A detailed mechanistic investigation of the previously reported ruthenium pseudo-dipeptide-catalyzed asymmetric transfer hydrogenation (ATH) of aromatic ketones was performed. It was found that the addition of alkali metals has a large influence on both the reaction rate and the selectivity, and that the rate of the reaction was substantially increased when THF was used as a co-solvent. A novel bimetallic mechanism for the ruthenium pseudo-dipeptide-catalyzed asymmetric reduction of prochiral ketones was proposed.

There is a demand for a larger substrate scope in the ATH reaction, and heteroaromatic ketones are traditionally more challenging substrates. Normally a catalyst is developed for one benchmark substrate, and a substrate screen is carried out with the best performing catalyst. There is a high probability that for different substrates, another catalyst could outperform the one used. To circumvent this issue, a multiple screen was executed, employing a variety of ligands from different families within our group’s ligand library, and different heteroaromatic ketones to fine-tune and to find the optimum catalyst depending on the substrate. The acquired information was used in the formal total syntheses of (R)-fluoxetine and (S)-duloxetine, where the key reduction step was performed with high enantioselectivities and high yield, in each case.

Furthermore, a new iron-N-heterocyclic carbene (NHC)-catalyzed hydrosilylation (HS) protocol was developed. An active catalyst was formed in situ from readily available imidazolium salts together with an iron source, and the inexpensive and benign polymethylhydrosiloxane (PMHS) was used as hydride donor. A set of sterically less demanding, potentially bidentate NHC precursors was prepared. The effect proved to be remarkable, and an unprecedented activity was observed when combining them with iron. The same system was also explored in the reduction of amides to amines with satisfactory results.