L'un des principaux freins au développement des piles à combustible de type PEMFC (Proton Exchange Membrane Fuel Cell) est lié aux phénomènes de dégradation des performances qui les pénalisent encore en termes de durée de vie. L'étude de ces phénomènes au niveau des composants de l’AME est un thème abordé aujourd'hui par de nombreuses équipes de recherche, mais une étude à une échelle d’un stack est nécessaire pour mieux comprendre les mécanismes en jeu. En effet, dans un stack les conditions de fonctionnement ne sont pas homogènes comme dans les cellules de laboratoire, notamment au niveau thermique. Ceci est particulièrement exacerbé dans les piles pour application automobile, dont la compacité contraint fortement la conception du circuit de refroidissement. De plus, les exigences en termes de démarrage à froid sont à prendre en compte, avec notamment la limitation de l'inertie thermique de l'empilement ou l'apparition d'hétérogénéités plus fortes pendant les phases transitoires.Ce travail de thèse se propose d'étudier l'effet d'hétérogénéités de température sur la performance d'une pile en application automobile et sa dégradation. L'étude est menée dans différentes conditions de fonctionnement: fonctionnement nominal, cyclage thermique et cyclage NEDC (New European Driving Cycles).Cette étude comporte une partie expérimentale, centrée sur des essais de vieillissement en pile et un travail sur le diagnostic électrochimique global et local. Elle est complétée par des expertises post-mortem des assemblages membrane-électrodes et des plaques testées. En parallèle, un travail de modélisation est mené pour relier les constatations expérimentales à une description des phénomènes en présence. L'influence du design des canaux de réactifs et de caloporteur sur le fonctionnement des piles est étudiée. Enfin, l’effet de la gestion thermique sur la dégradation des performances et sur la détérioration des composants de la pile est étudié. / One of the main challenges for Proton Exchange Membrane Fuel Cells development is the performance loss, which largely limits the durability. The study of the degradation phenomena of the different MEA components is a challenge addressed by many researchers, but a study at a stack scale is needed in order to better understand the ageing mechanisms. Indeed, in an industrial fuel cell the operating conditions are not homogeneous as for laboratory fuel cells, especially as regards thermal aspects. The heterogeneities are particularly emphasized for automotive fuel cells, because of the compactness constraint of the cooling circuit. Moreover, the requirements of cold start should be considered, as well as the inertial effects of the stacks and the increased heterogeneities during the driving cycles.In this work, the effects of the temperature heterogeneities and hot spots on the automotive fuel cell performances and degradations are investigated. The study is conducted in different conditions: nominal conditions, load/thermal cycling and New European Driving Cycles (NEDC).The work is composed of an experimental study, which consists of ageing tests on fuel cells and on-line diagnosis at both global and local scales. At the end of the tests, post-mortem analyses of the aged components are conducted. In parallel, a physic-based model is developed in order to predict the local temperature and humidity in the different components of the cell. Then, the impact of the reactive gases and cooling flow fields design on the thermal and water management of the cell is investigated. Finally, the experimental and modeling results are coupled in order to investigate the correlation between heat management, water management and degradations.

Hydrogen production by PEM-Water electrolysis is an environmentally benign and promising approach to store excess energy from renewable energy sources but facing drawbacks of high costs, mainly due to a harsh cell-environment. The aim of my Master Thesis was to reduce the costs of the most expensive cell component, the bipolar plate by surface engineering. Thin films of Ti, Zr and alloys thereof, as well as Nb and W have been vapor deposited by either cathodic arc deposition or magnetron sputtering in an industrial scale coating system. The nitrides, carbides, and pure metals from the previously mentioned transition metals were investigated by crosscut adhesion tests, interfacial contact resistance, electrochemical corrosion tests, scanning electron microscopy and energy dispersive X-ray spectroscopy. Highly promising thin film materials have been developed to functionalize the bipolar plates, enabling significant cost reductions of the PEMWE-cell.

D. Tech. (Department of Mechanical Engineering, Faculty of Engineering and Technology), Vaal University of Technology. / On account of their lightweight, low-cost, corrosion resistance, and good formability, conductive polymer composites (CPCs) are promising for the production bipolar plate (BP) for polymer electrolyte membrane fuel cell (PEMFC). However, a high conductive filler loading is needed to impart the required level of electrical conductivity to the insulating polymer matrix and as a consequence, the toughness of the plate deteriorates considerably. By using immiscible blend of polymers that have complementary hardness and ductility as matrix, with conducting multi-fillers of different morphologies, it is possible to optimize the matrix strength characteristics and favour the formation of conducting network to produce CPC meeting BP performance standards. Of course, a lot will depend on the formulation of the most favourable composition and production variables. In this regard, polypropylene-epoxy and polyethylene-epoxy blends, filled with zero- and two-dimensional carbon forms – graphite, carbon black (CB), and graphene (Gr) – were investigated over an extensive range of compositions and compression moulding pressures, in this study. Several compounding runs (using melt mixing), at different stages, followed by compression molding, were done. The goal is to obtain combination of composite formulation and processing conditions that will produce the most promising combination of properties for BP application. In the first stage of the investigations, by using thermogravimetric analysis, two-stage decomposition behavior of PP-epoxy and PE-epoxy blends was revealed, which confirms the immiscibility of PP and PE with epoxy resin. Scanning electron microscope (SEM) micrographs of the PP-epoxy and PE-epoxy blends revealed a co-continuous structure, which can be attributed to the close-to-symmetric composition of the blend and compatibilizers added. Preferential localization of synthetic graphite (SG), CB, and Gr in the polymer blends was also revealed by the SEM micrographs. This confirms the fact that CPCs based on PP-epoxy and PE-epoxy blends can be explored further. PP-epoxy and PE-epoxy blends filled with only SG, 30 – 80 wt %, were produced and characterized for their electrical conductivity and flexural properties. In-plane electrical conductivity ranged from 12.09 to 68.03 Scm-1 for PP-epoxy/SG and 11.68 to 72.96 Scm-1 for PE-epoxy/SG composites produced. These are higher than values reported for several single matrix polymer composites at similar filler loadings. With reference to the United States Department of Energy performance targets for BPs, PE-epoxy/SG composites performed better in terms of electrical conductivity, while PP-epoxy/SG composites exhibited better flexural properties. Thereafter, using SG and CB double filler, PE-epoxy/SG/CB composites performed better than PP-epoxy/SG/CB composites in terms of electrical conductivity, while PP-epoxy/SG/CB composites exhibited superior flexural properties than the PE-epoxy/SG/CB composites at similar filler loadings. However, with respect to the DOE targets, composites based on PP-epoxy blend exhibited a more promising combination of electrical conductivity and flexural properties than PE-epoxy blend matrix. PP-epoxy filled with SG/CB was studied further, by using graphene (Gr) as second minor filler. In-plane and through plane electrical conductivities as well as thermal conductivity and thermal diffusivity of the PP-epoxy/SG/CB/Gr composites increased as total filler content was increased from 65 to 85 wt%. It implies that more conductive networks between filler particles were formed. Also, flexural strength, flexural modulus, and impact strength decreased as the total filler content increased from 65 to 85 wt%. The reduced flexural properties could be due to increased agglomeration of CB and Gr, and poor filler wetting at higher filler loadings and low matrix material, which leads to the formation of microvoids and a reduction of the load bearing capacity of composites. With respect to the DOE targets, PP-EP/SG/CB/Gr composite with 80 wt% (i.e., PP/EP/73G/6.2CB/0.8Gr) filler has the best combination of properties. Further improvement in properties of the PP-EP/SG/CB/Gr composite with 80 wt% filler was achieved by molding at higher pressures. As molding pressure was increased from 4.35 to 13.05 MPa, in-plane electrical conductivity increased from 116.31 to 144.99 Scm-1, while flexural strength increased from 29.62 to 42.57 MPa, satisfying the performance requirement targets for bipolar plates.