Evaluation of ligand effects in the modified cobalt hydroformylation of 1-octene. Crystal structures of [Co(L)(CO)3]2 (L = PA–C5, PCy3 and PCyp3)

Abstract

A series of phosphine ligands with different electronic and steric properties were evaluated at fully modified conditions in cobalt catalysed hydroformylation of 1-octene. The steric demand of the ligands was based on the Tolman cone angle model covering a range of 132–175°. The electron donating ability was evaluated through the first order Se–P coupling constants as determined from the corresponding phosphine selenides covering a range of 672–752 Hz. Crystal structures of three phosphine modified cobalt dimers, [Co(CO)₃(L)]₂ (L = PA–C₅, PCy₃ and PCyp₃ with PA–C₅ = 1,3,5,7-tetramethyl-8-pentyl-2,4,6-trioxa-8-phosphatricyclo[3.3.1.1^3,7]decane), are reported. The Phoban and Lim ligands (Phoban = mixture of 9-phosphabicyclo[3.3.1 and 4.2.1]nonane, Lim = 4,8-dimethyl-2-phosphabicyclo[3.3.1]nonane) resulted in systems about twice as active as most of the other ligands investigated, these ligands have a high Lewis basicity with ¹J_Se–P values from 684–687 Hz. The linearity of the alcohol product in general decreased for the less electron donating ligands while no clear relationship was evident as a function of steric size. The parallel competing hydrogenation of 1-octene to octane varied from 9–15% for a cone angle range of 132–172°, but a sharp increase of up to 40% was observed for PA–C₅, PCy₃ and PCyp₃, all with cone angles > 169°. The catalytic behaviour provides evidence that is contrary to the dissociative substitution of CO by an alkene as the rate limiting step in all cases. For large symmetrical ligands, such as PA–C₅, PCy₃ and PCyp₃ the rate limiting step may move within the catalytic cycle and may now be situated at the carbonylation step where the chemoselectivity is also determined. The lack of clear correlation between the steric and electronic effect of the ligands and all catalytic parameters may serve as additional proof that the same system, especially in terms of the rate determining step, is not operative in all cases. The Phoban and Lim systems are superior with the highest reactivity and lowest alkene loss through hydrogenation. The unsymmetrical nature of the Phoban and Lim ligands may provide flexibility to adopt geometries inducing both high and low steric crowding, which may be a reason for its beneficial catalytic properties.