Engineering self-heated micro-reformer for micro-solid oxide fuel cell applications

The relevance of portable electronic devices has increased constantly in the lasts years leading to a growing interest in more efficient portable power sources. Among the proposed solutions, micro-solid oxide fuel cell power generators (µ-SOFC PG) are one of the most interesting alternatives to traditional Li batteries, due to their superior energy and power density, long lifetime and high reliability. µ-SOFC shares the same working principles of traditional SOFC, but the entire cell is based on thin films, drastically reducing the ohmic losses and allowing to operate at temperatures below 600 °C. Moreover, the system is integrated in Si technology, offering the possibility of using mainstream fabrication processes of Micro Electro-Mechanical Systems (MEMS) and, therefore, assuring low cost and a facile scale up to large scale production. One of the most important components of the µ-SOFC system is the fuel processing unit (µ-reformer), which allows working with liquid hydrocarbons fuels (Dimethyl Ether in this work), drastically increasing the energy density of the entire device. Nevertheless, some important issues related with fast start-up, stand-alone reforming of the fuel and low power consumption still need to be faced for allowing a reliable and effective integration into the µ-SOFC system. This thesis is devoted to improve this silicon based fuel processing unit, both by numerical modelling and by the fabrication and testing of a self-heated µ-reformer. The Si-based µ-reformer is integrated in a 1x1 cm2 Si substrate of 500 µm of thickness. The active part of the µ-reformer is composed by a micro-fabricated suspended platform, in which an array of more than 6000 microchannels and a metallic heater are located, which is intended to provide the thermal power necessary for the fuel processing. The microchannels (50 µm of diameter) are covered with two different catalysts, γ-phase alumina and platinum, which are expected to reform the hydrocarbons into hydrogen. After having followed the main fabrication steps of the Si-based µ-reformer, the deposition of these catalysts was successfully carried out by Atomic Layer Deposition (ALD). The thermal behaviour of the self-heated µ-reformers was then characterized by Raman in-situ measurements of the sample’s temperature as a function of the electrical power supplied. In this way, by following the shift of the main Si Raman mode, an estimation of the device temperature as function of the ohmic resistance of the heater was founded, also confirmed by a thermal FEM model. The devices show a fast start-up and a maximum temperature of 600 °C in vacuum, which is limited by the thermal losses from the suspended platform towards the Si substrate. After the thermal characterization of the substrate, the catalytic properties of these self-heated devices were assessed by feeding dimethyl ether and air to the µ-reformers and measuring the reformed gas flow by a micro-gas chromatography (µ-GC). The results showed that hydrogen was successfully produced by the self-heated µ-reformer but the percentage of conversion appeared to be low, probably due to a scarce Pt coverage of the device. Finally, a Finite Element Method (FEM) model of one microchannel has been developed in order to reproduce and improve the obtained results. The simulations showed a good match with previously obtained experimental results as a function of the µ-reformer’s temperature. An optimization analysis was carried out by varying so