The fate of branched and linear isomers in the rhodium-catalyzed hydroformylation of 3,4,4-trimethylpent-1-ene

Abstract

In nonreversible hydroformylations, the computational evaluation of regio- and stereoselectivities from the relative energy barriers of the transition states (TS) for the alkyl-Rh intermediate formation step is possible, provided all low energy conformers are considered. In contrast, in reversible hydroformylations, also the subsequent reaction steps need to be taken into account to shed some light on mechanistic details. Thus, an extensive comparison of branched (B) and linear (L) reaction pathways for the Rh-catalyzed hydroformylation of 3,4,4-trimethylpent-1-ene (a bulky chiral substrate), going from a number of reactant complexes to products, has been carried out to rationalize the experimental result that pointed to reaction reversibility, although the value of the regioselectivity ratio (B:L = 15:85), based on alkyl-Rh TS free energies, computed under the hypothesis of nonreversibility, was in satisfactory agreement with the experimental one (5:95). A density functional theory approach at the B3P86/6-31G* level coupled to effective core potentials for Rh in the LanL2DZ valence basis set has been employed. By comparing the activation free energies involved in the various steps for the different reactant adducts, interestingly a similar behavior along all the linear pathways is found: the alkyl-Rh formation TS presents the highest barrier; thus, the reaction is nonreversible for all the linear isomers that invariably proceed to yield the linear aldehyde. Conversely, the behavior is quite different along the branched pathways. While some branched isomers eventually produce the corresponding aldehydes, two of the others follow distinct competing pathways, because β-hydride elimination occurs (a) to the terminal olefin-Rh complex (the starting material) that reacts again with the original regioselectivity, increasing the linear fraction, when the CO addition and insertion TS are higher than the alkyl-Rh TS and (b) to the internal olefin-Rh complex when the CO addition and insertion TS are higher than the relevant β-hydride elimination TS, but not than the alkyl-Rh TS. The solvent effect on the reversible profile, evaluated either in the supermolecule approach by adding a benzene molecule to the calculations or in the IEF-PCM framework (e = 2.247), does not bring about any substantial change in the profile, leaving unaltered the conclusions reached.