Teoretické zkoumání zeolitů pomocí NMR: Od DFT po strojové učení

Lei, Chen

Solid-state nuclear magnetic resonance (ssNMR) spectroscopy is a highly valuable tool for experimentally characterizing zeolites, due to its high sensitivity and the detailed structural insight it provides from the spectra. However, interpreting ssNMR spectra is challenging, and accurately determining the structural features is often impractical, especially for the quadrupolar nuclei. Consequently, computational modelling of ssNMR spectra has become increasingly popular to complement experimental studies. Despite its usefulness, NMR simulations tend to be simplified to limited numbers of static calculations, and there is a lack of thorough studies under realistic conditions, taking into account factors such as temperature variations, dynamic evolution over time, appropriate Si/Al ratios to match experimental samples, and hydration effects. This thesis focuses on the computational study of 27 Al and 23 Na NMR related to active acid sites in zeolites, emphasizing the need for operando modeling to understand NMR spectra. Two approaches are proposed for operando modeling of NMR properties based on Density Functional Theory (DFT) and Machine Learning (ML) methods. The research contains three main parts. The first part combined experimental 27 Al NMR spectroscopy with ab initio calculations to reveal the...

guest :: login Digital Repository
		Search		Submit		Help		About