2 Search Results

The effect of two solvents, liquid water and 1,4-dioxane, has been studied from first principles on the hydrodeoxygenation of propionic acid over a Ni (111) catalyst surface model. A mean-field microkinetic model was developed to investigate these effects at a temperature of 473 K. Under all reaction conditions, a decarbonylation mechanism is favored significantly over a decarboxylation pathway. Although no significant solvent effects were observed on the decarbonylation rate, a substantial solvent stabilization of two key surface intermediates in the decarboxylation mechanism, CH₃CCOO and CH₃CHCOO, lead to a notable increase of the decarboxylation rate by two orders of magnitude inmore » liquid water and by one order of magnitude in liquid 1,4-dioxane. Furthermore, a significant solvent stabilization of the transition state of C-H bond cleavage of the α-carbon of CH₃CHCO, relative to the stabilization of the C-C bond cleavage of the α-carbon of CH₃CHCO, leads to a change in dominant pathway in the liquid phase environments. Finally, a sensitivity analysis shows that the C-OH bond cleavage of propionic acid and C-C bond cleavage of the α-carbon of CH₃CHCO are the most rate controlling states in the gas phase. In contrast, in solvents the dehydrogenation of CH₃CHCO becomes the most influential step. This shift in rate controlling state is attributed to the solvent effect on the dehydrogenation of CH₃CHCO, which is facilitated in aqueous phase. Altogether, it is likely that the investigated (111) facet of Ni is not active for the hydrodeoxygenation of propionic acid in neither the gas nor liquid phase and other Ni facets or phases must be responsible for the experimentally observed kinetics.« less

Computational catalyst discovery involves identification of a meaningful model and suitable descriptors that determine the catalyst properties. We study the impact of combining various descriptors (e.g., reaction energies, metal descriptors, and bond counts) for modeling transition-state energies (TS) based on a database of adsorption and TS energies across transition-metal surfaces for the decarboxylation and decarbonylation of propionic acid, a chemistry characteristic for biomass conversion. Results of different machine learning models for more than 1572 descriptor combinations suggest that there is no statistically significant difference between linear and nonlinear models when using the right combination of reactant energies, metal descriptors, andmore » bond counts. However, linear models are inferior when not including bond count and metal descriptors. In conclusion, when there are missing data for reaction steps on all metals, conventional linear scaling is inferior to linear and nonlinear models with proper choice of descriptors that are surprisingly robust.« less