MATERIALS SCIENCE AND ENGINEERING

QUANTITATIVE ASSESSMENT OF FUEL CELL CATALYST NANOPARTICLE SIZE DISTRIBUTIONS

A fuel cell is a device for generating electricity directly and very efficiently, from gaseous or liquid fuel, without combustion. Fuel cells are similar to batteries in that they also deliver electrical power from a chemical reaction. But unlike batteries, which store the chemical reactants within, fuel cells operate with reactants that are stored externally and must be supplied. There are several types of fuel cells. Presently, the two most promising technologies are proton-exchange membrane fuel cells (PEMFCs) and solid-oxide fuel cells (SOFCs).

In PEM fuel cells, the required chemical reactions, in which fuel molecules need to lose or accept electrons, are facilitated by the presence of an appropriate "catalyst" material. Usually, the catalyst consists of finely dispersed platinum nanoparticles. Since platinum is very expensive, a quest has begun for non-precious catalyst materials that can fulfill the same function more efficiently and at reduced costs. To assess the catalytic performance of novel catalyst nanoparticles, it is crucial to determine the active surface area of the particles reproducibly and with high reliably. The focus of present research, therefore, is to establish a quantitative method for determining the specific active catalyst surface area (area per unit mass). This requires to determine the size distribution of catalyst nanoparticles. A powerful method for this purpose is HRTEM (high-resolution transmission electron microscopy). Problems arise, however, from the fact that HRTEM images constitute two-dimensional projections of the true particle arrangements. Consequently, HRTEM images tend to show overlap of individual particle images in the projection. To deal with this problem, we have developed a procedure of digital image processing that recognizes individual particles by the orientation of their crystal lattice in HRTEM images (Fig. 1). This new method enables to determine particle size distributions from HRTEM images objectively and independent of an individual observer [1].

Fig. 1. HRTEM image of a platinum catalyst nanoparticle, overlaid with a color map indicating where the algorithm we have developed has detected a particle. The region of the particle is color-coded in green/blue, whereas the background (speckle pattern of originating from the amorphous carbon support) is color-coded in red.