Nickel and Molybdenum Nitrides As Oxygen Reduction Reaction Catalysts: Probing the Structure-Activity Relationship

Abstract

The development of active, stable, and low-cost catalysts for the oxygen reduction reaction (ORR) is necessary for the viability of polymer electrolyte membrane fuel cells (PEMFCs). Despite decades of study, the required mass loading of platinum-based catalysts remains prohibitively expensive and has limited the growth of fuel cell technology. Transition metal (TM) nitrides have recently gained attention as active ORR catalysts with promising stability in acidic electrolyte. Typically, TM nitrides are supported on high surface area carbon or nitrided-carbon, which can contribute significantly to overall activity and complicate understanding of intrinsic catalyst activity, including the active surface composition. In this study, we synthesize carbon-free molybdenum nitride (MoxN), nickel nitride (NixN), and bimetallic nitride thin films by reactive sputtering. These catalysts were found to be active in acid electrolyte, with onset potentials between 0.5 and 0.7 V vs RHE. Catalyst stability and selectivity to 4 electron ORR were tested in acid and base electrolyte. Activity, stability, and selectivity were found to depend on the catalyst structure. To better understand the effect of structure of on the catalyst surface and activity, we characterized the catalysts extensively. SEM showed that the films have a dense, pillar morphology. By varying the synthesis conditions, including temperature, N2 partial pressure, and substrate bias, we found that the crystallinity, crystal structure, and nitrogen-content of the films could be modified. Using grazing incidence X-ray diffraction (GIXRD), it was determined that the films were polycrystalline with different bulk structures and different nominal nitrogen contents (e.g. MoxN, 1≤x≤2). NEXAFS and cross-sectional TEM were also used to distinguish between similar crystal structures and to better understand the spatial distribution of crystal structures throughout the film based on the synthesis conditions. These differences in structure, for example between MoN and Mo2N, which are difficult to distinguish due to the possibility of many structures with similar XRD patterns, were correlated to changes in electrochemical activity, as shown in the accompanying figure. GIXRD shows that two of the catalysts appear to have the same structure in the surface and bulk, corresponding to nominal compositions of Mo2N and MoN. The catalyst which shows the MoN structure at the surface and Mo2N in the bulk is designated as Mo1.5N to approximately account for this gradient of N content. Cyclic voltammograms for MoxN in 0.1 M HClO4 show the highest ORR activity for this Mo1.5N. We hypothesize that the bulk structure of the catalyst, including the presence of a N gradient, affects the active surface and thus the ORR activity. Electrochemical failure mechanisms were probed through ex-situ XPS, SEM, and cross-sectional TEM characterization. Before testing, XPS showed that the surface of the films was a mixture of oxide and nitride, with the oxide attributed to air exposure. After stability tests, the films showed an increase in hydroxide and oxyhydroxide surface character, as well as loss of nitrogen. SEM and TEM revealed a corresponding change in morphology and loss of material. Together with the electrochemical stability measurements, these results indicate that catalyst activity loss is linked to loss of the active surface structure. Utilizing this carbon-free, well-defined catalyst morphology, we were thus able to gain insight into the structure and active surface of the material both before and after electrochemical testing. Figure Caption. (Left) GIXRD patterns for 3 MoxN catalysts at the surface and bulk. Vertical lines denote reference patterns. (Right) Cyclic voltammograms showing the ORR activity of the same three MoxN catalysts. Figure 1