Beyond energetics scaling for transition states of heterogeneous catalytic reactions
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Rational catalyst design is essential for meeting commercial process demands while considering sustainability and environmental impact. Computational methods in heterogeneous catalysis, propelled by accelerating computing ability and improving theoretical techniques, can offer fundamental insights into novel catalyst design and implementation. Translating atomistic level insights to measurable, ‘real-world’ quantities requires multiscale frameworks across different length and time scales. At the quantum level, density functional theory (DFT) is a popular electronic structure method used in heterogeneous catalytic modeling for predicting atomistic characteristics. However, DFT calculations may become too computationally intensive when dealing with large-scale catalyst screening or complex reaction networks. Thus, surrogate models, such as linear transition state scaling (TSS) and Brønsted-Evans-Polanyi (BEP) relationships, are initially constructed with DFT data and then used to bypass calculation of all energetic or kinetic parameters. Yet linear correlations beyond energetics have remained largely unexplored, especially for predicting vibrational frequencies and thermochemical quantities at transition states of heterogenous catalytic reactions. This thesis extends DFT-computed transition state correlations beyond energy scaling, rationalizes the correlations, and investigates implementation of such relationships in microkinetic models (MKMs). ☐ In Chapters 2 and 3, the transition state vibrational scaling relationship (TSVSR) is introduced for AHX (A = C, N, O) diffusions and dehydrogenations across low-index transition metal surfaces. Analogous to how a TSS/BEP relation would be built with DFT energies, TSVSRs using adsorbate-metal driven vibrational modes are first constructed between local minima and transition states of AHX diffusions. Using d-band theory and linear muffin orbital theory, the slopes of the TSVSRs are derived and predicted from corresponding energetics scaling, adsorbate geometries relative to the surface, and a reference DFT calculation. Uncertainty in slope predictions is quantified based on geometric variation across metals, and individual frequency scaling is extended to scaling of vibrational entropies, enthalpies, and Gibbs energies for AHX diffusions. For AHX dehydrogenations, the slopes of the TSS relationships are first derived and predicted using the Slater-Koster structure factors, geometric attributes, and the relevant orbital overlap between adsorbates and metal surfaces. The TSVSRs for AHX dehydrogenations are then built, and their slopes are predicted with a similar theoretical approach used for the AHX diffusions. Due to irregularities across metals and bond-breaking characteristics, AHX dehydrogenation TSVSRs are found to be highly dependent on the structural variation of the transition state. Scaling of entropies and enthalpies across the AHX homologous series demonstrates potential for estimating pre-factors and temperature contributions to DFT energies at transition states. ☐ In Chapter 4, practical applications of thermochemical scaling relationships are explored. With the aim of circumventing DFT calculations for complex / larger adsorbates and surface reactions, thermochemical data for adsorbates and transition states of C1-C6 n-alkane dehydrogenations on transition metals are revealed to linearly correlate with corresponding gas-phase species. Scaling is shown to largely depend on the construction of the chosen homologous series. To quantify uncertainty in scaling predictions, the thermochemical correlations are implemented in MKMs for ethane and propane hydrogenolysis on Ru(0001). Estimating thermochemistry based on scaling predictions can reduce time and effort when initially constructing an MKM, but for kinetically relevant steps, hierarchical refinement using DFT-computed data may still be needed.
Description
Keywords
Density functional theory, Transition state scaling, Brønsted-Evans-Polanyi, Microkinetic models