This paper presents concentration measurements of water vapour and methane, taken in-situ on an operational solid oxide fuel cell (SOFC) test rig using tunable diode laser spectroscopy (TDLS). Methane concentration measurements are presented for the TDLS system and are compared with concentration measurements taken using gas chromatography (GC). Furthermore, purge times for the SOFC gas-analysis system have been calculated using TDLS, which are measurements that cannot be obtained directly using GC. Finally, water vapour concentration measurements in the SOFC cathode are shown for different system operating conditions: a dry cathode cycle and during the introduction of water vapour. As GC cannot be used to measure water vapour in the SOFC cathode stream, a direct comparison cannot be made with the TDLS measurements.

A new in-line, real time gas analyser is described that uses tuneable diode laser spectroscopy (TDLS) for the measurement of methane in solid oxide fuel cells. The sensor has been tested on an operating solid oxide fuel cell (SOFC) in order to prove the fast response and accuracy of the technology as compared to a gas chromatograph. The advantages of using a TDLS system for process control in a large-scale, distributed power SOFC unit are described. In future work, the addition of new laser sources and wavelength modulation will allow the simultaneous measurement of methane, water vapour, carbon-dioxide and carbon-monoxide concentrations. (C) 2011 Elsevier B.V. All rights reserved.

The validity of two new approaches to tunable diode laser spectroscopy (TDLS) in the near-IR, namely the residual amplitude modulation approach and the phasor decomposition method, is investigated for application in industrial process monitoring where the operating temperatures and pressures are high and subject to significant change. Both techniques allow the recovery of absolute absorption profile line shapes and are completely calibration free, making them very attractive for online deployment in stand alone instrumentation in harsh environments where the calibration factors in conventional TDLS methods are subject to significant cumulative errors and drift. Currently established TDLS techniques, and indeed conventional gas composition analysis techniques, suffer from significant limitations when applied under these conditions, and there is a clear need for the development of a suitable alternative. The primary focus in this work is the analysis of water vapor in solid oxide fuel cell diagnostics where the operating temperatures range from 700degC to 950degC, the gas pressures are subject to change and the recovered signal levels are low. The 1391.7 nm overtone water vapor transition is interrogated over the above temperature range of interest at concentrations of 6%-50%, while the 1650.96 nm methane transition is also analyzed over a range of gas pressures at a fixed concentration of 1%. Excellent agreement between the experimentally recovered absorption line shapes and simulations based on parameters from the HITRAN (2004) database is observed; further evidence for the efficacy of the techniques is demonstrated through the accuracy of the gas concentration measurements which were achieved by curve-fitting absorption line shape simulations to the experimental data.

A novel reformer design has been demonstrated that converts the methane required for a multi kilowatt SOFC stack. Results show the influence of temperature and the benefits of operating at elevated pressure on the reforming-catalyst fundamental reaction kinetics. Due to the high heat demand of the steam reforming reaction, efficient heat transfer between the SOFC stack and the reforming catalyst is essential. Parameters such as the volume/surface area ratio, choice of catalyst, and catalyst metal loading are key to the design, and these have been determined through a combination of computer modelling and experimental measurements. The thermal properties of the unit have been evaluated over a range of temperatures and fuel compositions that simulate system operating-conditions in the final product.

A test system based around a thin-walled extruded solid electrolyte tube has been developed which enables the fuel reforming catalysis and surface chemistry occurring within solid oxide fuel cells to be studied under real operating conditions. It permits simultaneous monitoring of the catalytic chemistry and fuel cell performance, allowing a direct correlation between the cell output and the anode reforming characteristics. Using this system internal methane reforming over nickel/zirconia cermet anodes has been studied. The influence of the anode formulation, anode pretreatment, operating temperature and methane/steam ratio on the methane reforming characteristics, the nature and level of carbon deposition, and durability have been investigated under actual operating conditions. Pre-reducing the anodes in H2 at 1173 K leads to a more active reforming catalyst. Carbon deposited during reforming is removed from the anodes in two processes during temperature programmed oxidation. There is an increase in the temperature of both carbon removal processes and an increase in the population of the higher temperature state with increasing reforming temperature.

Nickel-based/yttria-stabilised zirconia anodes for solid oxide fuel cells (SOFCs) running on natural gas have been developed which show increased resistance towards carbon deposition and improved durability. Surface carbon formed on the anodes during reforming has been characterised using temperature programmed oxidation (TPO), The influence of anode composition and formulation, pre-treatment method, operating temperature and methane/steam ratio have been studied. Doping the nickel/zirconia anode with small quantities of molybdenum leads to a substantial reduction in the amount of carbon deposited. As current is drawn from the SOFC increased methane conversion occurs together with reduced carbon deposition.

A novel test system based on a thin-walled extruded tubular yttria-stabilised zirconia reactor has been developed, which can be used to investigate the fuel processing catalysis and surface chemistry occurring at nickel/zirconia anodes in solid oxide fuel cells, as well as the durability and electrochemical performance of the fuel cell, under actual operating conditions. In addition to enabling different anode formulations to be evaluated for their catalytic activity, electrical performance and durability, and permitting the surface chemistry to be studied, the system allows simultaneous monitoring of the catalytic activity and surface chemistry, and the cell performance, allowing direct correlation between the fuel cell performance and its reforming characteristics.

Internal methane reforming over nickel/zirconia cermet anodes has been studied in detail using a thin-walled extruded zirconia tubular SOFC reactor. The influence of anode formulation, anode pretreatment, operating temperature and methane/steam ratio on the reforming characteristics, resistance to carbon deposition and durability have been investigated under actual operating conditions. Post-reaction temperature programmed oxidation (TPO) has been used to determine the amount of carbon deposition and its strength of interaction with the anode. A 90 vol% nickel/zirconia anode shows higher activity than a 50 vol% Ni anode at higher reforming temperatures, and shows very good durability. Pre-reducing the anodes in H-2 at 1173 K leads to a more active reforming catalyst. Carbon is removed from the anodes in two processes during temperature programmed oxidation, suggesting two types of carbon species. As the reforming temperature increases there is an increase in the temperature of both carbon removal processes, and an increase in the population of the higher temperature state.

The catalytic oxidation of methane over supported nickel catalysts has been studied using conventional catalytic reactor measurements, temperature programmed reaction spectroscopy and gas pulsing experiments. The influence of support material, catalyst pre-treatment and operating temperature have been studied. The nature of the support material has a large influence on the subsequent activity and CO selectivity of the nickel catalysts. Temperature programmed measurements have been used to study methane activation, the surface reaction pathways and to evaluate the nature and level of any carbon species deposited during reaction. Temperature programmed oxidation reveals several types of carbon are formed on the catalyst during catalytic methane oxidation. Gas pulsing experiments have been carried out at different reaction temperatures to determine how the initial methane conversion, product selectivity and surface carbon yield vary as a function of contact time, and show that metallic nickel provides the active site for methane partial oxidation.

A novel test system based around an extruded solid electrolyte tube enables the study of both the catalytic chemistry of fuel processing anodes in working solid oxide fuel cells and the electrochemical performance of the cell under the same conditions, allowing a direct correlation to be made between the fuel cell performance and the reforming characteristics of the anode.

Nickel/zirconia fuel reforming anodes in solid oxide fuel cells (SOFCs) running on methane have been studied using a test cell based on a tubular SOFC. The anodes have been characterised using temperature-programmed reduction (TPR) which reveals that two distinct types of nickel oxide are present in the anode prior to reduction. The steam reforming activity and surface chemistry of two different nickel/zirconia anode formulations and a molybdenum doped nickel/zirconia anode have been studied. All three show good activity towards methane steam reforming. It is found that the quantity of carbon deposited on the anode during methane reforming is strongly affected by the operating temperature and the methane/steam ratio. The addition of small quantities of molybdenum leads to a significant reduction in the amount of carbon deposited, whilst having little effect on the reforming activity or cell performance. Temperature-programmed oxidation (TPO) has revealed that three types of carbon are formed on the anodes during high temperature reaction of methane. As current is drawn from the cell, increased methane conversion occurs together with reduced carbon deposition, through reaction via partial oxidation and oxidative coupling with the flux of oxygen ions through the solid electrolyte.

The activation energy for the removal of surface carbon formed by methane decomposition following high-temperature reforming, from a nickel/zirconia solid oxide fuel cell (SOFC) anode has been calculated using two methods based on temperature-programmed oxidation. It is found that there is a fairly good agreement between the two methods. In addition, it was observed that the addition of small quantities of lithium to the anode resulted in a significant lowering of the activation energy for surface carbon removal by about 50 kJ mol-1.

The dependence of the rate of hydrogenolysis of C2H6, C3H8, andn-C4H10on 0.3% Pt/Al2O3(AKZO CK303, EUROPT-3) has been determined at several temperatures, using short reaction pulses to minimise deactivation by carbon deposition. Rate maxima, which with C2H6occur at very low H2pressures, become broader and move to higher H2pressures as the chain-length or temperature increases. Criteria for selecting an appropriate rate expression with which to model the results is discussed in detail: one is selected that is based on the competitive chemisorption of the reactants, that of the alkane requiring the loss of more than one H atom to activate it. For each data set, optimum values of the constants of the rate equation are obtained by computation. Changes in the form of the kinetic curves are described by (i) values of the rate constantk1and the equilibrium constant for alkane chemisorptionKAthat increase with alkane chain length and with temperature, (ii) values ofKH, the adsorption coefficient for H2, that decrease with alkane chain-length but are not very temperature-dependent, and (iii) a decreasing degree of the needful alkane dehydrogenation as the chain-length increases. For C3H8andn-C4H10, true activation energies are respectively 82 and 76 kJ mol−1, and enthalpy changes for alkane chemisorption are 88 and 79 kJ mol−1. Apparent activation energies increase with H2pressure in consequence of the Temkin equation and show compensation effects. Rate dependences on H2pressure were measured at a single temperature using 0.3% Re-0.3% Pt/Al2O3(AKZO CK433, EUROPT-4): rates were faster than for Pt/Al2O3, and values ofk1,KA, andKHwere larger. Measurements were also made with catalysts partially deactivated by carbon deposition; slower rates were associated with lower values ofk1andKA. Changes in product selectivities with operating variables are recorded, but are only significant for PtRe/Al2O3, and forn-C4H10on Pt/Al2O3, where at 547 K the extent of internal C–C bond fission decreases as H2pressure increases. The difficulty of devising a simple rate expression to embrace all the experimental observations is discussed.

A catalytic test rig linked to a quadrupole mass spectrometer has been designed and built for the in situ study of tubular SOFC anodes. The apparatus allows electrochemical and catalytic measurements to be simultaneously made. Temperature Programmed Reduction (TPR) reveals two distinct types of NiO present in the anode prior to reduction: bulk NiO, and NiO closely bound to ZrO2. Doping the anode with small quantities of a transition-metal additive is found to affect the anode's reduction characteristics. The reduction of the anode has been studied under both open circuit and load conditions.

Surface carbon formed on Ni-cermet anodes in working SOFCs operating on methane has been characterised using Temperature Programmed Oxidation (TPO), and electrochemical measurements. It is found that the quantity of carbon deposited on the anode is significantly reduced by: lowering the operating temperature, increasing the steam/methane ratio, and introducing small quantities of additives such as molybdenum into the anode. TPO has revealed that three types of surface carbon species are formed during high temperature reaction of methane. Each of these species has been found to form at different rates on the Ni anode.

The hydrogenolysis of ethane, propane andn-butane has been studied on Pt/Al2O3(EUROPT-3) and PtRe/Al2O3(EUROPT-4) catalysts between 520 and 660 K by a thermal cycling technique using a 10-fold excess of H2. Analogous catalysts having higher metal concentrations have also been examined. With neither ethane nor propane does any progressive deactivation result from repeated thermal cycles, but withn-butane on Pt/Al2O3catalysts initial loss of activity alters product selectivities, specifically by decreasing the probability of breaking the internal C–C bond. After this, they behave reproducibly. With PtRe/Al2O3catalysts, product selectivities shown byn-butane are very different from those given by the Pt/Al2O3catalysts but are unaltered by the somewhat greater initial deactivation; they show some sample-to-sample variation, which is attributed to differences in surface Re concentrations. These results are explained by a model in which sites on low-index planes in very small Pt particles are those that are first inactivated through the formation of strongly held dehydrogenated species; when Re atoms are present they preferentially occupy these sites and the operation of bimetallic centres explains why deeper hydrogenolysis occurs in this case. Reproducible rates ofn-butane hydrogenolysis are obtained with Pt/Al2O3below about 600 K, but not with PtRe/Al2O3; use of short (1 min) reaction pulses overcomes this difficulty. Partially deactivated catalysts can, however, be studied in the continuous-flow mode.

The dependences of rate upon H-2 and alkane pressures have been determined for the hydrogenolysis of ethane, propane and n-butane on clean and coked Pt/Al2O3 and Pt-Re/Al2O3 catalysts. On clean surfaces, preserved by using only 1 min reaction periods, rates usually exhibited maxima as H-2 pressure was increased, the pressure at the maximum increasing with the size of the alkane; orders in the alkane measured at high H-2 pressure approximated to unity. Coking by reacting n-butane + H-2 at 713 K caused significant differences in the ways in which rates varied with H-2 pressure. Three rate equations have been tested, of which two give satisfactory fits. Constants for one of these, viz. rate = k1P(A)P(H)/(k2P(A) + P(H)k3) are presented, from which it appears that the reactive species may be formed from the reactant alkane by loss of four or five H atoms.