Ni-catalyzed migratory fluoro-alkenylation of unactivated alkyl bromides with gem-difluoroalkenes

A migratory fluoro-alkenylation of unactivated alkyl bromides is reported; the reaction is enabled by fluorine effects and involves an alkyl nickel chain-walking mechanism.


Introduction
Cross-electrophile coupling has found wide application in the construction of C-C bonds and serves as a powerful and reliable alternative to the classical nucleophile/electrophile procedures. 1 In particular, C(sp 2 )-C(sp 3 ) cross-coupling has been thoroughly explored using the combination of aryl/alkenyl and alkyl electrophiles, wherein b-hydride elimination is inhibited to minimize olen by-products. 2 Conversely, transition-metalcatalyzed remote functionalization exploits iterative b-hydride elimination and metal-H insertion, namely the "chain-walking" process, allowing introduction of functionality into an otherwise unreactive aliphatic position. 3 Recently, signicant progress in this eld has been achieved by the groups of Martin, 4 Zhu, 5 and Yin. 6 Remote arylation and carboxylation of unactivated alkyl halides have been well established (Scheme 1a). However, the reported reactions are still restricted to a relatively limited number of coupling partners, thus limiting the resulting molecular diversity of the products. A notable limitation remains: the corresponding alkenylation of unactivated alkyl halides has not been reported yet, although various alkenes are frequently found in pharmaceuticals or other functional molecules. The reason for the lack of a precedent probably lies in the potential addition of disassociated metal-H to the olenic coupling component and even the alkene product during the chain-walking process. 5,7 In this regard, the development of efficient and practical routes for metal-hydride-involving remote alkenylation has remained as a formidable challenge.
To circumvent the problem of promiscuous hydrometallation, an alkenyl coupling partner inherently inert to metal-H would be required in a successful migratory alkenylation. Recently, gem-diuoroalkenes have been demonstrated by Cao, Toste, Fu, our group, and others as efficient uoroalkenylating reagents to access a wide range of mono-uoroalkenes using radical or ionic manifolds. 8 The strong electron-withdrawing nature of the two uorine atoms renders gem-diuoroalkenes highly reactive toward nucleophilic attack on the uorinated sp 2 -carbon, and uoro-alkenylation could be accomplished through facile b-uoride elimination. Furthermore, we considered that the uorine-based polarization and pp interaction would make the uorinated olens less favored in the hydrometallation compared with the non-uorinated alkene intermediates and therefore amenable in the migratory Scheme 1 Migratory functionalization of unactivated alkyl halides. alkenylation reactions. 9 It could then be envisaged that gem-diuoroalkene 2 can coordinate to thermodynamically stable benzyl nickel V (generated through oxidative addition of the C-Br bond to the Ni 0 and single-electron reduction followed by chain-walking) to form intermediate VI, which undergoes regioselective migratory insertion to produce a carbonickelation adduct (VII). 5 Subsequently, b-uoride elimination could lead to the formation of the monouoroalkene 3. 10 Reduction of Ni-F (VIII) could then regenerate the catalytically active Ni 0 species (I) (Scheme 2). However, we anticipated that a few challenges would have to be overcome to realize this strategy: (1) a nickel catalyst must be able to mediate all the elementary steps including chain-walking and deuorinative coupling; (2) the resulting monouoroalkene product should be unreactive towards the Ni-H as well as Ni-alkyl species; (3) the migration of Ni-H along the carbon chain should be much faster than the alkenylation step to ensure good regioselectivity. In this report, we disclose such a convenient synthetic route for the migratory uoro-alkenylation of unactivated alkyl bromides with high regio-and stereocontrol (Scheme 1b). Remarkably, by virtue of this operationally simple methodology a variety of structurally privileged monouoroalkenes 11 could be obtained from easily available materials.

Results and discussion
To test our hypothesis, we selected gem-diuoroalkene 2a as an alkenylating reagent to react with 1-bromo-2-phenylethane 1a (Table 1). Aer careful evaluation of the reaction parameters, we found that a combination of inexpensive and bench-stable Ni(ClO 4 ) 2 $6H 2 O as a pre-catalyst, 6,6 0 -dimethyl-2,2 0 -bipyridyl (L1) as a ligand, and Mn as a reducing agent to generate Ni 0 in DMA at 25 C gave the benzylic uoro-alkenylation product 3a with Z-conguration in 28% NMR yield within 12 hours (entry 1). 12 Pleasingly, the product was obtained with excellent regioselectivity and stereoselectivity. Furthermore, it was found that introduction of MgCl 2 as an additive 13 gave an improved yield (44%), thus indicating that the Lewis acid has a positive effect on the reaction yield. Encouraged by this result, extensive screening of Lewis acid additives was conducted (Table S1 †), and Yb(OTf) 3 led to a favorable result (70% yield). The application of YbCl 3 afforded 3a in comparable yield (63%), reecting the importance of the ytterbium cation (entry 4). On the other hand, changing the reducing agents to Zn, B 2 pin 2 or HCOONa had a deleterious effect on the outcome of the reaction (Table S1 †). Moreover, Ni(ClO 4 ) 2 $6H 2 O proved to be the optimal catalyst aer examination of various nickel salts (Table S1 †). Subsequently, the ligand was further optimized. Increasing the steric prole of the substituent at the ortho position to the nitrogen in the bipyridyl scaffold (L2) hampered the reactivity of this alkenylation reaction, and no product was detected when the reaction was treated with bipyridyl (L3) as a ligand, suggesting the essential role of such substituents (entries 5 and 6). The structural analogue neocuproine (L4) also exhibited high catalytic efficiency while bathocuproine (L5) was less effective (entries 7 and 8). Slightly higher yields of 3a were obtained when the reaction was treated with fewer equivalents of Mn (entry 9). Of particular note, the process is readily scaled up, the reaction of 1a (2.5 mmol) with 2a (1.0 mmol) gave 3a in 70% yield (entry 10).
By using the optimized reaction conditions, the scope of this Ni-catalyzed migratory uoro-alkenylation of unactivated alkyl bromide with gem-diuoroalkenes was evaluated. As shown in Scheme 3, the present protocol shows a remarkably broad scope Scheme 2 Reaction design. with respect to the gem-diuoroalkene coupling partner. 1-Aryl-2-bromoethane was initially coupled with various aryl-gem-diuoroalkene derivatives to afford the uoro-alkenylation products (3a-r). The method tolerated a variety of substitution patterns on the phenyl group. In general, the reaction displayed a noticeable preference for substrates bearing electron-withdrawing groups, such as cyano, triuoromethyl, ester and ketone, which is in sharp contrast to radical-associated alkylation wherein electron-rich gem-diuoroalkenes were favored. 8i For example, para-substituted phenyl-gem-diuoroalkenes were converted into the corresponding products (3a, 3d, 3g, 3j and 3k), mostly in moderate to good yields. The use of metasubstituted phenyl-gem-diuoroalkenes was also explored, which led to moderate yields of products (3b, 3e and 3h). Notably, sterically demanding starting materials that bear orthosubstituents on the phenyl ring were amenable under the optimized reaction conditions; however, the yields of the desired products were reduced (3c, 3f and 3i). Chlorine and uorine substituents were also compatible in the present reaction (3l and 3m). Moreover, aryl-gem-diuoroalkenes with electron-donating substitution, including OMe, OTs and even the protic amide NHAc, were employed, and these transformations took place smoothly, leading to the compounds 3n-3p with moderate yields. Additionally, nitrogen-containing heterocycle derived substrates could also be converted by the catalytic system (3q and 3r). We next aimed to extend the scope of alkyl bromides. A range of b-bromoethylarenes having Me, OMe, OTBS, Cl and CF 3 groups on the phenyl ring were subjected to the alkenylation reaction with an acetyl or a methoxylcarbonyl phenyl-gem-diuoroalkene (2a and 2j). The products 3s-3z were obtained in useful yields with excellent regioselectivity and stereoselectivity. It is worth mentioning that an unprotected hydroxyl group was also tolerated, despite affording the product in decreased yield. To further expand the scope of the alkyl bromides, long range alkenylation was then examined. Alkenylation occurred favorably at the benzylic position, providing the corresponding uoro-alkenylated propane, butane and even pentane derivatives (3aa-ac). The yields decreased progressively as the carbon chain increased which might be attributed to the increased bulkiness of the benzylnickel intermediate. The yields of other related regioisomers were only slightly increased according to 19 F NMR analysis of the crude product. These results suggest that formation of the benzyl-Ni species is faster compared to the alkenylation step. Then the more challenging secondary alkyl bromide was subjected to an alkenylation reaction with 2j, and proved to be competent in this reaction, providing the expected benzylic uoro-alkenylation product with good regioselectivity. The relatively lower yield of 3aa from the secondary alkyl bromide (35% and 30% from (2-bromobutyl)benzene and (3-bromobutyl) benzene, respectively) than the primary one (40% from (4-bromobutyl)benzene) may indicate that the oxidative addition of the alkyl bromide to nickel is sensitive to the steric bulkiness, thus attenuating the reaction efficiency.
The synthetic utility of the uoro-alkenylation product was exemplied by further transformations of 3a (Scheme 4). Hydrogenation of the uoroalkene moiety contained within 3a was carried out (H 2 , Pd/C), yielding 4a. An epoxidation reaction using 3a also proceeded well to give uoroepoxide 4b with high yield (88%). In addition, dibromination of the C]C double bond with bromine was executed and the addition product 4c was formed uneventfully. Moreover, treatment of 3a with a base to eliminate HF furnished the synthetically useful trisubstituted Scheme 3 Substrate scope. See the ESI † for experimental details. Isolated yields are indicated. The regioisomeric ratio (rr, the ratio of the benzylic fluoro-alkenylation product to the other regioisomers) was determined by 19 F NMR analysis of the crude product. PMP ¼ pmethoxyphenyl. Ts ¼ tosyl. TBS ¼ tert-butyldimethylsilyl.
allene 4d in 70% yield. This sequence allows gem-diuoroalkenes to serve as a vinylidene source, and thereby an expedient migratory vinylidenation was achieved. Finally, we found that 3a could be converted into 1,2-diketone 4e by oxidation.
To gain preliminary insight into the unique uorine effects 14 that enable the migratory uoro-alkenylation reaction, several control experiments were then performed. First, we carried out the alkenylation reaction with a series of halogenated alkenes (Scheme 5a-c). As shown in Scheme 5a, gem-chloroalkene 5a and gem-dibromoalkene 5b did not lead to the desired haloalkenylation product. Second, we examined monohaloalkenes under the standard conditions. Not surprisingly, in these experiments no alkenylation product was detected again upon the consumption of the starting materials 7a-c (Scheme 5b). 15 The comparison with various halogenated alkenes highlights the prominent role of the two uorine atoms which entail the unique reactivity in migratory uoro-alkenylation. Regarding the C(sp 2 )-C(sp 3 ) bond formation, an alternative pathway involving oxidative addition of the C(sp 2 )-F bond to the benzyl-Ni followed by reductive elimination is also possible. 8c,e However, the 1-bromo-1-uoroalkene 9 failed to produce the uoro-alkenylation product which could tentatively rule out this mechanism (Scheme 5c). 16 The use of deuterium labelled alkyl bromide D 2 -1a gave rise to the deuterium-shi product D 2 -3a exclusively, which strongly supports idea that a process involving b-hydride elimination and reinsertion is operative in the present transformation (Scheme 5d). Note that no signicant further hydrogen/deuterium scrambling was found in D 2 -3a, revealing the thermodynamic preference of the benzylic-Ni intermediate in the migration process, which intrinsically dictates the regioselectivity of this transformation. Given the Nicatalyzed chain-walking process does not require a Lewis acid additive, 4-6 it is reasonable to attribute the role of Yb(OTf) 3 in activating gem-diuoroalkene towards the nucleophilic addition 17 or facilitating the reduction of Ni-F species. 18 To obtain additional insight into the inuence of Yb(OTf) 3 , reactions with a stoichiometric amount of Ni(ClO 4 ) 2 $6H 2 O/L1 were carried out (see the ESI † for details). In the reaction with Yb(OTf) 3 , the desired product (3a) was attained in 55% yield which is much higher than the ytterbium-free reaction (16%), implying that the Lewis acid should take effect in the nucleophilic addition step rather than the Ni-F reduction. To evaluate the ability of Ni-H to recognize the non-uorinated alkenes and gem-diuoroalkenes, styrene 10 was subjected to the uoro-alkenylation reaction with 1-bromopropane as the hydride source (Scheme 5e). 5,19 With the same catalytic system, the desired product 3a was formed in 40% NMR yield. This result clearly demonstrates the signicant impact of uorine substituents on the alkene moiety, which differentiate the two kinds of alkenes towards the Ni-H species as proposed in Scheme 1b.

Conclusions
In conclusion, a general, Ni-catalyzed migratory uoroalkenylation of unactivated alkyl bromides with gem-diuoroalkenes has been developed, providing ready access to diversely functionalized monouoroalkenes, which are valuable molecules in biological and materials science. More importantly, this work extends the boundaries of the highly attractive eld of remote functionalization of unactivated alkyl electrophiles since it represents the rst instance in which an alkene coupling partner has been used in a NiH-mediated process. It is also noteworthy that this C(sp 2 )-C(sp 3 ) bondforming reaction proceeds with excellent regio-and stereoselectivity under nonbasic conditions at room temperature, and an array of potentially reactive functional groups are tolerated. This journal is © The Royal Society of Chemistry 2019

Conflicts of interest
There are no conicts to declare.