Selective and direct oxygenation of benzylic C—H bonds is unarguably a fundamental and challenging transformation in modern organic chemistry and chemical industry. Due to the relatively high bond dissociation energy (BDE) of benzylic C—H bonds, classic thermodynamic oxidation procedure often needed harsh conditions,^[1] along with inevitable generation of certain amount of undesirable by-products. Consequently, endless efforts have been devoted to the development of efficient and selective oxidation processes.^[2] To this end, establishing mild chemical processes beyond the traditional oxidations is highly desirable, though still standing as a long-standing challenge. Over the past decades, visible light-driven photocatalysis has opened a new arena for molecular activation under mild conditions, revolutionizing the way for accessing new chemical transformations.^[3]

Up to now, a plethora of photo-oxidations of benzylic C—H bonds have been well documented.^[4] From 2000 to 2003, Fukuzumi’s group^[4a-4b] exemplified this type of transformation, in which xylenes could be smoothly oxidized to the corresponding monooxygenation product in the presence of 10-methyl-9-phenylacridinium (Ph-Acr^＋), but with limited substrate scope. Thereafter, further efforts have been made by König’s,^[4c] Wolf’s^[4d-4e] and Pandey’s^[4f] groups to improve the applicability and efficiency in the carbonylation of benzylic C—H bonds. However, the substrate scope of these photo-oxidation processes was generally limited to electron-rich arenes. In some cases, the selectivity issues were encountered with the generation of the malignant mixture of alcohol intermediate and the desired carbonyl product. In addition, other strategies relying on transition-metal catalysis or semiconductor-catalysis were also established, usually with cost issue, thus stumbling their practical applications in industrial field.^[5] Notably, oxygen gas was necessitated as an oxidant in most cases of these photocatalytic benzyl oxidation transformations, although air is the desirable oxidant with higher cost-effectiveness (Scheme 1a). From a sustainability point of view, it is highly demanding to comprehensively explore a metal-free, sustainable, and efficient strategy for photocatalytic benzyl oxidation, especially under air conditions.

Mechanistically, photocatalyst with high excited oxidation potential is required to trigger a single electron transfer (SET) from aromatic substrates due to their positive oxidation potential, which severely limited the functional group tolerance.^[6] Alternatively, photo-induced hydrogen atom transfer (HAT) process is not reliant on substrate redox potentials, offering a reliable pathway for activating benzylic C—H bonds.^[7] Therefore, designing an effective HAT process would be critical to the facile generation of benzyl radicals in the benzyl oxidation. Particularly, rational selection of photocatalyst simultaneously serving as a HAT activator would significantly upgrade the practicality and efficiency of the transformation.^[8] As part of our continued interest in photoredox catalysis,^[9] we herein reported a convenient photochemical strategy for the oxidation of benzylic C—H bonds to the corresponding carbonyl products with high chemoselectivity under air conditions (Scheme 1b). Furthermore, the developed protocol was successfully extended to the allylic oxidation of biologically important scaffolds with satisfactory reaction efficiency and regioselectivity.

To validate our hypothesis, we initiated our investigations using commercially available 1-ethylnaphthalene (1a, $E_{1 / 2}^{\mathrm{red}}$=－0.78 V vs Ag/AgNO₃ in MeCN) as the model substrate in the presence of a photocatalyst (Table 1). Pleasingly, the desired ketone product 2a was isolated in 36% yield, while using eosin Y (PC I) as photosensitizer, but along with 12% yield of alcohol 2a' as the byproduct (Table 1, Entry 1). Other organic photocatalyst gave inferior results (Table 1, Entries 2~4). Solvent screening showed that 1,4-dioxane was the optimal choice (Table 1, Entries 5~9). It was found that increasing the reaction concentration benefited this photo-reaction, yielding the desired ketone 2a in 61% yield (Table 1, Entry 10). Interestingly, lowering the illumination intensity dramatically improved the reaction efficiency (Table 1, Entry 11). In addition, control experiments further demonstrated that photocatalyst and light irradiation were requisite for this transformation (Table 1, Entries 12~13). Similar reaction results were observed while under O₂ atmosphere, but no product was obtained while under Ar conditions.

With the optimized reaction conditions in hand, the scope and limitation of this developed strategy were evaluated (Scheme 2). Ethylbenzenes bearing various functional groups, including ketone, halogen, NO₂, carboxylic acid, hydroxyl group, phenyl group or methoxy group, were tolerated well, delivering the desired aromatic ketones 2b~2l in moderate to good yields. To our delight, heteroaromatic substrate 2-ethylbenzofuran was amenable to this photocatalyzed transformation, producing the corresponding product 2m in moderate yield. But other heteroaryl alkanes such as 4-ethylpyridine were incompatible. In the case of diphenylmethane, the desired diphenyl ketones 2o and 2p were obtained in 53% and 84% yields, respectively. Furthermore, cyclic substrates were explored, which were oxidized readily to furnish 2q~2s smoothly. Ether was able to be oxidized as well, giving the corresponding ester product 2s in 43% yield. Toluene derivatives underwent this photo-oxygenation smoothly, producing the aldehydes 1u and 1v in an acceptable yield, respectively. Toluene substrate with strong electron-withdrawing substituent such as 4-NO₂-toluene exhibited quite low reactivity, only with 20% NMR yield of the corresponding aldehyde. Notably, in some cases, the overoxidized by- products were observed, thus resulting the low yield.

It is worth noting that this protocol was also applicable to the late-stage oxygenation of allylic C—H bonds in structurally complex biologically active steroids, affording the desired vinyl ketone 2w~2z in moderate yields (Scheme 3). It is worth noting that excellent regioselectivity and chemoselectivity were observed in these transfor-mations, although multiple reactive sites are involved in these substrates. However, alkyl alkynes such as but-1-yn- 1-ylbenzene were unable to give the desired product. Finally, the protocol was readily scaled up without loss in yield reaction efficiency, illustrating the synthetic potential of this protocol (Figure 1a).

In attempt to gain insight into the reaction mechanism, a series of experimental studies were performed (Figures 1b~1d). Under the standard conditions, addition of free radical inhibitors such as 2,2,6,6-tetramethylpiperidine oxide (TEMPO) or butylhydroxytoluene (BHT) inhibited the formation of target product, and the corresponding radical adduct could be found by high-resolution mass spectrometry (HRMS) (Figure 1b). The on-off experiments demonstrated that the reaction only proceeded with light irradiation. The possibility of free radical chain pathway was presumably ruled out (Figure 1c). Stern-Volmer experiments showed that 1-ethylnaphthalene (1a) had obvious quenching effect on the photocatalyst (Figure 1d).

On the basis of the mechanistic studies and previous reports,^[6-9] a photoinduced HAT process is proposed in Scheme 4. Initially, photocatalyst PC was excited to its excited state PC* under visible-light irradiation. Subsequently, a HAT process occurred between the excited state PC* and 1a, forming the benzyl radical A and PC-H. Capture of radical A by O₂ delivered the peroxy radical B, which could abstract the H-atom from PC-H to yield the intermediate C with regeneration of the photocatalyst. Final elimination of water produced the ketone product 2a.

In conclusion, an efficient and sustainable strategy for photocatalyzed oxidation of benzylic C—H bonds with high chemoselectivity has been established, affording a range of value-added carbonyl-bearing compounds. The well-designed HAT process is the key to the success of this transformation, which features mild conditions, eco- friendliness, and good functional group tolerance. The synthetic utility of the developed approach is well demonstrated by late-stage oxidation of bioactive complex steroids.