Open Access Article
Yuchen
Yang
a,
Jia-Long
Wu
a,
Sheng-Ye
Zhang
b,
Xu
Liu
a,
Teng
Sun
a,
Yanan
Zhao
*b and
Lijia
Wang
*a
aShanghai Engineering Research Center of Molecular Therapeutics and New Drug Development, Shanghai Frontiers Science Center of Molecule Intelligent Syntheses, East China Normal University, 3663 North Zhongshan Road, Shanghai 200062, China. E-mail: ljwang@chem.ecnu.edu.cn
bState Key Laboratory of Organometallic Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 345 Lingling Lu, Shanghai 200032, China
First published on 27th November 2025
Chiral vicinal amino alcohols are pivotal building blocks in organic synthesis and pharmaceutical research. Herein, we report a chiral copper-catalyzed two-step one-pot strategy for the efficient synthesis of diaryl chiral vicinal amino alcohols, featuring excellent diastereoselectivity (>99
:
1 dr) and enantioselectivity (up to 99.5% ee). A key discovery is the critical role of fluorinated counteranions, which significantly enhance both reactivity and stereocontrol—an underappreciated effect in copper-catalyzed asymmetric reactions. DFT calculations reveal that secondary Cu–F interactions, combined with the spatial confinement of hexafluorophosphate, fine-tune the catalytic chiral environment, enabling precise stereocontrol via modulation of π–π stacking and weak non-covalent interactions. This strategy exhibits broad substrate scope, accommodating diverse aryl, heteroaryl, and functionalized substituents, and allows gram-scale synthesis with facile deprotection to free amino alcohols. The mechanistic insights into counterion effects highlight counterion engineering as a powerful tool for optimizing asymmetric catalysis.
The strategy of synthesizing chiral vicinal amino alcohols via nucleophilic asymmetric ring opening of epoxides or aziridines with nitrogen or oxygen nucleophiles represents one of the most straightforward and atom-economical approaches. Among these strategies, a powerful one is the desymmetrization of meso-substrates, as depicted in Fig. 1a. Pioneering work in this area was reported by Jacobsen and coworkers in 1998, who described the first chromium-catalyzed enantioselective nitrogen-nucleophilic ring opening of meso-epoxides.9 Subsequently, Schneider et al.10 and Feng et al.11 independently developed asymmetric nitrogen-nucleophilic ring opening of meso-epoxides using chiral scandium(III) and indium(III) complexes, respectively. Alternatively, Toste et al.12 and List et al.13 utilized chiral phosphoric acid (CPA) catalysts to achieve the synthesis of chiral vicinal amino alcohols via alcohol-nucleophilic ring opening of meso-aziridiniums and meso-aziridines, respectively. More recently, Feng et al.14 and Wang et al.15 disclosed the oxygen-nucleophilic ring opening of meso-aziridines catalyzed by chiral magnesium complexes, further expanding the scope of this desymmetrization strategy. A complementary approach involves the kinetic resolution of racemic unsymmetric epoxides or aziridines (Fig. 1b). For example, Yamamoto and co-workers developed a chiral tungsten-catalyzed kinetic resolution of racemic 2,3-epoxy alcohols with aromatic amines for the construction of chiral vicinal amino alcohols.16 In a related study, Cao et al. reported an efficient chiral phosphoric acid-catalyzed water-nucleophilic ring opening of racemic aziridines, which also proceeds via a kinetic resolution pathway.17 Although the aforementioned strategies for synthesizing chiral vicinal amino alcohols are direct and efficient, they suffer from limitations: being restricted to desymmetrization of meso-type substrates and, for unsymmetric substrates, requiring more than 2 equivalents of racemic starting materials to proceed via kinetic resolution.
![]() | ||
| Fig. 1 Synthesis of chiral vicinal amino alcohols via asymmetric ring-opening of epoxides/aziridines. | ||
In 1991, Evans and co-workers first discovered the Cu(I)-catalyzed amination of trimethylsilyl enol ethers to afford N-Ts amino ketones, postulating that the reaction proceeds via an aziridine ring-opening rearrangement pathway analogous to Rubottom oxidation.18 This pioneering work initiated research into the asymmetric catalysis of this reaction.19 In 2018, Read de Alaniz et al. developed a highly efficient Cu(I)-catalyzed asymmetric electrophilic α-amination of silyl enol ether derivatives via the nitrosocarbonyl hetero-ene reaction (Fig. 1c)20 More recently, a landmark study by Matsunaga and colleagues demonstrated the chiral ruthenium-catalyzed amination of alkyl silyl enol ethers (Fig. 1c).19f However, no successful examples have been reported for diaryl-substituted silyl enol ether substrates. We conceived that developing a strategy involving asymmetric aziridination of silyl enol ethers with metal nitrenes,21 combined with one-pot stereoselective reduction, could provide an alternative and efficient pathway for the synthesis of chiral diaryl vicinal amino alcohols (Fig. 1d). Although the hexafluorophosphate anion (PF6−) is widely employed as a counterion in copper-catalyzed asymmetric reactions, its role in asymmetric catalytic processes has long been overlooked.22 This is likely due to the fact that PF6− is typically classified as a weakly coordinating anion (WCA) and was specifically designed to minimize interference with catalytic processes. Herein, we found that fluoro-containing counteranions can promote both the reactivity and enantioselectivity of the reaction. DFT calculations revealed that the interaction between a fluorine atom and the central copper metal can fine-tune the chiral environment of the catalytic center, thereby enabling the chiral copper-catalyzed one-pot synthesis of a series of diaryl chiral vicinal amino alcohols with excellent diastereo- and enantioselectivity.
:
4 dr (entry 7). Notably, during this process, we observed that the introduction of fluorine atoms into the counteranion significantly enhanced both the yield and enantioselectivity of the reaction. Using copper trifluoromethanesulfonate, the desired product 3a was afforded in 26% yield with 77% ee and a 90
:
10 dr (entry 8). In contrast, cuprous trifluoromethanesulfonate improved the yield of 3a to 44% with enhanced stereoselectivity (86% ee and 96
:
4 dr, entry 9). We further explored other fluorine-containing counteranions. The use of Cu(CH3CN)4BF4 led to a slight improvement in enantioselectivity to 88% ee (entry 10), while Cu(CH3CN)4PF6 resulted in further enhancements in both the yield and ee value, affording 3a in 87% yield with 96% ee and a 97
:
3 dr (entry 11). To further verify the influence of fluorine atoms on the catalytic center, we tested the counteranion BArF−, which might be spatially remote from the reaction center, and observed a significant decrease in both the yield and enantioselectivity of 3a (59% yield and 76% ee, entry 12). Subsequently, we examined the effect of the chiral backbone of the ligands on the reaction and found that, among the chiral bisoxazoline (BOX) ligands tested, the one with a tert-butyl backbone (L1) outperformed those with other chiral backbones, such as isopropyl, benzyl, phenyl, and indenyl (L2–L5, entries 13–16). Thus, we identified Cu(CH3CN)4PF6 as the optimal copper salt and L1 as the optimal chiral ligand. Notably, we also investigated the effects of different reducing agents on the diastereoselectivity of the product. Although a 97
:
3 dr can also be achieved when LiAlH4 is used as the reducing agent (entry 17), subsequent studies revealed that the diastereoselectivity of the reaction proved to be unsatisfactory when this reducing agent is applied to substrates with other substituents. Optimization of other reaction conditions can be found in the SI.
| Entry | [Cu] | L | Yieldb (%) | eec (%) | drd |
|---|---|---|---|---|---|
a Unless otherwise noted, all reactions were carried out with 1a (0.2 mmol), 2 (0.3 mmol), 4 Å MS (100 mg), and Cu salt/L (1 : 1.2, 10 mol%) in PhCl (2.0 mL) under a N2 atmosphere at 0 °C for 24 h, and then with Zn(BH4)2 (1.0 mmol, 1.0 M in THF), TBAF (0.6 mmol, 1.0 M in THF) and EtOH (1.0 mL) at 0 °C to rt for 4 h.
b Isolated yield.
c Determined by chiral HPLC.
d Determined by crude 1H NMR.
e LiAlH4 (1.0 mmol) was used instead of Zn(BH4)2.
|
|||||
| 1 | CuSO4 | L1 | N.D. | — | — |
| 2 | Cu3(PO4)2 | L1 | N.D. | — | — |
| 3 | CuCN | L1 | Trace | 59 | — |
| 4 | CuDPP | L1 | Trace | 68 | — |
| 5 | CuTC | L1 | Trace | 76 | — |
| 6 | CuBr·SMe2 | L1 | Trace | 49 | — |
| 7 | Cu(ClO4)2·6H2O | L1 | 32 | 75 | 96 : 4 |
| 8 | Cu(OTf)2 | L1 | 26 | 77 | 90 : 10 |
| 9 | CuOTf | L1 | 44 | 86 | 96 : 4 |
| 10 | Cu(CH3CN)4BF4 | L1 | 34 | 88 | 94 : 6 |
| 11 | Cu(CH3CN)4PF6 | L1 | 87 | 96 | 97 : 3 |
| 12 | Cu(BArF) | L1 | 59 | 76 | 97 : 3 |
| 13 | Cu(CH3CN)4PF6 | L2 | 34 | 71 | 93 : 7 |
| 14 | Cu(CH3CN)4PF6 | L3 | 55 | 78 | 96 : 4 |
| 15 | Cu(CH3CN)4PF6 | L4 | 51 | 78 | 94 : 6 |
| 16 | Cu(CH3CN)4PF6 | L5 | 32 | 70 | 96 4 |
| 17e | Cu(CH3CN)4PF6 | L1 | 80 | 97 | 97 : 3 |
|
|||||
This one-pot strategy for the synthesis of chiral amino alcohols exhibits good substrate adaptability. Under the optimized reaction conditions (Table 1, entry 8), we explored structurally diverse diaryl-substituted enol silyl ethers (Table 2). When R1 is 1-naphthyl and R2 is 2-naphthyl, the corresponding amino alcohol 3b was obtained in 93% yield with a 96
:
4 dr and 97% ee. Replacing the 2-naphthyl group with 5-benzofuranyl, 5-benzothienyl, or 5-benzodihydrofuranyl resulted in negligible changes in enantioselectivity, and the corresponding amino alcohols 3c–3e were still afforded with 97–98% ee. Substrates 1f and 1g bearing more complex 3-fluorenyl and 2-dibenzofuranyl substituents were also compatible with this reaction, achieving 97% ee and 96% ee, respectively. When R2 is a simple phenyl group, 3h was obtained in 96% yield with a 98
:
2 dr and 98% ee. For substrate 1h, we investigated the effect of different reducing agents on diastereoselectivity. NaBH4 only resulted in diminished diastereoselectivity (74
:
26 dr), which is presumably due to the fact that zinc metal might be more ready to form chelating interactions with the substrate, which facilitates the control of diastereoselectivity during the reduction process. Interestingly, an inversion of diastereoselectivity (20
:
80 dr) can be achieved when DIBAL-H is used instead of Zn(BH4)2. Detailed results are available in the SI. Introduction of halogen substituents at different positions of the benzene ring (including fluorine, chlorine, and bromine) enabled the synthesis of the corresponding chiral amino alcohols 3i–3n with 94–99% yields, 96
:
4 to >99
:
1 dr, and 93–96% ee. Incorporation of electron-withdrawing trifluoromethyl groups at the para- and meta-positions of the benzene ring led to a slight decrease in yield and enantioselectivity, giving 3o and 3p in 72% yield with 92% ee and 85% yield with 90% ee, respectively. Introduction of methyl groups at the ortho-, para-, and meta-positions of the benzene ring had almost no impact on the enantioselectivity of the reaction, and 3q–3s were obtained with 97–98% ee. Although the introduction of an electron-rich methoxy group at the ortho-position of the benzene ring caused a decrease in enantioselectivity (3t, 90% ee), the corresponding product was obtained with exclusive enantioselectivity when the methoxy group was introduced at the meta-position (3u, 99.5% ee). For the para-methoxy-substituted substrate (3v), the reaction activity decreased slightly (84% yield), while the diastereoselectivity remained exclusive (>99
:
1 dr) and the enantioselectivity showed no significant change (97% ee). Other substituents on the benzene ring were also well-tolerated. For example, the 3,4-dimethoxy-substituted product 3w was obtained in 73% yield with a 93
:
7 dr and 95% ee; the para-trifluoromethoxy-substituted substrate afforded the corresponding amino alcohol 3x in 67% yield with a 98
:
2 dr and 92% ee. Reactive functional groups such as carboxyl, protected alcohol and amine derivatives were tolerated, affording 3y–3aa with 90–95% ee. Bpin- and TMS-substituted substrates were also explored to give the corresponding 3bb and 3cc with 97–99% ee. Introduction of phenyl groups at the ortho-, meta-, and para- positions of the benzene ring led to the corresponding biphenyl amino alcohols 3dd–3ff with 93–95% ee. When the phenyl group was replaced with a heteroaromatic ring, the reaction system remained compatible, and the thiophene-substituted amino alcohol 3gg was synthesized in 96% yield with a 97
:
3 dr and 97% ee. The tert-butyl-aryl silyl enol ether was tried as substrate, and the desired product 3hh could be obtained in 91% ee and >99
:
1 dr with 31% yield.
a Unless otherwise noted, all reactions were carried out with 1 (0.2 mmol), 2 (0.3 mmol), 4 Å MS (100 mg), and Cu(CH3CN)4PF6/L1 (1 : 1.2, 10 mol%) in PhCl (2.0 mL) under a N2 atmosphere at 0 °C for 24 h, and then Zn(BH4)2 (1.0 mmol, 1.0 M in THF), TBAF (0.6 mmol, 1.0 M in THF) and EtOH (1.0 mL) were added at 0 °C to rt for 4 h; Isolated yield; dr value was determined by crude 1H NMR; ee value was determined by chiral HPLC.
b −10 °C.
c
2 was added in four portions.
|
|---|
|
On the other hand, the Ar1 group can also be extended in a variety of ways. For instance, naphthyl substrates substituted with fluorine, chlorine, or bromine atoms afforded the corresponding amino alcohols 3ii–3kk in 96–99% yields with excellent diastereoselectivity and 97% ee. Naphthyl substrates with electron-donating substituents also performed well, giving 3ll and 3mm with 97–98% ee. Other more complex substituents were well compatible with the reaction, and the dibenzofuranyl-substituted product 3nn was obtained with 94% ee. This method also allowed for the convenient preparation of various dimethyl-substituted phenyl amino alcohols 3oo–3qq as well as 1,2-diaryl-substituted amino alcohol 3rr with 95–97% ee. Meanwhile, the absolute configuration of product 3a was confirmed to be (1R, 2S) by X-ray diffraction analysis of its single crystal.
To gain insights into the fluoride benefited reaction mechanism, density functional theory (DFT) calculations were performed on the pathway of the asymmetric aziridination reaction between copper nitrenes and silyl enol ethers, based on the reported computational literature studies (calculation details and the reaction potential energy surface calculations are provided in the SI).23 DFT calculations revealed that the initiating species of the reaction is PF63[Cu]NTs, corresponding to the structure where the triplet nitrene and ligand are coordinated to Cu+, respectively. During the reaction, for the major and minor enantiomers, the rate-determining step involves the formation of the first N–C bond via the transition states (TS) with the highest energy barrier, PF63TS1 (ΔG≠ = 8.4 kcal mol−1) and PF63TS'1 (ΔG≠ = 11.0 kcal mol−1), which thus serves as the chirality determining step. The calculated ΔΔG≠ for this step is 2.6 kcal mol−1, consistent with the experimental 98% ee (Fig. 2a).
Notably, calculations indicated that in the initiating species PF63[Cu]NTs, one of the F atoms of the anion exhibit certain coordination interactions with Cu, with a Cu–F distance of 2.42 Å (significantly shorter than the sum of the van der Waals radii of Cu and F, 2.87 Å). Such secondary interactions enable the hexafluorophosphate ion to occupy the space above the PF63[Cu]NTs species, confining the Ts group to the region below the anion (Fig. 2a). The identification of these Cu–F secondary interactions corroborates the experimental observation that fluorinated counterions influence reaction activity and enantioselectivity. In the TS of the major enantiomer (PF63TS1), as the silyl enol ether substrate approaches the reaction center, the spatial confinement by the hexafluorophosphate ion induces a conformational change in the Ts group, facilitating its deformation and π–π stacking with the naphthyl ring of the silyl enol ether—this contributes to the lower energy of PF63TS1. As shown in Fig. 2a, the dihedral angle ∠N–S–C1–C2 (θ) changes from −91.7° in PF63[Cu]NTs to −75.3° in PF63TS1, corresponding to a 16.4° rotation of the benzene ring in the Ts group around the S–C1 bond. In contrast, no such π–π stacking is observed in PF63TS'1, and the variation in ∠θ is comparatively small.
The presence of the hexafluorophosphate ion also affects the conformation of the BOX ligand. IGMH (Independent Gradient Model based on Hirshfeld partition) analysis revealed two pairs of C⋯H weak interactions and two pairs of H⋯H weak attractive interactions between the benzene ring and the ligand in PF63TS1, which favor its lower energy barrier. In contrast, the weak interactions between the naphthyl ring and the ligand in PF63TS'1 are less pronounced. Based on the above IGMH analysis, the weak interaction sites in PF63TS1 are located at the ortho- and para-positions of the phenyl substituent. Thus, introducing an electron-donating substituent (e.g., methoxy) at the meta-position of the benzene ring is expected to enhance these electrostatic attractions, thereby lowering the energy barrier of PF63TS1 and increasing the energy barrier difference between PF63TS1 and PF63TS'1 (i.e., enhancing the ee value). Calculations showed that the energy barrier difference increases from 2.6 kcal mol−1 to 3.5 kcal mol−1, consistent with the experimental improvement in the ee value from 98% to 99.5% (Fig. 2b). Conversely, introducing an electron-withdrawing substituent (e.g., trifluoromethyl) at this position might have the opposite effect: the calculated energy barrier difference decreased to 2.0 kcal mol−1, and the experimental ee value dropped from 98% to 92% (Fig. 2b). These computational results are in good agreement with the experimental data, validating the reliability of the theoretical model.
This method provides a practical approach for the gram-scale synthesis of disubstituted chiral amino alcohols.
With 1,2-Dinaphthyl-substituted amino alcohol 3a, 1,2-diaryl-substituted amino alcohol 3rr, or the product 3h with two different substituents, the reaction can be successfully scaled up to the gram level, affording 79–94% yield with 3.42 g to 3.92 g (Fig. 3). Moreover, the resulting products can be isolated and purified by recrystallization to afford optically pure products with >99% ee and >99
:
1 dr. It is worth noting that these N-Ts protected amino alcohols can undergo mild and convenient deprotection under Red-Al conditions, thereby yielding the corresponding optically pure chiral amino alcohols 4a, 4rr, and 4h in 60–70% yields.
:
1 dr and 99.5% ee). A key innovation is the identification of fluorinated counteranions as critical regulators of reactivity and enantioselectivity—an underappreciated role in copper catalysis. DFT calculations revealed that secondary Cu–F interactions, coupled with the spatial confinement of hexafluorophosphate, fine-tune the catalytic chiral environment to enable precise stereocontrol, as validated by IGMH analysis. This method features broad substrate scope, gram-scale accessibility, and facile deprotection to free amino alcohols, offering a robust platform for accessing valuable chiral vicinal amino alcohols. Importantly, it highlights counterion engineering as a powerful tool for optimizing asymmetric catalytic performance.
CCDC 2478618 (3a) contains the supplementary crystallographic data for this paper.24
| This journal is © The Royal Society of Chemistry 2026 |