|Home | About | Journals | Submit | Contact Us | Français|
Addition of NbCl5, or NbBr5, to a series of magnesium, lithium, or potassium allylic or propargylic alkoxides directly provides allylic or allenic halides. Halogenation formally occurs through a metalla-halo-[3,3] rearrangement although concerted, ionic, and direct displacement mechanisms appear to operate competitively. Transposition of the olefin is equally effective for allylic alkoxides prepared by nucleophilic addition, deprotonation, or reduction. Experimentally, the niobium pentahalide halogenations are rapid, afford essentially pure E-allylic or allenic halides after extraction, and are applicable to a range of aliphatic and aromatic alcohols, aldehydes, and ketones.
Allylic halides are powerful, versatile electrophiles.1 The excellent electrophilicity stems from stereoelectronic interactions between the σ*C-X orbital and the adjacent π system2 which facilitates a range of efficient and predictable displacements.3 Numerous natural product syntheses have harnessed the excellent electrophilicity of allylic halides to overcome difficult displacements and challenging cyclizations.4
Allylic halide intermediates in total synthesis campaigns are frequently synthesized from aldehydes through olefination-reduction-halogenation sequences (Scheme 1, 1→3→4→2).5 The three-step sequence is necessitated in part because the requisite Wittig reagents suffer facile halide ejection,6 preventing a direct "halo-olefination," and partly because of the predictable conversion of primary allylic alcohols 4 to allylic halides 2 without rearrangement.7 In contrast, regioselective halogenation of secondary allylic alcohols is reagent8 and structure9 dependent with many reactions channeling through both SN2 and SN2' displacement manifolds.10
Direct halogenation of propargylic alcohols similarly affords mixtures of regioisomeric halides.10,11 Consequently a two step sequence of alcohol activation, usually sulfonylation, followed by SN2' halide displacement is typically employed to convert propargylic alcohols 5 to terminal allenic halides 6 (Scheme 2, 5→7→6).12 Subsequent transition metal catalyzed coupling allows a diverse range of bond constructions on these valuable synthetic partners.13
The inherent utility of allylic and allenic halides14 stimulated a direct15 synthesis from carbonyl and alcoholic precursors. Conceptually the transformation centers on a metalla-halo-[3,3] rearrangement16 predicated on metal oxide eliminations17 and the privileged nature of six-membered transition structures (Scheme 3).18 Addition of a vinyl metal bearing an appropriate halide was envisaged to access the allylic alkoxide 9 and trigger a concerted rearrangement to the corresponding allylic halide 10. As sporadically happens in chemical research,19 the same concept was being simultaneously pursued with allylic chlorotitanium alkoxides (9 MX=TiCl3).20 Mechanistic experiments with these titanium alkoxides implicated a stepwise ionization-halogenation sequence rather than a concerted rearrangement, although in principle tuning the metal oxophilicity and halogen nucleophilicity should favor a concerted halogen transfer.21
Publication of the pioneering titanium-based allylic chloride synthesis20 was closely followed by communication22 of a complementary niobium pentahalide procedure. Although preliminary, the use of niobium broadened the substrate scope and offered the promise of a general approach to both allylic and allenic halides through a concerted rearrangement. Complete details of these niobium pentachloride and pentabromide rearrangements are provided with an emphasis on: mechanistic insight; extended substrate scope to include allylic alcohols, aldehydes, enals, ketones, and enones; cascade reduction-halogenation and addition-halogenation strategies; and the synthesis of allenic bromides.
The metalla-halo-[3,3]-rearrangement strategy requires a metal halide capable of simultaneously activating the allylic alcohol, delivering a halogen in an SN2' displacement, and forming a stable metal oxide. Addition of vinylmagnesium bromide to 1-naphthaldehyde (1a),23 as with aldehydes in general,24 forms an allylic magnesium alkoxide but does not trigger an allylic rearrangement (Scheme 4, 1a→4a).25 Assuming that magnesium was insufficiently Lewis acidic, the allylic alcohol 4a was deprotonated with an organometallic base and the corresponding alkoxide 11a26 treated with one of a variety of a metal salts (Scheme 4). Screening numerous metal halides quickly identified the Lewis acidic, oxophilic, high-valent transition metals27 TiCl4, ZrCl4, NbCl4, and NbCl5 as being competent reagents.28 Sequential addition of KH and either TiCl4, ZrCl4, NbCl4, or NbCl5 to a THF solution of 4a afforded varying proportions of the allylic chloride 2a, unreacted allylic alcohol 4a, and 1,4-dichlorobutane arising from the chlorination of THF (Scheme 4, 4a→2a).
Niobium pentachloride was selected for further optimization because of a greater tolerance to ethereal solvents,29 the precedent for chlorinating aliphatic alcohols under forcing conditions,30 and the convenience of using commercial, anhydrous, powdered NbCl5.31 Employing the strong Lewis acid32 NbCl5 under basic conditions is unusual and contrasts with related reactions of high-valent metal halides33 which may well be promoted by adventitious acid produced by partial hydrolysis.34 The non-protic solvents CH3CN, DMF, PhCH3 and CCl4 were significantly inferior to THF whereas Et2O and t-BuOMe afforded comparable results. Reasoning that a bidentate ethereal solvent might be more effective led to the use of 1,4-dioxane which allowed essentially complete conversion to spectroscopically pure 2a in 10 minutes. Aqueous extraction readily removes the spent niobium salt providing pure chloride 2a upon concentration, a significant advantage over related phosphonium-based reagents.35
Deprotonating the naphthyl alcohol 4a with KH in dioxane followed by addition of niobium pentachloride affords the E-allylic chloride 2a in 98% yield. Bulb-to-bulb distillation does not change the spectral purity but decreases the yield to 70% as a result of polymerization in the still pot.36 Rapidly eluting the crude allylic chloride 2a through a short pad of silica or alumina decreases the purity with a concomitant 30–40% mass loss, presumably through irreversible adsorption on silica gel.37
An efficient displacement of the allylic chloride was pursued as an additional proof of reaction efficiency. Initially the anion of malononitrile was selected with the naïve hope of performing an in situ deprotonation-displacement sequence.38 No displacement occurs in dioxane39 whereas THF provided the substituted malononitrile 13a in 65% yield accompanied by 24% of the product of double malononitrile displacement (Scheme 4). Although the overall yield was high, attention was shifted to phenyl sufenylate as a potent nucleophile40 capable of only a single displacement. Not only is the overall chlorination-sulfenylation very efficient (Scheme 4, 2a→14a) but x-ray crystallography41 of the resulting sulfide 13a secured the unequivocal assignment of the olefin stereochemistry.42
The trans-stereochemistry of the allylic chloride 2a formally arises from a metalla-halo [3,3] rearrangement although concerted, ionic, and direct displacement mechanisms appear to operate competitively depending on the structure of the allylic alcohol (Table 1). Geraniol (4b), linalool (4c), and nerol (4d) generate mixtures of allylic chlorides accompanied by terpenyl chloride (2e). Principal conversion of geraniol (4b) to geranyl chloride (2c, Table 1, entry 1) likely occurs through a direct displacement30 whereas formation of terpenyl chloride (2e) requires a change in olefin stereochemistry. Isomerization of geraniol (4b) to linaloyl chloride (2b) is precedented43 and would facilitate an ionic or NbCl5-promoted cyclization to terpenyl chloride (2e).44
A signature of a concerted metalla-halo[3,3] chlorination, dictated by the cyclic transition structure, is preferential formation of an E-allylic chloride (Scheme 3). Linalool (4c) participates in the chlorination to afford some terpenyl chloride but mainly E-geranyl chloride (2c). The absence of Z- neryl chloride (2d) is consistent with a metalla-halo[3,3] rearrangement although the presence of multiple products from these terpenes strongly implies ionization as a significant pathway (Table 1, entry 2).45 Nerol (4d) affords a mixture of all four allylic chlorides with terpenyl chloride (2e) predominating, presumably because the Z-olefin geometry facilitates cyclization (Table 1, entry 3). Although these chlorinations can formally be viewed as metalla-halo-[3,3]-rearrangements, the formation of regioisomeric mixtures suggests that direct displacement and ionization mechanisms operate at least competitively and possibly predominate in some cases.
Chlorinating the cyclic allylic alcohol 4e and the acyclic alcohols 4f–4h with different propensities toward ionization confirm the presence of several competing chlorination mechanisms (Table 1, entries 4–7). Myrtenol (4e) exhibits a three-fold preference for the rearranged chloride 2f despite having a less stable exocyclic olefin compared to the endocyclic chloride 2g (Table 1, entry 4). Benzyl alcohol (4f) in which a metalla-halo-[3,3]-rearrangement is prevented, undergoes halogenation considerably slower and requires two equivalents46 of NbCl5 implying displacement via a bis-niobium complex.47 Activation of the alcohol by an adjacent π-system seems to be significant as the attempted chlorination of 4g was not successful (Table 1, entry 6). In contrast, the hydroxyalkenenitrile 4h bearing two adjacent π-systems affords exclusively the rearranged chloride 2j (Table 1, entry 7). Although several mechanisms appear to compete in the NbCl5 chlorination, the metalla-halo-[3,3]-rearrangement seems to dominate when olefin migration leads to a more stable allylic chloride.
The ability to chlorinate benzyl alcohol stimulated a direct nucleophilic addition-chlorination with aldehydes (Scheme 5). Adding BuMgCl or BuLi to benzaldehyde affords metal alkoxides 15a that are readily transformed into the secondary benzylic chloride 2k upon exposure to 2.5 equivalents of NbCl5. An analogous addition of the chlorine-containing Grignard reagent 1648 efficiently provides the dichloride 2l indicating that NbCl5 tolerates additional chlorination in the substrate (Scheme 5).
The sequential addition-chlorinations with benzaldehyde (Scheme 5) provided an excellent foundation for the direct halo-olefination of aldehydes and ketones (Table 2).49 In the optimized procedure, vinylmagnesium bromide50 was added to a THF solution51 of the aldehyde and then four volumes of dioxane and solid NbCl5 were added. After 10 min the crude chloride52 was isolated and subjected to phenylsulfenylate displacement in THF to afford the corresponding sulfide. In each case, 1H NMR analysis of the crude chloride and sulfide reaction mixture identified the E-alkene as the sole geometric isomer.
Aromatic aldehydes (1a–1d), enals (1e), and ketones (8f and 8g) are smoothly converted to the corresponding chlorides and sulfides (Table 2, entries 1–7). p-Cyanobenzaldehyde (1c) reacts sluggishly with NbCl5 whereas NbBr5 was more significantly reactive,53 a trend apparent in the addition-halogenations with aliphatic aldehydes and ketones (Table 2, entries 8–11).54 The substrates tolerate a nitrile group, a methoxy ether, and adjacent unsaturation in the aldehyde (Table 2, entries 3, 4, 5, respectively). Despite NbCl5 being able to cleave methyl ethers,55 the methyl ether-containing aldehyde 1d is smoothly converted to the allylic chloride provided that the temperature is lowered to 0 °C.56 Acetals are not well tolerated57 suggesting that the method is best suited to the synthesis of hydrocarbon scaffolds bearing limited heteroatom substituents.
The niobium-mediated halogenation of metal alkoxides is equally effective for the addition of organometallics to unsaturated carbonyl compounds (Scheme 6). Sequential addition of hexyllithium or PhMgCl and NbCl5 to acrolein (1l) efficiently provides the allylic chloride 2m and cinnamyl chloride (2n), respectively. Reducing ketone 8m with LiBH4 and adding NbCl5 affords the corresponding chloride that was displaced with sulfenate in an overall reductive-sulfenylation with translocation of the double bond. Collectively these addition-chlorination and reduction-chlorination sequences imply significant scope for halogenating allylic alkoxide intermediates.
The proclivity of allylic alcohols to participate in the formal metalla-halo-[3,3]-rearrangement stimulated expanding the substrate scope to include propargylic alcohols (Table 3). Sequential deprotonation and chlorination of the propargylic alcohol 5a afforded only a trace of the allenyl chloride at room temperature, whereas coaxing the reaction through heating caused considerable decomposition. NbBr5 proved to be a more efficient halogenating agent triggering a smooth rearrangement at room temperature (Table 3, entry 1).
The NbBr5 rearrangement is of reasonably broad scope and provides rapid access to synthetically versatile bromoallenes (Table 3).58 Secondary and tertiary propargylic alcohols react with similar efficiency in affording 1,2-disubstituted and 1,1,2-trisubstituted allenes, respectively (Table 3, entries 1–4 and entry 5). The bromination is equally applicable to alcohols with adjacent aromatic or aliphatic substituents (Table 3, compare entry 1 with entries 2–5). Formally, the allenyl bromide synthesis is envisaged through the hexacoordinate niobiate 1747 although full or partial ionization may occur during the olefin transposition. Internal delivery of the halogen to the olefin terminus is likely promoted by concomitant formation trichloroniobium oxide.59
Allylic and allenic halides are readily generated by adding NbCl5 or NbBr5 to allylic or propargylic alkoxides. The halogenation formally occurs through a metalla-halo[3,3] rearrangement although ionization and direct displacement mechanisms appear to operate competitively. The intermediate allylic alkoxides can be prepared by deprotonation or, equally as effectively, through an organometallic addition or reduction allowing the direct conversion of an aldehyde or ketone to the corresponding allylic chloride. Particularly useful is the direct "halo-olefination" of aromatic and aliphatic aldehydes by sequential addition of vinylmagnesium bromide and NbCl5 or NbBr5. Secondary or tertiary propargylic alcohols react similarly with NbBr5 to afford allenylic bromides. The halides are readily isolated in pure form through simple extraction and can be used in subsequent displacements without prior purification.
Potassium hydride (30% dispersion in mineral oil 1.2 equiv) was washed with hexane (3 mL) and then a dioxane solution of the allylic alcohol (1 equiv) was added dropwise. After 5 minutes solid NbCl5 (1.2 equiv) was added as a dry powder. After 10 min60 the reaction mixture was poured into 2 M HCl and then extracted with ethyl acetate.
A THF solution of vinyl magnesium bromide (0.7 M solution) was added to a 10 °C, THF solution (0.75–0.85 M) of the aldehyde. After 15 min. 1,4-dioxane (4 volumes relative to THF) and solid NbCl5 (2.5 equiv) were added sequentially. After 10 min the reaction mixture was poured into 2 M HCl, extracted with ethyl acetate, and the combined organic extract was then washed with brine and dried (Na2SO4).
A THF solution of vinyl magnesium bromide (0.7 M solution) was added to a 10 °C, THF solution (0.75–0.85 M) of the aldehyde or ketone. After 15 min. 1,4-dioxane (4 volumes relative to THF) and solid NbCl5 (1.2 equiv) were added sequentially. After 10 min the reaction mixture was poured into 2 M HCl, extracted with ethyl acetate, and the combined organic extract was then washed with brine and dried (Na2SO4). A THF solution (0.2 M) of the crude, essentially pure allylic chloride was added to a 10 °C, THF suspension of sodium hydride (2.0–2.5 equiv) to which had been added thiophenol (2.0–2.5 equiv). The reaction mixture was allowed to warm to rt over 16 h, poured into an aqueous sodium hydroxide solution (2% by weight) and then extracted with ethyl acetate. The combined organic extract was washed with brine, dried (Na2SO4), and concentrated under reduced pressure to furnish an oily residue that was purified by radial chromatography to furnish the pure sulfide.
A dioxane solution of the propargyllic alcohol (1 equiv.) was added to a dioxane solution of potassium hydride (1.2 equiv.). After 10 min, the solid niobium bromide (1.2 equiv.) was added. After 2 h, the reaction mixture was poured into HCl (2M), and the phases were separated. The aqueous phase was extracted with EtOAc, and then the combined organic phase was washed with brine, dried (NaSO4), and concentrated to provide an oil which was purified by radical chromatography.
Financial support from the National Institutes of Health (2R15AI051352-03) and in part from the National Science Foundation (CHE 0808996, 0421252 HRMS, 024872 x-ray, and 0614785 for NMR facilities) are gratefully acknowledged. Dr. Jin-Lei Yao is thanked for support in solving the x-ray structure of 14a.
Supporting Information Available. 1H NMR, and 13C NMR spectra for all new compounds, an ORTEP for 14a, and complete experimental details following the general procedures. This material is available free of charge via the Internet at http://pubs.acs.org.