. J. Chem. 65, 363 (1987).Racemization and radio-bromide exchange kinetics for I-phenylbromoethanes in acetonitrile and in nitromethane using tetrabutylammonium bromide are reported. The results, together with those previously reported for acetone solutions, provide direct empirical support for the ion-pair mechanism for nucleophilic substitution at saturated carbon. Changing the substituents on the phenyl from the 4-nitro through to the 3,4-dimethyl substrate and the solvent from acetone to the more polar acetonitrile and nitromethane shifts the transition state for bromide substitution from an early to a late stage of the equilibria series substrate intimate ion pair various solvated ion pairs free or dissociated ions. For all the substrates in acetone and, for the species giving the less stable carbocations, in acetonitrile and nitromethane, both racemizations and exchanges are bimolecular. In the latter solvents, the substrates giving the more stable carbocations show mixed kinetics. The ion-pair mechanism for nucleophilic substitution at saturated carbon was presented in some detail in the previous I paper in this series (8). It was pointed out that while there was I much direct empirical evidence for the involvement of ion pairs,
Iespecially for solvolysis in reactions at or near the unimolecular 1 extreme (3, 9), perhaps the best direct empirical evidence for I their involvement in bimolecular substitutions was provided in that paper. Chiral 1 -phenylbromoethanes were prepared (10) with the following substituents on the phenyl group: 4'-nitro, 4'-bromo, unsubstituted, 3'-methyl, 4'-methyl, and 3',4'-dimethyl. The rates of racemization of chiral substrate and of radio-bromidebromide exchange on racemic substrate were determined in acetone containing lithium bromide. If the Hughes-Ingold mechanisms apply, the rate constants of racemization and exchange should be equal for a unimolecular process whereas for a bimolecular one the ratio should be 2.00: 1 .OO since the inversion of one half of the substrate results in a racemic mixture. In fact Hughes and Ingold reported that, for the parent compound, 1 -phenylbromoethane itself, the rate of racemization was, within experimental error, indeed twice the rate of bromide exchange (2), an observation we previously confirmed (1 1). With the 4-methyl compound, by contrast, the racemization to exchange rate constant ratio (k,,,/k,,,~) was found to average 2.36: 1.00. Both racemization and exchange were bimolecular, first order in each of alkyl bromide concentration and bromide activity.2 These results are very difficult to accommodate with the Hughes-Ingold mechanisms but present no problem with the ion pair cheme (Scheme 1). Apparently a portion of the initially formed contact ion pair is being sufficiently solvent separated for some of the carbocations to turn over with respect to the bromide ion before collapsing to starting material with partial racemization. The observed rate ratio means that racemization via the internal pathway is more than one third as fast as...