(419h) Using Operando Characterization Techniques to Uncover the Structure and Composition of Active Sites during Oxygen Reduction Electrocatalysis

Heterogeneous platinum-based materials are the current state-of-the-art catalysts for proton-exchange membrane hydrogen fuel cells (PEMFC) that convert chemical energy (e.g. H₂) into electricity. Designing catalysts that reduce or replace Pt is one route towards reducing the cost of these systems. However, rational design of non-precious metal catalysts is limited by a poor understanding of the catalysts’ active site configuration during the oxygen reduction reaction (ORR). For example, in promising transition metal nitride (TMN) and metal-nitrogen-carbon (M-N-C) catalyst systems, the role of the metal, nitrogen, carbon, and oxygen is convoluted. Specifically, many TMN catalysts contain carbon or oxygen components that are unintentional, but imperative in activity. Similarly, the specific geometry and composition of each active site in M-N-C catalysts is so varied that identifying the most active species is difficult. Our work has focused on developing a catalyst template and specific operando cell to characterize the active site structure and composition for different TMN and M-N-C catalysts.

Many TMN catalysts unintentionally contain reactive carbon components due to pre-catalysis pyrolysis with a conductive carbon support. To overcome these complications, we optimized a reactive sputtering synthesis of carbon-free, thin film molybdenum nitride (Mo_xN) films, with a tunable bulk structure and composition, that are active and stable in 0.1 M perchloric acid. Using a specifically designed electrochemical cell for grazing incidence x-ray absorption and reflectivity, we have characterized changes in our Mo_xN surface under ORR conditions. These same surface changes are not apparent at similar non-ORR reducing conditions. By changing the x-ray incident angle, we were able to show that these changes were only apparent on the top several nanometers of a ~ 30 nm film. These results are the first step in developing a picture of the active site composition and geometry during catalysis.

M-N-C catalysts are typically activated via a pyrolysis step that forms a variety of M-N_x active sites. The unknown density and variety of active sites makes it almost impossible to define the optimal active-site geometry. Using a non-pyrolyzed Co-porphyrin catalyst (cobalt-tetrakis(4-carboxyphenyl)porphyrin, Co TCPP), we designed a fuel-cell-like setup that allowed us to use ambient pressure x-ray absorption spectroscopy to show direct adsorption of the oxygen onto the Co-metal center at ORR potentials. Furthermore, with ambient pressure x-ray photoelectron spectroscopy we saw no clear changes in the nitrogen species during the ORR. These results demonstrate the importance of the Co-metal center and give us a template for studying more complex pyrolyzed catalysts.

Together these studies highlight the development of operando characterization for the active surface during the ORR. With these characterization templates in hand we can develop a new appreciation for the active-site specific activity trends that have been previously obscured in either thin-film or nanoparticle configuration. This enhanced fundamental understanding will enable us to develop a new system for designing highly active electrocatalysts.