Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Direct methanol fuel cell with extended reaction zone anode : PtRu and PtRuMo supported on fibrous carbon Bauer, Alexander Günter 2008

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

Item Metadata

Download

Media
24-ubc_2008_spring_bauer_alexander_gunter.pdf [ 18.76MB ]
Metadata
JSON: 24-1.0058502.json
JSON-LD: 24-1.0058502-ld.json
RDF/XML (Pretty): 24-1.0058502-rdf.xml
RDF/JSON: 24-1.0058502-rdf.json
Turtle: 24-1.0058502-turtle.txt
N-Triples: 24-1.0058502-rdf-ntriples.txt
Original Record: 24-1.0058502-source.json
Full Text
24-1.0058502-fulltext.txt
Citation
24-1.0058502.ris

Full Text

DIRECT METHANOL FUEL CELL WITH EXTENDED REACTION ZONE ANODE: PtRu AND PtRuMo SUPPORTED ON FIBROUS CARBON  by ALEXANDER GUNTER BAUER B.A.Sc., UNIVERSITY OF ERLANGEN-NURNBERG, 2000 M.A.Sc., UNIVERSITY OF ERLANGEN-NURNBERG, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES (Chemical and Biological Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA  January 2008 0 Alexander Giinter Bauer, 2008  Abstract The direct methanol fuel cell (DMFC) is considered to be a promising power source for portable electronic applications and transportation. At present there are several challenges that need to be addressed before the widespread commercialization of the DMFC technology can be implemented. The methanol electro-oxidation reaction is sluggish, mainly due to the strong adsorption of the reaction intermediate carbon monoxide on platinum. Further, methanol crosses over to the cathode, which decreases the fuel utilization and causes cathode catalyst poisoning. Another issue is the accumulation of the reaction product CO2  (g)  in the anode, which  increases the Ohmic resistance and blocks reactant mass transfer pathways. A novel anode configuration is proposed to address the aforementioned challenges. An extended reaction zone (thickness = —100-300 izn) is designed to facilitate the oxidation of methanol on sites that are not close to the membrane-electrode interface. Thus, the fuel concentration near the membrane may decrease significantly, which may mitigate adverse effects caused by methanol cross-over. The structure of the fibrous electrode, with its high void space, is believed to aid the disengagement of CO2 gas. In this thesis the first objective was to deposit dispersed nanoparticle PtRu(Mo) catalysts onto graphite felt substrates by surfactant mediated electrodeposition. Experiments, in which the surfactant concentration, current density, time and temperature were varied, were conducted with the objective of increasing the active surface area and thus improving the reactivity of the electrodes with respect to methanol electro-oxidation. The three-dimensional electrodes were characterized with respect to their deposit morphology, surface area, composition and catalytic activity. The second objective of this work was to utilize the catalyzed electrodes as anodes for direct methanol fuel cell operation. The fuel cell performance was studied as a function of methanol concentration, flow rate and temperature by using a single cell with a geometric area of 5 cm 2 . Increased power densities were obtained with an in-house prepared 3D PtRu anode compared to a conventional PtRu catalyst coated membrane.  ii  Coating graphite felt substrates with catalytically active nanoparticles and the utilization of these materials, is a new approach to improve the performance of direct fuel cells.  iii  ^  Table of Contents ABSTRACT  ^ ii  TABLE OF CONTENTS ^ LIST OF TABLES  iv vii  ^  LIST OF FIGURES  .ix  NOMENCLATURE ^  .xvii  ABBREVIATIONS ^  .xx  ACKNOWLEDGEMENTS  ^xxi  1.0 INTRODUCTION ^  .1  1.1 PRODUCTION, PROPERTIES AND UTILIZATION OF METHANOL ^ 4  2.0 THEORETICAL BACKGROUND  ^7  ^2.1 THERMODYNAMICS OF ELECTROCHEMISTRY  .7  2.2 CURRENT-POTENTIAL CORRELATIONS ^  8  2.3 THE THREE ELECTRODE SETUP ^  10  2.4. REFERENCE ELECTRODES ^  11  2.5 FUNDAMENTAL ELECTROCHEMICAL RESEARCH TECHNIQUES ^ 12 2.5.1 Cyclic voltammetry ^  12  2.5.2 Chronopotentiometry ^  14  2.5.3 Chronoamperometry ^  15  3.0 LITERATURE REVIEW ^  16  3.1 THE DIRECT METHANOL FUEL CELL (DMFC) ^  16  3.2 METHANOL ELECTRO-OXIDATION REACTION MECHANISM ^  17  3.2.1 The bifunctional mechanism ^  21  3.2.2 The ligand effect ^  .22  3.2.3 Carbon monoxide oxidation studies ^  23  3.2.4 The influence of the Pt:Ru atomic ratio and temperature on catalyst activity ^26 3.2.5 Ternary and quaternary electrocatalysts ^  28  3.2.6 The influence of the carbon support on catalyst activity and utilization ^ 29 3.3 PREPARATION OF NANOSCALE ELECTROCATALYSTS ^ 3.3.1 Sol-gel method  30  ^ 30  3.3.2 Vapor deposition ^  31  3.3.3 Colloid method ^  31  3.3.4 Electrodeposition 3.3.4.1 Potentiostatic electrodeposition  .33 ^ ..33  iv  ^ 35  3.3.4.2 Pulsed potential deposition. 3.3.4.3 Pulsed current deposition ^ 3.3.5 Chemical reduction in microemulsion ^  36 .. 37  3.3.6 Chemical reduction in liquid crystalline electrolytes utilizing nonionic surfactants ^ 38 3.3.7 Electroless deposition ^  38  3.4 CARBON DIOXIDE DISENGAGEMENT FROM THE ANODE ^  38  3.5 METHANOL CROSS-OVER ^  42  3.6 ELECTRODE DESIGN  ^  3.6.1 Alternative electrode designs ^  4.0 RESEARCH OBJECTIVES AND NOVELTY ^ 5.0 EXPERIMENTAL METHODS ^ 5.1 ELECTRODEPOSITION ^  47 50  .53 .....57 57  5.1.1 Chemical and electrochemical cleaning of the catalyst surface ^  57  5.1.2 Electrode materials  58  5.1.3 Electrodeposition cell types  62  5.1.3.2 Glass beaker (100 ml volume) ^  62  5.1.3.3 Sandwich plating cell (3ml volume) ^  64  5.1.4 Surfactants and additives ^  64  5.2 CHARACTERIZATION OF THE CATALYST DEPOSITS ^  67  5.3 HALF-CELL ELECTROCHEMICAL EXPERIMENTS  68  5.3.1 Active surface area assessment by copper under potential deposition and stripping 5.4 FUEL CELL EXPERIMENTS  6.0 RESULTS AND DISCUSSION ^ 6.1 ELECTRODE PREPARATION ^ 6.1.1 Acid and surfactant concentration in the deposition bath ^  70 73  76 77 77  6.1.1.1 a) Deposit morphology obtained without surfactant ^  78  6.1.1.1 b) Deposit morphology obtained with surfactant ^  79  6.1.1.2 Comparison of alkaline and acidic electrodeposition of PtRu ^  82  6.1.2 Fundamental electrochemical studies: Correlation between peak current density and scan rate ^ 6.1.3 Factorial experimental design (I) (100 ml glass beaker cell)  84 87  6.1.3.1 Surfactant mediated galvanostatic PtRu co-deposition ^  87  6.1.3.2 Surfactant mediated galvanostatic sequential deposition of PtRu ^  90  6.1.3.3 Summary ^  .93  6.1.4 Pulsed current sequential deposition ^  96  6.1.5 Factorial experimental design (II) (3 ml sandwich plating cell)  99  6.1.6 Deposition temperature and catalyst activity  109  6.1.7 Pt, Ru and PtRu deposition studies by linear voltammetry  115  6.1.8 Comparison of PtRu and PtRuMo catalysts ^  117  6.1.8.1 Single step co-deposition ^  117  6.1.8.2 Two step co-deposition ^  120  6.1.9 Pt, Ru, Mo, PtRu and PtRuMo deposition studies by linear voltammetry ^  125  6.2 TAFEL PLOT ANALYSIS OF METHANOL OXIDATION ON PtRu AND PtRuMo ^  130  6.3 EXPLORATION OF REACTION ZONE PROTON CONDUCTIVITY ^  133  6.3.1 Nafion  134  6.3.2 Kelzan ^  137  6.3.3 Acidified Si gel ^  139  6.4 FUEL CELL EXPERIMENTS  143  6.4.1 Effect of sulfuric acid concentration (PtRu, serpentine flow)  143  6.4.2 Effects of cathode pressure and anolyte flow rate (PtRu, serpentine flow) ^  145  6.4.3 Anode compression ^  147  6.4.4 Comparison of the novel 3D anode and a conventional catalyst coated membrane  150  6.4.5 Presence of Mo in the 3D anode ^  155  6.4.6 Factorial fuel cell experiments ^  157  6.4.7 Deactivation behavior of the DMFC ^  .162  7.0 CONCLUSIONS  . 165  8.0 RECOMMENDATIONS ^  167  REFERENCES ^ APPENDIX A: ALTERING THE DEPOSITION CELL SETUP ^ APPENDIX B: PULSED CURRENT DEPOSITION ^  .  .169 183 .188  APPENDIX C: ASSESSMENT OF REPLICABILITY FOR PtRu AND PtRuMo DEPOSITION ^ 192 APPENDIX D: CALCULATION OF GAS HOLD-UP AND OHMIC DROP ^ 194 APPENDIX E: METHOD FOR VERIFYING REMOVAL OF TRITON X-100 FROM THE ELECTRODE SURFACE AFTER ELECTRODEPOSITION ^ 197  APPENDIX F: SAMPLE EDX AND AUGER SPECTRA ^  200  APPENDIX G: SUPPLEMENTAL FACTORIAL FUEL CELL TEST RESULTS ^202  vi  List of Tables Table 1-1: Properties of different fuel cell types ^  2  Table 2-1: Reference electrodes and their respective half cell potentials 12 Table 3-1: Mass and volumetric energy density of compressed hydrogen, liquid hydrogen, gasoline, and methanol [Zittel et al., 1996] ^ 16 Table 3-2: Experimentally determined rate constants for methanol adsorption in a 10 -2 M CH3OH-0.1 M H2SO4 solution at 298 K [Lamy et al., 2001] ^ 22 Table 3-3: CO desorption range for pure Pt(111) and two different fractions of Ru on a Pt(111) surface, T = 298 K [Waszczuk, et al., 2002] ^ 26 Table 3-4: Methanol adsorption peak current density as a function of the atomic ratio measured by voltammetry at —0.16 V vs. SHE. Electrolyte: 1 M CH3OH^26 0.5 M H 2 SO 4 , scan rate: 1 mV s -1 , T = 298 K [Iwasita, 2002] Table 3-5: Fundamental methanol oxidation studies: Effects of temperature and atomic composition ^ ...27 Table 3-6: Characterization of ternary and quaternary electrocatalysts ^ .29 Table 3-7: Examples of Pt electrodeposition in liquid crystalline media. Surfactant: Octaethylene-glycol-monohexadecyl-ether (C16E08) ^ 35 Table 3-8: Examples of pulsed current deposition of fuel cell catalysts particles ^ .37 Table 3-9: Examples of membrane modification ^ 46 Table 3-10: Examples of membrane modification (continued) ^ 47 Table 3-11: Comparison of alternative and conventional anode designs ^ 51 Table 5-1: Characteristics of graphite felt types in their uncompressed state ^ .59 Table 5-2: The effect of compression on GF-S6 graphite felt (uncompressed thickness = 2000 jAm) ^ 60 Table 5-3: The effect of compression on GF-S3 graphite felt ^ 61 Table 6-1: Preliminary deposition tests utilizing the 100 ml glass beaker cell and 78 GF-S6 ^ Table 6-2: Comparison between acidic and alkaline electrodeposition of PtRu 82 on GF-S3 at 298 K ^ Table 6-3: Deposition conditions and catalyst properties ^ 84 Table 6-4: Variables and levels of factorial experiment, T = 298 K ^ 88 Table 6-5: Selected experimental data from the factorial experiment investigating .88 galvanostatic PtRu co-deposition on GF-S6 ^ Table 6-6: Selected experimental data from the factorial. Investigation of .91 galvanostatic sequential deposition of Pt and Ru on GF-S6 ^ Table 6-7: Effect of the deposition parameters on the atomic bulk composition of the PtRu catalyst obtained by galvanostatic co-deposition ^ .95 Table 6-8: Factors and their levels used in the factorial design for electrodeposition....99 Table 6-9: Deposition parameters and catalyst properties for selected samples from the factorial experiment . 99 Table 6-10: PtRu electrodeposition on GF-S3 with and without Triton X-100, 298 K ^ 107  vii  Table 6-11: Deposition temperatures and resulting catalyst properties ^ 110 Table 6-12: Comparison of PtRu and PtRuMo prepared at 333 K, i =60 A r11 -2 , 40 °/0„,t Triton X-100, t =90 min ^ .117 Table 6-13: Binary and ternary catalysts obtained by depositing twice at 333 K for 90 minutes ^ .120 Table 6-14: Overview of Tafel slope values reported in the literature ^ .131 Table 6-15: Exchange current densities for methanol oxidation obtained from Fig. 6-53 ^ 131 Table 6-16: Protonic conductivity and Ohmic potential loss at 333 K as a function of electrolyte content (based on H2SO4 as a model protonic conductor), i = 3000 A m -2 , felt thickness of GF-S3 in compressed state = 150 1,1m ^ 133 Table 6-17: Approximate load, volume fraction and coating thickness for model proton conducting materials ^ 142 Table 6-18: Ionic conductivity of bulk H2SO4 and effective ionic conductivity within 3D anode matrix for different sulfuric acid concentrations (T = 333 K) ^ 145 Table 6-19: Peak power densities obtained with CCM or GDE type anodes at T = 333 K ^ 153 Table 6-20: Variables and their levels applied for factorial DMFC experiments ^ .157 Table 6-21: Peak power density response for factorial tests using PtRu and PtRuMo with serpentine and flow-by type anode end plates, respectively 158 Table 6-22: Methanol flux across the membrane at open circuit [Casalegno et al., 2007] ^ .161 Table A-1: Summary of electrodeposition procedures carried out to compare the use of one and two counter electrodes, respectively ^ 183 Table B-1: Electrodeposition parameters for comparison of galvanostatic and reverse pulsed electrodeposition, T = 298 K ^ 189 Table C-1: Assessment of replicability of the methanol oxidation performance for catalyzed GF-S3. Single step deposition in 40 % wt. Triton X-100 applying i = 60 A 111 -2 for 90 min (3 ml sandwich plating cell) ^ 192 Table C-2: Assessment of replicability of the methanol oxidation performance for catalyzed GF-S3. Two step deposition with 40 %,„, t Triton X-100 applying i = 60 A m 2 for 90 min twice (3 ml sandwich plating cell) ^ .193 Table D-1: Calculation of DMFC variables as a function of temperature and current 194 density ^ ^ 196 Table D-2: Theoretical maximum conversion of methanol Table G-1: Sequence of experiments carried out with PtRu anode, serpentine flow....202 Table G-2: Sequence of experiments carried out with PtRu anode, flow-by mode......205 Table G-3: Sequence of experiments carried out with PtRuMo anode, serpentine flow208 Table G-4: Sequence of experiments carried out with PtRuMo anode, flow-by mode..211  viii  List of Figures Fig. 1-1: DMFC polarization curve obtained at 333 K, anode: PtRu (-50 g m -2 ) electrodeposited on graphite felt (GF-S3) in aqueous media without additives, 1 M CH3OH-0.5 M H2SO4, 5 ml min -1 , 100 kPa(abs); cathode: Pt black (40 g m -2 ) dry 02, 500 ml min -1 STP, 200 kPa(abs) ^  .4  Fig. 2-1: Three electrode cell configuration ^ 11 Fig. 2-2: Cyclic voltammetry: Variation of potential over time and related current response for a reversible electrochemical reaction ^ 13 Fig. 2-3: Cyclic voltammogram representing a non-reversible electrochemical reaction ^ 14 Fig. 2-4: Chronopotentiometry: Current step and resulting potential response ^ 14 Fig. 2-5: Chronoamperometry: Potential step and resulting current response ^ .15 Fig. 3-1: Reaction pathways for methanol electro-oxidation ^ 20 Fig. 3-2: Dependence of CO oxidation potential on ruthenium surface coverage at 298 K. Electrolyte: 0.5 M H2SO 4 . Scan rate: 100 mV s -1 [Davies et al., 2002] ^ 25 Fig. 3-3: Shell stabilized metal core ^ 33 Fig. 3-4: Schematic model of hexagonal micelle assembly, model structure of resulting deposit and nanoporous Pt deposit on Au coated quartz crystal [Gollas et al., 2000] ^ 34 Fig. 3-5: Photographs of the surfaces of carbon cloth (a, b) and carbon paper (c) backing layers [Argyropoulos et al., 1999] ^ 40 Fig. 3-6: CO 2 evolution patterns in liquid feed DMFCs as a function of current density, flow rate and gas diffusion layer material [Argyropoulos et al., 1999] ^ ..41 Fig. 3-7: Simplified Nafion structure and proton transport [Choi et al., 2005] ^ .....43 Fig. 3-8: Schematic of proton transport for an H 9 0 5 + species according to the Grotthus mechanism. Letters a through d denote oxygen atoms [Agmon, 1995] ^44 Fig. 3-9: Exploded view of DMFC with conventional anode design ^ .48 Fig. 3-10: Schematic of conventional catalyst preparation method (a) and decal transfer method (b) [Song et al., 2005] ^ 49 Fig. 3-11: Titanium mesh electrode with electrodeposited PtRu nanoparticles (d p — 5 nm) [Allen et al., 2005] ^ 52 Fig. 3-12: Porous carbon substrate and carbon surface with PtRu nanoparticles (d p — 2-3 nm; most favorable C pore diameter = 25 nm) [Chai et al., 2004] ^ .52 Fig. 3-13: Micrograph of conductive polymer beads with Pt catalyst [Xie et al., 2005] ^ .52 Fig. 4-1: Exploded view of DMFC with novel 3D anode design ^ 54 Fig. 4-2: End plate with serpentine flow field ^ ..55 Fig. 4-3: Exploded view of flow-by anode schematic and photograph of flow-by type end plate ^ 56 Fig. 5-1: Scan in 0.5 M H 2 SO 4 at 100 mV s -1 and T = 298 K. PtRu on GF-S3, deposition parameters: 40 %.„„ t Triton X-100, 60 A m 2 , 90 min, 333 K ^ 58 Fig. 5-2: Micrograph of uncatalyzed graphite felt (GF-S6) ^ 59 Fig. 5-3: Micrographs of uncatalyzed graphite felt (GF-S3) ^ 60 Fig. 5-4: Micrograph of single graphite fiber (GF-S3) ^ 60 ix  Fig. 5-5: Schematic of 100 ml glass beaker cell containing a three electrode assembly used for electrodeposition 63 ^ Fig. 5-6: Working electrode as used in 2 counter electrode plating setup ^ 63 Fig. 5-7: Schematic and photograph of sandwich type plating cell ^ 64 Fig. 5-8: Molecular structure of Triton X-100 (n —10) ^ 65 Fig. 5-9: Phase diagram of Triton X-100/H 2 0 [Beyer, 1982, Alekseev et al., 1997] ^ 65 Fig. 5-10: Glass cell used for electrochemical testing ^ 69 Fig. 5-11: Superficial current density at 0 V vs. MSE as a function of CH3OH and H2SO 4 concentration. Scan rate = 5 mV s -I , T = 298 K ^ 70 Fig. 5-12: Blank scan and Cu stripping peak for a PtRu surface. Electrolytes: 0.1 M H 2 SO 4 and 2 mM CuSO4-0.1 M H2SO4. Scan rate = 10 mV s -1 , T = 298 K [Green, Kucernak, 2002] ^ 71 Fig. 5-13: Blank and stripping scans conducted at 298 K in 0.1 M H2SO4 and 2 mM Cu50 4 -0.1 M H2SO4, respectively. PtRu on GF-S3 (prepared with 40 %,,,,t Triton X-100, i = 60 A m -2 , t = 90 min, T = 333 K) ^ 72 Fig. 5-14: Blank and stripping scans conducted at 298 K in 0.1 M H 2 SO 4 and 2 mM CuSO 4 -0.1 M H2SO 4 , respectively. PtRu on GF-S3 (prepared with 40 %,,,, t Triton X-100, i = 60 A m 2 , t = 90 min, T = 298 K) ^ 72 Fig. 5-15: Fuel cell test setup and flow diagram ^ .73 Fig. 5-16: MEA components, backing layers and gasket used for fuel cell testing ^ .74 .  Fig. 6-1: Fiber surface (a) and enhanced view (b) of Pt Ru deposit obtained from 4.1 M HC1 electrolyte (1:1 PtRu atomic ratio), deposition at a superficial current density of 200 A 111 -2 applied for 30 minutes ^ 79 Fig. 6-2: PtRu deposit obtained in aqueous electrolyte (1:1 Pt:Ru atomic ratio) without HC1 present using a superficial current density of 200 A m -2 for 30 minutes ^ .79 Fig. 6-3: Coating obtained by electrodeposition in 4.1 M HC1 + Triton X-100 (40 %, t ) electrolyte containing Pt and Ru salt (4:1 Pt:Ru atomic ratio) at a superficial current density of 200 A r11 -2 applied for 30 minutes ^ .80 Fig. 6-4: Fiber surface (a) and enhanced view (b) of irregular coating obtained from aqueous solution without HC1 (1:1 atomic Pt:Ru ratio) containing 40 %,„ 4 Triton X-100 at a superficial current density of 200 A m 2 applied for 30 minutes ^ 81 Fig. 6-5: Electro-oxidation of methanol on PtRu electrodeposited in the presence of 50 °A m Triton X-100 on GF-S3: Comparison between acid (pH 1) and alkaline (pH 9) 1 83 electrodeposition conditions. 0.5 M CH 3 OH-0.1 M H2SO 4 , 5 mV s , T = 298 K ^ Fig. 6-6: Chronopotentiometry of methanol electro-oxidation on PtRu electrodeposited in the presence of 50 Vo wt Triton X-100 on GF-S3: Comparison between acid (pH 1) and alkaline (pH 9) electrodeposition conditions. 0.5 M CH 3 OH-0.1 M H2SO4, -2 10 A m ^ 84 Fig. 6-7: Cyclic voltammograms obtained at various scan rates in 0.1 M H2SO4 (inset) and 0.1 M H2504-0.5 M CH 3 OH, T = 298 K. ^ 85 Fig. 6-8: Cyclic voltammograms obtained at various scan rates 0.1 M H2504-0.5 M CH3OH, T = 333 K ^ 86 Fig. 6-9: Plot of peak current density vs. square root of the scan rate ^ 87 Fig. 6-10: Plot of peak potential vs. logarithm of the scan rate at 333 K ^ .87  x  Fig. 6-11: Particle distribution on the fiber surface (a) and enhanced view of deposit (b) prepared with 40 %,,„ t Triton X-100, applying 50 A m -2 for 30 minutes ^ 89 Fig. 6-12: Particle distribution on the fiber surface for the deposit prepared using 40 %wt Triton X-100, applying 100 A m -2 for 30 minutes ^ 90 Fig. 6-13: Particle distribution on the fiber surface for the deposit prepared with 80 % wt Triton X-100, applying 50 A m 2 for 60 minutes (for Pt and Ru, respectively) ^ 91 Fig. 6-14: SEM image of deposit prepared with 80 % wt Triton X-100, applying 50 A m -2 for 60 minutes (for Pt and Ru, respectively) (a), and enlarged view of crystallites (b) ^92 Fig. 6-15: Particle distribution on the fiber surface for the deposit prepared with 80 %„, t Triton X-100, applying 100 A m -2 for 15 minutes (for Pt and Ru, respectively) ^ 93 Fig. 6-16: Methanol electro-oxidation: Comparison between a catalyst prepared by co-deposition (40 % wt Triton X-100, 50 A m -2 , 30 min) and a catalyst prepared by sequential deposition (80 % wt Triton X-100, 50 A m -2 , 60 min for Pt and Ru, respectively). Linear sweep at 5 mV s -I , 298 K ^ 94 Fig. 6-17: Effect of surfactant concentration, deposition current density and time on the atomic Pt:Ru bulk ratio as determined by EDX for co-deposited PtRu on GF-S6 ^ 95 Fig. 6-18: SEM image of PtRu particles sequentially deposited with 40 %, 4 Triton X-100 utilizing a 0.1 s nucleation pulse (i = 300 A m -2 ) and a 15 minute growth current density (i = 50 A m -2 ) for Pt and a 10 minute current density for Ru (i = 50 A m -2 ) ^ 96 Fig. 6-19: SEM image of a deposit obtained by sequential deposition in aqueous solution without Triton X-100 utilizing a 0.1 s nucleation pulse (i = 300 A m 2 ) and a 15 minute lasting growth current density (i = 50 A m 2 ) for Pt and a 10 minute lasting current density for Ru (i = 50 A m 2 ) ^ .97 Fig. 6-20: Influence of Pt:Ru atomic ratio on methanol oxidation in 0.5 M CH 3 OH —0.1 M H2SO 4 at 298 K, scan rate = 5 mV s -1 . Sequential pulsed deposition. Samples were prepared by a 0.1 s nucleation pulse (i = 300 A m 2 ) and a 15 minute lasting growth current density (i = 50 A m 2 ) for Pt in both cases. For Ru a current density (i = 50 A m -2 ) lasting 7.5 minutes (4:1 Pt:Ru ratio), respectively 15 minutes (1:1 Pt:Ru ratio) was applied^ 98 Fig. 6-21: (a) Voltammogram of methanol electro-oxidation on selected catalyst samples from the factorial experiment. PtRu deposited on GF-S3. 0.5 M CH 3 OH-0.1 M H 2 SO 4 , -1 5 mV s , 298 K. (b) Enlarged view of the methanol oxidation onset. Inset shows comparison with blank scan (0.1 M H 2 SO 4 ) for the best performing sample (40 %wt 100 Triton X-100, 60 A m -2 , 90 min) ^ Fig. 6-22: Voltammogram of methanol electro-oxidation in terms of mass activity on selected catalyst samples from the factorial experiment. PtRu deposited on GF-S3. ... ..102 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s -I , 298 K ^ Fig. 6-23: Voltammogram of methanol electro-oxidation using catalyst i deposited on GFS3: Comparison of Pt and PtRu. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 298 K. Deposition parameters: 40 %,,,, t Triton X-100, 60 A m 2 , 90 min, 298 K ^ 102 Fig. 6-24: Chronoamperometric test at 0 V vs. MSE, 0.5 M CH 3 OH-0.1 M H2SO4, 104 298 K ^ Fig. 6-25: Influence of surfactant concentration, current density and deposition time on the atomic Pt:Ru bulk ratio as determined by ICP-AES ^ .105  xi  Fig. 6-26: SEM image of the PtRu electrodeposited fiber surfaces at 298 K: (a) 60 %,t Triton X-100, 20 A m 2 , 90 min at 298 K, d p 20-50 nm (b) 50 Vo wt Triton X-100, 40 A m 2 , 175 min at 298 K, dp 20-60 nm 2 (c) 40 %, t Triton X-100, 60 A m , 90 min at 298 K, dp 20-60 nm ^ ... 106 Fig. 6-27: Representative morphology of the PtRu electrodeposit on GF-S3 obtained without surfactant (see Table 6-10) ^ 108 Fig. 6-28: Representative morphology of the PtRu electrodeposit on GF-S3 obtained in the presence of Triton X-100 (micellar phase) (see Table 6-10) ^ 108 Fig. 6-29: Voltammogram of methanol oxidation on electrodeposited PtRu using a GF-S3 substrate: Comparison between the catalysts produced by electrodeposition with and without Triton X-100 present. 0.5 M CH3OH-0.1 M H2SO4, 5 mV s I , 298 K^ 109 Fig. 6-30: Methanol oxidation superficial current density at 298 K: Effect of 1 electrodeposition temperature; 0.5 M CH3OH-0.1 M H2SO4, 5 mV s ^ .110 Fig. 6-31: Methanol oxidation mass activity at 298 K: Effect of electrodeposition temperature; 0.5 M CH3OH-0.1 M H2SO4, 5 mV s ^ 111 Fig. 6-32: Methanol oxidation superficial current density at 343 K: Effect of 1 electrodeposition temperature; 0.5 M CH3OH-0.1 M H2SO4, 5 mV s ^ 112 Fig. 6-33: Methanol oxidation mass activity at 298 K: Effect of electrodeposition temperature; 0.5 M CH3OH-0.1 M H2SO4, 5 mV s 1 ^ .112 Fig. 6-34: Chronopotentiometry: Effect of electrodeposition temperature on the catalyst activity for methanol oxidation at 298 K, i = 50 A m -2 , 0.5 M CH3OH-0.1 M H2SO4...113 Fig. 6-35: Chronopotentiometry: Effect of electrodeposition temperature on the catalyst activity for methanol oxidation at 343 K, i = 50 A m 2 , 0.5 M CH3OH-0.1 M H2SO4...113 Fig. 6-36: Micrograph of PtRu electrodeposited on graphite felt at 333 K (see Table 6-11) ^ 114 Fig. 6-37: Voltammogram of pure Pt catalyst on GF-S3, 0.5 M CH3OH-0.1 M H2SO4, -1 5 mV s . Deposition parameters: 40 Vo ws Triton X-100, 60 A m -2 , 90 min, 333 K ^ 114 Fig. 6-38: Chronopotentiometry: Pure Pt catalyst on GF-S3, 0.5 M CH3OH-0.1 M H2SO4, i = 50 A m 2 . Deposition parameters: 40 %, t Triton X-100, 60 A m 2 , 90 min, 333 K^ 115 Fig. 6-39: Linear sweep voltammetry of Pt, Ru and PtRu electrodeposition on graphite felt in 0.1 M H2SO4: (a) without surfactant, (b) the effect of Triton X-100 concentration on the Pt deposition, (c) with 50 %,„- t Triton X-100. Scan rate = 5 mV s I , 298 K ^ 116 Fig. 6-40: Electro-oxidation of methanol on PtRu and PtRuMo electrodeposited on GF-S3 in the presence of 40 %„, t Triton X-100. 0.5 M CH3OH-0.1 M H2SO4, -1 5 mV s , 298 K ^ 118 Fig. 6-41: Electro-oxidation of methanol on PtRu and PtRuMo electrodeposited on GF-S3 in the presence of 40 %,,, t Triton X-100. 0.5 M CH3OH-0.1 M H 2 SO 4 , -1 ^118 5 mV s , 343 K Fig. 6-42: Electro-oxidation of methanol on PtRu and PtRuMo electrodeposited on GF-S3 in the presence of 40 °A m Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H2SO4, -2 i = 50 A m ^ 119 Fig. 6-43 HiRes SEM micrographs of PtRu (a) and PtRuMo (b) deposits on the graphite fiber surface obtained by double deposition at 333 K ^ ..121  xii  Fig. 6-44: Electro-oxidation of methanol on PtRu and PtRuMo double-electrodeposited on GF-S3 in the presence of 40 %,„ t Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H 2 SO 4 , -1 5 mV s ^ .122 Fig. 6-45: Electro-oxidation of methanol on pure Pt double-electrodeposited on GF-S3 in the presence of 40 %,,, t Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H 2 SO 4 , -1 5 mV s ^ 123 Fig. 6-46: Electro-oxidation of methanol on PtRu and PtRuMo double-electrodeposited on GF-S3 in the presence of 40 Vo wt Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H2SO4, -2 i = 50 A m ^ 124 Fig. 6-47: Electro-oxidation of methanol on Pt double-electrodeposited on GF-S3 in the presence of 40 Vo wt Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H2SO4, -2 i = 50 A m ^ 124 Fig. 6-48: Linear sweep voltammetry of Pt (a), Ru (b) and Mo (c) electrodeposition on graphite felt, 5 mV s -I , 298 K ^ 126 Fig. 6-49: Linear sweep voltammetry of Pt (a), Ru (b) and Mo (c) electrodeposition on graphite felt, 5 mV s -I , 333 K ^ .127 Fig. 6-50: Linear sweep voltammetry of PtRu and PtRuMo co-deposition on graphite felt, -1 5 mV s , 298 K ^ 128 Fig. 6-51: Linear sweep voltammetry of PtRu and PtRuMo co-deposition on graphite felt, 5 mV s 1 , 333 K ^ 129 Fig. 6-52: Linear sweep voltammetry with blank graphite felt in 0.1 M HC1 at different .130 surfactant concentrations, 5 mV s 1 , 333 K ^ Fig. 6-53: Tafel plots for PtRu (a) and PtRuMo (b) double-deposited on GF-S3 at 333 K, 0.5 M CH3OH-0.1 M H2 SO 4 ^132 Fig. 6-54: Micrograph of recast Nafion on GF-S6 type graphite felt ^ .134 Fig. 6-55: Voltammogramm comparing methanol electro-oxidation at 298 K with and without Nafion coating. 0.5 M CH 3 OH-0.1 M H2SO 4 , 5 mV s -1 . Substrate: GF-S6, thickness 2000 IAM ^ .135 Fig. 6-56: Chronopotentiometric scans on PtRu catalyzed GF-S3 with and without Nafion-1 M H 2 SO4 coating (-30 g m -2 ). 0.5 M CH3OH-0.1 M H 2 SO 4 , 5 mV s -1 . Prepared in 3 ml sandwich plating cell ^ ...136 Fig. 6-57: DMFC polarization curves obtained at 333 K; anode: 1 M CH 3 OH, 5 ml min -I , —100 kPa(abs); cathode: Pt black (40 g m -2 ) dry 02, 500 ml mid i STP, —200 kPa(abs)136 Fig. 6-58: Molecular structure of Kelzant Xanthan gum.. ^ .137 Fig. 6-59: SEM micrograph of graphite fibers coated with Kelzan film ^ 138 Fig. 6-60: Voltammogram measured in 0.5 M CH3OH — 0.1 M H 2 SO 4 , Scan rate = ^ .138 5 mV s -I . Effect of acidified Kelzan coating on the fibers Fig. 6-61: Distribution of dried silica gel on graphite fibers ^ ...139 ^ 139 Fig. 6-62: Distribution of dried silica gel on graphite fibers (continued) SO , Scan rate = Voltammogram measured in 0.5 M CH3OH — 0.1 M H 2 4 Fig. 6-63: 5 mV s -I . Effect of presence of acidified SiO 2 gel on the fibers ^ .140 Fig. 6-64: Acid leaching from SiO2 gel over time. Y-axis: Acid content in the felt normalized with respect to initial acid concentration, T = 298 K ^ 141  Fig. 6-65: DMFC polarization curves obtained at 333 K; anode: 1 M CH3OH, 5 ml min d , —100 kPa(abs); cathode: Pt black (40 g m -2 ) dry 02, 500 ml min d STP, —200 kPa(abs). Effect of acidified Si02 gel coating on the fibers ^ .142 Fig. 6-66: DMFC polarization curves obtained at 333 K; anode: 1 M CH 3 OH, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs). Effect of sulfuric acid concentration in anolyte ^ 144 Fig. 6-67: Variation of the cathode pressure; anode conditions: 1 M CH3OH-0.5 M H2 SO 4 , 2 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP ^ 146 Fig. 6-68: Effects of flow rate variation at 333 K; anode: 1 M CH3OH-0.5 M H2SO 4 , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs) ^ .147 Fig. 6-69: Effect of anode thickness on fuel cell performance using a serpentine flow field; anode: PtRu deposited on GF-S3, 1 M CH3OH-0.5 M H2SO4, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs); T = 333 K ^ 158 Fig. 6-70: Effect of anode thickness on fuel cell performance using a flow-by design; anode: PtRu deposited on GF-S3, 1 M CH 3 OH-0.5 M H2SO4, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs); T = 333 K ^ .149 Fig. 6-71: Fuel cell polarization experiments: Comparison of PtRu deposited on GF-S3 and commercial CCM design; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs); T = 333 K ^ .151 Fig. 6-72: Power density plots expressed in terms of mass activity; anode: 1 M CH3OH0.5 M H2SO4, 5 ml mind , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs); T = 333 K. ^ 152 Fig. 6-73: Polarization plots for PtRu prepared at 333 K using 40 %,,,, t Triton X-100, i = 60 A in -2 : Comparison of single step and a two-step codeposition; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs); T = 333 K ^ 154 Fig. 6-74: Power density plots expressed in terms of mass activity for PtRu prepared at 333 K using 40 Vo wt Triton X-100, i = 60 A ril -2 : Comparison of single step and a two-step codeposition; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs); T = 333 K ^ 155 Fig. 6-75: Fuel cell polarization tests comparing PtRu with. PtRuMo; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min d , —100 kPa(abs); cathode: dry 02, 500 ml min d STP, —200 kPa(abs) ^ .156 Fig. 6-76: Cube plot of peak power density as a function of methanol concentration, flow rate and temperature: PtRu; serpentine flow-field ^ ..159 Fig. 6-77: Cube plot of peak power density as a function of methanol concentration, flow rate and temperature: PtRu; flow-by mode.. ^ ..159 Fig. 6-78: Cube plot of peak power density as a function of methanol concentration, flow rate and temperature: PtRuMo; serpentine flow-field ^ .160 Fig. 6-79: Cube plot of peak power density as a function of methanol concentration, ..160 flow rate and temperature: PtRuMo; flow-by mode ^ Fig. 6-80: Fuel cell deactivation behavior of PtRu and PtRuMo; anode: 1 M CH3OH0.5 M H 2 SO4, 5 ml min d , —100 kPa(abs), serpentine flow; cathode: dry 02, 500 ml min d ^ 163 STP, —200 kPa(abs)  xiv  Fig. 6-81: Repeated fuel cell deactivation tests with a PtRuMo anode; anode: 1 M CH3OH-0.5 M H2SO4, 5 ml min -1 , —100 kPa(abs), serpentine flow; cathode: dry 02, 500 ml min -1 STP, —200 kPa(abs) .163 Fig. 6-82: Deactivation tests at 353 K using 2 M CH 3 OH-0.5 M H2SO4 at a flow rate of 2 ml min -1 ; anode: —100 kPa(abs); cathode: dry 02, 500 ml min -1 STP, —200 kPa(abs).164 Fig. A-1: SEM micrograph of PtRu catalyst in the center region of the felt obtained with 50 %,,,, t Triton X-100 and two counter electrodes ^ 184 Fig. A-2: SEM micrograph of PtRu catalyst in the center region of the felt obtained with 50 %,,t Triton X-100 and one counter electrode ^ .184 Fig. A-3: Electro-oxidation of methanol on PtRu electrodeposited in the presence of 50 °A m Triton X-100 on GF-S3: Comparison between electrodeposition with one and two counter electrodes. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s 1 , 298 K ^ ..185 Fig. A-4: Electro-oxidation of methanol on PtRu electrodeposited on GF-S3 in the presence of 50 %,, t Brij 56: Comparison between electrodeposition with one and two I counter electrodes. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 298 K ^ 185 Fig. A-5: Electro-oxidation of methanol on PtRu electrodeposited in the presence of TEAH (10 -3 M) on GF-S3: Comparison between electrodeposition with one and two counter electrodes. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 298 K ^ ...186 Fig. A-6: Micrograph showing substantial dendrite growth on graphite felt coated with PtRu utilizing TEAH ^ 186 Fig. B-1: Electro-oxidation of methanol on PtRu electrodeposited on GF-S6 by pulsed current plating in 0.5 M CH 3 OH. Effect of surfactant presence in plating bath. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 298 K ^ 188 Fig. B-2: Electro-oxidation of methanol on PtRu electrodeposited in pulsed or galvanostatic mode. 0.5 M CH3OH-0.1 M H2SO4, 5 mV s , 298 K ^ 190 Fig. B-3: Electro-oxidation of methanol on PtRu electrodeposited in pulsed or galvanostatic mode. Mass activity. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s -1 , 298 K^ 190 Fig. B-4: Electro-oxidation of methanol on PtRu electrodeposited at 298 K in pulsed or galvanostatic mode with 60 Vo wt Triton X-100. i = 50 A m ^ 191 Fig. B-5: Electro-oxidation of methanol on PtRu electrodeposited at 298 K in pulsed or galvanostatic mode without surfactant i = 50 A m2 ^ 191 Fig. E-1: DSC test of catalyzed GF-S3 fibers prepared without surfactant ^ 197 Fig. E-2: DSC test of catalyzed GF-S3 fibers prepared with 40 °A m surfactant ^ 198 Fig. E-3: DSC test of catalyzed GF-S3 fibers coated with bulk 40 % mit surfactant ^ 198 Fig. F-1: Example of EDX spectrum obtained in point scan mode for PtRu electrodeposited on GF-S3 with 40 Vo wt Triton X-100, 60 A m -2 , 2 x 90 min at 333 K..200 Fig. F-2: Example of Auger spectrum obtained for PtRu electrodeposited on GF-S3 with ...201 40 Vo wt Triton X-100, 60 A m -2 , 90 min at 333 K ^  XV  Fig. G-1: Polarization plots for PtRu, serpentine flow (runs #1-9) ^203 Fig. G-2: Polarization plots for PtRu, replicability assessment, serpentine flow ^204 Fig. G-3: Polarization plots for PtRu, flow-by mode (runs #1-9) ^ .....206 Fig. G-4: Polarization plots for PtRu, replicability assessment, flow-by mode ^ .207 Fig. G-5: Polarization plots for PtRuMo, serpentine flow (runs #1-9) ^ 209 Fig. G-6: Polarization plots for PtRuMo, replicability assessment, serpentine flow ^210 Fig. G-7: Polarization plots for PtRuMo, flow-by mode (runs #1-9) ^ 212 Fig. G-8: Polarization plots for PtRuMo, replicability assessment, flow-by mode.. ^ 213  xvi  Nomenclature A^Geometric area^  [m2]  AA^Pre-exponential  -  factor^  Acs^Cross sectional area^  [m2]  a^Activity of chemical species ^ [atm] or [M] as^Specific area^  [m2 in -3 ]  b^Tafel slope^  [mV dec d ]  c^Concentration^[mol m -3 ] or [M] or [mM] or [% wt] cb or Co^Bulk  concentration^  [mol m -3 ]  c s^Concentration in boundary layer near solid surface [mol m-3 ] D^(Effective) Diffusion coefficient ^[m2 s -1 ] D o^Diffusivity in bulk solution^ [m2 s -i ] d p^Particle diameter^  [nm]  d f^Diameter of single fiber^  [Ilm]  E^Electrode potential^[V vs. SHE] or [V vs. MSE] E i^Initial potential set for cyclic voltammetry ^[V vs. SHE] E2^  Highest potential set for cyclic voltammetry^[V vs. SHE]  E a^Activation energy^  [J moll  Ecen^Fuel cell potential^  [V]  E °^Standard potential^  [V]  E ° 1^Standard potential at temperature T1^[V] E °2^Standard potential at temperature T2^[V] Ee, anonde^Equilibrium  anode potential^ [V]  Ee, cathode^Equilibrium  cathode potential^[V]  Ee, cell^Equilibrium  cell potential^  Ee, electrode^Equilibrium  electrode potential ^[V]  [V]  Eoc^Open circuit voltage (full cell)^[V] E a °^Open circuit potential (anode) ^[V] E c °^Open circuit potential (cathode)^[V] F^Faraday constant (96485)^  [C mol -1 ]  xvii  [ml min -1 ]  F*  Anolyte flow rate  f  Electrolyte content in total electrode volume  G  Superficial gas load  HI  Liquid hold-up  i  Superficial current density  [A m 2 ]  io  Exchange current density  [A m 2 ]  i*  Local current density  [A m 2 ]  'kinetic  Kinetic current density  [A m -2 ]  iL  Mass transfer limiting current density  [A m-2 ]  1p  Peak current density  [A m 2 ]  k  Reaction rate constant  km  Mass transfer coefficient  [m s 1 ]  L  Superficial liquid load  [kg m2 s-1]  M  Electrochemical species  n  Number of electrons  N'  Gas evolution rate  nmax  Maximum number of electrons reacting per mol  [kg m" 2 s -1 ]  [mol s -1 ]  of methanol (6) p  (Peak) Power density  [W m -2 ]  R  Gas constant  [J ma i K -1 ]  T  Temperature  [K]  t  Time  [s] or [min]  to  Time of initiating current or potential step  [s]  uo  Velocity of single rising bubble  [m s' i ]  Ug  Superficial gas velocity  [m s 1 ]  14  Superficial liquid velocity  [m s 1 ]  V'  Volumetric gas flow rate  [m 3 s 1 ]  V'1  Volumetric liquid flow rate  [m 3 s -1 ]  x  Felt thickness  [m]  a s , ac  Anodic and cathodic transfer coefficients  Y  Activity or fugacity coefficient  xviii  6  Diffusion layer thickness  [ni]  AEOhm  Ohmic voltage drop  [V]  AG  Gibbs free energy change  [J moi l ]  AG °  Standard Gibbs free energy change  [J moll  Ali c  Enthalpy of combustion  [J moll  Al  Inter-electrode gap  [m]  AS  Entropy change  [J mol -1 K -1 ]  AS °  Standard entropy change  [J mo1 -1 K -1 ]  Ax  Compressed felt thickness  [In]  Ax 0  Uncompressed felt thickness  [m]  E  Gas hold-up  Co  Porosity of uncompressed felt  Ecf  Porosity of compressed felt  ETDmax  Maximum thermodynamic efficiency  Cop  Operational efficiency (including Faradayic efficiency) Surface coverage  11  Overpotential  [V]  us  Surface (kinetic) overpotential  [V]  Tic  Concentration overpotential  [V]  Ko  Electrolyte conductivity at 298 K  [S m -1 ]  Electrolyte conductivity at 353 K  [S m -1 ]  K  Electrolyte conductivity (gas hold-up correction)  [S m 1 ]  K  Effective electrolyte conductivity (felt porosity  [S m -1 ]  Ko  T  and liquid hold-up correction) 0  Surface coverage  o-  Electronic conductivity  v  Stoichiometric coefficient  u  Scan rate  [S m -1 ]  [mV s -1 ] or [V s -1 ]  xix  Abbreviations ABS^Absolute (referring to pressure) BET^Brunauer Emmett Teller adsorption isotherm CCM^Catalyst Coated Membrane CMC^Critical Micelle Concentration DMFC^Direct Methanol Fuel Cell EDX^Energy Dispersive X-Ray Diffraction FTIR^Fourier-Transform Infrared Spectroscopy GDE^Gas Diffusion Electrode ICP-AES^Inductively Coupled Atomic Emission Spectroscopy LEISS^Low Energy Ion Scattering Spectroscopy MCFC^Molten Carbonate Fuel Cell MEA^Membrane Electrode Assembly MSE^Mercury/Mercurous Sulfate Reference Electrode PAFC^Phosphoric Acid Fuel Cell PEM^Polymer Electrolyte or Proton Exchange Membrane RHE^Reversible Hydrogen Electrode SCE^Saturated Calomel Electrode SEM^Scanning Electron Microscopy SHE^Standard Hydrogen Electrode SOFC^Solid Oxide Fuel Cell TEAH^Tetraethylammonium Hydroxide TEM^Transmission Electron Microscopy XRD^X-Ray Diffraction UPD^Underpotential Deposition  XX  Acknowledgements This project was a collaborative effort and I am indebted to several people. My supervisors Professors ElOd Gyenge and Colin Oloman were a constant source of help, providing me with farsighted advice and suggestions to continue the progress of the experimental work. I greatly benefited from your extensive knowledge and experience and was inspired by your dedication and work ethic. I am also grateful for the recommendations by Professors David Wilkinson and Dan Bizzotto related to the experimental work and the revisions of this thesis. Thanks to Horace Lam and everyone at the departmental workshop for assistance with equipment and supplies and to my friends, current and former lab colleagues: Derek Lycke, Tommy Cheng, Vincent Lam, Flora Lo and Mohammed Atwan. I am very grateful for my wife's and my family's constant support over the past five years.  xxi  1.0 Introduction With an ever increasing energy demand by the world population and limited fossil fuel resources, efficient energy generation and use become urgent issues. Fuel cells may therefore play an important role in the near future as power sources in the transportation sector, for portable electronic items or stationary devices for industrial and domestic use. The theoretical maximum energy efficiency of a fuel cell is fairly high (e.g., 83 % for a H2/02 fuel cell at 298 K [Steele, 2001]), but in practice the performance is compromised by limitations that depend on the type of fuel cell and the operating conditions. A fuel cell produces heat and electricity by a controlled electrochemical reaction of a fuel with an oxidant. Such a process was first discovered and described in 1839 by Sir William Robert Grove for a hydrogen-oxygen cell with liquid sulfuric acid as the electrolyte [Grove, 1839]. Any type of fuel cell consists of two electrodes separated by an ion conducting electrolyte. In contrast to batteries, fuel cells do not function as energy storage devices but rather as energy converters and therefore rely on a controlled supply of reactants to the respective electrodes. Grove arranged two platinum electrodes with one end of each immersed in a container of aqueous sulfuric acid. The other ends were separately sealed in containers of oxygen and hydrogen. A current flowing between the electrodes was observed. His rationale was that if an electrical current can be utilized for electrolysis (i.e., splitting of water into H2 and 02 in an electrolyte) then this process could potentially be reversed. However, Grove's apparatus did not produce enough electricity to be useful. In 1932, Francis T. Bacon started exploring the development of fuel cells. He replaced the expensive platinum catalysts with nickel electrodes and used a less corrosive alkaline electrolyte. Alkaline fuel cells were used in the nineteen sixties to provide both electricity and drinking water onboard of manned space shuttles during the Gemini and Apollo programmes. Alkaline fuel cells have been operated for more than 65,000 hours in over 87 flights in manned space shuttles [Hirschenhofer et al., 1998]. In the late nineteen sixties both Siemens and Varta constructed converter stations stations on the basis of alkaline fuel cell technology for television transmitters in the power range of 25 and 100 W [Euler, 1974].  1  After the energy crisis in 1973 the MOONLIGHT-Programme was started in Japan where, in particular, the Phosphoric Acid Fuel Cell and the Molten Carbonate Fuel Cell were promoted [Blomen, Mugerwa, 1993]. The most common fuel cell types are listed in Table 1-1.  Table 1-1: Properties of different fuel cell types. Fuel cell type  Electrolyte  Temperature [K]  Energy Efficiency  Polymer electrolyte membrane  Pefluorosulfonic acid  353 - 393  40-50 %  (PEM)  based solid  Phosphoric acid  polymer  423 - 473  40-70 %  (PAFC)  electrolyte  Alkaline fuel cell  KOH matrix  423 - 473  70 %  Solid oxide (SOFC)  Y doped Zr oxide  873-1273  60 %  Moltent carbonate (MCFC)  Liquid Li,Na,K carbonates  873-1273  70 %  Polymer electrolyte (or proton exchange, (PEM)) membrane fuel cells can be operated with hydrogen gas, short chained alcohols, sodium borohydride, formic acid and other fuels. The proton transporting polymer electrolyte is utilized in the form of a thin (-25-175 gm), permeable membrane, which is commonly composed of a pefluorosulfonic acid compound (e.g., Nafion ®). The attachment of two electrodes (i.e., anode and cathode) to the membrane is referred to as the membrane electrode assembly (MEA). A Pt catalyst is employed on the anode and cathode. Since the Pt catalyst is sensitive to CO poisoning, hydrogen generated from e.g., reforming must be very pure (99.99 % vo l). Often a PtRu alloy anode catalyst is used, as it is more CO tolerant. CO tolerance is also essential when using carbon based fuels, since CO appears as a strongly adsorbed reaction intermediate on the Pt catalyst surface. Hydrogen powered PEM fuel cells are deemed most suitable for vehicles. When assessing the efficiency or emissions of any fuel cell one should include the fuel production (e.g., 'well to wheel efficiency' for cars) as well as the parasitic power consumption (balance of plant).  2  The main challenge for the implementation of the hydrogen fueling infrastructure is hydrogen storage, which requires high pressure (30 MPa at 298 K), low temperatures (-20 K) or the use of weight intensive metal hydrides. Furthermore a network of fueling stations would have to be implemented. Within the fuel cell itself carbon corrosion and peroxide induced membrane degradation can be significant problems. Furthermore, lowering the Pt catalyst load is desirable to reduce cost. High temperature fuel cells like MCFCs and SOFCs are more tolerant towards carbon monoxide poisoning. Internal reforming of hydrogen containing fuels is possible. On the other hand, the high temperature is a safety issue and limits the selection of materials and makes these types of fuel cells mainly suitable for stationary power generation. The solid electrolyte in the SOFC can degrade by forming cracks. The Direct Methanol Fuel Cell (DMFC, a variation of the PEM fuel cell) is considered to be a very promising power source for portable electronic devices, such as laptops, cameras or cell phones. Compared to hydrogen fuel cells, liquid feed DMFCs offer a simpler system design and operation, since no humidification or reforming of methanol to hydrogen is required. The main commercial potential is likely in the sector of portable electronic devices, but the application for vehicles has been proposed as well [McNicol et al., 1999]. A reduction of CO2 emissions by 44 % was estimated for a scooter operating with DMFCs compared to using a two-stroke gasoline powered engine [B. Lin, 1999]. A more detailed description of the DMFC is given in Section 3.1. Fuel cell performance is generally characterized by polarization measurements where the current is altered while the full cell potential is recorded (Fig. 1-1). Multiplying the current density with the respective potential yields the power density (expressed in this work as W m -2 ). The maximum performance of a fuel cell is therefore given by the peak power density. When no current is drawn the discrepancy between the theoretical open circuit voltage and the observed value is mainly due to fuel cross-over, which is significant in direct alcohol fuel cells for example. At low current densities the reaction is under kinetic control, while at higher current densities electronic and ionic resistances become the limiting factors. When the current density is further increased reactant mass transport limitation adversely affects the performance.  3  1.2  1.0 Mass Transfer Limitations  Ohmic Limitations  0.8 -  5— 0.6 ,  Ur 0.4 -  • • •  •  •  •  • •  0.2 -  0.0 0  ^  200  ^  400  ^  i [A rif2  600  ^  • 800  ]  Fig. 1-1: DMFC polarization curve obtained at 333 K; anode: PtRu (-50 g m -2 ) electrodeposited on graphite felt (GF-S3) in aqueous media without additives, 1 M CH 3 OH-0.5 M H2SO4, 5 ml min -1 , —100 kPa(abs); cathode: Pt black (40 g m -2 ) dry 0 2, 500 ml min -1 STP, —200 kPa(abs).  1.1 Production, Properties and Utilization of Methanol The most common feedstock for methanol production is methane, which is contained in natural gas. At pressures of 10 to 20 atm and high temperatures (-1123 K) methane reacts with steam on a nickel catalyst to produce synthesis gas according to the following reaction:  CH4+H20 —> CO + 3H 2^(1-1) Methane can also be partially oxidized with molecular oxygen to produce synthesis gas: 2CH 4 +0 2 ---> 2C0 + 4H 2^(1-2)  4  This reaction is exothermic and the heat can be used to promote the steam-methane reforming reaction. Combining the two processes is referred to as autothermal reforming. The ratio of CO and H2 can be adjusted by using the water-gas shift reaction (equation 1-3) to provide the appropriate stoichiometry for methanol synthesis: CO +H 2 0  4  -  .  CO 2 +H  .  2  (1-3)  The carbon monoxide and hydrogen then react on a second catalyst to produce methanol. The most widely used catalyst is a mixture of copper, zinc oxide and alumina and was first implemented by ICI in 1966 [Cheng, Kung, 1994]. At 50-100 atm and 523 K, it can catalyze the formation of methanol from carbon monoxide and hydrogen with high selectivity CO + 2H 2 > CH 3 0H -  (1-4)  The generation of synthesis gas from methane produces 3 moles of hydrogen for every mole of carbon monoxide, while the methanol synthesis consumes only 2 moles of hydrogen for every mole of carbon monoxide. One way of dealing with the excess hydrogen is to inject carbon dioxide into the methanol synthesis reactor, where it reacts to form methanol: CO2+3H2----> CH 3 OH +H 2 0  (1-5)  Although natural gas is the most economical and widely used source for methanol production, other feedstocks can be used. Where natural gas is unavailable, light petroleum products can be used instead. Methanol can also be produced by converting synthesis gas generated from coal gasification (Lurgi process, invented in 1930), which is about three times more expensive compared to obtaining it from natural gas [Stratton et al., 1982]. Biomass is another potential resource for methanol production. A study conducted by the European Union indicated that methanol can be produced by black liquor gasification in the periphery of pulp mills and sold at a price comparable to gasoline (referring to European gasoline prices) [http://www.iags.org/n052404t3.htm].  5  An approach that is still in the research stage is the production of methanol from CO 2 . CO 2 can be captured from fossil fuel burning plants or the atmosphere with e.g., KOH and potentially converted to methanol electrochemically (so called regenerative fuel cell) [Olah, Prakash, 1999]: CO2+2H20 ---> CH 3 OH + 1 .50 2  (1-6)  Methanol may cause blindness when ingested or absorbed through the skin. Ingesting more than 25-90 ml of methanol may be fatal, compared to 120-300 ml of gasoline [Olah et al., 2006]. Methanol is used to manufacture synthetic textiles, plastics, paints and adhesives as well as acetylsalicylic acid (Aspirin) for example. Methanol can be converted into hydrogen with a reformer. The hydrogen can be used to power PEM fuel cells after purification (i.e., CO removal). Methanol contains more hydrogen than liquid hydrogen per volume: 1 liter of liquid methanol contains 98.8 g of hydrogen at 298 K compared to 70.8 g in liquid hydrogen at 20 K [Olah et al., 2006].  6  2.0 Theoretical Background 2.1 Thermodynamics of Electrochemistry The standard (i.e., 298 K, a = 1) equilibrium potential E ° of an electrochemical system is related to the standard Gibbs free energy: AG°= —nFE°^  (2-1)  The Gibbs free energy of a system, which depends on temperature and the activities of the respective reactants and products is calculated as follows: AG = AG° +RT ln(1-1a, )^  (2-2)  The term in parenthesis is referred to as the reaction quotient and S, denotes the stoichiometric coefficient, which is positive for products and negative for reactants (as defined for an electro-reduction process). Ev ,M ,+ne = 0^ -  (2-3)  For example, the electro-oxidation of methanol on the anode can be expressed as: ^ (2-4) CO 2 + 6H+ + 6e — CH 3 0H — H 2 0 =-- 0 -  The full cell reaction can be expressed as: CO 2 + 2H 2 0 — CH 3 0H —1.50 2 ----- 0  ^  (2-5)  Combining equations (2-1) and (2-2) yields the Nernst equation, which relates the equilibrium cell potential to the standard equilibrium cell potential. E ,-=E° —RT1n(1 1a,')^ -  (2-6)  The activity for a gaseous species is approximated by their partial pressure ([atm]) at low pressures and at adequate dilution the activity of species in solution is expressed by their molar concentrations. The correlation between the standard potential and temperature is:  ( aEo^Aso aT a^nF^  (2-7)  Integration of Equation (2-7) and assuming that AS ° is constant yields a linear correlation between the standard equilibrium potential and temperature:  7  E°2 —E°1 =  AS° T 2 —T 1 ) nF  ^(  (2-8)  The equilibrium potential of a complete cell is described as: E e,cell —E e,cathode —E e,anode  (2-9)  The maximum thermodynamic efficiency is expressed by the ratio of the Gibbs free energy over the enthalpy of combustion: AG TD max —  (2-10)  The operating efficiency of a fuel cell depends on the amount of current that is drawn. Losses in efficiency increase with increasing current density while the power density increases up to a peak value and declines when the current density is further increased. The ratio of the voltage E that is obtained at a certain current during fuel cell operation and the equilibrium voltage E e determines the operating efficiency. To calculate the overall efficiency one must also account for the Faradaic efficiency, i.e., the ratio of electrons transferred over the maximum theoretical number of electrons transferred per molecule (e.g., 6 for methanol electro-oxidation): n E n. E e  op =^TD max  (2-11)  2.2 Current-Potential Correlations The difference between the equilibrium electrode potential and the actual electrode potential is the overpotential II. Its value expresses kinetic and mass transfer conditioned losses. 11 = E —E e,electrode (2-12)  The overpotential related to kinetic limitations is referred to as the surface or activation overpotential. In other words this overpotential is the driving force behind an electrochemical reaction. The following expression for an electrochemical reaction rate was presented by Butler, Erdey-Gruz and Volmer: ( „nl,"  ore  17,^17,  " T^R^]  (2-13)  8  At equilibrium the cathodic and anodic current density have the same value, i.e., the exchange current density, which is proportional to the apparent reaction rate constant. These parameters depend on the catalyst material, temperature and the reaction. The exchange current density also depends on species concentrations and temperature. The dependence of the reaction rate constant on temperature is defined by the Arrhenius equation: —E„)  k =A A e "  (2-14)  In cases of very large or very small overpotentials equation (2-13) can be simplified. For large overpotentials (e.g., 111, > 50 mV) one exponential term will become negligible as it approaches zero. Thus the so-called Tafel equation is obtained (1 electron transfer is assumed): (  s= b log  b= \  o  2.303RT F  (2-15)  Where lal^and lal =a, for anodic and cathodic reactions, respectively. The parameter b is called the Tafel slope. At high current densities b and i o can be obtained graphically from i s vs. log(i) plots. For small overpotentials (below —10mV) the exponential terms in the Butler-Volmer equation can be approximated with a Taylor series expansion. The resulting linear approximation is: (cead-ac)^nF  i =i„ ^Flis= i„ ^  RT^RT  (2-16)  Overpotentials caused by mass transfer constraints are called concentration overpotentials (m). It is assumed that there is a concentration gradient (c b in the bulk solution vs. c s near the electrode surface) across a thin diffusion layer with a thickness O. The current density is therefore linked to the reactant surface concentration by the Nernst diffusion layer model: i  — nFD(c b —e s )  =^  (2-17)  9  When the reactant transport to the electrode surface is much slower than the reaction on the surface, c s approaches zero. This leads to the expression for the mass transfer limited current density. – nFDc h  (2-18)  8 The term  D —  is equivalent to the mass transfer coefficient k m . The concentration  overpotential is related to i L as follows: ti c =  RT  ^ln(1–)^  ^nF^iL  (2-19)  When a reaction proceeds under 'mixed control', i.e., both kinetic and mass transport limitations affect the resulting current, the following equation applies: 1 ^1 ^1  (2-20)  i kinetic^L The potential required to overcome the resistance of the electrolyte between two electrodes installed in parallel at a distance Al is called the Ohmic drop and defines another term that accounts for a loss in the overall cell potential. AE Ohm=  iAl^  (2-21)  This expression is valid only for a uniform electrolyte. Subtracting all kinetic, mass transport related and Ohmic losses from the equilibrium cell potential yields the actual full cell potential: Ecel1=-E e,cell +17 s,cathode 17 s,anode +ri c,caihode 17 c,anode^Ohm  ^  (2-22)  2.3 The Three Electrode Setup Fundamental electrochemical studies, like the assessment of the performance of electrocatalysts for example, are carried out in an electrochemical cell. The cell contains an electrolyte (e.g., sulfuric acid), a working electrode, at least one counter electrode and a reference electrode placed close to the working electrode (Fig. 2-1). The reference electrode is often contained in a so-called Luggin capillary with an electrolyte having sufficient conductivity to reduce Ohmic losses that may distort the measurement.  10  Reference Electrode  Counter Electrode  Luggin Capillary  Fig. 2-1: Three electrode cell configuration. In conjunction with a potentiostat the cell configuration provides controlled current or controlled potential measurements.  2.4 Reference Electrodes The standard hydrogen electrode (SHE) often serves as a reference for half-cell potentials. Standard conditions are a pressure of 100 kPa(abs) for gases and a concentration of 1 M for solutes. The following reaction applies: 211+-1-2e  -  H 2^(2-23)  However, the electrode is not practical for experiments, as it requires a constant stream of H2 gas. Therefore a number of other reference electrode types are employed for most measurements (Table 2-1). Several chloride based reference electrodes can be used, but when studying Pt based catalysts in sulfuric acidic media the mercury/mercurous sulfate reference electrode is most suitable. The electrolyte is typically saturated aqueous K 2 SO4 and the measured potential is 0.642 V versus the standard hydrogen electrode at 298 K and 1 atm.  11  Table 2-1: Reference electrodes and their respective half-cell potentials. Reference electrode type^  Potential at 298 K and 1 atm with respect to SHE [V]  Standard hydrogen electrode^SHE^0.000 Reversible hydrogen electrode^RHE^0.00 (for p }{2 -= 1 00kPa(abs) and a(H + )= 1 M*) Saturated calomel electrode^SCE^0.242 Silver/Silver chloride electrode^Ag/AgC1^0.222 Mercury/mercurous sulfate electrode ^MSE^0.642 * RHE potential depends on partial pressure of  H2 and  pH of the electrolyte  (activity a = yc; y = activity coefficient (or fugacity coefficient for gaseous species), c = concentration).  2.5 Fundamental Electrochemical Research Techniques All techniques discussed in Section 2.5 are carried out using a variant of the three electrode setup described in Section 2.3.  2.5.1 Cyclic voltammetry This technique represents a non-steady state measurement of current resulting from an electrochemical process, such as the electro-oxidation of a chemical species. The electrode potential of the working electrode is increased at a constant rate (i.e., scan rate) from an initial value E1 to a maximum voltage E2 and then decreased at the same rate until the initial potential is reached [Vielstich et al., 2003]. The corresponding sweeps can be referred to as anodic and cathodic (i.e., the potential is increased from E 1 to E2 and vice versa). The graph on the right hand side in Fig. 2-2 describes a reversible electrochemical process, such as the conversion of ferrocyanide (Fe 2+ ) to ferricyanide (Fe 3+ ) on the anodic sweep and vice versa on the cathodic sweep. Fig. 2-3 shows a cyclic voltammogram for a non-reversible reaction, such as methanol electro-oxidation.  12  E Di vs. HE  t [ s]  1^  Fig. 2-2: Cyclic voltammetry: Variation of potential over time and related current  response for a reversible electrochemical reaction.  For electrochemical oxidations often the onset potential, marked by a distinct increase in current density, and the peak potential and peak current density are of interest. Observing the peak current density at different scan rates can give information about whether a reaction proceeds under diffusion control. A linear correlation between the peak current density and the square root of the scan rate indicates that the reaction is diffusion controlled. The following form of the Randles-Sevcik equation describes reactions that proceed under mixed diffusion and kinetic control in general [Bard, Faulkner, 2001]: 0 5 n (2.3Rly2 2 c I) 0 bF^  I P^1^  y2 V  (2-24)  Where u is the applied scan rate during the voltammetric measurement. An example for methanol electro-oxidation under diffusion control demonstrated by cyclic voltammetry is presented in Section 6.1.2.  13  rwg  CV  E  E[V vs. SHE]  4  E2  Fig. 2-3: Cyclic voltammogram representing a non-reversible electrochemical reaction.  2.5.2 Chronopotentiometry This method is considered valuable for catalysts that operate at a constant current density over time, as in certain fuel cell applications. The current is increased to a constant value at a time t o and the resulting electrode potential is measured [Bard, Faulkner, 2001] (Fig. 2-4). A decrease in catalyst activity, reflected in a potential increase, can result from poisoning or structural degradation of the catalyst, as well as reactant depletion or other changes occurring in the electrolyte.  E  A  t [s]  to  to  Fig. 2-4: Chronopotentiometry: Current step and resulting potential response.  14  2.5.3 Chronoamperometry  The current is measured at a discrete potential over time. The voltage is stepped up from e.g., the open circuit value at a time t = to (Fig. 2-5). The current response recorded after t = to indicates that a reaction is under diffusion control when there is a linear relationship between i and f  t [ s]  °  5.  ^  t [s]  to Fig. 2-5: Chronoamperometry: Potential step and resulting current response.  The corresponding current-time dependence is expressed by the Cottrell equation [Bard, Faulkner, 2001]: i = nFc 0  D 7T • t  (2-25)  When the reaction does not proceed under diffusion control, chronoamperometry can be carried out to determine the Tafel slope for the kinetic regime. Current densities at a quasi-steady state are measured at various potentials. Plotting the potential E versus log(i) yields the Tafel slope `17,' (see equation 2-15). Furthermore the exchange current density can be determined graphically in the same plot.  15  3.0 Literature Review 3.1 The Direct Methanol Fuel Cell (DMFC) Utilizing methanol for fuel cells is advantageous due to its low price (— $ 200-300 per ton [Olah et al., 2006]). The handling, storage and distribution of this short-chained alcohol are relatively uncomplicated compared to hydrogen. When considering methanol fuel cells for transportation it is interesting to note that the existing distribution network for gasoline could be used with minor modifications, e.g., making fueling equipment parts and tanks more corrosion resistant. One of the most important characteristics of methanol is its relatively high energy density, which is a key aspect for portable applications (Table 3-1). Table 3-1: Mass and volumetric energy density of compressed hydrogen, liquid hydrogen, gasoline, and methanol [Zittel et al., 1996]. Mass Energy Density^Volumetric Energy Density [Wh kg -1 ]^[Wh 1 -1 ] Hydrogen (298 K, 30 MPa)^33300^ 750 Liquid Hydrogen (20 K)^33300^ 2360 Gasoline^ 12700^ 8760 Methanol^ 5600^ 4420 Fuel^  The commonly employed membrane electrode assembly design is very similar to that in the hydrogen PEM fuel cell. The anode consists of a thin (-15-50 pm) catalyst layer where methanol reacts and a porous diffusion layer (-100-300 pm thick) for the transport of the reactants and products. However, alternative designs have been proposed to overcome the challenges that are specific to the DMFC (see Section 3.6.1). One major issue is the permeation (or cross-over) of unreacted methanol through the membrane, which causes a loss in efficiency, as the fuel cannot be fully utilized for electro-oxidation within the anode compartment. Furthermore, methanol may react with oxygen directly on the cathode, therefore causing a mixed potential and likely poisoning of the Pt catalyst. The electro-oxidation of methanol is sluggish, mainly due to the formation of strongly adsorbed reaction intermediates, especially carbon monoxide, on the platinum surface. To address this issue additional metals are employed. Ru is commonly accepted as the most 16  effective co-catalyst for a bimetallic composition. Often a bimetallic alloy is applied. Methanol electro-oxidation catalysis is the most extensively studied field related to direct methanol fuel cells and is further explained in Section 3.2. When methanol is oxidized completely carbon dioxide is produced, which is only partially soluble in the aqueous methanol solution that is fed to the anode. The resulting gas bubbles cause a decrease of the conductivity in the liquid phase, which implies a higher Ohmic potential loss. Therefore effective disengagement of carbon dioxide gas bubbles is desirable. The commercialization of DMFCs is in its early stages. Portable power supplies weighing 7.5 kg with capacities ranging from 600-1600 Wh per day are manufactured by the SFC (smart fuel cell) company. [http://www.efoy.de/index.php?option=com_content&task=view&id=13&Itemid=56]  3.2 Methanol Electro-Oxidation Reaction Mechanism The rate of the methanol electro-oxidation on platinum and platinum based alloys or nanoparticle catalysts has been studied intensely since the 1960s with the aim of utilizing methanol for direct fuel cells. Fundamental electrochemical studies (originated in the work of Petrii et al. [Petrii et al., 1968, Watanabe et al., 1975]) corroborated by direct methanol fuel cell tests showed that PtRu catalysts are more active than pure Pt. The Ru effect has been explained 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, [Iwasita, 2002]) and hydroxyl radicals (OH ad). Gasteiger et al. invoked a statistical model in conjunction with potentiodynamic and chronoamperometric experimental results to show that the Ru promotion is due to the formation of ()H ad on the Ru surface followed by the reaction of ()H ad with adjacent CO ad on Pt (bifunctional catalysis [Gasteiger et al., 1993]). Additionally, electronic effects induced by Ru on the Pt electron conduction band were proposed, that weaken the Pt—CO bond and decrease the activation barrier for CO surface diffusion by reducing back-donation and / or facilitate the dissociative adsorption of CH 3 OH on Pt. The complete methanol oxidation reaction and the reduction of oxygen as well as the overall cell reaction are shown below. The open circuit potentials refer to T = 298 K.  17  Anode:  ^  Cathode:  CH3 OH + H 2 O —> CO 2 + 6H + +6e -^E° = 0.04 V vs. SHE (3-1) 2O  2  + 6H + 6e - —> 3H 2 0^E° =1.23 V vs. SHE (3-2)  Overall:^CH3OH +  02 2 H 2 0 + CO2^E° =1.19 V vs. SHE (3-3)  Methanol electro-oxidation starts with the adsorption and (stepwise) oxidative dehydrogenation on platinum [Iwasita, 2002]. Which hydrogen atom is removed first depends on the electrode potential [Wieckowski et al., 2005]. Reactions (3-4) and (3-5) denote the dehydrogenation steps for the carbon-bonded and oxygen-bonded hydrogen atom, respectively. CH 3 0H (ad) —> CH 2 0H (ad) + H + + e -^(3-4) CH 3 OH (ad) ---> CH 3 0(ad) +^+ e -^(3-5)  After complete dehydrogenation strongly adsorbed carbon monoxide remains on platinum. CH3OH + Pt —> Pt —CO + 4H+ + 4e ^(3-6) -  The Pt-C bond is strengthened by back donation of electron density from the platinum d orbitals to the unoccupied 2,t* orbital of the CO molecule [Arico et al., 2001, Iwasita, 2002]. To oxidize adsorbed CO molecules, surface bonded oxygen as contained in OH groups formed from dissociative water adsorption, must be present [Friedrich et al., 2002]: ^  Pt +H 2 0 —> Pt — OH (ad) + H+ +e -  (3-7)  On pure Pt, water discharge (i.e., water splitting and formation of Pt-OH on the catalyst surface) requires high anodic potentials (> 0.7 V vs. SHE at T = 298 K) [Arico et al., 2001, Waszczuk et al., 2002]. A temperature increase of 20 K (in the range of 333 K to 373 K) resulted in a 0.05 V downshift of the CO stripping peak potential and the DMFC anode potential when monitored at a discrete methanol oxidation current density [Dinh et al., 2000]. Pt — CO + Pt — OH (ad) —> 2Pt + CO 2 +H + +1e - (3-8) Diffusion of surface bonded CO from platinum towards OH adsorbed on the adjacent sites was reported to be an essential step of the reaction mechanism  18  [Waszczuk et al., 2002]. Destabilization of the strongly adsorbed CO molecule is brought about by the presence of ruthenium due to the ligand effect, which will be described below. The reaction mechanism description given above is a simplification of the methanol oxidation process. In reality a number of intermediates may appear, such as formic acid or formaldehyde, and various reaction pathways are possible. The schematic shown in Fig. 3-1 indicates possible pathways and (intermediate) products. In this proposed mechanism the initially formed intermediates dehydrogenate further to form CHOH or CH 2 O. The unstable CHOH can react in different ways to form adsorbed CO. Carbon monoxide is converted to CO2 by reacting with adsorbed OH species, which originate from the dissociative chemisorption of water. Pt + H 2 O —> Pt —0H (ad) + H+ + e -  (3  -9)  Pt —00(ad)+ Pt —OH(ad) ----> 2Pt +CO2 + H+ +e -  (3-10)  The intermediate CH 2 O can desorb from the platinum surface. Otherwise it can be oxidized to formic acid or react with water to form the adsorbed complex H2C(OH)2. CH 2 O(ad) —* CHO(ad) + H+ + e -  (3-11)  CH 2 0(ad)  (3-12)  ---> CH 2 O „ (  )  CH 2 O(ad) + OH(ad) ---> HCOOH(ad) + H+ + e -  (3-13)  CH 2 0(ad)  (3-14)  + H 2 O --> H 2 C(OH) 2  19  HCOO H . 0 (  i 20  HCOOH  [CH 3 O]  -2©  0G  ZD'  +H20^H rio/r 2  CH 3OH^  2^ '^'  CH2O  i&  [CH 2 OH]  [CHO] [CHOH]  [ ] — Transient adsorbed intermediates  -1  ) I,^[COH]  Fig. 3-1: Reaction pathways for methanol electro-oxidation [Cao et al., 2005, Khazova et al., 2003, Arico et al., 2001].  Adsorbed formic acid and the complex H2C(OH)2 could also dehydrogenate resulting in 'direct' CO2 formation. HCOOH(ad) --> CO2 + 2H+ + 2e ^(3-15) -  H2  COH)/ 2 (ad ) --> CO2  +  4H+ + 4e  -  (3-16)  Specific intermediates were observed by employing in situ IR spectroscopy. By recording FTIR spectra the formation of HCOOH and CO was traced by Liu et al. [Liu et al., 2003]. The experimental data also indicated potential formation of weakly absorbed intermediates, such as HCHO, HCOOCH 3 and CH2(OCH3)2. However, these species could not be detected with absolute certainty. Linear and bridge bonded carbon monoxide were detected on sputtered PtRu alloy surfaces [Yajima et al., 2004] as well as on pure Pt [Lopes et al., 1991]. Formate intermediates can be linked to platinum by a single oxygen atom or two [Lopes et al., 1991].  20  The enhancement of CO removal by additional metal phases (preferably ruthenium for binary catalysts) is attributed to the bifunctional mechanism and the ligand effect, respectively.  3.2.1 The bifunctional mechanism The formation of surface hydroxides on ruthenium proceeds at potentials that are significantly lower than for platinum. At 298 K the onset of the adsorption of OH species, which originate from water splitting, occurs at –0.3 V vs. SHE for Ru and –0.7 V vs. SHE for Pt [Waszczuk et al., 2002]. Similar results were obtained by Ticianelli et al. who determined the presence of ()H ad at potentials as low as 0.25 V vs. SHE on PtRu alloy and pure Ru electrodes at 298 K [Ticianelli et al., 1989] Ru + H 2 O ---> Ru – OH(ad) + H + + e -  (3-17)  The bifunctional mechanism implies that the Ru sites provide the OH groups, which react with carbon monoxide adsorbed on Pt sites to form CO 2 [Liu, Norskov, 2001]. Ru – OH (ad) +Pt – CO - Ru + Pt + CO 2 + H + + e -  (3-18)  Studies incorporating radioactive C labeling or temperature controlled desorption under atmospheric and ultra high vacuum conditions showed that CO oxidation occurs at lower anodic potentials when Pt is combined with Ru compared to a catalyst that consist only of platinum (e.g., 0.3 V - 0.7 V vs. 0.5 V - 0.7 V (vs. SHE)) The more positive the potential, the faster oxidative desorption of CO occurs. Apart from Ru, several other elements including Sn, W and Mo showed enhancement of methanol oxidation and specifically CO oxidation due to co-catalytic effects when used as Pt based alloys or layers adsorbed on Pt [Arico et al., 2001, Goetz, Wendt, 1998]. Ru is the most effective element when added to Pt, because it provides labile-bonded oxygen at lower potentials than other metals. It has the highest H 2 O binding energy among all transition and coinage metals and chalcogens (e.g., Sn) and strongly favors the O-H bond scission. Water binding energies are 2.6 eV for Ru and 1.8 eV for Pt, respectively [Anderson et al., 1996].  21  3.2.2 The ligand effect The presence of ruthenium alters the electronic properties of adjacent platinum atoms, which may cause weakening of the Pt-C bond. This phenomenon is referred to as the ligand effect [Liu, Norskov, 2001]. The related model asserts that Ru changes the electronic structure of the Pt surface by interacting with the conduction band of Pt. Ru increases the electron density around Pt sites, leading to a weaker chemisorption of CO [Arico et al., 2001, Kang, Lee, 2000]. Electrochemical NMR studies indicated that presence of Ru adjacent to Pt sites leads to a reduction of the Fermi level local density of states of the metal surface and of the crucial 2n* orbital of the adsorbed CO [Tong et al., 2002, Babu et al., 2003]. Out of the 4-6 kcal moi l reduction of binding energy of CO due to Ru presence about 3-5 kcal moi l (i.e.,130-200 mV) were proposed to be related to the bifunctional mechanism, while the ligand effect accounts for about 1 kcal mo1 -1 (respectively 40 mV). Therefore the impact of the bifunctional mechanism was estimated to be about four times higher than that of the ligand effect [Waszczuk et al., 2002]. The ligand effect influences the adsorption of methanol, which is an electron rich polar molecule. Therefore the adsorption is affected by the altered electronic surface structure of platinum resulting from alloy formation with ruthenium. Table 3-2 displays apparent rate constants of methanol adsorption for pure platinum and binary alloy electrode surfaces. Rate constants were determined by monitoring the surface adsorbate coverage over time.  Table 3-2: Experimentally determined rate constants for methanol adsorption in a 10 -2 M CH 3 OH-0.1 M H2SO4 solution at 298 K [Waszczuk et al., 2001]. Potential [V vs. SHE]  0.2  0.2  0.3  0.3  Electrode  Pt  PtRu  Pt  PtRu  10 3 k ad [s - I ]  0.24  1.05  0.43  1.94  According to these measurements the adsorption rate constant was over four times higher when ruthenium was added compared to Pt. Electron transfer from ruthenium to  22  platinum creates a local field on the Pt surface sites, enhancing methanol adsorption and the C-H bond breaking of the CH 3 group (i.e., the positive pole). The presence of ruthenium significantly decreases the methanol adsorption potential [Iwasita, 2002]. Ruthenium deposited on a Pt(111) surface was found to cause a negative adsorption peak potential shift from 0.2 V to 0.16 V vs. SHE relative to the Pt single crystal. The adsorption peak potential was fairly consistent (i.e., —0.16 V vs. SHE) for various atomic Pt:Ru ratios. The lower the Pt fraction the more the peak current density was decreased. A low number of Ru surface sites is sufficient at 298 K, since methanol adsorbs on Pt sites exclusively (see also Table 3-4).  3.2.3 Carbon monoxide oxidation studies  Carbon monoxide is rapidly formed as an intermediate reaction product resulting from methanol dehydrogenation on Pt, which proceeds in exothermic steps [Kua, Goddard, 1999]. In-situ CO stripping voltammetry and voltammetric methanol oxidation studies carried out at temperatures ranging from 333 K to 373 K suggested that the oxidative removal of carbon monoxide from the electrode surface is the rate determining step during methanol electro-oxidation [Dinh et al., 2000]. CO stripping voltammetry utilizing radioactive labeling at 298 K revealed that other intermediates can still be present on platinum at potentials higher than 0.7 V vs. SHE at which all CO molecules are stripped from the surface, regardless of whether the active catalyst material of the electrode was pure Pt or PtRu [Waszczuk et al., 2001]. A study of CO oxidation at different Ru surface fractions in the top atomic layers of a Pt(111) single crystal proved the benefits of adding ruthenium, i.e., the CO oxidation potential was lowered. Ruthenium was deposited on platinum by means of metal vapor deposition followed by annealing at temperatures of up to 900 K, so that Ru was incorporated into the uppermost layers of the Pt surface [Davies et al., 2002]. The ruthenium fraction in the top 3-4 atomic surface layers was altered within a range of 0.18 and 1.34 as measured by XPS. Since this type of analysis corresponds to the top 3-4 atomic layers, it is possible to obtain a coverage that is greater than 1 monolayer. Low energy ion scattering spectroscopy (LEISS) was conducted to measure the ruthenium content in the top surface layer. For a low Ru content  (Oxps = 0.31  or °kiss = 0.16) distinct  23  oxidation peaks were observed at 0.62 V and 0.72 V vs. SHE (Fig. 3-2). Increasing the number of ruthenium sites led to a decrease of the high potential peak, which was reduced to a shoulder of the main peak located at 0.6 V for surface coverage values of Oxps  0.53 and O xps = 1.34 (resp. 0 leiss^0.18 and Oteiss = 0.39). A single CO oxidation  peak, was observed at 0.8 V on a pure Pt(111) surface. The oxidation peak at the lower potential was associated with the reaction taking place at the nearest neighbor Pt sites, which surround the Ru clusters while the second peak was attributed to oxidation on sites further away from the Ru surface atoms. Enhancement of the surface fraction of ruthenium decreases the number of remote sites. Therefore the high potential peak faded with increasing Ru content. Improved CO oxidation was explained based on the bifunctional mechanism. Langmuir-Hinshelwood kinetics, which are characterized by the reaction of two adsorbed chemical species on the catalyst surface, are possibly favored near Ru clusters. Furthermore, CO mobility between sites can be restricted when specifically adsorbed sulfate ions are present (e.g., as in H2SO4 electrolyte). The latter observation is in agreement with data reported elsewhere, suggesting that the adsorbed SO4 2- ions can partially inhibit the methanol oxidation [Iwasita, 2002, Tripkovic, Popovic, 1996]. Other researchers, who prepared the PtRu catalyst by means of spontaneous deposition of Ru on a Pt(111) single crystal carried out CO a d and CH3OH ad stripping under vacuum (temperature controlled desorption T = 110 K-800 K, 15 K s -1 , 10 4 Pa) as well as voltammetric and potential step tests at 298 K and atmospheric pressure in a three-electrode cell to investigate the oxidation of CO and CH 3 OH. The presence of a second CO oxidation peak was explained by slow surface diffusion of CO. The surface diffusion coefficient of carbon monoxide may depend on coverage and can be a function of site [Waszczuk et al., 2002]. Dynamic Monte Carlo simulations confirmed that slow diffusion of adsorbed CO towards remote ruthenium sites is a limiting step and responsible for the occurrence of multiple CO stripping peaks [Koper et al., 1999]. The slow surface diffusion of oxidant (i.e., the OH groups produced by water activation on Ru sites) to non-nearest neighbor sites represents another limiting factor responsible for the slower kinetics associated with the peak at the higher potential in the presence of SO4 2[Davies et al., 2002].  24  rt(111)  1 imA  0.05— au^62 DS Q4 ors s. $.^$ 63^16^E'V  -0.05—  $(xps) -= 0.18 ML; e(leiss) = 0.04ML  0.02  -0.02  e(xps) = 0.3 MIL; 0(kiss)=0.16ML  0.02 LeV  -0.02  0.02 -0. 0  O(xps) 0.53ML; 0(1ciss) =0.18ML 6-9^141  e(xps)= 1.34ML; (*kiss) 0.39ML 0.02 -0.02  Fig. 3-2: Dependence of CO oxidation potential on ruthenium surface coverage at 298 K. Electrolyte: 0.5 M H2SO4. Scan rate: 100 mV s -1 [Davies et al., 2002].  Carbon monoxide was adsorbed onto Pt-based catalysts in a bath of 0.01 M methanol mixed with 0.1 M H2SO 4 . Methanol was then rinsed off the sample and a linear sweep (0.1 mV s -1 ) from 0.16 V to 0.76 V vs. SHE was performed at 298 K while the CO surface coverage was monitored. The maximum coverage (F.) was set to 7.5 x 10  14  molecules per cm 2 at the beginning of each experiment. Table 3-3 contains the CO desorption potential for pure Pt black nanoparticle catalyst and Pt black nanoparticles decorated with Ru. Two different Ru surface coverages were examined. For both PtRu alloys the CO desorption proceeded at lower potentials relative to the Pt catalyst. The downshift was more pronounced at OR„ =- 0.38.  25  ^  Table 3-3: CO desorption range for pure Pt(111) and two different fractions of Ru on a  Pt(111) surface, T = 298 K [Waszczuk, et al., 2002]. CO desorption potential range [V vs. SHE] from 111 ,„=1 to F/F max <0.1 Pt black^Ru decorated nanoparticles^Pt nanoparticles ^ 0.36 - 0.56 0.26 - 0.46 OR,i=0.06 ^ 0.24 - 0.42 O Ru=0 .38 -  3.2.4 The influence of the Pt:Ru atomic ratio and temperature on catalyst activity  The rate of the electro-oxidation of methanol is strongly dependent on temperature, mainly due to the shift of the rate-limiting step from methanol adsorption and dehydrogenation to CO oxidation with increasing temperature [Waszczuk et al., 2002, Koper, 1999, Kua, Goddard, 1999]. Four different atomic ratios were tested to investigate the impact of the catalyst composition on the adsorption of methanol (Table 3-4).  Table 3-4: Methanol adsorption peak current density as a function of the atomic ratio  measured by voltammetry at —0.16 V vs. SHE. Electrolyte: 1 M CH 3 OH-0.5 M H 2 SO 4 , scan rate: 1 mV s -1 , T = 298 K [Iwasita, 2002]. Pt:Ru ratio i [A m -2 ]  6:1 42  3:1 40  1:1 20  1:3 16  The current density, which is proportional to the rate of adsorption, increased with the platinum content. Therefore, one can conclude that large quantities of platinum would account for the enhanced dissociative adsorption. The adsorption peak was detected at 0.2 V vs. SHE for Pt and at —0.16 V vs. SHE for the alloyed catalysts. The lowest Ru content yielded the highest adsorption rate. The electronic properties of platinum on the surface are modified by adjacent ruthenium sites, as to favor the adsorption of methanol [Waszczuk et al., 2001]. Since methanol does not adsorb on ruthenium at temperatures below 313 K [Iwasita, 2002], the presence of larger Ru fractions on the surface led to a decrease of the measured adsorption peak current density. Potentiostatic methanol oxidation studies were carried out by Gasteiger et al. to test different atomic PtRu ratios at 298 K and 333 K [Gasteiger et al., 1994].  26  Presence of multiple Pt sites surrounding one Ru site may be crucial for methanol adsorption. At 298 K a ruthenium content of 7 provides sufficient oxygenated species to oxidize strongly adsorbed CO molecules (Table 3-5). It is known that methanol does not adsorb on Ru at 298 K. Even though at temperatures above 313 K, the activity of ruthenium towards methanol adsorption and dehydrogenation is increased, it is still not as high as that of platinum. Therefore there has to be a balance of platinum and ruthenium content on the catalyst surface. Dickinson and co-workers carried out cyclic voltammetry and galvanostatic polarization measurements on platinum-rich (3:2 atomic ratio) catalysts, which performed better than equimolar compositions at 298 K and vice versa at 338 K [Dickinson et al., 2004] (Table 3-5). At 318 K the PtRu(1:1) catalyst yielded better performance for potentials greater than 0.64 V vs. SHE, while a 3:2 ratio was more favorable at lower potentials.  Table 3-5: Fundamental methanol oxidation studies: Effects of temperature and atomic composition. Preparation  Test method  method  Response  Arc-melting  Chronoamperometry i at 0.4 V vs. SHE  T  Catalyst  [K]  composition  0.5 M CH 3 OH  298  PtRu alloy  i = 0.1 A  + 0.5 M H 2 SO 4  333  (7 %,, o , Ru)  = 0.35 A M -2  298  PtRu alloy  = 0.05 A M -2  333  (46 % nu) , Ru)  = 0.8 A M -2  PtRu(1:1)  0.93 V vs. SHE  Electrolyte  [Gasteiger et al., 1994] Colloid method  Chronopotentiometry  1 M CH 3 OH  298  substrate: Toray  E at  + 1.5 M H2SO4  338  C-paper  80 A gptR. -I  298  PtRu load:  [Dickinson et al., 2004]  338  Result m -2  0.63 V vs. SHE PtRu(3:2)  0.83 V vs. SHE 0.68 V vs. SHE  20 g M -2 Colloid method  Voltammetry, 20 mV s-1  1 M CH 3 OH  substrate:  E at 10 A gPtRu  + 1.5 M H2SO 4  1  Vulcan XC-72 PtRu load:  E at 190 A gptRu  20 g M -2  [Dickinson et al., 2004]  1  318  PtRu(1:1)  0.44 V vs. SHE  PtRu(3:2)  0.29 V vs. SHE  PtRu(1:1)  0.94 V vs. SHE  PtRu(3:2)  1.04 V vs. SHE  27  3.2.5 Ternary and quaternary electrocatalysts Methanol oxidation occurs in several stages, including dehydrogenation, adsorption of CO like species and OH adsorption as well as reaction of CO ad with OH ad. Multifunctional catalysts are applied to increase the rates of each of the different reaction steps [Arico et al., 1995]. PtRuMo was identified as a promising ternary catalyst, as indicated in Table 3-6. Utilizing PtRuOs also yielded improvement compared to PtRu [Waszczuk et al., 2002]. Osmium adsorbs water at slightly more negative potentials than ruthenium, because it is more oxophilic. However, its solubility in the bulk of the platinum based alloy fcc phase is limited and lower than that of ruthenium [Gurau et al., 1998]. A high surface area Pt65Ru250s10 catalyst yielded a —3 times higher mass activity compared to a Pt50Ru50 at 333 K (see Table 3-6). The reaction proceeded in accordance with the bifunctional mechanism. Osmium handling is problematic, since it is highly toxic in its oxidized form ( 0 s04). Another critical requirement for enhanced catalytic activity is the ability of an alloyed metal to induce the C-H bond split. The Ir-C bond strength is close to that of the Pt-C bond [Gurau et al., 1998]. While the reaction rate for PtRu (1:1) and for the ternary catalyst was of order zero, the quaternary Pt47Ru290s20Ir4 catalyst showed a first order rate dependence on methanol concentration. The respective mass activity was several times higher than for the binary and ternary system (see Table 3-6). It was suggested that the high osmium content facilitates fast water adsorption relative to the methanol dehydrogenation. The open circuit potential was 120 mV higher than for the binary catalyst.  28  Table 3-6: Characterization of ternary and quaternary electrocatalysts. Preparation  Test method  Electrolyte  Catalyst  method  Response  Temperature  Composition  BOnnemann  Voltammetry, 10 mV s -1  method  CH 3 OH oxidation  1 M CH 3 OH  PtRu(1:1)  0.5 V vs. SHE  Vulcan XC-72  onset potential  + 0.5 M H 2 SO 4  PtRuMo(1:1:1)  0.4 V vs. SHE  support  mass activity  298 K  PtRu(1:1)  1 A gp, -1  [Neto et al., 2003]  at 0.6 V vs. SHE  PtRuMo(1:1:1)  6 A gp, -1  Electrodeposition  Voltammetry, 5 mV s -1  on polyaniline  + IR spectroscopy  Result  CO ad oxidation  1 M CH 3 OH  PtRu(1:1)  0.38 V vs. SHE  Potential  + 0.5 M HC1O 4  PtRuMo(1:1:1)  0.23 V vs. SHE  (complete) CH 3 OH  298 K  PtRu(1:1)  0.44 V vs. SHE  PtRuMo(1:1:1)  0.33 V vs. SHE  [Lima et al., 2004]  oxidation potential  Microemulsion  Chronopotentiometry  1 M CH 3 OH  PtRu(1:1)  0.59 V vs. SHE  4 g 111 -2 Pt on  E at 100 A M -2  + 0.5 M H 2 SO 4  PtRuMo(4.5:4.5:1)  0.66 V vs. SHE  Vulcan XC-72  298 K  [Zhang et al., 2004] H2 reduction  Chronoamperometry  1 M CH 3 OH  PtRu  2 A M -2  of metal halides  Current normalized to  + 0.5 M H2SO4  PtRuMoOx.  3 A M -2  and transition  active catalyst area  333 K  metal oxides,  at 0.5 V vs. SHE  glassy C support [Jusys et al., 2002] Reduction of metal  DMFC 100 kPa(abs)  Nafion 117  PtRu  salts with NaBH 4  1 M CH 3 OH, 12.5 ml min -1  333 K  PtRuOs  A gmetai I 22 A g ineta , 1  Anode Load:  cathode: air, 400 ml min -i  PtRuIrOs  108 A gineial-1  40 gmetal M 2  Mass activity at  on MEA by  0.3 V vs. SHE  decal method  (anode potential)  8  [Gurau et al., 1998]  3.2.6 The influence of the carbon support on catalyst activity and utilization  Carbon supported catalysts are advantageous, as the amount of metal can be reduced due to higher catalyst dispersion while the electronic conductivity of the electrode remains sufficient for practical applications [McNicol et al., 1999]. In fuel cells, catalyst particles are usually in the nanometer size range and are commonly supported on carbon to enlarge the surface area and reduce the noble metal cost [Neergat et al., 2002]. One 29  way to achieve enhanced dispersion is through chemical or electrochemical oxidation of the carbon support, which produces acidic surface oxides that enable complexing with metal cations [McNicol et al., 1999]. After reduction of these metal-surface complexes, near atomic dispersion and consequently increased reactivity was obtained. Cyclic voltammetry experiments indicated that the oxidized carbon modifies the platinum surface in a way that is similar to the ligand effect. The potentials for hydrogen and oxygen adsorption and reduction were shifted towards more negative values [McNicol et al., 1999]. The blocking of active sites by oxides on the carbon surface must be avoided to maintain the metal-support interaction. Carbon cloth, fiber or felt may also be utilized to achieve high catalyst dispersion [Coutanceau et al., 2004, Oloman et al., 1991]. Carbon nanofibers with surfaces that contain over 95 % edge sites represent a promising but expensive metal catalyst support material. The large fraction of edge sites is conducive to high reactivity when platinum group metals are present as a thin film on such an electrode surface [Terry et al., 2002]. Carbon nanofibers can also be grown on carbon microfibers, which can potentially provide catalyst substrates with exceptionally high surface areas [Sun et al., 2004].  3.3 Preparation of Nanoscale Electrocatalysts To increase the active surface area of the anode catalyst it is essential to reduce the particle size. Particle diameters of —5 nm are considered desirable for alloy catalysts [Coutanceau et al., 2004]. Different methods are available to obtain well dispersed metallic nanoparticles with a narrow size distribution. To prepare a conventional fuel cell electrode the catalyst particles are deposited on a carbon based gas diffusion electrode, or small carbon particles with catalyst dispersed on their surfaces are directly attached to the polymer electrolyte membrane as a thin film either by spray coating [Neergat et al., 2002] or hot pressing [Wilson, Gottesfeld, 1992].  3.3.1 Sol-gel method  The sol-gel method consists of four steps: formation of the hydrogel, aging of the hydrogel, removal of solvent, and heat treatment, which corresponds to catalyst activation. The formation of the hydrogel includes a hydrolysis step and a water or  30  alcohol condensation step. The sol-gel process itself results in a hydrated solid precursor or a hydrogel [Campanati et al., 2003]. The precursors for the synthesis consist of a metal surrounded by various reactive ligands. Metal alkoxides are often employed due to their reactivity with water. The morphology of the resulting solid structure can be modified as it depends on the conditions of the various steps involved (e.g., pH, aging temperature and time). Kim et al. synthesized PtRu particles of 10 nm diameter using a sol-gel method based on homogenous Pt and Ru sols derived from platinum acetylacetonate (Pt(C 5 H 7 0 2 ) 2 ) and ruthenium acetylacetonate (Ru(C 5 H 7 02)3) at 443 K followed by heat treatment in N2 containing 0.1 % mo i 02 at 573 K [Kim et al., 2003]. The active PtRu surface area was 139.6 m2 g -1 and power densities of up to 1400 W m -2 in a 25 cm 2 DMFC (T = 363 K, 1 M CH3OH, 70-110 g m -2 catalyst load) were obtained.  3.3.2 Vapor deposition  Metal vapor deposition can be applied to coat surfaces with zones (islands) or layers of catalytically active metals [Arps et al., 2001, Davies et al., 2000]. Ion sputtering can be used as an additional step in the electrode fabrication process to modify the surface by producing reactive edge atom sites [Hoster et al., 2001, Narayanan et al., 2001]. Sequential pulsed vapor deposition of platinum group metals offers the advantage of more uniform nucleation [Li, 2002]. The precursor and the reactants can be injected sequentially into the deposition chamber, which is claimed to produce deposits with higher purities. 3.3.3 Colloid method  PtRu particles of —2 nm in diameter with a narrow size distribution were obtained by reducing platinum and ruthenium chlorides with lithium borohydride (LiBH 4 ) in an organic phase [Lee et al., 2001]. In the organic phase, colloidal alloy particles are formed and deposited on the carbon particulate support. The reaction occurs in a slurry medium. The extent of alloy formation depends on the type of reducing agent. Borohydrides or hydrazine are commonly used. Particle dispersity can be modified by varying the concentration of added surfactants or alcohols.  31  The colloidal organosol method developed by BOnnemann and coworkers allows preparation of metal particles in the range of 1-10 nm [BOnnemann et al., 1991, 1998, 2002, 2004]. The first step is the addition of potassium triethylhydroborate to a solution of tetraoctylammonium bromate in THF. After 30 minutes of mixing at 298 K KBr precipitates. After 1 hour of mixing the solution is cooled at T = 273 K for 16 h and the precipitate is filtered. To obtain colloidal particles of a certain metal, e.g., platinum, the metal chloride suspended in THF is added slowly (over 3 hours) to the previously prepared tetraoctylammonium-triethyl-hydroborate solution while mixing. The reaction is continued under stirring overnight, thus producing the colloidal metal and hydrogen gas. The tetraoctylammonium-triethylhydroborate acts as a reducing agent of the metal chloride and as a protective shell around the formed metal particles, which prevents agglomeration (Fig. 3-3). Platinum particles of 1-5 nm diameter were obtained [BOnnemann et al., 2002]. The general reaction scheme for the preparation of platinum nanoparticles is [Bonnemann et al., 1991]: PtC1 2 + n(C 8 H 17 ) 4 BEt 3 H —> Pt + n(C 8 H17 )4 C1 + nBEt 3 + / 1 2 H2  (3-19)  Alloy particles can be produced in a similar way using a solution of two or more metal chlorides. Catalytic nanoparticles on carbon particulates were prepared by combining the received colloidal alloy particles with a suspension of Vulcan in THF to produce a supported fuel cell catalyst. The atomic ratio of the different metals in the deposit closely matched that of the respective metal ions in solution at the beginning of the procedure. BOnnemann and coworkers synthesized a variety of metals and alloys, such as PtSn particles of 1.3 nm diameter on silica substrates [BOnnemann et al., 1998]. Koffi et al. used a similar method to disperse PtCr with particle diameters of 4-5 nm onto carbon powder [Koffi et al., 2005].  32  Stabilizing Shell e.g., NR 4 +Cl Metal Core  Fig. 3-3: Shell stabilized metal core.  While this method works well in conjunction with carbon particle supports, experiments carried out by Lycke showed that applying such a colloid method for a graphite felt substrate yielded poor catalytic activity for the electro-oxidation of ethanol, as the obtained PtSn load was only 0.44 g  M -2  deposition facilitating current density of 62.5 A  [Lycke, Gyenge, 2007]. Employing a  M -2  resulted in a catalyst load of 1.8 g  M -2  and a fourfold increase of the ethanol oxidation peak current density. Furthermore, studies by Reetz et al. indicated that applying current in conjunction with a colloid method is beneficial with respect to particle size control [Reetz, Helbig, 1994]. For example the average particle diameter was decreased from 4.8 nm to 1.4 nm when the current density was raised from 1 A m -2 to 50 A m -2 . Another benefit of such a current assisted method is the absence of impurities, such as boron or metal hydrides. 3.3.4 Electrodeposition Electrodeposition allows tailoring deposit surface morphologies by selecting adequate experimental conditions with respect to temperature, deposition bath composition, cell potential and other variables. [Jayashree et al., 2005, Saber et al., 2003, Choo et al., 1995, Elliott et al., 1999, Allen et al., 2005]. 3.3.4.1 Potentiostatic electrodeposition Jayashree et al. electrodeposited Pt nanoparticles on a Au coated Si grid substrate in aqueous media [Jayashree et al., 2005]. The surface morphology varied significantly  33  when employing different deposition potentials. At -0.78 V vs. SHE, a smooth film was observed by SEM while mesoporous particles were obtained at -1.78 V vs. SHE. Electrodeposition can be employed in conjunction with particle size controlling electrolyte additives to synthesize metal nanoparticles. Liquid crystal templating was successfully applied to generate catalytic surfaces with hexagonal nanostructures [Elliot et al., 1999]. Nonionic surfactants were reported to form hexagonal arrangements of tubular micelles in aqueous media depending on concentration and temperature [Beyer, 1982, Alekseev et al., 1997]. The hexagonal nanostructure of such a templating electrolyte is a crucial feature for obtaining highly structured metal deposits (Fig. 3-4). The pore diameter can be adjusted depending on the chain length of the applied surfactant.  20 nin  [50 pm  Fig. 3-4: Schematic model of hexagonal micelle assembly, model structure of resulting deposit and nanoporous Pt deposit on Au coated quartz crystal [Gollas et al., 2000].  The surface morphology of the metal deposits was affected by the deposition potential. At 0.14 V vs. SHE, the hexagonal surface structure was obtained, while a nonuniform unstructured deposit resulted form depositing at 0.04 V vs. SHE (T = 298 K in both cases). At more negative potentials completely disordered deposits were obtained [Elliott et al., 1999]. An increase in temperature yielded a larger specific surface area  (Table 3-7).  34  Table 3-7: Examples of Pt electrodeposition in liquid crystalline media. Surfactant: Octaethylene-glycol-monohexadecyl-ether (Ci6E08). Substrate^Preparation^Specific surface Conditions^Roughness factor Au coated^298 K^ 32 m 2 g -1 quartz crystal^42 %„,t C I6 E0 8 , 29 %,„t H 2 PtC1 6^64.3 [Bartlett et al., 1999]^0.14  V vs. SHE, 0.3 C (Pt surface area enhanced by — 65 % compared to deposition without surfactant) Pt microdisc^298 K^ 6.5 m2 g -1 42 °/0„, t C 16 E0 8 , 29 °A m H2PtC1 6^210  [Elliott et al., 1999]^0.14  V vs. SHE, 6.4x10 -4 C m -2  Au plate  298 K  2.4 x 10 7 m  338 K  3.4 x 10 7 m -1  358 K  4.6 x 10 7 m -1  1  50 % wt CI6E08, 1.9 M H2PtC16 [Elliott et al., 1999]  0.14 V vs. SHE, 6.4x10 -4 C m -2  In this thesis most electrodeposition experiments were carried out in galvanostatic mode as opposed to potentiostatic deposition, since it is difficult to control the working electrode potential over the entire 3D structure. This problem becomes more severe when larger working electrodes (e.g., > 5cm 2 geometric area) are used. If the current efficiency is known, the deposition rate can be controlled by adjusting the current density when using galvanostatic deposition.  3.3.4.2 Pulsed potential deposition When a constant deposition potential is applied, relatively uncontrolled growth of electrocatalyst clusters can occur over time. It is possible to modify the chemical and physical characteristics of the deposit by adjusting the potential to selectively control the deposition process(es). Liu et al. presented a promising approach to achieve a narrow size dispersion range for metal particles by pulsed electrodeposition [Liu et al., 2001]. Particle growth and its dependence on time, diffusion and concentration of metal ions on the surface were described. The growth of a single particle over time is affected by whether it is located near other particles or isolated, in which case the particle would grow relatively 35  large with increasing deposition time as more ions would be accessible within its immediate surrounding. On the other hand, several densely distributed particles compete for available ions and therefore each grows to a smaller extent. In order to create uniformly dispersed Pt particles with a narrow size distribution a cathodic potential (e.g., -0.5 V vs. SHE) was applied for 5 ms to induce extensive nucleation, followed by a more positive potential (e.g., --0.1 V vs. SHE) for a longer time (e.g., 500 ms) to grow the previously formed nuclei. The deposition was carried out in an electrolyte containing 1 mM PtC16 2- in 0.1 M HC1. The mean particle diameter for Pt was 7 nm.  3.3.4.3 Pulsed current deposition This deposition method utilizes periodic current pulses of a rectangular wave form. In general, pulsed current electrodeposition can yield finer grain sizes and a more homogeneous surface morphology compared to galvanostatic deposition [Saber et al., 2003]. There are three characteristic variables related to pulsed current deposition, i.e., the on-time, off-time (or relaxation time) and the peak current density i p . At low peak current densities (such as 100 A m -2 ) the growth of existing crystals is favored while formation of fresh nuclei is limited. The reaction proceeds under kinetic control and therefore, there is sufficient time for the metal ions to diffuse and crystallize at stable positions on the surface such as existing nuclei. When the deposition current density is increased to within the mass transfer limited regime there is not enough time for the ions to reach stable surface sites and fresh nuclei are formed. Examples of pulsed current deposition techniques are summarized in Table 3-8. Saber et al. deposited Zn by pulsed current electrodeposition [Saber et al., 2003]. Operating at increased current densities was beneficial as it resulted in a high adion population and low adion surface mobility thus facilitating the formation of particles with relatively small diameters [Choo et al., 1995]. A soluble Zn anode was used in conjunction with a ZnC12 bath. The substrate was a low carbon steel disc.  36  ^  Table 3-8: Examples of pulsed current deposition of fuel cell catalysts particles. ^Deposit^Current^On-^Off-^Total^dp^Pt^Peak power Substrate^density^time^time^charge^area^density [A m -2 ]^[s]^[s]^[C m -2 ]^[nm]^[m2 g']^[w m -2 ] Pt^500^0.1^0.3^4x10-4^1.5^124.5^320 C black^200^galvanostatic^4x10-4^50^100.9^300 H2 FC  test, T = 343 K^  [Choi et al., 1998]  PtRu(4:1)^200^0.1^0.3^4.2x10-4^5-8^700 C cloth DMFC test: T = 343 K, anonde: 2 M CH 3 OH 200 kPa(abs), 20 g M -2 PtRu cathode: p02= 250 kPa(abs)^  [Coutanceau et al., 2004]  3.3.5 Chemical reduction in microemulsion  A mixture consisting of an organic phase, an aqueous phase, a surfactant and a cosurfactant is applied to form a microemulsion. The metal salt is dissolved in the aqueous regime, which is divided by the organic phase into discrete spaces of nanometer dimensions where the chemical reduction occurs. The microemulsion containing the metal salt is mixed with an identical water in oil microemulsion containing the chemical reducing agent in the aqueous domain, which is commonly hydrazine or sodium borohydride [Boutonnet et al., 1982, Lu, Wang, 2005]. The particle size obtained from the microemulsion method is dependent on the phase structure of the employed microemulsion, which in turn is a function of the salt concentration, composition, chemical nature of the different phases and temperature [Zhang, Chan, 2002, 2003]. An increase in salt concentration will lead to larger particles, while a decrease in water concentration or increasing temperature will result in smaller particles. Zhang et al. synthesized monodisperse PtCo and PtRu nanoparticles (3-5 nm diameter) with a cyclohexane/Triton X-100/isopropanol/water reverse microemulsion [Zhang, Chan, 2002, 2003].  37  3.3.6 Chemical reduction in liquid crystalline electrolytes utilizing nonionic surfactants  The nonionic surfactant Brij 76 (C5814111021) was applied at 298 K for templating to create a uniform nanostructured PtRu (1:1 ratio) surface with a specific area of 86 m 2 g -1 [Attard et al., 2001]. The metal salts were reduced by adding formaldehyde. The pore size can be adjusted by selecting a surfactant with an adequate chain length. Hydrocarbons that act as swelling agents could be added to enlarge pore diameters [Elliott et al., 1999]. The nonionic surfactant Triton X-100 can act as a reducing agent. Silver nanoparticles were produced at 293 K. The reduction in a 0.03 M AgNO3 bath with 50 Vo wt surfactant was carried out for 1 day (mean particle diameter = 3.4 nm), 5 days and 1 month (mean particle diameter = 18.3 nm) [Lee et al., 2002].  3.3.7 Electroless deposition  Deposits from metal salt solutions can be formed without using a power supply. A catalyst substrate (= electrode) is immersed into a solution, which contains a reducing agent functioning as the electron source. The reduction of metal ions and the oxidation of the reducing agent occur at the same electrode-electrolyte interface. The sites on the surface are statistically divided into anodic and cathodic sites. Therefore there is a flow of electrons between these different sites on the substrate surface. The spontaneous electrodeposition of Ru on a quartz crystal supported Pt electrode was carried out in a N2 purged 0.1 M HC1O4 + 0.5 M RuC13 bath [Vigier et al., 2001]. The surface coverage with Ru islands was calculated to be 10 % based on the mass change recorded by a quartz crystal microbalance and assuming 2D growth. The authors also suggested that Ru deposition was favored on Pt sites located remotely from PtRu sites.  3.4 Carbon Dioxide Disengagement from the Anode Accumulation of CO 2 ( g) in the anode assembly of liquid feed DMFCs should be avoided, since CO 2 ( g) can block active catalyst sites. Carbon dioxide bubble accumulation in the anode compartment also lowers the power output of the fuel cell by decreasing the effective ionic conductivity of the anode electrolyte, which implies an enhanced Ohmic  38  loss. Furthermore, the methanol concentration overpotential is increased, and the potential and current distribution in the anode can be highly non-uniform as a consequence. Increased presence of carbon dioxide gas also diminishes the liquid hold-up in the pores located within the diffusion and catalyst layers, thus obstructing the penetration of liquid reactants towards the catalyst sites. In fuel cells the catalyst is normally attached to a carbon support. Commonly this support consists of carbon particles like Ketjen Black or Vulcan XC-72 that are incorporated between the gas diffusion layer and the membrane. In this section the applicability of different carbon based gas diffusion layer materials for DMFCs is discussed with respect to their implications on CO 2 bubble formation and release. Carbon paper was found to be less suitable than carbon cloth [Argyropoulos et al., 1999]. In the carbon paper based anode assembly gas bubbles of 0.8-1.8 mm diameter were detected whereas gas bubbles of 0.6-0.8 mm diameter were observed when carbon cloth was used. The rough texture of the carbon paper causes a lower interfacial tension between the bubbles or slugs and the solid surface. Thus larger quantities of gas accumulated. The carbon paper had 20-50 vim pores and a significant number of blocked passages. The carbon cloth contained 50-100 i_tm pores and a mat area with smaller openings [Argyropoulos et al., 1999]. Fig. 3-5 shows the surface structure of: (a) unteflonized carbon cloth, (b) teflonized (20 %, t ) carbon cloth and (c) teflonized (20 %, t) carbon paper.  39  Fig. 3-5: Photographs of the surfaces of carbon cloth (a, b) and carbon paper (c) backing layers [Argyropoulos et al., 1999].  Gas bubbles moved against the liquid flow when the carbon cloth was used. This is due to gas evolution caused by the anode reaction and electro-osmotic transport of water and methanol. Parts of the carbon surface were made hydrophobic by a Teflon coating. CO 2 could be disengaged more effectively as channels with unobstructed gas flow were provided [Argyropoulos et al., 1999, Scott et al., 2001]. For a 270 cm 2 DMFC operating at 1000 A m -2 and a liquid flow rate of 1 1 min -I the gas fraction in the outlet reached nearly 40 %. Accumulation of bubbles had a strong detrimental effect on the fuel cell performance as opposed to fast moving smaller gas bubbles [Argyropoulos et al., 1999]. Spot design flow distributors were compared with parallel channel flow distributors. The parallel transport cannels were 2 mm deep, 2 mm wide and 30 mm long. The spot design had an array of square shaped openings arranged in parallel rows. Each spot was 2 mm deep, 1.5 mm wide and 1.5 mm long. The gap between two adjacent spots was 2 mm.  40  The spot design yielded better performance due to improved gas management compared to the parallel channel design. Current densities of up to 1500 A M -2 were applied. Increasing the liquid flow rate increased the fuel cell voltage at flow rates ranging from 8x10 -4 to 19x10 -3 m 3 s -I . Higher pressure resulting from the increased flow rate caused compression of the gas bubbles, or they were broken into smaller bubbles. The effects of the current density and the flow rate on the gas removal in a parallel channel flow bed is shown in Fig. 3-6 [Argyropoulos et al., 1999].  (a)  ^  (b)  ^ (d)  (e)  ^  (f)  Fig. 3-6: CO2 evolution patterns in liquid feed DMFCs as a function of current density,  flow rate and gas diffusion layer material [Argyropoulos et al., 1999].  Images (a) and (b) were obtained under the same conditions (i = 300 A m -2 , flow rate = 3.4x10 6 m 3 s -) ), but in experiment (b) a larger cell was used (226 cm 2 active superficial surface area). In the small unit (9 cm 2 active superficial surface area) gas slugs are formed whereas faster moving small bubbles appear in the large cell. In both cells a carbon cloth based gas diffusion layer was employed. Application of carbon paper instead of carbon cloth under the same conditions as in (a) caused a significantly higher accumulation of gas slugs (c). The image shown in Fig. 3-6(d) represents operation under the same conditions as in (a) but at a lower flow rate of 5.1x10 -8 m 3 s -1 , which results in a higher number of gas slugs. Gas formation at different current densities was compared applying a flow rate of 1.9x10 -5 m 3 s -I . A higher current density (800 A M -2 (e) as compared to 500 A M -2 (f)) led to formation of larger bubbles.  41  The evolution of CO2 gas bubbles can be simulated by the decomposition of H202 in the presence of a conventional PtRu catalyst. This approach simplifies the investigation of two phase flows in DMFC flow field designs, as the application of an electrical current is not required [Bewer et al., 2004]. The respective operating current densities are simulated by applying adequate H202 concentrations. Transparent parallel channel and grid flow field types were studied as well as different flow inlet and outlet manifolds. The grid flow field structure allows flow in the horizontal direction in addition to vertical flow. However, gas flow was occurring in the vertical direction exclusively. Compared to the channel design gas bubbles agglomerated to a lesser extent and observed maximum bubble sizes were 6 to 10 times smaller. The bubble velocity and transport were more homogenous in case of the grid design. The greatest fluctuations in flow velocity distribution occurred at the inlet and outlet. The splayed manifold design proved to be most suitable to address this fluctuation problem. To improve the CO2 disengagement, Nordlund et al. investigated the effect of polytetrafluoroethylene (PTFE) additive on the catalyst layer. Whilst the PTFE content (varying from 20 %,,t to 50 Vo wt) improved the gas disengagement, it also diminished the electronic conductivity of the catalyst layer (by about eight times) and compromised the utilization of the catalyst particles by interfering with the ionomer (Nafion) network [Nordlund et al., 2002].  3.5 Methanol Cross-Over One of the most detrimental effects on DMFC performance is caused by unreacted methanol that is transported through the membrane to the cathode. The fuel is crossing over by diffusion due to the concentration gradient between the anode and cathode [Barragdn et al., 2004], and because of electro-osmotic drag caused by protons migrating to the cathode. Potential losses on the cathode can occur, as the reduction of oxygen is hampered and the cathode can be poisoned. The osmotic drag also causes water to cross over, which may cause cathode flooding. In acidic aqueous media the protons are present as hydronium ions, i.e., H30 + , HSO2 + and H904 + species. Compared to ions of similar size the hydronium ions are transported very rapidly. The high mobility is explained based on the so-called Grotthus mechanism  42  where the transport is determined by the rate at which the hydrogen bond between a hydronium ion and a water molecule is formed [Choi et al., 2005] or broken [Agmon, 1995]. The protonic conductivity of Nafion is dependent on its nanostructure and water content as well as on temperature. A fully hydrated state is desirable for achieving sufficient proton transport. One distinguishes between bulk water in the center region of the pores where the proton mobility is relatively high and 'surface water' located near the arrays of SO3 - groups on the pore walls. Strong electrostatic attraction of the SO3 - groups causes the proton mobility to decrease near the pore walls. A detailed transport model was presented by Choi et al. [Choi et al., 2005]. The proton transfer along the pore wall (i.e., between neighboring sulfonic acid sites and the water molecules in between) can be described by the so-called hopping mechanism (Fig. 3-7).  Surface water  PTFE Fig. 3-7: Simplified Nafion structure and proton transport [Choi et al., 2005]. The hopping time from one water molecule to the next in the pore bulk regime (Grotthus mechanism) is about 10 -12 seconds. For comparison, surface diffusion or hopping along the pore walls is about 1000 times slower. Mass transport in the bulk water  43  of the pores is the third major driving force behind the proton transport through the membrane and proceeds almost 4 times faster than Grotthus diffusion.  (c}  Fig. 3-8: Schematic of proton transport for an H90 5 + species according to the Grotthus mechanism. Letters a through d denote oxygen atoms and lines denote oxygen-hydrogen bonds [Agmon, 1995].  Agmon critically reviewed common assumptions regarding the proton transport and emphasized that hydrogen bond cleavage is the rate limiting step for proton mobility. During the proton transport from one water molecule to the next the two water moieties  44  are equivalent in structure [Agmon, 1995]. Fig. 3-8 serves as a schematic describing the proton transport while the H904 + structure changes to HSO2 + (a) and vice versa (c). Since methanol is a polar molecule, the protons that are transported to the cathode side drag methanol with them. The drag coefficient for water was determined to be —2.5 H20/H + and the same value was used for methanol in model calculations for a study of methanol transport through Nafion membranes over a range of temperatures (303 K — 403 K) [Ren et al., 2000]. Elsewhere the methanol drag coefficient was reported to be 4 at 353 K for example [Dohle et al., 2002]. The carbon dioxide content in the cathode exhaust was exploited as a parameter that can indicate the degree of methanol cross-over [Dohle et al., 2002]. CO2 detected in the cathode exhaust is either formed by chemical oxidation of crossed-over methanol with the supplied oxygen or it permeated through the membrane from the anode side. To distinguish between the two sources of CO2 generation, additional half-cell measurements were carried out with an inert N2 gas stream present at the cathode. Since the methanol that crossed over was not entirely oxidized (presumably due to water presence in the cathode catalyst layer and diffusion layer pores), a catalytic burner was employed at the cathode outlet. The methanol permeation rate through Nafion 112 (50 iim thickness) was —2.7 times higher than through Nafion 117 (175 pm thickness) at 358 K when using 1 M aqueous methanol and an 0 2 feed at 250 kPa(abs). To address the cross-over problem membrane materials were structurally and chemically modified (Tables 3-9 and 3-10) [Gurau, Smotkin, 2002]; e.g., by adding silica [Miyake et al., 2001]. In general an increase in the silica content caused a decrease in both protonic conductivity and methanol permeation. It was speculated that incorporation of silica may increase the tortuosity of the methanol transport channels in the membrane [Jiang et al., 2006]. Other researchers explored the addition of phosphomolybdic acid or etching or sputtering of the membrane surface (Table 3-10). A multi-layer membrane where each layer was composed of two different polymers (one of them methyl-sulfonic acid based) was proposed [Son et al., 2006]. The material reduced the methanol crossover by a factor of 6 and 4 compared to Nafion 115 and 117, respectively. Nafion 112, 117, 1035 and 1135 (thickness: 51 iim, 183 i_tm, 89 pm and 89 ,irn) were compared by recording cyclic voltammograms at 298 K using a Pt electrode after  45  ^ ^  methanol was allowed to cross over for 1 h into a blank 0.5 M H2SO4 containing electrolyte [Ling, Savadogo, 2004]. The lowest resulting methanol oxidation currents were observed with Nafion 117, irrespective of the initial methanol concentration in the compartment on the opposite side of the membrane (0.5 M, 1.5 M or 3 M). The concentration of permeated methanol was determined as a function of cross-over time, membrane thickness and type. These studies corroborated the voltammetry results.  Table 3-9: Examples of membrane modification. ^ DMFC  Membrane^Peak power density ^ Parameters^ type [W m 2 ] 3 M CH3OH^298 K^Nafion 115^190 7 M CH 3 OH^  130  3 M CH 3 OH^ 1 % wt LDH*^140 150  7 M CH 3 OH^  3 M CH 3 OH^ 3 %wt LDH*^120 220  7 M CH 3 OH^ MEA: 9 cm 2 cross section, anode: PtRu(1:1) 50 g m -2 , cathode Pt black 50 g m -2 *Nafion / Mg 2+ Al' '(2:1 ) layered double hydroxide composite^  333 K^ 1 M CH 3 OH, 3 ml min -1^5  [Lee, Nam, 2006]  recast Nafion^550  % wt Si02^620  cathode: air, 200 ml min -1^10 % wt Si02^480 p = 100 kPa(abs) ^MEA: 5 cm 2 cross section, anode: PtRu(1:1) 40 g m -2 , cathode Pt black 40 g M-2^[Jiang  300  293 K, air (passive)^Nafion115 4.5 M CH 3 OH^10 tun Si02* MEA: anode: PtRu(1:1) 80 g m -2 , cathode Pt black 80 g  et al., 2006] 360  M -2  Fuel cell operated without pumps (CH 3 OH supply by diffusion)  [Kim et al., 2004] ^ 600 353 K, 2 M CH3OH, 2m1 min -1^Nafion 117  *Si0 2 film vapor deposited on Nafion 115 surface  cathode: 02, 500 ml min .1^38 lig g -1^620 ^ Si0 2 PWA* p0 2 = 160 kPa(abs) MEA: 4 cm 2 cross section, anode: Pt black 15 g m -2 , cathode: Pt black 20 g M -2 ^*Nafion / Si0 2 / phosphotungstic acid composite ^  [Xu et al., 2005]  46  In other work it was also shown experimentally that the diffusive methanol mass flow across a type 117 membrane was lower compared to a Nafion 115 membrane [Barragan et al., 2004].  Table 3-10: Examples of membrane modification (continued). DMFC  Membrane  parameters  type  Power density [w/ m 2 ]  353 K, 2 M CH3OH  Nafion 115  420 at 0.35 V  cathode: air, 300 ml mid i^0.5g PMA*^560 at 0.35 V p = 200 kPa(abs)^  [Sauk et al., 2005]  MEA: 1 cm2 cross section, anode: PtRu(1:1) 40 g m -2 , cathode: Pt black 40 g m -2 *Nafion / ployphenylene oxide composite with phosphomolybdic acid  368 K, 2 M CH 3 OH^Nafion 117^280 at 0.4 V p0 2 = 100 kPa(abs)^plasma etched^400 at 0.4 V Pt sputtered^800 at 0.4 V [Choi et al., 2001] 2  MEA: 9 cm cross section, anode: PtRu(1:1) 45 g  111 -2 ,  cathode: Pt black 16 g  111 -2  3.6 Electrode Design The conventional electrode in a DMFC consists of a catalyst layer and a gas diffusion layer (GDL) (Fig. 3-9). The fuel (i.e., CH 3 OH) is fed through the flow channels that are incorporated in the bipolar plate and penetrates the gas diffusion layer towards the active sites of the catalyst layer. The protons that result from methanol dehydrogenation and water splitting are transported through the polymer electrolyte membrane to the catalyst layer of the cathode where they react with oxygen to form H2O. The effluent on the anode side contains the end product CO 2 (formed in the catalyst layer) and unreacted methanol as well as intermediates, such as formic acid. The catalyst layer is typically prepared by spraying or brushing a catalyst ink, made by mixing of Nafion and noble metal particles, onto the GDL or the proton exchange membrane (PEM) and hot-pressing the catalyst onto the PEM [Lee et al., 2004] (Fig. 3-10). This binding method provides good adhesion and protonic conductivity. The GDL is typically prepared by mixing  47  tetrafluoroethylene (PTFE) with carbon powder and spreading the mixture onto a sheet of carbon cloth or carbon paper [Lee et al., 2004, Chen et al., 2004].  DMFC with Conventional Anode  02 (g)  CH 3 OH (I) Anode Reaction:^Cathode Reaction: CH 3 OH + H 2 O^CO2 + 6H+ + 6e 1.50 2 + 6H+ + 6e ^3H20 -  -  Fig. 3-9: Exploded view of DMFC with conventional anode design.  Song et al. employed the decal transfer method (DTM) (proposed first by Wilson and Gottesfeld [Wilson, Gottesfeld, 1992]) to prepare the catalyst layer instead of the conventional method [Song et al., 2005]. The DTM involves painting of a catalyst ink on a blank decal PTFE, followed by the transfer of the catalyst layer formed on the PTFE to the PEM by hot-pressing at 433-473 K (Fig. 3-10). The DTM method yielded better contact and lower resistance between the PEM and the catalyst layer. The thinner catalyst layer also allowed for improved mass transport of the reactive species. Compared to a conventionally prepared MEA the fuel cell peak power was significantly enhanced from 600 W M -2 to 1400 W m -2 when air was supplied to the cathode and increased from 700 W M -2 to 2300 W  M -2  when oxygen was fed. The catalyst loads and operating  conditions were: anode: 30 g PtRu m -2 , cathode: 30 g Pt m -2 , fuel: 1 M CH 3 OH at 1 ml min -I , air/0 2 pressure: 200 kPa(abs), 363 K.  48  brush  Att spray Diffusion layer  <-= Nation trrentbiilie  cathode  Dit itsion 4- Catalyst layer  MEA Hot press  (a) Nation  NaOH  spray PTFE blank  Catalyst ink preparation  reprotonate -c==  transfer  cry  Catalyst-coated membrane I i,SO 4 solution Na +-Nariot0' membrane + cat. layer NTT. film + cat. layer %b)^(CCM)  Fig. 3-10: Schematic of conventional catalyst preparation method (a) and decal transfer method (b) [Song et al., 2005].  49  3.6.1 Alternative electrode designs  Several types of novel electrode designs have been proposed for DMFC anodes (see summary of examples in Table 3-11). For instance, alternative DMFC anode models were obtained by modifying the fabrication process of the catalyst layer (see Section 3.6), using a titanium mesh catalyst support (Fig. 3-11) [Allen et al., 2005], fabricating porous carbon catalyst support (Fig. 3-12) [Chai et al., 2004] or using carbon nanocoils as a catalyst support material [Kim et al., 2004]. Other researchers prepared assemblies of latex spheres coated with electronically conductive polypyrrole and Pt catalyst on Au plates (Fig. 3-13). The resulting specific active catalyst area was almost twice as large as the area determined for a conventional E-Tek catalyst. Cyclic voltammograms obtained at 298 K and a scan rate of 20 mV s -1 yielded superficial current densities of 600 A  M-2  and 1000 A M-2 at 0.84 V vs. SHE for a commercial E-Tek  catalyst and for polymer bead supported Pt, respectively [Xie et al., 2005]. The Pt load was 2 g m -2 in each case. Alternative electrode designs have been proposed with the aim of providing a more effective transport of reactive species to the catalyst sites and improved disengagement of CO 2 gas bubbles. Therefore, the electrode should have sufficient porosity. A large specific area to allow deposition of highly dispersed catalyst particles is desirable. Other electrode designs for cross-over mitigation were proposed by Wilkinson et al. [Wilkinson et al., 1997]. To favor the reaction of the majority of the methanol content within the anode, catalyst particles were dispersed uniformly within the electrode matrix or concentrated on one surface of the electrode using a sufficiently high loading. Alternatively, an assembly of several catalyst layers parallel to the membrane was applied. In the latter case the main concept was to initiate substantial conversion of methanol at the first layer and further conversion of the remaining reactant content within the other layers closer to the membrane [Wilkinson et al., 1997]. For example five layers of porous electrically conductive sheet material were employed. Three of the layers contained catalyst particles and polymeric binder (i.e., the two outer layers and the layer in the center), while the other two layers did not contain catalyst particles or binder and therefore had a higher porosity to facilitate CO 2 gas disengagement. Sheet materials like carbon cloth or carbon fiber paper were stated to be preferred porous electrode materials.  50  Table 3-11: Comparison of alternative and conventional anode designs. MEA area  Reactants and  Alternative design  Peak power  catalyst load  flow rates  Conventional type  density  and composition  [W m-2]  9 cm 2  PtRu on Ti mesh  600  2 M CH 3 OH + 0.5 MH 2 SO 4  Fig. 3-11  10 g m PtRu(1:1)  12 ml min -1  GDE  cathode:  air, 1 1 mini 1 , 100 kPa(abs)  (Vulcan XC-72)  10 g M -2 Pt  T = 363 K  [Allen et al., 2005]  anode:  1 M CH3OH  cup-stacked carbon  900 (333 K)  nanotubes  1450 (363 K)  GDE  600 (333 K)  (Vulcan XC-72)  670 (363 K)  anode: -2  64 g t11 2 PtRu(1:1) cathode:  02  43 g m 2 PtRu(1:1)  700  [Kim et al., 2004] 2 cm 2  PtRu on porous  anode  2 M CH 3 OH  hexagonal C  30 g 111 -2 PtRu(1:1)  1 ml min -1  support Fig. 3-12  cathode: 50 g 11  2  02,  Pt  500 ml min -I  T = 343 K  2 cm 2  GDE (E-tek) PtRu on C  2 M CH3OH  nanocoils  20 g m 2 PtRu(1:1)  1 ml min -1  GDE  50 g m 2 Pt  02,  500 ml min -1  T = 333 K  1200  [Chai et al., 2004]  anode: cathode:  1700  2200 1500  (Vulcan XC-72) GDE (E-tek)  1200  [Hyeon et al., 2003] In another embodiment (suggested by Wilkinson et al.) gas flow channels were incorporated into the electrode structure. Further, an assembly of four layers of sheet material with catalyst dispersed on either surface of each sheet was described. In either case, the main concept was that a major fraction of the methanol should be consumed within the volume of the respective electrode structure before reaching the membrane, so that the remaining amount of unreacted methanol would be insufficient to cause detrimental effects. An increase in fuel utilization compared to a conventional anode  51  structure was observed by using a multi layer anode (e.g. from —60 % to —80 % at a cell voltage of 0.3 V) [Wilkinson et al., 1997].  Fig. 3-11: Titanium mesh electrode with electrodeposited PtRu nanoparticles (d p — 5 nm) [Allen et al., 2005]. la• *`  "4"  ,  •  -  Fig. 3-12: 3-12: Porous carbon substrate and carbon surface with PtRu nanoparticles (d p — 2-3 nm; most favorable C pore diameter = 25 nm) [Chai et al., 2004].  Fig. 3-13: Micrograph of conductive polymer beads with Pt catalyst [Xie et al., 2005].  52  4.0 Research Objectives and Novelty The use of catalyzed 3-D graphite felt anodes as opposed to the gas diffusion electrode (GDE) structure or a catalyst coated membrane (CCM) is a new approach for direct methanol fuel cells. The 3-D anode with electrodeposited catalyst nanoparticles on the fiber surface could replace the conventional catalyst layer - gas diffusion layer assembly in the fuel cell (Fig. 4-1). With catalyst particles dispersed throughout the graphite felt matrix the reaction zone is enlarged (active layer thickness = —100-300 ttm) compared to the conventional thin catalyst layer (-15-50 tun) - gas diffusion layer assembly (Fig. 3-9). Therefore a substantial fraction of methanol could react on active sites remote from the membrane, which could decrease the fuel cross-over to the cathode. However, sufficient protonic conductivity needs to be established to utilize these sites. The 3-D anode has an additional potential advantage with respect to CO2 gas disengagement. It is proposed that the extended reaction zone together with its high porosity (up to 95 %) and specific surface properties (e.g., gas/liquid/solid contact angle and surface roughness) may aid the disengagement of CO2( g ). The application of catalyzed graphite felt anodes may have several advantages compared to the approach of stacking several catalyst-diffusion layer assemblies in parallel to the membrane as outlined at the end of Section 3.6.1 [Wilkinson et al., 1997]: - Eliminates the contact resistance between the carbon fiber layers - More uniform dispersion of catalyst particles across anode thickness - Potentially lower catalyst load per geometric area. Electrodeposition is proposed as the method of choice to create active catalyst sites on the 3-D graphite felt substrate.  53  DMFC with Graphite Felt 3-D Anode  02 (g)  CH3OH (I) Anode Reaction:^ Cathode Reaction: CH 3 OH +^CO2 + 6H+ + 6e 1.50 2 + 6H+ + 6e ^3H20 -  -  Fig. 4-1: Exploded view of DMFC with novel 3D anode design. The general research objectives in this thesis are: - Surfactant mediated electrodeposition of nanoscale PtRu(Mo) catalyst on graphite felt substrates Analysis of the deposit morphology, weight, composition and active surface area - Assessment of the catalytic activity towards methanol electro-oxidation in half-cell experiments - Testing of the novel anode material in DMFC under various operating conditions.  The electrodeposition of catalytically active metals, such as Pt, Ru and Mo on the carbon substrate is carried out in an aqueous surfactant based liquid crystalline or micellar electrolyte. The purpose of employing such electrolytes is to provide uniformly dispersed deposits with nanoscale particle diameters. Electrodeposition of fuel cell catalysts in liquid crystalline or micellar electrolyte onto three-dimensional graphite felt substrates has not been previously reported in the literature by other research groups 54  [Bauer et al., 2006, 2007, Lycke, Gyenge, 2007]. Furthermore, there are no reports that describe the utilization of nanoparticulate catalyzed graphite felt DMFC electrodes or the application of such electrodes in electrochemical reactors in general. In this work novel methods towards the dispersion of nanoparticle catalyst onto 3D fibrous electrodes are devised. Such electrodes can potentially be quite versatile, e.g., for electrosynthesis reactors. The utilization of fibrous carbon cloth in conjunction with Pt based catalysts has been described in the literature (e.g., carbon cloth with PtRu particle coated carbon particulates and Teflon to form DMFC anodes [Coutanceau et al., 2004], or a carbon cloth based electrochemical cell to convert SO2 to H2SO4 [Lu, Ammon, 1982]). To promote the disengagement of CO 2 gas bubbles from the anode a different flow design is also investigated in conjunction with the 3D electrode. Bore For Pt Wire Reference Electrode  Fuel Outlet Fuel Inlet  Fig. 4 2: End plate with serpentine flow field (as supplied by Fideris Inc.). -  Instead of feeding methanol through serpentine channels as shown in Figs. 4-1 and 4-2, a flat end plate can be used (Fig. 4-3) where methanol is fed to the bottom part of the anode via a flow distributor (uncatalyzed graphite felt or plastic mesh) and directed to  55  flow parallel to the membrane, thus, promoting cocurrent upward gas-liquid flow. The modification is referred to as a flow-by design, which means the reactant flow is directed perpendicularly to the electric current. On the other hand, operating in flow-through mode means the current and reactant flows are parallel. The approach of operating the anode using a flow-by design is also novel with respect to DMFC technology. Further, the 3D electrode is expected to provide improved mass transfer of reactants and products compared to a conventional gas diffusion electrode or catalyst coated membrane.  Flow Collector Outle  , $ em2 Catalyzed Graphite Felt Inlet Flow Distributor  CH 3 OH  Fig. 4-3: Exploded view of flow-by anode schematic and photograph of flow-by type end  plate.  56  5.0 Experimental Methods  5.1 Electrodeposition The majority of deposition experiments carried out during this project involved the utilization of surfactants (predominantly Triton X-100). The graphite felt substrate was rinsed with methanol and distilled water and then dried in an oven at 353 K for 1 h. To form the electrodeposition baths, the surfactant solutions were mixed at 323 K for 10 min and then left to stand at 298 K for 1 h prior to use when depositing at 298 K. Otherwise the cell containing the well mixed solution was put in a water bath and the desired temperature was adjusted. H2PtC1 6 (Sigma-Aldrich) RuC1 3 (Alfa-Aesar) and MoC1 5 (Sigma-Aldrich) were used. The electrodeposition was carried out at selected deposition currents, surfactant concentrations and durations. The details regarding these variables are outlined in the respective chapters of the results section. After deposition the graphite felt was washed in methanol and deionized water at 333 K, sonicated in methanol for 5 min and rinsed with copious amounts of deionized or distilled water. The washing procedure was repeated three times. Then the sample was soaked in deionized water and cured in a nitrogen gas stream at 573 K for 1 h. (No significant amounts of surfactant residue were present on the substrate after washing as indicated in Appendix E.) Alternatively the samples were treated by applying a reducing potential of -0.8 V vs. MSE (mercury—mercurous sulfate reference electrode (Hg/Hg 2 SO 4 , K2SO4Std)) for 10 minutes at 298 K directly after depositing and washing.  5.1.1 Chemical and electrochemical cleaning of the catalyst surface  Before each experiment the catalyzed three-dimensional anodes were cleaned in a 1:1 v/v solution of H202 30 %,,,, t : H 2 SO 4 98 % wt at 298 K and electrochemically treated three times for 30 s each in 0.5 M H2504 at 298 K to oxidize organic residues from the surfactant at 0.5 V vs. MSE followed by reduction of possible surface oxides at —0.64 V vs. MSE, respectively. Afterwards, repeated cycling in 0.5 M H2SO 4 between —0.68 V and 0.5 V vs. MSE at 298 K was carried out at a scan rate of 100 mV s  57  (see example in Fig. 5-1) until a stable response was obtained. The final surface state of the electrode therefore corresponded to —0.68 V vs. MSE. Alternatively a reducing potential of -0.8 V vs. MSE was applied in 0.5 M H2SO4 for 10 minutes at 298 K. Substantial hydrogen gas evolution, which may also cause chemical reduction of surface oxides, was observed. The bath was stirred occasionally to remove hydrogen bubbles that adhered to or were trapped between the fibers. For fuel cell application the anodes of 5 cm 2 geometric area were pretreated according to the same method.  800 600 400 -  -200 -400 -600 ■  -0.8  -0.6  ,  -0.4^-0.2  0.0  0.2  0.4  0.6  0.8  E [V vs. MSE]  Fig. 5-1: Scan in 0.5 M H2SO 4 at 100 mV s -1 and T = 298 K. PtRu on GF-S3, deposition parameters: 40 % wt Triton X-100, 60 A m -2 , 90 min, 333 K.  5.1.2 Electrode materials Two different fibrous graphite substrates supplied by Test Solutions, Inc. / Electrolytica were used as electrode materials (GF-S6 and GF-S3, respectively). In the early stages of the project GF-S6 (Fig. 5-2) was used, starting with its full thickness as received (6000 tim). As the project progressed, a thickness of 2000 vtm was used for the majority of the experiments that were conducted with GF-S6. GF-S3 (thickness = 350 vim, Fig. 5-3) was used in the more advanced stage of the project. It proved to be more  58  robust, both chemically and mechanically. The porosity of the two materials was determined by measuring the volume of methanol displaced in a graded cylinder after immersion of the felt. The felt properties are listed in Tables 5-1 — 5-3. The specific surface area of GF-S3 can be estimated, since the fiber diameter is —20 um (see Fig. 5-4) and the surface of one single fiber of the same volume as occupied by the total amount of fibers in a specific volume of the felt can be calculated. For GF-S6 the manufacturer specified 420 g m -2 as the weight of the uncompressed felt per geometric area (obtained by BET N2 adsorption). Since the approximate fiber density is known (1500 kg m -3 [Oloman et al., 1991]) one can calculate the specific area. Furthermore, the fiber surface area per geometric area of the felt can be estimated (GF-S6, 2000 um: 21 m 2 mfelt 2 ; GF-S3, 350 um: 3.5 m 2 mfe i t 2 ). These numbers are relevant when assessing the surface roughness of the deposit. Table 5-1: Characteristics of graphite felt types in their uncompressed state.  Type  Thickness  Porosity  GF-S3 GF-S6 GF-S6 GF-S6  [fun] 350 6000 4000 2000  0.95 0.95 0.95 0.95  Specific surface area per volume [x10 3 111 2 111 -3 ] 10.0 10.7 10.7 10.7  Fig. 5-2: Micrograph of uncatalyzed graphite felt (GF-S6).  59  Table 5-2: The effect of compression on GF-S6 graphite felt (uncompressed thickness = 2000 Jim) Thickness Porosity [Pm] 300 500 1000  Electronic conductivity  Specific surface area per volume  [S m -1 ] 440 170 39  [x10 3 m 2 m -3 ] 66.7 40.0 20.0  [ 0.67 0.8 0.9 ]  ",s-:,•^•  •^- ,  „:141e44.  Fig. 5-3: Micrographs of uncatalyzed graphite felt (GF-S3).  60  Fig. 5-4: Micrograph of single graphite fiber (GF-S3). Table 5-3: The effect of compression on GF-S3 graphite felt. Thickness [1-1m]  Porosity [  Electronic conductivity [S m 1 ]  Specific surface area per volume [x103 m 2 m -3 ]  100 200 300  0.83 0.91 0.94  131 29 12  35.0 17.5 11.7  Graphite felt has been extensively used as a three-dimensional electrode in conjunction with gas—liquid flow, e.g., for 0 2 reduction to H202 and SO 2 reduction to S204 2- [Gyenge, Oloman, 2005, Oloman et al., 1990]. The porosity changes with compression according to the correlation below [Oloman et al., 1991]:  =1  AT  (1— 6.  C1 cf  Ax  °)  (5-1)  The electronic conductivity increases when the felt is compressed due to lower contact resistance between the fibers. This can be expressed as [Oloman et al., 1991]: Cr = 10 + 2800(1  S  1 55 `)  co  (5-2)  The contact resistance between the current collector (i.e., fuel cell end plate) and the fibers is also important. It became evident during fuel cell tests that adding a layer of  61  carbon cloth between the felt and the end plate is essential for providing sufficient electrical contact. The specific surface area of the compressed felt can be calculated when the fiber diameter is known [Hodgson, Oloman, 1999]: as=  4(1 —E,f ) df  (5-3)  It should be noted that carbon cloth and Toray carbon paper were also investigated as catalyst substrates. However, the observed methanol oxidation capabilities were poor compared to results obtained with catalyzed felt, mainly because insufficient catalyst loads were obtained with the deposition methods employed in this thesis work.  5.1.3 Electrodeposition cell types 5.1.3.1 Glass beaker (100 ml volume) The working electrode consisted of GF-S6 type graphite felt (4 cm 2 geometric area) glued with carbon particulate filled epoxy resin to a Ti feeder plate. A saturated calomel reference electrode was placed next to the working electrode. Fig. 5-5 shows a schematic of the three electrode assembly in the deposition solution. When one counter electrode is used and the side of the felt facing away from it is blocked by the current feeder the majority of the deposit is located on the side that faces the counter electrode and the lateral penetration of the deposit is very poor. Furthermore, the catalyst particles in the middle of the felt were relatively Pt rich. To mitigate these problems the working electrode was modified as shown in Fig. 5-6 and a second counter electrode was immersed in the electrolyte. The cell is otherwise identical to that shown in  Fig. 5-5. The Ti current feeder plate had a geometric area of 1 cm 2 and was glued to an extended part of the felt that was not exposed to the bath directly, although capillary forces may drag some of the electrolyte up into this section. The benefits of employing two counter electrodes are outlined in Appendix A.  62  Counter Electrode Working Electrode (Anode)^(Cathode) Working Electrode: SCE Reference Electrode Plexiglass Lid  Surfactant Solution  Platinized Titanium Plate  Graphite Felt  Carbon Filled ^Epoxy Glue 15 mm  Fig. 5-5: Schematic of 100 ml glass beaker cell containing a three electrode assembly used for electrodeposition.  Working Electrode  4__Ti Rod  Carbon Filled Epoxy Glue Platinized Ti Plate Deposition Bath Graphite Felt  Fig. 5-6: Working electrode as used in 2 counter electrode setup.  63  5.1.3.2 Sandwich deposition cell (3 ml volume)  In this cell design connections between the power supply and all three electrodes were established by stainless steel alligator clips. No reference electrodes were used. The total deposition bath volume is 3 ml with meshes, counter electrodes and GF-S3 type felt inserted. Fig. 5-7 shows both a schematic and a photograph of this deposition unit.  Inter-Electrode Gap = —150 ptm  22.5 mm  DC Power Supply ^Graphite Felt (Cathode) ^ Plastic Mesh Separator Plexiglass Cell Containing Metal Salt Solution Pla inized Ti Anodes  Fig. 5-7: Schematic and photograph of sandwich type deposition cell. 5.1.4 Surfactants and additives Polyoxyethylene iso-octyl phenyl ether (commercial name: Triton X-100) was selected for deposition in the liquid crystalline or micellar phase. The molecular structure is presented in Fig. 5-8. The nonionic surfactant consists of a hydrophobic iso-octyl phenyl group and a hydrophilic part, i.e., the chain of oxy-ethylene groups. Fig. 5-9 shows the phase diagram for Triton X-100/H 2 0 mixtures as a function of temperature and concentration [Beyer, 1982, Alekseev et al., 1997]. At 298 K the hexagonal phase is formed at surfactant concentrations between 40 %,„- t and 57 %.  64  0  —  CH 2  —  CH 2  OH  n  Fig. 5 8: Molecular structure of Triton X-100 (n –10). -  This mesophase can be modeled as a densely packed set of rodlike aggregates or microtubules forming a hexagonal pattern. Above 303 K solely the isotropic or micellar phase is present. At T < 278 K and a concentration of –70 °/0„, t the so called lamellar (or neat) phase is formed. At concentrations greater than 57 % in both inverted and planar micelles (fragments of the lamellar phase) are present in the isotropic phase.  T [K]  Hexagonal  303  293  i Isotropic  283 Lamellar  1^I^I^I^II^11 20^40^60^80  Fig. 5-9: Phase diagram of Triton X-100/H 2 0 [Beyer, 1982, Alekseev et al., 1997].  65  The presence of metal salts in the bath may influence the phase formation dependence on temperature and surfactant concentration; i.e., the phase diagram shown in Fig. 5-9 may not be accurate when adding e.g., Pt and Ru salts [Attard et al., 1998]. Apart from Triton X-100 other surfactants were tested. Polyethylene glycol octadecyl ether (C 1 8H37(OCH2CH2)nOH, n-10, commercial name: Brij 76 (also a nonionic surfactant)) has been successfully applied to create hexagonally shaped PtRu nanostructures on smooth surfaces [Attard et al., 2001]. In this thesis work a very similar configuration of Brij type surfactant was tested, i.e., Brij 56 (C16H33(OCH2CH2)100H. This agent was employed to fabricate e.g., mesoporous silica structures [Goncalves, Attard, 2003]. Detailed studies that describe the interaction of Triton X-100 with the carbon fiber surface could not be obtained from the literature. Adsorption experiments are typically carried out at concentrations close to the critical micelle concentration (2.3 x 10 -4 M at 293 K [Moraru et al., 1985]) and it was found that at concentrations above the CMC interactions between adsorbed surfactant molecules become significant compared to adsorbate-adsorbent effects [Gonzalez-Garcia et al., 2004]. The polyoxyethylene chain of a Triton X-100 molecule, which is oriented away from the adsorbate surface in a coil configuration, acts as a steric barrier and can therefore limit the accessibility of the substrate surface for other surfactant molecules. The surface area shielded by a single Triton X-100 molecule adsorbed on graphite was reported to be 0.63 nm 2 for surfactant concentrations close to the CMC [Moraru et al., 1985]. The surfactant concentration range used in this thesis was 0.64 M-1.28 M. It is unclear whether e.g., with 0.64 M (or 40 %„t) Triton X-100 a hexagonal micelle assembly is present at the graphite fiber surface, which would result in a highly structured deposit similar to that described in Fig. 3-5. Compared to a well defined smooth gold substrate used for liquid crystal templating ([Gollas et al., 2000]) the graphite surface may have more defects and it possibly contains oxide groups that may influence the interaction with the surfactant molecules and the micellar structures. Certain Brij type nonionic surfactants, that form cylindrical or spherical micelles in the bulk solution, form hemicylinders on graphite surfaces [Patrick et al., 1997]. On the other hand, Triton X-100 was found not to form hemicylinders at the graphite-solution interface, even though a cylindrical geometry can be formed in solution [Patrick, Warr, 2000].  66  5.2 Characterization of the Catalyst Deposits Scanning electron microscope images of the deposits were obtained at the Metallography lab at the Metals and Materials Engineering (MMAT) department at UBC, the UBC Bio-Imaging facilities, the department of physics at Simon Fraser University and at Surface Science Western (Ontario) to estimate the catalyst particle size and assess the degree of catalyst dispersion. Transmission electron microscopy (TEM) was not carried out due to difficulties regarding sample preparation (i.e., microtomy of single catalyzed fibers). Energy dispersive X-ray diffraction (EDX) was conducted at MMAT and at the department of physics at Simon Fraser University to indicate the atomic bulk composition of the catalyst. For each sample 5-20 measurements were typically carried out in point scan mode to obtain an average value. A representative spectrum is shown in Appendix F. The catalyst load and Pt:Ru(:Mo) bulk atomic ratio were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) of a solution obtained by digesting catalyzed felt in aqua regia. The analysis was carried out at Cantest (formerly Vizon Scitech) (Vancouver). XRD analysis is often carried out to investigate the crystal structure, particle diameter distribution and the extent of alloy formation. When XRD was carried out, no distinct peaks for Pt or Ru were observed for bulk sample pieces of catalyzed graphite felt. The available equipment could not be adjusted to study deposits on single fibers. Auger spectroscopy can theoretically be used to determine the surface composition of PtRu nanoparticles. However, there is an overlap of the C and Ru peak in the spectrum, which makes the quantitative determination of the Ru content impossible (see Appendix F).  67  5.3 Half-Cell Electrochemical Experiments The activity of the catalyst was assessed by standard electrochemical techniques, such as cyclic voltammetry, chronoamperometry and chronopotentiometry. Fig. 5-10 shows all the components of the testing cell, which holds a classical three electrode arrangement including two counter electrodes. The working electrode consisted of a Ti rod that was soldered (using Pb/Sn) to a 1 cm 2 Pt/Ti current feeder onto which the felt was attached by carbon powder filled epoxy glue. The current feeder was not immersed in the electrolyte, but there may have been some contact with the electrolyte due to capillary forces. Voltammograms of CH3OH oxidation on the fully immersed Pt/Ti feeder were recorded to determine the contribution of the feeder to the overall generated current. The peak -2  current density for CH3OH oxidation was 10 A m at 0.2 V vs. MSE with a scan rate of -1^ 5 mV s , which isi about forty times lower than the current densities recorded at the same parameters with a catalyzed felt attached. (A PtRu catalyst supported on GF-S3 was prepared in the sandwich deposition cell at 298 K with 40 c/0„ 4 Triton X-100 applying a superficial current density of 60 A m -2 for 90 minutes.) Taking also into account that in the actual setup the epoxy glue covered the feeder surface, the contribution of the Pt/Ti feeder to the observed electrochemical response can be considered negligible. Alternatively the felt was immersed by holding it with a stainless steel alligator clip that was soldered to the Ti rod current feeder. Catalyzed felt was tested by immersing 1 cm  2  of its geometric area without compression. A Parstat 2263 potentiostat operated by PowerSuite TM software was employed to control the current or voltage, respectively. Two graphite rod counter-electrodes (total area = 20 cm 2 ) on opposite sides of the working felt electrode were used in conjunction with a mercury—mercurous sulfate reference electrode (Hg/Hg 2 SO 4 ,  K2SO4std)•  All  electrode potentials in the following sections are reported with respect to the mercury— mercurous sulfate reference electrode (MSE) unless stated otherwise. To facilitate the comparison with other published data in the literature the potential of the MSE was measured separately against the standard calomel electrode (Hg/Hg 2 C1 2 , KClstd, SCE): 0.42 V at 298 K.  68  Graphite Rod Counter Electrode':  M SE  Referent lectrod  LUggin  Cr,pillary  Catat3,rze41 Graphite Felt  Fig. 5-10: Glass cell used for electrochemical testing. The electrolytes were typically 0.1 M H 2 SO 4 (Fisher Scientific) for blank experiments, and 0.5 M CH 3 OH (Fisher Scientific) in 0.1 M H 2 SO 4 for the majority of the methanol oxidation studies. The low acid concentration was selected to minimize the specific adsorption of sulfate ions on the catalyst surface, which can compete with methanol adsorption. The IR correction of the measured data was carried out based on conductivity measurements of 0.1 M H2SO4-0.5 M CH3OH at 298 K (5 S m -1 ) and 343 K (10 S m -1 ). The gap between the Luggin capillary (containing 1 M H 2 SO4) and the working electrode was —2 mm. Fig. 5-11 shows the superficial current density without IR correction at 0 V vs. MSE as a function of methanol and sulfuric acid concentration recorded for PtRu deposited onto GF-S3 at 298 K felt using 40 %„ t Triton X-100, 60 A  M -2  and a deposition time of  90 minutes. The voltammograms were recorded at 298 K and 5 mV s -I . The highest currents were observed when applying 0.5 M or 1 M H2SO4. At lower acid concentrations the Ohmic drop must be accounted for, whilst higher H 2 SO 4 concentrations (e.g., 2 M) are not suitable due to enhanced SO 4 2- adsorption on the electrode surface.  69  Since most fundamental half-cell CH 3 OH electro-oxidation studies in the literature were performed in either 0.1 M H2SO 4 or 0.1 M HC1O4, in the present work 0.1 M H2504 was selected and IR correction was applied.  350  300 (.1)  250  > 200 O 150  E < 100 50  2.0 0.5 1.0  [H2so4J^0  . 5^ 0.0 0.  0  Fig. 5-11: Superficial current density at 0 V vs. MSE as a function of CH3OH and H2SO4  concentration. Scan rate = 5 mV s  1  ,  T = 298 K.  5.3.1 Active surface area assessment by copper under potential deposition and stripping  The catalytically active surface area on the felt had to be determined. Oxidative stripping of adsorbed hydrogen or CO monolayers utilizing cyclic voltammetry are established methods for calculating the active surface area of Pt [Vielstich et al., 2003]. However, the presence of Ru causes difficulties and therefore these methods are not suitable for the binary catalyst. Bock and McDougall proposed the utilization of oxalic acid to assess the surface area and distinguish between Pt and Ru surface fractions quantitatively [Bock, McDougall, 2003]. This procedure did not work properly when testing catalyzed 3D graphite felt electrodes. Green and Kucernak presented a method based on the under potential deposition and subsequent stripping of a copper monolayer (Fig. 5-12) [Green, Kucernak, 2002]. The  70  authors used sweep rates of 10 or 100 mV s -1 in their work. However, using these scan rates in conjunction with the 3D electrodes did not yield interpretable responses, i.e., no peaks or shoulders were observed. As a first step a full cyclic sweep was performed at 0.5 mV s -I between -0.68 V and 0.27 V vs. MSE in 0.1 M H2SO4 at 298 K to provide a blank background scan. By setting the potential to -0.48 V for 300 s in the same electrolyte the surface oxides were reduced. A monolayer of copper was then deposited by immersing the pretreated felt in 2 mM CuSO 4 -0.1 M H2SO4 and depositing at -0.38 V vs. MSE for 300 s. Then a linear oxidative sweep from -0.38 V to 0.27 V vs. MSE (0.5 mV s -1 ) was applied to strip the Cu off the surface. All these steps were carried out at 298 K. It is assumed that one Cu atom deposits per Ru or Pt surface atom. A charge of 4.2 C m -2 is associated with Cu stripping from the surface of either of these metals. Therefore the active catalyst area can be estimated. Copper does not deposit on carbon when using this underpotential deposition method [Cheng, Gyenge, 2006].  )^I 0.2 0.4 0.8 0.8 E [V vs. SHE] Fig. 5-12: Blank scan and Cu stripping peak for a PtRu surface. Electrolytes: 0.1 M H2SO4 and 2 mM CuSO4-0.1 M H2SO 4 . Scan rate = 10 mV s -I , T = 298 K [Green, Kucernak, 2002].  In most cases one stripping peak was observed in the Pt region (— -0.2 — +0.17 V vs. MSE) for the graphite felt based electrodes discussed in this thesis (Fig. 5-13). Only one of the investigated samples showed a response in the region where  71  the Cu desorption from the Ru surface takes place (Fig. 5-14). The atomic Pt:Ru bulk ratio of this sample was 1:1 (see also Section 6.1.5). Based on the voltammogram shown in Fig. 5-14 the Pt surface area is — 55 % of the total PtRu area. For all other samples the Pt:Ru surface area ratio is unknown.  ,■, blank •••• Cu stripping 40  PtRu 0.5 my s -1  E [V vs. MSE]  Fig. 5-13: Blank and stripping scans conducted at 298 K in 0.1 M H 2 SO 4 and  2 mM CuSO 4 -0.1 M H2SO4, respectively. PtRu on GF-S3 (prepared with 40 %,, t Triton X-100, i = 60 A m -2 , t = 90 min, T = 333 K).  E [V vs. MSE]  Fig. 5-14: Blank and stripping scans conducted at 298 K in 0.1 M H2SO4 and  2 mM CuSO 4 -0.1 M H2SO 4 , respectively. PtRu on GF-S3 (prepared with 40 %, 4 Triton X-100, i = 60 A tn -2 , t = 90 min, T = 298 K).  72  5.4 Fuel Cell Experiments The novel anode material was tested in a full cell assembly. Connecting the electrode to a Nafion membrane by hot pressing was not viable, since the felt was crushed when applying conventional hot-pressing conditions (T = 393 K, 100 kg cm 2 ). The anode was installed in the single fuel cell testing system in conjunction with the respective end plates that are shown in Figs. 4-2 and 4-3. The torque applied to each of the bolts used for the full cell assembly was 10 Nm. Each bolt had a diameter of 6 mm. Nafion 117 membranes with conventional Pt black (40 g m -2 ) cathode catalysts were utilized (Lynntech Inc.). The performance of the cell, depending on temperature, anolyte flow rate and methanol concentration was tested by changing only one variable at a time, and subsequently by a factorial experimental design approach. The degree of compression of the anode material is also a crucial factor due to its implications regarding resistance and porosity and is discussed in Section 6.4.3 [Oloman et al., 1991]. Oxidant Exhaust.^ + CO 2  H2O ^ t Fuel Cell Test Station Load Bank  Alf 02  Supply  Effluent Analysis / Disposal  t PC Operating FC Power Software  F Thermocouple  Cathodé Oxidant Return + H 2 O  -.  Fuel Pre-Heater  Anode  ^ Fuel Reservoir  Current Voltage Peristaltic Pump  Hot Plates  Fig. 5-15: Fuel cell test setup and flow diagram.  73  In Fig. 5-15 the fuel cell testing system is described. Methanol is fed in either circulating or single pass mode. The bulk methanol container is preheated using a hot plate and the flow is preheated again downstream of the pump by a water bath. One heating rod is inserted in each end plate through a centrally located bore. The temperature is monitored by a thermocouple, that is attached to the anode endplate by a screw connection. The fuel cell performance of the catalyzed felt anodes was evaluated employing a 5 cm 2 geometric area (2.24 cm x 2.24 cm) experimental DMFC with gold plated stainless steel end plates having serpentine type flow channels or a flat design for the flow-by mode (see Fig. 4-3). Carbon cloths were used as backing layers on the anode and cathode side, respectively to ensure electrical contact of the entire electrode area. Temperature, oxidant flow rate, cathode pressure and the current density were set by a fuel cell test station (Fideris Inc.) that was controlled by FC Power TM software. Dry 02 was supplied at a flow rate 500 ml min 1 STP. The anolyte, consisting of 1 M CH 3 OH in 0.5 M H2SO 4 in most cases, was supplied at near ambient pressure by a peristaltic pump.  Gasket  FIAT^Carbon Cloth  Catalyzed Graphite Felt  :Nation I I 'Membrane -  Fig. 5-16: MEA components, backing layers and gasket used for fuel cell testing.  74  Typical MEA components and backing layers as well as a Teflon gasket are shown in Fig. 5-16. The uncatalyzed ELAT microporous layer (-400 id111 thick when not  compressed) was used as a diffusion layer on the cathode with the main purpose of providing sufficient electrical contact between the cathode catalyst layer and the endplate. The material contains Teflon to mitigate flooding. Similarly, on the anode side a carbon cloth (Textron, —250 pm uncompressed thickness) was applied as a backing layer between the catalyzed graphite felt and the endplate, since otherwise the contact resistance would have been too high. For testing of the catalyzed GF-S3 electrodes typically one or two Teflon gaskets were used on the anode and one gasket was used on the cathode. The thickness of one gasket was 127 pm in its uncompressed state.  75  6.0 Results and Discussion Outline In the early stages of this thesis the main focus was on achieving high catalyst particle dispersion both on fiber surfaces and through the electrode thickness while also trying to lower the noble metal load. Exploratory deposition experiments revealed the effects of acidity and surfactant presence in the electrolyte on the deposit morphology and load. Factorial deposition experiments were carried out at 298 K to study the impact of altering the surfactant concentration, deposition current density and time on the particle diameter, catalyst load, PtRu bulk atomic ratio and catalytic activity. Reducing the felt thickness and the deposition bath volume led to a more homogenous catalyst distribution and enhanced catalytic activity. In general better results were obtained when utilizing GF-S3 type graphite felt compared to GF-S6. The thickness of GF-S3 (350 pm) is also more suitable for fuel cell applications compared to GF-S6 (2000-6000 lArn). Increasing the deposition temperature to 333 K resulted in a higher superficial methanol oxidation current density. Carrying out two subsequent deposition steps further increased the methanol electro-oxidation performance, albeit at the expense of having substantially increased catalyst loads. PtRuMo compositions having higher catalytic activity than PtRu were identified. The superficial current density (i [A m -2 ]) was reported with respect to the geometric area of the catalyzed felt electrode. Mass activity data was based on the Pt load per geometric electrode area. Pt catalysts were prepared under the same conditions as the best performing PtRu samples and tested for comparison. The active area of the pure Pt catalysts was characterized by hydrogen adsorption and stripping. The catalyzed graphite felt anode was compared with a conventional catalyst coated membrane in a 5 cm 2 experimental DMFC, showing a significant improvement of the power output. Most fuel cell experiments were carried out with an anode feed that contained 0.5 M sulfuric acid, since a solid proton conducting material that would allow a more practical aqueous methanol feed could not be identified. By applying the sulfuric acid supporting electrolyte the entire catalyst surface was exposed and therefore accessible for methanol adsorption and oxidative dehydrogenation. When Nafion is used  76  a fraction of the catalyst surface is blocked. The incorporation of Nafion in Pt-based nanoparticle-inks was reported to cause a 40 % decrease in oxidation current density and 13 % decrease in active surface area [Jayashree et al., 2005, McGovern et al., 2003]. The effect of felt compression on fuel cell performance was assessed by varying the anode gasket thickness. Furthermore, factorial experiments were carried out to gauge the effects of altering the methanol concentration, flow rate and operating temperature. Applying a conventional end plate with a serpentine flow field was compared with using a flat end plate in flow-by mode. The degradation of the fuel cell performance over time was studied as well.  6.1 Electrode Preparation 6.1.1 Acid and surfactant concentration in the deposition bath  All electrochemical experiments discussed in Section 6.1.1 were carried out at 298 K and 100 kPa(abs). The glass beaker cell design as shown in Fig. 5-5 (1 counter electrode) was employed. The electrodeposition bath for the experiments listed in Table 6-1 was obtained from a stock solution containing 4.1 M HC1 and 17 mM Pt and Ru salts (for a 1:1 atomic ratio in solution, 65 mM Pt salt was used for a 4:1 ratio). Therefore the acid concentration and metal salt concentrations were decreased (e.g., to 2.5 M and 10 mM, respectively, when the stock solution was mixed with the surfactant, e.g., 40 Vo wt Triton X-100). The GF-S6 felt having an uncompressed thickness of 6000 pm and a geometric area of 4 cm2 was used. The catalyst load was determined by weighing the dry felt before and after deposition. The surfactant concentration was selected based on work by Gollas et al. [Gollas et al., 2000]. Using both 1:1 and 4:1 Pt:Ru ratios in HC1 resulted in a 4-5 times reduced catalyst load when Triton X-100 was added. Both the lower HCl concentration and partial blocking of the graphite surface and / or mass transport limitation caused by the surfactant may be responsible for this effect.  77  Table 6-1: Preliminary deposition tests utilizing the100 ml glass beaker cell and GF-S6. Atomic Pt:Ru ratio in solution 1:1 1:1 4:1 4:1 1:1 1:1 1:1 1:1 1:1 1:1 1:1  [Triton X-100] rowd 0 40 0 40 0 40 40 40 40 40 40  i [A m -2 ] 200 200 200 200 200 200 200 400 400 100 100  t [min] 30 30 30 30 30 30 60 30 60 60 120  [HC1] [M] 4.1 2.5 4.1 2.5 -  PtRu load [g m -2 ] 190 40 420 95 45 16 24 28 39 89 112  The current efficiency was decreased from —21 % to —5 % when the surfactant was used in conjunction with HC1 and a 1:1 Pt:Ru ratio in solution. The catalyst load was further decreased when the deposition was carried out without using HC1. Since hydrogen evolution competes with the electrodeposition of the metals, especially at higher current densities, deposition current densities of 100 A m -2 or less were applied in subsequent experiments.  6.1.1.1 a) Deposit morphology obtained without surfactant Employing a deposition electrolyte composed of Pt and Ru salts (1:1 atomic ratio) in 4.1 M HC1 and a superficial electrodeposition current density of 200 A minutes yielded a thick deposit of 190 g  M2  M -2  applied for 30  (geometric surface area) with a rough  surface (Fig. 6-1). The coating contained agglomerates of —100-500 nm diameter.  Fig. 6-2 displays the deposit formed (45 g m 2 ) under the same conditions except that HC1 was not present in the deposition solution. A thick and fairly smooth coating was obtained, which is unlikely to be suitable for catalytic reactions. Furthermore, poor catalyst penetration into the center region of the graphite felt was observed. Clearly, in the absence of surfactant the acidic solution was the more beneficial electrolyte for the same deposition parameters, since a higher surface roughness and better penetration of the deposit into the felt were obtained.  78  Air  Fig. 6-1: Fiber surface (a) and enhanced view (b) of Pt Ru deposit obtained from  4.1 M HC1 electrolyte (1:1 PtRu atomic ratio), deposition at a superficial current density of 200 A 111 -2 applied for 30 minutes.  0.479pm  Fig. 6-2: PtRu deposit obtained in aqueous electrolyte (1:1 Pt:Ru atomic ratio) without  HC1 present using a superficial current density of 200 A m 2 for 30 minutes.  6.1.1.1 b) Deposit morphology obtained with surfactant First, a 2.5 M HC1 solution containing Pt and Ru (4:1 ratio) and 40 %,„- t Triton X-100 was used. The ruthenium content, deposition time and current density were the same as  79  described in Section 6.1.1.1 a). The smallest agglomerates were estimated to be —250-500 nm in diameter (Fig. 6-3). A continuous liquid crystalline phase was not obtained in this case. Instead, a macroscopic phase separation was observed. This could be due to insufficient mixing. In the acid electrolyte, the presence of the surfactant had no beneficial effect on the structure of the deposit, as it looks fairly similar to the coating shown in Fig. 6-1. This might be due to the observed phase separation, which could mean that there was no surfactant present near the graphite fibers. Next, deionized water containing Pt and Ru salts (1:1 Pt:Ru atomic ratio) was mixed with Triton X-100 (40 % wt ). The deposition was repeated as described before. The obtained catalyst load was 16 g m -2 . The deposit morphology is shown in Fig. 6-4. Agglomerates of —500 nm diameter were obtained consisting of particles having a —20100 nm diameter.  Fig. 6-3: Coating obtained by electrodeposition in 2.5 M HC1 with Triton X-100 (40 % wt) electrolyte containing Pt and Ru salt (4:1 Pt:Ru atomic ratio) at a superficial current density of 200 A m -2 applied for 30 minutes.  The enlarged image shows a distinct surface roughness, which is due to the presence of surfactant during deposition (Fig. 6-4 (b)). In the absence of acid, the surfactant addition had a more pronounced effect on the deposit surface structure (compare Fig. 6-4 and Fig. 6-2).  80  The nonuniformity of the surface deposit coverage can be explained by insufficient mixing of the two phases (i.e., surfactant and water) at 298 K in this case. Therefore a continuous liquid crystalline phase was not obtained. To improve mixing in subsequent experiments the solution was heated to 323 K, mixed vigorously at that temperature, and then cooled to 298 K. Both acidity and surfactant presence contributed to an increase in the surface roughness. Presence of acid (i.e., enhanced electrolyte conductivity) facilitated the deposition throughout the felt thickness. The qualitatively high surface roughness obtained with the Triton X-100 containing electrolyte was considered promising.  Fig. 6-4: Fiber surface (a) and enhanced view (b) of irregular coating obtained from aqueous solution without HC1 (1:1 atomic Pt:Ru ratio) containing 40 %,t. Triton X-100 at  a superficial current density of 200 A in -2 applied for 30 minutes.  81  It is important to note that electrodeposition on GF-S3 type felt using 4.1 M HCl without surfactant yielded very poor catalytic activity compared to electrodeposition with 40 Vo wt Triton X-100 (without HC1) when a deposition charge density of 32.4 C cm -2 (geometric area) was applied in both cases. The superficial current density for methanol electro-oxidation was —100 times lower at -0.2 V vs. MSE and 343 K when the electrodeposition was carried out in 4.1 M HC1. The electrodeposition in concentrated HC1 baths was not investigated further in conjunction with GF-S3. Further discussion related to electrodeposition on GF-S3 focused on the application of Triton X-100.  6.1.1.2 Comparison of alkaline and acidic electrodeposition of PtRu  For this comparison the sandwich cell (Fig. 5-7), GF-S3 and Pt and Ru salt concentrations of 65 mM each were used. Electrodeposition of Pt and PtRu in alkaline media was reported in the literature (e.g., at 338 K yielding Pt crystallites of 6 to 15 nm diameter) [Djokic, 2000, Kim et al., 2003]. Table 6-2 shows the results of electrodepositing in acidic and alkaline (1.8 M (NH4)2HPO4 with NH4OH added to adjust the pH) media in the presence of Triton X-100 at 298 K.  Table 6-2: Comparison between acidic and alkaline electrodeposition of PtRu on GF-S3 at 298 K. Surfactant concentration [%vvt] pH = 1 50 pH = 9  50  Superficial current density [A m 2 ] 40  Deposition time [min] 80  PtRu load [g m 2 4.0  1.3:1  40  120  14.6  114:1  Atomic Pt:Ru ratio of deposit ]  The alkaline solution inhibited the Ru deposition producing, therefore, a Pt rich catalyst, with a Pt:Ru ratio of 114:1 compared to 1.3:1 produced under acid conditions. The pH value of 1 is due to the proton content provided by the H 2 PtC1 6 salt. Figs. 6-5 and 6-6 contain corresponding voltammetric and chronopotentiometric data. The onset of methanol oxidation at 298 K occurred at a lower potential (i.e., —0.35 V vs. MSE) in the case of the catalyst prepared in the acidic bath, since its ruthenium content was significantly higher (Fig. 6-5). The catalyst prepared in alkaline media containing a very  82  low fraction of Ru induced methanol oxidation only at potentials greater than —0.2 V vs. MSE. Chronopotentiometry at 298 K and 333 K using a 10 A m 2current step showed that the anode potentials were between 30 and 100 mV more positive for the electrode prepared under alkaline conditions (Fig. 6-6). 250 - 500 - 400  200  300  E  200  150 -  - 100 0  100 -0.6^-0.4^-0.2^0.0^0.2^0.4  0,  50 -  E [V vs. MSE]  E 0  .............................................. -  •■ acidic plating bath •••• alkaline plating bath -50 -0.8^-0.6^-0.4^-0.2^0.0^0.2  ^  0.4  ^  06  E [V vs. MSE]  Fig. 6-5: Electro-oxidation of methanol on PtRu electrodeposited in the presence of 50 %„,t Triton X-100 on GF-S3: Comparison between acid (pH 1) and alkaline (pH 9) 1 electrodeposition conditions. 0.5 M CH3OH-0.1 M H2SO4, 5 mV s , T = 298 K.  These results support the documented co-catalytic effect of Ru on methanol oxidation [Liu, Norskov, 2001, Oldfield et al., 2002, Gasteiger et al., 1994, Jusys et al., 2002] and clearly show the superiority of the anodes produced in acidic (pH —1) conditions. Hence, anode catalyst preparation was carried out in acidic media for subsequent methanol electro-oxidation and fuel cell experiments.  83  •  -0.05  -0.10 -  -0.15 -  E co  4 20  • ■••■••■•••■•• •••■•• ■■••■••••••  -  '  -0.30 pH = 1 T = 298 K ^ pH = 9 T = 298 K ^ pH = 1 T = 333 K 01. 1. ••••••• ^••^pH=9 T =333 K  -0.35 -  -0.40 -10  0  10  20  30  40  50  60  70  t CS]  Fig. 6-6: Chronopotentiometry of methanol electro-oxidation on PtRu electrodeposited in  the presence of 50 Vo wt Triton X-100 on GF-S3: Comparison between acid (pH 1) and alkaline (pH 9) electrodeposition conditions. 0.5 M CH 3 OH-0.1 M H 2 SO 4 , 10 A m 2 . 6.1.2 Fundamental electrochemical studies: Correlation between peak current density and scan rate  The peak current density and electrode potential dependence on the can rate was studied for a sample prepared at 298 K (pH = —1, see Table 6-3) using a 2000 um thick GF-S6 type felt and the 100 ml deposition cell with 2 counter electrodes. The respective voltammograms are shown in Figs. 6-7 and 6-8. The peaks observed during the reverse scans at —0.0-0.1 V vs. MSE in Fig. 6-7 and at —0.15-0.3 V vs. MSE in Fig. 6-8 indicate the oxidation of adsorbed intermediates formed during the anodic sweep. Table 6-3: Deposition conditions and catalyst properties.  [Triton X-100] %,,„t 50  i [A m -2 ] 40  t [min] 30  PtRu load [g m -2 ] 5  Atomic Pt:Ru ratio of deposit 1:2  dp [nm] —10-50  A linear relationship between the square root of the scan rate and the peak current density is observed (referring to anodic sweep) (Fig. 6-9), thus implying that the reaction  84  ^  is controlled by diffusion of methanol, in agreement with the Randles-Sevcik equation (2-24). For the three high scan rates there is also a linear dependence of the peak potential on the logarithm of the scan rate at 333 K (Fig. 6-10). The following equation applies [Bard, Faulkner, 2001]:  D  E p =E: 333K + 2 [1.04 — log(— D – 2^log k + log u]^(6-1) —  The effective diffusion coefficient at 333 K was 3.2 x 10 -9 m 2 s -I using Do equal to 3.5 x 10 -9 m 2^[Lide et al., 2006] and co of 0.95:  ^D  =D  0  E  0  (6-2)  "^  The estimated (apparent) Tafel slope value b at 333 K was 365 mV dec -I based on the plot shown in Fig. 6-10.  60 -  5 mV s  50 mV s  40 MOO  OMM100  200 mV s  OOOOOOOOOOOOO  ••••• •: 41.•••  .  16 ° ••■•11 r / e.•.• 'S^ * . • % •1• 81.4 fo:^.•  -1^ -1  40 e .•  .  • IOW • • ow • • NIS • •  -20 -  I^\  ^..:*r4.• •  ^•• 20 E  0%  -1^  100 mV s MM4PO  011. •^•  -1^  .  • -0.6^40.4 40  4,.  • •°' •  -0.2^0.0  20  .•  E  -40 -  < 0  -60 -  -20  -0.6  ^  0.2^0.4  E [V vs. MSE]  ...••• •  ...... ••  O.: •  ..... .........  -0.4^-0.2^0.0^0.2  ^  0.4  ^  06  E [V vs. MSE]  Fig. 6-7: Cyclic voltammograms obtained at various scan rates in 0.1 M H2SO4 (inset) and 0.1 M H2SO4-0.5 M CH 3 OH, T = 298 K.  85  5 mV s' 50 mV s-1 -1 100 mV s-1 -1 200 mV s-1 -1  -0.6^-0.4^-0.2^0.0  ^  0.2  ^  0.4  E [V vs MSE]  Fig. 6-8: Cyclic voltammograms obtained at various scan rates 0.1 M H 2 SO 4 -0.5 M CH 3 OH, T = 333 K. Since the methanol concentration is known, the Randles-Sevcik equation (2-24) can be applied to estimate the number of electrons transferred per molecule of methanol during the anodic sweep. Instead of the theoretical maximum of 6 electrons only 1 electron was transferred at the peak potential, indicating the formation of intermediates, which might undergo further oxidation at higher potentials. Moreover, based on equation (6-1) and Fig. 6-10 the heterogeneous reaction rate constant (k) was 2.4 x 10 -7 m s -1 . The anode open circuit potential at 333 K was 0.027 V vs. MSE.  86  ^ ^  250 ^  200  N  -  150  C  100 -  50 -  0^ 00  0.1  0.2  0.3  ll"  0.4  05  [(V s -1 )°1  Fig. 6-9: Plot of peak current density vs. square root of the scan rate. 0.42 0.40 0.38 0.36 ui  slope = b/2  0.34 0.32 0.30 0.28 -1.4  ^  -1.3  ^  -1.2  ^  -1.0^-0.9  ^-0.8^  -0.7  ^  -0.6  log (I) [V s - 1 ])  Fig. 6-10: Plot of peak potential vs. logarithm of the scan rate at 333 K.  6.1.3 Factorial experimental design (I) (100 ml glass beaker cell)  6.1.3.1 Surfactant mediated galvanostatic PtRu co-deposition The effect of surfactant concentration, deposition current and deposition time on particle size and dispersion was studied using a factorial design approach. The substrate was GF-S6 type felt having a geometric area of 4 cm 2 and an uncompressed thickness of 6000 µm. A single counter electrode was used. The selected variables and their levels are 87  outlined in Table 6-4. The atomic ratio of Pt:Ru in the deposition electrolyte was 1:1 in all cases and the total metal salt concentration was 40 mM. The weight fraction of Triton X-100 was varied between 40 %„ t and 80 Vo wt . The total noble metal load could be estimated by weighing the felt / current feeder assembly before and after the deposition.  Table 6-4: Variables and levels of factorial experiment, T = 298 K. c [%„t] i [Am -2 ] t [min]  Low Middle High 40 60 80 50 100 75 30 60 120  Table 6-5: Selected experimental data from the factorial experiment investigating galvanostatic PtRu co-deposition on GF-S6. Deposition mode  PtRu load  40 Vo wt Triton X-100, 50 A m -2 , 30 min  [gm -2 ] 3  40 Vo wt Triton X-100, 100 A m 2 , 30 min  5  Atomic Pt:Ru ratio of deposit  dp [nm]  5:1  50  5:1  120  Fig. 6-11 shows the deposit morphology when all the variables were set at their 'low' levels. This sample showed the best results as far as particle size and dispersion are concerned (see also Table 6-5). Fig. 6-12 shows another sample prepared under the same conditions except that the current density was 100 A m 2 .  88  Fig. 6-11: Particle distribution on the fiber surface (a) and enhanced view of deposit (b) prepared with 40 %,,„t Triton X-100, applying 50 A  M-2  for 30 minutes.  For the rest of the factorial experiments, in general, deposits with less uniform catalyst dispersion on the fiber surfaces were obtained. Current densities and deposition times at their high levels yielded larger particles and increased formation of agglomerates.  89  Fig. 6-12: Particle distribution on the fiber surface for the deposit prepared using 40 °/0„, t Triton X-100, applying 100 A  M-2  for 30 minutes.  6.1.3.2 Surfactant mediated galvanostatic sequential deposition of PtRu Platinum was deposited first followed by ruthenium deposition at the same surfactant concentration, current density and deposition time. All variables and levels are the same as used for the co-deposition described in Section 6.1.3.1. The deposition times listed in Table 6-4 represent the deposition interval for each metal (Pt and Ru). A uniform catalyst dispersion as shown in Fig. 6-13 was obtained at a surfactant concentration of 80 % and a constant current density of 50 A m -2 applied for lh for each metal deposition (see also Table 6-6). Irregular particle growth was observed with 80 %,, t surfactant (Fig. 6-14).  90  Fig. 6-13: Particle distribution on the fiber surface for the deposit prepared with 80 %,„- t  Triton X-100, applying 50 A 111 -2 for 60 minutes (for Pt and Ru, respectively). Table 6-6: Selected experimental data from the factorial. Investigation of galvanostatic  sequential deposition of Pt and Ru on GF-S6. Deposition mode (t for each deposition step)  PtRu load [g  Atomic Pt:Ru ratio of deposit  dp [nm]  80 %,,,t Triton X-100, 50 A m -2 , 60 min  7.5  4:1  60  80 % ^Triton X-100, 100 A 111 -2 , 15 min  6  5:1  90  m21  Increasing the deposition current density to 100 A m -2 yielded a higher degree of agglomeration in spite of the shorter deposition time (Fig. 6-15). This shows the impact of the current density on particle diameter and agglomeration. Employing all variables at their high levels yielded a thick coating at the top layers of the graphite felt and a very non-uniform deposit with a significant fraction of uncovered areas on the inner fibers.  91  Fig. 6-14: SEM image of deposit prepared with 80 % wt Triton X-100, applying 50 A M-2 for 60 minutes (for Pt and Ru, respectively) (a), and enlarged view of crystallites (b).  92  Fig. 6-15: Particle distribution on the fiber surface for the deposit prepared with 80 %,, t  Triton X-100, applying 100 A m -2 for 15 minutes (for Pt and Ru, respectively).  6.1.3.3 Summary In the case of co-deposition the lowest selected surfactant concentration (i.e., 40 °A m) combined with the lowest current density and deposition time provided the smallest particle diameters and (qualitatively) the highest degree of particle dispersion at 298 K. The sequential deposition yielded the most favorable particle diameters and dispersion when using a surfactant concentration of 80 % and each metal was deposited for 60 minutes at the lowest current density (i.e., 50 A m -2 ). The presence of surfactant promoted deposition of catalyst particles across the entire thickness of the felt. The samples with the most evenly dispersed catalyst deposits from the co— and sequential deposition factorial experiments were analyzed using cyclic voltammetry. Linear sweeps at 5 mV s -I were conducted at 298 K. The electrolyte contained 0.5 M CH 3 OH in 0.1 M H2SO4. The activity was expressed through the measured current normalized with respect to the platinum load (i.e., mass activity (Fig. 6-16)). The inset describes the same measurements expressed as the superficial current density. The sample obtained by co-deposition showed higher activity.  93  18 ^  40  16 30  14 20  12 10  10 ^0  Codeposition  8 -  6-  40 %„, 50 A m -2 ^ 0.3^ 0.2^ 0.1^0.0^0 1^ 30 min • •^Sequential E [V vs. MSE]^ Deposition -  -  42-  *****  ******  80 %„, 60 A m -2 60 miA • •• ********  ********** ***************  0 -0.4  ^  -0.3^-0.2^-0.1  ^  0.0  ^  01  E [V vs. MSE]  Fig. 6-16: Methanol electro-oxidation: Comparison between catalysts prepared by codeposition (40 %,,t Triton X-100, 50 A m -2 , 30 min) and a sample prepared by sequential deposition (80 % wt Triton X-100, 50 A m -2 , 60 min for Pt and Ru, respectively). Linear sweep at 5 mV s -1 , 298 K  For sequential deposition, Table 6-6 shows that increasing the current density by 100 %, while reducing the deposition time to 25 %, yielded a 50 % growth of the average particle diameter at a surfactant concentration of 80 % wt . Thus, the particle growth is a stronger function of current density than of deposition time. Using high surfactant concentrations in the isotropic phase yielded formation of dendrites. Furthermore, the relatively low conductivity of the deposition solution in this regime is another minor disadvantage. It is unlikely that alloying occurs to any extent during sequential deposition, which may also in part explain the higher activity observed for the codeposited catalyst. The obtained EDX data shows the atomic ratio for each co-deposited sample (Table 6-7).  94  Table 6-7: Effect of the deposition parameters on the atomic composition of the PtRu catalyst obtained by galvanostatic co-deposition. [Triton X-100]  i  t  Atomic Pt:Ru  Nal 40 40 40 40 80 80 80 80 60  [A m -2 ]  [min]  bulk ratio  100 50 100 50 100 50 100 50 75  120 120 30 30 120 120 30 30 60  2:1 3:1 5:1 5:1 6:1 5:1 5:1 4:1 3:1  The following graph (Fig. 6-17) shows general trends indicated by the factorial data with co-deposited PtRu. When the surfactant concentration was increased the atomic Pt:Ru ratio was increased in most cases.  5  Zrt  •  4  •  •  •  6  •  •  •  •  •  •  •  •  0.  3  •  2  •  •  •  •  •  • •  •  40^50^60^70^80  Triton X-100]  50^60^70^80^90 100^40^60^80^100^120  i [A m -2  ]^  t [min]  Fig. 6-17: Effect of surfactant concentration, deposition current density and time on the atomic Pt:Ru bulk ratio as determined by EDX for co-deposited PtRu on GF-S6.  95  6.1.4 Pulsed current sequential deposition  In this section a method is described where platinum is deposited at 298 K for a very short time at a relatively high current density (nucleation pulse) to produce a large number of nuclei. Then a much smaller current density is applied to grow the Pt particulates. This approach is similar to that described by Liu et al. [Liu et al., 2001]. Afterwards Ru is deposited at the same low constant current density and 298 K. Sequential deposition experiments were conducted utilizing a 0.1 s nucleation pulse with a current density of 300 A  111 -2  followed by 50 A m -2 of current density continuously  applied for 15 minutes for platinum to grow existing nuclei. Ruthenium was deposited afterwards at a current density of 50 A  111 -2  for 10 minutes to decorate existing Pt  particles. Each metal was deposited at 298 K in a 40 % wt surfactant solution, i.e., in the liquid crystalline phase. The substrate was 6000 [tm thick GF-S6 with 4 cm -2 geometric area. A single counter electrode was used. Fig. 6-18 shows the distribution of catalyst particles on the fibers in the top layer of the felt.  Fig. 6-18: SEM image of PtRu particles sequentially deposited with 40 ^Triton X-100  utilizing a 0.1 s nucleation pulse (i = 300 A 111 -2 ) and a 15 minute growth current density (i = 50 A m -2 ) for Pt and a 10 minute current density for Ru (i = 50 A m -2 ).  The estimated average particle diameter was —200 nm. A higher concentration of particles was found at the top surface of the felt compared to the center region of the felt. When an aqueous solution without surfactant was applied under the same deposition  96  conditions, a thick and highly nonuniform coating (Fig. 6-19) was obtained on the top layers of the graphite felt.  Fig. 6-19: SEM image of a deposit obtained by sequential deposition in aqueous solution without Triton X-100 utilizing a 0.1 s nucleation pulse (i = 300 A m -2 ) and a 15 minute lasting growth current density (i = 50 A m -2 ) for Pt and a 10 minute lasting current density for Ru (i = 50 A m 2 ) .  There was no penetration of catalyst particles into the felt. This underlines the importance of the surfactant, which acts as a leveling agent. The influence of the ruthenium content on activity was tested by preparing samples with a 4:1 and 1:1 Pt:Ru atomic bulk composition. During preparation of both samples platinum was deposited first, utilizing a 0.1 s nucleation pulse at 300 A m 2 followed by a 15 min growth pulse (50 A m -2 ). Ruthenium was deposited afterwards at a current density of 50 A m -2 for 7.5 min and 15 min, respectively. The obtained atomic Pt:Ru ratios were determined by EDX to be 4:1 (Ru deposition for 7.5 minutes) and 1:1 (Ru deposition for 15 minutes). Fig. 6-20 shows the activity of both samples. In this case the platinum weight was determined by titration with ferrocyanide to be about 4 g m -2 (geometric area). To determine the weight of deposited platinum a measured amount of ferrocyanide ([Fe(CN)6] 4- ) was added to the residual solution after deposition. Ferrocyanide reacts with Pt(IV) according to [Saxena, 1966]: 2 H 2 PtC1 6 + 2 Fe(CN)6^[Fe(CN)6]2 + 4 1-1 ± + 12 Cl ^(6-3) -  97  The unreacted ferrocyanide was titrated with cerium sulfate to determine the residual Pt content: [Fe(CN)6]1- + Ce 4+ -- [Fe(CN)6] + Ce 3+  (6-4)  Since the initial Pt content is known the difference in Pt concentration in the solution determined by titration after deposition corresponds to the catalyst load on the felt. This method does not work when Ru is present and could therefore only be applied for the sequential deposition process. The inset in Fig. 6-20 contains the same data expressed as superficial current density. The sample with the lower ruthenium fraction showed higher activity at 298 K, which is consistent with data reported in the literature [Gasteiger et al., 1993]. Continuous PtRu co-deposition yielded better performance than pulsed sequential deposition, in particular on a mass activity basis (compare Fig. 6-16 and Fig. 6-20).  12 - 30  10 ;22•Øaamy,  ... -?"  8-  20  - io  6-  4-  cli E <  0  -  0.3^-0.2^-0.1  0.0  E [V vs. MSE]  2• Pt:Ru = 4:1 Pt:Ru = 1:1  o -  0.4^-0.3^-0.2^- 0.1  ^  0.0  ^  01  E [V vs. MSE]  Fig. 6-20: Influence of Pt:Ru atomic ratio on methanol oxidation in 0.5 M CH3OH0.1 M H 2 SO 4 at 298 K, scan rate = 5 mV s -1 . Sequential pulsed deposition. Samples were prepared by a 0.1 s nucleation pulse (i = 300 A rn -2 ) and a 15 minute lasting growth current density (i = 50 A rn 2 ) for Pt in both cases. For Ru a current density (i = 50 A m -2 ) lasting 7.5 minutes (4:1 Pt:Ru ratio), respectively 15 minutes (1:1 Pt:Ru ratio) was applied.  98  6.1.5 Factorial experimental design (II) (3 ml sandwich plating cell) A second set of factorial experiments with three variables at two levels plus a center 3  point, (i.e., 2 +1 matrix) was carried out to examine the effects of surfactant concentration, current density and time on the catalyst particle diameter range, composition and electro-catalytic activity of the three-dimensional GF-S3 based anodes with respect to methanol oxidation. The anodes were prepared at 298 K in the 3 ml sandwich cell under the conditions summarized in Table 6-8. Pt and Ru salts were supplied at a concentration of 65 mM for each metal. Compared to the glass beaker cell the required plating bath volume was decreased by a factor of 33, so that overall less metal salt was required. The inter-electrode gap is also much smaller (-150 iim vs. —1 cm), which decreases the Ohmic drop. Cyclic voltammograms of CH3OH electro-oxidation on selected samples from the factorial design are shown in Fig. 6-21 and 6-22. Table 6-9 contains the respective catalyst loads and bulk atomic Pt:Ru ratios.  Table 6-8: Factors and their levels used in the factorial design for electrodeposition. High Centre Variable^Range Low [Triton X-100] 40 60 50 %,t 40 i A m -2 20 60 min 260 175 t 90 Table 6-9: Deposition parameters and catalyst properties for selected samples from the factorial experiment. PtRu electrodeposited onto GF-S3 in the sandwich cell at 298 K PtRu load [Triton X-100] i t [g m-2] [min] [A m -2 ] rowti 4.5 20 90 60 9.2 40 60 90 260 7.0 20 40 7.0 175 50 40  Atomic Pt:Ru ratio of deposit 1.3:1 1:1 1.3:1 1.1:1  Increasing the deposition current density decreased the Pt:Ru bulk atomic ratio, irrespective of surfactant concentration and deposition time (Table 6-9). This indicates that catalysts with different Pt:Ru atomic ratio can be prepared by varying the current density. This general observation is also in agreement with results found in factorial (I). 99  The obtained Pt:Ru ratios were lower compared to using the glass beaker cell with one counter electrode. a) 1000 - ••■■ 60% Triton X-100 20 A  M -2 90 min •••• 40% Triton X-100 60 A M -2 90 min 40% Triton X-100 20 A m -2 260 min •• 50% Triton X-100 40 A M -2 175 min 800 -  600 -  •  • • • •  400 -  ••  •  200 •  •  •........ • • • • -0.6  ^  -0.4  ^  -0.2^0.0  ^  0.2  ^  0.4  E [V vs. MSE]  b)^E [V vs. MSE]  -0.60 -0.55 -0.50 -0.45 -0.40 -0.35 -0.30  120 ^100  ..,4o ^/ , .^/ 30  0.1 M FI 2 SO 4^.  ^80  E  0.5 M CH 2 OH +^ •:4 1 -^.^ 20  ,^ #•• / ••  -^•••• 0.1 M H2SO4^10,^ . ..................... ..,  /^ de •• •• •^/^• • • I^40  60 N.--,^ E 40 < 20 -  •••••• ....... •• ^/  •'^ #9  •  ^•  ......^  0•  ••  •  00  ••••••••• 0 -6.11‘61 ao ••••■•• -20 -^  ■••• 60% Triton X-100 20 A m -2 90 min •••• 40% Triton X-100 60 A m -2 90 min ••• 40% Triton X-100 20 A M -2 260 min •••• 50% Triton X-100 40 A m -2 175 min  -40 ^ -0.6^-0.4^-0.2^0.0^0.2 E [V vs. MSE]  Fig. 6-21: (a) Voltammogram of methanol electro-oxidation on selected catalyst samples from the factorial experiment. PtRu deposited on GF-S3. 0.5 M CH 3 OH-0.1 M H2SO4, -1 5 mV s , 298 K. (b) Enlarged view of the methanol oxidation onset. Inset shows comparison with blank scan (0.1 M H2SO4) for the best performing sample (40 %„,t Triton X-100, 60 A 111-2 , 90 min). 100  Figs. 6-21 (a) and 6-22 show that deposition in 40 °A m Triton X-100 at 60 A m 2 for -1 90 min yielded the highest superficial current density and mass activity (e.g., 110 A gpt at 0 V vs. MSE), whereas electrodeposition under the centre-point conditions provided -1 the worst performance (e.g., 10 A gp t at 0 V vs. MSE). For comparison, potentiostatic measurements published in the literature using PtRu(1:1) dispersed onto 1 cm 2 Toray carbon paper electrodes (immersed in 1.5 M CH 3 OH-1 M H 2 SO 4 ) yielded a mass activity of 30 A gp t 1 at 0 V vs. MSE and 298 K [Neergat et al., 2002]. At a potential of –0.2 V vs. MSE, relevant for practical fuel cell applications, the highest mass activity for CH3OH electro-oxidation was about 30 A gpt 1 . For all samples the methanol oxidation onset was observed between –0.4 V and –0.35 V vs. MSE (Fig. 6-21 (b)). The plots indicate that adsorption and dehydrogenation of methanol start at –0.6 V vs. MSE. The pronounced peak recorded at –0.5 V vs. MSE for the best performing sample may in part be due to oxidation of adsorbed hydrogen, since a peak is observed at the same potential in 0.1 M H2SO4 (inset of Fig. 6-21 (b)) without methanol. However, in the presence of methanol the peak current density at –0.5 V vs. MSE is seven times higher, suggesting that the dissociative adsorption and electro-oxidation of methanol is the main reaction. It is well known that methanol is adsorbed and dehydrogenated in this potential range [Iwasita, 2002, Waszczuk et al., 2001]. CO oxidation occurs at higher potentials as it necessitates adsorption of OH species on Ru surface sites (i.e., at —0.4 V vs. MSE [Gojkovic et al., 2003]). The methanol oxidation current response showed no peak at potentials more positive than –0.2 V vs. MSE, indicating the absence of either diffusion limitation or poisoning by adsorbates. This type of current–potential response was also reported for CH 3 OH oxidation on PtRu obtained by pulse electrodeposition on glassy carbon [Wei, Chan, 2004].  101  ^  180 160 140 -  a  a)  ■•■ 60% Triton X-100 20 A m -2 90 min •••• 40% Triton X-100 60 A m -2 90 min IM 40% Triton X-100 20 A m -2 260 min •• • 50% Triton X-100 40 A mr 2 175 min  •  120 -  ••  < 100 -  • •  •  • • •  •  •  E  40 20 0-20 -0.6  ^  -0.4  ^  -0.2^0.0  ^  0.2  ^  04  E [V vs. MSE]  Fig. 6-22: Voltammogram of methanol electro-oxidation in terms of mass activity on selected catalyst samples from the factorial experiment. PtRu deposited on GF-S3. 0.5 M CH301-1-0.1 M H2SO4, 5 mV s -1 , 298 K.  ^1000  .•^160.-^ 1^ 40 tr.^.• . :. •:^120 Q < :^100 /..^.• 800 -^• 80 . 5^: .^ T.)^• - 60 co -^  Pt •••• PtRu  CO •  40 to m• N 600 -^• • ^ 0.1^•^1100000."....20 E•  E^kaalasa.s*  ^-0.6^-0.4^-0.2^0.0^0.2^0 4  o ...  400 -^E [V vs. MSE]^:  •  200 -  -0.6  -0.4  •  -0.2^0.0  0.2^0.4  E [V vs. MSE]  Fig. 6-23: Voltammogram of methanol electro-oxidation using catalyst deposited on GFS3: Comparison of Pt and PtRu. 0.5 M CH3OH-0.1 M H2SO4, 5 mV s 1 , 298 K. Deposition parameters: 40 %,„ 4 Triton X-100, 60 A m -2 , 90 min, 298 K.  102  For comparison, a pure Pt catalyst was also prepared at 298 K (40 Vo wt Triton X-100, 60 A m 2 , 90 min). The obtained load was 8 g m -2 and the specific surface area was 59 m 2 gpt- 1 . Compared to the PtRu sample prepared under the same conditions, the onset of methanol oxidation on pure Pt occurred at a higher potential (—. -0.15 V vs. MSE compared to — -0.3 V vs. MSE). At 0 V vs. MSE the superficial current density obtained with the PtRu catalyst was 12 times higher and the Pt mass normalized current density was 17 times higher compared to the Pt catalyst (Fig. 6-23). Quantitative comparison of voltammetry results published by different research groups is not straightforward due to different catalyst preparation methods, substrates, and testing conditions (e.g., scan rate, concentration, temperature). Mass activity is hereby suggested as the most applicable parameter for comparison of the results as it quantifies the catalyst utilization. The potentials reported in the literature were converted to V vs. MSE. Bock et al. reported a mass activity of 3 A gp t 1 at 293 K corresponding to a potential of —0.2 V vs. MSE for a PtRu (1.2:1) alloy powder dispersed on Au. The onset of methanol oxidation was observed at —0.3 V vs. MSE [Bock et al., 2004]. Dubau and coworkers investigated nanoparticles produced by a colloidal method and dispersed onto a glassy carbon substrate. The obtained electrodes were scanned at 1 mV s I using 1 M CH3OH in 0.5 M H 2 SO4. At 298 K and —0.2 V vs. MSE the mass activity reported for -1 PtRu (1:1) was 40 A g pt [Dubau et al., 2003], which is about the same as for the best performing sample from the present factorial experiment. Electrodeposition of PtRu(1:1) -1 on a Ti mesh resulted in a mass activity for CH3OH oxidation of 45 A gp t at —0.2 V vs. MSE and 298 K [Allen et al., 2005].  103  70 60 70„, Triton X-100 20 A m -2 90 min 60 -  ^ 40 %,, Triton X-100 60 A m -2 90 min ^ ' 40 7.„,, Triton X-100 20 A m -2 260 min  50 a) <  Z.  IMM••■•••■••^  60 % Triton X-100 40 A m -2 175 min  40 -  -  .> -b.^30 c.) co  g 10  1/4^  0  3Warm‘sAm‘sione armreowdbma lamp as Arararbsoiares war  0  20  40^60  ,^ . 80^100^120  140^160  180^200  t [s]  Fig. 6-24 Chronoamperometric test at 0 V vs. MSE, 0.5 M CH 3 OH-0.1 M H 2 SO 4 , 298 K. The voltammetry results were corroborated by chronoamperometry (Fig. 6-24).The current—time responses recorded at 0 V vs. MSE and 298 K show no loss of activity over 180 s for all the samples. The highest obtained mass activity was 24 A gp t 1 . The best performing catalyzed felt anode had the highest specific PtRu surface area, as determined 2^-1 by the Cu underpotential deposition method, i.e., 129 m gptRu , whilst the rest of the 2^-1 samples were characterized by PtRu specific surface areas ranging from 30-70 m gptRu • -2  This also explains the high peak current density of 40 A m for the methanol adsorption peak at —0.5 V shown in Fig. 6-21 (b). Fig. 6-25 summarizes the general trends regarding the bulk Pt:Ru ratio dependence on the deposition parameters.  104  1.8  1.6 -  •  3 re  z^1.4 E ^.`■ -CI:  ,:'  E^1.2 -  E  •  •  •  •  •  •  •  •  •  •  • •  • •  • • •  1.0  •  •  •  0.8 30^40^50^60^2)^33^40^50^60^100^150^200^250  [Yo Triton X-100]  ^  i [Ani2]^  t [min]  Fig. 6-25: Influence of surfactant concentration, current density and deposition time on the atomic Pt:Ru bulk ratio as determined by ICP-AES.  Compared to experiments conducted with GF-S6 where the glass beaker cell was used for electrodeposition (Fig. 6-15) a narrower distribution of the atomic ratios and lower Pt:Ru ratios were obtained. This might be due to a number of factors, such as different graphite substrates and the smaller inter-electrode gap in the sandwich cell configuration, impacting the potential distribution in the two counter electrode cell setup. Fig. 6-26 shows typical catalyst deposit morphologies obtained with Triton X-100 forming micellar or liquid crystalline phases. Particles and agglomerates varying in diameter from —20-60 nm were observed. To further emphasize the importance of adding the nonionic surfactant to the deposition bath, the effect of Triton X-100 on catalyst properties using acidic electrolyte and GF-S3 had to be demonstrated (conditions listed in Table 6-10).  105  S4700 111W 82mm x70.2k SE(U)  Fig. 6-26: SEM image of the PtRu electrodeposited fiber surfaces at 298 K: (a) 60 N4 Triton X-100, 20 A m 2 90 min at 298 K, d p z-; 20-50 nm ,  (b) 50 N„,t Triton X-100, 40 A m 2 ,175 min at 298 K, d p 20-60 nm (c) 40 Nit Triton X-100, 60 A m 2 90 min at 298 K, d p 20--60 nm. ,  106  Table 6-10: PtRu electrodeposition on GF-S3 with and without Triton X-100, 298 K. PtRu electrodeposited onto GF-S3 in the sandwich cell at 298 K [Triton X-100]^i^t^PtRu load^Atomic Pt:Ru [g m -2 ] ratio of deposit MA^[A m-2 ]^[min]^ 7:1 60 0 260 73 60 260 1.5:1 60 6.3 Applying surfactant decreased both the PtRu load and the Pt:Ru ratio. The lower Pt:Ru ratio of 1.5:1 can be considered more beneficial for fuel cell applications [Gasteiger et al., 1994]. Figs. 6-28 and 6-29 show a comparison of the deposit morphologies. Without Triton X-100 a dense coating composed of large (up to —0.5 [tm) continuous agglomerates was obtained (Fig. 6-28). Electrodepositing in micellar solution resulted in highly dispersed particles and agglomerates in the range of approximately 10— 60 nm (Fig. 6-29). The catalyst properties shown in Table 6-10 and Fig. 6-27 and 6-28 had a profound effect on the activity of the PtRu deposit towards methanol oxidation as revealed by the cyclic voltammograms (Fig. 6-29). At potentials greater than —0.3 V the mass activity was significantly higher in the case where the deposition was carried out with surfactant, indicating superior weight-normalized performance of the nano-sized catalyst. This could be at least partly attributed to the higher weight specific surface area, as determined by the Cu underpotential deposition and stripping method, which was 30 m 2 gptR u 1 for the deposit obtained with surfactant, and 3 m 2 gptR u -1 without surfactant, respectively. However, the superficial current density (Fig. 6-29 inset) was higher in the case of the catalyst prepared without surfactant due to the over ten times higher catalyst load. This indicates that while the nano-sized catalyst is utilized efficiently, the total catalyst amount and / or its distribution throughout the three-dimensional electrode should be improved to increase the superficial current density.  107  Fig. 6-27: Representative morphology of the PtRu electrodeposit on GF-S3 obtained  without surfactant (see Table 6-10).  S47001.0kV 5.0nin x150k SE(M) al 105 Fig. 6-28: Representative morphology of the PtRu electrodeposit on GF-S3 obtained in  the presence of Triton X-100 (micellar phase) (see Table 6-10).  108  ^  80 •500 - 400  60 -^  300  c11" E  200  40 .  - 100 0  5  ••  -0.4^-0.2^0.0^0.2^0.4  -^E [V vs. MSE]  ^20  .  co°  .•.-  .•• .•••  .•.•  E ......... .  0"• 60 % Triton X-100 60 A m -2 260 min •••• 60 A m2 260 min without surfactant  -20 -0.6^-0.4^-0.2^0.0  ^  0.2  ^  0.4  E [V vs. MSE]  Fig. 6-29: Voltammogram of methanol oxidation on electrodeposited PtRu using a GF-S3 substrate: Comparison between the catalysts produced by electrodeposition with and without Triton X-100 present. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s 1 , 298 K.  6.1.6 Deposition temperature and catalyst activity This section describes an approach to alter the catalyst load and Pt:Ru bulk atomic ratio as well as the active area by varying the electrodeposition temperature. The following parameters were kept constant for an investigation regarding the influence of the electrodeposition bath temperature on the catalyst properties: - surfactant concentration = 40 %,„- t - current density = 60 A rT1 -2 - deposition time = 90 minutes - Pt and Ru salt concentration = 65 mM each These parameters were determined to be favorable for co-deposition based on factorial experiments at 298 K as described in Section 6.1.5. The substrate was GF-S3 and the sandwich type cell was used. Three temperatures above 298 K were selected (T = 323 K, 333 K and 343 K). The resulting catalyst properties are summarized in Table 6-11.  109  •  Table 6-11: Deposition temperatures and resulting catalyst properties. T^PtRu load^Atomic Pt:Ru^Active area [K]^[g m -2 ] ratio of deposit^[m2 gpt -l]^[m2 mfeit 298 323 333 343  9.2 3.7 5.5 3.9  1:1 1.4:1 1.4:1 1.1:1  198 201 160 62  2]  1180 530 650 470  The PtRu load was decreased compared to plating at 298 K in every case. The voltammetric data presented in Figs. 6-30 and 6-31 show that the highest superficial current density was obtained when the electrodeposition was carried out at 333 K. The corresponding peak at -0.47 V indicates the pronounced dehydrogenation of methanol compared to other samples.  300  E  deposition deposition ▪ deposition ■••=.• ••••^deposition  at at at at  T T T T  = = = =  298 323 333 343  K K K K  200 -  100 -  • -0.6  -0.5  -0.4  -0.3  -0.2  E [V vs. MSE]  Fig. 6-30: Methanol oxidation superficial current density 298 K: Effect of -1 electrodeposition temperature; 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s . The higher methanol oxidation superficial current density obtained with the sample prepared at 333 K relative to the electrodes prepared at 323 K and 343 K can be explained by the larger active area and higher catalyst load. The same trends have been obtained when the methanol electro-oxidation was studied at 343 K (Figs. 6-32 and  6-33). The electrode prepared at 333 K was most active with respect to methanol electro-oxidation as also shown by chronopotentiometry (Figs. 6-34 and 6-35).  110  Based on the lower atomic bulk ratios it can be assumed that the samples prepared at 298 K and 343 K have a lower Pt surface fraction that is available for methanol adsorption compared to the catalyst obtained at 333 K.  Fig. 6-36 shows a representative deposit of the catalyst prepared at 333 K. Agglomerates of —100 nm in diameter were observed while most particle diameters measured —10-50 nm. For comparison, a pure Pt catalyst was prepared at 333 K (40 Vo wt Triton X-100, 60 A m -2 , 90 min). The obtained load was 6 g  M -2  and the specific area was  512 m 2 p t 111-2 felt or 85 m 2 gpt-1 . The onset for the methanol electro-oxidation was observed at — -0.3 V at 343 K. The onset for PtRu prepared at 333 K was shifted to —0.45 V at 343 K. The shift to the lower potential indicates that water splitting and consequently CO oxidation occurs at a lower potential when Ru is present. The superficial current density at -0.2 V and 343 K was —15 times higher for PtRu compared to Pt (both prepared at 333 K) and the Pt mass specific current density was increased by a factor of 22 (compare  Fig. 6-32 and Fig. 6-37). Further, the anode potential required to maintain a current density of 50 A m -2 at 343 K was decreased by 0.24 V with PtRu compared to pure Pt (compare Fig. 6-35 and Fig. 6-38).  deposition deposition deposition •■= • • •■• • .■ ••^deposition  120 - ^  100  at at at at  T T T T  = 298 K = 323 K = 333 K = 343 K  .I  '."..  0.  <  80 -  :•?; o co  60 -  ii9 ?  i t i Al%^•  I: I.  40 -  .. •  20  ...."••°. '.‘  •  • ..^  ,•••••,•  / • 0 : I • "•*........••••°°  ..............................  ...•••■•••••■•••■ •  •  : ...  ". •■••  -0.6^-0.5^-0.4^-0.3  ^  -0.2  E [V vs. MSE]  Fig. 6-31: Methanol oxidation mass activity 298 K: Effect of electrodeposition temperature; 0.5 M CH3OH-0.1 M H2SO 4 , 5 mV s 1 .  111  600 1=1■  / / / /  deposition at T = 298 K deposition at T = 323 K deposition at T = 333 K deposition at T = 343 K  •• ■ •• =1•1• •  200 -  oo  -0.6  ^  -0.5  ^  -0.4  ^  -0.3  ^  -0.2  ^  -0.1  E [V vs. MSE]  Fig. 6-32: Methanol oxidation superficial current density 343 K: Effect of electrodeposition temperature; 0.5 M CH 3 OH-0.1 M H2SO 4 , 5 mV s 1 .  200 -  deposition at T = 298 K deposition at T = 323 K deposition at T = 333 K M10••^•••••••^deposition at T = 343 K  ••• / r  .—.  •8: 150 a) < Z'  t.')^100 cc N N  is E  50 -  -0.6  ^  -0.5  -0.4^-0.3  -0.2  ^  -0.1  E [V vs. MSE]  Fig. 6-33: Methanol oxidation mass activity at 343 K: Effect of electrodeposition 1 temperature; 0.5 M CH 3 OH-0.1 M H2SO 4 , 5 mV s .  112  -0.20  r  "14.10,1 1,  ..... IllallalbagaD OD OW 1 • falaDaeal aal• a 1 eillaaaaaa GIMP  I  -0.25 -  I I  r  ,  I  I  -0.30 -  .., °°°°°°°°° a " °°°°° . .... .., . al •  CD  2 ui  > -0.35 -  • •  Li' -0.40 -  -0.45 -  •  •• • • •  •  • -  • •11, • .. • • . • 00 . 0 6 o a 9 • a. 10 al • • a  .  ..  deposition at T = 323 K deposition at T = 333 K •• • deposition at T = 343 K  • a  0  ^  50  ^  100  ^  150  t Es]  Fig. 6-34: Chronopotentiometry: Effect of electrodeposition temperature on the catalyst activity for methanol oxidation at 298 K, i = 50 A m -2 , 0.5 M CH 3 OH-0.1 M H 2 SO 4 .  -0.34  -0.36 -  -0.38 ril U) -0.40 -  2 ui >  -0.42 -  w -0.44 -  -0.46 -  -0.48  0  deposition at T = 323 K deposition at T = 333 K deposition at T = 343 K ■^ ' 50^100^150  t Es]  Fig. 6-35: Chronopotentiometry: Effect of electrodeposition temperature on the catalyst activity for methanol oxidation at 343 K, i = 50 A m -2 , 0.5 M CH 3 OH-0.1 M H2SO4.  113  Fig. 6-36: Micrograph of PtRu electrodeposited on graphite felt at 333 K (see Table 6-11).  1"o',  (-11  K -298K •••• 343 K -0.1^0 0  E [V vs. MSE]  Fig. 6-37: Voltammogram of pure Pt catalyst on GF-S3, 0.5 M CH3OH-0.1 M H2SO 4 , 5 mV s . Deposition parameters: 40 %„ t Triton X-100, 60 A m -2 , 90 min, 333 K.  114  -0.05  -0.10 -  r^  •••••••• ei••••••••••• •••••••• • • • • •••• ^ •••••••••••• • •••••••••• •  -0.20 -  -0.25 -  IA  • 298 K •••• 343 K  S.  0  ^  50  ^  100  ^  150  t [s] Fig. 6-38: Chronopotentiometry: Pure Pt catalyst on GF-S3, 0.5 M CH3OH-  0.1 M H 2 SO 4 , i = 50 A I11 -2 . Deposition parameters: 40 % wt Triton X-100, 60 A m -2 , 90 min, 333 K.  6.1.7 Pt, Ru and PtRu deposition studies by linear voltammetry  To better understand the effect of Triton X-100 on the electrodeposition and -I especially on the Pt:Ru ratio, linear voltammetry at a sweep rate of 5 mV s was carried 2  out at 298 K using 1 cm pieces of graphite felt (GF-S6, 2000 i_tm thickness). Fig. 6-39 (a) shows that without surfactant the cathodic superficial current density  was low at potentials more positive than —0.7 V vs. MSE indicating a low Ru deposition rate. At potentials more negative than —0.7 V vs. MSE the current efficiency for Ru deposition was lowered by competing H2 evolution. During Pt deposition higher cathodic currents were measured with peak potentials at —0.57 and —0.80 V vs. MSE due to Pt nucleation and growth, respectively. Furthermore, the co-deposition of PtRu yielded a current response similar to that of the pure Pt deposition, indicating that in the absence of surfactant mainly Pt was deposited. Thus, the voltammetry data supports the observation that electrodeposition in the absence of surfactant yields a platinum rich catalyst (see Table 6-10).  115  0 ....................................... .......  -40  -60  -80  MP. 1=1, 1M•  1 mM Pt 1 mM Ru 1 mM Pt + 1 mM Ru without surfactant  -100 -1.0  -0.8^-0.6^-0.4  -0.2  00  E [V vs. MSE]  • an. • • os• • • 4mo • • 40= • •  ow • • ■mr • • am • • • • ■::....001100  E  1 mM Pt without surfactant •••• 1 mM Pt 1 Triton X-100 -- 1 mM Pt 10 ./•„t Triton X-100 •■•••• 1 mM Pt 50 %,,„ Triton X-100 -100 -1.0  -0.8^-0.6^-0.4  -0.2  00  E [V vs. MSE]  E  E [V vs. MSE]  Fig. 6-39: Linear sweep voltammetry of Pt, Ru and PtRu electrodeposition on graphite felt in 0.1 M H 2 SO 4 : (a) without surfactant, (b) the effect of Triton X-100 concentration on the Pt deposition, (c) with 50 % wt Triton X-100. Scan rate = 5 mV s I , 298 K.  116  Increasing the surfactant concentration diminished the Pt deposition current as shown in Fig. 6-39 (b). Interestingly, when 50 Triton X-100 was applied the current for Pt and Ru deposition was almost identical (Fig. 6-39 (c)). In the case of Ru, a peak and a shoulder were observed at —0.67 and —0.57 V vs. MSE, respectively. In this regime the current was more negative than for Pt deposition. The current measured during codeposition in the presence of 50 % wt Triton X-100 at potentials more positive than —0.7 V vs. MSE was virtually the sum of the deposition currents of each of the single metals. This indicates that both metals are deposited at substantial rates, thus potentially yielding a catalyst with approximately equal fractions of Pt and Ru. The atomic Pt:Ru ratio of 1.1:1 obtained by using 50 % wt surfactant (see Table 6-9) is in agreement with this assumption. During co-deposition with 50 % wt Triton X-100 the magnitude of the cathodic current was lower compared to deposition without surfactant in spite of ten times higher Pt and Ru ion concentration (compare Fig. 6-39 (a) and (c)), except for potentials more negative than --0.87 V vs. MSE, where a significant fraction of the measured current is presumably due to hydrogen evolution.  6.1.8 Comparison of PtRu and PtRuMo Catalysts 6.1.8.1 Single step co-deposition Mo was reported in the literature as a promising co-catalyst in conjunction with PtRu (see Section 3.5.2). The deposition method outlined in Section 6.1.6 (deposition at 333 K) was used to compare PtRu (1:1 in solution) and PtRuMo (1:1:05 in solution).  Table 6-12 contains information regarding the resulting catalyst composition and active area. The atomic ratio was determined by ICP-AES and EDX. The two methods yielded considerably different bulk ratio values for the binary catalyst. The response signal for Pt is stronger than for Ru when conducting EDX.  Table 6-12: Comparison of PtRu and PtRuMo prepared at 333 K, i =60 A m -2 , 40 %,t Triton X-100, t = 90 min. Catalyst PtRu load Atomic Pt:Ru ratio of deposit ^Active Pt area [g m -2 ] ICP^EDX^[m2 gpt i ^[m2 mfeit -2 ] ]  PtRu^5.5^1.4:1^2.4:1^160^650 PtRuMo^15^1.3:1:0.03^2:1:0.1^71^545  117  -0.6  ^  -0.5  ^  -0.4  ^  -0.3  ^  -0.2  E [V vs. MSE]  Fig. 6-40: Electro-oxidation of methanol on PtRu and PtRuMo electrodeposited on GF-S3 in the presence of 40 %,„, t Triton X-100. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 298 K.  •  PtRu  ••• • PtRuMo T = 343 K •  I^  I^  I^  I^  ,  ^-0.7^-0.6^-0.5^-0.4^-0.3^-0.2  E [V vs. MSE]  Fig. 6-41: Electro-oxidation of methanol on PtRu and PtRuMo electrodeposited on -1 GF-S3 in the presence of 40 %,„4 Triton X-100. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 343 K.  118  The voltammetric scans indicated no significant differences in terms of methanol oxidation superficial current density at 298 K (Fig. 6-40). The Pt load normalized mass activity observed for PtRu was substantially higher compared to PtRuMo (Figs. 6-40 and  6-41). The chronopotentiometry responses shown in Fig. 6-42 suggest that the electrode coated with the ternary catalyst performs better compared to PtRu, in particular at 343 K. Mo is believed to have bifunctional properties similar to Ru, i.e., it may facilitate adsorption of OH groups [Bolivar et al., 2003]. Samjeske et al. explained the benefits of Mo based on the assumption that the activation barrier for the oxidation of weakly adsorbed CO is lowered due to an oxygen spillover effect [Samjeske et al., 2002]. It is noteworthy that supplying Pt, Ru and Mo using a 1:1:1 ratio in solution yielded no improvement compared to results presented in this section. On the other hand, using a 1:1:0.25 ratio resulted in a lower catalytic activity.  -0.30 -0.32 -0.34 -0.36  ruin  -  - 0.38 -  2 -0.40 ui  ••  ••• e .. •I / • a•• . 11, I'NINA 8 11".... g  •  ••  > -0.42 -  ^. uJ -0.44 -  •••••••••••••••••••••••• ••••••••••••••••••••." ••••••••••••••••••••  wIlla.  ••••■ ■•••■•■■■•■ ■■••■•••••••■••••••• Am  "1"•4• 011■••••••••••••••••••••■■••■•••••••••••••••■•••■•••wmb  -0.46 PtRu T = 298 K ^ PtRuMo T = 298 K ^ - PtRu T = 343 K ■■•• ^••^••^PtRuMo T = 343 K  -0.48 -0.50 -0.52 0  ^  50  ^  100  ^  150  t [s] Fig. 6-42: Electro-oxidation of methanol on PtRu and PtRuMo electrodeposited on GF-S3 in the presence of 40 %,„, t Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H2SO4, -2  i=50 Am .  119  6.1.8.2 Two-step co-deposition Extending the deposition time did not necessarily improve the catalytic activity. Alternatively a second deposition step with a fresh deposition bath was carried out after cleaning the catalyzed GF-S3 felt. The depositions discussed in this section utilize 40 % wt Triton X-100, a current density of 60 A m -2 and metal salt concentrations of 65 mM (Pt and Ru salts, respectively) and 32.5 mM (Mo salt). After electrodeposition in galvanostatic mode for 90 minutes and cleaning, the catalyzed felt was dried in an oven at 333 K and the electrodeposition was carried again for 90 minutes. The resulting catalysts properties are described in Table 6-13. The load and Mo content were increased compared to single step codeposition (Table 6-12).  Table 6-13: Binary and ternary catalysts obtained by depositing twice at 333 K for 90 minutes. Catalyst^PtRu load^Atomic Pt:Ru ratio of deposit^Active Pt area Eg m. 2 ] ICP^EDX^[m2 gPt 1 ]^[m2 mfett 2 ] PtRu^43^1.4:1^2.5:1^23^712 PtRuMo^52^1:1:0.3^1.7:1:0.2^8^254 The porous catalyst coating consisted of particles and agglomerates with approximate diameters ranging from 10-100 nm for both catalysts (Fig. 6-43). Compared to the sample prepared by a single electrodeposition step, the voltammetric response curves obtained at 298 K showed minor improvement for PtRuMo and a significantly increased methanol oxidation current for PtRu (compare Fig. 6-40 and Fig. 6-44). At 298 K the PtRu catalyst yielded higher oxidation currents than PtRuMo  (Fig. 6-44). This can be explained in part by the larger Pt surface area of the PtRu catalyst (Table 6-13) available for methanol adsorption. It has been established that at low temperatures (e.g., 298 K) methanol adsorption and dehydrogenation does not occur on Ru sites. This was also shown by first principle quantum mechanics calculation of binding energies and heats of formation [Kua, Goddard, 1999]. It is unknown, however, whether methanol adsorbs on Mo, which is thought to have bifunctional properties similar to Ru [Bolivar et al., 2003]. Further, it may lower the activation barrier for COad 120  oxidation [Samjeske et al., 2002]. Increasing the temperature from 298 K to 343 K had a pronounced effect on the Mo containing catalyst at potentials more positive than —0.35 V vs. MSE. The Pt load was the same for both catalysts (i.e., —31 g m -2 ). A pure Pt catalyst that was prepared under the same conditions. The specific area was 17 m 2 gp t-I (620 m2 pt m -2 felt) and the load was 36 g m -2 . At 343 K and 0 V vs. MSE the superficial current density was —15 times higher for PtRu compared to Pt and the respective mass activity was increased by a factor of —9 (compare Fig. 6-44 and Fig. 6-45).  54700-KJC 5.0kV 5.7mm x60.0k SE(M) 7/19/06 ^500nm  54700-KJC 5.0kV 5 7mm x110k SE(M) 7119/06  Fig. 6-43: HiRes SEM micrographs of PtRu (a) and PtRuMo (b) deposits on the graphite fiber surface obtained by double deposition at 333 K.  121  1600 1400 MMOOMMOOMMOO  4,  PtRu T = 298 K PtRuMo T = 298 K PtRu T = 343 K PtRuMo T = 343 K  1200 -  —. loos -  ,  •4 E  < 800 -  /  600 400 -  i  ,...... . ......-  ..0*  .,  6  S  .r..............'"  200 -  sr  .........nle arRakiouwib  -0.6^-0.5^-0.4  -0.1  -0.3^-0.2  00  E [V vs. MSE]  60 PtRu T = 298 K PtRuMo T = 298 K 50 - ^ .. PtRu T = 343 K a••••••••••^PtRuMo T= 343 K  /e ' .  ••  .  to to  to^20 to  O  E 10 -  Pee.  r f  1  l ...°.. • ................... •  ^ j"4114::1; -0.6^-0.5^-0.4^-0.3^-0.2^-0.1^0 0  E [V vs. MSE]  Fig. 6-44: Electro-oxidation of methanol on PtRu and PtRuMo double-electrodeposited on GF-S3 in the presence of 40 Vo wt Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H2SO 4 , -1 5 mV s .  122  •^  200  r_  .*^ 4 150 -3  •  ••  p .•  -2 ▪  • co RI^• E • 1^.• • • • •  100 -  50  -0.6^-0.4^-0.2 E [V vs. MSE]  0.0 •  •  •  •  •  0 -0.6  ^  -0.4^-0.2  ^  0.0  E [V vs. MSE]  Fig. 6-45: Electro-oxidation of methanol on pure Pt double-electrodeposited on GF-S3 in the presence of 40 %,„t Triton X-100 at 333 K. 0.5 M CH 3 OH-0.1 M H 2 SO 4 , 5 mV s 1 .  Additionally, chronopotentiometry for methanol oxidation was performed on the two catalysts at 298 K and 343 K, respectively (Fig. 6-46). At 298 K and 50 A m -2 the anode potential of the PtRuMo catalyst was more positive (by about 60 mV). However, at 343 K an approximately 10 to 30 mV lower anode potential was obtained with the Mo containing catalyst, while having the same Pt load for both catalyst formulations. Thus, there is a strong interaction effect between the presence of Mo and temperature. Addition of Mo to the PtRu catalyst formulation appeared to be beneficial only at higher temperatures such as 343 K and above. The positive interaction effect between temperature and Mo content seems to be in agreement with the observation made by Dickinson et al. regarding Ru, in that a higher Ru content (i.e., PtRu ratio of 1:1 vs. 1.5:1) enhanced the anode performance at high temperatures (e.g., 338 K) [Dickinson et al., 2004]. This is related to the rate determining step shift from CH3OH adsorption/dehydrogenation at low temperatures to the reaction of CO ad with °H ad at high temperatures. The anode potential required to maintain a current density of 50 A  IT1 -2  123  at 343 K was decreased by 0.24 V for PtRu compared to pure Pt (compare Fig. 6-46 and Fig. 6-47).  -0.25  .000.004,„.mogyecoodllommemek",..2:01414. -0.30 -  -0.35 -  2 ei -0.40 ass • aso111.41111••■•■••••••••■•••• • wee  -0.45 -  ••••••■•• ••■•• • NO •  -0.50 I■ • • OM • • IM1 • •  PtRu T = 298 K PtRuMo T = 298 K PtRu T = 343 K PtRuMo T = 343 K  -0.55 0  50  100  150  t rsi Fig. 6-46: Electro-oxidation of methanol on PtRu and PtRuMo double-electrodeposited  on GF-S3 in the presence of 40 Triton X-100 at 333 K. 0.5 M CH 3 OH-0.1 M H2SO 4 , -2  i 50 A m . -0.14  -0.16 -  U  -0.18 -  2 -0.20 -  *".•••■••••■■■ft....••••■•••••,  •owne."•••■• ••••,.......Me...~..~.  -0.22 -  -0.24 • 298 K ^ 343 K -0.26 50  ^  100  ^  150  t [s] Fig. 6-47: Electro-oxidation of methanol on Pt double-electrodeposited on GF-S3 in the  presence of 40 % wit Triton X-100 at 333 K. 0.5 M CH3OH-0.1 M H2SO4, i = 50 A  M 2  .  124  6.1.9 Pt, Ru, Mo, PtRu and PtRuMo deposition studies by linear voltammetry  Linear cathodic sweeps were carried out at 5 mV s -1 and temperatures of 298 K and 333 K to obtain information on the influence of the surfactant concentration on the electrodeposition of the three metal ion species (Pt iv+ , Rum+ and mo m+) onto 1 cm 2 pieces of GF-S3. The scans for the single metal depositions are shown in Figs. 6-48 and 6-49, while Fig. 6-50 contains the data for the co-deposition of PtRu and PtRuMo at 298  K. The concentration for each metal ion species was 1 mM. At potentials more negative than -0.8 V vs. MSE, hydrogen evolution competed with the electroreduction of the metal ions, in particular at surfactant concentrations ranging from 0 %,,,, t to 10 % wt . In each case the current was decreasing with increasing surfactant concentration at potentials below -0.4 V vs. MSE. Irrespective of the surfactant concentration and temperature, the lowest deposition currents were measured for Mo, while Pt electrodeposition yielded the highest currents. At 333 K the Pt deposition onset was observed at — -0.1 V vs. MSE while the deposition of Mo and Ru commenced at potentials lower than — -0.35 V vs. MSE (Fig. 6-49). In the absence of surfactant, for all three species essentially two cathodic  waves can be distinguished. At potentials more positive than -0.8 V vs. MSE, primarily the electrodeposition of Pt, Ru and Mo takes place, while at more negative potentials the secondary reaction of H2 evolution gains significance, thereby, lowering the deposition current efficiency. In the potential domain of interest, i.e., between —0.4 V and —0.8 V vs. MSE where the metal deposition occurs with high current efficiency, the largest cathodic current was obtained for PtC16 2 ". Thus, in the absence of Triton X-100 the electrodeposition of Pt is favored forming a Pt rich catalyst on the graphite felt. The gradual increase of Triton X-100 concentration from 0 to 40 % wt resulted in a significant decrease of the Pt deposition current density.  125  2  a) 1 mM PtC16 2-  .•••• • ••■• ••••^ • -4 - .• ••••• .... .... •• . ..• -6 - /dr -^...... ••• ■, ...• •••  0 %w t Triton X-100  1 %wt Triton X-100  -10  10 ./0„,,t Triton X-100 ....••^•••■■••^40 0/0wt Triton X-100  -12 -1.0^-0.8^-0.6^-0.4^-0.2  ^  00  E [1./ vs. MSE]  2  0-  E  -4  :  0 %wt Triton X-100 1 %wt Triton X-100  -8  10 %wt Triton X-100  11=••■••■••■••^40 %wt Triton X-100 -  10 -1.0  -0.8  -0.6  -0.4  -0.2  00  E [V vs. MSE]  0 %,,,t Triton X-100 1 Triton X-100 10 %wt Triton X-100 40 %wt Triton X-100 -1.0  -0.8  -0.6  -0.4  -0.2  00  E [V vs. MSE]  Fig. 6-48: Linear sweep voltammetry of Pt (a), Ru (b) and Mo (c) electrodeposition on graphite felt, 5 mV s -1 , 298 K.  126  0 Yowt Triton X-100 1 YowtTriton X-100 10 Triton X-100 40 %,Nt Triton X-100  E [V vs. MSE]  4 2-  b) 1 MM RuCI3  0-  ..... .....  -8  • 0 Yam Triton X-100  -10 -  1 %wt Triton X-100 10 Vowt Triton X-100  -12 ■••■■••■••  40 %wt Triton X-100  -14 -1.0  ^  -0.8  ^  -0.6  ^  -0.4  ^  -0.2  ^  00  E [V vs. MSE]  0 % w t Triton X-100 1 % w t Triton X-100 10 % w t Triton X-100 40 % w t Triton X-100  E [V vs. MSE]  Fig. 6-49: Linear sweep voltammetry of Pt (a), Ru (b) and Mo (c) electrodeposition on  graphite felt, 5 mV s -1 , 333 K.  127  -2  1 mM Pt C16 2 ' + 1 m M RuCI3  -4 -6  ...••••••  ...• .•^-----e.  -8  ry E  -10  •  -12  •.  -14 -  •  -16 0 %wt Triton X-100 -18 -  1 %wt Triton X-100 10 %wt Triton X-100  -20 MIN•••••■••■••  40 %wt Triton X-100  -22 -1.0  ^  -0.8  ^  -0.6  ^  -0.4  ^  -0.2  ^  0.0  E [V vs. MSE]  0  1 m M PtC16 2- + 1 mM RuCI3 + 1 mM MoCI5  -5 -  E _10 _  -15 0 Vowt Triton X-100 ^ -20  1 %wt. Triton X-100 10 % wt. Triton X-100 .■,••,...••••••••^40 % wt. Triton X-100  -25 -1.0  ^  -0.8  ^  -0.6^-0.4  ^  -0.2  ^  00  E [V vs. MSE]  Fig. 6-50:  Linear sweep voltammetry of PtRu and PtRuMo co-deposition on graphite felt,  5 mV s, 298 K.  At 40 %,„ 4 Triton X-100 the Pt deposition polarization curve at either 298 K or 333 K became linear up to —0.8 V vs. MSE with a large dE/di ratio. This is due to the large crystallization overpotential as a result of low ad-atom surface diffusivity on the surfactant-covered surface [Fischer, 1969], leading to isolated nuclei formation followed by restricted growth of nuclei to three-dimensional crystallites and possible coalescence into larger aggregates (nucleation-coalescence mechanism for electrodeposition). The effect of surfactant on the electrodeposition current of either Ru or Mo was less pronounced compared to Pt irrespective of temperature. This indicates that the  128  electroreduction kinetics of both Ru and Mo ions are slow, hence, their deposition current density was less affected by ad-atom diffusion limitation and restricted growth effects. 10  0-  1 mM PtC16 2 " + 1 mM RuCI3 .^  ..••• •^ .6••• . •••• •  a* • . .• . •^............. •" .• " .  AI "^  -20  - /^  = -30  / .• 1.'  ;  -40  . •41  .  .  -10 -  E ct  .gte ... ... . • ..  ..  .  ••  I. .•  0 % wt . Triton X-100 1 % wt . Triton X-100  -50  ^ . 10 % wt Triton X-100 40 Yo wt . Triton X-100  -60 -1.0  -0.8  -0.6  -0.4  -0.2  00  E [V vs. MSE]  20  0  -20  E  0 % wt Triton X-100  -60  1 Yo wt Triton X-100 ^ . 10 % wt Triton X-100 ..,••••••••■••^40 % wt Triton X-100  -80 -1.0^-0.8^-0.6^-0.4  -0.2  00  E [V vs. MSE]  Fig. 6-51: Linear sweep voltammetry of PtRu and PtRuMo co-deposition on graphite felt, 5 mV s, 333 K.  The voltammograms obtained at 333 K for co-deposition of PtRu and PtRuMo, respectively, are shown in Fig. 6-51. In the case of both PtRu and PtRuMo, without surfactant present in the deposition bath the co-deposition current at potentials between —0.6 V and —0.8 V vs. MSE is fairly close to the sum of the individual deposition currents of the constituent elements, therefore, a Pt-rich catalyst would be generated. With 40%,t Triton X-100 the co-deposition current is again approximately equal to the sum of the individual deposition currents of the respective elements, and the atomic 129  ratio of the elements can be controlled due to the low Pt deposition current,. Therefore, the presence of surfactant was crucial to control the Pt:Ru and Pt:Ru:Mo atomic ratio in the deposit, by selectively lowering the Pt deposition current density compared to Ru and Mo. The blank experiments carried out at 333 K (Fig. 6-52) show that without metal cations the observed current densities were much lower and indicate hydrogen evolution. The surfactant itself appears to be electrochemically non-reactive.  1 %wt Triton X-100 10 % wt Triton X-100 40 % wt Triton X-100  E [V vs. MSE]  Fig. 6-52: Linear sweep voltammetry with blank graphite felt in 0.1 M HC1 at different 1 surfactant concentrations, 5 mV s , 333 K.  6.2 Tafel Plot Analyis of Methanol Oxidation on PtRu and PtRuMo PtRu(Mo) electrodes, prepared as described in Section 6.1.8.2, were employed in this study to determine the apparent Tafel slopes and exchange current densities for methanol oxidation at 333 K, 343 K and 353 K. The current density responses required for the calculation of the apparent Tafel slopes were measured in chronoamperometric mode after 180 seconds at each potential. During each measurement the current density became constant after approximately 20 — 40 seconds. In the low potential regime, between -0.46 V and -0.37 V vs. MSE, apparent Tafel slopes (b1) ranged from 89 to 120 mV dec -1 for PtRu and from 111 to 155 mV dec .! for PtRuMo (depending on temperature). In the high potential range bh values between 190 and 244 mV dec -1 and 215 and 255 mV dec -1 were calculated for PtRu and PtRuMo, respectively (Fig. 6-53). It is proposed that the shift  130  from the low value to the higher value for the Tafel slope is due to an increased influence of mass transport limitations at the higher potentials. Tafel slopes that were reported in the literature are shown in Table 6-14. Exchange current densities obtained for the catalyzed 3D electrodes are listed in Table 6-15. An exchange current density of 0.08 A m -2 was observed for PtRu dispersed on single-wall carbon nanotubes at 298 K using 1 M H2SO4-2 M CH 3 OH [Liu et al., 2007]. Data obtained with Vulcan supported PtRu and 0.5 M H 2 SO 4 -2 M CH 3 OH at 333 K revealed an exchange current density of 1.5 A m -2 [Schmidt et al., 1999].  Table 6-14: Overview of Tafel slope values reported in the literature. T  Electrolyte  [K]  Anode potential (Range)  Tafel slope  [V vs. MSE]  [mV dec -1 ]  Reference  313  0.1 M HC10 4 -1 M CH 3 OH  -0.34 - -0.24  105-130  [Gojkovic et al., 2003]  298  0.1 M HCIO 4 -0.6 M CH 3 OH  -0.50 - -039  60  [Chrzanowski,  -0.39 - -0.21  200  Wieckowski, 1998]  313  0.5 M H 2 SO 4 -1 M CH 3 OH  -0.30  155  [Khazova et al., 2003]  333  0.5 M HCI0 4 -1 M CH 3 OH  -0.28 - -0.20  110  [Roth et al., 2002]  333  0.5 M H 2 SO 4 -0.5 M CH 3 OH  -0.30 - 0.20  180  [Gasteiger et al., 1994]  80 (Pt only) 333  0.5 M H 2 SO 4 -0.5 M CH3 OH  -0.3^- 0.15  195  [Schmidt et al., 1999]  373  2 M CH 3 OH (DMFC)  -0.44  100  [Arico et al., 2001]  Table 6-15: Exchange current densities for methanol oxidation obtained from Fig. 6-53. i 0 [A m -2 ] PtRu 333 K^0.01 343 K^0.04 353 K^0.16 PtRuMo 333 K^0.06 343 K^0.20 353 K^0.63  131  T = 333 K, bi = 89 mV dec', b h = 190 mV dec"' T = 343 K, bi = 106 mV dec', b h = 229 mV dec' T = 353 K, bi = 120 mV dec"', b h = 244 mV dec -1  0 6^0.8^1.0^1.2^1.4^1.6^1.8^2.0^2.2^2.4^2.6  log (i[A rn -2  28  ])  -0.15  -0.20 -  •  T = 333 K, IN = 111 mV dee, b h = 215 mV dee  O  T = 343 K, bi = 132 mV dec"', b h = 215 mV dec'  ♦  T = 353 K, IN = 155 mV dec', b h = 255 mV dec'  -0.25 -  b)  1" 0 -0.30 • 6  > ?..  -0.35  u.1 -0.40  -0.45  -0.50 08  ^  1 .0  ^  1.2^1.4^1.6^1.8^2.0  log (i[A m -2  ^  2.2  ^  2.4  ^  2.6  ^  28  ])  Fig. 6-53: Tafel plots for PtRu (a) and PtRuMo (b) double-deposited on GF-S3 at 333 K, 0.5 M CH 3 OH-0.1 M H2SO4.  132  6.3 Exploration of Reaction Zone Proton Conductivity The values summarized in Table 6-16 indicate that the bulk electrolyte conductivity for the extended reaction zone anode in an operating DMFC should be at least equivalent to that of 0.5-1 M sulfuric acid and the electrolyte volume fraction (1) should be higher than 0.5 to avoid excessive Ohmic losses (i.e. AEoh. < 45 mV).  Table 6-16: Protonic conductivity and Ohmic potential loss at 333 K as a function of  electrolyte content (based on H 2 SO 4 as a model protonic conductor), i = 3000 A m -2 , felt thickness of GF-S3 in compressed state = 15011m. [H2SO4] [M] 0.1 0.5 1 2 3 4 5  KO  [S 111-1 ] 5.6 28.7 55.0 95.8 118.2 124.5 119.7  =  (f 0.1) [S M.1 ] 0.2 0.9 1.7 3.0 3.7 3.9 3.8  K  AEoh m [MV] 2250 500 265 150 122 115 118  K  (f=0.25) [S 111-1 ] 0.7 3.6 6.9 12.0 14.8 15.6 15.0  AEoh m [my] 643 125 65 38 30 29 30  (f=0.5) [S 111-1 ] 2.0 10.1 19.4 33.9 41.8 44.0 42.3  K  AEcth m [mV] 225 45 23 13 11 10 11  K=K0 333 K(0 1 5 [Prentice, 1991]  K  (f=0.75) [S 111-1 ] 3.6 18.6 35.7 62.2 76.8 80.9 77.7  (6-5)  fi* dx A^= Ohm  0  ^i •  AX  (6-6)  K^K  Assumption: Felt thickness Ax = 150 i_tm under compression.  Various materials were explored with the aim of establishing a uniform 'solid' proton conducting network to replace the sulfuric acid based liquid supporting electrolyte. Protonic conductivity is essential to utilize all catalyst sites present in a three-dimensional electrode. Samples containing all three essential components, i.e., the felt, catalyst and proton conductor were tested by voltammetry or chronopotentiometry and in some cases in a fuel cell assembly. Nafion was used without further modification and in acidified form. Furthermore Kelzan, an industrial grade Xanthan gum material was investigated in  AEOhm [mV] 125 24 13 7 6 6 6  acidified form. In another set of experiments silica based gels containing H 2 SO 4 were employed.  6.3.1 Nafion Nafion is the most commonly used proton transporting material for PEM fuel cells. For the conventional DMFC anode design adding Nafion to the catalyst layer provides sufficient protonic conductivity since the active sites are located close to the membrane. However, one can argue that depending on the thickness of the Nafion layer a certain limitation regarding transport of methanol to the active sites may be imposed. The inclusion of Nafion in Pt based nanoparticle-inks was reported to lead to a 40 % decrease in oxidation current density and 13 % decrease in active surface area [Jayashree et al., 2005]. Pieces of graphite felt were coated after the catalyst deposition step by dipping into 1 %,t Nafion solution for 1 minute followed by air drying at ambient conditions and recasting the film at 393 K for 30 minutes. Nafion coated fibers without catalyst are shown in Fig. 6-54. The approximate coating thickness is presented in Table 6-17.  Fig. 6-54: Micrograph of recast Nafion on GF-S6 type graphite felt.  134  The data shown in Figs. 6-55 and 6-56 was obtained with samples prepared by depositing PtRu at i = 20 A m 2 for 60 min at 298 K and i = 60 A m -2 for 30 min at 333 K, respectively. The surfactant concentration was 40 Triton X-100 in each case.  .....•••••••••  •  • 30 g m -2 Nafion •••• without Nafion -1.5 -0.6  ^  -0.4  ^  -0.2^0.0  ^  0.2  ^  0.4  ^  06  E [V vs. MSE]  Fig. 6-55: Voltammogramm comparing methanol electro-oxidation at 298 K with and without Nafion coating. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s -I . Substrate: GF-S6, thickness 2000 p.m.  A significant decrease of the methanol oxidation current density was observed at E > -0.2 V vs. MSE (Fig. 6-55), which implies that the Nafion coating imposed a substantial barrier for the transport of methanol to the catalyst sites. Further, the hydrogen evolution observed below -0.2 V vs. MSE without Nafion also indicates that Nafion was blocking the Pt surface. When Nafion was used in acidified form (catalyzed felt was dipped for 1 minute in 1 M H 2 SO 4 / 1 % Nafion solution) the decrease in performance was still profound compared to the uncoated sample (Fig. 6-56). For a comparison of the different Nafion coatings in a fuel cell, PtRu was deposited on GF-S3 using i = 60 A m -2 for 260 min at 298 K without surfactant. The load was — 70 g m -2 and the Pt:Ru bulk ratio was —7:1. Fig. 6-57 shows that a marginal improvement in performance was observed when the Nafion containing coatings were applied compared to utilizing the catalyst with an aqueous methanol feed with no proton conducting 135  material present. The graph also shows a significant increase in cell performance when methanol is fed in conjunction with 0.5 M sulfuric acid. -0.10 -0.15 -  t I I  -0.20 • N  -0.25 -  6  -0.30 -  .......,•••/.•••10••••■•••••••■•••••■••••• 111111• • •■•••1111 ••■•  -0.35 T = 343 K Nafion / 1M H2SO4  -0.40 -  T = 343 K without Nafion T = 298 K Nafion / 1M H2SO4  ▪ -0.45 -  •■•••••••••■■•• 0  ^  50  T = 298 K without Nafion  ^  100  ^  150  t [s]  Fig. 6-56: Chronopotentiometric scans on PtRu catalyzed GF-S3 with and without Nafion-1 M H2SO4 coating (-30 g m -2 ). 0.5 M CH 3 OH-0.1 M H2SO4, 10 A m -2 . Prepared in 3 ml sandwich plating cell.  0.7  0.6 g'v V V  0.5 -  V  V V  ■—■ 0.4 -  o  Ur  0.3 -  V V  v O  •  0.2 -  0.1 -  0.0  0  ^  100  ^  200  •  no coating  O  30 g m -2 Nafion  ♦  30 g m -2 Nafion/1 M H2SO4  o  no coating; anode feed with 0.5 M H2SO4  ^  300^400  i [A m -2  ^  500  ^  600  ]  Fig. 6-57: DMFC polarization curves obtained at 333 K; anode: 1 M CH3OH, 5 ml min -I , —100 kPa(abs); cathode: Pt black (40 g m -2 ) dry 02, 500 ml min STP, —200 kPa(abs).  136  6.3.2 Kelzan Kelzan is applied in industrial processes to thicken and stabilize aqueous suspensions, emulsions, and foams. The molecular structure of this agent is presented in Fig. 6-58. The average molecular formula is  C32 71148 9026 8  (molar mass = 869.3 g mo1 -1 ). Mixtures  containing the material can have pseudoplastic (shear thinning) qualities. A Kelzan containing solution is viscous when quiescent and becomes thinner when it is mixed. To form a model protonic conductor, Kelzan was mixed with 1 M H 2 SO 4 for 10 minutes at 333 K forming a foam or paste like substance. Catalyzed graphite felt was immersed in the mixture, which was then allowed to cool to 298 K. GF-S3 with PtRu electrodeposited at 298 K with 60 % wt Triton X-100, i = 60 A M-2 and t = 260 min was used. The load of the acidic Kelzan coating was —40 g m -2 . The approximate coating thickness is presented in Table 6-17.  Fig. 6-58: Molecular structure of Kelzan Xanthan gum. [http://www.cpkelco.com/xanthan/industrial/molecularstructure.html]  Fig. 6-59 shows a micrograph of fibers coated with Kelzan. The voltammogram in Fig. 6-60 indicates a minor decrease in the methanol electro-oxidation current density at 333 K. At 298 K a strong inhibition of the reaction was observed, which was most likely due to mass transfer constraints.  137  Fig. 6-59: SEM micrograph of graphite fibers coated with Kelzan film.  100 40 30  80 -  • • •  20 (9"  E 60 -10  40 -  0.6^-0.4^-0.2^0.0^0.2^0.4  E [V vs. MSE] 20 -  •  •  •  •  ••  •  •. • I.  • • • • •  •  20 • • •  •• • 40 g Kelzan •••• without Kelzan T = 333 K  -20  •^ -0.8  -0.6  -0.4^-0.2^0.0^0.2  0.4  06  E [V vs. MSE]  Fig. 6-60: Voltammogram measured in 0.5 M CH3OH — 0.1 M H2SO4, Scan rate = 5 mV s -1 . Effect of acidified Kelzan coating on the fibers.  The increase in temperature may favor diffusion of methanol through the Kelzan coating. Further, it is possible that the coating dissolved to a certain extent at 333 K, making the catalyst surface more accessible for the reactant.  138  6.3.3 Acidified Si gel  Fumed silica particles (Degussa Aerosil A 300, BET surface area = 300 +/-30 m 2 g -I , average particle diameter = 7 nm) were mixed with sulfuric acid for 10 minutes at 333 K and then cooled to 298 K to obtain a gel with an acidity equivalent to 1 M H 2 SO 4 . The gel was then applied onto catalyzed GF-S3 with a spatula resulting in an approximate load of 50 g m -2 . The approximate coating thickness is presented in Table 6-17.  Fig. 6-61: Distribution of dried silica gel on graphite fibers.  Fig. 6-62: Distribution of dried silica gel on graphite fibers (continued).  139  ^  ^-  The silica based composition had to be dry before the SEM analysis, since the microscope operates under vacuum conditions. The micrographs shown in Figs. 6-61 and 6-62 show the distribution of the dried gel within the fiber matrix.  700 500 .  600 -  - 400  61" .... E^ft.. * <^ - 200 —^•• .0  300 500 -  ur 100^....  400 E < 300 200 -  .  .  ..  0  .. So  o  Se 9.  r  -0.6^-0.4^-0.2^0.0^0.2^0.4 .::. E [V vs. MSE] 1 % w t Si gel  100 -  •••• without Si gel •^ -0.6  ^  -0.4  T = 333 K  ^  -0.2  ^  0.0  ^  0.2  ^  04  E [V vs. MSE]  Fig. 6-63: Voltammogram measured in 0.5 M CH3OH — 0.1 M H2SO4, Scan rate = 5 mV s -1 . Effect of acidified SiO2 gel coating on the fibers.  The voltammograms in Fig. 6-63 indicate a strong limitation imposed by the gel at 298 K while the loss in performance was less pronounced at 333 K. The catalyst was prepared at 298 K utilizing 60 Vo wt Triton X-100, i = 60 A 111 -2 and t = 260 min. To assess the extent of sulfuric acid leaching, 5 cm 2 pieces of GF-S3 were coated with gel and immersed in a beaker containing 100 ml deionized water. Samples of 5 ml volume were taken from the bulk solution over time and titrated against NaOH using phenolphthalein to back-calculate the amount of sulfuric acid remaining within the gel coating the catalyzed felt. A stronger leaching effect was observed during the first 100 minutes (Fig. 6-64). Similarly, significant leaching can also be expected for the acidified Kelzan coating that is described in Section 6.3.2.  140  • test 1 0 test 2  1.0  0  0  0  II  0.9  0  4 0  UL  •  0  I .". 0.8  • 0  M  0.7  0.6 0  50^100^150  ^  200  ^  250  t [min]  Fig. 6-64: Acid leaching from Si0 2 gel over time. Y-axis: Acid content in the gel coating normalized with respect to initial acid concentration, T = 298 K. The fuel cell polarization graph (Fig. 6-65) shows slightly enhanced performance due to the presence of the acidified Si gel. The anode catalyst was prepared at 298 K utilizing GF-S3, 60 Vo wt Triton X-100, i = 60 A m -2 and t = 260 min. In Table 6-17 parameters for the different proton conducting materials are summarized. The coating thickness relates to the approximate average coverage of a single fiber. In conclusion, the application of acidified gels or foams is not practical for DMFC anodes, due to the leaching effect. Leaching would probably be more pronounced if the gel or foam coated anode was exposed to liquid flow compared to a quiescent aqueous solution. Further, the liquid flow would probably wash out the gel or foam.  141  0.60 • without gel O with gel 0.55  0.50 -  2 0.45 -  0  •  8 W 0.40 '.  0  •  0.35 0  •  0.30 -  O  0.25 ^ 0^20^40^60^80  100  i [A m " 2 ]  Fig. 6-65: DMFC polarization curves obtained at 333 K; anode: 1 M CH 3 OH, 5 ml min -1 , —100 kPa(abs); cathode: Pt black (40 g rn -2 ) dry 02, 500 ml mid i STP, —200 kPa(abs). Effect of acidified Si gel.  Table 6-17: Approximate load, volume fraction and coating thickness for model proton conducting materials. Substrate: Protonic conductor:  GF-S6, 2000 um Nafion (5%w0  GF-S3, Nafion (P/o wt) 1 M H2SO4  350 um Kelzan^Si02 gel 1 M H2SO4^1 M H2SO4  Load [g m -2 ]  150  30  40  50  Volume fraction at 50 % felt compression  10%  < 10%  23%  32%  Coating thickness [um]  4  <2  11  17  142  6.4 Fuel Cell Experiments All graphite felt based anodes used for measurements outlined in Section 6.4 were prepared by electrodeposition on GF-S3 in the sandwich plating cell at 333 K, using 40 cYo wt Triton X-100 and a deposition current density of 60 A m -2 . The electrodeposition was carried out in two steps, as described in Section 6.1.8.2, and the deposition time for each step was 90 minutes. The loadings were 43 g m -2 and 52 g  M -2  for PtRu and  PtRuMo, respectively (Table 6-13). Methanol was fed in circulating mode unless stated otherwise. Supporting sulfuric acid electrolyte was used in all cases, and the effect of acid concentration in the anolyte was assessed, which is described in the next section. Furthermore, the effects of altering the cathode pressure and anolyte flow rate were studied. The performance depending on the degree of the catalyzed felt anode compression was assessed in conjunction with serpentine flow and in flow-by mode. Both flow field types were also applied for a comparison of the novel 3D anode design with a conventional catalyst coated membrane. Half-cell experiments indicated that the addition of Mo to the PtRu catalyst can improve the performance at increased temperature (Section 6.1.8). Therefore the effect of Mo presence in the anode was studied by carrying out fuel cell tests at 333 K and 353 K. Based on the aforementioned studies, which were conducted to find favorable conditions for variables, such as cathode pressure and anode thickness (i.e., compression), factorial experiments were conducted to assess the influence of altering the methanol concentration, anolyte flow rate and temperature on the fuel cell power output. The selected response variable for comparison was the peak power density. Furthermore, the fuel cell was operated at constant current densities over time using 3D graphite felt based anodes with PtRu and PtRuMo catalyst, to study the deactivation.  6.4.1 Effect of sulfuric acid concentration (PtRu, serpentine flow) Different anolyte acid concentrations were investigated. The most favorable anolyte compositions contained either 0.5 M or 1 M H 2 SO 4 (Fig. 6-66). With 0.25 M acid the protonic conductivity was insufficient.  143  0.6  •  0_.__^2_ 25 M H2SO4 _4  •  ♦  0.5 M H2SO4  ^  1 M H2SO4 2_ _4  ••  •  2 M H2SO4  ■  2_ _4 5 M H2SO4  • • a • • • ■^•  ■  a •  6  • •• ■^ ■  a  • •  •  •  0.2 -  a  •  2  •  •  •  0.0 0  1000  2000  3000^4000  i [A nf 2  5000  ]  1600 1400 1200 -  •  0.25 M H2SO4  ♦  0.5 M H2SO4  o  2_ SO4 4 1 M H2  •  2 M H2SO4  ■  2_ _4 5 _M_H2SO4  1000  E 0-  6 a •  800 -  6000  a  2  • • •  a^•  •  • •  ••  • • •• •  600 -  a  400 -  ■• ■  200 -  0  0  1000^2000^3000^4000  [A m' 2  ^  5000  ^  6000  ]  Fig. 6-66: DMFC polarization curves obtained at 333 K; anode: 1 M CH 3 OH, 5 ml min -1 , —100 kPa(abs); cathode: dry 0 2 , 500 ml min -1 STP, —200 kPa(abs). Effect of sulfuric acid concentration in anolyte.  At higher acid concentrations (2 M and 5 M) the power output was below the values obtained with either 0.5 M or 1 M acid, which could possibly be explained by the  144  increased diffusion rate of protons to the cathode, which causes fuel cross-over and possibly cathode flooding. Furthermore, the increase in anolyte viscosity at higher acid concentrations may decrease the methanol mass transport to the active sites. Also, specific adsorption of sulfate ions on the catalyst surface can compromise the adsorption of methanol. Table 6-18 shows the ionic conductivity at 333 K for the selected sulfuric acid concentrations and the effective conductivity based on 83 % electrode porosity (see also equation (6-5) and Table 5-3).  Table 6-18: Ionic conductivity of bulk H2SO4 and effective ionic conductivity within 3D anode matrix for different sulfuric acid concentrations (T = 333 K). [1-12SO4]  KO  K  [M]  [S M -1 ]  [S M-1]  0.25 0.5 1 2 5  12 29 55 96 120  9 22 42 73 91  6.4.2 Effects of cathode pressure and anolyte flow rate (PtRu, serpentine flow) The cathode was operated at two pressure settings at 333 K and also at near ambient pressure at 298 K. Oxygen was fed at 500 ml min -1 STP. The higher open circuit voltage and overall performance enhancement observed with 200 kPa(abs) is due to a combination of reduced fuel cross-over and higher 02 electroreduction rate (Fig. 6-67). The power density curve at 298 K is similar to that obtained for an air breathing DMFC at 295 K with the same MEA geometric area (5 cm 2 ) and similar anode catalyst loading (40-50 g m -2 ) [Lu, Wang, 2006]. The cathode catalyst loading was 10 g m -2 . The passive fuel cell was operated with 2 M methanol at a flow rate of 0.1 ml min -I . The peak power density of 200 W M -2 was obtained at 1100 A m -2 . The effect of flow rate variation was investigated at 333 K. A decrease in open circuit voltage, which can be an indication for increased cross-over [Qi, Kaufman, 2002], was observed with flow rates increasing from 2 ml min -1 to 10 ml min -1 (see Fig. 6-68). The values for E oe were 0.67 V, 0.66 V and 0.62 V for 2, 5 and 10 ml min -I , respectively. It is proposed that at a flow rate of 5 ml min -1 the mass transport of  145  methanol to the active sites was improved compared to a flow rate of 2 ml min -I , whereas the fuel cross-over cross-over effect was prevalent at 10 ml min 1 due to the increased anode pressure.  • 0.6  2  • ■ •  I  1 •  • •  ••  • •  0.4- •  t^• LIP^•  •  0.2 -^•  200 kPa(abs), 333 K 100 kPa(abs), 333 K 100 kPa(abs), 298 K  • ■ ^• ■  • 0.0  • •  0^1000^2000^3000^4000^5000 i [A m -2  • ■ ♦  1200 -  200 kPa(abs), 333 K 100 kPa(abs), 333 K 100 kPa(abs), 298 K  1000 -  • •  800 -  ]  • ■  •  •  • •  •• 400  •• •  200  •• •  V  0I 0  1000  2000  3000  4000  5000  i [A nil  Fig. 6-67: Variation of the cathode pressure; anode conditions: 1 M CH3OH-  0.5 M H 2 SO 4 , 2 ml min -I , —100 kPa(abs); cathode: dry 02, 500 ml min d STP  146  S  • 2 ml min d  0.6 0O 5 v• 0^ •  ml  min d  •  • 10 ml min d  Ci  g v  0  •  O  •  •  0  •  0.2  0  •  •  0.0 0  1000  2000  3000^4000  i [A m -2  1600 1400 -  •  2 ml min d  O  5 ml min d  •  10 ml min d  ,^. 5000^6000  ]  0 O  1200 -  O  cr. E 1000 —  •  •  0  •  •  • •  g •  800 O. 600 400 -  ;  200 0  4• •^ 0  1000  ^  2000  ^  3000^4000  ^  5000  ^  , 6000  i [A till  Fig. 6-68: Effects of flow rate variation at 333 K; anode: 1 M CH3OH-0.5 M H 2 SO 4 ,  —400 kPa(abs); cathode: dry 0 2 , 500 ml min -1 STP, —200 kPa(abs). 6.4.3 Anode compression  By selecting the number of gaskets, placed between the anode end plate and the membrane, the thickness of the anode and therefore the degree of compression can be  147  adjusted. One gasket is 127 [im thick in its uncompressed state. The compression was controlled by the number of gaskets (i.e., 1, 2 or 3) that are placed between the membrane and the end plate, enclosing the anode.  •  0• )  0.6 -  2  • 100 gin 0 200 p.m  •  oo  •  •  •  0.4 -  •  O  •  0.2 -  0.0  0  O  ^  1000  ^  2000^3000  ^  4000  ^  5000  i [A m -2 ]  1600 • 100 gm 0 200 ;./m  1400 -  •  •  •  1200 1000 -  •  O  •  800 600 -  •  400 200 0  n o  0 0^1000  2000  3000  4000  5000  i [A m -2 ]  Fig. 6-69: Effect of anode thickness on fuel cell performance using a serpentine flow field; anode: PtRu deposited on GF-S3, 1 M CH3OH-0.5 M H2SO4, 5 ml —100 kPa(abs); cathode: dry  02,  500 ml min -1 STP, —200 kPa(abs); T = 333 K.  148  • ♦ ^  0.6 -  !I  100 gm 200 gm 300 gm  •  ,-... 0.4 > 7  g  • •  0.2 -  •  0  0.0 0  1000  2000  3000  i [A m -2  4000  5000  6000  ]  1600  •  1400 -  0  I  1200 -  •  1000 -  !I 0  800  •  600 -  II  •  •  400 -  • 100 gm 200 gm ♦ ^ 300 gm  200 0 • 0  • • 1000  2000  3000  4000  5000  6000  i [A nfl  Fig. 6-70: Effect of anode thickness on fuel cell performance using a flow-by design; anode: PtRu deposited on GF-S3, 1 M CH3OH-0.5 M H2SO 4 , 5 ml min -I , —100 kPa(abs); cathode: dry 02, 500 ml min -I STP, —200 kPa(abs); T = 333 K.  The torque applied to each of the end plate connecting bolts was 10 Nm in each case (bolt diameter = 6 mm). Fig. 6-69 shows that in the case of the serpentine flow field one  149  gasket, imposing an anode thickness of 100  IIM,  provided better performance. However,  when the flow-by type end plate was utilized the highest power output was obtained with 2 gaskets (Fig. 6-70). The serpentine flow field provided a larger electrode thickness locally (the electrode can 'expand' into the —1mm deep channels), which may explain why a gasket thickness of —100 yielded better results. Using a gasket thickness of —100 Jim did not work for the flow-by end plate, since it resulted in a partial disintegration of the anode material. The dependence of the porosity and electronic conductivity of the felt (GF-S3) on the anode thickness are listed in Table 5-3.  6.4.4 Comparison of the novel 3D anode and a conventional catalyst coated membrane Fuel cell polarization curves were obtained for the GF-S3 supported PtRu catalyst, as well as for a commercially available PtRu unsupported catalyst coated membrane (CCM) of comparable load and composition (40 g m  2  ,  1:1 atomic ratio) (Lynntech Inc.).  Fig. 6-71 compares the performance of a conventional catalyst coated membrane with PtRu dispersed on GF-S3 at 333 K. Methanol was fed using the serpentine flow field or the flow-by mode. The in-house prepared electrode provided a significantly better performance compared to the catalyst coated membrane at current densities higher than 1000 A m -2 . On a per Pt mass basis the power densities were similar at current densities below 2000 A M -2 (Fig. 6-72). The open circuit voltage was 0.67 V for the commercial electrode and 0.7 V for the graphite felt supported PtRu anode, irrespective of the flow field type. The carbon fiber structure of the anode substrate acts as a diffusion barrier with respect to methanol, thus decreasing the cross-over. This effect was also exploited by Lam and coworkers. To build up a 3D structure, multiple gas diffusion layers, each with a PtRu loading of 10 g m -2 on one side, were stacked in parallel [Lam et al., 2007]. The higher open circuit cell voltage observed with the graphite felt anode is also in part due to a higher degree of catalyst dispersion resulting in a larger active area per geometric graphite support area. In the kinetic regime (e.g., up to —1000 A m -2 ) very similar performances were observed for the compared electrode types. At 3000 A m 2 and  150  333 K, however, the fuel cell power density was enhanced by —40 % with the graphite felt anode; e.g., from 870 W m -2 (i.e., peak power for the CCM) to 1200 W m -2 .  •  iv  2  •• v  i •  0.4-  •  CCM, PtRu(1:1) 40 g m -2, serpentine  o  CCM, PtRu(1:1) 40 g m 2 , flow-by  •  GF-S3, PtRu(1.4:1) 43 g m -2 , serpentine  ^  GF-S3, PtRu(1.4:1) 43 g m 2 , flow-by  V  i •G  • V Y^•  Y  T-3 W  Y  0.2 -  0.0  0  ^  1000  ^  2000  ^  3000  v V  •  0  •  V  G  ^  4000  ^  5000  i [A ITC 2]  1800 1600 1400 -  -2  •  CCM, PtRu(1:1) 40 g m  , serpentine  o  CCM, PtRu(1:1) 40 g m2 , flow-by  •  GF-S3, PtRu(1.4:1) 43 g m -2 , serpentine  ^  GF-S3, PtRu(1.4:1) 43 g m2 , flow-by  1200 -  v  V  v  •  v^•  •  •  • i •o o • 0• 0 Q • 600 -  vQ  400 -  Q 0•  200 -  0  •1 0  0 1000^2000^3000  4000  5000  i [A m 2 ]  Fig. 6-71: Fuel cell polarization experiments: Comparison of PtRu deposited on GF-S3 and commercial CCM design; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min -I , —100 kPa(abs); cathode: dry 02, 500 ml min -I STP, —200 kPa(abs) ; T = 333 K.  151  60  50 -  40 -  •  CCM, PtRu(1:1) 40 g m -2 , serpentine  O  CCM, PtRu(1:1) 40 g m -2 , flow-by  •  GF-S3, PtRu(1.4:1) 43 g m 2 , serpentine  •  GF-S3, PtRu(1.4:1) 43 g m 2 , flow-by  o o •ov • •  30  •6  0.  20  •  •  •  •  0  • •  O  10 -  0 0^1000^2000^3000  i [A m -2  ^  4000  ^  5000  ]  Fig. 6-72: Power density plots expressed in terms of mass activity; anode: 1 M CH3OH0.5 M H2SO4, 5 ml min -1 , —100 kPa(abs); cathode: dry 02, 500 ml min 1 STP, —200 kPa(abs); T = 333 K.  The highest power output observed with the extended reaction zone anode was about 1500-1550 W m -2 . Both the novel and the conventional anode type performed slightly better when methanol was fed in flow-by mode as opposed to using the serpentine flow field. For comparison, Table 3-11 shows power density data from the literature that was obtained by employing alternative electrode designs with PtRu catalysts. For supported (e.g., Vulcan XC- 72) PtRu catalysts of 20 to 50 g m  -2  load in GDE or  CCM configuration peak power outputs at 333 K between 250 W m 2 600-750 W m 2 ,  and 1,000 W m -2 were reported (see Table 6-19). Therefore, it can be stated that the extended reaction zone anode yielded a better performance compared to conventional anodes with comparable loading and similar anode operating conditions. This observation can be most likely attributed to the higher degree of utilization of the nanoparticle catalyst, as shown also by Lycke and Gyenge for direct ethanol fuel cells [Lycke, Gyenge, 2007].  152  ^ ^  Table 6-19: Peak power densities obtained with CCM or GDE type anodes at T =333 K.  Anode [CH 3 OH]^Flow rate [M]^[ml min -1 ] CCM 1 10 N/A CCM 1 CCM 1 5 2 2 GDE GDE 1 2 1 5 GDE 1 5 GDE  Oxidant 02 Air Air 02 Air Air Air  Cathode Flow rate [ml min -1 ] 200 N/A 720 N/A 350 200 100  Pressure [kPa(abs)] N/A 100 100 250 100 100 100  Peak power density [w m 2  Reference  280 1150 1000 250 600-750 800 850  [Tang et al., 2007]  ]  [DuPont*] [Etek **] [Coutanceau et al., 2004] [Baglio et al., 2006] [Jiang et al., 2006] [Gurau, Smotkin, 2002]  [*http://dupont.com/fuelcells/pdf/dfc502.pdf - 3:1 0 2/CH 3 OH stoichiometric ratio, Fuel cell temperature: T = 343 K; ** http://www.etek-inc.com/pdfs/MEASeries12D-W.pdf]  Moreover, although not proven experimentally here, it is proposed that the extended reaction zone lowers the methanol cross-over rate. Thereby, the negative effect of the mixed cathode potential is to some extent mitigated. Fig. 6-73 shows a comparison between two 3D PtRu electrodes, prepared by single  step and two-step electrodeposition, respectively. The serpentine flow field was used. The open circuit voltage was lower for the PtRu catalyst prepared by single step codeposition (0.66 V vs. 0.7 V). This indicates higher activity and better dispersion for the doubledeposited catalyst. The higher power density on a per area basis is due to the higher PtRu loading (43 g m 2 vs. 5.5 g m -2 ) and larger active surface area (712 m 2 111 2 reit vs. 650 m 2 111 2 reit) for the double-deposited catalyst (Fig. 6-73). The highest obtained Pt mass normalized power density was about three times greater for the 3D anode prepared by single step codeposition (Fig. 6-74).  0.6 -  •0  E  •  0  •  •  0.4 -  0  •  •  -21  ti.1  o ••  0  o  o  •  0  0  0.2 •  PtRu(1.4:1) 5.5 g m  ° PtRu(1.4:1) 43 g m 0.0 0  ^  1 0 00  ^  -2 , deposition time = 90 min  -2 , deposition time = 2 x 90 min  ^ ^ 2000^3000 4000 5000  i [A rif' 2  ]  1600 O  1400 -  0 0  1200 -  ^  O  O O  600 -^ 400 -  •  200 -  0•  •  o• •  •  PtRu(1.4:1) 5.5 g m -2 , deposition time = 90 min  ° PtRu(1:1) 43 g m 2 , deposition time = 2 x 90 min  • 0 0  ••••  1000  ^  2000^3000  i [A  ^  4000  ^  5000  m] -2  Fig. 6-73: Polarization plots for PtRu prepared at 333 K using 40 % wt Triton X-100, i = 60 A m"2 : Comparison of single step and a two-step codeposition; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min d (serpentine flow), —100 kPa(abs); cathode: dry 0 2, 500 ml min d STP, —200 kPa(abs); T = 333 K.  154  160  •  140 120 -  • •  100 80 -  PtRu(1.4:1) 5.5 g m 2 , deposition time = 90 min  0 PtRu(1.4:1) 43 g m -2 , deposition time = 2x 90 min  •  40 20 -  •  •  60 -  •  •• •  •  oe  • 0  O  o  O  0  0  0  0  0  -)^  0^1000^2000^3000  i [A m -2  ^  4000^5000  ]  Fig. 6-74: Power density plots expressed in terms of mass activity for PtRu prepared at 333 K using 40 %,„4 Triton X-100, i = 60 A  111 -2 :  Comparison of single step and a two-step  codeposition; anode: 1 M CH 3 OH-0.5 M H2SO4, 5 ml min -1 , —100 kPa(abs); cathode: dry 0 2 , 500 ml min -1 STP, —200 kPa(abs); T = 333 K.  6.4.5 Presence of Mo in the 3D anode Fig. 6-75 shows that the addition of Mo to the anode catalyst formulation was beneficial for the fuel cell power output at 353 K, a trend that is in agreement with the voltammetric and chronopotentiometric experiments (see Figs. 6-44 and 6-46). In these experiments the end plate with the serpentine flow field was employed. At 333 K and 353 K the open circuit voltages were 0.70 V and 0.74 V without Mo and 0.72 V and 0.77 V when using PtRuMo, respectively. At a current density of 5500 A 111-2 and 353 K the power density with the PtRuMo catalyst was 2200 W  fil-2  while the binary PtRu catalyst yielded 1925 W ril-2 . As discussed before, the improved performance obtained with Mo at 353 K can be explained based on its bifunctional properties, which are similar to Ru [Bolivar et al., 2003] and / or oxygen spillover, which may lower the activation barrier for CO ad s oxidation [Samjeske et al., 2002].  155  0.8 PtRu T = 333 K PtRu T = 353 K PtRuMo T = 333 K PtRuMo T = 353 K  • O ♦ ^  4 0.7  'i  0.6 -  v v  o  5v i o v v v o • o v v 5^o^v o^v • o^o^v II^ o • 5^o 5  E7, 0.5 4  u.lu 0.4 -  •  0.3 -  0.2  1000  0  2000  3000  4000  5000  6000  i [A m -2 ]  2500  2000 -  • O ♦ ^  PtRu T = 333 K ^ PtRu T = 353 K ^ PtRuMo T = 333 K ^ PtRuMo T = 353 K  v  v v^ ^ 0 0  v 0 V 0 ^  o  0  •  5  5  •  II  500 -  Ig ^0  e  ■^  I^  ■^  I^  I  0^1000^2000^3000^4000^5000  ^  6000  i [A m -2 ]  ^Fig. 6-75:^Fuel cell polarization tests comparing PtRu with. PtRuMo; anode:  1 M CH 3 OH-0.5 M H 2 SO4, 5 ml min -1 , —100 kPa(abs); cathode: dry  02,  500 ml min -I  STP, —200 kPa(abs).  156  6.4.6 Factorial fuel cell experiments  Factorial experimental design is a method to determine both individual and synergistic effects of various factors on target responses. The effects of varying the methanol concentration, anolyte flow rate and temperature on DMFC performance were studied by carrying out full (2 3 +1) factorial experiments. PtRu and PtRuMo catalyzed graphite felt anodes, prepared as described in Section 6.1.8.2, were tested in conjunction with serpentine flow and flow-by mode, respectively. Based on observations discussed in Section 6.4.3 one anode gasket was used for serpentine flow whereas two gaskets were  used for flow-by operation to fix the anode thickness and compression. The variables of the factorial at their high and low levels as well as the center point conditions are listed in Table 6-20. The obtained peak power density was selected as a response variable. For  each catalyst and flow design the first experiment was carried out at center point conditions. The other tests were conducted in a random sequence. Replicate runs were conducted after completion of the full factorial. Methanol was fed in single pass mode. Appendix G contains the sequence for each set of experiments including replicates. The  dry 02 flow rate was 500 ml min -1 STP and the cathode pressure was 200 kPa(abs) in all cases.  Table 6-20: Variables and their levels applied for factorial DMFC experiments.  Methanol^Anolyte^Fuel cell  concentration flow rate temperature Level^c^F*^T [M]^[ml min-1 ]^[K] + 2 10 353 0 1.25 6 333 0.5 2 313 The obtained peak power densities are summarized in Table 6-21 and presented in cube plot form in Figs. 6-76 — 6-79. In general, PtRu yielded better performance when the temperature was set at the low level compared to PtRuMo and vice versa. Operation with the flow-by type end plate improved the average power output compared to the serpentine flow field.  157  Table 6-21: Peak power density response for factorial tests using PtRu and PtRuMo with serpentine and flow-by type anode end plates, respectively.  c^F*^T^ P [W 111-2 ] [A]^[ml min -1 ]^[K]^PtRu^PtRuMo Serpentine Flow-by Serpentine Flow-by -1^-1^-1^580^600^480^500 1^-1^-1^870^880^690^720 -1^1^-1^630^660^540^570 1^1^-1^900^920^780^810 Replicate^1^1^-1^840^870^720^600 0^0^0^1500^1560^1500^1550 Replicate^0^0^0^1440^1500^1400^1450 Replicate^0^0^0^1380^1400^1350^1380  Standard deviation for center point experiments: ^± 3.4 %^± 4.4 %^± 4.5 %^± 4.8 % -1^-1^1^1750^1740^1980^2040  1^-1^1^2170^2240^2310^2380 Replicate^1^-1^1^2030^2100^2100^2240 -1^1^1^1680^1620^1860^1800  1^1^1^1960^1890^1980^2170 Average p [AV ni -2 ]^1338^1346^1368^1393 For each set of catalysts and end plate designs the highest peak power density was obtained with the flow rate at the low level and temperature as well as concentration at the high level. The peak power density under these conditions was observed at —7000 A M -2 (see Appendix G) for both catalysts and flow modes, which means 18 % of the supplied methanol and 2.4 % of the supplied oxygen were consumed (assuming a complete 6 electron transfer reaction) per one pass. It is proposed that the decrease of the peak power output with an increase of the flow rate at 353 K is in part due to increased cross-over. The methanol permeability of the membrane may increase with temperature and the higher flow rate causes an increase in anode pressure, both factors causing an increase in cross-over. An experimental investigation of methanol cross-over was recently carried out using fuel cell testing at current densities up to 3000 A m -2 [Casalegno et al., 2007]. The selected temperature range was the same as for the factorial study. A CCM with loadings of 40 g M -2 was used (PtRu(1:1) on the anode and Pt black on the cathode). One of the conclusions was that the methanol cross-over was mainly caused by diffusion, and therefore it is strongly affected by temperature.  158  Response: p [W m -2] PtRu (serpentine flow) 1680  1960  1750 353(+)  a)  900  Ca  a)  E 580 313(-) 0.5(-)Me0H concentration [M] 2(+)  Fig. 6-76: Cube plot of peak power density as a function of methanol concentration, flow rate and temperature: PtRu; serpentine flow-field. Response: p [W m -2] PtRu (flow-by) 1620  1890  1740 353(+)  2240 1560  a)  z Co a)  920  660  10(+)^  4\  E a)  \1/:4  600 313(-) 0.5(-) Me0H concentration [M] 2(+)  880 2(-) ^  4$'  Fig. 6-77: Cube plot of peak power density as a function of methanol concentration, flow rate and temperature: PtRu; flow-by mode.  159  Response: p [W m  PtRuMo (serpentine flow)  2  - ]  1860  2170  1980 353(+)  a)  780 10(+)  Ea)  .(ky  480 313( )  690 2(-)  0.5(-) Me0H concentration [M] 2(+)  0  Fig. 6-78: Cube plot of peak power density as a function of methanol concentration,  flow rate and temperature: PtRuMo; serpentine flow-field. Response: p^m PtRuMo (flow-by)  -  2  1  1800  2170  2040 + 353  810  z  a)  + 10^ .<(  E  \?:  t1/49  500 - 313  720 -2 - 0.5 Me0H concentration [m] + 2  0 \-\\°  Fig. 6-79: Cube plot of peak power density as a function of methanol concentration,  flow rate and temperature: PtRuMo; flow-by mode.  160  At open circuit conditions an increase in cross-over was observed with increasing temperature, anolyte flow rate and methanol concentration (see Table 6-22).  Table 6-22: Methanol flux across the membrane at open circuit [Casalegno et al., 2007].  T [K] 313 333 353  Methanol flow rate [x 10 -4 mol s -i ]  Methanol concentration [M]  (Methanol concentration = 1 M) 1.8^2.4 3.6 0.70^0.75 0.80 1.35^1.40 1.50 2.20^2.40 2.50  (Flow rate = 2.4 x 10 -4 mol s -I ) 1 1.5 2.8 2.10 4.40 1.40 3.60 5.90 2.30  (x 10 -3 mol I/1 -2 S -1 )  (x 10 -3 mol 111 -2 S-1)  Other researchers examined the methanol conversion and cell voltage dependence on methanol concentration and flow rate at a constant current density (i.e., 1000 A m -2 ) [Gurau, Smotkin, 2002]. Gas diffusion electrodes with 40 g m -2 unsupported Pt black and PtRu were employed on the cathode and anode, respectively. The anolyte flow rate and methanol concentration were varied between 0.15-5 ml min -I and 0.5-2 M, respectively. The selected temperatures were 313 K, 333 K and 353 K. Irrespective of temperature and concentration, the fuel conversion was increased by decreasing the flow rate. In single pass mode the highest methanol conversion (34.5 % for each temperature) was obtained with a feed concentration of 1 M. The performance decrease observed at concentrations above 1 M at 353 K was attributed to increased fuel cross-over due to methanol permeability of the membrane compared to 313 K, where the 2 M methanol concentration was more favorable [Gurau, Smotkin, 2002]. In the present work, at 313 K the PtRu catalyzed graphite felt anode performed better at 10 ml min -I compared to 2 ml mid i (Figs. 6-76 and 6-77). This trend is not in agreement with the observations made by Gurau and Smotkin. Perhaps improved lateral dispersion within the 3D electrode due to the higher flow rate increased the reactant mass transport to the active sites. Appendix G contains the related polarization curves in Figs. G-1 and G-3.  161  6.4.7 Deactivation behavior of the DMFC Methanol was fed in single pass mode for all deactivation tests. A constant current density of 4000 A m -2 was applied for 6 hours to monitor the PtRu(Mo) catalyst deactivation during fuel cell operation at 353 K (Fig. 6-80). The results show that the performance of the two catalyzed GF-S3 electrodes was fairly similar after 300 minutes, even though the ternary catalyst performed better at the beginning of the experiment. Pt poisoning by adsorbed CO as well leaching of Ru and especially Mo each may have contributed to the observed deactivation. Furthermore, three repeated experiments were performed with PtRuMo. After the first and second experiment the DMFC anode was rinsed with distilled water in situ and the cathode was purged with air to remove water and methanol. The results indicate that purging of the cathode contributed to a partial recovery of the cell performance (Fig. 6-81). For comparison, in the case of a DMFC operated at 383 K with air and 0.4 M CH 3 OH, a cell voltage drop after 16 h from 0.4 V to 0.32 V at 4000 A  M -2  was reported  [Knights et al., 2004]. At 2000 A M -2 the loss was less pronounced (i.e., 0.52 V at t = 0 h vs. 0.50 V at t = 16 h). When a constant load was applied, water accumulated at the cathode over time causing gas diffusion limitations. To mitigate this effect load cycling was employed, i.e., every 30 minutes the load was removed from the cell for 30 seconds. This strategy resulted in good performance stability, e.g., when 2000 A m  -2  were applied  the degradation rate was 13 i.tV h -1 over a period of 1590 h. After 1500 h the cell voltage was —0.45 V compared to 0.52 V at the beginning of the experiment [Knights et al., 2004]. Deactivation tests were carried out based on the peak power operating conditions from the factorial fuel cell tests that were discussed in Section 6.4.6. A constant current density of 7000 A m -2 was applied (Fig. 6-82). The deactivation rates for PtRu in conjunction with serpentine flow were fairly similar (-26 mV h -I ) for the experiments shown in Figs. 6-80 and 6-82.  162  0.50  • PtRu 0 PtRuMo  0.45  0 0 0  i = 4000 A ri T = 353 K  0^ 0 0.40  ^0 • •^• • • • •^ OO  •^  0 0  •• 0.30 -  O0 • 0  •  • 0  0.25 -  0.20  2  O  0^100^200^300  t [min] Fig. 6-80: Fuel cell deactivation behavior of PtRu and PtRuMo; anode: 1 M CH3OH0.5 M H 2 SO 4 , 5 ml min -1 , -100 kPa(abs), serpentine flow; cathode: dry 02, 500 ml min -I STP, -200 kPa(abs).  0.50  0.45 M  • 0.40 -  •  O6 • •  •  •  •  1 St run  O  2 nd run 3 rd run  •  00 • 0  ♦  •  i = 4000 A m -22 T = 353 K  0^ •^ 0^• • • •0 • 0 ♦ ♦^•  • 9^• •  9o  0.30 -  O  •  0•  • o•  0.25 -  0.20 0^100^200  ^  •  300  t [min]  Fig. 6-81: Repeated fuel cell deactivation tests with a PtRuMo anode; anode: 1 M CH3OH-0.5 M H2SO4, 5 ml min -I , -100 kPa(abs), serpentine flow; cathode: dry 02, 500 ml min -I STP, -200 kPa(abs).  163  0.34 ^ • PtRu serpentine v •^ 0 PtRu flow-by 0.32 •^•^v^ • PtRuMo serpentine • 0^0 •^v^ 0.30 - •^0 • 0^v^ v PtRuMo flow-by • • 0 • v 0^v^i = 7000 A rn 4 0.28 -^ T = 353 K 0 • • 0.26 -^ •^0 • v • 0 v • •^• 0 v 0.24 -^ • 0•0v •^• 0.22 -^ • • v 0 • 0 0.20 -^ • • v 0 • v 0.18 -^ • • 0.16 -^ • 0  ^  100^200^300 t [min]  Fig. 6-82: Deactivation tests at 353 K using 2 M CH 3 OH-0.5 M H2SO4 at a flow rate of 2 ml min -1 ; anode: —100 kPa(abs); cathode: dry 02, 500 ml min -1 STP, —200 kPa(abs).  164  7.0 Conclusions This thesis work was conducted to assess the utilization of catalyzed graphite felt as novel 3D anodes for liquid feed direct methanol fuel cells. Two variants of graphite felt were employed. Surfactant assisted electroeposition onto 350 Jim thick GF-S3 type felt produced electrodes with higher activity and more uniformly dispersed catalyst particles compared to the thicker (2000 - 6000 [an) GF-S6 type felt. Catalyst particle diameters ranging from 10 nm to 50 nm were observed, as well as agglomerates with diameters between —100 nm and —200 nm for the deposits formed on GF-S3. Factorial electrodeposition experiments were carried out. Deposition with 40 % t. Triton X-100 at a superficial current density of 60 A  M-2  for 90 minutes at 298 K yielded  the highest active specific catalyst area and consequently the best performance (e.g., superficial current density of —200 A  M -2  at 298 K and —420 A m -2 at 343 K  (PtRu 1:1, 9.2 g m 2 ) at -0.2 V vs. MSE as determined by cyclic voltammetry). Increasing the electrodeposition bath temperature to 333 K while maintaining the same deposition current density, time and surfactant concentration improved the performance of the 3D electrode with respect to methanol electro-oxidation compared to depositing at 298 K. Furthermore, carrying out two consecutive electrodeposition procedures, while using a fresh deposition bath both times, increased the methanol oxidation current density significantly (i = —380 A m -2 at 298 K and —780 A m -2 at 343 K for E = -0.2 V vs. MSE (PtRu 1.4:1, 43 g m 2 ); see also Appendix C for a discussion of replicability). This was due to the substantially increased catalyst load. DMFC operation at 4000 A  M -2  and 333 K with the 3D electrode using liquid  electrolyte (1 M CH 3 OH-0.5 M H 2 SO 4 ) yielded a power density of —1400 W M 2 compared to —800 W m 2 for a conventional CCM, while the PtRu catalyst load and composition were similar (43 g m-2 , 1.4:1 vs. 40 g m-2 , 1:1). The corresponding open circuit voltage was also increased from 0.67 V to 0.7 V, indicating a reduction in crossover due to the implementation of the 3D electrode. Chronopotentiometry in a half-cell setup and fuel cell tests showed that PtRuMo yielded superior performance compared to PtRu at 343 K and 353 K. For example, the potential required to maintain a superficial current density of 50 A  M2  at 343 K was 165  decreased from -0.43 V vs. MSE (PtRu(1.4:1), 43 g m -2 ) to -0.45 V vs. MSE (PtRuMo(1:1:0.3), 52 g m 2 ). Fuel cell experiments revealed that at a current density of 4000 A 111-2 and 353 K the power density obtained with the PtRuMo catalyst was 1800 W m-2 , while the binary catalyst yielded 1640 W m 2 . The platinum loading was identical for both catalysts (i.e., 31 g m -2 ). The fuel cell performance with a 3D anode depends on the degree of anode compression in conjunction with the type of employed flow field. For the serpentine flow field an electrode compressed to a thickness of 100 um (including the compressed carbon cloth backing layer) yielded a higher power output compared to a thickness of 200 um. The opposite trend was observed when tests were carried out in flow-by mode. Sulfuric acid supporting electrolyte was essential to exploit the extended reaction zone of the catalyzed graphite felt. In principle the entire catalyst surface was accessible for methanol adsorption and oxidation. An acid concentration of 0.5-1 M provided sufficient protonic conductivity, whereas increasing the acid concentration to 2 M and 5 M caused a decrease of the power density. This behavior could be explained by several phenomena, such as (i) increased cross-over of protons to the cathode, thereby enhancing the transport of methanol through the membrane, (ii) increased viscosity lowering the mass transport of methanol to the active sites and (iii) competing adsorption of sulfate ions on active sites. The factorial fuel cell tests revealed that operating with PtRuMo yielded a higher peak power output compared to PtRu with the temperature adjusted to its high level value and vice versa. Utilizing the flow-by design to distribute the anode flow improved the performance compared to application of a basic serpentine flow field in most cases. The combined effects of the electrocatalyst reactivity, anode design and operation were crucial for performance enhancement.  166  8.0 Recommendations Apart from PtRuMo other ternary and also quaternary catalyst formulations can be explored (e.g., PtRuIr(Os) [Gurau et al., 1998]). However, the stability and long term performance of such catalysts has to exceed that of binary catalysts.  The substrate surface and consequently the active area of the catalyst deposit can be enhanced significantly by growing carbon nanofibers on the graphite felt structure [Sun et al., 2004].  The effluent of the DMFC anode should be analyzed by on-line gas chromatography to assess the methanol conversion, the amount of CO2 formed and the formation of incomplete oxidation products (e.g., CH2O or HCOOH).  The catalyst deactivation can be studied by conducting half cell experiments (e.g., at a constant current over several hours) in a flow cell with a constant methanol supply. This approach may be useful to assess the impact of anode catalyst degradation (compared to cathode deactivation) on the fuel cell performance.  - The extent of methanol cross-over during operation with the 3D anode must be thoroughly compared against conventional CCM and GDE structures.  - The anode two-phase flow should be visualized by using transparent end plates and tracers. This may also provide insight into flow design improvement.  A portable DMFC has to function by utilizing air at atmospheric pressure for its oxidant supply. Therefore a cathode end plate allowing for 'passive' operation needs to be designed and tested.  167  The described electrodeposition methods should be modified and applied for carbon fiber paper substrates. These substrates may be practical for DMFC operation, since a solid ionomer can be used to provide protonic conductivity. The carbon paper matrix could be pretreated to facilitate nucleation and deposition.  168  References Agmon, N., The Grotthus Mechanism, Chemical Physics Letters 244 (5-6) (1995) 456 Alekseev, A., Evmenenko, G.A., Dembo, A.T., Osmotic Pressure and Structure of Aqueous Solutions of Triton X-100, Colloid Journal 59 (3) (1997) 267 Allen, R.G., Lim, C., Yang, L.X., Scott, K., Roy, S., Novel Anode Structure for the Direct Methanol Fuel Cell, Journal of Power Sources 143 (1-2) (2005) 142 Anderson, A.B., Grantscharova, E., Seong, S., Systematic Theoretical-Study of Alloys of Platinum for Enhanced Methanol Fuel-Cell Performance, Journal of the Electrochemical Society 143 (7) (1996) 2075 Argyropoulos, P., Scott, K., Taama, W.M., Gas Evolution and Power Performance in Direct Methanol Fuel Cells, Journal of Applied Electrochemistry 29 (6) (1999) 661 Argyropoulos, P., Scott, K., Taama, W.M., Carbon Dioxide Evolution Patterns in Direct Methanol Fuel Cells, Electrochimica Acta 44 (20) (1999) 3575 Arico, A.S., Poltarzewski, Z., Kim, H., Morana, A., Giordano, N., Antonucci, V., Investigation of a Carbon-Supported Quaternary Pt-Ru-Sn-W Catalyst for Direct Methanol Fuel Cells, Journal of Power Sources 55 (2) (1995) 159 Arico, A.S., Srinivasan, S., Antonucci, V., DMFCs: From Fundamental Aspects to Technology Development, Fuel Cells 1 (2) (2001) 133 Arps, J.H., Cavalca, A.C., Murthy, M., Fuel Cell Membrane Electrode Assemblies with Improved Power Outputs and Poison Resistance, US Patent Application 20010033960 (2001) Attard, G.S., Edgar, M., GOltner, C.G., Inorganic Nanostructures from Lyotropic Liquid Crystal Phases, Acta Materialia 46 (3) (1998) 751 Attard, G.S., Leclerc, S.A.A., Maniguet, S., Russell, A.E., Nandhakumar, I., Bartlett, P.N., Mesoporous Pt/Ru Alloy from the Hexagonal Lyotropic Liquid Crystalline Phase of a Nonionic Surfactant, Chemistry of Materials 13 (5) (2001) 1444 Babu, P.K., Kim, H.S., Oldfield, E., Wieckowski, A., Electronic Alterations Caused by Ruthenium in Pt-Ru Alloy Nanoparticles as Revealed by Electrochemical NMR, Journal of Physical Chemistry B 107 (31) (2003) 7595 Baglio, V., Di Blasi, A., Modica, E., Creti, P., Antonucci, V., Arico, A.S., Electrochemical Analysis of Direct Methanol Fuel Cells for Low Temperature Operation, International Journal of Electrochemical Science 1 (2006) 71  169  Bard, A.J., Faulkner, L.R., Electrochemical Methods, Fundamentals and Applications, John Wiley & Sons, New York (2001) 231 163 305 236 Bauer, A., Gyenge, E.L., Oloman, C.W., Electrodeposition of Pt-Ru Nanoparticles on Fibrous Carbon Substrates in the Presence of Nonionic Surfactant: Application for Methanol Oxidation , Electrochimica Acta 51 (2006) 5356 Bauer, A., Gyenge, E.L., Oloman, C.W., Direct Methanol Fuel Cell with Extended Reaction Zone Anode: PtRu and PtRuMo Supported on Graphite Felt, Journal of Power Sources 167 (2007) 281 Barragan, V.M., Ruiz-Bauz6., C., Villaluenga, J.P.G., Seoane, B., Transport of Methanol and Water through Nafion Membranes, Journal of Power Sources 130 (1-2) (2004) 22 Bewer, T., Beckmann, T., Dohle, H., Mergel, J., Stolten, D., Novel Method for Investigation of Two-Phase Flow in Liquid Feed Direct Methanol Fuel Cells Using an Aqueous 1120/02 Solution, Journal of Power Sources 125 (2) (2004) 1 Beyer, K., Phase Structures, Water Binding and Molecular Dynamics in Liquid Crystalline and Frozen States of the System Triton X-100-D20. A Deuteron and Carbon NMR Study, Journal of Colloid and Interface Science 86 (1) (1982) 73 Blomen, L.J.M., Mugerwa, M.N., Fuel Cell Systems, New York (1993) 280 Bock, C., McDougall, B., Novel Method for the Estimation of the Electroactive Pt Area, Journal of the Electrochemical Society 150 (8) (2003) E377 Bock, C., McDougall, B., LePage, Y., Dependence of CH3 OH Oxidation Activity for a Wide Range of PtRu Alloys, Journal of the Electrochemical Society 151 (8) (2004) Al269 Bolivar, H., Izquierdo, S., Tremont, R., Cabrera, C.R., Methanol oxidation at Pt/Mo0,-/MoSe2 thin film electrodes prepared with exfoliated MoSe 2 , Journal of Applied Electrochemistry 33 (12) (2003) 1191 BOnnemann, H., Brij oux, W., Brinkmann, R., Dinjus, E., Joussen, T., Korall, B., Formation of Colloidal Transition Metals in Organic Phases and Their Application in Catalysis, Angewandte Chemie International Edition 30 (1991) 1312 BOnnemann, H., Britz, P., Vogel, W., Structure and Chemical Composition of a Surfactant- Stabilized Pt3Sn Alloy Colloid, Langmuir 14 (23) (1998) 6654 BOnnemann, H., Nagabhushana, K.S., Advantageous Fuel Cell Catalysts from Colloidal Nanometals, Journal of New Materials for Electrochemical Systems 7 (2) (2004) 93  170  BOnnemann, H., Richards, R., Manufacture of Heterogeneous Mono- and Bimetallic Colloid Catalysts and Their Applications in Fine Chemical Synthesis and Fuel Cells, Synthetic Methods of Organometallic and Inorganic Chemistry 10 (2002) 209 Boutonnet, M., Kizling, J., Stenius, P., The Preparation of Monodisperse Colloidal Metal Particles from Microemulsions, Colloids and Surfaces 5 (1982) 209 Campanati, M., Fornasari, G., Vaccari, A., Fundamentals in the Preparation of Heterogeneous Catalysts, Catalysis Today 77 (2003) 299 Cao, D., Lu, G.Q., Wieckowski, A., Mechanisms of methanol decomposition on platinum: A combined experimental and ab initio approach, Journal of Physical Chemistry B 109 (23) (2005) 11622 Casalegno, A., Grassini, P., Marchesi, R., Experimental Analysis of Methanol Cross-Over in a Direct Methanol Fuel Cell, Applied Thermal Engineering 27 (2007) 748 Chai, G.S., Yoon, S.B., Yu, J.S., Choi, J.H., Sung, Y.E., Ordered Porous Carbons with Tunable Pore Sizes as Catalyst Supports in Direct Methanol Fuel Cell, Journal of Physical Chemistry B 108 (22) (2004) 7074 Chrzanowski, W., Wieckowski, A., Surface Structure Effects in Platinum/Ruthenium Methanol Oxidation Electrocatalysis, Langmuir 14 (1998) 1967 Chen, J., Matsuura, T., Hori, M., Novel Gas Diffusion Layer with Water Management Function for PEMFC, Journal of Power Sources 131 (1-2) (2004) 155 Cheng, T.T., Gyenge, E.L., Electrodeposition of Mesoscopic Pt-Ru on Reticulated Vitreous Carbon from Reverse Emulsions and Microemulsions: Application to Methanol Electro-Oxidation, Electrochimica Acta 51 (2006) 3904 Cheng, W.-H., Kung, H.H., Methanol Production and Use, Marcel Dekker, Inc., New York New York (1994) 2 56 Choi, K.H., Kim, H.S., Lee, T.H., Electrode Fabrication for Proton Exchange Membrane Fuel Cells by Pulse Electrodeposition, Journal of Power Sources, 75 (1998) 230 Choi, W.C., Kim, J.D., Woo, S.I., Modification of Proton Conducting Membrane for Reducing Methanol Crossover in a Direct Methanol Fuel Cell, Journal of Power Sources 96 (2001) 411 Choi, P., Jalani, N.H., Datta, R., Thermodynamics and Proton Transport in Naflon Proton Diffusion Mechanisms and Conductivity, Journal of the Electrochemical Society 152 (3) (2005) E123  171  Choo, R.T.C., Toguri, J.M., El-Sherik, A.M., Erb, U., Mass Transfer and Electrocrystallization Analyses of Nanocrystalline Nickel Production by Pulse Plating, Journal of Applied Electrochemistry 25 (4) (1995) 384 Coutanceau, C., Rakotondrainibe, A.F., Lima, A., Gamier, E., Pronier, S., Leger, J.-M., Lamy, C., Preparation of Pt—Ru Bimetallic Anodes by Galvanostatic Pulse Electrodeposition: Characterization and Application to the Direct Methanol Fuel Cell, Journal of Applied Electrochemistry 34 (1) (2004) 61 Davies, J.C., Hayden, B.E., Pegg, D.J., The Modification of Pt(110) by Ruthenium: CO Adsorption and Electro-Oxidation, Surface Science 467 (1-3) (2000) 118 Davies, J.C., Hayden, B.E., Pegg, D.J., Rendall, M.E., The Electro-Oxidation of Carbon Monoxide on Ruthenium Modified Pt(1 1 1), Surface Science 496 (1-2) (2002) 110 Detor, A.J., Schuh, C.A., Tailoring and Patterning the Grain Size of Nanocrystalline Alloys, Acta Materialia 55 (2007) 371 Dickinson, A.J., Carrette, L.P.L., Collins, J.A., Friedrich, K.A., Stimming, U., Performance of Methanol Oxidation Catalysts with Varying Pt:Ru Ratio as a Function of Temperature, Journal of Applied Electrochemistry 34 (10) (2004) 975 Dinh, H.N., Ren, X., Garzon, F.H., Zelenay, P., Gottesfeld, S., Electrocatalysis in Direct Methanol Fuel Cells: In-Situ Probing of PtRu Anode Catalyst Surfaces, Journal of Electroanalytical Chemistry 491 (1-2) (2000) 222 Djokic, S.S., Electrodeposition of Platinum from an Alkaline Solution, Plating and Surface Finishing 87(5) (2000) 145 Dohle, H., Divisek, J., Mergel, J., Oetjen, H.F., Zingler, C., Stolten, D., Recent Developments of the Measurement of the Methanol Permeation in a Direct Methanol Fuel Cell, Journal of Power Sources 105 (2) (2002) 274 Dubau, L., Coutanceau, C., Gamier, E., Leger, J.-M., Lamy, C., Electrooxidation of Methanol at Platinum—Ruthenium Catalysts Prepared from Colloidal Precursors: Atomic Composition and Temperature Effects, Journal of Applied Electrochemistry 33 (5) (2003) 419 Elliott, J.M., Attard, G.S., Bartlett, P.N., Coleman, N.R.B., Merckel, D.A.S., Nanostructured Platinum Films — Effects of Electrodeposition Conditions on Film Properties, Owen, J.R., Chemistry of Materials 11 (12) (1999) 3602 Elliot, J.M., Attard, G.S., Bartlett, P.N., Owen, J.R., Ryan, N., Singh, G., A Generic Approach to Electrodes with Novel Controllable Continuous Nanoarchitectures, Journal of New Materials for Electrochemical Systems 2 (4) (1999) 239  172  Elliott, J.M., Birkin, P.R., Bartlett, P.N., Attard, G.S., Platinum Microelectrodes with Unique High Surface Areas, Langmuir 15 (22) (1999) 7411 Entina, V.S., Petrii, O.A., Electrooxidation of Methanol on PtRu and Ru Electrodes at Various Temperatures, Elektrokhimiia 4 (1968) 678 Euler, K.J., Entwicklung der Elektrochemischen Brennstoffzellen (Development of Electrochemical Fuel Cells), Munich (1974) 24-27 Fischer, H., Electrocrystallization of Metals under Ideal and Real Conditions, Angewandte Chemie International Edition 8 (1969) 108 Friedrich, K.A., Geyzers, K.P., Dickinson, A.J., Stimming, U., Fundamental Aspects in Electrocatalysis: From the Reactivity of Single-Crystals to Fuel Cell Electrocatalysts, Journal of Electroanalytical Chemistry 524-525 (3) (2002) 261 Gasteiger, H.A., Markovic, N., Ross Jr., P.N., Cairns, E.J., Methanol Electrooxidation on Well-Characterized Platinum-Ruthenium Bulk Alloys, Journal of Physical Chemistry 97 (46) (1993) 12020 Gasteiger, H.A., Markovic, N., Ross Jr., P.N., Cairns, E.J., Temperature Dependent Methanol Electro Oxidation on Well Characterized PtRu Alloys, Journal of the Electrochemical Society 141 (7) (1994) 1795 Goetz, M., Wendt, H., Binary and Ternary Anode Catalyst Formulations Including the Elements W, Sn and Mo for PEMFCs Operated on Methanol or Reformate Gas, Electrochimica Acta 43 (24) (1998) 3637 Gojkovic, S.Lj., Vidakovic, T.R., Durovic, D.R., Kinetic Study of Methanol Oxidation on Carbon-Supported PtRu Electrocatalyst, Electrochimica Acta 48 (24) (2003) 3607 Gollas, B., Elliot, J.M., Bartlett, P.N., Electrodeposition and Properties of Nanostructured Platinum Films Studied by Quartz Crystal Impedance Measurements at 10 MHz, Electrochimica Acta 45 (22-23) (2000) 3711 Goncalves, M., Attard, G., Nanostructured Mesoporous Silica Films, Reviews on Advanced Materials Science 4 (2003) 147 Gonzalez-Garcia, C.M., Gonzales-Martin, M.L., Gonzales, J.F., Sabio, E., Ramiro, A., Ganan, J., Nonionic Surfactants Adsorption onto Activated Carbon. Influence of the Polar Chain Length, Powder Technology 148 (2004) 32 Green, C.L., Kucernak, A., Determination of the Platinum and Ruthenium Surface Areas in Platinum-Ruthenium Electrocatalysts by Underpotential Deposition of Copper. I. Unsupported Catalysts, Journal of Physical Chemistry B 106 (5) (2002) 1036  173  Green, C.L., Kucernak, A., Determination of the Platinum and Ruthenium Surface Areas in Platinum-Ruthenium Electrocatalysts by Underpotential Deposition of Copper. II. Effect of Surface Composition on Activity, Journal of Physical Chemistry B 106 (5) (2002) 11446 Grove, W.R., On Voltaic Series and the Combination of Gases by Platinum, Phil. Mag. XIV (1839) 127 Gurau, B., Smotkin, E.S., Methanol Crossover in Direct Methanol Fuel Cells: A Link between Power and Energy Density, Journal of Power Sources 112 (2) (2002) 339 Gurau, B., Viswanathan, R., Liu, R., Lafrenz, T.J., Ley, K.L., Smotkin, E.S., Reddington, E., Sapienza, A., Chan, B.C., Mallouk, T.E., Sarangapani, S. , Structural and Electrochemical Characterization of Binary, Ternary, and Quaternary Platinum Alloy Catalysts for Methanol Electro-Oxidation, Journal of Physical Chemistry 102 (49) (1998) 9997 Gyenge, E.L., Oloman, C.W., The Surfactant-Promoted Electroreduction of Oxygen to Hydrogen Peroxide. Reactor Engineering Aspects, Journal of the Electrochemical Society 152 (3) (2005) D42 Hirschenhofer, J.H., Stauffer, D.B., Engelmann, R.R., Klett, M.G., Fuel Cell Handbook, Morgantown, West Virginia (1998) 1-21 Hodgson, I., Oloman, C., Pressure Gradient, Liquid Hold-Up and Mass Transfer in a Graphite Fibre Bed with Cocurrent Upward Gas-Liquid Flow, Chemical Engineering Science 54 (1999) 5777 Hoster, H., Iwasita, T., Baumgaertner, H., Vielstich, W., Current Time Behaviour of Smooth and Porous PtRu Surfaces for Methanol Oxidation, Physical Chemistry Chemical Physics 3 (3) (2001) 337 Hyeon, T., Han, S., Sung, Y.-E., Park, K.-W., Kim, Y.-W., High-Performance Direct Methanol Fuel Cell Electrodes Using Solid-Phase-Synthesized Carbon Nanocoils, Angewandte Chemie International Edition 42 (36) (2003) 4352 Iwasita, T., Electrocatalysis of Methanol Oxidation, Electrochimica Acta 47 (22-23) (2002) 3663 Jayashree, R.S., Spendelow, J.S., Yeom, J., Rastogi, C., Shannon, M.A., Kenis, P.J.A., Electrooxidation of Formic Acid on Electrodeposited Pt and Pt/Pd Catalyst Structures, Electrochimica Acta 50 (24) (2005) 4674 Jiang, J., Kucernak, A., Electrooxidation of Small Organic Molecules on Mesoporous Precious Metal Catalysts I: CO and Methanol on Platinum, Journal of Electroanalytical Chemistry 533 (1-2) (2002) 153  174  Jiang, J., Kucernak, A., Electrooxidation of Small Organic Molecules on Mesoporous Precious Metal Catalysts II: CO and Methanol on Platinum-Ruthenium Alloy, Journal of Electroanalytical Chemistry 543 (2) (2003) 187 Jiang, R., Kunz, H.R., Fenton, J.M., Composite Silica/Nafion Membranes Prepared by Tetraethylorthosilicate Sol-Gel Reaction and Solution Casting for Direct Methanol Fuel Cells, Journal of Membrane Science 272 (2006) 116 Jiang, R., Kunz, H.R., Fenton, J.M., Multilayer Structure Membranes with Sulfonated Hydrocarbon Methanol Barrier for Direct Methanol Fuel Cells, Journal of the Electrochemical Society 153 (8) (2006) A1554 Jusys, Z., Kaiser, J., Behm, R.J., Composition and Activity of High Surface Area PtRu Catalysts Towards Adsorbed CO and Methanol Electrooxidation- - A DEMS study, Electrochimica Acta 47 (22-23) (2002) 3693 Jusys, Z., Schmidt, T.J., Dubau, L., Lasch, K., Jorissen, K., Garche, J., Behm, R.J., Activity of PtRuMeO x (Me = W, Mo or V) Catalysts towards Methanol Oxidation and their Characterization , Journal of Power Sources 105 (2002) 297 Kang, D.B., Lee, C.K., Electronic Properties of Ru/Pt(111) Alloy Surface: A Theoretical Study of H2 O Adsorption, Bulletin of the Korean Chemical Society 21 (2000) 87 Khazova, O.A., Mikhailova, A.A., Skundin, A.M., Tuseeva, E.K., Havranek, A., Wippermann, K., Kinetics of Methanol Oxidation on Supported and Unsupported Pt/Ru Catalysts Bonded to PEM, Fuel Cells 2 (2) (2003) 99 Kim, S.-H., Choi, H.-G., Nam, J.-D., Lee, Y., Cho, S.-M., Lee, E.-S., Le, J.-K., Chung, C.-H., Characteristics and Performance Evaluation of Pt-(Ru) Electrode Prepared by Periodic-Reverse Electrodeposition in Nafion-Contained Electrolyte, Abstract 996, 204th Meeting of the Electrochemical Society, Inc. (2003) Kim, C., Kim, Y.J., Kim, Y.A., Yanagisawa, T., Park, K.C., Endo, M., High Performance of Cup-Stacked Carbon Nanotubes as a Pt/Ru Catalyst Supported for Fuel Cell Applications, Journal of Applied Physics 96 (11) (2004) 5903 Kim, D., Scibioh, M.A., Kwak, S., Oh, I.-H., Ha, H.Y., Nano-Silica Layered Composite Membranes Prepared by PECVD for Direct Methanol Fuel Cells, Electrochemistry Communications 6 (2004) 1069 Kim, J.Y., Yang, Z.G., Chang, C.C., Valdez, T.I., Narayanan, S.R., Kumta, P.N., A SolGel-Based Approach to Synthesize High-Surface-Area Pt-Ru Catalysts as Anodes for DMFCs, Journal of the Electrochemical Society 150 (11) (2003) A1421 Knights, S.D., Colbow, K.M., St-Pierre, J., Wilkinson, D.P., Aging Mechanisms and Lifetime of PEFC and DMFC, Journal of Power Sources 127 (2004) 127  175  Koffi, R.C., Coutanceau, C., Gamier, E., Leger, J.-M., Lamy, C., Synthesis, Characterization and Electrocatalytic Behaviour of Non-Alloyed PtCr Methanol Tolerant Nanoelectrocatalysts for the Oxygen Reduction Reaction (ORR), Electrochimica Acta 50 (2005) 4117 Koper, M.T.M., Lukkien, J.J., Jansen, A.P.J., van Santen, R.A., Lattice Gas Model for CO Electrooxidation on Pt-Ru Bimetallic Surfaces, Journal of Physical Chemistry B 103 (26) (1999) 5522 Kua, J., Goddard, W.A., Oxidation of Methanol on 2nd and 3rd Row Group VIII Transition Metals (Pt, Ir, Os, Pd, Rh, and Ru): Application to Direct Methanol Fuel Cells, Journal of the American Chemical Society 121 (47) (1999) 10928 Lam, A., Wilkinson, D., Zhang, J., A Novel DMFC Electrode Assembly based on a Membraneless Architecture and Advanced 3D Anodes, 58 th Annual Meeting of the International Society of Electrochemistry, Banff, Canada, Spetmber 13 th 2007 Lee, K., Nam, J.-D., Optimum Ionic Conductivity and Diffusion Coefficient of IonExchange Membranes at High Methanol Feed Concentrations in a Direct Methanol Fuel Cell, Journal of Power Sources 157 (2006) 201 Lee, M.H., Oh, S.-G., Suh, K.-D., Kim, D.-G., Sohn, D., Preparation of Silver Nanoparticles in Hexagonal Phase Formed by Nonionic Triton X-100 Surfactant, Colloids and Surfaces A 210 (2002) 49 Lee, H.K., Park, J.H., Kim, D.Y., Lee, T.H., A Study on the Characteristics of the Diffusion Layer Thickness and Porosity of the PEMFC, Journal of Power Sources 131 (12) (2004) 200 Lee, S.A., Park, K.W., Kwon, B.K., Sung, Y.E., Method of Preparing Platinum Alloy Electrode Catalyst for Direct Methanol Fuel Cell Using Anhydrous Metal Chloride, US Patent Application 20010027160 Li, W., Sequential Pulse Deposition, US Patent Application 2002110991 Lide, D.R., CRC Handbook of Chemistry and Physics, 87 th edition, CRC Press, Boca Raton FL (2006); http://vvww.hbcpnetbase.com/ Lima, A., Hand, F., Leger, J.-M., Oxidation of Methanol on Pt, Pt—Ru, and Pt—Ru—Mo Electrocatalysts Dispersed in Polyaniline: An in situ Infrared Reflectance Spectroscopy Study, Russian Journal of Electrochemistry 40 (2004) 326 Lin, B., Conceptual Design and Modeling of a Fuel Cell Scooter for Urban Asia, Princeton University, School of Engineering and Applied Sciences, Dep. of Mechanical and Aerospace Engineering (1999)  176  Ling, J., Savadogo, O., Comparison of Methanol Crossover among four Types of Nafion Membranes, Journal of the Electrochemical Society 151 (10) (2004) A1604 Liu, H., Favier, F., Ng, K., Zach, M.P., Penner, R.M., Size-Selective Electrodeposition of Meso-Scale Metal Particles: A General Method Electrochimica Acta 47 (2001) 671 Liu, Z., Ling, X.Y., Guo, B., Hong, L., Lee, J.Y., Pt and PtRu Nanoparticles Deposited on Single-Wall Carbon Nanotubes for Methanol Electro-Oxidation, Journal of Power Sources 167 (2007) 272 Liu, P., Norskov, J.K., Kinetics of the Anode Processes in PEM Fuel Cells - The Promoting Effect of Ru in PtRu Anodes, Fuel Cells 1 (3-4) (2001) 192 Liu, F., Yan, M., Zhou, W., Jiang, Z., In Situ Transmission Difference FTIR Spectroscopic Investigation on Anodic Oxidation of Methanol in Aqueous Solution, Electrochemistry Communications 5 (2003) 276 Lopes, I.M., Beden, B., Hahn, F., Leger, J.-M., Lamy, C., On the Nature of the Adsorbates Resulting from the Chemisorption of Methanol at a Platinum Electrode in Acid Medium. An EMIRS Study, Journal of Electroanalytical Chemistry and Interfacial Electrochemistry 313 (1-2) (1991) 323 Lu, G.Q, Wang, C.Y., Development of High Performance Micro DMFCs and a DMFC Stack, Journal of Fuel Cell Science and Technology 3 (2006) 131 Lycke, D., Gyenge, E.L., Electrochemically Assisted Organosol Method for Pt-Sn Nanoparticle Synthesis and In-Situ Deposition on Graphite Felt Support: Extended Reaction Zone Anodes for Direct Ethanol Fuel Cells, Electrochimica Acta, 52 (2007) 4287 Lu, W.P.-T., Ammon, R.L., Carbon Cloth Supported Electrode, US Patent 4349428 (1982) Lu, C.H., Wang, H.C., Microemulsion-Mediated Synthesis and Electrochemical Characterization of Nanosized LiNio 25C0 0 .5Mno 2502 Powders, Journal of the Electrochemical Society 152 (6) (2005) C341 McGovern, M.S., Garnett, E.C., Rice, C., Masel, R.I., Wieckowski, A., Effects of Nafion as a Binding Agent for Unsupported Nanoparticle Catalysts, Journal of Power Sources 115 (2003) 35 McNicol, B.D., Rand, D.A.J., Williams, K.R., Direct Methanol—Air Fuel Cells for Road Transportation, Journal of Power Sources 83 (1999) 15  177  Miyake, N., Wainright, J.S., Savinell, R.F., Evaluation of a Sol-Gel Derived Nafion/Silica Hybrid Membrane for Polymer Electrolyte Membrane Fuel Cell Applications II Methanol Uptake and Methanol Permeability, Journal of the Electrochemical Society 148 (8) (2001) A898 Montgomery, D.C., Design and Analysis of Experiments, John Wiley & Sons, New York (1997) 228 Moraru, V.N., Ocharenko, F.D., Kobylinskaya, L.I., Karmazina, T.V., Adsorption of OxyEthylated Nonionic Surfactants and its Influence on the Stability of the Aqueous Dispersion of Graphite, Colloid Journal 46 (6) (1984) 1148 Murphy, T.D., Design and Analysis of Industrial Experiments, Chemical Engineering (June 6) (1977) 168 Narayanan, S.R., Jeffries-Nakamura, B., Chun, W., Ruiz, R.P., Valdez, Sputter-Deposited Fuel Cell Membranes and Electrodes, T., US Patent 6,171,721 (2001) Neto, A.O., Franco, E.G., Arico, E., Linardi, M., Gonzales, E.R., Electro-Oxidation of Methanol and Ethanol on Pt-Ru/C and Pt-Ru-Mo/C Electrocatalysts Prepared by Bônnemann's Method, Journal of the European Ceramic Society 23 (2003) 2987 Neergat, M., Leveratto, D., Stimming, U., Catalysts for Direct Methanol Fuel Cells, Fuel Cells 2 (1) (2002) 25 Nordlund, J., Roessler, A., Lindbergh, G., The Influence of Electrode Morphology on the Performance of a DMFC Anode, Journal of Applied Electrochemistry 32 (3) (2002) 259 Olah, G.A., Goeppert A., Prakash, G.K.S., Beyond Oil and Gas: The Methanol Economy, Wiley-VCH, Weinheim, Germany (2006), 191 245 201 200 Olah, G.A., Prakash, G.K.S., Recycling of Carbon Dioxide into Methyl Alcohol and Related Oxygenates for Hydrocarbons, US Patent 5,928,806 (1999) Oloman, C., Lee, B., Leyten, W., Electrosynthesis of Sodium Dithionite in a Trickle-Bed Reactor, Canadian Journal of Chemical Engineering 68 (1990) 1004 Oloman, C., Matte, M., Lum, C., Electronic Conductivity of Graphite Fiber Fixed-Bed Electrodes, Journal of the Electrochemical Society 138 (8) (1991) 2330 Patrick, H.N., Wan, G.G., Manne, S., Aksay, I.A., Self-Assembly Structures of Nonionic Surfactants at Graphite—Solution Interfaces, Langmuir 13 (1997) 4349 Patrick, H.N., Wan, G.G., Self-Assembly Structures of Nonionic Surfactants at Graphite— Solution Interfaces. 2. Effect of Polydispersity and Alkyl Chain Branching, Colloids and Surfaces A 162 (2000) 149  178  Prentice, G., Electrochemical Engineering Principles, Prentice Hall, Eaglewood Cliffs, New Jersey (1991) Qi, Z., Kaufman, A., Open Circuit Voltage and Methanol Crossover in DMFCs, Journal of Power Sources 110 (2002) 177 Reetz, M.T.,^Helbig,^W.,^Size-Selective Synthesis^of Nanostructured Transition Metal Clusters Clusters, Journal of the American Chemical Society 116 (1994) 7401 Ren, X., Springer, T.E., Zawodzinski, T.A., Gottesfeld, S., Methanol Transport Through Nafion Membranes. Electro-osmotic Drag Effects on Potential Step Measurements, Journal of the Electrochemical Society 147 (2) (2000) 466 Roth, C., Martz, N., Halm, F., Leger, J.-M., Lamy, C., Fuess, H., Characterization of Differently Synthesized Pt-Ru Fuel Cell Catalysts by Cyclic Voltammetry, FTIR Spectroscopy, and in Single Cells, Journal of the Electrochemical Society 149 (11) (2002) E433 Saber, K., Koch, C.C., Fedkiw, P.S., Pulse Current Electrodeposition of Nanocrystalline Zinc, Materials Science and Engineering A 341 (1-2) (2003) 174 Samjeske, G., Wang, H., LOffler, T., Baltruschat, H., CO and Methanol Oxidation at PtElectrodes Modified by Mo, Electrochimica Acta 47 (22-23) (2002) 3681 Sauk, J., Byun, J., Kim, H., Composite Nafion/Polyphenylene Oxide (PPO) Membranes with Phosphomolybdic Acid (PMA) for Direct Methanol Fuel Cells, Journal of Power Sources 143 (2005) 136 Saxena, O.C., New Titrimetric Methods for Palladium and Platinum, Talanta 13 (1966) 862 Scott, K., Agyropopoulos, P., Yiannopoulos. P., Taama, W.M., Electrochemical and Gas Evolution Characteristics of Direct Methanol Fuel Cells with Stainless Steel Mesh Flow Beds, Journal of Applied Electrochemistry 31 (8) (2001) 823 Schmidt, T.J., Gasteiger, H.A., Behm, R.J., Methanol Electrooxidation on a Colloidal PtRu-Alloy Fuel-Cell Catalyst, Electrochemistry Communications 1 (1999) 1 Sigrist,^L.,^Dossenbach,^O.,^Ibl,^N.,^On^the^Conductivity and Void Fraction of Gas Dispersions in Electrolyte Solutions, Journal of Applied Electrochemistry 10 (2) (1980) 223  179  Son, H.D., Cho, M.S., Nam, J.-D., Cho, S.M., Chung, C.H., Choi, H.-G., Lee, Y., Depression of Methanol-Crossover Using Multilayer Proton Conducting Membranes Prepared by Layer-by-Layer Deposition onto a Porous Polyethylene Film, Journal of Power Sources 163 (1) (2006) 66 Song, S.Q., Liang, Z.X., Zhou, W.J., Sun, G.Q., Xin, Q., Stergiopoulos, V., Tsiakaras, P., Direct Methanol Fuel Cells: The Effect of Electrode Fabrication Procedure on MEAs Structural Properties and Cell Performance, Journal of Power Sources 145 (2) (2005) 495 Steele, B.C.H., Material Science and Engineering: The Enabling Technology for the Commercialisation of Fuel Cell Systems, Journal of Materials Science 36 (2001) 1053 Stratton, A., Hemming, D.F., Teper, M., Methanol Production from Natural Gas or Coal, International Energy Agency Report, p , IEA Coal Research, London, England (1982) Sun, X., Li, R., Stansfield, B., Dodelet, J.P., Desilets, S., 3D Carbon Nanotube Network Based on a Hierarchical Structure Grown on Carbon Paper Backing, Chemical Physics Letters 394 (4-6) (2004) 266 Tang, H., Wang. S., Pan, M., Jiang, S.P., Ruan, Y., Performance of Direct Methanol Fuel Cells Prepared by Hot-Pressed MEA and Catalyst-Coated Membrane (CCM), Electrochimica Acta 52 (11) (2007) 3714 Terry, R., Baker, K., Rodriguez, N.M., Graphite Nanofiber Catalyst Systems for Use in Fuel Cell Electrodes, US Patent 6,485,858 (2002) Ticianelli, E., Beery, J.G., Paffett M.T., Gottesfeld, S., An Electrochemical, Ellipsometric, and Surface Science Investigation of the PtRu Bulk Alloy Surface, Journal of Electroanalytical Chemistry 258 (1) (1989) 61 Tong, Y.Y., Kim, H.S., Babu, P.K., Waszczuk, P., Wieckowski, A., Oldfield, E., An NMR Investigation of CO Tolerance in a Pt/Ru Fuel Cell Catalyst, Journal of the American Chemical Society 124 (2002) 468 Tripkovic, A.V., Popovic, K.D., Oxidation of Methanol on Platinum Single-Crystal Stepped Electrodes from (110) Zone in Acid-Solution, Electrochimica Acta 41 (15) (1996) 2385 Vielstich, W., Lamm, A., Gasteiger, H.A., Handbook of Fuel Cells — Fundamentals, Technology and Applications, Volume 2: Electrocatalysis, John Wiley & Sons, Chichester, UK (2003), chapter 14 p 153-162 and chapter 21 p 306-307  180  Vigier, F., Gloaguen, F., Leger, J.-M., Lamy, C., Electrochemical and Spontaneous Deposition of Ruthenium at Platinum Electrodes for Methanol Oxidation: An Electrochemical Quartz Crystal Microbalance Study, Electrochimica Acta, 46 (28) (2001) 4331 Waszczuk, P., Lu, G.-Q., Wieckowski, A., Lu, C., Rice, C., Masel, R.I., UHV and Electrochemical Studies of CO and Methanol Adsorbed at Platinum/Ruthenium Surfaces, and Reference to Fuel Cell Catalysis, Electrochimica Acta 47 (22-23) (2002) 3637 Waszczuk, P., Wieckowski, A., Zelenay, P., Gottesfeld, S., Coutanceau, C., Leger, J.-M., Lamy, C., Radioactive Labeling Method for Studies of Adsorption on Fuel Cell Nanoparticle Electrodes, Journal of Electroanalytical Chemistry 511 (1-2) (2001) 55 Watanabe, M., Motoo, S., Electrocatalysis by Ad-Atoms Part H Enhancement of the Oxidation of Methanol on Platinum by Ruthenium Ad-Atoms, Journal of Electroanalytical Chemistry 60 (3) (1975) 267 Wei, Z.D., Chan, S.H., Electrochemical Deposition of PtRu on an Uncatalyzed Carbon Electrode for Methanol Electrooxidation, Journal of Electroanalytical Chemistry 569 (1) (2004) 23 Wilkinson, D.P., Johnson, M.C., Colbow, K.M., Campbell, S., Method and Apparatus for Reducing Reactant Crossover in an Electrochemical Fuel Cell, US Patent 5,672,439 (1997) Wilson, M.S., Gottesfeld, S., Thin-Film Catalyst Layers for Polymer Electrolyte Fuel Cell Electrodes, Journal of Applied Electrochemistry 22 (1) (1992) 1 Xie,^F.,^Tian,^Z.,^Meng,^H.,^Shen,^P.K.,^Increasing^the Three-Phase Boundary by a Novel Three-Dimensional Electrode, Journal of Power Sources 141 (2) (2005) 211 Xu, W., Lu, T., Liu, C., Xing, W., Low Methanol Permeable Composite Nafion/Silica/PWA Membranes for low Temperature Direct Methanol Fuel Cells, Electrochimica Acta 50 (2005) 3280 Yajima, T., Uchida, H., Watanabe, M., In-Situ ATR-FTIR Spectroscopic Study of ElectroOxidation of Methanol and Adsorbed CO at Pt-Ru Alloy, Journal of Physical Chemistry B 108 (2004) 2654 Zhang, X., Chan, K.Y., Microemulsion Synthesis and Electrocatalytic Properties of Platinum-Ruthenium Nanoparticles, Chemistry of Materials 15 (2) (2003) 451 Zhang, X., Chan, K.Y., Microemulsion Synthesis and Electrocatalytic Properties of Platinum-Cobalt Nanoparticles, Journal of Materials Chemistry 12 (4) (2002) 1203  181  Zhang, X., Zhang, F., Chan, K.-Y., Synthesis of Pt—Ru—Mo Ternary Metal Nanoparticles by Microemulsions, Their Characterization and Electrocatalytic Properties, Journal of Materials Science 39 (18) (2004) 5845  Referenced websites http://wwwl.eere .energy.gov/hydrogenandfuelcells/fuelcells/images/fc_diagrampem.gif http://www. iags. org/n052404t3.htm W. Zittel, R. Wurster, HyWeb: Knowledge - Hydrogen in the Energy Sector (1996): http ://www. hydro gen. org/Knowl edge/w-i-energiew-eng2 .html http ://www. efoy . de/index .php?option=com_content&task=view&id=13 &Itemid=56 http://www. cpkelco.com/xanthan/industrial/molecular_structure.html http ://dupont. com/fuel cell s/pdf/dfc502 .pdf http://vvww. etek-inc .com/pdfs/MEASeries12D-W.pdf  182  APPENDIX A: ALTERING THE DEPOSITION CELL SETUP Benefits of utilizing two counter electrodes for deposition To gauge the effects of using two counter electrodes, three different deposition bath compositions were prepared as listed in Table A-1. GF-S6 substrates with a thickness of 4000 pm and a 4 cm 2 geometric area were used in conjunction with the 100 ml beaker cell. Both sides of the felt were exposed to the solution and a more uniform current distribution was likely established in the 2 counter electrode mode. As a result a more uniform distribution of catalyst across the felt thickness was obtained (i.e., less agglomeration in the outer layers and more deposit in the middle layers) (Figs. A-1 and A-2). Cyclic voltammograms consistently showed the significantly enhanced electrode activity due to the modified deposition arrangement (Figs. A-3 - A-5). Table A-1: Summary of electrodeposition procedures carried out to compare the use of one and two counter electrodes, respectively. Deposition bath^i^t composition^[A m -2 ]^[min]  T [K]  PtRu load [g 111-2]  Atomic Pt:Ru ratio of deposit  50 Vo wt Triton X-100^50 10 2 M H2PtC16 10 -2 M RuC1 3  30  298  11.8 8.7*  4:1 3:1*  50 Vo wt Brij 56^50 10 2 M H2PtC1 6  30  298  13.5  4:1  9.2*  3:1*  37  5:1  22*  4:1*  10 -2 M RuCl3 10 -3 M TEAH^100 0.073 M sulfamic acid  10  343  1.2*10 -3 M H 2 PtC1 6 1.2*10 -3 M Ru(N0)(NO3)x(OH)y (x+y ;----, 3) * load and Pt:Ru ratio obtained with 1 counter electrode  183  S4700 5.0kV 4.9mm x130k SE(U,O)  ^  400nm  Fig. A-1: SEM micrograph of PtRu catalyst in the center region of the felt obtained with 50 °A m Triton X-100 and two counter electrodes.  S4700 10 OkV 4. 9mm x1 80k SE(U.())  ^  3CICI n m  Fig. A-2: SEM micrograph of PtRu catalyst in the center region of the felt obtained with 50 %, Triton X-100 and one counter electrode.  184  -  80  60 -  40  E 20 -0.3^-0.2 E [V vs. MSE] • 1 counter electrode •••• 2 counter electrodes Catalyst prepared with 50 % Triton X-100  •  -20  -0.6^-0.5^-0.4^-0.3^-0.2  -0.1^0 0  E [V vs. MSE]  Fig. A-3: Electro-oxidation of methanol on PtRu electrodeposited in the presence of 50 cYo wt Triton X-100 on GF-S3: Comparison between electrodeposition with one and two -1 counter electrodes. 0.5 M CH3OH-0.1 M H 2 SO 4 , 5 mV s , 298 K.  50  40 -  30 -  20  -  E != 10 0 ■N• 1 counter electrode •••• 2 counter electrodes  -10  Catalyst prepared with 50 %,..„ Brij 56  -20 -0.6  -0.5  -0.4  -0.3  -0.2^-0.1  00  E [V vs. MSE]  Fig. A-4: Electro-oxidation of methanol on PtRu electrodeposited on GF-S3 in the presence of 50 %,„ 4 Brij 56: Comparison between electrodeposition with one and two counter electrodes. 0.5 M CH3OH-0.1 M H2SO4, 5 mV s , 298 K.  185  200  150 -  - 3 cri < -2  100 rw,  .••• g^ . •• 0 ••  0 0  .-•  'E  ..  ..^  g . .• •  ^". ^..  IC, 50 -  ..".  -0.5^-0.4^-0.3^-0.2..-0!? ••• E. Q/.yg..MeEl •■• 1 counter electrode • • •• 2 counter electrodes  .••*  Catalyst prepared with 10'3 M TEAH at 343 K  -50 -." -0.6^-0.5  -0.4  -0.3^-0.2  -0.1^0 0  E [V vs. MSE]  Fig. A-5: Electro-oxidation of methanol on PtRu electrodeposited in the presence of TEAH (10 -3 M) on GF-S3: Comparison between electrodeposition with one and two counter electrodes. 0.5 M CH3OH-0.1 M H2SO 4 , 5 mV s , 298 K.  Fig. A-6: Micrograph showing substantial dendrite growth on graphite felt coated with PtRu utilizing TEAH.  186  Using Triton X-100 yielded slightly better performance compared to Brij 56. Depositing at a high current density and increased temperature with TEAH (tetraethylammonium hydroxide, a cationic surfactant) resulted in relatively high methanol oxidation current density (with respect to the electrode geometric area) (Fig. A-5). However, the catalyst load was —2.5-3 times higher compared to the deposits  obtained with nonionic surfactants, in part due to the higher deposition current density. Another negative aspect is the formation of dendrites, which are likely to break off from the fiber surface due to compression or when exposed to gas / liquid flow in a fuel cell (Fig. A-6).  The use of a second counter electrode increased the catalyst load in all cases and in particular when the nonionic surfactants were used. The overall bulk atomic Pt:Ru ratio did not change substantially. However, the ruthenium content in the middle section was increased and the atomic ratio distribution across the thickness of the felt became more uniform. It is unlikely that the increased load alone was causing the improved performance. The higher degree of catalyst dispersion through the felt thickness was probably the more important factor.  187  APPENDIX B: PULSED CURRENT DEPOSITION  Application of pulsed current codeposition in acidic media  Lamy and coworkers produced an alternative anode design by dispersing carbon particles onto a layer of carbon cloth and depositing PtRu in 0.5 M sulfuric acid by applying pulsed current [Lamy et al., 2004]. The method was adapted using a 2000 pm thick GF-S6 substrate and applying the same on- and off-times (0.1 s and 0.3 s) and current density (200 A m -2 ). The electrodeposition time (i.e., total on-time) was 15 minutes.  - 1 M H 2 SO 4 .... 1 M H 2 SO 4 40 %,,,, Triton-X 100 -0.6  -0.2  0.0  02  E [V vs. MSE]  Fig. B-1: Electro-oxidation of methanol on PtRu electrodeposited on GF-S6 by pulsed current deposition in 0.5 M CH 3 OH. Effect of surfactant presence in plating bath. -1 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s , 298 K. The obtained PtRu atomic ratio was approximately 4:1 at a noble metal load of 25 g m -2 . The addition of the surfactant may have provided enhanced catalyst dispersion resulting in higher catalytic activity (Fig. B-1).  188  Application of alternating anodic and cathodic pulses: Reverse pulse deposition  Deposition utilizing recurrent cathodic and anodic pulses was reported to be a suitable method to control the particle size [Kim et al., 2003]. The purpose of applying an anodic pulse following each cathodic deposition pulse is to desorb cations from the substrate surface and thus minimize the extent of particle growth. GF-S3 was used in either the micellar phase (60%, t Triton X-100) or in an aqueous electrolyte without surfactant. The cathodic pulse times that are reported in the literature (e.g. 20 ms [Detor, Schuh, 2007]) were found to be too short to obtain a suitable catalyst load on the graphite fiber substrate. The absolute values for the current densities applied during the pulsed deposition were 60 A m-2 (cathodic, 180 s) and 20 A m -2 (anodic, 30 s). The accumulated deposition time was identical with the deposition time for galvanostatic mode (260 min), which was used for comparison. The resulting catalyst loads and compositions are listed in Table B-1. In each case galvanostatic deposition proved to be more effective, as indicated by the better performance observed by CV and CP (Figs. B-2—B-5).  Table B-1: Electrodeposition parameters for comparison of galvanostatic and reverse pulsed electrodeposition, T = 298 K. Electrodeposition mode  Atomic Pt:Ru ratio of deposit  PtRu load [g m 2 ]  without surfactant  reverse pulse galvanostatic  10:1 7:1  40 73  with surfactant  reverse pulse galvanostatic  4:1 1.5:1  2.6 6.3  189  ^ ^  600 300 -^  500 400  s  t  cr  • 300 <^.1.s• 200^.;  200  C■1  E  'Is  100 ••• •• 0• •• • :•• -0.6^-0.4^-0.2^0.0^0.2^0.40.100 -^ E [V vs. MSE] •••■•  ••  0 0••  •s  r•:'  with surfactant • tcath = 180s t ano d = 30 s ••• • galvanostatic  11#•.'  0- •  -0.6^-0.4  -0.2  0.0  0.2  0.4  E [V vs. MSE]  Fig. B-2: Electro-oxidation of methanol on PtRu electrodeposited in pulsed or galvanostatic mode. 0.5 M CH 3 OH-0.1 M H 2 SO 4 , 5 mV s 1 , 298 K.  5 8^ rn  OS  4a.  -0.4^-0.2^0.0^0.2 E [V vs. MSE]  04•  ••••  with surfactant  0  •  ?  -0.6  418•  -0.4  tcath = 180s t ano d = 30 s • •••• galvanostatic plating  • -0.2  0.0  0.2  0.4  E [V vs. MSE]  Fig. B-3: Electro-oxidation of methanol on PtRu electrodeposited in pulsed or galvanostatic mode. Mass activity. 0.5 M CH 3 OH-0.1 M H2SO4, 5 mV s 1 , 298 K.  190  -0.16 -0.18 • ...  -0.20 -0.22 -  V) -0.24 -  t .^90 . ...••■••••••••••••••••••••••••• 00•10.1••••••••••••••••••••••••• '0  0  ui -0.26 Lu  -0.28 -0.30 -  •  • • •  tcath = 180s t ano d = 30 s T = 298 K  -0.32 -  •• •• galvanostatic^T = 298 K ^ tcath = 180s t ano d = 30 s T = 343 K  -0.34 -  - galvanostatic^T = 343 K -0.36  0  ^ ^ 50  100  ^  150  ^  200  t [s] Fig. B-4: Electro-oxidation of methanol on PtRu electrodeposited at 298 K in pulsed or  galvanostatic mode with 60 %,, t Triton X-100. i = 50 A m -2 .  -0.15  -0.20 -  r eb sb...... ••• •••••• ail.  F.1 CO  -0.25 -  war • am. aver a • woo.__  ^lib *vamp&  =I  ,  2  ui -0.30 Lu  -0.35 -  ..• • • 0000000000 • •  -0.40 -  tcath = 180s t ano d = 30 s T = 298 K galvanostatic T = 298 K tcath = 180s tano d = 30 s T = 343 K •••••■•••••••••^galvanostatic^T = 343 K  -0.45 0^50^100^150  ^  200  t [s]  Fig. B-5: Electro-oxidation of methanol on PtRu electrodeposited at 298 K in pulsed or  galvanostatic mode without surfactant. i = 50 A m -2 .  191  APPENDIX C: ASSESSMENT OF REPLICABILITY FOR PtRu AND PtRuMo DEPOSITION  Table C-1: Assessment of replicability of the methanol oxidation performance for catalyzed GF-S3. Single step deposition in 40 % wit Triton X-100 applying i = 60 A  111 -2  for  90 min (3 ml sandwich cell). Chronopotentiometry response:  Cyclic voltammetry  E [V vs. MSE]^response:  i [A In -2 ] at -0.2 V vs. MSE  depending on current density i= 50 Am -2^i= 100 A m -2 298 K^343 K^298 K^343 K PtRu deposition at 298 K  standard deviation  PtRu deposition at 333 K  standard deviation  PtRuMo deposition at 333 K  standard deviation  scan rate = 5 rnV s -1^sample  298 K^343 K^# 195 420 1 410 2 200 425 190 3 420 4 180  -0.21 -0.18 -0.16 -0.22  -0.32 -0.31 -0.3 -0.33  -0.14 -0.13 -0.12 -0.16  -0.27 -0.26 -0.25 -0.28  2.38E-02  1.12E-02  1.48E-02  1.12E-02  7.4  5.5  -0.31 -0.32 -0.32 -0.33  -0.39 -0.37 -0.37 -0.38  -0.26 -0.25 -0.24 -0.24  -0.34 -0.32 -0.33 -0.32  320 300 310 290  510 490 520 500  7.07E-03  8.29E-03  8.29E-03  8.29E-03  11.2  11.2  -0.34 -0.32 -0.34 -0.32  -0.41 -0.38 -0.41 -0.39  -0.26 -0.28 -0.25 -0.27  -0.35 -0.36 -0.35 -0.36  290 280 290 260  560 530 540 560  1.00E-02  1.30E-02  1.12E-02  5.00E-03  12.2  13.0  1 2 3 4 1 2 3 4  Tables Cl and C2 show a satisfactory level of reproducibility.  192  Table C-2: Assessment of replicability of the methanol oxidation performance for catalyzed GF-S3. Two step deposition with 40 % wt Triton X-100 applying i = 60 A m -2 for 90 min twice (3 ml sandwich cell). Cyclic voltammetry  Chronopotentiometry response:  E [V vs. MSE]^response:  i [A m -2 ] at -0.2 V vs. MSE  depending on current density i = 50 A m -2^i = 100 A m -2 298 K^343 K^298 K^343 K 2 step PtRu deposition at 333 K standard deviation  2 step PtRuMo deposition at 333 K standard deviation  scan rate = 5 mV s l^sample  298 K^343 K^# 790 1 380 750 2 370 380 780 3 4 370 780  -0.34 -0.33 -0.35 -0.34  -0.43 -0.42 -0.40 -0.41  -0.28 -0.27 -0.29 -0.27  -0.36 -0.35 -0.35 -0.34  7.07E-03  1.12E-02  8.29E-03  7.07E-03  5.0  15.0  -0.29 -0.30 -0.28 -0.28  -0.45 -0.44 -0.45 -0.43  -0.22 -0.20 -0.22 -0.23  -0.39 -0.37 -0.38 -0.36  300 290 280 280  810 760 790 760  8.29E-03  8.29E-03  1.09E-02  1.12E-02  8.3  21.2  193  1 2 3 4  APPENDIX D: CALCULATION OF GAS HOLD-UP AND OHMIC DROP  These calculations were carried out under the following assumptions: -  Flow-by mode 1 M CH 3 OH-0.5 M H2SO4, flow rate = 5 ml min -1  -  Anode thickness including compressed carbon cloth backing layer = 300 vim  -  Cross section perpendicular to methanol flow, Acs = 6.6 x10 -6 m2  -  Anode porosity = 0.85  -  Solubility of CO2 = 14.6 mol M -3 (333 K) and 9.7 mol m -3 (353 K)  -  Anode pressure -100 kPa(abs)  Table D-1: Calculation of DMFC variables as a function of temperature and current  density. i  1000  333 K 3000  5000  1000  0.00E+00  3.80E-08  8.60E-08  1.50E-09  4.90E-08  9.70E-08  0.00E+00  5.76E-03  1.30E-02  2.27E-04  7.42E-03  1.47E-02  1.26E-02  1.26E-02  1.26E-02  1.26E-02  1.26E-02  1.26E-02  0.00E+00  3.36E-01  7.61E-01  1.33E-02  4.34E-01  8.58E-01  H  0.00  0.25  0.43  0.01  0.30  0.46  [S M -1 ]  2.90E+01 1.24E+01 0.00E+00  2.90E+01 1.24E+01 9.20E-03  2.90E+01 1.24E+01 2.07E-02  3.30E+01 1.21E+01 3.71E-04  3.30E+01 1.21E+01 1.19E-02  3.30E+01 1.21E+01 2.34E-02  0.85 2.27E+01 0.01  0.85 1.32E+01 0.07  0.85 8.86E+00 0.17  0.85 2.41E+01 0.01  0.85 1.31E+01 0.07  0.85 9.37E+00 0.16  [A m -2 [m3, s -1 I ]  V' ul  [m s -1 ] [m s -i ]  u g /(ul+uo)  [ ]  ug  E Ko  T  L G  [kg rr1 -2 S -1 ] [kg M -2 S -1 ]  Ecf K*  [S M -1 1  AEohm  [ ]  N  353 K 3000  5000  The parameters listed in Table D-1 are explained below: Gas evolution rate:  N=iS A '  nF  (D-1)  Volumetric gas flow rate:  RT V' = N' P arm  (D-2)  194  Superficial gas velocity:  V' u=^ g Acs  (D-3)  Flow velocity of liquid phase:  u1= ^  (D-4)  ACS  Gas hold-up:  (D-5)  1—c^u chu o [Sigrist et al., 1980]: (uo = 0.0045 m s -1 (rising velocity of single bubble))  Conductivity  Temperature correction for e.g., 353 K (Casteel-Amis): \ 1 03^/  -4 /^\ 2^1.03 exp 4.3x10^35.34) — ^ (c —35.34) 35.34)^ 35.34  K T = 1484.2  (D-6)  (sulfuric acid concentration: c [Vo wt]) Correction for gas hold-up: T  K* - "=K 0 -  (  1—c  (D-7)  \ 1 + 0.5c  [Sigrist et al., 1980] Correction accounting for the porosity of the compressed felt (E c f— 0.85) and the liquid hold-up: K  = K *(H ,s cf )15^(D-8) .  Ohmic drop:  It is assumed that the current does not change across the thickness of the anode and that all current is in the ionic phase: fi*dx A^0  Ohm  K  i • 6x  (D-9)  195  The conversion of methanol was estimated assuming complete oxidation to CO2 (6 electron reaction, methanol feed concentration: 1 M) (Table D-2).  Table D -2: Theoretical maximum conversion of methanol. Flow rate [mlmin 1 ]  2 5 10 20  CH3OH conversion as a function of current density [A m 1000 2.6% 1.0% 0.5% 0.3%  2000 5.2% 2.1% 1.0% 0.5%  3000 7.8% 3.1% 1.6% 0.8%  4000 10.4% 4.2% 2.1% 1.0%  5000 13.0% 5.2% 2.6% 1.3%  6000 15.5% 6.2% 3.1% 1.6%  2 ]:  7000 18.1% 7.3% 3.6% 1.8%  8000 20.1% 8.3% 4.2% 2.1%  196  APPENDIX E: METHOD FOR VERIFYING REMOVAL OF TRITON X-100 FROM THE ELECTRODE SURFACE AFTER ELECTRODEPOSITION  Differential scanning calorimetry (or DSC) is a thermoanalytical technique in which the difference in the amount of heat required to increase the temperature of a sample and a reference are measured as a function of temperature. The temperature was increased linearly at a rate of 10 K per minute. A PtRu on GF-S3 sample prepared at 333 K using 60 A ni 2 and a deposition time of 90 minutes without surfactant was compared with a sample that was prepared under the same conditions with 40 °A m Triton X-100 (Figs. E-1 and E-2, respectively). Both samples were cleaned by washing and sonication as described in Section 5.1. In addition another (`dummy') sample of graphite fibers coated with a liquid crystalline mixture of 40 Vo wt Triton X-100 and water was prepared. The latter sample was tested without further treatment or cleaning (Fig. E-3).  DSC ^ mW  Temp C  300.00  200.00  100.00  0.00  ^  10.00  ^  20.00  ^  30.00  Time [min]  Fig. E-1: DSC test of catalyzed GF-S3 fibers prepared without surfactant.  197  Temp  C  300.00  200.00  100.00  0.00^  10.00^ Time [min]  20.00  30.00  Fig. E-2: DSC test of catalyzed GF-S3 fibers prepared with 40 %„ t surfactant. DSC mW  Temp C  2.00300.00 1.00-  0.00 200.00  -1.00-  -2.00 -  100.00  -3.00 0.00  10.00^ Time [min]  20.00  30.00  Fig. E-3: DSC test of catalyzed GF-S3 fibers coated with bulk 40 Vo„ t surfactant. The data shown in Figs. E-1 and E-2 indicate that the cleaning procedure provides adequate surfactant removal, as no discrepancy was observed qualitatively between the samples prepared with and without surfactant at temperatures below 250 C (523 K) where evaporation of Triton X-100 remains would occur. The increase in heat at —473 K,  198  required to maintain a steady temperature, is probably due to the heat absorption by the graphite fibers. Heat is released as the liquid crystalline structure is disintegrated at T = —313 K and therefore less heat is required to increase the temperature of the sample with bulk amounts of surfactant, which explains the negative trend with respect to the required wattage added to the system (Fig. E-3). At temperatures above 373-393 K Triton X-100 evaporates.  199  APPENDIX F: SAMPLE EDX AND AUGER SPECTRA ]-, Element^11t.%^At% C K 0 K PtM RuL  75 31 93 38 5 17 4 82 15 11 1 15 4.41 0 65  J J  Pt  AOJ^  110^  nr,  500^6.00^700^300  9.00^3000  Fig. F-1: Example of EDX spectrum obtained in point scan mode for PtRu electrodeposited on GF-S3 with 40 %," Triton X-100, 60 A m -2 , 2 x 90 min at 333 K.  200  PtRu Pt  80000  70000  60000  si g 50000  2  Ru + C 40000  30000  20000L 100 200 300 400 500 600 700 800 900 1000 1100 1200 1330 1400 1500 1600 1700 1800 1900 2000 2100 2200 Kinetic Energy (eV)  Fig. F-2: Example of Auger spectrum obtained for PtRu electrodeposited on GF-S3 with  40 °/0„,t Triton X-100, 60 A in -2 , 90 min at 333 K. Fig. F-2 shows the overlap of the Ru and C peaks that makes the quantitative  assessment of the catalyst surface Ru content difficult.  201  APPENDIX G: SUPPLEMENTAL FACTORIAL FUEL CELL TEST RESULTS  For each experiment described in APPENDIX G the 02 flow rate was 500 ml min -1 STP. The cathode pressure was 200 kPa(abs) and the sulfuric acid concentration fed to the anode was 0.5 M. The anode pressure was —100 kPa(abs). The utilized PtRu(Mo) catalysts were prepared as described in Section 6.1.8.2.  PtRu, serpentine flow Table G-1: Sequence of experiments carried out with PtRu anode, serpentine flow. run # 1  2 3 4 5 6 7 8 9 10 11 12 13  [CH3OH] [M] 1.25 2 0.5 0.5 2 2 2 0.5 0.5 1.25 2 2 1.25  flow rate [ml min - 1 ] 6 2 2 10 10 2 10 2 10 6 2 10 6  T [K] 333 353 313 353 353 313 313 353 313 333 353 313 333  response: p iw m -2 1 (ctr pt.)  (ctr pt.)  (ctr pt.)  replicate replicate replicate replicate  1500 2170 580 1680 1960 870 900 1750 630 1440 2030 840 1380  The corresponding polarization plots are given in Figs. G-1 and G-2.  202  0.8 -  •  1.25 M, 6 ml min d , 333 K  •  2 M, 10 ml min d , 353 K  A 2 M, 2 ml min d , 353 K  • 0.6  0.5 M, 10 ml min d , 353 K  El 0.5 M, 2 ml mind , 353 K  • li  A  = a, 0.4 -  1  or  a  •61 •  2M, 10 ml mind , 313 K  •  2M, 2 ml min d , 313 K  •  0.5 M, 10 mi min d , 313 K  • •0  0A  t  0.2 -  •  .^, A • U •  0 5 M^2 ml min d , 313 K  7,  t  •A •  61^  •  0 .0 2000  0  8000  4000^6000  i [A m -2]  2500 -  2000 -  •  1.25 M, 6 ml min 1 , 333 K  •  2 M, 10 ml min d , 353 K  A 2 M, 2 ml min d , 353 K •  0.5 M, 10 ml min 1 , 353 K  O  0.5 M, 2 ml min d , 353 K  ^  2M, 10 ml min -1 , 313 K  A  •  •  A^A  •^•  ^  2 M, 2 ml min -1 , 313 K  A  •  0.5 M, 10 ml min -1 , 313 K A  O N 1500 -  0.5 M, 2 ml min -1 , 313 K A  li^ li^•• • • • •  I  C  a 1000  •  -  Q 7  500 -  0  • • 0  A  e 0  ^  t  T  2000^4000^6000^8000  i [A m -2]  Fig. G-1: Polarization plots for PtRu, serpentine flow (runs #1-9).  203  0.8 0.7 0.6  •  1.25 M, 6 ml min d , 333 K  •  1.25 M, 6 ml min d , 333 K (rep. 1)  •  1.25 M, 6 ml min d , 333 K (rep. 2'  •  2 M, 2 ml min d , 353 K  A 2 M, 2 ml min d , 353 K (rep.)  0.5  A  0.4  •  2 M, 10 ml min d , 313 K  •  2 M, 10 ml min d , 313 K (rep.)  Lu° 0.3  8•  0.2  •^ A •  ^A  0.1 0.0 0  ^  2000  ^  4000^6000  i [A m -2 ]  2500 -  2000 -  •  1.25 M, 6 ml min -1 , 333 K  •  1.25 M, 6 ml min d , 333 K (rep. 1)  •  1.25 M, 6 ml min -1 , 333 K (rep. 2)  •  2 M, 2 ml min-1 , 353 K  •  2 M,^2 ml min -1 , 353 K (rep.)  •  2 M, 101,11min -1 , 313 K  •  2 M, 10 ml min -1 , 313 K (rep.) A  •  • A  O  E a 1000 -  0  A  •  • o 0  •  •  0  ■  0  •  A  • A  1500 -  500 -  •  •  A ^  2000  ^  4000  ^  6000  ^  8000  i [A n1 2]  Fig. G-2: Polarization plots for PtRu, replicability assessment, serpentine flow.  204  PtRu, flow-by mode Table G-2: Sequence of experiments carried out with PtRu anode, flow-by mode.  run # 1 2 3 4 5 6 7 8 9 10 11 12 13  [CH 3 01-1] [M] 1.25 2 0.5 2 2 0.5 2 0.5 0.5 1.25 2 2 1.25  flow rate [ml min -1 ] 6 10 2 2 2 2 10 10 10 6 2 10 6  T [K] 333 313 313 353 313 353 353 313 353 333 353 313 333  (ctr pt.)  (ctr pt.) replicate replicate replicate (ctr pt.) replicate  response: p [W m-21 1560 920 600 2240 880 1740 1890 660 1620 1500 2100 870 1400  The corresponding polarization plots are given in Figs. G-3 and G-4.  205  •  1.25 M, 6 ml min d , 333 K  • 2 M, 10 ml min d , 353 K A 2 M, 2 ml min d , 353 K  A  •  0.5 M, 10 ml min d , 353 K  ^  0.5 M, 2 ml min d , 353 K  •  2M, 10 ml min d , 313 K  ^  2M, 2 ml min d , 313 K  •  0.5 M, 10 ml min d , 313 K  O A  0.5 M, 2 ml mind , 313 K  • •  A  • • •^•  A  O  i [A m -2]  •  1.25 M, 6 ml min  d , 333 K  • 2 M, 10 ml min -1 , 353 K 2500 - A 2 M,^2 ml min -1 , 353 K  2000 -  r-. 1500 -  '11  E  •  0.5 M, 10 ml min -1 , 353 K  ^  0.5 M, 2 ml min d , 353 K  •  2 M, 101.111min -1 , 313 K  ^  2 M,^2 ml min -1 , 313 K  •  0.5 M, 10 ml min -1 , 313 K 6'  •  0.5 M, 2 ml min d 313  Q 1000 -  A  A  A O  6  •  •  A  •  • • • ■  A  •  ^  0  500 -  0  8 0  ^  2000  ^  4000  ^  6000^8000  i [A mr 2]  Fig. G-3: Polarization plots for PtRu, flow-by mode (runs #1-9).  206  0.8  0.6  •  1.25 M, 6 ml min d , 333 K  O  1.25 M, 6 ml min d , 333 K (rep. 1)  O  1.25 M, 6 ml min d , 333 K (rep. 2)  •  2 M, 2 ml min d , 353 K  A 2 M,^2 ml min-1 , 353 K (rep.)  •  0.4 -  2 M, 10 mi min d , 313 K  •  A  El 2 M, 10 ml min d , 313 K (rep.)  A  8  g  •  ZI)^ • o^M  0.2 -  A  a^t  e o  lb  0.0 8000  4000^6000  2000  0  i [A m -2 ]  2500 -  •  1.25 M, 6 ml min d , 333 K  •  1.25 M, 6 ml min d , 333 K (rep. 1)  •  1.25 M, 6 ml min d , 333 K (rep. 2)  •  2M, 2 mi min d , 353 K  •  A 2 M,^2 ml min d , 353 K^(rep.)  2000 -  •A  •  2M, 10 ml min d , 313 K^•  •  •  2 M, 10 ml min d , 313 K (rep.)  A  A  •^ • •  ^A  1500  E  8^•  o^•  •  A  0 •  a 1000 -  •  500 -  0 0  ^  2000  ^  4000  ^  i [A m -2  6000  ^  8000  ]  Fig. G-4: Polarization plots for PtRu, replicability assessment, flow-by mode.  207  PtRuMo, serpentine flow Table G-3: Sequence of experiments carried out with PtRuMo anode, serpentine flow.  run #  [CH3OH] [M]  flow rate [ml min -l ]  T [K]  1 2 3 4 5 6 7 8 9 10 11 12 13  1.25  6 10 10  333  0.5 2 0.5 2 0.5 2 2 0.5 1.25 2 2 1.25  2 10  313 353 313 353 353 313 313 353 333 353 313  6  333  2 2 2 2 10 10 6  response: p LW in -2 1  (ctr pt.)  (ctr pt.)  replicate replicate replicate (ctr pt.) replicate  1500 540 2170 480 2310 1980 690 780 1860 1400 2100 720 1350  The corresponding polarization plots are given in Figs. G-5 and G-6.  208  ^  •  0.8 -  •  1.25 M, 6 ml min d , 333 K  •  2 M, 10 ml min d , 353 K  • •  ts  2 M, 2 ml min d , 353 K 0.5 M, 10 ml min d , 353 K 0.5 M, 2 ml min d , 353 K  0.6 i  A A Ir^W^  2 M, 10 mi min d , 313 K  •  2M, 2 ml min d , 313 K  ••  • •  0.4 -  •  0.5 M, 10 mi min d , 313 K  0^i^Oi 0.5 M, 2 mi min d , 313 K •  *^ • 0^  II • 4 •^ W •^ W • t^ v^•^ •  v  •^• • o^v^  0.2 -  W^4 •  t  0.0 ^ 0^2000^4000^6000^8000  i [A n 2 ]  2500 -  •  1.25 M, 6 ml min d , 333 K  •  2M, 10 ml min -1 , 353 K  •  2 M, 2 ml min d , 353 K  • o  2000 -  • ^ •  N  O  isoo  A 6 A^ •^A 2M, 10 mi min 1 , 313 K^ • 0^0 2 M, 2 ml min d , 313 K• o^•^ D • A^ •  •  -  500 -  •  -  0.5 M, 10 ml min d , 313 K •^ 0.5 M, 2 ml min -1 313 K I:3 •^• • L ••  • I^ •  C  1000  0.5 M, 10 ml min -1 , 353 K  0.5 M, 2 ml min -1 , 353 K^A^  I  e^a  • • •  ■  •  •V  0  0 0  2000  4000  6000  8000  i [A mr 2]  Fig. G 5: Polarization plots for PtRuMo, serpentine flow (runs #1-9). -  209  ^  0.8  •  1.25 M, 6 ml min d , 333 K  0.7^  0 1.25 M, 6 ml min d , 333 K (rep. 1)  0.6^  • 2 M, 2 ml min d , 353 K  O  1.25 M, 6 ml min d , 333 K (rep. 2)  A 2 M, 2 ml min d , 353 K (rep.) 0.5  ^  0.4  8^A^■ 2 M, 10 ml min d , 313 K . -1 Ot 2 M, 10 ml mint , 313 K (rep.) e5^t • A  (!)  w  •  A^  0.3  •  A^ ^  •0  0.2  •  A  g  0.1 0.0 0  ^  2000  ^  4000^6000  ^  8000  i [A m -2 ]  2500 -  2000 -  •  1.25 M, 6 ml min d , 333 K  O  1.25 M, 6 ml min d , 333 K (rep. 1)  O  1.25 M, 6 ml min d , 333 K (rep. 2)  •  2 M, 2 ml min d , 363 K  •^• 6 2 M, 2 ml min d , 353 K (rep.)^ •^6^A • 2M, 10 ml min d , 313 K^ O  • A  2 M, 10 ml min d , 313 K (rep.) •  A  A  E  1500 -  A  500 -  •  • 8  • •  A  1000 -  A  •  • 0  •  0 0^2000^4000^6000  ^  8000  i [A m -2]  Fig. G-6: Polarization plots for PtRuMo, replicability assessment, serpentine flow.  210  PtRuMo, flow-by mode Table G-4: Sequence of experiments carried out with PtRuMo anode, flow-by mode. run # 1 2 3 4 5 6 7 8 9 10 11 12 13  [CH3OH] [M] 1.25 2 0.5 2 0.5 2 0.5 0.5 2 1.25 2 2 1.25  flow rate [ml min -1 ] 6 10 2 2 2 10 10 10 2 6 2 10 6  T [K] 333 353 353 313 313 313 353 313 353 333 353 313 333  response: p iw in -2 L_ (ctr pt.)  replicate replicate replicate (ctr pt.) replicate  (ctr pt.)  1550 2170 2040 720 500 810 1800 570 2380 1450 2240 600 1380  The corresponding polarization plots are given in Figs. G-7 and G-8.  211  •^  0.8  •  1.25 M, 6 mi min d , 333 K  •  2 M, 10 ml min d , 353 K  A 2M, 2 mi min d , 353 K  0.6  ^  0.5 M,^2 ml mind , 353 K  •  2 M,^10 ml min -1 , 313 K  6^Q  V^2M,^2 ml min d , 313 K  E,7 0.4 12  •  A  • 0 I  3  A • ii •  V  ♦  o  2000  •  •  0.2  ^  i 0 •  V E  0  0.5 M, 10 mi min -1 , 353 K  Q  •  0.0  •  ^  T t5  ♦  0.5 M, 10 ml min -1 , 313 K  <> 0 0.5 M,^2 ml min d , 313 K  8^A  • •  •0^A •^•0 •• ■  •  •  T  4000^6000  ^  8000  i [A rt 2]  •  1.25 M, 6 ml min d , 333 K  •  2M, 10 ml min d , 353 K  zs 2 M, 2 mi min d , 353 K 2500 - •  0.5 M, 10 ml min d , 353 K  ^  0.5 M, 2 ml min d , 353 K  •  2M, 10 ml mind , 313 K  2000 - ^  2 M, 2 ml min d , 313 K  ♦  0.5 M, 10 ml min d , 313 Ki  O  0.5 M, 2 ml min d , 313 KO  A^■ •0 • I  .--. 1500 ,•,  E  °- 1000 -  I •  500 -  0  e 0  •v  e^a•0  0  2000  A ^A A A^•^A •0 ^^• • ■^•^^ • • ••^■ •  •v ^•v 0 ■ 4000^6000  8000  i [A rn -2]  Fig. G 7: Polarization plots for PtRuMo, flow-by mode (runs #1-9). -  212  ^  0.8 0.7 0.6•  2  NI,  2  •  1.25 M, 6 ml min d , 333 K  o  1.25 M, 6 ml min d , 333 K (rep. 1)  O  1.25 M, 6 ml min d , 333 K (rep. 2)  ml  t^  A  min d , 353 K  2 M, 2 ml min d , 353 K (rep.)  g^t^• 2 M, 10 ml min d , 313 K 2^ g^A^^t 2 M, 10 ml min d , 313 K (rep.) 0.4^• 0.5^  ^  •  0.3  M^  t  M^  ^  •  0.2^  ^  t  • A  • A  ■ ^  0.1 0.0 0  ^  2000  ^  4000^6000  ^  8000  i [A m 2] -  2500 -  2000 -  •  1.25 M, 6 ml min d , 333 K  O  1.25 M, 6 ml min -1 , 333 K (rep. 1)  O  1.25 M, 6 ml min -1 , 333 K (rep. 2)  2 M, 2 ml min -1 , 353 K^ • • 6 2 M, 2 ml min -1 , 353 K (rep.)^•^A^ •  • ^  •  A  ";.E  1500 -  a 1000  -  A  2 M, 10 ml min -1 , 313 K^ 6^ -1 2 M, 10 ml min , 313 11(rep.)  •  • •  O  a  0  • 6  •  0^•  A  ■  A^  0  ■  500  0 0  2000  4000  6000  8000  i [A M -2 ]  Fig. G-8: Polarization plots for PtRuMo, replicability assessment, flow-by mode.  213  

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.24.1-0058502/manifest

Comment

Related Items