Abstract

A new generation of polymer electrolyte membrane fuel cells (PEMFCs) has started to enter the public market ahead with Toyota’s model Mirai in 2015. Further on, several automakers also announced small scale production of fuel cell electric vehicles (FCEVs). Simultaneously the hydrogen infrastructure is aimed to be developed further in order to support FCEVs’ market entry. In the last decade, PEMFC research has focused on characterization method development which should support the understanding of the catalyst layer (CL). Especially the CL’s microstructure is not yet completely investigated. Due to the materials complexity, an accurate description of catalyst dispersions appears to be impossible. Hence, the present study focuses on experimental investigation of the catalytic ink particle size distribution and the corresponding catalyst layer morphology. An electrochemical evaluation was subsequently performed in order to categorize the CLs into sufficient or insufficient cell behavior. Therefore, a consecutive approach of parameter variation was carried out. First, the manufacturing process of cathodic electrodes was investigated by selective analysis of catalyst layers resulting from diverse processed catalyst dispersions (mixing time variation). Second, the ionomer to carbon ratio (I/C) was varied in order to evaluate the pore network development and finally the particle size distribution was directly tuned with the goal to create differentiated catalyst layers. As a result, the mixing time and the particle size distribution tuning revealed detectable variations within the CL microstructure which were directly measurable with electrochemical testing. Hence, monomodal catalytic ink particle size distributions with a maximal size of 1 µm presented a 50 % reduced film thickness and slit shaped pores within the CL leading to a direct cell breakdown during electrochemical evaluation. In contrast, polymodal catalyst dispersions with a maximal size of 10 µm yielded well performing catalyst layers related to a sufficient pore and ionomer network which manage gas and proton transport, respectively. These cell behaviors are predictable by using the identified catalyst dispersion and catalyst layer characteristics which include the state of the catalytic dispersion Df, the modality of the particle size distribution, the maximal particle size of the distribution xmax, the hysteresis loop of the resulting CL, and the specific surface area of the CL. A good cell performance was achieved when the catalyst dispersion presented a polymodal distribution below particle sizes of 10 µm, the absence of Mie scattering, and H2 type hysteresis loop for the resulting catalyst layer. On the other hand, poor cell performances are created with monomodal particle size distributions below 1 µm maximal size resulting in Mie scatter. Hence, a high particle packing density is observed within the CL so that slit shaped pores are created which induce a H4 type hysteresis loop and an early cell breakdown.