Metal-Free Photoredox Catalyzed Cyclization of O-(2,4-Dinitrophenyl)oximes to Phenanthridines

A metal-free visible-light photoredox-catalyzed intermolecular cyclization reaction of O-2,4-dinitrophenyl oximes to phenanthridines was developed. In this study, the organic dye eosin Y and i-Pr2NEt were used as photocatalyst and terminal reductant, respectively. The oxime substrates were transformed into iminyl radical intermediates by single-electron reduction, which then underwent intermolecular homolytic aromatic substitution (HAS) reactions to give phenanthridine derivatives.

The intramolecular homolytic aromatic substitution (HAS) reactions of iminyl radicals have shown advantages in the synthesis of phenanthridine derivatives and other N-containing heterocycles. One of the pathways to iminyl radicals is the N-O bond cleavage of O-acyl or aryl oximes under the UV or microwave irradiation at high temperature reported by Walton [19][20][21][22][23] and colleagues (Scheme 1A). Recently, s visible light photocatalytic strategy for the conversion of N-containing compounds though a N-radicals and radical ion intermediates pathway was proved to be a mild and general tool in radical reactions [24]. By taking advantage of the single-electron redox potential of a photoexcited catalyst Ir(ppy) 3 , Yu and co-workers [25,26] found that the acyl oximes could be converted to iminyl radical intermediates which were able to undergo intramolecular homolytic aromatic substitution to give phenanthridines (Scheme 1B). More recently, Leonori and co-workers developed a photoredox cyclization of iminyl [27] and amidyl radicals [28] derived from electron-poor aryloximes and aryloxy-amides, and this activation mode was applied in the synthesis of dihydropyrrole and lactam derivatives.
In the context of our study on biological active phenanthridine derivatives [11,[29][30][31][32][33][34], we focused our attention on the development of facile, efficient and environmental-friendly synthetic method for phenanthridines and related compounds [35,36]. Drawing inspiration from the work of Walton, Yu and Leonori, we speculated that a visible-light photoredox catalyzed single electron reduction of electron-poor O-phenyl oximes 2 (Scheme 1C) to iminyl radicals might be followed by the generation of phenanthridines.

Results and Discussion
Among the commonly available electron-poor O-aryl oximes, O- (2,4-dinitrophenyl) oxime has the highest E1/2 red potential value of −0.55 V [27], which is suitable for SET with the excited state of commonly used photocatalysts such as Ru(bpy)3Cl2.6H2O (E1/2 *II/I = +0.77 V vs SCE), Ir(ppy)3 (E1/2 *III/IV = −1.73 V vs. SCE) and the organic dye eosin Y (E1/2 *EY/+EY = −1.11 V vs. SCE) [37]. In this study, as shown in Table 1, O- (2,4-dinitrophenyl) oxime (3a, Table 1) was used as model substrate. Ru(bpy)3Cl2·6H2O was firstly selected as photocatalyst and acetonitrile was used as solvent. After 24 h of reaction under visible light irradiation, only traces of phenanthridine (4a) could be detected along with the recovered starting compound 3a. The addition of the terminal reductant i-Pr2NEt was necessary to quench the visible light excited Ru II * and give Ru I species, a stronger electron donor (E1/2 II/I = −1.33 V vs. SCE) [37] that could accelerate the single electron transfer (SET) process between the substrate and photocatalyst. As it can be seen in entry 2, target compound 4a was isolated in 32% yield when i-Pr2NEt was used. Next, DMSO (entry 3) and DMF (entry 4) were screened as reaction solvents, respectively, which demonstrated that DMF was suitable for this type of radical cyclization reaction. Further photocatalyst screening showed that the replacement of Ru(bpy)3Cl2.6H2O with Ir(ppy)3 (E1/2 *III/IV = −1.73 V vs. SCE) [37] could afford phenanthridine 4a in 28% yield with part of starting material 3a being recovered (entry 5). As we anticipated, the combined use of i-Pr2NEt and Ir(ppy)3 significantly increased the yield of 4a to 51% (entry 6). It should be noted that 2-phenyl benzonitrile was detected as the major byproduct in entries 1-6. When the organic dye eosin Y (E1/2 *EY/+EY = −1.11 V vs. SCE) [27] was used instead of Ir(ppy)3, an obviously increased yield of compound 4a was observed (75%, entry 7). The addition of i-Pr2NEt could give compound 4a in 74% yields after 12 h of reaction (entry 8). No conversion of substrate was observed when the reaction was carried out in darkness (entry 9). Interestingly, compound 4a was obtained in 8% yield when the reaction was carried out under visible light irradiation in the absence of photocatalyst (entry 10). According to Leonori's studies [27], we suggested that a simple tertiary amine i-Pr2NEt would be able to reversibly interact with the 2,4-dinitrobenzene motif of 3a to give an electron donor-acceptor complex. Visible light irradiation could initiate a SET process of this complex to give the radical ion pair which would successively undergo fragmentation to give iminyl radical. Without visible light excitation, i-Pr2NEt could not initiate the cyclization reaction (entry 11). Finally, a trace of target compound 4a could be detected by simply heating a solvent of 3a in DMF at 100 °C for 6 h (entry 12). The generation of trace In the context of our study on biological active phenanthridine derivatives [11,[29][30][31][32][33][34], we focused our attention on the development of facile, efficient and environmental-friendly synthetic method for phenanthridines and related compounds [35,36]. Drawing inspiration from the work of Walton, Yu and Leonori, we speculated that a visible-light photoredox catalyzed single electron reduction of electron-poor O-phenyl oximes 2 (Scheme 1C) to iminyl radicals might be followed by the generation of phenanthridines.

Results and Discussion
Among the commonly available electron-poor O-aryl oximes, O-(2,4-dinitrophenyl) oxime has the highest E 1/2 red potential value of −0.55 V [27], which is suitable for SET with the excited state of commonly used photocatalysts such as Ru(bpy) 3 [37]. In this study, as shown in Table 1, O-(2,4-dinitrophenyl) oxime (3a, Table 1) was used as model substrate. Ru(bpy) 3 Cl 2 ·6H 2 O was firstly selected as photocatalyst and acetonitrile was used as solvent. After 24 h of reaction under visible light irradiation, only traces of phenanthridine (4a) could be detected along with the recovered starting compound 3a. The addition of the terminal reductant i-Pr 2 NEt was necessary to quench the visible light excited Ru II * and give Ru I species, a stronger electron donor (E 1/2 II/I = −1.33 V vs. SCE) [37] that could accelerate the single electron transfer (SET) process between the substrate and photocatalyst. As it can be seen in entry 2, target compound 4a was isolated in 32% yield when i-Pr 2 NEt was used. Next, DMSO (entry 3) and DMF (entry 4) were screened as reaction solvents, respectively, which demonstrated that DMF was suitable for this type of radical cyclization reaction. Further photocatalyst screening showed that the replacement of Ru(bpy) 3 Cl 2 .6H 2 O with Ir(ppy) 3 (E 1/2 *III/IV = −1.73 V vs. SCE) [37] could afford phenanthridine 4a in 28% yield with part of starting material 3a being recovered (entry 5). As we anticipated, the combined use of i-Pr 2 NEt and Ir(ppy) 3 significantly increased the yield of 4a to 51% (entry 6). It should be noted that 2-phenyl benzonitrile was detected as the major byproduct in entries 1-6. When the organic dye eosin Y (E 1/2 *EY/+EY = −1.11 V vs. SCE) [27] was used instead of Ir(ppy) 3 , an obviously increased yield of compound 4a was observed (75%, entry 7). The addition of i-Pr 2 NEt could give compound 4a in 74% yields after 12 h of reaction (entry 8). No conversion of substrate was observed when the reaction was carried out in darkness (entry 9). Interestingly, compound 4a was obtained in 8% yield when the reaction was carried out under visible light irradiation in the absence of photocatalyst (entry 10). According to Leonori's studies [27], we suggested that a simple tertiary amine i-Pr 2 NEt would be able to reversibly interact with the 2,4-dinitrobenzene motif of 3a to give an electron donor-acceptor complex. Visible light irradiation could initiate a SET process of this complex to give the radical ion pair which would successively undergo fragmentation to give iminyl radical. Without visible light excitation, i-Pr 2 NEt could not initiate the cyclization reaction (entry 11). Finally, a trace of target compound 4a could be detected by simply heating a solvent of 3a in DMF at 100 • C for 6 h (entry 12). The generation of trace cyclization product in entry 12 was possibly ascribed to the intermolecular nucleophilic substitution because N-O bond in 3a was weak and the 2,4-dinitrophenoxy motif was a suitable leaving group in the substitution reaction. cyclization product in entry 12 was possibly ascribed to the intermolecular nucleophilic substitution because N-O bond in 3a was weak and the 2,4-dinitrophenoxy motif was a suitable leaving group in the substitution reaction. Having developed a photoredox transition-metal-free radical cyclization as shown in Table 1, entry 8, we decided to explore the scope of substituent groups on the aryl ring of O-(2,4-dinitrophenyl) oximes 3. As shown in Table 2, when R 1 were electron-donating groups such as methoxyl, methyl, 2,4-dimethyl and chloro atoms, the target compounds 4b-4e were isolated in moderate yields (46%-56%), which were lower than that of 4a. Interestingly, when group R 1 was replaced by a trifluoromethyl group, an obviously increased yield of compound 4f was observed. It was suggested that electron-poor phenyl ring A of substrate 3 was much more suitable for the present HAS reaction. When the A ring of substrate 3 was 3-methyl)-substituted (compound 3g) the HAS reaction provided 4ga and 4gb in a ratio of 2:1 with total yield of 47%. Similar experimental results could be observed when ring A was 3,4-dimethoxyl-substituted (3h), and target compounds 4ha and 4hb were isolated in 58% total yield with a ratio of 2:1. We next turned to explore the scope of substituent group R 2 on ring B of substrate 3. When R 2 were 4′,5′-dimethoxy groups, target compounds 4i-4l were isolated in 40%-51% yield. Further exploration showed that changing R 2 to 4-F (3m-3p) or 4-Me (3q-3r) had no apparent effects on the yield of the phenanthridine derivatives, and target compounds 4m-4r were isolated in 40%-57% yield.
It was worth noting that 2-phenylbenzonitrile derivatives 5 were detected as byproducts as shown in Scheme 2. We speculate that these nitriles were produced by a competing hydrogen atom transfer (HAT) process. In order to avoid the HAT process, O-2,4-dinitrophenyl acetophenone oximes 6 were evaluated as substrates in the radical cyclization reaction to give corresponding 6-methyl phenanthridines 7 (Table 3).
Ru(bpy) 3  Having developed a photoredox transition-metal-free radical cyclization as shown in Table 1, entry 8, we decided to explore the scope of substituent groups on the aryl ring of O-(2,4-dinitrophenyl)oximes 3. As shown in Table 2, when R 1 were electron-donating groups such as methoxyl, methyl, 2,4-dimethyl and chloro atoms, the target compounds 4b-4e were isolated in moderate yields (46%-56%), which were lower than that of 4a. Interestingly, when group R 1 was replaced by a trifluoromethyl group, an obviously increased yield of compound 4f was observed. It was suggested that electron-poor phenyl ring A of substrate 3 was much more suitable for the present HAS reaction. When the A ring of substrate 3 was 3-methyl)-substituted (compound 3g) the HAS reaction provided 4ga and 4gb in a ratio of 2:1 with total yield of 47%. Similar experimental results could be observed when ring A was 3,4-dimethoxyl-substituted (3h), and target compounds 4ha and 4hb were isolated in 58% total yield with a ratio of 2:1. We next turned to explore the scope of substituent group R 2 on ring B of substrate 3. When R 2 were 4 ,5 -dimethoxy groups, target compounds 4i-4l were isolated in 40%-51% yield. Further exploration showed that changing R 2 to 4-F (3m-3p) or 4-Me (3q-3r) had no apparent effects on the yield of the phenanthridine derivatives, and target compounds 4m-4r were isolated in 40%-57% yield.
It was worth noting that 2-phenylbenzonitrile derivatives 5 were detected as byproducts as shown in Scheme 2. We speculate that these nitriles were produced by a competing hydrogen atom transfer (HAT) process. In order to avoid the HAT process, O-2,4-dinitrophenyl acetophenone oximes 6 were evaluated as substrates in the radical cyclization reaction to give corresponding 6-methyl phenanthridines 7 (Table 3).                                               As shown in Table 3, the metal-free photoredox-catalyzed cyclization of O-2,4-dinitrophenyl acetophenone oximes 6 provided 6-methylphenanthridines 7 in excellent yield (85%-92%). Yields of target compounds 7a-7f were not obviously affected by the R 1 group.
Based on the above experimental results in this study and previous work reported by Leonori [27], a reaction mechanism could be proposed, as shown in Scheme 3. In photocatalyst cycle I, visible light excited eosin Y* was reduced by i-Pr 2 NEt to eosin Y •− , which was a more powerful reductant that could reduce substrate 3 to radical anion A. The fragmentation of radical anion A led to phenoxyl anion B and iminyl radical intermediate C [27]. Cyclization of radical C through HAS process gave radical D which was further deprotonated by phenoxyl anion B to radical anion E. At this stage, radical anion E was involved in photocatalyst cycle II and was oxidized by excited eosin Y* to target compound 4 along with the generation of Eosin Y •− which was able to reduce substrate 3 and led to the generation of ground state eosin Y to complete photocatalyst cycle II. 6e 7e 85% 6f 7f 90% a Reaction conditions: 3 (0.5 mmol), photocatalyst (2 mol %), i-Pr2NEt (3.0 equiv.), solvent (5.0 mL), 25 W compact fluorescent light bulb, under nitrogen atmosphere for 12 h.

General Information
All reactions were carried out under a nitrogen atmosphere unless otherwise stated. 1 H-NMR (400 MHz) and 13 C-NMR (100 MHz) spectra were obtained at 25 • C with CDCl 3 as solvent and TMS as internal standard on a Bruker AVANCE III 400 M NMR instrument (Bruker, Swiss). HRMS data were obtained in the ESI mode on a 6530 Q-TOF/MS system (Agilent, Singapore). For flash chromatography silica gel (200-300 mesh) was employed (Qingdao Haiyang Chemical Co., Ltd., Qingdao, China). The 1 H-NMR and 13 C-NMR spectrum of compounds 4 and 7 are available at the Supplementary Materials.

Representative Experimental Procedure for Visible Light Promoted Synthesis of Phenanthridines 4 and 7
A solution of O-(2,4-dinitrophenyl) oximes 3 or 6 (0.5 mmol), 1.5 eq of i-Pr 2 NEt, 2 mol % eosin Y in DMF (5 mL) was firstly bubbled with nitrogen for 10 min and then irradiated with a 25 W household compact fluorescent lamp. After 16 h of reaction, the resulting mixture was poured into water (50 mL) and then extracted with EtOAc (20 mL × 3). The combined organic solution was then washed with water (20 mL × 3). The organic layers were washed with brine and dried over MgSO 4 . The solvent were removed via vacuo and the residue was purified by flash column chromatography (SiO 2 ) with petroleum ether/EtOAc (8:1) to give target compounds 4 or 7.