Investigation of the effect of CO2 bubbles and slugs on the performance of a DMFC by means of laser-optical flow measurements

A single-channel DMFC is constructed that allows for flow measurements at the anode side as well as detailed time-resolved cell-voltage measurement. The coherence between flow and bubble clogging and slug movement can be investigated without parasitic effects like flow shortcuts through the gas diffusion layer (GDL) between neighbouring channels, as in serpentine or parallel-channel configurations. Optical access is granted to the anode side by a transparent foil, which is necessary for the application of the laser-optical velocity measurement technique (microparticle image velocimetry, mu PIV). Small fluorescent particles are added to the fluid, which are illuminated by a laser. The particle movement can be optically detected using a microscope, and transferred to a planar velocity field. Hence, the appearance and evolution of CO2 bubbles can be qualitatively and quantitatively investigated. The analysis of the velocity structure around a CO2 bubble or a moving slug allows a deeper understanding of the coherence of fluid motion, channel blockage, and cell performance. In addition to the flow analysis, a time-resolved measurement of the cell voltage is performed. The results clearly indicate that the cell power increases when huge bubbles reduce the free cross-section area of the channel. Methanol is forced into the GDL, i.e. methanol is continuously convected to the catalyst layer and is oxidised to CO2. Hence, the fuel consumption increases and the cell performance rises. When the huge bubble is released from the GDL and forms a moving slug, the moving slug effectively cleans the channel from CO2 bubbles on its way downstream. Since the channel cross-section is not severely diminished by the bubbles at this stage, the methanol flow is no longer forced into the GDL. The remaining amount of methanol in the GDL is oxidised. The cell power decreases until enough CO2 is produced to eventually form bubbles again that significantly reduce the free cross-section of the channel, and the process starts again. (C) 2012 Published by Elsevier Ltd.