Author: Ageo Meier de Andrade
This thesis is built around two pillars. One is heterogeneous catalysis in the broader context of green chemistry. The focus here is on identifying catalytically active materials suitable for the valorization of renewable feedstocks. The second pillar deals with materials modelling itself, both its role to help identify the features responsible for certain desired material properties and the assessment of model quality and how to overcome challenges when modelling complex systems.
Density functional theory (DFT) has become a standard method in heterogeneous catalysis and materials science, as it generally combines good accuracy with an affordable computational cost. The choice of density functional (in relation to the system under study) strongly affect the accuracy of the DFT results. In this thesis, a subset of functionals have been tested and validated with respect to their ability to predict structural and energetic properties of single- and multi-component materials. It is shown that the inclusion of dispersion corrections by computing the nonlocal correlation self-consistently, as done in the vdW-DF-cx functional, increases the accuracy of computed results in relation to experimental data.
In the evaluation of the functionals, the computed properties were rationalized in terms of (i) the reduced density gradient distribution, unique for each material, and (ii) the exchange enhancement factor, unique for each density functional and dependent on the reduced density gradient distribution. Moreover, a tool is presented that can guide researchers towards the most appropriate density functional for the problem in question. This involves a protocol that brings DFT results into better agreement with experiment.
Heterogeneous catalysts are complex and catalyst research is often performed using model experiments and calculations. Here descriptors have important roles to play. Two descriptors for catalytic activity have been scrutinized in this thesis. The first is the work function of metal surfaces. Here, it is shown that the adsorption of selected ad-atoms on Ni surfaces provides a route to control the metal’s work function over a wide energy range. The second descriptor is the difference in stability between the enol and keto tautomers of a model lignin molecule on a model metal catalyst surface in the context of lignin depolymerization. The aim is to explore the reasons underlying the relative stabilities and to enhance the preference for enol in the keto-to-enol tautomerization. The modelling results show that a mixed PdPt alloy surface stabilizes the enol tautomer, suggesting that this could be an active catalyst for lignin depolymerization.
Doctorate thesis, Acta Universitatis Upsaliensis, 2021. , p. 74