Date of Award

Fall 2025

Document Type

Open Access Dissertation

Department

Chemical Engineering

First Advisor

Andreas Heyden

Abstract

In recent years, researchers have sought to design efficient chemical processes for plastics upcycling. However, on the computational investigation front, several limitations stymie the development of efficient, representative models for plastics upcycling. Some of such limitations include the large computational cost associated with modeling large molecules, inherent errors in ab initio calculations introduced by the functional used and disparity in the experimental and simulation environments. This work focuses on addressing these limitations.

In the first study, we introduced a methodology for computing accurate degrees of rate control (DRCs), one of the easy-to-implement rubrics for model reduction, using direct sensitivity analysis driven by automatic differentiation (AD). We successfully applied the AD-driven sensitivity analysis to the ethane hydrogenolysis on Pt(111) model. In addition, we demonstrated its superiority to the common finite differences (FD) method. The second study outlined the application of uncertainty quantification using Bayesian statistics and inference to account for inherent errors in DFT computation results. This was used to elucidate how tin doping of platinum catalysts affects the propane dehydrogenation (PDH) chemistry on four different PtSn skin alloys, namely Pt3Sn/Pt(100), PtSn/Pt(100), Pt3Sn/Pt(111), and Pt2Sn/Pt(211), using the BEEF-vdW functional with its 2000 ensembles to capture uncertainties. With slightly better evidence, Bayesian inference on the calibrated models suggested that the stepped surfaces Pt2Sn/Pt(211) may be the active site for the PDH reaction. The last study was aimed at reconciling the differences in activity ofΒ adsorbates and TSs in gas-phase ab initio calculations and melt-phase reaction environment. Melt-phase (βˆ†βˆ†πΊπ΄π‘‘π‘ π‘œπ‘Ÿπ‘π‘Žπ‘‘π‘’ π‘”π‘Žπ‘ β†’π‘™π‘–π‘ž ) effects were computed for selected adsorbates and TSs over Pt(111) surface model at 573 K in an environment of C36H74 polyethylene (PE) surrogate melt using a hybrid QM/MM free-energy perturbation technique. Intermolecular interactions were found to have the most significant contribution to βˆ†βˆ†πΊπ΄π‘‘π‘ π‘œπ‘Ÿπ‘π‘Žπ‘‘π‘’ π‘”π‘Žπ‘ β†’π‘™π‘–π‘ž , and adsorbates/TSs with longer carbon backbones were found to be more destabilized by the PE melt. Lastly, a physics-based descriptor (𝐹) equation was parametrized forΒ (ΔΔGAdsorbategasβ†’liq) based on factors such as the adsorbate size, its van der Waals interactions with the melt-phase and its distance from the catalyst surfaces relative to the polymer height distribution.

Rights

Β© 2025, Olajide Hezekiah Bamidele

Download

Available for download on Friday, December 31, 2027

Share

COinS