Organophotoredox 1,6-Addition of 3,4-Dihydroquinoxalin-2-ones to para-Quinone Methides Using Visible Light

An organophotoredox 1,6-radical addition of 3,4-dihidroquinoxalin-2-ones to para-quinone methides catalyzed by Fukuzumi’s photocatalyst is described under the irradiation of a HP Single LED (455 nm). The corresponding 1,1-diaryl compounds bearing a dihydroquinoxalin-2-one moiety (20 examples) are obtained with good to excellent yields under mild reaction conditions. Several experiments have been carried out in order to propose a reaction mechanism.

T he conjugate addition of nucleophiles to electrondeficient alkenes is one of the most important synthetic methodologies for the formation of C−C bonds in organic synthesis. 1−3 In contrast, the radical addition (Giese reaction) to electron-deficient alkenes is less investigated. 4−6 In this context, the 1,6-addition 7−9 is much less studied than the 1,4addition that is pivotal for synthetic organic chemistry. Nevertheless, in recent years, para-quinone methides have become important substrates for the development of 1,6conjugate additions. 10−12 para-Quinone methides are organic molecules that contain a carbonyl group and an exo-methylene moiety connected to cyclohexadiene, and display intrinsically high reactivity as versatile Michael acceptors driven by aromatization. Despite the significant advances in the field of 1,6-conjugate additions thanks to the versatility of paraquinone methides, if we compare the nucleophilic versus the radical 1,6-addition reactions, we could conclude that the radical version is scarcely explored.
Since the development of visible-light photoredox catalysis has allowed the generation of organic radicals under mild reaction conditions, 13−17 impressive achievements have been made in radical functionalization reactions. Accordingly, several radical 1,6-additions have been reported using paraquinone methides as electron-deficient acceptors mediated by visible-light. 18,19 For example, photocatalytic fluoroalkylation reactions using sodium sulfinates 20 or difluoroalkylating reagents 21 have been described, as well as alkylation reactions using cyanoalkylation reagents, 22 4-substituted Hantzsch esters, 23 or carboxylic acids. 24−27 Moreover, a photocatalytic 1,6-radical acylation reaction had been reported using simple carboxylic acids, triphenylphospine, and iridium photocatalyst. 28 Regarding the rich chemistry of α-aminoradicals 29,30 for conjugate additions, amines such as glycine 26 or anilines 31 have been used as precursors to describe the radical 1,6-addition with para-quinone methides. These reactions represent a convenient strategy for the synthesis of 2,2-diarylethylamines, 32 an important motif that widely exists in drugs and natural products. Despite these successful examples, these reports are limited to acyclic amines. As a part of our continuing interest in the development of synthetic approaches for the generation of α-amino radicals from other tertiary amines such as 3,4-dihydroquinoxalin-2-ones, 33−39 we envisioned that these cyclic amines could be suitable α-amino radical precursors which undergo a 1,6-radical addition with para-quinone methides using photocatalysis (Scheme 1). Furthermore, 1,4-dihydroquinoxalinones are an interesting class of nitrogen heterocycles which are present in many molecules with biological activities such as antiviral, 40 anticancer 41 or anti-inflammatory compounds. 42 Accordingly, the functionalization of this class of nitrogen heterocycles is significant for medicinal and pharmaceutical chemistry.
First, we evaluated the reaction using Ru(bpy) 3 Cl 2 as photocatalyst (entry 1). With these conditions, we obtained product 3aa in 72% yield determined as a mixture of diastereoisomers (1.2:1). After we decided to evaluate organophotocatalysts in order to increase the yield of product 3aa. When Eosin Y (entry 2) or 2,4,6-triphenylpyrylium tetrafluoroborate (entry 3) were used as photocatalysts, the efficiency of the reaction was worse, and 3aa was gained with much lower yield. A complex reaction mixture was observed when 4-CzIPN (2,4,5,6-tetrakis(9H-carbazol-9-yl) isophthalonitrile) 43 was used, while product 3aa was not observed when 9,10-phenanthrenedione 44,45 was tested (entry 4 and 5, respectively). Delightfully, we could quantify by 1 46 was employed. After, we proceeded to evaluated different solvents (entries 7−11) with [Mes-Acr-Me][BF 4 ] photocatalyst. When DMF was used as solvent, we could observe only 41% yield of 3aa, after 24 h of irradiation (entry 7). To our delight, when the reaction was performed in dichloromethane (DCM), the product 3aa was found in quantitative yield after only 9 h of irradiation (entry 8). However, the reaction did not proceed at all in toluene, probably due to the low solubility of both photocatalyst and 3,4-dihydroquinoxalin-2-one 1a in this solvent (entry 9). Other chlorinated solvents such as 1,2dichloroethane (DCE) and chloroform, were also tested obtaining high yields for product 3aa, but the performance of DCM as solvent was slightly better. The variation of the equivalents of 1a (entry 12) or 2a (entry 13) did not improve the yield of the reaction. The use of Et 3 N-deactivated silica gel as stationary phase allowed us to purify product 3aa without observing decomposition, and 3aa was isolated in 99% yield (entry 8). Additionally, control experiments showed that the photocatalyst, visible-light irradiation, and an inert atmosphere are essential for the success of this transformation (entries 14− 16). Moreover, product 3aa was not observed when the reaction was performed under oxygen atmosphere or in the presence of 1.5 equiv of the radical scavenger TEMPO (entry 17).
After establishing the optimized reaction conditions to carry out the photocatalytic 1,6-addition reaction of 3,4-dihydroquinoxalin-2-one 1a to para-quinone methide 2a, we wanted to explore the generality of this transformation. First, the versatility of the cyclic amines was investigated. Different substituted 3,4-dihydroquinoxalin-2-ones with different electronic and steric properties were tested in the reaction with para-quinone methide 2a and the corresponding addition products 3aa−3ia could be obtained with good to excellent yields (Scheme 2). Initially, we studied the effect of different substituents at the aminic nitrogen (R 1 ) of 3,4-dihydroquinoxalin-2-one 1. The presence of a more electron-rich benzylic substituent such as the para-methoxybenzyl group resulted in the corresponding product 3ba with an excellent 99% yield, comparable with that of compound 3aa. Similarly, the presence of a methyl or CH 2 CO 2 Me group at this nitrogen of the dihydroquinoxalin-2-one moiety was allowed, and the corresponding products 3ca and 3da, were obtained in 91 and 81% yield, respectively. In any case, we did not observe the product functionalized at exocyclic CH 2 of amines 1. When we tested the reaction with N-4 unprotected quinoxalin-2-one derivative 1e, we isolated N-alkylated product 4ea in 44% yield after 15 h. This product corresponds to the 1,6-aza-conjugate addition reaction to para-quinone methide 2a. Actually, we confirmed that this reaction should be mediated by visible light, since if it is performed in the dark, product 4aa was only isolated in 11% yield after 3 days. To our delight, 3,4-dihydroquinoxalin-2-one bearing an electron-donating (Me) or electron-withdrawing (F) group at different positions of the parent aromatic ring furnished the corresponding phenols 3fa and 3ga in good to   Subsequently, the scope and limitation of para-quinone methides 2 were explored (Scheme 3). Initially, we envisioned that it would be of interest to carry out this photochemical reaction with all the regioisomeric MeO-substituted paraquinone methides at the aromatic ring (2b−2d). Independently of the position of methoxy group, we could isolate the corresponding products with excellent yields (86−97%). Next, we evaluated the incorporation of electron-withdrawing groups such as halogens (Cl or Br), NO 2 , or CN on the benzene ring of the para-quinone methide 2, and we observed that the presence of these groups had no remarkable impact on the reaction and the corresponding products (3ae − 3ah) were obtained very high yields. Moreover, the reaction tolerates para-quinone methides bearing different hydroxyl groups protected with tert-butyldimethylsilyl or acetyl groups. Besides, a para-quinone methide with an alkyl group (Me) at the electrophilic position was tolerated under the optimized reaction conditions providing the expected product (3ak) in quantitative yield. Finally, we demonstrated the utility of our protocol for the late-stage functionalization of structurally diverse pharmaceutically relevant substances using a sophisticated para-quinone methide 2l resulting from the incorporation of the indomethacin core, a nonsteroidal antiinflammatory drug. This derivative was subjected to our organophotoredox 1,6-radical addition protocol furnishing the desired dihydroquinoxalin-2-one derivative 3al bearing the indomethacin scaffold in 79% yield.
To gain insight into the mechanism of the reaction, we first examined the reduction potential values of each component in the reaction mixture. According to the literature, [Mes-Acr-Me]* + has a reduction potential of +1.88 V (vs SCE) from its T 1 excited state and a reduction potential of +2.18 V (vs SCE) from its S 1 excited state. 47,48 Curiously, since [Mes-Acr-Me] + does not exhibit reductive abilities, it can only participate in reductive quenching cycles. Regarding both substrates, the reduction potential of 3,4-dihydroquinoxalin-2-one 1a was already determined by us, 35 and it was +0.80 V (vs SCE). The reduction potential of para-quinone methide 2a was determined by Tang, Cai, and co-workers, and it was found to be −1.18 V (vs SCE). 27 Hence, according to these data, the most probable pathway involves a single electron transfer between the excited state of [Mes-Acr-Me] + and 1a. To prove this thermodynamic assumption, we decided to perform steady-state luminescence quenching experiments. The study of the luminescence quenching of [Mes-Acr-Me] + by 2a was already reported in the bibliography by Ao, Liu, and coworkers. 23 They found that para-quinone methide 2a was not able to quench the excited state of [Mes-Acr-Me] + . Therefore, we only tested the ability of 3,4-dihydroquinoxalin-2-one 1a to  quench the excited photocatalyst. Luminescence quenching experiments are summarized in Figure 1A. 49 According to these studies, 3,4-dihydroquinoxalin-2-one 1a could quench the photoexcited [Mes-Acr-Me] + effectively, and therefore, we can establish a Stern−Volmer constant (K SV ) of 127 M −1 ( Figure 1B). Additionally, to confirm the participation of a closed photoredox catalytic cycle and exclude a radical chain process, we determined the quantum yield of the process. 49 We found out that the quantum yield of our photochemical reaction is as low as Φ = 0.040 ± 0.004, showing that the participation of a chain mechanism is unlikely.
With all this information, we were able to postulate a plausible reaction mechanism for our photochemical protocol (Scheme 4). Dihydroquinoxalin-2-one 1a, can be engaged in a single electron transfer (SET) with the excited state form of [Mes-Acr-Me]* + which results after the irradiation with 455 nm light. The SET results in the formation of the corresponding radical cation I, which can suffer the loss of a proton at its α position to generate the nucleophilic α-amino radical II. This carbon centered radical II is nucleophilic enough to react with the electrophilic exocyclic carbon of paraquinone methide 2a in a 1,6-fashion. The product of this radical 1,6-addition may be O-centered radical III. Taking into account the oxidative potential of the radical intermediate 50 the phenoxyl radical III could readily oxidize it, via SET, into [Mes-Acr-Me] + , 51,52 and yield alkoxide IV. Finally, a proton transfer over IV affords the desired product 3aa.
In summary, we have developed a 1,6-radical addition of 3,4-dihydroquinoxalin-2-one derivatives with several paraquinone methides using visible-light organophotoredox catalysis. Our methodology provides a rapid and efficient access to functionalized phenols bearing a dihydroquinoxalin-2-one moiety under mild reaction conditions and simple operational protocol using the irradiation of HP single LED of 455 nm. Also a series of experiments have been carried out in order to gain insights into the reaction mechanism. ■ ASSOCIATED CONTENT

Data Availability Statement
The data underlying this study are available in the published article and its Supporting Information.
Complete experimental procedures, photochemical setup, quantum yield determination, characterization of new products and 1 H and 13