UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of a Swiss-roll mixed-reactant fuel cell Aziznia, Amin 2013

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

Item Metadata

Download

Media
24-ubc_2014_spring_aziznia_amin.pdf [ 11.56MB ]
Metadata
JSON: 24-1.0166838.json
JSON-LD: 24-1.0166838-ld.json
RDF/XML (Pretty): 24-1.0166838-rdf.xml
RDF/JSON: 24-1.0166838-rdf.json
Turtle: 24-1.0166838-turtle.txt
N-Triples: 24-1.0166838-rdf-ntriples.txt
Original Record: 24-1.0166838-source.json
Full Text
24-1.0166838-fulltext.txt
Citation
24-1.0166838.ris

Full Text

DEVELOPMENT OF A SWISS-ROLL MIXED-REACTANT FUEL CELL  by Amin Aziznia  M.A.Sc., Sharif University of Technology, 2008 B.A.Sc., Iran University of Science and Technology, 2006   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTROAL STUDIES (Chemical and Biological Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2013  ? Amin Aziznia, 2013  ii  Abstract Capital and operating costs of fuel cell systems must be reduced before they can be competitive with conventional energy conversion technologies. This dissertation concerns the development of an unconventional fuel cell aimed at meeting that challenge. Presented here, for the first time, is a novel cylindrical Swiss-roll mixed-reactant fuel cell (SR-MRFC) that eliminates expensive and failure-prone components of conventional fuel cells. The proof-of-concept of the SR-MRFC was performed both in monopolar and bipolar architectures. In the monopolar case 3D anodes with platinum or with osmium catalysts were coupled to a gas-diffusion MnO2 cathode in a 20?10-4 m2 single-cell SR-MRFC, operated with a two-phase mixture of 1 M NaBH4/2M NaOH(aq) + O2(g). Instead of a Nafion? membrane, a porous diaphragm was employed. At 323 K, 105 kPa(abs), the peak superficial power densities of the SR-MRFC with the platinum and osmium anode catalysts were up to respectively 2230 and 1880 W m?2 with good performance stability during 3 hr continuous operation. These values are the highest power densities ever reported for MRFCs operating under similar conditions and match the highest reported values for conventional dual chamber PEM direct borohydride fuel cells.  Scale up of the single-cell SR-MRFC to 100?10-4 m2 and 200?10-4 m2 gave corresponding peak superficial power densities of 900 and 700 W m-2, while the 20?10-4 m2 bipolar reactors produced peak volumetric power densities of 267 and 205 kW m-3.  iii  This work also explored the feasibility of electroreduction of N2O on Pt and Pd in the cathode of a MRFC to generate electricity from N2O in the tail gases of industrial processes. Here the SR-MRFC was operated using two-phase fuel + oxidant mixtures of 1 M NaBH4 / 2M NaOH(aq) + N2O(g) and 0.5 M CH3OH/2 M NaOH(aq) + N2O(g). At 323 K, 105 kPa(abs) the peak superficial power densities for the mixed NaBH4- and MeOH-N2O systems were respectively 340 W m-2 (Pt anode/Pd cathode) and 38 W m-2 (PtRu anode/Pd cathode). This work demonstrates for the first time that co-generation of electricity and abatement of N2O may potentially compete with thermochemical processes of N2O capture currently under development.  iv  Preface The following papers, presentations and patent application resulted from the work presented in this dissertation:  A version of Chapters 2 and 3 is published in:  1. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?Platinum- and membrane-free Swiss-roll mixed-reactant alkaline fuel cell?, ChemSusChem 6 (2013) 847-855. 2. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?A Swiss-roll liquid-gas mixed-reactant fuel cell?, Journal of Power Sources 212 (2012) 154-160. 3. A.Aziznia, E.L.Gyenge, C.W.Oloman, ?A Swiss-roll mixed-reactant fuel cell?, Hydrogen & Fuel Cells 2013 (HFC2013) Vancouver, BC, Canada. June 16-19th 2013. 4. A.Aziznia, E.L.Gyenge, C.W.Oloman, ?Development of a Swiss-roll mixed-reactant direct borohydride fuel cell ?, 222th Meeting of the Electrochemical Society (ECS), Honolulu, HI, USA, October 7-12th, 2012. 5. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?An osmium electrodeposited anode for the Swiss-roll mixed-reactant fuel cell?, The 62nd Canadian Chemical Engineering Conference, Vancouver, BC, Canada, October 14-17th, 2012. 6. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?Development of a novel Swiss-roll mixed-reactant fuel cell: Application for direct borohydride fuel cells?, the 61st  v  Canadian Chemical Engineering Conference, London, Ontario, Canada, October 23-26th, 2011. A version of Chapters 3 and 4 was submitted as:  7. A.Aziznia, C.W.Oloman, E.L.Gyenge ?Experimental advances and mathematical modeling of the Swiss-roll mixed-reactant direct borohydride fuel cell?, submitted (2013). A version of Chapters 2 and 6 is in preparation as:  8. A.Aziznia, C.W.Oloman, E.L.Gyenge ?Scale-up and bipolar operation of the Swiss-roll mixed-reactant direct borohydride fuel cell?, to be submitted (2013). For the items 1 to 8 above, all of the experiments were performed by Amin Aziznia and manuscripts were co-authored with El?d L. Gyenge and Colin W. Oloman. A version of Chapter 5 is published in:  9. A.Aziznia, A.Bonakdarpour, E.L.Gyenge, C.W.Oloman, ?Electroreduction of nitrous oxide on platinum and palladium: Toward selective catalysts for methanol-nitrous oxide mixed-reactant fuel cells?, Electrochimica Acta 56 (2011) 5238?5244.  10. A.Aziznia, E.L.Gyenge, C.W.Oloman, ?Kinetic analysis of electroreduction of nitrous oxide over polycrystalline Pd and Pt catalysts?, 218th Meeting of the Electrochemical Society (ECS), Las Vegas, NV, USA, October 10-15th, 2010.  vi  11. A.Aziznia, E.L.Gyenge, C.W.Oloman, ?Electrochemical abatement of nitrous oxide (N2O) in a Swiss-roll mixed-reactant fuel cell?, to be submitted (2013). For the items 9 and 10 all of the experiments were performed by Amin Aziznia with laboratory recommendations of Arman Bonakdarpour. The manuscripts were co-authored with Arman Bonakdaropour, El?d L. Gyenge and Colin W. Oloman. For the item 11, all of the experiments were performed by Amin Aziznia and manuscripts were co-authored with El?d L. Gyenge and Colin W. Oloman. Part of Chapter 2 of this dissertation filed as a United States Provisional Patent Application: 12. A.Aziznia, C.W.Oloman, E.L.Gyenge ?Apparatus and method for feeding an electrochemical reactor?, U.S. Provisional Application, US 61/869,053 (2013). Where the intellectual property is co-developed by Amin Aziznia, El?d L. Gyenge and Colin W. Oloman. All of the experiments for this intellectual property application were performed by Amin Aziznia. The application was co-authored by Colin W. Oloman, El?d L. Gyenge, and Larry Kyle (Patent Attorney). Other contributions during Mr. Aziznia?s tenure as a Ph.D. student at the University of British Columbia included a collaboration with the Center for Emerging Energy Technologies of the University of New Mexico on development of platinum-free electrocatalysts for the Swiss-roll fuel cell presented in this dissertation and which resulted in:  vii  13. A.Serov, A.Aziznia, P.H.Benhangi, K.Artyushkova, P.Atanassov, E.L.Gyenge, ?A novel borohydride-tolerant oxygen electroreduction catalyst for mixed-reactant Swiss-roll direct borohydride fuel cells?, Journal of Material Chemistry A 1 (2013), 14384-14391. 14. A.Serov, A.Aziznia, P.H.Benhangi, K.Artyushkova, P.Atanassov, C.W. Oloman, E.L.Gyenge, ?A highly selective platinum-free oxygen cathode for the Swiss-roll mixed-reactant direct borohydride fuel cell?, 224th Meeting of the Electrochemical Society (ECS), San Francisco, CA, USA, October 27th-November 1st, 2013. Moreover, Mr. Aziznia was invited to present the commercialization opportunity of the work presented in this dissertation at: 15. A. Aziznia, S. Kjellander ?Commercialization opportunities of the Swiss-roll mixed-reactant fuel cell?, UBC Sauder School of Business Entrepreneurship Luncheon, the Fairmont Waterfront Hotel, Vancouver, BC, Canada, May 13th, 2013.  viii  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ....................................................................................................................... viii List of Tables ................................................................................................................................ xi List of Figures .............................................................................................................................. xii List of Symbols ........................................................................................................................... xiv List of Abbreviations ................................................................................................................. xix Glossary ...................................................................................................................................... xxi Acknowledgements ................................................................................................................... xxii Dedication ................................................................................................................................. xxiv Chapter 1: Introduction ................................................................................................................1 1.1 Background ................................................................................................................ 1 1.2 Mixed-Reactant Fuel Cell (MRFC) Concept ............................................................. 2 1.3 Swiss-Roll Electrochemical Reactors ........................................................................ 3 1.4 Principles of Fuel Cell Design ................................................................................... 5 1.4.1 Material and Energy Balance ............................................................................. 6 1.4.2 Voltage Balance ................................................................................................. 7 1.4.3 Electrochemical Thermodynamics ..................................................................... 8 1.4.4 Reaction Rate and Electrode Kinetics .............................................................. 10 1.4.5 Ohmic Losses ................................................................................................... 15 1.4.6 Mass Transport ................................................................................................. 17 1.4.7 Three-Dimensional (3D) Electrodes ................................................................ 18 1.4.8 Mixed-Potential Theory ................................................................................... 21 1.4.9 Multiphase Liquid/Gas Electrochemical Systems ............................................ 22 1.4.10 Shunt Current ................................................................................................... 24 1.4.11 Figures of Merit ................................................................................................ 25 1.5 Research Approach .................................................................................................. 27 1.5.1 Objective .......................................................................................................... 27 1.5.2 Fuel Cell System, Reactant and Electrocatalyst Selection ............................... 28 1.5.2.1 Fuel Cell Type........................................................................................... 28 1.5.2.2 Fuel ........................................................................................................... 29 1.5.2.3 Oxidant ...................................................................................................... 31 1.5.2.4 Electrocatalysts ......................................................................................... 32 1.5.3 Thesis Layout ................................................................................................... 33 1.5.4 Literature Review  ............................................................................................ 36 1.6 Summary of Literature Review ................................................................................ 58 Chapter 2: Swiss-Roll Fuel Cell Design and Components .......................................................60 2.1 Introduction .............................................................................................................. 60 2.2 The Monopolar Swiss-Roll ...................................................................................... 60 2.2.1 Three-Dimensional (3D) Anode ....................................................................... 61  ix  2.2.1.1 Carbon Cloth Substrate ............................................................................. 61 2.2.1.2 Pt and PtRu Anode Electrocatalysts ......................................................... 63 2.2.1.3 Electrodeposited Os .................................................................................. 64 2.2.2 Gas Diffusion Cathode ..................................................................................... 65 2.2.3 Separator ........................................................................................................... 67 2.2.4 Current Collector/Fluid Distributor .................................................................. 68 2.2.5 Two-Phase Flow Feeding System .................................................................... 70 2.2.6 Swiss-Roll Fuel Cell ......................................................................................... 71 2.2.6.1 Reactor A: Small- and Medium-Scale ...................................................... 71 2.2.6.2 Reactor B: Large-Scale ............................................................................. 73 2.3 Bipolar Swiss-Roll Design and Components ........................................................... 74 2.3.1 Arrangement 1: Multi-Layer Roll .................................................................... 75 2.3.2 Arrangement 2: Rolls-in-Series ........................................................................ 76 2.4 Fuel Cell Test Apparatus ......................................................................................... 77 Chapter 3: Single-Cell Swiss-Roll Reactor for DBFCs ............................................................80 3.1 Introduction .............................................................................................................. 80 3.2 Anode Electrocatalyst .............................................................................................. 81 3.2.1 Pt and PtRu ....................................................................................................... 81 3.2.2 Electrodeposited Os .......................................................................................... 83 3.2.3 Borohydride Hydrolysis Activity of Pt, Os and OsO2 ...................................... 95 3.3 Cathode Electrocatalyst ........................................................................................... 98 3.4 Current Collector/Fluid Distributor ....................................................................... 101 3.5 Separator: Comparison of Nafion? and Viledon? Porous Diaphragm .................. 104 3.6 Effect of Nozzle Sprayer Feeding .......................................................................... 106 3.7 Swiss-Roll Performance Comparison with Conventional DBFCs ........................ 108 3.8 Summary ................................................................................................................ 109 Chapter 4: Mathematical Modeling of the Swiss-Roll Single-Cell ........................................112 4.1 Introduction ............................................................................................................ 112 4.2 Equilibrium Electrode Potentials ........................................................................... 113 4.3 Activation Overpotentials ...................................................................................... 115 4.4 Mixed-Potentials .................................................................................................... 116 4.5 Mass-Transfer Limited Current Density ................................................................ 116 4.6 Ohmic Losses ......................................................................................................... 119 4.7 Modeling Approach ............................................................................................... 121 4.8 Model Predictions .................................................................................................. 124 4.9 Summary ................................................................................................................ 127 Chapter 5: Single-Cell Swiss-Roll with N2O Oxidant .............................................................129 5.1 Introduction ............................................................................................................ 129 5.2 Nitrous Oxide Reduction on Pt .............................................................................. 131 5.3 Effect of Methanol Oxidation on N2O Reduction on Pt ........................................ 138 5.4 Nitrous Oxide Reduction on Pd ............................................................................. 140 5.5 Effect of Methanol on N2O Reduction on Pd ........................................................ 145 5.6 The Swiss-Roll Mixed Methanol-N2O Fuel Cell ................................................... 147  x  5.7 The Swiss-Roll Mixed Borohydride-N2O Fuel Cell .............................................. 151 5.8 Summary ................................................................................................................ 153 Chapter 6: Scale-Up and Bipolar Operation of the Swiss-Roll Cell .....................................155 6.1 Introduction ............................................................................................................ 155 6.2 Single-Cell Scale-Up from 20?10-4 m2 to 200?10-4 m2 ......................................... 155 6.3 Bipolar Operation................................................................................................... 161 6.3.1 Arrangement 1: Multi-Layer Roll .................................................................. 161 6.3.2 Arrangement 2: Rolls-in-Series ...................................................................... 163 6.4 Summary ................................................................................................................ 166 Chapter 7: Conclusions and Recommendations for Future Work .......................................168 7.1 The Swiss-Roll Mixed-Reactant Architecture ....................................................... 168 7.2 Proof-of-Concept of the Swiss-Roll for Direct Borohydride MRFCs ................... 168 7.3 Scale-Up and Bipolar Operation of the SR-MRFC ............................................... 171 7.4 Nitrous Oxide as an Oxidant for Mixed-Reactant Fuel Cells ................................ 172 7.5 Recommendations .................................................................................................. 174 7.6 Contribution to Knowledge.................................................................................... 178 7.6.1 An Innovative Mixed-Reactant Fuel Cell Design and Feeding Device ......... 178 7.6.2 Electrocatalysis Study and Electrode Modification ....................................... 179 7.6.3 Electroreduction of N2O ................................................................................. 179 Bibliography ...............................................................................................................................181 Appendices ..................................................................................................................................193 Appendix A Table of Potential Fuels and Oxidants for the MRFC ........................................ 194 Appendix B Experimental Procedure ..................................................................................... 196 B.1 Activity of MnO2 toward the BOR and ORR ................................................. 196 B.2 Activity and Selectivity of Pt and Pd toward N2O Reduction ........................ 198 B.3 Activation of carbon Cloth Substrate ............................................................. 200 B.4 Preparation of Pt and PtRu Anode ................................................................. 200 B.5 Preparation of the Electrodeposited Os 3D Anode ........................................ 200 B.6 Electrode Characterization ............................................................................. 202 B.7 NaBH4 Hydrolysis Rate Measurements ......................................................... 203 B.8 Pre-Treatment and Conditioning of Nafion? Membrane ................................ 204 B.9 Electrochemical Impedance Spectroscopy (EIS) Measurements ................... 204 Appendix C Dimensioned Drawings of the Swiss-Roll Cell and Components ...................... 205 C.1 Reactor A ........................................................................................................ 205 C.2 Reactor B ........................................................................................................ 207 C.3 Multi-Layer Bipolar Reactor .......................................................................... 210 C.4 Rolls-in-Series Bipolar Reactor ...................................................................... 212 C.5 Sprayer Nozzle ............................................................................................... 214 Appendix D Standard Error Calculation ................................................................................. 215   xi  List of Tables Table  1-1 Selected sets of fuel and oxidant .................................................................................. 33 Table  3-1 Elemental composition of the carbon cloth  ................................................................. 85 Table  3-2 Performance comparison of conventional DBFCs and the Swiss-roll MRFC ........... 109 Table  4-1 Variables and parameters used in the mathematical modeling .................................. 122 Table  5-1 Temperature dependence of the N2O reduction kinetics parameters ......................... 136  Table A-1 Potential fuels and oxidants for the MRFC in acidic solutions ................................. 194 Table A-2 Potential fuels and oxidants for the MRFC in basic solutions .................................. 195    xii  List of Figures Figure  1-1 Examples of mixed feed stack designs.......................................................................... 3 Figure  1-2 Cross section of sandwich construction of rolled electrode .......................................... 4 Figure  1-3 Illustration of the ?jelly-roll or ?Swiss-roll? design in batteries ................................... 5 Figure  1-4 Example of a fuel cell V-I polarization curves ............................................................. 8 Figure  1-5 Sample plot of Erdey-Gr?z?Butler?Volmer (EBV) model of electrode kinetics ....... 13 Figure  1-6 Tafel plot; high field approximation of EBV equation ............................................... 14 Figure  1-7 Configurations for 3D electrodes ................................................................................ 19 Figure  1-8 Example of the multiphase flow (L/G) patterns in electrochemical reactores ............ 24 Figure  1-9 Concept of the cylindrical ?Swiss-roll? mixed-reactant fuel cell ............................... 27 Figure  1-10 Flow diagram of the thesis ........................................................................................ 34 Figure  1-11 Schematic cross-section of Dyer?s device................................................................. 41 Figure  1-12 A series of thin-film fuel cells ................................................................................... 43 Figure  1-13 Schematic view of strip cell concept ......................................................................... 45 Figure  1-14 Schematic view of MRFT design introduced by CMR ............................................. 47 Figure  1-15 Examples of perforated MEAs of CMR fuel cell...................................................... 48 Figure  1-16 Schematic view of MEAs with flow distributor in a CMR fuel cell stack ............... 49 Figure  1-17 Schematic view of an oxygen impermeable barrier layer ......................................... 52 Figure  2-1 Cross section of sandwich construction for monopolar Swiss-roll ............................. 62 Figure  2-2 SEM micrograph of the woven carbon cloth .............................................................. 63 Figure  2-3 SEM micrograph of cathode GDEs............................................................................. 67 Figure  2-4 SEM micrographs of Viledon? and Scimat? 720/20 .................................................. 69 Figure  2-5 Photographs of the Swiss-roll components ................................................................. 70 Figure  2-6 Conceptual illustration of the SR-MRFC with feed sprayer nozzle ........................... 71 Figure  2-7 Photograph (right) and illustration (left) of Reactor A ............................................... 73 Figure  2-8 Exploded illustration of Reactor B and components ................................................... 74 Figure  2-9 Bipolar construction of the Swiss-roll in the multi layer roll arrangement ................. 75 Figure  2-10 Illustration of rolls-in-series bipolar arrangement for 5 Swiss-roll in series ............ 77 Figure  2-11 Experimental apparatus of the Swiss-roll mixed-reactant fuel cell .......................... 78 Figure  2-12 Photograph of the SR-MRFC experimental apparatus.............................................. 79 Figure  3-1 Effect of anode electrocatalysts on the Swiss-roll MRFC .......................................... 82 Figure  3-2 Performance comparisons of Os anodes using treated and un-treated carbon cloth ... 84 Figure  3-3 XPS narrow scan of Os electrodeposited anodes ........................................................ 87 Figure  3-4 SEM images of prepared Os anode ............................................................................. 89 Figure  3-5 TEM images of an Os agglomerate on carbon ............................................................ 90 Figure  3-6 Effect of the Os anode loading on the SR-MRFC....................................................... 91 Figure  3-7 Performance comparison between Os and Pt anodes in the SR-MRFC ..................... 94 Figure  3-8 NaBH4 hydrolysis rate of Pt, Os and OsO2 ................................................................. 98 Figure  3-9 Effect of cathode electrocatalysts on the Swiss-roll single-cell MRFC. ................... 100 Figure  3-10 Photograph of two different cathode current collector/fluid distributors................ 102 Figure  3-11 Effect of cathode current collector/fluid distributor and oxidant (oxygen/air) ....... 103  xiii  Figure  3-12 Comparison of Nafion? 112 and Viledon? in the SR-MRFC ................................. 105 Figure  3-13 Effect of sprayer nozzle on performance of the single cell SR-MRFC .................. 107 Figure  4-1 Simulation of the SR-MRFC: effect of hydrophilic pore volume............................. 125 Figure  4-2 Simulation of the SR-MRFC: effect of anode hydrophilic pore volume .................. 126 Figure  4-3 Simulation of the SR-MRFC and comparison with experimental data..................... 127 Figure  5-1 Cyclic voltammetry of N2O reduction on Pt ............................................................. 131 Figure  5-2 Cathodic wave of CVs for reduction of N2O on Pt electrode ................................... 132 Figure  5-3 Cathodic sweep rate variation of Pt cyclic voltammetry for reduction of N2O ........ 133 Figure  5-4 Sweep rate dependence of the peak potential for the reduction of N2O on Pt .......... 135 Figure  5-5 Temperature dependence of the rate constant for the reduction of N2O on Pt ......... 136 Figure  5-6 Cyclic voltammetry for reduction of N2O on Pt in the presence of methanol .......... 139 Figure  5-7 Impact of rotation rate on reduction of N2O on Pd ................................................... 143 Figure  5-8 Impact of temperature on N2O reduction on Pd ........................................................ 145 Figure  5-9 Effect of methanol concentration on reduction of N2O on Pd .................................. 147 Figure  5-10 Performance of the MeOH-N2O SR-MRFC ........................................................... 149 Figure  5-11 Galvanostatic stability of the MeOH-N2O SR-MRFC with PtRu anode ................ 150 Figure  5-12 Performance of the NaBH4-N2O SR-MRFC ........................................................... 152 Figure  5-13 Galvanostatic stability of the SR-MRFC NaBH4-N2O ........................................... 153 Figure  6-1 Geometric aspect ratios of the three electrodes for bipolar operation ...................... 156 Figure  6-2 Polarization curves of the single cell SR-MRFC at 3 geometrical surface area ....... 158 Figure  6-3 Effect of flow rate on a 200?10-4 m2 single-cell SR-MRFC ..................................... 160 Figure  6-4 Comparison of single-cell and dual-cell bipolar Swiss-roll MRFC .......................... 162 Figure  6-5 Effect of liquid flow in a dual-cell bipolar Swiss-roll MRFC .................................. 163 Figure  6-6 Bipolar operation of the SR-MRFC for 2, 3 and 5 rolls-in-series ............................. 165 Figure  6-7 Effect of fuel flow rate on 5 cells in series arrangement of SR-MRFC .................... 166  Figure B-1 H cell used for half-cell electrochemical tests .......................................................... 197 Figure B-2 Linear voltammetry and Tafel plot of MnO2 GDE ORR in 2 M NaOH .................. 198 Figure B-3 Schematic of the electrodeposition cell. ................................................................... 201  Figure C-1 Photograph of the Swiss-roll reactor ........................................................................ 205 Figure C-2 Dimensioned drawing of Reactor A middle rod for monopolar operation .............. 206 Figure C-3 Photograph of Reactor B and components ............................................................... 207 Figure C-4 Dimensioned drawing of Reactor B components ..................................................... 208 Figure C-5 Dimensioned drawing of Reactor B components; graphite block ............................ 209 Figure C-6 Dimensioned drawing of Reactor B components; SS middle rod ............................ 210 Figure C-7 Dimensioned drawing of middle rod for bipolar operation ...................................... 211 Figure C-8 Photograph and schematic illustration of (5 cells) rolls-in-series reactor ................ 212 Figure C-9 Dimensioned drawing of rolls-in-series components; CPVC Cap ........................... 213 Figure C-10 Dimensioned drawing of rolls-in-series components; CPVC Adaptor .................. 213 Figure C-11 Sprayer nozzle used in Reactor B ........................................................................... 214   xiv  List of Symbols    Pre-exponential factor (m s-1)   Electrode specific surface area (m2 m-3)    Activity of component i             Activity of component i (in the product side of a reaction)             Activity of component i (in reactant side of a reaction)    Cathodic Tafel slope (V dec-1)   Concentration (mol m-3)    Specific heat capacity at constant pressure (J mol-1 K-1)            Inlet concentration of sodium borohydride (mol m-3)        Inlet concentration of oxygen in air (mol m-3)           Inlet concentration of oxygen in pure O2 (mol m-3)       Bulk dissolved concentration of N2O in 0.1 M NaOH (mol m-3)            Inlet concentration of NaOH (mol m-3)          Concentration of oxidized species at surface as a function of t (mol m-3)    Concentration of component i (mol m-3)   Diffusion coefficient (m2 s-1)       BH4- diffusion coefficient (m2 s-1)          Effective BH4- diffusion coefficient (m2 s-1)      Diffusion coefficient of N2O (m2 s?1)        Carbon cloth fiber diameter (m)      Equilibrium electrode potential at P1 (VSHE)      Equilibrium electrode potential at P2 (VSHE)    Standard electrode potential (VRHE)     Standard electrode potential at temperature T (VRHE)        Standard electrode potential at 298 K (VSHE)            Standard potential of BOR at 298 K (VSHE)            Equilibrium potential of BOR at 298 K (VSHE)            Standard potential of ORR at 298 K (VSHE)            Equilibrium potential of ORR at 298 K (VSHE)       Total operating voltage of reactor (V)    Anode potential (VSHE)    Apparent activation energy (J mol -1)     Standard electrode potential of anode (VSHE)       Anode open circuit potential (VMMO)     Equilibrium potential of anodic reaction (VSHE)    Cathode potential (VSHE)     Standard electrode potential of cathode (VSHE)  xv        Cathode open circuit potential (VSHE)     Equilibrium potential of cathodic reaction (VSHE)        Standard cell voltage (V)           Cell voltage at current density of j (V)            Peak potential in voltammogram (VRHE)   Faradic constant (96,485 C mol?1)              Standard Gibbs free energy of component i (J mol-1)       Total cell current (A)   Superficial current density (A m-2)    Cathodic current density (A m-2)      Limiting current density of component i (A m-2)       Local limiting current density of BH4- oxidation (A m-2)     Overall limiting current density of BOR in 3D anode (A m-2)     Limiting current density of O2 cathode (A m-2)          Current density of BOR on MnO2 (A m-2)        Current density of BOR on Pt (A m-2)           Limiting current density of ORR in 3D anode (A m-2)          Current density of ORR on MnO2 (A m-2)        Current density of ORR on Pt (A m-2)           Apparent exchange current density of BOR on MnO2 (A m-2)         Apparent exchange current density of BOR on Pt (A m-2)           Apparent exchange current density of ORR on MnO2 (A m-2)         Apparent exchange current density of ORR on Pt (A m-2)    Exchange current density (A m-2)    Limiting current density (A m-2)         Anodic current density (A m-2)           Cathodic current density (A m-2)    Current density in kinetics controlled region (A m-2)      Net current density (A m-2)    Peak current density (A m-2)    Kohlrausch coefficient (ohm-1 m7/2 mol-3/2)    Local mass transfer coefficient of anode (m s-1)    Global cathode mass transfer coefficient (m s-1)      Mass transfer coefficient of component i (m s-1)    Cathodic reaction rate constant (m s?1)  o Standard heterogeneous rate constant (m s?1)   Characteristic length (m)    Thickness of component n (m)   Molecular weight (kg kmol-1)   Total number of electron exchanged  xvi        Number of electron transferred in rate determining step      Number of electron transferred in BOR reaction (8)      Number of electron transferred in ORR reaction (4)       Number of electron transferred in rate determining step of BOR       Number of electron transferred in rate determining step of ORR        Inlet molar rate of component i (mol s-1)        Molar rate of component i generated by Faradaic reaction (mol s-1)         Outlet molar rate of component i (mol s-1)            Molar rate of consumption in thermochemical reaction (mol s-1)            Molar rate of consumption in electrochemical reaction (mol s-1)   Oxidant pressure (kPa(abs))     Partial pressure of oxygen in feed (kPa(abs))         Pressure at state 1,2,?. (kPa(abs))     Superficial power density (W m-2)     Volumetric power density (W m-3)    Rate of heat transfer into the system (W)   Universal gas constant (8.314 J mol?1 K?1)          Total electronic plus contact resistance (Ohm m2)      Total Ohmic resistance (?)    Sherwood number   Temperature (K)    Teragram (109 kg)   Time (s)    Electro-active thickness of 3D electrode (m)   Applied electrode potential (VRHE)     Rate of shaft work to the system (W) Greek Letters     Standard entropy change of reaction (J K-1mol-1)       Entropy change of BOR (J K-1mol-1)       Entropy change of ORR (J K-1mol-1)            Contact resistance voltage drop (V)              Ionic Ohmic voltage drop across CCL (V)                Ionic Ohmic voltage drop across CCL (V)         Total Ohmic voltage drop (V)       Ohmic drop over separator (V)      Gibbs free energy change of formation (J mol-1)         Gibbs free energy change of formation at 298 K (J mol-1)      Gibbs free energy change of anodic reaction (J mol-1)      Gibbs free energy change of cathodic reaction (J mol-1)        Gibbs free energy change of reaction (J mol-1)  xvii         Change in number of moles of the gaseous species in the reaction (mol)    Voltage drop (V)      Gibbs free energy change of formation (J mol-1)         Gibbs free energy change of formation at 298 K (J mol-1)   Dimensionless charge transfer coefficient        Charge transfer coefficient of BOR on Pt          Charge transfer coefficient of BOR on MnO2          Charge transfer coefficient of ORR MnO2        Charge transfer coefficient of ORR on Pt   Symmetry factor    Anodic charge transfer coefficient    Cathodic charge transfer coefficient    Thickness of cell component i      Carbon cloth thickness (m)      Gas diffusion electrode thickness (m)    Nernst diffusion layer thickness (m)      Separator thickness (m)   Component porosity      ACL (carbon cloth) porosity      GDE Porosity    Cathode porosity    Diaphragm porosity    Electronic conductivity of component i ( mho m-1)    Electronic conductivity of 3D matrix ( mho m-1)    Anodic overpotential (V)    Cathodic overpotential (V)       Concentration overpotential (V)       Compressor efficiency   Ionic conductivity (mho m-1)         Effective ionic conductivity of GDE (mho m-1)        Effective ionic conductivity of electrolyte at 298K (mho m-1)        Effective ionic conductivity of electrolyte at 323K (mho m-1)     Effective ionic conductivity of electrolyte at T (mho m-1)         Effective ionic conductivity of separator (mho m-1)      Effective ionic conductivity of electrolyte (mho m-1)    Molar ionic conductivity of ion (or cell component) i (mho m-1)           Stoichiometry coefficient of oxidized or reduced component i    Stoichiometry coefficient of component i   Sweep rate (V s-1)      Inlet enthalpy flow (W); Ref. condition of the elements at std. state, 298K       Outlet enthalpy flow (W); Ref. condition of the elements at std. state, 298K  xviii     Molar conductivity (mho m-1 mol-1)     Molar conductivity at infinite concentration (mho m-1 mol-1)   Overpotential (V)   Volume fraction of liquid in ACL   Scan rate (V s?1)   Hydrophilic pore fraction of GDE   Liquid volumetric flow rate (m3 s-1)   Electric potential (V)  xix  List of Abbreviations 3D 3-dimensional ACL Anode catalyst layer BOR Borohydride oxidation reaction  CCL Cathode catalyst layer CE Counter electrode CMR Compact mixed-reactant CPVC Chlorinated polyvinyl chloride CV Cyclic voltammetry DBFC Direct borohydride fuel cell DMFC Direct methanol fuel cell ECS Electrochemical Society EBV Erdey-Gr?z?Butler?Volmer FTIR Fourier transform infrared spectroscopy GDE Gas diffusion electrode GDL Gas diffusion layer GTE General Telephone & Electronics Corporation IUPAC International Union of Pure and Applied Chemistry MEA Membrane electrode assembly MMO Mercury/mercury oxide  MR-DMFC Mixed-reactant direct methanol fuel cell MRFBFC Mixed-reactant flow-by fuel cell MRFC Mixed-reactant fuel cell MRFT Mixed-reactant flow-through NBL Nafion? barrier layer NCBL Nafion?-carbon barrier layer OCV Open circuit voltage OCP Open circuit potential ORR Oxygen reduction reaction PEM Proton/polymer exchange membrane PGM Platinum group metal PhD Doctor of philosophy PTFE Polytetrafluoroethylene RDE Rotating disc electrode RE Reference electrode RHE Reversible hydrogen electrode rpm Revolutions per minute SEM Scanning electron microscopy SHE Standard hydrogen electrode  SOFC Solid oxide fuel cells SPE Solid-polymer-electrolyte  xx  SR-MRFC Swiss-roll mixed-reactant fuel cell SS Stainless steel STEM Scanning transmission electron microscopy TEM Transmission electron microscopy UBC University of British Columbia UTC United Technologies Corporation VOC Varsity Outdoor Club WE Working electrode XPS X-ray photoelectron spectroscopy XRD X-ray diffraction analysis   xxi  Glossary Current density: Unless otherwise specified, all of the current densities in this dissertation are superficial (geometrical) current density (A m-2) Power density: Unless otherwise specified, all of the power densities in this dissertation are superficial (geometrical) power density (W m-2)  xxii  Acknowledgements Many, many people have helped me along my journey to complete this dissertation, but some deserve special mention.  First and foremost, I thank my family, specially my mother for her everlasting love and support. Mom, I would not be where I am today without your encouragement and support. Thank you Ardi, my brother and best friend, for being with me all the time. I express my deepest gratitude to my advisors Profs. Colin Oloman and El?d Gyenge. Thank you for the never-ending support, patience, and wisdom. Thank you for teaching me to be a critical thinker, a creative researcher, and most importantly, a good person. Certainly my experience at UBC goes beyond this thesis and will be most valuable to me because of the relationship established with Profs. Oloman and Gyenge. Colin, I must express a second thank-you for your continuous encouragement of my backcountry adventures which served to sustain the soul and made me a stronger person.  I thank my PhD committee, Prof. Kevin Smith and Dr. Jiujun Zhang, for their advice, critical thoughts, and insight into my research. I gratefully thank Prof. Thomas Hellman and Deven Dave at UBC Sauder School of Business for introducing me into the world of technology entrepreneurship. I thank Prof. David Wilkinson and his research group for supporting my PhD in any possible way. Special thanks to the University of British Columbia, Clean Energy Research Center, NSERC, and the Electrochemical Society for their generous support, awards, funding, travel grants and fellowships.  xxiii  Thank you my colleagues, mentors and dearest friends: Dr. Vincent Lam and Dr. Arman Bonakdarpour for your friendship, mentorship, our conversations, and our laughter. I must sincerely thank my fellow team members at UBC for the good time we spent together: Pooya, Sanam, Amir, Andrew, Winton, Amin, and Anna. Thank you CHBE community, in particular Head Prof. Peter Englezos, Doug Yuen, Richard Ryoo, Ivan Leversage, and Helsa Leong. Thank you Charanjit Bains, 6th floor custodian, for making my afterhours work enjoyable. Thank you my friends and fellow graduate students for making CHBE a great place to be in: Hooman, Fahimeh, Babak, James, Saad, Kevin, Dave, Hafiz, Hassan, and Iwan. Thank you Vancouver friends; Ali, Pedram, Nima, Aziz, Simin, and Kamal for all of the good times we had together. I cannot thank enough the Varsity Outdoor Club (VOC) of UBC. This club rocks! My peer mountaineers at VOC, thank you for our trips, and instructing me all sorts of backcountry skills: from touring skiing and glacier travel to rock-climbing and alpine mountaineering. Thank you British Columbia, The Best Place on Earth indeed, for the endless sources of backcountry exploration, tons of powder snow, and the Coast Mountains.  Thank you Canada for your generosity and offering me a new path of life.  Last, but not least, thank you Sona for being my love and partner. Your constant support has been a blessing. Thank you for being my compliment. Thank you for believing that I could be better than I was, and better than I am. Thank you for your endless love.  xxiv  Dedication To my mother 1  Chapter 1: Introduction 1.1 Background Worldwide demand for primary energy is projected to grow at the average rate of 1.8 % per year over the period of 2000-2030 [1]. This increased demand will be met largely by fossil fuels that emit greenhouse gasses and other pollutants. Concerns with fossil fuels have captured the attention of engineers, scientists and policy makers world-wide regarding the development of new sources and improvements in the efficiency of generation, conversion, and storage of energy. Due to their relatively high thermodynamic efficiency and low emissions, fuel cells are under development to meet energy conversion demands in the future. Fuel cells have the potential to be used in products ranging from portable devices such as mobile phones and laptops, through automotive applications like cars, buses and ships, to heat and power generators in stationary applications for the domestic and industrial sectors [2]. The commercialization of fuel cell technology hinges on socio-economic ?pull? and technology ?push?. Both driving forces for technology adoption depend on a balance between performance and cost, respecting society, the economy and the environment. It is well-known that the present capital and operating costs of fuel cell power systems must be reduced before they can be competitive with conventional energy conversion technologies. For this purpose, cheaper electrocatalysts and fuel cell designs are needed and the entire fuel cell systems should  2  be reduced in size, weight and complexity to meet the demands of many potential applications. These technical and economic challenges are the driving force of development of unconventional fuel cell systems such as mixed-reactant fuel cells (MRFCs).  1.2 Mixed-Reactant Fuel Cell (MRFC) Concept In a conventional fuel cell, the fuel and oxidant flow in separate streams, kept apart by an ion conducting membrane that divides the cell into discreet anode and cathode chambers. The single-cells are stacked in series electric connection using bipolar flow-field plates that provide most of the stack weight and volume. The membrane and bipolar plates contribute respectively 15-68 % and 10-25 % to the stack cost, depending on the intended application and stack design [3,4]. By comparison to the conventional fuel cell, in a MRFC a mixture of fuel and oxidant flows through the cell as a single stream. The mixed-feed concept allows for a variety of conventional and unconventional cell stack designs. Some previously proposed configurations of MRFCs are depicted in Figure  1-1 [5]. For instance, using conventional sandwich-type cells, the mixed-feed can flow either parallel with adjacent cells (Figure  1-1 A), or through the plane of each cell via perforations (Figure  1-1 B). An unconventional stack geometry made possible using mixed-feed is the ?strip cell? (Figure  1-1 C), in which anode and cathode are arrayed in alternating fashion on the same side of an electrolyte layer. The strip cell approach allows a compact high voltage stack due to its small cell pitch, but requires micron-scale electrode widths to avoid high Ohmic resistance from in-plane current flow [5].  3   Figure ?1-1 Examples of mixed feed stack designs: (A) sandwich cells with in-plane flow (B) sandwich cells with through plane flow (C) strip cell [5] Simplification of MRFCs systems is possible because they can operate without the gastight structures within the stack that are required for sealing, manifolding, and separate reactant delivery in conventional fuel cells. Corresponding simplifications may also be realized in the balance of plant. As a result, MRFCs could potentially provide relatively low cost fuel cell systems with high volumetric power density. 1.3 Swiss-Roll Electrochemical Reactors Electrochemical reaction engineering provides several different reactor configurations which can be chosen dependent upon the process being considered. The so-called ?Swiss-roll? or ?jelly-roll? electrochemical reactor, which is comprised of concentric spiral working and counter electrode coils separated by an electronically non-conducting membrane or screen (as in Figure  4   1-2), has been shown to have superior electrode packing density and electrochemical reaction capacity in comparison with more conventional stacked plate designs in electro-synthesis [6?8]. In this architecture, the shapes and material structures of electrodes and insulators enable axial and/or radial flow of an electrolyte through an electrode roll with a high ratio of electrode surface to reactor volume.  Figure ?1-2 Cross section of sandwich construction and schematic of rolled electrode sandwich (1) working electrode (2) secondary electrode (3) separator cloth [7] The Swiss-roll design also has been used in the majority of cylindrical rechargeable batteries (e.g. Ni-Cd and Li-ion). As shown in Figure  1-3, a sandwich made of an electronically insulating separator, a thin layer of an anode and cathode are laid down, then rolled up and inserted into a hollow cylinder casing. The battery is then sealed and metal contacts are attached for current collection/feeding. The design also has been used for primary (non-rechargeable) batteries, although most primary batteries use the conventional rod-paste-tube design.  5   Figure ?1-3 Illustration?of?the??jelly-roll?or??Swiss-roll??design?in?rechargeable?batteries; by courtesy of Encyclopaedia Britannica? Inc., Copyright 2007; used with permission  1.4 Principles of Fuel Cell Design There are certain governing relations for the design of any electrochemical system, regardless of their intended application. These relations are the material balance, the energy and the voltage and charge balance and thermodynamics. Together these relations determine the distribution of composition, temperature, potential and the electric current in an electrochemical system.   6  1.4.1 Material and Energy Balance The material and energy balances for electrochemical systems are analogous to those for thermochemical systems, except that they involve respectively Faradaic reactions and Joule heating. The material balance allows calculation of changes in the amount and composition of electrolyte (and electrodes) in the reactor. An overall balance for species i in the process is written as:                                                                 (1-1)  Where        and          are molar rate of respectively generation and consumption of component i in both thermochemical (subscript: 1) and electrochemical reaction (subscript: 2). For an electrochemical system with continuous flow at steady-state, the energy balance relates the electric energy input or output to the change in enthalpy of the process streams. The overall energy balance is:                                        (1-2) Where    is the net heat rate into the system (W) and     is the shaft work into the system including the energy output by the fuel cell, -EcellIcell  (W) [9].  7  1.4.2 Voltage Balance The voltage balance is unique to electrochemical systems because electrochemical reaction rates are both potential and temperature dependent. Basically, the voltage balances coupled with charge balance provide the potential and current distribution in an electrochemical system. The cell voltage balance is expressed as:                                      (1-3) The equilibrium potentials (Ece and Eae) depend only on composition and temperature and can be calculated from the Nernst equation. The jRohm term express the total Ohmic potential drops in the electrodes, electrolyte, connectors and separator components of an electrochemical cell. The design of cells with low electrical resistance is a major issue in electrochemical engineering. Clearly, the internal resistance is lowered by, for instance decreasing the inter-electrode gap and increasing the electrolyte conductivity (see further). The overpotential terms (?c and ?a) are defined as the differences between the applied potentials (Ec and Ea) and equilibrium potentials (Ece, and Eae) at each electrode: ?c=Ec- Ece and ?a = Ea- Eae. The overpotentials are dependent on electrode material, electrolyte composition, temperature and current density thus they contain contributions from electron transfer and mass transport processes.  8   Figure ?1-4 Example of a fuel cell V-I polarization curves indicating various regions voltage losses  Figure  1-4 shows an example of a typical fuel cell polarization curve which is separated into three regions where different voltage losses are dominant. As shown in Figure  1-4, the open circuit voltage is the actual cell potential at zero current density and, due to the establishment of mixed-potentials, may not be the same as the expected equilibrium cell voltage (see Section 1.4.8). 1.4.3 Electrochemical Thermodynamics According to the IUPAC convention, the half-cell reaction of electrochemical cells is written with the oxidized species on the left and the reduced species on the right:                                 (1-4)  9  The change in Gibbs free energy of the half-cell reaction can be calculated from Gibbs free energy of formation of components.                                            (1-5) The half-cell equilibrium potential at standard state i.e. 298 K and unit activity of reactants and products, expressed relative to the standard hydrogen electrode (SHE), can be derived from the following relationships for cathodic and anodic reactions:                      (1-6)                      (1-7) The standard electrochemical cell potential is the full cell equilibrium voltage at standard state and it can be expressed as the difference between the standard half-cell equilibrium potentials of the cathode and anode, or alternatively calculated from Gibbs free energy of the complete reaction:                             (1-8) The equilibrium potential at non-standard conditions can be calculated by the Nernst equation:                                                        (1-9)  10  The effect of temperature on the equilibrium electrode potential can be estimated from the integrated Gibbs-Helmholtz equation:                                               (1-10) Where    is the entropy change of the reaction and is approximately constant for relative small temperature changes.  The effect of pressure on the equilibrium electrode potential may be expressed as:                                     (1-11) 1.4.4 Reaction Rate and Electrode Kinetics Electrode reactions are heterogeneous catalytic processes in which intrinsic kinetics are a function of (i) reactant concentration (ii) electrode material (iii) potential (iv) temperature, and (v) electrolyte composition. The kinetic rate of a 1st order reaction (for example an electroreduction reaction) involving only one oxidized and one reduced species may be expressed by:                        (1-12) Where          is the concentration of the oxidized species at the surface of the electrode as a function of time. The specific reaction rate of a single electrode,   , is related to the electric current by Faraday? law:  11                   (1-13) Combining Eqs. (1-12) and (1-13), an expression for the cathodic current density can be obtained:                          (1-14) The    is the thermochemical reaction rate constant and its potential and temperature dependency can be estimated by an Arrhenius type equation [10]:                                (1-15) Where                           (1-16) Where    is the apparent thermochemical activation energy and    is the standard heterogeneous rate constant of the electrochemical reaction which depends on the intrinsic electrocatalytic properties of the electrode. Fast electrode kinetics is typically characterized by    ~10-2-10-1 m s-1 (e.g. Cd2+/Cd on Hg) whilst sluggish electrode kinetics is indicated by    ~?10-11 m s-1 (e.g. O2 reduction to H2O) [10]. The variable   is the symmetry factor with values between 0 and 1. The fundamental phenomenological equation of electrode kinetics was developed in the late 1920s and 1930s mainly by T. Erdey-Gr?z and M. Volmer [11,12] with independent  12  contribution from J. Butler [13?15]. Earlier in 1905, J. Tafel proposed, based on experimental studies, that the current density increases exponentially with the overpotential [16]. The Erdey-Gr?z?Butler?Volmer (EBV) equation expresses the net current density as a function of overpotential for a single reaction at a single electrode as the difference between the anodic and cathodic components of current as shown in Equation (1-17):                                     (1-17) Transfer coefficient,  , is related to symmetric factor:                    (1-18)                      (1-19) Where clearly:                    (1-20) Where      is the number of electron exchanged in the rate determining step of the reaction. The exchange current density,  , is a measure of the electrocatalytic properties of the electrode surface and expressed by the standard heterogeneous rate constant, and the bulk activities of species. Considering multiple species that affect the rate, the exchange current density can be written as:                        (1-21)  13  Where   is a power expressing the concentration/pressure dependence of the activity for species i. The exchange current density depends on the catalytic nature of electrode/electrolyte interface and also is a function of the temperature. Values of exchange current density for electrode reactions at 298 K commonly lie in the range of 1 to 10-8 A m-2 depending on the electrode material and the redox reaction [17].  Figure ?1-5 Sample plot of Erdey-Gr?z ? Butler ? Volmer (EBV) model of intrinsic electrode kinetics Figure  1-5 shows an exemplary graph of the EBV equation. When the absolute value of the overpotential is lower than about 0.025 V, both the anodic and cathodic contributions to the total current are significant and the j vs. ? relationship approaches linearity. If the absolute value of the overpotential is higher than 0.1 V either the cathodic direction or the anodic term is negligible and the EBV equation is reduced to Tafel equation:  14                              (1-22)                            (1-23) Where                      (1-24) Similar analysis can be done for anode polarization:                      (1-25) The Tafel slope depends on the catalytic nature of electrode/electrolyte interface and also is a function of the temperature. Figure  1-6 shows an exemplary graph of the Tafel equation.  Figure ?1-6 Tafel plot Eq. (1-23); high field approximation of EBV equation  15  1.4.5 Ohmic Losses The overall resistance of an electrochemical cell is simply the sum of the electronic and ionic resistance of each component in the cell i.e. cathode, anode, separator (membrane), current feeders and collectors, and contact resistances. It is important to relate the individual resistances to physical and electrochemical parameters of the components and materials such as thickness, resistivity, ionic conductivity, and porosity. The total Ohmic resistance is                           (1-26) For cell?s components utilizing ionic conductive liquid electrolyte, the effect of several factors such as concentration, temperature and inert phase void fraction on ionic conductivity must be considered. For aqueous solutions of single electrolytes the ionic conductivity may be determined form molar conductivity and is given by [18]:                 (1-27) For strong electrolytes at low concentrations, such as salts, and strong acids and bases, molar conductivity can be estimated with Kohlrausch equation [19]:                      (1-28)  16  Where     is the molar conductivity at infinite dilution (or limiting molar conductivity) and KK is the Kohlrausch coefficient, which depends on the nature of the specific salt in solution. Values of molar ionic conductivity of electrolytes, Kohlrausch coefficients, and electronic resistivity of electrode materials can be found in common physical chemistry handbooks [18].  In certain cases where non-conductive phases (solid particles or gas bubbles) are present in the electrolyte an effective specific conductivity must be determined. The effective ionic conductivity in a porous structure where total porosity is a combination of the void fraction occupied by gases and the electrolyte, a Bruggeman type equation may be used:                      (1-29) The ionic Ohmic drop across the porous separator can be written in terms of the effective ionic conductivity of the electrolyte in the separator and the current density:                             (1-30) In a diaphragm of non-conducting material saturated by electrolyte, the effective ionic conductivity can be estimated by the Maxwell equation [9]:                             (1-31)  17  1.4.6 Mass Transport In many practical electrochemical systems, the rates of electrode reaction are constrained by mass transport of reactants/products to and/or from the electrode surface. The voltage loss due to the reactant depletion in the diffusion boundary layer can be calculated by [10]:                                    (1-32) The current density under pure mass transfer control is the maximum current density which can be supported by that reactant. If migration flux is negligible the mass transfer limiting current density is given by:                       (1-33) Knowledge of the mass transfer coefficient assists in the development and operation of the electrochemical reactors with mass transfer or mixed-control. The mass transfer coefficients depend on many structural and operational conditions such as cell and electrode geometry, physiochemical property of the electrolyte, rate of gas evolution as well as temperature and hydrodynamic conditions. Values of the mass transfer coefficients may be determined via four methods [17]:  (i) From known dimensionless groups of mass transport correlations:                      (1-34)  18  In which      is estimated from the Sherwood number. Correlations for determining mass transfer coefficients under various hydrodynamic and geometric conditions are extensively discussed in the literature [20?22]. (ii) By measurements of certain species conversion (for instance electrolysis of NaBr [20]) with electrochemical reactor operation under pure mass transport control, and the application of design equations and material balances throughout the reactor.  (iii) By determination of the Nernst diffusion layer thickness:                    (1-35) In this method, measurement of the Nernst diffusion layer thickness is difficult. Optical techniques have been introduced to observe the concentration boundary layers in some cases [17]. (iv) By determination of the limiting current density by direct measurements and then using Equation (1-36):                         (1-36) It should be noted that methods (i) to (iii) measure averaged mass transport coefficient values over the electrode surface (i.e. ?global? or ?macroscopic? mass transfer coefficient). 1.4.7 Three-Dimensional (3D) Electrodes Each of the components in the voltage balance for an electrochemical reactor may be position dependent, as determined by the configuration of the cell(s). In the Swiss-roll design,  19  since the cathode and anode are parallel and close together the passage of the current through the electrolyte cannot cause significant non-uniformity of potential (or current) distribution due to geometric effects. However, non-uniformity can result from a change of the solution conductivity along the inter-electrode gap owing to changes in composition and temperature or from the voltage drop through the electrode and electrolyte in three-dimensional (3D) porous electrodes. The latter effects should be given proper consideration because Swiss-roll (fuel cell) reactors will use 3D electrodes with long and thin current feeders.  Three dimensional (3D) electrodes present an electro-active surface extending in the direction of current. Examples of 3D electrodes are particulate beds, porous plates, and fiber cloths of felt. Superficial current density on a 3D electrode is obtained by integrating the local Faradic current, on the real electrode surface, across the electrode thickness, and it can be orders of magnitude higher than the current density on a comparable 2D electrode. There are many possible configurations for 3D electrodes but two extreme types, with respect to directions of electric current and electrolyte, are flow-by and flow-through as illustrated in Figure  1-7.  Figure ?1-7 Configuration for 3D electrodes (A) flow through: current flow parallel to electrolyte flow (B) flow-by: current flow perpendicular to electrolyte flow  20  The majority of electrochemical reactors with porous 3D electrodes, including the Swiss-roll MRFC, employ the flow-by configuration (Figure  1-7 B) in which electric current flow is perpendicular to electrolyte flow. Potential distribution in a 3D electrode for a flow-by configuration is given by [9]:                               (1-37) The above equation may sometimes be solved analytically with the proper boundary conditions, but for many practical cases the solution is best obtained by numerical methods. An important consequence of Eq.(1-37) is that the potential gradient limits the electro-active thickness of a 3D electrode to a relatively thin regime facing the counter-electrode. For example, assuming a plug flow reactor under pure mass transport control the electro-active thickness of a 3D electrode matrix of infinite conductivity (i.e. conductivity of electrode material >> conductivity of the electrolyte) is given by [9,17]:                         (1-38) Where    is maximum allowed potential drop across the electrode. Equation (1-38) shows that a thicker electro-active zone will be achieved by use of high effective electrolyte conductivity of electrolyte and/or by a relatively low specific surface area and/or limiting current density. However, thin 3D electrodes with a relatively high specific surface area and mass transfer capacity are desirable features of the Swiss-roll cell. Therefore, it is important to maintain a high effective ionic conductivity inside the electrodes. Furthermore additional factors  21  also have an effect on the electro-active thickness and potential/current distribution. For instance, gas evolution at the electrodes may affect mass transport and cause Ohmic losses, as well as reducing the active electrode area by partially blocking the 3D electrode porosity. Knowledge of the effects of electrolyte conductivity, porosity, specific surface, and thickness, along with multi-phase transport phenomena within porous media, such as capillary pressure, wicking rate, convection and diffusion of reactants into the electro-active zones of the 3D electrodes are necessary in the design of Swiss-roll fuel cell reactors. 1.4.8 Mixed-Potential Theory Establishment of mixed-potentials on the cathode and/or anode is expected in many electrochemical systems, but is particularly important in the mixed-reactant systems. In order to design a MRFC, knowledge of the mixed-potential theory for understanding the mixed-feed electrode behavior is required. Similar to mixed-potential theory in corrosion studies [10,23], the theory is based on the fact that an anodic and a cathodic reaction can occur simultaneously on the same electrode. The anodic reaction produces electrons that are discharged to the electrode, and the cathodic reaction consumes electrons from the same electrode.  Calculating the mixed-potential and corresponding partial current densities is possible by solving the simultaneous electrochemical kinetic rate equations. If there is no net current flowing from this electrode into the external circuit (i.e. the electrode is at open circuit) then the rate of the anodic electron production at this electrode must be equal to the cathodic electron consumption, i.e.:  22                             (1-39) By substituting the EBV equation in Eq (1-39) one can obtain the mixed-potential at the open circuit. The open circuit voltage is different from the expected equilibrium cell voltage because of establishment of mixed-potentials at the electrodes. If, however, there is a net flow of electrons from the electrode (anodic reaction in fuel cell), then the anodic production of electrons must be greater than the cathodic consumption at this electrode, by an equivalent amount:                                 (1-40) The initial assumption of this theory is that anodic and cathodic reactions occur independently of each other on the same surface. Therefore, for example, at a given potential, the anodic reaction rate is unaffected by the presence or absence of the cathodic reaction. The main requirement for this assumption to hold is that no surface blocking occurs by either of the reactants. If this assumption is valid, then the net external current for a mixed-feed electrode can be predicted from the individual single-feed electrode currents (or reaction rates).  1.4.9 Multiphase Liquid/Gas Electrochemical Systems Mixed-reactant fuel cells like many other electrochemical processes involve continuous multiphase flow reactions and phase transformations like gas evolution and/or non-reactive phase formation. The mechanisms of many of these phenomena, and their interaction with the electrical double layer and their effect on mass transfer, Ohmic losses and cell performance are not well  23  understood. Often only empirical models of such phenomena are available for use in cell design and engineering [24?27].  Similar to thermochemical reactors, multiphase flow in electrochemical cells may result in substantial pressure gradient. Another important characteristic of multiphase flow is liquid hold-up in the reactor that reflects the liquid/solid interface behavior. The liquid hold-up in electrochemical reactors affects the effective ionic conductivity of electrolyte and consequently the current and potential distribution, pressure drop, and shunt currents. Some knowledge of multiphase fluid dynamics, including capillary effects, is important for reactor design. Few pressure drop and liquid hold-up correlations are available for fixed-bed multiphase flow electrochemical reactors composed of 3D electrodes [17,27].  Multiphase systems also affect the operation mode of the electrochemical reactors. For instance fixed-bed electrochemical reactors with liquid/gas feed may be operated with co-current flow in either the down-flow (i.e. trickle bed) [28] or up-flow modes and, as illustrated conceptually in Figure  1-8, over a range of flow regimes from liquid continuous (bubble flow) to gas continuous [27].  24   Figure ?1-8 Example of the multiphase flow (L/G) patterns in electrochemical reactors In these multiphase flow cases, the feeding method and device is an important factor affecting the pressure drop, reactant distribution and consequently reactor performance. Dispersing the multiphase fluid feeds in the feed stream and/or reactive zone of the electrochemical reactor may be obtained by various fluid contacting devices, such as spray nozzles, and in-line mixers. In the case of a gas/liquid MRFC system (e.g. a liquid fuel and gaseous oxidant) this dispersion can be obtained by a spray nozzle producing droplets of liquid dispersed in the gas (L/G) in the form of a micron-range mist. 1.4.10 Shunt Current In multi-cell electrochemical devices having a plurality of cells in series and having a common electrolyte, e.g. circulating through the cells, shunt current losses (also known as current bypasses) occur as a result of conductive paths through the electrolyte. These shunt  25  current losses may also occur under open circuit conditions, and cause undesired discharge of electrochemical devices. Also, corrosion of the electrodes and/or other components may occur, reactants may unnecessarily be consumed and excess thermal losses may result. In the Swiss-roll configuration using a liquid electrolyte (e.g. aqueous NaOH), shunt current is an issue to consider. Various modifications have been made to reduce or eliminate shunt currents in electrochemical reactors [29]: ? Electrolyte interruption techniques ? Insertion of gas bubbles into the electrolyte solution to reduce or break up the conductive path through the electrolyte  ? Having braided electrolyte inlet and outlet passages to and from the cells  ? Geometric re-design e.g. decreasing the cross-sectional area and increasing the length in the channels and manifold ? Applying protective current that counters the electrical voltage gradient in the manifold. 1.4.11 Figures of Merit The maximum thermodynamic efficiency of a fuel cell can be calculated by the following expression [17]:                                                   (1-41)  26  For reactions that involve water, it is important to note that the heat of formation,     for liquid and vapor are different. The higher heating value refers to the production of liquid water and the lower heating value refers to the production of water vapor. The electric power output for a power source per unit volume is defined as:                                        (1-42) Where     is the volumetric power density (W m-3). In case anode and cathode have the same geometrical area (A), which is common in most of the fuel cells, the superficial power density can be defined as:                                                        (1-43) The specific power of a fuel cell (W kg-1) is defined as:                                                                (1-44) Electric energy (W h) and specific energy (Wh kg-1) are defined as:                                     (1-45)                                                                  (1-46) In the above equations care should be taken to qualify exactly what is included in the weight and volume terms. The volume and weight term for fuel cells is usually measured for the  27  full packaged system including the conversion system (fuel cell stack), plus the fuel and fuel tank and relevant auxiliary components. Note that a fuel cell with no fuel has no energy, so to accurately estimate these terms the amount of fuel consumed over a period of time must be known. 1.5 Research Approach 1.5.1 Objective The idea of using a cylindrical Swiss-roll reactor for dual and single chamber fuel cells was proposed by Oloman [30], but to the Author?s knowledge, the Swiss-roll configuration for fuel cells has not previously been experimentally investigated.   Figure ?1-9 Concept?of?the?cylindrical??Swiss-roll??mixed-reactant fuel cell  28  The main objective of this research is to develop an innovative Swiss-roll mixed-reactant cylindrical fuel cell (as shown conceptually in Figure  1-9) and to investigate its performance by a combination of experimental and mathematical modeling approaches. This objective may be reached by: ? Adopting a system in which the fuel and oxidant are in separate phases  ? Selecting appropriate reactants and electrocatalysts ? Exploiting mass transfer effects to constrain undesired reactions ? Providing uniform fluid distribution to the electrodes  ? Providing for high mass transfer rates for selected reactants, with consequent high specific and/or volumetric power density ? Separator selection and eliminate the ion exchange membrane ? Scale-up and bipolar operation ? Mathematical modeling of the Swiss-roll fuel cell 1.5.2 Fuel Cell System, Reactant and Electrocatalyst Selection Many liquid fuels have been investigated in direct fuel cells, such as methanol, ethanol, formic acid, dimethyl ether, hydrazine, sodium borohydride (solution), while oxygen (air) and hydrogen peroxide are the most common oxidants. Nitrous oxide (N2O) is also an interesting oxidant because of its high reduction potential to N2, as well as a potent greenhouse effect in Earth?s atmosphere and destructive role in the ozone layer. Table A-1 and A-2 (Appendix A)  29  compares the half-cell reactions, known active electrocatalysts, full cell standard potentials and other selected properties for fuels, oxidants and their combination in acidic and basic solutions. 1.5.2.1 Fuel Cell Type Alkaline fuel cells are amongst the most mature fuel cell technologies, having been in use since the mid 1960s in NASA programs to provide electrical power for onboard systems and drinking water. Alkaline fuel cells can use non-noble cathode catalysts, thereby significantly reducing capital cost [31]. Also, the oxygen reduction reaction is faster under basic conditions (compared to acidic conditions) and good internal thermal management is achieved with a circulating liquid electrolyte. Further, a high fuel cell OCV can be achieved in alkaline conditions (Appendix A, Table A-2) and most importantly alkaline systems have the potential of lower manufacturing and operating costs. Whereas the separator/membrane and other desired components for the acid fuel cells are expensive, the equivalent in the alkaline system is relatively cheap. Moreover, corrosion of the reactor components and testing systems is a challenge in acid fuel cells. Alkaline systems are less prone to corrosion and degradation, and generally easier to run and operate. Carbonation of the electrolyte could be an issue for alkaline fuel cell and battery technology using carbonaceous organic fuels (such as methanol), and operation of air. However, it has been successfully addressed in the literature by stripping CO2 from the air feed and/or modifying the porous electrode structure creating larger pores that are less prone to clogging.  30  Therefore alkaline systems are selected for this research, but it should be noted that operation of the Swiss-roll MRFC is not limited to alkaline fuel cells and can include acid fuel cells by appropriate modifications. 1.5.2.2 Fuel Among fuel candidates, methanol has high volumetric energy density (4820 kWh m-3, 100 wt.%), and it can be generated from a number of sources like natural gas, coal, or biomass. Methanol is relatively cheap, readily available, easily stored and handled, and soluble in aqueous electrolytes. Direct methanol fuel cells (DMFCs) have received the most attention compared to other types of direct fuel cells and significant improvements have been made in the performance of the DMFC in the literature and also by previous research at UBC [32?34]. Moreover, for Pt based catalysts, faster kinetics for methanol oxidation is reported compared to ethanol, dimethyl ether and many other organic fuels [35]. Borohydrides and their various derivatives have been intensely researched for alternative energy related applications as either hydrogen storage compounds or as ?electrochemical fuels? [1,36]. Direct borohydride-oxygen fuel cells (DBFC), where borohydride (such as NaBH4) is directly supplied to the anode as an alkaline solution, possess two important advantages compared to other direct liquid fuel cells such as direct methanol, ethanol or formate. These advantages of the DBFC are: i) high theoretical gravimetric energy density of the fuel (9.3 kWh kg?1 of NaBH4), and ii) the inherent absence of carbon in the fuel. The latter implies that the DBFC is a zero carbon emission device and furthermore the inherent absence of CO (a well-known intermediate formed during electrooxidation of fuels such as methanol, ethanol and  31  formic acid, acting as catalytic poison [37?39]) could facilitate the development of durable and cheap anode. Furthermore, previous research at UBC on the borohydride oxidation electrocatalysts has resulted in significant improvement in DBFC performances with both Pt and non-Pt catalysts. With respect to the DBFC design, most of the published literature employed the conventional dual-chamber proton exchange membrane (PEM) technology in a single-cell configuration. While this set-up is adequate for laboratory scale catalyst research purposes, the plate-and-frame PEM fuel cell stack design, imported unchanged from the hydrogen-oxygen fuel cell research, poses several challenges for the scale-up and stack design of alkaline DBFCs. Some of these challenges are: PEM durability in the concentrated alkaline electrolyte, need for both gas-tight and liquid-tight sealing, use of heavy and expensive bipolar flow-field plates that must withstand the concentrated alkaline solution and need for fairly complex stack manifolds to assure uniform distribution of the alkaline borohydride solution to each anode in the stack with low pressure drop. There have been very few publications addressing any of these very important issues concerning the conventional DBFC technology and its scale-up. Therefore, the SR-MRFC design for DBFCs represents an innovative approach. 1.5.2.3 Oxidant Oxygen (air) is the most abundant and practical oxidant for the fuel cells. From the point of view of oxygen electroreduction, the alkaline electrolyte offers the possibility of using non-platinum cathode catalysts. It has been well-documented in the literature that the oxygen  32  reduction reaction (ORR) in alkaline electrolytes is catalyzed by non-platinum group (non-PGM) catalysts such as Ag, MnO2 and various activated and doped carbons [31,40?42]. Hydrogen peroxide is a strong oxidizing agent but it is unstable in alkaline solutions and therefore was not selected for this work. Also it would be useful to develop a N2O electrochemical destruction process to lower the emissions of this greenhouse gas. Platinum and palladium were shown to have activity for the electroreduction of nitrous oxide [43,44] and therefore is selected as electrocatalysts for N2O reduction this work. After a consideration of the candidate fuel/oxidant/catalyst systems set out in Appendix A Table A-2, the present research focused on the matrix of fuel and oxidant combinations set out in Table  1-1.    33  Table ?1-1 Selected sets of fuel and oxidant  O2 N2O CH3OH was not studied in this work CH3OH + 3N2O +2OH- ? 3N2 + 3H2O + CO3-2 Eo=1.84 V at 298 K NaBH4 NaBH4 + 2O2 ? NaBO2 + 2H2O Eo= 1.64 V at 298 K NaBH4 + 4N2O ? 4N2 + 2H2O + NaBO2 Eo=2.18 V at 298 K 1.5.2.4 Electrocatalysts In this thesis, different catalysts will be investigated to find sufficiently selective and active catalysts for the fuel/oxidant combinations of Table  1-1. Example catalysts are Os, Pt, and PtRu for the anode, plus Pd, MnO2 and Ag for the cathode. Osmium and Os alloys were investigated by current and previous research members in the Fuel Cell and Applied Electrochemistry Laboratory at UBC [45?48] and shown to be promising for borohydride electrooxidation. However, as the performance of osmium catalysts for MRFCs was unknown they were included in the present study of the SR-MRFC. Platinum is conventionally used for oxygen cathode catalysts due to its chemical stability and high activity. However, platinum is one of the most expensive catalysts and is active toward many reactions of fuels and oxidants. Consequently, investigations should be conducted to either lower the Pt loading or replace it with cheaper and comparatively more selective alternative cathode catalysts. For instance, Ag is an effective and fairly selective oxygen reduction catalyst in alkaline fuel cells. Manganese dioxide (MnO2) is also a good oxygen reduction catalyst in alkali and is reported to have a low activity toward the electrooxidation and hydrolysis of borohydride, making it a good candidate for the ORR in MRFC borohydride systems[49].  34  To design and develop a Swiss-roll MRFC, certain catalyst kinetics/selectivity and stability analyses are required. Here the half-cell electrochemical techniques such as cyclic voltammetry (CV) and rotating disc electrode (RDE) studies were used to investigate the activity/selectivity of the catalysts. 1.5.3 Thesis Layout  Figure ?1-10 Flow diagram of the thesis   35  The objectives of this research were met by the combination of literature review, experimental work and theoretical analysis outlined in the flow diagram of Figure  1-10 and elaborated in the individual Chapters 1 to 7: ? Chapter 1: Overview of electrochemical reaction/reactor theory and the development of mixed-reactant fuel cell technology. ? Chapter 2: Description of the Swiss-roll architecture and components.  ? Chapter 3: o Proof of principle testing of the direct borohydride SR-MRFC with respect to performance metrics such as the superficial power density. o The results of short term experiments on the performance of the SR-MRFC. o Comparison of the figures of merit of the SR-MRFC with those of conventional DBFCs in which the fuel and oxidant streams are separated by an ion exchange membrane. ? Chapter 4: o A mathematical model for the direct borohydride SR-MRFC, developed by considering the reaction kinetics and mass transport of reactants in each region of the fuel cell, incorporating the mixed-potential losses. ? Chapter 5: o Results of electro-analytical measurements to deduce the electrochemical kinetics of N2O reduction on Pt and Pd in aqueous alkaline solution.   36  o Electrochemical characterization of N2O reduction performed with methanol present in the electrolyte, for potential use in mixed-reactant MeOH-N2O fuel cells.  o The performance of the SR-MRFC for direct methanol-N2O and borohydride?N2O fuel cell chemistries. ? Chapter 6:  o The scale-up of SR-MRFC for a single cell DBFCs from 20?10-4 m2 to 200?10-4 m2 superficial electrode area. o Construction and results from the operation of two innovative arrangements of the SR-MRFC for a DBFC. ? Chapter 7:  o Concluding remarks and highlights of the thesis. o Recommendation for future work in the development of SR-MRFCs. o Contribution to knowledge.   37  1.5.4 Literature Review * The so-called mixed-reactant fuel cell (MRFC) was introduced around 50 years ago (1961). One of the first configurations of a MRFC was patented by Gr?neberg and co-workers in 1961 [50]. They tested a fuel cell stack with an alkaline electrolyte (anion membrane or KOH) and stochiometric H2/O2 gas feed (i.e. H2+0.5O2) which first was depleted of O2 at a fairly selective cathode (e.g. C, Ag, Au). Subsequently, the depleted gas mixture, below the flammability limit, was exposed to an anode electrocatalyst (e.g. Pt, Pd) that promoted hydrogen oxidation in the temperature range of 293-423 K. The fuel cell produced a total power of 85 W at 333 K for three 0.35 ? 0.60 m plates. This equates to a power density of 135 W m-2. The main aim of that work was to utilize the energy from oxy-hydrogen gas (oxygen and hydrogen in a proportion of 1:2 by volume) produced as by-products in nuclear power plants [50].  In a later variant, Goebel, Struck and Vielstich [51] operated a conventional hydrogen-oxygen cell with oxy-hydrogen gas (9 vol.% O2-91 vol.% H2) at room temperature using hydrophobic sintered-metal electrodes of area 8?10-4 m2, 3?10-3 m apart and 6 M KOH electrolyte. The hydrogen-oxygen mixture was supplied at the same composition to each electrode. The oxygen cathode was silver or sintered nickel, and the hydrogen anode was catalyzed by 0.4?10-2 kg m-2 (0.4 mg cm-2) of Pd-Pt alloy (atomic ratio 4:1). That cell delivered                                                   * Please see Glossary. Unless otherwise specified, all of the current and power density terms in this thesis refer to the superficial (geometrical) current and power density with units of respectively A m-2 and W m-2.  38  an open circuit voltage (OCV) of 1.0 V and a current density of 40 A m-2 at 0.4 V. Under load the cell attained a steady terminal voltage within minutes, which remained constant for the several hours. It was claimed that an increase in operating temperature considerably improved the power output, however there was no report of any temperature range [51].  Around the same time (1961), Grimes et al.[52] from the Allis-Chalmers Company introduced an early demonstration of the feasibility of liquid-phase (methanol-hydrogen peroxide) MRFCs and performed some experiments on a fuel cell stack consisting of 40 cells using bipolar electrodes with a Pt anode and Ag cathode. The liquid feed was a mixture of hydrogen peroxide (0.3 M) and methanol (1 M) in a 0.5?7 M potassium hydroxide solution. The stack delivered power up to 40 A at 15 V (600 W) at 353 K. In single-cell tests using a solid anion exchange membrane, an OCV of 0.41 V was measured in the mixed-feed mode, compared to 0.81 V when methanol and hydrogen peroxide were supplied separately to the anode and cathode, respectively. Analysis of reaction products determined that the net reaction in the cell was the oxidation of methanol to potassium formate, possibly either by direct electrochemical reaction for which the standard cell potential is 1.88 V at 298 K, or via disproportionation of a chemically generated methanal (formaldehyde) intermediate, for which the standard cell potential would be 0.94 V. Quantitative comparison of the reaction products with the charge passed by the cell indicated that significant direct chemical reaction, in contrast to electrochemical oxidation, occurred between the reactants. This chemical short circuiting was primarily catalyzed by the Pt electrode [52].  39  Also in 1961, Eyraud [53] described another gas-phase mixed oxygen-hydrogen device in which a micro-porous alumina support flooded with condensed moisture from the humid gas mixture acted as a film electrolyte. The original paper is published in French but that paper is cited many times in related articles. It can be concluded that the Ni-Al2O3-Pd cell was operated with OCV varying from -0.35 to +0.6 V, depending upon gas composition (no power was drawn from this cell). In 1963, Schulze [54] at the Massachusetts Institute of Technology continued the work of Grimes [52] on mixed-feed methanol fuel cells as his PhD thesis, but used oxygen and air instead of hydrogen peroxide oxidant. He tested an alkaline DMFC with either Ag or carbon cathode and Pt anode catalysts in the temperature range of 322-351 K. Silver was found to be a fairly selective cathode catalyst. Adversely, platinum anode was not selective. A porous gas diffusion type platinum anode was observed inefficient since a large non-Faradaic reaction occurred between methanol and oxygen. A perforated Pt gauze electrode was more efficient, due to physical selectivity for methanol: i.e. the gauze was wetted with methanol and electrolyte, which excluded the gaseous oxygen from platinum surface. With Ag cathode, the fuel utilization efficiency for 160 A m-2 current was 95% compared to 80% for non-perforated Pt [54]. In 1965, van Gool described a further variant of the MRFCs, the surface-migration cell [55]. In this geometry, two closely spaced selective electrodes (anode and cathode) are positioned on the same side of an insulating substrate with a film electrolyte between them. While such geometry could be operated with separate feeds of hydrogen and oxygen, van Gool suggested that the close electrode spacing (10-6 m: 1?m) would make gas separation impracticable. For the  40  selective anode electrocatalyst he suggested that a metal with a stable hydride and unstable oxide (e.g. Pt, Pd, Ir) would be an appropriate starting point for experimentation; for the selective cathode he proposed a metal, such as silver with unstable hydride and stable oxide. No experimental data were reported by van Gool [55].  Fourteen years later (1981) a single-cell and strip cell geometry of van Gool?s [55] mixed-reactant surface-migration cell were patented by Louis et al. [56] of United Technologies Corp. In the UTC approach, a supported thin-film (3?10-6 m, 3 ?m) alumina electrolyte was employed, along with closely spaced (3-4 ?10-4 m gap) supported Pt anode and SrRuO3 cathode. Each electrode was 5?10-6 m (5 ?m) thick and with an area of 10-4 m2. With a single humidified mixed gas feed of 4 vol.% O2, 4 vol. % H2 (by volume) in nitrogen, an OCV of 0.67 V was obtained at room temperature and, at peak power, a current density of 8.2 ?10-2 A m-2 at 0.39 V was generated (pressure is not mentioned in the patent [56]). They also described a surface strip cell geometry where multiple pairs of surface electrodes are interconnected electrically in series and in which several of such layers are connected in parallel. Alternative electrolyte and selective anode and cathode electrocatalysts (e.g. zirconia, LaCo0.5Ru0.5O3, LaMnO3) were also suggested. In 1990, Moseley and Williams [57] briefly described a similar Pt-oxide-Au surface-migration cell, which they had tested as a sensor in various gas mixtures at room temperature. Using a porous metal (W, Sn) oxide as substrate for the sputtered metal electrodes and for condensation of an aqueous film electrolyte, they discovered that this cell generated an OCV of up to 0.5 V in humid air at room temperature. Adding small amounts (ca. 1 vol.%) fuel gas, such as H2, CO, NH3 or ethanol vapor to the gas mixture, they found that the OCV of the cell  41  increased. The authors reasoned that the OCV response in ambient air was a result of the different mixed-potentials established by oxygen reduction and metal oxidation at the respective gold and platinum electrodes. Introduction of an additional fuel gas, they believed, mainly affected the mixed-potential at the Pt electrode not the Au. No power was drawn from this cell [57]. Also in 1990, Dyer from Bell Communications Research described a quartz-supported thin-film mixed-reactant fuel cell that operated with active but apparently non-selective electrodes [58?61]. The schematic cross-section of his variant is shown in Figure  1-11 [58].   Figure ?1-11 Schematic cross-section?of?Dyer?s?device?[58] As with the systems described by Eyraud et al [53], van Gool [55] and Louis et al [56], Dyer used a hydrated alumina film as a membrane (thickness of 5 ?10-7 m). In this system, however, the electrodes, which are positioned on either side of the alumina film, can be identical and are either Pt or Pd. As shown in Figure  1-11, only one electrode of the thin-film cell is exposed to the gaseous fuel (H2, CH4, etc) and oxidant (air, O2) mixture, while the other side is supported by an impermeable substrate. Dyer found, in his initial experiments with hydrogen-air gas mixtures,  42  that the outer Pt electrode was the anode, while the inner Pt electrode was the cathode. The open circuit voltage varied over a range of -0.2 V to 1.1 V, depending on gas composition, with OCV > 0.95 V when the gas mixture consisted of at least 50vol% H2.  Perhaps the most important factor contributing to the magnitude and polarity of the observed OCV is that the inner Pt electrode was treated in boiling water to convert an initial < 5?10-8 m coating of Al metal to boehmite phase alumina (?-AlOOH). It is likely that in this preparation process the Pt surface itself was oxidized. Compared to Pt metal, Pt-oxide was reported a reasonably selective O2 cathode [62]. A secondary contribution to the OCV may be from the differential oxygen reduction/electrode oxidation reactions occurring in the presence of a hydrous electrolyte film (Moseley and Williams [57] observed potentials up to 0.5 V) [58,61]. An alternative explanation, based on the relative ability of the two electrodes to catalyze the formation of hydrogen peroxide (the inner electrode being more active), was given by Ellgen [63], which provided the untested basis for a 1991 patent claiming a Pt-Pd alloy as a preferred cathode, simultaneously active toward chemical decomposition and electrochemical reduction of H2O2, and an anode that is inhibited toward peroxide formation but active toward hydrogen oxidation [63]. Mallouk [64] praised Dyer?s configuration in a review and described some explanation for the magnitude and polarity of the OCV for Dyer?s device in series (Figure  1-12). He believed that Dyer?s microstructure solid state fuel cell would be useful in many applications such as sensors, electric vehicle propulsion or space programs [64].  43  Dyer was also able to draw ~10-50 W m-2 from his H2/O2 Pt-Pt cell. He proposed that this thin-film fuel cell would be particularly suitable as an integrated power source in planar electronic circuits and, if deposited on a flexible substrate, could also be packaged in a compact stack form (e.g. the Swiss-roll design) to replace conventional batteries, potentially with the fuel-oxidizer mixture being supplied in liquid form [60,61].   Figure ?1-12 A series of thin-film?fuel?cells,?after?Dyer?s?design,?operating? in a mixture of hydrogen and oxygen [58,64] A few months after Dyer?s paper [58] was published in Nature, Gottesfeld [62] criticized Dyer?s device. He believed that there are potential problems with Dyer?s configuration for mixed-gas fuel cells. First of all, a mixture of H2 and O2 in the ratios considered favorable by Dyer is explosive; at safer ratios, the device performance would be limited by transport to the inner Pt electrode. Gottesfeld believed that progress towards desirable fuel cell applications is more likely to come from systems with separated fuel and oxygen streams and not MRFCs [62]. By changing the outer electrode to Ni, Dyer succeeded to reverse the polarity of his cell. This reflects Gottesfeld?s criticism [62] and Gruneberg?s original mixed-reactant cell of the 1961  44  which featured selective depletion of oxygen as the gas-phase mixture was exposed first to a hydrogen-inert cathode and then to a hydrogen-active anode [50]. In 1990, Taylor [65] (also from Bell Communications) continued Dyer?s work to improve fuel cell efficiency. Taylor had found that an unintended side reaction was occurring on the thin film solid electrolyte device resulting in the reaction of hydrogen and oxygen to form water. Taylor?s work states that a hydrogen-permeable but oxygen-impermeable barrier is needed to prevent this loss in efficiency. An efficiency improvement was suggested, involving patterning the outer electrode and coating it with a sub-micrometer-thick gas permeable ionic conducting membrane made of pseudoboehmite (fuel permeable, oxygen impermeable barrier) to increase anode selectivity to fuel and the permeability of the cell to oxygen. Other coatings such as Nylon, polysulfone, polytrifluorochlorethylene or polypropylene were also mentioned as suitable polymer for this application. No experimental data were documented in that patent [65].  Continuing research on low temperature MRFCs, in 2001 Barton et al. [66] introduced a strip cell MRFC. The strip cell arrangement, in which strips of anode and cathode material are alternated over strips of membrane electrolyte on the same side of a non-conducting support film, is shown in Figure  1-13.  45   Figure ?1-13 Schematic view of strip cell concept [66] Three electrochemical configurations were tested at 353 K and ambient pressure to provide a proof-of-concept and to compare the performance to a conventional DMFC. In the first orientation, for testing anode selectivity, a cell of 0.32 ?10-4 m2 area and membrane electrode assembly (MEA) with a PtRu black anode, Pt-black cathode and Nafion? electrolyte was operated with aqueous methanol (1 M) fed to the anode and hydrogen evolved on the cathode (which also served as a reference electrode). For mixed-reactant testing, 5?10-5 m3s-1 # * of either dry nitrogen or air was fed to the anode with the methanol.  In the second arrangement, for testing the cathode selectivity, MEAs were prepared using either FeTMPP or RuSeMo. They also used Teflon to enhance the hydrophobicity of the air                                                   * Authors of some of the reviewed literature did not specify the operating and/or measurement conditions such as pressure and temperature. Wherever some or all of these conditions were not specified in the source, a # sign was put next to the reported values or results.  46  cathode. For mixed-reactant testing, 5?10-8 m3s-1 of 1 M methanol solution was added to the air. Hydrogen was oxidized on the anode. Also, a mathematical design study by the authors provided an introduction to some of the design consideration when exploiting strip cell geometry. The effect of electrolyte-membrane thickness, electrode widths, and electrode spacing of a simple strip cell DMFC with selective electrodes was analyzed and the best parameters found to reduce Ohmic losses and increase electrode utilization. For the electrode materials considered, the following geometry is reported appropriate [66]: ? An anode width of about 10-3 m  ? A membrane thickness 1 to 2 times the average electrode width ? An electrode spacing one tenth the average electrode width In 2002, Priestnall et al. from Scientific Generics Ltd. and CMR Fuel Cells PLC, introduced the concept of a compact mixed-reactant (CMR) fuel cell and discussed its application in the DMFC, SOFC and H2/PEM technology [67?72].  47   Figure ?1-14 Schematic view of (A) conventional bipolar MEA fuel cell  and (B) MRFT design introduced by CMR [71] As illustrated in Figure  1-14, in a CMR fuel cell or so-called mixed-reactant flow-through (MRFT) architecture, the anode, cathode, and electrolyte membrane are perforated. The fuel and oxidant mixture can flow through the anode, membrane, and cathode in the same direction as the current flow. The respective selective catalysts were coated onto the porous surfaces of the anode and cathode. Perforated Nafion? was used as the electrolyte separator since the MRFT design required the porous separator to allow the flow of the reactants axially through the cell (Figure  1-15).  48   Figure ?1-15 Examples of perforated MEAs of CMR fuel cell [71] A gas diffusion layer was also provided for distributing reactants over the surfaces of the electrodes. The anode, separator, cathode, and gas diffusion layer are all sandwiched together in a cell whose overall thickness can be about 2?10-4 m. A plurality of these cells can be assembled into a stack, with gas diffusion layers separating the anode of each cell from the cathode of the adjacent cell. The CMR has developed a methanol/air mixed-reactant cell (Figure  1-16) with power densities of about 250 W m-2 using a Nafion? membrane and Ru-based cathode catalyst [72].  49   Figure ?1-16 (A) Schematic view of MEAs with flow distributor in a CMR fuel cell stack (B) a 7-cell bipolar stack (1.28 ?10-3 m? / MEA), stack volume 1.6 ?10-5 m3, nominal power is 2.5 and/or 3.8 W (depending on cathode catalyst) [71] In 2003, Shukla and co-workers [73] reported the performance of a DMFC with a mixed-reactant feed (1 M methanol and air at 373 K and ambient pressure) in conjunction with some design considerations. A DMFC was assembled with a MEA composed of Nafion?, PtRu as the anode and Pt as the cathode catalyst and total geometrical area of 25?10-4 m2.  50  Their result suggested that CO2 removal from the anode would be critical to further improve the cell performance. From a material balance Shukla and co-workers [73] found the ratio of volumetric flow rate of product carbon dioxide gas to feed methanol solution for anode is ca. 550:                                                                           (1-47) In 2004, Scott et al. [74,75] systematically investigated the mixed-reactant direct methanol fuel cells (MR-DMFCs) with MEAs comprising a carbon-supported PtRu anode and two different categories of methanol-tolerant oxygen-reduction cathode catalyst (a) macrocyclic complexes, namely transition-metal tetra-methyl phenylporphyrins (TMPPs) such as FeTMPP, CoTMPP, FeCoTMPP and (b) carbon-supported selenium containing a ruthenium-based cluster catalyst (RuSe), with both Nafion? 117 and Nafion? 112 membranes. The geometrical area was 25?10-4 m2. They used both air and pure oxygen as oxidant and performed several experiments feeding either mixed or single reactant to both cathode and anode separately. The galvanostatic polarization data for the selective reactant and mixed-reactant anode tests shows that there was no significant difference between the mixed-feed anode with methanol plus air and mixed-feed anode with methanol plus nitrogen, which is akin to the findings of Barton et al [66]. Galvanostatic polarization data for the MR-DMFCs with PtRu anode and various selective cathodes (loading of 1?10-2 kg m-2: 1 mg cm-2) catalysts were obtained operating the cells at 363 K with 1 M methanol plus 1.66?10-5 m3s-1 # oxygen feed on both the anode and the cathode. The performance curves for the MR-DMFCs employing 10-2 kg m-2 of FeTMPP,  51  CoTMPP, FeCoTMPP, and RuSe at the cathode showed that the best performance was obtained with RuSe, with a maximum power density of ca. 300 W m-2. The RuSe-based catalysts gave better performance than any of the TMPP catalysts. Scott et al [74] also investigated the performance at 362 K for MR-DMFCs employing varying amounts of RuSe at the cathode from 2?10-2 to 3?10-2 kg m-2 (2-3 mg cm-2). The highest performance, with the maximum output power density of about 500 W m-2, was obtained for the MR-DMFC with the RuSe loading of 2.5?10-2 kg m-2 (2.5 mg cm-2), operating with mixed methanol and oxygen. Increasing the cathode catalyst loading to 3?10-2 kg m-2 (3 mg cm-2) decreased the power density from 480 to 350 W m-2. The effect of oxidant (air or oxygen) was also investigated on the 2.5?10-2 kg m-2 RuSe cathode catalyst. A maximum output power density of 200 W m-2 was obtained when operating the cell with methanol plus air, compared to 480 W m-2 for oxygen and methanol under the same conditions.  Also in 2005, Jerome et al. [76] introduced the idea of incorporating molecular sieves to adjust the mass transfer selectivity of electrodes in MRFCs. Some possible selective membranes are discussed in the patent application with no experimental evaluation. These selective membranes are diatom and zeolite-based selective molecular screens, and carbide-derived carbon based membranes. In 2005, Barton and co-workers [5] reported further advances of MR-DMFCs. They tested four different categories of cathode catalysts; laccase (copper containing oxidase enzymes that are found in many plants, fungi and microorganisms), CoFeTPP (pyrolyzed metal porphyrin), RuX (metal chalcogenide) and conventional Pt using thin-film rotating disk electrodes at 313 and  52  333 K and 900 rpm over a range of methanol concentrations up to 10 M. All three catalysts demonstrated significant methanol tolerance compared to platinum. The most active catalyst in terms of current density per unit catalyst mass was laccase, which demonstrated an order of magnitude higher specific activity [5]. However, enzymatic laccase catalyst is inactive at temperatures above 323 K and it is difficult to be implemented in a gas-phase cell. The CoFeTPP and RuX catalysts show little sensitivity to temperature. The FeCoTPP is most active in the absence of methanol, but shows a higher sensitivity to methanol concentration than does RuX. The activity of both catalysts is reduced by an order of magnitude in the presence of 10 M methanol.  Barton and co-workers [5] also tried to improve the anode mass transfer selectivity. As shown in Figure 1-17, a hydrophilic barrier layer is added between the reactant distribution layer and catalyst layer of the anode to act as an oxygen barrier.  Figure ?1-17 Schematic view of an oxygen impermeable barrier layer between  reactant distribution layer and catalyst layer on the anode side [5] The hydrophilic layer provides a selective permeability to the liquid fuel in mixed-feed (also known as mass transfer selectivity based on mass transfer effects). Anode polarization curves for unmodified and barrier modified PtRu under mixed-reactant feed conditions showed that in the absence of a barrier layer, the anode is depolarized by the introduction of air in the  53  feed, the anode potential shifting positive by 0.46 V. However, introduction of air into the feed to the barrier-modified electrode did not lead to a significant shift of the anode potential. In 2006, Meng and co-workers [77] reported several electrocatalysts for a selective cathode in a mixed-reactant alkaline fuel cell. A selective Ag-W2C electrocatalyst for oxygen reduction was developed by the intermittent microwave heating method. Both the W2C and Ag-W2C showed catalytic activity for oxygen reduction in alkali. At first, W2C, Ag and Pt were tested as cathode catalyst in a standard three-electrode cell with separate anode and cathode compartments, in oxygen saturated 1 M KOH at 298 K, at an unspecified pressure. Platinum foil and Hg/HgO electrodes were used as the counter and reference electrodes, respectively. The results showed that W2C is active for oxygen reduction; however, the overpotential was higher than in case of Ag or Pt cathodes. The introduction of Ag into W2C resulted in an enhanced activity, evidenced by a positive shift in the ORR onset potential and an increased current density. The Ag-W2C gave a similar performance to Pt. The Ag-W2C catalyst showed no activity for methanol oxidation in 1 M KOH/1 M CH3OH solution in the presence and absence of O2, implying that the Ag-W2C catalyst is inert towards methanol oxidation. The effect of different fuels on electrode performance was further examined in oxygen-saturated solution, containing different concentrations of methanol, ethanol, isopropanol, and glycerol. Oxygen reduction on the Ag-W2C/C catalyst was reported stable in the methanol-containing solution up to a methanol concentration of 1 M. The onset potential and the magnitude of the current for oxygen reduction were almost the same with or without methanol [77].  54  Several attempts have been made to improve the activity and methanol tolerance of platinum cathodes, by alloying Pt with transition metals such as Cr, Fe, Co, Ni, etc. and, therefore, improve performance of fuel cells with the Pt alloy cathodes [78]. In 2006, Yuan et al [78] synthesized and characterized various carbon-supported PtFe catalysts and tested them for oxygen reduction in half-cells and in DMFCs using voltammetric and steady-state polarization measurements with PtRu anodes. By using the PtFe cathodes, methanol oxidation was partially suppressed and higher net oxygen reduction currents were achieved. The authors attributed the improvements to the modification of the structural, electronic and chemical properties of Pt by the Fe.  Continuing research on Pt-free methanol tolerant cathode electrocatalysts, in 2007, Cheng and co-workers [79] tried to enhance the activity of RuSe cathodes for the ORR through the addition of tungsten, molybdenum and rhodium. Although these alloy catalysts suffer from lower activity than Pt under the typical fuel cell conditions, it was reported that they have some promising properties such as adequate activity for oxygen reduction and high methanol tolerance [79]. All above transition metal-modified catalysts exhibited better performance than that of RuSe alone, but the tungsten modification performed the best [79]. For a conventional PEM DMFC with a PtRu anode catalyst, active area of 4?10-4 m2, Nafion? 117, 1 M methanol, oxidant: O2 ambient pressure and 363 K the peak power densities were 238, 219 and 213 W m-2 for the DMFCs with the RuSe0.20W0.29, RuSe0.20Mo0.25 and RuSe0.22Rh0.22 cathodes, respectively. The peak power densities for the DMFC with the Pt cathode was distinctly higher in the similar condition but dropped drastically with increasing methanol concentration, e.g. from 495 to 35 W m-2 as the  55  methanol concentration increases from 2 to 10 M. In contrast, for the DMFC with the RuSe0.20W0.29 cathode, the power density slightly increased from 238 to 264 W m-2 as the methanol concentration increased from 1 to 4 M; a further increase in methanol concentration to 10 M led to only a 23% decrease in power density. This demonstrated that the Ru-based catalysts have higher methanol tolerance than the Pt cathode. Nevertheless, the high methanol tolerance of the Ru-based catalysts hardly compensates their lower activity, compared to Pt. Later, Cheng et al. [80] reported the performance of their fuel-tolerant Ru-based cathode electrocatalyst in a MRFC (but with conventional PEM architecture) using formic acid, methanol and ethanol as fuel and oxygen, air and hydrogen peroxide as oxidant in 0.5 M H2SO4. They investigated the influence of fuel and oxidant conditions, feeding patterns and cathode electrocatalysts. For carbon supported Pt and Ru in a two phase mixed-feed (methanol+oxygen) flowing in both anode and cathode, the performance of Ru cathodes was better than Pt. Addition of Se and W to Ru-based cathode catalysts increased the performance and the highest performance was achieved using the RuSeW cathode. Peak power densities for methanol MRFCs with Ru-based cathodes were between 110 and 135 Wm?2. In summary, the performance of the cathodes are as follows: RuSe0.2W0.28 (135 W m?2) > RuSe (124 W m?2) > Ru (11 W m?2) > Pt (6.5W m?2) The operating conditions were MEA anode: carbon-supported PtRu (1.52?10-2 kg m?2, 1.52 mg cm?2), Nafion?117, feed: 1 M methanol + O2 at 363 K and ambient pressure cell with active area 4?10-4 m2. These results are in agreement with their previous reports in which Ru-based catalysts have been shown to give better methanol tolerance than Pt and the addition of Se  56  and W enhanced the catalyst activity for oxygen reduction [73,75,79?82]. By investigating the effect of fuel, it was reported that formic acid gave superior performance (peak power density) compared to methanol and ethanol. Peak power densities followed the following order for 1 M solutions of formic acid (150 W m?2) > methanol (135 W m?2) > ethanol (~110 W m?2) in 0.5 M H2SO4, reflecting the activity and selectivity for fuel oxidation of the PtRu anode. The fuel concentration effect of the best fuel (formic acid) was also investigated between 1 to 10 M. The best performance was achieved at 2 M formic acid with 0.5 M H2SO4 in the feed and the worst performance at 10 M, with peak power densities respectively 160 and 6.5 Wm-2. In 2007, Zeng et al [83] tried to change the mass transfer selectivity of the Pt and PtRu electrocatalysts by changing the hydrophobic properties of the electrodes in a bipolar plate-free fuel cell stack for mixed two phase ethanol/oxygen feed in a KOH solution. The electrode was prepared using carbon paper as current collector and 10 kg m?2 carbon powder as a diffusion layer on the current collector. The active area of a typical electrode was 0.79 ?10-4 m2 containing a 2?10-2 kg m?2 (2 mg cm-2) Pt (50 wt. % Pt/C) and 2?10-2 kg m?2 (2 mg cm-2) PtRu (Pt:Ru=1:1 at.). The carbon and the electrocatalysts were sprayed on the PTFE treated carbon paper. The separator was a 1.5?10-4 m thick polypropylene membrane with about 66% porosity. A solution of 0.5 M ethanol and 2 M KOH was fed along with bubbling oxygen gas at 298 K and ambient pressure. The hydrophobicity of the electrodes was adjusted by changing the amount of PTFE and Nafion? in the components (carbon paper, carbon powder sub-layer and electrocatalysts layer). It was clear that the hydrophobicity of the electrode plays an important role for the selective access of the gaseous O2 and liquid alcohol [83].  57  In 2008, Kothandaraman et al. [84] tried to improve the mass transfer selectivity of the PtRu anodes by addition of a Nafion? barrier layer (NBL) between the catalyst layer and gas diffusion layer (GDL). The NBL, similar to the previous work of Barton and co-workers [5], is a barrier layer comprised of reconstituted Nafion?, hot pressed between the anode catalyst layer and reactant diffusion layer. The NBL reduced the infiltration of oxygen from the mixed-feed (methanol+air) while allowing the permeation of aqueous methanol. From half-cell tests they reported that the NBL effectively limits oxygen diffusion at ambient pressure to the anode catalyst layer without significantly limiting the methanol oxidation rate for a MEA of 5?10-4 m2 using PtRu black and Pt as anode and cathode catalyst, respectively. Only the anode was fed mixed methanol (8 M solution) and air (9.33?10-7 #).  Ilicic et al.[85] introduced a mixed-reactant direct liquid redox PEM fuel cell in which a mixed-reactant aqueous solution containing methanol, water, Fe3+/Fe2+ redox couple and acid is fed to the cathode. The methanol permeates from the cathode to the anode through the Nafion? membrane where it is oxidized to produce CO2 and water. This novel configuration is based on the high permeability of methanol through Nafion? membranes. A peak power density of 260 W m-2 was observed at 363 K and a methanol concentration of 1 M, operating with a catholyte composition of 0.81 M FeNH4(SO4)2, 0.09 M FeSO4, and 0.5 M H2SO4. In 2011, a novel mixed-reactant rectangular flow-by fuel cell (MRFBFC) was developed by Oloman [30]. In the flow-by architecture, the fluid mixture flows generally parallel to each anode and cathode in the fuel cell, whereas it is orthogonal to the current flow. The MRFBFC can be operated in the monopolar or bipolar mode. Operating at 333 K/ 120 kPa (abs) on a feed  58  of aqueous sodium formate/sodium hydroxide mixed with air this reactor gave peak superficial and volumetric power densities respectively up to 700 W m-2 and 200 kW m-3.  Recently, novel direct fuel cell designs have been introduced with configurations inspired by the MRFC principle. Two such designs are the microfluidic fuel cell [86] and the membraneless fuel cell with an advanced 3D anode [87]. Microfluidic fuel cells use micro-channels in a parallel co-laminar flow configuration, without a physical barrier to separate the anode and cathode [86]. Convective mixing between the liquid fuel and the liquid oxidant is suppressed by operation with laminar flow at low Reynolds number. Alternatively, the membraneless/3D anode architecture of Lam et al. [87] avoids the use of a selective separator by introducing the fuel to the back face of the 3D anode, in which the fuel is converted before it reaches the inter-electrode gap adjacent to a conventional oxygen cathode. Using a PtRu anode and Pt cathode, this 3D electrode system demonstrated a power density of 60 W m-2 at 298 K/101 kPa (abs) with 1 M methanol/0.5M H2SO4 and a passive air cathode. Beginning in 1990, high-temperature mixed-reactant solid oxide fuel cells (SOFCs) were introduced by Wang [88] and developed as ?single-chamber? SOFCs extensively by Hibino and co-workers in different aspects i.e. selective electrocatalysts and reactor engineering [89]. Hibino et al. have demonstrated single-chamber SOFC?s with superficial power density up to 6440 W m-2 and operation temperatures as low as 473 K using hydrocarbons-air mixture [89,90]. Volumetric power densities of Hibino?s reactors were not reported but it can be assumed that single-chamber SOFCs have a compact design and can reach significant volumetric power densities compared to conventional dual chamber SOFCs. The detailed review of mixed-reactant SOFCs is beyond the  59  scope of this dissertation, which is focused on the development of a low-temperature (<473K) MRFC. 1.6 Summary of Literature Review  The review of the low temperature MRFC research and the range of work carried out in this area demonstrate that the mixed-reactant approach is applicable to gaseous, liquid and multi-phase reactants operating over a wide range of temperatures and fuel cell chemistries. To the best of Author?s knowledge, there is no investigation on mixed-reactant direct borohydride fuel cells. Relative to conventional fuel cells the key advantages of mixed-reactant systems identified by the various workers in this field can summarized as follows ? More compact designs are possible because manifolding is simplified ? Higher volumetric power densities are possible due to compact stacking ? Lower fabrication cost Key disadvantages of mixed-reactant systems are as follows: ? Lower energy efficiency ? Need for selective electrodes ? More complicated fluid hydrodynamics of reactants Although several developments are reported during last 10 years, fully selective catalysts are not available for MRFCs and those selective electrocatalysts that do exist are not as active as conventional fuel cell electrocatalysts, especially for oxygen reduction. Apart from the intrinsic  60  electrode kinetics, selectivity can be promoted by fluid dynamic and capillary effects designed for the selective transport of reactants into their corresponding electrodes.  Regarding design and engineering of the MRFC reactor architecture, previous research has primarily focused on operating conditions, electrode design and selective catalysts. Few novel reactor designs have been proposed and/or tested. The combination of multi-phase reactants, electrode modifications and selective catalysts, with the implication of a compact cylindrical ?Swiss-roll? reactor represents a new and innovative approach for mixed-reactant systems. The work presented in this thesis will explore this novel architecture and may lead to significant reductions in size and cost to assist the commercialization of fuel cells. Furthermore, this work is the first investigation on a mixed-reactant DBFC of any architecture.   61  Chapter 2: Swiss-Roll Fuel Cell Design and Components 2.1 Introduction This Chapter introduces the Swiss-roll architecture for MRFCs, based on the concept of the Swiss-roll electrosynthesis reactors (Chapter 1, Section  1.3). The monopolar and bipolar designs of the Swiss-roll mixed-reactant fuel cell (SR-MRFC) are described in the following sub-sections, along with their components and the fuel cell test system.  2.2  The Monopolar Swiss-Roll  Figure  2-1 shows the conceptual Swiss-roll reactor components in monopolar mode. A flexible sandwich of anode, cathode, separator and metal mesh current conductor /fluid distributor is rolled around an electronically conductive mandrel that feeds current to the cathode. The outer surface of the roll exposes a counter-electrode for electronic contact with the inside of a cylindrical vessel that houses the fuel cell and acts as the current collector. A liquid/gas mixture of fuel and oxidant is passed longitudinally through the reactor (i.e., parallel to the central mandrel). As the two fluids move along the porous cell both the anode and the cathode are exposed to fuel and oxidant. In this situation the performance of the reactor depends primarily on the electrode selectivities for the desired oxidation and reduction reactions and secondarily on the electrical conductivity (electronic and ionic) of the cell components. The reaction selectivity may be due to intrinsic electrode kinetic and/or mass transfer effects that combine to determine the rates of the anode and cathode processes.   62   Figure ?2-1 (A) Cross section of sandwich construction for monopolar Swiss-roll (B) Schematic perspective view of the Swiss-roll electrodes after rolling 2.2.1 Three-Dimensional (3D) Anode 2.2.1.1 Carbon Cloth Substrate The essential feature of the Swiss-roll reactor is a flexible sandwich of electrodes, separators and current collector/fluid distributors rolled around an axis. Thus, the substrate for the 3D anode has to fulfill a number of requirements such as mechanical flexibility coupled with mechanical strength, high electronic conductivity and hydrophilicity for preferential retention of aqueous electrolyte/fuel solutions. Comparing various fibrous carbon materials including unwoven graphite felt, carbon paper and carbon cloth, a flexible woven carbon cloth was chosen as substrate for the Pt, PtRu and electrodeposited Os anode. Figure  2-2 shows the overall morphology of the woven carbon cloth with uncompressed thickness of 350?10-6 m (350 ?m).  63   Figure ?2-2 SEM micrograph showing the morphology of the woven carbon cloth used as substrate for the 3D anode Physico-chemical treatment of carbon supports and substrates (sometimes referred to as activation) is commonly employed to impart specific surface characteristics (e.g., rougher surface, and/or modification of the oxygen-containing functional groups on the surface) that could improve the adhesion and dispersion of the deposited catalyst nano-particles, and also to modify the surface hydrophilicity [91]. Among different oxidizing agents employed for carbon activation (e.g., HNO3, H2O2, and (NH4)2S2O8), treatment with HNO3 was reported to have a great influence on both surface area and porosity development [92] and is used in the present work. Details of the experimental protocol for treatment of carbon cloth treatment with HNO3 are given in Appendix B.3.  64  2.2.1.2 Pt and PtRu Anode Electrocatalysts Platinum and PtRu are well-studied electrocatalysts for the direct methanol fuel cells [32?34,93,94] and therefore were chosen for the mixed methanol-N2O fuel cell tests of this thesis. The positive effect of Ru has been well explained in the literature on the basis of the Langmuir?Hinshelwood rate determining step model for methanol oxidation involving the surface reaction between two adsorbed species, i.e. methanol dehydrogenation products (mostly COad) [95] and hydroxyl radicals (OHad) [93,94]. Regarding the DBFCs, the electrooxidation of BH4? has been extensively investigated during the last decade. The electrooxidation mechanism involves a number of competitive electrocatalytic and thermocatalytic pathways, that can diminish the total number of electrons in the oxidation from eight to as low as four, depending on the electrocatalyst, electrode potential, temperature and OH?/BH4? concentration ratio. Investigation of the BH4? oxidation mechanism on Pt, and Pt-alloys (with Ir, Ni, Au) using techniques such as static and rotating disk electrode voltammetry, electrochemical quartz crystal microbalance and in-situ FTIR, shed some light on the adsorption of BH4? and the role of adsorbed intermediates with a general formula BHx(OH)y,ad [96?99]. The dissociative adsorption of BH4? on Pt, confirmed also by density functional theory (DFT) calculations [100,101], leading to two parallel pathways for anodic current generation, namely, oxidation of BHy,ad (with y =1,...,3) and oxidation of Had. The number of electrons transferred on smooth Pt varies between approximately six and four, corresponding to anode potentials between -1 and -0.3 and above -0.1 VSHE at 298 K, respectively [99].  65  Apart from losses associated with the electrooxidation mechanism, the non-Faradaic hydrolysis of BH4? (See Chapter 3, Section 3.2.1) catalyzed by surfaces such as Pt and Ru [102?104], generates hydrogen that may be oxidized in-situ and/or released as a gas that interferes with transport in the anode. As first described by Lam and Gyenge [105] the problem of hydrogen gas hold up in the anode can be alleviated by the use of a 3D anode structure that promotes the electrooxidation of hydrogen while allowing 2-phase (gas/liquid) flow to reduce the effect of gas occlusion on the cell voltage. In their work Lam et al. showed that, under otherwise similar conditions, the peak power density of a DBFC increased from 800 to 1300 W m?2 when the anode was changed from a conventional catalyst coated membrane-electrode to a 3D anode of catalyzed carbon felt [105]. The detailed experimental protocols for preparation of Pt and PtRu (Pt:Ru 1:1 at.) electrodes are reported in Appendix B.4. 2.2.1.3 Electrodeposited Os Gyenge and co-workers demonstrated that osmium nanoparticles are promising candidates as alternatives for platinum electrocatalysts in DBFCs because Os costs significantly less than platinum and exhibits high catalytic activity for the electrooxidation of borohydride [45?48]. Experiments with DBFCs using an Os catalyst-coated Nafion? 117 membrane revealed that at 298 K, in both the kinetic and mass-transport regions of the polarization curves, that borohydride electrooxidation on Os was superior to commercial Pt and PtRu. The superior performance of Os was attributed to lower H2(g) generation, due to the thermochemical hydrolysis of NaBH4, compared to Pt and PtRu [45?48].   66  In 2012, a surfactant-assisted galvanostatic method for osmium nanoparticle catalyst deposition on fibrous graphite substrates was developed by Lam et al.[46]. A conventional direct borohydride/oxygen fuel cell equipped with a Nafion? 117 membrane and Os/AvCarb?P75 anode with 1.7?10-2 kg m-2 (1.7 mg cm-2) Os and a 4?10-2 kg m-2 (4 mg cm?2) Pt black cathode, generated a peak power density of 1090 W m-2 at O2 446 kPa (abs) (4.4 atm) and 333 K, with good performance stability [46]. The combination of the 3D architecture of the fibrous graphitic support with osmium nanoparticle catalysts created an efficient extended reaction zone within the anode for alkaline borohydride oxidation.  Thus, in attempts toward a platinum-free Swiss-roll fuel cell, the present work included experiments on anodes prepared by surfactant-assisted electrodeposition of osmium nanoparticles into a woven carbon cloth substrate (Section 2.2.2.1). The first investigation of osmium nanoparticle activity is reported both in a direct borohydride?oxygen SR-MRFC and towards the hydrolysis of borohydride in alkaline solutions, in comparison to the behavior of a conventional platinum-based electrocatalyst. The detailed experimental protocols for Os electrodeposition and characterization are given respectively in Appendix B.5 and B.6. 2.2.2 Gas Diffusion Cathode The gas-diffusion cathodes used in this work were commercial products obtained from Gaskatel GmbH and used as is. The cathodes consisted of Teflonated Ag or MnO2 (for oxygen cathode) and Pd (for N2O cathode) catalyst layer pressed onto gold-plated Ni mesh (Figure  2-3). According to information provided by Gaskatel GmbH, catalyst loadings for Ag, MnO2 and Pd  67  were respectively 15?10-2 kg m-2 (15 mg cm-2), 20?10-2 kg m-2 (20 mg cm-2) and 2?10-2 kg m-2 (2 mg cm-2).  Figure  2-3 shows the SEM micrograph of Ag (Figure  2-3 A and C) and MnO2 (Figure  2-3 B) ORR cathode GDEs used in the SR-MRFC. The GDEs are made in the form of 3D high specific area PTFE-bonded porous structures where the Ag and MnO2 oxides are dispersed in the porous matrix. The PTFE binder in the electrodes has two functions; (a) it improves the mechanical stability of the structure and (b) suppresses cathode flooding by forming a web of PTFE fibers that creates the desired hydrophobic pore system for O2 gas to penetrate the structure.  Figure ?2-3 SEM micrograph of cathode GDEs (A) Ag (B) MnO2 (C) cross-section SEM of Ag GDE  68  2.2.3 Separator Two separators are used in this work: (a) two layers of hydrophilic microporous polypropylene (Scimat? 720/20, one layer thickness: 150 ?m; because of relatively low thickness of one layer, two layers were employed to prevent the short circuiting of the electrodes after) and (b) one layer of a hydrophilic polyolefine diaphragm (Viledon? FS2227E, Freudenberg Nonwovens with thickness 215 ?m, porosity 66%, and predominant pore radius 10 ?m) [106]. Viledon? shows good wetting characteristics (wicking rate: 0.0075 m min-1 in 30% KOH at 298 K) and high electrolyte adsorption (0.180 kg m-2, 30% KOH). The fibrous structure of the Scimat? and Viledon? separators used in this study are shown in Figure  2-4. The fibers are neutral with no functional groups. Both separators are only permeable diaphragms that prevent the cathode and anode from short-circuiting while allowing the transport of ions in the electrolyte, i.e. BH4?, OH?, and Na+. In this situation, it is necessary for the cathode electrocatalyst to have high ORR selectivity in the presence of BH4?.  69   Figure ?2-4 SEM micrograph of (A) and (B) Viledon? FS2227E (C) and (D) Scimat? 720/20 A Nafion? 112 membrane separator (DuPont) is used for comparison to the performance of Viledon?. The detail of experimental protocol of pre-treatment and conditioning of Nafion? 112 membrane before test in the SR-MRFC is reported in Appendix B.8. 2.2.4 Current Collector/Fluid Distributor The metal current collectors also act as fluid distributors and provide a path for the reactants to reach the anode and cathode catalyst layers. Examples of suitable materials for the collector/fluid distributor are metal mesh, expanded metal sheet and metal foam. As the current collector/fluid distributor in this work, a metal mesh with matching size of the connected electrode is used for both anode and cathode of the Swiss-roll cell as shown in Figure  2-5.   70  For monopolar operation, a gold plated mesh 40, 2.2?10-4 m (0.22 mm) thick, 304 stainless steel (SS) screen [thickness of gold layer: 4?10-6 m (4?m)] was placed on the 3D anode. On the cathode side, either a plain SS mesh 40 or a 316L expanded SS mesh 6, 7.6?10-4 m (0.76 mm) (Dexmet Corp., S67800/4SS (316L) 23-284DBA) was used. Further in Chapter 3, the effect of the two cathode current collectors/distributors are discussed (see Figure  3-9, Chapter 3). Figure  2-5 shows the photograph the Swiss-roll components.  Figure ?2-5 Photographs of Swiss-roll components  71  2.2.5 Two-Phase Flow Feeding System The Swiss-roll MRFC is operated with a two phase liquid (fuel)/gas (oxidant) mixture. The dispersion of the two phase fluid is an important aspect of Swiss-roll design that affects the performance of the reactor. The reactant distribution, mass transfer rates, temperature profile and parasitic power consumption are some of the engineering factors that depend on the fluid feeding device of two phase flow reactants. The selectivity of electrode reactions and consequent performance of the fuel cell also depends in part on the fluid dynamics in the reactor and the uniformity of the reactant mixture, as shown conceptually in Figure  2-6.   Figure ?2-6 Conceptual illustration of the SR-MRFC with feed sprayer nozzle  72  After initial work with a simple gas/liquid mixing tee, a gas driven liquid spray nozzle was adopted to feed the Swiss-roll reactor. The feed sprayer provides a means to enhance the performance of the Swiss-roll MRFC by dispersing the multi-phase fluid feeds in the form of a mist into the reactive zone of the reactor. Two different nozzle sprayers were used for Reactor A and B. The spray nozzles specifications (dimensions, flow rate, pressure drop, angle, droplet size, spray pattern, etc) are described in Appendix C. 2.2.6 Swiss-Roll Fuel Cell The single cell Swiss-roll MRFC investigation was carried out with three electrode geometrical areas: (i) small; 20?10-4 m2: [2 cm ? 10 cm] (ii) medium; 100?10-4 m2: [10 cm ? 10 cm] and (iii) large 200?10-4 m2: [10 cm ? 20 cm]. The small- and medium-scale Swiss-rolls were tested in Reactor A. The large-scale Swiss-roll was tested in Reactor B. Dimensioned drawings of the Swiss-roll reactors are presented in Appendix C. 2.2.6.1 Reactor A: Small- and Medium-Scale A sandwich of the Swiss-roll components is rolled around a stainless steel rod/current collector which was in electronic contact only with the cathode. Figure  2-5 A shows a photograph of the Swiss-roll components before rolling. At the end of the roll the outmost exposed anode layer made a tight press fit when slid inside the gold-plated stainless steel cylinder, serving as the Swiss-roll container and anode current collector. The reactor was covered with a 10-2 m (1 cm) thick layer of thermal insulation. Figure  2-7 show the assembled Reactor A without insulation.   73   Figure ?2-7 Photograph (right) and illustration (left) of Reactor A 2.2.6.2 Reactor B: Large-Scale Large-scale tests of Swiss-roll fuel cell were carried out in Reactor B as shown in Figure  2-8 which is made of a split graphite block current collector and a transparent polycarbonate shell. The Swiss-roll components were rolled around a SS central middle rod, inserted between the two halves of the graphite block and slid into the transparent polycarbonate reactor shell.    74   Figure ?2-8 (A) Exploded illustration of Reactor B and components; 1: polycarbonate flange, 2: polycarbonate reactor, 3: Two split graphite reactors 4: SS middle rod 5: Feed sprayer (B) Assembled illustration of Reactor B and components (C) photograph of the split graphite Reactor B showing the Swiss-roll inside. 2.3 Bipolar Swiss-Roll Design and Components A single fuel cell is only capable of producing an OCV of about 1 V, depending on the electrocatalysts, fuel, oxidant and operating conditions. Therefore it is necessary to connect and operate individual cells in series to obtain more practical voltages. Such a multi-cell reactor ?stack? can be configured with several bipolar cells in series or with series/parallel connections to tailor the voltage, current, and power output. The number of cells in one stack depends on the  75  application. In this work, two bipolar configurations of the SR-MRFC are demonstrated, as a first step in developing a compact reactor with N-bipolar cells in series. 2.3.1 Arrangement 1: Multi-Layer Roll Figure  2-9 shows the conceptual Swiss-roll reactor components in the multi-layer roll bipolar mode for two cells. For dual-cell bipolar operation, the same 3D anode and gas-diffusion cathode as for monopolar operation (Section 2.2.1 and 2.2.2) were employed. As shown in Figure  2-9 A a bipole current collector/fluid distributor (stainless steel mesh 40) is placed between the cathode of first cell and the anode of the adjacent cell to provide the necessary electronic connection between two cells. A dual cell sandwich is rolled around the axial current collector and slid inside of the Swiss-roll container/anode current collector.  Figure ?2-9 (A) Bipolar construction of Swiss-roll in the multi layer roll arrangement (two cells) (B) schematic perspective view of the Swiss-roll electrodes after rolling in bipolar arrangement for two cells  76  2.3.2 Arrangement 2: Rolls-in-Series In this configuration, the bipolar SR-MRFC comprises a number of single-cell Swiss-rolls arranged on an axis. Figure  2-10 shows a reactor with 5 rolls-in-series. The axis of the bipolar reactor is an SS rod with insulating (PTFE) inserts at some points (as shown by arrows in Figure  2-10 B). The outer reactor vessel is also segmented with electrical insulating materials (Figure  2-10 B). The electrode rolls are arranged in a way that there is an electronic gap (either in the middle rod or in the outer reactor vessel) between two adjacent cells. The mixed-reactants are fed from the top of the first roll and are forced to flow through each roll within the bipolar arrangement. The electrical connection to the row of the electrode rolls can be either parallel or series. In Figure  2-10, the series arrangement for 5 cells is shown. A photograph of the bipolar reactor is shown in Figure  2-10 C.  To properly connect the rolls-in-series the electrodes need to be rolled differently. In the first roll, the electrodes are rolled in a way that the cathode is in contact with the middle rod and anode is electronically connected to the outer vessel. In the adjacent roll, the order of the cathode-anode is reverse i.e. the anode is electronically connected to the middle rod, and cathode is electronically connected to the outer reactor vessel. Figure  2-10 B shows the polarity of the cells and direction of the electrical currents in the 5-cell series. It should be noted that the 5 cells in series is an example and this configuration potentially can be adopted for any arbitrary number of rolls. This study investigated the bipolar operation of 2, 3 and 5 cells in the rolls-in-series configuration.   77   Figure ?2-10 (A) Illustration of Rolls-in-Series bipolar arrangement for 5 Swiss-roll in series (B) same, showing the polarity and direction of electrical current (C) photograph of assembled rolls-in-series bipolar arrangement for 5 cells 2.4 Fuel Cell Test Apparatus As shown in the flow sheet in Figure  2-11, the fuel cell was supplied with a continuous downward two-phase flow of metered alkaline fuel solution and oxidant gas. The liquid feed was pumped by a Masterflex? peristaltic pump. The oxidant flow rate was monitored by a FiderisTM fuel cell test station. The 2-phase reactant feed mixture was preheated to a controlled temperature. The pressure (P) and temperature (T) across the reactor were measured by digital pressure gauges and thermocouples, respectively. The gas feed was fixed in all experiments at 10 standard liters per minute (SLPM) while the liquid flow ranged from 3 to 32 mL min-1.  78    Figure ?2-11 Experimental apparatus of Swiss-roll mixed-reactant fuel cell Figure  2-12 shows the photograph of experimental apparatus. Experimental runs consisted of fixing the process conditions and measuring the fuel cell polarization curve. In all polarization experiments, except the tests for stability and liquid flow effects in the bipolar mode, a new anode and cathode was used. Each polarization curve took 5 min to complete (from OCV to cell voltage lower than 0.1 V). Regarding calculation of figures of merit, current drawn from the reactor is divided by the geometrical surface area of electrodes to calculate the superficial current density (A m-2). The superficial current densities are multiplied by the total voltage of the cell(s) to calculate superficial power densities. For calculation of volumetric power densities (W  79  m-3), the product of current and total reactor voltage is divided by the total volume occupied by the Swiss-roll(s) between the middle rod and reactor vessel wall. More information about the Swiss-roll fuel cells and experimental procedure can be found in Appendix B and C.  Figure ?2-12 Photograph of the SR-MRFC experimental apparatus  80  Chapter 3: Single-Cell Swiss-Roll Reactor for DBFCs* 3.1 Introduction Chapter 2 introduced the novel Swiss-roll architecture for mixed-reactant fuel cells in both the monopolar and bipolar modes. As a proof-of-concept, the monopolar Swiss-roll mixed-reactant fuel cell (SR-MRFC) was tested in a direct borohydride/oxygen system with the results presented here. These tests examined the performance of the SR-MRFC in short term experiments with several variables, as well as the voltage stability in extended operation under galvanostatic conditions.  Finally, the figures of merit of the single cell SR-MRFC are compared to those of conventional DBFCs in which the fuel and oxidant streams are separated by an ion exchange membrane. The fuel cell work reported in this Chapter used only Reactor A, with an active superficial electrode area of 20?10-4 m2 (20 cm2).                                                    * A version of Chapter 3 is published as:  1. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?Platinum- and membrane-free Swiss-roll mixed-reactant alkaline fuel cell?, ChemSusChem 6 (2013) 847-855. 2. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?A Swiss-roll liquid-gas mixed-reactant fuel cell?, Journal of Power Sources 212 (2012) 154-160.  81  3.2 Anode Electrocatalyst 3.2.1 Pt and PtRu  Figure  3-1 shows polarization curves for the BH4-/O2 SR-MRFC comparing Pt with PtRu as anode catalysts. The cathode catalyst was MnO2 and two layers of Scimat? 700/20 were employed as separator (see Chapter 2 Section 2.2.3; because of relatively low thickness of one layer, two layers were employed to prevent the short circuiting of the electrodes during the rolling). The electrode reaction selectivity is paramount for the operation of any mixed-reactant system. In this regard, Chatenet et al.[49] reported that Mn-based electrocatalysts were selective toward the ORR in the presence of BH4?. Moreover, the different hydrophobic-hydrophilic balance of the electrodes, namely, the hydrophilic 3D anode and hydrophobic gas-diffusion cathode, probably helps to separate the oxidation and reduction reactions and suppress mixed-potentials.  82   Figure ?3-1 Effect of anode electrocatalysts on the Swiss-roll single-cell monopolar MRFC anode Pt or PtRu (Pt:Ru at. 1:1), loading: 0.8 mg cm-2 ? cathode: MnO2 (loading 15 mg cm-2), feed: 1 M NaBH4+2 M NaOH (8 mL min-1), O2 (10 SLPM), 323 K, 105 kPa(abs), separator: two layers of Scimat? 700/20 The results in Figure  3-1 show that Pt gave superior performance over the entire polarization curve, with a peak power density of 640 W m-2, whereas under the same conditions PtRu generated only 280 W m-2. From replicated runs with single cells using a Pt anode and MnO2 cathode the standard error of the mean peak power density was ?11 W m-2. The open-circuit voltage (OCV) was 0.86 V for Pt and 0.64 V for PtRu. Based on the standard cell potential for the direct oxidation of BH4? by O2 (i.e., 1.64 V at 298 K) a higher OCV might be expected. The loss of OCV is presumably due to the H2 evolution and oxidation in the anode, combined with mixed-potentials on the electrodes from fuel (BH4? and H2) oxidation and O2 reduction. In particular, the development of a mixed-potential on the anode cannot be fully prevented because dissolved O2 will always reach the catalytic sites in the 3D electrode.  83  Regarding the hydrogen evolution, Ru is a more active catalyst than Pt for the BH4? hydrolysis [105]: (aq)222(aq)4 BO4HO2HBH ?? ???        (3-1) Therefore, the anodic reaction will be a combination of direct borohydride oxidation and hydrogen oxidation according to Eq. (3-2):                    E?298 K = -0.83 VSHE   (3-2) Furthermore, at current densities above 1100 A m?2 the higher rate of H2 evolution on PtRu causes mass transfer limitation for the BH4?, which is manifested in the corresponding sharp drop of the cell voltage (Figure  3-1). 3.2.2 Electrodeposited Os Among the alternatives to Pt electrocatalysts in DBFCs, Gyenge and co-workers demonstrated that Os nanoparticles are promising candidates due to the combination of significantly lower cost and high catalytic activity for the electrooxidation of BH4? [46,48,100,105]. Recently, a surfactant assisted galvanostatic method for Os nanoparticle catalyst deposition on fibrous graphite substrates (e.g., AvCarb?P75) has been developed by Lam et. al [46] and it was used for preparing the Os 3D anode in this work (Section  2.2.1.3 and Appendix B.5). Figure  3-2 shows the comparative performance in a direct borohydride/oxygen SR-MRFC operated at 293 K, of Os anodes electrodeposited on the untreated (as received) and on a HNO3- 84  treated carbon cloth substrate, respectively. The Os loading in both cases was 0.12?10-2 kg m-2 (0.12 mg cm-2). The Os anode prepared by electrodeposition on the untreated carbon cloth generated a lower peak power density in the SR-MRFC compared to the anode that employed the HNO3 treated (activated) substrate (Figure  3-2). This result is due to the early onset of a BH4- mass transfer limitation in the porous anode with the untreated substrate, which is manifested in the significant voltage drop at current densities above 1000 A m-2, created by a high BH4- concentration overpotential. On the other hand, the polarization curve of the SR-MRFC with the Os/HNO3-treated anode substrate is virtually free of mass transfer limitation at current densities as high as 4000 A m-2.  Figure ?3-2 Performance comparisons in a direct borohydride/oxygen SR-MRFC of electrodeposited Os anodes using HNO3-treated and un-treated (as received) carbon cloth substrates; Os loading: 0.12?10-2 kg m-2 (0.12 mg cm-2), cathode: MnO2 gas-diffusion electrode; mixed-feed: 1 M NaBH4, 2 M NaOH (12 mL min-1) and O2 10 SLPM, 105 kPa (abs) 293 K  85  To understand the effect of HNO3 treatment on the Os 3D anode mass transfer performance exhibited in Figure  3-2, the carbon cloth substrates were analyzed by XPS before and after HNO3-treatment (Table  3-1). The elemental analysis of the untreated carbon cloth reveals a high Al content on the surface (12.99 at.%), present as Al2O3, that was most likely formed during the carbon cloth manufacturing process. Hence, the oxygen content of the untreated cloth comes almost exclusively from Al2O3 (Table  3-1) whereas, the carbon surface is virtually free of oxygen containing functional groups and therefore it is highly hydrophobic  The HNO3 treatment produced two important changes. First, it completely removed the Al2O3 from the carbon cloth and second it oxidized the carbon surface. The oxygen content of the HNO3-treated carbon cloth (7.22 at.%, Table  3-1) is due to oxygen containing functional groups formed on the carbon surface. Table ?3-1 Elemental composition (at.%) of the carbon cloth before and after pre-treatment with HNO3 Sample C O Si Al Plain carbon cloth 64.1 21.5 1.4 13.0 HNO3-treated carbon cloth 91.7 7.2 1.1 - The carbon surface oxidation with HNO3 typically generates carboxyl, hydroxyl, ketone and ether groups [92]. These functional groups impart hydrophilicity to the carbon surface and they could act as preferential nucleation sites for Os electrodeposition, hence enhancing the dispersion of Os nanoparticles. Based on Figure  3-2, the most important effect of HNO3 treatment is the improved hydrophilicity and possibly porosity of the carbon cloth that enhances  86  the wetting and wicking of the alkaline NaBH4 solution in the porous anode, thus generating superior polarization performance in the SR-MRFC in the BH4- mass transfer controlled regime. An XPS analysis was also performed on the Os electrodeposited 3D anode. Figure  3-3 shows the narrow scan XPS spectra for the Os 3D anode before and after 3 hr of functioning as the anode in the borohydride/oxygen SR-MRFC. In Figure  3-3 A, for the Os before use in the fuel cell, three Os compounds were detected. The peaks at 50.9 and 53.6 eV are assigned to the spin-orbit-split doublet of Os 4f7/2 and 4f5/2, respectively, characteristic of metallic Os(0). The binding energies at 51.9 and 54.6 eV are assigned to OsO2 4f7/2 and 4f5/2 [107?110]. Osmium dioxide (OsO2) in the fresh electrode is likely formed by hydrolysis of the OsCl62- precursor during the electrodeposition [111]. The binding energies at 53.2 and 55.9 eV can be assigned to 4f7/2 and 4f5/2 of hexachloroosmate(IV) (OsCl62-) precursor [112]. The electrodeposition bath consisted of a solution of ammonium hexachloroosmate (IV) ((NH4)2OsCl6), as described in Appendix B.5.  87   Figure ?3-3 XPS narrow scan of Os electrodeposited anodes (A) as-prepared (i.e., before operation in a SR-MRFC) (B) after 3 hr operation in the SR-MRFC In Figure  3-3 B for the Os after 3 hours of use in the fuel cell, no hexachloroosmate(IV) (OsCl22-) species are detected, showing that they are washed from the anode during operation. However, other than Os and OsO2, there is a new component present in the spectrum, with binding energies of 54.6 and 57.3 eV, respectively. These binding energies can be assigned to the dioxoosmate (VI) ion, OsO2(OH)42- [113]. The formation of OsO2(OH)42- on the anode during the  88  MRFC operation could be attributed to the presence of O2 at the anode, generating a redox cell involving O2 reduction to OH- coupled with the oxidation of either Os or OsO2 to OsO2(OH)42-. A major concern about using Os compounds is formation of a highly toxic Os oxide: OsO4. It is well-known in organometallic chemistry that aqueous alkaline solution of sodium sulfide, sodium sulfite or sodium borohydride can reduce OsO4 to less hazardous forms like Os(0) and OsO2 [114]. Therefore, the alkaline borohydride solution of DBFCs is quenching any trace amount of OsO4 if it would form during the electrodeposition or operation of the SR-MRFC. Figure  3-4 shows low and high magnification SEM images of the HNO3-treated carbon cloth and the Os electrodeposited cloth before use in the fuel cell. The carbon fiber surface is smooth but it contains several micro-cracks (Figure  3-4 A and B). The Os nanoparticles, generated by surfactant assisted electrodeposition form porous agglomerates on the carbon fiber surface with diameters equal to and less than 50 nm (Figure  3-4 C and D). The morphology of the catalyst layer is favorable for the present application because of the combination of the high surface area of the nanoparticles and the hydrophilic open pore structure of the HNO3-treated carbon cloth.  89   Figure ?3-4 (A) and (B) SEM images of HNO3-treated carbon cloth substrate at low and high magnification; (C) and (D) SEM images at low and high magnification of the as-prepared Os anode obtained by surfactant assisted electrodeposition on the HNO3-treated carbon cloth (Appendix B.5), Os loading; 0.32?10-2 kg m-2 (0.32 mg cm-2) The fine structure of the electrodeposited Os was further analyzed using TEM and STEM (Figure  3-5). The larger Os clusters are composed of Os nanoparticles with an average diameter of 4 nm (see Appendix B.6). Comparing Figure 3-4 D and Figure 3-5 A, it can be observed that some of the Os clusters detach from the carbon felt surface during operation. Diffraction pattern analysis (Figure  3-5 B inset) shows that the Os nanoparticles are amorphous. The same observation was also made in case of Os thin films produced by differential potential pulse deposition [45].  90   Figure ?3-5 TEM images of an Os agglomerate on the HNO3-treated carbon fibre surface produced by surfactant assisted electrodeposition; the electrode was analyzed after 3 hr operation in the SR-MRFC Figure  3-6 show results from an investigation of the effect of Os anode catalyst loading on the polarization curve of the borohydride/oxygen SR-MRFC, as well as the temporal stability of the fuel cell over 220 minutes continuous operation. Here the Os loading on the carbon cloth was incrementally increased from 0.12?10-2 kg m-2 (0.12 mg cm-2) to 0.32?10-2 kg m-2 (0.32 mg cm-2) by repeating the electrodeposition on a carbon cloth up to four times (Appendix B.5). The peak power density increases from 1400 to 1880 W m-2, with Os loadings of respectively 0.12?10-2 kg m-2 and 0.32?10-2 kg m-2 (Figure  3-6 A). However, the corresponding Os-mass specific power densities decreased from: 1167 W gOs-1 to 562 W gOs-1. Similar trends were reported by Lam et al. for the effect of Os loading on an AvCarbTM substrate used in a DBFC with a Nafion? 117 membrane [46]. Increasing the Os loading enhances the carbon fiber surface coverage by the Os catalyst, which leads to higher peak power densities per geometric anode  91  area (Figure  3-6 A). However, at higher catalyst loading the mass specific electrocatalytic area is reduced, which explains the lower Os-mass specific power density.  Figure ?3-6 (A) Effect of the Os anode loading on the SR-MRFC polarization curves; Anode substrate: HNO3-treated carbon cloth, 323 K; (B) Temporal stability of the SR-MRFC using the Os anode with loading of 0.32 mg cm-2, constant current density of 4500 A m-2; Other conditions are identical to Figure ?3-2   92  The temporal stability of the borohydride/oxygen SR-MRFC is shown over a 220 min period under a constant current density of 4500 A m-2, using the anode with 0.32 ?10-2 kg m-2 (0.32 mg cm-2) Os (Figure  3-6 B). This current density corresponds to the peak power density of 1880 Wm-2. The cell voltage drop is equal to or less than 5?10-3 V hr-1 (5 mV hr-1) indicating good stability for the borohydride/oxygen mixed-reactant system using an Os anode and MnO2 gas-diffusion cathode. The temporal stability shown in Figure  3-6 B is corroborated by previous results obtained on either Os/AvCarbTM P75 [46] or Os/Vulcan? XC-72 anodes [47]. These findings imply that there is relatively little Os catalyst poisoning during BH4- electrooxidation, while the observed small performance degradation is due rather to degradation of the cathode gas-diffusion electrode. The anode catalyst poisoning mechanism during BH4- electrooxidation has been much less investigated, in contrast to other direct systems, such as the direct methanol or ethanol fuel cells where anode poisoning has been extensively studied [39,115,116]. Recently, Finkelstein et al. studied the poisoning of Pt and Au electrodes during BH4- oxidation and showed that the application of high potential pulses in the platinum oxide region is an effective method for restoring the electrocatalytic activity [97]. However, whether the Os anode is poisoned or not by boron oxide or oxyhydroxide type of species has not been investigated yet and further research is required in this area. While the stability of Os for BH4- oxidation presented by Figure  3-6 B is promising, for practical applications the fuel cell durability should be investigated up to 5,000 h of operation in combination with start-up/shut-down cycles and a variety of testing conditions (e.g., fuel starvation, wide temperature range, etc.). At present, with respect to direct borohydride fuel cells  93  of any configuration there is no published information on comprehensive durability testing. Future research should address the comparative durability of mixed-reactant and conventional PEM borohydride fuel cells with Os anode catalysts. Furthermore, it is of interest to compare the Os anode catalyst with Pt, both at the same loading. Figure  3-7 presents the alkaline borohydride SR-MRFC polarization curves using Os and Pt anodes, each with a metal loading of 0.32?10-2 kg m-2 (0.32 mg cm-2).  Both Os and Pt were applied onto HNO3-treated carbon cloth. The Pt/Vulcan XC-72 was sprayed onto the cloth using an ink containing 30 wt.% Nafion? (Appendix B.4), whereas the Os anode was produced by surfactant assisted electrodeposition and the Os catalyst layer did not contain Nafion?. The two anode catalysts showed virtual identical performance, with an open circuit voltage of 0.86 V and peak power densities of 1866 and 1880 W m-2, for Pt and Os, respectively.    94   Figure ?3-7 Performance comparison between Os and Pt anodes in the SR-MRFC at 323 K, metal loading: 0.32 mg cm-2, anode substrate: HNO3-treated carbon cloth. Other conditions the same as Figure ?3-2 The similar fuel cell polarization performances of the mixed-reactant fuel cells with either Os or Pt anodes (both at the same loading of 0.32?10-2 kg m-2 (0.32 mg cm-2)) suggests that these two electrocatalysts produce similar anodic charge transfer overpotentials. Hence, the overall cell polarization behavior presented by Figure  3-7 is determined rather by the cathode performance (both charge and mass transfer related) and by the effective ionic conductivity of the separator and electrolyte. The mathematical modeling presented in Chapter 4 explores this polarization behavior and ionic losses of the SR-MRFC components in more detail. It is also important to note that in Figure 3-7, the surface roughness and specific surface area of the Pt/C and electrodeposited Os are different which affects the electrocatalytic activity  95  and therefore the SR-MRFC performance and NaBH4 hydrolysis. For a more accurate comparison, information on the crystallography, surface roughness and specific surface area of the electordepsited Os and Pt/C must be determined. While this information is widely reported for Pt/C, there is very few in depth information in the literature on the Os surface characterization and the corresponding catalytic activity for BH4? electrooxidation. With respect to the surface roughness and specific surface area determination, Ignaszaka and Gyenge have estimated the electrochemical surface area of the Os catalyst from the anodic stripping charge of the under-potential deposited (upd) hydrogen [45]. They reported that electrochemical surface area of the Os films increases with decrease of the number of deposition pulses because of the agglomeration of the Os particles over longer deposition times. At lower number of pulses a rougher and more porous Os deposit is generated, which is reflected in the higher specific surface area [45]. 3.2.3 Borohydride Hydrolysis Activity of Pt, Os and OsO2 In addition to the four times lower cost of Os, another advantage of this catalyst compared to Pt is related to the thermocatalytic H2 evolution rate. Borohydride hydrolysis (3-3) is an important secondary reaction in DBFCs that can compromise the fuel cell performance and could cause safety concerns due to H2 gas accumulation in the mixed- reactant system: NaBH4(aq) + 2H2O(l) ? 4H2(g) + NaBO2(aq)     (3-3)  The rate of thermocatalytic BH4- hydrolysis generating H2 gas must be kept low in the mixed-reactant DBFC for four main reasons: a) to avoid the accumulation of H2 to explosive  96  levels when mixed with the O2 in the reactor, b) to promote the Faradaic efficiency for BH4- conversion, c) to prevent the loss of effective electrolyte conductivity due to H2 hold-up and d) to prevent gas blinding of the 3D anode surface. To further examine the Os electrocatalyst, the NaBH4 hydrolysis catalytic activities of Pt, Os and OsO2 are compared under similar conditions at both 298 K and 323 K (Figure  3-8). Details of the experimental procedure for hydrolysis activity measurements are presented in Appendix B.7. Both Os and OsO2 were investigated because the electrodeposited surface contains both species (Figure  3-3). The OsO2 was formed by two main routes: hydrolysis of the OsCl62- precursor during electrodeposition [111] and exposure of metallic Os to O2 in the SR-MRFC reaction zone.  As shown in Figure  3-8 A, without catalyst NaBH4 is very stable in strong alkaline solutions. The half-life time of BH4- in 2 M NaOH at 298 K has been previously reported to be 430 days [117?119]. In Figure  3-8, Pt shows the highest hydrolysis rate, with a half-life times of respectively 180 min and 25 min at 298 and 323 K. The OsO2 shows a moderate activity toward NaBH4 hydrolysis with half-life times of respectively 410 min and 63 min at 298 K and 323 K. Interestingly, Os(0) has virtually no activity toward BH4 hydrolysis at either 298 K or 323 K (Figure  3-8). The hydrolysis rates of BH4- in 2 M NaOH with and without Os(0) are the same. Thus, in the case of the electrodeposited Os anode, only the OsO2 component causes some hydrolysis and H2 gas evolution (Eq. 3-1). However, even with OsO2 in the 3D anode catalyst composition, the H2 evolution rate is much lower than that on Pt (Figure  3-8).  97  It is important to note that in hydrolysis measurements (Figure 3-8), particle size and surface area of the investigated commercial powders catalysts i.e. Pt black, Os and OsO2 have an effect on NaBH4 hydrolysis activity. The commercial catalysts used in these experiments have different particle size distribution and surface area, and therefore a direct quantitative comparison of their hydrolysis activity may not be accurate. This experiment was performed only to qualitatively compare hydrolysis activity of Pt, Os and OsO2 catalysts. In addition to considerations of particle size and surface area of catalysts, some other experimental conditions must also be taken into account for an accurate quantitative hydrolysis comparison such as mass transfer effects of H2 bubbles formed as the result of BH4- hydrolysis and/or changes in the surface structure of catalysts over time due to of strong reducing effects of BH4-.  98   Figure ?3-8 NaBH4 hydrolysis rate in a solution composed of 4.3 %wt. initial NaBH4 content in 2 M NaOH. Catalysts: commercial powders of unsupported Pt, Os and OsO2 each at 5?10-5 kg (50 mg) loading, temperature: (A) 298 K, (B) 323 K; detail methodology is presented in Appendix B.7 3.3 Cathode Electrocatalyst For the cathode, two electrocatalysts were investigated: Ag and MnO2. Both are good ORR electrocatalysts in alkaline media [120?122]. In contrast to MnO2, Ag is also a good electrocatalyst for borohydride oxidation [49,123,124] and it would not seem acceptable in the  99  cathode of a borohydride/oxygen MRFC. Figure  3-9 compares the performance of Pt?MnO2 and Pt?Ag anode?cathode electrocatalyst pairs using one layer of Viledon? separator. The open circuit voltage with the Ag is slightly lower than that with the MnO2 cathode i.e., 0.75 V vs. 0.86 V, confirming a higher mixed-potential on Ag. However, at current densities above about 2500 A m?2 the cell voltage was significantly higher with the Ag cathode, with a peak power density of 2500 W m?2, compared to 1600 W m?2 for MnO2. For the SR-MRFC with Pt-Ag electrode pair, a superficial current density of as high as 10,000 A m-2 was observed (Figure  3-9). The differences between the two cathodes of Figure  3-9 could be due to a lower ORR activation polarization and higher electronic conductivity of Ag compared to MnO2 in the gas diffusion electrode. Successful utilization of the non-selective Ag electrocatalyst implies the importance of the structural and physiochemical engineering of electrodes for MRFC application. The hydrophobic-hydrophilic balance of the GDE helps to separate the two electrode reactions and suppress mixed-potentials of the borohydride oxidation reaction in the Ag cathode. Therefore, by proper engineering of electrodes, electrocatalysts with no intrinsic electro-kinetic selectivity can be utilized for the SR-MRFC, and the reactor can potentially be applicable to a variety of fuel cell chemistries.  100   Figure ?3-9 (A) Effect of cathode electrocatalysts on the Swiss-roll single-cell MRFC; anode 3D Pt (0.8 mg cm-2), cathode GDE MnO2 or Ag (B) Galvanostatic stability at 2500 A m?2. Feed: 1 M NaBH4+2 M NaOH (12 mL min-1), O2 (10 SL/min), 323 K, 105 kPa(abs), separator: one layer of Viledon?, cathode current collector: mesh 40. Figure  3-9 B shows the short term stability of the SR-MRFC with a constant current density of 2500 A m?2 at 323 K. For the SR-MRFC with Pt-Ag electrodes, the cell voltage decrease was below the detection limit of testing instrument (i.e., 100 ?V) over a 3 hour continuous operation, indicating excellent stability. For the SR-MRFC with Pt-MnO2, a cell  101  voltage decrease of about 7?10-3 V h-1 was observed which is encouraging but not satisfactory for commercial application.  These promising preliminary stability data warrant further and more detailed durability investigations. In this regard, Wagner et al [41] studied the long term stability of the PTFE-bonded Ag electrodes over 5,000 hr operation at 1000 A m-2, 343 K and observed a 20 ?V h-1 decrease in the electrochemical performance. Silver corrosion at current densities lower than 250 A m-2 [125], partial decomposition of PTFE, and changes in hydrophobic-hydrophilic pore system due to the decrease in the electrode surface roughness are the main factors contributing to degradation of the Ag GDE [41].  A review of publications up to 2013 indicates that 2500 W m-2 and 2230 W m-2, for a Pt-Ag and Pt-MnO2 electrode pair, respectively, are higher than any peak power density ever reported for low temperature (< 373 K) mixed-reactant fuel cells and match the highest power densities of DBFCs with conventional PEM architecture. Since MnO2 has an initial fixed cost advantage over Ag, the MnO2 gas diffusion cathode was used with the 3D Pt carbon cloth anode in the subsequent studies of separators, current collectors, feeding system, and in bipolar reactors. 3.4 Current Collector/Fluid Distributor In the MRFC with gas/liquid flow in contact with a GDE, flooding may be a major problem. In the present case, involving gaseous oxidants, flooding occurs when the reactant gas  102  phase in the gas diffusion cathode is replaced by liquid, which prevents transport of the oxidant to the catalyst surface and severely limits the cathode performance. Some methods proposed to suppress cathode flooding are: (a) adjusting the hydrophobicity of the cathode (b) finding a proper ratio of gas/liquid load. In addition to engineering the GDE structure and operating conditions, the dimensions and physical properties of the current collector(s)/fluid distributor(s) play an important role in controlling electrode flooding and/or reactant mass transport. As shown in Figure  3-10, SS mesh 40 has a smaller opening compared to the expanded SS mesh 6.   Figure ?3-10 Photograph of two different cathode current collector/fluid distributors used: SS mesh 40 (left) and expanded mesh 6 (right). As expected, with otherwise similar conditions, the SR-MRFC performance with expanded SS mesh 6 current collector/fluid distributor is superior to SS mesh 40 at current densities higher than 2000 A m-2 where mass transfer effects are dominant (Figure  3-11). This can be explained by the lower Laplace pressure in the larger pores of the expanded mesh 6, which allows for higher gas hold-up and improved oxygen mass transfer to the cathode. In  103  contrast, when the mesh 40 was used with smaller openings, severe oxygen mass transfer limitation occurred as a result of high liquid hold-up in the fine pores of the fluid distributor. The peak power density of the SR-MRFC with a Pt-MnO2 electrode pair and expanded SS mesh 6 fluid distributor reached 2200 W m-2.  Figure ?3-11 (A) Effect of cathode current collector/fluid distributor: SS mesh 40 and expanded SS mesh 6 (B) Effect of oxidant (oxygen and air) on the Swiss-roll single-cell MRFC; anode (Pt), cathode GDE MnO2, expanded mesh 6 current collector/fluid distributor. Other conditions are the same as Figure ?3-9 A. Figure  3-11 B compares the performance of the SR-MRFC with an expanded SS mesh 6 current collector/fluid distributor and Pt-MnO2 electrode pair under oxygen and air with  104  otherwise similar condition. The OCV with air and pure oxygen are respectively, 0.83 V and 0.86 V. When operated with air, the SR-MRFC can reach a current density of 3500 A m-2 before it faces mass transfer limitations. No significant mass transport limitation was observed with oxygen, even at a current density as high as 6000 A m-2.  3.5 Separator: Comparison of Nafion? and Viledon? Porous Diaphragm  Figure  3-12 A is a comparison in performance of a single cell SR-MRFC with 1 layer each of Nafion? 112 and Viledon?. The open circuit voltages for Nafion? and Viledon? were respectively ca. 0.91 V and 0.86 V, whereas the corresponding peak power densities were 1930 W m-2 and 2230 W m-2. The difference in performance is attributed to the combined effects of mixed-potentials and ionic conductivity, which depends on the properties of the separators. In particular, the peak power density appears to depend mainly on the internal resistance of the fuel cell, which for the Nafion? and Viledon? were respectively 0.042 ? (or 8.4?10-5 ? m2) and 0.027 ? (or 5.4?10-5 ? m2), (Figure  3-12 B inset). Nafion? 112 typically costs in the range 700-1300 $ m-2 while Viledon? costs 3 $ m-2. Therefore, in addition to significant cost reduction and higher stability, the porous polymer separator (Viledon?) offers lower Ohmic resistance when compared to the PEM. The desired ionic conductivity of the Viledon? separator relies on all the ions present in the electrolyte, in contrast to the Nafion? that selectively transfers cations only (i.e. Na+). The effective ionic conductivity of 2 M NaOH electrolyte solution in porous separator (42.9 mho m-1 at 323 K, see Chapter 4, Table  4-1) is significantly higher than that of the Nafion? 112 (2.6 mho m-1, 2 M KOH at 299 K [126]).   105  Similar superior performance of the hydrophilic polymer separators over PEMs has been shown in conventional DBFCs architectures with selective electrodes [127].  Figure ?3-12 (A) Comparison of Nafion? 112 and Viledon? in the SR-MRFC with 2 M NaOH. (B) Electrochemical impendence spectrum of the SR-MRFC with Nafion? 112 and Viledon?, inset: the high frequency points on real-axis. Pt anode, MnO2 cathode, SS expanded mesh 6 fluid distributor. Other conditions are the same as Figure ?3-9 A. Furthermore, as shown by Figure  3-12 B the impedance spectra are strongly influenced by the separator type, i.e., either cation exchange membrane (Nafion? 112) or porous diaphragm  106  (Viledon?). In case of the latter, the spectrum is of a semi-circle shape at lower frequencies, which is very similar to the spectra presented by Bidault et al. for oxygen (or air) gas diffusion cathodes in alkaline media and is described by the parallel combination of a constant phase element and resistor [31,128], whereas Nafion? 112 impedance spectra is indicative of a Warburg-type impedance imposed by the ion transport in the membrane pores as discussed and modeled by Samec et al [129]. The comparative fuel cell polarization curves with the two different separators (Figure  3-12 A), support the impedance analysis by showing a higher Ohmic potential drop in the case of Nafion? 112, whereas when Viledon? was used the Ohmic potential drop was lower but the potential drop was more pronounced in the kinetic region. Hence, the maximum power density, which occurs at current densities in the Ohmic controlled region of the polarization curve, was higher for Viledon? reaching 2230 W m-2 (Figure  3-12 A). 3.6 Effect of Nozzle Sprayer Feeding The SR-MRFC was tested with and without presence of sprayer nozzle. For the SR-MRFC without sprayer nozzle, the two phase (L/G) mixed-feed was fed directly to the reaction zone through a ?? SS tube which was centered in the Reactor A, 0.02 m above the Swiss-roll. Appendix C.5 provides more information on the sprayer nozzle used in Reactor A. The performance of SR-MRFC with and without the feed sprayer nozzle is illustrated in Figure  3-13. As shown here, a higher open circuit potential (0.86 V vs. 0.78 V) and also approximately 2.5 fold increase in maximum power density (2230 W m-2 vs. 860 W m-2) is attributed to the use of the feed spray nozzle, which distributes the fuel more uniformly over the anode and facilitates access of the oxidant gas to the cathode of the MRFC. This uniform distribution of fuel and  107  oxidant also suppresses flooding and mixed-potential on the cathode. Except for the presence and absence of the sprayer feed nozzle, all operating and reactor conditions were kept constant for Figure  3-13. Fluid visualization in the SR-MRFC was not possible with the present apparatus and reactors. However, for the sprayer nozzle, a uniform two phase droplet flow (mist) and without the sprayer nozzle an annular flow with droplets (Section 1.4.9) can be expected. This preliminary result of the feeding device in the SR-MRFC is intriguing and warrants further investigations through a concerted experimental, visualization and computational fluid dynamics.  Figure ?3-13 Effect of sprayer nozzle on performance of the single cell SR-MRFC; conditions are the same as Figure ?3-12 A  108  3.7 Swiss-Roll Performance Comparison with Conventional DBFCs Finally, Table 3-2 compares the peak power densities for various conventional PEM DBFCs reported in the literature with those from the Swiss-roll MRFC of this work. Volumetric power densities of the PEM DFBCs are typically not available from the source documents. The peak power densities obtained for the Swiss-roll cell design match the highest results obtained with conventional PEM DBFC operating under similar conditions of temperature and pressure.    109  Table ?3-2 Performance comparison of conventional DBFCs (for a complete review see [36]) and the Swiss-roll MRFC Anode loading /  mg cm-2 Cathode loading /  mg cm-2 Separator Oxidant T/PO2 K/kPa(abs) Peak power density (W m-2) Ref.  Pt (0.8) Ag (10) Viledon?  O2 323/105 2500 This work SR-MRFC Pt (0.8) MnO2 (15) Viledon? O2 323/105 2200 This work Pt (0.8) MnO2 (15) Viledon? O2 323/105 2200 This work Os (0.3) MnO2 (15) Viledon?  O2 323/105 1900 This work Pt (0.8) MnO2 (15) Viledon? Air 323/105 1100 This work         Au (2) Pd (2) Nafion?-117 O2 358/101 656 [130] Convnetional PEM design Os (1) Pt black (4) Nafion?-117 O2 333/348 690 [47] Pt (1) Pt (1) Nafion?-212 O2 333/200 1000 [131] PtRu (1) Pt black (4) Nafion?-117 O2 333/348 1300 [105] Ni (1) Pt (1) Nafion?-212 O2 333/200 1500 [131] Ni (2) Pt (2) Nafion?-117 O2 358/101 405 [130]  Pd (2) Pt (2) Nafion?-117 O2 358/101 896 [130] Pt (2) Pt (2) Nafion?-117 O2 358/101 513 [130] Ni-Pt (1) Pt (1) Nafion?-212 O2 333/200 2210 [131] Pt-Ru (1) Pt (1) Morgan? O2 333/101 1490 [132] Ag (2) Pt (2) Nafion?-117 O2 358/101 436 [130] Ag/Ti (2) Pt (2) Nafion?-117 O2 358/101 500 [133] Au (2) Ag (2) Nafion?-117 O2 358/101 328 [130] Au (2) Ni (2) Nafion?-117 O2 358/101 354 [130] Au (2) Pt (2) Nafion?-117 O2 358/101 722 [78] Au (2) FeTMPP (2) Nafion?-117 O2 358/200 653 [134] Au/Ti (2) Pt (2) Nafion?-117 O2 358/101 814 [133] Ni (167) Pt (1) Nafion?-212 Air 298/101 400 [135] Pd (1.08) Pt (0.3) Nafion?-117 Air 298/NA 194 [136] NiPd (20) Pt (1) Nafion?-112 Air 323/NA 2500 [137] Pt-Ni  Non-Pt (NA) Morgan? Air 293/NA 1150 [138] 3.8 Summary Chapter 3 reports, for the first time, the performance of the Swiss-roll single-cell mixed-reactant fuel cell (SR-MRFC). The monopolar operation of the SR-MRFC for the borohydride-oxygen system, using a liquid/gas mixture of 1 M NaBH4/2 M NaOH and O2 and air was discussed. A 3D anode with Pt, PtRu and electrodeposited Os on carbon cloth was coupled with a gas-diffusion Ag or MnO2 cathode. The effects of anode and cathode electrocatalysts, cathode- 110  side fluid distributor type, separator, feed sprayer and temporal stability were investigated. At 323 K and atmospheric pressure, a peak power of density of 2500 W m?2 was achieved with Pt-Ag anode-cathode pair (Pt loading: 0.8 mg cm-2). Under similar conditions but for an Os-MnO2 pair of electrodes (Os loading: 0.32 mg cm-2), a peak power density of 1880 W m-2 was achieved. The comparative fuel cell polarization curves and electrochemical impedance spectroscopy with the two different separators (Nafion? 112 PEM and Viledon? porous diaphragm) revealed a higher Ohmic potential drop in the case of Nafion? 112. When Viledon? was used the Ohmic potential drop was lower but the potential drop was more pronounced in the kinetic region. The maximum power density, which occurs at current densities in the Ohmic controlled region of the polarization curve, was higher for Viledon? reaching 2230 W m-2 (with Pt-MnO2 anode cathode) compared to ca. 1900 W m-2 with Nafion? 112 under otherwise similar conditions. Nafion? membranes typically cost in the range 700-1300 $ m-2 while Viledon? costs ca. 3 $ m-2. Therefore, in addition to significant cost reduction and higher stability, the porous polymer separator (Viledon?) offers lower Ohmic resistance and a higher SR-MRFC performance when compared to the PEMs. The performance of the SR-MRFC is significantly improved with the feed sprayer nozzle, as shown by a higher open circuit potential (0.86 V vs. 0.78 V) and also approximately 2.5 fold increase in maximum power density (2230 W m-2 vs. 860 W m-2 with Pt-MnO2 anode-cathode electrode pair). The effect of the feed spray nozzle is attributed to the distribution of the fuel more uniformly over the anode and facilitating access of the oxidant gas to the cathode of  111  the MRFC. This uniform distribution of fuel and oxidant also suppresses flooding and mixed-potential on the cathode. Lastly, the durability of the SR-MRFC was investigated over continuous galvanostatic operation up to 3 hours at 323 K with single-pass feed in the cell. The cell voltage decrease with Pt-Ag at 2500 A m-2 was below the detection limit of the testing instrument (i.e., 100 ?V) over the entire 220 min test time, indicating excellent stability. A cell voltage decrease of about 5?10-3 V h-1 at a constant current density of 2500 A m-2 is observed with Os-MnO2 electrode pair, which is encouraging but not satisfactory for commercial application. While a complete analysis of the SR-MRFC degradation is beyond the objective of this dissertation, it is recommended to be the focus of further studies. Ultimately, for commercial application durability has to be demonstrated for 5,000 to 20,000 hours of continuous operation with a degradation rate of only a few ?V hr-1.  112  Chapter 4: Mathematical Modeling of the Swiss-Roll Single-Cell* 4.1 Introduction Mathematical modeling is important in the design and understanding of any electrochemical system. However, with respect to mixed-reactant fuel cells of any kind, there have been very few modeling studies published [73]. To the best of Authors? knowledge, by a review up to 2013, there has been no published work on the modeling of MRFCs based on the mixed-potential theory. In the present chapter, a mathematical model of the SR-MRFC is presented for the mixed borohydride ? oxygen system based on the mixed-potential theory. In contrast to other type of direct fuel cells, very few numerical models have been developed specifically for DBFCs [139?141]. Verma and Basu [140] presented a simplified model ignoring important cathode activation and mass-transfer overpotentials. A more detailed mathematical model was developed by Shah et al [141] taking into account the mixed-potential on the anode due to hydrogen evolution on the anode. However, all the previous models neglected the mixed-potential losses at the electrodes induced by simultaneous borohydride oxidation and oxygen reduction reactions (i.e., BOR and ORR). Thus, previous models implicitly assumed perfectly selective anode and cathode electrocatalysts. Mixed-potential losses are important, not only for the SR-MRFC, but also for conventional DBFC architectures that utilize unselective                                                   * A version of Chapter 4 is submitted as:  1. A.Aziznia, C.W.Oloman, E.L.Gyenge, ?Experimental advances and mathematical modeling of the Swiss-roll mixed-reactant direct borohydride fuel cell?, (2013).  113  electrodes. At the cathode of a PEM DBFC, borohydride can cross-over the electrolyte membrane, similar to methanol cross-over in direct methanol fuel cells, lowering the Faradaic efficiency and the operating voltage of the DBFC [142]. Mixed-potential losses are even more vital in the SR-MRFC due to the mixed-reactant operation of the fuel cell and therefore, must be considered.  The model developed here embraces intrinsic kinetic, mass transport and mixed-potentials effects in the electrodes, along with the ionic and electronic resistance of the cell. However the model is limited as it applies only to an isobaric and isothermal single cell SR-MRFC operating at steady-state with a fixed reactant composition. These conditions are close to those of the present experimental work but would not apply in a larger scale fuel cell where pressure, temperature and composition gradients may be substantial. Further, this model assumes equipotential current feeders (i.e. zero voltage drops along the metal mesh) and neglects the electrooxidation of H2 from the thermocatalytic decomposition of BH4-.  4.2 Equilibrium Electrode Potentials At the anode of a DBFC, borohydride ions are oxidized on the electrocatalyst surface ideally in an eight-electron reaction:                                               (4-1) Therefore, the equilibrium potential of Eq. (4-1) depends only on composition and temperature and can be calculated from the Nernst equation according to Eq. (4-2) where the activities of the species are approximated with the molar concentration:  114                                                              (4-2) As discussed, the non-Faradaic hydrolysis of BH4- (Eq. 4-3) catalyzed by surfaces such as Pt and Ru, generates hydrogen that may be oxidized in-situ (Eq. 4-4) and/or released as gas bubbles. The H2 gas bubbles interfere with mass transport in the anode and also lower the effective anolyte ionic conductivity in the porous anode [143].                            (4-3)                                          (4-4) In the presence of borohydride hydrolysis, the electrooxidation mechanism at the anode is complicated and it is not fully understood at this time. Shah et al. presented a model involving both hydrogen evolution and borohydride oxidations on the anode [141]. In the present model, the mixed-potential due to borohydride oxidation and oxygen reduction on the anode surface is considered. However, for the sake of simplicity, the contribution of the competitive hydrogen evolution and oxidation is neglected. Furthermore, the mixed-potential on the MnO2 gas diffusion cathode due to the BOR and ORR, is also taken into account (see further). The oxygen reduction reaction (ORR) either on the cathode or on the anode (where it contributes to the mixed-potential) is given by Eq. (4-5):                                          (4-5) The equilibrium ORR potential is expressed by Eq. (4-6):  115                                                     (4-6) 4.3 Activation Overpotentials For the primary anodic and cathodic reactions, the dependence of activation overpotentials on current density is determined by the Erdey-Gr?z ? Butler?Volmer (EBV) equations:                                                                                (4-7)                                                                                (4-8)                                                                                       (4-9)                                                                                      (4-10) A series of half-cell experiments with cathode (i.e. commercial MnO2 GDE) and electrolyte identical to those used in the fuel cell experiments were performed to obtain the pertinent cathode kinetic parameters utilized in the mathematical model. The experimental procedure and electrochemical measurement results are presented in Appendix B.1. Table  4-1 shows the values of the cathode measured electrode-kinetic parameters, along with the anode kinetic parameters, which were obtained from the literature.   116  4.4 Mixed-Potentials In the SR-MRFC, the potentials at the anode and cathode are mixed-potentials due to the simultaneous occurrence of the borohydride oxidation and oxygen reduction reactions. Considering the EBV equation for BOR and ORR, the anode and cathode mixed-potentials at net zero current density (i.e., EaOCP, EcOCP) are calculated from Eq. (4-11) and (4-12):                                                                                                                                                            (4-11)                                                                                                                                                                       (4-12) 4.5 Mass-Transfer Limited Current Density The mass-transfer limited current density is defined as:                       (4-13) For the anode, considering that the carbon cloth is the only region imposing mass transport limitations and that no convective flow occurs in the pores of the carbon cloth, the local mass transport limiting current density of BH4- to ACL can be estimated by [144]:                              (4-14)  117  Assuming the local mass transport coefficient,    to be 5?10-6 m s-1, the local borohydride limiting current density is approximately 3800 A m-2.  The SR-MRFC uses the 3D electrodes with a relatively high specific surface area and mass transfer capacity (Section 1.4.7). The electro-active thickness of 3D carbon cloth can be determined by Eq. (1-32):                           (1-33) Where         is the effective electrolyte conductivity in ACL. Due to the generation of H2 bubbles at the anode surface (as a result of the borohydride hydrolysis over Pt) the carbon cloth contains a volume fraction of gas. Therefore, to estimate         with a Bruggeman type of equation, it is necessary to assume the liquid volume fraction in the porous anode, ?:                               (4-15) Where       is the electrolyte conductivity at operating temperature of T (discussed in Section 4.6, see Eq. (4-22)). It is not possible to measure the volume fraction of liquid in the carbon cloth during operation. Assuming a liquid volume fraction of 0.8 in the pores of ACL, with a BH4- concentration of 1 M and 0.1 V voltage drop through anode electrolyte in mass transfer limited regime, the electro-active thickness of the SR-MRFC was determined to be 150?10-6 m. The total limiting current density of BH4- in porous 3D anode can be then calculated by:  118                           (4-17) Where   is the electrode specific surface area (m2 m-3) and for the carbon cloth anode substrate can be estimated by [9]:                             (4-18) With      of 0.9 and         of 10?10-6 m the specific surface area of the 3D anode was determined 4?104 m2 m-3. Therefore, the overall limiting current density of the 3D anode is estimated as ca. 23000 A m-2. For the cathode GDE, the O2 mass transfer coefficient was estimated by comparing the polarization performance of the SR-MRFC under air and oxygen at similar operating conditions (Chapter 3, Figure 3-11 B). The concentration of O2 in pure oxygen feed and in air was determined by the ideal gas law at an operating pressure of 105 kPa(abs) and assuming 21 % O2 by volume for air. The limiting current density of the SR-MRFC operated with air was determined and the mass transport coefficient of the commercial cathode GDE was then calculated using:                           (4-19) The mass transport parameters values for the model are shown in Table  4-1.  119  4.6 Ohmic Losses The Ohmic losses in this model include electronic and contact resistances of the current collectors and electrodes, and ionic losses in the separator, 3D anode, and the cathode GDE. The ionic Ohmic drop across the porous separator can be written in terms of the effective ionic conductivity of the NaOH in the separator and the current density:                             (4-20) In a diaphragm of non-conducting material saturated by electrolyte, the effective ionic conductivity of the NaOH can be estimated by the Maxwell equation [9]:                             (4-21) The effect of the temperature on the ionic conductivity of NaOH is estimated by [9]:                                 (4-22) In the cathode GDE, both hydrophobic and hydrophilic pore networks are present. Hydrophilic pores allow the penetration of the electrolyte into the electrode and transport of the ions to or from the reaction zone, whereas the hydrophobic pore network is required for the transport of the oxygen to the reaction zone. Therefore, to estimate the effective ionic conductivity of the electrolyte in the cathode GDE with a Bruggeman type of equation, it is necessary to assume the hydrophilic pore volume fraction, ?:                               (4-23)  120  The total current density in the GDE is a combination of the electronic and ionic current densities:                    (4-24) Usually pertinent differential equations must be solved to find the electronic and ionic solution current densities. However, in this model it is assumed that the entire cathode GDE is electrochemically active and the electronic and ionic current densities are equal:                    (4-25) A detailed analysis for the above simplifying assumption has been described by Shah et al. [141]. Therefore, the ionic Ohmic loss of the GDE then can be calculated with:                                       (4-26) The total ionic potential drop in the anode can be estimate with:                                    (4-27) The effective ionic conductivity of the electrolyte in the anode with a Bruggeman type of equation was described in Section 4.5 take into account the liquid volume fraction of the porous anode, ? (assuming that ACL contains H2 gas bubbles due to BH4- hydrolysis on Pt anode):                              (4-28)  121   The voltage loss due to the all of the electronic plus contact resistances can be calculated using Ohm?s law:                                             (4-29) The total electronic plus contact resistance of the SR-MRFC components is measured according to a method described in the Appendix B.9 and is reported in Table  4-1. 4.7 Modeling Approach At non-zero current densities, conservation of charge demands that:                                        (4-30) Each of the current densities in Eq.(4-30) can be expressed as, charge transfer and mass transport current densities, using:                                                                                        (4-31)                                                                                           (4-32)                                                                                                (4-34)  122                                                                                                (4-33) The cell voltage is calculated from:                                                       (4-35) All terms in Eq. (4-35) have been defined in previous sections. At a given cell voltage (Ecell), Equations (4-30 to 4-35) can be solved numerically and the SR-MRFC polarization curve is simulated. Variables and parameters of the Swiss-roll fuel cell components used in the model are shown in Table  4-1. Table ?4-1 Variables and parameters used in the mathematical modeling of the SR-MRFC (Reactor A: 20?10-4 m2) Symbol Definition Value Unit Note Operating conditions     P Oxidant pressure 105 kPa(abs)  T Temperature 323 K           O2 inlet concentration 39 mol m-3         Air inlet concentration 8 mol m-3  ? Liquid flow rate 2?10-6 m3 s-1            Inlet concentration of NaBH4 1000 mol m-3             Inlet concentration of NaOH 2000 mol m-3  Anodic related                 Standard potential of BOR at 298 K -1.24 VSHE [145]            Standard potential of BOR at 298 K -1.40 VMMO        Anode open circuit potential -1.14 VMMO measured*       Entropy of BOR 357.8 J mol-1 K-1 [146]        BOR charge transfer coefficient on Pt 0.4 - fitted          BOR charge transfer coefficient on MnO2 0.5 - fitted           BOR exchange current density on MnO2 4.26 A m-2 Appx. B.1  123          BOR exchange current density on Pt 60 A m-2 [147]      Number of electron transfer in BOR per mole BH4-  8 -        Number of electron transfer in rate determining step of BOR 2 - assumed      ACL (carbon cloth) thickness 300?10-6 m SEM      ACL (carbon cloth) porosity 0.9 - assumed       BH4- diffusion coefficient in 2 M NaOH 3.63?10?9 m2 s?1 313 K [148] ? Volume fraction of liquid in      0.8 - assumed dcloth Carbon cloth fiber diameter 10?10-6 m SEM    Local mass transfer coefficient of anode 5?10-6 m s-1 [149]     Overall limiting current density of 3D anode 23000 A m-2 calc. Cathodic related                Standard potential of the ORR at 298 K 0.40 VSHE, 298 K [145]            Standard potential of the ORR at 298 K 0.21 VMMO        Anode open circuit potential -0.24 VMMO measured*       Entropy of the ORR -387 J mol-1 K-1 [146]          ORR charge transfer coefficient: MnO2 0.6 - fitted        ORR Charge transfer coefficient: Pt 0.3 - [150]           ORR apparent exchange current density: MnO2 0.49 A m-2 Appx. B.1         ORR apparent exchange current density: Pt 2.1?10-3 A m-2 [150]      Number of electron transfer in ORR per mole O2 4 -        Number of electron transfer in rate determining step of ORR 1 - assumed      Gas diffusion electrode thickness 3?10-4 m SEM       GDE Porosity 0.5 - assumed   Hydrophilic pore fraction of      0.65  - assumed    Global cathode mass transfer coefficient  10-3 m s-1 fitted     Limiting current density of O2 cathode  15000 A m-2 calc. Separator & electrolyte     ?e298 K  Ionic conductivity of electrolyte at 298 K 32.7 mho m-1 [9] ?eeff/sep Effective ionic conductivity of separator 12.1 mho m-1  ?d Diaphragm porosity 0.66 -  ?sep Separator thickness 1.8?10-4 m  Swiss-roll components     ? Rcontact+ Relectronic Total electronic plus contact resistance 4?10-6 Ohm m2 Appx. B.9 * Note: Open circuit potentials of electrodes were measured by an apparatus described in Appendix B.1 using a Hg/HgO reference electrode (MMO). Since in strong alkaline solution (i.e. 2 M NaOH in this work) it is not common to convert potentials to VRHE, the open circuit potentials and subsequently overpotentials were reported as VMMO.   124  4.8 Model Predictions In this section the effect of important component variables of the Swiss-roll i.e. anode, and cathode on performance of SR-MRFC were simulated and are briefly discussed. An important variable in this model is the hydrophilic pore volume fraction in the cathode GDE, ?, (Section 4.6). Figure  4-1 shows that the variable ? has a significant influence on the SR-MRFC polarization. As ? increases from 40% to 85%, the Ohmic potential drop in the cathode GDE decreases, improving tremendously the performance, e.g. at 0.4 V for ? =40% the current density is about 2000 A m-2, whereas at ? =85% the current density is 8125 A m-2 with peak power density of 3300 W m-2. At current densities higher than 8000 A m-2 the O2 mass transfer limitation become apparent.  125   Figure ?4-1 Simulation of SR-MRFC performance: the effect of cathode hydrophilic pore volume simulation; ?=0.8, other simulation condition as in Table ?4-1. Increasing the cathode hydrophilic pore volume fraction above 85% is not beneficial due to the high likelihood of flooding by the alkaline borohydride solution that would diminish dramatically the O2 mass transfer rate and compromise the electrocatalytic selectivity. On the other hand, the ACL losses are relatively small compared to equivalent losses in GDE (Figure  4-2). The latter, shown in Figure  4-2, agrees with the polarization simulations performed by Shah et al. [141]. Unlike methanol, borohydride electrooxidation does not result in a gaseous product. The gas fraction in the ACL is a result of borohydride hydrolysis. Therefore electrocatalysts with low and/or zero hydrolysis activity like Os(0) can potentially improve the performance by lowering the ionic Ohmic drop in the ACL.  126   Figure ?4-2 Simulation of SR-MRFC performance with different ??(anode?hydrophilic pore volume due to the H2(g) evolution by BH4- hydrolysis);?=65%; other simulation conditions as in Table ?4-1. To better understand the polarization behavior of the cell equipped with a Viledon? separator, model predictions were made using the variables and parameters presented in Table 4-1 and are compared to experimental data presented in Chapter 3. Comparing the experimental results with the model predictions a very good fit is obtained for ?=0.8 and ? = 65% (Figure  4-3).   127   Figure ?4-3 Simulation of SR-MRFC performance and comparison with experimental?data:?simulation??=65% and??=0.8, other parameters as in Table ?4-1; experimental condition as Figure ?3-12 A with Viledon? as separator. 4.9 Summary To gain insights on the SR-MRFC, a simple mathematical steady-state, isobaric, isothermal, and fixed reactant composition model was developed which takes into account kinetics, mass transfer and Ohmic losses in the reactor based on the mixed-potential theory.  The simulated results confirm that the ionic Ohmic potential losses represent a significant portion of the total potential losses, which is consistent with previous modeling reports [141] and experimental observations for the conventional PEM architecture with either O2 or H2O2 oxidants [135,151,152]. Modeling simulations suggest that proper engineering of the cathode gas diffusion electrode to provide an optimum hydrophilic-hydrophobic balance is important to the development of a high performance SR-MRFC. These results demonstrate that further improving  128  the cathode GDE should focus on i) increasing the hydrophilic pore volume fraction to about 85%, and ii) enhancing the selectivity for the ORR. Moreover, increasing the liquid volume fraction in the ACL (by suppressing H2(g) formation due to BH4- hydrolysis; for instance by employing Os(0) anode) can further improve the performance of the SR-MRFC.  It is important to note that the many variables of the present model may interact with each other. A sensitivity analysis of interacting major variables such as the anode specific surface, separator thickness, BH4- concentration, temperature, O2 pressure, NaOH concentration in a statistical experimental design would be essential to further develop the present model for understanding the complex interactive nature of the variables in increasing the performance of the Swiss-roll mixed-reactant fuel cell.  129  Chapter 5: Single-Cell Swiss-Roll with N2O Oxidant* 5.1 Introduction In Chapter 1 nitrous oxide (N2O) was identified as a potential gaseous oxidant in mixed-reactant fuel cells. This observation was made on the basis of the poor homogeneous reactivity of N2O with typical fuels, its the relatively low solubility in water (2.43 ? 10-2 M) and its high equilibrium reduction potential to N2 [145]:                                   E?298 K = 0.94 VSHE (5-1)                                  E?298 K = 1.76 VSHE (5-2) The first two factors are important in the MRFC to prevent the consumption of fuel in non-Faradaic reactions and to constrain access of N2O to the hydrophilic anode, where it can engage in mixed-potential electrode processes causing a loss of energy efficiency. The high reduction potential presents the tantalizing prospect of an alkaline borohydride fuel cell with theoretical open circuit voltage around 2 V.                                                    * A version of Chapter 5 is published as:  1. A.Aziznia, A.Bonakdarpour, E.L.Gyenge, C.W.Oloman, ?Electroreduction of nitrous oxide on platinum and palladium: toward selective catalysts for methanol-nitrous oxide mixed-reactant fuel cells?, Electrochimica Acta 56 (2011) 5238?5244.   130  Apart from these considerations, N2O is interesting because it is a potent greenhouse gas, so reducing its emission levels while generating electricity in a fuel cell is a desirable objective. Nitrous oxide is generated in large quantities as a by-product of industrial processes such as production of Nylon and nitric acid. For instance N2O emissions from nitric acid production worldwide have been estimated at about 400,000 tones, equivalent to 120 million tones of CO2 annually, making nitric acid manufacturers the largest single industrial-source of N2O [153]. Currently a number of N2O abatement technologies such as the BASF N2O catalytic destruction, Norsk Hydro Agri thermal destruction and Uhde EnviNOx? are under development, but these high temperature (600-1000 K) processes are energy intensive and require complicated thermo-chemical plants which face many technological and economical challenges [154]. Finally, the use of N2O as a fuel cell oxidant is a novel concept whose investigation may open new avenues for fuel cell technology. As with most fuel cells, electrode kinetics is a critical issue for N2O reduction. The utility of N2O in a mixed-reactant fuel cell will depend primarily on the intrinsic kinetics of N2O reduction at the cathode in the presence of a fuel and of fuel oxidation at the anode in the presence of N2O. This Chapter reports work on the selection of electrocatalysts and their subsequent application in the SR-MRFC for the alkaline methanol/N2O and alkaline borohydride /N2O systems.  Details of experimental protocol for this Chapter are presented in Appendix B.2. All the measured electrode potentials in Chapter 5 are referenced versus the reversible hydrogen electrode (RHE).  131  5.2 Nitrous Oxide Reduction on Pt  Figure  5-1 shows the cyclic voltammograms (CV) of the polycrystalline Pt electrode under N2O(g) purged and N2(g) purged (inset) conditions in 0.1 M NaOH at 295 K. The N2O reduction current density is independent of sweep direction and the onset potential of N2O reduction is about 0.5 V. The reduction of N2O occurs only in the potential region where hydrogen is under-potentially adsorbed on the surface of the Pt (Figure  5-1 inset). The reduction peak currents occur at a potential close to the peak potential for adsorbed hydrogen formation (Hads) which is about 0.2 V. These observations confirm that the Hads is essential in the electroreduction of N2O on Pt.  Figure ?5-1 Cyclic voltammetry of polycrystalline Pt electrode (a) N2O, 0.01 V s-1, 1600 rpm 0.1 M NaOH, 295 K, 101 kPa (abs); (inset) same measurement with N2. The arrows indicate sweep directions  132  Figure  5-2 A shows the impact of rotation speed on the N2O cathodic scans. Figure  5-2 B presents the reduction peak current density (jp) versus the electrode rotation speeds. For electrode rotation speed of less than about 900 rpm jp increases, whereas for rotation speed greater than 900 rpm jp decreases, possibly due to ejection of weakly adsorbed N2O from the Pt surface.   Figure ?5-2 (A) Cathodic wave of CVs for reduction of N2O on polycrystalline Pt electrode (B) jp vs. rotation speeds; 0.010 V s-1, 0.1 M NaOH, 295 K, 101 kPa (abs)  Figure  5-3 shows the cathodic reduction currents of N2O for various potential sweep rates (?) at 0 rpm. The peak potential of N2O reduction, designated hereafter as Ep(irrev), occurs in the  133  Hads region and shifts to lower potentials as ? increases. The reduction peaks at higher potentials (0.6-0.8 V) are due to reduction of Pt surface oxides and hydroxides.  Figure ?5-3 Cathodic sweep rate variation of polycrystalline Pt cyclic voltammetry for reduction of N2O, 0 rpm, 0.1 M NaOH, 295 K, 101 kPa (abs)  Under mixed kinetic and diffusion control, the peak potential of an irreversible cathodic wave shows a logarithmic dependence on the sweep rate [155,156]: ?????? ??????????? ?log21loglog2152.0 00)( kDbbEE irrevp     (5-3) The Tafel slope, b, is given by:   FRTb ?3.2?          (5-4)  134  The peak current density for an irreversible cathodic wave is expressed by [155,156]:                                    (5-5) The variables and their respective units in Eq. (5-3) to (5-5) are given in the Nomenclature. The temperature dependence of E0 is negligible between 295 and 333 K (ca. -1?10-3 V K-1). The diffusion coefficient of N2O in 0.1 M NaOH,     , was assumed to be same as that of N2O in water i.e., 1.8?10-9 m2 s-1 at 295 K [157] and its temperature dependence between 295 and 333 K was estimated using the Stokes-Einstein equation. The N2O bulk concentration,      , was assumed to be equal to the solubility of N2O in water which is 2.43 ? 10-2 M at 295 K and 101 kPa(abs) [158]. The temperature dependence of the N2O solubility in water was also taken into account using the data from Ref. [158].  A plot of Ep(irrev) vs. log ? gives the Tafel slope, and the intercept yields the standard heterogeneous rate constant, k0, (Figure  5-4 A). The number of electrons exchanged can be estimated from the jp vs. v1/2 dependence (Figure  5-4 B).  135   Figure ?5-4 Sweep rate dependence of the peak potential Ep (A) and current jp (B) for the reduction of N2O on Pt in 0.1 M NaOH, 295 K, 101 kPa (abs)  As presented in Table  5-1, the Tafel slopes showed an increasing trend from 0.111?0.019 V dec-1 to 0.150?0.014 V dec-1 when temperature increased from 295 K to 333 K, while the electron transfer coefficient remained nearly constant.    136  Table ?5-1 Temperature dependence of the N2O reduction kinetics parameters; polycrystalline Pt, 0 rpm, 0.1 M NaOH, 101 kPa(abs)  T (K)  295 305 323 333 ?b? / V dec-1 0.111?0.019 0.131?0.028 0. 137?0.018 0.150?0.014 ? 0.5 0.5 0.5 0.4 k? / m s-1 [1.7?1.1]?10-10 [3.7?1.5]?10-8 [1.5?1.7]?10-7 [5.7?2.3]?10-7 The temperature dependence of the standard reaction rate constant k0 follows the Arrhenius Equation (5-6). From the Arrhenius plot of k0 shown in Figure  5-5, an activation energy of 107?17 kJ mol-1 is obtained for electrochemical reduction of N2O on Pt. RTEaAek ??0           (5-6)  Figure ?5-5 Temperature dependence of the heterogeneous rate constant for the reduction of N2O on Pt in 0.1 M NaOH at 101 kPa (abs) Regarding the number of electrons exchanged, for low sweep rates, between 0.005 V s-1 and 0.05 V s-1 (Figure  5-4 B), applying Eq. (5-5) the number of the electrons transferred in  137  electroreduction of N2O on the Pt surface was 0.8 and 1.3 for temperatures of 295 K and 333 K, respectively. Thus, at low sweep rates the N2O reduction on Pt is a one-electron process. However, for sweep rates higher than 0.050 V s-1 (Figure  5-4 B) the number of electrons exchanged was 0.2 and 0.6, for 295 K and 333 K, respectively. This indicates that the peak current is no longer controlled by diffusion of electroactive species to the surface, hence, the high sweep rate result cannot be explained by a purely electron transfer rate determining step. Based on the experiments performed here, a mechanism for N2O electroreduction on Pt is proposed: As shown by Figure  5-3, scanning in the cathodic direction the reduction of N2O starts only after the Pt surface hydroxide reduction wave has been completed, forming fresh Pt surface sites (5-7). The Pt surface hydroxide reduction wave widens with increasing scan rate extending between 0.8 and 0.4 V at 0.5 V s-1 (Figure  5-3).  ?? ???? OHPteOHPt 22         (5-7) The broad Pt surface hydroxide reduction wave has been explained by the participation of both strongly and weakly adsorbed OHads species [159]. The relative rate of formation for the two types of OHads species depends on the state of the Pt surface (e.g., pre-reduced, oxidized, etc.) and its crystallographic features. Regarding N2O, the first step is charge transfer associated adsorption (sometimes referred to as electrosorption):  138  ?? ?? adsONeON 22          (5-8) The electrochemical step (5-8) is followed by the surface reaction between N2O?ads and Hads, the latter formed by underpotential deposition: ?? ??? OHgNHON adsads )(22        (5-9) Reactions (5-8) and (5-9) can explain the trends observed both in static and rotating disk electrode voltammetry. On a static electrode at scan rates higher than 0.050 V s-1 the reduction peak is no longer completely controlled by the N2O diffusion but the surface reaction (5-9) becomes a limiting step as well. Thus, the calculated number of electrons decreases from one to as low as 0.2 at 295 K. On the rotating electrode, the decreasing trend of peak current density with electrode rotation rate (Figure  5-2 B) can be explained by considering that in Eq. (5-8) N2O must be adsorbed on the surface for the reduction to occur, and the weakly adsorbed N2O is ejected from the surface at high rotation rates. 5.3 Effect of Methanol Oxidation on N2O Reduction on Pt The effect of methanol oxidation on the electroreduction of N2O was also investigated. In fuel cells methanol crossover to the cathode by diffusion and electro-osmotic drag generates a mixed cathode potential. This problem is even more significant for mixed-reactant fuel cells since both fuel (methanol) and oxidant (N2O) are simultaneously exposed to the electrocatalyst therefore, a high catalytic selectivity is desired.  139   Figure ?5-6 Cyclic voltammetry for reduction of N2O on polycrystalline Pt electrode in the presence of 0.5 and 1 M methanol; 0.01 V s-1, 1600 rpm, 0.1 M NaOH, 295 K, 101 kPa (abs); (inset) same. The arrows indicate sweep directions  Figure  5-6 shows the cyclic voltammograms of polycrystalline Pt electrode with N2O in 0.1 M NaOH containing 0.5 and 1 M methanol. On the anodic portion, the onset potentials for methanol oxidation in both 0.5 and 1 M solutions coincide with the beginning of OHads formation on polycrystalline Pt. The methanol oxidation current increases almost linearly over the potential region of 0.6 to 1.1 V for both concentrations. The increase in the oxidation current is followed by an abrupt decrease upon reaching the peak current at the potentials corresponding to the irreversible Pt oxide formation [160]. The methanol oxidation peak current increased proportionally with methanol concentration (Figure  5-6). The presence of N2O had virtually no  140  effect on methanol oxidation when compared to scans identical to those shown by Figure  5-6 but recorded in the absence of N2O. On the cathodic scan, as can be seen in Figure  5-6 inset, the presence of methanol adversely affects the N2O reduction in the potential range of 0 to 0.4 V. Due to the adsorbed intermediates produced during the oxidation of methanol, specific active sites on Pt related to the N2O reduction are poisoned. Also, in the potential region of the 0.2 V to 0.6 V, the reduction current density was higher in the 1 M methanol than in the 0.5 M methanol containing electrolyte. The higher reduction current in this region for 1 M methanol solution can be attributed to the reduction of some intermediate Pt-(CHO)ads on the Pt surface [161].  Based on Figure  5-6, Pt shows good selectivity toward methanol oxidation in the N2O-methanol mixture and it could function as a component of the anode catalyst for the mixed-reactant fuel cell. Therefore, a cathode catalyst is required with selectivity toward N2O reduction in the N2O ? methanol mixed system. 5.4 Nitrous Oxide Reduction on Pd Figure  5-7 A shows the N2O reduction behavior on the polycrystalline Pd electrode in 0.1 M NaOH. The Pd voltammogram in 0.1 M NaOH in the absence of N2O (Figure  5-7 A inset) has a characteristic reduction wave starting at about 0.87 V where surface PdO reduces to Pd. It is evident from Figure  5-7 A that N2O electroreduction on Pd begins at a potential of about 0.70 V, which coincides with the PdO reduction plateau (Figure  5-7 A inset). Thus, similarly to Pt, only the oxide/hydroxide ?free? surface is catalytic toward N2O reduction. However, an important  141  difference between the Pd and Pt electrodes is the +0.2 V more positive N2O onset reduction potential on the former, which points toward kinetically more facile electroreduction on Pd.  Moreover, contrary to Pt, with Pd on the cathodic scan there are no distinct peaks associated with underpotential Hads formation, due to the high affinity of Pd for hydrogen intercalation (Figure  5-7 A, inset). It was shown by Martin and Lasia [162] that the cathodic and anodic peak currents associated with underpotential hydrogen adsorption and desorption on Pd diminished greatly with cycling to the point of not being detectable after about ten cycles. They attributed this to surface contamination from impurities present in 0.1 M NaOH [162]. This result raises an interesting question whether on Pd the same N2O electroreduction mechanism takes place as on Pt, involving the underpotentially formed Hads (Eq. 5-9). Based on Figure  5-7 A and Figure  5-2, it can be clearly seen the N2O reduction voltammograms are very different on Pd compared to Pt. On Pd there was no N2O diffusion controlled peak. Furthermore, as shown in Figure  5-7 A, increasing the rotation rate of the Pd electrode from 900 rpm to 3600 rpm had little impact on the N2O reduction current density, suggesting that neither mass transfer of N2O nor ejection of weakly adsorbed N2O (as observed for the case of Pt in Figure  5-2) play a role here.  Therefore, it is proposed that on Pd surface the following mechanism prevails: adsONON 22 ?           (5-10) ?? ???? OHgNeOHON ads 2)(2 222        (5-11)  142  N2O is strongly adsorbed on Pd as compared to Pt. Furthermore, the fact that Hads does not play a key role allows the reduction onset potential on Pd to be more positive in accordance with reaction (5-11). As depicted in Figure  5-7 B, the Tafel plot of N2O reduction on polycrystalline Pd is linear up to -0.40 V. At 295 K the values of Tafel slope, charge coefficient transfer (?) and exchange current density (j0) are: 0.084?0.007 V dec-1, 0.7 and [5.3?2.5]?10-5 A m-2, respectively.    143   Figure ?5-7 (A) Impact of rotation rate on reduction of N2O on Pd at 0.010 Vs-1 in 0.1 M NaOH; Inset: Pd in 0.1 M NaOH at 0 rpm (B) Tafel plot for N2O reduction on Pd generated from the 1600 rpm data, 295 K, 101 kPa (abs) Figure  5-8 A presents the temperature dependence of the N2O reduction current density on Pd in the range 295 K to 333 K. At potentials lower than about 0.4 V increasing the electrolyte temperature from 318 to 333 K significantly decreased the reduction current densities (Figure  5-8). At low cathodic potentials the N2O adsorption step could be rate limiting (Eq. 5-10). Therefore, the temperature effect can be explained in light of Eq. (5-10) by two phenomena that  144  lead to lower surface coverage by N2Oads. First, the higher the temperature the lower is the solubility of N2O in water. The reported solubilities of N2O in water at 295 K, 313 K and 333K at 101 kPa (abs) are respectively: 2.43 ? 10-2 M, 1.69 ? 10-2 M and 0.74?10-2 M, [158]. That means the solubility of N2O decreases by about 70 % from 295 to 333 K. Second, is the effect of temperature on the adsorption-desorption equilibrium due to the entropy factor of chemisorption (Eq. (5-10)) [163]. Since in the vast majority of cases adsorption is exothermic, the adsorption equilibrium constant decreases with an increase of temperature, lowering the N2Oads surface coverage. These two phenomena result in a corresponding drop in current density from about 270 to 90 A m-2 at 0.1 V (Figure  5-8 A). To determine the activation energy of the N2O reduction on polycrystalline Pd the data at 0.65 V was used (Figure 5-8 B) and fitted to Eq. (5-12) [164]:  ? ?RETj aUk3.21log??????????????????????         (5-12) Where jk is the current density in the kinetic region and U is the applied potential.  Figure  5-8  B presents the linear variation of log(jk) vs. 1/T. The apparent activation energy at 0.65 V was 27?8 kJ mol-1, almost four times lower than the value on Pt.  145   Figure ?5-8 (A) Impact of temperature on N2O reduction on Pd (B) Temperature dependence of the N2O reduction kinetic current density at 0.65 V. Polycrystalline Pd, 1600 rpm, 0.1 NaOH, 101 kPa (abs), sweep rate: 0.010 V s-1 5.5 Effect of Methanol on N2O Reduction on Pd Figure  5-9 shows the impact of methanol concentration on the reduction of N2O on Pd at 295 K. The electrooxidation of methanol on polycrystalline Pd in the absence of N2O results in two current peaks, one during the cathodic and other in the anodic sweeps (Figure  5-9 inset). During the anodic sweep, the oxidation peak starts around 0.8 V and corresponds to the oxidation of  146  chemisorbed species. The sharp oxidation peak in the cathodic sweep occurs at 0.7 V where reduction of PdO to Pd initiates. This sharp peak can be associated with removal of carbonaceous species such as Pd-COOHads, Pd-(CHO)ads which are not completely oxidized during the anodic scan. With 0.5 and 1.0 M methanol in the electrolyte the onset potential for N2O reduction shifts from 0.62 V to 0.42 and 0.52 V, respectively, indicating somewhat lower activity of the Pd electrode in the presence of methanol. However, in contrast to the Pt, methanol did not dramatically affect the N2O electroreduction and N2O can be effectively reduced in the presence of methanol (Figure  5-9). The methanol oxidation peaks in 1 M methanol are greatly diminished in the presence of N2O (compare Figure  5-9 and the inset). This shows that N2Oads on Pd blocks the chemisorption and oxidation of methanol.  147   Figure ?5-9 Effect of methanol concentration on reduction of nitrous oxide on polycrystalline Pd, 1600 rpm, 0.1 M NaOH, sweep rate: 0.010 V s-1, 295 K, 101 kPa (abs) Inset: 1 M methanol in 0.1 M NaOH with N2; other conditions are the same Based on the data presented here, Pt and Pd were chosen respectively for anode and cathode of the SR-MRFC using N2O as an oxidant (Section 5.6 and 5.7). Furthermore for the mixed MeOH-N2O system, a PtRu anode catalyst was also investigated to lower the methanol oxidation overpotential. It can be expected that PtRu alloy maintains the same good selectivity as Pt in suppressing the N2O reduction. 5.6 The Swiss-Roll Mixed Methanol-N2O Fuel Cell The Swiss-roll MRFC tests were performed using a 3D Pt or PtRu anode and Pd gas-diffusion cathode. The Pt and PtRu anodes were manufactured in-house by a standard technique  148  described in Appendix B.4. The commercial Teflonated Pd gas-diffusion cathode (Pd loading of 2 mg cm-2) was obtained from Gaskatel GmbH and used as is. The apparatus and other testing details are described in Chapter 2. The mixed-reactant fuel cell experiments with N2O as the oxidant were carried out over the temperature range of 295 to 333 K. Figure 5-10 shows the performance of the Swiss-roll mixed CH3OH?N2O alkaline fuel cell with a Pt anode and Pd cathode at 298 and 333 K. Due to the improved kinetics of methanol oxidation and N2O reduction, temperature has a positive effect on the performance, manifested in increasing the peak power density from 12 W m-2 to 22 W m-2 at 298 and 333 K, respectively. The open circuit voltage (OCV) is ca. 0.41 V while the expected theoretical equilibrium cell potential for a methanol-nitrous oxide system is ca. 1.85 V at 298 K.  To improve polarization performance of the methanol-N2O fuel cell system toward more practical voltages, a PtRu anode catalyst, which is well-known to have lower methanol oxidation overpotential [32,34,93,94], was investigated. Comparing Pt and PtRu anode electrocatalysts coupled with a Pd cathode (Figure 5-10 B), PtRu gave superior results over the entire polarization curve, with a peak power density of 38 W m?2 at 322 K, whereas under the same conditions Pt generated only 22 W m?2. The OCV of the mixed methanol-N2O fuel cell with a PtRu anode is ca. 0.5 V, about 0.1 V higher than for the Pt anode at similar conditions. The positive effect of Ru on OCV of DMFCs has been well explained in the literature on the basis of the Langmuir?Hinshelwood rate determining step model for methanol oxidation involving the surface reaction between two adsorbed species, i.e. methanol dehydrogenation products (mostly COad [95]) and hydroxyl radicals (OHad) [32,34,93,94].  149    Figure ?5-10 Performance of the MeOH-N2O SR-MRFC (A) Effect of tempreture, Pt anode (B) Comparison of PtRu (Pt:Ru 1:1) and Pt (loading 0.8 mg cm-2) cathode: Pd GDE (2 mg cm-2) at 323 K; Feed: 3 mL/min 0.5 M MeOH+2M NaOH, oxidant: N2O 10 SLPM. P=105 kPa(abs) Concerning the fuel cell temporal stability, Figure  5-11 depicts the performance of the methanol-N2O SR-MRFC over 3 hours continuous operation at a constant current density of 100 A m-2 and shows a voltage drop of about 20?10-3 V hr-1. This high rate of degeneration is  150  probably due to wetting of the cathode by the methanol solution since the analogous BH4-/N2O system described next showed a voltage drop of only about 2?10-3 V hr-1 at 400 A m-2 (Figure 5-13).   Figure ?5-11 Galvanostatic stability of the SR-MRFC MeOH-N2O with PtRu anode at 100 A m-2; cathode: Pd GDE (2 mg cm-2); feed: 3 mL/min 0.5 M MeOH+2M NaOH, oxidant: N2O 10 SLPM. T=323 K, P=105 kPa(abs) As mentioned previously, the commercial gas-diffusion cathode is PTFE-treated for hydrophobicity to suppress cathode flooding by the fuel solution. However, in the mixed-reactant system, liquid fuel may undesirably enter the porous GDE by convective and capillary effects. The capillary action of methanol solution is beneficial in direct methanol fuel cells for fuel delivery to the hydrophilic anode [165]. However, since methanol tends to wet PTFE, the performance of the cathode is compromised by uptake of the fuel solution, which both inhibits oxidant transport and sets up a mixed-potential in the GDE. In this respect, to improve the  151  performance of the liquid(fuel)/gas(oxidant) mixed-reactant system consideration should be given to: (a) selecting fuels for a high capillary pressure in the anode and low capillary pressure in the cathode (b) controlling the capillary action of fuel and oxidant by engineering the contact (wetting) angles, pore dimensions, porosity and thickness of the electrodes. 5.7 The Swiss-Roll Mixed Borohydride-N2O Fuel Cell Figure  5-12 shows the performance of the mixed NaBH4-N2O SR-MRFC with a Pt anode and Pd cathode at 323 K. The OCV is ca. 0.6 V, which is about 0.2 V higher than for the methanol-N2O system. A peak power density of 340 W m-2 was achieved which is over ten times higher than the peak power density of the methanol-N2O SR-MRFC with Pt-Pd anode-cathode electrocatalysts.    152   Figure ?5-12 Performance of the NaBH4-N2O SR-MRFC; Pt anode (0.8 mg cm-2) and Pd cathode (2 mg cm-2) GDE, feed: 3 mL min-1 1 M NaBH4+2M NaOH, oxidant: N2O 10 SLPM P=105 kPa(abs). Concerning the fuel cell temporal stability, Figure  5-13 depicts the performance of the borohydride-N2O SR-MRFC over 3 hours continuous operation at a constant current density of 400 A m-2. It was shown in Figure  5-11, the stability of the methanol-N2O over 3 hr operation at 100 A m-2 is poor and a voltage drop of more than 20 mV hr-1 was observed. On the other hand, the voltage loss of the borohydride-N2O (Figure  5-13) at the constant current density of 400 A m-2 was equal to or less than 2 mV hr-1. Thus, the stability of the BH4--N2O system is much better than the CH3OH-N2O SR-MRFC.  153   Figure ?5-13 Galvanostatic stability of the SR-MRFC NaBH4-N2O at 400 A m-2; cathode: Pt anode (0.8 mg cm-2), Pd GDE (2 mg cm-2). 3 mL/min 1 M NaBH4+2M NaOH, oxidant: N2O 10 SLPM. T=323 K, P=105 kPa(abs) other condition as Figure ?5-12 5.8 Summary The kinetics of N2O electroreduction in the absence and presence of methanol was studied by static and rotating electrode voltammetry between 295 and 333 K on polycrystalline Pt and Pd electrodes in 0.1 M NaOH. In the absence of methanol, the reduction of N2O on Pd is more facile than on Pt as shown by the approximately four times lower apparent activation energy and lower Tafel slope (Pt: 0.111?0.019 V dec?1 , Pd: 0.084?0.007 V dec?1 at 295 K). Two different electroreduction mechanisms are proposed for Pt and Pd with and without participation of under-potential deposited hydrogen, respectively. The selectivity of Pt and Pd electrodes toward both N2O electroreduction and methanol (0.5 and 1 M) oxidation at 295 K was also investigated. Platinum based electrocatalysts are promising candidates for the anode of a  154  mixed-reactant CH3OH?N2O fuel cell due to inhibition of N2O reduction by chemisorbed methanol. Pd on the other hand is a selective cathode electrocatalyst since N2O reduction takes place fairly actively in the presence of methanol, while methanol oxidation is inhibited. The Swiss-roll mixed-reactant fuel cell was operated using a two-phase mixture of alkaline methanol and/or sodium borohydride solution, as fuel, and N2O gas as the oxidant. For mixed CH3OH-N2O system, a significantly lower peak power density (38 W m-2) was achieved at 323 K and 105kPa(abs) with a PtRu (anode) and Pd (cathode) and showed a voltage drop of about 20?10-3 V hr-1. This high rate of degeneration is probably due to wetting of the cathode by the methanol solution since the analogous BH4-/N2O system showed a voltage drop of only about 2?10-3 V hr-1 at 400 A m-2 and 323 K. For the mixed NaBH4/N2O system, a peak power density of 340 W m-2 was achieved at 323 K and 105kPa(abs) with a Pt (anode) and Pd (cathode).  155  Chapter 6: Scale-Up and Bipolar Operation of the Swiss-Roll Cell  6.1 Introduction Developments in electrocatalysts, electrode engineering, and the innovative Swiss-roll design presented in the previous chapters have brought dramatic improvements in the performance of a small-scale mixed-reactant DBFC. Power densities of about 2000 W m-2 were achieved with platinum- and membrane-free single-cell alkaline Swiss-roll mixed-reactant DBFCs using a geometric electrode area of 20 ?10-4 m2, for a maximum power output of 4 W. This Chapter describes the scale-up of a single?cell reactor and preliminary work on the development of a multi-cell bipolar reactor based on the Swiss-roll configuration. 6.2 Single-Cell Scale-Up from 20?10-4 m2 to 200?10-4 m2 Scale-up is an important issue for the potential commercialization of the fuel cell systems but there is little information available on it for MRFCs. Scaling up the SR-MRFC leads to a better understanding of the various interacting factors such as the geometry and properties of cell components, the 2-phase flow regime, mass transport, kinetics and reactant distribution in the reactor.  The single-cell Swiss-roll MRFC scale-up investigation was carried out with three electrode geometrical areas (described in Chapter 2) in Reactors A and B and shown in Figure  6-1.  156   Figure ?6-1 Geometric aspect ratios of the three electrodes Figure  6-2 compares the performance data for the three fuel cells operating at 298 K under similar conditions. It is apparent that although the OCVs of the ?medium? and ?large? cells are comparable to that of the ?small?- cell (ca. 0.86 V), the performance of the larger cells normalized per electrode geometrical area is substantially inferior to that of the ?small? cell (Figure 6-2 B). The peak power densities for the small-, medium- and large cells were respectively ca. 2000 W m-2, 900 W m-2, and 700 W m-2. The most significant performance drop in the scale-up was seen where the electrode dimensions increased from 0.02?0.10 m to 0.10?0.10 m (Figure  6-1). An increase in height of the electrode (from 0.02 m to 0.10 m) showed a more severe negative effect on performance i.e. 2000 W m-2 vs. 900 W m-2, compared to the increase in the length of the electrodes (from 0.10 m to 0.20 m), which resulted in a performance  157  drop from 900 W m-2 to 700 W m-2. A non-uniform axial distribution of reactants, especially the liquid, could explain this significant performance drop on the single cell scale-up performance. Combined with this is the problem of non-uniform current and voltage distribution in either axial or radial directions [7]. Moreover, the SR-MRFC was operated at constant temperature of the feed. The actual operating temperature inside the reactor may be different (and non-uniform) due the Joule heating especially for the large-scale reactor. This non-uniform temperature distribution inside the SR-MRFC may contribute to the scale-up difficulties.  158   Figure ?6-2 Single cell SR-MRFC with three geometrical electrode surface areas (A) Polarization curves (B) Superficial polarization curve;. Feed: 1 M NaBH4, 2 M NaOH: 12 mL/min O2: 10 SLPM, P=105 kPa (abs), anode: Pt, cathode: MnO2 Gaskatel. T=298 K. Reactor A for 20?10-4 m2 and Reactor B with 100?10-4 m2 or 200?10-4 m2 electrode area Elevated operating temperature due to Joule heating and an increase in surface area raises the rate of H2 gas produced by borohydride decomposition over the Pt anode. It can be assumed that some of this hydrogen gas was trapped in the pores of carbon cloth and partially oxidized,  159  but could also cause severe borohydride mass transfer limitations [143]. One possible explanation for poor performance in the larger cells is that the anode could not deal with the volume of H2 gas produced and consequent restriction of the liquid access to the electrode.  A scale-up study performed by Scott et al. [144] on a 204 ?10-4 m2 DMFC showed that the high rate of CO2 gas produced at the anode (as the result of methanol oxidation) severely restricted the liquid fuel transport to the anode [144]. There, flow visualization of the DMFC revealed that gas formation manifested as slug and/or channel flow, depending on the manifold design and anode substrate (Toray carbon paper or woven carbon cloth). Under identical conditions, for the Toray carbon fiber, large gas slugs were formed in the electrode which tended to the surface of the paper and blocked the flowfield paths and methanol delivery channels. In the case of woven carbon cloth, the flow regime was visualized as bubbly, with relatively small gas bubbles showing a tendency to coalesce [144]. An analogous gas evolution flow regime can be expected in the Swiss-roll DBFC with the carbon cloth Pt anode, due to the high activity of Pt to hydrolyze the BH4- to H2 gas. These observations point to the importance of capillary effects and utilization of electrocatalysts with little or no activity toward BH4- hydrolysis (such as Os(0)) in the scale-up and development of the SR-MRFC. The effect of liquid flow rate in a 200?10-4 m2 single cell SR-MRFC is shown in Figure  6-3. The increase in liquid flow rate from 5 to 12 mL min-1 increases the performance. The higher liquid flow rate may serve to reduce the buildup of H2 gas in the anode, create a more uniform BH4- distribution in the reactor, and to lower the borohydride conversion, thus  160  increasing the average concentration of BH4- in the reactor. Further increase of liquid flow rate to 22 mL min-1 did not improve the performance.  Figure ?6-3 Effect of flow rate on a 200?10-4 m2 single-cell SR-MRFC, conditions same as Figure ?6-2  These scale-up results are intriguing and require further investigation through a concerted experimental and theoretical approach with 2-phase flow visualization [144], mass transport, pressure drop, flow dispersion [20] and capillary effects in the SR-MRFC. These studies were beyond the objective of the present work.  161  6.3 Bipolar Operation 6.3.1 Arrangement 1: Multi-Layer Roll Figure  6-4 shows the polarization curve of the Swiss-roll MRFC composed of two cells in series in multi-layer roll arrangement (Figure  2-9, Chapter 2) each cell having a geometrical area of 20?10-4 m2. The open circuit voltage was 1.20 V, while the OCV of a single cell was 0.86 V. The fact that the dual cell OCV is not near two times that of a single cell can be explained by the establishment of shunt currents, which are a common occurrence in multi-polar electrochemical reactors when there is an ionically conductive pathway between cells [166] (e.g. anode of cell 1 connected with the cathode of cell 2). In bipolar fuel cells shunt current lowers both the OCV and the operating voltage. Thus the peak volumetric power densities for single cell and dual cell were respectively 199 kW m-3 and 267 kW m-3.   162   Figure ?6-4 Comparison of single-cell monopolar and dual-cell bipolar operation of the Swiss-roll MRFC; anode (Pt) ? MnO2 cathode; feed: 1 M NaBH4+2 M NaOH (8 mL min-1), O2 (10 SL min-1), 323 K, 105 kPa(abs). Separator: 2 layers of Scimat?. Geometrical surface area of each cell 2?10-4 m2; superficial current density is normalized per area of one cell. Figure  6-5 illustrates a critical feature of the bipolar Swiss-roll MRFC respecting fluid flow through the cells in the multi-layer roll. In this reactor, the bipolar plates of a conventional fuel cell are replaced by a SS mesh 40 (Chapter 2 Section 2.3.1) that form the main passages for flow of the two-phase fuel-oxidant mixture through each cell, while providing electronic contact between adjacent cells. The effectiveness of this design depends on maintaining a high gas to liquid hold up ratio, and consequent low effective ionic conductivity in the bipole current collector/fluid distributors. Ionic conduction in the bipoles allows shunt currents that essentially by-pass the Faradic processes and compromise the reactor performance. This result is clear in  163  Figure  6-5, where the bipolar reactor voltage at 1500 A m-2 goes from about 0.8 to 0.6 V as the gas to liquid volumetric flow ratio drops from around 2000 to 700. Such fluid dynamic effects are important in scaling up the Swiss-roll MRFC.  Figure ?6-5 Effect of liquid flow in a dual-cell bipolar Swiss-roll MRFC; constant current density of 1500 A m-2, liquid flow rate changed every 5 min, oxygen gas flow 10 SLPM, 323 K, 105 kPa(abs), other condition as Figure ?6-4 6.3.2 Arrangement 2: Rolls-in-Series As mentioned in Section  6.3.1, scale-up of the multi-layer roll configuration is limited by severe liquid shunt currents. In the following a rolls-in-series arrangement is described which potentially can overcome the shunt current limitation observed in the multi-layer arrangement. As described in Chapter 2, in the rolls-in-series configuration, the bipolar SR-MRFC is composed of a number of single-cell Swiss-rolls arranged on a common axis (See Figure  2-10, Chapter 2).  164  Figure  6-6 shows the polarization curves of the Swiss-roll MRFC composed of 2, 3 and 5 cells in series for the rolls-in-series configuration. The open circuit voltage for a single cell was 0.86 V while for 2 cells it was 1.5 V. By comparison in the previous multi-layer roll arrangement a dual cell operation gave an OCV of only 1.2 V. The higher OCV of the dual cell in this arrangement shows the smaller effect of shunt currents. The OCV of 3 and 5 cells in the rolls-in-series arrangement was respectively 2.5 V and 3.8 V. This clearly demonstrates that ionic shunt current paths are suppressed in the rolls-in-series arrangement. The peak volumetric power densities in the SR-MRFC for 2, 3 and 5 cells in series with Pt 3D anode-MnO2 gas diffusion cathode were respectively 129 kW m-3, 188 kW m-3, and 205 kW m-3. For 5 cells at a current density higher than 1500 A m-2, significant reduction in the power and a sharp fall in cell voltage is observed, which is in part due to mass transfer limitations associated with fluid distribution in the 5-rolls in series. For calculating the superficial power densities, the total power drawn from the stack is divided by the geometrical surface area of one cell (20 cm2 which is equal for all of the cells) and is divided by number of cells (i.e. 2, 3 or 5) in the stack. The peak superficial power densities in the SR-MRFC for 2, 3 and 5 cells in series were respectively 350, 510, 550 W m-2.  165   Figure ?6-6 Bipolar operation of SR-MRFC for 2, 3 and 5 rolls in series (A) current-voltage curves (B) current-superficial (per one cell) power curve (C) current-volumetric power curve, geometrical electrode surface area of each cell is 2?10-4 m2: 0.02 m?0.10 m, feed: 1 M NaBH4, 2 M NaOH: 12 mL min-1 O2: 10 SLPM, Pinlet=115 kPa (abs), Tfeed=323 K, anode: Pt, cathode: MnO2. To investigate the potential fuel starvation in the 5 rolls-in-series, the effect of liquid flow rate is reported in Figure  6-7. In contrast to multi-layer roll arrangement, increasing the fuel flow rate from 8 mL min-1 to 20 mL min-1 increases the performance of the SR-MRFC but a further increase to 32 mL min-1 does not change the performance. A similar effect of the liquid flow rate was reported by Ponce-de-Leon et al [20], in flow dispersion and mass transport studies on a divided industrial scale filter-press redox flow reactor with 5 bipolar electrodes.  166   Figure ?6-7 Effect of fuel flow rate on 5 cells in series arrangement of SR-MRFC; conditions as in Figure ?6-6 B From these observations it is clear that the performance of the rolls-in-series arrangement is controlled by complex 2-phase fluid dynamic phenomena that govern the distribution of reactants in the system. The promising results with the axial rolls-in-series arrangement warrants further investigations to characterize the reaction environment and to improve the performance of the bipolar SR-MRFC. 6.4 Summary In this Chapter the scale-up of a single?cell reactor and preliminary work on the development of a multi-cell bipolar reactor based on the Swiss-roll configuration was presented.  167  Comparison of the performance data for the three Swiss-roll fuel cells with increasing electrode surface area (20?10-4 m2, 100?10-4 m2 and 200?10-4 m2) revealed that the OCVs of the ?medium? and ?large? cells are comparable to that of the ?small? cell (ca. 0.86 V), but the performance of the larger cells is substantially inferior to that of the small cell. The peak power densities for the small-, medium- and large-scale cells were respectively ca. 2000 W m-2, 900 W m-2, and 700 W m-2. The most significant performance drop in the scale-up was seen when the electrode height increased from 0.02 m to 0.10 m. A non-uniform axial distribution of reactants, especially the liquid, could explain this significant performance drop on the single cell scale-up performance. Further improvements may be investigated by modifying the fluid distributors (e.g. using expanded mesh 6) and reactor feeding system as well as the electrode design. Further, results of two innovative arrangements for bipolar operation of SR-MRFC were introduced: multi-layer roll and rolls-in-series. In multi-layer roll arrangement, the peak volumetric power densities for single cell and dual cell were respectively 199 kW m-3 and 267 kW m-3. It was observed that the multi-layer roll stack is significantly affected by shunt currents. On the other hand, experiments with up to 5 cells in the rolls-in-series arrangement revealed that this arrangement is less affected by establishment of shunt paths. The peak volumetric power densities in the SR-MRFC for 2, 3 and 5 Swiss-rolls in rolls-in-series arrangement with Pt 3D anode-MnO2 gas diffusion cathode were respectively 129 kW m-3, 188 kW m-3, and 205 kW m-3.   168  Chapter 7: Conclusions and Recommendations for Future Work 7.1 The Swiss-Roll Mixed-Reactant Architecture Previous research in the area of mixed-reactant fuel cells (MRFCs) has primarily focused on operating conditions, electrode design and selective catalysts. Few novel reactor designs have been proposed and/or tested. The combination of multi-phase reactants, electrode modifications and selective catalysts, with the employment of a compact cylindrical ?Swiss-roll? reactor represents a new and innovative approach for mixed-reactant systems. The work presented in this thesis explored this novel architecture for the first time in NaBH4-O2, CH3OH-N2O and NaBH4-N2O mixed-systems.  7.2 Proof-of-Concept of the Swiss-Roll for Direct Borohydride MRFCs The performance of the Swiss-roll mixed-reactant fuel cell (SR-MRFC) NaBH4-O2 fuel cell is investigated by a combination of experimental studies and mathematical modeling. Peak power densities of up to 2500 W m?2 were obtained for a Pt-Ag anode-cathode pair at 323 K and 105 kPa(abs) using Viledon? porous diaphragm with encouraging durability that warrants further engineering improvements and optimization for development of a compact, light-weight, device. To the best of Author?s knowledge, the power density of 2500 W m-2 is higher than any low temperature (T < 373 K) mixed-reactant fuel cell so far reported in the literature and is among the highest ever reported for conventional dual chamber and PEM direct borohydride fuel cells.  169  Furthermore, a platinum- and membrane-free mixed-reactant alkaline borohydride solution/oxygen fuel cell with Swiss-roll design using an Os 3D anode and a MnO2 gas?diffusion cathode is presented. The anode was prepared by surfactant-assisted galvanostatic electrodeposition on woven carbon cloth, producing Os-OsO2 clusters (diameter about 50 nm) composed of 4 nm average nanoparticle size. For a single cell SR-MRFC a peak power density of up to 1880 W m-2 was achieved at 323 K and 105 kPa(abs) with an Os loading of 0.32 ?10-2  kg m-2 (0.32 mg cm-2). The performance was identical to a SR-MRFC operated under the same conditions but equipped with a commercial Pt/Vulcan XC-72 anode catalyst, with identical loading, sprayed onto the carbon cloth. Since the price of Os is about one quarter that of Pt the membrane-free and Pt-free features of SR-MRFC could significantly lower the cost of borohydride fuel cells.  Instead of a polymer electrolyte membrane (PEM), the SR-MRFC is employing a hydrophilic porous diaphragm (Viledon?) as an electronic separator. Nafion? membrane is an expensive component of direct fuel cells which typically costs in the range of 700-1300 $ m-2 while Viledon? costs about 3 $ m-2. In addition to this significant cost reduction, Viledon? porous polymer separator offers higher stability and lower Ohmic resistance when compared to the PEM. The desired ionic conductivity of the Viledon? separator relies on all the ions present in the electrolyte, in contrast to the Nafion? that selectively transfers cations only (i.e. Na+). The effective ionic conductivity of 2 M NaOH electrolyte solution in porous separator (42.9 mho m-1 at 323 K) is significantly higher than that of the Nafion? 112 (2.6 mho m-1, 2 M KOH at 299 K [126]).  170  Investigation of the H2 generation by thermocatalytic BH4- hydrolysis on Pt, Os(0), and OsO2 revealed that Os(0) is not catalytic for the reaction even at 323 K, while Pt and OsO2 showed BH4- half-life times at 323 K of about 25 min and 63 min, respectively. The low thermocatalytic activity of the Os-OsO2 anode for BH4- hydrolysis compared to Pt is an advantageous feature here because it reduces operational difficulties and safety concerns related to H2 gas evolution and accumulation in the mixed-reactant fuel cell. The mathematical modeling illustrates the complex interactive effects of many variables on the performance of the SR-MRFC and confirms that the ionic Ohmic potential losses represent a significant portion of the total potential losses in the reactor. The latter finding is consistent with previous modeling reports [141] and experimental observations for the conventional PEM DBFCs architecture with either O2 or H2O2 oxidants [135,151,152]. Modeling simulations suggest that proper engineering of the cathode gas diffusion electrode to provide an optimum hydrophilic-hydrophobic balance is important to the development a low Ohmic resistance and high performance SR-MRFC. The durability of the SR-MRFC was investigated over continuous galvanostatic operation up to 3 hr at 323 K with single-pass feed for different electrodes used in this work and voltage drops of about 5?10-3 V hr-1 (5 mV hr-1) were observed with the MnO2 cathode which is encouraging but not satisfactory for commercial applications. With Ag GDE, the stability result was excellent over 220 min continuous operation with no detectable performance drop (i.e. 100 ?V). While a complete analysis of the SR-MRFC degradation is beyond the objective of this dissertation, it is recommended to constitute the focus of further studies. Ultimately, for  171  commercial application durability has to be demonstrated for 5,000 to 20,000 hours of continuous operation with a degradation rate of only a few ?V h-1. Regarding DBFCs, the product is an alkaline NaBO2 solution. The solubility of NaBO2 in 2 M NaOH at 298 K is about 2 M [167,168]. The recycling of NaBO2 to NaBH4 either thermo or electrochemically [169], is not feasible at present. Unequivocally, if the recycling of NaBO2 becomes a reality it will increase the attractiveness of this technology. Carbonation of the electrolyte and cathode GDE could be an issue for any alkaline fuel cell and battery technology. However, it has been successfully addressed in the literature by stripping CO2 from the air feed [170] and/or modifying the porous electrode structure to create larger pores that are less prone to clogging. The three-dimensional anode used in this work has larger pores compared to conventional gas diffusion electrodes therefore, anode fouling is less likely. 7.3 Scale-Up and Bipolar Operation of the SR-MRFC In single cell scale-up from 20?10-4 m2 to 200?10-4 m2, although the OCVs are comparable (ca. 0.86 V), the SR-MRFC performance is inferior. The geometrical dimensions and aspect ratios of the electrodes have significant effect on the scale-up performance of the SR-MRFC. Non-uniform distribution of reactants, especially the liquid fuel, affects the cell performance and imposes limitations on the design of the electrodes in the scale-up of the SR-MRFC. An increase in geometrical surface area increases the H2 gas produced due to borohydride decomposition over the Pt anode in the SR-MRFC. In addition to safety concerns regarding non-Faradaic hydrogen generation over Pt in the SR-MRFC, the gas produced restricts  172  the mass transport of fuel to the anode in the larger surface area SR-MRFCs. Further scale-up studies are recommended using electrocatalysts with no or limited hydrolysis activity such as Os(0). The proof-of-concept experiments on the two bipolar arrangements emphasize the importance of understanding the two?phase flow transport phenomena in the SR-MRFC. The multi-layer roll arrangement can provide higher volumetric power densities, compared to rolls-in-series configuration, but a challenge in development of the SR-MRFC in this arrangement was found to be the shunt currents. Shunt or bypass currents in liquid electrolytes, are the result of electric potential gradients along common electrolyte paths. The rolls-in-series arrangement, on the other hand, is less affected by shunt currents, but presumably suffers from a non-uniform reactant and electric potential distribution.  Overall the scale-up of the SR-MRFC depends on complex 2-phase flow hydrodynamics of the liquid/gas reactants which affect mass transfer of the reactants to the electrodes, pressure drop, reactant distribution and conversion, and temperature variations in the SR-MRFC caused by Joule heating. 7.4 Nitrous Oxide as an Oxidant for Mixed-Reactant Fuel Cells This work also explored the feasibility of electrochemical reduction of a potent greenhouse gas, N2O, in the cathode of a mixed-reactant fuel cell to generate electricity from unwanted N2O in the tail gases of industrial processes. It was shown that on a polycrystalline Pt surface in an aqueous alkaline electrolyte, the electroreduction of N2O is an irreversible surface  173  reaction involving weakly adsorbed N2O and the underpotential adsorbed hydrogen (Hads). This result suggests that is Hads an essential participant in N2O reduction on Pt. An electrochemical-chemical sequential mechanism was proposed for N2O reduction where the chemical steps involved the participation of Hads. The presence of methanol in the electrolyte adversely affected the activity of Pt for electroreduction of N2O. However, the activity of Pt towards methanol oxidation remains significant in the presence of the N2O, suggesting Pt based catalysts as promising candidates for the anode of a mixed-reactant methanol-N2O fuel cell. For polycrystalline Pd, a different mechanism for N2O reduction was observed involving strongly adsorbed N2O on Pd sites. The electroreduction on Pd does not involve Hads, which affords a more positive N2O reduction onset potential compared to Pt. Nitrous oxide reduction takes place at Pd sites when surface PdO begins to reduce at about 0.8 VRHE. Furthermore, in contrast to the Pt, the presence of methanol did not dramatically affect the activity of the Pd for N2O electroreduction, while methanol oxidation at 1 M concentration was inhibited by N2Oads. These observations imply that Pd or Pd-based catalysts are good candidates for the electroreduction of N2O, that could be active and relatively selective in the cathodes of mixed-reactant methanol-N2O fuel cells.  It was shown, for the first time, that it is feasible to electrochemically reduce N2O in the cathode of a Swiss-roll mixed-reactant fuel cell to generate electricity. As proof-of-principle for this waste-to-energy approach, the SR-MRFC was operated using two-phase mixtures of alkaline sodium borohydride and methanol solutions as fuel, with N2O gas as the oxidant. A peak power density of 340 W m-2 was achieved for a borohydride-N2O system at 323 K and 105kPa(abs)  174  using a Pt (anode) and Pd (cathode) with encouraging short-term stability. Although the power densities reported here are low, this work demonstrates that co-generation of electricity and abatement of N2O may potentially compete with thermochemical processes of N2O capture currently under development. For instance, nitric acid production plants consume large quantities of NH3 part of which could potentially be used in a mixed NH3-N2O fuel cell (with              , Appendix A, Table A-2) for co-generation of electricity and N2O abatement. Finally, it is noteworthy that the electrocatalysts, components and operating conditions of the Swiss-roll mixed-reactant system described in this work are not optimized and further improvement in the cell performance may be possible. 7.5 Recommendations Developments in electrocatalysts, electrode engineering, and the innovative Swiss-roll design presented in this thesis have brought improvements in the performance of a small-scale DBFC. Power densities of about 2000 W m-2 were achieved with platinum- and membrane-free alkaline SR-MRFCs. These power densities, although attractive, are substantially lower than those obtained with conventional H2 fuel cells; approximately 10,000 W m-2 [127]. In the future research and development of the application of SR-MRFC for DBFC and similar direct fed fuel cells, several issues have to be addressed including electrocatalyst and electrode engineering, mass transport, thermal management, gas management and 2-phase flow hydrodynamics, capillary effects, current collection, flow distribution, and energy efficiency.  175  ? In respect to a Pt-free system for the DBFCs, it is recommended to evaluate the SR-MRFC with an Os 3D anode and Ag GDE to achieve a higher performance and stability.  ? Regarding the electrodepsoted Os anodes, it is recommended to evaluate the effect of surface roughness and specific surface area of the electrodeposited Os on the BOR electrocatalytic activity, the SR-MRFC performance, and NaBH4 hydrolysis. To establish a more accurate comparison with Pt/C, more in depth information on the crystallography, surface roughness and specific surface area of the electordepsited Os must be determined. ? Experimental and modeling results in this work highlighted the importance of the cathode GDE engineering. To increase the performance of the SR-MRFC further research should focus on the design of borohydride-tolerant Pt-free GDEs including Fe-aminoantipyrine [171], EuO2 [172], iron phthalocyanine [173] and perovskite-type metal oxide (LaNiO3, LaCoO3) [127] electrocatalysts. ? The scale-up results are intriguing and require further investigations through a concerted experimental and theoretical approach on two-phase flow visualization, mass transport, pressure drop and flow dispersion. Very few computation fluid dynamic studies have been conducted for the understanding of the electrolyte transport in alkaline fuel cells [174]. Understanding the two-phase flow transport phenomena can improve the design of the SR-MRFC. ? It is recommended to use the expanded mesh 6 as bipole fluid distributor in the multi-layer roll design for lowering the shunt effect of this bipolar arrangement  176  ? While a complete analysis of the SR-MRFC degradation was beyond the objective of this dissertation, it is recommended as the focus of further studies. Extensive SR-MRFC stability testing in more severe conditions including several shut-down and start-up, fuel starvations tests and longer operating time is strongly recommended. The investigation of alternative non-precious metal cathode catalysts with superior stability compared to MnO2 is strongly recommended. Ultimately, for commercial application durability has to be demonstrated for 5,000 to 20,000 hrs of continuous operation with a degradation rate of only a few ?V h-1. ? In spite of the fairly good selectivity of both Pt and Pd, the operating cell voltage of the mixed-reactant N2O fuel cells is impractically low. Future investigations should focus on Pt- and Pd-alloy electrocatalysts that would lower the fuel oxidation and N2O reduction overpotential, while providing the same fair selectivity observed for Pt and Pd. On the cathode side the N2O reduction onset potential in the presence of methanol could be increased to 0.7 VRHE, which is the value for Pd in the absence of methanol, possibly by using Pd alloys.  ? It is recommended that the activity and selectivity of Os 3D anode toward N2O reduction is investigated. This is important for the development of a mixed NaBH4-N2O SR-MRFC with Os-Pd anode and cathode electrocatalysts. ? The mixed-reactant Swiss-roll fuel cell design is not limited to either the NaBH4-O2 system, methanol-N2O or to alkaline fuel cells in general. Further research on the SR-MRFC is recommended for a variety of systems in acid media as well, such as  177  methanol-transition metal redox couple (e.g., Fe(III)/Fe(II) [175], hydrogen-transition metal redox couple [176], and alkaline NH3-N2O fuel cells. ? The model presented in this dissertation is highly simplified, and is of limited usefulness with regard to practical cell design and interpretation of results.  Improvement and extension of the model to practical scale conditions with pressure, temperature and composition gradients, followed by use of the model in design and optimization studies is strongly recommended. ? For the present model, the assumption that the electronic and ionic current densities in the 3D electrodes are equal, is a highly simplified approximation. It is well known that in porous electrodes the ionic current is converted to electronic current through the depth of the pore. Therefore, it is recommended that a better description of the electronic and ionic current densities distribution in porous electrodes to be taken into consideration.  ? An important feature that was not included in the experiments of this dissertation is the use of reference electrodes.  It is recommended to place a reference electrode inside the working SR-MRFC. This would allow more meaningful data relevant to identifying the performance limiting electrode, and allowing for more complete characterization of the cells and the electrode reactions including an accurate determination of mixed-potential losses. ? The energy efficiency of the laboratory-scale SR-MRFC in this study is significantly low because external power sources have been used for the feed heater and gas compressor.  Clearly, the energy efficiency of the SR-MRFC must be increased before  178  it can be used as a viable commercial system. To increase the energy efficiency it is recommended that the following key areas to be investigated: (i) design and employment of a nozzle sprayer with lower pressure drop (ii) using lower gas flow rates and (iii) eliminating the need for heater and recycling the heat of reaction for preheating the feed. 7.6 Contribution to Knowledge 7.6.1 An Innovative Mixed-Reactant Fuel Cell Design and Feeding Device Previous research in the area of MRFCs has primarily focused on operating conditions, electrode design and selective catalysts. Few novel MRFC designs have been proposed and/or tested.  ? The combination of multi-phase reactants, electrode modifications and selective catalysts, employed in a compact cylindrical ?Swiss-roll? reactor represents a new and innovative approach for mixed-reactant systems. ? Two innovative bipolar architectures are introduced for the Swiss-roll design with preliminary experiments on 1,2,3 and 5 cell in series. ? A two phase flow feed sprayer nozzle is used for atomizing a liquid/gas mixture into an electrochemical reactor. The addition of sprayer nozzle showed promising improvement in the uniform distribution of reactants to electrodes. Although reactant distribution in multi-phase electrochemical reactors is important, it has received little attention in the literature. To the best of Author?s knowledge, introduction of a sprayer nozzle for  179  atomizing two phase flow of reactants in electrochemical reactors is inventive and has not been addressed in the literature, and therefore is subject of a patent application. 7.6.2 Electrocatalysis Study and Electrode Modification Presented in this thesis, the activity and selectivity of some important catalysts for anodes and cathodes of direct alkaline fuel cells are investigated. ? Certain kinetic parameters for MnO2 GDE cathode for oxygen reduction and borohydride oxidation reaction are determined and reported, which is important in electrochemistry and catalyst technology. ? For the first time, a 3D anode with electrodeposited Os nanoparticulates is used in a mixed-reactant direct borohydride fuel cell and its performance compared to a commercial Pt catalyst. ? The activity of Os, OsO2, and Pt catalysts for borohydride hydrolysis was measured and compared. Such a comparison was not available in the literature. 7.6.3 Electroreduction of N2O A related aspect of this work was an investigation into the electroreduction of a potent green house and ozone depleting gas, nitrous oxide (N2O). For the first time, N2O is electrochemically reduced in the cathode of a mixed-reactant fuel cell.  180  ? For the first time, N2O electroreduction in alkali is investigated over Pt and Pd electrocatalysts with and without presence of MeOH in the electrolyte. Two different mechanism of N2O electroreduction are proposed. ? This investigation provides kinetic information (e.g. Tafel slopes, rate constants, activation energies) of N2O electroreduction which were not available in the literature. Such information would be useful in design of electrochemical processes which could use N2O as an oxidant for the cathodic reaction and possibly for the reduction of this environmentally destructive gas from Earth?s atmosphere by electrochemical processes. ? For the first time, a mixed-reactant fuel cell is operated using two-phase fuel + oxidant mixtures of 1 M NaBH4 / 2M NaOH(aq) + N2O(g) and 1 M MeOH/2 M NaOH(aq) + N2O(g).   181  Bibliography [1] D. M. F. Santos and C. A. C. Sequeira, ?Sodium borohydride as a fuel for the future? Renewable and Sustainable Energy Reviews, vol. 15, no. 8, pp. 3980?4001, Oct. 2011. [2] J.H. Wee, ?Which type of fuel cell is more competitive for portable application: Direct methanol fuel cells or direct borohydride fuel cells?? Journal of Power Sources, vol. 161, no. 1, pp. 1?10, Oct. 2006. [3] K. Jayakumar, S. Pandiyan, N. Rajalakshmi, and K. S. Dhathathreyan, ?Cost-benefit analysis of commercial bipolar plates for PEMFC?s? Journal of Power Sources, vol. 161, no. 1, pp. 454?459, Oct. 2006. [4] D. J. L. Brett and N. P. Brandon, ?Review of materials and characterization methods for polymer electrolyte fuel cell flow-field plates? Journal of Fuel Cell Science and Technology, vol. 4, no. 1, pp. 29?44, Feb. 2007. [5] S. C. Barton, W. Deng, J. W. Gallaway, S. Levendovsky, T. S. Olson, P. Atanassov, M. Sorkin, A. Kaufman, and H. F. Gibbard, ?Mixed-feed direct methanol fuel cell: Materials and design solutions? ECS Transactions, 2006, vol. 1, no. 6, pp. 315?322. [6] P. M. Robertson and N. Ibl, ?Electrolytic recovery of metals from waste waters with the Swiss-roll cell? Journal of Applied Electrochemistry, vol. 7, no. 4, pp. 323?330, Jul. 1977. [7] P. M. Robertson, P. Berg, H. Reimann, K. Schleich, and P. Seiler, ?Application of the ?Swiss-roll? electrolysis cell in vitamin-C production? Journal of The Electrochemical Society, vol. 130, no. 3, pp. 591?596, 1983. [8] P. M. Robertson, ?The variation of current density and electrode potential with electrode resistance and its role in cell design? Electrochimica Acta, vol. 22, no. 4, pp. 411?419, Apr. 1977. [9] C. W. Oloman, Electrochemical Processing for the Pulp & Paper Industry. Romsey, Hants, UK: The Electrochemical Consultancy, 1996. [10] E. L. Gyenge, CHBE577 Fuel Cells & Electrochemical Systems Course Notes, 2012. [11] T. Erdey-Gruz and M. Volmer, Z. Physik. Chem., vol. 150A, p. 203, 1930. [12] T. Erdey-Gruz and H. Wickie, Z. Physik. Chem., vol. 162A, p. 53, 1932. [13] J. A. . Butler, Trans.Faraday Soc., vol. 19, p. 729, 1924. [14] J. A. . Butler, Trans.Faraday Soc., vol. 19, p. 734, 1924. [15] J. A. . Butler, Trans.Faraday Soc., vol. 28, p. 379, 1932. [16] J. Tafel, Z. Physik. Chem., vol. 50, p. 641, 1905. [17] F. Walsh, A First Course in Electrochemical Engineering, Hants, England: The Electrochemical Consultancy, 1993, p. 381. [18] D. R. Lide, Ed., CRC Handbook of Chemistry and Physics, 89th (Inte. Boca Raton, FL: CRC Press/Taylor and Francis, 2009. [19] P. Atkins and J. de Paula, Elements of Physical Chemistry, 6th ed. New York: W. H. Freeman, 2009. [20] C. Ponce-de-Le?n, G. W. Reade, I. Whyte, S. E. Male, and F. C. Walsh, ?Characterization of the reaction environment in a filter-press redox flow reactor? Electrochimica Acta, vol. 52, no. 19, pp. 5815?5823, May 2007.  182  [21] J. R. Selman and C. Tobias, Advances in Chemical Engineering Volume 10, vol. 10, New York: Elsevier Academic Press, 1978. [22] E. L. Gyenge, ?Dimensionless numbers and correlating equations for the analysis of the membrane-gas diffusion electrode assembly in polymer electrolyte fuel cells? Journal of Power Sources, vol. 152, pp. 105?121, Dec. 2005. [23] N. Perez, ?Mixed-Potential Theory? in Electrochemistry and Corrosion Science, Boston: Kluwer Academic Publishers, 2004, pp. 155?1666. [24] K. Scott, ?Industrial Electrochemical Synthesis Processes: Recent Developments in Reactor Design? Developments in Chemical Engineering and Mineral Processing, vol. 1, no. 2?3, pp. 71?117, May 2008. [25] H. Vogt, ?The rate of gas evolution of electrodes: An estimate of the efficiency of gas evolution from the supersaturation of electrolyte adjacent to a gas-evolving electrode? Electrochimica Acta, vol. 29, no. 2, pp. 167?173, Feb. 1984. [26] R. Anderson, L. Zhang, Y. Ding, M. Blanco, X. Bi, and D. P. Wilkinson, ?A critical review of two-phase flow in gas flow channels of proton exchange membrane fuel cells? Journal of Power Sources, vol. 195, no. 15, pp. 4531?4553, Aug. 2010. [27] I. Hodgson and C. Oloman, ?Pressure gradient, liquid hold-up and mass transfer in a graphite fibre bed with cocurrent upward gas-liquid flow? Chemical Engineering Science, vol. 54, no. 23, pp. 5777?5786, May 1999. [28] C. Oloman, ?Trickle bed electrochemical reactors? Journal of The Electrochemical Society, vol. 126, no. 11, p. 1885, Nov. 1979. [29] R. J. Bellows, P. G. Grimes, and M. Zahn, ?Shunt current elimination and device? US Patent US4197169A 1981. [30] C. W. Oloman, ?Mixed-reactant flow-by fuel cells? British Patent GB 2474202, 2012. [31] F. Bidault, D. J. L. Brett, P. H. Middleton, and N. P. Brandon, ?Review of gas diffusion cathodes for alkaline fuel cells? Journal of Power Sources, vol. 187, no. 1, pp. 39?48, Feb. 2009. [32] A. Bauer, E. L. Gyenge, and C. W. Oloman, ?Direct methanol fuel cell with extended reaction zone anode: PtRu and PtRuMo supported on graphite felt? Journal of Power Sources, vol. 167, no. 2, pp. 281?287, May 2007. [33] A. Bauer, E. L. Gyenge, and C. W. Oloman, ?Direct methanol fuel cell with extended reaction zone anode? ECS Transactions, vol. 3, pp. 1271?1277, 2006. [34] A. Bauer, E. L. Gyenge, and C. W. Oloman, ?Electrodeposition of Pt?Ru nanoparticles on fibrous carbon substrates in the presence of nonionic surfactant: Application for methanol oxidation? Electrochimica Acta, vol. 51, no. 25, pp. 5356?5364, Jul. 2006. [35] Z. Liu and L. Hong, ?Electrochemical characterization of the electrooxidation of methanol, ethanol and formic acid on Pt/C and PtRu/C electrodes? Journal of Applied Electrochemistry, vol. 37, no. 4, pp. 505?510, Jan. 2007. [36] J. Ma, N. A. Choudhury, and Y. Sahai, ?A comprehensive review of direct borohydride fuel cells? Renewable and Sustainable Energy Reviews, vol. 14, no. 1, pp. 183?199, Jan. 2010.  183  [37] S. Basri, S. K. Kamarudin, W. R. Daud, and Z. Yaakub, ?Nanocatalyst for direct methanol fuel cell? International Journal of Hydrogen Energy, vol. 35, no. 15, pp. 7957?7970, Aug. 2010. [38] N. V. Rees and R. G. Compton, ?Sustainable energy: A review of formic acid electrochemical fuel cells? Journal of Solid State Electrochemistry, vol. 15, no. 10, pp. 2095?2100, Apr. 2011. [39] E. L. Gyenge, ?Electrocatalytic Oxidation of Methanol, Ethanol and Formic Acid? in PEM Fuel Cell electrocatalysts and Catalyst Layer, J. Zhang, Ed. Springer, 2008. [40] A. Serov and C. Kwak, ?Review of non-platinum anode catalysts for DMFC and PEMFC application? Applied Catalysis B: Environmental, vol. 90, no. 3?4, pp. 313?320, 2009. [41] N. Wagner, M. Schulze, and E. G?lzow, ?Long term investigations of silver cathodes for alkaline fuel cells? Journal of Power Sources, vol. 127, no. 1?2, pp. 264?272, Mar. 2004. [42] A. Verma, A. Jha, and S. Basu, ?Manganese dioxide as a cathode catalyst for a direct alcohol or sodium borohydride fuel cell with a flowing alkaline electrolyte? Journal of Power Sources, vol. 141, no. 1, pp. 30?34, Feb. 2005. [43] A. Kudo and A. Mine, ?Electrocatalysis for N2O reduction on metal electrodes? Journal of Electroanalytical Chemistry, vol. 408, pp. 267?269, 1996. [44] A. Kudo and A. Mine, ?Electrocatalytic reduction of nitrous oxide on metal and oxide electrodes in aqueous solution? Applied Surface Science, vol. 122, pp. 538?542, 1997. [45] A. Ignaszak and E. Gyenge, ?Differential potential pulse deposition of amorphous osmium thin films and electrocatalytic activity for borohydride oxidation in alkaline media? Electrochimica Acta, vol. 95, pp. 268?274, Apr. 2013. [46] V. W. S. Lam, D. C. W. Kannangara, A. Alfantazi, and E. L. Gyenge, ?Electrodeposited osmium three-dimensional anodes for direct borohydride fuel cells? Journal of Power Sources, vol. 212, pp. 57?65, Aug. 2012. [47] V. W. S. Lam and E. L. Gyenge, ?High-performance osmium nanoparticle electrocatalyst for direct borohydride PEM fuel cell anodes? Journal of The Electrochemical Society, vol. 155, no. 11, p. B1155, 2008. [48] M. Atwan, D. O. Northwood, and E. L. Gyenge, ?Evaluation of colloidal Os and Os-Alloys (Os?Sn, Os?Mo and Os?V) for electrocatalysis of methanol and borohydride oxidation? International Journal of Hydrogen Energy, vol. 30, no. 12, pp. 1323?1331, Sep. 2005. [49] M. Chatenet, F. Micoud, I. Roche, E. Chainet, and J. Vondr?k, ?Part 2: ORR kinetics of sodium borohydride direct oxidation and oxygen reduction in sodium hydroxide electrolyte? Electrochimica Acta, vol. 51, no. 25, pp. 5452?5458, Jul. 2006. [50] G. Gruneberg, W. Wicke, and E. Justi, ?Generation of electrical energy? British Patent GB994448, 1962. [51] Goebel, Struck, and Vielstich, ?Fuel Cells: Modern Processes for the Electrochemical Production of Energy English translation by D.J.G. Ives? in Fuel Cells: Modern Processes for the Electrochemical Production of Energy English translation by D.J.G. Ives, New York: Wiley/Inter Science, 1965, pp. 374?376. [52] P. G. Grimes, B. Fielder, and J. Adam, Annual Power Sources Conference, 1961, pp. 29?32.  184  [53] C. Eyraud, J. Lenoir, and M. Gery, ?Fuel cells utilizing the electrochemical properties of absorbates? Compt. rend. Acadi. des Sci. Paris, vol. 252, pp. 1599?1600, 1961. [54] S. R. Schulze, ?Mixed-feed Methanol-Oxygen Fuel Cells Schulze? PhD Thesis, Massachusetts Institute of Technology, May 1967. [55] W. van Gool, ?The possible use of surface migration in fuel cells and heterogenoues catalysis? Philips Res. Repts., vol. 20, pp. 81?93, 1965. [56] G. A. Louis, J. M. Lee, D. L. Maricle, and J. C. Trocciola, ?Solid electrolyte electrochemical cell? U.S. Patent Application 4,248,941, 1981. [57] P. Moseley and D. Williams, ?Sensing reducing gases? Nature, vol. 346, p. 23, 1990. [58] C. K. Dyer, ?A novel thin-film electrochemical device for energy conversion? Nature, vol. 343, no. 6258, pp. 547?548, May 1990. [59] C. K. Dyer, ?Primary source of electrical energy using a mixture of fuel and oxidizer? US Patent US4863813 A, May 1989. [60] C. K. Dyer, ?Compact fuel cell and continuous process for making the cell? US Patent US4988582 A, May 1991. [61] C. K. Dyer, ?Modular fuel cell assembly? US Patent US5094928 A, May 1992. [62] S. Gottesfeld, ?Thin film fuel cells? Nature, vol. 345, p. 673, 1990. [63] P. C. Ellgen, ?Devices providing electrical energy from fuel/oxygen mixtures? US Patent US5162166 A, May 1992. [64] T. E. Mallouk, ?Miniaturized electrochemistry? Nature, vol. 343, no. 6258, pp. 515?516, May 1990. [65] T. M. Taylor, ?Efficiency enhancement for solid-electrolyte fuel cell? US Patent US5102750 A, May 1992. [66] S. C. Barton, T. Patterson, E. Wang, T. F. Fuller, and A. C. West, ?Mixed-reactant strip cell direct methanol fuel cells? Journal of Power Sources, vol. 96, no. 2, pp. 329?336, 2001. [67] M. A. Priestnall, V. P. Kotzeva, D. J. Fish, and E. M. Nilsson, ?Compact mixed-reactant fuel cells? Journal of Power Sources, vol. 106, pp. 21?30, 2002. [68] M. A. Priestnall, M. J. Evans, and M. S. P. Shaffer, ?Mixed-reactant fuel cells with flow through porous electrodes? US Patent Application US20030165727 A1, May 2003. [69] M. A. Priestnall, M. J. Evans, and M. S. P. Shaffer, ?Mixed-reactant fuel cells? US Patent Application US20080063909 A1, May 2008. [70] M. A. Priestnall, M. J. Evans, and M. S. P. Shaffer, ?Mixed-reactant fuel cells? US Patent Application US20030165727 A1, May 2004. [71] C. Gibbs, A. Jouvray, B. Lin, D. Papageorgopoulos, J. Fairless, and M. Hogarth, ?Novel mixed-reactant flow through DMFC stack employing RuSe ORR catalyst designed with computational fluid dynamics? in 3rd MEA Manufacturing Symposium, Ohio, USA 2007. [72] Fuel Cell Report Industry, ?Fuel Cell Stack Technology? 2004. [73] A. K. Shukla, C. L. Jackson, K. Scott, and G. Murgia, ?A solid-polymer electrolyte direct methanol fuel cell with a mixed-reactant and air anode? Journal of Power Sources, vol. 111, no. 1, pp. 43?51, 2002.  185  [74] K. Scott, A. K. Shukla, C. L. Jackson, and W. R. Meuleman, ?A mixed-reactants solid-polymer-electrolyte direct methanol fuel cell? Journal of Power Sources, vol. 126, no. 1?2, pp. 67?75, Feb. 2004. [75] A. K. Shukla, R. K. Raman, and K. Scott, ?Advances in mixed-reactant fuel cells? Fuel Cells, vol. 5, no. 4, pp. 436?447, Dec. 2005. [76] A. Jerome, ?Mixed-reactant molecular screen fuel cell? US Patent US20050058875 A May 2005. [77] R. Kothandaraman, W. Deng, M. Sorkin, A. Kaufman, H. F. Gibbard, and S. C. Barton, ?Methanol anode modified by semipermeable membrane for mixed-feed direct methanol fuel cells? Journal of The Electrochemical Society, vol. 155, no. 9, pp. B865?B868, 2008. [78] H. Meng, M. Wu, X. X. Hu, M. Nie, Z. D. Wei, and P. K. Shen, ?Selective cathode catalysts for mixed-reactant alkaline alcohol fuel cells? Fuel Cells, vol. 6, no. 6, pp. 447?450, Dec. 2006. [79] W. Yuan, K. Scott, and H. Cheng, ?Fabrication and evaluation of Pt?Fe alloys as methanol tolerant cathode materials for direct methanol fuel cells? Journal of Power Sources, vol. 163, no. 1, pp. 323?329, Dec. 2006. [80] H. Cheng, W. Yuan, and K. Scott, ?The influence of a new fabrication procedure on the catalytic activity of ruthenium-selenium catalysts? Electrochimica Acta, vol. 52, no. 2, pp. 466?473, Oct. 2006. [81] H. Cheng, W. Yuan, and K. Scott, ?A liquid-gas phase mixed-reactant fuel cell with a RuSeW cathode electrocatalyst? Journal of Power Sources, vol. 183, no. 2, pp. 678?681, 2008. [82] H. Cheng, W. Yuan, K. Scott, D. J. Browning, and J. B. Lakeman, ?The catalytic activity and methanol tolerance of transition metal modified-ruthenium?selenium catalysts? Applied Catalysis B: Environmental, vol. 75, no. 3?4, pp. 221?228, Sep. 2007. [83] K. Scott, W. Taama, D. R. A. Haslar, G. Britain, and J. Cruickshank, ?Performance of a direct methanol fuel cell? Journal of Applied Electrochemistry, vol. 28, no. 3, pp. 289?297, 1998. [84] R. Zeng and P. K. Shen, ?Selective membrane electrode assemblies for bipolar plate-free mixed-reactant fuel cells? Journal of Power Sources, vol. 170, no. 2, pp. 286?290, Jul. 2007. [85] A. B. Ilicic, D. P. Wilkinson, and K. Fatih, ?Advancing Direct Liquid Redox Fuel Cells: Mixed-Reactant and In Situ Regeneration Opportunities? Journal of The Electrochemical Society, vol. 157, no. 4, pp. B529?B535, 2010. [86] E. Kjeang, N. Djilali, and D. Sinton, ?Microfluidic fuel cells: A review? Journal of Power Sources, vol. 186, no. 2, pp. 353?369, Jan. 2009. [87] A. Lam, D. Wilkinson, and J. Zhang, ?Novel approach to membraneless direct methanol fuel cells using advanced 3D anodes? Electrochimica Acta, vol. 53, no. 23, pp. 6890?6898, Oct. 2008. [88] D. Y. Wang, D. T. Kennedy, B. W. MacAllister, ?Method and device for gaseous fuel cell operation? US Patent US 5100742 May-1992.  186  [89] M. Yano, A. Tomita, M. Sano, and T. Hibino, ?Recent advances in single-chamber solid oxide fuel cells: A review? Solid State Ionics, vol. 177, no. 39?40, pp. 3351?3359, Jan. 2007. [90] I. Riess, ?On the single chamber solid oxide fuel cells? Journal of Power Sources, vol. 175, no. 1, pp. 325?337, Jan. 2008. [91] M. Polovina, B. Babi?, B. Kaluderovi?, and A. Dekanski, ?Surface characterization of oxidized activated carbon cloth? Carbon, vol. 35, no. 8, pp. 1047?1052, Jan. 1997. [92] B. K. Pradhan and N. K. Sandle, ?Effect of different oxidizing agent treatments on the surface properties of activated carbons? Carbon, vol. 37, no. 8, pp. 1323?1332, Jan. 1999. [93] M. Watanabe and S. Motoo, ?Electrocatalysis by ad-atoms? Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, vol. 60, no. 3, pp. 267?273, Apr. 1975. [94] B. D. McNicol and R. T. Short, ?The influence of activation conditions on the performance of platinum/ruthenium methanol electrooxidation catalysts surface enrichment phenomena? Journal of Electroanalytical Chemistry and Interfacial Electrochemistry, vol. 81, no. 2, pp. 249?260, Aug. 1977. [95] T. Iwasita, ?Electrocatalysis of methanol oxidation? Electrochimica Acta, vol. 47, no. 22?23, pp. 3663?3674, Aug. 2002. [96] E. Gyenge, M. Atwan, and D. Northwood, ?Electrocatalysis of Borohydride Oxidation on Colloidal Pt and Pt-Alloys (Pt-Ir, Pt-Ni, and Pt-Au) and Application for Direct Borohydride Fuel Cell Anodes? Journal of The Electrochemical Society, vol. 153, no. 1, pp. A150?A158, 2006. [97] D. A. Finkelstein, N. Da Mota, J. L. Cohen, and H. D. Abru?a, ?Rotating disk electrode (RDE) investigation of BH4? and BH3OH? electrooxidation at Pt and Au: Implications for BH4? fuel cells? The Journal of Physical Chemistry C, vol. 113, no. 45, pp. 19700?19712, Nov. 2009. [98] B. Molina Concha, M. Chatenet, E. A. Ticianelli, and F. H. B. Lima, ?In situ infrared (FTIR) study of the mechanism of the borohydride oxidation reaction on smooth Pt electrode? The Journal of Physical Chemistry C, vol. 115, no. 25, pp. 12439?12447, Jun. 2011. [99] V. W. S. Lam, D. C. W. Kannangara, A. Alfantazi, and E. L. Gyenge, ?Electrochemical quartz crystal microbalance study of borohydride electrooxidation on Pt: the effect of borohydride concentration and thiourea adsorption? The Journal of Physical Chemistry C, vol. 115, no. 6, pp. 2727?2737, Feb. 2011. [100] M. C. S. Esca?o, E. Gyenge, R. L. Arevalo, H. Kasai, M. Clare, and S. Esca, ?Reactivity descriptors for borohydride interaction with metal surfaces? The Journal of Physical Chemistry C, vol. 115, no. 40, pp. 19883?19889, Oct. 2011. [101] G. Rostamikia and M. J. Janik, ?Direct borohydride oxidation: mechanism determination and design of alloy catalysts guided by density functional theory? Energy & Environmental Science, vol. 3, no. 9, p. 1262, Aug. 2010. [102] G. Guella, B. Patton, and A. Miotello, ?Kinetic features of the platinum catalyzed hydrolysis of sodium borohydride from 11b NMR measurements? Journal of Physical Chemistry C, vol. 111, no. 50, pp. 18744?18750, Dec. 2007.  187  [103] J. S. Zhang, W. N. Delgass, T. S. Fisher, and J. P. Gore, ?Kinetics of Ru-catalyzed sodium borohydride hydrolysis? Journal of Power Sources, vol. 164, no. 2, pp. 772?781, Feb. 2007. [104] Y. Shang, R. Chen, and G. Jiang, ?Kinetic study of NaBH4 hydrolysis over carbon-supported ruthenium? International Journal of Hydrogen Energy, vol. 33, no. 22, pp. 6719?6726, Nov. 2008. [105] V. W. S. Lam, A. Alfantazi, and E. L. Gyenge, ?The effect of catalyst support on the performance of PtRu in direct borohydride fuel cell anodes? Journal of Applied Electrochemistry, vol. 39, no. 10, pp. 1763?1770, Apr. 2009. [106] Freudenberg Nonwovens, ?Viledon? FS2227E Technical Data Sheet? http://www.freudenberg-nw.com/, 2012. [107] Y. Hayakawa, K. Fukuzaki, S. Kohiki, Y. Shibata, T. Matsuo, K. Wagatsuma, and M. Oku, ?X-ray photoelectron spectroscopy of highly conducting and amorphous osmium dioxide thin films? Thin Solid Films, vol. 347, no. 1?2, pp. 56?59, Jun. 1999. [108] A. Hamnett and B. J. Kennedy, ?Bimetallic carbon supported anodes for the direct methanol-air fuel cell? Electrochimica Acta, vol. 33, no. 11, pp. 1613?1618, Nov. 1988. [109] R. Liu, H. Iddir, Q. Fan, G. Hou, A. Bo, K. L. Ley, E. S. Smotkin, Y.-E. Sung, H. Kim, S. Thomas, and A. Wieckowski, ?Potential-dependent infrared absorption spectroscopy of adsorbed CO and X-ray photoelectron spectroscopy of arc-melted single-phase Pt, PtRu, PtOs, PtRuOs, and Ru electrodes? The Journal of Physical Chemistry B, vol. 104, no. 15, pp. 3518?3531, Apr. 2000. [110] C. K. Rhee, M. Wakisaka, Y. V. Tolmachev, C. M. Johnston, R. Haasch, K. Attenkofer, G. Q. Lu, H. You, and A. Wieckowski, ?Osmium nanoislands spontaneously deposited on a Pt(111) electrode: an XPS, STM and GIF-XAS study? Journal of Electroanalytical Chemistry, vol. 554?555, pp. 367?378, Sep. 2003. [111] T. Jones, ?Electrodeposition of osmium? Metal Finishing, vol. 100, no. 6, pp. 84?90, Jun. 2002. [112] L. E. Cox and D. M. Hercules, ?A study of some potassium hexachlorometallate complexes using electron spectroscopy? Journal of Electron Spectroscopy and Related Phenomena, vol. 1, no. 3, pp. 193?207, Jan. 1972. [113] D. L. White, S. B. Andrews, J. W. Faller, and R. J. Barrnett, ?The chemical nature of osmium tetroxide fixation and staining of membranes by X-ray photoelectron spectroscopy? Biochimica et Biophysica Acta (BBA) - Biomembranes, vol. 436, no. 3, pp. 577?592, Jul. 1976. [114] M. G. Richmond, ?Annual survey of ruthenium and osmium for the year 1993? Coordination Chemistry Reviews, vol. 141, pp. 63?152, 1995. [115] P. Ferrin, A. U. Nilekar, J. Greeley, M. Mavrikakis, and J. Rossmeisl, ?Reactivity descriptors for direct methanol fuel cell anode catalysts? Surface Science, vol. 602, no. 21, pp. 3424?3431, Nov. 2008. [116] Y. Chung, C. Pak, G.-S. Park, W. S. Jeon, J.-R. Kim, Y. Lee, H. Chang, and D. Seung, ?Understanding a degradation mechanism of direct methanol fuel cell using TOF-SIMS and XPS? Journal of Physical Chemistry C, vol. 112, no. 1, pp. 313?318, Jan. 2008.  188  [117] V. G. Minkina, S. I. Shabunya, V. I. Kalinin, V. V. Martynenko, and A. L. Smirnova, ?Stability of alkaline aqueous solutions of sodium borohydride? International Journal of Hydrogen Energy, vol. 37, no. 4, pp. 3313?3318, Feb. 2012. [118] V. Minkina, S. Shabunya, V. Kalinin, V. Martynenko, and A. L. Smirnova, ?Long-term stability of sodium borohydrides for hydrogen generation? International Journal of Hydrogen Energy, vol. 33, no. 20, pp. 5629?5635, Oct. 2008. [119] S. Amendola, ?A safe, portable, hydrogen gas generator using aqueous borohydride solution and Ru catalyst? International Journal of Hydrogen Energy, vol. 25, no. 10, pp. 969?975, Oct. 2000. [120] I. Roche, E. Chainet, M. Chatenet, and J. Vondrak, ?Carbon-supported manganese oxide nanoparticles as electrocatalysts for the oxygen reduction reaction (ORR) in alkaline medium: physical characterizations and ORR mechanism? Journal of Physical Chemistry C, vol. 111, no. 3, pp. 1434?1443, Jan. 2007. [121] D. Sepa, M. Vojnovi, and A. Damjanovict, ?Oxygen reduction at silver electrodes in alkaline solutions? Electrochimica Acta, vol. 15, pp. 1355?1366, 1970. [122] E. L. Gyenge and J.-F. Drillet, ?The electrochemical behavior and catalytic activity for oxygen reduction of MnO2?C?Toray gas diffusion electrodes? Journal of The Electrochemical Society, vol. 159, no. 2, pp. F23?F34, 2012. [123] V. Hern?ndez-Ram rez, A. Alatorre-Ordaz, M. de L. Y pez-Murrieta, J. G. Ibanez, C. Ponce-de-Le n, and F. C. Walsh, ?Oxidation of the borohydride ion at silver nanoparticles on a glassy carbon electrode (GCE) using pulsed potential techniques? ECS Transactions, 2009, vol. 20, no. 1, pp. 211?225. [124] M. H. Atwan, D. O. Northwood, and E. L. Gyenge, ?Evaluation of colloidal Ag and Ag-alloys as anode electrocatalysts for direct borohydride fuel cells? International Journal of Hydrogen Energy, vol. 32, no. 15, pp. 3116?3125, 2007. [125] J. E. Schroeder, D. Pouli, and H. J. Seim, Fuel Cell Systems-II, vol. 90. Washington DC: American Chemical Society, 1969, pp. 93?101. [126] H. Hou, S. Wang, W. Jin, Q. Jiang, L. Sun, L. Jiang, and G. Sun, ?KOH modified Nafion? 112 membrane for high performance alkaline direct ethanol fuel cell? International Journal of Hydrogen Energy, vol. 36, no. 8, pp. 5104?5109, 2011. [127] X. Yang, Y. Liu, S. Li, X. Wei, L. Wang, and Y. Chen, ?A direct borohydride fuel cell with a polymer fiber membrane and non-noble metal catalysts.? Scientific reports, vol. 2, p. 567, Jan. 2012. [128] F. Bidault, D. J. L. Brett, P. H. Middleton, N. Abson, and N. P. Brandon, ?An improved cathode for alkaline fuel cells? International Journal of Hydrogen Energy, vol. 35, no. 4, pp. 1783?1788, Feb. 2010. [129] Z. Samec, A. Troj?nek, and J. Langmaier, ?Negative impedance of the Nafion? membrane between two electrolyte solutions? Journal of The Electrochemical Society, vol. 145, no. 8, p. 2740, Aug. 1998. [130] H. Cheng, K. Scott, and K. Lovell, ?Material aspects of the design and operation of direct borohydride fuel cells? Fuel Cells, vol. 6, no. 5, pp. 367?375, Oct. 2006.  189  [131] X. Geng, H. Zhang, W. Ye, Y. Ma, and H. Zhong, ?Ni?Pt/C as anode electrocatalyst for a direct borohydride fuel cell? Journal of Power Sources, vol. 185, no. 2, pp. 627?632, Dec. 2008. [132] N. Duteanu, G. Vlachogiannopoulos, M. R. Shivhare, E. H. Yu, and K. Scott, ?A parametric study of a platinum ruthenium anode in a direct borohydride fuel cell? Journal of Applied Electrochemistry, vol. 37, no. 9, pp. 1085?1091, Jul. 2007. [133] H. Cheng and K. Scott, ?Investigation of Ti mesh-supported anodes for direct borohydride fuel cells? Journal of Applied Electrochemistry, vol. 36, no. 12, pp. 1361?1366, Sep. 2006. [134] H. Cheng and K. Scott, ?Investigation of non-platinum cathode catalysts for direct borohydride fuel cells? Journal of Electroanalytical Chemistry, vol. 596, no. 2, pp. 117?123, Nov. 2006. [135] B. H. Liu, Z. P. Li, K. Arai, and S. Suda, ?Performance improvement of a micro borohydride fuel cell operating at ambient conditions? Electrochimica Acta, vol. 50, no. 18, pp. 3719?3725, Jun. 2005. [136] C. Celik, F. G. B. San, and H. I. Sarac, ?Effects of operation conditions on direct borohydride fuel cell performance? Journal of Power Sources, vol. 185, no. 1, pp. 197?201, 2008. [137] Z. P. Li, B. H. Liu, J. K. Zhu, and S. Suda, ?Depression of hydrogen evolution during operation of a direct borohydride fuel cell? Journal of Power Sources, vol. 163, no. 1, pp. 555?559, 2006. [138] R. Jamard, A. Latour, J. Salomon, P. Capron, and A. Martinent-Beaumont, ?Study of fuel efficiency in a direct borohydride fuel cell? Journal of Power Sources, vol. 176, no. 1, pp. 287?292, 2008. [139] A. E. Sanli, M. L. Aksu, and B. Z. Uysal, ?Advanced mathematical model for the passive direct borohydride/peroxide fuel cell? International Journal of Hydrogen Energy, vol. 36, no. 14, pp. 8542?8549, Jul. 2011. [140] A. Verma and S. Basu, ?Experimental evaluation and mathematical modeling of a direct alkaline fuel cell? Journal of Power Sources, vol. 168, no. 1, pp. 200?210, May 2007. [141] A. Shah, R. Singh, C. Ponce-de-Le n, R. G. Wills, and F. C. Walsh, ?Mathematical modelling of direct borohydride fuel cells? Journal of Power Sources, vol. 221, pp. 157?171, Jan. 2013. [142] B. H. Liu and S. Suda, ?Influences of fuel crossover on cathode performance in a micro borohydride fuel cell? Journal of Power Sources, vol. 164, no. 1, pp. 100?104, Jan. 2007. [143] A. Aziznia, C. W. Oloman, and E. L. Gyenge, ?A Swiss-roll liquid?gas mixed-reactant fuel cell? Journal of Power Sources, vol. 212, pp. 154?160, Aug. 2012. [144] K. Scott, W. M. Taama, and P. Argyropoulos, ?Engineering aspects of the direct methanol fuel cell system? Journal of Power Sources, vol. 79, no. 1, pp. 43?59, May 1999. [145] S. G. Bratsch, ?Standard electrode potentials and temperature coefficients in water at 298.15 K? Journal of Physical and Chemical Reference Data, vol. 18, no. 1, p. 1, Jan. 1989. [146] D. D. Wagman, W. H. Evans, V. B. Parker, R. H. Schumm, and I. Halow, ?The NBS tables of chemical thermodynamic properties: Selected values for inorganic and C1 and C2  190  organic substances in SI units? Journal of Physical and Chemical Reference Data, vol. 11, no. Supplement 2, Jan. 1982. [147] B. Molina Concha and M. Chatenet, ?Direct oxidation of sodium borohydride on Pt, Ag and alloyed Pt?Ag electrodes in basic media? Electrochimica Acta, vol. 54, no. 26, pp. 6130?6139, Nov. 2009. [148] M. Chatenet, M. B. Molina-Concha, N. El-Kissi, G. Parrour, and J.-P. Diard, ?Direct rotating ring-disk measurement of the sodium borohydride diffusion coefficient in sodium hydroxide solutions? Electrochimica Acta, vol. 54, no. 18, pp. 4426?4435, Jul. 2009. [149] K. Scott, W. M. Taama, S. Kramer, P. Argyropoulos, and K. Sundmacher, ?Limiting current behaviour of the direct methanol fuel cell? Electrochimica Acta, vol. 45, no. 6, pp. 945?957, Dec. 1999. [150] E. H. Yu, K. Scott, and R. W. Reeve, ?Electrochemical reduction of oxygen on carbon supported Pt and Pt/Ru fuel cell electrodes in alkaline solutions? Fuel Cells, vol. 3, no. 4, pp. 169?176, Dec. 2003. [151] B. H. Liu, Z. P. Li, J. K. Zhu, and S. Suda, ?Influences of hydrogen evolution on the cell and stack performances of the direct borohydride fuel cell? Journal of Power Sources, vol. 183, no. 1, pp. 151?156, Aug. 2008. [152] R. K. Raman and A. K. Shukla, ?A direct borohydride/hydrogen peroxide fuel cell with reduced alkali crossover? Fuel Cells, vol. 7, no. 3, pp. 225?231, Jun. 2007. [153] M. C. E. Groves and A. Sasonow, ?Uhde EnviNOx? technology for NOX and N2O abatement: a contribution to reducing emissions from nitric acid plants? Journal of Integrative Environmental Sciences, vol. 7, no. sup1, pp. 211?222, Aug. 2010. [154] J. P rez-Ram  rez, F. Kapteijn, K. Sch?ffel, and J. . Moulijn, ?Formation and control of N2O in nitric acid production, where do we stand today?? Applied Catalysis B: Environmental, vol. 44, no. 2, pp. 117?151, Aug. 2003. [155] E. Gyenge, ?Electrooxidation of borohydride on platinum and gold electrodes: implications for direct borohydride fuel cells? Electrochimica Acta, vol. 49, no. 6, pp. 965?978, Mar. 2004. [156] E. Gileadi, Electrode Kinetics for Chemists, Chemical Engineers and Materials Scientists. New York: Wiley/Inter Science, 1993, p. 413. [157] A. Tamimi, E. B. Rinker, and O. C. Sandall, ?Diffusion Coefficients for Hydrogen Sulfide, Carbon Dioxide, and Nitrous Oxide in Water over the Temperature Range 293-368 K? Journal of Chemical Engineering Data, vol. 39, no. 2, pp. 330?332, 1994. [158] R. F. Weiss and B. A. Price, ?Nitrous oxide solubility in water and seawater? Marine Chemistry, vol. 8, no. 4, pp. 347?359, Feb. 1980. [159] A. V. Tripkovic, K. D. Popovic, and J. D. Lovic, ?A time effect in the early stages of a surface oxidation of a Pt(111) plane in alkaline solution? J. Serb. Chem. Soc., vol. 66, no. 11?12, pp. 825?833, 2001. [160] A. . Tripkovi?, S. ?trbac, and K. D. Popovi?, ?Effect of temperature on the methanol oxidation at supported Pt and PtRu catalysts in alkaline solution? Electrochemistry Communications, vol. 5, no. 6, pp. 484?490, Jun. 2003.  191  [161] A. V. Tripkovi?, K. ?. Popovi?, J. D. Mom?ilovi?, and D. M. Dra?i?, ?Kinetic and mechanistic study of methanol oxidation on a Pt(110) surface in alkaline media? Electrochimica Acta, vol. 44, no. 6?7, pp. 1135?1145, Nov. 1998. [162] M. H. Martin and A. Lasia, ?Hydrogen sorption in Pd monolayers in alkaline solution? Electrochimica Acta, vol. 54, no. 22, pp. 5292?5299, Sep. 2009. [163] E. T. Voltcheva, ?Influence of the entropy factor on the equilibrium and kinetics of chemisorption? Reaction Kinetics and Catalysis Letters, vol. 5, no. 2, pp. 85?91, Jun. 1976. [164] A. B. Anderson, J. Roques, S. Mukerjee, V. S. Murthi, N. M. Markovic, and V. Stamenkovic, ?Activation energies for oxygen reduction on platinum alloys: Theory and experiment? The Journal of Physical Chemistry B, vol. 109, no. 3, pp. 1198?1203, Jun. 2005. [165] S. K. Kamarudin, W. R. W. Daud, S. L. Ho, and U. A. Hasran, ?Overview on the challenges and developments of micro-direct methanol fuel cells (DMFC)? Journal of Power Sources, vol. 163, no. 2, pp. 743?754, Jan. 2007. [166] A. T. Kuhn and J. S. Booth, ?Electrical leakage currents in bipolar cell stacks? Journal of Applied Electrochemistry, vol. 10, no. 2, pp. 233?237, Mar. 1980. [167] C. R. Cloutier, E. L. Gyenge, and A. Alfantazi, ?Physicochemical Properties of Alkaline Aqueous Sodium Metaborate Solutions? Journal of Fuel Cell Science and Technology, vol. 4, no. 1, pp. 88?98, 2007. [168] S. Suda, ?Aqueous borohydride solutions? in in Handbook of Fuel Cells: Fundamentals, Technology, Applications, 1st ed., W. Vielstich, A. Lamm, and H. A. Gasteiger, Eds. John Wiley & Sons, Ltd., 2003, p. 3826. [169] E. L. Gyenge and C. W. Oloman, ?Electrosynthesis attempts of tetrahydridoborates? Journal of Applied Electrochemistry, vol. 28, no. 10, pp. 1147?1151, Oct. 1998. [170] J.-F. Drillet, F. Holzer, T. Kallis, S. M?ller, and V. M. Schmidt, ?Influence of CO2 on the stability of bifunctional oxygen electrodes for rechargeable zinc/air batteries and study of different CO2 filter materials? Physical Chemistry Chemical Physics, vol. 3, no. 3, pp. 368?371, Jan. 2001. [171] A. A. Serov, A. Aziznia, H. P. Benhangi, K. Artyushkova, P. Atanassov, and E. L. Gyenge, ?A novel borohydride-tolerant oxygen electroreduction catalyst for mixed-reactant Swiss-roll direct borohydride fuel cells? Journal of Materials Chemistry A, vol. Accepted, 2013. [172] X. Ni, Y. Wang, Y. L. Cao, X. P. Ai, H. X. Yang, and M. Pan, ?A highly efficient and BH4?-tolerant Eu2O3-catalyzed cathode for direct borohydride fuel cells? Electrochemistry Communications, vol. 12, no. 5, pp. 710?712, May 2010. [173] J. Ma, J. Wang, and Y. Liu, ?Iron phthalocyanine as a cathode catalyst for a direct borohydride fuel cell? Journal of Power Sources, vol. 172, no. 1, pp. 220?224, 2007. [174] G. Zhou, L.-D. Chen, and J. P. Seaba, ?CFD prediction of shunt currents present in alkaline fuel cells? Journal of Power Sources, vol. 196, no. 20, pp. 8180?8187, Oct. 2011. [175] A. B. Ilicic, M. S. Dara, D. P. Wilkinson, and K. Fatih, ?Improved performance of the direct methanol redox fuel cell? Journal of Applied Electrochemistry, vol. 40, no. 12, pp. 2125?2133, Sep. 2010.  192  [176] K. Fatih, D. P. Wilkinson, F. Moraw, A. Ilicic, and F. Girard, ?Advancements in the direct hydrogen redox fuel cell? Electrochemical and Solid-State Letters, vol. 11, no. 2, p. B11, Feb. 2008. [177] J. Moulder, W. Stickle, K. Sobol, and K. Bomben, Handbook of X-ray Photoelectron Spectroscopy. Eden Prairie, MN USA: Perkin-Elmer Corp., 1992. [178] W. D. Davis, L. S. Mason, and G. Stegeman, ?The heats of formation of sodium borohydride, lithium borohydride and lithium aluminum hydride? Journal of the American Chemical Society, vol. 71, no. 8, pp. 2775?2781, Aug. 1949. [179] D. M. F. Santos and C. A. C. Sequeira, ?Sodium borohydride determination by measurement of open circuit potentials? Journal of Electroanalytical Chemistry, vol. 627, pp. 1?8, 2009. [180] D. A. Lyttle, E. H. Jensen, and W. A. Struck, ?Simple volumetric assay for sodium borohydride? Analytical Chemistry, vol. 24, no. 11, pp. 1843?1844, Nov. 1952. [181] M. V. Mirkin and A. J. Bard, ?Voltammetric method for the determination of borohydride concentration in alkaline aqueous solutions? Analytical Chemistry, vol. 63, no. 5, pp. 532?533, Mar. 1991.  193  Appendices 194  Appendix A  Table of Potential Fuels and Oxidants for the MRFC Table A-1 Potential fuels and oxidants for the MRFC, with reaction equations and selected properties in acidic solutions ([H+] =1.0 mol kg-1), T=298 K, P=101.3 kPa (abs) Fuel                                                Oxidant Oxygen/Air O2 + 2H+ + 2e? ? H2O  E?=1.23 VSHE Example catalysts: Pt H2O2 H2O2 + 2H+ + 2e? ? 2H2O  E?=1.78 VSHE Example catalysts: Pt, Au, Ni Nitrous Oxide N2O+2H++2e-? N2+H2O  E?=1.77 VSHE Example catalysts: Pt, Pd Methanol CO2 + 6H+ + 6e- ?CH3OH + H2O E?= 0. 03 VSHE Energy Density: 4820 (100% wt.) Wh L-1 Example catalysts: Pt, PtRu CH3OH + 3/2O2 ? 2H2O + CO2  Eo=1.20 V CH3OH + 3H2O2 ? 5H2O + CO2 E?=1.75 V CH3OH + 3N2O ? 2H2O + CO2+ 3N2 E?=1.75 V Ethanol 2CO2 + 12H+ + 12e- ?C2H5OH + 3H2O E?= 0. 104 VSHE Energy Density: 6280 (100% wt.) Wh L-1 Example catalysts: Pt, PtSn C2H5OH + 3O2 ? 3H2O + 2CO2 E?=1.12 V C2H5OH + 6H2O2 ? 9H2O + 2CO2 E?=1.67 V C2H5OH + 6N2O ? 3H2O + 2CO2+ 6N2 E?=1.67 V Formic Acid CO2 + 2H+ + 2e- ?HCOOH  E?= -0.13 VSHE Energy Density: 1750 (88 % wt.) Wh L-1 Example catalysts: Pt, Pd HCOOH + 1/2O2 ? H2O + CO2 E?=1.36 V HCOOH+ H2O2 ? 2H2O + CO2 E?=1.90 V HCOOH + N2O ? H2O + CO2+ N2 E?=1.90 V Ammonia N2 + 6H+ + 6e-?2NH3  E?= 0.09 VSHE Example catalysts: Pt, Au, Ag, Ni 2NH3 + 3/2O2 ? N2 + 3H2O E?=1.14 V 2NH3 + 3H2O2 ? 6H2O + N2 E?=1.87 V NH3 + 3/2N2O ? 3/2H2O + 2N2 E?=1.86 V     195  Table A-2 Potential fuels and oxidants for the MRFC, with reaction equations and selected properties in basic solutions ([OH-] =1.0 mol kg-1), T=298 K, P=101.3 kPa (abs)  Fuel                                                Oxidant Oxygen/Air 2O2 + 4H2O + 8e? ? 8OH?  E=0.40 VSHE Example catalysts: Pt, Ag, MnO2 H2O2 Nitrous Oxide N2O + H2O + 2e- ?N2+2OH- E=0.94 VSHE Example catalysts: Pt, Pd Methanol CO32- + 6H2O + 6e-?CH3OH + 8OH-  E= -0.91 VSHE Example catalysts: Pt, PtRu CH3OH + 3/2O2+2OH- ? CO32- + 3H2O  E?= 1.31V not stable CH3OH + 3N2O +2OH- ? 3N2 + 3H2O + CO32- E?=1.85 V Ethanol  2CO32- + 11H2O + 12e-?C2H5OH + 16OH-  E= -0.55 VSHE Example catalysts: Pt, PtSn C2H5OH + 3O2+4OH- ? 2CO32-+5H2O E?= 0.95 V C2H5OH + 6N2O+4OH- ? 6N2 + 5H2O + 2CO32- E?=1.49 V Formate CO32- + 2H2O + 2e- ?HCOO- + 3OH- E= -0.93 VSHE Example catalysts: Pd, Pt HCOO- + OH- + 1/2O2 ? CO32- + H2O  E?= 1.33 V HCOO- + N2O + OH- ? N2 + H2O + CO32- E?=1.87 V Sodium borohydride NaBO2 + 6H2O + 8e- ?NaBH4 + 8OH- E= -1.24 VSHE Example catalysts: Pt, PtRu, Os, Ni, NaBH4 + 2O2 ? NaBO2 + 2H2O  E?= 1.64 V NaBH4 + 4N2O ? 4N2 + 2H2O + NaBO2 E?=2.18 V Ammonia N2 + 6H2O + 6e- ?2NH3 + 6OH- E= -0.73 VSHE Example catalysts: Pt, Au, Ag, Ni  NH3 + 3/4O2 ? 1/2N2 + 3/2H2O  E?= 1.13 V NH3 + 3/2N2O ? 2N2 + 3/2H2O E?=1.67 V  196  Appendix B  Experimental Procedure B.1 Activity of MnO2 toward the BOR and ORR The electrocatalytic activity and selectivity of the MnO2 GDE toward NaBH4 oxidation and the ORR was evaluated and the kinetics parameters were used in Chapter 4. The half-cell experiments were carried out in an asymmetric H-type glass cell (Figure B-1) with a gas and an electrolyte chamber which are kept tightly together by two screws. The working electrode (GDE), with geometrical surface area of 3.1?10-4 m2 (3.1 cm2), was placed between these two chambers and sealed with a Neoprene gasket. The gas chamber has an inlet and outlet for the gas flow. The whole assembly was placed in a water bath for carrying out experiments at different controlled temperatures. A WaveNow potentiostat (Pine Research Instruments) was used for the steady state and cyclic voltammetry measurements. A Hg/HgO oxide reference electrode (XR400, Radiometer Analytical) with 0.1 N KOH filling solution, and a platinized titanium counter electrode (Gold Plating Services) was used. Before measurements, N2 was sparged into the electrolyte for about 60 min to remove the dissolved oxygen from the electrolyte. Then O2 was sparged for about 30 min. During the measurements, a flow of O2 at ambient pressure (101 kPa (abs)) was maintained over the electrolyte surface and the gas diffusion chamber to hold a constant concentration of O2 in the solution and diffusion layer.  197   Figure B-1 H cell used for half-cell electrochemical tests Figure B-2 shows the BOR and ORR activity of MnO2 GDE in 2 M NaOH and 323 K. Figure B-2 A shows cyclic voltagramms of the same electrode under N2 in 2 M NaOH with presence of 1 M NaBH4. It can be seen that MnO2 GDE is active toward BH4- oxidation with onset potential of -0.6 VMMO. Therefore, the establishment of a mixed-potential is expected at the cathode. The ORR onset potential is about -0.1 VMMO without the presence of the BH4- (Figure B-2 B). A Tafel plot of the BOR and ORR on the MnO2 GDE is shown in Figure B-2 insets. The apparent exchange current density of BOR and ORR is determined to be respectively 0.49 and 4.26 A m-2 at 323 K and is reported in Table  4-1. Other electrochemical parameters of Pt and MnO2 GDE used in this study i.e. charge transfer coefficients and number of electron transfers, were fitted by comparing the measured electrode potential at OCV with Equations (4-11) and (4-12).  198   Figure B-2 Linear voltammetry and Tafel plot (inset) of (A) anodic scan of BOR on MnO2 GDE  in 2M NaOH+1 M NaBH4 (B) MnO2 GDE ORR in 2 M NaOH. Conditions: 5 mV s-1, 333 K, P=101.3kPa (abs). Note: EeBOR=-1.40 VMMO and EeORR=0.21 VMMO. B.2 Activity and Selectivity of Pt and Pd toward N2O Reduction The electrochemical setup consisted of a computer-controlled CHI potentiostat (1100A), a modulated speed rotator and an adjustable speed controller from Pine Instruments. The working electrode was a Pine Instruments rotating disk electrode with a Teflon disk insert tip. The disks used as electrodes were a 5?10-3 m (5 mm) diameter polycrystalline Pt (Pine Instruments) and a 5?10-3 m (5 mm) diameter polycrystalline Pd (Accumet Materials). Both  199  disks were first polished to a 0.25 ?m finish using alumina (Cypress Systems polishing kit) then thoroughly rinsed and sonicated in de-ionized water. The three?electrode cell setup, used for all the experiments, consisted of a double jacketed two-compartment glass cell, Pt or Pd rotating disc electrode as working electrode, Hg/HgO with 0.1 N KOH filling solution as reference electrode (Radiometer), a smooth platinum mesh as counter electrode and 0.1 M NaOH as electrolyte. In this study, N2 (99.999 vol%), O2 (99.99 vol%) and N2O (99.99 vol%) gases were obtained from Praxair Company. Methanol (99.5 wt%, ACS reagent) and NaOH (99.8 wt%) were obtained from Fischer Scientific. Electrolytes were prepared using ultrapure water (18 M?). Before the measurements, N2 was sparged into the electrolyte for about 60 minutes to remove dissolved oxygen in the electrolyte. Then N2O was sparged for about 30 minutes. During the measurements, a flow of N2O at ambient pressure (101 kPa(abs)) was maintained over the electrolyte surface to hold a constant concentration of N2O in the solution. The electrolyte temperature was controlled from 295 to 333 K by cycling water, with the desired temperature, in the cell?s jacket using a temperature-controlled water bath and pump (Fischer Scientific, 3016HD). Prior to N2O reduction measurements, the electrode was cycled between -0.05 and 1.5 VRHE at 0.100 V s-1 for about 200 cycles, in an N2-sparged electrolyte at 295 K, until steady-state voltammograms were obtained. The cyclic voltammetry measurements were performed between -0.08 V and 1.2 V using sweep rates ranging from 0.005 to 0.500 V s-1 and rotation speeds between 0 and 3600 rpm.  200  B.3 Activation of carbon Cloth Substrate Prior to electrodeposition of the Os or spraying Pt or PtRu catalysts, the woven carbon cloth substrates (described in Chapter 2) were pretreated in 1 M HNO3 at 363 K for 1 hr. The substrates were then thoroughly rinsed with 18 M? deionized water and air-dried at 333 K in an oven for 2 hr. B.4 Preparation of Pt and PtRu Anode The Pt and PtRu anodes used in this Swiss-roll MRFC were manufactured in-house by standard techniques. Catalysts inks made of Pt/C (with 50 wt% Pt) or PtRu/C (with 20 wt% PtRu, 1:1 atom ratio) mixed with 30 wt% Nafion? (from a 5 wt% Nafion? in lower alcohols solution, Sigma-Aldrich Inc.) were sprayed by a CNC moving table equipped with an air sprayer gun onto a 0.1 m?0.1 m woven carbon cloth (purchased from Fuel Cell Earth LLC, thickness of 3.5?10-4 m). Platinum and PtRu catalyst loadings were 0.8?10-2 kg m-2 (0.8 mg cm-2) on the anode. B.5 Preparation of the Electrodeposited Os 3D Anode The Os electrodeposition was carried out on the HNO3-treated carbon cloth substrate according to a surfactant-assisted method introduced recently by Lam et al.[46]. The electrodeposition media consisted of 10 mM (NH4)2OsCl6 (ammonium hexachloroosmate(IV)) solution (99.99% trace metal basis, Alfa Aesar) prepared with 18 M? deionized water. To control the deposit morphology and structure, 12.5 %vol. Triton-X 102 (Sigma Aldrich) non-ionic surfactant was added to the mixture. The electrodeposition media, composed of Triton-X  201  102 and the (NH4)2OsCl6 solution, was transferred to a Ti container with a total volume of 140 mL. The solution was heated and stirred on a hot plate at 341 K for 30 min prior to deposition. Figure B-3 shows the schematic and components of the electrodeposition cell.  Figure B-3 Schematic of the electrodeposition cell. 1: perforated platinized Ti counter electrodes (anode), 2: polycarbonate cell frames, 3: Ti frames acting as current feeders to the carbon cloth substrate, 4: woven carbon cloth cathode (deposition substrate), 5: Ti cell holder box The woven carbon cloth substrate was sandwiched between two Ti frames acting as current feeders. The counter electrodes were two perforated platinized Ti plates (purchased from Gold Plating Services) placed at a distance of 1 cm from the carbon cloth substrate. The  202  sandwiched assembly was then slid into the Ti box containing 120 mL of the electrodeposition solution. The polycarbonate cell was designed such that 0.02 m ? 0.2 m (2 cm?10 cm) area of the carbon cloth was exposed to the deposition solution. Electrodeposition was carried out at 341 K at a constant current density of 40 A m-2 for 30 min. The electrodeposition media was slowly stirred by a magnetic stirrer during the deposition process. At the end of the deposition, the substrate was removed and soaked in boiling acetone for 30 min. Afterwards the substrate was dried in air at 295 K. The electrodeposited Os anode was tested for fuel cell performance the same day. When more than one electrodeposition steps were performed on the same substrate, the substrate was washed in boiling acetone for 15 min between depositions and 30 min after the final deposition. B.6 Electrode Characterization The morphology and structure of the electrodes in this diessertation were characterized by X-ray photoelectron spectroscopy (XPS) (Leybold Max 200 and Kratos AXIS Ultra), scanning electron microscopy (SEM) (Hitachi S-4700 and Hitachi S-4500 field emission scanning electron microscopes), scanning TEM (FEI Tecnai G2 200 kV), and inductively coupled plasma optical emission spectrometry (ICP-OES) for Os loading determination (Maxxam Inc.).  Statistical size-distribution histograms for the resulting nanoparticles were produced from TEM images (software: ImageJ, sample size: 67). For TEM analysis of the electrodeposited Os anodes, the electrode was scratched on a weighting paper and then dispersed in ethanol by  203  sonication for 15 min before drop casting on a carbon-coated copper grid. For a more accurate deconvolution of the XPS spectrum, a commercial Os powder (200 mesh 99.8% metal basis, Alfa Aesar) was also analyzed as a reference. A nonlinear least-square curve-fitting procedure (software: XPS PEAK4.1) was applied in the Os 4f region. The area ratio between the 4f7/2 and 4f5/2 peaks for Os was fixed at the theoretical value of 1.33. The binding energy difference was fixed at 2.72 eV [113,177]. B.7 NaBH4 Hydrolysis Rate Measurements To investigate the BH4- hydrolysis rate, an apparatus was developed based on the gasometric measurements of the H2 generated by the complete decomposition of BH4- with HCl according to Equation (B-1): NaBH4 (aq) + HCl (aq)+2H2O (l)? 9/2H2 (g)+ NaBO2(ag)+ NaCl(aq)   (B-1) The volume of water displaced by H2 gas was measured and compared to a calibration curve in the range of 0-5 %wt. NaBH4. This method shows excellent reproducibility and accuracy for measurement of BH4- concentrations, competitive or in some cases superior to other analytical methods such as acid-base titration [178], potentiometric method [179], iodate method [180] and voltammetric measurements [181]. To investigate the BH4- hydrolysis rate on Pt, OsO2 and Os(0), 50 mg of the following commercial catalysts: Pt black (HiSPECTM, Alfa Aesar), Os powder (200 mesh 99.8% metal basis, Alfa Aesar), and OsO2 (Os 83% min, Alfa Aesar) were added separately to 100 mL of 4.3 % wt. NaBH4 (Aldrich) and 2 M NaOH (Fischer) solution at controlled temperatures of 298 K and at 323 K, respectively. Every 5 minutes, a 3 mL sample  204  was taken from the solution and fully decomposed by injection of 20 mL 6 N HCl. The concentration of BH4- was then determined by comparison of the evolved H2 gas volume with respect to the calibration curve. B.8 Pre-Treatment and Conditioning of Nafion? Membrane Before the experiments in the SR-MRFC, the Nafion? 112 was conditioned in boiling 3% H2O2, rinsed thoroughly and boiled in deionzied water (DI), followed by conditioning in boiling 0.5 M H2SO4, each step for 30 min. The conditioned Nafion? membrane was then rinsed thoroughly and stored in DI water until used in SR-MRFC experiments. B.9 Electrochemical Impedance Spectroscopy (EIS) Measurements To determine the Ohmic resistance of the SR-MRFC, electrochemical impedance spectroscopy (EIS) was performed using a PARSTAT 2263 potentiostat (Princeton Applied Research). The EIS was measured at OCV with amplitude of 10 mV in the frequency range of 105 Hz to 10-2 Hz. For finding the contact resistances, the Swiss-roll was assembled without the separator (short-circuited from inside) and EIS was performed while there was no reactant or electrolyte inside the reactor. Since there was no ionic conductivity inside the reactor, the measured impendence was related to total electronic plus contact resistance of the components.   205  Appendix C  Dimensioned Drawings of the Swiss-Roll Cell and Components It is important to note that in Appendix C, units of the Swiss-roll and its components dimensions are in inches. C.1 Reactor A  Figure C-1 Photograph (right) and illustration (left) of Reactor A    206   Figure C-2 Dimensioned drawing of Reactor A middle rod for monopolar operation  207  C.2 Reactor B  Figure C-3 (A) photograph of Reactor B and components; polycarbonate flange, 2 polycarbonate shell, two split graphite blocks, SS middle rod, feed sprayer  (B) Illustration of Reactor B  208   Figure C-4 Dimensioned drawing of Reactor B components; polycarbonate shell (units in inch)  209   Figure C-5 Dimensioned drawing of Reactor B components; graphite block  210   Figure C-6 Dimensioned drawing of Reactor B components; SS middle rod C.3 Multi-Layer Bipolar Reactor Multi-layer bipolar experiments were carried out in Reactor A (Appendix C.1) but with a different middle rod as shown in Figure C-7:  211   Figure C-7 Dimensioned drawing of middle rod for bipolar operation of multi-layer roll bipolar reactor; reactor was same as Reactor A  212  C.4 Rolls-in-Series Bipolar Reactor  Figure C-8 Photograph and schematic illustration of (5 cells) rolls-in-series reactor components; Cap, Nipple-1, Nipple-2, End, Adopter  213   Figure C-9 Dimensioned drawing of rolls-in-series components; CPVC Cap  Figure C-10 Dimensioned drawing of rolls-in-series components; CPVC Adaptor   214  C.5 Sprayer Nozzle  Figure C-11 Sprayer nozzle used in Reactor B: 1/4 JX6MPL1AC purchased form BEX    215  Table C-1 Specification of sprayer used in Reactor A Air pressure / psig Liquid capacity in GPH Optimum pattern Sprayer angle / ? Sprayer range / in 18-19 12-17 20 0.1 40 0.2 100 0.3 400 0.4 Reference: McMaster Carr website http://www.mcmaster.com/ Product Number: 32885K51 Table C-2 Specification of JPL1 BEX sprayer used in Reactor B Air pressure / psig Air capacity / SCFM Liquid capacity in GPH Optimum pattern Gravity Head Sprayer angle / ? Sprayer range / in 18? 12? 6? 18-19 12-17 10 0.47 0.63 0.55 0.46 20 0.66 0.73 0.66 0.60 40 1.06 0.87 0.81 0.76 60 1.48 0.98 0.92 0.88 Reference: BEX website http://www.bex.com/pdf/JPL99C.pdf Appendix D  Standard Error Calculation The standard error of the mean,  , was calculated according to:                          (D-1) Where    is the data value in measurement number i,    is the mean (or average) of a set of data values (sum of all of the data values divided by the number of data values), and N is the number of data points in the sample population, which in this dissertation, is equal to 3. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0166838/manifest

Comment

Related Items