(376j) Multi-Scale Simulations with Linear Scaling Relations and Microkinetic Models of Aqueous Phase Reforming of Sugar Alcohols

With rising interest in the bio-refinery industry comes interest in converting biomass to useful chemicals and fuels. Specifically, there is interest in converting sugar alcohols, such as glycerol, methanol, ethanol and ethylene glycol, some of which currently exist in surplus. Currently, techniques exist to convert these sugar alcohol species to H₂, CO, and CO₂ gas, as well as liquid products such as acids and ketones. Among those techniques, aqueous phase reforming features low operation temperatures. Further, it is carried out in aqueous conditions, which are the same conditions as the sugar alcohols in surplus would be supplied. However, challenges in catalyst cost, selectivity and activity impede the broader expansion of this process. Design of new catalysts requires substantial knowledge of the reaction mechanism and operation conditions, which is challenging due to the large reaction networks, consisting of hundreds to thousands of possible reactions, and the aqueous environment itself, which complicates the mechanistic steps and also impedes experimental and computational attempts at observing the catalysis. In this work, we develop computational methods to overcome these challenges. To understand the influence of the water environment on the energetics and mechanism, we develop a multi-scale simulation strategy that employs force-field molecular dynamics (FFMD) and density functional theory (DFT). In this strategy, DFT is used to optimize surface structures and calculate catalytic energies, and FFMD is used to simulate the liquid water solvation. This strategy enables the capture of both energetic and mechanistic influences of water on the catalysis. To overcome the large number of potential reactions, we derive scaling relations, which are linear equations that correlate thermodynamic and kinetic properties of a catalyst. The scaling relations thus allow reaction energetics to be estimated based on a relatively smaller number of DFT calculations, which significantly improves the computational tractability. Finally, we use microkinetic modeling to examine reaction hierarchy in great detail. Our results are in excellent agreement with experimental results from the literature. This work serves as an important step for developing models for liquid phase heterogeneous catalysis, and ultimately, to design catalysts for such conditions.