(363ah) Multiscale and Particle-Size Modeling of Hydrogenolysis of Light Hydrocarbons on Pt Catalysts

Catalytic plastic waste upcycling has gained attention as a profitable approach to plastic waste management. Pt catalysts have been shown to be active for hydrogenolysis of Polyolefin and polypropylene into useful products including waxes, lubricants, and fuels. Understanding the chemistry of light alkane hydrogenolysis using density functional theory (DFT) and microkinetic model (MKM) can help facilitate the understanding of plastic deconstruction. Many computational studies, however, limit the investigation of hydrogenolysis chemistry to individual Pt facets, i.e., Pt(100), Pt(111) and Pt(211), hence failing to describe how the interaction between the facets affects the hydrogenolysis chemistry. This work uses DFT calculations, machine learning, and descriptor-based relations to develop a coupled multiscale CSTR reactor and particle model to investigate the kinetics and mechanism of ethane, propane, and butane hydrogenolysis on different sizes of Pt catalysts. We observed that the hydrogenolysis of a hydrocarbon on Pt first proceeds through an α - β dehydrogenation to form an olefin on Pt(111) and Pt(100) terrace sites. The olefin products then desorb and re-adsorb on Pt(211) edge sites to get further dehydrogenated and finally undergo C-C cleavage through the CH3C* species for the case of ethane and other deep dehydrogenated species for propane and butane. Particle size analysis shows a decreasing hydrogenolysis activity and constant selectivity with increase in particle size, agreeing with experimental studies. We observe that the TOF of hydrocarbon activation decreases and selectivity to methane increases with increase in hydrocarbon conversion. In the limit of ~0% conversion, olefin desorption is preferred over hydrogenolysis reactions. This work also studies factors influencing the product distribution in the case of propane and butane. The multiscale and particle model used in this work helps to reconcile computational and experimental observations and clearly details the insufficiency of single facet non-reactor based microkinetic modeling for investigating the hydrogenolysis of hydrocarbons.