(597f) A Combined DFT and Classical Force-Field Approach for Modeling Kinetics of Acid-Catalyzed Reactions in Mixed Solvents.

The kinetics of acid-catalyzed dehydration reactions of biomass in aqueous solution are impacted by addition of aprotic solvents. Electronic structure and atomistic modeling can help connect solvent composition and structure to elementary step reaction energetics, helping to explain experimentally observed kinetics and facilitate rational optimization of solvent properties. In this presentation, we will present a computational approach (dubbed DFT-FF) that combines static DFT calculations of reaction energetics with a classical force-field description of solvent dynamics to model acid catalysis in the liquid phase. Our method development begins with determining the stable hydrated structure of the excess proton. Between the two well-known Eigen and Zundel motifs, our DFT-FF approach predicts the Zundel cation is 3.2 ± 1.2 kJ/mol lower in free energy. This consequently guides our choice for proton structure used in simulating acid-catalyzed reactions. We modeled the acid-catalyzed dehydration for two biomass model molecules, t-butanol and fructose, in three solvent mixtures: water/DMSO, water/γ-valerolactone and water/acetonitrile. The solution-phase activation barrier was determined by adding the DFT-calculated gas-phase activation barrier to the difference in solvation free energies of the initial and transition states as calculated in force-field molecular dynamics. The calculated relative barriers (mixture vs. pure water) exhibit great agreement in both values and trends with experimental rate measurements (Mellmer et. al. Nature 2018), validating our method. Our approach explicitly describes the dynamic solvent structure about each intermediate and transition state along the reaction path. We analyze these structures to provide a link between preferential clustering in mixed solvents and reaction kinetics. We decompose the activation free energy into its enthalpy and entropy contributions. The successful demonstration of our DFT-FF approach in modeling condensed-phase reaction kinetics provides further opportunities to model reaction kinetics in even more complex solvent systems, as well as in heterogeneous systems such as at electrochemical interfaces.