The asymmetric hydrogenation of quinolines is one of the most convenient route to synthesize enantiomerically pure tetrahydroquinoline derivatives which are important organic synthetic intermediates and structural units of alkaloids and biologically active compounds ^[1]. The first successful homogeneous asymmetric hydrogenation of quinolines was described by Zhou and co-workers using a Ir catalytic system ^[2]. Quinolines with different substituents were investigated in their following works ^{[3, 4, 5, 6, 7]}. Chan’s group subsequently described that the air-stable and recyclable Ir-P-Phos and Ir- diphosphinite are efficient catalysts for the asymmetric hydrogenation of quinolines ^{[8, 9, 10]}. Reetz and co-workers also found that BINOL-derived diphosphonites with achiral P-ligands as additive catalyzed the same reaction ^[11]. Despite these important advances, a problem in the reported reaction systems was that good results were only obtained at a low substrate-catalyst ratio. The problem with a high catalyst loading was attributed to the formation of an irreversible iridium dimer and trimer during the reaction which would be pathways for catalyst deactivation ^[12]. Therefore, developing a highly efficient catalytic system for the asymmetric hydrogenation of quinolines is still needed.

Several attempts to avoid iridium dimer formation were tried (Scheme 1). One method is the introduction of bulky substituents onto the chiral ligands to inhibit the formation of inactive dimmers and trimers, thus promoting the catalytic activity (Scheme 1.1). Fan’s group synthesized a series of dendritic dendrimer G_nDenBINAP ligands by the condensation of the dendritic wedges G_n-COOH with (S)-5,5’-diamino BINAP (S)-1 ^[13]. The G_nDenBINAP ligands gave much higher catalytic activity than BINAP through the steric shielding effect of the bulky dendritic wedges to reduce dimerization (Scheme 1.1(a)). Zhou and coworkers reported that introducing bulky groups onto the coordination phosphorus atoms effectively blocked the formation of inactive dimer and trimer species and improved the activity of the Ir catalysts ^[14] (Scheme 1.1(b)).

Scheme 1. Strategies for inhibiting the dimerization of the Ir-complex in the hydrogenation of quinolone.

Recently, we reported the synthesis of a series of CMPs by embedding the chiral BINAP ligand into CMP networks. These solid materials are efficient catalysts for the asymmetric hydrogenation of β-keto esters ^[15]. Because all the BINAPs are embedded on the pore structure of the networks separately, we envisioned that these heterogeneous catalysts ^{[16, 17, 18]} with an inherent isolation effect can perform the asymmetric hydrogenation of quinoline with high activity through preventing the formation of dimers and trimers. (Scheme 1.2) Here, we reported the results.

Scheme 2. Structure of BINAP-CMPs.

A typical strategy for synthesizing BINAP-CMP-3D-2 was as follows: 1,3,5,7-Tetrakis(4-ethynylphenyl) adamantine, (R)- 4,4-DibromoBINAPO, tetrakis-(triphenylphosphine) palladium, and copper iodide were dissolved in a mixture of dioxane and Et₃N. The reaction mixture was heated to 80 °C and stirred for 72 h under a nitrogen atmosphere. The precipitate was filtered and washed with solvent. After reducing with HSiCl₃, BINAP-CMP-3D-2 was obtained.

We began by examining the hydrogenation of 2- methylquinoline 1a with an iridium catalyst in the presence of different BINAP-CMPs. We were pleased to find that the hydrogenation catalyzed by the BINAP-CMPs proceeded smoothly and afforded the product with moderate enantioselectivity (Table 1, entries 1-4). In sharp contrast, the homogeneous hydrogenation catalyzed by BINAP gave trace product under same reaction conditions (Table 1, entry 5). Moreover, the catalytic activity gradually increased with increasing surface area and pore volume of the BINAP-CMPs. CMPs with larger surface area and pore volume benefit the diffusion of substrate and product, which makes it easier to access the active center and improve the activity.

Table 1 Comparison of catalytic activity of the synthesized BINAP-CMPs. Entry BINAP-CMPs ABET a (m2/g) Vtotal b (cm3/g) Conversion c (%) TOF d (h-1) ee e (%) 1 BINAP-CMP-S 170 0.37 9 180 48 2 BINAP-CMP-M 318 0.46 10 200 50 3 BINAP-CMP-3D-1 279 0.76 12 240 54 4 BINAP-CMP-3D-2 398 0.77 17 340 54 5 BINAP — — 5 100 70 Reaction conditions: quinaldine 0.25 mmol, CH2Cl2 3 mL, [Ir(COD)Cl]2 0.000125 mmol, I2 0.05 mmol, 1.5 MPa H2, 25 °C, 0.5 h. a Surface area calculated from the N2 adsorption isotherm using the BET method. b Total pore volume at p/p0 = 0.99. c Determined by 1H NMR analysis of the crude product. d Average TOF over the reaction time. e Determined by HPLC analysis with a Chiralpak OJ-H column.

Table 1 Comparison of catalytic activity of the synthesized BINAP-CMPs.

To obtain higher enantioselectivity, the effects of the solvent, temperature and pressure were investigated by using BINAP-CMP-3D-2 as the catalyst (Table 2). A series of organic solvents were investigated. CH₂Cl₂ was the best in terms of both conversion and enantioselectivity (Table 2, entries 1-8). The enantioselectivity of the reaction was slightly increased at low temperature with incomplete conversion (Table 2, entry 13). Also, the reaction did not reach completion under a low pressure(Table 2 entry 12).

Table 2 Asymmetric hydrogenation of quinaldine (1a) catalyzed by the BINAP-CMP-3D-2 catalyst. Entry Solvent Conversion a (%) ee b(%) 1 THF 99 34 2 CH2Cl2 99 60 3 MeOH 88 29 4 toluene 99 40 5 benzene 99 38 6 dioxane 99 39 7 CH2Cl2/THF 1:2 99 56 8 CH2Cl2/toluene 1:2 99 51 9 c,d CH2Cl2 99 69 10 c,e CH2Cl2 99 70 11 c CH2Cl2 99 70 12 c,f CH2Cl2 93 70 13 c,g CH2Cl2 95 73 14 c,h CH2Cl2 99 69 Reaction conditions: quinaldine 0.25 mmol, CH2Cl2 3 mL, [Ir(COD)Cl]2 0.000125 mmol, I2 0.05 mmol, 4 MPa H2, 25 °C, 18 h. a,b See Table 1. c [Ir(COD)Cl]2 0.00125 mmol, 2 h. d 6.5 MPa H2. e 5 MPa H2. f 1.5 MPa H2. g 0 °C. h 50 °C.

Table 2 Asymmetric hydrogenation of quinaldine (1a) catalyzed by the BINAP-CMP-3D-2 catalyst.

Using the optimized reaction conditions, the limit of catalyst activity was also investigated. The BINAP-CMP-3D-2 catalyst was found to be highly effective even at a high substrate/catalyst ratio although the enantioselectivity decreased with a low catalyst loading (Table 3).

Table 3 Minimum amount of Ir/BINAP-CMP-3D-2 catalyst that can be used. Entry Time (h) Substrate/Catalyst Conversion a (%) ee b(%) 1 2 100:1 99 70 2 2 500:1 99 64 3 2 1000:1 99 60 4 d 6 2000:1 99 53 5 e 24 5000:1 93 42 Reaction conditions: quinaldine (1a) 0.25-5 mmol, CH2Cl2 3 mL, I2 0.05 mmol, 4 MPa H2, 25 °C. a,b See Table 1.

Table 3 Minimum amount of Ir/BINAP-CMP-3D-2 catalyst that can be used.

Using the optimal conditions, we demonstrated the scope of quinoline for this heterogeneous hydrogenation (Table 4). All substituted quinolines used here were hydrogenated with good conversion and modest enantioselectivity. The reaction was relatively insensitive to the length of the 2-alkylated side chain of the quinoline. It was interesting that with increasing length of the side chain of R₂, the enantioselectivity obtained with BINAP-CMP-3D-2 increased from being slightly less than to be obviously higher than the enantioselectivity of the homogeneous Ir/BINAP catalyst (Table 4, entries 1-5). The conversion decreased with the increasing of steric hindrance (Table 4, entries 4-6), which may be due to that bulky substituted quinolines suffer more diffusion resistance when diffusing through the pore structure of the polymer. With 6-substituted quinolines as the substrate (Table 4, entries 7-9), low enantioselectivity with complete conversion were observed.

Table 4 Catalytic asymmetric hydrogenation of quinoline derivatives. Entry R1/R2 (substrates) Conversion a (%) ee b(%) Configuration c Ir/BINAP-CMP-3D-2 Ir/BINAP Ir/BINAP-CMP-3D-2 Ir/BINAP 1 H/Me (1a) 99 24 70 72 R 2 H/Et (1b) 99 23 77 78 R 3 H/n-Pr (1c) 99 16 78 74 R 4 H/n-Bu (1d) 80 18 77 75 R 5 H/i-Pr (1e) 83 13 79 75 R 6 H/Ph (1f) 84 34 44 50 R 7 F/Me (1g) 99 23 63 77 S 8 Me/Me (1h) 99 16 67 76 R 9 MeO/Me (1i) 97 9 70 78 R Reaction conditions: quinoline 0.25 mmol, CH2Cl2 3 mL, [Ir(COD)Cl]2 0.00125 mmol, I2 0.05 mmol, 4 MPa H2, 25 °C, 2 h. a,b See Table 1. d The absolute configuration is assigned by comparison of the HPLC retention time with those reported in the literature data.

Table 4 Catalytic asymmetric hydrogenation of quinoline derivatives.

The recyclability of the BINAP-CMP-3D-2 catalyst was also investigated (Table 5). Upon completion of the reaction, theIr/BINAP-CMP-3D-2 catalyst was easily recovered by centrifugation and a regular filter. The colorless filtrate from the asymmetric hydrogenation of quinaldine (1a) did not afford any additional product, showing the heterogeneous nature of the reaction system. After washing with CH₂Cl₂ and heating under vacuum, the solid catalyst was reused for the next cycle. We did not see any significant deterioration in the activity for the recovered catalyst even after five cycles.

Table 5 Recycling experiment of the BINAP-CMP-3D-2 catalyst. Cycle 0 1 2 3 4 5 6 Conversion a (%) 99 99 99 99 99 99 90 ee b(%) 72 72 71 70 70 70 69 Reaction conditions: quinaldine 0.5 mmol, CH2Cl2 3 mL, [Ir(COD)Cl]2 0.00125 mmol, I2 0.05 mmol, 4 MPa H2, 25 °C, 2 h. a,b See Table 1.

Table 5 Recycling experiment of the BINAP-CMP-3D-2 catalyst.

In summary, we synthesized a series of BINAP-CMPs with different surface areas, which gave high activity and modest enantioselectivity in the Ir-catalyzed asymmetric hydrogenation of quinolines. The catalytic activity was related to the structural properties of the BINAP-CMPs. The CMP with a higher surface area and pore volume gave better asymmetric hydrogenation results. Through preventing the formation of dimmers by the spatial isolation effect of the CMP, the BINAP-CMPs gave much higher activity than the homogeneous BINAP ligand.