A Multi-Scale Modeling Approach to Study and Design PEM Fuel Cells Catalyst Layers with Non-Precious Metals

Abstract

Building a catalyst layer includes many elements that make understanding and studying it a complex problem. These systems have different elements with multiscale, multiphase and multicomponent issues, as Figure 1 illustrates. Typical PEMFC catalyst layers are composed of three elements: porous carbon to provide electron and gas/liquid transport, ionomer to provide protonic transport and catalyst to facilitate the reduction or oxidation reactions. The replacement of Pt with non-precious-metal-based catalysts (NPMCs) for the ORR is considered to be one key to cost-effective power generation in PEMFCs. Accordingly, developing NPMCs with high performance for ORR (1-4) has been a focal point of research. A major complication in analysis of the NPMC systems under study lies in the unknown nature of the active site. It is difficult to identify structural and electronic parameters based on most characterization techniques. Many questions arise when compared with Pt-based systems such as: Which elements do Pt and NPMC catalyst layers have in common? What makes one metal more suitable than another in the NPMC? What specific catalyst layer composition and structure are optimal? How do we (or can we) correlate energetics, structure and transport issues? Modeling catalyst layers is a complicated task because we need to take into account the phenomena occurring over different time and size scales as represent in Figure 1: (i) heterogeneous electrochemical reactions in the microscale; (ii) proton, electron and gases transport in the mesoscale; and (iii) gases and water transport through membranes and porous media in the macroscale (5-10). The first steps are to identify and classify the main macroscopic parameters that affect the performance by using complex and adequate experimental data sets and to reduce the degrees of freedom for fitting. Once the parameters are identified, we can work at the right scale, with the appropriate physics and computational methodology that the macroscopic results point out in order to complete the picture and resolve the unknowns. The aim of this study is to extend previous results from a 1D model in which we identified the active area and permeability as the most distinguishing elements between Pt-based and NPM-based catalyst layers. On one hand, the active area enters into macroscopic models convolutes with the intrinsic activity of the catalyst site. The latter is related to the specific catalytic properties of the active site; understanding it requires a deep study of the catalyst nature, which is not well known due to the pyrolysis of NPM materials. In this case, a coupled modeling approach using DFT (Density Functional Theory) and MD (Molecular Dynamics) techniques could give us insight in the active site behavior. On the other hand, transport properties and differences can be study by using CFD (Computational Fluid Dynamics) tools to compare fluid transport and local distributions within the catalyst layers with parameters obtained via DFT-MD simulations. This contribution will summarize the previous modeling of transport and describe calculations to probe the final configuration after high temperature treatment of porphyrin and iron porphyrin carbon structures. We aspire to describe interactions of the porphyrin with the carbon substrate during the heating process and give insight into the active site final configuration from a theoretical perspective by combining MD with DFT calculations. Acknowledgement We gratefully acknowledge the support of the NSF EPSCoR program and Colciencias for support of this work.