Beom Ho
Kim‡
,
Byeongdo
Roh‡
,
Dong Jun
Kim
and
Hong Geun
Lee
*
Department of Chemistry, Seoul National University, Seoul 08826, Republic of Korea. E-mail: hgleee@snu.ac.kr
First published on 11th June 2025
The development of efficient methods for C(sp3)–C(sp3) bond formation remains a longstanding challenge in synthetic chemistry, especially in palladium catalysis employing sterically bulky electrophiles. In this study, we present a novel approach for achieving C(sp3)–C(sp3) cross-coupling via photoinduced palladium catalysis, employing cyclopropanols as masked C(sp3)-nucleophiles. Leveraging the unique reactivity of photoexcited palladium, this protocol enables radical-mediated C(sp3)–C(sp3) coupling across a broad range of substrates including sterically hindered and functionally diverse alkyl halides under mild conditions. This method significantly expands the extent of palladium-catalysed cross-coupling for bond construction between sp3-hybridized carbon units, providing streamlined access to structurally complex C(sp3)-rich frameworks that are crucial for medicinal chemistry.
Photoinduced palladium catalysis has recently emerged as a powerful tool to overcome the limitations of traditional methods that require the activation of C(sp3)-based electrophiles.7 Excited-state palladium complexes, generated by visible light irradiation, can activate C(sp3)-based precursors to form hybrid alkyl radical palladium species via single-electron oxidative addition (Fig. 1B, right). Subsequently, it serves as a valuable radical-based reaction partner for challenging bond-forming transformations.8–11 Nonetheless, the applications of photoinduced palladium catalysis for the formation of C(sp3)–C(sp3) bonds are still limited and the reactivity of hybrid alkyl radical palladium with C(sp3)-based nucleophilic coupling partners remains rare and elusive (Fig. 1C).12
Recognizing the critical importance of introducing sp3-hybridized carbon nucleophiles, the potential of cyclopropanols as latent homoenolate equivalents was investigated in the context of photoinduced palladium catalysis. Cyclopropanols serve as three-carbon synthons by undergoing ring-opening processes and have been utilized as valuable synthetic building blocks to accomplish the formation of various carbon–carbon bonds.13,14 In the present study, cyclopropanols were utilized as masked C(sp3)-based nucleophiles, facilitating the efficient formation of C(sp3)–C(sp3) bonds in conjunction with photoinduced palladium catalysis to achieve the activation of challenging electrophiles under mild reaction conditions (Fig. 1D). This approach underscores the potential of photoinduced palladium catalysis to induce an unprecedented bond formation event with a formally nucleophilic counterpart.
Entrya | Deviation from standard conditions | Yieldb |
---|---|---|
a Reaction conditions: 1a (0.1 mmol), 2a (2.0 equiv.), Pd(OAc)2 (10 mol%), rac-BINAP (20 mol%), Cs2CO3 (2.0 equiv.) in DCM (0.6 mL), rt, N2, 18 h, irradiated with blue LEDs (456 nm). b Determined by 1H NMR using 1,3,5-trimethoxybenzene as an internal standard; isolated yields given in parentheses. | ||
1 | None | 90 (85) |
2 | No rac-BINAP | 45 |
3 | XantPhos instead of rac-BINAP | 41 |
4 | t-Bu XantPhos instead of rac-BINAP | 15 |
5 | DPEPhos instead of rac-BINAP | 28 |
6 | PdI2, PdCl2(PhCN)2, PdCl2, Pd2dba3 | 74, 60, 39, 6 |
7 | Li2CO3, Na2CO3, K2CO3, CsOAc, CsOPiv | 30, 7, 7, 54, 36 |
8 | TEA, DIPEA | 28, 28 |
9 | No Pd(OAc)2 | <5 |
10 | No Cs2CO3 | 8 |
11 | No light irradiation | <5 |
12 | Pd/L = 5/10 mol% | 53 |
13 | 2.0 mmol scale | (88) |
![]() |
After establishing the optimal reaction conditions, we began to evaluate the applicability of the developed method (Table 2). Initially, we investigated the functional group tolerance of our system with a variety of substituted cyclopropanols. Cyclopropanols containing an electron-donating group, such as methyl (3ab), methoxy (3ac), or dimethylamino groups (3ad), were found to be effective reaction partners to afford the cross-coupling products. Moreover, cyclopropanols with an electron-withdrawing group (3ae) also underwent the desired transformation smoothly. Halogen-substituted substrates (3af, 3ag) were also compatible with the reaction conditions, indicating the orthogonal characteristics of the protocol with aryl halides. Cyclopropanol pronucleophiles with an extended π-system, such as naphthyl (3ah) or vinyl (3ai) moieties, also participated in the reaction with high efficiency. Notably, further addition reaction did not occur to the corresponding α,β-unsaturated carbonyl product, demonstrating the mildness of the developed protocol (3ai). Moreover, medicinally relevant heterocycles, including indole (3aj), pyrrole (3ak) or thiophene (3al), were well-tolerated during the transformations. Finally, various aliphatic cyclopropanols bearing cyclopentyl (3am), tetrahydropyranyl (3an), n-pentyl (3ao) or benzylic (3ap) substituents were also efficiently coupled with electrophiles, affording C(sp3)-rich ketones. In contrast, β-substituted cyclopropanols failed to participate effectively, presumably due to increased steric hindrance (ESI Table S5†).
a All reactions were performed on a 0.2 mmol scale. |
---|
![]() |
Subsequently, we surveyed the generality of the method in terms of the electrophilic reaction partner to find that a wide range of alkyl halides — primary, secondary, or tertiary — can readily undergo the desired transformation. Various sterically hindered tertiary halides performed well in this transformation (3bb–3bd). Additionally, cyclic secondary alkyl halides containing cyclopentyl (3be), cyclohexyl (3bf), tetrahydropyranyl (3bg) or piperidinyl (3bh) moieties provided the cross-coupling products with good to excellent yields. Additionally, an acyclic secondary alkyl halide (3bi) was shown to be compatible with the protocol. Lastly, primary alkyl halides were also successfully engaged in the photoinduced palladium catalysis, demonstrating the generality of the method. Primary alkyl halides substituted with arene (3bj), tetrahydrofuran (3bk) or silyl ether (3bl) were viable substrates for the reaction. Also, synthetically useful functional groups such as carboethoxy (3bm), cyano (3bn) or boronic acid pinacol ester (3bo) groups were shown to be compatible with the reaction conditions. Notably, alkyl substrates containing multiple carbon–halogen bonds afforded the desired product in a chemo-selective manner (3bp). Additionally, alkyl halides bearing α-heteroatom (3bq) or α-carbonyl (3br) substituents were efficiently coupled with cyclopropanols.
The applicability of the developed method toward complex bioactive architectures was also evaluated (Scheme 1A). A series of drug molecules and natural products, such as naproxen (4a), adapalene (4b) and cholic acid (4c), could be conveniently converted to the corresponding cyclopropanols. By subjecting the complex cyclopropanol pronucleophiles to the developed reaction protocol, the extended ketone derivatives of the parent precursors could be successfully prepared. Additionally, complex alkyl halides derived from bioactive molecules, such as indomethacin (4d), menthol (4e), or cholesterol (4f), were also proved to be competent coupling partners in this transformation, affording C(sp3)-rich architectures.
To further highlight the synthetic utility of the protocol, various downstream transformations of the β-alkylated ketone products were attempted (Scheme 1B). A substituted δ-keto ester (5aa), which could be conveniently prepared using the developed method, was subjected to either enantioselective carbonyl or imine reduction conditions to yield chiral δ-hydroxy (5ab) or δ-amino ester products (5ae), respectively (paths a and b). Subsequent acid- or base-mediated cyclization afforded lactone (5ac) and lactam (5af) products in enantiopure form. Through this divergent synthetic sequence, pharmaceutically important scaffolds could be readily synthesized, which would otherwise be difficult to access with conventional reactivity of ester enolates. Moreover, the carboxyl group of the product could be further exploited as a functional handle to furnish more complex molecular architectures. Another substituted δ-keto ester product (5ba) underwent straightforward hydrolysis to give the corresponding δ-keto acid (5bb). The reactive intermediate could be advanced to the amino ketone product (5bd) by Curtius rearrangement (path c). Alternatively, the carboxylic acid intermediate could be converted to its redox-active ester form (5be) to undergo a further electron-donor-acceptor (EDA) complex-mediated Giese addition (5bf, 5bg) (path d).16 Importantly, through the use of carboxylate as a secondary functional handle, it was possible to construct product structures that are either difficult to access using a halide-based precursor, due to their susceptibility to elimination (5bd), or resistant to undergo the desired transformation under the developed conditions (5bg).
The synthetic versatility of the developed method was further examined by evaluating its chemical orthogonality (Scheme 1C). Cyclopropanols bearing either a bromine (1q) or a boronic acid ester (1r) on its aryl substituent were subjected to the developed reaction conditions to furnish the cross-coupling products (6aa, 6ba) (Scheme 1C, a). The coupling products with an existing functional handle could be successfully converted into more complex biaryl ketones (6ab, 6bb) by the thermally-induced palladium-catalysed C(sp2)–C(sp2) cross-coupling. Analogously, alkyl halide electrophiles containing identical functional handles could be further elaborated in a similar manner (Scheme 1C, b). These examples demonstrate that the developed method is an orthogonal strategy that can be coupled with the conventional cross-coupling reactions based on the Pd(0)/Pd(II) cycle.
To gain mechanistic insight into the developed synthetic strategy, a number of control experiments were conducted (Scheme 2). The presence of 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), a radical scavenger, significantly inhibited reactivity, and the cross-coupled product (3ap) was not formed at all (Scheme 2A). Instead, TEMPO adducts derived from both cyclopropanol and alkyl halide were observed. At this point, it can be speculated that both the formally nucleophilic homoenolate, originating from the cyclopropanol precursor, and the alkyl counterpart participate in the catalytic cycle in the form of radical species. Furthermore, the incorporation of electron-rich or -deficient olefins, such as p-methoxystyrene or 2-benzylidenemalononitrile, as radical acceptors led to the formation of radical addition products from the alkyl halide precursors (7a and 7b) (Scheme 2B). In addition, competition experiments based on the simultaneous usage of the tertiary and secondary alkyl bromide substrates yielded only the cross-coupled product derived from the tertiary halide (3ab), suggesting the involvement of a thermodynamically more stable tertiary radical (Scheme 2C). These results reinforced the involvement of intermediary alkyl radical species.
Based on the mechanistic experiments and existing literature precedents,14c,d,17 a plausible reaction mechanism is depicted (Scheme 2D). Initially, alkyl halides undergo a single electron transfer mediated by a photoexcited Pd(0) complex A to generate hybrid alkyl Pd(I) radical B1, which is in equilibrium with ionic oxidative addition intermediate B2. Then, ligand exchange of the halide ligand with cyclopropanol results in the formation of a Pd(I) cyclopropoxide species C. The subsequent β-carbon elimination step produces palladium(I) homoenolate intermediate D. Finally, C(sp3)–C(sp3) bond-forming radical-induced reductive elimination occurs, yielding the β-alkylated ketone product and regenerating the Pd(0) catalyst. Indeed, alkyl palladium(I) complex D leads to the formation of alkyl palladium(II) complex D2 and hybrid β-keto radical palladium(I) species D1, which are in equilibrium with the corresponding palladium(I) intermediate, as supported by mechanistic investigations (Scheme 2A). The presence of the coupled product was detected when a bicyclic cyclopropanol was used as a mechanistic probe; nevertheless, the possibility of a radical ring-opening pathway cannot be completely ruled out at this stage (ESI Table S5†).
Footnotes |
† Electronic supplementary information (ESI) available: Experimental procedures, characterization data for newly synthesized compounds and other experimental details. See DOI: https://doi.org/10.1039/d5sc02325d |
‡ These authors equally contributed to this work. |
This journal is © The Royal Society of Chemistry 2025 |