Reactions catalyzed by transition metal complexes almost always entail binding of one or more reactants to the metal center, and nearly every corner of the "chiral pool" has been picked over in efforts to develop enantioselective catalysts. As reported by Alfred Werner in 1911-1912, salts of the formally D3-symmetric [Co(en)3]3+ trication (en = ethylenediamine) were among the first chiral inorganic compounds to be resolved into enantiomers. These air- and water-stable complexes are substitution-inert, so for 100 years they languished without application in organic synthesis. We then showed that when they are rendered soluble in organic media by lipophilic anions such as fluorinated tetraarylborates BArf-, they become potent catalysts for a variety of carbon-carbon and carbon-heteroatom bond forming reactions.These involve substrate activation by hydrogen bonding to the coordinated NH2 units (pKa ca. 15), a "second coordination sphere" mechanism. Only modest enantioselectivities are obtained with [Co(en)3]3+ 3BArf- or related chromium, rhodium, iridium, and platinum salts. However, high enantioselectivities are achieved when the three en ligands are replaced by the 1,2-diphenyl analogues (S,S)- or (R,R)-H2NCHPhCHPhNH2. Here only one BArf- anion is required to solubilize the trication, so a number of mixed-salt catalysts (2X-BArf-) have been evaluated. Alternatively, a dimethylamino group can be tethered to the backbone of one en ligand, providing bifunctional catalysts that obviate any need for an external base. Interestingly, the counteranions modulate the enantioselectivities somewhat. However, catalysts with chiral anions do not significantly outperform benchmark catalysts with achiral anions. Cagelike chiral hexaaminecobalt(III) complexes known as sepulchrates and sarcophagines, which feature secondary NH donor atoms, can also serve as catalysts, but the enantioselectivities are very low.In a spinoff application, certain salts are found to be superb "chiral solvating agents", leading to distinct sets of NMR signals for enantiomers of chiral analytes with Lewis basic functional groups. Loadings of 10-25 mol % generally suffice, providing the best way of assaying the enantiomeric purities of a host of compounds. Also, mixtures of several chiral compounds can be simultaneously analyzed. It is not surprising that complexes that perform well in chiral recognition phenomena also excel as enantioselective catalysts.In this Account, the stereochemical properties of the preceding complexes are treated, as well as arcana generally known only to specialists in the field. These include the use of charcoal for equilibrating configurations of the cobalt stereocenter and Sephadex for separating enantiomers and diastereomers. Other types of metal-containing hydrogen-bond-donor catalysts are briefly surveyed (noncoordinating NH units can also be effective), including several developed by other groups. However, the mechanisms of enantioselection in all of these transformations remain obscure. The optimum diastereomer and anion set varies from reaction to reaction, suggesting a "phenotypic plasticity" that allows adaption to a variety of processes.
Read full abstract