Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Design of durable and efficient catalysts for the electro-oxidation of methanol
(USC Thesis Other)
Design of durable and efficient catalysts for the electro-oxidation of methanol
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
I Design of Durable and Efficient Catalysts for the Electro-Oxidation of Methanol By Dan Fang A DISSERTATION SUBMITTED TO THE FACULTY OF THE USC GRADUATE SCHOOL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF CHEMISTRY UNIVERSITY OF SOUTHERN CALIFORNIA LOS ANGELES, CALIFORNIA MAY 2019 II Contents Dedication ...................................................................................................................................... VI Acknowledgements ....................................................................................................................... VII Summary ......................................................................................................................................... 1 Chapter 1 Introduction ................................................................................................................... 4 1.1 Background ....................................................................................................................... 4 1.2 Direct Methanol Fuel Cell ................................................................................................. 8 1.2.1 Proton-Exchange Membrane Direct Methanol Fuel Cell - Working Principle............ 10 1.2.2 Direct Methanol Fuel Cell Components and Assembly .............................................. 13 1.3 Catalysts for the Electro-Oxidation of Methanol ........................................................... 14 1.3.1 Platinum-Ruthenium Catalysts ......................................................................................... 14 1.3.2 Platinum Based Bi-functional Catalysts ............................................................................ 16 1.4 Sputter Deposition Technology ........................................................................................... 18 Chapter 2 Experimental Techniques and Analysis Methodologies .............................................. 21 2.1 Physical Characterization .................................................................................................... 21 2.1.1 Scanning Electron Microscopy ......................................................................................... 21 2.1.2 Energy Dispersive Analysis of X-rays ................................................................................ 21 2.1.3 X-ray Diffraction ............................................................................................................... 21 2.1.4 X-ray Photoelectron Spectroscopy ................................................................................... 22 III 2.2 Half-cell Electrochemical Measurements ...................................................................... 22 2.3 Full Cell Test ........................................................................................................................ 23 Chapter 3 Platinum-Tantalum Catalysts Prepared by Sputter Deposition ................................... 25 3.1 Preparation of Platinum-Tantalum Electrodes by Co-sputter Deposition .......................... 25 3.2 Physical Characterization of Co-sputter Platinum-Tantalum Catalysts .............................. 27 3.3 Electrochemical Characterization of Co-Sputtered Platinum-Tantalum Catalysts ............. 33 3.3.1 Electrochemical Surface Area .......................................................................................... 34 3.3.2 Methanol Oxidation Activity ............................................................................................ 38 3.3.3 Catalyst Stability ............................................................................................................... 43 3.3.4 Platinum-Tantalum Catalyst Methanol Oxidation Mechanism ....................................... 44 3.4 Summary ............................................................................................................................. 50 Chapter 4 Catalysts based on Co-sputtered Platinum-M (M=Niobium, Titanium, Zirconium) .... 52 4.1 Co-sputtered Platinum-Niobium Catalysts ......................................................................... 52 4.1.1 Preparation of Platinum-Niobium Catalysts by Co-sputter Deposition ........................... 52 4.1.2 Physical Characterization of Platinum-Niobium Catalysts ............................................... 53 4.1.3 Electrochemical Characterization of Platinum-Niobium Catalysts .................................. 58 4.1.4 Effects of Platinum to Niobium Ratio ............................................................................... 66 4.1.5 Co-sputtered Platinum-Niobium Catalyst Methanol Oxidation Mechanism ................... 69 4.2 Platinum-Titanium Catalysts ............................................................................................... 70 IV 4.2.1 Preparation of Platinum-Titanium by Co-Sputter deposition .......................................... 70 4.2.2 Physical Characterization of Platinum-Titanium Catalysts .............................................. 71 4.2.3 Electrochemical Characterization of Platinum-Titanium Catalysts .................................. 75 4.2.4 Effects of Platinum to Titanium Ratio .............................................................................. 82 4.3 Platinum-Zirconium Catalysts ............................................................................................. 85 4.3.1 Preparation of Platinum-Zirconium by Co-sputter Deposition ........................................ 85 4.3.2 Physical Characterization of Platinum- Zirconium Catalysts ............................................ 86 4.3.3 Electrochemical Characterization of Platinum- Zirconium Catalysts ............................... 90 4.3.4 Effects of Platinum to Zirconium Ratio ............................................................................ 98 4.4 Comparison of Platinum-M (M=Tantalum, Niobium, Titanium, Zirconium) Catalysts ..... 101 4.4.1 Electrochemical Surface Area ........................................................................................ 102 4.4.2 Methanol Oxidation Activities ........................................................................................ 102 Chapter 5 Properties of Membrane Electrode Assemblies Employing Sputter-deposited Electrocatalysts for Methanol Oxidation .................................................................................... 106 5.1 Membrane Electrode Assembly Preparation .................................................................... 107 5.2 Membrane Electrode Assembly Performance of Co-sputter Pt-Ta Catalysts ................... 107 5.3 Effects of Carbon Catalyst Supports .................................................................................. 110 5.3.1 The Effect of CNT Coating Techniques ........................................................................... 111 5.3.2 The Effect of CNT Loading .............................................................................................. 116 V 5.4 Effects of Pressure during Sputtering. .............................................................................. 118 5.5 Effects of Heating Substrate during Sputter Deposition ................................................... 124 5.6 Effects of Catalysts Wettability ......................................................................................... 126 5.7 Effects of Amount of Nafion Ionomer. .............................................................................. 130 5.8 The Optimized Preparation of Co-sputtered Pt-Ta/CNTs electrodes ............................... 131 Chapter 6 Platinum-Ti 0.7Ru 0.3O 2 Catalysts ................................................................................... 133 6.1 Introduction ....................................................................................................................... 133 6.2 Pt-Ti 0.7Ru 0.3O 2 Catalyst Synthesis ....................................................................................... 134 6.2.1 Hydrothermal Method for Synthesizing Ti 0.7Ru 0.3O 2 ...................................................... 135 6.2.2 Polyol Method synthesizing Pt-Ti 0.7Ru 0.3O 2 .................................................................... 135 6.3 Results and Discussion ...................................................................................................... 136 6.3.1 Ti 0.7Ru 0.3O 2 Nanoparticles ............................................................................................... 136 6.3.2 Pt-Ti 0.7Ru 0.3O 2 Catalysts .................................................................................................. 138 6.4 Summary ........................................................................................................................... 142 Reference ................................................................................................................................ 143 VI Dedication I dedicate this thesis to my parents, whose love, unselfish support and examples over many years laid the foundation for the discipline and application necessary to complete this work. VII Acknowledgements Thank you, Professor Narayan, for your mentorship over the past years I have been at University of Southern California. I have learned so much from you, not only the fundamentals and intricacies of electrochemistry, but also efficient and safe laboratory practices, creative problem- solving methodologies, and most importantly, to always have patience and perseverance. I am truly grateful to have had this chance to work with someone as passionate, knowledgeable, and understanding as you. Thank you, all the members of the Narayan group, for making our research group such an enjoyable place to be. Thank you for always being respectful and professional colleagues while we worked together in the lab. Thank you for being such a great group of friends outside the lab. I had a lot of fun travelling with you guys all over LA and across the states. I could not have asked for better group members. Thank you, Professor Surya Prakash, Dr. Robert Aniszfeld, and the Loker Hydrocarbon Institute, for all the support I received while I worked on my projects. Thank you for always letting me use your facilities, equipment, and resources all these years. Thank you for allowing me to be a part of such a welcoming scientific community. Thank you, all my families and friends, for always being there for me. Thank you for making me feel as though I’ve never left home. Thank you for always supporting me and helping me get through the difficult of times. 1 Summary This work presented in this thesis focuses on designing electrocatalysts for methanol oxidation and understanding the catalytic processes. The goal is to achieve more durable and efficient electrocatalysts for direct methanol fuel cells. In Chapter 1, I have discussed the growth of energy conversion technology and the necessity of energy storage technology. I have discussed different types of electrochemical energy systems. I have described the working principle of the direct methanol fuel cell and methanol oxidation catalysts. I have reviewed the challenges of methanol oxidation catalysts and what other researchers have done to solve these problems. In Chapter 2, I have described the experimental techniques and methodologies that were used in performing the research. These experimental techniques include various physical and electrochemical characterization techniques and the methodologies for catalyst preparation including the hydrothermal method, polyol method and sputter deposition method. In Chapter 3, I have described the properties of a series of thin film platinum-tantalum (Pt-Ta) catalysts prepared by the sputter deposition method. Characterization of these thin film catalysts was performed using scanning electron microscopy, X-ray diffraction, energy dispersive analysis of X-rays and X-ray photoelectron spectroscopy. Assessment of the methanol oxidation activity of platinum-tantalum catalysts was carried out using half-cell experiments in three-electrode cells. The Pt-Ta catalyst with Ta:Pt atomic ratio of 1: 0.298 showed electrochemical area- specific activity comparable to commercially- available PtRu/C catalysts. We also found that the atomic ratio of platinum to tantalum in the catalyst affects the platinum-normalized methanol oxidation 2 activity. This suggests that Pt-Ta catalyst has a different methanol oxidation mechanism compared to platinum. The surface oxides species were found to activate the water molecules and hence facilitate the process of removing carbon monoxide from the platinum sites. In Chapter 4, I have described as to how the co-sputtered Pt-M (M=Nb, Ti, Zr) catalysts were synthesized by co-sputtering. The physical properties and electrochemical properties of the co- sputtered Pt-M (M=Nb, Ti, Zr) catalysts have been analyzed. The effects of Pt to M ratio are also discussed in this chapter. We have also compared Pt-Nb, Pt-Ti, Pt-Zr with the Pt-Ta catalysts we discussed in Chapter 3. Among all the catalysts, co-sputtered Pt-Ta catalysts shows the best methanol oxidation activity and stability. In Chapter 5, I have discussed the factors that affect the catalyst materials utilization when we apply co-sputtered Pt-Ta catalysts to a full fuel cell consisting of two electrodes. In these experiments, we keep the amount of catalyst consistent and aim at achieving the highest methanol oxidation current. Firstly, we increased the surface area of existed carbon substrate that was used in the previous half-cell experiments by coating the carbon nanotubes on a carbon fiber composite paper. The factors that affected the final carbon structure including the pre- treatment of CNTs, the coating techniques and the loading of CNTs were investigated. Secondly, I explored the effect of sputter pressure on the utilization of sputter materials. Thirdly, the effects of heating the sputter substrate during deposition to improve atom surface atom mobilities, has been discussed. Fourthly, the wettability of the resulting electrodes were studied. Finally, the effect of changing the ratio of Nafion ionomer to the catalyst materials was investigated. In Chapter 6, I have discussed the preparation and properties of Pt-Ti 0.7Ru 0.3O 2 for use as an anode catalyst. I have examined the use of titanium oxide-ruthenium oxide solid solutions as 3 catalyst supports that also act as a co-catalyst. Ti 0.7Ru 0.3O 2 as was synthesized by the hydrothermal process and then platinum was deposited using the polyol method. The structure of the synthesized catalyst was investigated by XRD, SEM and EDX. Results of steady-state voltammetry suggested that the electrocatalytic activity and stability of the newly synthesized catalysts compared well with the commercially--available carbon-supported Pt-Ru catalyst. The Tafel slopes suggested that the bifunctional mechanism for methanol oxidation was operative on the new catalyst. The research has met the original goals by demonstrating that the durable and efficient electrocatalyst alternatives to the commercially--available platinum-ruthenium catalysts can be achieved by combing platinum with tantalum, niobium and titanium using the sputter-deposition method. Also, catalytic supports based on the mixed oxides of titanium and ruthenium are promising. 4 Chapter 1 Introduction 1.1 Background Throughout history, energy has been the foundation for world development and is now the life blood of the world’s economy. Fig 1-1 shows that the annual world consumption of energy stayed lower than 1,000 million tonnes of oil equivalents (MTOE) before the 20th century. There was a dramatic increase to 13,000 MTOE between the year 1900 to 2010. Before the year 2040, the total energy consumption will reach 17000 MTOE/year as shown in Fig. 1-2. The contribution from renewable energy (grey area in Fig.1-2) is expected to increase dramatically. Right now, most of the energy still comes from fossil fuels which causes the increase in carbon dioxide levels in the atmosphere. Hence taking fully advantage of renewable energy, for example, using those fuels that can be produced renewably using carbon dioxide and sunlight, will be an important part of the effort to reduce carbon emissions. 5 Figure 1-1 The growth in the world’s total energy consumption split by sources from 1800 and into 2013 1 Figure 1-2 Projected global energy consumption from 1990 to 2040, by energy source (in million metric tons of oil equivalent) 2 One of the major limitations of renewable energy conversion technology is the intermittency of its source. Solar energy is only available during the day and wind energy is only available when the wind is blowing and this is highly variable. Yet the demand for energy is constantly present. To meet this constant demand of energy, energy storage technology must support energy conversion technology by providing energy when the demand of exceeds the supply. Another 6 major limitation of renewable energy conversion technology is its portability. The demand for energy is not always in locations connected to the power grid. Energy is required in many applications such as transportation and portable devices. Equipping such applications with renewable energy conversion technology is impractical. Thus, energy storage technology can be used to store converted energy and supply it at the location where it is needed. Fig. 1-3 is a Ragone plot that shows the characteristics of several electrochemical energy storage systems in terms of their specific power and specific energy. 3 Capacitors store energy within a temporarily generated electric field caused by the separation of charges. Capacitors can store only a small amount of energy, but this energy can be delivered very quickly. Batteries have reasonable specific energy and specific power. They are the energy storage technology of choice for portable devices like laptops, because they can be scaled to fit in a portable device to deliver the required power and energy for the devices while maintaining the ease of rechargeability. However, increasing demand for energy for such devices necessitate the use of systems that have a very high specific energy content. 7 Figure 1-3 Ragone plots of specific power versus specific energy for the various electrochemical energy storage devices and internal combustion engine Among the various of the energy technologies, fuel cells have the highest specific energy. Fuel cells are efficient, clean, quiet, safe and reliable energy conversion devices in which the chemical energy of a fuel is transformed directly to electrical energy through the electrochemical reaction with oxygen. Fuel cells have the potential reduce carbon emissions because of their high energy conversion efficiency. In fuel cells, the free energy change of the fuel oxidation reaction can be harnessed at constant temperature. Thus, unlike combustion engines, the efficiency of fuel cells is not limited by the Carnot limit for heat engines. Due to their scalable nature a wide range of applications including portable devices, transportation, and stationary applications (from a few W to MW) can be supported by fuel cell technology. 8 Typically, an energy storage device is tailored towards the application based on their properties. Fuel cells are generally limited in how quickly they can deliver energy. According to US Department of Energy Fuel Cell Technologies Office, the current market transformation efforts focus on several key early market applications including specialty vehicles, emergency backup power and prime power for critical loads. To meet the requirement of high power density in applications such as vehicles, a fuel cell can have a battery as a system component to store the electricity it’s generating. Thus, fuel cells and batteries work together to meet the combined high energy demand and power demand of devices. 1.2 Direct Methanol Fuel Cell Ever since its successful demonstration, direct methanol fuel cells (DMFC) have been considered a promising solution for providing electrical power source. 4 Direct-methanol fuel cell is a type of fuel cell that uses an aqueous solution of methanol as the fuel to produce electricity by electrochemical reaction of methanol and oxygen. 5 The methanol electro-oxidation was first observed by E. Muller in the early 1920 6 , and the concept of methanol fuel cells was first investigated 30 years later by Kordesch, Marko 7 and Pavela 8 . Alkaline electrolytes were initially used for methanol fuel cells; the search for the active anode and cathode catalysts mainly led to nickel or platinum for the methanol oxidation reaction MOR and silver for the oxygen reduction process 8 9 . Subsequent efforts by Hitachi and others led to acid-based methanol fuel cells. In the 1990s, the configuration of the acid-free version developed by JPL and USC triggered a renewed interest in this technology. Their contributions include: development of advanced catalyst materials, high-performance fuel cell membrane 9 electrode assemblies, compact fuel cell stacks, and system designs. 10 The 300-watt engineering prototype of a Direct Methanol Fuel Cell system that was developed by a team of scientists at NASA's Jet Propulsion Laboratory is shown in Fig. 1-4. Figure 1-4 A 300-watt engineering prototype of a Direct Methanol Fuel Cell system for defense applications developed by a team of scientists at NASA's Jet Propulsion Laboratory. Image credit: NASA/JPL-Caltech Fuel cells in which methanol is reformed to carbon dioxide and hydrogen before using in a fuel cell is called the indirect methanol fuel cell (IMFC), and is a subcategory of proton-exchange membrane-based methanol fuel cells. The advantages of IMFC are higher efficiency, smaller cell stacks and minimal water management. The tradeoff is that IMFC systems operate at hotter temperatures and therefore need more advanced heat management and insulation. A small 25 10 Watt IMFC system developed by UltraCell for the United States military is commercially- available. Larger systems (350W to 8 MW) are also available from SerEnergy for multiple applications, such as power plant generation, backup power generation and battery range extension. Yet another application is improving the performance of a heavy duty electric vehicle. Methanol fuel cells do not have many of the fuel storage problems typical of hydrogen-based fuel cell systems. Firstly, methanol has a higher volume energy density than hydrogen 11 . Methanol is also easier to transport and supply to the public using our current infrastructure because it is a liquid, like gasoline 12 . Besides, methanol is readily obtained today from methane, coal, and also by reduction of carbon dioxide. The “George Olah Carbon Dioxide (CO 2) to renewable methanol plant” is located at the Svartsengi geothermal power station in the town of Grindavik, Iceland is producing 5 million liters of methanol per year from carbon dioxide. Liquid methanol is already in use as an additive in transportation fuel in China. Methanol fuel cells can be expected to enable the “Methanol Economy” concept promoted by the late Nobel Laureate, Dr. George Olah. 1.2.1 Proton-Exchange Membrane Direct Methanol Fuel Cell - Working Principle The membrane electrode assembly (MEA) is at the core of the direct methanol fuel cell. The MEA consists of a proton exchange membrane (PEM) sandwiched between two catalytic electrodes. The catalytic electrodes sustain the electro-oxidation of methanol and the electro-reduction of oxygen. State-of-art direct methanol fuel cell systems are based on the use of platinum- ruthenium as a catalyst for the electro-oxidation of methanol and platinum as catalyst for the electro-reduction of oxygen 13 . Typically, the catalysts consisting of finely–divided precious metal 11 powders are coated on a porous carbon fiber paper such as the ones available from Toray Inc., Japan. The porous carbon paper allows for the reactants and products to reach the surface of the catalyst and also conducts electrons to the catalytic surface. The proton exchange membrane usually consists of Nafion-based material, a sulfonated tetrafluoroethylene-based fluoropolymer that conducts protons. The proton exchange membrane serves as an electrolyte between two electrodes. As shown in Fig. 1-5, a 3% solution of methanol in water is used as the fuel. The methanol solution is circulated past the negative electrode where methanol oxidation occurs. Methanol reacts with water producing carbon dioxide, protons and electrons. The carbon dioxide produced is removed as a gas, and more methanol is fed into the cell to maintain the desired concentration of methanol. The standard reduction potential of methanol oxidation reaction is 0.02 V. The protons released in the reaction are transported to the other electrode through the proton exchange membrane. Air is fed to the positive electrode. The oxygen from the air reacts with protons and electrons to produce water. The standard reduction potential of oxidation reduction is 1.23V. 12 Figure 1-5 Direct methanol fuel cell The overall cell reaction in the direct methanol fuel cell is that methanol reacts with oxygen and produces water and carbon dioxide. The overall cell voltage is 1.21 V. These reactions are summarized in Table 1-1. Table 1-1 Electrochemical reactions in the direct methanol fuel cell Electrode/Cell Reactions Standard Electrode Potential Negative CH 3OH + H 2O → CO 2 + 6H + + 6e - E o = 0.02 V Positive 3/2O 2 + 6H + + 6e - → 3H2O E o = 1.23 V Cell CH 3OH + 3/2 O 2 → CO 2 + 2H 2O E o cell = 1.21 V 13 1.2.2 Direct Methanol Fuel Cell Components and Assembly A direct methanol fuel cell unit hardware in Fig. 1-6 consists of various components other than the membrane electrode assembly (MEA) to allow for continuous operation. These additional components include teflon gaskets, graphite plates, and current collecting plates. The gaskets maintain the compression of the cell and provides a fluid-tight seal. The graphite plates make electric connection to the electrodes of the MEA. The flow field pattern machined into the graphited plates distributes the reactant gas or liquid. Gold-coated copper current collector plates allow the connection of the fuel cell to the external loads. Figure 1-6 Direct methanol fuel cell full-cell assembly 14 1.3 Catalysts for the Electro-Oxidation of Methanol 1.3.1 Platinum-Ruthenium Catalysts Platinum is well-known as a key component in the catalysts for reactions involving carbon monoxide. 14 However, the platinum surface is also easily poisoned by carbon monoxide because of the large free-energy of adsorption. It is also widely accepted that adsorbed carbon monoxide CO ads on pure platinum is the main poisoning species formed during partial oxidation of organic molecules, 15 methanol for example. 16 To solve the poisoning problem of pure platinum, a variety of Pt-based bi-functional catalysts have been studied. 17,18 Thus far, the best commercial anode catalyst is platinum-ruthenium on carbon (PtRu/C). This catalyst consists of highly-dispersed crystallites of a platinum-ruthenium alloy (PtRu) supported by high-surface-area carbon, Vulcan XC 72. These catalysts are available from Johnson-Matthey Inc. The most common bulk atomic ratio of Pt:Ru is 1:1. The total metal loading on the carbon ranges from 10 wt.% to 80wt.%. In this type of catalyst, the roles of platinum and ruthenium are quite distinct and hence it is referred to as a bi-functional catalyst. 19 Platinum facilitates the dissociative chemisorption of the methanol molecule to form adsorbed carbon monoxide and hydrogen. Similarly, ruthenium facilitates the dissociative chemisorption of the water molecule to form adsorbed hydroxyl and hydrogen atoms. These adsorbed carbon monoxide and hydroxyl species recombine on the surface through an electrochemical reaction to form carbon dioxide and protons. Protons are also generated by the electrochemical oxidation of the adsorbed hydrogen atoms. These reactions on the catalyst surface are schematically depicted in Fig 1-7. 15 Figure 1-7 Methanol oxidation on platinum ruthenium catalyst The chemical equations below represent the processes occurring on the bi-functional catalyst 18 , the overall methanol oxidation is in Table 1-1 Methanol Dissociation Pt + CH 3OH → Pt(CH 3OH) ads → Pt(CH 3O) ads + H + + e - Pt(CH 3O) ads → Pt(CH 2O) ads + H + + e - Pt(CH 2O) ads → Pt(CHO) ads + H + + e - Pt(CHO) ads → Pt(CO) ads + H + + e - Eq. 1-1 Eq. 1-2 Eq. 1-3 Eq. 1-4 Water Dissociation Ru + H 2O → Ru-(H 2O) ads → Ru-(OH) ads + H + + e - Eq. 1-5 Surface Recombination Pt(CO) ads + Ru-(OH) ads → Pt + Ru + CO 2 + H + + e - Eq. 1-6 Although Pt-Ru catalyst is thus far the most active methanol oxidation catalyst, 20 to achieve reasonable performance, there are several disadvantages faced with Pt-Ru. Firstly, a large precious metal loading is required to achieve high reaction rates. It is estimated that approximately 60-80 grams of platinum and ruthenium are required for an efficient 1kW fuel cell. Secondly, the metallic ruthenium in the Pt-Ru catalyst is prone to oxidation and dissolution as ruthenium hydroxide (Fig. 1-8), limiting the durability of the catalyst to 500-1000 hours. Due to the accelerated degradation of ruthenium 21 , the catastrophic degradation of a 400-Watt, 80-cell 16 direct methanol fuel cell stack that happened in only six months was observed by Valdez et al. 22 Consequently, the large-scale commercialization of direct methanol fuel cells has not occurred, although smaller size products are available on the market for various portable applications for example, the products from Smart fuel cells and Oorja Inc. Figure 1-8 Pourbaix diagrams of tantalum and ruthenium 1.3.2 Platinum Based Bi-functional Catalysts One of the challenges faced in the commercialization of methanol fuel cells is achieving cheaper and more stable methanol oxidation catalysts. Earlier research has showed that tantalum, niobium, titanium and zirconium are potential alternatives for the ruthenium in methanol oxidation catalysts. 23 24 Since the oxygen vacancies in tantalum oxide surface 25 can facilitate the chemisorption of water 26 we can expect tantalum to fulfill the role of ruthenium in bi-functional catalysts. Niobium and zirconium have properties similar to tantalum and can be potentially 17 useful. In addition, unlike ruthenium, tantalum 27 , titanium 27 and niobium 28 are not prone to dissolution under anodic conditions because of the passivating layer of the oxide. Tantalum, for example, as per the Pourbaix diagram in Fig. 1-8 stays as an oxide at the pH and potentials of methanol oxidation. Thus, tantalum is more stable than ruthenium in the direct methanol fuel cell environment (red circle in Fig. 1-8). Finally, the cost of tantalum and niobium at $156/kg and $165/kg respectively, is much lower than that of ruthenium, at $7000/kg and therefore a significant reduction in cost can be achieved. The oxidation of carbon monoxide and methanol oxidation also have some similarities. They both involve the removal of adsorbed CO from platinum sites. A carbon monoxide tolerant platinum- tantalum alloyed catalyst was invented by UTC Power LLC at 1991. According to the study of Atsushi Ueda et al. 23 , the addition of TaOx and NbOx to the catalytic Pt anode enhances the electrochemical oxidation of CO. Role of TaOx and NbOx was proposed to relax the strong CO adsorption on the surface of Pt. A concerted mechanism for the enhancement of the CO oxidation activity of Pt was also proposed in which the relaxation of the strong CO adsorption and the supply of activated oxygen species were assumed. Thus, the focus of my research was to determine if the advantages of tantalum, niobium and zirconium could be exploited for the benefit of direct methanol fuel cells. Tantalum(V) oxide-encapsulated platinum has been shown to have improved methanol oxidation performance due to improved surface area. No bifunctional behavior of Pt-Ta catalyst was demonstrated in this research. 29 A tantalum-modified platinum electrode was prepared and studied. 30 In their study, the researchers proposed the bi-functional mechanism of tantalum 18 modified platinum. However, no experimental proof of bi-functional catalysis was shown in the research. In another research, high throughput thin film Pt-M alloys were prepared by a single sputter deposition. The methanol oxidation catalytic activities of the alloys were studied and showed the potential of Pt-Ta as a methanol electrooxidation catalyst. 24 31 Zhang and his colleague observed niobium dioxide facilitates methanol electrooxidation on Pt/C catalyst by synergistic effect. 32 Sasaki and Adzic found that monolayer of Pt deposited on niobium oxide nanoparticles indicated that NbO 2 holds considerable promise as a support for a Pt monolayer for methanol oxidation. 33 Therefore, we set a goal for our research to investigate the benefit of tantalum, niobium, titanium and zirconium for the electro-catalysis of methanol oxidation. Of the methods of preparation of catalysts, sputter-deposition of thin films presented several advantages that will be reviewed here. 1.4 Sputter Deposition Technology Sputter deposition is a physical vapor deposition method of preparing thin films by sputtering. This involves sputtering atoms from one or multiple targets onto the substrate such as a silicon wafer as shown in Fig. 1-9. 34 An argon ions in a plasma are accelerated towards a metal target and the collision of the energetic argon ions with the target results in sputtering of the atoms from the target. Sputtered atoms ejected from the target have a wide energy distribution, typically up to tens of eV. 35 The sputtered ions can fly from the target and impact energetically on the substrates or the walls of the vacuum chamber. At higher gas pressures, the ions collide with the gas atoms, reaching the substrate or vacuum chamber wall after condensation while 19 moving. 36,37 The entire range from high-energy ballistic impact to low-energy thermalized motion is accessible by changing the background gas pressure. The sputtering gas is often an inert gas such as argon. Reactive gases such as nitrogen and oxygen can also be used to sputter compounds. The compounds with reactive gases can be formed on the target surface, in-flight or on the substrate depending on the process parameters. The availability of many parameters that control sputter deposition make it a complex process, but also allow experts a large degree of control over the composition, growth and microstructure of the film. Figure 1-9 Inside the sputter deposition chamber There are several advantages of sputter deposition method. 38,39 An important advantage of sputter deposition is that even materials with very high melting points are easily sputtered while evaporation of these materials in a resistance evaporator or Knudsen cell is problematic or impossible. Because of the low substrate temperatures used, sputtering is an ideal method to deposit contact metals for thin-film transistors. Sputtered films typically have a better adhesion Substrate Target 2 Target 1 20 on the substrate than evaporated films. Sputtering sources contain no hot parts (to avoid heating they are typically water cooled) and are compatible with reactive gases such as oxygen. Sputtering can be performed top-down while evaporation must be performed bottom-up. Advanced process such as epitaxial growth is possible. Some disadvantages of the sputtering process are that the process is more difficult to combine with a lift-off for structuring the film. This is because the diffusional transport characteristic of sputtering, makes a full shadow impossible. Thus, one cannot fully restrict where the atoms go, which can lead to contamination problems. Also, active control for layer-by-layer growth is difficult compared to pulsed laser deposition and inert sputtering gases are built into the growing film as impurities. One of the most important commercial applications of sputter deposition is in the production of computer hard disks. Sputtering is used extensively in the semiconductor industry to deposit thin films of various materials in integrated circuit processing. Thin antireflection coatings on glass for optical applications are also deposited by sputtering. Another familiar application of sputtering is low-emissivity coatings on glass, used in double-pane window assemblies. A large industry has developed around tool bit coating using sputtered nitrides, such as titanium nitride, creating the familiar gold colored hard coat. Sputtering is also used as the process to deposit the metal layer during the fabrication of CDs and DVDs. Hard disk surfaces use sputtered CrOx and other sputtered materials. Sputtering is one of the main processes of manufacturing optical waveguides and is another way for making efficient photovoltaic solar cells. 21 Thus, sputter-deposition presents several attractive attributes for the preparation of thin film multi-component catalyst materials that can be applied to a substrate. Chapter 2 Experimental Techniques and Analysis Methodologies 2.1 Physical Characterization 2.1.1 Scanning Electron Microscopy A JEOL JSM 7001 scanning electron microscope (SEM) was used to study the morphology of the carbon nanotube coated carbon nanofiber composite and the morphology of the sputtered platinum-M catalyst. 2.1.2 Energy Dispersive Analysis of X-rays The Pt and Ta metallic composition was determined by Scanning Electron Microscopy – Energy Dispersive X-ray (SEM-EDX) analysis using a Kevex Quantum Detector with an IXRF digital pulse processing analyzer microscope operated at 10 kV. 2.1.3 X-ray Diffraction The sputtered catalysts were characterized ex situ by XRD with Rigaku Ultima IV diffractometer (using Co Kα radiation, λ = 1.7902 Å) to determine their phase compositions and approximate chemical compositions. The crystallite size was calculated using the Scherrer equation. 40 B(2ϴ) = Kλ/Lcosϴ Eq. 1-7 22 Where B is full-width at half maximum (FWHM), L is the crystallite size, λ is the X-ray wavelength, θ is the Bragg angle, and K is the Scherrer constant. The constant K depends on how the breadth B is determined and the actual shape of the crystallite. The most common value of constant K is 0.94. 2.1.4 X-ray Photoelectron Spectroscopy The surface composition of the catalysts were studied using X-ray photoelectron Spectroscopy (XPS) with a magnesium X-ray source (1253.6 eV, SPECS XPS at NETL). The XPS data was analyzed by using CASA software and all the XPS data was corrected based on the carbon peak at 284.6 eV. 2.2 Half-cell Electrochemical Measurements Half-cell electrochemical studies were carried out in a three-electrode cell at room temperature. 0.1 M perchloric acid solution was used as the electrolyte, a mercury/mercury sulfate (MSE) (Hg|Hg 2SO 4, 1M H 2SO 4) electrode was used as the reference electrode, and a Pt wire was used as a counter electrode. A 2cm x 2cm Toray paper electrode and glassy carbon disk electrode (RDE) with a diameter of 5mm (Pine Instruments, Change-Disk Electrode AFE5TQ050) were both used as working electrodes. For the RDE electrode, an ink was prepared by sonicating the catalyst in the mixture (1 mg/ml) consisting of water, isopropyl alcohol and Nafion (5% solution of Nafion 1100 EW ionomer, Sigma-Aldrich). Subsequently, the appropriate amount of the ink was placed on the working electrode dropwise and air dried at 85 °C.) An Ametek-PAR-VMC-4 Potentiostat / Galvanostat /Frequency Response Analyzer was used to perform the electrochemical characterization. 23 Figure 2-1 Half-cell experiment setup Copper underpotential deposition was done by holding electrode in 0.005M CuSO 4 and 0.1M H 2SO 4 solution at a potential slightly higher than -0.4V then linear change the voltage toward the positive direction. Assuming the charge required to oxidize a monolayer of Cu adsorbed on each metal surface is 0.42micron colum/cm 2 . The electrochemical surface area of the electrode is calculated from the following equation: ECSA = Q Cu / (0.42mC/cm 2 ) 2.3 Full Cell Test The MEAs for full cell testing were fabricated with methanol oxidation catalysts as the anode, Nafion 117 as the proton-exchange membrane, and Pt black as the cathode catalyst. On the anode of co-sputtered Pt-Ta MEA, Pt 0.77-Ta 0.23 anode electrode was prepared by co-sputtering Pt 24 at 40W and Ta at 45W onto a 5cm by 5cm carbon fiber composite paper (Avcarb MGL 190) spray- coated with carbon nanotubes of 0.008 mg/cm 2 . The background pressure during sputterdeposition was maintained at 30 mTorr with an argon flow rate at 139 sccm. The total catalyst loading at the anode was 0.15 mg/cm 2 of which the platinum loading was 0.04 mg/cm 2 . A thin coating of Liquion solution (5% Nafion ionomer solution) was sprayed over the Pt 0.77-Ta 0.23 anode electrode to make a continuous proton conducting path between the catalyst materials and membrane. On the anode of PtRu/C MEA, PtRu/C catalyst layers were coated on carbon fiber composite paper (Avcarb MGL 190) using catalyst inks containing PtRu/C, Liquion, isopropanol and water. The mass ratio of PtRu to Liquion solution is 1:5. The loading of platinum was 0.04mg/cm 2 , which was the same as the anode of co-sputtered Pt-Ta MEA. On the cathode side, to prevent cathode polarization and methanol crossover from affecting the results, cathode catalyst layers were coated on Toray carbon paper (TGPH-060, 20% teflonized) using catalyst inks containing platinum black and water. A Pt black loading of 2 mg/cm 2 was remained constant throughout the course of the studies. The MEA was pressed at 140 ºC under 1000 pounds for 15 minutes to bond the coated anode and cathode to either side of the Nafion 117 membrane. The MEA was assembled in fuel cell test hardware (Electrochem Inc.). Silicone gaskets provided the sealing. The anode was fed with 1M methanol water solution flowing at 1.1 L/min. The total circulating amount of methanol was 2 liters. The cathode was fed with air flowing at 2L/min and the pressure was set at 10 psig. The methanol solution was circulated through the cell at 60°C for 1 hour before testing. Cells were tested with oxygen at 10 psi and 1 M methanol at 60°, 80°C and 90°C. 25 Chapter 3 Platinum-Tantalum Catalysts Prepared by Sputter Deposition In this chapter we describe the properties of thin film platinum-tantalum (Pt 1-x-Ta x, 0<x<1) catalysts that were prepared by sputter-deposition. Physical and structural characterization of these thin film catalysts was performed using scanning electron microscopy (SEM), X-ray diffraction (XRD), energy dispersive analysis of X-ray fluorescence (EDAX) and X-ray photoelectron spectroscopy (XPS). The methanol oxidation activity of platinum-tantalum catalysts was studied by electrochemical polarization methods in the half-cell configuration as shown in Fig. 2-1. We found that Pt 0.77-Ta 0.23 catalyst showed similar electrochemical area specific activity compared to commercially- available PtRu/C catalysts. We also found that changing the atomic ratio of platinum to tantalum affected the catalytic activity for methanol oxidation. This observation suggested that Pt 1-x-Ta x catalyst had a different methanol oxidation mechanism compared to platinum. The surface oxides species on tantalum could activate water molecules and hence facilitate the process of removing carbon monoxide from platinum sites. We discuss the preparation, characterization and mechanistic aspects in detail in the following sections. 3.1 Preparation of Platinum-Tantalum Electrodes by Co-sputter Deposition Platinum-tantalum electrodes consisting of various ratios of the two metals (designated as Pt 1-x- Ta x) were prepared in a custom-built multisource sputter deposition system shown in Fig.1-9. Each of the Pt 1-x-Ta x electrodes was prepared by sputter-depositing platinum and tantalum onto a 2cm by 2cm carbon fiber composite paper (Avcarb MGL 190). The main chamber of the sputter- 26 deposition apparatus was pumped down to 1E-6 Torr before sputtering. Argon gas was allowed to flow into the chamber to regulate the pressure during sputtering. Then platinum (99.95% purity, Plasmaterials) and tantalum (99.95% purity, Plasmaterials) were co-deposited from separate magnetron sputter sources. Five minutes of pre-sputtering was performed to clean the target surface and remove any possible contamination. During pre-sputtering the carbon paper substrate was shielded from the target. During co-sputtering, the platinum cathode power was kept at 40 W while tantalum cathode power was varied between 20 to 140 W in steps of 20 W to yield the various atomic ratios that were desired. The background pressure during deposition was maintained at 10 mTorr by maintaining the argon flow rate at 35 sccm. The substrate holder was rotated at 20 rpm to ensure lateral compositional homogeneity of the binary-metal films. The deposition rates were measured with a quartz crystal monitor. Figure 3-1 Co-sputter Pt 0.77-Ta 0.23 electrode on carbon fiber composite substrate Fig. 3-1 is Pt 0.77-Ta 0.23 is a typical example of the co-sputtered Pt 1-x-Ta x electrodes. The grey color is from carbon fiber composite substrate Avcarb MGL 190. The darkened square is the part of the substrate that was covered by co-sputtered Pt-Ta. The geometric area of co-sputtered Pt 1-x-Ta x was 2cm x 2cm. 27 3.2 Physical Characterization of Co-sputter Platinum-Tantalum Catalysts Figure 3-2 SEM images of AvCarb MGL 190 (a, b) and co-sputtered Pt 0.77-Ta 0.23 (c, d) on AvCarb MGL 190 SEM Fig.3-2a and b are SEM images of the AvCarb MGL 190 substrate. AvCarb MGL 190 is a carbon fiber composite paper. Fig. 3-2a shows that the carbon fiber composite substrate is rather porous formed by a random arrangement of carbon fibers about 10 micron in diameter. Fig. 3- 2c, d are typical SEM images of the co-sputtered Pt-Ta electrodes. There was no significant a b c d 28 difference between the co-sputter deposited electrode (Fig. 3-2c, d) and the substrate (Fig. 3-2a, b), except for the slightly increased brightness of Fig. 3-2c, d caused by the back-scattering of electrons from the highly conductive layer of metal. The SEM study confirmed that a thin film of co-sputtered Pt 1-x-Ta x catalyst layer covered the entire surface of the AvCarb MGL substrate, evenly. Figure 3-3 EDX elemental mapping of (a) platinum and (b) tantalum Table 3-1 The tantalum to platinum atomic ratio of the co-sputter Pt-Ta electrodes Tantalum Sputter Power/Watt 20 40 45 60 Ta:Pt Atomic Ratio 0.132 0.264 0.298 0.957 Tantalum Sputter Power/Watt 80 100 120 140 Ta:Pt Atomic Ratio 1.34 1.55 1.69 1.83 a b 29 In Figs.3-3 a, b we present the EDX elemental mapping of platinum and tantalum, respectively. The distribution of platinum and tantalum is uniform and cover the whole carbon nanofiber substrate. The EDX also indicates that the Ta to Pt atomic ratio is 28.8% to 71.2% in this co-sputter Pt-Ta electrode. The Ta:Pt atomic ratio of the various co-sputtered Pt-Ta electrodes prepared, are shown in Table 3-1. In all the samples, the sputter time and sputter power of platinum were the same, the only difference was the sputter power applied to the tantalum target. The amount of platinum is the same in all the samples while the amount of tantalum linearly increases with the sputter power of tantalum. So we should expect the atomic ratio of Ta:Pt to linearly increase with the increasing sputter power on the tantalum target. We indeed find that the Ta:Pt atomic ratio measured by EDX increases with increasing the sputter power of tantalum as indicated by the data in Fig. 3-4. Figure 3-4 The effect of varying the sputter power of tantalum on Ta:Pt atomic ratio in the co- sputter Pt 1-x-Ta x catalysts 30 Profilometer - The thickness of the co-sputter Pt 0.77-Ta 0.23 catalyst measured using a profilometer is 15 nm. Because of the sensitivity of the profilometer, the error with the thickness measurement on the thinner electrodes is expected to be bigger. The thickness of sputtered material is related to sputter time and sputter power. We were able to estimate the thickness of very thin electrodes from thicker co-sputtered Pt-Ta electrodes prepared under the same condition except that the sputter power of tantalum was different. Using the atomic ratio of Ta:Pt shown in Chart 3-2, and assuming the density of Pt is 21.45 g/cm³ and the density of Ta 2O 5 is 8.2 g/cm³, the volume ratio of tantalum oxide to platinum oxide was estimated and shown in Chart 3-2. From the volume ratio and the thickness of Pt-Ta catalyst sputtered with 140W applied on tantalum target and 40W applied on platinum target, we estimated the thickness of all the co-sputtered Pt-Ta electrodes. The results of the thickness estimates are shown in Table 3-2. Tantalum Sputter Power/Watt 20 40 45 60 Ta 2O 5 : Pt Volume Ratio 0.37 0.74 0.83 1.1 Catalyst Layer Thickness 6.6nm 13nm 15nm 20nm Tantalum Sputter Power/Watt 80 100 120 140 Ta 2O 5 : Pt Volume Ratio 1.5 1.8 2.2 2.6 Catalyst Layer Thickness 26nm 33nm 39nm 46nm Table 3-2 The tantalum oxide to platinum volume ratio and catalyst thickness of the co-sputter Pt-Ta electrodes XRD - The XRD patterns of the co-sputtered Pt 0.77-Ta 0.23 are presented in Fig. 3-5. We found platinum face-centered cubic phase with prominent reflections from the (100) and (110) crystal 31 faces, and orthorhombic Ta 2O 5 in the co-sputtered Pt 0.77-Ta 0.23 sample. The average crystallite size of both platinum and tantalum oxide were estimated using the Scherrer equation 40 , which was described in Section 2.1.3. B(2ϴ) = Kλ/Lcosϴ Eq. 1-7 Take the Pt 0.77-Ta 0.23 catalyst for example, the calculation of the crystallite size based on Ta 2O 5 (001) peak was as follows: L(Ta 2O 5) = Kλ / (Bcosϴ) = 1 x 0.154nm / [(3π / 180) cos(22 o )] = 3.17nm Eq. 3-1 The average crystallite size of platinum was 7.15 nm and that of Tantalum oxide is 3.17 nm as estimated by the method shown above. One particle may have one or more crystallites. Therefore, these numbers may not reflect the real particle size. Figure 3-5 XRD patterns of co-sputter Pt 0.77-Ta 0.23 32 XPS - In Fig 3-6 and Table 3-3, we present XPS spectra for the Pt-Ta samples. We find that the Ta-4f peak confirms the oxidation state of tantalum in Pt-Ta is 5+. However, in the co-sputtered Pt-Ta, Ta 4f binding energy shifts to 25.75eV compared to that of the tantalum oxide electrode prepared by sputter deposition, which is 26.81 eV. On the other hand, the binding energy associated with the Pt-4f peak of co-sputter Pt-Ta sample is 71.47eV, exceeds that of sputter Pt which is 71.35eV. This suggests a partial charge transfer from Pt to Ta. However, the Ta 4f binding energy shifts to a greater extent than the Pt 4f binding energy. This shift is because the atomic ratio Ta: Pt is 28.8% to 71.2%, and thus one tantalum atom is surrounded by more Pt, so the resultant effect of Pt-Ta interaction on Ta atoms is more pronounced than on the Pt atoms. Figure 3-6 Co-sputter Pt-Ta XPS spectra, (a) Ta-4f and (b) Pt-4f binding energy, XPS spectra were corrected using carbon spectra as a standard Pt4f 7/2 Ta 4f 7/2 Sputter Pt Co-sputter Pt-Ta Sputter Ta Co-sputter Pt-Ta Binding Energy 71.35eV 71.47eV 26.81eV 25.75eV b a Ta-4f Pt-4f 33 Table 3-3 Binding energy of co-sputter Pt-Ta, sputter Pt and sputter Ta In the electro-oxidation of methanol, the strong adsorption of CO on Pt poisons the active catalytic sites hence inhibits the reaction. This strong adsorption comes from a dual electron donation, first from the σ bonding orbital of CO orbital to the Pt 5d orbital and then from the back–donation of the Pt 5d orbital to the CO anti–bonding 2π* CO orbitals, as illustrated by the Blyholder model 41 in Fig 3-7. When the d-electron density decreases via charge transfer to TaO 7/2, the Pt-CO binding energy would be expected to decrease. Thus, Ta is expected to reduce the strength of adsorption of CO on platinum. Figure 3-7 The molecular bonding of CO and Pt according to the Blyholder model 3.3 Electrochemical Characterization of Co-Sputtered Platinum-Tantalum Catalysts The electrochemical characterization of the Pt-Ta electrode was performed in a half-cell in Fig. 2- 1. The measurement was performed by varying the electrode potential and measuring the current response in a 0.1 M perchloric acid solution de-gassed to be free of oxygen by saturation 34 with argon. For the surface area studies, the solution was free of methanol. For the methanol oxidation studies, 1M methanol was used. The results of these tests are analyzed in the following sections. 3.3.1 Electrochemical Surface Area A typical cyclic voltammogram of an electrode that contains platinum exhibits several pairs of peaks corresponding to adsorption/desorption of hydrogen- and oxygen-containing species. For the platinum electrode studied in this research, the oxidation and reduction peaks in Fig. 3-8 at - 0.6V vs MSE correspond to the electro-sorption and electro-desorption of hydrogen on the platinum sites. Using the charge under the hydrogen desorption peak, the electrochemical active surface area values of the electrode were estimated assuming a monolayer adsorption of hydrogen atom on platinum sites and the charge density is 210 micro-coulomb /cm 2 . 42 We found that the electrochemical surface area of platinum for the co-sputtered Pt-Ta catalyst was 5 times that of the platinum catalyst sputtered under the same conditions. 35 Figure 3-8 The cyclic voltammetic scans of co-sputter Pt 0.77-Ta 0.23 and sputtered Pt at a scanning rate of 200mV/s in 0.1 M perchloric acid This increase in surface area cannot be attributed to the increased volume added by tantalum to the catalyst for the following reason. According to platinum to tantalum atomic ratio from EDX elemental composition analysis, and the density of Pt and Ta 2O 5 being 21.45 g/cm 3 and 8.2 g/cm 3 , respectively, the estimated volume ratio of Pt to Ta 2O 5 is 1:0.83. The total volume of co-sputtered Pt-Ta can be expected to increase only about 1.83 times of the sputtered Pt sample with the same amount of Pt. The total platinum electrochemical surface area, however, is 5 times of sputtered Pt. Thus we conclude that the morphology of Pt deposition has been changed by tantalum co-sputter deposition interrupting the growth process for the platinum. We are proposing that in the sputtered platinum catalyst, the platinum particles could grow on top of each of the platinum particles whereas in the Pt-Ta structure, such growth is not possible allowing many more uncovered platinum sites to be present as shown in Fig. 3-9. The increased electrochemical active surface area of co-sputter Pt-Ta catalyst is also an indication of a better utilization of platinum materials. 36 Figure 3-9 Illustrations of (a) Sputter Pt catalyst (b) Co-sputter Pt 0.77-Ta 0.23 catalyst on a carbon substrate Besides Pt 0.77-Ta 0.23, we varied the atomic ratio between platinum and tantalum by changing the sputter power for tantalum and keeping the sputter power for platinum constant. The atomic ratio of Pt: Ta was measured by EDX elemental analysis (Table 3-1). We found that the Pt electrochemical surface area changes with varying the atomic ratio of Pt:Ta inspect that the amount of platinum are the same in all electrodes. As shown in Fig. 3-10, the electrochemical surface area of platinum first increases with increasing Ta:Pt atomic ratio and reaches a maximum at Ta : Pt atomic ratio equals 0.364 and then starts decreasing. The maximum platinum surface area on a 2 cm x 2 cm was 23.24 cm 2 . This process is illustrated in Fig. 3-11. With increasing amount of Ta, the ECSA first increases because the Pt particles are better dispersed in the Pt-Ta catalyst (Fig.3-11 b) compared to the aggregated Pt particles in Pt catalysts; at the maximum electrochemical surface area (Fig.3-11 c), tantalum particles help platinum particles to disperse to the largest extent meanwhile all the platinum particles are still connected to the carbon b a 37 substrate or other platinum particles. With further increase of Ta as in Fig.3-11d, some of the Pt particles are isolated by the oxide particles. The poor electronic conductivity of the oxide could cause the decrease of access to the platinum particles and hence a decrease in the electrochemically active surface area of platinum. Figure 3-10 The effect of Pt 1-x-Ta x catalysts composition on the electrochemical surface area 38 Figure 3-11 Illustrations of co-sputter Pt 1-x-Ta x catalysts with variant amount of tantalum 3.3.2 Methanol Oxidation Activity The electro-oxidation of methanol was studied at the sputter-deposited electrodes by slow-scan voltammetry and the results are shown in Fig. 3-12. The methanol oxidation currents under these conditions was measured for the co-sputtered Pt-Ta electrode, co-sputtered Pt-Ru electrode, sputtered Pt electrode and the commercial PtRu/C (powder) by anodically polarizing the a b c d 39 electrodes in 1M methanol solutions with 0.1M perchloric acid from -0.5V to 0V vs MSE at the scanning rate of 1mV/s. Figure 3-12 pseudo-steady state methanol oxidation mass activities of co-sputtered Pt 0.77-Ta 0.23, co-sputtered Pt-Ru, sputtered Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol Onset Potential – The onset potential is known as the potential at which a sharp increase in oxidation/reduction current is noted. Onset potential of methanol oxidation during an anodic scan is one of the parameters for assessing the catalyst activity. 43 Here we define the potential under which the oxidation current equals 3% of the current at 0V vs MSE as the onset potential. We found that the commercial PtRu/C catalyst has a slightly higher onset potential compared with co-sputtered Pt-Ru catalyst, while both these values of onset potential are close to -0.3V vs 40 MSE. The onset potential of co-sputtered Pt-Ta electrode is -0.15V and that for the sputtered Pt is -0.125V. The onset potential for methanol oxidation reaction is indicative of the potential when water adsorption coverage begins to increase. Under these conditions, the adsorbed intermediate from water, namely, -OH, could react with adsorbed -CO on the platinum sites leading to an oxidation current. These onset potential values are consistent with the water dissociation reaction usually happens around -0.3V vs MSE on ruthenium, -0.1V on the platinum sites. 44 17 The results in Fig. 3-13 indicate that the onset potential of Pt 1-x-Ta x decreases with increasing the amount of tantalum in the catalysts. It’s very likely that adding tantalum could promote methanol oxidation. As discussed in Chapter 1, oxide’s ability for water dissociation allows it to work together with platinum as a bi-functional catalyst. Thus, based on the onset potential observed for co-sputtered Pt 1-x-Ta x, we believe tantalum could facilitate methanol oxidation and it is very likely that it could dissociate water at a potential as low as -0.55V. 41 Figure 3-13 Methanol oxidation onset potentials of co-sputter Pt 1-x-Ta x catalysts and commercial PtRu/C catalyst Area Specific Catalytic Activity – The methanol oxidation area specific activity is defined as the observed methanol oxidation current divided by the electrochemically active platinum surface area as determined from the hydrogen-desorption studies. In Fig. 3-14, the area specific activity of Pt 0.77-Ta 0.23 increases with electrode potential at a greater rate compared to PtRu/C, and the activity soon exceeds that of PtRu/C at -0.12 V vs MSE as we anodically polarize the electrode towards positive values of potentials. The Pt electrode exhibited a low oxidation current as it was poisoned by the strong adsorption of CO adsorption on Pt sites. Once it reaches substantially positive electrode potentials, >0V vs. MSE, the water dissociation reaction is facilitated on the Pt sites and the adsorbed CO can be removed from the catalytic active surface to allow for further dissociation of methanol. The methanol oxidation activities of co-sputtered Pt-Ta catalyst at - 0.1V and 0V are compared with PtRu/C and Pt in Table 3-4. 42 Figure 3-14 pseudo-steady state methanol oxidation ECSA activities of co-sputter Pt-Ta, sputter Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol Table 3-4 Co-sputter Pt-Ta catalyst electrochemical area specific activities compared to sputter Pt and PtRu/C catalysts at -0.1 V and 0 V vs MSE Co-sputtered Pt-Ta catalyst electrochemical area specific activity At -0.1V vs MSE 2.8 times of Pt 1.2 times of PtRu/C At 0V vs MSE 2.3 times of Pt 3.2 times of PtRu/C 43 We also found that the methanol oxidation activity of Pt-Ta catalyst is affected by the atomic ratio of Pt to Ta in the catalysts. In Fig. 3-15, the activity first increases with increasing Ta: Pt ratio because more and more platinum catalytic sites are in contact with tantalum sites so that more platinum sites become bifunctional Pt-Ta catalysts; then the catalytic activity decreases because some of the Pt particles are isolated by the semi-conductive oxide particles. Figure 3-15 The effect of Pt 1-x-Ta x catalysts composition on the methanol oxidation ECSA activity 3.3.3 Catalyst Stability To test the stability of Pt-Ta electrode, we carried out steady state polarization tests at a relatively positive potential of -0.1V vs MSE as shown in Fig. 3-16. Co-sputtered Pt-Ta catalyst showed similar electrochemical area specific activity compared to PtRu/C after holding at -0.1V vs MSE for 1 hour. The slow stabilization of the oxidation current of co-sputtered Pt-Ta is most likely 44 caused by surface reconstruction. Also, this similar phenomenon was observed from all the sputtered electrodes including co-sputtered Pt-Ru electrodes. The rate of current change of co- sputtered Pt-Ta catalyst at the end of one hour was 0.0029% per second and that of co-sputter Pt-Ru catalyst was 0.0115% per second, which is much faster compared to co-sputtered Pt-Ta catalysts. Figure 3-16 Electrode electrochemical stability test by holding electrodes at -0.1V vs MSE for 1 hour 3.3.4 Platinum-Tantalum Catalyst Methanol Oxidation Mechanism Tafel plots and Tafel slopes of co-sputtered Pt-Ta, sputter Pt and PtRu/C electrodes are shown in Fig 3-17 and Table 3-4. The Tafel slope of Pt-Ta and Pt over the range of -0.3V < V < -0.15V vs 45 MSE potential region indicates the similar kinetic behavior, which is a one electron transfer process with surface recombination as rate-determining step in Eq. 1-6. 45 The Tafel slope of PtRu/C in the range of -0.3V < V < -0.15V vs MSE potential region is higher than in the range - 0.15V < V < 0V vs MSE potential region, which could indicate a change in the mechanism. Earlier research showed that the limiting currents of PtRu/C in this potential region are not mass transfer controlled and cannot be enhanced by changing the rotating speed of the rotating disk electrode in a half-cell experiment. 46 This suggests the formation of new adsorbed intermediates. Jusys et al. have reported detection of methyl formate in this potential region. 47 Figure 3-17 Methanol oxidation reaction Tafel plots of co-sputter Pt-Ta, sputter Pt and PtRu/C Electrode Tafel slope Potential Range Co-sputtered Pt-Ta 119 mV/dec -0.3V < V < 0V 46 Sputtered Pt 117 mV/dec -0.3V < V < 0V PtRu/C 143 mV/dec 376 mV/dec -0.3V < V < -0.15V -0.15V < V < 0V Table 3-4 Tafel slopes of each Tafel plot in Fig 3-16 Here we are showing the derivation of the Tafel Equation for methanol oxidation. We assume that methanol dissociation and water dissociation are equilibrium and surface recombination is the rate determining step. For methanol dissociation step Pt + CH 3OH ⇌ Pt-CO + 4H + + 4e - , we have: 𝑘 1 ( 1 − 𝑃𝑡 − 𝐶𝑂 ) 𝐶 𝐶𝐻 3 𝑂𝐻 𝑒 𝑥𝑝 − 1 ( 𝐸 − 𝐸 𝑟 1 ) 𝐹 𝑅𝑇 = 𝑘 − 1 𝑃𝑡 − 𝐶𝑂 𝐶 𝐻 + 4 exp ( 1 − 1 ) ( 𝐸 − 𝐸 𝑟 1 ) 𝐹 𝑅𝑇 Eq. 3-2 α 1 is the charge transfer coefficient of methanol dissociation step. k 1 and k- 1 are the forward and backward rate constant of methanol dissociation step. ϴ Pt-CO is the fractional coverage of CO on platinum sites. C CH3OH is the concentration of methanol. C H+ is the concentration of the protons. R is the universal gas constant. F is the Faraday constant. E r1 is the equilibrium potential of the methanol dissociation step. Let 𝑘 1 𝑘 − 1 = 𝐾 𝑒𝑞 1 Eq. 3-3 𝐾 𝑒𝑞 1 is equilibrium constant for methanol dissociation step. 𝑃𝑡 − 𝐶𝑂 1 − 𝑃𝑡 − 𝐶𝑂 = 𝑘 1 𝑘 − 1 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 + 4 𝑒 𝑥𝑝 − ( 𝐸 − 𝐸 𝑟 1 ) 𝐹 𝑅𝑇 = 𝐾 𝑒𝑞 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 + 4 𝑒 𝑥𝑝 − ( 𝐸 − 𝐸 𝑟 1 ) 𝐹 𝑅𝑇 Eq. 3-4 𝐾 𝑒𝑞 = 𝑒 𝑥𝑝 − 𝐺 𝑎𝑑 𝑠 1 𝑅𝑇 Eq. 3-5 47 ΔG ads1 is the Gibbs energy of CO adsorption on platinum sites. Therefore, 𝑃𝑡 − 𝐶𝑂 1 − 𝑃𝑡 − 𝐶𝑂 = 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 + 4 𝑒 𝑥𝑝 − 𝐺 𝑎𝑑 𝑠 1 + 𝐹 ( 𝐸 − 𝐸 𝑟 1 ) 𝑅𝑇 Eq. 3-6 𝑃𝑡 − 𝐶𝑂 = 𝐶 𝐶𝐻 3 𝑂𝐻 𝑒 𝑥𝑝 − 𝐺 𝑎𝑑 𝑠 1 + 𝐹 ( 𝐸 − 𝐸 𝑟 1 ) 𝑅𝑇 𝐶 𝐻 + 4 + 𝐶 𝐶𝐻 3 𝑂𝐻 𝑒 𝑥𝑝 − 𝐺 𝑎𝑑 𝑠 1 + 𝐹 ( 𝐸 − 𝐸 𝑟 1 ) 𝑅𝑇 Eq. 3-7 Similar, for water dissociation step Ru + H 2O ↔ Ru-OH + H + + e - , The fractional coverage of CO on platinum sites is: 𝑅𝑢 − 𝑂𝐻 = 𝐶 𝐻 2 𝑂 𝑒 𝑥𝑝 − 𝐺 𝑎𝑑 𝑠 2 + 𝐹 ( 𝐸 − 𝐸 𝑟 2 ) 𝑅𝑇 𝐶 𝐻 + + 𝐶 𝐻 2 𝑂 𝑒 𝑥𝑝 − 𝐺 𝑎𝑑 𝑠 2 + 𝐹 ( 𝐸 − 𝐸 𝑟 2 ) 𝑅𝑇 Eq. 3-8 For the surface combination step Pt-CO + Ru-OH → Pt + Ru + CO 2 + H + +e - 𝐼 = 𝑘 3 𝑃𝑡 − 𝐶𝑂 𝑅𝑢 − 𝑂𝐻 𝑒 𝑥𝑝 − 3 ( 𝐸 − 𝐸 𝑟 3 ) 𝐹 𝑅𝑇 = 𝑘 3 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝑒 𝑥 𝑝 − 𝐺 𝑎 𝑑 𝑠 1 + 𝐺 𝑎 𝑑 𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) + 3 𝐹 ( 𝐸 − 𝐸 𝑟 3 ) 𝑅𝑇 𝐶 𝐻 + 5 + 𝐶 𝐻 + 𝐶 𝐶𝐻 3 𝑂𝐻 𝑒 𝑥 𝑝 − 𝐺 𝑎 𝑑 𝑠 1 + 𝐹 ( 𝐸 − 𝐸 𝑟 1 ) 𝑅𝑇 + 𝐶 𝐻 + 4 𝐶 𝐻 2 𝑂 𝑒 𝑥 𝑝 − 𝐺 𝑎 𝑑 𝑠 2 + 𝐹 ( 𝐸 − 𝐸 𝑟 2 ) 𝑅𝑇 + 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝑒 𝑥 𝑝 − 𝐺 𝑎 𝑑 𝑠 1 + 𝐺 𝑎 𝑑 𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) 𝑅𝑇 Eq. 3-9 Taking the natural logarithm of the above equation, 𝑙𝑛𝐼 = − 𝐺 𝑎 𝑑 𝑠 1 + 𝐺 𝑎 𝑑 𝑠 2 − 𝐹 ( 𝐸 𝑟 1 + 𝐸 𝑟 2 + 3 𝐸 𝑟 3 ) 𝑅𝑇 − 2 + 3 𝑅𝑇 𝐹𝐸 + 𝑙𝑛 𝑘 3 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝐶 𝐻 + 5 − 𝑙𝑛 ( 𝐶 𝐻 + 5 + 𝐶 𝐻 + 𝐶 𝐶𝐻 3 𝑂𝐻 𝑒 𝑥𝑝 − 𝐺 𝑎 𝑑 𝑠 1 + 𝐹 ( 𝐸 − 𝐸 𝑟 1 ) 𝑅𝑇 + 𝐶 𝐻 + 4 𝐶 𝐻 2 𝑂 𝑒 𝑥𝑝 − 𝐺 𝑎 𝑑 𝑠 2 + 𝐹 ( 𝐸 − 𝐸 𝑟 2 ) 𝑅𝑇 + 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝑒 𝑥𝑝 − 𝐺 𝑎 𝑑 𝑠 1 + 𝐺 𝑎 𝑑 𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) 𝑅𝑇 ) Eq. 3-10 Under the following three assumptions: 48 𝐶 𝐻 + 5 ≪ 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝑒𝑥𝑝 − 𝐺 𝑎 𝑑𝑠 1 + 𝐺 𝑎 𝑑𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) 𝑅𝑇 Eq. 3-11 𝐶 𝐻 + 𝐶 𝐶𝐻 3 𝑂𝐻 𝑒𝑥𝑝 − 𝐺 𝑎 𝑑𝑠 1 + 𝐹 ( 𝐸 − 𝐸 𝑟 1 ) 𝑅𝑇 ≪ 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝑒𝑥𝑝 − 𝐺 𝑎 𝑑𝑠 1 + 𝐺 𝑎 𝑑𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) 𝑅𝑇 Eq. 3-12 𝐶 𝐻 + 4 𝐶 𝐻 2 𝑂 𝑒𝑥𝑝 − 𝐺 𝑎 𝑑𝑠 2 + 𝐹 ( 𝐸 − 𝐸 𝑟 2 ) 𝑅𝑇 ≪ 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝑒𝑥𝑝 − 𝐺 𝑎 𝑑𝑠 1 + 𝐺 𝑎 𝑑𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) 𝑅𝑇 Eq. 3-13 We have: 𝑙 𝑛 𝐼 = − 𝐺 𝑎𝑑 𝑠 1 + 𝐺 𝑎𝑑 𝑠 2 − 𝐹 ( 𝐸 𝑟 1 + 𝐸 𝑟 2 + 3 𝐸 𝑟 3 ) 𝑅𝑇 − 1 + 3 𝑅𝑇 𝐹𝐸 + 𝑙𝑛 𝑘 3 𝐶 𝐶𝐻 3 𝑂𝐻 𝐶 𝐻 2 𝑂 𝐶 𝐻 + 5 − 𝑙𝑛 𝐶 𝐶𝐻 3 𝑂𝐻 + 𝐺 𝑎𝑑 𝑠 1 + 𝐺 𝑎𝑑 𝑠 2 + 𝐹 ( 2 𝐸 − 𝐸 𝑟 1 − 𝐸 𝑟 2 ) 𝑅𝑇 Eq. 3-14 Tafel Slope equals 2 . 3 𝑅𝑇 𝐹 ( 1 − 3 ) For a Tafel slope that equals 120 mV/dec, α 3 = 0.5. For any deviation from this number, for example PtRu/C, it is caused by the charge transfer coefficient of surface combination step deviating from 0.5. We also found that changing the atomic ratio of platinum to tantalum affected the Tafel Slope for methanol oxidation as shown in Fig 3-18. The Tafel slope at low Ta:Pt ratio is 120 mV/dec, suggests that the rate determining step is a one-electron transfer step. As previously discussed, this rate determining step is the surface recombination step. At higher Ta:Pt ratio, Tafel slope first increases to 300mV/dec. The most likely rate determining step is methanol dissociation. 45 Further increase in the Ta:Pt ratio results a Tafel slope around 500mV. The co-sputtered Pt-Ta electrode at this ratio mainly consists of Ta 2O 5 and TaO 2 as shown in Fig 3-19. The insulating property of these oxides result in the anomalously high Tafel slope. 49 Figure 3-18 The effect of Pt 1-x-Ta x catalysts composition on the methanol oxidation Tafel slope Figure 3-18 XRD of co-sputter Pt-Ta electrode with 1.8:1 Ta:Pt atomic ratio As we discussed in Section 3.3.2, electrochemical active surface area specific activity of co- sputtered Pt-Ta is higher than Pt. This suggests the increased MOR current is because increased catalytic activity and not a mere increase in surface area. This observation is consistent with the changes in area specific activity with varying ratio of Pt : Ta. Further, the change in onset potential 50 also suggests that Ta is playing an active role in enhancing the kinetics of the methanol oxidation reaction. Based on the metal oxide’s ability to activate the water molecule, we propose the bi- functional mechanism of Pt-Ta catalysts. Methanol Dissociation: Pt + CH 3OH → Pt-H + Pt-CO Eq. 3-15 Water Dissociation: TaOx + H 2O → TaOx-H + TaOx-OH Eq. 3-16 Surface Recombination: Pt-CO + Ta-OH → Pt + Ta + CO 2 + H + + e - Eq. 3-17 Figure 3-14 Co-sputtered Pt-Ta catalyst as a bi-functional catalyst for methanol oxidation 3.4 Summary In this chapter, a series of thin film platinum-tantalum (Pt-Ta) catalysts were prepared by the sputter deposition method. Characterization of these thin film catalysts was performed using scanning electron microscopy, X-ray diffraction, energy dispersive X-ray and X-ray photoelectron spectroscopy. Assessment of the methanol oxidation activity of platinum-tantalum catalysts was achieved through half-cell experiments. The Pt 0.77-Ta 0.23 showed similar electrochemical area specific activity compared to the commercially-available PtRu/C catalysts. We also found that 51 the atomic ratio of platinum to tantalum in the catalyst affects the platinum-normalized methanol oxidation activity. This observation suggests that Pt-Ta catalyst has different methanol oxidation mechanism compared to platinum. This difference is because the surface oxides activate water molecules and hence facilitate the process of removing carbon monoxide from platinum sites. 52 Chapter 4 Catalysts based on Co-sputtered Platinum-M (M=Niobium, Titanium, Zirconium) In this chapter, catalysts based on co-sputtered Pt 1-x-M x (M=Nb, Ti, Zr, 0<x<1) were synthesized in the same way as the co-sputtered Pt-Ta catalysts described in Chapter 3. We have analyzed the physical and electrochemical properties of these co-sputtered Pt 1-x-M x (M=Nb, Ti, Zr) thin films. We have also discussed the effect of atomic ratio of platinum to the second metal. We have compared Pt-Nb, Pt-Ti, Pt-Zr with the Pt-Ta catalyst. While the Pt-Nb, Pt-Ti and Pt-Zr exhibited catalytic activity for methanol oxidation, their performance and durability did not exceed that of Pt-Ta catalysts. 4.1 Co-sputtered Platinum-Niobium Catalysts 4.1.1 Preparation of Platinum-Niobium Catalysts by Co-sputter Deposition Co-sputtered Pt-Nb electrodes with different Pt:Nb atomic ratio were prepared in the custom- built multisource sputter deposition system described earlier in Chapter 1. Each of the Pt 1-x-Nb x electrodes was generated by co-depositing platinum and niobium onto a 2cm by 2cm carbon fiber composite paper (Avcarb MGL 190). Platinum and niobium were co-deposited from separate magnetron sputter sources (99.95% purity, Plasmaterials). The main chamber was pumped down to 1E-6 Torr before sputtering. A five-minute pre-sputtering step was used to clean the target surface and remove any possible contamination. The platinum cathode power was kept at 40 W while niobium cathode power was varied in steps of 20W from 20 to 140W for the various samples. The samples were designated as Pt 1-x-Nb x. The background pressure during 53 deposition was maintained at 10 mTorr by a flow of argon at the rate of 35 sccm. The substrate holder was rotated at 20 rpm to ensure the lateral compositional homogeneity of the binary- metal films. The deposition rates were measured with a quartz crystal monitor. The amount of Pt was the same on all the electrodes. The loading of platinum based on the geometric area of the electrode was calculated to be 0.026mg/cm 2 . 4.1.2 Physical Characterization of Platinum-Niobium Catalysts The carbon fiber paper electrodes coated with a thin film of the catalyst prepared by ]co-sputter deposition of platinum and niobium had the following physical properties: Scanning Electron Microscopic Images – Fig.4-1(a)(b) are the images of the AvCarb MGL 190 carbon fiber composite substrate as a blank sample and Fig.4-1(c)(d) are images of the co- sputtered Pt 0.85-Nb 0.15 electrodes. These photographs are representative of all the Pt 1-x-Nb x electrodes that we prepared. The texture of the carbon fiber remained unchanged after co- sputter deposition. The dimension of the carbon fiber was not noticeably changed either. The only differences observed between the co-sputtered Pt 0.85-Nb 0.15 electrode Fig.4-1(c)(d) and the substrate Fig.4-1(c)(d), are the slightly increased brightness of Pt 0.85-Nb 0.15 electrodes caused by increased scattering of electrons from the highly conductive metallic coating. These observations suggested that a very thin layer of co-sputtered Pt 0.85-Nb 0.15 catalyst evenly covered the AvCarb MGL substrate. 54 Figure 4-1 SEM images of (a), (b) AvCarb MGL 190 and(c), (d)co-sputter Pt 0.85-Nb 0.15 on AvCarb MGL 190 Compositional Analysis by EDX – Fig.4-2 (a)(b) show the elemental distribution of platinum and niobium in Pt 0.95-Nb 0.05. The distribution of platinum and niobium appeared to be uniform. The underlying porous structure of the carbon fiber paper is recognizable from EDX mapping and was similar to the SEM image in Fig.4-2 (c). Both platinum and niobium evenly covered the carbon nanofiber substrate. The EDX measurements allowed us to determine the niobium to platinum a b c d 55 atomic ratio of the co-sputter Pt 0.95-Nb 0.05 electrode shown in Fig. 4-2 was 0.047:1. The amount of deposited niobium increased with the sputter power. Since the amount of platinum in every sample was the same, the atomic ratio of niobium to platinum in each sample increased linearly with sputter power. This linear relation between niobium and platinum as shown in Fig. 4-3 based on data in Table 4-1 could be used to determine the Pt:Nb ratio for the samples sputtered at other power levels. a b c 56 Figure 4-2 EDX elemental mapping of (a) platinum and (b) niobium and (c) the corresponding SEM image Figure 4-3 Linear relation between niobium sputter power vs. Nb:Pt atomic ratio in Nb-Pt electrodes Table 4-1 The Atomic ratio of Nb:Pt of Pt-Nb electrodes Nb sputter Power/Watt 0 20 80 140 Nb:Pt atomic ratio 0 0.047 0.23 0.41 Surface Analysis by X-ray Photoelectron Spectroscopy – Fig. 4-4 (a) XPS spectra of Nb-4f confirmed that the chemical state of niobium in co-sputtered Pt-Nb is Nb 5+ . In the co-sputtered y = 0.0029x 0 0.2 0.4 0.6 0 50 100 150 Nb:Pt Atomic Ratio Niobium Sputter Power (Watt) Nb:Pt Atomic Ratio vs Nb Sputter Power 57 Pt 0.9-Nb 0.1, the Nb 4f binding energy shifted to 206.7 eV while the 4f binding of the niobium pentoxide electrode prepared by sputter niobium was 207.5eV. On the other hand, Pt-4f binding energy of co-sputter Pt 0.9-Nb 0.1 catalyst, which is 71.40eV, exceeds that of sputter Pt, which is 71.35eV. This suggests a partial charge transfer from Pt to Nb. Nb 4f binding energy shifts more than Pt 4f binding energy because the atomic ratio of Nb:Pt is 1: 9, as a result, the effect of the interaction between platinum and niobium on each niobium atom is stronger than on each Pt atom. Figure 4-4 Co-sputtered Pt 0.9-Nb 0.1 XPS spectra, (a) Nb-4f and (b) Pt-4f binding energy, XPS spectra were corrected using carbon spectra as a standard Table 4-2 Binding energy of co-sputter Pt 0.9-Nb 0.1, sputter Pt and sputter Nb Pt4f 7/2 Nb4f 7/2 Sputter Pt Co-sputter Pt-Nb Sputter Nb Co-sputter Pt-Nb a b Nb 4f Pt 4f 58 Binding Energy 71.35eV 71.40eV 207.5eV 206.7eV In methanol oxidation reaction, the strong binding interaction between CO and Pt poisons the active catalytic sites hence inhibits methanol oxidation reaction. This strong interaction comes from a dual electron donation, first from the bonding σ CO orbital to Pt 5d orbital and then from the back–donation of the Pt 5d orbital to the anti–bonding 2π*CO orbitals 41 as described in Chapter 3. The d-electron density decreases via electron transfer to Nb, and as a result, the Pt- CO binding energy would decrease. 4.1.3 Electrochemical Characterization of Platinum-Niobium Catalysts The electrochemical characterization of the co-sputtered Pt-Nb electrodes was performed in a three-electrode half-cell. The measurement was performed by varying the electrode potential and measuring the current response in an argon saturated 0.1 M perchloric acid solution containing. The methanol concentration was 1M. The amount of Pt is the same in all the electrodes. The loading of platinum divided by the geometric surface area of the electrode is 0.026mg/cm 2 . Electrochemical Surface Area – The oxidation and reduction peaks at -0.6V to -0.3V vs MSE in Fig. 4-5 correspond to hydrogen adsorption and desorption at platinum sites. Using the hydrogen desorption peak, the electrochemical active surface areas of the electrode were estimated assuming a monolayer adsorption of hydrogen atom on platinum sites and the charge density is 210 micro-Coulomb/cm 2 . 42 The platinum electrochemical surface area of the co-sputtered Pt 0.85- 59 Nb 0.15 catalyst was 3.8 times that of the platinum catalyst sputtered under the same conditions although both samples had the same amount of platinum. Figure 4-5 The cyclic voltammetry curve of co-sputter Pt 0.85-Nb 0.15 at a scanning rate of 200mV/s in 0.1 M perchloric acid without any methanol Take Pt 0.85-Nb 0.15 catalyst for example, according to Pt:Nb atomic ratio from EDX elemental composition analysis, and assuming that the density of Pt is 21.45 g/cm³ and the density of Nb 2O 5 is 4.6 g/cm³, the estimated volume ratio of Pt to Nb 2O 5 is 1:0.7 The total volume of co-sputtered Pt 0.85-Nb 0.15 will be 1.7 times of the sputter Pt sample, which is not comparable to the actual value of the surface area of Pt 0.85- Nb 0.15 catalyst (3.8 times that of the platinum catalyst). We believe that the morphology of Pt deposition has changed due to the co-sputtering of niobium. Unlike in 60 the sputtered platinum catalysts, which allowed the platinum particles to pile up, the co- sputtered niobium acts as a catalyst support. Consequently, the resulting structure has a greater platinum surface area as shown in Fig. 4-6. The increased electrochemical active surface area is also an indication of a better utilization of platinum materials. Figure 4-6 Illustrations of (a) Sputter Pt catalyst structure and (b) Co-sputter Pt 0.85-Nb 0.15 catalyst structure on a carbon substrate Onset Potential – The methanol oxidation currents as a function of electrode potential were measured by slow-scan voltammetry for the co-sputtered Pt 0.9-Nb 0.1 electrode, co-sputtered Pt- Ru electrode, sputtered Pt electrode and PtRu/C electrode. The electrodes were anodically polarized from -0.5V to 0V vs MSE at the scan rate of 1mV/s as shown in Fig.4-7. Commercial PtRu/C catalyst has a slightly higher onset potential compared with co-sputtered Pt-Ru catalyst. However, both were around -0.3V vs MSE. The onset potential of co-sputtered Pt-Nb electrode a b 61 was -0.25V and sputtered Pt was -0.15V. The onset potential of methanol oxidation reaction is related to the potential at which the water activation begins. It is at this potential that the surface intermediates M-OH would react with M-CO on the platinum sites to produce a continuous oxidation current. Water dissociation reaction usually happens around -0.3V vs MSE on ruthenium, -0.15V on the platinum sites. The observations in Fig. 4-7 are consistent with these values. Based on the onset potential difference between co-sputtered Pt 0.9-Nb 0.1 and sputtered Pt, it appears that niobium plays a role in the methanol oxidation catalysis process and it starts dissociating water at a potential lower than pure platinum. In other words, niobium beneficially affects the rate of methanol oxidation relative to pure platinum. Figure 4-7 Results of slow scan voltammetry of methanol oxidation presented as mass activities of co-sputtered Pt 0.9-Nb 0.1, co-sputtered Pt-Ru, sputtered Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol 62 Mass Activity – In Fig. 4.7, the methanol oxidation currents are divided by the mass of platinum. The commercial Pt-Ru/C catalyst has the highest oxidation current in -0.3V to -0.1V region. In - 0.1V to 0V region the activity of co-sputtered PtRu exceeds that of the PtRu/C electrode. Compared to the differences between co-sputtered Pt 0.9-Nb 0.1 and the sputtered Pt electrode, the methanol oxidation polarization curves of PtRu/C and co-sputtered Pt-Ru are quite similar. The methanol oxidation current of co-sputtered Pt 0.9-Nb 0.1 electrode is lower than the electrodes containing Ru in all the -0.3V to 0V potential region, but the activity is significantly improved compared to the Pt electrode. At -0.1V, the methanol oxidation current of co-sputtered Pt-Nb electrode is 8.3 times of sputtered Pt electrode. At 0V, the methanol oxidation current of co- sputtered Pt 0.9-Nb 0.1 electrode is 6.2 times of sputtered Pt electrode. As mentioned earlier, the ECSA of co-sputtered Pt 0.9-Nb 0.1 is 3.8 times of sputtered Pt electrode. Thus, the effect of increased surface area is not the only reason for the increased methanol oxidation current. Again, these observations suggested that the rate of methanol oxidation on co-sputtered Pt 0.9-Nb 0.1 electrode is significantly different from the sputtered Pt electrode. Electrochemistry Surface Area Activity – In Fig. 4-8 and Table 4-3, the methanol oxidation activity is presented as the current divided by the ECSA of platinum. The surface area of co- sputtered Pt 0.9-Nb 0.1 and sputtered Pt electrodes were estimated from the charge required for electrochemical hydrogen desorption. The surface area of PtRu/C was estimated from copper underpotential deposition because ruthenium also adsorbs hydrogen. As we anodically polarize the electrode, although PtRu/C has a lower onset potential, the activity of co-sputtered Pt 0.9-Nb 0.1 electrode ECSA methanol oxidation activity increases more steeply compared to PtRu/C, exceeding the value for PtRu/C at -0.12 V vs MSE. Also, the ECSA activity of co-sputtered Pt 0.9- 63 Nb 0.1 is higher than Pt in the whole potential region. This result suggests that the niobium plays an active role in the electrocatalysis. The interaction of niobium with platinum could be through facilitating the water dissociation step and lowering Pt-CO binding energy. Figure 4-8 Results of slow-scan voltammetry of methanol oxidation on co-sputtered Pt 0.9-Nb 0.1, sputtered Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol, scan rate 1mV/s. Table 4-3 Comparison of electrochemical area specific activity of co-sputtered Pt 0.9-Nb 0.1 catalyst and sputtered Pt and PtRu/C catalysts at -0.1 V and 0 V vs MSE. Co-sputter Pt 0.9-Nb 0.1 catalyst electrochemical area specific activity At -0.1V vs MSE 2.6 times of Pt 1.1 times of PtRu/C At 0V vs MSE 1.9 times of Pt 2.7 times of PtRu/C 64 Tafel Slope - Tafel plots of co-sputtered Pt 0.9-Nb 0.1, sputter Pt and PtRu/C electrodes are shown in Fig. 4-9 and the Tafel slope are shown in Table 4-4. The chemical equations below represent the methanol oxidation process that occurs on the catalyst Methanol Dissociation Pt + CH 3OH → Pt(CH 3OH) ads → Pt(CH 3O) ads + H + + e - Pt(CH 3O) ads → Pt(CH 2O) ads + H + + e - Pt(CH 2O) ads → Pt(CHO) ads + H + + e - Pt(CHO) ads → Pt(CO) ads + H + + e - Eq. 1-1 Eq. 1-2 Eq. 1-3 Eq. 1-4 Water Dissociation Ru + H 2O → Ru-(H 2O) ads → Ru-(OH) ads + H + + e - Eq. 1-5 Surface Recombination Pt(CO) ads + Ru-(OH) ads → Pt + Ru + CO 2 + H + + e - Eq. 1-6 In the previous research of methanol electro-oxidation, 17,18,19,20,21,45 when the surface recombination is rate determining, the expected Tafel slope is 120 mV/dec, sometimes the it varies between 115mV/dec and 135mV/dec. A Tafel slope higher than this range means methanol dissociation is the rate determining step. The slope of Pt at -0.3V < V < -0.15V vs MSE potential region is very close to the theoretical value predicted for one-electron transfer reaction as rate determining step. The Tafel slope of Pt 0.9-Nb 0.1 and PtRu/C at -0.3V < V < -0.5V vs MSE potential region indicates the rate determining step could be a surface recombination step. The Tafel slope of PtRu/C at -0.15V < V < 0V dramatically increased. Earlier research carried out on RDE electrode in a half-cell experiment showed that methanol oxidation in this potential region is not a mass-transfer controlled process. 46 This suggested the formation of new adsorbed intermediates. Jusys et al. have reported detection of methyl formate in this potential region. 47 65 Figure 4-9 Methanol oxidation reaction Tafel plots of co-sputter Pt 0.9-Nb 0.1, sputter Pt and PtRu/C Table 4-4 Tafel slopes of co-sputter Pt 0.9-Nb 0.1, sputter Pt and PtRu/C Electroder Tafel slope Potential Range Co-sputtered Pt 0.9-Nb 0.1 150 mV/dec -0.3V < V < 0V Sputtered Pt 117 mV/dec -0.3V < V < 0V PtRu/C 143 mV/dec 376 mV/dec -0.3V < V < -0.15V -0.15V < V < 0V 66 4.1.4 Effects of Platinum to Niobium Ratio We prepared a series of co-sputtered Pt-Nb electrodes with the same amount of platinum but different amounts of niobium by maintaining the sputter rate of platinum and changing the sputter rate of niobium for each electrode. The atomic ratio of Pt:Nb was measured by EDX elemental analysis and shown in Fig.4-10. We found out that the Pt electrochemical surface area, methanol oxidation activity and Tafel slope change with the atomic ratio of Pt:Nb. Figure 4-10 The effect of niobium to platinum atomic ratio on the electrochemical surface area Electrochemical Surface Area – As shown in Fig. 4-11, the electrochemical surface area of co- sputter Pt-Nb first increases with increasing Nb : Pt atomic ratio, after it reaches the maximum at 0.17:1 Nb:Pt atomic ratio and starts decreasing and then levels off. We hypothesize the effect of composition using the schematic representations in Fig. 4-10. With increasing the amount of Nb, the ECSA first increases because Pt particles are better dispersed in the Pt-Nb catalyst (b) compared to the aggregated Pt particles in Pt catalysts (a). Upon further increasing the Nb content (d), some of the Pt particles could become isolated by the semi-conductive oxide 67 particles, so the ECSA of Pt starts decreasing. The ECSA reaches the maximum (c) at a point in between (b) and (d), when at this ratio the arrangement of Pt particles and Nb 2O 5 particles exhibit connectivity of the platinum particles and their dispersion. The volume ratio of Nb 2O 5 to Pt at the maximum of ECSA is 0.54:1 and the atomic ratio of Pt: Nb is 1:0.17. Thus, we may expect a threshold to be reached at this atomic ratio and thickness, beyond which further addition of niobium and platinum does not result in an increase in the accessible area of platinum. Figure. 4-11 Illustrations of co-sputter Pt-Nb catalysts with different Nb:Pt atomic ratio b a c d 68 Methanol Oxidation Electrochemical Surface Area Activity – In Fig.4-12, The methanol oxidation activity of co-sputtered Pt-Nb first increases with increasing Nb: Pt ratio because more and more platinum catalytic sites are in contact with the niobium sites so that the CO absorbed on platinum sites could be removed by the adsorbed OH formed on the niobium sites. The activity reaches a maximum when platinum particles and niobium oxide particles are mixed to the most extent at the atomic ratio of 0.12:1 Nb : Pt and then the activity decreases because some of the Pt particles are isolated by the semi-conductive oxide particles. Figure 4-12 Effect of Nb:Pt atomic ratio on methanol oxidation activity Tafel Slope – Effect of Nb:Pt atomic ratio on methanol oxidation Tafel Slopes is shown in Fig. 4- 13. Tafel slope at low Nb:Pt ratio is 150 mV/dec, suggests that the rate determining step is a one- electron transfer step. As discussed in the previous sections, this Tafel slope corresponds to the surface recombination step. With further increase in the Nb:Pt ratio, the Tafel slope gradually increases to 200mV/dec. Thus the mechanism of methanol oxidation is likely to have changed. 69 Figure 4-13 Effect of Nb:Pt atomic ratio on methanol oxidation Tafel Slopes 4.1.5 Co-sputtered Platinum-Niobium Catalyst Methanol Oxidation Mechanism Several observations suggest that Pt-Nb is a bi-functional catalyst. Firstly, the onset potential of co-sputtered Pt-Nb is lower than Pt. Secondly, the ECSA activity of co-sputtered Pt-Nb is higher than that of co-sputtered Pt. Thirdly, the ECSA activity changes with varying Pt:Nb Ratio. Based on the metal oxide’s ability to activate the water molecule and previous research on the behavior of niobium oxides, we propose the bi-functional mechanism of Pt-Nb catalyst. The rate determining step of Pt-Nb catalyst with a Nb:Pt ratio less than 0.2 is likely the surface combination step. This mechanism is consistent with other bifunctional catalysts such as Pt-Ru and co-sputtered Pt-Ta catalysts because the Tafel slope values observed for these catalysts are similar as noted below. 70 Methanol Dissociation: Pt + CH 3OH → Pt-H + Pt-CO Water Dissociation: Nb + H 2O → Nb-H + Nb-OH Surface Recombination: Pt-CO + Nb-OH → Pt + Nb + CO 2 + H + +e - 4.2 Platinum-Titanium Catalysts 4.2.1 Preparation of Platinum-Titanium by Co-Sputter deposition Platinum-titanium electrodes consisting of various ratios of the two metals (designated as Pt 1-x- Ti x) were prepared in a custom-built multisource sputter deposition system shown in Fig.1-9. Each of the Pt 1-x-Ti x electrodes was generated by sputter-depositing platinum and titanium onto a 2cm by 2cm carbon fiber composite paper (Avcarb MGL 190). The main chamber of the sputter- deposition apparatus was pumped down to 1E-6 Torr before sputtering. Argon gas was allowed to flow into the chamber to regulate the pressure during sputtering. Then platinum (99.95% purity, Plasmaterials) and titanium (99.95% purity, Plasmaterials) were co-deposited from separate magnetron sputter sources. A five-minute pre-sputtering process was performed to clean the target surface and remove any possible contamination. During pre-sputtering the carbon paper substrate was shielded from target. During co-sputtering, the platinum cathode power was kept at 40 W while titanium cathode power was varied between 20 to 140 W in steps of 20 W to yield the various atomic ratios that were desired. The background pressure during deposition was remained 10 mTorr by maintaining the argon flow rate at 35 sccm. The substrate holder was rotated at 20 rpm to ensure lateral compositional homogeneity of the binary-metal films. The deposition rates were measured with a quartz crystal monitor. The amount of platinum 71 in all the electrodes is the same. The loading of platinum divided by the geometric surface area of the electrode is 0.026mg/cm 2 . 4.2.2 Physical Characterization of Platinum-Titanium Catalysts The thin film methanol oxidation electrodes prepared by co-sputter deposition of platinum and titanium had the following physical properties: SEM – Fig.4-14 a are the SEM photographs of the carbon fiber composite substrate AvCarb MGL 190 as a blank sample. Fig.4-14 b is SEM images of the co-sputtered Pt 0.94-Ti 0.06 catalyst. This sample is typical of the Pt-Ti electrodes that we prepared with various Ti:Pt ratio. The texture of the carbon fiber was maintained after co-sputter deposition. The size of the carbon fiber did not have a noticeable change either. The only differences observed between the co-sputtered Pt 0.94- Ti 0.06 electrode Fig.4-14 a and the substrate Fig.4-14 b, were the slightly increased brightness of Fig.4-14 b caused by the highly conductive metal coating layer. This observation suggested that a very thin layer of co-sputtered Pt 0.94-Ti 0.06 catalyst covered the AvCarb MGL substrate evenly. a b 72 Figure 4-14 SEM images of (a) AvCarb MGL 190 and (b)co-sputter Pt 0.94-Ti 0.06 on AvCarb MGL 190 EDX – Fig.4-15 a, b are the EDX elemental mapping of platinum and titanium in Pt 0.94-Ti 0.06 electrode. The distribution of platinum and titanium are not identical. The porous structure of the carbon fiber paper is recognizable from platinum elemental mapping and is similar to the SEM image Fig.4-15 c. Platinum evenly covers the carbon nanofiber substrate. On the other hand, titanium does not distribute the same way. Instead of following the carbon nanofiber pattern in as Pt (Fig.4-15a), it spreads everywhere (Fig.4-15b). The atomic ratio of some of the co-sputtered Pt-Ti catalysts are shown in Table 4-5. 73 Figure. 4-15 EDX elemental mapping of (a) platinum and (b) titanium and (c) the corresponding SEM image of the Pt 0.94-Ti 0.06catalyst Table 4-5 The Atomic ratio of Ti:Pt of Pt-Ti electrodes Ti sputter Power/Watt 0 20 40 80 140 Ti:Pt atomic ratio by EDX 0 0.09 0.063 0.47 0.60 XPS – Fig 4-16 (b) co-sputtered Pt 0.94-Ti 0.06 XPS spectra of Ti p 3/2 confirmed the chemical state of titanium as Ti 4+ . Co-sputtered Pt 0.94-Ti 0.06, Ti p 3/2 binding energy shifted to 458.1 eV while the Ti b a c 74 p 3/2 binding of the titanium oxide electrode prepared under the same condition is 458.8eV. Since the atomic ratio of Ti:Pt is 6:94, if platinum atoms and titanium atoms do have electronic interaction, the effect on Ti p 3/2 binding energy could shift much more than Pt 4f binding energy. This might be the reason that Pt-4f binding energy of both co-sputter Pt 0.94-Ti 0.06 71.35V and sputter Pt 71.33V, were almost the same. Figure. 4-16 Co-sputtered Pt 0.94-Ti 0.06 XPS spectra, (a) Ti-4f and (b) Pt-4f binding energy, XPS spectra were corrected using carbon spectra as a standard Table 4-6 Binding energy of co-sputtered Pt 0.94-Ti 0.06, sputter Pt and sputter Ti Pt4f 7/2 Ti P 3/2 Sputter Pt Co-sputter Pt 0.94-Ti 0.06 Sputter Ti Co-sputter Pt 0.94-Ti 0.06 Binding Energy 71.35eV 71.33eV 458.8eV 458.1eV Pt-4f Ti - p a b 75 4.2.3 Electrochemical Characterization of Platinum-Titanium Catalysts The electrochemical characterization of the co-sputtered Pt-Ti electrodes was performed in a half-cell. The measurement was performed by varying the potential and measuring the current response in an Argon saturated 0.1 M perchloric acid solution. A methanol concentration of 1M was used in the methanol oxidation studies. The amount of Pt was the same in all the electrodes. The loading of platinum based on the geometric surface area of the electrode was 0.026mg/cm 2 . Electrochemical Surface Area – In Fig 4-17, by integrating the hydrogen desorption peak between -0.6V to -0.3V and assuming a monolayer adsorption of hydrogen atom on platinum sites and the charge density is 210 microcoulombs/cm 2 , the electrochemical active surface area values of both electrodes were estimated. The platinum electrochemical surface area of the co- sputtered Pt 0.94-Ti 0.06 catalyst was 1.53 times that of the platinum catalyst sputtered under the same conditions with the same amount of platinum. According to the Pt:Ti atomic ratio from EDX elemental composition analysis, and assuming that the density of Pt is 21.45 g/cm³ and the density of TiO 2 is 4.23 g/cm³, the estimated volume ratio of Pt to TiO 2 is 1:0.97. The total volume of co-sputter Pt 0.94-Ti 0.06 is 1.97 times that of the sputtered Pt sample. 76 Figure 4-17 The cyclic voltammetry curve of co-sputter Pt 0.94-Ti 0.06 and sputter Pt at a scanning rate of 200mV/s in 0.1 M perchloric acid As mentioned in Chapter 3, the ECSA of the sputtered Pt electrode is linearly related to it volume under the assumption that Pt atoms are not densely packed when sputtered at 40 W for less than 240 seconds. If we assume Pt and TiO 2 particles have the same size and deposition morphology, the total ECSA of Pt and TiO 2 will also be linearly related to the total volume of the electrode. The total ECSA of the co-sputtered Pt-Ti electrode will 1.97 times of the sputter Pt electrode. Under the same assumption, the ratio of surface area of Pt and Ti should be the same as the atomic ratio Ti:Pt=0.468:1 measured by EDX. As a result, the platinum ECSA of this Pt-Ti electrode will be 1.97 x 1/1.468=1.34. This number is similar to the ECSA calculated from hydrogen desorption charge. This suggests that the morphology of Pt deposition did not change much with and without co-sputtered Ti as shown in Fig 4-18. -0.006 -0.004 -0.002 0 0.002 0.004 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 I/A E/V vs MSE CV 200mV/s Sputter Pt 77 Figure 4-18 Illustrations of (a) Sputter Pt catalyst structure and (b) Co-sputter Pt-Ti catalyst structure on a carbon substrate Onset Potential – The pseudo-steady state methanol oxidation current on co-sputtered Pt 0.94- Ti 0.06 electrode, co-sputtered Pt-Ru electrode, sputtered Pt electrode and PtRu/C electrode were measured by anodically polarizing the electrodes from -0.5V to 0V vs MSE at the scanning rate of 1mV/s and the results are shown in Fig. 4-19. Commercial PtRu/C catalyst has a slightly higher onset potential compared with co-sputtered Pt-Ru catalyst. Both are around -0.3V vs MSE. The onset potential of co-sputtered Pt 0.94-Ti 0.06 electrode is -0.2V and sputter Pt is -0.15V. The onset potential on methanol oxidation reaction is the potential when water activation begins. Thus, the surface intermediate -OH could react with adsorbed -CO on the platinum sites to produce a continuous oxidation current. Based on the similar onset potentials of co-sputtered Pt 0.94-Ti 0.06 and sputtered Pt, it is hard to conclude whether titanium sites activate water to a significantly greater extent than platinum or not. a b 78 Figure 4-19 Pseudo-steady state methanol oxidation mass activities of co-sputtered Pt 0.94-Ti 0.06, co-sputtered Pt-Ru, sputtered Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol Mass Activity – In Fig. 4-19 the methanol oxidation currents are divided by the mass of platinum. The commercial Pt-Ru/C catalyst has the highest oxidation current in -0.3V to -0.1V region. In - 0.1V to 0V region co-sputtered PtRu exceeds PtRu/C electrode. Compared to co-sputtered Pt 0.94- Ti 0.06 and sputtered Pt electrode, the methanol oxidation polarization curves of PtRu/C and co- sputtered Pt-Ru are similar. The methanol oxidation current of co-sputtered Pt 0.94-Ti 0.06 electrode is lower than the electrodes containing Ru and is only slightly improved compared to the Pt electrode. Electrochemical Surface Area Activity – In Fig. 4-20, we have presented the methanol oxidation currents divided by the ECSA of platinum. The surface area of co-sputtered Pt 0.94-Ti 0.06 and 0 0.01 0.02 0.03 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 I(A) / mgPt E (V) vs MSE Methanol Oxidation Mass Activity Sputter Pt PtRu/C Co-sputter Pt-Ru Co-sputter Pt-Ti 79 sputtered Pt electrodes were estimated from hydrogen desorption charge. The surface area of PtRu/C was estimated from copper underpotential deposition. As we anodically polarized the electrode towards a positive potential, although PtRu/C has a lower onset potential, the methanol oxidation activity of the co-sputtered Pt-Ti electrode based on ECSA exceeded that of PtRu/C at -0.09 V vs MSE. The Pt electrode first got poisoned because of the CO adsorption on Pt sites. Once it reached a higher potential, water dissociation reaction starts happening on the Pt sites. Once OH ads is formed, the adsorbed CO could be removed from the catalytic active surface and the dissociative adsorption of methanol could continue happening. The possible ways of ensuring this happens includes facilitating the water dissociation step and lowering the Pt-CO binding energy. Figure 4-20 Pseudo-steady state methanol oxidation voltammetry and ECSA activities of co- sputtered Pt 0.94-Ti 0.06, sputtered Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol 80 Table 4-7 Co-sputtered Pt 0.94-Ti 0.06 catalyst electrochemical area specific activities compared to sputtered Pt and PtRu/C catalysts at -0.1 V and 0 V vs MSE Co-sputtered Pt 0.94-Ti 0.06 catalyst electrochemical area specific activity At -0.1V vs MSE 2.1 times of Pt 0.9 times of PtRu/C At 0V vs MSE 1.4 times of Pt 2.0 times of PtRu/C Tafel Slope - Tafel plots and Tafel slopes of co-sputtered Pt-Ti, sputtered Pt and PtRu/C electrodes are shown in Fig. 4-21 and Table 4-8. The slope of the sputtered Pt between -0.3V and -0.15V vs MSE potential region is very close to the theoretical value predicted for a one-electron transfer reaction as the rate determining step. The Tafel slope of Pt 0.94-Ti 0.06 and PtRu/C at -0.3V < V < - 0.5V vs MSE potential region indicates the rate-determining step could be the surface recombination step. The Tafel slope of PtRu/C at -0.15V < V < 0V is much higher compare to the one electron transfer process. 46 As discussed in Section 3.3.4, this suggested the formation of new adsorbed intermediates in this potential range. 81 Figure. 4-21 Methanol oxidation reaction Tafel plots of co-sputtered Pt-Ti, sputtered Pt and PtRu/C Table 4-8 Tafel slopes of co-sputtered Pt 0.94-Ti 0.06, sputtered Pt and PtRu/C Electrode Tafel slope Potential Range Co-sputtered Pt 0.94-Ti 0.06 153 mV/dec -0.3V < V < 0V Sputtered Pt 117 mV/dec -0.3V < V < 0V PtRu/C 143 mV/dec 376 mV/dec -0.3V < V < -0.15V -0.15V < V < 0V 82 4.2.4 Effects of Platinum to Titanium Ratio We prepared a series of co-sputter Pt-Ti electrodes with the same amount of platinum but different amounts of titanium by maintaining the sputter rate of platinum and changing the sputter rate of titanium for each electrode. The atomic ratio of Pt:Ti was measured by EDX elemental analysis. We found out the factors that change with varying the atomic ratio are Pt electrochemical surface area, methanol oxidation activity and Tafel slope. Figure 4-22 The effect of Ti: Pt atomic ratio on the electrochemical surface area Electrochemical Surface Area – The electrochemical active surface area of co-sputtered Pt-Ti electrodes with different Ti:Pt atomic ratio in Fig.4-22 do not show an obvious trend as Pt-Ta and Pt-Nb electrodes that we discussed previously. The sample with 0% of Ti is sputtered Pt electrode. The maximum ECSA occurs at the Ti:Pt atomic ratio of 0.09. The average ECSA of the Pt-Ti electrodes was 8.86cm 2 and the variance was 2.53cm 2 . Atoms ejected from cathode escape with energies of 10 to 50 eV, which is 10-100 times the energy of evaporated atoms. This additional energy (together with bombardment by other ions) 83 provides sputtered atoms with additional surface mobility. Sufficient surface mobility of Pt and Ti atoms allow them to form the most stable structure at the surface. The excess Ti atoms will diffuse to the bulk. So that a layered film may be formed. The composition of the surface layer always has a relatively stable composition hence the surface atomic ratio of platinum and titanium doesn’t have a trend as we increase the amount of titanium. Methanol Oxidation Electrochemical Surface Area Activity - The methanol oxidation activities of co-sputtered Pt-Ti electrodes with different Ti:Pt atomic ratios in Fig.4-23 do not follow a trend either. These observations can be attributed to the surface Ti:Pt ratio not being the same as the bulk ratio. The maximum activity occurs when the Ti:Pt atomic ratio equals 0.352. The average ECSA activity of co-sputtered Pt-Ti electrodes is 0.097mA/cm 2 at 0V vs MSE and at -0.1V it is 0.022mA/cm 2 as shown in Table 4-9. Figure 4-23 Effect of Ti:Pt atomic ratio on methanol oxidation ECSA activity 84 Co-sputter Pt-Ti catalyst average electrochemical area specific activity At -0.1V vs MSE 1.4 times of Pt 0.60 times of PtRu/C At 0V vs MSE 1.1 times of Pt 1.6 times of PtRu/C Table 4-9 Co-sputtered Pt-Ti catalysts average electrochemical area specific activities compared to sputtered Pt and PtRu/C catalysts at -0.1 V and 0 V vs MSE Tafel Slopes – Different from ECSA -based methanol oxidation activity, the Tafel slope of the co- sputtered Pt-Ti catalyst increases with increasing the percentage of titanium in the catalysts. Although the ratio of Ti:Pt in the surface layer is not the same as in the bulk, the Tafel slope shows a trend. Another factor that’s affected by the bulk composition is the conductivity of the catalyst. From XPS analysis we know that titanium exists as TiO 2 in the catalyst. TiO 2 is wide band gap semi-conductor. The more Ti in the bulk, the less conductive the electrode will be. The electronic conductivity of the co-sputter Pt-Ti could also affect the Tafel slopes. 85 Figure 4-24 Effect of Ti:Pt atomic ratio on Tafel Slopes 4.3 Platinum-Zirconium Catalysts 4.3.1 Preparation of Platinum-Zirconium by Co-sputter Deposition Platinum-zirconium electrodes consisting of various ratios of the two metals (designated as Pt 1- x-Zr x) were prepared by co-sputter-deposition in a custom-built physical vapor deposition system shown in Fig.1-9. Each of the Pt 1-x-Zr x electrodes was generated by sputter-depositing platinum and zirconium onto a 2cm by 2cm carbon fiber composite paper (Avcarb MGL 190). The main chamber of the sputter-deposition apparatus was pumped down to 1E-6 Torr before sputtering. Argon gas was allowed to flow into the chamber to regulate the pressure during sputtering. Then platinum (target: 99.95% purity, Plasmaterials) and zirconium (target: 99.95% purity, Plasmaterials) were co-deposited from separate magnetron sputter sources. A five-minute pre- sputtering run was performed to clean the target surface and remove any possible contamination. During pre-sputtering the carbon paper substrate was shielded from target. During co-sputtering, the platinum cathode power was kept at 40 W while the zirconium cathode power was varied in steps of 20 W over the range of 20 to 140 W to yield the various atomic ratios that were desired. The background pressure during deposition was remained 10 mTorr by maintaining the argon flow rate at 35 sccm. The substrate holder was rotated at 20 rpm to ensure lateral compositional homogeneity of the binary-metal films. The deposition rates were measured with a quartz crystal monitor. The amount of platinum in all the electrodes is the same. The loading of platinum divided by the geometric surface area of the electrode is 0.026mg/cm 2 . 86 4.3.2 Physical Characterization of Platinum- Zirconium Catalysts The thin film methanol oxidation electrodes prepared by co-sputter deposition of platinum and zirconium had the following physical properties: SEM –Figs.4-25 a, b are the microscope photographs of the carbon fiber composite substrate as a blank sample. Fig.4-25 c, d are SEM images of the co-sputtered Pt 0.95-Zr 0.05 electrode. The texture of the carbon fiber is maintained after co-sputter deposition. The size of the carbon fiber doesn’t have noticeable change either. The only differences observed between the co-sputtered Pt 0.95-Zr 0.05 electrode and the substrate is that the image of Pt 0.95-Zr 0.05 is brighter because of the highly conductive metal coating layer. This observation suggests a very thin layer of co-sputtered Pt-Zr catalyst covered the AvCarb MGL substrate evenly. 87 Figure 4-25 SEM images of (a),(b) AvCarb MGL 190 and(c),(d)co-sputtered Pt 0.95-Zr 0.05 on AvCarb MGL 190 EDX – The results in Figs.4-26 a, b are the distribution of platinum and zirconium in Pt 0.95-Zr 0.05 electrode. The distribution of platinum and zirconium are identical. The porous structure of the carbon fiber paper Fig.4-26 c is recognizable from EDX mapping. Both platinum and zirconium evenly cover the carbon nanofiber substrate. a b c d 88 Figure 4-26 EDX elemental mapping of (a) platinum and (b) zirconium and (c) the corresponding SEM image XPS –The Zr-3d binding energy of sputtered Zr Fig. 4-27a at 182.6 eV confirmed that the chemical state of zirconium is Zr 4+ . The Zr-3d binding energy in co-sputtered Pt 0.83-Zr 0.17 shifted to 181.6 eV. On the other hand, Pt-4f binding energy of co-sputter Pt 0.83-Zr 0.17 sample, which is 71.41eV, exceeds that of sputtered Pt, which is 71.35eV. These shifts suggested a partial charge transfer from Pt to Zr. Zr 3d binding energy shifts more than Pt 4f binding energy because the atomic ratio a b c 89 of Zr:Pt is 17:83, as a result, the interaction between platinum and zirconium on each zirconium atom is more significant than on each platinum atom. Figure 4.27 Co-sputter Pt 0.83-Zr 0.17 XPS spectra, (a) Zr-3d and (b) Pt-4f binding energy, XPS spectra were corrected using carbon spectra as a standard Table 4-10 Binding energy of co-sputter Pt 0.83-Zr 0.17, sputter Pt and sputter Zr Pt4f 7/2 Zr3d 5/2 Sputter Pt Co-sputter Pt 0.83-Zr 0.17 Sputter Zr Co-sputter Pt 0.83-Zr 0.17 Binding Energy 71.35eV 71.41eV 182.6 eV 181.6 eV In the methanol oxidation reaction, the strong binding interaction between CO and Pt poisons the active catalytic sites hence inhibits methanol oxidation reaction. This strong interaction Pt-4f Zr-3d a b 90 comes from a dual electron donation, first from the bonding σ CO orbital to Pt 5d orbital and then from the back–donation of the Pt 5d orbital to the anti–bonding 2π* CO orbitals as described in Chapter 3. The d-electron density decreases via electron transfer to Zr 3d, as a result, the Pt-CO binding energy would decrease. 4.3.3 Electrochemical Characterization of Platinum- Zirconium Catalysts The electrochemical characterization of the co-sputtered Pt 0.87-Zr 0.13 electrodes were performed in a half-cell. The measurement was performed by varying the voltage and measuring the current response in an Argon saturated 0.1 M perchloric acid solution. A methanol concentration of 1M was used in the methanol oxidation reaction. The amount of Pt is the same in all the electrodes. The loading of platinum divided by the geometric surface area of the electrode is 0.026mg/cm 2 . Electrochemical Surface Area – The oxidation and reduction peaks at -0.6V vs MSE in Fig 4-28 corresponded to hydrogen adsorption and desorption at platinum sites. Using the hydrogen desorption peak, the electrochemical active surface area values of the electrode were estimated assuming a monolayer adsorption of hydrogen atom on platinum sites and the equivalent charge density as 210 microcoulombs/cm 2 . The platinum electrochemical surface area of the co- sputtered Pt 0.87-Zr 0.13 catalyst was 3.7 times that of the platinum catalyst sputtered under the same conditions with the same amount of platinum. 91 Figure 4-28 The cyclic voltammetric response of co-sputtered Pt 0.87-Zr 0.13 at a scanning rate of 200mV/s in 0.1 M perchloric acid Onset Potential – The pseudo-steady state methanol oxidation currents for co-sputtered Pt 0.87- Zr 0.13 electrode, co-sputtered Pt-Ru electrode, sputtered Pt electrode and PtRu/C electrode were measured by anodically polarizing the electrodes from -0.5V to 0V vs MSE at the scan rate of 1mV/s and the results are shown in Fig. 4.29. Commercial PtRu/C catalyst had a slightly higher onset potential compared with co-sputtered Pt-Ru catalyst. Both were around -0.3V vs MSE. The onset potential of both co-sputtered Pt 0.87-Zr 0.13 electrode and sputtered Pt is -0.15V. The onset potential on methanol oxidation reaction is the potential when water activation begins. When the intermediate -OH ads can react with -CO ads on the platinum sites then a continuous oxidation current can be sustained. Hence the similar onset potential of Pt electrode and Pt-Zr electrode suggests water activation starts at the same potential on both electrodes. 92 Figure 4-29 pseudo-steady state methanol oxidation mass activities of co-sputter Pt 0.87-Zr 0.13, co-sputter Pt-Ru, sputter Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol Mass Activity – In Fig. 4.29, the methanol oxidation currents are presented after dividing by the mass of platinum. The commercial Pt-Ru/C catalyst has the highest oxidation current in -0.3V to -0.1V region. . In the region -0.1 V to 0 V the activity of the co-sputtered PtRu exceeds that of PtRu/C electrode. Compared to co-sputtered Pt 0.87-Zr 0.13 and sputtered Pt electrode, the methanol oxidation polarization curves of PtRu/C and co-sputtered Pt-Ru are similar. The methanol oxidation current of co-sputtered Pt 0.87-Zr 0.13 electrode is lower than the electrodes containing Ru in the region of -0.3V to 0V. But the currents are significantly improved comparing to co-sputtered Pt electrode. At -0.1V, methanol oxidation current of co-sputtered Pt 0.87-Zr 0.13 electrode is 11 times of sputtered Pt electrode. At 0 V, methanol oxidation current of co- 93 sputtered Pt 0.87-Zr 0.13 electrode is 4.7 times of sputter Pt electrode. This could be caused by either increased electrochemical surface area or increased methanol oxidation activities or both. Methanol Oxidation Specific Activity based on ECSA – In Fig. 4-30, the methanol oxidation currents are divided by the ECSA of platinum. The surface area of co-sputtered Pt 0.87-Zr 0.13 and sputtered Pt electrodes were estimated from hydrogen desorption charge. The surface area of PtRu/C was estimated from copper underpotential deposition. As we anodically polarize the electrode towards a positive potential, PtRu/C has better activity before -0.08V. Co-sputtered Pt 0.87-Zr 0.13 electrode activity exceeds PtRu/C after that. The activity of co-sputtered Pt 0.87-Zr 0.13 is higher than Pt electrode once it reaches a higher potential in the whole potential region. This proved that the higher mass activity of co-sputtered Pt 0.87-Zr 0.13 compared to Pt in that potential region is not only because of the higher surface area of Pt 0.87-Zr 0.13 electrode but by also enhancements of other catalytic pathways. The results also suggest introducing zirconium to the catalytic system could affect the methanol oxidation mechanism of the pure Pt electrode. The possible ways include facilitating the water dissociation step after reaching a potential higher then -0.15V vs MSE and lowering Pt-CO binding energy. 94 Figure 4-30 ECSA activities measured by pseudo-steady state polarization of co-sputtered Pt 0.87- Zr 0.13, sputtered Pt and PtRu/C in 0.1 M perchloric acid solution with 1M methanol Table 4-11 ECSA of co-sputtered Pt 0.87-Zr 0. catalyst compared to sputtered Pt and PtRu/C catalysts at -0.1 V and 0 V vs MSE Co-sputtered Pt 0.87-Zr 0.13 catalyst electrochemical area specific activity At -0.1V vs MSE 1.3 times of Pt 0.54 times of PtRu/C At 0V vs MSE 2.9 times of Pt 4.1 times of PtRu/C 95 Tafel Slope - Tafel plots and Tafel slopes for co-sputtered Pt 0.87-Zr 0.13, sputtered Pt and PtRu/C electrodes are shown in Fig. 4-31 and Table 4-12. The Tafel slope of Pt is close to the theoretical value with a one-electron transfer step as the rate determining step. The Tafel slope of Pt 0.87- Zr 0.13 and PtRu/C in the region -0.3V < V < -0.15V vs MSE potential indicates that the rate determining step could be a surface recombination step. The Tafel slope of PtRu/C at -0.15V < V < 0V is much higher than expected for the surface combination rate-determining step. This suggested the formation of new adsorbed intermediates. 46 Figure 4-31 Methanol oxidation reaction Tafel plots of co-sputter Pt 0.87-Zr 0.13, sputter Pt and PtRu/C Electrode Tafel slope Potential Range 96 Co-sputter Pt 0.87-Zr 0.13 132 mV/dec 78 mV/dec -0.3V < V < -0.15V -0.15V < V < 0V Sputter Pt 117 mV/dec -0.3V < V < 0V PtRu/C 143 mV/dec 376 mV/dec -0.3V < V < -0.15V -0.15V < V < 0V Table 4-12 Tafel slopes of co-sputter Pt 0.87-Zr 0.13, sputter Pt and PtRu/C Stability – We found all the results of co-sputtered Pt-Zr electrodes are not repeatable. The methanol oxidation current of co-sputter Pt 0.87-Zr 0.13 significantly decreases on the second day of testing as shown in Fig.4-32. The methanol oxidation current in the second-day of testing is only 23% of the original current. In addition, EDX analysis shows that the Zr:Pt atomic ratio decreased from 0.21 to 0.10 during electrochemical experiments. The Pourbaix diagram of Zr 48 in Fig.4-33 reveals the reason. The experimental environment is at a pH=1, and the electrode potential of interest is in the range of 0<E<0.65 V vs. NHE. In this region, zirconium oxide dissolves as ZrO 2+ ion in the solution. The decreased methanol oxidation current is therefore attributable to a large portion of zirconium in co-sputter Pt-Zr dissolved and so a large fraction of the Pt-Zr electrode became pure Pt. Furthermore, part of the co-sputtered Pt-Zr40 electrode layer peeled off after electrochemical measurements as shown in Fig.4-34. This phenomenon is because the dissolution of zirconium oxide destroyed the layered structure and caused the loss of active catalyst materials. This is another reason that the methanol oxidation current decreased as shown in Fig.4-32. Although zirconium could enhance the activity of platinum catalysts, zirconium is not a suitable catalyst material in direct methanol fuel cell because it is not stable in the DMFC environment. 97 Figure 4-32 Methanol oxidation current of co-sputter Pt 0.87-Zr 0.13 before and after electrochemistry tests Figure 4-33 The Peourbaix diagram of zirconium at 298K adopted from Chen et al. 98 Figure 4-34 SEM images of (a), (b) co-sputter Pt 0.87-Zr 0.13 on AvCarb MGL 190 before electrochemistry tests and (c),(d) co-sputter Pt 0.87-Zr 0.13 on AvCarb MGL 190 after electrochemistry tests 4.3.4 Effects of Platinum to Zirconium Ratio We prepared a series of co-sputtered Pt-Zr electrodes with the same amount of platinum but different amount of zirconium by maintaining the sputter rate of platinum and changing the sputter rate of zirconium for each electrode. The atomic ratio of Pt: Zr was measured by EDX a b c d 99 elemental analysis. We found out the factors that change with varying the atomic ratio are Pt electrochemical surface area, methanol oxidation activity and Tafel slope. Electrochemical Active Surface Area – As shown in Fig. 4-35, the electrochemical surface area of co-sputterd Pt-Zr electrodes increases with increasing Zr:Pt ratio, this is another evidence of zirconium dissolution in the acidic solution. The more zirconium in the electrode, the more porous the structure it will create after zirconium gets dissolved, hence the more the platinum surface that will be exposed. The highest surface area is the electrode that had originally the largest amount of zirconium. The platinum electrochemical surface area of it is 5.5 times of the sputtered Pt electrode which contains the same amount of platinum. Figure 4-35 The effect of Zr: Pt atomic ratio on the electrochemical surface area Electrochemical Surface Area Methanol Oxidation Activity – The fact that Zr:Pt atomic ratio affects the methanol oxidation ECSA activity shown in Fig. 4-36 suggests that zirconium affects the methanol oxidation current of platinum catalyst in a way not only through increased 100 electrochemical surface area, but also by the ECSA activity. In other words, zirconium affects the mechanism of methanol oxidation of platinum catalysts. Figure 4-36 The effect of Zr: Pt atomic ratio on ECSA methanol oxidation activity Tafel Slope – In Fig. 4-37 the Tafel slope of co-sputtered Pt-Zr electrode starts at 80mV/dec, was much lower than that of the Pt electrode. As the ratio of Zr:Pt increases, the Tafel slope also increases, and ends with a Tafel slope very close to that of the Pt electrode. This result confirmed that zirconium affects the mechanism of methanol oxidation of platinum catalysts. 101 Figure. 4-37 The effect of Zr: Pt atomic ratio on Tafel slope There are a couple of evidences that suggest that Pt-Zr has superior catalytic activity to platinum. First, the ECSA activity of co-sputtered Pt-Zr is higher than that of sputtered Pt. Besides, the ECSA activity changes with varying Pt: Zr Ratio. Based on the metal oxide’s capability of activating water molecule and previous research, we propose the bi-functional mechanism to be operative on the Pt-Zr catalysts. Methanol Dissociation: Pt + CH 3OH → Pt-H + Pt-CO Water Dissociation: Zr + H 2O → Zr-H + Zr-OH Surface Recombination: Pt-CO + Zr-OH → Pt + Zr + CO 2 + H + +e - 4.4 Comparison of Platinum-M (M=Tantalum, Niobium, Titanium, Zirconium) Catalysts A series of platinum-M (M=tantalum, niobium, titanium and zirconium) were prepared by sputter deposition method. The properties including electrochemical surface area, methanol oxidation 102 activities and stabilities will be compared between Pt-M catalysts and the differences will be discussed in this section. 4.4.1 Electrochemical Surface Area As mentioned earlier, hydrogen adsorption/desorption does not happen on M sites. In terms of electrochemical surface area, the role of M nanoparticles is to improve the total electrochemical surface area of Pt-M. So that the Pt:M atomic ratio that gives the highest ECSA is different for each metal. The smaller the M particle is, the fewer particles are needed to achieve the maximum ECSA. However, the smaller the M particle is, the more compact the Pt-M can get. Therefore, small M particle will result in small maximum ECSA. Our observation is consistent with this trend as shown in Table 4-13, tantalum particle has the largest size, niobium particle in the middle, and titanium has the smallest one. The maximum ECSA follows the order Ta > Nb > Ti and M: Pt atomic ratio of the maximum is Ta > Nb > Ti. Zirconium is not stable in the experiment measuring ECSA, so will not be included in this discussion. Table 4-13 The maximum electrochemical surface area of Pt-M catalysts Tantalum Niobium Titanium Maximum ECSA 23.2 cm 2 17.1 cm 2 11.6 cm 2 M : Pt atomic ratio 0.36 0.17 0.088 4.4.2 Methanol Oxidation Activities Both platinum-tantalum and platinum-zirconium catalyst have relatively high electrochemical surface area specific activities at 0V vs MSE for methanol oxidation among the platinum-M (M = 103 tantalum, niobium, titanium and zirconium) electrodes made by sputter deposition as shown in Fig. 4-39. Platinum-tantalum is more active at potentials lower than 0 V and has a lower onset potential. So we conclude that Pt-Ta is the most active among the Pt-M catalysts made by co- sputter deposition. Figure 4-39 The electrochemical surface area activity of co-sputter platinum-M (M = tantalum, niobium, titanium and zirconium) electrodes for methanol oxidation The oxidation of carbon monoxide and methanol oxidation share some similarities. They both involve the removal of absorbed carbon monoxide from platinum sites. According to Atsushi Ueda et al.’s research, the addition of TaO x and NbO x to the catalytic platinum anode enhances the electrochemical oxidation of carbon monoxide. 23 And they found that Pt-TaO x has better activity than Pt-NbO x. This is consistent with our finding. They also believe that the carbon 104 monoxide adsorption on the platinum surface is inhibited by the presence of TaO x which should weaken the Pt-CO bond. This is again consistent with our observation. In Table 4-14, the shift of Pt4f binding energy caused by tantalum (0.12eV) is higher than niobium (0.05V). The shift caused by titanium is even lower, which lead to lower methanol oxidation activity as we discovered in the half cell experiments. Although most of zirconium dissolves in the acidic solution, the shift of Pt4f binding energy caused by zirconium (0.06eV) is stronger than that caused by niobium. Table 4-15 shows the shift of binding energy of M (M=Ta, Nb, Ti, Zr) caused by platinum. The shift is in the same order (Ta > Zr > Nb > Ti) as that of Pt4f shift caused by M. Based on these evidence, we believe the strength of the interactions between platinum and metal M is in the order of Ta > Zr > Nb > Ti. Table 4-14 The Pt 4f 7/2 binding energy of electrodes prepared by sputter deposition Sputter Pt Co-sputtered Pt-Ta Co-sputtered Pt-Nb Co-sputtered Pt-Ti Co-sputtered Pt-Zr Pt4f 7/2 Binding Energy (BE) 71.35eV 71.47eV 71.40eV 71.33eV 71.41eV ΔBE = BE co-sputter Pt-M – BE sputter Pt 0 eV 0.12eV 0.05eV 0.02eV 0.06eV Table 4-15 The binding energy of metal M, M = Ta, Nb, Ti, Zr in the electrodes prepared by sputter deposition Ta 4f7/2 Nb 4f7/2 Ti p3/2 Zr 3d5/2 105 BE M M=Ta, Nb, Ti, Zr Sputter Ta Co-sputter Pt-Ta Sputter Nb Co-sputter Pt-Nb Sputter Ti Co-sputter Pt-Ti Sputter Zr Co-sputter Pt-Zr 26.81eV 25.75eV 207.5eV 206.7eV 458.8eV 458.1eV 182.6 eV 181.6 eV ΔBE M= BE sputter M – BE co-sputter Pt-M 1.06eV 0.8eV 0.7eV 1.0eV 4.5 Summary Among all the catalysts, co-sputtered Pt-Ta catalysts showed the highest methanol oxidation activity and stability. In the series of co-sputtered Pt-Ta examined in Chapter 3, co-sputter Pt- Ta40 has the highest methanol oxidation activity and good stability in half-cell experiments. In the next chapter, we will focus on the co-sputtered Pt-Ta40 catalyst. Full cell experiments will be carried out and several factors that affect the performance of direct methanol fuel cell will be discussed. 106 Chapter 5 Properties of Membrane Electrode Assemblies Employing Sputter-deposited Electrocatalysts for Methanol Oxidation As discussed in the previous chapter, among the co-sputtered catalysts, Pt-Ta, Pt-Nb, Pt-Ti, and Pt-Zr, Pt-Ta catalysts showed the highest methanol oxidation activity combined with durability. In the series of co-sputtered Pt-Ta catalysts examined in Chapter 3, co-sputtered Pt 0.77-Ta 0.23 had the highest methanol oxidation activity in half-cell experiments. In this chapter, we focus on co- sputtered Pt 0.77-Ta 0.23 catalyst. We prepared membrane electrode assemblies (MEAs) with both Pt-Ta and commercial PtRu/C catalysts. These MEAs consisted of a proton-conducting membrane sandwiched between two catalytic layers. The design and preparation of these MEAs have been described in Section 3.1. The performance of these MEAs as direct methanol fuel cells at different temperature values was tested in the full cell mode. In this chapter, we will also discuss the factors that affect the catalyst materials utilization when we apply co-sputtered Pt-Ta to prepare an MEA to be operated in a fuel cell. In these experiments, we keep the amount of catalyst constant and aim at achieving the highest methanol oxidation current. To achieve this, we increased at first the surface area of the carbon substrate that was used in the previous half-cell experiments. This surface area enhancement was achieved by coating the carbon nanotubes (CNTs) on carbon fiber composite paper. We investigated the factors that affect the final morphology of the carbon layer including the pre-treatment of CNTs, the coating techniques and the loading of CNTs. Secondly, we explored the effect of the chamber pressure during sputter-deposition on the utilization of sputtered materials. Thirdly, the effect of heating the sputtered substrate during deposition to improve atom surface mobilities was 107 investigated. Fourthly, the wettability of the resulting electrodes were studied. Finally, the effect of changing the amount of the ionomer added to the catalyst layer relative to the amount of carbon was also investigated. 5.1 Membrane Electrode Assembly Preparation MEAs consisted of co-sputtered Pt 0.77-Ta 0.23 as anode, Nafion 117 as the proton-conducting membrane and Pt black as cathode. The anode was prepared by co-sputtering Pt at 40W and Ta at 45W onto a 5cm by 5cm carbon fiber composite paper (Avcarb MGL 190) coated with carbon nanotubes. The background pressure during deposition was remained at 30 mTorr by maintaining the argon flow rate at 139 sccm. The total catalyst loading of the anode was 0.15 mg/cm 2 . A thin coating of Nafion ionomer was sprayed over the anode to make a continuous proton conducting path between the catalyst materials and membrane. To prevent cathode polarization and methanol crossover from affecting the results, cathode catalyst layers were applied as an ink formed by combining Pt-black and Nafion ionomer. The catalyst loading on the cathode was 2 mg/cm 2 of Pt black and maintained constant throughout the course of the studies. The MEA was pressed at 140ºC at 1000 pounds for 15 minutes. 5.2 Membrane Electrode Assembly Performance of Co-sputter Pt-Ta Catalysts Figure 5-1 shows cell voltages and cell power of the direct methanol fuel cells with co-sputtered Pt-Ta and PtRu/C as anode catalysts, respectively. The details of the cell assembly were described in Chapter 1. The catalyst loading values for Pt-Ta and Pt-Ru/C were similar; the loading of platinum in PtRu/C electrode was calculated to be 0.6706 mg, while that of co-sputter Pt-Ta electrode was measured to be 0.69 mg (0.028mgPt/cm 2). Both cells were operated at three 108 different temperature values, 90°C, 80°C and 60°C, in 1 M methanol with 2 lpm oxygen at 10 psi. The geometric area of the electrodes was 25 cm 2 . a b c d e f 109 Figure.5-1. The cell voltage and cell power of the direct methanol fuel cells with co-sputtered Pt- Ta and PtRu/C as anode catalysts at 90°C (a)(b), 80°C (c)(d) and 60°C (e)(f), in 1 M methanol with 2 lpm oxygen at 10 psi. The maximum power densities of both Pt-Ta and PtRu/C at different operating temperature are shown in Table 5-1. The maximum power density and open circuit potential of both Pt-Ta and PtRu/C MEAs increase with operating temperature. The maximum power of MEA with Pt-Ta is higher than with PtRu/C at all the temperatures we tested out. The best performance for both PtRu/C and Pt-Ta were attained at 90°C. For co-sputtered Pt-Ta, a cell voltage of 300mV was realized at 139 mA/cm 2 /mg Pt (95.6 mA/cm 2 ), and the cell gives a voltage of 183 mV at 300 mA/cm 2 . The maximum power density achieved by Pt-Ta was 82 mW/cm 2 /mg Pt (57mW/cm 2 with a platinum loading of only 0.026mg/cm 2 ), about 2 times of that of a state-of-the-art PtRu/C catalyst. This result is comparable to C. K. Witham’s PtRu MEA made by sputter deposition. 49 In his work, a power density of 60 mW/cm 2 was achieved with 0.07 mg/cm 2 Pt loading and 75mW/cm2 was achieved with 0.02 mg/cm 2 Pt loading. The maximum current density of Pt-Ta was 778mA/cm 2 /mg Pt, 2.7 times that of PtRu/C. Table 5-1 Maximum power densities and current densities of Pt-Ta and PtRu/C at 90°C, 80°C and 60°C MEAs 90°C 80°C 60°C Co-sputter Pt-Ta Maximum Power Density mW/cm 2 /mg Pt 82 59 21 Maximum Current Density 778 635 314 110 mA/cm 2 /mg Pt PtRu/C Maximum Power Density mW/cm 2 /mg Pt 45 25 6.0 Maximum Current Density mA/cm 2 /mg Pt 289 180 50.1 We report excellent MEA performance for co-sputtered Pt-Ta in Section 5.1 by optimizing the MEA in different ways. In the following sections, we will discuss the factors that affected the performance of the cell and how we improved them. 5.3 Effects of Carbon Catalyst Supports In our membrane electrode assemblies, the Pt-Ta catalyst is co-sputterdeposited onto to the carbon substrate. As a result, the surface area of the Pt-Ta catalyst is affected by the surface of the carbon substrate. Instead of coating the catalyst directly on the substrate as what is commonly practiced with PtRu/C, we used carbon nanotubes to increase the surface area of the carbon fiber composite paper. This strategy of utilizing co-sputtered catalysts is shown in Fig. 5- 2. With the same amount of co-sputtered Pt-Ta catalyst, the total surface area of the co-sputter Pt-Ta catalyst was increased 3 times with the carbon nanotube substrate. The surface areas of the sputter-deposited catalysts are shown to be affected by the coating techniques and the loading of CNTs. These factors will be discussed in this section. 111 Figure. 5-2 The new strategy of utilizing co-sputtered catalyst materials by coating carbon fiber composite paper substrate with CNTs before sputter deposition 5.3.1 The Effect of CNT Coating Techniques Carbon Nanotube Heat-treatment – Two electrodes consisting of co-sputtered Pt-Ta were prepared using the same strategy shown in Fig. 5-2. One of the electrode substrates was coated with CNTs that had been heated at 400°C in air for 6 hours and was labeled co-sputter Pt-Ta/H- CNT. The other electrode substrate was coated with CNTs without any heat-treatment and has been labeled co-sputtered Pt-Ta/CNT. We found the electrochemical surface area of co- sputtered Pt-Ta/H-CNT is 2 times that of the co-sputtered Pt-Ta/CNT. The electrochemical surface areas of both the electrodes were calculated from the charge for hydrogen desorption ( Fig. 5-3). Although the SEM images of the CNT coated carbon fiber composite paper show no obvious difference before and after heat treatment, TGA analysis from literature showed the heat- treatment of CNT at 400°C in air caused a 10-20% weight loss due to the oxidation and removal 112 of the disordered carbon. 50 In conclusion, the 400°C heat-treatment of CNTs in air yielded a higher ECSA for the co-sputtered Pt-Ta catalysts. In the subsequent experiments, all the carbon nanotubes were heat-treated. Figure 5-3 Cyclic voltammogram of co-sputtered Pt-Ta/CNT and co-sputtered Pt-Ta/H-CNT in 0.1M perchloric acid solution with 200mV scanning rate. Carbon Nanotube Dispersing – Multi-walled carbon nanotubes are not only strong and flexible but also very cohesive. Carbon nanotubes are generally available as dry material. They are difficult to disperse in liquids. To utilize the nanotubes to their maximum potential, ultrasound is an effective method to obtain discrete carbon nanotubes. We made a dilute carbon nanotubes ink with isopropanol (1mg CNT in 10 ml isopropanol). Then we used powerful ultrasonic dispersing for different periods. The CNTs in the electrode labeled co-sputtered Pt-Ta/H- 113 CNT0.25h were dispersed with ultrasonicator for 0.25 hour. The CNTs in the electrode designated as co-sputtered Pt-Ta/H-CNT1h were dispersed using ultrasonication for 1 hour. We then compared the electrochemical surface area of two co-sputtered Pt-Ta electrodes. The ECSA of co-sputtered Pt-Ta/H-CNT1h is 1.75 times of that of co-sputtered Pt-Ta/H-CNT0.25h. As shown in Fig. 5-4, the carbon nanotubes are better separated after 1 hour’s ultrasonic dispersion (b) compared to 0.25 hour (c). The carbon nanotubes without ultrasonic dispersion (a) are big chunks of aggregated carbon nanotubes. In conclusion, in order to achieve a high surface area of the co- sputtered Pt-Ta for the MEA, the one-hour ultrasonic dispersing of the carbon nanotubes is a necessary process. In all the later experiments, the carbon nanotubes were ultrasonic dispersed for one hour to achieve a good dispersion. 114 Figure 5-4 SEM images of co-sputtered Pt-Ta on carbon fiber composite paper coated with CNTs: (a) CNTs were not pre-treated (b) CNTs were ultrasonic dispersed for 15 minutes, (c) CNTs were ultrasonic dispersed for 1 hour Carbon Nanotube Coating Methods – After dispersing the CNTs, we tried to find a way to coat CNTs evenly on the carbon fiber composite paper. First, CNT ink was coated by brush. However, instead of being coated on the carbon fiber composite paper, CNTs stayed on the brush. As shown in Fig.5-5 (a), there was barely any CNT left on the substrate even after brush-coating multiple times. Then we also tried adding catalyst ink drop by drop with pipette. As shown in Fig.5-5 (b), a b c 115 the CNT settled as numerous small circles because of the interfacial tension of the carbon fiber composite paper. We also added Nafion to the CNT ink in order to make CNTs sticky to the substrate surface and thus there was more CNT on the surface of the substrate. But an even coating of CNTs was not achieved after either single brush-coating or multiple coats. Figure. 5-5 : 2 cm x 2 cm carbon fiber composite paper coated by (a) coating CNT ink with brush for multiple times, (b) adding CNT ink drop by drop with pipette, (c) coating CNT and Nafion ink with brush, (d) coating CNT and Nafion ink with brush multiple times Finally, spray-coating proved to be the best method. The CNT ink without any Nafion was sprayed on the carbon paper. The spray was almost like fog so that the effect of surface tension was much smaller compared to adding drop by drop with pipette. Besides, without any tool touching the paper and taking away CNTs, multiple layers of CNTs were sprayed on the substrate and an even coating of CNTs was achieved as shown in Fig. 5-6. In the subsequent experiments, all the electrode substrates were prepared in this manner. a b c d 116 Figure. 5-6 2cm x 2cm carbon fiber composite paper coated with CNTs by spraying method, CNT loading is 0.2mg/cm2 5.3.2 The Effect of CNT Loading The electrochemical surface area of co-sputtered Pt-Ta electrodes with different CNT loadings in Fig. 5-7 was calculated from the hydrogen desorption charge. The electrochemical surface area first increases with increasing CNT loading, then reaches the maximum ECSA at the CNT loading of 0.008mg/cm 2 . The ECSA of the co-sputtered Pt-Ta45 at this CNT loading is 17 times of the geometric surface area of the electrode and is 4 times of the ECSA of a co-sputtered Pt-Ta electrode without CNT coated on the carbon paper. As shown in Fig. 5-8 a (CNT loading 0.005 mg/cm 2 ) and b (CNT loading 0.008 mg/cm 2 ), the carbon nanotubes increased the surface area of carbon fiber paper by creating a porous structure on the surface. The structure appeared more porous as more carbon nanotubes are coated on the carbon paper. In Fig.5-8 (b), CNTs cover 117 more of the surface of carbon fiber paper compared to In Fig.5-8 (a), so the surface area of the resulting carbon substrate in Fig.5-8 (b) is also larger and so also the Pt-Ta catalyst deposited on the substrate. However, too much CNTs will not only cover carbon fibers but also block the pores in the carbon fiber paper as shown in In Fig.5-8 (c). As a result, the total surface area of carbon substrate started to decrease after reaching the maximum in Fig.5-7. From this series of experiments we concluded that a carbon nanotube loading of 0.008 mg/cm 2 will lead to the largest surface area of carbon substrate and we used this loading of carbon nanotube to make all the electrodes mentioned below. Figure. 5-7 Electrochemical active surface area of co-sputtered Pt-Ta electrodes with different levels of CNT loading. 118 Figure 5-8 SEM images of co-sputter Pt-Ta on carbon fiber composite paper that were coated with different loading of CNTs (a) 0.005 mg/cm 2 (b) 0.008 mg/cm 2 (c) 0.05 mg/cm 2 5.4 Effects of Pressure during Sputtering. Among the different thin film methods shown in Fig.5-9, chemical vapor deposition has the best step coverage. To fully take advantage of the porous carbon substrate, we want to improve the step coverage of sputter deposition. One way is to increase the sputter pressure, the other way a b c 119 is to increase the temperature of the substrate. We will show our results and discuss about these two methods in section 3.5 and 3.6. Figure 5-9 The step coverage difference of various thin film deposition methods including chemical vaper deposition (A), sputter deposition (B, C) and evaporation deposition (D) In this section, co-sputtered Pt-Ta electrodes were prepared under various values of pressure during sputtering, 10, 29 and 50 mTorr. The sputter pressure was controlled by controlling the argon gas flow rate during the sputter deposition process. The chamber pressure was desired to be in the range between 5 to 100 mTorr in order to form a plasma. In this range, the chamber pressure increases with increasing argon gas flow rate and an empirical linear relation between argon flow rate and chamber pressure was as shown in Fig. 5-10. With this relation, we are able to estimate the argon flow rate needed to achieve a certain sputter pressure. 120 Figure 5-10 The linear relation between argon flow rate and the resulting chamber pressure during co-sputtering of platinum and tantalum experiments We found that the methanol oxidation current of the co-sputtered Pt-Ta electrode is affected by the pressure during sputtering. As shown in Fig. 5-11, the electrodes prepared under 50 mTorr has the highest methanol oxidation mass activity. y = 0.1662x + 3.6733 0 10 20 30 40 50 60 0 100 200 300 Chamber Pressure (mTorr) Ar Flow Rate (sccm) Ar Flow Rate vs Chamber Pressure 121 Figure 5-11. Methanol oxidation mass activity of co-sputtered Pt-Ta electrodes prepared under three different values of pressure during sputtering. To understand why the methanol oxidation current increases with sputtering pressure, we studied the electrochemical surface area of the co-sputter-deposited Pt-Ta electrodes. We used two methods, hydrogen adsorption/desorption method, which has been mentioned in the previous chapters, and the double-layer capacitance method. In Fig 5-12, both the hydrogen adsorption, desorption regions and the double layer region area indicated in the cyclic voltammogram of the platinum electrode. Hydrogen adsorption/desorption only happens on the platinum sites, and rarely on the tantalum sites. Therefore, the electrochemical surface area estimated by this method is primarily that of platinum. On the other hand, double layer charging and discharging happen both on platinum and tantalum sites. Therefore, the double-layer capacitance reflects both platinum and tantalum surface sites of the catalysts. 122 Figure 5-12 The cyclic voltammogram of a platinum electrode in acidic solution 51 Figure 5-13 Double layer current vs scanning rate of cyclic voltammetry 123 To estimate the double-layer capacitance of the electrode, we took the currents in the double layer region and plotted them versus the scan rates, and the slope of the linear fitting is the double layer capacitance of the electrodes in Fig. 5-13. The double-layer capacitance of the three Pt-Ta electrodes sputtered under 10, 29 and 50 mTorr are listed in Table 5-1 together with their electrochemical surface area estimated from hydrogen deposition area. Table 5-1 Electrochemical surface areas and double layer capacitances of co-sputter Pt-Ta electrode prepared under different sputter pressure 10 mTorr 29 mTorr 50 mTorr ECSA estimated from Hydrogen Desorption Area 48 cm 2 49 cm 2 66 cm 2 Double-layer Capacitance 0.001F 0.002F 0.004F In Table 5-1, both platinum electrochemical surface area and double layer capacitance increased with increasing sputter pressure. This is because the flux of the sputtered material reaching the substrate depends on the sputter pressure. Higher pressure leads to more collisions and more scattering. At higher pressure the atoms are thermalized and this results in more isotropic flux of atoms and hence better coverage. On the other hand, lower pressures result in a more directed flow which results in less uniform films. From Table 5-1 we also found that the effect of changing sputter pressure is greater on the double layer capacitance compared to the hydrogen desorption on platinum. This observation 124 suggests the effect of pressure on the electrochemical surface area of tantalum is greater than that on the electrochemical surface area of platinum. 5.5 Effects of Heating Substrate during Sputter Deposition We soaked the CNT coated carbon nanofiber composite paper substrate at 200 o C for 900 seconds before co-sputtering platinum and tantalum. The rest of the experiment procedure was consistent with co-sputtered Pt-Ta/CNTs electrode. This electrode was labeled as co-sputter Pt- Ta/CNTs_200deg. The intent was to increase the surface mobility of the sputtered platinum and tantalum atoms when they reach the heated substrate. So that the platinum and tantalum atoms could spread better on the substrate and get into the porous structure. However, we found the methanol oxidation current decreased because of raising substrate temperature as shown in Fig. 5-14. By estimating the electrochemical surface area of the electrodes, we found that the decreased methanol oxidation current is caused by decreased surface area as shown in Fig. 5-15. The electrochemical surface area calculated from hydrogen desorption charge of co-sputter Pt- Ta/CNTs_200deg decreased to 49% compared with co-sputtered Pt-Ta/CNTs without heating. The possible reason is that once the surface mobility of platinum and tantalum atoms increase, and the atoms tend to aggregate. Depositing metal on a high temperature substrate is comparable to the sintering process, which usually produces larger particles and smaller surface area. And in this case the sintering effect is greater than the atoms spreading effect caused by increasing surface atom mobility. 125 Figure 5-14 Methanol oxidation mass activity of co-sputter Pt-Ta electrode Figure 5-15 Cyclic voltammetry of the co-sputtered Pt-Ta/CNT and co-sputtered Pt- Ta/CNT_200deg electrodes -0.8 -0.6 -0.4 -0.2 0 1E-7 1E-5 1E-3 1E-1 E(V) vs MSE I/mgPt (A/mg) Methanol Oxidation Mass Activity Sputter Pt PtRu/C Co-sputter PtTa/CNTs Co-sputter PtTa/CNTs_200 deg -0.02 -0.01 0 0.01 0.02 -1 -0.5 0 0.5 1 I (A) E (V) vs MSE Polarization Curve 200mV/s Co-sputter Pt-Ta/CNT Co-sputter Pt-Ta/CNT 200 degree 126 Figure 5-16 Methanol oxidation electrochemical surface area activity of co-sputtered Pt-Ta/CNTs and co-sputtered Pt-Ta/CNTs_200deg electrodes. Although the methanol oxidation current of co-sputtered Pt-Ta/CNTs_200deg electrode decreased because of the temperature of the substrate, we found out the methanol oxidation activity did not change ( Fig. 5-16). This finding suggests that heating the substrate only affected the catalyst surface area. 5.6 Effects of Catalysts Wettability In an MEA test, the wettability of the electrodes affects the methanol oxidation current. Because the methanol oxidation process requires good contact of the catalyst layer to both methanol and water. Triton X-100 (C 14H 22O(C 2H 4O) n) is a nonionic surfactant that has a hydrophilic polyethylene oxide chain. It is soluble at 25 °C in water and could improve the wettability of the electrode. In 127 order to achieve the best wettability for our catalyst layer in the fuel cell environment, we tried two ways to increase the wettability of the co-sputtered Pt-Ta/CNTs electrode. Firstly, by adding Triton X-100 (0.1g/L) to the 0.1M perchloric acid solution in which the electrode was tested out. The solution was stirred for 15 minutes before electrochemistry measurement. Secondly we first wetted the electrode with methanol and then quickly removed it from methanol solution and put it in water at 80 degrees Celsius. This way of wetting the electrode is similar to the real direct methanol fuel cell environment, where the temperature is around 80 degrees Celsius and methanol concentration is 1M. The wettability test was done in a half cell. The cyclic voltammetry of the co-sputtered Pt-Ta/CNT electrode wetted by the two methods mentioned above are shown in Fig. 5-17. A greater electrochemical surface area indicated better wettability. We found out that the electrochemical surface area of the electrode without wetting agent is the smallest one. Triton X100 could improve the wettability of the electrode, but the same level of wettability could be achieved by soaking in methanol and hot water. 128 Figure 5-17 Cyclic voltammetry of the co-sputtered Pt-Ta/CNT electrode after wetting by different methods However, the greater electrochemical surface area achieved with Triton surfactant did not make the methanol oxidation current of the co-sputter Pt-Ta catalysts greater, on the contrary, Triton X100 caused methanol oxidation activity loss as shown in Fig. 5-18. This might because the adsorption of the surfactant interfered with the adsorption of methanol on the platinum sites and therefore affected the methanol oxidation process. 129 Figure 5-18 Methanol oxidation electrochemical surface are of the co-sputtered Pt-Ta/CNT electrode wet by different methods Later, we washed the Pt-Ta electrode several times with water to remove Triton, and we also used a freshly made acid solution replacing the solution containing Triton x100 in the half cell experiment. The methanol oxidation current of the electrode recovered immediately as shown in Fig. 5-19. This confirmed that Triton interfered with the methanol oxidation reaction as noted in the surface area measurements. 130 Figure 5-19 Methanol oxidation current of the co-sputter Pt-Ta/CNT electrode wet by different methods 5.7 Effects of Amount of Nafion Ionomer. In every membrane-electrode assembly, the contact of catalysts to the Nafion membrane is achieved by first coating Nafion ionomer solution on the catalyst layer then applying heat and pressure forthe catalyst to bond with the Nafion membrane. A bad contact between the catalyst and Nafion materials could affect the proton conductivity between the catalyst layer and membrane and hence affect the cell power density. In our experiment, after Pt-Ta sputter deposition, Nafion solution was spray-coated on the Pt-Ta electrode, followed by a drying in air at 80 o C. The ratio of the sprayed Nafion ionomer to catalysts was found to be important. Too less of the Nafion ionomer will not guarantee a good contact of catalyst and Nafion membrane. Too many Nafion molecules in the catalyst layer will also block the access of the catalyst and prevent the 131 methanol from diffusing to the active sites on the catalysts. Both these situations will lower the direct methanol fuel cell power output. We tested two MEAs made with different mass ratios of PtTa/CNT to Liquion (Nafion ionomer soluiton), 1:2.5 and 1:5, the results are shown in Fig. 5-20. The one with 1:5 catalyst to Nafion ionomer had a lower maximum power density. This suggested that this electrode had too much Nafion ionomer that prevented the methanol mass transfer at high current density. Figure 5-20 Cell power of the direct methanol fuel cells with co-sputtered Pt-Ta as anode catalysts at 60°C and 80°C in 1 M methanol with 2 lpm oxygen. 5.8 The Optimized Preparation of Co-sputtered Pt-Ta/CNTs electrodes In this chapter, we explored several factors that affected the methanol oxidation current of the co-sputtered Pt-Ta catalyst in a direct methanol fuel cell membrane-electrode-assembly. We found the optimized procedures of making co-sputtered Pt-Ta/CNT electrode as per the following: 132 1. Heat treat CNT at 400 o C for 6 hours 2. Make a dilute carbon nanotubes ink with isopropanol (1mg CNT in 10 ml isopropanol) 3. Powerful ultrasonic dispersing for 1 hour 4. Spray the mixture on the carbon fiber composite paper 5. Co-sputter-deposit Pt and Ta at 280 sccm Ar flow rate 6. Spray Nafion ionomer solution (Liquion) on the co-sputter Pt-Ta/CNTs electrode, the mass ratio of Pt-Ta/CNT to Liquion is 1:2.5. 7. Air-dry at 80 o C The anode in the MEA was prepared by the above steps to achieve the performance reported in Section 5.1. 133 Chapter 6 Platinum-Ti 0.7 Ru 0.3 O 2 Catalysts 6.1 Introduction As discussed in the previous chapters, direct methanol fuel anode catalysts are faced the challenges of stability and cost. We had discussed how Pt-Ta catalysts can meet this challenge. In this chapter, we discuss results from a different approach using oxide support catalysts to solve the above-mentioned problems. Specifically, we explored the use Ti 0.7Ru 0.3O 2 nanoparticles as a support for the catalytic Pt nanoparticles as shown in Fig. 6-1 a. Figure 6-1 Structures of a) commercialized PtRu/C catalyst and b) Pt- Ti 0.7Ru 0.3O 2 catalyst Ti 0.7Ru 0.3O 2 is a well-known electrocatalyst used for chlorine evolution in the chlor-alkali industry for several decades now. 52,53, Combining titanium dioxide with ruthenium oxide increases the oxidative stability of these ruthenium oxide catalysts. Thus, a combination of highly stable TiO 2 forming a robust solid solution with RuO 2, which also noted for its good proton conductivity, 54,55 is used as an approach to develop a multifunctional support material for platinum that could enhance the methanol oxidation catalytic activity and durability. Ti 0.7Ru 0.3O 2 provides various advantages in terms of its high surface area, and the interactions between Pt and the co-catalytic metals in the support, leading to highly dispersed and well-anchored Pt catalyst particles. The a b 134 Ti 0.7Ru 0.3O 2 support acts as a co-catalyst, due to the ability of Ti 0.7Ru 0.3O 2 to dissociate water, and thus Pt-Ti 0.7Ru 0.3O 2 works as a bifunctional methanol oxidation catalyst. A research group from Taiwan synthesized Ti 0.7Ru 0.3O 2 as a support for Pt by the hydrothermal method and microwave- assisted polyol method. 56 The half-cell results showed the potential of using these materials as catalyst supports in direct methanol fuel cells and polymer electrolyte membrane fuel cells. In this chapter, we have examined the use of titanium oxide-ruthenium oxide solid solutions as co-catalyst supports. We prepared Pt-Ti 0.7Ru 0.3O 2 for use as an anode catalyst. Ti 0.7Ru 0.3O 2 as a co-catalyst support was synthesized in a one-step hydrothermal process and then platinum was deposited using the polyol method. The structure of the synthesized catalyst was investigated by XRD, SEM and EDX. Results of steady-state voltammetry suggested that the electrocatalytic activity and stability of the newly-synthesized catalysts compared well with the commercially- available carbon-supported Pt-Ru catalyst. The Tafel slopes suggested that the well-known bifunctional mechanism for methanol oxidation was operative on the new catalyst. 6.2 Pt-Ti 0.7Ru 0.3O 2 Catalyst Synthesis The Pt-Ti 0.7Ru 0.3O 2 Catalyst was prepared by two steps. First the Ti 0.7Ru 0.3O 2 nanoparticles were prepared using a one-step hydrothermal method. Then platinum nanoparticles were deposited on Ti 0.7Ru 0.3O 2 by the polyol method. The synthesis strategy is shown in Fig. 6-2. Figure 6-2 The synthesis strategy of Pt-Ti 0.7Ru 0.3O 2 catalysts 135 6.2.1 Hydrothermal Method for Synthesizing Ti 0.7Ru 0.3O 2 Ti 0.7Ru 0.3O 2 nanoparticles were prepared using a one-step hydrothermal method at low temperature as a low-energy consuming fabrication technique. 57 Hydrothermal method is a method of synthesis of single crystals that depends on the solubility of minerals in hot water under high pressure. The crystal growth is performed in an apparatus consisting of a steel pressure vessel called an autoclave, in which a precursor is supplied along with water. Advantages of the hydrothermal method is the ability to make crystalline materials in fine particulate form without the occurrence of sintering encountered in high-temperature methods. A 30mL an aqueous solution containing 12.00 mM ruthenium chloride (RuCl 3), 28.00 mM titanium chloride (TiCl 4) were heated at 200ºC in a 100 mL pressure tube for 2 hours and then cooled to room temperature. The sample was washed with water and collected by centrifugation several times until the washings showed a pH of 7. The precipitates were dried at 80ºC in a vacuum oven overnight. 6.2.2 Polyol Method synthesizing Pt-Ti 0.7Ru 0.3O 2 Platinum nanoparticles were anchored over the Ti 0.7Ru 0.3O 2 support using the polyol method. Tge polyol method is widely adopted to prepare carbon-supported Pt-Ru nanoparticles. The “polyol” is usually ethylene glycol or polyethylene glycol. The polyol is both a reducing agent and a protectant. The long chain molecule adsorbed onto the surface of the growing particles prevents aggregation of the nanoparticles. The reaction mixture consisted of 60 mg Ru 0.3Ti 0.7O 2 and 0.0337g NaOH added to 30mL ethylene glycol with magnetic stirring and ultrasonic treatment to form a homogeneous solution. The 136 solution was then heated to 180ºC in an oil bath and 2 mL of an aqueous solution of 0.05M H 2PtCl 6 was quickly added. After 10 min of reaction time, the solution was cooled to room temperature. The catalyst was then washed with water. The obtained catalyst was denoted as Pt-Ti 0.7Ru 0.3O 2. 6.3 Results and Discussion 6.3.1 Ti 0.7Ru 0.3O 2 Nanoparticles X-Ray Diffraction - The structure of the Ti 0.7Ru 0.3O 2 nanoparticles was verified by X-ray diffraction. By indexing these XRD patterns using the PDF standard card, we confirmed that Ti 0.7Ru 0.3O 2 is present principally as a single-phase solid solution with the TiO 2 being in the anatase (PDF#01- 086-1157) form (Fig. 6-3). No signal corresponding to phase separation between ruthenium and titanium oxide was detected. The average particle size of Ti 0.7Ru 0.3O 2 catalyst calculated by the Scherrer Equation in the same way as in Section 3.2 was 4.33 nm. Figure 6-3 X-Ray diffraction patterns of the Ti 0.7Ru 0.3O 2 Nanoparticles. Energy-Dispersive X-ray Spectroscopy - The EDX analysis of the Ti 0.7Ru 0.3O 2 (Fig.6-4 Inset) showed a uniform distribution of titanium and ruthenium. These results suggested that titanium 137 and ruthenium are not phase separated. The EDX results of Ti 0.7Ru 0.3O 2 (Fig.6-4) showed the Ti : Ru atomic ratio to be 76.21 : 23.79, which closely agreed with the expected atomic ratio of 7:3. Figure 6-4 EDX elemental analysis of Ti 0.7Ru 0.3O 2 nanoparticles. The inset shows the element map for (a) Ti (b) Ru Cyclic Voltammetry - The cyclic voltammograms of Ti 0.7Ru 0.3O 2 catalyst obtained in the 0.1 mol/L perchloric acid solution with 10 mV/s scanning rate Fig. 6-5(a) showed stable currents caused by the double layer capacitance. No significant oxidation or reduction current was observed in the range of -0.6-0.1V vs MSE. On the other hand, commercially-available platinum-ruthenium shown in Fig. 6-5(b) had increasing oxidation current when anodically scanned and increasing reduction current when scanning cathodically. The current corresponded to the redox reactions of Ru 4+ /Ru 3+ and Ru(0). This revealed the instability of RuO 2 in the potential range of interest. Therefore, it was promising to use a more stable structure Ti 0.7Ru 0.3O 2 as the catalyst support, to enhance the stability of the bifunctional catalyst. 138 Figure 6-5 Cyclic voltammograms obtained in 0.1 mol/L perchloric acid solution with 10 mV/s scanning rate (a) 2.0 mg Ti 0.7Ru 0.3O 2 catalyst (b) 8.3 mg commercially- available ruthenium oxide 6.3.2 Pt-Ti 0.7Ru 0.3O 2 Catalysts X-Ray Diffraction - The similar XRD patterns of Ti 0.7Ru 0.3O 2 and Pt/Ti 0.7Ru 0.3O 2 (Fig.6-6) suggested that the polyol process by which Pt nanoparticles were anchored to the Ti 0.7Ru 0.3O 2 the support, does not change the crystal structure or cause any phase separation of the Ti 0.7Ru 0.3O 2 solid solution. No platinum signal appeared on the powder XRD pattern of Pt/Ti 0.7Ru 0.3O 2. It suggested the platinum nanoparticles synthesized by polyol method are very fine particles and most likely well adhered to the Ti 0.7Ru 0.3O 2 support preventing any sintering of the platinum particles. These findings were further confirmed by the EDX mapping Pt element and Ru element (Fig.6-7 Inset). The EDX evaluation of Pt/Ti 0.7Ru 0.3O 2 catalyst (Fig.6-7) showed the Pt: Ru atomic ratio to be 1: 2.15. a b 139 Figure 6-6 X-Ray diffraction patterns of the catalysts support and the catalyst with various ratios of platinum to ruthenium. The XRD patterns of pure platinum is for reference. Figure 6-7 EDX elemental analysis of Pt-Ti 0.7Ru 0.3O 2 catalysts. The inset shows the element map for Ru (a) Pt (b) Methanol Oxidation Activity - In Fig. 6-8 we show the methanol oxidation catalytic activity of catalysts with various Pt:Ru atomic ratio and also for the PtRu/C catalyst. The onset potential for 140 methanol oxidation on the bi-functional catalyst is roughly -0.35~-0.25vs MSE while that for a platinum catalyst is approximately -0.15V vs MSE. It is widely accepted that PtRu/C is a bi- functional catalyst. 58,59 The fact that methanol oxidation currents of PtRu/C and Pt- Ti 0.7Ru 0.3O 2 catalysts both occurred at -0.3V vs MSE suggested that Pt-Ti 0.7Ru 0.3O 2 does not behave as platinum catalysts, and exhibiting a bi-functional mechanism as in PtRu/C. However, a physical mixture prepared by combining platinum black and Ti 0.7Ru 0.3O 2 had an onset potential similar to platinum. This is because there is limited inteaction between platinum black and Ti 0.7Ru 0.3O 2 and thus the mixture does not behave similar to Pt-Ti 0.7Ru 0.3O 2 prepared by the polyol method. Figure 6-8 Steady state polarization data of Pt-Ti 0.7Ru 0.3O 2 catalysts with various Pt:Ru atomic ratio and PtRu/C catalyst obtained in 1 mol/L methanol in 1M perchloric acid. The currents were measured by holding electrodes at each potential for 300 seconds 141 Pt-Ti 0.7Ru 0.3O 2 catalyst with a Pt:Ru atomic ratio of 1:2 exhibited the highest activity among all the Pt-Ti 0.7Ru 0.3O 2 catalysts. This catalyst had a smaller proportion of Pt compared to the commercial PtRu/C catalyst which has a Pt:Ru ratio of 1:1. This difference could be caused by the arrangement of the platinum and ruthenium atoms in Pt-Ti 0.7Ru 0.3O 2 and PtRu/C catalyst. In a Pt- Ti 0.7Ru 0.3O 2 catalyst, if the amount of Pt is too much, Ru sites will be covered by Pt atoms, which will hinder water activation. Hence the methanol oxidation rate will decrease. On the other hand, insufficient amount of Pt particles will not take full advantage of the Ru sites, which is also not an effective catalytic structure. The methanol oxidation activity can also be further optimized by changing the Pt:Ru atomic ratio. Tafel Slope - The Tafel slopes of Pt-Ti 0.7Ru 0.3O 2 catalyst, PtRu/C catalyst and the mixture of Pt black and Ti 0.7Ru 0.3O 2, obtained from steady states current in 1M methanol (Fig. 6-7) are presented in Table 6-1. The Tafel slopes suggested that the well-known bifunctional mechanism for methanol oxidation was operative on the new catalyst as well. With further optimization of this type of catalyst we expected to lower the Pt content and enable practical use of these materials in fuel cells. Table 6-1 Tafel slope of Pt-Ti 0.7Ru 0.3O 2 catalyst, PtRu/C catalyst and the mixture of Pt black and Ti 0.7Ru 0.3O 2 nanoparticles Formulation Tafel slope / mV dec -1 PtRu/C, Pt:Ru=1:2 108.9 Pt-Ti 0.7Ru 0.3O 2, Pt:Ru=1:2 112.6 Pt-Ti 0.7Ru 0.3O 2, Pt:Ru=1:1 164.5 Pt-Ti 0.7Ru 0.3O 2, Pt:Ru=2:1 208.7 142 Mixture of Pt black and Ti 0.7Ru 0.3O 2, Pt:Ru=1:2 158.1 Catalyst Stability. The stability of the Pt-Ti 0.7Ru 0.3O 2 catalyst was investigated by holding potential at 0V vs MSE. Fig.6-8 showed that once the methanol oxidation current reached the steady state, the current stayed constant for 20000 seconds, after which the test was terminated. This test yielded a preliminary indication of the stable nature of the Pt-Ti 0.7Ru 0.3O 2 catalyst. Figure 6-8 Pt/Ti 0.7Ru 0.3O 2 polarization curve at 0 V vs MSE, obtained in 1 mol/L methanol 6.4 Summary In this chapter, we have examined the use of titanium oxide-ruthenium oxide solid solutions as co-catalyst supports. We prepared Pt-Ti 0.7Ru 0.3O 2 for use as an anode catalyst. The structure of the synthesized catalyst was investigated by XRD, SEM and EDX analysis. Results of steady-state voltammetry suggested that the electrocatalytic activity and stability of the newly synthesized catalysts compared well with the commercially--available carbon-supported Pt- 143 Ru catalyst. The Tafel slopes suggested that the well-known bifunctional mechanism for methanol oxidation was operative on the new oxide supported platinum catalyst. Reference iiiiiiivv (1) The developments in the world’s total energy consumption split on sources as from 1800 and into 2013 https://damnthematrix.wordpress.com/2014/11/page/2. (2) Projected global energy consumption from 1990 to 2040, by energy source (in million metric tons of oil equivalent) https://www.statista.com/statistics/222066/projected- global-energy-consumption-by-source/. (3) Chen, G. Z. Supercapacitor and Supercapattery as Emerging Electrochemical Energy Stores. International Materials Reviews. 2017, pp 173–202.. (4) McGrath, K. M.; Prakash, G. K. S.; Olah, G. A. Direct Methanol Fuel Cells. J. Ind. Eng. Chem. 2004, 10 (7), 1063–1080. (5) Cameron, B. D. S.; Hards, G. A.; Harrison, B.; Potter, R. J. Direct Methanol Fuel Cells. Platin. Met. Rev. 1987, 31 (3), 173–181. (6) Aricò, A.; Baglio, V.; Antonucci, V. Direct Methanol Fuel Cells: History, Status and Perspectives. Electrocatal. Direct Methanol Fuel Cells From Fundam. to Appl. 2009. (7) Behret, H. Fuel Cells and Their Applications. Berichte der Bunsengesellschaft für Phys. Chemie 2018, 100 (11), 1922. (8) Pavela, T. O. Ann. Acad. Sci. Fennicae AII. Ann. Acad. Sci. Fenn. AII 1954, 59, 7–11. 144 (9) Justi, E.W. and Winsel, A. W. Brit. Patent 821, 688, 1955. (10) Surampudi, S.; Narayanan, S. R. R.; Vamos, E.; Frank, H.; Halpert, G.; LaConti, A.; Kosek, J.; Prakash, G. K. S.; Olah, G. A. A. Advances in Direct Oxidation Methanol Fuel Cells. J. Power Sources 1994, 47 (3), 377–385. (11) Bagkar, N. C.; Chen, H. M.; Parab, H.; Liu, R.-S. Electrocatalysis of Direct Methanol Fuel Cells; 2009. (12) Olah, G. A. The Methanol Economy. Chem. Eng. News Arch. 2003, 81 (38), 5. (13) Whitacre, J. F.; Valdez, T.; Narayanan, S. R. Investigation of Direct Methanol Fuel Cell Electrocatalysts Using a Robust Combinatorial Technique. J. Electrochem. Soc. 2005, 152 (9), A1780. (14) Beden, B.; Lamy, C.; de Tacconi, N. R.; Arvia, A. J. The Electrooxidation of CO: A Test Reaction in Electrocatalysis. Electrochimica Acta. 1990, pp 691–704. (15) Oudar, J. Sulphur Poisoning of Metals. “Model Experiments on Single Crystals.” Stud. Surf. Sci. Catal. 1982, 11 (C), 255–268. (16) Beden, B.; Hahn, F.; Lamy, C.; Léger, J. M.; de Tacconi, N. R.; Lezna, R. O.; Arvia, A. J. Chemisorption of Methanol on Different Platinum Electrodes (Smooth and Rough Polycrystalline, Monocrystalline, and Preferentially Oriented), as Studied by EMIRS. J. Electroanal. Chem. 1989, 261 (2 PART 2), 401–408. (17) Wasmus, S.; Küver, A. Methanol Oxidation and Direct Methanol Fuel Cells: A Selective Review. J. Electroanal. Chem. 1999, 461 (1–2), 14–31. 145 (18) Aricò, A. S.; Srinivasan, S.; Antonucci, V. DMFCs: From Fundamental Aspects to Technology Development. Fuel Cells 2001, 1 (2), 133–161. (19) Hogarth, M. P.; Hards, G. A. Direct Methanol Fuel Cells. Platin. Met. Rev. 1996, 40 (4), 150– 159. (20) Moura, A.; Fajín, J.; Mandado, M.; Cordeiro, M. Ruthenium–Platinum Catalysts and Direct Methanol Fuel Cells (DMFC): A Review of Theoretical and Experimental Breakthroughs. Catalysts 2017, 7 (2), 47. (21) Piela, P.; Eickes, C.; Brosha, E.; Garzon, F.; Zelenay, P. Ruthenium Crossover in Direct Methanol Fuel Cell with Pt-Ru Black Anode. J. Electrochem. Soc. 2004, 151 (12), A2053. (22) Valdez, T. I.; Firdosy, S.; Koel, B.; Narayanan, S. R. Investigation of Ruthenium Dissolution in Advanced Membrane Electrode Assemblies for Direct Methanol Based Fuel Cell Stacks. ECS Trans. 2006, 293–303. (23) Ueda, A.; Yamada, Y.; Ioroi, T.; Fujiwara, N.; Yasuda, K.; Miyazaki, Y.; Kobayashi, T. Electrochemical Oxidation of CO in Sulfuric Acid Solution over Pt and PtRu Catalysts Modified with TaOxand NbOx. In Catalysis Today; 2003; Vol. 84, pp 223–229. (24) Gregoire, J. M.; Tague, M. E.; Cahen, S.; Khan, S.; Abruña, H. D.; Disalvo, F. J.; Van Dover, R. B. Improved Fuel Cell Oxidation Catalysis in Pt1-XTaX. Chem. Mater. 2010, 22 (3), 1080– 1087. (25) Ramprasad, R. First Principles Study of Oxygen Vacancy Defects in Tantalum Pentoxide. J. Appl. Phys. 2003. 146 (26) McCafferty, E.; Wightman, J. P.; Mcca, E.; Wightman, J. P. Determination of the Concentration of Surface Hydroxyl Groups on Metal Oxide Films by a Quantitative XPS Method. Surf. Interface Anal. 1998. (27) McCafferty, E. Introduction to Corrosion Science; 2010. (28) Asselin, E.; Ahmed, T. M.; Alfantazi, A. Corrosion of Niobium in Sulphuric and Hydrochloric Acid Solutions at 75 and 95 °C. Corros. Sci. 2007, 49 (2), 694–710. (29) Park, K.-W.; Sung, Y.-E. Pt Nanostructured Electrode Encapsulated by a Tantalum Oxide for Thin-Film Fuel Cell. J. Vac. Sci. Technol. B Microelectron. Nanom. Struct. 2004, 22 (6), 2628. (30) Masud, J.; Alam, M. T.; Awaludin, Z.; El-Deab, M. S.; Okajima, T.; Ohsaka, T. Electrocatalytic Oxidation of Methanol at Tantalum Oxide-Modified Pt Electrodes. J. Power Sources 2012, 220, 399–404. (31) Tague, M. E.; Gregoire, J. M.; Legard, A.; Smith, E.; Dale, D.; Hennig, R.; DiSalvo, F. J.; Bruce van Dover, R.; Abruna, H. D. High Throughput Thin Film Pt-M Alloys for Fuel Electrooxidation: Low Concentrations of M (M = Sn, Ta, W, Mo, Ru, Fe, In, Pd, Hf, Zn, Zr, Nb, Sc, Ni, Ti, V, Cr, Rh). J. Electrochem. Soc. 2012, 159 (12), F880–F887. (32) Zhang, N.; Zhang, S.; Gao, Y.; Yin, G. Niobium Dioxide Facilitating Methanol Electrooxidation on Pt/C Catalyst by Synergistic Effect. Fuel Cells 2013, 13 (5), 895–902. (33) Sasaki, K.; Adzic, R. R. Monolayer-Level Ru- and NbO[Sub 2]-Supported Platinum Electrocatalysts for Methanol Oxidation. J. Electrochem. Soc. 2008, 155 (2), B180. (34) Rossnagel, S. Sputtering and Sputter Deposition. In Handbook of Thin Film Deposition 147 Processes and Techniques; 2001; pp 319–348. (35) Depla, D.; Mahieu, S.; Greene, J. Sputter Deposition Processes. Handb. Depos. Technol. Film. coatings 1991, 281, 253–296. (36) Wasa, K. 2 - Sputtering Phenomena. In Handbook of Sputtering Technology (Second Edition); Wasa, K., Kanno, I., Kotera, H., Eds.; William Andrew Publishing: Oxford, 2012; pp 41–75. (37) Westwood, W. D. Sputter Deposition Processes. MRS Bull. 1988, 13 (12), 46–51. (38) Scherer, M. Magnetron Sputter-Deposition on Atom Layer Scale. Vak. Forsch. und Prax. 2009, 21 (4), 24–30. (39) Rossnagel, S. M. Thin Film Deposition with Physical Vapor Deposition and Related Technologies. J. Vac. Sci. Technol. A 2003, 21, S74–S87. (40) Patterson, A. L. The Scherrer Formula for X-Ray Particle Size Determination. Phys. Rev. 1939, 56 (10), 978–982. (41) Blyholder, G. Molecular Orbital View of Chemisorbed Carbon Monoxide. J. Phys. Chem. 1964, 68 (10), 2772–2777. (42) Trasatti, S.; Petrii, O. A. Real Surface-Area Measurements In Electrochemistry. Pure Appl. Chem. 1991, 63 (5), 711–734. (43) Dongsheng, G.; Gongxuan, L. Dependence of Onset Potential for Methanol Electrocatalytic Oxidation on Steric Location of Active Center in Multicomponent Electrocatalysts. J. Phys. Chem. C 2007. 148 (44) Kakati, N.; Maiti, J.; Lee, S. H.; Jee, S. H.; Viswanathan, B.; Yoon, Y. S. Anode Catalysts for Direct Methanol Fuel Cells in Acidic Media: Do We Have Any Alternative for Pt or Pt-Ru? Chemical Reviews. 2014. (45) Vidaković, T.; Christov, M.; Sundmacher, K. Rate Expression for Electrochemical Oxidation of Methanol on a Direct Methanol Fuel Cell Anode. J. Electroanal. Chem. 2005, 580 (1), 105–121. (46) Li, L.; Xing, Y. Methanol Electro-Oxidation on Pt-Ru Alloy Nanoparticles Supported on Carbon Nanotubes. Energies 2009, 2 (3), 789–804. (47) Jusys, Z.; Kaiser, J.; Behm, R. J. Composition and Activity of High Surface Area PtRu Catalysts towards Adsorbed CO and Methanol Electrooxidation - A DEMS Study. Electrochim. Acta 2002, 47 (22–23), 3693–3706. (48) Chen, X., Bai, X., Deng, P., Peng, D. Liu, X. Potential-PH Diagram of Zr-H2O System at the Increased Temperatures. Rare Met. Mater. Eng. 2004, 33 (7), 710–713. (49) Witham, C. K.; Chun, W.; Valdez, T. I.; Narayanan, S. R. Performance of Direct Methanol Fuel Cells with Sputter-Deposited Anode Catalyst Layers. 2000, 3 (11), 497–500. (50) Behler, K.; Osswald, S.; Ye, H.; Dimovski, S.; Gogotsi, Y. Effect of Thermal Treatment on the Structure of Multi-Walled Carbon Nanotubes. J. Nanoparticle Res. 2006, 8 (5), 615–625. (51) Kumsa, D. W.; Bhadra, N.; Hudak, E. M.; Kelley, S. C.; Untereker, D. F.; Mortimer, J. T. Electron Transfer Processes Occurring on Platinum Neural Stimulating Electrodes: A Tutorial on the i(V e) Profile. J. Neural Eng. 2016, 13 (5). 149 (52) GRAMSE, E. L.; DIAMOND, L. H. Literature of the Chlor-Alkali Industry. In Literature of Chemical Technology; Advances in Chemistry; AMERICAN CHEMICAL SOCIETY, 1968; Vol. 78, p 1. (53) Janssen, L. J. J.; Starmans, L. M. C.; Visser, J. G.; Barendrecht, E. Mechanism of the Chlorine Evolution on a Ruthenium Oxide/Titanium Oxide Electrode and on a Ruthenium Electrode. Electrochim. Acta 1977, 22 (10), 1093–1100. (54) Cao, H.-Z.; Lu, D.-H.; Zheng, G.-Q. Novel Ruthenium-Titanium Oxide Electrode Prepared by the TiO2 Template. Dongbei Daxue Xuebao/Journal Northeast. Univ. 2012, 33 (SUPPL.2). (55) Hummelgård, C.; Gustavsson, J.; Cornell, A.; Olin, H.; Bäckström, J. Spin Coated Titanium- Ruthenium Oxide Thin Films. Thin Solid Films 2013, 536, 74–80. (56) Thanh Ho, V. T.; Pillai, K. C.; Chou, H. L.; Pan, C. J.; Rick, J.; Su, W. N.; Hwang, B. J.; Lee, J. F.; Sheu, H. S.; Chuang, W. T. Robust Non-Carbon Ti0.7Ru0.3O2support with Co-Catalytic Functionality for Pt: Enhances Catalytic Activity and Durability for Fuel Cells. Energy Environ. Sci. 2011, 4 (10), 4194–4200. (57) Yoshimura, M.; Byrappa, K. Hydrothermal Processing of Materials: Past, Present and Future. J. Mater. Sci. 2008, 43, 2085–2103. (58) Gonzalez, E. R.; Mota-Lima, A. Catalysts for Methanol Oxidation; 2013; Vol. 9789400777. (59) Iwasita, T. Electrocatalysis of Methanol Oxidation. Electrochim. Acta 2002, 47 (22–23), 3663–3674. 150
Abstract (if available)
Abstract
This work presented in this thesis focuses on designing electrocatalysts for methanol oxidation and understanding the catalytic processes. The goal is to achieve more durable and efficient electrocatalysts for direct methanol fuel cells. Direct methanol oxidation catalysts Pt₁₋ₓ-Mₓ (M = Ta, Nb, Ti, Zr, 0 < x < 1) was prepared using co-sputter deposition. Characterization of these thin film catalysts was performed using scanning electron microscopy (SEM), energy dispersive X-ray (EDX) and X-ray photoelectron spectroscopy (XPS). Assessment of the methanol oxidation activity of Pt₁₋ₓ-Mₓ catalysts were achieved through half-cell experiments. Among all the Pt₁₋ₓ-Mₓ catalysts, Pt₀.₇₇-Ta₀.₂₃ catalyst showed best electrochemical area specific activity which is comparable to platinum ruthenium alloy on carbon (PtRu/C) catalysts. Pt₁₋ₓ-Taₓ, Pt₁₋ₓ-Nbₓ and Pt₁₋ₓ-Zrₓ catalysts work as bifunctional methanol oxidation catalysts. The surface oxides species activates water molecules and hence facilitate the process of removing carbon monoxide from platinum sites. The membrane electrode assembly (MEA) of Pt₀.₇₇-Ta₀.₂₃ catalyst was tested at 60℃, 80℃ and 90℃. The power density achieved at 90℃ was 82 mW/cm²/mg Pt, which is 1.82 times of PtRu/C catalyst with similar platinum loading. This is the first a methanol oxidation catalyst containing tantalum being tested in an MEA.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The modification of catalysts and their supports for use in various fuel cells
PDF
Design and modification of electrocatalysts for use in fuel cells and CO₂ reduction
PDF
Integrated capture and conversion of carbon dioxide from air into methanol and other C1 products
PDF
Studies on direct oxidation formic acid fuel cells: advantages, limitations and potential
PDF
Understanding the mechanism of oxygen reduction and oxygen evolution on transition metal oxide electrocatalysts and applications in iron-air rechargeable battery
PDF
Electrocatalysts for direct liquid-feed fuel cells and advanced electrolytes for lithium-ion batteries
PDF
Integrated carbon dioxide capture and utilization: catalysis enabled carbon-neutral methanol synthesis and hydrogen generation
PDF
Reforming of green-house gases: a step towards the sustainable methanol economy; and, One-step deoxygenative fluorination and trifluoromethylthiolation of carboxylic acids
PDF
Studies on direct methanol, formic acid and related fuel cells in conjunction with the electrochemical reduction of carbon dioxide
PDF
Small organic molecules in all-organic redox flow batteries for grid-scale energy storage
PDF
Novel methods for functional group interconversions in organic synthesis and structural characterization of new transition metal heterogeneous catalysts for potential carbon neutral hydrogen storage
PDF
PSSA-PVDF semi-IPN blends for direct methanol fuel cells
PDF
Electrochemical pathways for sustainable energy storage and energy conversion
PDF
Methanol synthesis in a membrane reactor
PDF
Development of polystyrene sulfonic acid-polyvinylidene fluoride (PSSA-PVDF) blends for direct methanol fuel cells (DMFCS)
PDF
Design, synthesis, and study of polypyridine based molecular and heterogenized molecular electrocatalysts for CO₂ reduction
PDF
Sustainable continuous flow syntheses of colloidal inorganic nanoparticle catalysts
PDF
2-nitrophenyl-α-trifluoromethyl carbinols as smart synthons for novel fluoroorganics
PDF
Exploring new frontiers in catalysis: correlating crystal chemistry and activity in layered silicates
PDF
Design of nanomaterials for electrochemical applications in fuel cells and beyond
Asset Metadata
Creator
Fang, Dan
(author)
Core Title
Design of durable and efficient catalysts for the electro-oxidation of methanol
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Chemistry
Publication Date
03/05/2019
Defense Date
03/05/2019
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
catalysts,DMFC,MEA,methanol electro-oxidation,niobium,OAI-PMH Harvest,platinum,tantalum
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Narayan, Sri (
committee chair
), Prakash, Surya (
committee member
), Shing, Katherine (
committee member
)
Creator Email
dancao1115@gmail.com,danfang1115@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c89-130090
Unique identifier
UC11675708
Identifier
etd-FangDan-7132.pdf (filename),usctheses-c89-130090 (legacy record id)
Legacy Identifier
etd-FangDan-7132.pdf
Dmrecord
130090
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Fang, Dan
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
catalysts
DMFC
MEA
methanol electro-oxidation
niobium
platinum
tantalum