Transition state structure in cycloaddition reactions as a function of ring size and geometry

Abstract

The relative energies of the stepwise one-bond and synchronous two-bond reaction pathways are compared at the semi-empirical AM1 and PM3 SCF and ab initio MP2 levels for the closed shell π2s+π4s cycloaddition reaction between the NO2+ ion and ethyne. The intrinsic bias towards the stepwise route for this reaction is estimated at 11 kcal mol–1 for PM3 and rather more for AM1. There is reasonable agreement between the AM1 and PM3 geometries and those obtained using ab initio MCSCF methods for the π2s+π4s reaction between ethene and butadiene, the π2s+π2a cyclodimerisation of ethene, and the π2s+π2a reaction between ethene and ketene. The location of two second-order saddle points for the latter reaction, where the antarafacial component is located on either the ethene or the ketene, allows the antarafacial stabilisation via the carbonyl group to be estimated as 27 kcal mol–1. The effect of E/Z isomerism and of antarafacial components on the geometries and energies of the synchronous stationary points for a range of larger ring cycloadditions is established at the AM1 and PM3 levels. The prediction by Mclver that such reactions will become increasingly asynchronous as the ring size increases is only weakly manifested at the closed shell RHF SCF level, but is more prominent at the spin unrestricted UHF level. The crossover between synchronous and asynchronous bond formation probably occurs for ten-membered rings.