Introduction

C–C bonds are the basic skeleton of organic compounds and the direct functionalization via C–C bond deconstruction is quite meaningful for synthesis of complex molecules1,2. Cyclopropanes are important building blocks. The ring-opening of cyclopropanes driven by the release of ring strain has been widely applied in total synthesis3,4. Donor–acceptor cyclopropanes (DACs), which are activated by vicinal electron-donating and electron-withdrawing groups are predisposed to ring-opening under Lewis acid catalysis due to the inherent electronic bias5,6,7, whereas non-activated cyclopropanes, which are more regular in nature, are reluctant to ring-opening due to insufficient electronic bias. There are two methods for ring-opening of non-activatied cycolpropanes. One relies on oxidative addition by transition metals. However, these reaction are limited to ring-opening rearrangement or cycloaddition reactions and requires specific directing groups for regioselective ring-opening functionalization8. The other one relies on electrophilic activation with Lewis acidic species. However, most of the transformations are limited to electrophilic addition reactions9,10,11,12,13,14,15,16,17. Ring-opening functionalization of arylcyclopropanes initiated through the single electron oxidation followed by the yield of corresponding radical cations was discovered in the 1970s18,19. In recent years, this strategy was further applied in 1,3-aminofunctionalization, 1,3-oxoamination, and 1,3-oxochlorination of arylcycropropanes in the presence of oxidants or light. However, in most cases, its large scalability application was not studied and arylcycropropanes with electron-withdrawing groups could not be well compatible due to their high oxidative potential (Fig. 1a)20,21,22,23,24.

Fig. 1: 1,3-difunctionalization of cyclopropanes based on radical cation mechanism.
figure 1

a Thermochemical and photochemical strategy. b Shono’s work. c This work: Electrochemical 1,3-difunctionalization of arylcyclopropanes (PC photoredox catalyst).

Full size image

Organic electrochemistry is reviving due to their effortlessness of scalability, avoidance of stoichiometric oxidants or reducing agents, and flexible reaction tunability25. Various redox reactions have been achieved by the comsumption of traceless electrons under constant potential or current conditions26,27,28,29,30,31,32,33,34,35,36,37,38. As a main part of preparative electrosynthesis, anode processes such as C–H functionalization, oxidative coupling, decarboxylation, and olefin functionalization has been developed39,40,41,42,43,44,45,46,47,48. However, electrochemical oxidative C−C bond cleavage/functionalization are rarely developed due to the inertness and weak electronic bias of C−C bonds, which are always encumbered by other bonds8,48,49. Pioneering work was disclosed by Shono and coworkers who reported anodic oxidation of arylcyclopanes in methanol50. But only six examples were presented in this report (Fig. 1b). Our design for electrochemical C−C bond cleavage/functionalization based on the following mechanistic proposal (Fig. 1c): Firstly, arylcyclopropane is oxidized to a radical cation by anode, which results in the weakening of the Cα–Cβ bond, as the BDE of Cα–Cβ bond decreases more than 30 kcal/mol from the neutral cyclopropane to the corresponding radical cation51. Then the radical cation undergoes three-electron SN2 reaction to generate a benzyl radical52. Different from reported thermochemical and photochemical strategy, the benzyl radical can further lose one electron at anode and converted to a benzyl carbonium under electrochemical conditions43. The following nucleophilic attack to the benzyl carbonium can finally yield the 1,3-difunctionalization product. Fluorinated products could be prepared by employing Et3N·3HF as nucleophilic fluorine source53,54,55,56,57,58. In this work, we develop the electrochemical 1,3-difluorination, 1,3-oxyfluorination, and 1,3-dioxygenation of arylcyclopropanes with a high chemoselectivity and regioselectivity by the strategic choice of nucleophiles. Moreover, a wide variety of arylcyclopropanes with electron-donating and electron-withdrawing groups could be converted to the 1,3-difunctionalized molecules by following this protocol.

Results

Investigation of reaction conditions

We began our investigation by exploring the selective synthesis of 1,3-difluorination, 1,3-oxyfluorination, and 1,3-dioxygenation products from phenylcyclopropane. After extensive screening of various conditions (for more details, see Supplementary Table 1–5), with the use of Et3N·3HF as a fluorine source, 1,3-difluorination product 2 was obtained in 77% yield by conducting the electrolysis under constant current of 16 mA in an undivided cell equipped with platinum plate as both anode and cathode (Fig. 2, Entry 1). 1,3-oxyfluorination product 3 was obtained in 47% yield with the concomitant formation of other three 1,3-difuntionalization products under the existence of both Et3N·3HF and MeOH (Fig. 2, Entry 2). The yield of 1,3-oxyfluorination product 3 increased slightly when the reaction temperature decreased to 0 °C (Fig. 2, Entry 3). An obviously improved yield was observed by using carbon cloth as anode materials (Fig. 2, Entry 4). The influence of the concentration of phencyclopropane to the reaction could be neglected (Fig. 2, Entry 5). 1,3-oxyfluorination product 3 was finally observed in 76% yield by adjusting the ratio of Et3N·3HF/MeOH (Fig. 2, Entry 6). The good selectivity of the 1,3-oxyfluorination product 3 was speculated to be controlled by kinetics. At the first step, the reaction rate between MeOH and arylcyclopropane radical cation is larger than the reaction rate between Et3N·3HF and arylcyclopropane radical cation, which possibly due to the lower nucleophilicity of Et3N·3HF than MeOH57. Therefore, arylcropropane radical cation mainly reacts with MeOH instead of Et3N·3HF. According to the rate constant equation proposed by Mayr et al, the reaction rate of benzyl carbonium with MeOH or Et3N·3HF are both very fast and determined by diffusion rate59,60, so the fluorination is the major process in the second step because of the excess amount of fluorine source compared with methanol in the reaction system. Further condition screening demonstrated that 1,3-dioxygenation products could be obtained in 95% yield under the electrolysis in MeOH with Bu4NBF4 as supporting electrolyte (Fig. 2, Entry 7).

Fig. 2: Optimization of the reaction conditionsa.
figure 2

aReaction conditions: electrode surface (1.5 × 1.5 cm2, J = 7.1 mA/cm2), 16 mA, undivided cell, 1 (0.5 mmol), PhCF3 (4.8 mL), 2 h, 2.4 F mol−1 (based on 1). bGC Yield using biphenyl as internal standard. c 1 (0.25 mmol), 1 h. dWithout PhCF3, Bu4NBF4 (0.25 mmol) was added as supporting electrolyte (CC Carbon cloth).

Full size image

Substrate scope

To investigate the substrate scope of the 1,3-difluorination reaction, we tested a wide range of arylcyclopropanes under the optimized condition (Fig. 3). Phenylcyclopropane gave 75% NMR yield (2). Methanesulfonyl group no matter at para or meta position was well amenable to this protocol in good yields (6 and 7). Halogen groups such as F, Cl, Br, and I were also compatible with this transformation (8–12), which provided an opportunity to further transformation through coupling reactions. Ortho substituted arylcyclopropane was also suitable reaction partner, furnishing the desired 1,3-difluorination product with 58% yield (13). Arylcyclopropanes with electron-withdrawing groups (such as ester, amide, acetyl, cyano, nitro, and aldehyde groups) were also amenable, furnishing the desired products in 42–56% yields by prolonging the reaction time (14–20), although these substrates have higher oxidative potential. However, Arylcyclopropanes with electron-donating groups (such as –OMe, SMe, OPh, and NHBoc) were not amenable in this transformation (for more details, see Supplementary Fig. 3). Trifluoromethyl and oxytrifluoromethyl substituted arylcyclopropanes gave the corresponding products in 49% and 53% yields, respectively (21–22). TMS group was retained in this transformation with 52% yield (23). Alkyl group such as tert-butyl group substituted arylcyclopropanes underwent the reaction in lower yield of 40% (24). 3-Methyl-4-ester substituted arylcyclopropane provided the desired product with 54% yield (25). Moreover, heterocyclic arylcyclopropane such as thienyl cyclopropane was also suitable for this reaction (26). Then, the feasibility of 1,1-disubstituted cyclopropanes and 1,2-disubstituted cyclopropanes were also tested. 1,1-disubstituted cyclopropanes were good reaction partners to this reaction, giving the 1,3-difluorination products with tertiary-carbon-fluorine bonds formation in good yields (27–29). Trans-1,2-diphenylcyclopropane provided 1,3-difluorination products with 61% yield and 1.3:1 diastereoselectivity (30) since 1,2-diphenylcyclopropane radical cation was once determined to be open geometries51. Tri-substituted cyclopropane was also tolerated in this transformation (31). Moreover, arylcyclopropanes containing complex nature product scaffolds were also compatible in this transformation. l-menthol scaffold substrate provided the desired product in 86% yield (32). 5α-cholestan-3β-ol scaffold substrate provided the coresponding product in 65% yield with 10% starting material recoverd (33). Protected sugar group was also compatible with 82% yield was obtained (34). Androsterone scaffold substrate gave the desired product in 33% yield with 36% starting material recoverd (35). The compatibility of these complex substrates further confirms the practicality of this transformation.

Fig. 3: Substrate scope of 1,3-difluorination reaction.
figure 3

Reaction condition: Cyclopropanes (0.25 or 0.5 mmol), Et3N·3HF (1.2 mL), PhCF3 (4.8 mL), 16 mA, 40 min–20 h, isolated yields, electrode surface (1.5 × 1.5 cm2, J = 7.1 mA/cm2), undivided cell. aNMR yield using 1-fluoronaphthalene as internal standard. b0.5 mL DCE was added. c0.1 mmol scale, 2–6 h. s.m. starting material. dLarge-scale synthesis condition: carbon cloth anode, nickel foam cathode, cyclopropanes (6 or 8 mmol), Et3N·3HF (2.4 mL), PhCF3 (9.6 mL), 25–75 h, isolated yields. For more details, see Supplementary Experimental section.

Full size image

The scalability of this method was demonstrated by using carbon cloth as anode and nickel foam as cathode instead of expensive platinum electrode with equivalent amount of Et3N·3HF (Fig. 3). The yields increased for most of the tested substrates when operated at gram scale. The yields increased from 75% to 90% for para-methanesulfonyl substituted phenylcyclopropane (6). The yields increased to 65% and 63% for ester (14) and oxytrifluoromethyl (22) substituted arylcyclopropanes. The yields would decrease slightly from 67% to 63% only when it came to 4-bromocyclopropylbenzene (10).

The scope of the 1,3-oxyfluorination reaction was also explored, as shown in Fig. 4. Para-bromine substituted phenylcyclopropane was compatible to give the target product in moderate yield (36). Both electro-donating group and electro-withdrawing groups could be tolerated in this transformation (37–40). Meta-trifluoromethyl substituted arylcyclopropane generated 1,3-oxyfluorination product 39 in 53% yield. This 1,3-oxyfluorination product could be potentially converted to fluorinated Cinacalcet following the reported methods61,62. Subsequently, the scope of the alcohols were examined as well. Alcohols such as EtOH, n-BuOH and long-chain primary alcohols (n-C6H13OH and n-C8H17OH) were transformed into desired products with good efficiency (41–44). 2-Chloro-1-ethanol afforded the desired product in 40% yield (45). Secondary alcohols (i-PrOH and cyclohexanol) were suitable substrates (46–47). Sterically hindered tertiary alcohol was also tolerant (48). In addition to alcohols, ethers could also participate in this reaction (Fig. 4). 1,3-Oxyfluorination could proceed smoothly when using ether as both oxygen nucleophiles and solvent (49–51). Cyclic ether such as tetrahydrofuran gave interesting product 49 in 43% yields accompanying with 14% 1,3-difluorination product. 1,2-Dimethoxyethane gave 1,3-oxyfluorination product 50 in 44% yield. The reaction system was separated into two phases and not conductive when using tetrahydropyran as both reactant and solvent. To solve the problem of solubility and conductivity, DCE was added to enhance the solubility and 1,3-oxyfluorination product 51 was obtained in 36% yield. Notably, C–O etheric bond was broken in these transformations. We proposed that oxonium is the key intermediate for cleavage of C–O etheric bond (for more details, see Supplementary Fig. 5)63,64. Finally, the scope of the 1,3-dioxygenation reaction was investigated (Fig. 4). Arylcyclopropanes with electron-donating groups (–OMe, –OPh) gave 74% and 93% yields, respectively (52–53). Para tert-butyl substituted arylcyclopropane underwent the reaction in 85% yield (54). Both TMS and halogen groups could be tolerated (55–56). The electronically deficient arylcyclopropane could also give 1,3-dioxygenation products with high yields (57–58).

Fig. 4: Substrate scope of 1,3-oxyfluorination and 1,3-dioxygenation reaction.
figure 4

aCyclopropanes (0.25 mmol), Et3N·3HF (0.8 mL), ROH (0.2–0.75 mL), PhCF3 (4.8 mL), 16 mA, 1–2 h, isolated yields. bCyclopropanes (0.5 mmol), Et3N·3HF (1.2 mL), ether (4.8 mL), 4 h, isolated yields. c One milliliter of DCE was added. dCyclopropanes (0.25 mmol), Bu4NBF4 (0.25 mmol), MeOH (6 mL), 16 mA, 50 min−1h, isolated yields. For more details, see Supplementary Experimental section.

Full size image

Discussion

To gain more insights into the aforementioned transformation, several mechanistic studies were conducted. Redox potentials of the cyclopropanes were tested by cyclic voltammetry experiments (Fig. 5a). The alkyl substituted cyclopropane 1c has a high oxidation potential and no oxidation peak can be observed under 3 V. In contrast, arylcyclopropanes even with strong electron withdrawing group (−NO2) have relatively lower oxidative potentials (Ep/2 = 1.96 V vs. Ag/AgCl/KClsat for 1 and Ep/2 = 2.32 V for 1a vs. Ag/AgCl/KClsat). These results illustrate that aryl group is crucial for the oxidation of the cyclopropane substrates. The charge distribution of phenylcyclopropane radical cation has been studied by DFT calculation (Fig. 5b). The results show that distal C atom in cyclopropane motif possess partial positive charge and are the potential nucleophilic attack sites. Considering the reaction between benzyl radical and dioxygen was very fast (rate constant 2.8 × 109 L mol−1 s−1)65, the reaction was conducted in the dioxygen atmosphere in order to trap the possible benzyl radical intermediate. The detection of the oxygenation products is highly inductive of the formation of the benzyl radical during the reaction (Fig. 5c). In addition, the existence of benzyl radical was also evidenced by the trapping experiment using BrCCl366,67. Furthermore, the electrolysis of phenylcyclopropane and Et3N·3HF in CH3CN resulted in the isolation of amidation product 59, which suggested the involvement of benzyl carbonium intermediate during the reaction (Fig. 5d)68,69.

Fig. 5: Mechanistic studies.
figure 5

a Cyclic voltamemetry studies. b Electrostatic potential surface and charge of phenylcyclopropane radical cation (1·+). c Trap of benzyl radical intermediate. d Trap of benzyl carbonium intermediate.

Full size image

In conclusion, we have developed a electrochemical C–C bond cleavage of arylcyclopropanes, enabling 1,3-difunctionalization of arylcyclopropanes to yield 1,3-difluorination, 1,3-oxyfluorination, and 1,3-dioxygenation products. Neither additional oxidant or catalyst were needed in this transformation. Productive gram-scale 1,3-difluorination reaction was conducted by using stoichiometric amount of commercial available Et3N·3HF as fluorine source. Mechanistic studies show that arylcyclopropane radical cation and benzyl carbonium play paramount role in this reaction. This study provides a simple strategy for constructing 1,3-difunctionalized molecules.

Methods

General procedure (2)

An oven-dried undivided three-necked bottle equipped with a stir bar. The bottle was equipped with platinum plate (15 mm × 15 mm × 0.3 mm) as both the anode and cathode and then charged with argon gas in glove box. Phenylcyclopropane (0.5 mmol), Et3N·3HF (1.2 mL) and PhCF3 (4.8 mL) were added. The reaction mixture was stirred and electrolyzed at a constant current of 16 mA at 25 °C for 2 h. The reaction was diluted with water. The organic layer was extracted with CH2Cl2, dried with anhydrous Na2SO4, filtered, and concentrated under reduced pressure. The pure product was obtained by flash column chromatography on silica gel. Full experimental details can be found in the Supplementary Methods.