Use of Cyclic Allylic Bromides in the Zinc–Mediated Aqueous Barbier–Grignard Reaction

The zinc–mediated aqueous Barbier–Grignard reaction of cyclic allylic bromide substrates with various aldehydes and ketones to afford homoallylic alcohols was investigated. Aromatic aldehydes and ketones afforded adducts in good yields (66–90%) and with good diastereoselectivities. Non–aromatic aldehydes also reacted well under these conditions, but only poor yields were obtained with non–aromatic ketones. Regioselectivity was high when some substituted cyclic allylic bromides were investigated.


Introduction
The synthesis of homoallylic alcohols via the reaction of cyclic and acyclic allylic organometallic compounds with aldehydes and ketones is an important process in synthetic organic chemistry [1]. Unfortunately, however, the process typically requires the use of toxic and/or water-sensitive organometallic compounds. Recently a "greener" allylation method has been developed in which the reaction is carried out in aqueous media under Barbier-type conditions using allylic halides and metals such as zinc, tin, and indium [2]. The method is simple, avoids the handling of toxic and watersensitive reagents, and generally affords good-to-high yields of products. In addition, examples of impressive regio-, diastereo-, and (more recently) enantioselectivities have also been reported [2,3]. While acyclic allylic halides have been subjected to detailed investigations in this regard, the reactivity of cyclic allylic halides-with only a few exceptions [4]-have escaped attention despite the fact that the cycloalkenyl-substituted methanols generated from these reactions are of considerable synthetic importance [1b,5]. We have therefore undertaken a systematic study of the reactivity of a series of cyclic allylic bromides under Barbier-type conditions with the intention of determining the feasibility and scope of the method, as well as the regio-and diastereoselectivity of the process.

Results and Discussion
Addition of 2 equivalents of 3-bromocyclohexene (1b) as a solution in THF to a rapidly stirred mixture of zinc metal (2 eq) and benzaldehyde (1 equiv) in saturated aqueous NH 4 Cl resulted in rapid consumption of the zinc metal in a mildly exothermic reaction. The addition product 2b (R = C 6 H 5 , R' = H) was isolated in 87% yield ( Table 1).
The 1 H-NMR spectra of both the crude and purified adduct suggested the presence of diastereomeric products. Gas chromatographic analysis indicated an 83:17 ratio of the erythro and threo isomers, respectively, identified by comparison of their 1 H-and 13 C-NMR spectra with literature data [5a]. The major by-product from this reaction was a dimeric compound formed from Wurtz-type coupling of the starting bromide [6].
Bromide 1b reacted with substituted benzaldehydes to afford adducts in good yields and stereoselectivities (see Table 1). In all cases, erythro adducts were obtained as the predominant diastereomer as determined by comparison of their 1 H-and/or 13 C-NMR spectra with those of the same or similar compounds [5a-b,7]. Reaction with the non-aromatic aldehydes heptaldehyde and isobutyraldehyde, however, afforded very low stereoselectivity although the mixtures of diastereomers were obtained in reasonable yield. The change in reactivity from aromatic to non-aromatic aldehydes was also marked by a change in diastereoselectivity from that favoring formation of erythro adducts to that favoring threo adducts. A reasonable yield of addition product was also obtained with the aromatic ketones acetophenone and benzophenone, and with good diastereoselectivity in the case of acetophenone [8]. However, the non-aromatic ketones 3-pentanone and acetone afforded only poor yields of adduct (12% and 28% yield, respectively) and were not further pursued.  In order to determine the scope of the method, a series of cyclic allylic bromides (1a-d) that differed in ring size was investigated. Utilizing the same experimental protocol as was used with 1b, with the exception that three equivalents of bromide and zinc were used rather than two, good yields of addition products were obtained with tolualdehyde as substrate in all cases (Table 1). The stereoselectivity of the process was found to be nearly independent of the size of the bromide ring. The reactivity of bromide 1a towards hydrolysis under the aqueous reaction conditions required conducting the reaction at 0 °C rather than the usual room temperature to limit competing formation of 2cyclopenten-1-ol.
Reaction of bromide 3 with benzaldehyde could have conceivably afforded adduct 4 or regioisomer 5 by way of the two possible isomeric organometallic intermediates (see below). We found, however, that standard reaction conditions yielded only adduct 4 in 61% yield as a 90:10 mixture of erythro/threo diastereomers, respectively, resulting from reaction at the more highly substituted allylic site [7]. This finding is in agreement with those previously reported for acyclic allylic bromides in which reaction was also generally observed to occur at the more substituted allylic position [2].
Similarly, reaction of bromide 6 with tolualdehyde afforded adduct 7 as the sole product in 82% yield as a 91:9 mixture of diastereomers [9]. The major diastereomer of 7 is tentatively assigned the erythro configuration based on two observations: 1) the chemical shift of the carbinol proton of the major diastereomer was found at lower field in the 1 H-NMR (δ 4.67) relative to that of the minor isomer (at δ=4.52) as was observed for the erythro isomers of the other aromatic adducts 2a-d, and 2) the GC retention time of the major isomer was shorter than that of the major, and it was observed in all cases examined by us that the erythro isomer consistently eluted from a carbowax GC column prior to that of the threo isomer. Given the potential for conversion of compound 7 to ketone 8 via oxidative cleavage of the exocyclic double bond, this reaction presents itself as a possible aqueous-based synthetic alternative to the conventional aldol reaction. Studies directed towards the exploitation of this potential route are underway in our labs.

Conclusions
Cyclic allylic bromides behave admirably in the zinc-mediated aqueous Barbier-Grignard reaction towards aromatic and non-aromatic aldehydes, as well as towards aromatic ketones to afford good yields of the corresponding homoallylic alcohols. Good diastereoselectivities were observed for aromatic aldehyde and ketone substrates.

Acknowledgments
Acknowledgment is made to the Donors of The Petroleum Research Fund, administered by the American Chemical Society, for partial support of this research. We also thank Berry College for partial support of this work. The FT-NMR upgrade used in this work was made possible by NSF DUE ILI-IP grant # 9750684.

General
Unless otherwise indicated, reagents were obtained from commercial suppliers, and used without further purification. Column chromatography was conducted on Merck grade 60 (230-400 mesh) silica gel. 1 H-and 13 C-NMR spectra were recorded in CDCl 3 at 60 and 15 MHz, respectively. Gas chromatography was conducted using a 6' × 1/8" stainless steel column packed with 10% carbowax 20M on 80/100 Chromosorb W AW. Bromides 1a and 1c-1d were synthesized via standard allylic bromination of the corresponding alkenes utilizing N-bromosuccinimide followed by distillation at reduced pressure [10].

General Procedure for the Reaction of Cyclic Allylic Bromides with Aldehydes and Ketones.
A solution of the cyclic bromide (2 or 3 mmol) in THF (2 mL) was added dropwise to a rapidly stirring mixture of aldehyde or ketone (1 mmol), saturated aqueous NH 4 Cl (1 mL) and zinc metal (2 or 3 mmol, see text). An immediate reaction was observed to take place with loss of the zinc metal. The mixture was stirred 3 h, filtered to remove excess zinc and precipitated salts, and the organic layer separated. The aqueous layer was washed with ether (3 × 1 mL). The combined organics were dried over Na 2 SO 4 , filtered, and concentrated. Reaction products were purified by column chromatography (SiO 2 ), eluting with a suitable hexane/ethyl acetate solvent mixture. The results are summarized in Table 1. NOTE: For the compounds below, the erythro and threo diastereomers were inseparable by column chromatography. Data is provided for the major (erythro) isomer, but where distinct differences were discernable, information for the threo isomer is included in square brackets (i.e., [ ]) immediately following the corresponding data for the erythro isomer.