Open Access Article
Ran
Song
a,
Jiaxing
Li
a,
Lin
Li
a,
Huaixi
Mu
a,
Wei
Feng
a,
Daoshan
Yang
ab,
Wen
Si
*a and
Jian
Lv
*a
aKey Laboratory of Optic-electric Sensing and Analytical Chemistry for Life Science, MOE, College of Chemistry and Molecular Engineering, Qingdao University of Science and Technology, Qingdao, 266042, P. R. China. E-mail: lvjian@iccas.ac.cn; siwen@qust.edu.cn
bNational Engineering Research Center of Low-Carbon Processing and Utilization of Forest Biomass, Nanjing Forestry University, Nanjing 210037, China
First published on 14th August 2025
Developing applicable methods to forge linkages between sp3- and sp3-hybridized carbons is of great significance in drug discovery. Here, a radical decarboxylative C(sp3)–C(sp3) cross-coupling of γ-methylidene-δ-valerolactones (GMDVs) and alkyl carboxylic acids was developed by combining photocatalysis and palladium catalysis, providing the target products in good to excellent yields (up to 97% yield). The reaction exhibited a broad substrate scope and wide functional group compatibility.
γ-Methylidene-δ-valerolactones (GMDVs), first established as versatile C4-synthons through Shintani and Hayashi's seminal [4 + n] cycloaddition chemistry in 2007,6,7 have remained underexplored as coupling partners compared to conventional organometallic reagents. Our prior investigations revealed the unique reactivity profile of these cyclic systems, demonstrating their capacity as O-allylation reagents for phenolic substrates through Csp3–O bond formation under mild conditions.8a Very recent advances in our group further expanded this paradigm through the development of a Pd0/CuII/VB2-catalyzed N-allylation protocol with aniline derivatives under blue LED irradiation, achieving efficient construction of allylamine architectures.8b Despite these advances, the direct engagement of GMDVs in Csp3–Csp3 bond-forming processes has remained conspicuously absent from the synthetic toolbox. To address this gap, we herein disclose a pioneering dual decarboxylative strategy that synergistically combines GMDVs with alkyl carboxylic acids under palladium/photoredox dual catalysis (Scheme 1B).
To validate our hypothesis, we commenced the investigation by employing GMDVs 1a and 9H-fluorene-9-carboxylic acid 2 as model substrates in a palladium-catalyzed system (Table 1). The initial reaction utilizing Pd(PPh3)4 (5 mol%), Ir(dFMeppy)2(dtbbpy)(PF6) (PC-1, 2 mol%) as a photocatalyst, and K2CO3 (4 equiv.) in DMSO under 370 nm UV irradiation (10 W LED) at ambient temperature under a nitrogen atmosphere afforded the Csp3–Csp3 coupling product 3a in 82% isolated yield (entry 1). Control experiments revealed critical dependencies: partial suppression of 3a formation occurred in the absence of a photocatalyst (entry 2), while exclusive generation of the Csp3–O byproduct 3a′ was observed under dark conditions (entry 3). Subsequent systematic optimization identified DMA as the superior solvent (96% yield, entry 7 vs. entries 1, 5 and 6), outperforming DMF (78%) and maintaining solvent superiority over DMSO (82%). The reaction demonstrated strict base dependence, with Cs2CO3 (64%), Na2CO3 (78%), KH2PO4 (34%), NaH2PO4 (39%), KOAc (60%) and NaOAc (60%) proving inferior to K2CO3 (entry 1 vs. Table S1, entries 2–7). Notably, stoichiometric analysis revealed the necessity of excess 2 (4 equiv.) and K2CO3 (4 equiv.), as reduced equivalents diminished yields (entry 9), while substitution of K2CO3 and 2 with potassium carboxylate K[2] completely abolished reactivity (entry 8), strongly implicating in situ-generated H2O or free carboxylic acid as crucial proton donors. Photocatalyst screening confirmed PC-1's superiority over alternative photocatalysts (entry 7 vs. entries 10–12), with 370 nm UV irradiation proving optimal among the tested wavelengths (entry 7 vs. entries 13–15). Temporal analysis revealed exceptional reaction efficiency, achieving 94% yield within 2 hours (entry 16), though incomplete conversion (76%) resulted from premature termination at 1 hour (entry 17).
| Entrya | Variation from standard conditions | Yieldb (%) |
|---|---|---|
| a Standard reaction conditions: under a nitrogen atmosphere, 1a (0.1 mmol), 2 (0.4 mmol), Pd(PPh3)4 (5 mol%), Ir(dFMeppy)2(dtbbpy)(PF6) (2 mol%), and K2CO3 (0.4 mmol) in DMSO (1.0 mL), 370 nm (10 W) at room temperature under N2 for 12 h. b Isolated yield. c np = no product; by-product 3a′ was obtained in 7% yield. d DMA (1.0 mL) as solvent. nr = no reaction. | ||
| 1 | None | 82 |
| 2 | No photocatalyst | 32 |
| 3 | In the dark | npc |
| 4 | Other bases instead of K2CO3 | 34–78 |
| 5 | DMF instead of DMSO | 62 |
| 6 | THF | 52 |
| 7 | DMA | 96 |
| 8 | K[2] instead of 2/K2CO3 | nr |
| 9d |
2/K2CO3 (1 : 1, 3.0 equiv.) |
64 |
| 10d | PC-2 instead of PC-1 | 62 |
| 11d | PC-3 | 78 |
| 12d | PC-4 | 38 |
| 13d | 390 nm instead of 370 nm | 93 |
| 14d | 427 nm | 75 |
| 15d | 456 nm | 54 |
| 16d | 2 h | 94 |
| 17d | 1 h | 76 |
With the optimized conditions in hand, we next investigated the substrate scope of the decarboxylative radical C(sp3)–C(sp3) cross-coupling reaction between GMDVs 1 and alkyl carboxylic acid 2. At first, we varied the GMDVs 1 and the results are summarized in Table 2. We were delighted to find that a wide variety of structurally diverse GMDVs bearing either electron-donating groups (Me and OMe), electron-withdrawing groups (Cl and Br) or electron-neutral groups (H) at various positions of the aryl ring smoothly reacted with 9H-fluorene-9-carboxylic acid 2, affording the corresponding cross-coupling products 3a–j in moderate to excellent yields (entries 1–10). Aliphatic GMDV 3k did not work well (entry 11), but heteroaromatic 3-thiophene was also tolerated in the reaction, affording the target product 3l in good yield (entry 12). Gratifyingly, ethyl ester and cyano substrates were also well tolerated, affording the target product 3m in 91% yield and product 3n in 97% yield (entries 13 and 14), respectively. Unfortunately, no desired product was observed when di-phenyl-substituted GMDV was introduced into the reaction (entry 15).
a
| Entrya | R1 | R2 | 3 | Yieldb (%) |
|---|---|---|---|---|
| a Standard reaction conditions: under a nitrogen atmosphere, 1 (0.1 mmol), 2 (0.4 mmol, 4.0 equiv.), Pd(PPh3)4 (5 mol%), Ir(dFMeppy)2(dtbbpy)(PF6) (2 mol%), and K2CO3 (0.4 mmol, 4.0 equiv.) in DMA (1.0 mL), blue LEDs (370 nm, 10 W) at room temperature for 2 h. b Isolated yield. c nd = not determined. | ||||
| 1 | Ph | CO2Me | 3a | 94 |
| 2 | 2-ClC6H4 | CO2Me | 3b | 86 |
| 3 | 2-BrC6H4 | CO2Me | 3c | 40 |
| 4 | 2-MeC6H4 | CO2Me | 3d | 81 |
| 5 | 2-MeOC6H4 | CO2Me | 3e | 85 |
| 6 | 4-MeC6H4 | CO2Me | 3f | 75 |
| 7 | 4-MeOC6H4 | CO2Me | 3g | 78 |
| 8 | 3,4-Me2C6H3 | CO2Me | 3h | 72 |
| 9 | 3,4-(MeO)2C6H3 | CO2Me | 3i | 89 |
| 10 | 3,4-(OCH2O)C6H3 | CO2Me | 3j | 70 |
| 11 | Bn | CO2Me | 3k | 25 |
| 12 | 3-Thiophene | CO2Me | 3l | 90 |
| 13 | Ph | CO2Et | 3m | 91 |
| 14 | Ph | CN | 3n | 97 |
| 15 | Ph | Ph | 3o | ndc |
Furthermore, different alkyl carboxylic acids were examined. As depicted in Scheme 2, the reaction proceeded well when primary carboxylic acids bearing electron-withdrawing groups (F, Cl, Br, and CF3), electron-donating groups (OMe and Ph) or electron-neutral groups (H) were applied, affording the desired products 5a–h in moderate yield and a byproduct 5′ was observed. Moreover, substrates bearing various heterocycles, such as indole 4l or thiophene 4m, could also work, affording the target products 5l andm in slightly lower yields. In addition, secondary benzylic-type carboxylic acids 4i andj also exhibited good reactivity. To our delight, commercially available and commonly-used anti-inflammatory drugs, such as naproxen 4k, were tolerated in this radical cross-coupling to afford the corresponding products. Finally, tertiary carboxylic acids were also investigated. To our satisfaction, the reactions could work well when flurenol 4n and 9-methylfluorene-9-carboxylic acid 4o were used as substrates, and the target products could be obtained in up to 50% yield. Unfortunately, the reaction did not work with inactivated aliphatic carboxylic acid 3p.
Under the standard conditions, the developed decarboxylative cross-coupling reaction of GMDVs 1a (1.08 g, 4.4 mmol) could be performed on a gram scale, affording the product 3a in 86% yield (Scheme 3A). In addition, a sunlight-accelerated experiment was carried out, affording the desired product 3a in 78% yield (Scheme 3B). To further confirm the structures of products 3 and 5, compound 3a was hydrolyzed with LiOH in a mixture of methanol and water to obtain carboxylic acid 6 in 89% yield. The solid structure of compound 6 (CCDC 2358852) was unambiguously assigned by X-ray crystallography (Scheme 3C-a).10 Treatment of compound 6 with DIBAL-H in THF at room temperature led to the reduction of the carboxyl group, providing alcohol 7 in 91% yield (Scheme 3C-b), which could be further oxidized with PCC in DCM to afford aldehyde 8 in 53% yield (Scheme 3C-c). Moreover, compound 3a could also react with 4-ClC6H4OH to afford product 9 in 69% yield (Scheme 3C-d).
To investigate the mechanism of this transformation, several control experiments were performed. Initially, three radical trapping experiments were conducted. Under the standard conditions, the addition of (2,2,6,6-tetramethylpiperidin-1-yl)oxidanyl (TEMPO), 2,6-di-tertbutyl-4-methylphenol (BHT) and 1,1-diphenylethylene to the reaction solution significantly suppressed these transformations, and the radical trapping intermediates 10–13 were detected by HRMS, respectively (Scheme 4A). These results, as well as the formation of byproduct 5′, indicated the presence of allyl and benzyl radicals in this reaction. Stern–Volmer quenching experiments indicate that a combination of K2CO3 and 2 quenches the photocatalyst (Fig. S8). Interestingly, when the combination of K2CO3 and 2 was replaced by fluorene-9-carboxylic acid 2 or its potassium salt, this reaction did not work, suggesting that K2CO3 should be used as a proton shuttle between 2 and Pd(II)-1,4-zwitterionic complexes. In addition, it was found that this reaction also worked in the absence of a photosensitizer, albeit in lower yield. This result indicates that in addition to Ir(III) photooxidants, Pd(I) species could also oxidize the carboxylate of 2 to generate benzyl radicals but with lower reactivity. Finally, we performed an on–off experiment. The reaction progressed smoothly upon irradiation with light (Fig. S9), while the reaction paused during the “light-off” periods. These results, as well as the quantum yield (1.73%), suggest that the present transformation was not a radical-chain process.
Based on the above mechanistic studies and previous reports,9 a mechanism for this decarboxylative cross-coupling is proposed (Scheme 4B). This transformation starts with the palladium-catalyzed decarboxylation of GMDVs to generate Pd(II)-1,4-zwitterionic complexes A, which are easily protonated to form intermediate B by the combination of K2CO3 and 2. The allyl Pd(II) intermediate B undergoes homolysis to form hybrid allyl Pd(I) radical C by utilizing light energy,10 followed by the generation of free allyl radical D and Pd(I). On the one hand, the generated Pd(I) complex slowly reacted with carboxylate E from fluorene-9-carboxylic acid 2 to regenerate the Pd(0) catalyst and benzyl radical F. On the other hand, a single electron transfer (SET) between carboxylate E and excited Ir(III) released radical F, along with the generation of Ir(II), which much faster reacted with Pd(I) to regenerate the Pd(0) catalyst and Ir(III) photosensitizer. Finally, the generated allyl radical D coupled with benzyl radical F to form the desired product 3a.
CCDC 2358852 contains the supplementary crystallographic data for this paper.11
| This journal is © the Partner Organisations 2025 |