Introduction

Wacker oxidation, that is, the reaction of Pd-catalyzed oxidation of alkenes into high value-added and synthetically versatile carbonyls, is a central transformation in chemistry1,2,3,4 (Fig. 1a). This classical transformation is well established for oxidation of terminal olefins. However, in this reaction internal olefins are relatively unactivated without reliance upon suitable coordinating groups5,6,7,8,9. Moreover, the internal olefins, particularly unbiased internal olefins, commonly provide inseparable regioisomers with comparable yields5,6,7,8,9. Actually, the regioselectivity issue of oxidation of internal olefins represents another longstanding challenge in Wacker chemistry1,2,3,4,5,6,7,8,9. Consequently, these challenges inherently undermine the utility of Wacker chemistry because the vast majority of alkenes are unbiased internal olefins readily accessible from petroleum and renewable resources such as seed oils10 and through well-established synthetic routes such as carbonyl olefination11 and olefin metathesis12.

Fig. 1
figure 1

Nickel-catalyzed Wacker-type oxidation at remote sp3 C–H sites. a Classical Wacker oxidation vs. remote Wacker-type oxidation. b Examples of biologically active organic molecules that contain an aromatic ketone motif. c Mechanistic rationale for the nickel-catalyzed Wacker-type oxidation at remote sp3 C–H sites

Full size image

Remote functionalization that allows direct bond formation at a distal and specific position other than the initial reactive site is a significant challenge13,14,15,16,17,18,19,20,21, particularly in remote Wacker-type oxidation. This is because the remote Wacker-type oxidation would involve transition-metal hydride addition of an alkene into the organometallic intermediate transition-metal alkyl, followed by a sequential β-hydride elimination/migratory-insertion iteration process reaching a specific remote sp3 C–H position where oxidation occurs to liberate the desired single carbonyl product13,14; unfortunately, both the reductive transition-metal hydride and the unstable organometallic intermediate are commonly sensitive to oxidative conditions. Moreover, alkenes are more reactive than alkanes owing to the exposed π-bonding electrons, leading to oxidation of carbon–carbon double bonds in preference to the remote sp3 C–H. Indeed, there were few clues in the literature implying the exceptional difficulty for the development of this protocol22,23,24,25,26,27,28,29,30,31,32,33. For example, isomerization of olefinic alcohols enabled by transition-metal hydride species leads to the ketone products, which requires migration of the olefinic unsaturations toward the tethered alcohols22,23,24,25,26,27,28,29. Alternatively, internal olefins isomerization/hydroborations or isomerization/hydrosilylations followed by oxidations can realize remote oxidations30,31,32,33,34. While these methods are efficient, they all require prefunctionalized starting materials. Apparently, straightforward oxidation of olefins’ remote sp3 C–H to valuable ketones serves as a more practical approach (For an example of Pd-catalyzed tandem isomerization–Wacker oxidation of allyl arenes, see ref. 35.).

We previously identified an iron catalysis system enabling highly efficient Wacker-type oxidation36. This oxidation transformation involves iron hydride in situ generated from iron(II) and hydrosilane to react with an alkene producing the adduct alkyl iron intermediate that gives carbon-centered radical via single-electron transfer37,38, followed by generation of the iron peroxide complex in the presence of oxygen, ultimately to release the desired carbonyl product36. This work led us to attempt remote Wacker-type oxidation with an analogous process. Apparently, this new strategy would require the discovery of a catalyst enabling migration of an olefinic unsaturation toward a specific position where oxidation occurs. Encouragement in this regard was found in the work of inexpensive nickel-catalyzed remote sp3 C–H functionalizations through a chain-walking mechanism via iterative β-hydride elimination/nickel-hydride species migratory insertion yielding a nickel alkyl sequences31,32,39,40,41,42,43,44. Inspired by these investigations, we questioned whether we could utilize a nickel catalyst for the challenging remote oxidation of unactivated sp3 C–H sites in the presence of a priori more reactive carbon–carbon double bonds.

On the basis of the above-mentioned results, we hypothesize that the remote Wacker-type oxidation proceeds through a single electron transfer (SET) process. To achieve requisite selectivity and avoid the formation of regioisomeric oxidation products, this remote Wacker-type oxidation process should preferably have sufficient kinetic and thermodynamic favorability. In this context, aryl-substituted olefins selected as substrates would be the optimal choice because they can generate stable benzyl radical intermediates and give the corresponding conjugated aryl ketone products. Notably, aryl-substituted olefins that are widely present in readily available and renewable plants of the family Anacardiaceae and Ginkgo biloba fruits10,45 undergo such a transformation to deliver aryl ketones, a privileged scaffold in medicinal chemistry and pharmaceutical agents (Fig. 1b).

Here we develop an alternative disconnection, in which internal olefins, including unbiased internal olefins, can be easily converted to single regioisomeric ketones with high activity in the absence of coordinating groups through remote Wacker-type oxidation. This method is highlighted by the successful employment of ideal ambient air as the sole oxidant, natural abundance of nickel as the catalyst, and a byproduct of the silicone industry, polymethylhydrosiloxane (PMHS) as the hydride source under room temperature allowing for effective oxidation of challenging alkenes with excellent selectivity.

Results

Design principle

Details of our design principle are outlined in Fig. 1c. Nickel hydride I, formed in situ from the reaction of L[NiII]X with hydrosilane, undergoes hydrometalation of olefin to provide alkyl nickel intermediate II40,41,42,43, followed by β-hydride elimination to give a new metal-bonded hydride with a concomitant migration of the double bond III. Subsequently, re-addition of the metal hydride generates a new alkyl–Ni species IV. Iterative the β-hydride elimination/migratory insertion sequences achieve a chain-walking process, delivering a critical benzyl nickel intermediate V39,40,41,42,43. Then the reaction of V with oxygen would form a thermodynamically favored benzyl radical VII, followed by generation of a nickel peroxide complex VIII46,47,48,49,50. Finally, VIII would undergo homolytic cleavage of O–O and C–H bond breaking to liberate the thermodynamically more stable product aryl ketone 2, while providing the LNiII-OH species, which could regenerate the nickel hydride I catalyst by hydride transfer in the presence of a hydrosilane50,51.

To probe this potential, our initial studies tested the isomerization–Wacker-type oxidation of 4-allylanisole (1a) with NiCl2/L1 (neocuproine) catalyst system using PhSiD3 as hydrosilane and EtOH as solvent and was found to allow the formation of desired aryl ketone 2a-d bearing similar degrees of D-incorporation at all positions of its hydrocarbon chain (Fig. 2a), suggesting that the process proceeds via Ni-D hydrometalation and subsequent β-hydride elimination/migratory insertion sequences and is bidirectional39,40,41,42,43. Furthermore, when olefin 1p was subjected to the reaction conditions under N2, a mixture of olefins originating from olefin isomerization was observed by gas chromatographic–mass spectrometric (MS) analysis and subjected to the oxidation reaction conditions to produce only one regioconvergent aryl ketone 2p (Fig. 2b), indicating that olefin isomerization is independent of an oxidant and chain walking precedes oxidation event. In the following oxidation section, a benzyl nickel intermediate undergoing SET oxidation to form a benzyl radical is supported by a radical capture experiment where addition of a radical inhibitor Galvinoxy to the reaction mixture of 1a resulted in total inhibition of the oxidation transformation and intercepted the corresponding benzyl radical to provide 2a’ based on high-resolution MS (electrospray ionization) analysis (Fig. 2c). Collectively, these results demonstrate the feasibility of the elementary steps in our design principle.

Fig. 2
figure 2

Mechanistic studies. a Isotopic labeling experiment. b Intermediates analysis and oxidation of olefin isomers. The ratio of the mixture of alkenes was determined by GC/MS. c Radical-trapping experiment

Full size image

Encouraged by the above results, we systematically evaluated various parameters with exposure of 1a to a nickel species in the presence of hydrosilane as a hydride source under aerobic conditions at room temperature and found that optimal conditions were achieved by using a combination of NiBr2 (5.0 mol%) and bench-stable L1 (6.0 mol%) as the catalyst and PMHS (3.0 equiv) as the hydride source in EtOH under ambient air to yield the desired aryl ketone 2a in 80% yield as a single regioisomer (see Supplementary Table 1). Notably, compared with L1, in 1,10-phenanthroline L2 lack of encumbered two methyl groups at C2 and C9 led to no desired oxidation product. Additionally, in line with our design principle, nickel catalyst, PMHS, and air are all essential for oxidation. Without any one of the three elements, no oxidation was possible.

Substrate scope

With the optimized conditions established, the scope of the nickel-catalyzed remote and proximal Wacker-Type oxidation was investigated (Fig. 3). Allyl benzenes bearing electron-donating or/and electron-withdrawing groups could be oxidized efficiently to afford the corresponding aryl ketones 2a–2m in good-to-excellent yields. The reaction of 1a could be run on a 5 mmol scale to give a similar yield (78%). Although steric hindrance on the aryl ring had a deleterious effect on the yield, a synthetic useful yield could also be obtained (2c). Particularly noteworthy is the tolerance of bromo, chloro, fluoro, and iodo groups (2e–2h and 2m), which can be used as prefunctionalities for further transformations. Additionally, methoxy, trifluoromethyl, cyano, and trimethylsilyl substitutes and esters were unaffected under the normal reaction conditions. Gratifyingly, the transformation can move beyond the simple aryl group. For instance, allyl-substituted naphthalene and indole were viable reagents in this reaction to build-up aryl ketones (2n–2o). Compared with allylbenzene (1d), homoallylbenzene (1p) could achieve an identical result in a longer reaction time (6 h). Even further increasingly distal functionality, such as a thiophene ring three methylene and a phenyl ring four methylene units away (1r–1s), provided the desired products in 82% and 88% yields, respectively. Notably, internal olefins such as 1t–1v, 1s’, 1s”, and 1r’ are not successfully oxidized through conventional Wacker-type reaction but are suitable substrates for oxidation via the present nickel catalysis. A variety of styrenes could also undergo direct oxidation under the reaction conditions to provide the corresponding products 2w–2y, 2a, and 2t in good-to-excellent yields.

Fig. 3
figure 3

Substrate scope for the nickel-catalyzed Wacker-type oxidation. Reported yields are for the isolated products; asterisk “*” denotes the initial position (right side) of the unsaturated carbon atom of a red double bond

Full size image

A particularly noteworthy aspect of this protocol is its amenability to late-stage synthetic applications (Fig. 4). For instance, remote Wacker-Type oxidation of a Vitamin E succinate derivative proceeded smoothly to give the desired product 4a in 82% yield. Eugenol peracetyl-glucoside 3b having activity against the fungi52 was also oxidized in 82% yield (4b). In addition, the steroidal substrates 3b–3f synthesized from estrone, epiandrosterone, lithocholic acid, and glycyrrhetinic acid, respectively, were subjected to the oxidation procedure, affording the desired products in 62–86% yields (4b–4f). Notably, oxidatively labile functional groups such as a hydroxy group and a carbon–carbon double bond conjugated with a highly electron-withdrawing substitute remained intact (4e–4f). Furthermore, cashew nutshell liquid, a raw material, is directly obtained from the shell of the cashew nut and contains cardol, cardanol, and anacardic acid. It undergoes decarboxylation, vacuum distillation, hydrogenation, and methylation to deliver 3g53. With our protocol, 3g could be transformed to 4g in 65% yield with ideal selectivity (Fig. 4b).

Fig. 4
figure 4

Late-stage oxidation of complex molecules. Reported yields are for the isolated products. a Complex medicinally relevant molecules as the substrates. b Cashew nutshell liquid (CNSL) as the substrate; the yellow shape is analogous to a drop of cashew nutshell liquid

Full size image

Ultimately, a major benefit of this mild, nickel-catalyzed aerobic oxidation is its viability to demonstrate regioconvergent oxidation of mixtures of alkenes to produce single-regioisomer aryl ketones. To demonstrate this potential, an equimolecular mixture of three olefin regioisomers was subjected to the standard reaction conditions to selectively generate a single product 2s in 88% yield (Fig. 5a). This finding showcased the potential of the catalytic platform for conversion of cheap and abundant petroleum-derived alkenes that are often mixtures of regioisomers to value-added products containing single regioisomers. In addition, elimination reactions frequently result in a mixture of at least two olefin regioisomers that are difficult to separate. For example, aliphatic tosylate 5a underwent elimination in alcoholic MeONa solution54 to give a mixture of two regioisomers 1p’ and 1p” (molar ratio, 4.1:1) that could be directly transformed to a single-regioisomer aryl ketone 2p in 78% yield by using our protocol (Fig. 5b). Similarly, alcohol 5b dehydration55 generated a mixture of olefins 1s’ and 1s” with a 1.5:1 molar ration, which was subjected to the reaction conditions and produced only one regioconvergent oxidation product 2s in 62% yield (Fig. 5c). Neither purification nor isolation of the intermediate olefinic isomers was necessary, showing the robustness of our method.

Fig. 5
figure 5

Regioconvergent Ni-catalyzed remote Wacker-type oxidation of mixtures of alkenes. a Regioconvergent oxidation of mixtures of alkenes to produce a single-regioisomer aryl ketone. b Regioconvergent oxidation of a mixture of alkenes from an elimination reaction of an alkyl tosylate. c Regioconvergent oxidation of a mixture of alkenes from dehydration of an alcohol. Reported yields are for the isolated products. The ratios of mixtures of alkenes were determined by GC/MS

Full size image

Discussion

In summary, we have developed a conceptually new approach to selective oxidation of remote sp3 C−H bonds to produce ketones. The transformation proceeds through the nickel-catalyzed alkene isomerization to transfer the unit of unsaturation to the most thermodynamically favored position in the molecule and subsequent oxidation using ambient air as the sole oxidant at room temperature. The mild, expeditious, and operationally simple protocol allows efficient remote oxidation of terminal olefins and unactivated internal olefins with excellent functional-group tolerance and regio- and chemo-selectivity. Even unrefined mixtures of olefins can undergo regioconvergent oxidation to selectively produce a single product, which is not currently accessible. Notably, the protocol is particularly useful for late-stage oxidation of structurally complex substrates, which offers a unique approach to conceptualize the remote oxidation disconnections in organic synthesis.

Methods

General procedure

A 25 mL flask was charged with NiBr2 (2.8 mg, 0.0125 mmol), neocuproine (3.2 mg, 0.0150 mmol), olefin (0.25 mmol), EtOH (2 mL), and PMHS (170 μL, 0.75 mmol). The reaction mixture was stirred in an open air atmosphere at room temperature until the reaction was complete (observed by thin layer chromatography). The resulting reaction solution was directly purified by column chromatography (petroleum ether/ethyl acetate) on silica gel to afford the corresponding product.

Synthesis and characterization

See Supplementary Methods for general information about chemicals and analytical methods, synthetic procedures, 5 mmol-scale synthesis, regioconvergent oxidation of mixtures of olefins (see Supplementary Figures 6–8), and characterization of products. For 1H and 13C nuclear magnetic resonance data, see Supplementary Figures 9–40.

Optimization

See Supplementary Table 1.

Mechanistic studies

See Supplementary Figures 1 and 2 (isotopic labeling experiment), Supplementary Figure 3 (intermediates analysis and oxidation of olefin isomers), and Supplementary Figure 4 and 5 (interception of radical intermediate).