Towards More Predictable and Accessible Palladium-catalyzed C–H Activation
The doctoral work overall addresses the issue of making palladium-catalyzed C–H activation more interpretable, predictable and accessible.
Chapter 1 explores the combination of multivariate linear regression (MLR) with nondirected arene C–H olefination. Experimental screening of various ligands informed DFT-parameterized MLR models which successfully enabled the regioselectivity predictions for monodentate pyridine ligands and the identification of a more effective ligand combination. For reaction yields decision trees were deemed suitable for dividing datasets into inactive and active regions, thereby allowing for the in-silico filtering of non-reactive ligands.
Chapter 2 applies MLR to predict a more selective bidentate N-acylamino acid ligand in a regiodivergent thiophene alkynylation. Additional mechanistic studies revealed a ligand-dependent Curtin–Hammett scenario and limitations of DFT or CCSD(T) for qualitative selectivity predictions. This could be traced back to shortcomings of implicit solvation models, namely the inadequate truncation of intramolecular dispersion effects. The crucial role of silver in catalyst regeneration and suppression of alkyne homocoupling, and the beneficial role of the solvent in facilitating substrate coordination was also elucidated.
Chapter 3 translates a C–H deuteration protocol for arenes into an M.Sc. laboratory course. Students applied advanced analytical and experimental techniques and were taught key concepts in C–H activation and NMR spectra simulation. The course was successfully implemented using only commercially available materials and serves as an accessible entry point for newcomers to C–H activation.
Rights
Use and reproduction:
Please note that individual components of the publication may be subject to other licensing or copyright conditions.