Visible-Light-Mediated Three-Component Cascade Sulfonylative Annulation.

Visible-light-promoted cascade radical cyclization for the synthesis of sulfonylated benzimidazo/indolo[2,1-a]iso-quinolin-6(5H)-ones has been reported. The reaction provides transition-metal-free and expeditious access to sulfonylated polyaromatics. The use of sodium metabisulfite as a SO2 surrogate and the rapid generation of molecular complexity using a three-component photochemical protocol are the salient features of this reaction manifold.

Introduction

Sulfonylated molecules are ubiquitous in a large selection of pharmaceuticals, agrochemicals, and synthetic intermediates.[1−4] Traditional routes for sulfonylation rely on sulfinic acids and their salts,[5−10] sulfonylhydrazides,[11−13] tosyl chloride,[14−17] or the oxidation of sulfides and sulfoxides.[18] Typically, sulfonylation processes require harsh oxidizing conditions, high temperatures, and/or equivalent amounts of additives, leading to issues of scalability and limited substrate scope. DABSO is an interesting SO2 surrogate[19−21] but is expensive and tedious to synthesize.[22] In this context, sodium metabisulfite offers an alternative method to incorporate SO2, as it is readily available and inexpensive.[23] We envisioned that sulfonyl radical formation from the incorporation of SO2 into photochemically generated aryl radicals would provide an alternate method toward sulfonylated derivatives.[24−30] Due to their lower redox potential (0 V vs SCE), diazonium salts are very useful aryl surrogates.[20,31]

Multicomponent cascade reactions involving radical intermediates have emerged as efficient and ecofriendly[32−35] pathways for the synthesis of substituted benzimidazo/indolo[2,1-a]iso-quinolin-6(5H)-ones. Yu and coworkers have reported a silver-catalyzed decarboxylative radical cyclization[36] and perfluoroalkylation.[37] Subba Reddy et al. have reported an interesting acylation/cyclization methodology.[38] Sun et al. have developed the synthesis of THF-incorporated benzimidazo/indolo[2,1-a]iso-quinolin-6(5H)-ones.[39] Recently, a visible-light-promoted tri- and difluoroalkylation/cyclization cascade was reported by Guo and coworkers.[40] The Adiyala group has demonstrated a deaminative alkylation/cyclization continuous flow.[41] Xu and coworkers have reported a ketone-catalyzed photochemical synthesis of imidazoisoquinolinone derivatives.[42] An electrochemical radical cyclization was developed by the Lei group employing Mn catalysis.[43] In the context of sulfonylative cyclizations, Wang et al.,[44] Xia et al.,[45] and Yang et al.[46] have employed sulfonylhydrazides as a SO2 surrogate (Scheme a), and Li and coworkers demonstrated sulfonylation as well as carbamoylation under transition-metal-free conditions.[47] Gao and coworkers demonstrated visible-light-mediated sulfonylation using sulfonyl chlorides (Scheme b).[14] Sodium metabisulfite has also been explored as a SO2 surrogate.[48−50] Xie and coworkers recently incorporated SO2 (using K2S2O5) into N-propargylindoles toward the assembly of 9H-pyrrolo[1,2-a]indoles.[51] He and coworkers reported a four-component tandem reaction using Na2S2O5 as a SO2 precursor for the synthesis of sulfonylated quinoxalin-2(1H)-ones.[49] We envisioned a visible-light-promoted three-component sulfonylative annulation toward sulfonylated benzimidazo/indolo[2,1-a]iso-quinolin-6(5H)-ones involving acrylamides, aryl diazonium salts, and Na2S2O5 (Scheme c).

Scheme 1

(a–c) Synthetic Strategies for Benzimidazo/indolo[2,1-a]iso-quinolin-6(5H)-ones

Results and Discussion

For the initial studies, we chose the benzimidazole derivative 1a as the model substrate, and the results are outlined in Table . In the preliminary reaction in a dichloroethane (DCE) solvent, 67% yield of the product 3a was obtained when 1.5 equiv of the phenyl diazonium salt was used (Table , entry 1). On increasing and decreasing the amount of Na2S2O5 (to 2 and 4 equiv, respectively), diminished yields were observed (Table , entries 2 and 3). No improvement in the yield was observed when we evaluated solvents such as DMF, MeCN, and MeOH, which provided the product in 41, 46, and 35% yields, respectively (Table , entries 4–6). Employing acetone as the solvent did not lead to any measurable product formation. When the reaction was performed in dichloromethane (DCM), 64% yield of 3a was obtained (Table , entry 9). We selected DCE as the solvent for further optimization. Evaluation of other photocatalysts (rose bengal, 4CzIPN, rhodamine B, and rhodamine 6G) also failed to enhance the yield. On increasing the amount of Eosin Y from 2 to 5 mol %, we obtained 81% yield of 3a (Table , entry 12). We discovered that very slow product formation occurred in the absence of the photocatalyst (28% yield in 24 h; Table , entry 13). The reaction did not yield any product when performed in the dark in the presence of the photocatalyst. Further, no product was obtained when the reaction was performed in the absence of both the photocatalyst and light (Table , entry 15).

Table 1

Optimization Studiesa

entry	photocatalyst (mol %)	solvent	yield
1	Eosin Y (2)	DCE	67
2b	Eosin Y (2)	DCE	48
3c	Eosin Y (2)	DCE	42
4	Eosin Y (2)	DMF	41
5	Eosin Y (2)	MeCN	46
6	Eosin Y (2)	MeOH	35
7	Eosin Y (2)	THF	21
8	Eosin Y (2)	acetone	trace
9	Eosin Y (2)	DCM	64
10	rose bengal (2)	DCE	54
11	4CzIPN (2)	DCE	trace
12	Eosin Y (5)	DCE	81
13		DCE	28
14d	Eosin Y	DCE	trace
15e		DCE	NR

The reaction was performed with 0.18 mmol of 1a, phenyldiazonium salt 2a in the presence of the photocatalyst (mol %), and Na2S2O5 (equiv) dissolved in 2 mL of a solvent and irradiated with blue LED strips.

2 equiv of Na2S2O5 was employed.

4 equiv of Na2S2O5 was employed.

Reaction vial wrapped by Al foil (in the dark).

Reaction performed in the absence of light and a photocatalyst.

After establishing the optimized reaction conditions, we evaluated the generality of this reaction. As depicted in Figure , a variety of phenyldiazonium salts underwent the transformation smoothly. Diazonium precursors featuring electron-donating substituents p-methyl (3b) and p-methoxy (3c) provided the products in 69 and 68% yields, respectively. In the case of m-methyl (3d), we have obtained a 76% yield of the product. Meta- and para-bromo diazonium salts furnished the corresponding products 3e and 3f in good yields (71 and 70% yields, respectively). Certain electron-withdrawing substituents performed well such as the para-acylated diazonium salt, which afforded the product 3g in a 60% yield. Ortho-substituted diazonium precursors generally afforded diminished yields, and this effect appears to be independent of the electronic disposition of the substituent. Ortho-CF3- and o-F-substituted products 3h and 3i were obtained in 56 and 37% yields, respectively. The naphthyl derivative- and the 2-Ph-substituted products 3j and 3 k were obtained in 37 and 48% yields, respectively.

Figure 1

Reaction scope: evaluation of diazonium salts.

In the next stage, we explored various substituted benzimidazoles as outlined in Figure . Among the 2-aryl-substituted benzimidazoles, a variety of electronically distinct substitutions were tolerated well on the phenyl ring. The parent derivative 3l was obtained in a 75% yield. Electron-withdrawing substituents were tolerated well in the para position as exemplified by the fluoro- and cyano-substituted products 3m (68% yield) and 3n (90% yield). The m-OMe derivative afforded a regioisomeric mixture of products in a 76% yield (3o). In line with the expected effect of such substituents on the regioselectivity, the m-Br derivative resulted in a single regioisomer in a 71% yield (3p).

Figure 2

Reaction scope: evaluation of 2-aryl-benzimidazoles and indoles.

The ortho-chloro-substituted product (3q) was furnished in an 80% yield, while the naphthyl-substituted derivative (3r) was obtained in a 70% yield. The 2-thiophene-substituted benzimidazole precursor resulted in the product 3s in a good yield (67%). The dichloro benzimidazole precursor afforded a 55% yield of the corresponding product 3y. We also evaluated indolyl methacrylate precursors, which furnished products 3t–v in up to an 88% yield (Figure ). The nitro-substituted indolyl precursor afforded a 67% yield of 3z. We discovered that precursors with a free phenolic −OH group, internal olefins, and acrylate derivatives did not afford the desired product. We also performed the synthesis of 3t starting with 1 mmol of the precursor and obtained the product in an 86% yield.

We performed some preliminary experiments to obtain insights about the reaction mechanism. In an attempt to trap radical intermediates, we performed an experiment in the presence of TEMPO and discovered that the reaction was completely suppressed (Scheme ). Although we were unable to isolate the TEMPO adduct, we detected the presence of species 4 upon GC–MS analysis of the reaction mixture. When BHT was employed with the intent to trap radicals, the original product was obtained in only a 20% yield, although no BHT adduct could be isolated. We also performed a potential competition experiment by adding 1,1-diphenylethylene to the reaction under otherwise identical reaction conditions. We observed that only a trace amount of the product formed, and the species 5 and 6 could be detected upon the GC–MS analysis of the reaction mixture, indicating that the phenyl and phenylsulfonyl radicals were trapped by the olefin (Scheme ). We designed a radical clock experiment wherein the O-allylated phenyldiazonium salt (2w) was employed under standard conditions. The product 3w was obtained in a 63% yield, implying that the intramolecular trapping of the phenyl radical by the allyl double bond preceded the alkylsulfonate formation and subsequent steps. The results obtained from the above reactions provide evidence to indicate that the reaction likely follows a radical pathway. We also determined that the Eosin Y fluorescence was being quenched by the diazonium salt 2a.

Scheme 2

(a,b) Mechanistic Investigation

Based on abovementioned experimental observations and literature information,[31,52,53] we proposed a plausible mechanism of the reaction as depicted in Scheme . Initially, the phenyl radical is generated from the diazonium (E1/2 = −0.2 V)[54−56] by the oxidative quenching of the photoexcited catalyst (E1/2 = −1.11 V),[57] which is trapped by either Na2S2O5 or SO2 (thermally generated from Na2S2O5)[58] to form the phenylsufonyl radical B. This radical intermediate further reacts with the precursor 1a to generate the alkyl radical C, which is trapped by the aryl ring to generate the intermediate D. Finally, the photocatalytic cycle is being terminated by the oxidation of D to afford E. Alternatively, E could be generated by the single electron oxidation with the diazonium salt via chain propagation,[59] which then results in the final product 3a.

Scheme 3

Proposed Mechanism

In conclusion, we have developed a photochemical cascade cyclization that results in the formation of two C–S bonds and one C–C bond in a single operation. A collection of sulfonylated benzimidazo/indolo[2,1-a]iso-quinolin-6(5H)-ones were accessed in an expeditious manner utilizing this transformation that does not require transition metals, oxidants, or additives and employs a readily available SO2 surrogate. The preliminary potential of this reaction manifold to access even more complex moieties was demonstrated through a functionalized diazonium precursor, and further studies in this area are ongoing in our labs.