Asymmetric benzylic C(sp3)−H acylation via dual nickel and photoredox catalysis

Asymmetric C(sp3)−H functionalization is a persistent challenge in organic synthesis. Here, we report an asymmetric benzylic C−H acylation of alkylarenes employing carboxylic acids as acyl surrogates for the synthesis of α-aryl ketones via nickel and photoredox dual catalysis. This mild yet straightforward protocol transforms a diverse array of feedstock carboxylic acids and simple alkyl benzenes into highly valuable α-aryl ketones with high enantioselectivities. The utility of this method is showcased in the gram-scale synthesis and late-stage modification of medicinally relevant molecules. Mechanistic studies suggest a photocatalytically generated bromine radical can perform benzylic C−H cleavage to activate alkylarenes as nucleophilic coupling partners which can then engage in a nickel-catalyzed asymmetric acyl cross-coupling reaction. This bromine-radical-mediated C−H activation strategy can be also applied to the enantioselective coupling of alkylarenes with chloroformate for the synthesis of chiral α-aryl esters.

C hiral α-aryl ketones are versatile building blocks and represent important pharmacophores existing in many drug molecules such as ibuprofen and naproxen 1,2 . Although numerous enantioselective approaches for preparing quaternary α-aryl ketones have been reported [3][4][5] , asymmetric methods to access more commonly encountered tertiary variants remain limited presumably owing to the lability of tertiary stereocenters 6 . Nevertheless, transition-metal catalyzed asymmetric couplings of aryl organometallic reagents with α-bromo ketones [7][8][9] , benzylic zinc reagents with thioesters 10 , benzylic chlorides with acid chlorides under reductive conditons 11 , and aryl alkenes with activated carboxylic acids in the presence of a hydrosilane 12,13 have been disclosed in seminal studies by Fu, Maulide, Reisman, and Buchwald, respectively (Fig. 1a). Despite this impressive progress, it remains highly desirable to develop complementary methods that use feedstock functional groups to avoid sensitive organometallic reagents, preformed organohalides, and stoichiometric reductants.
Recently, our laboratory reported a direct enantioselective C (sp 3 )−H acylation of N-alkyl benzamides for the synthesis of αamino ketones; wherein, a chiral nickel catalyst could engage photocatalytically generated α-amino radicals and in situactivated carboxylic acids in acyl cross-couplings 38 . We reasoned that this strategy could be applied to the asymmetric benzylic C−H acylation of alkylarenes to address the challenges described above for the synthesis of α-aryl ketones via radical C (sp 3 )−H functionalization [36][37][38] . Despite that initial progress 38 , no examples of enantioselective benzylic C−H acylation have been reported. In addition, there is an increasing demand for the development of benzylic C−H functionalization reactions for the synthesis of high value-added molecules from simple alkylarenes [39][40][41][42][43][44][45][46][47][48][49][50] . In this work, we report an enantioselective benzylic C−H acylation of alkylarenes with in situ-activated carboxylic acids enabled by nickel and photoredox dual catalysis (Fig. 1c, top).

Results
Reaction design. The proposed catalytic cycle for this benzylic acylation is shown in the bottom of Fig. 1c. It has been reported that single-electron oxidation of bromide anion by photoexcited photocatalyst can generate bromine radical (E 1/2 [Ir(III*/II)] = +1.21 V vs SCE in CH 3 CN; E 1/2 ox [Br − /Br·] = +0.80 V vs SCE in DME) [51][52][53][54] . According to the literature precedent and our previous mechanistic experiments 38,51-54 , we hypothesize that the catalytic reaction is initiated by oxidative addition of Ni(0) catalyst I to an in situ-activated carboxylic acid to afford Ni(II) species II. Subsequent trapping of prochiral benzylic radicals generated from the bromine-radical-mediated HAT process provides Ni(III) complex III, which undergoes reductive elimination to yield the desired product and Ni(I) species IV. A recent computational study of nickel-catalyzed cross-coupling of photoredox-generated benzylic radicals suggested that reductive elimination is the stereochemistry-determining step 55 . Finally, SET between Ni(I) species IV and reduced photocatalyst regenerates the Ni(0) catalyst I and ground-state photocatalyst to close  Reaction optimization. Our investigation began with an exploration of reaction conditions for the coupling of 4ethylbiphenyl and 3-phenylpropanoic acid (Table 1). Based on previously reported elegant strategies and our recent conditions for carboxylic acid activation in ketone synthesis 38,56-60 , dimethyl dicarbonate (DMDC) was chosen as the activating agent to generate mixed anhydride in situ from carboxylic acids. After an extensive study of reaction parameters (also see Supplementary Table 1), we were delighted to find that a simple chiral nickel/bis (oxazoline) catalyst and a known Ir-photocatalyst could provide the acylation product in 85% yield and 94% ee (entry 1). An attractive feature of this transformation is that only commodity chemicals are involved in this reaction. From the standpoint of commercial availability, carboxylic acids are perhaps the most ubiquitous functional group. The reaction could be also performed at room temperature with similar efficiency (entry 2). The use of a nickel source free of bromide led to almost no product formation (entry 3). Interestingly, the addition of NaBr was found to restore the reaction with comparable outcome, showcasing the crucial role of bromide anion in the catalytic cycle (entry 4). The use of Boc 2 O instead of DMDC provided the desired product in 93% ee, but with poor yield (entry 5). Replacing the Irphotocatalyst with a ketone triplet sensitizer, which has been employed in nickel/photoredox catalyzed C(sp 3 )−H functionalization reactions 18,23,36 resulted in no product formation (entry 6). Running the reaction in the absence of NH 4 Cl, which has been previously employed to facilitate the formation of a mixed anhydride 56,60 , led to a significantly lower yield (entry 7). Control experiments revealed nickel, photocatalyst, and light are indispensable for product formation (entry 8). The use of other acyl surrogates in place of the in situ combination of carboxylic acid, DMDC, and NH 4 Cl did not provide improvements (entry 9).
Other chiral ligands such as L1 and L2 delivered acylation products with similar enantioselectivities, albeit in diminished yields (entry 10).
Late-stage functionalization. Given the particularly broad functional group tolerance of our method, we sought to demonstrate the utility of this operationally convenient method in the late-stage functionalization of medicinally relevant molecules (Fig. 3) 61 . Specifically, acylation of benzylic C−H bonds of drugs such as ibuprofen, fenoprofen, ketoprofen, and naproxen, provided corresponding drug analogs in good yields and enantioselectivities (47−51). Employing menthol and amino acid derivatives as alkylbenzene coupling partners led to good diastereoselectivities (52−55). With oxaprozin, stearic acid, oleic acid, 2,4-D, and lithocholic acid derivatives as acyl donors, the acylation proceeded with good stereoselectivity (56−61).
Gram-scale synthesis and parallel synthesis. To demonstrate the scalability of the present method, two 20.0 mmol scale reactions were performed in a common flask to produce 5.35 g of chiral ketone product 8, and 9.13 g of lithocholic acid derivative 60 with excellent stereoselectivity and good yield (Fig. 4a). To further demonstrate the synthetic utility, two types of drug analogs derived from (S)-flurbiprofen and artesunate were prepared in parallel with high yields and excellent stereoselectivities (Fig. 4b).
More than 100 mg of product was obtained in all cases. It is noteworthy that the labile peroxide subunit in artesunate was tolerated particularly well under mild conditions. This powerful  method enables the streamlined synthesis of drug analogs, providing attractive opportunities for the rapid exploration of structure-activity relationships in drug discovery 62 , as well as complementing the existing methods for the synthesis chiral αaryl ketones 7−13 .
Mechanistic observations. We next performed preliminary mechanistic studies for this newly developed method (Fig. 5). The primary kinetic isotope effect was observed in intermolecular parallel and competition experiments, which suggested that C−H cleavage significantly contributed to the rate-determining step (Fig. 5a). When the reaction was performed in the presence of an electron-deficient alkene, the benzylic acylation was completely inhibited, and a racemic adduct 71 was obtained in 58% yield (Fig. 5b, top). This observation supported the benzylic radical might be involved in the catalytic cycle. Moreover, in the absence of nickel catalyst and in situ-generated acyl electrophile (Fig. 5b,  bottom), the addition of 1.5 equiv of NaBr to the coupling of 4ethylbiphenyl with electron-deficient alkene led to the adduct 71 in 16% yield, which suggested photochemical oxidatively generated bromine radical was likely involved in the acylation reaction.
Rational expansion. Finally, we questioned whether this bromine-radical-mediated C−H cleavage strategy could be applied to the synthesis of α-aryl esters rather than α-aryl ketones [63][64][65] . Indeed, replacing the in situ-generated mixed anhydride with commercially available phenyl chloroformate led to a number of α-aryl esters in good yields and selectivities under similar conditions (Fig. 6). Chiral ligand L2 proved to be optimal for this transformation.

Discussion
In summary, a direct enantioselective benzylic C(sp 3 )−H acylation for the synthesis of α-aryl ketones has been developed. Several attractive features are noteworthy. First, both coupling

Data availability
The data that support the findings of this study are available within the article and its Supplementary Fig. 6 Rational expansion for the synthesis of α-aryl esters. All data represent the average of two experiments. Unless otherwise noted, reactions were conducted on a 0.5 mmol scale under stated conditions. a In place of the stated conditions, the reaction was conducted at −40°C with 5.0 equiv ethylbenzene.