• alcohols;
  • density functional calculations;
  • heterogeneous catalysis;
  • hydrogen transfer;
  • racemization;
  • ruthenium


Thumbnail image of graphical abstract

Which mechanism? Dissociation of a CO ligand from alkoxide complex 1 to give intermediate 2 seems to be a viable pathway for racemization according to DFT calculations. The highest Gibbs free energy barrier is 22.6 kcal mol−1 for this pathway. Very fast alkoxide exchange can occur in 2 with essentially no barrier. The calculations indicate an alternative low-energy pathway for alkoxide exchange between 1 and 1′ without CO dissociation with a barrier of only 13 kcal mol−1.

Two possible pathways of inner-sphere racemization of sec-alcohols by using the [RuCl(CO)25-pentaphenylcyclopentadienyl)] catalyst (1) have been thoroughly investigated by means of density function calculations. To be able to racemize alcohols, catalyst 1 needs to have a free coordination site on the metal. This can be achieved either by a η5[RIGHTWARDS ARROW]η3 ring slippage or by dissociation of a carbon monoxide (CO) ligand. The η5[RIGHTWARDS ARROW]η3 ring-slip pathway was found to have a high potential energy barrier, 42 kcal mol−1, which can be explained by steric congestion in the transition state. On the other hand, CO dissociation to give a 16-electron complex has a barrier of only 22.6 kcal mol−1. We have computationally discovered a mechanism involving CO participation that does not require η5[RIGHTWARDS ARROW]η3 ring slippage. The key features of this mechanism are 1) CO-assisted exchange of chloride for alkoxide, 2) alcohol–alkoxide exchange, and 3) generation of an active 16-electron complex through CO dissociation with subsequent β-hydride elimination as the racemization step. We have found a low-energy pathway for reaction of 1 with potassium tert-butoxide and a pathway for fast alkoxide exchange with interaction between the incoming/leaving alcohol and one of the two CO ligands. We predict that dissociation of a Ru-bound CO ligand does not occur in these exchange reactions. Dissociation of one of the two Ru-bound CO ligands has been found necessary only at a later stage of the reaction. Though this barrier is still quite high, our results indicate that it is not necessary to cross the CO dissociation barrier for the racemization of each new alcohol. Thus, the dissociation of a CO ligand is interpreted as a rate-limiting reaction step in order to create a catalytically active 16-electron complex.