Increasing carbon emissions and insecurities in oil supply have led to heightened interest in hydrogen powered fuel cells. Preferably, the cell runs on hydrogen gas, though due to the sensitivity of the catalytic components in the fuel cell to carbon monoxide, the hydrogen must be extremely pure (typically <50 ppm CO). Due to a lack of hydrogen infrastructure, it is envisaged that a medium term solution will be the reforming of more conventional fuels such as gasoline. The gas mixture produced however, contains impurities such as CO, CO₂ and CH₄. Purification may be achieved using palladium membranes, which allow selective permeation of hydrogen. This thesis describes the research carried out in conjunction with Johnson Matthey on thin (typically 7.5 μm) palladium/silver alloy membranes supported on both ceramic and stainless steel porous tubular substrates. Extensive experimental flow testing has been performed to assess the effect of temperature, feed composition, including wet feeds, and membrane thickness on the hydrogen purification properties. An existing Fortran based model was validated and revised to accurately account for the effects of operating conditions such as temperature and carbon monoxide concentration. This work provided excellent correlation between experimental and simulated results. The validated and improved model was incorporated in the design of a hydrogen refuelling station in Aspen Plus and the palladium membrane requirements assessed to supply 650 fuel cell vehicles per day. The system incorporated a steam reformer, membrane clean-up module, water trap and high pressure compressor for hydrogen storage at 1000 bara. Operating conditions such as system pressure, fuel feed and steam to carbon ratio were investigated and adjusted to optimise the overall system efficiency. An efficiency of 52% was achieved with a steam to carbon ratio of SCR = 2.5. A membrane requirement of 6000 standard tubes was found to provide a 90% hydrogen recovery efficiency.

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the Degree of Doctor of Philosophy. Johannesburg, 2015 / Bimetallic platinum nanoparticles were synthesized for application as anode catalysts for low temperature fuel cells such as direct methanol fuel cells (DMFCs). Two distinct synthesis procedures were used; namely conventional synthesis with post-synthesis heat treatment, and secondly polyol microwave-irradiation without further heat-treatment. The aim was to synthesize interesting and novel bimetallic nanostructures and relate their shape and morphologies to their methanol oxidation reaction (MOR) activities and their CO tolerance. Due to the high cost of the conventional synthesis processes as well as their use of harmful solvents, microwave-irradiation was explored as a possible synthesis procedure. It is a greener and more environmentally friendly approach with possibilities of mass production of the nanoparticles. For both the synthesis procedures, the reducing agent, the precursor salts, surfactants, pH of the solution and molar ratios were varied to determine the effect on the shape, size and ultimately the electrocatalytic activities of the Pt-Co and Pt-Ni nanoparticles. For the conventional synthesis procedure, the main parameter of comparison was the strength of the reducing agents, where NaBH4 and N2H4 were used under the same reaction conditions. In this study, the strength of the reducing agent affected the properties of the Pt-Co and Pt-Ni nanoparticles, such that, the stronger the reducing agent, the higher the degree of alloying and the more electrocatalytically active the materials. The drawback in the conventional synthesis was however low current outputs, in the microamps range, which necessitates a need to explore other synthesis procedures. Microwave-irradiation was thus used as an alternative synthesis procedure in an attempt to produce more active bimetallic platinum nanoparticles. Different reaction parameters were changed in this process to optimize the synthesis process, namely the pH of the solution, the amount of surfactant and the Pt-Ni molar ratio. In changing the reaction parameters, there was an observed change in the structure of the nanoparticles, with an average size in the order of 5 nm and different MOR activities. Furthermore, it was found that the activity was highest for the optimum amount of PVP and NaOH concentration of 500 mg and 1.0 M NaOH. In general, the MW synthesized nanoparticles achieved current values in the microamps to amps range, making it a more attractive synthesis procedure compared to the conventional method. The CO tolerance of the materials is an important aspect, as one of the main drawbacks of the commercial application of fuel cells is the propensity of Pt to get poisoned by CO during the methanol dissociation process. Therefore CO stripping measurements were performed on the MW-irradiated catalysts. The catalysts produced in this work showed good resistance towards CO. In general, the behaviours of the catalysts were dependent on the amount of surfactant and the molar ratio of the starting solution. The mechanism of CO tolerance in this case was determined as the bifunctional model, where the Ni-oxide and Ni-hydroxide species donate O to the electrooxidation of CO to CO2. In conclusion, the study of microwave-irradiated bimetallic nanoparticles performed here, resulted in highly active catalysts, which are even more active than commercial Pt/C nanoparticles.

The replacement of internal combustion engines used for transportation by polymer electrolyte membrane fuel cells (PEMFCs) is one goal of the future since they are clean, quiet, energy efficient and capable of quick start-up. At present, fuel cells are receiving much attention at both fundamental research, and technology development levels, but cost is the main factor that hinders the commercialisation of PEMFCs. In order to reduce cost, a better, fundamental description of fuel cell operation than is presently available is required. The operation of PEMFCs simultaneously involves electrochemical reactions, current distribution, fluid mechanics, multicomponent multiphase mixtures, and heat transfer processes. It is important to have a comprehensive mathematical model to provide improved understanding of the interactions between various electrochemical and transport phenomena in PEMFCs in order to aid in the design and optimisation of fuel cells. This thesis describes research at developing such a comprehensive model.

Proton Exchange Membrane Fuel Cells (PEMFCs) are widely regarded as the next generation of portable power production devices, with uses ranging from powering automotive vehicles to laptops and smartphones. PEMFCs convert oxygen and hydrogen into water and usable electricity and have no moving parts, meaning that they can reach overall efficiencies of 60%. However current Polymer Electrolyte Membranes only work efficiently below 80 C and at high humidity. At this low temperature, CO poisoning of the Pt electrocatalysts means that only high-grade fuel (low CO concentration ≤ 2 ppm) and high catalyst loading are required. This means that the overall cost of a PEMFC is prohibitively expensive. To dramatically the reduce cost and increase the efficiency of a PEMFC, new membranes are required which work at 120 C, at which point CO poisoning is no longer a dominating issue. In this thesis, the synthesis of novel organic/inorganic hybrid polyurethane Polymer Electrolyte Membranes (PEMs) with covalently bound phosphonic acid moieties (PA) made from cheap source materials have been reported, which, for the first time, demonstrate high conductivities at high temperatures, for example a PEM made from triethoxysilylpropyl isocyanate, polyethylene glycol, 4,4’-methylene diphenyl diisocyanate and PA displayed a conductivity of 3  10-2 S cm-1 at 120 C and 100% RH. The membranes also display good mechanical, thermal and chemical stability making them ideal candidate PEMs for the use in PEMFCs. However further work needs to be done to reduce the thickness of the membranes from their current thickness of 200 m to just 20-30 m, which would dramatically increase their efficiency when used in a PEMFC, by reducing the Area Specific Resistance and increasing the output (usable) power.

Currently, electrical generation from solid carbon (biomass, coal) is conducted at low efficiency (~35 %) compared to other power sources. The Direct Carbon Fuel Cell (DCFC) is a technology capable of electro-oxidising elemental carbon for the production of electricity at a projected 80 % efficiency. This improvement has significant benefits for the reduction of greenhouse gas emissions. The research status of the DCFC technology is in early stages, with no practical continuous or stacked designs having been established. The sole concept for a continuous anode has been based on particulate carbons, these designs suffer from poor carbon polarisation and a lack of fuel versatility. This work focusses on the development of a continuous, monolithic anode for a direct carbon fuel cell. A monolithic anode has the benefit of acting both as fuel and current collector. This concept achieves improved fuel polarisation and also avoids the pumping of hot molten carbonate mixtures, and the corrosion issues associated with a separate metallic anode. In this regard, a parallel was drawn with the aluminium production industry in the Söderberg electrode. This technology allows for the continuous pyrolysis and extrusion of carbonaceous mixtures into solid carbon anodes. This project simulated the process of Söderberg electrodes through isostatic compression of pine sawdust in a novel, bespoke heated press, designed and built in-house. This apparatus also allowed for the live monitoring of resistance during heating. The formation factors of pyrolysis temperature, applied load and particle size were studied. The anodes formed in these processes were subjected to various characterisation methods and a practicality assessment made. The electrochemical properties of each anode were also assessed in a novel, bespoke DCFC apparatus, again designed and built in-house. It was found that the anodes formed were of a suitable BET surface area (300 – 450 m2 g-1), possessed high microporosity and were of a tensile strength comparable to industrial Söderberg electrodes. Electrochemical tests found the anodes to produce OCV values near the theoretical value for carbon electro-oxidation (1.01 V). A maximal power of 7.87 mW cm-2, at 0.58 V was achieved using an anode formed at 620°C, 12.3 N applied load and with a mixed particle size.

<p>In this work a modified polyol method was developed to synthesize in-house catalysts. The method was modified for maximum delivery of product and proved to be quick and efficient as well as cost effective. The series of IH catalysts were characterized using techniques such as UV-vis and FT-IR spectroscopy, TEM, XRD, ICP and CV.</p>

Higher temperature (~ 200°C) operation for proton exchange membrane (PEM) fuel cells would have several advantages including enhanced electrochemical kinetics, useful heat recovery, and improved catalyst tolerance for contaminants. Conventional perfluorosulfonic acid membranes (PFSA), such as Nafion show a dramatic decrease in proton conductivity at temperatures above 80°C. For this reason, there has been an increasing effort toward the development of stable, higher temperature membranes with acceptable proton conductivity. This work is directed toward the development of Nafion free membranes for direct hydrocarbon PEM fuel cells containing zirconium phosphate as the proton conductor component. Hence, composite membranes composed of zirconium-phosphate (ZrP), a solid proton conductor, which was precipitated within the voids of a porous polytetraflouroethylene (PTFE) support were synthesized. Amorphous-like zirconium phosphate (ZrP) powder was synthesized in this work. ZrP was prepared by precipitation at room temperature via reaction of ZrOCl2 with H3PO4 aqueous solutions. The proton conduction properties of ZrP powder were studied under the processing conditions found in direct hydrocarbon fuel cell. Our experimental results showed that the ZrP powder processed at 200°C possess a proton conductivity that is greater by one order of magnitude than the oven-dried samples at 70°C. Thereby, it was possible to avoid the normal decrease in conductivity with increasing temperature by having sufficient water in the vapor phase. This thesis reports the first synthesis of composite ZrP/PTFE/Glycerol (GLY) membranes. Glycerol (GLY) was introduced into the pores of PTFE with the ZrP proton conductive material using the successive wetting/drying technique. These membranes had reasonable values of proton conductivities (0.045 S cm-1), approaching that of Nafion (0.1 S cm-1) at room temperature. Samples of these composite membranes were processed at the inlet conditions of a propane fuel cell, at 200°C. Experimental results showed that the proton conductivity remained almost unchanged. This thesis also describes and reports the first synthesis of sulphur “S” or silicon, Si–modified zirconium phosphate (ZrP), porous polytetrafluoethylene (PTFE) and, glycerol (GLY) composite membranes. It was aimed at the substitution of a minor amount of phosphorus “P” in the ZrP by (S or Si) in the ZrP to modify the proton conduction properties. The modification was performed by adding a certain amount of silicic acid or sulphuric acid into phosphoric acid then proceeding with the precipitation in situ. A high proton conductivity, of 0.073 S cm-1,i.e. 73% of that of Nafion, was observed for the Si–ZrP/PTFE/GLY composite membrane.

Microbial fuel cells (MFCs) are alternative energy resources that generate electricity from organic matter, where microorganisms such as the Geobacter species oxidize organic waste and transfer electrons to an electrode. Mathematical models are used to study biofilm processes, in hopes of developing MFCs into commercial applications. Existing biofilm models are based on Nernst-Monod type expressions, and are restricted to studying extracellular electrochemical/microbiological components, separated from the metabolic behavior of microorganisms. In this thesis, a model was developed combining extracellular biofilm conditions, with the intracellular metabolic fluxes of microorganisms under spatial heterogeneities (electron donor/acceptor levels) across the biofilm. This model predicts biofilm processes under varying extracellular conditions (presence/absence of NH4+, shear stress in continuous mode MFCs), and intracellular conditions (ATP maintenance fluxes); and also provides a preliminary evaluation of the pH changes across the biofilm. A sensitivity analysis based on the cell density and the biofilm conductivity was also conducted.

In this work, a MFC fabrication procedure including two non-conventional techniques (partial baking and cap-sealing) were employed for the development of an up-scaled microfluidic fuel cell (MFC). Novel carbon-based electrode materials were employed, including carbon foam, fibre, and cloth, the results from which were compared with traditionally-employed carbon paper. The utilization of carbon cloth led to 15% of the maximum power that resulted from carbon paper; however, carbon fibre led to a 24.6% higher power density than carbon paper (normalized by electrode volume). When normalized by projected electrode area, the utilization of carbon foams resulted in power densities up to 42.5% higher than that from carbon paper. The impact of catalyst loading on MFC performance was also investigated, with an increase from 10.9 to 48.3 mgPt cm-2 resulting in a 195% increase in power density.

Microbial fuel cells (MFCs) are alternative energy resources that generate electricity from organic matter, where microorganisms such as the Geobacter species oxidize organic waste and transfer electrons to an electrode. Mathematical models are used to study biofilm processes, in hopes of developing MFCs into commercial applications. Existing biofilm models are based on Nernst-Monod type expressions, and are restricted to studying extracellular electrochemical/microbiological components, separated from the metabolic behavior of microorganisms. In this thesis, a model was developed combining extracellular biofilm conditions, with the intracellular metabolic fluxes of microorganisms under spatial heterogeneities (electron donor/acceptor levels) across the biofilm. This model predicts biofilm processes under varying extracellular conditions (presence/absence of NH4+, shear stress in continuous mode MFCs), and intracellular conditions (ATP maintenance fluxes); and also provides a preliminary evaluation of the pH changes across the biofilm. A sensitivity analysis based on the cell density and the biofilm conductivity was also conducted.