Fan
Gao
,
Tian
Wang
and
Xiaoyu
Yan
*
Key Laboratory of Advanced Light Conversion Materials and Biophotonics School of Chemistry and Life Resources, Renmin University of China, Beijing, 100872, China. E-mail: yanxy@ruc.edu.cn
First published on 22nd November 2024
Cyclopropanes are ubiquitous and key structural motifs in commercially available drugs and bioactive molecules. Herein, we present regio-selective acylation of aryl cyclopropanes with cooperative photoredox and N-heterocyclic carbene catalysis. This approach involves a deconstruction–reconstruction strategy via γ-chloro-ketones as intermediates and fulfills the formal C(sp3)–H functionalization of cyclopropanes.
Direct C(sp3)–H functionalization of strain cyclopropane frameworks has been recognized as an economic and simple strategy to access various cyclopropane scaffolds. Yu, Gaunt, Xu and others have reported powerful strategies to synthesize monofunctionalized cyclopropane derivatives3via the coordination interaction of transition-metals (Pd or Ir) and various directing groups (DGs) such as carboxylic acids,3f primary amines,3gN-aryl carboxamides,3iN-triflamides,3eN-aryl-aminomethyl,3h carboxamide,3c,d and ether3j (Scheme 2a).
Those available protocols show good regioselectivity, introducing various functional groups to the β-position of DGs, while few directed C(sp3)–H functionalized examples at the α-position3l have been achieved. Besides requiring tedious processes to introduce appropriate directing groups (DGs) and costly transition-metals as catalysts, those aforesaid methodologies usually require harsh reaction conditions, like heating to keep high regio- and stereoselectivity.
Due to their high ring strain energies (ca. 27.5 kcal mol−1), cyclopropane derivatives easily undergo deconstruction of the coplanarity ring, and have been identified as versatile and powerful C3 units in synthesis. Diverse strategies catalyzed by Lewis acids,4 transition-metals (Rh, Ni, Pd, Fe),5 visible light6 and electricity7 have been developed to produce [3 + n] annulation products (Scheme 2b). Meanwhile, 1,3-difunctionalization of cyclopropanes has also been achieved, generating acyclic products with introduction of two distinct functional groups.8 We envisioned that, with suitable functional groups as leaving groups, 1,3-elimination reactions are viable, which will reconstruct the cyclopropane skeleton. The two-step strategy of deconstructive 1,3-difunctionalization and 1,3-elimination would fulfill the formal C(sp3)–H functionalization of cyclopropanes and avoid the preinstallation of directing groups. With the rapid development of radical N-heterocyclic carbene (NHC) catalysis9 and our contribution in this area,9,10 herein, we disclose a cooperative NHC and photoredox9h,11 catalyzed acylation of aryl cyclopropanes, which involves formal C–H functionalization that has been achieved via a deconstruction–reconstruction strategy with γ-chloro-ketones as intermediates. Either aromatic or aliphatic acyl groups can be selectively introduced to the α-position of aryl groups. Meanwhile, this method could be extended to esterification of cyclopropanes with carbonate esters.
Studer's group has achieved the 1,3-difunctionalization of aryl cyclopropanes catalyzed by a cooperative NHC and organophotoredox catalyst, generating various γ-aroyloxy ketones.11d To facilitate reconstruction of cyclopropanes, we used acyl imidazole as the acyl source to avoid introducing the poor-leaving ester group. Hence, we first started our investigation by using 1-anisoylcyclopropane (1a) and N-anisoylimidazole (2a) as the model substrates in the presence of the triazolium-type NHC-1 as the organo-catalyst, [Ir(dF(CF3)ppy)2(dtbbpy)]PF6 (PC-1) as the photocatalyst, LiCl as an additive, K2CO3 as the base and anhydrous DMF as the solvent at room temperature under an Ar atmosphere and the irradiation of 20 W blue LEDs. After 48 h of reaction, the desired tertiary cyclopropane product 3aa was obtained in 78% yield (Table 1, entry 1). When we reduced 2.0 equivalents of K2CO3 to 0.2 equivalents, a 50% yield of 3aa was obtained, accompanied by the 1,3-difunctionalization product 3aa′ in 26% yield (entry 2). This indicates that product 3aa′ was transformed into product 3aa after increasing the equivalent of base and that product 3aa′ was the key intermediate. Screening different organo-catalysts (entries 3–6) showed that the yields of 3aa decreased with other NHCs. Other photocatalysts (entries 7 and 8) like 4CzIPN also showed decreased yields. Lower yields were obtained in CH3CN and DMSO (entries 9 and 10). Subsequently, several bases and additives were explored (entries 11–15), and the results showed that LiCl and K2CO3 were the most suitable additive and base in this reaction system. Control experiments indicated that the NHC catalyst, photocatalyst and visible light were critical for this reaction (entries 16–18). Finally, a higher concentration for substrates gave a better yield of 85% (entry 19).
Entry | Variation from “standard conditions” | Yieldb3aa [%] | Yieldb3aa′ [%] |
---|---|---|---|
a All reactions were performed by using 1a (0.1 mmol), 2a (2.0 equiv.), PC-1 (2 mol%), NHC-1 (20 mol%), K2CO3 (2.0 equiv.), LiCl (2.0 equiv.), and anhydrous DMF (0.05 M) under blue LEDs (440 nm, 20 W), stirred at room temperature and in Ar for 48 h. b Yield was determined by 1H NMR. c Isolated yield was given in parentheses. | |||
1 | None | 78 | 0 |
2 | K2CO3 (0.2 equiv.) instead of K2CO3 (2.0 equiv.) | 50 | 26 |
3 | NHC-2 instead of NHC-1 | 11 | 12 |
4 | NHC-3 instead of NHC-1 | 8 | 26 |
5 | NHC-4 instead of NHC-1 | 24 | 0 |
6 | NHC-5 instead of NHC-1 | 0 | 0 |
7 | PC-2 (5 mol%) instead of PC-1 | 32 | 0 |
8 | PC-3 instead of PC-1 | 11 | 0 |
9 | MeCN instead of DMF | 21 | 20 |
10 | DMSO instead of DMF | 42 | 0 |
11 | Cs2CO3 instead of K2CO3 | 57 | 0 |
12 | K3PO4 instead of K2CO3 | 29 | 22 |
13 | Na2CO3 instead of K2CO3 | 37 | 0 |
14 | Me4NCl instead of LiCl | 50 | 0 |
15 | KCl instead of LiCl | 15 | 0 |
16 | Without NHC-1 | 0 | 0 |
17 | Without PC-1 | 0 | 0 |
18 | In the dark | 0 | 0 |
19 | DMF (0.1 M) instead of DMF (0.05 M) | 85(79)c | 0 |
![]() |
With the optimized reaction conditions in hand, we explored a series of acyl imidazoles 2 (Table 2). Benzoyl imidazoles with different substituents bearing electron-donating or electron-withdrawing groups at the para- or meta-position proceeded smoothly to afford the desired tertiary cyclopropanes 3aa–3aj in moderate to good yields (33–81%). In general, higher yields were obtained for benzoyl imidazoles with electron-donating groups while lower yields were obtained for benzoyl imidazoles with strong electron-withdrawing groups. A low yield was obtained for 3ad, which is due to dehalogenation. For benzoyl imidazoles with ortho-substituents, corresponding products were obtained in 41–47% yields due to the effect of steric hindrance. Different aromatic substituted acyl imidazoles like 2-thienyl and 2-naphthyl were also found to be suitable in this reaction, affording corresponding products 3ao and 3ap in 26% and 62% yields, respectively. We speculate that the oxidation of thienyl under reaction conditions results in the low yield for 3ao. NHC-catalyzed radical acylations were usually limited to aryl acyl substrates, while aliphatic substrates were challenged.11e Gratifyingly, we found that aliphatic acyl imidazoles could also react smoothly with cyclopropane 1a, leading to the formation of desired products 3aq–3ar in high yields (71–77%). To demonstrate the high functional group tolerance and broad substrate scope of acyl imidazoles, late-stage functionalization of a bioactive molecule derived from ketoprofen was explored, generating the corresponding α-acylated cyclopropane derivative 3as in moderate yield (48%).
a All reactions were performed by using 1a (0.1 mmol), 2 (2.0 equiv.), PC-1 (2 mol%), NHC-1 (20 mol%), K2CO3 (2.0 equiv.), LiCl (2.0 equiv.), and anhydrous DMF (0.1 M) under blue LEDs (440 nm, 20 W), stirred at room temperature and in Ar for 48 h, isolated yields were given. b dr was determined by 1H NMR. |
---|
![]() |
The scope of aryl cyclopropanes was also investigated (Table 2). For aryl cyclopropanes bearing various substituents at the para-positions of the aryl moiety such as benzyloxy, disubstituted methoxy, cyclopropyl and phenyl, the corresponding products 3ba–3ea were obtained in moderate yields (32–55%). Notably, it could selectively afford mono-acylated cyclopropane derivatives 3da from a substrate bearing two cyclopropyl groups at 1,4-positions of the benzene ring. A low yield was obtained for 3fa, which is due to acylation at the naphthalene ring.12 Finally, to address the limitation of regioselectivity and diastereoselectivity, an unsymmetric cyclopropane, 1-ethoxy-2-phenylcyclopropane was explored to produce 3ga in 30% yield, showing excellent regioselectivity albeit moderate diastereoselectivity (dr = 78:
22).
To demonstrate the ease and practicality of this method, a “one-pot, two-step” process was performed starting from carboxylic acid 4, generating 3aa in 80% yield (compared to 79% when starting from N-anisoylimidazole 2a) (Scheme 3). NHC-catalyzed radical esterification reactions have been developed recently.11g We also extended this deconstruction–reconstruction strategy for the formal esterification of cyclopropanes. With diethyl dicarbonate 5 as the esterification reagent, 1a was converted to α-esterified cyclopropanes 6 in 74% yield. Furthermore, 6 can be easily converted to amide product 7, which has the skeleton of lumacaftor (Scheme 4).
To gain a deep insight into the mechanism for this transformation, several control experiments were subsequently conducted. Initially, radical scavengers (2,2,6,6-tetramethylpiperidin-1-yl)oxidanyl (TEMPO) and 2,6-di-tert-butyl-4-methylphenol (BHT) were employed respectively, which clearly inhibited the reaction (Scheme 5a). These results indicated that a radical process might be involved. Moreover, the intermediates A and B were successfully detected by high-resolution mass spectrometry (HRMS), implying that the NHC-derived ketyl radical and alkyl radical were generated in this transformation. Subsequently, Stern–Volmer fluorescence quenching experiments were carried out (for details see ESI†). As shown in Scheme 5b, the obvious linear relationships and different slopes between the fluorescence intensities and the concentrations of 1a and 9 suggested that the excited photocatalyst was more favorable to be quenched by acyl azolium 9.
![]() | ||
Scheme 5 Control experiments and linear relationship between I0/I and the concentration of 1a and 9. |
Based on these results of control experiments and previous investigations, we proposed the following mechanism (Scheme 6). The excited photocatalyst [Ir(dF(CF3)ppy)2(dtbbpy)]PF6 (PC*) (Eox = −0.89 V vs. SCE)13 was oxidatively quenched by acyl azolium 9 (E1/2 = −0.81 V vs. SCE)14 to generate NHC-derived ketyl radical intermediate III and PC˙+. Then, the substrate 1a (E1/2 = + 1.35 V vs. SCE) was oxidized by PC˙+ (EPC˙+/PC = + 1.69 V vs. SCE) to regenerate ground photocatalyst PC and intermediate I. Subsequently, chloride ions reacted with intermediate I to produce alkyl radical intermediate II, which coupled with the NHC-derived ketyl radical intermediate III to produce γ-chloro-ketones as intermediates IV. Finally, the targeted product V was afforded via nucleophilic substitution in the presence of a base.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4sc06355d |
This journal is © The Royal Society of Chemistry 2025 |