Reaction optimization. The feasibility of this concept was first investigated in a bromine-catalyzed hydroacylation of styrene with two equivalents of hexanal. After an extensive evaluation (Supplementary Table S1), we established that a combination of NBS (15 mol%) in PhF (0.025 M) at 80 oC under blue LED irradiation was optimal and produced the desired ketone product 3a in 66% isolated yield. No decarbonylated product was detected at this elevated temperature (80 oC). Other bromine sources such as Br2,40 CBr441, and BnBr42 also generated the desired product, albeit with lower efficiency. However, (n-Bu)4N-Br failed to promote this transformation. The solvent has a large impact on the reaction outcome and the formation of the product was observed only in aromatic organic solvents (e.g., PhCl, 33% yield; benzene, 46% yield). Use of other non-aromatic solvents such as MeCN, DMF, and EtOAc gave no product. A lower reaction temperature (50 oC) resulted in a significantly decreased yield of 16%. The yield of product can be improved to 71% by adding NBS in two portions sequentially (10 mol% NBS followed by additional 5 mol% NBS after 36 hours). Control experiments showed that light was essential for this transformation. Replacing NBS with PhSSPh could not lead to the formation of product under the optimized conditions (Supplementary Table S1).
Substrate scope. With the optimal conditions in hand, we investigated the substrate scope of the hydroacylation of vinyl arenes 2. As demonstrated in Fig. 2a, a broad scope of aliphatic aldehydes was effective and provided the ketone products in moderate to good yields. Primary aldehydes which are readily available feedstocks, participated smoothly in hydroacylation reactions (3a-3f). Some sensitive functionalities such as a terminal ester (3e) and alkyne (3f) were tolerated in this radical transformation. Secondary aldehydes, either acyclic (3g), cyclic (3h), or heterocyclic (3i), were all viable coupling partners. Compared to aliphatic aldehydes, aromatic aldehydes were less reactive, producing the corresponding products (3j-3l) in moderate yields (42-56%). The generality of this hydroacylation with respect to the vinyl arene component was subsequently investigated. A wide range of functionalities, including aryl bromides (3m, 3n), acetate (3o), cyanide (3q), tosylate (3r), and acid (3s) were tolerated. 2-Vinylnaphthalene was also a suitable substrate for the hydroacylation, delivering product 3t in good yield (75%). In addition, phosphine oxides could also be converted to corresponding organophosphorus compounds (3u, 3v). When pivaldehyde and diphenylacetaldehyde were applied to this protocol, decarbonylation was detected, and no desired product could be obtained (Supplementary Fig. S1and S2). Moreover, no reaction occurred when acrylates 4 were used in place of vinyl arenes 2, owing to the mismatched polarity between the radical adduct and HBr.
Multi-component radical cascade reactions play a privileged role in organic synthesis and enable direct construction of complex and diverse molecular scaffolds.7, 43-44 The failure of hydroacylation with acrylates 4 and the radical nature of this PRC led us to search for a possible bromine-catalyzed three-component coupling reaction. We speculated that the bromine-induced nucleophilic acyl radicals from aldehydes would preferentially add to electron-deficient acylates in the presence of vinyl arenes. The generated a-acyl radical was relatively electrophilic and would be expected to couple with vinyl arenes instead of undergoing an unfavorable HAT with HBr. The intended three-component PRC reactions were realized effectively with coupling aldehydes 1, unsaturated esters 4 and vinyl arenes 2, generating densely functionalized 1,4-dicarbonyl compounds in a highly selective manner (Fig. 2b). Such useful products would be difficult to access otherwise, and related reports usually require excess oxidants for quenching the nucleophilic radical adduct.45-46 Primary (5a-5c), secondary (5d-5f), and aromatic (5g-5j) aldehydes were all effective substrates for this cascade transformation, leading to 1,4-dicarbonyl products in moderate to good yields (41-79%). To further demonstrate the diversity of this transformation, acrylates 4 containing various functionalities, such as benzyl, trifluoromethyl, ether, chloride, and isobornyl groups participated smoothly to afford the corresponding products (5k-5p) in good yields (64-75%). Various vinyl arenes were applied in this protocol which illustrated good tolerance of functionalities (5q-5v). Functional groups such as the aryl chloride (5q) and bromide (5u) provided handles for further diversification.
In addition, we hypothesized that the bromine-induced nucleophilic acyl radicals may trap a deuterium atom from D-Br in a polarity-matched manner, thus realizing deuteration of aldehydes. The D-Br could be formed by rapid H/D exchange between HBr and D2O. The intended deuteration of aromatic aldehydes proceeded well in the presence of 25 equivalents of D2O using solely NBS as the catalyst under blue LED irradiation. All electron-deficient and electron-rich aromatic aldehydes delivered products in generally good to excellent isolated yields and deuterium incorporation (Fig. 2c). Functionalities ranging from electron-withdrawing substituents such as halides, trifluoromethyl groups, and boronate esters (6a-6e), to electron-donating substituents, such as phenyl (6f), alkyloxy (6g-6j, 6n), diphenylamino (6k), tert-butyl (6l), and methylthio (6m) groups, were tolerated well. However, aliphatic aldehydes were less effective in the H/D exchange process, and only low deuterium incorporations in the aldehyde C-H bond (6r-6t) were achieved. This may be due to the keto-enol tautomerization of aliphatic aldehydes in the presence of excess D2O,47 which was suggested by the deuteration of the α-positions of the formal groups. The successful H/D exchange also confirmed the formation of acyl radicals under light-promoted conditions. Notably, no deuteration was detected at the benzylic positions.
Further synthetic applications. Encouraged by the broad substrate scope of the bromine-based PRC protocols, the potential of this strategy for the late-stage site-selective functionalization of complex molecules was subsequently investigated (Fig. 3a). Aldehydes derived from natural products such as (+)-fenchol and lithocholic acid, participated in hydroacylation smoothly to afford products 7 and 8 in moderate yields. H/D exchange of derivatives from L-menthol, epiandrosterone, and cholesterol resulted in high deuterium incorporation with excellent chemoselectivity of the aldehydic C-H bonds. No tertiary-carbon functionalization could be detected in above transformations, even though Br radical has been reported as a suitable HAT agent for tertiary C-H bonds.11,37 The synthetic value of this method was further demonstrated by the diversification of the synthesized products (Fig. 3b). D-labeled alkenes can be easily obtained from deuterated benzaldehydes by Horner-Wadsworth-Emmons olefination.48 Treatment of 1,4-diketones that generated from the radical cascade reaction with NaBH4 followed by TsOH produced lactones 12 in good yields albeit in low diastereoselectivity.49
Mechanistic elucidation with supporting evidence. A series of control experiments were performed to reveal more hints concerning the mechanisms of the aforementioned transformations (Fig. 4a). A deuterium-labeling experiment was conducted by adding D2O to the hydroacylation reaction. This resulted in a 55% yield of deuterium-containing 1-phenyl-3-octanone 3a’, where deuteration took place at the α-positions of the carbonyl group and benzylic positions. When 3a was treated with D2O under the same conditions, deuteration was only detected at the α-positions of the carbonyl group (3a’’). These results indicate that keto-enol tautomerism47 promoted the deuteration at the α-positions of the carbonyl group, and this is consistent with the results of deuterated aldehydes 6r-6t (Fig. 2c). The 26% deuteration of 3a’ at the benzylic positions suggested a HAT between a generated benzylic radical and H/D-Br. No product could be detected under the standard hydroacylation conditions in the presence of TEMPO, supporting a radical pathway. The light on/off experiment indicated that photolysis is essential for the reformation of Br radicals to sustain the radical chain process. UV-Vis measurements indicate that both NBS and Br2 absorb and can be activated by light in the range of 400 to 600 nm (Supplementary Fig. S7 and S8).
In light of all experimental data and previous literature reports,35-38,50 plausible mechanistic pathways for photo-mediated bromine-catalyzed PRC reactions are proposed and illustrated in Fig. 4b. Even though the homolysis of NBS to bromine radicals could be induced by blue light irradiation,37 this process might be accelerated by heating under our optimal conditions. The generated bromine radicals can recombine to form Br2, which explains the requirement of constant light irradiation for the homolysis of Br2 back to the desired Br radicals. The Br radical abstracts a hydrogen atom from aldehyde 1 to produce an acyl radical I. When vinyl arenes 2 were utilized to trap the acryl radical, the resulting benzylic radical II can subsequently undergo a polarity-favored HAT with HBr to produce ketone product 3. Alternatively, in the presence of acrylate 4, the nucleophilic acyl radical I will preferentially add to electron-deficient acrylates rather than to vinyl arenes. In this way, an electrophilic radical adduct III is formed, which is polarity-mismatched with H-Br and will couple instead with vinyl arene 2. The resulting benzylic radical IV finally can trap a hydrogen atom from HBr to give the 1,4-dicarbonyl product 5. In the deuteration reaction, the acyl radical I will abstract the deuterium atom from D-Br, which is generated from the rapid H/D exchange between D2O and HBr.
Density functional theory (DFT) calculations were conducted to further support the proposed mechanism of hydroacylation. The calculated potential energy surface (PES) was illustrated in Fig. 5a. The reaction started with a process of bromine radical abstracting the hydrogen atom from propionaldehyde to deliver INT1, which was an easy process without transition state (Fig. 5b). The subsequent radical addition to styrene led to a radical adduct INT2 with an energy barrier of 11.6 kcal/mol. The HAT between hydrogen bromide and INT2 gave an energy barrier of 7.0 kcal/mol to deliver the product. The pathway of INT2 abstracting the hydrogen atom from propionaldehyde was unlikely due to a higher energy barrier of 23.6 kcal/mol.
In summary, we have disclosed for the first time, to the best of our knowledge, that bromine can be utilized as a polarity-reversal catalyst under visible-light irradiation. Hydroacylation of vinyl arenes, a three-component 1,2-difunctionalization of acrylates, and a deuteration of aldehydes were achieved in an atom- and step-economic and highly selective manner. The key to the success relies on the constant light-irradiation which induces and maintains bromine radicals for an active chain process. Compared to conventional PRC, this photo-mediated bromine-based PRC is distinguished by its green characters which stem from its no requirement of any chemical initiators.