Angew. Chem. Int. Ed. 53, 5903–5906 (2014)

Credit: WILEY

Computationally predicting the structure of molecules with desired reactive properties is not only intellectually satisfying, but can also be a big saver of time and resources: in particular, avoiding the need for numerous 'trial-and-error' screenings is a significant attraction. Such predictions, however, still pose a big challenge to chemists — successful examples of molecules with pre-designed reactivity, followed by their synthesis and subsequent experimental validation, are rare.

It is in this context that Franziska Schoenebeck and co-workers from RWTH Aachen University and ETH Zürich began studying reductive elimination of trifluoromethyl groups. Reductive elimination of ArCF3 from palladium(II) centres is the final step in the Pd-catalysed synthesis of trifluormethylarenes, but it is known to be a challenging process. It depends largely on the properties of the coordinated ligand — only a handful of ligands are capable of prompting such chemistry and all have large bite angles. Smaller bite-angle compounds had proved ineffective, including the simple ligand 1,2-bis(diphenylphosphino)ethane (DPPE). Previous analysis by Schoenebeck and co-workers had explained this as a destabilization of the reductive elimination transition state by the bulky phenyl substituents on DPPE, rather than being directly related to the bite angle. Modification at this position could therefore be a fruitful route to avoid transition-state destabilization and, ideally, to also destabilize the starting complex.

Computational results showed that any replacement for phenyl on DPPE should be small. Coupled with a desire to destabilize the starting complex, this led to the idea of using CF3 not just as a leaving group but also to replace the phenyl groups of DPPE — this was proposed to cause destabilization by electrostatically repelling the leaving groups. Calculations on this hypothetical structure predicted a barrier to activation similar to those of successful wide bite-angle ligands. The next step was to make the complex and see if it worked. After some optimization the Pd(II) complex with the designed ligand plus phenyl and CF3 leaving groups was synthesized and clean reductive elimination to PhCF3 was observed.