This item has the following downloads:
1 STUDY OF PLATINUM FAMILY CATALYSTS IN DIRECT METHANOL FUEL CELL A ND PARAHYDROGEN INDUCED HYPERPOL ARIZATION APPLICATIONS By WEI CHENG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Wei Cheng
3 To my loving father, mother, family, and friends for their love, sacrifice, and full supp ort
4 ACKNOWLEDGMENTS The completion of this work would not be possible without guidance from my committee members Dr. Helena Hagelin Weaver, Dr. William Lear, Dr. Mark Orazem and Dr. Peng Jiang They have supported, guided, and encouraged me with a l ot of patience through my Ph.D. program. I would like to take this opportunity to express my thankful gratitude to them. First, I would like to express my deepest gratitude to my advisor, Dr. Helena Hagelin Weaver, for her excellent guidance, caring, pati ence knowledge, and dedication to teach ing and for providing me with an excellent atmosphere for perform ing my research. She was always supportive and positive when I met difficulties in my research, and that really made my PhD study in UF an enjoyable on e. I would like to thank Dr. William Lear for giving me the opportunity to work in the excellent direct methanol fuel cell project, and for his numerous supportive guidance on the fuel cell project. I would like to thank Dr. Orazem for guidance on electroc hemical theory and experimentation, as well as for being so supportive and giving me access to the equipment in his lab w ithout which Section 2 (Chap 3 of the dissertation) could not be accomplished. I would like to sincerely thank Dr. Peng Jiang for his guidance and serv e on my supervi sory committee, for years of supportive guidance on my research and life during my entire st a y at the University of Florida. I would also thank Dr. Clifford Russell Bowers from the chemistry department for giving me an oppo rtunity to working on the heterogeneous parahyd rogen reaction project. W ithout his guidance, knowledge and scientific insight, it would not have been possible to complete Section s 3, 4 (Chap 4, 5 of the dissertation.
5 I also express my special thanks to Dr. Luke M. Neal, who coached me on most of the catalysis laboratory techniques, and who provided much knowledge through excellent discussion s and conversations Give thanks to Ronghui Zhou from Dr. Bowers lab, with whom I collaborate d closely on the heteroge neous parahydrogen reaction project. Without all the NMR measurements performed by Ronghui, Section 3, 4 (Chap 4, 5 of the dissertation) would not have been possible I would also like to thank my lab partners, Justin Dodson, Trent Elkin s, and Haibing Zhe ng for their help in numerous laboratory technique and academic discussions. I would also like to thank the research group and our collaborators past and presen t in the DMFC project : Dr. Jim Fletcher, Dr. Phillip Cox, Dr. C.C. Kuo, Dr. Sydni Credle, Dr. Rafe Biswas, Shyam Prasad, Fenner Colson, Matt Inman, Xin Zhan, Quanning Li, Jason Harrington, John Crittenden, Praneeth Pillarisetti, Anupam Patil, Phil Bailey. Special thanks go insightful discussion on fuel cell electrochemistry. Also, I would like to express my sincere gratitude to the Chemical Engineering department at the University of Florida giving me this opportunity to purs u e my PhD degree in this wonderful atmosphere, and my great thanks to t he Department of Energy the American Chemical Society, and the Petroleum Research Fund for their financial support for my research. Finally, I would like to thank my parents for their love, support, encouragement and sacrifice for letting me, their only c hild to leav e home for so long to pursue my PhD
6 study at the University of Florida. T hanks to all my friends for being so supportive during my PhD study.
7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 11 LIST OF ABBREVIATIONS ................................ ................................ ........................... 15 ABSTRACT ................................ ................................ ................................ ................... 17 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 19 1.1 Overview of Platinum Catalysts ................................ ................................ ........ 19 1.2 Pt based Catalysts in Direct Methanol F uel Cell Applications ........................... 20 1.2.1 Pt based Catalysts in DMFC Anode Application ................................ ..... 21 1.2.2 Pt based Catalysts in DMFC Cathode Applicatio n ................................ .. 23 1.3 Pt based Catalysts in Parahydrogen Induced Polarization Applications ........... 24 1.4 Ir based Catalysts in Parahydrogen Induced Po larization Applications ............ 26 1.5 Objective and Motivations ................................ ................................ ................. 27 1.5.1 Commercial Anode Catalyst Screen Study for An Open Cathode Desig n DMFC ................................ ................................ ................................ 27 1.5.2 Investigation of Contamination Effects of Methyl Ethyl Ketone on A Platinum /Carbon Electrocatalyst for Oxygen Reduction Activity .................. 27 1.5.3 Platinum and Iridium based Catalysts in Parahydrogen Introduced Hyperpolarization Applications ................................ ................................ ...... 28 2 COMPARATIVE STUDY OF COMMERCIAL ANODE CATALYSTS IN AN OPEN CATHODE DIRECT METHANOL FUEL CELL ................................ ............ 29 2.1 Background ................................ ................................ ................................ ....... 29 2.1.1 Types of Fuel Cell ................................ ................................ .................... 29 2.1.2 Anode Catalyst in DMFC ................................ ................................ ......... 34 2.1.2 Objectives ................................ ................................ ................................ 35 2.2 Experimental Methods ................................ ................................ ...................... 36 2.2.1 Membrane Exchange Assembly Configuration ................................ ........ 36 2.2.2 Fuel Cell Testing Experimental Setup ................................ ..................... 37 2.2.3 Electro chemical Characterization Experimental Setup ............................ 38 126.96.36.199 MEA initial performance measurement ................................ .......... 38 188.8.131.52 Single cell overall polariz ation measurement ................................ 39 184.108.40.206 Anode polarization measurement ................................ ................... 40 220.127.116.11 Cathode polarization measurement ................................ ............... 40
8 18.104.22.168 Single cell resistance measurement ................................ ............... 41 22.214.171.124 Methanol crossover measurement ................................ ................. 41 2. 2.3.7 Accelerated low voltage anode durability test ................................ 42 2.2. 3 8 Calculation of electrochemically active surface area calculation .... 43 2.3 Results and Discussion ................................ ................................ ..................... 45 2.3.1 Initial Performance of MEAs ................................ ................................ .... 45 2.3.2 Single Cell Polarization and Power Density Curves ................................ 47 2.3.3 Anode Polarization ................................ ................................ .................. 48 2.3.4 Relative Electrochemical Active Surface Area ................................ ......... 48 2.3.5 Methanol Crossover ................................ ................................ ................ 49 2.3.6 Cathode Polarization ................................ ................................ ............... 50 2.3.7 MEA Resistance ................................ ................................ ...................... 51 2.4 Summary ................................ ................................ ................................ .......... 51 3 THE IMPACT OF METHYL ETHYL KETONE CONTAMINATION OF A PLATINUM/CARBON ELECTROCATALYST ON ITS OXYGEN REDUCTION ACTIVITY ................................ ................................ ................................ ................ 68 3.1 Background ................................ ................................ ................................ ....... 68 3.2 Experimental Methods ................................ ................................ ...................... 71 3.2.1 Electrocatalyst and Electrolyte ................................ ................................ 71 3.2.2 Electrode Preparation ................................ ................................ .............. 71 3.2.3 Electrochemical Measurement ................................ ................................ 72 3.3 Res ults and Discussion ................................ ................................ ..................... 73 3.3.1 Cyclic Voltammetry Studies on the DMFC Cathode Catalyst in the Presence of MEK Contamination ................................ ................................ .. 73 3.3.2 Oxidation Reduction Reaction Kinetics for the DMFC Cathode Catalyst in the Presence of MEK Contamination ................................ ........... 76 126.96.36.199 ORR polarization curves ................................ ................................ 76 188.8.131.52 ORR kinetically current density calculation ................................ .... 77 184.108.40.206 Oxygen reduction reaction kinetics ................................ ................ 80 3.3.3 Discussio n ................................ ................................ ............................... 84 3.4 Summary ................................ ................................ ................................ .......... 86 4 KINETIC STUDY OF PROPYLENE HYDROGENATION OVER A PLATINUM CATALYST USING AN IN SITU HYPERPOLARIZED NMR TECHNI QUE ............ 97 4.1 Background ................................ ................................ ................................ ....... 97 4.2 Experimental Methods ................................ ................................ .................... 101 4.2.1. Cataly st Preparation ................................ ................................ ............. 101 4.2.2 Chemisorption Measurement ................................ ................................ 102 4.2.3 NMR Studies Based on Parahydrogen Induced Polarization (PHIP) ..... 102 4.3 Results and Discussion ................................ ................................ ................... 104 4.3.1 Catalyst Characterization ................................ ................................ ...... 104 4.3 .2 1 H NMR Spectra during Propylene Hydrogenation with Parahydrogen 105 4.3.3 NMR Signal Enhancement of PHIP and Pairwise Hydrogen Addition Selectivity ................................ ................................ ................................ .... 107
9 4. 4 Summary ................................ ................................ ................................ ........ 109 5 KINETIC STUDY OF PROPYLENE HYDROGENATION OVER IRIDIUM CATALYST USING AN IN SITU HYPERPOLARIZAED NMR TECHNIQUE ........ 113 5.1 Background ................................ ................................ ................................ ..... 113 5.2 Experimental Methods ................................ ................................ .................... 115 5.2.1. Catalyst Preparation ................................ ................................ ............. 115 5.2.2 Inductively coupled Plasma atomic Emission Spectrometry ................. 115 5.2.3 Temperature Programmed Reduction Measurements ........................... 115 5.2.4 Chemisorption Measurement ................................ ................................ 116 5.2.5 Transmission Electron Microscopy (TEM) Measurement ...................... 117 5 .2.6 NMR Studies Based on Parahydrogen Induced Polarization (PHIP) ..... 117 5.3 Results and Discussion ................................ ................................ ................... 118 5.3.1 Catalyst Characterizat ion ................................ ................................ ...... 118 5.3.2 Temperature Dependence of NMR Signal Enhancement Factor and Pairwise Hydrogen Addition Selectivity ................................ ....................... 119 5.3.4 NMR Sig nal Enhancement Factor and Pairwise Hydrogen Addition Selectivity with Respect to Hydrogen Concentration ................................ ... 120 5.4 Summary and Discussion ................................ ................................ ............... 121 6 CONCLUSIONS ................................ ................................ ................................ ... 129 LIST OF REFERENCES ................................ ................................ ............................. 132 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 139
10 LIST OF TABLES Table page 2 1 Power density of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVADT at a current density of 120 mA cm 2 ............... 66 2 2 Power density loss rate for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVADT at a current density of 120 mA cm 2 ................................ ................................ ................................ ..................... 66 2 3 Anode potential for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVADT at a current density of 120 mA cm 2 .... 66 2 4 Cathode potential for DMFCs with JM ELE 147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVADT at a current density of 120 mA cm 2 .... 66 2 5 Initial and post test performance for DMFCs with JM ELE147, JM ELE170, Tanaka, an d Cabot anodes during the 144 hours LVADT at current density of 120 mA cm 2 ................................ ................................ ................................ ....... 67 2 6 Degradation in different components for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during the 144 hours LVADT at current density of 120 mA cm 2 ................................ ................................ ....................... 67 3 1 H udp and Platinum electrochemically active surface area for a thin film Hispect11100 60 wt% Pt/C electrode in Nitrogen saturate d 0.5 mol L 1 H 2 SO 4 solution contain dissolved MEK at various concentrations. Pt catalyst loading 60 g cm 2 GC ;150 cm 3 N 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 50 C; sweep rate: 20 mV/s ................................ ................................ ................................ ...... 96 3 2 Parameters obtained from Levich plots of the limiting diffusion current density j f for ORR for a thin film Hispect11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution in the presence and absence of dissolved MEK in various concentr ations ................................ ........................... 96
11 LIST OF FIGURES Figure page 2 1 Conventional design of a close cathode DMFC system with its water recycling system (Courtesy of Dr.J ames Fletcher, University of North Florida). ................................ ................................ ................................ .............. 54 2 2 Simplified design of an open cathode DMFC system with liquid barrier layer. (Courtesy of Dr.James Fletcher University of North Florida ). ........................... 55 2 3 Schematic of a DMFC stack. ................................ ................................ .............. 56 2 4 Initial durability test of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabo t anodes Operating condition: cell temperature: 5 0 C, cell voltage 0.35 V Anode feed: 0.8 M CH 3 OH solution flow rate 2 mL min ............................. 57 2 5 Initial life/stabilized power density and initia l operational power density loss for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes a) Comparison of beginning life and stabilized power densities. b) Comparison of operational power density loss for the first hour of operation. The power densi ty loss was calculated by subtracting stabilized power density of the specific DMFC from the initial power density. Operating condition: cell temperature: 5 0 C, cell voltage 0.35 V Anode feed: 0.8 M CH 3 OH solution flow rate 2 mL min ................................ ................................ ........................... 58 2 6 Polarization curves and power densities of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test Polari zation test conditions: cell temperature: 5 0 C, anode feed: 0.8 M CH 3 OH solution flow rate 2 mL min cathode feed: air, flow rate 2 SLPM. ................................ ................................ ... 59 2 7 Anode polarization curves and power den sities of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test Polarization test conditions: cell temperature: 5 0 C, anode feed: 0.8 M CH 3 OH solution flow rate 2 mL min cathode fee: hydrogen, flow rate 0.3 SLPM. ................................ ....................... 60 2 8 Anode relative electrochemical active surface areas of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test Data calculated at 120 mA cm 2 Test conditions: cell temperature: 5 0 C, anode feed: 0.8 M CH 3 OH solution flow rate 2 mL min cathode fee: hydrogen, flow rate 0.3 SLPM. ..................... 61 2 9 Methanol crossover of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated d urability test. Polarization test condition: cell temperature: 50 C, anode
12 feed: 0.8 M CH 3 flow rate 1 SLPM. ................................ ................................ ............................... 62 2 10 Cathode polariz ation curves of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test. Test conditions: cell temperature: 50 C, anode air, flow rate 2 SLPM. ................................ ................................ ................................ ....... 63 2 11 Impedance spectrum of DMFCs with JM ELE147, JM ELE170, Tanaka, Cabot anodes during a course of 144 hours low voltage accelerated durability test. Operating con dition: cell temperature: 50 C, anode feed: 0.8 rate 0.3 SLPM, test frequency range: 10 kHz to 0.1 Hz with an AC amplitude of 20 mV. ................................ ................................ ................................ ............ 64 2 12 Resistance of DMFCs with JM ELE147, JM ELE170, Tanaka, Cabot anodes during a course of 144 hours low voltage accelerated durability test Operating condition: cell temperature: 5 0 C, anode feed: 0.8 M CH 3 OH solution flow rate 2 m L min cathode feed: hydrogen gas, flow rate 0.3 SLPM, test frequency range: 10 kHz to 0.1 Hz with an AC amplitude of 20 mV. ................................ ................................ ................................ ..................... 65 3 1 CV curves for a thin film Hispect11100 60 wt% Pt/C electrod e in nitrogen saturated 0.5 mol L 1 H 2 SO 4 solution containing dissolved MEK at various concentrations. Pt catalyst loading 60 g cm 2 GC ;150 cm 3 N 2 satd. 0.5 mol L 1 H 2 SO 4 solution at 50 C; sweep rate: 20 mV/s ................................ ................... 88 3 2 Rotating disk electrode polarization curves for a thin film Hispect11100 60 wt% Pt/C electrode in the presence and absence of dissolved MEK at various concentrations. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s, rotation rate: 1600 rpm. .................... 89 3 3 Oxygen reduction reaction polarization curves for a thin film Hispect11100 60 wt% Pt/C electrode in the presence and abs ence of dissolved MEK in various concentrations and at various rotating rates. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; sweep rate: 20 mV/s. ................................ ................................ ................................ .................. 90 3 4 Koutechi Levich plots of the limiting diffusion current density j d for ORR for a thin film Hispect11100 60 wt% Pt/C electrode in the presence and absence of dissolved MEK at various concentrations. Pt catalyst loading 30 g cm 2 GC ; 150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; sweep rate: 20 mV/s. ...... 91 3 5 Tafel plots for mass transfer corrected ORR kinetic current density for a thin film Hispect11100 60 wt% Pt/C in the presence a nd absence of dissolved MEK in various concentrations. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotation rate: 1600 rpm. ....... 92
13 3 6 Corres ponding changes to the mass transfer corrected ORR kinetic current density j k at different overpotentials on a thin film Hispect11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution as a function of dissolved MEK concentration. Pt cata lyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotation rate: 1600 rpm. ....... 93 3 7 Turnover frequencies (TOF) for ORR on a thin film Hispect11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution at 25 C as a function of potentials without or with presence of different concentration of MEK in electrolyte. Calculated from the kinetic current densities used in Fig 3 5 using Eq 3 7. Pt ca talyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotation rate: 1600 rpm. ............................. 94 3 8 Corresponding changes to the turnover frequencies (TOF) at differe nt overpotentials on a thin film Hispect11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution as a function of dissolved MEK concentration. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotation rate: 1600 rpm. ................................ 95 4 1 1 H NMR spectra of propylene hydrogenation with a) parahydrogen and b) normal hydrogen over Pt/TiO 2 catalyst at 150 C and 1atm. Gas flow rate (total 300 mL/m in): N 2 /p H 2 /Propylene=120/150/30 mL/min. Data were taken under steady state conditions at 1atm. (NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab in the UF Chemistry Department) .. 110 4 2 3 D plots of 1H NMR spectra during propylene hydrogenation with parahydrogen over Pt/TiO 2 catalyst at different temperatures. Gas flow rate (total 300 mL/min): N 2 /p H 2 /Propylene=120/150/30 ml/min. Data were taken under steady sta te conditions and 1atm. (NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab in the UF Chemistry Department) ................................ ................................ ................................ ...... 111 4 3 Pairwise selectivities for propylene hydrogena tion over Pt/TiO 2 catalyst at different temperatures. Two sets of reactions were carried out with n H 2 or p H 2 respectively. The CH 2 group of the product propane was integrated for all calculations, and the integral from n H 2 was subtracted from that of p H 2 and the result was treated as pure PHIP contribution. Enhancement factors were calculated based on the ratio between the PHIP and thermal integrals with n H2. Pairwise selectivities were calculated based on the ratio of experimental enhancement factor over theoretical enhancement factor. Gas flow rate (total 300 mL/min): N2/p H2/Propylene=120/150/30 mL/min. Pressure : 1 atm. lab.) ................................ ................................ ................................ .................. 112 5 1 Hydrogen temperature programmed reduction profile for Ir/TiO2 catalyst. heating rate at 10C per minute up to a temperature of 500C ......................... 124
14 5 2 TEM image of Ir/TiO 2 catalyst. 100,000X magnification. ................................ ... 124 5 3 1 H NMR spectra of propylene hydrogenation with a) parahydrogen and b) normal hydrogen over Ir/TiO 2 catalyst at 150 C. Gas flow rate (total 300 ml/ min): N 2 /p H 2 /Propylene=120/150/30 mL/min. Data were obtained under steady state conditions. (NMR data obtained by Ronghui Zhou through ................................ ................................ .. 125 5 4 3 D plo ts for 1 H NMR spectra measured during propylene hydrogenation with parahydrogen over Ir/TiO 2 catalyst at different temperature. Gas flow rate (total 300 mL/min): N 2 /p H 2 /Propylene=120/150/30 mL/min. Data were taken under steady state conditions (NMR data o btained by Ronghui Zhou through ................................ ................................ .. 126 5 5 Pairwise selectivities for propylene hydrogenation over Ir/TiO 2 catalyst at different temperatures. Two sets of reactions were carried out with n H2 or p H 2 respectively. The CH 2 group of the product propane was integrated for all calculations. Integral for n H 2 was subtracted from that of p H 2 which was treated as pure PHIP contribution. Enhancement factors were calculated based on the ratio between the PHIP and thermal integrals of n H 2 Pairwise selectivity was calculated based on the ratio of experimental enhancement factor over theoretical enhancement factor. Gas flow rate (total 300 mL/min): N 2 /p H 2 /Propylen e=120/150/30 mL/min. (Produced from NMR data obtained ......................... 127 5 6 Pairwise selectivities for propylene hydrogenation over Ir/TiO 2 a t different hydrogen partial pressures at 150 C. Two sets of reactions were carried out with n H 2 or p H 2 respectively. The CH 2 group of the product propane was integrated for all calculations. The integral from n H 2 was subtracted from that of p H 2 and the result was treated as pure PHIP contribution. Enhancement factors were calculated based on the ratio between the PHIP and thermal integrals of n H 2 Pairwise selectivity was calculated based on the ratio of experimental enhancement factor over theoret ical enhancement factor. Gas flow rate (total 300 mL/min): Flow rate of propylene=30 mL/min. (Produced from NMR data obtained by Ronghui Zhou through collaboration ................................ ................................ ........................ 128
15 LIST OF ABBREVIATIONS ACL An ode catalyst layer AFC Alkaline fuel cell AGDL Anode gas diffusion layer ALTADENA adiabatic longitudinal transport after dissociation engenders net alignment C CL Cathode catalyst layer CFC Cathode flow channel CGDL Cathode gas diffusion layer CV Cyclic Vo ltammetry DEMS D ifferential electrochemical mass spectrometric DMFC Direct methanol fuel cell EASA Electrochemical active surface area H FC Hydrogen fuel cell H upd Hydrogen underpotential deposition ICP AES inductively coupled plasma atomic emission spectro scopy LBL Li quid barrier layer LSV Li near sweep voltammetry LVADT L ow v oltage accelerated d urability test MCFC C arbonate fuel cell MEA Membrane electrolyte assembly M EK Methyl ethyl ketone MEM membrane NMR Nuclear magnetic resonance OCV Open circuit voltag e
16 ORR Oxygen reduction reaction PAFC P hosphoric a cid fuel cell PASADENA parahydrogen and s ynthesis allow dramatically enhanced nuclear alignment PEM FC P olymer electrolyte m embrane fuel cell PHIP parahydrogen induced polarization RDE Rotating disk electrod e rEASA Relative electrochemical active surface area SCE Saturated calomel electrode SLPM Standard liter per minute SOFC Solid o xide fuel cell TCD thermal conductivity detector TEM Transmission electron microscopy TOF Turnover frequency TPR temperature pr ogrammed reduction
17 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STUDY OF PLATINUM FAMILY CATALYSTS IN DIRECT METH ANOL FUEL CELL AND PARAHYDROGEN INDUCED HYPERPOLARIZATION APPLICATIONS By Wei C heng August 2013 Chair: Helena Hagelin Weaver Major: Chemical Engineering Platinum is an important and widely used metal, which finds many applications in various catalytic processes. In this study, platinum and iridium catalysts were investigated in two different applications: as electrocatalysts in a direct methanol fuel cell and as hydrogenation catalysts in parahyrogen induced polarization (PHIP), an important NMR signal enhancement technology. The success of direct methanol fuel cell (DMFC) technology depends largely on the platinum electro catalyst. A low voltage accelerated durability test method, together with various other el ectrochemical measurements, was employed to study the anode/cathode degradation in four commercial anode catalysts. It was found that two anode catalysts from Johnson Matthey Co. demonstrated higher anode activity than the two anode catalysts from Tenaka Co. and Cabot Co Between the two Johnson Matthey catalysts, the stabilized DMFC anode catalyst ELE170, demonstrated the best degradation resistance properties among the four anode catalysts investigated. Efforts have also been devoted to determining the effects of contaminants, such as methyl et hyl ketone (MEK), on the oxygen reduction reaction (ORR) in the direct methanol fuel
18 cell cathode reaction. In this study, MEK was proved to negatively affect the ORR at the DMFC cathode. The rotating disk electrode data show ed that the rate of oxygen red uction reaction can decrease by ~15 40% in terms of kinetic current density (corrected for transport limitation). This can translate into a ~15 30 mV overpotential penalty upon introduction of 1 M 1 mM MEK into an acidic electrolyte. This is the first re port on the effects of MEK contamination on the ORR kinetics in the DMFC cathode. P arahyrogen induced polarization has been employed in homogenous processes, as an important technology to enhance NMR signal s. PHIP requires p airwise hydrogen addition, where the two hydrogen atoms come from a single parahydrogen (nuclear spin isomer of hydrogen) molecule When the two hydrogen atoms from parahydrogen are added pairwise and retain their spin correlation, the NMR signal is enhanced up to several orders of magn itude. PHIP provides a unique spectroscopic tool to study pairwise hydrogen addition kinetics In this research, PHIP was used to study propylene hydrogenation over both platinum and i ridium nanoparticles supported on TiO 2 and the PHIP effect was observed for the first time on iridium supported on TiO 2 catalyst. Pairwise hydrogen addition reaction selectivity was found to be temperature dependent for both catalysts.
19 CHAPTER 1 INTRODUCTION 1.1 Overview of Platinum Catalysts P latinum the most expensive of t he widely traded precious metals is of vital importance and interest to the modern evolving chemical industry for catalysis purpose s [ 1 2 ] Because it was f irst discovered as a contaminant in South American gold mines in 1557, the metal was named after the Spanish term platina which is literally translated into "little silver" [ 3 ] Platinum is very resistant to corrosion, high temperature, tarnish and wear, which make it ideal for fine jewelry. Although it was primarily used to adulterate gold in the 16 th to 17 th centuries platinum gained its own sta tus and name as a desirable and decorative metal during the earl y 18 th century in France [ 4 ] From then on platinum became an important precious metal in the jewelry industry, as well as in the storage of wealth. Bearing so many similarities with gold, platinum earned another precious metal commodity. However, the decorative use of platinum metal has not been the driving force in the huge demand for this metal. With so many extraordinary characteristics of its ow n, such as low reactivity remarkable resistance to corrosion even at high temperatures stable electrical properties [ 5 6 ] platinum is now widely utilized for industrial purposes. After the Second Wo rld War, the price and demand for platinum escalated due to its important industrial application in catalysts. Since the I970s, the automotive industry has become the largest demand for platinum. To accommodate the regulations on air quality, catalytic converters on automobile were introduced to reduce the polluting gas in automobile exhaust Platinum i s one of the most widely used catalysts in converters,
20 allowing complete oxidation of low concentrations of unburned hydrocarbons [ 7 ] Platinum catalysts are also used in nuclear reactions, in crude oil refining process, and in many other petrochemical processes [ 8 ] PtO 2 known as Adams' catalyst is used as a hydrogenation catalyst, specifically for vegetable oils [ 9 ] Also, platinum is used extensively in many other applications including laboratory equipment [ 10 ] electrical contacts [ 11 ] and electrodes [ 12 ] In this study, two applications for platinum based catalysts were investigated: 1) Pt based catalyst s in direct met hanol fuel cell research; 2) Pt based catalyst s for the parahydrogen induced polarization reaction. 1.2 Pt based Catalysts in Direct Methanol Fuel Cell Applications As early as 1838, platinum was used by William Robert Grove (1811 1896) in his dev elop ment of an improved crude fuel cell battery [ 13 ] Fuel cells are devices for generating electric power directly from chemical fue ls They s har e many propertie s of batteries silent operation, eliminating moving parts and use of an electrochemical reaction to generate power However, unlike batter ies fuel cells requires no recharging and will run endlessly as long as they are supp lied with fuel and have functional components Of the various kinds of fuel cell s the proto n exchange membrane fuel cell (PEMFC) which typically employs platinum catalysts has attracted the most research interest PEM FC fuel cells c an run on hydrogen or methanols, which are the two most common ly used fuels. Because of their potential to replace the internal combustion engine in vehicles, hydrogen fuel cells are being studied in effort to reduce transportation related air pollution. Direct methanol fuel c ells (DMFC) are especially suitable for applications as portable power source s because they operate at low
21 temperature (typically 50 60 C ) and because methanol is a liquid fuel which is easy to store, transport and handle in small fuel cell device s It ha s been acknowledged that the success of DMFC largely depends on two key component s: the membra ne and the electro catalyst. Pt based catalysts are the most used for both anode and cathode reactions in DMFC designs D ue to their desirable catalytic and elect rical characteristics [ 14 16 ] 1 2 .1 Pt based Catalysts in DMFC Anode Application There has been s teady progress has been made in various aspects of fuel cell rese arch, including catalys ts electrolytes, electrode structure, gas diffusion layer design and engineering effort s in manufacturing. However, the DMFC has been considered the most difficult fuel cell technology due to methanol crossover (methanol fuel from t he anode side through the proton exchange membrane to the cathode catalyst) and slow anode kinetics for methanol oxidation [ 16 ] The latter issue can be overcome on ly by employing better anode catalysts. The two main goals for DMFC anode catalyst research are improved activity and increased durability. To improv e anode activity, decades of extensive research efforts have been devoted to finding better materials for D MFC anodes. U nfortuna tely, it seems that very few electrode materials are suitable for direct methanol oxidation electro catalysis, especially in the acidic DMFC environment and at electrode/membrane interface. Traditionally, the active component in DMFC a node catalysts is platinum. However, employing pure platinum in DMFC would trigger the problem of CO poisoning Adsorbed CO, an intermediate in methanol oxidation, occupies the active Pt catalyst surface sites and cause s low activity. When ruthenium metal is incorporated with platinum the methanol oxidation kinetics are improved considera b ly to a practical level,
22 since the ruthenium help s in the oxidation of CO species in a region of low potential [ 17 ] Other binary Pt based catalysts have been investigated to improve DMFC anode activit y such as P tRu, PtOs, PtSn, PtW and PtMo, B ut platinum/ruthenium alloys showed the best activities and du rability in methanol oxidation, and are thus the most practical anode catalysts for the DMFC. At the current stage of development, non noble metal DMFC anode catalysts are not yet feasible [ 18 ] To improve the durability of Pt based anode catalyst in DMFC, considerable research effort ha s been devoted to understand ing the underlying mechanism causing the performance degradation, and a number of deactivation pathways ha ve been identified leading to Pt Ru anode catalyst degradation. The deterioration of anode catalysts can be due to a decrease in the number of catalyt ically active sites or to the alternating electronic properties caused by metal particle agglomeration [ 19 20 ] poisoning, or oxidation [ 21 ] In addition, r uthenium crossover has been found to contribute to anode performance degradation, as well as to contamination of the cathode catalyst in DMFC s [ 22 ] Ruthenium contamination of the cathode catalyst decrease s the number of available active Pt sites for the oxygen reduction half reaction, and it decrease s the [ 22 ] .P latinum in the PtRu anode catalyst can also be driven to the cathode compartment by the electrical field gradient, which damage s the structure and integrity of anode catalyst [ 23 ] Mitigation of the anode component crossover contamination could be achieved by developing stable PtRu anode catalysts, novel fuel cell designs, or more Ru tolerant cathode catalysts.
23 1 2 2 Pt based Catalysts in DMFC Cathode Application As p rove n by decades of research efforts in search of an ideal cathode catalyst for DMFC purpose s platinum has the highest catalytic activity for oxygen re du ction of any of the pure metals [ 24 ] Platinum supported on conductive carbon serves as the state of art cathode catalyst in low temperature DMFCs. Although platinum catalysts are the best and most commonly employed DMFC cathode catalysts, they still fac e many challenges. One of the significant problems hindering the large scale applic ation of DMFC technology is cathode catalyst performance loss during extended operation. C athode degradation is predominantly sintering and dissolution of platinum based cathode catalysts under high dynamic operation environment [ 25 ] and ruthenium crossover from the anode catalyst [ 22 ] Also, oxidation of the cathode catalyst support (the conductive carbon) [ 26 ] and defect formatio n in proton exchange membranes [ 27 ] are other DMFC cathode performance degradation pathways. Another problem with Pt cathode catalysts is methanol crossover. Meth anol adsorbed on active Pt sites on the cathode catalyst c ould react directly with oxygen, thereby wast ing fuel and result ing in reduced cathode potential, production of additional water, and increase d oxygen consumption. Moreover, as mentioned above, the Pt active surface would be poisoned by carbon monoxide, which is an intermediate product of methanol electro oxidation. The methanol crossover problem can be mitigated by using electrolytes allowing lower methanol permeability, or by employing new cathode catalysts with higher methanol tolerance without sacrific e of oxygen reduction activity [ 24 28 ]
24 In addition many compounds have been identified as poison s to cathode catalyst s including CO, CO 2 NH 3 H 2 S, NO x S O x other sulfur compounds and chloride ions [ 24 29 ] which are either normally present in air or can be introduced by manufacturing process es of various fuel cell components These contaminants should be removed before entering the fuel cell. 1.3 Pt based Catal ysts in Parahydrogen Induced Polarization Applications In n uclear magnetic resonance (NMR) nuclei with nonzero spin absorb and emit electromagnetic radiation when placed in a magnetic field. The frequency of the resonant radiation depends on the magnetic field and the electronic environment surrounding the nuclei. Since first described and measured in molecular beams by Isidor Rabi in 1938 [ 30 ] NMR has found numerous applicatio ns in modern scienc e, including molecular structure elucidation, investigating solids [ 31 ] and advanced medical imaging [ 32 ] However, NMR suffers from inherently low sensitivity due to low po pulation difference between the probed nuclear spin states in a magnetic field. This issue has hindered application of NMR in probing low concentration species. A range of techniques have been developed to increase the NMR signal, including dynamic nuclear polarization [ 33 ] optical pumping [ 34 ] and parahydrogen induced polarization (PHIP) [ 35 ] with the latter being directly related to catalysis. In PHIP parahydrogen, a spin isomer of H 2 with ant iparallel spins ,is chemically introduced into targeted substrate molecule loading to a dramatic NMR signal enhancement by a factor of 10 4 to 10 5 [ 36 ] The nature of PHIP reaction requires addition of two hydrogen atoms from one parahydrogen molecule to a nonsymmetrical substrate molecule [ 37 ] in order to achieve a hyperpolarized NMR signal. Originally, the PHIP reaction was achieved by
25 employing homogeneous hydrogenation reactions, such as those involving the famous 3 ) 3 ) [ 38 ] mplex (Ir(CO)Cl(PPh 3 ) 2 ) [ 39 ] Homogenous reactions can preserv e the correlation between the two proton nuclei of a parahydrogen molecule. However, the homogeneous catalysts us ed in th e r eaction mixture cause separation problem s, precluding further application of the reaction product in novel NMR techniques. Supported heterogeneous metal catalysts, predominantly used in industrial hydrogenation processe s are not expect ed to produce PHIP e ffects, as the reaction is likely to destroy the original correlation of the two nuclear spins. More specifically, molecular hydrogen would easily dissociate on the catalyst surface upon chemisorption, and the adsorbed hydrogen atoms would diffuse over the metal surface, or dissolve into the bulk of the metal, and thus lossing correlation. However, recently Koptyug et al. [ 40 ] PPh 3 ) 3 ) supported on either modified silica gel or polymer could hydrogenat e styrene into ethylbenzene to enhance the observed NMR signal Kovtunov et al. [ 41 ] de monstrated that platinum supported on alumina or titanium oxide catalysts the PHIP reaction The use of heterogeneous catalysts in PHIP reaction is a novel field which has emerg ed over the last decade, with platinum supported on metal oxide catalysts as th e most studied heterogeneous catalyst up to now. The structure sensitivity and the catalyst support effects on the hydrogenation of propylene over Pt catalysts supported on a range of commonly used metal oxide supports have been investigated using the PHIP technique [ 42 ] The study indicated that different types of active Pt catalyst
26 surface sites are responsible for pairwise and non pairwise hydrogen addition to C = C double bond s during propylene hydrogenation. The successful use and the advantag es of pt based heterogeneous catalysts for dramatically NMR signal enhancement has a number of i mportant potential applications, for example in NMR imag ing [ 43 44 ] A nd another important application is a method for a unique type of reaction product labeling as only product g enerated from pairwise addition of parahydrogen can be detected with enhanced NMR signal This kind of labeling technique could be useful in studying the mechanism s of hydrogenation reactions. 1.4 Ir based Catalysts in Parahydrogen Induced Polarization Ap plications Recent research has shown that transition metal catalyst s other than platinum can also demonstrated the PHIP effect using a Au(III) Schiff base complex catalyst immobilized within a metal organic framewo rk material IRMOF 3 [ 45 ] and using Pd(0) nanoparticles embedded in an ionic liquid phase supported on activated carbon fibers [ 46 ] Iridium containing compounds are also commonly used as homogenous catalyst s 3 ) 2 ) [ 39 ] Ir(CO)(dppe)Br and Ir(CO)(dppe)(CN) [ 47 ] have been used in PHIP reaction s as homogenous catalysts. In addition, small clusters of irid ium (Ir4 and Ir6) (as well as of larger aggregates of these metals) on oxide supports (gamma Al 2 O 3 MgO, and La 2 O 3 ) [ 48 ] have shown catalytically activity in t he hydrogenation reactions of propylene. However, to the best of our knowledge, there is no reported study on PHIP reaction s involving heterogeneous Iridium catalysts.
27 1.5 Objective and Motivations In the present study, different Pt based catalysts were investigat ed for use in two applications: 1) as anode and cathode catalysts in direct methanol fuel cell applications; and 2) as supported metal particle catalysts in parahydrogen induced hyperpolarization applications. In addition, one other transition metal in platinum family, iridium, was inve stigated as a supported catalyst on titanium oxide in PHIP applications. 1.5.1 Commercial Anode Catalyst Screen Study f or A n Open Cathode Design DMFC The first part of this study focus ed on evaluating the performance and degradation of four commercial ano de catalysts in an open cathode design DMFC single stack T echnical considerations for anode catalyst choice for the open cathode design DMFC were proposed based on the research result. 1.5.2 Investigation o f Contamination Effect s of Methyl Ethyl Ketone o n A Platinum /Carbon E lectrocatalyst f or Oxygen Reduction Activity The second focus in this work is understanding to what extent the presence of one type of wetting agent m ethyl ethyl ketone, would influence the oxygen reduction reaction and thus the perfo rmance of a typical Pt/Carbon DMFC cathode catalyst. The thin film rotating disk electrode (RDE) method was employed to study the ORR kinetics of a DMFC cathode Pt/Carbon catalyst upon MEK contamination. The results from this study provide important knowle dge of the effect s of trace MEK contamination on DMFC cathode s leading to suggest ion to mitigat e performance l osses and degradation phenomena
28 1.5.3 Platinum a nd Iridium based Catalysts i n Parahydrogen Introduced Hyperpolarization Applications P latinum and iridium supported on titanium (IV) oxide were used as model heterogeneous catalyst s to study the kinetics of heterogeneous propylene hydrogenation with parahydrogen. The temperature dependence of NMR signal enhancement and the pairwise hydrogen additi on reaction s were investigated over the Pt/TiO 2 and Ir/TiO 2 catalyst s
29 CHAPTER 2 COMPARATIVE STUDY OF COMMERCIAL ANODE CATALYSTS IN AN OPEN CATHODE DIRECT METHANOL FUEL CELL 2.1 Background Since the Welsh Physicist William Grove developed the first crude fuel cells in 1839 [ 13 ] fuel cells h ave attracted substantial attention from academia and industry all across the world for their great potential as an economic, efficient, and environmentally friendly means to convert chemical energy into electric energy. Fuel cells can efficiently harvest energy that is preserved in chemical bonds and convert it into electrical energy. C onventional power plant s or other heat engine based technologies introduce several steps to convert chemical energy to electricity and comes with the efficiency limitation established by Carnot In contrast, fuel cells do the conversion in one step, giving a higher energy conversion efficiency than any other known device For the same power output, fuel cells using propane or methanol emit less carbon dioxide compare to trad itional heat engines using fossil fuel s and no greenhouse gas es are emitted when hydrogen is used [ 49 ] In addition efficiency is independent of unit size, and can be scaled down from very large to very small units without efficiency loss 2.1.1 Types of Fuel Cell Fuel cells can be classified into many different types according to cat alyst, material separating anode and cathode, and type of fuel utilized. There are five types of fuel cell s most commonly studied : a lkaline fuel cell (AFC) p hosphoric a cid fuel cell (PAFC), m olten c arbonate fuel cell (MCFC) and solid o xide fuel cell (SOF C) and p olymer electrolyte m embrane fuel cell (PEM FC ) [ 50 ]
30 The a lkaline fuel cell (AFC) which can achieve efficienc ie s as high as 70%. AFC use the aqueous potassium hydroxide as electrolyte, to accelerate cathode reaction kinetics and platinum or non precious metal s su ch as nickel, as catalysts, for a power output s of 300 W 5 kW. Two noticeable limitation s of AFC s a re that high purity hydrogen and oxygen are needed as feeds, and the tolerance for atmospheric carbon dioxide is poor, as CO 2 would react with the KOH electrolyte. Due to the se limitations, the application s of AFC s are very restricted [ 51 54 ] The p hosphoric a cid fuel cell (PAFC) uses phosphoric acid as electrolyte, typically operating at temperature s ranging from 180 210 C with power output s between 50 and 200 kW The efficiency for PAFC itself can reach about 40%, and can be increased to about 60 70% if us ed in a combined heat and power system. Due to the acid nature of the electrolyte, PAFC s overcome the CO 2 tolerance problem in AFC s thereby broaden ing its fuel feed selections. However, the carbon monoxide concentration must be maintained below 0.5% in the fue l feed, to avoid catalyst poisoning The acid electrolyte also cause s corrosion problem s for fuel cell components. Because pt based catalysts must be emplo yed in PAFC s large scale application s are limited due to the high cost of platinum usage [ 51 54 ] M olten c arbonate fuel cells use l ithium carbonate sodium carbo nate or/a nd potassium carbonate solution as the electrolyte, and can achieve an efficiency as high as 60 85%, while operating at 620 660 C High operation temperature gives the advantage of higher energy efficiency and more choice of fuel/catalyst options. MCFC s c an use hydrogen, carbon mo noxide, natural gas, propane or diesel as fuel s reaching a
31 power output range of 10 kW 2 MW. MCFC s are mainly used as stationary power generator s One disadvantage of MCFC s is severe corrosion to fuel cell components due to the high temperature [ 51 54 ] The s olid o xide fuel cell (SOFC) is promising in large stationary power generator applications, with outputs up to 2 MW SOFC s u se a solid oxide as electrolyte, and operate at very high temperatures (500 1 000 C ) with high efficiency (60 85%). SOFC s use non precious metal catalyst s including a Co or Ni based anode catalyst, and a Sr promoted LaMnO 3 cathode catalyst. High operati on temperature enables SOFC to use a large range of different fuels, including light hydrocarbon s, such as methane, propane and butane However, high operation temperature which cause s corrosion and mechanical problem s for various SOFC components, i s one of the major disadvantages of this type of fuel cell [ 51 54 ] P olymer electrolyte m embrane fuel cell s are characterized by use of a proton conducting polymer base d membrane as electrolyte. Depending on fuel types PEMFC s can be further classified into hydrogen fuel cell s and direct methanol fuel cell s For the hydrogen fuel cell, the an ode and cathode reaction s are Anode reaction: (2 1) Cathode reaction: In the hydrogen fuel cell (HFC), hydrogen and oxygen f low through the gas channels on bipolar plates, and reach the anode/cathode sites respec tively. The anode catalyst break s dow n hydrogen molecul es into protons and electrons. The protons exist in the form of hydrates and are transported into the cathode side through the proton exchange membrane. To fulfill the proton conduction process, at the same time the
32 electrons from hydrog en pass through the external circuit and reach the cathode side where they reduce oxygen to oxide. The proton and reduced oxygen react at the cathode catalyst surface, and produce water at the same time. The HFC typically operate s at temperature s of 60 8 0 C making them convenient in vehicle and portable power device applications due to simplification of heat management. The theoretical efficiency of the combined heat and power system for an H FC can reach approximately 90%. However, due to the limitation s of all types of polarization processes, the typical realistic efficiency for HFC is about 60% in transportation application s and around 35% in portable power supply application. The typical power output for HFC ranges from a few watts to 100 kW. The HFC demonstrate s the highest power density (300 1000 mW cm 2 ) among all types of fuel cells and c an be turned on/off rapidly, making it very suitable for portable power supply and transportation applications. However, the HFC relies on expensive noble metal c atalysts, requires an accessory system for hydrogen storage and water management, and has limited tolerance t o carbon monoxide and sulfur contamination In particular, the necessary accessories for hydrogen st orage and water management make it very difficu lt to minimize HFC size to adapt to the requirement for portable power supply applications [ 51 54 ] The d irect methanol fuel cell (DMFC) is another type of PEMFC that uses methanol as liquid fuel, and it especially suitable for developing portable power source devices. Unlike hy drogen, methanol can be stored as a liquid, which is reasonably safe and convenient for storage, transporting, and filling into fuel cell d evices. In the DMFC, methanol is supplied to the anode side, where it is catalytically broken down into proton s electron s and carbon dioxide. As discussed above for the HFC, proton and
33 electrons travel to the cathode where they with oxyge n and form water. The reactions at anode and cathode are : Anode reaction: (2 2 ) Cathode reaction: Another advantage is that methanol i s a mature chemical product, which can be easily produced from coal, gasoline, or biomass. Moreover, methanol has four times the volume energy density compared to hydrogen gas at a pressure of approximately 69 MPa in a state of the art containment system un der development [ 55 ] The shortcomings associated with DMFC are methanol oxidation intermediate species which can possibly poison the electrode [ 56 ] slow met hanol oxidation reaction kinetics on the anode catalyst [ 57 ] and methanol crossover problem [ 28 ] The methanol crossover costs significant power loss an d fuel efficiency loss es Another disadvantage of DMFC is water management. In most conventional design s, the water produc ed in the cathode reaction is condensed and recycled by a dedicated water recycle system downstream from the cathode and t he recycled water is returned to the anode feed. The conventional design of water management system requires a bulk y accessory water cycle system and the associated extra power consumption. Also, if the water recycling system is not functioning well, water overflow o n the cathode side can occurs flooding the cathode catalyst layer pores, and inhibit ing oxygen transport from the cathode gas diffusion layer to the cathode catalyst layer. Poor water management not only decrease s the net power output and energy density o f DMFC, but also possibly shut s down the cathode reaction by water flooding.
34 Despite the disadvantages, DMFC is a very promising candidate for overcoming the difficulty of hydrogen storage and transportation in PEMFC large scale applications, especially fo r portable power supply purpose s 2.1.2 Anode Catalyst i n DMFC Platinum ruthenium based catalysts are the most feasible and prevalent anode catalysts currently used in DMFC system s due to their high activity in the methanol oxidation reaction and their hi gh carbon monoxide tolerance [ 16 18 ] The ruthenium in the alloy catalyst has two functions: 1) in teraction with th e d orbital electrons of platinum to help prevent Pt poisoning by CO ; and 2) formatting active oxygen species on the ruthenium surface, thereby facilitat ing oxidation of intermediate s formed during methanol oxidation, further improv ing methanol oxidation activity and catalyst poison tolerance. An atomic ratio of Pt to Ru of 1:1 h as been considered to be the best composition for the most CO tolerant DMFC anode catalyst [ 58 ] For better utilization of platinum, t he optimum size of the Pt Ru catalyst particles in a DMFC is o n the order of nanometer s in the range of 2 6 nm [ 16 59 ] Due to the slow anode reaction kinetics, the catalyst loading typically ranges between 2 to 8 mg cm 2 in order to maintain the desired performance and life span. I mprovement s in anode DMFC catalysts is m ai nly directed into two pathways: improv ed performance and decreased cost. Improvement of DMFC performance can be further specified as improv ed activity, reliability and/or durability. One known pathway for anode catalyst degradation is particle sintering. For nano particle anode catalyst s th e specific surface energy increase s as the particle size decrease s [ 60 ] thus, the smaller particles are intrinsically more eas ily to agglomerate d /sinter ed [ 61 ] Once the anode catalyst particles sinter into a larger size, the electrochemical of active surface
35 area (EASA) decrease s accordi ngly, and the performance of the DMFC also degrade s as a result To overcome the sintering effect, improv ement in the support for stabilizing nano particle catalysts is one promising research direction. To decrease the anode catalyst cost, development of n on noble metal anode catalysts co uld be the most promising route. Unfortunately, PtRu based anode catalysts are currently the only feasible option s in practice. In order to lower anode catalyst cost, better PtRu form ulation and support design are other met hods to improve precious metal utilization 2.1.2 Objective s DMFC is a very promising technology for portable power suppl ies However, there are many problems that need to be tackled before its commercialization. Currently, t he main problem with DMFC s, is the need to increase anode catalyst activity/durability, decrease methanol crossover and improve water management. Figures 2 1 and 2 2 show closed cathode and open cathode DMFC stack designs, respectively For the closed cathode design, the water condens er, water storage device, and water pump are us ed to recycle the water to the anode feed This bulky water recycle system increase s the overall DMFC weight and volume, and decreases the net power output by consuming energy in the water recycling effort. Ho wever, an open cathode design removes bulky water recycle system. The open cathode design of DMFC stacks is achieved by addi ng a hydrophobic liquid barrier layer (LBL) in the membrane electrolyte assembly ( MEA ) between the cathode catalyst layer (CCL) and the cathode gas diffusion layer (CGDL). The LBL allows passive recovery of the water produced in cathode reaction back to the anode side through the membrane The present research focus on : 1) test ing the performance and degradation of four commercial ano de catalysts in the open cathode design DMFC single stack ; 2)
36 investigating the effect of anode choice on cathode performance and degradation in the open cathode design DMFC single stack ; 3) mak ing technical suggestions for anode catalyst choice for the op en cathode design DMFC based on research data. 2.2 Experimental Methods 2.2.1 Membrane Exchange Assembly Configuration DMFC single cell membrane exchange assembly used in this work has an active cell area of 15.5 cm 2 The membrane electrolyte assemblies ( M EA s) were m anufactured by our collaborating research group at the University of North Florida. Figure 2 3 shows the configuration of the DMFC MEA design. The MEA consists of nine different functional layers, which incorporate the unique DMFC open cathode d esign allowing advanced water management. Each layer in the MEA is responsible for its dedicated task. The b acking plate, anode flow channel (AFC) would provide s necessary mechanical support from the anode side It also create s an optimized anode flow cha nnel to e nsure an adequate and stable anode fuel supply throughout the adjacent anode catalyst layer in a gas liquid two phase flow condition. The a node gas diffusion layer (AGDL) is designed to allow the carbon dioxide generated in anode reaction to diffu se back to the AFC and flow out of the DMFC system. Anode catalysts are evenly deposited onto the a node catalyst layer (ACL) is to catalytically facilitate the anode reaction. Membrane (MEM) is responsible for conducting proton flow from the anode side t o the cathode side and inhibiting electron transfer thr ough the membrane .The c athode catalyst layer (CCL) is next to the membrane, and is evenly embedded with cathode catalyst particle s where the cathode reduc tion reaction takes place. The l iquid barrier layer (LBL) is a unique functional design dedicated to reduce the water loss from
37 the cathode reaction to the open atmosphere. The hydrophobic nature of LBL retain s most of water molecules generated in the cathode reaction, and force s them to diffuse back to the anode through the small channels in membrane. The c athode gas diffusion layer (CGDL) contains diffusion channels for air from the cathode flow channels getting into the CCL. The b ipolar plate is place d between the two MEAs, with one side anode flow channel and one side for the cathode flow channel, thereby enabling the connection of multiple MEAs in series to form a DMFC stack. The c athode flow channel (CFC) is similar in function to the AFC to optimize the air flow and transport from the CFC into th e CCL. In this study, four commercial DMFC anode catalysts were used to manufacture the MEAs. ELE147 (referred to as ELE147), and ELE170 (referred to as ELE170) from the Johnson Matthey Company, a DMFC anode catalyst from theTanaka Company (referred to as Tanaka), and a DMFC anode catalyst from the Cabot Company (referred to as Cabot). The cathode catalysts in this study were the same for all MEAs : Hispect 11100 from Johnson Matthey Company. 2.2.2 Fuel Cell Testing Experimental Setup The DMFC sing le cell test rig consists of a thermocouple electric current collector plates, a closed container for sealing the cathode inlet, a peristaltic pump for the anode fuel solution supply, and a heating element for achieving d esired temperature. The thermocouples were positioned at the anode flow stack outlet which thermostatically controlled the stack to an accuracy of 0.1C through a heating patch adhered to the outside of the current collector. A commercial electronic load regulated the output current and voltage o f the DMFC single cell test rig, and the mass flow controllers were employed to control gas flow rates through the cathode/anode The
38 data were acquired in real time by a customized LabVIEW code and a data acquisition system from National Instruments. All MEAs were received from the University of North Florida and conditioned at the University of Florida before testing. During MEA single cell assembling, the MEA was held and sealed tightly via t wo gaskets screw ed on to the graphite plates engraved as either anode flow channel or cathode flow channel. Two gold plated brass plates were pressed in close contact with the outside surface of the graphite plates for collecting the electric current. 2.2.3 Electrochemical Characterization Experimental Setup Electroch emical characterization of the single DMFC MEA was performed using the commercial electronic load, a Solartron 1260 i mpedance/gain phase a nalyzer ( Solartron Electronic Group UK), and a Solartron 4480 p otentiostat a nalyzer ( Solartron Electronic Group UK). 220.127.116.11 MEA initial performance measurement Each MEA was subjected to a break in process for activating and conditioning the MEA into proper stable working condition. The anode channel was first purged with hydrogen to remove any air in the channel ; then hydrogen was pass ed through the anode channel to activ ate the catalyst surface. Thereafter, nitrogen was again purged through the anode side for removal of any residual hydrogen, before a 0.8 M methanol solution was pass ed through the anode side for MEA co nditioning purpose s At the same time, the cathode was continuously purged with nitrogen to prevent any crossover reaction on the cathode side. Then the tempera ture of the MEA was gradually increased to 80 C in order to hydrate the membrane properly, whil e the anode was provided with 0.8 M methanol solution and the cathode was under nitrogen protection.
39 drop from 0.8 V to 0.35 V at a rate of 10 mV/min, with the single c ell voltage would b e held at 0.35 V for 10 minutes. This was followed by a catalyst rejuvenation procedure to reverse degradation effects on the cathode catal yst side. The cell temperature wa s controlled at 50 C and anode flow was 2 mL /min with a 0.8 M aq ueous solution o f methanol, while cathode flow wa was cycled for 8 hours. The cell current density at 0.35 V was monitored by a Labview controlled fuel cell test station during the entire cycling proc ess to measur e the continuous performance of the MEA at the beginning of its life time 2.2.3. 2 S ingle cell overall polarization measurement For the s i ngle DMFC MEA polarization curve test, a solution of 0.8 mol/L methanol was prepared and pumped into the a node flow inlet at a constant flow rate of 2 mL/min. Compressed air (Airgas UAS, LLC) at 1 standard liter per minute (SLPM) was purged through the cathode. A polarization curve for the fuel cell c an be characterized by three major polarization regions, kn own as the activation, ohmic, and concentration over potential [ 62 ] In the low current density region, the activation overpotential of the DMFC dominated the polarization cur ves. At the medium current density region, the cell voltage show s a linear loss with the increas ing current density attribute d to the oh mic overpotential The high current density region concentration overpotential is observed At a given current density, a higher cell voltage indicate s higher cell performance and power output. In this study, polarization curves were measured to understand the cell performance. The current density was controlled at a specified value, from 0 mA/cm 2 to the maximum 150 mA/cm 2 which still maintained the MEA voltage above 0.2 V to prevent damage to the MEA.
40 2.2.3. 3 A node polarization measurement The a node polarization curve is a plot of current density versus anode potential. It is important to resolve anode polarization from overall polarization in order to assess the performance of the anode individually. It is noteworthy that the potentials obtained in anode polarization experiment are total voltage s consist ing the anode potential and the resistance loss. A n IR (resistance loss) corrected anode potential is calculated according to the equation 2 3 : (2 3 ) in which E total is the potential obtained directly, while purging hydrogen through the cathode to set the cathode potential to z ero and measuring the potential at various current densities. IR is the resistance loss, which is calculated by multiply ing the current by the membrane resistance. V corrected is an actual anode polarization potential after r esistance loss compensation. Th e anode polarization meas urements were performed by a Solartron 1480 MultiStat. H ydrogen was pass ed through the cathode side of the single cell at the reference electrode and 0.8M methanol solution was pass ed through the anode side where it was connected to the working electrode. The applied voltage on the working electrode was set in a range from 0.01 V to 0.55 V with a scanning rate of 2 mV/s. The limitation of the maximum current was set a t 4 A for electrical safety protection 2.2.3. 4 C athode polariza tion measurement T he cathode polarization potential E cathode can also be obtained by simple calculation as shown in equation 2 4 :
41 (2 4) The overall cell potential V cell is obtained from overall polarization curve ; IR is the resistance loss ; and E anode is the anode corrected v oltage as described above 18.104.22.168 S ingle cell re sistance measurement Impedance spectroscopy was employed to measure the resistance of the single DMFC cell. Impedance spectroscopy impos es a small sinusoidal ( AC ) voltage or current signal of known amplitude and frequency, and the ratio and phase relation of the AC voltage and current signal response corresponding to the complex impedance, Z(w) This measurement w as performed using a Solartron 1260 i mpedance/gain phase analyzer (Solartron Electronic Group, UK), while pass ing h ydrogen through cathode at a r ate of 0.3 SLPM for set ting a potential baseline, and passing 0.8 M met hanol solution at a rate of 2 mL /min through the anode side. The temperature of DMFC was maintained at 50 C and the test frequency rang e from 10 kHz to 0.1 Hz with AC amplitude of 20 mV as the parameter for resistance measurements A plot of Z (w) vs Z (w) was prepared and the cell part of Z(w)), when the phase angle is zero ( =0 ), and therefore the cell behaves in a purely resist ive manner 2.2.3. 6 M ethanol crossover measurement One intrinsic and almost inevitable prob lem associated with DMFC is the methanol crossover in which the methanol solution would diffuse s f rom the anode to the cathode side. Methanol crossover in DMFC operation cause s low er power output due to chemical oxidation of methanol at the cathode catalyst, giv ing rise to several
42 problems: (1) electrode depolarization; (2) m ixed potential s resulting in a lower open circuit voltage (OCV) ; ( 3 ) cathode catalyst poisoning by CO (an intermediate of methanol oxidation); an d (4) water flooding on the cathode, impair ing the oxygen transport to the cathode catalyst layer and damag ing the cathode reaction. Mor eover the overall fuel utilization efficiency of the fuel cell is lowered when there is excessive methanol crossover [ 28 ] In this study, crossover methanol was assumed to react completely at the cathode side N itrogen was used to purge the fuel cell cathode at 1.00 SLPM, while 0.8M methanol solution passed through the fuel cell anode at 2.00 mL/ min A Solartron 4480 potentiostat analyzer (Solartron Electronic Gr oup, UK) controlled by CorrWare software from Scribner Associates was used to perform a linear sweep voltammetry (LSV) between 0 and 0.8 V The resulting plot of working electrode current vs working electrode potential was used to derive the methanol cro ssover flu x from Faraday's Law. 22.214.171.124 A ccelerated low voltage anode durability test I nvestigation of the DMFC durability is a time consuming and potentially ineffi cient task [ 63 ] simply because the targeted design lifetime for our DMFC is more than 3000 hours. Studying the durability of a DMFC anode catalyst under normal operatin g condition s is very difficult if not possible, due to experimental time span. Therefo re, a more realistic experimental method for probing anode catalyst durability is required to fulfill the task. In this study, a low voltage cycling technique was developed to purposely create a harsh environment for the anode catalyst and induce an accel erated degradation, a so called l ow v oltage accelerated d urability test (LVADT) The degradation resistance ability of DMFC was measure d by operating a designed 144 hour long LVADT During the single cell operation, methanol solution
43 (0.3M, 1.25mL /min) was delivered to anode si d e at room temperature and atmospheric pressure via a p eristaltic pump, while purified air was fed to the cathode site at a flow rate of 2 SLPM. During the experiment, the gas flow, cell temperature and load density were carefully con trolled by a NI LABVIEW program connect ed to the Fuel Cell Test Station. The overall voltage of single cell stack was controlled to drop from 0.8 V to 0.05 V at a rate of 20 mV/min. Then the single cell voltage was held at 0.05 V for 10 minutes, followed b y a catalyst rejuvenation procedure to reverse degradation effects on the cathode catalyst side. Moreover, the system was checked for gross leaks every time the MEA was changed, followd by b reak in cycling to ensure new MEA was operated in normal Before starting 144 hours LVADT test, 8 h ours of normal cycling was performed to observe the initial performance of each MEA, which is very important when evaluat ing qualitie s of different kinds of MEA. A series of electrochemical measurement s including impedanc e measurement, methanol crossover measurement, anode /overall polarization measurement were obtained at intervals of 8 h ours in normal cycling, and after 48 h ours 96 h ours 144 h ours of low voltage cycling in order to obtain performance and degradation information of the single fuel cell 2.2. 3 8 Calculation of e lectrochemical ly active surface area calculation To investigate the change in the EASA of the anode catalysts the EASA was calculated by a modified Bulter Volmer equation [ 64 ] during the different periods of LVADT. A general starting point for the development of a relationship between an anode polarization and anode catalyst active surface area for a partic ular current density is the Butler Volmer equation (Equation 2 5).
44 (2 5 ) where I is the actual anode reaction current density ; A is the electrode active surface area; i 0 represents the e xchange current density on the catalyst surface, which is an intrinsic parameter for a catalyst, defined as the value of the current density at zero net current where the reaction is at an equilibrium ; is the symmetry factor which is a constant parameter for anode electrochemical reaction with a value between 0 and 1 ; is the anode overpotential which can be calculated by subtracting the actual anode potential E by theoretical anode equilibrium potential E o ; F stands for the F arad ay constant ; R is the idea gas constant and T is the operation al temperature to Equation 2 5 was applied since a current density of 120 mA cm 2 was chosen for catalyst active surface area calculation, where the pol arization is considered to be high enough for adapting the Since in the condition, is large Equation 2 5 can be simplified to (2 6) For the identical MEA before and after LVADT, the current density of 120 mA cm 2 was cho sen for calculation, and was substit uted into Equation 2 6 : (2 7)
45 w here the current density I is identical for fresh and degraded MEA. Consider ing the overpotential definition for fresh a nd degraded anodes: ; Equation 2 7 can be rewritten as (2 8) Therefore, Equation 2 8 can be use d to calculate relative electrode active surface area regarding fresh and degraded MEAs. In this study, all relative electrode active surface area s of MEAs with different anode s w ere normalized to the percentage of the beginning life electrode active surfa ce area of ELE147 anode MEA, which were arbitrarily set to 1. All electrochemical measurements and LVADT described in this section were performed on two MEAs for each type of anode catalysts independently, and same trend during degradation tests were obse rved. One set of data obtained from one MEA for each type of anode catalysts were presented as result. 2.3 Results and Discussion 2.3 1 Initial Performance of MEAs Before carrying out accelerated the voltage anode durability test, an 8 hour performance tes t was used to measure the initial performance of each MEA with different anodes. The MEAs initial performance s w ere measured by monitoring the current density of the MEA at 0.35 V. Fig 2 4 shows the initial performance of MEAs with J ohnson M atthey ELE147, JM ELE170, Tanaka, Cabot anodes. From Fig 2 4, all MEAs demonstrate d higher performance at the very beginning of life, and most of the
46 performance degradation happened within the first 60 minutes, after which the performance plateaued at a relatively stab ilized value. This is observ ed be cause at the very beginning li fe of MEA, various components are not conditioned or stabilized in actual working condition s and a significant performance loss occurs due to the adaption of the MEA to working condition s For example, during conditioning potential hold and rejuvenation cycling, any loosely deposited Pt or Ru catalyst from for anode may detach from the catalyst support, and Ru could migrate across the membrane and deposit on the cathode side, thus caus ing perfo rmance degradation [ 22 ] Figure 2 5 (a) shows the initial power dens ities of the various MEAs and the stabilized power density value observed during the initial 8 hour performance test. The initial power density is defined as the average power density over the first ten minutes of operation right after the is defined as the average power density across the plateau region as shown in Fig 2 4. Initial power density loss in Fig 2 5 (b) corresponds to the loss bet ween initial power density and stabilized power density. Tanaka anode MEA demonstrated the lar gest initial power density loss: 7.25 mW cm 2 ELE147 and Cabot anode MEAs als o showed significant power loss: 5.11 mW cm 2 and 3.07 mW cm 2 respectively. Howev er, ELE170 anode MEA appeared to be the most stable MEA for initial performance loss, with a loss of only 0.34 mW cm 2 This study indicates that ELE170 is the most stable anode during the first few hours of fuel cell operation and conditioning, which m ay due to its having the least unstable components in anode catalyst composition. Tanaka and ELE147 all showed significant initial performance degradation during the first hour at the very beginning
47 operating life implying that those two anode catalysts con tain relatively unstable anode catalyst s during the beginning life conditioning process. 2.3 2 Single Cell Polarization and Power Density Curves Figure 2 6 shows the current density voltage polarization curves and power density vo ltage curves recorded unde r current control, at various periods of potential cycling. The data demonstrate that the loss in fuel cell performance throughout the polarization curve occurred during the entire time range investigated in LVADT. From Table 3 1, the ELE147 DMFC started w ith the best fresh DMFC performance, with a power density of 49.16 m W cm 2 at a cm 2 which is about 10 % higher than performance of other three DMFCs. ELE170, Cabot and Tanaka started with almost the same initial performance, a power density around 40 m W cm 2 However, at the end of 144 hours LVADT, and ELE 147 showed almost the same, or slightly worse performance compared to ELE170 both around 40 m W cm 2 Tanaka DMFC and Cabot DMFC showed a power density of 38.36 m W cm 2 and 35.97 m W cm 2 respectively, which wa s much lower than the ELE147 and EL E170 end of life performance. ELE147 demonstrated the most rapid performance loss during entire course of 144 hours LVADT (Table 3 2), which is 66.32 W cm 2 h 1 The ELE170 experienced the least power loss during LVADT, correspond ing to a power density l oss rate 24.95 W cm 2 h 1 for the entire time range. Cabot and Tanaka anode DMFCs showed moderate performance degradation rate s 49.90 W cm 2 h 1 and 41.59 W cm 2 h 1 respectively. From the overall polarization and power density curves, ELE170 DMFC s tands out as the candidate with best degradation resistance, and thus the best end of life performance among all the DMFCs. ELE147 st arted with the highest initial of life
48 performance, but had the most rapid performance degradation rate among all four anod es investigated. Tanaka and Cabot anodes DMFCs both started with a moderate initial life performance, but with a significant performance degradation rate of 40 50 W cm 2 h 1 they ended up with the worst performance at the end of the LVADT. 2.3 3 Anode Polarization The anode polarization curves of the DMFC s with four anodes after different LVADT times are shown in Fig. 2 7 With durability test time increasing, the anode potential at the same current density increase d to a larger positive value i ndicat ing that the activity of the anode catalyst s ha d degraded for all four DMFCs Table 2 3 shows the anode potential evolution during the durability test time frame at a curre nt density of cm 2 Tanaka and ELE170 anode demonstrated very similar degradation pattern s 1 h 1 degradation rate s across entire 144 hours durability test, wh ile ELE147 anode showed the largest degradation rate of 0.14 1 and Cabot anode showed the least degrad ation 1 2.3 4 Relative Electrochemical Active Surface Area In order to better access the anode catalyst degradation, relative electrochemical active surface area s (rEASA) w ere calculated for anode activity across the LVADT. Fig 2 8 shows rEASA s normalized to the fresh ELE147 anode EASA, all of anodes during durability test. The rEASA of all anodes decrease d during the durability test, until ELE147 anode show ing the largest rESAS both at fresh condition s and at end of durability test condition s Across the entire durability test, the rEASA value decrease d at the order of ELE147>ELE170>Tanaka>Cabot, which again confirmed that ELE147 has the best anode active surface area, as well as the bes t anode acti vity. However, the rEASA degradation p ercentage s decreased in the order of ELE147 > T anaka = ELE170
49 > Cabot, showing 33.0%, 29.9%,29.9% and 18.8% loss of beginning life rEASA for each anode respectively. It is generally accepted that the elec trochemical active surface area for DMFC anode catalyst s is an important parameter for anode performance [ 65 ] as larger EASA indicates that more catalyst sites are available for anode reaction. Previous research has shown that the performance degradation of PEMFC is largely due t o the E A SA loss of the electrode catalysts [ 66 ] Different mechanisms were discovered to contribute to the EASA loss, including the agglomeration of Pt Ru catalyst particl es, dissolution/leaching of ruthenium species from the anode catalyst, detachment of catalyst particles from the carbon support, and etching of the carbon support itself [ 25 65 ] T he degradation mechanism for Pt based DMFC catalysts was also found to depend on various working conditions and the dissolution /leaching of ruthenium mechanism m ay be more prevalent duri ng load cycling In this study, ELE147 was found to experience the largest rEASA loss, indicat ing that ELE147 m ay have the largest ruthenium dissolution/leaching from the anode and this may also contribute to 2.3 5 Methanol Crossover Methanol crossover was measured by linear sweep voltammetry Me thanol crossover current densities for all four DMFC s across the entire LVADT are shown in Fig 2 9. It was found that in all time periods sampled during the durability te st for all DMFCs, methanol crossover current density always stay ed around 50 mA cm 2 indicating that no noticeable de gradation in the proton exchange membrane occurred during the course of LVADT Thus, we can safely assume that methanol crosso ver effect i s same during entire time range of LVADT for all DMFCs.
50 2.3 6 Cathode Polarization In order to investigate the cathode performance during LVADT, the relationship between the cathode potential and the current density for all DMFCs were compared As shown i n Fig 2 10, the cathode potential of all DMFCs decreased at the same current density as the LVADT time increased which is evidence for cathode degradation occurrence. At the current density of 120 mA cm 2 ,ELE147 and Tanaka DMFC started with higher cathode potential s than ELE170 and Cabot However, ELE170 DMFC stands out with the best cathode performance at the end of the entire LVADT, while Cabot DMFC show ed the lowest end of test cathode performance. Considering the cathode degrad ation rate, ELE147 DMFC showed the greatest 1 while Cabot and Tanaka DMFC degraded at a moderate rate of 1 1 r espectively. Again, ELE170 DMFC showed a lmost no cathode degradation for the entire durability test. Considering that iden tical cathode s w ere utilized in all four DMFCs with different anodes, the difference behavior of cathode degradation most likely was rooted in the interaction with other components in the DMFC, especiallythe anode. There is some known mechanism for anode c athode interaction could affect cathode degradation significantly, such as ruthenium crossover. It has been reported that the amount of oxidized Ru species generated by the oxidation of crossed Ru is closely correlated to the performance degradation [ 23 ] The observation of significant cathode degradation for ELE170 m ay be largely due to ruthenium crossover from the anode to the cathode side.
51 2.3 7 MEA Resistance Single MEA resistance of four anodes DMFC was monitored by impedance spectrum test throughout durability test (Fig2 12). At all time periods sampled during the durability test, MEA resistance for ELE147, ELE170 and Cabot DMFC were relatively constant, and stay ed around 0.3 cm 2 These values agreed with the methanol crossover measurements further indicating that no noticeable degradation in the proton exchange membrane occurred during the course of LVADT. Tanaka DMFC showed a relatively larger change in impedance, from 0.3 cm 2 to 0.5 cm 2 The common mechanism for membrane degradation includes mechanical failure, thermal degradation, chemical radical induced degradation and trace metal induced degradation [ 27 ] For Tanaka DMFC, the reason for relatively noticeable change in resistance is still not clear. 2.4 Summary The anode electrode durability is an important factor determining the performance and degradat ion of DMFC. In this study, we examined four commercial anode catalysts Johnson Matthey ELE147, ELE170, Cabot and Tanaka anode DMFC catalyst in one unique DMFC configuration with an open cathode design. One low voltage accelerated durability test method, employing a low voltage cycling and rejuvenation technique, was developed to evaluate the durability of DMFC anodes. For better comparison of before test and post test performance of anodes, the DMFC overall power output, cathode and anode potential s are listed in Table 2 5, and degrad ation rate data for various fuel cell components are summarized in Table 2 6. ELE147 DMFC has the best fresh anode performance, cathode performance and power output (Table 2 5). However, ELE147 also demonstrates the largest output degrad ation
52 rate, anode degradation rate and cathode degradation rate The post test overall power output of ELE147 DMFC decreased by about 20%, which was slighter lower than post test power output ELE170 DMFC. ELE170 DMFC showed excellent durabilit y among all four DMFC anode catalysts, in terms of initial power loss rate, overall power output degradation rate, anode degradation rate and cathode degradation rate for the entire durability test time range (Table 2 6). The excellent degradation resistan ce indicates that the ELE 170 anode is the most stable among the four anode catalysts investigated. Cabot and Tanaka DMFC s both started with a moderate power output similar to that of ELE170 DMF C, but they both degraded more rapidly than ELE170 in terms o f overall power output, and cathode degradation rat e. Cabot and Tanaka both ended with low post test power output. An e xtra stable anode would not only possess very reliable anode activity/ durability, but also demonstrate very positive effect s on the dur ability performance on the other components in the DMFC. For instance, the stab i l ity of ELE170 m ay contribute to the very small cathode degradation rate for ELE170 DMFC, since one of the most known cathode degradation mechanism, ruthenium crossover, would be greatly reduced if the anode is stable. The above results demonstrate that ELE170 is the best anode candidate among the four anode s investigated for the DMFC configuration with the un ique open cathode design, with a stable anode performance an d overall durability when it is incorporated into a DMFC. Based on data obtained from this study, ELE170 was recommended to be
53 the choice for anode catalyst to be used in the open cathode design direct methanol fuel cell.
54 Figure 2 1. Conventional design of a close cathode DMFC system with its water recycling system (Courtesy of Dr.James Fletcher University of North Florida).
55 Figure 2 2 Simplified design of a n open cathode DMFC system with liquid barrier layer. (Co urtesy of Dr.James Fletcher University of North Florida).
56 Figure 2 3. S chematic of a DMFC stack.
57 Figure 2 4. Initial durability test of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes Operating condition: cell temperature : 5 0 C, cell voltage 0.35 V Anode feed: 0.8 M C H 3 OH solution flow rate 2 mL min
58 Figure 2 5 Initial life/stabilized power density and initial operation al power density loss for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes a ) Comparison of beginning life and stabilized power densit ies b ) Comparison of operati on al power density loss for the first hour of operation. The power density loss was calculated by subtracting stabilized power density of the specific DMFC from the initial power density Operating condition: cell temperature: 5 0 C, cell voltage 0.35 V A node feed: 0.8 M CH 3 OH solution flow rate 2 mL min
59 Figure 2 6. Polarization curve s and power densities of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test Polarization test condition s: cell temperature: 5 0 C, a no de feed: 0.8 M CH 3 OH solution flow rate 2 m L min cathode fee d : air, flow rate 2 SLPM.
60 Figure 2 7. Anode polarization curve s and power densit ies of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage ac celerated durability test Polarization test condition s: cell temperature: 5 0 C, a node feed: 0.8 M CH 3 OH solution flow rate 2 m L min cathode fee: hydrogen, flow rate 0.3 SLPM.
61 Figure 2 8. Anode relative electrochemical active surface areas of DM FCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test Data calculated at 120 mA cm 2 Test conditions: cell temperature: 5 0 C, a node feed: 0.8 M CH 3 OH solution flow rate 2 m L min cathode fee: hydrogen, flow rate 0.3 SLPM.
62 Figure 2 9. Methanol crossover of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test. Polarization test condition: cell temperature: 5 0 C, anode feed: 0.8 M CH 3 OH solution flow rate 2 m L min SLPM.
63 Figure 2 10 Cathode polarization curves of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during a course of 144 hours low voltage accelerated durability test. Test conditions: cell temperature : 5 0 C, anode feed: 0.8 M CH3OH solution rate 2 SLPM.
64 Figure 2 11 Impedance spectrum of DMFCs with JM ELE147, JM ELE170, Tanaka, Cabot anodes during a course of 144 hours low voltage accelerated durabi lity test. Operating condition: cell temperature: 5 0 C, anode feed: 0.8 M CH3OH solution flow rate 2 mL d : hydrogen gas, flow rate 0.3 SLPM, test frequency range: 10 kHz to 0.1 Hz with an AC amplitude of 20 mV.
65 Figure 2 12. Resista nce of DMFCs with JM ELE147, JM ELE170, Tanaka, Cabot anodes during a course of 144 hours low voltage accelerated durability test Operating condition: cell temperature: 5 0 C, a node feed: 0.8 M CH 3 OH solution flow rate 2 mL min cathode feed: hydrogen gas, flow rate 0.3 SLPM, test frequency range: 10 kHz to 0.1 Hz with an AC amplitude of 20 mV.
66 Table 2 1 Power density of DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVADT at a current density of 120 mA cm 2 LVADT Time (h ours) Power density of DMFCs during the course of LVADT ( m W cm 2 ) ELE147 ELE170 Cabot Tanaka Fresh 49.16 44.36 43.16 44.36 48 44.36 43.16 39.57 39.56 96 43.16 40.77 37.17 38.36 144 39.57 40.77 35.97 38.36 Table 2 2 Power density loss rate for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes d uring 144 hours LVADT at a current density of 120 mA cm 2 LVADT Time Range Power density degradation rate for DMFCs ( W cm 2 h 1 ) ELE147 ELE170 Cabot Tanaka 0 48 hours 99.79 24.98 74.94 99.79 48 96 hours 25 .00 49.96 49.96 24.98 96 144 hours 74.17 0 24 .79 0 0 144 hours 66.32 24.98 49.9 41.59 Table 2 3 Anode potential for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVADT at a current density of 120 mA cm 2 LVADT Time Range (hours) Anode Potential for DMFCs ( V ) ELE1 47 ELE170 Cabot Tanaka Fresh 0.38 3 0.390 0. 400 0.401 48 0.400 0.40 1 0.408 0.406 96 0.402 0.402 0.413 0.41 6 144 0.404 0.409 0.417 0.42 1 Table 2 4 Cathode potential for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during 144 hours LVAD T at a current density of 120 mA cm 2 LVADT Time Range (hours) Cathode Potential for DMFCs ( V ) ELE147 ELE170 Cabot Tanaka Fresh 0.829 0.789 0.798 0.810 48 0.804 0.799 0.784 0.792 96 0.797 0.788 0.765 0.792 144 0.774 0.790 0.756 0.780
67 Table 2 5 In itial and post test performance for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during the 144 hours LVADT at current density of 120 mA cm 2 ELE147 ELE170 Cabot Tanaka Fresh o verall p ower output (mW cm 2) 49.16 44.36 43.16 44.36 Fresh Anod e potential (V) 0.383 0.39 0 0.4 0 0.401 Fresh Cathode potential (V) 0.829 0.789 0.798 0.81 Post test o verall p ower output (mW cm 2) 39.57 40.77 35.97 38.36 Post test Anode potential (V) 0.404 0.409 0.417 0.421 Post test Cathode potential (V) 0.774 0.79 0 0.756 0.78 0 Table 2 6 Degradation in different components for DMFCs with JM ELE147, JM ELE170, Tanaka, and Cabot anodes during the 144 hours LVADT at current density of 120 mA cm 2 Degradation Type ELE147 ELE170 Cabot Tanaka Initial power loss rat e (mW cm 2 h 1 ) 5.119 0.346 3.071 7.257 Overall power output degradation rate ( W cm 2 h 1 ) 66.32 24.95 49.9 0 41.59 Anode degradation rate (mV h 1 ) 0.146 0.132 0.138 0.118 Cathode degradation rate (mV h 1 ) 0.382 0 .000 0.292 0.208
68 CHAPTER 3 THE I MPACT OF METHYL ETHYL KETONE CONTAMINATION OF A PLATINUM/CARBON ELECTROCATALYST ON ITS OXYGEN REDUCTION ACTIVITY 3 .1 Background For more than several decades fuel cells have been praised for their prospect as efficient, clean means to convert chemical ene rgy into electric energy in many important applications, such as stationary power station s vehicles [ 67 ] and portable power sources [ 68 ] The theoretical efficiency of a fuel cell is higher than the efficiency of a traditiona l internal combustion engine, electrical power plant at the same power output [ 69 ] Certain type of fuel cell, particularly direct methanol fuel cell s has higher power density than traditional lithium batteries for mobile electronic device application s [ 70 ] in fuel cells such as CO, CO 2 NH 3 H 2 S, NO x S O x sulfur compounds and carbon hydrogen compounds, can cause performa nce damage, or permanent destruction to the anodes cathodes membranes, and conponents [ 71 ] [ 72 ] All contaminants listed above could originate from the manufacturing process of fuel cell fuel, such as CO, CO 2 and CH 4 or from other small organic molecule produced from hydrogen during reforming process [ 73 ] Contaminants c an also come from the air, which is commonly used as the oxidation reactant for fuel cell cathode reaction s For example, NO x S O x are commonly present in the atmos phere due to widely existing air pollution [ 74 ] Some undesired ions or organic compounds such as greases, coming from the system components are also found to be so urces of fuel cell contaminant ion [ 75 ] Due to the hydrophobic nature of many commercial gas diffusion layer (GDL materials, GDL s ha ve very low affinity for hig h ly polar solvents, including water. During
69 the MEA manufacturing process, in order to enable adequate contact between the aqueous platinum catalyst ink and the hydrophobic GDL surface for depositing catalyst i nk, wetting agent chemicals are widely used to modify GDL surface properties and improve gas diffusion electrode performance [ 76 ] .Due to the nature of the working environment in the fuel cell, wetting agent c hemicals need to be stable and electro inactive. In other words, the wetting agent is not expected to participate in the electrochemical reactions in the fuel cell, and should not be a poison to the fuel cell catalyst. One commonly used wetting agent is m e thyl ethyl ketone (MEK) [ 77 ] D espite t he large effort made in past decades on electrochemical reactions of adsorbed organic species on platinum electrodes [ 78 81 ] many studies have been devoted to analysis of adsorbate configurations in electrocatalytic reactions of simple organic molecules, such as carbon monoxide and methanol. However, knowledge of the oxidation of MEK on Pt based fuel cell catalyst s is v ery limited due to the nature and complexity of its electrochemical decomposition. The electrochemical response of ketones has not b een completely understood on Pt based catalyst s Zinola et al [ 82 ] used differential electrochemical mass spectrometric (DEMS) techniques to elucidate mechanistic details and to monitor product formation of MEK electro oxidation on P t surface s The oxidation/redu ction of MEK adsorbates formed in the potential range of 0.1 to 0.6 V was investigated in their study, and they found that adsorption of MEK takes place in this potential range, reaching a maximum Pt surface coverage by MEK ad sorbed species at 0.2 V. The reduction products of MEK were d etected to be mainly butane, as well as trace amount of methane, ethane and propane. Carbon dioxide was also detected during the oxidation of the organic residue originating from MEK
70 adsorbates n ot reduced during cycling in the hydrogen region and 16% of initial ly adsorbed species from MEK were found to be remaining on the Pt surface after a few cyclic voltammetry cycles. However, to our best knowledge, the effect of the presence of MEK species o n ORR activity on DMFC cathode s has not been studied. The understanding of the kinetic limitations of the oxygen reduction reaction due to the presence of MEK contamination on the fuel cell cathode catalyst is a problem of practical significance. In spite of extensive effort in the past few decades by fuel cel l researchers to develop non Pt based catalyst s for l ow temperature air cathodes, Pt based catalyst s remain the best known electrocatalyst s for the oxygen reduction reaction (ORR) in DMFC application s The oxygen reduction reaction on different platinum crystal surface s has been extensively investigated [ 83 85 ] It is well known that the reaction rates of the OR R on Pt surface are very sensitive to surface str ucture and contamination, due to absorbates [ 86 ] such as hydroxyl species and chloride ions [ 87 ] In this study, the main goal is to establish to wh at extent the presence of the wetting agent m ethyl ethyl ketone influences the oxygen reduction reaction performance of a typical Pt/Carbon DMFC cathode catalyst. Quant itative kinetic measurements of the ORR of a DMFC Pt/Carbon cathode catalyst upon MEK im purity introduction were performed using the thin film rotating disk electrode (RDE) method. This technique enables the accurate calculation of kinetic current s of electrochemical reactions on the catalyst without influence from transport limitation s It a lso allows precise determination of kinetic data, such as Tafel slopes and turnover frequenc ies Although this study was conducted in a liquid electrolyte, similar reduced activity can be expected in a DMFC. The result from this study will add to the under standing of the effects of trace amount s
71 of MEK contamination on DMFC cathode catalyst performance, and will indicat e the necessity of modifying catalyst ink formulation s in order to reduce performance losses and degradation phenomen a 3 .2 Experimental Me thods 3 .2.1 Electrocatalyst and Electrolyte Hispect11100 (70% Pt/C) electrocatalyst (Hispect11100, Johnson Matthey, USA) was chosen as a model cathode catalyst for the investigation of MEK contamination effect s on ORR kinetics. The catalyst was used as rec eived without further treatment. A 0.5 M sulfuric acid solution (Fisher Scientific, USA) was used as electrolyte for the entire course of study. Ultrahigh purity nitrogen and oxygen (Airgas,USA) were used for solution deaeration and oxygen saturation res pectively. 3.2.2 Electrode Preparation Electro catalyst suspension ink was prepared by mixing 20 mg of catalyst with 3.84 mL methanol (Acros Organic,USA), 0.08 mL of 5 wt% Nafion (Dupont, USA) and 16 mL DI water. The mixture was agitated continuously in an ultrasonic bath for 15 minutes to achieve a uniform suspension. A glassy carbon rotation disk electrode (Pine Instruments, 0.196 cm 2 ) was used as the substrate for the supported eletrocatalyst. Before depositing the catalyst ink, the electrode surface was polished to mirror finish using an Al 2 O 3 water suspension, particle size 0.05 m, and then cleaned ultrasonically in water for 15 minutes. A 10 L aliquot of ink suspension (corresponding to 6 g Pt metal) was placed on the R DE by micropipette and left ov ernight to dry completely and give a constant metal loading of ~30 g/cm 2 (geometric). The coating was attached to the glassy carbon surface.
72 3 .2. 3 Electrochemical Measurement The el e ctrochemical measurements were conducted in a glass cell for rotating e lectrodes with water jacket The working electrode was an RDE with a disk made of glassy carbon, the counter electrode was a carbon rod electrode, and the reference electrode was a saturated calomel e lectrode (SCE), separated from the main glass cell compa rtment by a reference electrode salt bridge (Pine Instrument, USA) to avoid chloride contamination of the studied solutions The RDE was screwed onto the Pine Instruments AFSAR 2 electrode rotator for rotating speed control. The cell was thermostated at th e desired temperature s by a n i sotemp cooling/h eating r ecirculating c irculator (Fisher Scientific, USA). A Gamry reference 600 potentiostat ( Gamry Instruments, Inc. UAS) was connected to the electrochemical reaction cell and used to perform all the measure ments. Cyclic Voltammetry Special att ention was paid to the cleanliness and purity of the electrode surface during cyclic voltammetry measurement s In order to produce a clean surface t he e lectr ode was immersed in deaerated (h igh purity nitrogen, Airgas USA) 0.5M sulfuric acid s olution, and cycled between 0 and + 1.3 V at 50 mV/s for 10 times until stabilized ove rlapping curves were obtained. Approximately 1 50 m L of a 0.5 M sulfuric acid electrolyte solution was vigorous bubb led with high purity nitrogen gas for 1 h our prior to begin the electrochemical test with continuous purg ing during the experiment. The cyclic voltammetry was obtain ed between 0 and +1.3 V at 20 mV/s for four cycles. Typically after the first cycle, all successive cycles overlap ped w ith each other, and the last three cycles of CV were averaged For investigating the contamination effect of different concentration s of MEK on CV, a controlled amount of diluted MEK solution was injected into the sulfuric acid electrolyte using a pipette, to
73 give a desired concentration of MEK After changing the MEK concentration in solution, the electrolyte was continuously bubbled with nitrogen for 30 minutes, in order to equilibrate the electrode with the newly introduced impurities and to k eep oxygen from dissolving in the solution. The charge due to adsorbed hydrogen was determined from the hydrogen underpotential deposition region [ 88 ] Oxygen reduction reaction The electrode was immersed in sulfuric acid solution previously purged with high purity oxygen (Airgas, USA) for 30 min at 25 C The cyclic voltammograms were recorded be tween 0.05 and 1.2 V at a sweep r ate of 20 mV/s for four cycles at each rotation speed, and the last three cathodic scans at each experimental condition were averaged and used for ORR data analysis. The rotation speeds of 400, 625, 900, 1600, and 2500 rounds per minutes were used for the Koutechi Levich relationship study. For investigating contamination effect s of different concentration of MEK on ORR, the desired MEK concentration was introduced into the electrolyte as described in the previous paragraph 3.3 Results and Discussion 3.3 1 Cyclic Voltammetry Studies on the DMFC Cathode Catalyst in the Presence of MEK Contamination Cyclic voltammetry (CV) was employed as the first approach to probe how the fundamental electrode reactions on a DMFC cathode catalyst are affected in the prese n ce of MEK contamination in an acid electrolyte. Figure 3 1 presents steady state voltammogram s for the Hispect11 100 cathode catalyst coated RDE with different concentration s of MEK contamination. After the concentration of MEK was adjusted, the voltammon grams typically overlapped with each other after 1 2 cycles, indicat ing rapid onset of a steady state at the Pt surface.
74 The solid blue line represents the base voltammogram with a Pt loading of 60 g/cm 2 On the anodic sweep there are three distinct reg ions: hydrogen desorption, 0 to 0.4 V, double layer charging, 0.4 to 0.75 V ( Note: double layer charging occurs over the entire voltage range applied, but it is the sole process involved over this potential range), and surface oxidation, 0.75 to 1.3V. The f eatures in the hydrogen adsorption region can be rationalized since the carbon supported Pt particles are normally cubo octahedral crystallites, mainly consisting of Pt(111) facets with additional Pt(100) facets and low coordination Pt atoms. The H adsorpt ion/desorption peak at ~0.12 V are in agree ment with data reported for Pt(110) single crystals [ 89 90 ] and m ay be associated with low coordination atoms at the edge s and corner s of Pt particles. The H adsorption peaks at 0.18 V could also be associated with Pt(110) sites [ 89 91 ] Furthermore, all the characteristic peaks for H adsorption/desorption were observed for all voltammograms of MEK contaminated electrolyte at various concentrations. Progressive changes in the voltammogram profile were observed as a function of increasing MEK concentration. Hydrogen underpotential deposition (H upd ) charging current and double layer charging current were found to decrease with increasing MEK concentration, which suggests that a portion of the Pt catalyst surface is blocked by MEK. The fraction of blocked Pt surface area can be estimated by calculating the area under the hydrogen adsorption/desorption current region between 0.05 and 0.4 V, assuming a hydrogen adsorption stoichiome try of one. In order to minimize influence from the residual oxygen present in solution, the electrochemical ly active area of Pt, namely SPt, was calculated from an a veraged electric charge for the hydrogen desorption/adsorption B oth the positive going an d negative going potential scan s from
75 0.05 to 0.40 V in cyclic voltammetry at the sweep rate of 20 mV/s at 25 C were used The hydrogen adsorption/desorption charge can be determined by Q H =( Q Total Q DL ), where Q total is the total charge transfer in the hydrogen adsorption/desorption region and Q H is the capacitive charge from both Pt and carbon support double layer charging [ 92 ] The value of Q H can be used to calculate the Pt active surface area by assuming that a charge of 0.22 mC cm 2 corresponds to a monolayer of adsorbed hydrogen atoms [ 93 ] The comparison s of Q H and Pt active surface areas in various concent ration of ME K contamination are shown in Table 3 1. The value of Q H of fresh catalyst is 2.42 mC/cm 2 corresponding to a Pt surface concentration of 2.5110 8 mol pt cm 2 assuming a monolayer of adsorbed hydrogen atoms and Faraday constant 96,485 C mol 1 Th e Pt surfa ce concentration calculated from Pt loading (60 g pt cm 2 ) and dispersion (13%) is 3.6910 8 mol pt cm 2 Based on Pt surface concentration from hydrogen s tripping and dispersion calculation, we can reasonably estimate that roughly 68% of Pt surface atoms a re electrochemically active. The electrochemical active surface area (EAS) for Pt catalyst was decreased by 13%, 21% and 37% in the presence of 10 M, 100 M and 1000 M MEK respectively. Pt catalysts are known for being inefficient in breaking down C C bond s but the decrease in Pt EAS with addition of MEK reveal that MEK and/or its electro oxidation product can block active Pt surface sites and cause catalyst poisoning. This finding agree s with previous work by Zinola et al [ 82 ] who found that MEK reache d its maximum surface coverage on Pt electrode at 0.2 V, which is within the H upd region. The magnitude of the Pt oxide formation peak starting from the double layer region in the positive going anodic sweep increases with increasing MEK concentration.
76 This can be explained by the retardation of oxide formation by MEK related adsorbed species. The magnitude of the Pt oxide reduction current peak at around 0.72 V in the cathodic sweep of cyclic voltammetry decrease d as a function of MEK concentration providing additional evidence that absorbent from MEK contamination decreased the number of Pt acti ve surface sites, and the number of available surface Pt atoms for forming Pt oxide oxidation and reduction current decrease d in agree ment with the observation in the Hupd region. The Pt oxide reduction did not show a shift in peak position indicat ing that the reaction potential of the Pt surface may retain the original characteristics after MEK product deposition 3.3.2 Oxidation Reduction Reaction Kinetics f or the DMFC Cathode Catalyst in the Presence o f MEK Contamination 3 3 2 1 ORR polarization curves Fig 3 2 illustrates the RDE polarization curves for the ORR on DM FC cathode catalyst electrode with increasing MEK concentration (0 1000 M MEK, 1600 rpm, 20 mV s 1 25 C only cathodic sweep shown). From Fig 3 2, the entire polarization curve shifted progressively toward higher overpotential value s with increasing con centration of MEK in the electrolyte. However, loss in ORR activity did not seem to be highly sensitive to increasing concentrations of MEK above 10 M, as demonstrated in the small progressive changes among oxide reduction RDE curves for MEK concentration s between 10 1000 M. Fig 3 3 presents the ORR polarization curves (same experimental condition s as in Fig 3 2) for the DMFC Hi spect11100 catalyst electrode recorded in oxygen saturated 0.5 mol L 1 sulfuric acid electrolyte at different rotation rates. The ORR is under mixed
77 diffusion and kinetic control in the potential range between 0.9 a nd 0.6 V (cathodic sweep data were used in Fig 3 3), followed by a sole diffusion control region. Apparently, the diffusion controlled limiting current density for ORR wa s achieved for various rotating rates at cathodic sweeps. As expected, higher rotating rate g ave a more robust diffusion of reactant and product species and result ed in a higher ORR current. Another phenomenon which needs to be pointed out is that the ORR was faster during the cathodic sweep than that during the anodic sweep (data not shown). During the cathodic sweep, the Pt oxide species on surface would be reduced to Pt metal This phenomenon can be explained by more rapid ORR on reduced compared to than on oxidized Pt surface s which is intrinsic characteristic of Pt electrodes. 3 3 2 2 ORR kinetically current density calculation To quantitatively determine the kinetically limited current density of the electrochemical catalyst, j k at different contami nation condition s the measured apparent current density j can be described by the following equation [ 88 ] : (3 1) where j k is the kinetically limited current density of the electrochemical catalyst in the absence of mass transfer effects, and j d is the diffusion limited current densit y and j f is the current due to mass t ransfer through the Nafion polymer/solution interface. It is important to note that equation 3 1 assum e a first order reaction for oxygen [ 94 ] reduction in the kinetic current in the mixed activ ation diffusion region. The diffusion limited current densit y, j d is given by : (3 2)
78 or in short form (3 3) Equation 3 2 is the conventional Levich equation n is the number o f electrons transferred in the half reaction (mol 1 ) F is the Faraday constant (C mol 1 ) S is the electrode area (cm 2 ) D is the diffusion coefficient (cm 2 s 1 ) is the angular rota tion rate of the electrode (rad s 1 ) is the kinematic viscosity (cm 2 s 1 ) and C is the reactant concentration in the solution (mol cm 3 ) and B is the slope of the plot of j d 0.5 j f can be given by the equation (3 4) where D f C f and f are the diffusion coefficient of oxyge n, the concentration of oxygen, and the thickness of the Nafion polymer film, respectively. However, the effect of the mass transfer through the polymer/solution interface is significant only when the electrode is covered by a thick Nafion polymer film Since the kinetic current density j k is independent of rotation rate relationship between j d 1 and 0 .5 can be expected. Fig 3 4 presents the Levich plo t for limiting current density, j lim 1 and 0 .5 Clearly, limiting current densities, with or without the MEK presence in electrolyte, were governed by the Koutechi Levich relationship, as evidenced by the linear relationship in the j lim 1 0 .5 plot. A least squares linear fit algorithm was applied to the Koutechi Levich plots for DMFC catalyst electrode at various MEK concentrations. Therefore the v alue of B can be obtained from the slope
79 (Table 3 2). Values of B, ranging from 6.64 6.82 m A cm 2 rpm 0.5 are nearly identical with and without presence of different MEK concentrations, indicating that the number of electron s transferred per oxygen mole cu le in ORR is not changing in the presence of MEK contamination. From Eq 3 1, the diffusion limiting current density can be expressed as: (3 5) If the thickness of the Nafion film f can be reduced to the extent that j f becomes much larger than j k and j d then the 1/ j f will be a negligible contribution In this study, the Nafion used in c atalyst suspensi on preparation wa s in sufficiently small amount, and is not expected to be a significant factor contributing to the limiting current, while only 0.04 L of 5% Nafion solution was deposited into a 0.196 cm 2 glassy carbon surface. The film thickness of Nafi on can be estimated to be ~0.0 45 m by using the density of dry commercial Nafion membranes (~2 g cm 3 [ 95 ] the contribution of th e film diffusion resistance j f to the measured current j is negligible when the thickness of Nafion film is less than 1 m for Pt/C catalyst s. In our study, the j f value wa s ~50 mA cm 2 for non contaminated electrolyte, and it decreased to ~36 mA c m 2 for electrolyte with the highest MEK concentration. T he value of j f at all condition s wa s much higher than the measured j (0 6.7 m A cm 2 ). Especially in the potential region of ORR study 0.7 to 0 .9 V vs. SHE, the measured current density is typically less than 3 mA cm 2 which is much smaller than j f Therefore, we conclude that the contribution of the film diffusion resistance to the measured current density should be negligible, and the equation for measured apparent current density, j can be simp lified as:
80 (3 6) At any potential, the cathodic kinetic curre nt densities can be corrected for mass transport in solution using Eq 3 6 to calculate the mass corrected current density (see below). It is also note worth y point ing out that th e ohmic loss of electrolyte is negl igible in [ 96 ] the resistance of a small electrode disk in an electrolyte solution can be calculated according to the radius of the rotating disk electrode and the conductivity of the electrolyte. (3 7) electrolyte, and a is the radius of the disk. In t his study, the conductivity of 0.5 M H 2 SO 4 solution was 2 00 mS cm 1 [ 97 ] an d the radius of disk electrode wa s 0.25 cm, which yield ing a resistance with a value of current observed wa s less than 7 mA cm 2 while the ohmic voltage drop of electrolyte solution can be calculated to be 6 8 mV at this current density. The ohmic drop of the electrolyte wa s significantly small er than the overpotential of rotating disk electrode and therefore can be neglect ed 3 3 2 3 Oxygen r eduction r eaction k inetics The parallel behaviour of the Koutechi Levich lines in Fig 3 4 and in the previous literature study on Pt/C electrode in impuri ty free electrolyte does not necessarily translate into an unaltered reaction mechanism in the presence of MEK contamination. One of the criteria to check this probability is the Tafel plot slope. The Tafel plots for
81 mass transport corrected (using Eq 3 6) ORR are presented in Fig 3 5 for both impurity free and MEK contaminated conditions. The Tafel slope changes continuously in the studi ed potential range, as reported in previous literature on supported Pt catalyst s [ 98 ] The Tafel slopes do not show any evidence of significant changes in the mechanis m of ORR in the overpotential region investigated (0.6 to 0.85 V) upon MEK contamination. Tafel slopes were fitted through the linear part in Tafel plots between 0.62 and 0.70 V for both fresh and MEK contaminated conditions. For clean electrode, the Tafel slope was found to be 1 47 .7 mV decade 1 which is practically close to Tafel slopes for ORR on smooth Pt reported in the literature, with a value of 120 mV decade 1 [ 99 101 ] Tafel slopes were found to be increase more negative with the concentration of MEK present in electrolyte, with values of 157 2 mV decade 1 1 62 5 mV decade 1 and 1 66 3 mV decade 1 for 10 M MEK, 100 M MEK and 1000 M MEK respectively. The reason for change of the Tafel slope from 60 at low currents to 120 mV decade 1 at high currents at the smooth Pt surfac e was explained by the change from Temkin to Langmuir adsorption conditions (decrease in the surface coverage) [ 100 ] A s imilar reason for the change in oxygen adsorption condition on Pt surface due to MEK contamination could possibly be responsible for the observed increase of Tafel slope with increasing MEK contaminant concentration. It is reasonable to assume that the oxi dation product of MEK adsorbed on the Pt surface could interfere with oxygen adsorption on adjacent Pt atoms, which would change the ORR kinetics and result in a noticeable change in Tafel slope. The significant drop in ORR kinetics at the Pt catalyst as a result of MEK contamination is readily evident from the Tafel plots in Fig 3 5. This drop can translate
82 to a 30 mV increase in terms of overpotential at the current density of 1 mA cm 2 Fig 3 6 presents the mass transfer corrected ORR kinetic current de nsity at different overpotentials upon introduction of MEK contamination. Up to 40% ORR kinetic current density loss was found at a potential of 0.65 V for the DMFC cathode catalyst electrode when the MEK concentration reache d 1 mM as compared to those mea sured on the clean electrolyte. This finding agrees with the H upd based electrochemically active surface area loss during introduction of MEK contamination, which showed ~37% ESA loss between fresh electrolyte and electrolyte containing 1 mM MEK. From Fig 3 6, dramatic kinetic current density loss commence d even at the very low concentration of MEK, as 26% kinetic current loss was already tak ing place when MEK concentration was only 10 M in electrolyte (at 0.65 V). However, the kinetic current loss was no t proportional to the MEK concentration, and it decrease d in a much slower manner as MEK concentration increase d and this is evident at all potential s investigated in Fig 3 5. Therefore, we conclude that the Pt/C electrode was very sensitive to MEK pollu tion, and m ost kinetic activity loss occurred with the introduction of a trace concentration of MEK. Continuously increasing the MEK contamination concentration would further degrade the Pt/C electrode performance, but not in a rapid manner. The oxygen red uction reaction is known to proceed either in a four electron pathway to water or in a two electron pathway to hydrogen peroxide. The f our electron pathway is generally considered as the major reaction route on polycrystalline platinum electrodes in impuri ty free acidic aqueous electrolytes [ 102 103 ] The four electron pathway is also the major reaction pathway f or Pt single crystals, except for potentials within the hydrogen adsorption/desorption region for Pt (111) and Pt(110) surfaces,
83 where quantitative oxygen is reduced in a two electron pathway [ 98 104 ] Since the carbon supported platinum DMFC cathode catalyst consist s of cube octahedral nanocrystallites with a large portion of Pt(100) and Pt(111) for particle size s above 2nm [ 105 ] w e can assume that the ORR on Pt/C catalyst (d pt =4.6 nm) in this study resembles the behavior mostly of Pt(111) and Pt(100). U.A. Paulus et al ha ve investigated the ORR pathways on Pt/Vulcan catalyst s deposited on a glass y carbon rotating disk electrode, and it was found t he hydrogen peroxide formation wa s negligible above 0.6 V, and the ORR proceed ed exclusively via a complete four electron path way [ 105 ] Therefore, we assume that in the potential region above 0.6V in this study, all the ORR in contami nation free electrolyte proceeded in a four electron pathway manner. Closely follow ing the method used by T.J.Schmidt et al [ 106 ] the turnover frequency (TOF) at different potentials was calculated from the kineti c current densities obtained (cathodic sweep of cyclic voltammetry in oxygen saturated electrolyte, rotating rate 1600 rpm) according to the formulation: (3 8 ) in which TOF is the absolute reaction rate, defined as number of oxygen molecules being reduced per second and per active surface Pt atom, i k is the kinetic density in unit s of mA cm 2 n is the number of electrons parti cipat ing in the ORR reaction (n=4 for potential above 0.6 V as explained above ), F is the Faraday constant, and N s is the number of active surface Pt atoms in the unit of atoms cm 2 (data listed in Table 3 1). Fig 3 7 presents the TOF as a function of pote ntial for ORR in electrolyte with and without MEK contaminant. An ORR activity decrease per active Pt surface atom of about 30 40% was observed for the Hispect11100 cathode catalyst electrode upon
84 introduction of different concentration of MEK contaminati on. To look more closely at the activity loss rate associated with MEK concentration, plot of TOF vs. MEK contamination concentrat ion at different potentials are shown in Fig 3 8. Again, most of the ORR activity loss (60 70% out of total loss between impur ity free and the highest MEK contamination condition) was found to occur after introducing trace concentration of MEK in solution ( 10 M MEK in acid solution at all potential s studied ) This observation agree with the finding from the Tafel plot study, wh ich confirms that even trace amount s of MEK c an impair ORR activity considerably. In terms of TOF, such activity loss translate s into a 27% loss at 0.6 V when 10 M MEK was introduced into acid electrolyte. A s imilar degree of activity of loss was observed at all potential range (0.6 0.8 V) investigated. 3.3 3 Discussion According to the experimental results, the presence of trace amount s of MEK caused a destructive influence on ORR at the Pt/C DMFC cathode catalyst electrode This translated into a 30 mV increase in terms of overpotential, or 50% TOF loss when the MEK concentration reache d 1 mM. Loss of Pt active surface area upon introduction of MEK impurity w as observed from CV in nitrogen enriche d electrolyte. Therefore, the ad sorption of MEK and its re duction /oxidation product was proposed to be one mechanism attribut ing for the loss of active surface Pt sites and altered ORR activity. Surface absorbed species have been known to block active surface metal sites and impair ORR activity on Pt electrode s Previous work by T.J. Schmidt et al. [ 87 ] has shown that the onset of ORR is shifted to higher overpotentials with increasing chloride contamination conce ntrati on in the electrolyte, with roughly a constant potential decrease of one order of magnitude for each order of magnitude increase in chloride
85 anion concentration. In our study, MEK impurities impair ORR in a much gentle manner compare to effect on ORR intro duced by chloride ions, which suggests that MEK and its associated absorbed species have a relatively weak interaction with the Pt surface. This agrees with the finding by Zinola [ 82 107 ] that majority of initially adsorbed species of MEK at 0.2 V would be reduced in the hydrogen adsorption/desorption potential region, and only 16% of the initially adsorbed specie s would remain unreduced on the Pt surface. It wa s also found that trace level of methane, ethane and propane were possibly formed as electro oxidation/reduction product s of MEK on Pt surface. Even though Zinola et al ation into the mechanism of MEK adsorption and electrochemical reaction on a Pt surface, the lack of information about surface absorbate species left the reaction pathway and product of MEK on Pt surface unclear. It has been suggested [ 107 ] that the adsorption of MEK on a P t surface is related solely to the functional group, >C=O group, and that adsorption occurs in the potential range between 0.1 0.6 V [ 82 ] The oxidation of MEK absorbate residue occurs in the Pt oxide formation potential range [ 107 ] The electroadsorption of organic substance with double C=C or C=O bonds on noble metals in acidic electrolyte s could trigger a process leading to formation of CO like surface species [ 108 ] A CO covered polycrystalline Pt electrode cannot be fully oxidized by oxygen ; therefore the remaining CO can essen tially block the Pt sites for ORR activity [ 86 ] Also, oxidation of CO on Pt surface s has been associated with a significant amount of peroxide formation, attack s various fuel cell components and lead s to performance loss.
86 It is known that Pt surface adsorbates c an interact with ORR reactan ts. For example, it is know n that the reversible adsorption of hydroxyl ions (OH ) on Pt suppresses the ORR kinetics [ 86 ] The interaction between MEK associated adsorbates on the Pt surface with ORR reactants or products may also be expected. 3.4 Summary The oxygen reduction reaction on a typ ical carbon supported Pt DMFC fuel cell cathode catalyst in the presence of different concentration s of methyl ethyl ketone was investigated in this study using a rotating disk electrode method. The cyclic voltammetry study in nitrogen enriched electrolyte show ed that the Pt e lectro active surface area decreased ~13 38% upon exposure to 1 M 1 mM MEK contaminated electrolyte. The rotating disk data show ed that the rate of oxygen reduction reaction can decrease by ~15 40% in terms of kinetic current density (corrected for transport limitation), translating into ~15 30 mV overpotential penalty upon introduction of 1 M 1 mM MEK into the acidic electrolyte. In addition, turnover frequency analysis demonstrated that TOF value dropped by ~30 40% after 1 M 1 mM MEK was added to the electrolyte solution. The data in this study and available literature information suggest a proposed mechanism c ausing ORR activity loss by MEK. MEK adsorbate species act as a site block er which reduce the number of available active Pt surface sites for ORR activity. MEK concentration on the order of 0.7 ppm (w/w) would result in a fuel cell voltage loss of 15 mV, which would equally affect the open circuit potential. The presented data demonstrated that MEK is indeed a pollutant for ORR on Pt/C based DMFC cathode catalyst s and could similarly cause performance degradation of DMFCs. The utilization of MEK in fuel cell manufacture should be given
87 special attention, either by complete removal of MEK, or by replacing MEK with a more ORR activity friendly component.
88 Figure 3 1. CV curves for a thin film Hispec t11100 60 wt% Pt/C electrode in n itrogen saturated 0.5 mol L 1 H 2 SO 4 solution contai ning dissolved MEK at various concentrations Pt catalyst loading 60 g cm 2 GC ;150 cm 3 N 2 sat d 0.5 mol L 1 H 2 SO 4 solution at 50 C ; sweep rate: 20 mV/s
89 Figure 3 2 Rotating disk electrode polarization curves for a thin film Hispec t11100 60 wt% Pt/C electrode in the presence and absence of dissolved MEK at various concentrations Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C ; 20 mV/s rotat ion rate : 1600 rpm.
90 Figure 3 3 Oxygen reduction reaction polarization curves for a thin film Hispec t11100 60 wt% Pt/C electrode in the presence and absence of dis solved MEK in various concentrations and at various rotating rates. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C ; sweep rate: 20 mV/s.
91 Figure 3 4 Koutechi Levich plots of the limiting diffusion current density j d for ORR for a thin film Hispec t11100 60 wt% Pt/C electrode in the presence and absence of dissolved MEK at various concentrations Pt catalyst loading 30 g cm 2 GC ; 150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C ; sweep rate: 20 mV/s.
92 Figure 3 5. Tafel plots for mass transfer corrected ORR kinetic current density for a thin film Hispect11100 60 wt% Pt/C in the presence and absence of dissolved MEK in various concentrations. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 soluti on at 25 C; 20 mV/s; rotati on rate : 1600 rpm.
93 Figure 3 6 Corresponding changes to the mass transfer corrected ORR kinetic current density j k at different overpotentials on a thin film Hispec t11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution as a function of dissolved MEK concentration Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotati on rate : 1600 rpm.
94 Figure 3 7 Turnover frequencies (TOF) for ORR on a thin film Hispec t11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution at 25 C as a function of potentials without or with presence of different concen tration of MEK in electrolyte. Calculated from the kinetic current densities used in Fig 3 5 using Eq 3 7. Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotati on rate : 1600 rpm.
95 Figure 3 8 Corresponding changes to the turnover frequencies (TOF) at different overpotentials on a thin film Hispec t11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution as a function of dissolved MEK concentration Pt catalyst loading 30 g cm 2 GC ;150 cm 3 O 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 25 C; 20 mV/s; rotati on rate : 1600 rpm.
96 Table 3 1 H udp a nd Platinum electrochemical ly active surface area for a thin film Hispec t11100 60 wt% Pt/C electrode in Nitrogen saturated 0.5 mol L 1 H 2 SO 4 solution contain dissolved MEK at various concentrations Pt catalyst loading 60 g cm 2 GC ;150 cm 3 N 2 sat. 0.5 mol L 1 H 2 SO 4 solution at 50 C ; sweep rate: 20 mV/s 0 M MEK 1 0 M MEK 1 0 0 M MEK 100 0 M MEK Q H (mC cm 2 ) 2.4 2 1.99 1.9 6 1.8 4 Pt EAS ( m 2 g 1 ) 11.4 4 9.9 9 8.99 7.1 9 Active Pt atoms surface concentration (mol pt cm 2 ) 2.5110 8 2.0610 8 2.0310 8 1.9110 8 Table 3 2. Parameters obtained from Levich plots of the limiting diffusion current density j f for ORR for a thin film Hispec t11100 60 wt% Pt/C electrode in Oxygen saturated 0.5 mol L 1 H 2 SO 4 solution in the presence and absence of dissolved MEK in vari ous concentrations Fresh 1 0 M MEK 1 0 0 M MEK 100 0 M MEK B (m A cm 2 rpm 0.5 ) 6.64 6.82 6.70 6.66 j f (mA cm 2 ) 49.74 49.59 41.31 36.51
97 CHAPTER 4 KINETIC STUDY OF PROPYLENE HYDROGENATION OVER A PLATINUM CATALYST USING AN IN SITU HYPERPOLARIZED NMR TE CHNIQUE 4 .1 Background Nuclear magnetic resonance (NMR) has been actively used in numerous scientifi c fields due to its capability to provide information about the structure s of chemical and biological system s, as well as dynamic processes such as diffusi on and fluid dynamics. However, the nature of low sensitivity of NMR, due to the small population differences between nuclear spin energy levels at thermal equilibrium, requires high concentration s of analyte for successful and reliable measurement s. Unfor tunately, NMR technique still suffers from low sensitivity, which excludes it from many application s, such as detection of biomarkers or reaction intermediates often existing in low concentration. During the past three decades, an emerging technique enabl es NMR to be applied to in detection of low concentration species. Bowers and Weitekamp first suggested that significantly enhanced NMR signals of molecules could be achieved by addition of hydro gen atoms from parahydrogen (a spin isomer of the H 2 molecule with its two proton spins aligned antiparallel ) into the target molecule of interest [ 109 ] The term hyperpolarization has been accepted to describe this phenomenon. This effect is known as parahydrogen induced polarization (PHIP) [ 109 ] The first experimenta l observation of PHIP was also reported by Bowers and Weitekamp [ 110 ] A 100 200 fold enhancem ent of NMR signal observed when parahydrogen was added to acrylonitrile to form hyperpolarized propionitrile (CH 3 CH 2 CN). Eisenshmid et al. [ 35 ] and Eisenberg [ 111 ] found that the nuclear polarization from hydrogen gas enriched with parahydr ogen could be transferred to unsaturated organic molecules through a catalytical
98 hydrogenatio n reaction. Each hydrogen atom in a hydrogen molecule (H 2 ) consists of one proton and one electron, and each proton has a magnetic moment associated with its spin. Therefore, the spins of the two protons in a hydrogen molecule couple to form a triplet state known as orthohydrogen, and a singlet state known as parahydrogen [ 11 2 113 ] Orthohydrogen and parahydrogen possess different prope rties in an NMR spectrum. When parahydrogen enriched hydrogen gas is used during a hydrogenation reaction, the resulting product can exhib it hyperpolarized signals, as long as the spins are not decoupled, in NMR spectra. The main advantage of PHIP technology is the strong signal enhancement of the target product in NMR (on the order of 10 3 or higher) [ 114 ] rendering this a powerful and versatile method for NMR detection of products. For the PHIP reaction to occur, both hydrogen atoms from the former para hydrogen nuclei must be added to the reactant molecule without lose of the spin correlation. The PHIP effect produce s magnified antiphase signal patterns in NMR spectrum from two the former para hydrogen nuclei in the resulting product. In this manner, resulting products containing para hydrogen nuclei can be detected and characterized [ 115 ] Because t his kind of reaction enhance s the signal sensitivity of the products in NMR, which enable s para hydrogen induced polarization to be a powerful analytical tool with various potential applications in chemistry [ 116 ] materials science [ 117 118 ] biology [ 119 ] and medicine [ 33 ] In early research, the PHIP reactions mostly employed homogenous catalysts, such as 3 ) 3 ) [ 38 ] 3 ) 2 ) [ 39 ] because homogenous hydrogena tion reactions were more likely to preserve the
99 spins of the two hydrogen nuclei from a parahydrogen molecule in the same resulting product molecule. The success of PHIP reactions using homogeneous catalysts demonstrated the ability of PHIP to enhance NMR signal s for detect ion of low concentration species. However, the utilization of homogeneous catalyst s in PHIP has obvious drawbacks, such as the presence of dissolved catalyst in the reaction mixture. Heterogeneous catalysts, i n contrast to homogenous cat alysts are macroscopic particles which are not dissolved in the liquid They thus provide a solid phase which can be separated relatively easily from the solvent or gas phase reagents or products The PHIP reaction provides a unique reaction product label ing technique. The enhanced NMR signal can only be obtained when two hydrogen atoms from a parahydrogen molecule are added to a substrate in a pairwise manner. In contrast to homogeneous catalysts, supported metal catalysts, are not expected to give enhan ced PHIP signals, since the hydrogen atoms from parahydrogen can dffuse on the surface of the catalyst after dissociation. However, PH IP has recently been observed using heterogeneous catalyst s suc h as highly dispersed platinum on metal oxide supports [ 42 ] Therefore, conditions for pairwise parahydrogen addition can be achieved in heterogeneous alkene hydrogenation reactions. Despite numerous studies of hete rogeneous hydrogenation of unsaturated compounds, the mechanism of PHIP hydrogenation reaction on supported metal catalysts is still not well understood. The occurrence of PHIP in heterogeneous reactions provides several ad vantages over the homogeneous rea ction s, including production of gaseous hyperpolarized product s better control of the reaction process, and easier separation of
100 product s from the catalyst. More than 90% of all commercial chemical processes involve heterogeneous catalysts. Therefore, con siderably mature technology from existing industrial chemical process technology could be used for a heterogenous PHIP reaction including continuous reaction through a catalyst bed reactor instead a batch reactor. The first reported study of the PHIP effe ct on supported heterogeneous metal Al 2 O 3 for propylene hydrogenation in both the PASADENA (parahydrogen and s ynthesis allow dramatically enhanced nuclear alignment ) and ALTADENA (adiabatic longitudinal transport after dissociation engender s net alignment) PHIP experiments [ 41 ] It was tentatively suggested that numerous surface adsorbed species, including reagent, intermediates and products, would create a confined local surface area which would force two H atoms from the same parahydrogen to be adsorb ed close to each other on the metal catalyst surface. Another study [ 42 ] investigated the dependence of PHIP signal enhancement on Pt particle size on different oxide supports. The results indicated that the hydrogenation reaction s (non pairwise) are mainly associated with closely packed (1 1 1) and (1 0 0) planes of the platinum surface, and pairwise hydrogen addition is more likely to occur on corner or kink platinum atoms in small Pt particles (<3 nm). Also, platinum catalysts supported on TiO 2 were found to give higher selectivity toward pairwise parahyd rogen addition in propylene hydrogenation compared to platinum catalysts supported on Al 2 O 3 ZrO 2 and SiO 2 Thus, metal support interactions also influence pairwise parahydrogen addition. There is only one study on the kinetics of heterogeneous propylene h ydrogenation with parahydrogen [ 120 ] In February 2013, Salnikov et al. [ 120 ] reported that for non
101 pairwise hydrogen addition, the reaction order of propylene hydrogenation over Pt/Al 2 O 3 catalyst is 0.1, while for pairwise addition, the reaction order is 0.7. In this study, platinum supported on titanium (IV) oxide wa s employed to study the kinetics of heterogeneous propylene hydrogenati on using parahydrogen and NMR. In addition, t he temperature dependence of NMR signal enhancement was found for the PHIP effect on Pt/TiO 2 catalyst. 4 .2 Experimental Methods 4.2.1 Cata lyst P reparation The c atalyst used in this study, platinum supported on titanium (IV) oxide ( Pt/ TiO 2 ), was prepared by precipitation of platinum hydroxide onto the supports from an aqueous solution of h exachloro platinic acid hexahydrate (H 2 Pt Cl 6 6H 2 O) (Al fa Aesar) by controlling solution pH with sodium hydroxide. More specifically for preparation of Pt/ TiO 2 catalyst 1.98 g of TiO 2 support (Alfa Aesar) was dispersed in 100 mL of deionized water under constant stirring. H 2 Pt Cl 6 6H 2 O (0.05 3 g) was dissolved in 5 mL deionized H 2 O and was poured into the support/water mixture. Platinum hydroxide was deposited onto the support by dropwise addition of a 2.5 mM NaOH solution until the pH value of the catalyst dispersion mixture reached 11. The catalyst dispersion solution was then aged over night under continuous stirring, and subsequently titrated with diluted acetic acid to pH 7. Then, the mixture was filtered, re dispersed in deionized water and filtered a second time after stirring overnight. The aging step w as necessary to prepare a reproducible and active catalyst and washing was necessary to remove residual sodium ions and any chloride contaminant as these can accumulate on the surface and reduce activity. The
102 re dispersed catalyst was dried at 105 C o ver night, and then calcined at 3 50 C for 3 hours. 4.2.2 Chemisorption Measurement Chemisorption is powerful technique for measuring supported metal catalyst surface area Carbon monoxide (CO) chemisorption is wide ly used to determine the surface area of supp orted platinum catalyst s, the CO molecules adsorb on platinum metal atoms on the catalyst surface and if the stoichiometry of CO to Pt atoms on the surface is known the number of Pt surface atoms can be calculated. In this study, c hemisorption measurem ents to characterize the active surface area, dispersion and particle size of platinum on the TiO 2 support was performed on a ChemBET 3000 from Quantachrome Instruments. The fresh catalysts were first oxidized using 5% oxygen in helium at 170 C for 0.5 h and outgassed in helium before the catalyst was reduced in 5% hydrogen in nitrogen at 170 C for 0.5 h The oxidation and reduction cycle yield a the catalyst having a well defined reduced. The catalyst was outgassed in helium at 170 C for another 15 minu tes in order to remove all physically a d sorbed hydroge n on the catalyst surface before cooling to room temperature. The optimized temperature of 170 C was chosen for performing the hydrogen reduction based on the information obtained from temperature prog rammed reduction (TPR) measurement s which indicate d that most of the Pt species was reduced at 170 C The catalyst w as then subjected to CO adsorption measurements to determine the Pt surface area. 4.2.3 NMR Studies B ased on Parahydrogen Induced Polariza tion (PHIP) All NMR experiments were performed using a Bruker Avance 400 MHz NMR spectrometer. A U tube reactor cell was constructed using a glass tubing and placed
103 on the top of the magnet. The r eaction temperature was controlled by a tube furna ce, and monitored with a K type thermocouple The thermal couple was inserted into the outlet side of the gas flow in the reactor bed and separated from the catalyst using a plug of glass wool. A 50 m g sample of the catalyst was used and placed at the bottom of the U tube reactor. Prior to experiments, the catalyst was reduced and activated under constant h ydrogen flow (100 m L /min) for 30 minutes at 17 0 C and then purged with nitrogen (300 m L /min) for 15 minutes at 17 0 C to remove any physically adsorbed hydro gen. Following the activation step the reactor was set to the designated experimental temperature under the same nitrogen flow The precise deliv ery of reaction gas mixture of nitrogen (carrier gas), hydrogen (either normal or para hydrogen ) and propylen e was accomplished with mass flow controllers (Alicat Scientific Inc. ). Parah ydro gen ( p H 2 ) was produced by passing normal h ydrogen ( n H 2 ) through a copper coil filled with activated cha rcoal and immersed in a liquid nitrogen (77 K) D ewar yielding a mixture consisting of 50% parahydrogen and 50% orthohydrogen isomers (hereafter referred to as p H 2 ) The gas mixture flowed through the reactor at elevated temperature (50 35 0 C) and then into the NMR tube positioned in the NMR probe. Since the reaction is exothermic, the reactor temperature increased upon switching from nitrogen to the re action gas mixture. To reduce heat generation and the temperature fluctuations, a s ignificant portion of nitrogen wa s introduced into the reaction gas mixture to both di ssipate the heat and serve as the carrier gas. However, nitrogen dilution also significantly reduced the magnitude of signal enhancement. Obviously, with a lower concentration of nitrogen a larger hyperpolarization effect have be en achieved NMR spectra w ere acquired with 128 scans using a 90 degree pulse
104 when steady state reaction condition s are reached T h e reactor setup wa s automated and controlled with a custom designed LabVIEW program. As the reaction was performed in a low magnetic field of outside o f the superconducting NMR magnet the resulting spectra exhibited the ALTADENA type polarization pattern. V ery reproducible ALTADENA signals could be obtained through the use of precision m ass flow controllers (MFCs) and a thermocouple inserted in to the c atalyst bed. The accurate temperature measurement is a key advantage for ALTADENA, over PASADENA [ 120 ] in which the temperature of the catalyst must be measure d using an external thermocouple located in the probe outside the NMR tube Th e latter is a significant problem for exothermic reactions as it is difficult to determine the actual reaction temperature Catalysts were re activated between each measurement using the same hydrogen reduction method described previously to provide a fresh metal surface. The collected NMR data were used to calculate the values of the selectivity and activity of PHIP under steady state reaction condition s 4.3 Results and Discus sion 4.3 1 Catalyst Characterization By titrating a kno wn amount of adsorbate gas (CO ) into a str eam of inert makeup gas ( helium) th e titrated CO gas stream flows over the sample surface and is adsorb ed onto the exposed Pt surface until the Pt surface is saturated with CO The amount of CO gas in the makeup gas helium stream not adsorbed on the Pt surface is measured by a thermal conductivity detector (TCD). By subtracting the known amount of unabsorbed CO from the TCD signal from the total amount of titra ted CO, the absorbed
105 CO gas can be quantitatively calculated. Estimates of the Pt particle sizes were made from these CO adsorption measurements using equation 4 1 [ 1 21 ] (4 1) w here i s the average particle size; Sav is the average stoichiometry: CO/Pt=1; k is the shape factor, 5 in this study, which is related to a cube with one side attached to the support and five sides exposed to the environment ; Vm is the molar volume; Na is the Avogadro's number; is the metal density; Vg is the volume of gas adsorbed; and Cm is the surface density of metal atoms, and a value of1.25 10 15 atoms/cm 2 was used i n this study. Similarly, the platinum metal surface area ( ) can be calculated according to equation 4 2. (4 2) Based on 1% nominal platinum metal loading and CO adsorption measurement, using Eq (4 1) and Eq (4 2), the dispersion and mean Pt metal particle size for 1% Pt on TiO 2 catalyst were obtained, with value of 89.72% and 10.52 nm respectively. 4.3 2 1 H NMR Spectra during Propylene Hydrogenation with Parahydrogen For hydrogenation of propyle ne over Pt/TiO 2 catalyst, different partial pressures of parahydrogen, propylene and nitrogen carrier gas were used while maintaining the same total gas flow rate. Keeping the constant gas flow rate ensure s th at travel time of the resulting product from th e U tube reactor to the NMR detector remained th e same across all experiments. T herefore the relaxation factor for different experiments remained the same allowing the NMR signal s resulting from different reactant gas mixtures to be compared
106 The NMR sig nal of the propane product was considered to be proportional to the reaction rate as we were detecting the the NMR signal in a continuously flowing product gas stream. The reaction rate, can be treated as (4 3 ) where I is the NMR peak integral value of p olarized 1 H NMR signals of the CH 2 and CH 3 group s in propane product, C is the conversion c oefficient is t h e detection time of a single scan and is the r elaxation factor Fig 4 1 shows the ALTADENA polarized 1 H NMR signals of the CH 2 and CH 3 group s of propane produce d by propylene hydrogenation using both normal hydrogen and parahydrogen. For both normal hydrogen and parahydrogen hydrogenation reaction s the reagent mixture flow rates f or nitrogen, hydrogen and propylene gas were 120, 150, 30 mL/min respectively. The p olarized 1 H NMR signals of the CH 2 and CH 3 group s were successfully detected following the parahydrogen hydrogenation reaction, at their characteristic chemical shift s as shown by numbers in Fig 4 1 T he NMR peak positions for hydrogen in either propylene (resulting from hydrogenation of propyne, present as an impurity), or product propane are labeled with a number on the NMR spectra The intensities of the hyper polarized signals are strongly enhanced relative to those obtained with normal hydrogen and propylene reaction, as evidenced by pronounced peaks of both CH 2 ( labeled as peak 2 in Fig 4 1 and CH 3 peaks (labeled as peak 1 in Fig 4 1 ) on the resulting product propane. The Pt/TiO 2 catalyst clearly lead s to ALTADENA signals as parahydrogen flow through the reaction bed.
107 4.3 3 NMR Signal Enhancement of PHIP and Pairwise Hydrogen Addition Selectivity We also observed that the PHIP enhanced NMR signal is sensitive to tempera ture. Fig 4 2 shows the ALTADENA polarized 1 H NMR signals at different temperature for rea ctant mixture flow rates of 120, 150, 30 mL/min for nitrogen, parahydrogen and propylene gas, res pectively. As the temperature increased from 100 to 250 C the NMR signals for both the CH 2 and CH 3 group s of propane co ntinuously increased monotonically. PHIP spectra were used in this study to investigate the selectivity to pair wise addition of hydrogen in both normal hydrogen and parahydrogen hydrogenation reactions. The propane signal from 1 H NMR spectra can be treated as a summation of both non polarized (I non ) and polarized (I polar ) signals. The p olarized signal (I polar ) is the absolute area of th e polarized peak after substraction of the non polarized (I non ) peak area The experimental enhancement factors can be calculated according to Eq 4 4 : (4 4 ) The theoretical enhancement factor may be calculated using the equation [ 120 122 ] (4 5 ) w here is the Boltzmann constant is temperature, is the redu ced Planck constant is the gyromagnetic ratio for 1 H is the fraction of p H 2 in the m ixture of supplied hydrogen, is the number of equivalent protons, which equals 2 for CH 2 group; is the ratio of the integrals of the signal intens ity patterns for the entire multiplet of the
108 thermally polarized and hyperpolarized CH 2 group in ALTADENA experiment and is 1 in this study In our experiment, B 0 is 9.4 Tesla, is 0.5, is 2, and is 1. The selectivity f or pairwise hydro gen addition ( ) w as estimated as the ratio of the experimental ( ) and theoretical ( ) enhance ment factors, Eq (4 6 ) (4 6 ) Fig 4 3 presents the pairwise selectivities for propy lene hydrogenation over Pt/TiO 2 catalyst s at different temperature s P arawise selectivity exhibited an in crease with temperature rising from 50 150 C and reaches a maximum value plateau between 150 250 C before decreasing over the temperature range bet ween 250 C and 350 C According to the nature of the heterogeneous catalyst reaction, the enhancemen t factor and pairwise addition reflect s the overall apparent heterogeneous reaction mechanism, which is dictated by a combination of the intrinsic reactio n mechanism, surface species adsorption and desorption, and transport processes For the low temperature range (50 150 C ), the reaction rate for pairwise addtion is likely to be controlled by the kinetic reaction rate. In the middle temperature range (150 250 C ), the signal enhancement and pairwise selectivity reache d a platea suggest ing that the pairwise parahydrogen reaction is more likely controlled by a diffus ion effect, which means that transport of a certain species participating in the reaction be c a me the rate limiting step. In high temperature range (250 350 C ), decreas ed pairwise selectivities were observed as the temperature increased possibly be explained by rapid diffusion of the hydrogen on the s urface, thereby decreasing the pairwise addit ion. The detailed reaction mechanism, including the reaction pathway dominating pairwise hydrogen
109 addition in propylene hydrogenation need further investigation to reveal the reaction limiting step in each temperature range. 4. 4 Summary T his wo rk repres ents the study of heterogeneous propylene hydrogenation with parahydrogen employing platinum supported on titanium dioxide. It was found that the NMR signal enhancement depends on the reaction temperature, and different rate limiting steps for the pairwise parahydrogen addition were suggested for three different temperature ranges due to the observed dependency of pairwise selectivity on temperature. However, t he detailed reaction mechanism of h eterogeneous propylene hydrogenation with parahydrogen remains to be elucidated.
110 Figure 4 1 1 H NMR spectra of propylene hydrogenation with a) parahydrogen and b) normal hydrogen over Pt/TiO 2 catalyst at 150 C and 1atm Gas flow rate (total 300 m L /min): N 2 /p H 2 /Propylene=120/150/30 m L /min. Data were taken under steady state conditions at 1atm (NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab in the UF Chemistry Department)
111 Figure 4 2 3 D plots of 1H NMR spectra during propy lene hydrogenation with parahydrogen over Pt/TiO 2 catalyst at different temperatu res. Gas flow rate (total 300 mL /min): N 2 /p H 2 /Propylene=120/150/30 ml/min. Data were taken under steady state conditions and 1atm (NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab in the UF Chemistry Department)
112 Figure 4 3 Pairwise selectivities for propylene hydrogenation over Pt/TiO 2 catalyst at different temperature s Two sets of reactions were carried out with n H 2 or p H 2 respectively. The CH 2 group of the product propane was in tegrated for all calculations, and the i ntegral from n H 2 wa s subtracted from that of p H 2 and the result was treated as pur e PHIP contribution E nhancement factors were calculated bas ed on the ratio between the PHIP and thermal integrals with n H2. Pairwise selectivit ies w ere calculated based on the ratio of experimental enhancement factor over theoretical enhancement factor. Gas flow rate (total 300 m L /min): N2/p H2/Propylene=120/150/ 30 mL /min. Pressure : 1 atm. (NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab )
113 CHAPTER 5 KINETIC STUDY OF PROPYLENE HYDROGENATION OVER IR I DIUM CATALYST USING AN IN SITU HYPERPOLARIZAED NMR TECHNIQUE 5 .1 Background Parahydro gen induced polarization (PHIP) is a method for increasing the sensitivity of nuclear ma gnetic resonance (NMR) signals, thereby, extending the NMR technique to potential applications requiring detection of species present at low concentration PHIP can sig nificantly increase the population differences between nuclear spin energy levels, as a result of addition of two hydrogen atoms from parahydrogen with correlated nuclear spins to a substrate molecule through catalytic reaction. This method c an provide an enhancement of NMR signal s by several orders of magnitude [ 35 109 110 ] As the PHIP reaction requires pairwise addition of both atoms from the hydrogen molecule of parahydrogen to a substrate (to retain the spin correlation), the catalyst employed is a key factor in this type of reaction. As discussed in th e previous chapter, PHIP reaction s have traditionally utilized homogeneous hydrogenation catalysts [ 123 ] However, an obvious problem of employing homogenous cat alysts in PHIP reactions in potential biomedical applications is the presence of dissolved homogenous catalysts, which are difficult to separat e from the hydrogenation product. One promising approach to o vercom e this difficult is to use heterogeneous catal ysts in a gas solid phase PHIP reaction. Recently, PHIP signal enhancement was discovered using a heterogeneous catalyst. T he first results of this type w ere reported in 2008 by Kovtunov et al. [ 41 ] in propylene hydrogenation over Pt/Al 2 O 3 catalyst To our knowledge, only two kind of metals (Pt [ 41 ] Pd [ 124 125 ] ) have been studied as active supp orted metal catalysts for PHIP activity
114 As discussed in chapter 4, supported p latinum i s the most investigat ed heterogeneous catalyst of PHIP reaction s Platinum metal particle size effects [ 42 ] and the pairwise hydrogen addition reactio n order have been reported [ 120 ] For palladium catalyst s it was found that PHIP effect over depend s heavily on the support [ 124 ] Palladium supported on TiO 2 exhibit strongly polarized peaks for CH 3 and CH 2 groups of propane produced through propylene hydrogenation, while palladium supported on ZrO 2 and SiO 2 shows only thermally polarized signals. Iridium compounds are commonly used as homogenous catalyst s for PHIP h 3 ) 2 ) [ 39 ] and Ir(CO)(dppe)Br and Ir(CO)(dppe)(CN) [ 47 ] have demonstrated activity in PHIP reaction s Also, immobilized iridium complexes using [Ir(COD)Cl] 2 anchored on silica gels demonstrated catalytic activity towards generating a PHIP effect in the gas phase hydrogenation of propylene. Despite previous studies revealing that homogeneous iridium compounds are effective catalysts for PHIP reaction s, there has been no study reported regarding PHIP reaction s over heterogeneous iridiu m metal catalysts. In this study, we present the study of propylene hydrogenation reaction by PHIP using Ir on TiO 2 as the catalyst. T his is the first time the PHIP reaction effect was has been observed and studied on an iridium metal heterogeneous catalys t. Here we report our studies of the propylene hydrogenation reaction on Ir/TiO 2 using the ALTADENA protocol [ 126 ] where reactions a re carried out in the low magnetic field followed by transport into a high field for NMR detection. The discovery of PHIP effect on heterogeneous iridium c atalyst is an important contribution to the existing pool of heterogeneous catalysts that are active to PHIP.
115 5 .2 Experimental Methods 5.2.1 Catalyst Preparation The c atalyst used in this study iridium metal particle s supported on titanium (IV) oxide (T iO 2 ), was prepared by precipitation of h exachloroiridic acid hexahydrate (H 2 Ir Cl 6 6H 2 O) (Alfa Aesar) onto the supports from an aqueous solution by controlling solution pH using sodium hydroxide with the s ame method as described in Chapter 4.2.1 More specifically, 1.98 g of TiO 2 support (Alfa Aesar) was dispersed in 100 mL of deionized water under constant stirring. H 2 Ir Cl 6 6H 2 O (0.05 6 g) was dissolved deionized H 2 O and add ed to the support/water mixture drop wise. The mixture was titrated with a 2.5 mM NaOH solution until the pH value reache d 11. The catalyst dispersion solution was then aged over night under continuous stirring before it was titrated with acetic acid solution to pH 7. The mixture was d ispersed in deionized water again, and filtered a second time after stirring overnight. The redispersed catalyst was filtered and dried at 105 C overnight, and then calcined at 450 C for 3 hours. 5.2.2 I nducti vely coupled P lasma atomic E mission S pectrometry The actual metal loading s of prepared Ir catalysts were characterized by inductively coupled plasma atomic emission spectroscopy (ICP A ES) ( Perkin Elmer Optima 3200 RL ) In short, Ir/TiO 2 catalyst was fused with sodium peroxide at 500 C for 1 hour followed by dissolv ing in water and neutraliz ation with hydrochloric acid. The iridium species concentration in the fusion solution was analyz ed by ICP AES and then translated into actual iridium metal loading in the catalyst. 5.2.3 Temperature Programmed Reduction Measurements Temperature Programmed Reduction (TPR) measurement for Ir/TiO 2 catalyst s were performed using a Quantachrome ChemBET 3 000 instrument. For the
116 measurement, 200 mg of the catalyst was loaded in a quartz tube reactor and secured with plug of quartz wool at the gas flow outlet end. The temperature in the catalyst bed was monitored by an Omega K type thermocouple. The reductio n was performed at a heating rate of 10C per minute up to a temperature of 500C. A 5% H 2 /N 2 mixture at a total flow rate of 70 SCCM was used for reduction until no further hydrogen adsorption on the catalyst was observed. 5.2.4 Chemisorption Measurement In this study, c hemisorption measurements to characterize the active surface area, dispersion and particle size of the Ir/TiO 2 catalyst was performed on a ChemBET 3000 from Quantachrome Instruments. The fresh catalysts were first oxidized using 5% oxygen in helium at 350 C for 0.5 h and then outgassed in helium, and then the catalyst was reduced in 5% hydrogen in nitrogen at 350 C for 0.5 h The catalyst was outgassed in helium at 35 0 C for another 15 minutes in order to remove all physically ad sorbed h ydroge n on the catalyst surface before cooling to room temperature. The optimized temperature of 35 0 C was chosen for hydrogen reduction based on the information obtained from temperature programmed reduction (TPR) measurement s, which indicated that most of Ir species was reduced at 35 0 C The catalyst w as then subjected to CO adsorption measurements to determine the Ir surface area. Estimates of the Ir particle sizes were made from th e CO adsorption measurements using equation 4 1 [ 121 ] in which is the average stoichiometry: CO/Ir= 1 [ 127 128 ] ; k is the shape factor, and a value of 5 was chosen in this study, and is the surface density of metal atoms, assumed to be 1.16 10 15 atoms/cm 2 in
117 this study. Similarly, the iridium metal surface area c an be calculated according to equation 4 2 5.2.5 Transmission Electron Microscopy ( TEM ) Measurement Thin film s of Ir/TiO 2 solutions (dispersed in isopropanol) w ere drop cast onto 400 mesh ultrathin carbon film supported copper grids (TED PELLA,INC, USA). TEM micrographs were obtained using a JEOL Model JEM 2010F System. 5.2.6 NMR Studies Based o n Parahydrogen Induced Polarization ( PHIP ) A Bruker Avance 400 MHz NMR spectrometer was employed for all NMR experiments. A glass U tube reactor cell was place d on the top of the magnet and a tube furnace equipped with a K type thermocouple was used to control the reactor temperature. The thermocouple was placed in the outlet side of the gas flow in the reactor bed separat ed from the catalyst with a glass wool plug A 50 m g sample of the catalyst was used for each experiment Prior to experiments, the catalyst was reduced and activated under constant h ydrogen flow (100 m L /min) for 30 minutes at 350 C followed by purging with nitrogen to remove any physical ly adsorbed hydrogen. P recise control of reaction gas flow rates of nitrogen (as carrier gas), hydrogen (either normal or para hydrogen ) and propylene was achieved b y using mass flow controllers (Alicat Scientific Inc. ). Parah ydro gen ( p H 2 ) was produced in situ by passing normal h ydrogen ( n H 2 ) through a copper coil filled with activated cha rcoal and immersed in a liquid nitrogen (77 K) D ewar to produce a mixture consisting of 50% parahydrogen and 50% orthohydrogen isomers (referred this mixture as p H 2 ) The gas mixture flowed through the reactor at elevated temperatures (50 350 C) and then into the NMR tube positioned in the NMR probe. Nitrogen wa s introduced into the reaction gas mixture as the carrier gas and heat dissipat e r. NMR spectra were acquire d with 128 scans using a 90 degree
118 pulse when steady state reaction condition s were reached in the reactor T h e reactor setup was automatically controlled with a custom designed LabVIEW program. As the PHIP reactions were performed in a low magnetic field of and the product was transferred to the high magnetic field for detection the resulting spectra exhibited the ALTADENA type polarization pattern. F o r ALTADENA experiments, this reactor setup well controlled, very reproducible data w ere obtained due to t he precise control of m ass flow controllers (MFCs) and the reactor temperature. The measured NMR spectra were used to calculate the values of the steady state selectivity and the activity of PHIP after the reaction reached constant condition. 5.3 Results a nd Discussion 5.3.1 Catalyst Characterization Temperature programmed reduction ( TPR ) measurement s (Fig 5 1) provid ed further information about the Ir/TiO 2 catalyst. Result s showed that two hydrogen r eduction peaks associated with iridium were detected. O ne peak centered at 428K, which was assigned to the Ir species, while the other peak centered at 535K, was assigned to the Ir species interact ing strong ly with the TiO 2 support. This observation of two separate hydrogen reduction peaks agrees well with resul ts of previous work [ 129 ] Starting from 570K, a broad hydrogen reduction peak appeared and this peak was attributed to partial reduction of the TiO 2 support [ 129 ] Based on the TPR result s we used a hydrogen reduction treatment temperature of 350 C at which most of the Ir species would be re duced to metal but a minimum a mount of the T iO 2 support would be reduced, for both CO adsorption measurement s and hydrogen pretreatment for catalyst activation between hydrogenation reactions.
119 ICP analysis show ed th at the actual loading of metal i ridium supported on TiO 2 catalyst wa s 0 .54% (w/w%). Based on this metal loading and CO adsorption measurement, using Eq (4 1) and Eq (4 2), the dispersion and mean iridium metal parti cle size were calculated, with value of 70% and 1.17 nm respectively. Fig 5 2 shows a typical TEM image of the Ir/TiO 2 catalyst. The i r idium particles are difficult to identify by TEM since the ir idium metal loading and particle size in our catalyst are particularly small (~0.5% metal loading, ~1.1 nm in diameter respectively). Also, both iridium particles and t he support TiO 2 itself are crystalline, making it more difficult to distinguish metal particle s from support by crystal structure patterns. Small particle s of iridium metal were highly dispersed on the outermost TiO 2 surface, as can be seen refer ring to t he arrow in Fig 5 2 5.3 2 Temperature Dependence of NMR Signal Enhancement Factor and Pairwise Hydrogen Addition Selectivity For hydrogenation of propylene over Ir/TiO 2 catalyst at different temperature s the total flow rate of parahydrogen or normal hy drogen propylen e and carrier gas nitrogen was maintain ed The same gas flow rate s ensure d th at the travel time of the product from U tube reactor to NMR detector would be the same for all measurements resulting a constant signal relaxation rate for NMR signal (refer Eq 4 3) The NMR signal of propane product was considered to be proportional to the PHIP reaction rate as discussed in C hapter 4. As shown in Fig 5 3 peaks for both the CH 2 ( labeled as peak 2 in Fig 5 3 ) and CH 3 ( labeled as peak 1 in Fig 5 3 ) from resulting product propane molecule were pronounced and detected, providing strong evidence for the occurrence of PHIP effect during proplynene hydrogenation over Ir/TiO 2 catalyst. Fig 5 4 shows the ALTADENA
120 polarized 1 H NMR signals of the CH 2 grou p of propane at different temperature s. The flow rates of nitrogen, p arahydrogen and propylene gas we re 120,150, 30 mL/min respectively. The p olarized 1 H NMR signals of the CH 3 group and CH 2 group from propane were successfully detected, with characterist ic peaks (labeled as 1 and 2 in Fig 5 3 respectively). Fig 5 5 presents the pairwise selectivities for propylene hydrogenation over Ir/TiO 2 catalyst measured at a series of temperature s P airwise selectivit ies (also enhancement factor s) increase with t emperature over the 50 350 C range Comparing the pairwise hydrogen addition reactions over Ir/ TiO 2 and Pt/ TiO 2 catalyst, significant differences were found. At the same temperature, Ir/ TiO 2 demonstrated higher pairwise selectivities, as well as NMR signa l enhancement factors than that of Pt/ TiO 2 catalyst, suggesting that Ir/ TiO 2 is more active than Pt/ TiO 2 catalyst towards pairwise hydrogen addition in propylene hydrogenation. P airwise addition selectivity also increas es with temperature between 50 350 C for the iridium catalyst, while pairwise selectivity for Pt/TiO 2 shows a broad maximum between 150 250C before decreasing at h igher temperature (compare Fig 4 3 to Fig 5 5) It is possible that the observed difference of parahydrogen pairwise addition b etween two catalysts is due to differences in the metal support interactions The importance of metal support interactions in TiO 2 supported platinum catalyst is well known, and the influence on the PHIP activity has been discussed in the literature [ 124 ] 5.3 4 NMR Signal Enhancement Factor and Pairwise Hydrogen Addition Selectivity with Respect to Hydrogen Concentration The selectivity in hydrogenation reactions has been shown to be hydrogen concentration dependent. Studies of the relationship between pairwise selectivit ies and
121 hydrogen concentration, can provide information about the pairwise hydrogen addition hydrogenation reaction mechanism [ 130 ] For hydrogenation of propyle ne over Ir/TiO 2 catalyst, the partial pressure s of parahydrogen or normal hydrogen, propylene and nitrogen carrier gas nitrogen were varie d while maintaining the same total gas flow rate, thereby ensur ing the same travel time of resulting product and a con stant signal relaxation rate for NMR signal The NMR signal of the propane product was considered to be proportional to the PHIP reaction rate Th e simplified calculation of experimental NMR signal enhancement factor ( ) and pairwise addition selectivity ( ) neglect ed nuclear spin relaxation of the hyperpolarized products during travel time from the reaction site and NMR detection point. Relaxation would reduce the detected hyperpolarization signal, and thus the actual pairwise select ivity values we re expected to be larger than the observ ed values. Fig 5 6 presents pairwise addition selectivities for propylene hydrogenation over Ir/TiO 2 at different hydrogen partial pressure ( at 150 C ) P airwise a ddition selectivit ies also enhancement factor s, increase d in the hydrogen partial pressure range between 0.05 to 0.25 atm P airwise addition selectivit ies remain practically constant between 0.25 to 0.7 atm. In the hydrogen partial pressure range investig ated (0.05 0.7 atm), the enhancement factor and pairwise addition selectivity we re relatively insensitive to hydrogen concentration. 5.4 Summary and Discussion This chapter presents the first observation of the PHIP eff ect in propylene hydrogenation usin g a supported iridium catalyst It was found that on TiO 2 supported iridium is an effective catalyst for PHIP in propylene hydrogenation. The signal
122 enhancement factor s and hydrog en pairwise addition selectivities were found to increase with increasing tem perature over the range of temperatures investigated (50 350 C). Also, the signal enhancement factor and hydrogen pairwise addition selectivity were found to be relatively insensitive to hydrogen concentration. Heterogeneous catalysts such as supported m etal nanoparticles are often known to have several different types of active sites which can operate in parallel with significantly different catalytic behavior. The p airwise pathway has been reported to occur more re adily on sites with lower coordination numbers such at edges or corners [ 42 ] The possible mechanisms [ 124 ] leading to pairwise addition of parahydrogen over Ir/TiO 2 catalyst include : 1) a pure ly statistical effect in which a small por tion of parahydrogen molecules w ould add to propylene in a pairwise m anner; 2) the presence of other substance absorbed on the catalyst metal surface thereby creat ing a confined active area that force two H atoms from parahydrogen to stay close together on the metal surface; 3) certain type s of active sites, such as co r ners, and edges, which can operate as isolated active centers for pairwise addition of parahydrogen; 4) physically adsorbed pa rahydrogen (H H covalent bond not broken) participates in a pa irwise manner in the hydrogenation reaction. At present the mechanism of pairwise hydrogen addition is still not well understood. Further studies of reaction orders and activation energy is for both hydrogen pairwise addition and non pairwise addition ar e under way Also, f urther catalyst optimization and theoretical investigation using Density Functional Theory (DFT) to determine the underlying reaction mechanism are in progress. H eterogeneous
123 PHIP could provide continuously flowing hyperpolarized fluid s for potential applications in diffusion and novel imaging [ 131 ]
124 Figure 5 1 Hydrogen temperature programmed reduction profile for Ir/TiO2 catalyst. heat ing rate at 10C per minute up to a temperature of 500C Figure 5 2 TEM image of Ir/TiO 2 catalyst. 100,000X magnification.
125 Figure 5 3 1 H NMR spectra of propylene hydrogenation with a) parahydrogen and b) normal hydrogen over Ir/TiO 2 catalyst at 1 50 C. Gas flow rate (total 300 ml/min): N 2 /p H 2 /Propylene=120 /150/30 mL /min. Data were obtained under steady state conditions (NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab )
126 Figure 5 4 3 D plots for 1 H NMR spectra mea sured during propylene hydrogenation with parahydrogen over Ir/TiO 2 catalyst at different temperature. Gas flow rate (total 300 m L /min): N 2 /p H 2 /Propylene=120/150/30 m L /min. Data were taken under steady state conditions (NMR data obtained by Ronghui Zhou t hrough collaboration with Dr. Bowers lab )
127 Fig ure 5 5. Pairwise selectivit ies for propylene hydrogenation over Ir/TiO 2 catalyst at different temperature s Two sets of reactions were carried out with n H2 or p H 2 respectively. The CH 2 group of the p roduct propane was integrated fo r all calculations. Integral for n H 2 wa s subtracted from that of p H 2 which was treated as pure PHIP contribution. Enhancement factors were calculated based on the ratio between the PHIP and thermal integrals of n H 2 Pai rwise selectivity was calculated based on the ratio of experimental enhancement factor over theoretical enhancement factor. Gas flow rate (total 300 mL /min): N 2 /p H 2 /Propylene=120/150/30 m L /min. (Produced from NMR data obtained by Ronghui Zhou through coll aboration with Dr. Bowers lab )
128 Fig ure 5 6 P airwise selectivit ies for propylene hydrogenation over Ir/TiO 2 at different hydrogen partial pressure s at 150 C Two sets of reactions were carried out with n H 2 or p H 2 respectively. The CH 2 group of th e product propane was integrated for all calculations. The i ntegral from n H 2 was subtracted from that of p H 2 and the result was treated as pure PHIP contribution. Enhancement factors were calculated based on the ratio between the PHIP and thermal integr als of n H 2 Pairwise selectivity was calculated based on the ratio of experimental enhancement factor over theoretical enhancement factor. Gas flow rate (total 300 mL /min): Flow rate of propylene= 30 mL /min. (Produced from NMR data obtained by Ronghui Zhou through collaboration with Dr. Bowers lab )
129 CHAPTER 6 CONCLUSIONS This study cover ed two relatively different topics: 1) platinum catalysts in direct methanol fuel cell applications; and 2) platinum and iridium catalysts in parahydrogen induced pola rization measurements Several conclusions can be made from this research, and each of the four chapters provided a slightly different focus on these topics. Chapter 2 presented a comparison study of four commercial direct methanol fuel cell anode catalys ts in terms of performance and durability in an open cathode design fuel cell. This project was designed to facilitate a technical choice for anode catalyst employed in an actual prototype direct methanol fuel cell with a unique open cathode design. The lo w voltage accelerated durability test revealed that the Johnson Matthey (JM) ELE170 anode catalyst had the best durability among the four anode catalysts tested. The JM ELE147 anode had the best beginning of life performance, but demonstrated the largest d egradation rate among the four anode catalysts investigated. Cabot and Tanak a anodes both demonstrated more rapid degradation compared with the ELE170. Based on the durability test, it was shown that ELE170 is the best anode candidate among the four anode catalysts investigated in the open cathode design fuel cell setup. Chapter 3 focus sed on understanding the effect of methyl ethyl ketone (MEK), a wetting agent commonly used in the manufactur ing of direct methanol fuel cell s on the cathode catalyst i.e the oxygen reduction reaction (ORR) o f the direct methanol fuel cell. A rotating disk electrode experimental set up and various electrochemical methods
130 were employed to investigate the ORR at the cathode ca talyst when MEK was present. The rotating disk electrode data show ed that upon introduction of 1 M 1 mM MEK into the acidic electrolyte the rate of oxygen reduction can be decreased by ~15 40% in terms of kinetic current density (corrected for transport limitation), translat ing to an overpotential lo ss of ~15 30 mV In addition, turnover frequency analysis demonstrated that the TOF dropped by ~30 40% after 1 M 1 mM MEK was added to the electrolyte solution To explain the contamination effects it was proposed that the MEK adsorbate act s as a site blocking species which reduce s the number of available active Pt surface sites for ORR activity. This study is the first report on the contamination effect of MEK on DMFC cathode s and mitigation procedures were suggested to avoid performance loss due to contamination. Chapter s 4 and 5 focus on the studies of heterogeneous propylene hydrogenation with parahydrogen over titanium dioxide supported platinum and iridium catalysts respectively. PHIP is an important technique for enhancing the NMR signal and ov ercoming its inherently low sensitivity. The PHIP reaction over heterogeneous catalysts was recently discovered, and there have been very few studies reported i n this field. Chapter 4 presented the study of heterogeneous propylene hydrogenation with parahy drogen over a Pt/TiO 2 heterogeneous catalyst and Chapter 5 presented the first observation of PHIP effect s in propylene hydrogenation over a heterogeneous iridium catalyst It was found that the NMR signal enhancement depends on the reaction temperature Because this temperature dependence, different rate limiting steps for pairwise parahydrogen addition were suggested for the three different temperature ranges. The Ir/TiO 2 heterogeneous catalyst also showed the dependency of NMR signal
131 enhancement and pai rwise hydrogen selectivity on temperature, with both increasing with temperature between 50 350C. NMR signal enhancement and pairwise hydrogen selectivity were found to be relatively insensitive to hydrogen concentration change (0.05 0.7 atm). The mechani sm of pairwise hydrogen addit ion on a heterogeneous catalyst is still not well understood at this time. A few suggestions can be made for continued work to develop an understanding of the pairwise hydrogen addition mechanism in the PHIP reaction : 1) inves tigation of reaction kinetics, such as reaction orders and activation energies, for both pairwise hydrogen addition and non pairwise hydrogen addition over heterogeneous catalysts; 2 ) Propose possible surface reaction steps and use Density Functional Theor y (DFT) to develop theoretical values of kinetic parameters of pairwise hydrogen addition for the proposed reaction steps. By comparing theoretically predicted kinetic parameters with experimental value s, validation of the proposed m echanism models can be achieved; 3 ) Continue heterogeneous catalyst optimization and characterization to reveal relationships between metal catalyst properties and pairwise hydrogen addition in the PHIP reaction. Catalyst metal particle size, crystal planes, metal support effect s, and binary/ternary mixtures of active metals can influence the pairwise hydrogen addition reaction and should be investigated. Understanding the relationship between various catalysts and pairwise hydrogen addition kinetics w ill provide important infor mation about the underl ying reaction mechanisms, and will ultimately contribute to the develop ment of PHIP into a mature NMR signal enhancing technology for various NMR applications.
132 LIST OF REFERENCES  D.T. Thompson, Platinum Meta ls Review 48 (2004) 169 172.  H.U. Blaser, H.P. Jalett, M. Muller, M. Studer, Catal Today 37 (1997) 441 463.  Encyclopdia Britannica Online, Encyclopdia Britannica Inc., Web, 24 April 2012.  N.D. Klerk, Platinum Metals Review 54 (2010) 120 121.  D.R. In Lide, CRC Handbook of Chemistry and Physics, CRC Press, New York, 2007.  B.D.A. Craig, David S;, Handbook of corrosion data, ASM International, Russell Township, Geauga County, Ohio, 1995.  A. Fornalczyk, M. Saternus, Metalurgija 50 (2011 ) 261 264.  T. Yamanaka, Int Chem Eng 11 (1971) 127 &.  R.E. Krebs, "Platinum". The History and Use of our Earth's Chemical Elements, Greenwood Press, 1998.  A. Wold, W. Kunnmann, A. Ferretti, R.J. Arnott, Inorg Chem 3 (1964) 545 &.  Z.G. Yan g, G.G. Xia, P. Singh, J.W. Stevenson, J Power Sources 155 (2006) 246 252.  Rosenber.B, L. Vancamp, T. Krigas, Nature 205 (1965) 698 &.  G.J.K. Acres, J Power Sources 100 (2001) 60 66.  A.S. Arico, S. Srinivasan, V. Antonucci, Fuel Cells 1 (200 1) 133 161.  B. Fan, Y.G. Guo, L.J. Wan, Prog Chem 22 (2010) 852 860.  S. Wasmus, A. Kuver, J Electroanal Chem 461 (1999) 14 31.  D.R. Rolison, P.L. Hagans, K.E. Swider, J.W. Long, Langmuir 15 (1999) 774 779.  H.S. Liu, C.J. Song, L. Zhang, J.J. Zhang, H.J. Wang, D.P. Wilkinson, J Power Sources 155 (2006) 95 110.  D.A. Blom, J.R. Dunlap, T.A. Nolan, L.F. Allard, J Electrochem Soc 150 (2003) A414 A418.
133  M.K. Jeon, K.R. Lee, K.S. Oh, D.S. Hong, J.Y. Won, S. Li, S.I. Woo, J Power Source s 158 (2006) 1344 1347.  C. Eickes, P. Piela, J. Davey, P. Zelenay, J Electrochem Soc 153 (2006) A171 A178.  P. Piela, C. Eickes, E. Brosha, F. Garzon, P. Zelenay, J Electrochem Soc 151 (2004) A2053 A2059.  Y.S. Chung, C. Pak, G.S. Park, W.S. J eon, J.R. Kim, Y. Lee, H. Chang, D. Seung, J Phys Chem C 112 (2008) 313 318.  E. Antolini, T. Lopes, E.R. Gonzalez, J Alloy Compd 461 (2008) 253 262.  P.J. Ferreira, G.J. la O', Y. Shao Horn, D. Morgan, R. Makharia, S. Kocha, H.A. Gasteiger, J Elec trochem Soc 152 (2005) A2256 A2271.  C.A. Reiser, L. Bregoli, T.W. Patterson, J.S. Yi, J.D.L. Yang, M.L. Perry, T.D. Jarvi, Electrochem Solid St 8 (2005) A273 A276.  A. Collier, H.J. Wang, X.Z. Yuan, J.J. Zhang, D.P. Wilkinson, Int J Hydrogen Energ 31 (2006) 1838 1854.  A. Heinzel, V.M. Barragan, J Power Sources 84 (1999) 70 74.  X.S. Zhao, G.Q. Sun, L.H. Jiang, W.M. Chen, S.H. Tang, B. Zhou, Q. Xin, Electrochem Solid St 8 (2005) A149 A151.  I.I. Rabi, J.R. Zacharias, S. Millman, P. Kusc h, Phys Rev 53 (1938) 318 318.  J. Jokisaari, Y. Hiltunen, Mol Phys 50 (1983) 1013 1023.  P. Mansfield, A.A. Maudsley, Brit J Radiol 50 (1977) 188 194.  A. Abragam, M. Goldman, Rep Prog Phys 41 (1978) 395 467.  J. Brossel, A. Kastler, Cr He bd Acad Sci 229 (1949) 1213 1215.  T.C. Eisenschmid, R.U. Kirss, P.P. Deutsch, S.I. Hommeltoft, R. Eisenberg, J. Bargon, R.G. Lawler, A.L. Balch, J Am Chem Soc 109 (1987) 8089 8091.  T. Jonischkeit, K. Woelk, Adv Synth Catal 346 (2004) 960 969. [37 ] S.K. Hasnip, S.A. Colebrooke, C.J. Sleigh, S.B. Duckett, D.R. Taylor, G.K. Barlow, M.J. Taylor, J Chem Soc Dalton (2002) 743 751.  J.A. Osborn, F.H. Jardine, J.F. Young, Wilkinso.G, J Chem Soc A (1966) 1711 &.
134  L. Vaska, J.W. Diluzio, J Am Chem Soc 83 (1961) 2784 &.  I.V. Koptyug, K.V. Kovtunov, S.R. Burt, M.S. Anwar, C. Hilty, S.I. Han, A. Pines, R.Z. Sagdeev, J Am Chem Soc 129 (2007) 5580 5586.  K.V. Kovtunov, I.E. Beck, V.I. Bukhtiyarov, I.V. Koptyug, Angew Chem Int Edit 47 (2008) 149 2 1495.  V.V. Zhivonitko, K.V. Kovtunov, I.E. Beck, A.B. Ayupov, V.I. Bukhtiyarov, I.V. Koptyug, J Phys Chem C 115 (2011) 13386 13391.  L.S. Bouchard, K.V. Kovtunov, S.R. Burt, M.S. Anwar, I.V. Koptyug, R.Z. Sagdeev, A. Pines, Angew Chem Int Edit 4 6 (2007) 4064 4068.  V.V. Telkki, V.V. Zhivonitko, S. Ahola, K.V. Kovtunov, J. Jokisaari, I.V. Koptyug, Angew Chem Int Edit 49 (2010) 8363 8366.  K.V. Kovtunov, V.V. Zhivonitko, A. Corma, I.V. Koptyug, J Phys Chem Lett 1 (2010) 1705 1708.  K.V. Kovtunov, V.V. Zhivonitko, L. Kiwi Minsker, I.V. Koptyug, Chem Commun 46 (2010) 5764 5766.  C.E. Johnson, R. Eisenberg, J Am Chem Soc 107 (1985) 6531 6540.  A.M. Argo, J.F. Odzak, J.F. Goellner, F.S. Lai, F.S. Xiao, B.C. Gates, J Phys Chem B 110 ( 2006) 1775 1786.  M.W. Ellis, M.R. Von Spakovsky, D.J. Nelson, P Ieee 89 (2001) 1808 1818.  L. Carrette, K.A. Friedrich, U. Stimming, Chemphyschem 1 (2000) 162 193.  R. O'Hayre, S.W. Cha, W. Colella, F.B. Prinz, Fuel Cell Fundamentals, Wiley, 2 009.  C. Spiegel, Designing and Building Fuel Cells, Mcgraw hill, 2007.  J. Larminie, A. Dicks, Fuel cell systems explained, J. Wiley, 2003.  G. Xiao, Fuel Cell Technology, Publishing House of Electronics Industry, Beijing, China, 2009.  D. N. Prater, J.J. Rusek, Appl Energ 74 (2003) 135 140.  G.J.K. Acres, J.C. Frost, G.A. Hards, R.J. Potter, T.R. Ralph, D. Thompsett, G.T. Burstein, G.J. Hutchings, Catal Today 38 (1997) 393 400.  P.S. Kauranen, E. Skou, J. Munk, J Electroanal Chem 40 4 (1996) 1 13.
1 35  H.N. Dinh, X.M. Ren, F.H. Garzon, P. Zelenay, S. Gottesfeld, J Electroanal Chem 491 (2000) 222 233.  S.Q. Song, Z.X. Liang, W.J. Zhou, G.Q. Sun, Q. Xin, V. Stergiopoulos, P. Tsiakaras, J Power Sources 145 (2005) 495 501.  E. Gui lminot, A. Corcella, F. Charlot, F. Maillard, M. Chatenet, J Electrochem Soc 154 (2007) B96 B105.  J. Xie, D.L. Wood, K.L. More, P. Atanassov, R.L. Borup, J Electrochem Soc 152 (2005) A1011 A1020.  D.L. Wood, Y.S. Yi, T.V. Nguyen, Electrochim Acta 43 (1998) 3795 3809.  J. Xie, D.L. Wood, D.M. Wayne, T.A. Zawodzinski, P. Atanassov, R.L. Borup, J Electrochem Soc 152 (2005) A104 A113.  R.F. Mann, J.C. Amphlett, B.A. Peppley, C.P. Thurgood, J Power Sources 161 (2006) 775 781.  Y.Y. Shao, G.P Yin, Y.Z. Gao, P.F. Shi, J Electrochem Soc 153 (2006) A1093 A1097.  J.G. Liu, Z.H. Zhou, X.X. Zhao, Q. Xin, G.Q. Sun, B.L. Yi, Phys Chem Chem Phys 6 (2004) 134 137.  J. St Pierre, D. Wilkinson, Aiche J 47 (2001) 1482 1486.  D.L. Trimm, Z.I. O nsan, Catal Rev 43 (2001) 31 84.  A. Schafer, J.B. Heywood, M.A. Weiss, Energy 31 (2006) 2064 2087.  J.H. Wee, J Power Sources 173 (2007) 424 436.  C.S. Macomber, H. Wang, K. O'Neill, J. Christ, G. Bender, B. Pivovar, H.N. Dinh, Abstr Pap Am Ch em S 242 (2011).  H. Tawfik, K. El Khatib, Y. Hung, D. Mahajan, Ind Eng Chem Res 46 (2007) 8898 8905.  B. Balasubramanian, A.L. Ortiz, S. Kaytakoglu, D.P. Harrison, Chem Eng Sci 54 (1999) 3543 3552.  N. Fann, C.M. Fulcher, B.J. Hubbell, Air Qua l Atmos Hlth 2 (2009) 169 176.  J. St Pierre, D.P. Wilkinson, S. Knights, M.L. Bos, J New Mat Elect Syst 3 (2000) 99 106.  J. Twidell, A.J. Appleby, Philos T R Soc A 354 (1996) 1693 1693.
136  Y.M. Volfkovich, V.E. Sosenkin, Russ Chem Rev+ 81 (2012 ) 936 959.  D. Rolle, J.W. Schultze, Electrochim Acta 31 (1986) 991 1000.  N.K. Tumanova, Y.K. Delimarskii, K.M. Taranishina, Ukr Khim Zh+ 48 (1982) 994 996.  V.A. Grinberg, Y.B. Vasilev, Sov Electrochem+ 27 (1991) 542 548.  J.O.M. Bockris, K.T. Jeng, J Electroanal Chem 330 (1992) 541 581.  C.F. Zinola, J.L. Rodriguez, M.C. Arevalo, E. Pastor, J Electroanal Chem 454 (1998) 161 172.  M.P. Hyman, J.W. Medlin, J Phys Chem B 109 (2005) 6304 6310.  T. Lopes, J. Chlistunoff, J.M. Sansi nena, F.H. Garzon, Int J Hydrogen Energ 37 (2012) 5202 5207.  N.P. Subramanian, T.A. Greszler, J. Zhang, W. Gu, R. Makharia, J Electrochem Soc 159 (2012) B531 B540.  N.M. Markovic, T.J. Schmidt, V. Stamenkovic, P.N. Ross, Fuel Cells 1 (2001) 105 11 6.  T.J. Schmidt, U.A. Paulus, H.A. Gasteiger, R.J. Behm, J Electroanal Chem 508 (2001) 41 47.  D.R. Lawson, L.D. Whiteley, C.R. Martin, M.N. Szentirmay, J.I. Song, J Electrochem Soc 135 (1988) 2247 2253.  K. Yamamoto, D.M. Kolb, R. Kotz, G. Le hmpfuhl, J Electroanal Chem 96 (1979) 233 239.  B.E. Hayden, A.J. Murray, R. Parsons, D.J. Pegg, J Electroanal Chem 409 (1996) 51 63.  E. Herrero, J.M. Feliu, A. Aldaz, J Electroanal Chem 368 (1994) 101 108.  T.J. Schmidt, H.A. Gasteiger, G.D. Stab, P.M. Urban, D.M. Kolb, R.J. Behm, J Electrochem Soc 145 (1998) 2354 2358.  S.L. Gojkovic, S.K. Zecevic, R.F. Savinell, J Electrochem Soc 145 (1998) 3713 3720.  V.S. Murthi, R.C. Urian, S. Mukerjee, J Phys Chem B 108 (2004) 11011 11023.  E Higuchi, H. Uchida, M. Watanabe, J Electroanal Chem 583 (2005) 69 76.
137  J. Newman, J Electrochem Soc 113 (1966) 501 &.  G. Hunter, R.M.M.C.R.C.V.B. G. Hunter, Sensors, Actuators, and Microsystems (General) 213th ECS Meeting, Electrochemical Soci ety, 2008.  N. Markovic, H. Gasteiger, P.N. Ross, J Electrochem Soc 144 (1997) 1591 1597.  L. Gancs, B.N. Hult, N. Hakim, S. Mukerjee, Electrochem Solid St 10 (2007) B150 B154.  A. Damjanovic, D.B. Sepa, M.V. Vojnovic, Electrochim Acta 24 (197 9) 887 889.  D.B. Sepa, M.V. Vojnovic, A. Damjanovic, Electrochim Acta 25 (1980) 1491 1496.  E. Yeager, Electrochim Acta 29 (1984) 1527 1537.  E. Yeager, J Mol Catal 38 (1986) 5 25.  B.N. Grgur, N.M. Markovic, P.N. Ross, Can J Chem 75 ( 1997) 1465 1471.  U.A. Paulus, T.J. Schmidt, H.A. Gasteiger, R.J. Behm, J Electroanal Chem 495 (2001) 134 145.  U.A. Paulus, A. Wokaun, G.G. Scherer, T.J. Schmidt, V. Stamenkovic, V. Radmilovic, N.M. Markovic, P.N. Ross, J Phys Chem B 106 (2002) 4181 4191.  C.F. Zinola, A.M.C. Luna, J Colloid Interf Sci 219 (1999) 260 274.  G.E. T Engel, D.D Eley, H Pines, P.B Weisz (Eds.), Advances in Catalysis, Academic Press, New York, 1979.  C.R. Bowers, D.P. Weitekamp, Phys Rev Lett 57 (1986) 2 645 2648.  C.R. Bowers, D.P. Weitekamp, J Am Chem Soc 109 (1987) 5541 5542.  R. Eisenberg, Accounts Chem Res 24 (1991) 110 116.  C.R. Bowers, D.P. Weitekamp, Solid State Nucl Mag 11 (1998) 123 128.  C.R. Bowers, Encyclopedia of Nuclear Magnetic Resonance, John Wiley & Sons, Chichester, 2002.  P. Hubler, R. Giernoth, G. Kummerle, J. Bargon, J Am Chem Soc 121 (1999) 5311 5318.  S.B. Duckett, C.L. Newell, R. Eisenberg, J Am Chem Soc 116 (1994) 10548 10556.
138  Y. Cohen, L. Avram L. Frish, Angew Chem Int Edit 44 (2005) 520 554.  C.P. Grey, N. Dupre, Chem Rev 104 (2004) 4493 4512.  P.S. Pregosin, P.G.A. Kumar, I. Fernandez, Chem Rev 105 (2005) 2977 2998.  H.C. Keun, Pharmacol Therapeut 109 (2006) 92 106.  O.G. Salnikov, K.V. Kovtunov, D.A. Barskiy, V.I. Bukhtiyarov, R. Kaptein, I.V. Koptyug, Appl Magn Reson 44 (2013) 279 288.  L.M. Neal, D. Hernandez, H.E. Hagelin Weaver, J Mol Catal a Chem 307 (2009) 29 36.  C.R. Bowers, Encyclopedia of Magnetic Reson ance, John Wiley & Sons, Ltd, 2002.  I.V. Skovpin, V.V. Zhivonitko, R. Kaptein, I.V. Koptyug, Appl Magn Reson 44 (2013) 289 300.  K.V. Kovtunov, V.V. Zhivonitko, I.V. Skovpin, D.A. Barskiy, I.V. Koptyug, Topics in current chemistry (2012).  K.V. Kovtunov, I.E. Beck, V.V. Zhivonitko, D.A. Barskiy, V.I. Bukhtiyarov, I.V. Koptyug, Phys Chem Chem Phys 14 (2012) 11008 11014.  M.G. Pravica, D.P. Weitekamp, Chem. Phys. Lett. 145 (1988) 255 258.  G. Corro, R. Gomez, React Kinet Catal L 9 ( 1978) 325 330.  S.J. Tauster, S.C. Fung, R.T.K. Baker, J.A. Horsley, Science 211 (1981) 1121 1125.  P. Reyes, D. Salinas, C. Campos, M. Oportus, J. Murcia, H. Rojas, G. Borda, J.L.G. Fierro, Quim Nova 33 (2010) 777 780.  P.G.M. Zoltn Pal, Hydrogen Effects in Catalysis: Fundamentals and Practical Applications, CRC Press, 1988.  L.S. Bouchard, S.R. Burt, M.S. Anwar, K.V. Kovtunov, I.V. Koptyug, A. Pines, Science 319 (2008) 442 445.
139 BIOGRAPHICAL SKETCH Wei Cheng was born in Xuancheng China in 19 83 He earned both his Bachelor of Science and Master of Science degrees both in the Department of Chemical Engineering at the University of Science and Technology in Shanghai in 200 4 and 200 7 respectively, under t he guidance of Dr. Bo Fang After serving in the Engineering Unit at BASF (Shanghai) Company for three months, Dr. Cheng began his doctoral studies in Chemical Engineering at the University of Florida (UF), under the guidance of Dr. Helena Halogen Weaver. Dr. Cheng is interested in c atalysis, fuel cell, and other energy related chemical engineering domains. He hopes to continue contributing his knowledge and skills learned at UF to the progress of chemical engineering research and the chemical industry