(45e) Orienteering in an Uncharted Chemical Space: Searching for an Optimal Bimetallic Nanocatalyst

Nanoparticles (NPs) have received tremendous attention as catalysts due to their high specific surface area and abundance of unique active sites. When designing such a catalyst, the Sabatier principle tells us that it must be tuned to adsorb key intermediates neither too strongly nor too weakly. For this reason, it is imperative to screen candidate NPs for their adsorption behavior. This task is conventionally accomplished either through time-consuming experiments, or computationally-expensive ab-initio calculations, which are primarily limited to highly idealized periodic surfaces. This presents a bottleneck in materials discovery, one which can be overcome through the development of rapid methods to predict adsorption behavior. To this end, we present a universal model of small molecule adsorption to NPs [1], which is able to accurately capture the adsorption behavior of metal NPs of any size, shape, or composition with minimal computational cost. By combining this adsorption model with our new genetic algorithm approach to NP structure prediction, we are now able to efficiently traverse the vast chemical space of bimetallic NPs and identify possible catalytic NPs to feed experiments. Overall, this work leverages cutting-edge chemical modeling techniques, such as first principles calculations, machine learning and statistical regression to significantly accelerate materials discovery and enable the rapid identification of promising bimetallic nanocatalysts.

1. Dean, J.; Taylor, M.; Mpourmpakis, G. Unfolding Adsorption on Metal Nanoparticles: Connecting Stability with Catalysis. Science Advances 5 (9), eeax5101.