.92 15.49 1.25 43.34 0.00 0.879 34.71 11.04 1.21 53.05 0.00 0.879 34.30 11.67 1.25 52.78 0.00 0.981 37.29 17.94 1.05 43.72 0.00 0.981 37.95 18.33 1.03 42.70 0.00 0.981 38.11 17.30 1.00 43.59 0.00 1.133 34.90 12.14 1.28 51.68 0.00 1.133 35.15 11.36 1.27 52.16 0.06 1.235 38.85 15.64 1.23 44.27 0.00 1.235 39.83 15.20 1.15 43.80 0.03 1.235 39.15 15.02 1.18 44.65 0.00 1.387 38.23 13.37 1.39 47.01 0.00 1.387 38.74 13.81 1.35 46.10 0.00 1.489 37.51 9.69 1.84 50.97 0.00 1.489 37.22 9.33 1.88 51.54 0.02 1.489 37.34 8.68 1.81 52.17 0.00 1.489 37.91 9.13 1.83 51.13 0.00 1.489 37.97 9.12 1.85 51.06 0.00 1.489 37.40 8.88 1.84 51.88 0.00 1.641 33.79 12.67 1.77 51.77 0.00 1.641 34.52 13.56 1.76 50.16 0.00 2.327 37.39 9.00 1.78 51.83 0.00 2.327 38.11 8.84 1.77 51.28 0.00 2.327 38.91 8.48 1.73 50.88 0.00 2.327 38.23 8.25 1.72 51.80 0.00   290  FBMR Temperature (°C) 0.01 m 413 0.32 m 472 0.52 m 460 0.78 m 509 1.08 m 503 1.29 m 509 1.59 m 544 Bed Average 500 2.26 m 491  Time Averaged Conditions P (kPa abs) 802 Pm (kPa abs) 25 Feed C3H8 (g/h) 135.2  Feed Water (g/h) 829.5 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 1.668 Permeate 2 1.661 Permeate 3 0.793 Permeate 4 0.625 Permeate 5 0.977 Permeate 6 1.519  Experiment: Propane 4.b Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 30.39 36.13 0.47 31.60 1.41 0.219 31.22 35.06 0.48 32.30 0.94 0.371 31.93 30.05 0.58 36.28 0.95 0.371 34.11 27.84 0.62 37.25 0.17 0.473 30.65 34.63 0.45 31.62 0.12 0.473 29.74 38.31 0.36 29.34 0.03 0.473 31.52 36.25 0.37 30.48 0.00 0.625 34.69 32.93 0.46 30.49 0.00 0.625 34.23 33.95 0.46 30.08 0.00 0.727 34.55 30.05 0.96 33.16 0.30 0.727 36.55 26.00 0.94 36.22 0.00 0.879 30.96 19.66 1.00 46.36 0.00 0.879 30.07 21.45 1.07 47.21 0.00 0.981 35.88 25.54 0.95 35.77 0.00 0.981 36.14 25.05 0.96 36.23 0.00 1.133 30.99 18.90 1.30 47.63 0.00 1.133 31.26 17.87 1.31 48.61 0.95 1.235 35.63 24.41 1.08 38.65 0.23 1.235 35.14 22.51 1.11 40.25 1.00 1.387 34.41 20.00 1.21 43.31 0.00 1.387 33.13 21.03 1.27 43.68 0.88 1.489 32.62 18.72 1.38 45.83 0.00 1.489 32.84 18.20 1.42 46.09 0.00 1.489 32.40 17.88 1.39 46.93 0.00 1.489 32.71 17.84 1.53 46.62 0.00 1.489 33.14 17.39 1.48 46.86 1.13 1.641 30.99 20.70 1.52 45.82 0.00 1.641 31.37 21.35 1.53 45.46 0.28 2.327 35.99 17.53 1.62 44.81 0.05 2.327 36.40 16.44 1.64 45.32 0.20 2.327 35.27 16.62 1.66 46.10 0.05   291  FBMR Temperature (°C) 0.01 m 419 0.32 m 473 0.52 m 457 0.78 m 510 1.08 m 501 1.29 m 509 1.59 m 544 Bed Average 499 2.26 m 490  Time Averaged Conditions P (kPa abs) 603 Pm (kPa abs) 25 Feed C3H8 (g/h) 124.78  Feed Water (g/h) 612.58 Feed H2 (slpm) 0   Permeate H2 (slpm) Permeate 1 1.522 Permeate 2 1.415 Permeate 3 0.649 Permeate 4 0.481 Permeate 5 0.828 Permeate 6 1.314  Experiment: Propane 5.a Sample Gas Dry Composition (%) Stream Location (m) H2 CH4 CO CO2 C3H8 0.219 28.56 36.52 0.52 29.80 0.28 0.219 29.73 32.13 0.54 31.89 0.24 0.219 26.45 39.31 0.57 29.39 0.14 0.371 35.32 30.70 0.67 25.51 0.13 0.371 42.03 24.77 0.64 25.36 0.14 0.371 37.55 25.32 0.58 27.89 0.01 0.371 35.38 28.86 0.57 27.51 0.05 0.473 28.58 45.87 0.32 23.94 0.00 0.473 35.68 37.30 0.30 24.66 0.00 0.473 29.60 45.29 0.34 23.40 0.00 0.473 32.25 41.43 0.33 24.35 0.00 0.625 31.55 35.73 0.50 28.34 0.00 0.625 33.56 33.68 0.48 28.00 0.00 0.625 30.03 38.55 0.52 27.87 0.00 0.727 30.03 30.04 1.10 32.79 0.00 0.727 33.13 27.12 1.01 32.58 0.00 0.727 31.23 28.91 1.07 32.41 0.00 0.879 26.01 23.57 1.11 39.43 0.00 0.879 28.94 19.90 1.05 40.14 0.00 0.879 26.80 25.45 1.14 37.57 0.00 0.981 29.43 27.99 1.07 34.72 0.00 0.981 29.37 28.99 1.05 34.32 0.00 1.133 25.79 21.25 1.40 41.82 0.00 1.133 28.25 16.74 1.26 42.31 0.00 1.235 30.54 24.49 1.28 35.48 0.00 1.235 31.93 22.59 1.21 36.23 0.00 1.387 28.49 22.60 1.40 38.29 0.00 1.387 30.11 18.62 1.32 39.29 0.00 1.387 28.89 21.99 1.33 38.29 0.00 1.489 28.09 17.14 1.70 42.17 0.00 1.489 28.20 16.17 1.70 42.96 0.00 1.489 28.31 15.51 1.65 43.21 0.19 1.489 28.37 16.36 1.67 43.08 0.17 1.489 28.43 18.03 1.68 41.92 0.00 1.489 28.77 15.95 1.77 42.47 0.00 1.489 28.98 14.22 1.69 44.18 0.10 1.641 27.50 19.19 1.69 41.56 0.00 1.641 27.40 21.01 1.77 39.98 0.00 1.641 27.54 17.70 1.72 42