To test our proposal, we reacted (3-bromopropyl)benzene (1a) with methyl acrylate (2a) and 1-((trimethylsilyl)methyl)piperidine (3a) in the presence of catalytic amounts of NiBr2 and Ir(ppy)2(bpy)PF6 with 4,4’-di(CO2Me)bpy (L1) as ligand in CH3CN. After 24 hours irradiation with 1.5 W blue LED, we could only detect 6% yield of desired product 4. Besides, a large amount of side product 5 arising from homocoupling of 1a and trace amount of two-component cross-coupling product 6 were detected (Table 1, entry 1). Other bipyridyl ligand such as L2, L3 or L4 could only give a low yield less than 20% (entries 2-4). To our delight, terpyridyl ligand L5 largely promoted the reactivity and the yield of 4 could be increased up to 58% (entry 5). Finally, 4’-phenyl-substituted terpyridine L6 gives the best result (entry 6). A slightly increased yield was further observed when adding 12 mol% of L6 (entry 7). The reaction displayed an excellent regioselectivity without any other regioisomers detected. Control experiments revealed that nickel, photocatalyst, and visible light irradiation are indispensable for the dialkylation process (entries 8-10). Of note, in the absence of nickel catalyst, Giese-adduct of 2a and 3a via the in-situ formed α-amino radical was mainly detected (entry 10). Other parameters including photocatalysts, nickel sources and solvents also have significant impacts on the reactivity and the outcome distribution (see SI for details). The results showed that Ir(ppy)2(bpy)PF6 is better than other Ir-based photocatalysts. Replacing NiBr2 to other Ni(II) sources led to lower yields. Other solvents including DMF, DMAc, NMP, and THF all resulted in a largely decreased yield of the desired product.
Table 1. Nickel/photoredox dual-catalyzed alkyl-aminoalkylation of alkenes: conditions optimization[a]
[a] Without otherwise noted, all reactions were conducted on a 0.1 mmol scale with 1a:2a:3a = 1:2:2, Ir(ppy)2(bpy)PF6 (1 mmol%), NiBr2 (10 mmol%), L (10 mmol%), CH3CN (0.1 M) at room temperature under 1.5 W blue LED irradiation for 24 h. [b] Corrected GC yield with n-decane as the internal standard. [c] L6 (12 mmol%), and the molar ratio of 1a:2a:3a = 1:2:1.5. [d] Without photocatalyst. [e] Without light. [f] Without Ni-catalyst.
Having established optimal reaction conditions, we turned our attention to investigate the substrate scope. First, a wide range of unactivated primary alkyl bromides were surveyed and all delivered desired products in moderate to good yields with an excellent regioselectivity (Figure 2). Alkyl bromides bearing various functional groups at a distal position such as OTBS, F, Cl, thioether, CN, ester, alkene, ether, OAc, boronate, sulfamide, and sulfone were well tolerated (9, 10, 13-22, 29, 30), which provides a further synthetic platform for derivatization. Moreover, a variety of heterocycles such as furan (24), benzo[d][1,3]dioxole (25), thiophene (26), and benzo[d]thiazole (27) were incorporated smoothly. α-branched alkyl bromides (31, 32) were also valid substrates. Interestingly, as opposite to literature report,[28] our reaction with more reactive alkyl bromides such as BrCH2CN, BrCH2COOEt, or BnBr cannot afford the desired outcome (see SI for unsuccessful substrates). Instead, a huge amount of homocoupling product of alkyl bromide was detected in each case. Suggesting that the performance of reactivity-matched alkyl halides to make a balance between nickel catalytic cycle and photoredox cycle is key for success of this dual catalytic dialkylation reaction.
Next, secondary and tertiary alkyl bromides were examined (Figure 3). A series of acyclic (33, 34) and cyclic secondary alkyl bromides including carbocycles (35-37) heterocycles (38, 39) and bridged system (40) were reacted with good yields. Regarding tertiary alkyl bromides, both acyclic and cyclic tertiary alkyl bromides containing ester, ether, NTs group were successfully converted to the desired regioisomerically pure products (41-51). The reaction can scale up to 3 mmol without an obviously diminished yield (57% vs 64% for compound 41). The performance of sterically hindered alkenes has proven to be a fundamental challenge in alkene difunctionalization. Gratifyingly, our protocol can tolerate 1,1-disubstituted alkenes[40] and allow for generating complex structures bearing two skipped (1,3) all-carbon quaternary centers (46-51). However, 1,2-disubstituted alkenes and styrenes are reluctant to the reaction (see SI for unsuccessful substrates). It should be noted that our protocol for alkene dialkylation is simultaneously compatible with primary, secondary and tertiary alkyl halides, which have not yet been realized in literatures. Finally, alkyl bromides derived from naturally occurring products or drug intermediates (52-55) were successfully incorporated.
Furthermore, we examined the scope generality with respect to α-silylamines (Figure 4). Various pharmaceutically privileged cyclic alkyl amines including morpholine (56), thiomorpholine (57), piperidines (58-61), tetrahydroisoquinoline (62), piperazine (63, 64) were successfully introduced into target molecules. The structure of the products was confirmed unambiguously by X-ray diffraction analysis of compound 57 and 69. Of particularly attractive, this protocol enables a simultaneous introduction of two medicinally relevant cyclic amine moieties into products to prepare a series of β-amino acid derivatives with otherwise inaccessible structural features. When using a α-branched silylamine, the reaction could enable access to α,β-disubstituted β-amino-acid derivative bearing two contiguous tertiary stereocenters in a 5:1 diastereoselectivity (65). The relative stereochemistry for the major diastereomer was assigned as shown by 1H NMR analysis. Finally, we tried to extend the alkene scope to other Giese acceptors. Various N-aryl or N-alkyl substituted acrylamides even with a free NH group reacted smoothly (69-72). Other alternative Giese acceptors include acrylonitrile (73) and enone (74).
The reaction also displays a good orthogonal reactivity toward alkyl and aryl halides. Substrate bearing both an alkyl halide and aryl halide reacted chemoselectively at alkyl site to afford β-amino acid ester 75 under the standard condition. While the three-component aryl-aminoalkylation proceeded site-specifically with aryl iodide using our previous condition[41] to afford γ-amino acid ester 76 with alkyl bromide functionality untouched (Figure 5-1). Of particularly note, the two reactions display an opposite regioselectivity as the α-aminoalkyl group added at a different position of alkene. Moreover, competitive experiments between 1o/2o, 2o/3o, and 1o/3o alkyl bromides show that the reactivity order of alkyl halides is 3o>2o>1o, which is consistent with the order of stability of free alky radicals (Figure 5, eq. 2-4).
In order to gain insight into the reaction mechanism, several control experiments were conducted. First, The reaction was inhibited when TEMPO [(2,2,6,6-tetramethylpiperidin-1-yl)oxyl] was added (Figure 6-1). Radial clock experiment with (bromomethyl)cyclopropane afforded exclusive ring-opening product (Figure 6-2). Both results together indicated a radical pathway involved in the reaction. Control experiment with 1-methylpiperidine failed to deliver the outcome with or without a base, indicating the indispensable role of TMS group (Figure 6-3). Moreover, Stern-Volmer fluorescence quenching experiments (see SI for details) revealed that the reductive quenching of the excited state of photocatalyst by α-silylamine is more efficient over other reaction components (Figure 6-4). In-situ ESI-MS experiment revealed that the iminium ion species 1-methylenepiperidin-1-ium was detected by mass spectroscopy analysis (Figure 6-5).
Based on the experimental results and our previous reports[41,42], two plausible pathways were proposed (Figure 7). Reductive quenching of *IrIII (E1/2*III/II = +0.66 V vs SCE)[43] by α-silylamine (E1/2ox = +0.4~+0.8 V vs SCE)[44-45] generates an α-amino radical, which could be easily further oxidized to iminium ion species owing to its much lower oxidative potential (E1/2ox = -1.03 V)[46]. The formation of iminium ion was confirmed by In-situ ESI-MS analysis. On the other hand, Ni0 activates alkyl bromide via SET to generate a more reactive alkyl radical and NiIX species. Addition of the alkyl radical to the olefin affords a stabilized α-carbonyl radical, which combines with the NiIX intermediate to form NiIIX. α-Amino radical recombination with NiIIX forms NiIIIX, followed by reductive elimination affords the desired product. Alternatively, Mannich-type reaction of Ni-enolate with iminium ion could also deliver the product. Final SET between NiI [Ered(NiI/Ni0) = -1.17 V vs SCE][47] and the reduced state of PC (E1/2III/II = -1.51 V vs SCE)[43] closed both catalytic cycles.