The Brønsted relationship, hydroxide solvent isotope effects, and isotope exchange in E1cB elimination of 9-(dimethylaminomethyll)fluorene

Abstract

A comparison is reported between Bronsted coefficients for catalysis by amine bases (β) and from kinetic and equilibrium measurements of hydroxide solvent isotope effects: k_OD/k_OH=(K_OD/K_OH)^βi, for E1cB elimination of 9-(dimethylaminomethyl)fluorene and its conjugate acid to form dibenzofulvene. As in related reactions, agreement between values is poor (β_i 0.31; β > 0.65), and better agreement is sought by separating the isotope effects into contributions from the hydrogens of the hydroxy-group and its solvation shell, making use of kinetic measurements in 1 :1 H₂O–D₂O mixtures. Following Gold, β values are defined from fractionation factors for the transition state (ϕ₁ and ϕ₂) and hydroxide ion (ϕ_a and ϕ_b) as ϕ₁=ϕ_a^{1 –βa} and ϕ₂=ϕ_b^{1 –βb} for the hydroxy and solvating hydrogens, respectively. It is found that the measured β values are very sensitive to the values of ϕ_a and ϕ_b, and since these themselves are poorly defined experimentally it appears that evaluation of a β value providing a practical measure of transition state structure is at present not possible. On the other hand, the measurements in 1 : 1 H₂O–D₂O provide a constraint on ϕ_a and ϕ_b in that solutions for the transition state fractionation factors ϕ₁ and ϕ₂ exist only outside the ranges ϕ_a 0.45–1.43 and ϕ_b 0.99–0.67. This result is sensitive to experimental error but not to the fractionation model chosen for the transition state. An attempt to compare hydroxide with methoxide solvent isotope effects in MeOH and MeOD was prevented by the occurrence of isotope exchange between solvent and substrate in MeOD. Kinetic analysis of the competing reactions of exchange and elimination in MeOD, however, offers confirmation of the mochanism of reaction and of the involvement of carbanion and zwitterion intermediates in the elimination.