Sequential C–O decarboxylative vinylation/C–H arylation of cyclic oxalates via a nickel-catalyzed multicomponent radical cascade

A selective, sequential C–O decarboxylative vinylation/C–H arylation of cyclic alcohol derivatives enabled by visible-light photoredox/nickel dual catalysis is described. This protocol utilizes a multicomponent radical cascade process, i.e. decarboxylative vinylation/1,5-HAT/aryl cross-coupling, to achieve efficient, site-selective dual-functionalization of saturated cyclic hydrocarbons in one single operation. This synergistic protocol provides straightforward access to sp3-enriched scaffolds and an alternative retrosynthetic disconnection to diversely functionalized saturated ring systems from the simple starting materials.


Introduction
In the last decade, the nickel-catalyzed cross-coupling reaction has emerged as a powerful technique to construct C-C bonds in chemical synthesis. 1 Notably, nickel-catalyzed multicomponent reactions, that allow for the formation of multiple C-C and/or C-heteroatom bonds thus enabling the streamlined synthesis of complex molecular scaffolds in a single operation, are particularly attractive due to their good compatibility and unique selectivity. 2 Signicant progress has been achieved in the area of catalytic multicomponent difunctionalization of unsaturated systems, enabling the simultaneous, one-pot installation of two functionalities over double bonds or triple bonds via nickel catalysis (Fig. 1). 3,4 In contrast, there has been a lack of reports in which saturated hydrocarbons have been manipulated for cross-coupling at more than one reaction site in one single operation. 5 This catalytic strategy would simultaneously install two sp 3 C-C bonds in saturated hydrocarbons, readily available and abundant building blocks in organic synthesis; nevertheless, the realization of such a strategy requires the ability to overcome the relative inertness of saturated hydrocarbons and to control the selectivity in the presence of multiple similar chemical bonds.
Alcohols have been widely employed as salient synthetic building blocks in chemical synthesis. Notable advances have been made in selective functionalization of sp 3 C-H bonds of alcohols through the hydrogen-atom-transfer (HAT) process. 6,7 Moreover, alcohols can function as latent alkylating agents through transition-metal-catalyzed ipso-C-O activation 8 or homolytic C-O cleavage. 4j,k,9,10 These catalytic approaches enable the facile construction of diverse structural motifs from abundant alcohols. Nevertheless, one single transformation that combines ipso-C-O functionalization and remote C-H functionalization of alcohols, which would provide an attractive platform for the synthesis of highly functionalized sp 3 -rich scaffolds, 11 has not been reported yet. We herein report a selective, cascade C-O bond vinylation/C-H bond arylation of alcohols that achieves site-selective dual carbofunctionalization of simple abundant aliphatic alcohols via a photoredox/nickel 12enabled multicomponent radical cascade process. This strategy relies on a synergistic combination of alkyl oxalate decarboxylation, 1,5-hydrogen atom transfer (1, of the vinyl radical, and C(sp 3 )-C(sp 2 ) cross-coupling. Our original design for this radical cascade strategy is outlined in Fig. 1. Specically, we envisioned that a radical addition of alkylalkyne would give rise to a s-type vinyl radical I, which would prefer to undergo an intramolecular 1,5-HAT process to yield a nucleophilic alkyl radical species II. 13 Subsequent nickel-mediated coupling of alkyl radial II with aryl halides would forge a C(sp 3 )-Ar bond, 12 and nally lead to a sequential two-site functionalization of oxalates, i.e. C-O decarboxylative vinylation & C-H arylation (Fig. 1). We are particularly interested in exploring cyclic alkyl oxalates to construct highly functionalized cycloalkanes, 14 due to two considerations: (i) this synergistic cascade process would undoubtedly expedite the synthesis of challenging aliphatic rings, and more importantly, would provide a novel retrosynthetic disconnection for complex saturated ring systems from simple starting materials; (ii) conformation of cyclic substrates might be benecial for the intramolecular 1,5-HAT migration of vinyl radical species. 15

Results and discussion
We began our investigations by employing tertiary cyclic oxalate 1 as a model substrate to test the possibilities (Table 1).
Pleasingly, we found that irradiation of a DMSO solution of 1, aliphatic alkyne 2, and 4-bromobenzonitrile 3 did afford the desired 1,3-vinylarylation product 4 in 84% yield in the presence of catalytic Ir[dF(CF 3 )ppy] 2 (dtbbpy)PF 6 , NiCl 2 (Py) 4 , 4,4 0 -di-tertbutyl-2,2 0 -dipyridyl (dtbbpy), and bis(4-methoxyphenyl) methanone (Table 1, entry 1). The countercation of oxalates impacted the reaction efficiency to a small extent, while the use of cesium oxalates afforded higher yields than the corresponding Li, Na, K salts (entries 2-4). Further evaluation indicated that both Ni(II) and Ni(0) catalysts were able to promote the desired transformation, with the precatalyst NiCl 2 (Py) 4 proving optimal (entries [5][6][7][8][9]. Control experiments demonstrated that the photocatalyst, the nickel catalyst, and visible light were all essential to this synergistic cascade process, as the desired 1,3-vinylarylation products were not observed in the absence of any of these components (entries [10][11][12]. Nevertheless, low conversion (27% yield) was still observed in the absence of the bipyridine ligand (entry 13). The use of bis(4methoxyphenyl)methanone as an additive proved to be slightly benecial to the reaction efficiency (entry 14). 13 With the optimal conditions in hand, we next turned our attention to evaluating the applicability of substrates as well as the potential limitations of this dual functionalization protocol. As shown in Scheme 1, a number of tertiary cyclic oxalates, readily prepared from the corresponding alcohols, could undergo the sequential C-O vinylation/C-H arylation with excellent site-selectivity, installing both vinyl and aryl functionalities onto the skeletons of alcohols under redox-neutral and mild conditions (products 4-14, 36-77% yields). A number of substituents including alkyl, ketal, ketone, and alkene on the cyclohexyl alcohols were tolerated in this dual catalytic cascade system, furnishing multi-substituted cyclic alkanes in a single operation under mild conditions (products 5-9, 36-77% yields). Nonetheless, steric hindrance of the substituents was found to have a considerable effect on the reaction efficiency: installing a methyl group at the aor gposition, or replacing methyl with ethyl at the ipso position of cyclohexyl oxalates resulted in decreased efficiency (products 6-8, 36-54% yields). Pleasingly, saturated O-, S-, and Nheterocyclic oxalates turned out to be viable substrates, yielding the vinyl/aryl-disubstituted saturated heterocycles with good efficiency (products 10-13, 50-76% yields). Moreover, cyclopentyl oxalates were also competent substrates, delivering the 1,3-difunctionalized cyclopentane products in synthetically useful yields (product 14, 40% yield).
Unfortunately, cyclic oxalates with larger or smaller ring sizes (e.g. 4-and 7-membered cyclic oxalates), bicyclic oxalates, and linear oxalates were unsuccessful substrates for this cascade protocol, probably due to the less favorable conformation (for unsuccessful oxalates, see Section 5, page S62 in the ESI †). Interestingly, the reaction of 4-methylenecyclohexyl oxalate with aryl bromides under optimal conditions afforded the 1,5-vinylarylation products in moderate yields (products 15-18, 58-64% yields). Excellent chemoselectivity was observed in this case, with no observations of 1,3-difunctionalized products. We reasoned that allyl-Ni species, generated via intramolecular 1,5-HAT followed by nickel trapping, underwent a selective coupling with aryl bromide at the terminal position, probably due to steric hindrance, to afford the cyclohexene product. This protocol represents a new and efficient platform to construct highly functionalized saturated heterocycles, important structural scaffolds for bioactive molecules, from simple starting materials. Next, we examined the scope with respect to the alkyne component (Scheme 1). Pleasingly, a wide range of terminal alkylalkynes could be efficiently employed in this cascade protocol, yielding the corresponding trans-alkenes in moderate to good yields (products 19-29, 45-74% yields). Notably, aliphatic alkynes tethered with complex molecules, exemplied by oxaprozin and dehydroabietic acid, also worked with moderate efficiency, further demonstrating the amenability of this synergistic strategy for late-stage manipulations (products 27 and 29, 53% and 50% yield, respectively). Nevertheless, internal alkynes proved to be inefficient substrates, most of which remained intact under the standard conditions. Finally, we explored the scope of aryl halides in this multicomponent transformation (Scheme 1). Aryl bromides containing electron-withdrawing substituents, including aldehydes, ketones, esters, nitriles, sulfones, and tri-uoromethylates, are competent coupling partners under optimal conditions, delivering the desired products with good efficiency (products 30-35, 53-70% yields). The mild conditions allow for the good tolerance of these important functionalities. This reaction is amenable to heteroaryl halides, selectively installing pyridines and benzothiazoles into cyclohexanes with moderate yields (products 36-38 and 41-42, 50-77% yields). Additionally, (hetero)aryl iodides/chlorides also participated in this sequential C-O/C-H dual functionalization process smoothly (products 39-42, 52-61% yields). Electron-rich aryl halides were applicable coupling partners, albeit with decreased efficiency (products 39-40).
To further demonstrate the synthetic application of our cascade protocol, we have performed several transformations by utilizing the alkene functionality (Scheme 2). The double bond of compound 41 readily underwent selective hydrogenation with H 2 in the presence of a Pd/C catalyst (product 43). Epoxidation of compound 30 with m-CPBA gave epoxide 44 in 65% yield. Furthermore, ozonolysis of 30 led to the formation of aldehyde 45, which could subsequently be oxidized to the corresponding carboxylic acid 46, or be reduced to the related alcohol 47 in good yield.
To shed some light on the potential reaction pathway, we have conducted several mechanistic experiments (Scheme 3). Reaction of ethynylcyclopropane with oxalate 1 and aryl bromide 3 gave the allene product 48, presumably generated via a ring-opening/cross-coupling process, indicating the involvement of vinyl radical species (Scheme 3A). Initially, we assumed that the 1,5-HAT process could be related to the subtle conformation of oxalates. We also found that the expected 1,5-HAT process is reliant on the nature of the vinyl radicals (s-type vs. ptype). For instance, competitive experiments between aryl and aliphatic alkynes showed that arylalkynes exhibited higher reactivity to afford the exclusive formation of 1,2-alkylarylation products, 4j and the 1,5-HAT/coupling product was not observed in this case (Scheme 3B). Regarding the coupling step, we Scheme 2 Synthetic manipulations of products. (i) Pd/C (10 mol%), H 2 (4 atm), THF, rt; (ii) m-CPBA (2 equiv.), CH 2 Cl 2 , rt; (iii) O 3 , CH 2 Cl 2 , À78 C; Me 2 S (10 equiv.), rt; (iv) NaClO 2 (5 equiv.), H 2 O 2 (4 equiv.), MeCN, 0 C to rt; (v) NaBH 4 (10 equiv.), MeOH, 0 C to rt. See the ESI † for experimental details. This journal is © The Royal Society of Chemistry 2020 Chem. Sci., 2020, 11, 4904-4910 | 4907 prepared the ligated Ar-Ni(II) complex 50, 16 and found that the stoichiometric reaction of the Ar-Ni(II) complex with alkyne and oxalate didn't form the desired 1,3-disubstituted cycloalkane product 51, suggesting that Ar-Ni(II) might not be a reactive intermediate for this cascade transformation (Scheme 3C). Nevertheless, the Ar-Ni(II) complex was able to catalyze the synergistic cascade reaction, giving the 1,3-disubstituted cyclohexane 33 in 30% yield in acetone (Scheme 3D).
On the basis of these experimental results as well as literature precedents, 12,17 a plausible mechanism for this photoredox/ nickel-catalyzed dual functionalization is depicted in Scheme 4. A thermodynamically feasible SET event between the photoexcited Ir-catalyst B and oxalate C would generate a tertiary alkyl radical D via decarboxylation, followed by a radical addition of alkyne to give rise to a s-type vinyl radical E. The resulting vinyl radical E would go through an intramolecular 1,5-HAT to selectively activate the sp 3 C-H of oxalates and to afford a nucleophilic, secondary alkyl radical species G. Subsequent interception of alkyl radical G by Ni(0) H would generate an alkyl-Ni(I) species I, which then undergoes an oxidative addition with aryl halide J to yield an (alkyl)(aryl)Ni(III) intermediate K. 17 This high-valent Ni(III) complex K would undergo a feasible reductive elimination to forge the C(sp 3 )-Ar bond and furnish the nal product L as well as Ni(I) species M. Finally, a SET event between Ni(I) M and the reduced Ir(II) F would regenerate Ni(0) H and the ground state Ir(III) A to close these two catalytic cycles. At this stage, we could not preclude another pathway that proceeds via oxidative addition of aryl bromide to Ni(0) followed by interception of Ar-Ni(II) N by alkyl radical species to afford the same (alkyl)(aryl)Ni(III) K.

Conclusions
In summary, we have developed a sequential C-O decarboxylative vinylation/C-H arylation of cyclic oxalates via photoredox/nickel dual catalysis. This synergistic protocol enables efficient and selective assembly of both vinyl and aryl functionalities onto saturated cyclic hydrocarbons in one single operation under mild and redox-neutral conditions, providing a new and complementary retrosynthetic method for densely functionalized saturated cyclic hydrocarbons. The mild conditions allow for excellent compatibility of functional groups and substrate scope in the oxalates, alkynes, and (hetero)aryl halides.

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