Chiral alcohols are key structural units in many pharmaceutical molecules and fine chemicals. For instance, aprepitant (a drug used to treat chemotherapy-induced nausea and vomiting), duloxetine (a drug for depression), crizotinib (a drug for non-small cell lung cancer), and ezetimibe (a drug for hypercholesterolemia) all contain chiral alcohol structures (Figure 1).^[1-3] In recent years, a series of synthetic strategies for chiral alcohols, including asymmetric oxidation,^[4-5] have been developed. Among various approaches to chiral alcohol synthesis, asymmetric catalytic hydrogenation of prochiral ketones has been recognized as one of the most attractive methods due to its high atom economy and straightforward, efficient reaction pathway.

Within this research domain, catalysts formed by tridentate chiral ligands and transition metals demonstrate excellent chiral induction capability. In the 1990s, the Ru/ diphosphine-diamine catalytic system developed by Noyori et al.^[6-8] enabled the highly efficient asymmetric hydrogenation of ketones, which has since become a landmark contribution to this field. Following this breakthrough, a range of tridentate ligands featuring novel structural motifs have been developed in succession. In 1998, Zhang et al.^[9] reported the Ru complex of the N-N-N tridentate ligand ambox, which could facilitate the asymmetric hydrogenation of simple ketones under relatively elevated temperatures, achieving an enantiomeric excess (ee) value of 98%. In 2010, the same group^[10] further designed the Ru-Indan- ambox catalyst, optimizing the reaction conditions to ambient temperature while maintaining an ee value of 97%. In 2011, Zhou et al.^[11] developed a class of chiral tridentate phosphine-nitrogen-nitrogen ligands (SpiroPAP). The complex formed by these ligands with Ir exhibited ultrahigh catalytic activity in the hydrogenation of acetophenone, with a turnover number (TON) reaching up to 4550000, thus establishing a record within this domain. In 2020, the group^[12] further designed sterically hindered SpiroPNP ligands, which successfully enabled the highly stereoselective synthesis of various chiral dialkyl alcohols with an ee value exceeding 99%.

Despite the notable successes of the aforementioned systems, chiral P-N-O tridentate ligands that possess structural diversity, facile synthesis, and both high activity and high enantioselectivity still remain relatively scarce, with relatively limited reports on systematic investigations in this regard.^[13-17] Between 2009 and 2010, the Liu group^[18-20] reported a series of P-N-O ligands with novel structures, which demonstrated favorable catalytic performance in asymmetric Baylis-Hillman reactions. Nevertheless, systematic studies exploring the application of such high-performance P-N-O ligands in the hydrogenation of simple aromatic ketone substrates have yet to be reported.

Building upon this background, the present study performed structural modifications on this class of P-N-O ligands, resulting in the development of a series of P-N-O tridentate ligands (Figure 2), which were subsequently applied to the Ir-catalyzed asymmetric hydrogenation of simple aromatic ketones. The influences of the electronic effects and steric hindrance of the ligands on catalytic performance were systematically examined, while the substrate scope of the catalytic system was assessed—with the goal of providing a new catalytic system for the green and highly efficient synthesis of chiral aryl alcohols.

Ligands L1~L7 were first synthesized, and subsequently, with acetophenone (1a) as the model substrate, the performance of the catalytic systems formed by these ligands with iridium in asymmetric hydrogenation reactions was systematically evaluated, while the reaction conditions were optimized.

Initially, the ligand structures were screened. As shown in Entry 1 of Table 1, the unmodified ligand L1 afforded the best reaction performance, achieving 99% conversion and 93% enantioselectivity. Subsequently, different substituents were introduced onto the benzene ring of the phenolic hydroxyl group to investigate electronic and steric hindrance effects. The results indicated that none of the modifications—whether introducing electron-withdrawing groups (e.g., L5, L6, L7), electron-donating groups (e.g., L2, L3), or increasing the steric hindrance of substituents (e.g., L4)—resulted in catalytic performance superior to that of L1 (Entries 2~6). In particular, introducing a methoxy group (L3) resulted in a significant decrease in both catalytic activity (76% conversion) and enantioselectivity (82% ee, Entry 3). This suggests that slight changes in ligand structure exert a significant impact on catalyst performance, with L1 being the optimal structure among this series of ligands.

After L1 was identified as the dominant ligand, base additives were further screened. The results (Entries 8~15) revealed that the nature of the base is crucial to both reaction activity and enantioselectivity. When Na₂CO₃ was used, the reaction barely proceeded (only 6% conversion, Entry 9). In contrast, with t-BuONa, t-BuOK, Cs₂CO₃, LiOH, KOH, NaOH, or CH₃ONa, most systems maintained high conversion, but enantioselectivity decreased to varying degrees (ee values ranging from 52% to 87%).

Ultimately, Ba(OH)₂ was identified as the optimal base additive, as it could simultaneously ensure high reaction activity and high enantioselectivity.

In solvent screening (Table 2, Entries 1~13), it was found that alcoholic solvents are critical for achieving efficient hydrogenation. Aprotic solvents such as dichloromethane (DCM), 1,2-dichloroethane (DCE), toluene, dimethylformamide (DMF), and tetrahydrofuran (THF) all led to catalyst deactivation (Table 2, Entries 1~3, 5~6). Among alcoholic solvents, 95% ethanol exhibited the best performance, increasing the ee value to 94.4% while maintaining conversion at >99% (Table 2, Entry 9). The use of methanol or isopropanol resulted in a slight decrease in enantioselectivity (Table 2, Entries 4, 7). Notably, the addition of an appropriate amount of water to ethanol facilitated the improvement of enantioselectivity; however, excessively high content of water (e.g., ≥15%) led to significant decreases in both activity and selectivity (Table 2, Entries 11~13). This underscores the practical advantage of employing inexpensive and readily available 95% ethanol as the solvent.

The influence of hydrogen pressure was then investigated (Entries 14~15). The reaction performed best under a hydrogen pressure of 5.0 MPa; either increasing (6.0 MPa) or decreasing (3.0 MPa) the pressure resulted in a slight decrease in enantioselectivity. Further decreasing the reaction temperature to 0 ℃ (Entry 16) increased the ee value to 95% without affecting catalytic activity. Exploration of catalyst loading (Entries 17~18) showed that when the substrate-to-catalyst molar ratio (S/C) was increased to 1000, no significant decrease in catalytic activity and en-antioselectivity was observed. However, when the S/C was further increased to 2000, both conversion and enantioselectivity decreased significantly. Finally, a gram-scale experiment conducted under optimized conditions (Entry 19, substrate: 0.96 g) successfully afforded the target product with a 99% isolated yield and 95% ee, demonstrating the potential of this catalytic system for practical applications.

After establishing the optimal conditions for the model substrate acetophenone (1a), we sought to investigate the substrate generality of this catalytic system. To establish a unified and efficient protocol for evaluating various substrates while ensuring high conversion, the conditions were fine-tuned. It was found that increasing the catalyst loading to 1.0 mol% (while maintaining the reaction temperature at 0 ℃) could significantly enhance the reaction efficiency, affording excellent conversion for the vast majority of substrates. Accordingly, a catalyst loading of 1.0 mol% and a temperature of 0 ℃ were adopted as the standard conditions for investigating the substrate scope. The reaction results are summarized in Table 3.

It was demonstrated that the system exhibited excellent catalytic activity and enantioselectivity for aromatic ketones with different alkyl chain lengths (yields>99%, ee values of 93%~96%). Both the position and electronic nature of substituents on the benzene ring significantly influenced the reaction outcomes. For substrates with ortho-substituents, both reaction activity and enantioselectivity decreased significantly (2d~2f), which may be attributed to the steric hindrance effect induced by the ortho substituents, hindering the effective coordination of the substrate to the chiral catalyst’s active site. In contrast, substrates with meta- or para-substituents—whether electron-donating groups (e.g., methyl, methoxy) or electron-withdrawing groups (e.g., fluoro, chloro, bromo, trifluoromethyl)—afforded good to excellent ee values (92%~97%). Among these, the para-fluorine-substituted substrate (2m) performed best with an ee value of 97%.

The catalytic system was also applicable to other heteroaromatic ketones, such as 2-acetonaphthone (2q), 2-acetylthiophene (2r), and 2-acetylfuran (2s). Although the enantioselectivities for the hydrogenation of 2-acetona- phthone and 2-acetylfuran were not satisfactory, the system exhibited high catalytic activity for 2-acetylthiophene (99% yield) and achieved an ee value of 81%.

Notably, 3-(dimethylamino)-1-(thiophen-2-yl)propan-1- one (2u) could also undergo hydrogenation smoothly, affording (S)-3-(dimethylamino)-1-(thiophen-2-yl)propan-1- ol—a key precursor of the drug duloxetine^[21]—in 98% yield and 90% ee. The successful application of this method further highlighted its practical value and potential in the synthesis of important chiral pharmaceutical intermediates. It should be noted that the catalytic system exhibited poor performance when applied to the catalytic hydrogenation of non-aromatic ketone substrates. For instance, no significant catalytic activity was observed for any of the investigated substrates, such as 1-cyclohexylethan-1-one and 1-cyclohexylpropan-1-one.

An efficient catalytic system based on the chiral binaphthyl P-N-O tridentate ligand L1 and iridium was successfully developed in this study for the asymmetric hydrogenation of simple aromatic ketones. Through systematic optimization, optimal reaction conditions—emp- loying Ba(OH)₂ as the base and cost-effective, green 95% ethanol as the solvent. For the acetophenone model substrate, this system exhibited a conversion rate of over 99% and an enantioselectivity of 95%. Ligand screening results indicated that the parent structure L1 showed the best performance, and any modification at the ortho-position of the phenolic hydroxyl group led to a decrease in catalytic efficiency. This reveals the critical role of ligand structural precision in determining catalytic performance. It also highlights the unique advantages of the ligand synthetic route adopted in this study, including its conciseness and the achievement of high enantioselectivity without complex modifications. Further substrate scope studies demonstrated that the catalytic system had good generality for a series of aromatic ketones and heteroaromatic ketones with different electronic properties and substitution positions, achieving a maximum enantioselectivity of 97%. Notably, the synthetic practicality and potential of this method were fully verified by the successful implementation of a gram- scale experiment and its application in the asymmetric synthesis of (S)-duloxetine precursor—a key pharmaceutical intermediate—with 98% yield and 90% ee, thereby directly confirming the system’s application potential in the synthesis of important drug intermediates.

In summary, this study not only provides a novel and practical catalytic system for the green and efficient synthesis of chiral aryl alcohols by utilizing an environmentally friendly, low-cost ethanol solvent and a concise ligand design, but also further highlights the great application prospects of chiral P-N-O tridentate ligands in the field of asymmetric catalysis.