(327b) Theoretical Investigation of the Hydrodeoxygenation of Levulinic Acid to g-Valerolactone over Ru Catalysts

Due to diminishing reserves of fossil fuels and a renewed emphasis on green technology for energy production, there is a desire for development of second generation bio-fuels that possess high energy density and offer low net carbon emissions. In the past decades massive efforts were given to conversion of lignocellulosic biomass into usable fuel and levulinic acid (LA) was found to be one of the promising biomass derived platform molecules that can be readily produced and converted into various fuel derivatives.  One conversion of significant interest has been the aqueous phase hydrodeoxygenation of LA to g-valerolactone (GVL) since GVL is a similarly versatile platform molecule to LA that is at the same time less water soluble than LA which facilitates its separation from the aqueous solution LA are produced in.  Conversion of LA to GVL requires a complex hydrodeoxygenation (HDO) process over transition metal catalysts whose mechanism is experimentally difficult to study due to concurrent catalyst deactivation and mass transfer limitations at practical operating conditions.

In this project, we studied the mechanism of the heterogeneous HDO of LA to GVL over Ru (0001) model surfaces from first principles.  First, we studied various elementary processes at the gas-solid interface with PBE-D3 level of theory.  Then, we applied our novel implicit solvation method for solid surfaces (iSMS) to determine elementary reaction and activation free energies in aqueous phase environments. Finally, we developed a microkinetic model to predict reaction rates, activation barrier and reaction orders.  A challenge during this investigation has been that we significantly underpredicted turnover frequencies in comparison to experimental observations.  We identified as the reason for our underprediction the large number of sites occupied by an individual adsorbed LA molecules while only a small number of free sites are available for reaction. A reasonable agreement with experimental observations could be obtained when considering lateral interactions and forcing LA to adsorb on fewer sites.