Photoredox Imino Functionalizations of Olefins

1Small‐molecule nitrogen heterocycles constitute the core of many drugs and agrochemicals and are one of the key epitopes evaluated in biological screenings.1 As a result, the invention of methods enabling their rapid construction is a topic of continuous scientific endeavour. An overarching goal in the design of these methodologies is the identification of divergent approaches whereby simple starting materials are converted into a broad array of products which contain different functional groups and efficiently tap into chemical space.

2Nitrogen radicals are a class of versatile synthetic intermediates, and their reactivity encompasses a series of powerful bond‐forming reactions, like intramolecular cyclization and 1,5‐H abstraction.2 In recent years, owing to the power of visible‐light photoredox catalysis3 in realizing single‐electron‐transfer (SET) processes,4 the chemistry of nitrogen‐radicals has witnessed a remarkable resurgence in scientific interest.5

3We have developed a class of electron‐poor O‐aryl oximes and aryloxy amides which allows, upon reductive SET, access to iminyl6 and amidyl7 radicals to use in radical hydroaminations (Scheme 1 A). However, despite all our efforts, we have not been able to harness this reactivity mode for the development of much sought after, but also more challenging, amino‐functionalization processes.2i, 8 Herein, we describe our work towards this goal, which has led to the development of a powerful platform for the divergent assembly of polyfunctionalized nitrogen heterocycles (Scheme 1 B).9

4While facing the difficulties of developing cascade photoredox imino functionalizations of electron‐poor O‐aryl oximes,6 we realized that an inherent chemical reactivity issue was associated with the envisaged catalytic cycle. The nitrogen radicals A are electrophilic species and undergo facile exo‐trig cyclizations (Scheme 1 C).10 However, upon ring closure, they generate the nucleophilic carbon radical B, which benefits from polar effects11 in the reaction with SOMOphiles X−Y. This process leads to the formation of the imino‐functionalized product C and the radical Y.. This radical, even though it is not involved in any bond‐forming event, controls the fate of our photoredox manifold. In fact, owing to the required polarization of X−Y for efficient reaction with B, Y. is expected to be electrophilic. This species generates an electronic mismatch in our first‐generation photoredox cycle, whereby [PC.+ + Y. → PC + Y+] is a endothermic process (Scheme 1 D, left). Therefore, we reasoned that should the iminyl A be generated by oxidative SET from the visible‐light‐excited photocatalyst (*PC), a facile SET between Y. and PC⋅− ought to take place, thus ensuring catalytic activity (Scheme 1 D, right). An additional challenge associated with the realization of this approach concerns the nature of the SOMOphile X−Y.12 In fact, many of these species are excellent SET quenchers for *PC, and they have been exploited in many powerful atom/group‐transfer reactions and they can lead to the formation of onium ions.12b As a result, we would require the oxidative generation of the iminyl radical, its intramolecular cyclization, and the following intermolecular radical reaction to supersede any other possible photoredox and ionic pathway.

5Inspired by the pioneering work of Forrester13 and Zard,2h we speculated that the oxime I might serve as a traceless (n+σ+σ) electrophore and undergo a sequence of two β‐scissions upon in situ deprotonation (II) and SET oxidation (Scheme 2 A). This step would deliver an iminyl radical in conjunction with the release of CO2 and HCOH (III→IV).14 As α‐hydroxy acids (e.g., 1) have been employed in many photoredox decarboxylations,12c, 15 we were confident that such a substrate design would be feasible. Surprisingly, electrochemical analysis of the model substrate 2 a revealed its oxidation potential to be almost outside the range for photoredox oxidation (E 1/2 ox=2.10 V vs. SCE; Scheme 2 B). In fact, previous activation methods required the use of strong oxidants either under forcing conditions or with the formation of the corresponding Barton ester.2h,13

6Hence, we speculated that the addition of organyl groups at the methylenic position might inductively increase the electron density at the carboxylate, and lower its E 1/2 ox. Pleasingly, the addition of a methyl (2 b) and a phenyl (2 c) group lowered the E 1/2 ox by 0.25 and 0.3 V, respectively. More importantly, the substrate 2 d, which was prepared by condensation of acetophenone with a commercially available hydroxylamine, displayed E 1/2 ox=1.65 V (vs. SCE), a value which can be reached by some organic photoredox catalysts.16

7Having identified a suitable electrophore, we prepared the substrate 3 a and begun our investigations by looking at developing an oxidative hydroimination (Scheme 3). To accomplish this goal, we selected the methyl acridinium perchlorate 4 (E* 1/2=+2.06 V vs. SCE)3c, 17 as the photoredox catalyst and evaluated a range of solvents and bases under blue LED irradiation. Pleasingly, in the presence of an aryl‐disulfide as H‐atom relay catalyst,18 and 2,6‐lutidine as the base, 5 a was obtained in 81 % yield, which, to the best of our knowledge, represents the first fully catalytic radical hydroimination of olefins. Control experiments established the requirement for 4 and continuous irradiation and the calculated quantum yield19 Φ=0.09 supports the photoredox nature of the process.16

8We next turned our attention to defining the capacity for imino functionalization in the presence of external coupling partners. By employing NCS, diethyl‐bromo‐malonate, NIS, and NFSI, we developed olefin imino halogenations in excellent to good yield (5 b–e). Interestingly, the use of Selectfluor as a radical fluorinating agent20 provided 5 e in considerably lower yield. Imino iodinations and imino fluorinations have not been reported previously. The use of arylsulfonyl azide and DEAD as coupling partners enabled the preparation of substrates with two nitrogen‐based functional groups in different oxidation states [imine and azide (5 f) or hydrazine (5 g)] and with orthogonal reactivity. This aspect opens the potential for further functionalization and application in click chemistry ligation. As thioether motifs are frequently found in pharmaceutical and agrochemical compounds,21 we evaluated their introduction as part of our strategy. Pleasingly, we were able to form the imines 5 h,i which incorporated S−CF3 and S−Ph groups. In this case, CsOBz12h proved to be uniquely effective in promoting good conversions. The formation of 5 h is particularly relevant because the introduction of the S−CF3 group in organic molecules is a significant but challenging task in medicinal chemistry.22 This reactivity was also successfully extended to the introduction of a Se functionality (5 j). Finally, we evaluated this approach for the installation of organic groups based on the cascade formation of C−N and C−C bonds. By using a Michael acceptor, we obtained the product 5 k, which represents a radical imino Michael cascade. Furthermore, by using hypervalent IBX reagents,23 we have developed unprecedented radical imino cyanation (5 l), olefination (5 m), and alkynylation (5 n) processes. The successful formation of 5 l is interesting, as recent literature reports described the cyanating reagent to be able to react with only C(sp3)‐radicals α to a heteroatom.12i The reagent used for the formation of 5 m 24 has not been used before in radical manifolds, and makes this example the first use of IBX reagents as vinylation partners in free‐radical processes.

9Regarding the optimization of these 14 different imino functionalizations, we were able to use 4 as the photocatalyst across all the systems, but the base, solvent, and reaction concentration had to be adjusted to maximize the reaction efficiency. In general, we have found that the inorganic bases Cs2CO3 and K2CO3, and the solvents CH2Cl2 and toluene, provided the best results.16 We have also conducted additional electrochemical Stern–Volmer and quantum yield (Φ) studies, which corroborate our working hypothesis for the reaction mechanism.16 According to these studies we believe that, in general, the iminofunctionalizations proceed by a single photoredox catalytic cycle starting with the SET oxidation of 3 a. However, in the case of the IBX reagents (imino‐cyanation 5 l and alkynylation 5 m), the quantum yields 2<Φ<5 suggest that very short‐lived productive radical‐chain processes might be operating together with the main photoredox manifold. It is worth mentioning, that none of these processes provided the desired products using our first generation iminyl radical precursors. This outcome clearly highlights the importance of polar effects and redox balance in the development of cascade reactions of nitrogen radicals.

10Having developed a platform for the divergent assembly of poly‐functionalized nitrogen heterocycles, we expanded the reaction scope. As shown in Scheme 4, we were able to implement substrates containing terminal olefins (3 b), which lead, via the intermediacy of a primary radical, to the imines 6 a–h. Pyridine substituents were tolerated, thus providing the nicotine analogues 6 i–k. Substrates containing disubstituted olefins (3 d,e) were also successfully engaged, thus proving that the strategy is amenable to the formation and functionalization of secondary benzylic (7 a–e) and α‐ester (7 f–j) radicals. By embedding the olefin into a cycle (3 f), we generated polyfunctionalized bicyclic molecules in good yield and good to moderate diastereoselectivity, thus favoring the all‐syn isomer (7 j–q). We also expanded the scope of coupling partners in terms of Michael acceptors (8 a,b) and IBX reagents for imino olefination (8 c) and alkynylation (8 d,e). Substrates prepared from α‐ketoacids could also be employed leading to the formation of proline‐like products (8 f–i). Lastly, this radical platform was evaluated as an enabling tool for the late‐stage imino functionalization of biologically active molecules. Therefore, we tested the structurally complex and densely functionalized morphane derivative thevinone (9),25 which was successfully engaged in the formation of products derived from imino azidation (10 a), imino amination (10 b), and imino selenation (10 c) of the olefin moiety.

11In conclusion, we have developed a general method for the fast and divergent assembly of polyfunctionalized nitrogen heterocycles. The reaction involves the organo‐photoredox generation of iminyl radicals by oxidative SET of a traceless electrophore and a subsequent cyclization/functionalization cascade with a broad range of SOMOphiles. This strategy generates useful building blocks in a single step, and its application to a number of more complex examples highlights its broad applicability.

12The authors declare no conflict of interest.