Aniline derivatives serve as essential structural scaffolds in the pharmaceutical, agrochemical, and materials science domains (Figure 1).^[1-3] Among these derivatives, acylated anilines can serve as versatile building blocks for the synthesis of various important heterocyclic compounds,^[4-5] while difluoroalkylated anilines act as bioisosteres, offering unique advantages in improving metabolic stability and lipophilicity.^[6-11] Traditional method for alkylation and acylation of anilines, the Friedel-Crafts reaction, is often constrained by the stoichiometric amounts of strong Lewis acids and the protection of amino group, typically affording ortho-products or ortho/para mixtures.^[12-14] Thus, achieving para-selective difluoroalkylation and acylation remains a challenging endeavor.

To address these limitations, several protocols were devised. Lewis acid catalysis^[15-17] and σ-activation^[18-19] strate- gies were applied to achieve para-selective difluoroalkylation (Scheme 1, A). Besides these thermal reactions, several visible-light-mediated photoredox catalysis protocols were devised (Scheme 1, B).^[20-23] In contrast to rapid develop-ment of difluoroalkylation, para-acylation of anilines is premature. Several thermal versions of these reactions employing aromatic ketones,^[24] benzyl alcohols,^[25] toluene^[26] and sulfoxonium ylides^[27-28] as acylating reagents were devised for para-acylation of anilines (Scheme 1, C). These approaches required elevated temperatures and prolonged reaction times. Notably, the group of Song^[29-36] has recently achieved a series of progresses in difluorocarbene chemistry, offering new strategies for the defluorinative acylation process under mild conditions.

Consistent with our interest in remote C—H functionalization,^[37-44] herein, we report an efficient para-C—H acylation and difluoroalkylation of aniline derivatives (Sche- me 1, D). Compared with the aforementioned existing methods, this work exhibits distinct advantages including mild reaction conditions and short reaction times. The mild reactions were induced by ruthenium catalysis with good functional group tolerance, which enabled the late-stage functionalization of complicated molecules. The reaction direction towards acylation and difluoroalkylation might be controlled by the electronic nature of the anilines. Mechanism investigation revealed that Ru(0) species initiated the catalytic cycle.

Our study was commenced by investigating the reaction of N,N-diethylaniline 1a with ethyl bromodifluoroacetate 2a (Table 1). Gratifyingly, the catalytic system with [Ru- Cl₂(p-cymene)]₂ (5.0 mol%) and K₂CO₃ (0.4 mmol, 2.0 equiv.) in N,N-dimethylformamide (DMF) successfully afforded 3a in 60% isolated yield (Table 1, Entry 1). The reaction vanished in the absence of either light or the ruthenium catalyst, indicating the crucial role of the photocatalyst and light (Table 1, Entry 2). Inert gas protection was also required, indicating a plausible radical pathway (Table 1, Entry 3). Replacing ruthenium catalysts with RuCl₃ or [RuCl₂(benzene)]₂ provided lower yields (Table 1, Entries 4 and 5). Notably, Ru₃(CO)₁₂ afforded 57% yield and shortened the reaction time to 2 h (Table 1, Entries 6 and 7). Thus, both Ru₃(CO)₁₂ and [RuCl₂(p-cymene)]₂ could be optimal catalysts. Further investigation of bases and temperature revealed that K₂CO₃ and 60 ℃ were optimal condition (Table 1, Entries 8 and 9 vs Entry 1).

With the optimized conditions, the generality of this para-C—H acylation was explored (Scheme 2). meta-Sub- stituted anilines with electron donating groups were well- tolerated, affording the desired products (3a~3d) in 41%~87% yields. Either increasing or reducing the steric hindrance on amino group decreased the reaction efficiency (3e and 3f). Tetrahydroquinoline and julolidine, common in fluorescent dyes and bioactive drugs,^[45-55] were also viable substrates, underscoring the method applicability (3g and 3h). Diverse difluoroalkyl bromides were next evaluated (3i~3p). Various α-difluorobromo esters were accessible to para-C—H acylation of anilines. Notably, medicinally relevant heterocycles such as benzoxazole and benzothiazole also underwent successful acylation. Inspired by the good functional group compatibility, this protocol was applied to the late-stage functionalization of natural products and bioactive molecules. α-Difluorobromoesters derived from 1-adamantanol (3q), monoterpenoid compounds (3r~3t), cetyl alcohol (3u), and steroids (3v, 3w) were shown to be compatible with our system. More specifically, sensitive functional groups such as alkenes (3v and 3w) and ketone (3w) were tolerated.

During the investigation of substrate scope, 1i with electron-withdrawing group at meta-site surprisingly afforded the para-difluoroalkylated aniline derivative 4a (Table 2, Entry 2). Therefore, optimization of the para-difluoro- alkylation was carried out (Table 2). After tedious screening of various factors, the desired product 4a could be obtained in 52% yield using 45 W 440~450 nm LED as light source and KHCO₃ as the base at 40 ℃ for 0.5 h (Table 2, Entry 1). Similar to para-C—H acylation, Ru catalysts, light and inert gas protection are crucial, indicating the same radical pathway (Table 2, Entries 3 and 4). Among different ruthenium catalysts, Ru₃(CO)₁₂ afforded the same yield with [RuCl₂(p-cymene)]₂ (Table 2, Entry 7 vs Entries 5 and 6). Subsequent screening identified 40 ℃ and KHCO₃ with 0.5 h as the optimal combination (Table 2, Entries 8~10).

With the optimized reaction conditions in hand, the scope of N,N-disubstituted anilines and difluoromethyl bromides was explored (Scheme 3). Substrates containing electron- withdrawing groups at the meta position were tolerated and afforded the corresponding products in 14%~41% yields (4a~4d). The low yields of 3 and 4 can be attributed to the decomposition of these products. Both α-difluorobromo amides and esters were accessible to para-C—H acylation of anilines (4e~4j). Furthermore, the protocol could be extended to the late-stage functionalization of natural product derivatives. Alkyl bromides derived from natural products such as 1-adamantanol (4k), monoterpenoid com- pounds (4l~4n), and steroids (4o) were all viable substrates, demonstrating its utility.

To explore the mechanism, a series of control experiments were carried out. The radical scavenger 2,2,6,6-tetra- methylpiperidinooxy (TEMPO) and 1,1-diphenylethylene completely inhibited the reaction, providing adduct 5. Furthermore, addition of ¹⁸O-labeled H₂O resulted in the detection of ¹⁸O in 3a, indicating that the carbonyl oxygen of 3a would be from H₂O either in the environment or in situ generated in the reaction (Scheme 4B). To further elucidate substituent effect of anilines on carbonylation vs difluoroalkylation, the difluoroalkylation product 4a to para-car- bonylation conditions was subjected. The absence of acylated product 3a indicates that the reaction time is unrelated and the electronic nature of the substrates would be crucial to control and suppress the hydrolysis of difluoromethyl group. To further evaluate the plausible pathway of para- acylation, difluoroalkyl compound 4p was synthesized following a literature method^[56] and completed its separation by preparative chromatography within 20 min. Notably, 4p is decomposed under the conventional column chromatography. When treated with one equivalent of water, 4p decomposed rapidly to give defluorinated acylation product 3e (Scheme 4C). These results further demonstrate the in situ generated water caused the conversion of 4p to 3e. To determine how the Ru complex triggers the redox cycle, Stern-Volmer experiments were conducted among the Ru catalysts and substrates (Figure 2). Notably, N,N- diethylaniline 1a quenched [RuCl₂(p-cymene)]₂ more efficiently than ethyl bromodifluoroacetate 2a, indicating that the reduction of [RuCl₂(p-cymene)]₂ with 1a would trigger the redox catalytic cycle with generation of low valence ruthenium species. Further Stern-Volmer experiments of effective Ru(0) catalyst, Ru₃(CO)₁₂, with 1a and 2a support this result: 2a exhibited better efficiency to quench Ru₃(CO)₁₂ than 1a. Therefore, Ru(0) can be considered as a plausible intermediate generated from reduction of [RuCl₂- (p-cymene)]₂ and initiate the redox cycle.

Based on the experimental results and previous reports,^[57-58] a possible reaction mechanism is proposed (Scheme 5). Under visible light irradiation, the reduction of [RuCl₂(p-cymene)]₂ with 1a occurs to afford Ru(0) species 8, which is excited by irradiation and reduces 2a to generate radical 10, which adds to 1a to form 11. 11 is oxidized by Ru(II) species 7 to generate the Ru(0) catalytic species 8 and 12. 12 undergoes deprotonation to yield difluoroalkylated product 13. 13 undergoes defluorination to generate intermediate 14 with extra stabilization from the electron- donating group on aniline substrates. 14 may undergo nucleophilic addition with H₂O, followed by a second defluo- rination to afford the target product 3a.

A visible light-mediated ruthenium-catalyzed protocol for the para-acylation and difluoroalkylation of aniline derivatives was developed. This mild photochemical method enables the late-stage functionalization of natural products and bioactive molecules. ¹⁸O labeling experiments revealed in situ generated water provided the oxygen of carbonyl group in acylation. Further mechanism investigation indicates Ru(0) species initiated catalytic cycle.