The isomerization of organic isocyanides to cyanides has sporadically piqued interest over the past 45 years. 1 In the gas phase, evidence has been put forth for a unimolecular sigmatropic pathway, 2 whereas in solution, sigmatropic, free radical, and cationic chain mechanisms can occur, depending on the conditions. 1a,3 Silyl cyanides have been shown to exist in equilibrium with the isocyanide form, exchanging intermolecularly in the liquid phase and intramolecularly in dilute solution. 4 Because of severe instability, the few compounds known in which isocyanide is bonded to a heteroatom 5 (O, S, N, P) have not been candidates for isomerization studies.We recently reported the stable isomers 1 and 2. 6 Here we report a detailed kinetic study of the isomerization of 1 in which we postulate an unusual dicationic intermediate in a chain mechanism.Isomerization of 1 to its more stable isomer 2 takes place slowly in acetonitrile (90% conversion in 96 h at 100°C), but not in benzene (<2% conversion under the same conditions). 6 Failure of the reaction in benzene is consistent with the inability of this low dielectric constant solvent to support ionic intermediates. The bromo analogue 3, when present, also plays an important role in the isomerization (see below).The kinetics of the transformation of 1 to 2 was studied by 31 P NMR spectroscopy in CD 3 CN. The reaction clearly does not follow first-order kinetics, in that plots of ln([1] t ) against time are not linear; further, and even more telling, the approximate first-order rate constants increase with [1] 0 . These data rule out a mechanism in which 1 dissociates into its component ions which then recombine to form 2. After reflection on the nature of the variation of [1] t and other possible kinetic schemes, we decided to attempt to fit the kinetic data to the mathematical form for the reaction sequence presented in eqs 1-4, Scheme 1. Note that the reverse of eq 1 is not included because eq 4 takes precedence owing to the greater thermodynamic stability of 2. Although this scheme appears to resemble that of a standard chain reaction, with eqs 1 and 4 being the respective initiation and termination steps and eqs 2 and 3 the propagation steps with the intermediates P 2+ and CN -, respectively, there is a major difference. One intermediate does not undergo a reaction to regenerate the other; rather, each regenerates itself (eq 2). Such a reaction step has been previously postulated for the isomerization of tritylisocyanide. 3b The postulate of a bicyclic P 2+ intermediate in Scheme 1 suggests structure 4, which is isoelectronic/isostructural with boratrane 5 whose structure has been determined. 7With the steady-state approximation for the concentrations of the P 2+ and CN -intermediates, these expressions result:The expression for the reaction rate, with the further assumption that the kinetic chains are long (i.e., that the rate of propagation is much larger than initiation), is the following:According to this analysis, the kinetic data should follow 3/2-order kinet...