Acetylene−Dicobaltcarbonyl Complexes with Chiral Phosphinooxazoline Ligands: Synthesis, Structural Characterization, and Application to Enantioselective Intermolecular Pauson−Khand Reactions

Abstract

The reaction of the phenylacetylene−dicobalthexacarbonyl complex (2) with the 4-R-2-(2-diphenylphosphinophenyl)oxazolines 1 (R = Ph) and 4 (R = CH2CH2SCH3) leads to the selective formation of the chelated complexes 3 and 5, respectively. On the other hand, the tert-butyl-substituted phosphinooxazoline 6 acts as a monodentate ligand, and its reaction with several 1-alkyne-derived complexes (2,7−10) affords readily separable mixtures of the diastereomer nonchelated complexes 11a,b−15a,b. The interconversion rate between diastereomeric pairs is dependent on the steric bulk of the alkyne substituent, and neither 3 nor 5 epimerize at room temperature. The structures of both kinds of complexes have been ascertained by a combination of spectroscopical (IR, NMR), X-ray diffraction, and chiroptical methods; this has allowed the development of a practical procedure for the establishment of the absolute configuration of the chiral alkyne−dicobaltcarbonyl complexes obtained by the selective substitution of a carbon monoxide on one of the diastereotopic cobalt atoms. The intermolecular Pauson−Khand reaction of the chelated complexes 3 and 5 with norbornadiene respectively affords the (+) and (−) enantiomers of expected enone adduct 25, but in low enantiomeric excesses. Contrary to that, the tertiary amine N-oxide-promoted intermolecular Pauson−Khand reactions of nonchelated complexes 11a,b−13a,b give the corresponding norbornadiene- or norbornene-derived adducts both in high yields (85−99%) and enantioselectivities (93−97% enantiomeric excess), in what constitutes a substantial improvement over preexisting procedures for this reaction. The possibility of achieving chiral induction in the Pauson−Khand reaction of symmetrical alkynes (via the corresponding dicobaltpentacarbonyl complexes with ligand 6) has been demonstrated for the first time. An enantioselectivity mnemonic rule and a mechanistic model that explains the observed asymmetric sense of induction have been developed, and have been found to be in agreement with the results of model semiempirical molecular orbital calculations.