Clear-cut difference in the rearrangement of 1-bromo-2-(2-phenyl-3-butenyl)benzene under anionic or radical conditions

The o -(3-butenyl)phenyl system bearing a phenyl group on the 2-position of the side chain was studied as a potential mechanistic probe for distinguishing between radical and carbanion intermediates. The Bu 3 SnH reduction of 1-bromo-2-(2-phenyl-3-butenyl)benzene, 4 , yields mainly the 1,5-cyclization product, 6 , with a preference for the trans-isomer. Treatment of 4 with Mg or t -BuLi leads to double-bond isomerization and yields ( E )-1,2-diphenyl-2-butene, 8 , along with 3,4-diphenyl-1-butene, 5 .


Introduction
Mechanistic studies of chemical and biochemical processes often include reactions of radical probe substrates.6][7][8] However, cyclization of the o- (3-butenyl)phenyl anion 2c has to be considered (Scheme 1).The reaction of 1 with 2 equiv of n-BuLi yields 42% of cyclization product after 5 minutes at room temperature. 9This anionic rearrangement was confirmed later by a series of electrochemical experiments. 10Consequently, tests using the radical probe 1 have to be carried out under conditions where the anion does not cyclize, at low temperature, or by increasing the proton-donating ability of the solvent with adequate additives. 3,10,11This mechanistic drawback is not specific to radical probe 1.Several probes based on the cyclopropylcarbinyl radical ring-opening have been applied in studies of P-450 oxidation of hydrocarbons. 12However, the corresponding carbocations lead to the same rearranged products.Newcomb et al., by substituting the cyclopropyl ring with appropriate substituents, designed a hypersensitive mechanistic probe to differentiate between radical and carbocation intermediates. 13,14In mechanistic studies involving 6-halo-1-heptenes, this information was provided by the proportion of cis-and trans-isomers of the 1,2-dimethylcyclopentane obtained.6][17][18] Cohen et al. used such cis-/trans-ratios of cyclization products to show that unsaturated alkylzinc iodides cyclize by a radical rather than a polar mechanism.They found that probably all of the organozinc cyclizations are radical chain reactions initiated by traces of oxygen. 19These findings appear as an additional warning against deceptive rearrangements observed with radical probes. 20he aim of our work was to design a new radical probe for distinguishing between o-(3butenyl)phenyl radical 2r and the corresponding carbanion 2c intermediates.On the basis of the results obtained with 6-halo-1-heptenes, our probe design was first based on the simple idea that substitution on the allylic position of the butenyl chain with a phenyl group could induce a difference in the stereochemistry of cyclization according to the nature of the intermediate.

Scheme 3
These results conform to Beckwith's studies on the o-(3-butenyl)phenyl radical bearing substituents on the 2-position of the butenyl chain.These radicals show a preference for the formation of trans-disubstituted products and give also small amounts of 6-membered rings (endo products). 29,30With a larger excess of Bu3SnH (5 equiv., 0.46 M) the ratio 5/6/7 becomes 15/77/8.The application of the steady-state theory shows that when Bu3SnH is in large excess, The quantities [C] and [U] are the final concentrations of cyclized and uncyclized products, respectively, and kc is the sum of the rate constants for cyclization. 31he value kH is the rate constant for the reaction of the aryl radical with Bu3SnH.Taking kH = 10 9 s -1 (80 °C), 5 cyclization of radical 5r occurs with a rate constant kc of about 2.6 x 10 9 s -1 at 80 °C. 32This value is slightly higher than the previous determination of the rate constant for cyclization of 2r (kc = 10 9 s -1 , 80 °C). 5,6,29,31A solution of benzene, radical probes 1 (1 equiv., 0.05 M), 4 (1 equiv., 0.05 M), AIBN (0.19 equiv) and an excess of Bu3SnH (0.95 M, added last) was heated under reflux for 5 h.After work up, we examined the ratios [C]/[U] for each structure.These ratios were respectively 61/39 and 73/27 for 1 and 4. Assuming that kH has the same value for alkenylphenyl radicals 2r and 5r, we confirm that 5r cyclizes faster than 2r.This rate-enhancing effect of the phenyl group was expected.Beckwith, in the same way, found an increase of kc when the 3-butenyl chain was substituted with a methyl or an ester group on the 2-position. 29he carbanion chemistry of 4 was first studied under metal-halogen exchange reaction conditions.The reaction of 4 with t-BuLi (2.2 equiv.) in THF generates, after quenching with water, a mixture of hydrocarbons 5 and 8 in a ratio 5/8 = 34/66 (Scheme 4).No cyclized product 6 was formed (<1%).Instead, double bond isomerization takes place, which is in surprising comparison with the cyclization observed with the aryllithium reagent formed from 1.This migration is reminiscent of the results described by Wilson et al., 33 in which reaction of 3,4diphenyl-1-butene, 5, at 0 °C with 1 equiv. of n-BuLi in a hexane/THF mixture yields, after quenching, 78% of 8 and 17% of the starting derivative, 5.

Scheme 4
The lithium-halogen exchange is an extremely fast process.Coupling products may be avoided by the use of two equivalents of t-BuLi.The second equivalent reacts rapidly with the t-BuBr formed, even at very low temperature, to give 2-methylpropene, 2-methylpropane and lithium halide. 346][37][38][39] However the interpretation of this kind of competition can be far from straighforward. 40Under our reaction conditions, the total conversion of 4, and the low yields of incorporation of the t-Bu group observed (≤1-20%, see the Experimental Section) suggest that the excess of t-BuLi mainly quenches the t-BuBr formed, rather than deprotonating the allylic position.The formation of the rearranged product 8 could then be explained by an intramolecular 1,4-proton transfer leading to the delocalized carbanion 8Li (Scheme 5).Previous work proposes that such a proton transfer leading to an allyllithium intermediate could compete effectively with the cyclo-isomerization of an aryllithium. 41ntramolecular 1,4-proton transfer was also observed in the 5-hexenyl anion so that, under favorable conditions, the 5-hexenyl probe detects both alkyl radical and anion intermediates. 42ntramolecular proton transfers can be very fast, the proximity of reactive centers and the angle of attack playing an important role. 43

Scheme 5
Another way of generating carbanionic species was to study the reaction of 4 with Mg.This reaction, in THF at room temperature, yields a mixture of 5 and 8 in a ratio of 5/8 = 91/9 (Scheme 6).No cyclized product 6 was formed (<1%).In a series of works, Bickelhaupt's group showed [45][46][47][48] that the reaction of aryl halides with magnesium involves reactive aryl carbanion intermediates whose presence is evidenced by well identified by-products formed in yields of 4-85%.The detailed interpretation of these results involves an aryl bromide radical anion whose cleavage provides an aryl radical which is reduced extremely rapidly into a carbanion.Combination of this carbanion with MgBr + leads to the Grignard reagent. 49The productdistribution depends on the rate of the reaction trapping the carbanion versus the barrier to the association of the carbanion with Mg II . 50Such competition, applied to the reaction of 4, is indicated in Scheme 6.Thus, considering Bickelhaupt's works, the rearranged product 8 would originate from a carbanion intermediate rather than the Grignard reagent itself.In both cases, this experiment indicates that the aryl carbanion species formed from 4 lead to the rearranged product 8.

Conclusions
These results show that the radical probe derived from 4 behaves in a specific way.The presence of the phenyl ring in the αposition to the exocyclic unsaturation increases considerably the acidity of the proton α to this double bond.As a consequence, the rate of proton abstraction becomes higher than the rate of cyclization of the aryl carbanion.Under such conditions, the use of probe 4 to distinguish between radical and carbanionic route would be based directly on the nature of the formed products rather than on subtle stereochemical difference in yield of cyclized products.This work opens the way to a new family of radical probes designed for the study of reactions where carbanions and radicals are possibly involved.