Visible-Light-Mediated Generation of Nitrogen-Centered Radicals: Metal-Free Hydroimination and Iminohydroxylation Cyclization Reactions

Abstract

The formation and use of iminyl radicals in novel and divergent hydroimination and iminohydroxylation cyclization reactions has been accomplished through the design of a new class of reactive O-aryl oximes. Owing to their low reduction potentials, the inexpensive organic dye eosin Y could be used as the photocatalyst of the organocatalytic hydroimination reaction. Furthermore, reaction conditions for a unique iminohydroxylation were identified; visible-light-mediated electron transfer from novel electron donor–acceptor complexes of the oximes and Et₃N was proposed as a key step of this process.

Nitrogen-centered radicals (NCRs) are a versatile class of intermediates that have wide applications in the synthesis of N-containing molecules (Scheme 1 A).[1] However, the difficulties associated with their generation have significantly thwarted their use in synthetic chemistry. In fact, established methods often rely on the homolysis of difficult-to-construct N–X bonds and require the use of toxic and hazardous reagents at elevated temperatures.[1a, 2] The development of a mild, selective, and general method to catalytically generate NCRs from readily available precursors would enable the facile construction of many N-heterocycles, which are privileged motifs in natural products and therapeutic agents.[3]

Scheme 1.

Nitrogen-centered radicals and divergent functionalization processes developed in this work.

Photoredox catalysis has emerged as a powerful technique through which single electron transfer (SET) reactions can be performed under mild conditions.[4] MacMillan[5] and co-workers have developed an asymmetric visible-light-mediated amination of aldehydes by enamine catalysis, and the groups of Sanford,[6] Lee,[7] Yu,[8] and Luo[9] have reported the photoredox generation of phthalimidyl and saccharyl radicals and their use in Minisci-type reactions. The groups of Zheng[10] and Knowles[11] have developed a method for the photoredox generation of diaryl and aryl alkyl aminium radical cations and employed them in C–N bond-forming reactions.

Drawing inspiration from the work of Forrester,[12] Narasaka,[13] and Walton,[14] we speculated that appropriately functionalized O-aryl oximes could serve as general, bench-stable NCR precursors that could deliver iminyl radicals upon photoredox activation under mild conditions.[15] Such an approach would clearly benefit from the facile synthesis of aryl oximes, and we hoped that the high structural modularity of the O-aryl hydroxylamines would allow us to identify substrates that do not require the use of transition-metal-based photocatalysts.[16] Herein, we describe the successful implementation of this approach and the development of novel, transition-metal-free, visible-light-mediated hydroimination and iminohydroxylation cyclization reactions (Scheme 1 B).

The guiding principle of our photoredox NCR synthesis capitalized on the evidence that electron-poor aromatic compounds have reduction potentials compatible with SET reduction by visible-light-excited photocatalysts,[17] as shown by MacMillan and co-workers.[18] Our envisaged photoredox iminyl NCR generation was initiated by the visible-light-promoted excitation of a photocatalyst (PC→*PC)[19] followed by SET reduction of the aryl unit of oxime A to give radical anion B (Scheme 2 A). A fragmentation leading to phenoxide C and the desired NCR D was anticipated to occur next owing to the low bond dissociation energy of the N–O bond.[20] At this stage, we decided to test the viability of this activation mode by combining it with an intramolecular cyclization to synthesize valuable five-membered N-heterocycles.[21] After 5-exo-trig cyclization, the C-centered radical E was expected to abstract a H atom from 1,4-cyclohexadiene (CHD)[20b] to give the desired product F and radical G, which regenerates the photocatalyst by SET, closing the catalytic cycle. As 5-exo-trig cyclizations occur rapidly (k_c≈9×10³ s⁻¹ at RT),[22] we expected that the reaction yields would correlate with the efficiency of the photoredox system.

Scheme 2.

Proposed photoredox cycle and electrochemical studies. EY=eosin Y, ppy=2-phenylpyridine.

The SET between the visible-light-excited photocatalyst and the aryl oxime became a focal point. As the reduction potentials of many photocatalysts are known,[4a] we started our investigations by evaluating the redox profiles of various aryl oximes with the goal of identifying the most suitable/active substrates. Analysis of oximes 1 a–1 g by cyclic voltammetry revealed irreversible reduction profiles that are in accordance with the expected fragmentation process. According to our electrochemical scale (Scheme 2 B), almost all of the examined oximes are expected to undergo SET reduction by *Ir^III; whereas only the nitro-substituted substrates 1 a–1 d have E_1/2^red potentials suitable for SET with the excited state of the organic dye eosin Y.[23]

Based on these results, we selected oximes 2 a–2 c as representative substrates for the evaluation of the proposed radical cyclization reaction (Table 1). To our delight, visible-light irradiation of 2 a–2 c in the presence of [Ir(ppy)₃], cyclohexadiene, and K₂CO₃ gave pyrroline 3 a in good to excellent yields (entries 2, 4, and 6). As predicted by the electrochemical studies, 2 a and 2 b furnished 3 a when eosin Y was used as the photocatalyst (entries 3 and 5), thus setting the stage for a fully organocatalytic photoredox hydroimination cyclization.

Table 1.

Optimization of the hydroimination cyclization

Entry	Substrate[a]	Photocatalyst	Solvent	Yield [%][b]
1	2 a	[Ir(ppy)3]	DMF	81
2	2 a	eosin Y	DMF	68
3	2 a	eosin Y	acetone	93[c]
4	2 b	[Ir(ppy)3]	DMF	53
5	2 b	eosin Y	DMF	15
6	2 c	[Ir(ppy)3]	DMF	91
7	2 c	eosin Y	DMF	7

[a] 2 a: Ar=2,4-(NO₂)₂C₆H₃; 2 b: Ar=4-(NO₂)C₆H₄; 2 a: Ar= 4-(CN)C₆H₄. [b] Determined by ¹H NMR analysis; see the Supporting Information for control experiments. [c] 2,4-Dinitrophenol (8-H) was also isolated; see the Supporting Information.

The substrate scope was evaluated with a focus on 2,4-dinitro-substituted aryl oximes owing to four favorable aspects: 1) The required hydroxylamine is commercially available, 2) these oximes are typically purified by crystallization, 3) their photoredox reactions do not require a transition-metal catalyst, and 4) the products can be purified by a simple acid–base wash (no chromatographic purification needed on the way from the ketone to the final product).

A broad range of oximes with diverse electronic and steric properties participated efficiently in the visible-light-promoted process (Scheme 3). Bicyclic heterocycles were also obtained in good yields as well as products arising from the cyclization onto di- and trisubstituted olefins.

Scheme 3.

Reaction scope. Yields of isolated products after acid–base wash are given. Yields determined by NMR spectroscopy are given in parentheses. [a] 1 mmol scale. [b] Reaction run in DMF. Boc=tert-butyloxycarbonyl.

Intrigued by the low reduction potential and LUMO energy[24] of the 2,4-dinitro-substituted aryl oxime 1 a, and inspired by the reports of Kochi,[25] Cossy,[26] and Melchiorre[27] on SET, we wondered whether a complementary activation mode could be exploited for the generation of NCRs by visible-light irradiation. As illustrated in Scheme 4 A, we speculated that a simple tertiary amine would be able to reversibly interact with 2 a to give an electron donor–acceptor complex H.[28] Visible-light irradiation should then initiate a SET process to give the radical ion pair J.[29] Fragmentation to give D, 5-exo-trig cyclization, and H atom abstraction would deliver pyrroline 3 a. By using the Rehm–Weller equation for electron transfer [ΔG_ET= 0.24 F(E_1/2−E_1/2^2a)−ΔE_excit+ΔE_coul],[30] the process was calculated to be exergonic (ΔG≈−30 kcal mol⁻¹), which indicates a very favorable SET. UV/Vis spectroscopy data further corroborated this proposal. When a CH₃CN[31] solution of 2 a was treated with Et₃N, a bathochromic shift was observed, which indicates the formation of a donor–acceptor complex (Scheme 4 B). The formation of such complexes has not been studied extensively, prompting us to evaluate the strength of this key interaction. By using Job’s method, the 2 a/Et₃N stoichiometry in the complex was confirmed to be 1:1, and titration experiments gave an association constant of K≈22 m⁻¹ (Scheme 4 B). TD-DFT calculations [CAM-B3LYP/6-311++G(d,p) in CH₃CN] confirmed that absorption at approximately 440 nm is due to a transition from the nitrogen lone pair to the π* orbital of the aromatic unit of the oxime.[24] Exposure of 2 b and 2 c to Et₃N (up to 10 equiv) did not lead to significant bathochromic shifts, which suggests that there is limited or no donor–acceptor complex formation.[24]

Scheme 4.

A) Proposed visible-light-mediated electron transfer via an electron donor–acceptor complex for the hydroimination of O-aryl oximes. B) UV/Vis studies. C) Initial findings. D) Proposed reaction mechanism for the iminohydroxylation cyclization.

Encouraged by the UV/Vis studies, we decided to evaluate the ability of 2 a to undergo the proposed visible-light- and Et₃N-mediated SET process. Irradiation of a solution of 2 a, Et₃N, and cyclohexadiene in CH₃CN furnished the desired product 3 a (47 %) together with iminoalcohol 4 a (41 %; Scheme 4 C). The unforeseen formation of 4 a opened the way to the development of the first visible-light-mediated iminohydroxylation cyclization reaction. By simply excluding cyclohexadiene from the reaction mixture, the yield of 4 a was increased to 85 %. Other amines were evaluated, and they also selectively provided 4 a, albeit in lower yields. As suggested by the UV/Vis studies, substrate 2 b gave the desired product in low yield whereas 2 c did not react.[24]

The formation of 4 a raised additional questions about the underlying mechanism and the origin of the oxygen atom in the final product (Scheme 4 D). The involvement of adventitious O₂ or H₂O was excluded by running the reaction under rigorously moisture- and oxygen-free conditions.[24] In contrast to the hydroimination cyclization, 2,4-dinitrophenol (8-H) was not formed, but we obtained 2-NO-4-NO₂-C₆H₃OH (10-H). This observation indicates a unique trifunctional role of the aromatic unit of the O-aryl oximes, which sequentially serves as a sensitizer, an electron acceptor, and an oxidant. Initial rate kinetics revealed the reaction to be first order in 2 a and to display saturation behavior in Et₃N (1st order at 0<[Et₃N]<1 equiv and zero order at [Et₃N]>1 equiv). Based on these findings, we propose the following mechanism: Fast and reversible binding of Et₃N and 2 a gives intermediate 5, which undergoes SET upon visible-light excitation to give the dipolar species 6. Fragmentation and 5-exo-trig cyclization give the C-centered radical 7 and the stable phenoxide 8 (pK_a≈4). Subsequent oxidation by attack of the radical onto the NO₂ group[32] leads to 9, and successive N–O bond homolysis furnishes 10 and the O-centered radical 11, which undergoes a fast hydrogen atom abstraction.[24]

With this very simple optimized procedure in hand, the scope of the iminohydroxylation was evaluated with the aryl oximes 2 a and 2 e–2 y. All examined substrates reacted well and provided the desired iminoalcohols 4 a–4 u in good to high yields (Scheme 5). Bicyclic products could be obtained, and substrates containing di- and trisubstituted olefins also reacted well, giving access to products containing up to three contiguous stereogenic centers.

Scheme 5.

Scope of the iminohydroxylation cyclization reaction. [a] 2 mmol scale.

In conclusion, we have developed a divergent strategy for the hydroimination and iminohydroxylation cyclization of unactivated olefins. Electrochemical studies facilitated the identification of a very reactive class of O-aryl oximes that obviate the need for a transition-metal photocatalyst and undergo organocatalytic hydroimination cyclizations. The unprecedented ability of the aryl unit to sequentially act as a sensitizer, electron acceptor, and oxidant enabled the development of a unique Et₃N- and visible-light-mediated iminohydroxylation cyclization. Future studies will focus on applying this method to other nitrogen-centered radicals and on developing asymmetric variants of the hydroimination and iminohydroxylation cyclizations.

Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

miscellaneous_information