Introduction

Chiral phosphines have found numerous applications in asymmetric catalysis used as ligands and catalysts, both on laboratory scale and in industrial processes, providing invaluable access to chiral molecules in a range of areas, including drugs and polymers as well as agrochemicals synthesis1,2,3,4,5,6. Thus the construction of chiral phosphines has been a field of intense research interest and development. During the infancy of the field, Knowles and co-workers successfully introduced the highly active DIPAMP ligand for rhodium catalyst in the first practical and highly enantioselective synthesis of l-DOPA7,8. Noyori and co-workers successfully introduced the BINAP in Ru-catalyzed asymmetric hydrogenation9,10,11 casting a profound influence on asymmetric catalysis, thus leading to the demand for synthetic methods for accessing to chiral phosphine molecules. Among the chiral phosphines developed in the past decades, the exploited axial-chirality, C-stereogenic center, or planar chirality, are frequently used as the basic chiral elements constituting chiral phosphine ligands such as BINAP and DIOP. For an increasing number of asymmetric catalytic syntheses, combinations of the different chiral elements have proven advantageous. For instance, TangPhos and DuanPhos featuring C-stereogenic centers and P-chiral phosphorus atoms, have found great applications in asymmetric catalysis12.

On the other hand, chiral monodentate biaryl phosphines have attracted intense attention as chiral ligands over the past decades. For example, MOP13 and KenPhos14 with axial chirality, SegePhos15 with point chirality, BI-DIME16 with P-stereogenic center (Fig. 1a). However, phosphines containing both axial chirality and a P-stereogenic center have been less studied. In this regard, only a few types of such ligands were available due to the lack of general and efficient methods. Buchwald successfully prepared P-chirogenic binaphthyl-substituted monophosphines showing great application potential17. Cramer and coworkers disclosed the efficient and highly enantioselective C–H arylations of phosphine oxides with o-quinone diazides leading to the construction of monodentate chiral phosphorusIII compounds with biaryl ligand backbones and chirality on P-atom (Fig. 1a)18. He and Liu reported the construction of P-stereogenic center and N–C axis enabled by asymmetric allylic substitution-isomerization19. Thus, the development of efficient protocols, providing access to structurally diverse phosphines containing both P- and axial chirality is a highly desirable goal.

Fig. 1: Diversity-oriented synthesis of structurally diverse monodentate biaryl phosphines containing both P- and axial chirality.
figure 1

a Monodentate biaryl phosphines with various chiral elements. b Transition-metal catalyzed achiral alkynylation of the C–P bond. c Enantioselective C–P bond arylation (our previous work). d Diversity-oriented synthesis of structurally diverse P- and axially chiral monodentate biaryl phosphines (this work).

Full size image

Due to the great interest and importance of axially chiral biaryl compounds20,23,24,25,26,27,1, entries 1–6, for details, see Supplementary Information). To our delight, the reaction underwent smoothly to deliver the desired product 3a in 86% yield and 91% ee using chiral phosphoramidite L6 as the ligand (Table 1, entry 6). Further optimization of the additives indicated that CuBr was the best, providing 3a in 90% yield and 91% ee (Table 1, entry 7). Other solvents were then surveyed but failed to improve the reactivity or enantioselectivity (Table 1, entries 9–12). The bases were also evaluated indicating that K2CO3 or K3PO4 gave better enantioselectivity but with lower yield (Table 1, entries 13 and 14). Further investigations revealed that 2.5 mol% catalyst loading with CuBr (1 equiv) was optimal providing the desired product 3a in 83% yield and 93% ee at 45 °C for 36 h (Table 1, entry 18).

Table 1 Optimization studies for enantioselective alkynylation of C–P bond with terminal alkynes
Full size table

Reaction scope of enantioselective alkynylation of C–P bond with terminal alkynes

With the optimal conditions in hand (Table 1, entry 18), we next set out to explore the substrate scope of this reaction by examining derivatives of biaryl phosphonium salts 1 and terminal alkynes 2. Because of the product’s sensitivity to air, they were more conveniently isolated as the corresponding air-stable phosphine sulfides or boranes by workup with S8 or BH3·SMe2 respectively. Overall, all the tested substrates gave the expected MOPs in high yields with excellent atroposelectivities and enantioselectivities (Fig. 2). A wide range of terminal alkynes with different substituents or substitution patterns were evaluated under the standard conditions (3c–3s). Aryl terminal alkynes that bear a substituent at the ortho, para, or meta positions were all well tolerated, and the corresponding products were furnished in high yields with high ee (3c–3p). The 2-ethynylnaphthalene was also tolerated, providing the desired product 3q in 55% yield and 94% ee. Then 2-ethynylthiophene and 3-ethynylthiophene were evaluated, and the corresponding products were furnished in high yields with high ee (3r, 3s). In addition, the TMS-alkynyl product 3t could also be afforded in 65% yield and 92% ee with bis(trimethylsilyl)acetylene as the coupling partner. Furthermore, several substituents were also well tolerated, such as fluoride (3d, 3p), chloride (3e, 3o), bromide (3c), OMe (3g, 3l, 3n), NH2 (3i), OH (3j), and TMS (3t) functionality.

Fig. 2: Substrate scope of enantioselective alkynylation of C–P bond.
figure 2

Reaction conditions: 1 (0.20 mmol), 2 (2.0 equiv), [Pd(Allyl)Cl]2 (2.5 mol%), L6 (11 mol%), CuBr (1.0 equiv), Cs2CO3 (2.0 equiv), 2-Me-THF (2.0 mL) at 45 °C for 36 h, then S8 (5 equiv) or BH3·SMe2 (2 equiv, 10 M in Me2S), unless otherwise stated. Isolated yields were reported. ee values of the major isomers are shown and determined by chiral HPLC analysis. dr values were determined by 1H NMR analysis of the crude reaction mixtures; a60 h; b1 (0.05 M); cthe reaction was conducted at 60 °C; dwith bis(trimethylsilyl)acetylene as the coupling partner; e0.10 mmol scale.

Full size image

To further expand the generality of this strategy, a wide range of phosphonium salts 1 with different substituents were tested affording the corresponding products in high yields and excellent enantioselectivities of up to 97% ee (3u–3ac). The introduction of substituents at 7- and 7’-position on the aromatic ring of naphthalene was also applicable providing the desired product 3ae in 63% yield with 95% ee. Furthermore, double-substituted phosphonium salts at 6- and 6’-postions were also investigated, and the corresponding products were observed in high yields with high ee values (3af–3aj). In addition, the hindered biphenyl phosphonium salts were also tolerated providing the desired products in moderate to good yields and high ee values (3ak–3al). Finally, the reaction was scaled up to the 2 mmol scale and the corresponding product 3a was afforded an 82% yield with 92% ee. High diastereoselectivities (>25:1 dr) were obtained for most cases. The absolute configuration of the product was confirmed by single-crystal X-ray diffraction analysis (Figs. 2, 3c, CCDC 2227153).

Fig. 3: Substrate scope of enantioselective silylation of C–P bond.
figure 3

Reaction conditions: 1 (0.20 mmol), R3Si-Bpin 4 (2.0 equiv), [Pd(Allyl)Cl]2 (2.5 mol%), L6 (11 mol%), CuCl (0.75 equiv), Cs2CO3 (2.0 equiv), 2-Me-THF (2.0 mL) at 60 °C for 12 h, then S8 (5 equiv) or BH3·SMe2 (2 equiv, 10 M in Me2S), unless otherwise stated. Isolated yields were reported. ee values of the major isomers are shown and determined by chiral HPLC analysis. dr values were determined by 1H NMR analysis of the crude reaction mixtures; athe reaction was conducted at 45 °C for 36 h; b0.10 mmol scale.

Full size image

Reaction scope of enantioselective silylation of C–P bond

After the enantioselective alkynylation of the C–P bond was achieved, we became interested in develo** an enantioselective silylation of the phosphonium salt 1 providing the corresponding MOPs 5 containing both P- and axially chirality. As important building blocks, organosilicon compounds have found numerous applications in organic synthesis, material science, and medicinal chemistry owing to their high stability and low toxicity37,38. Therefore, the development of efficient protocols for the construction of C–Si bond is of great interest and desirable. In addition, the introduction of silicon groups into a chiral ligand or catalyst can tune the steric phenomenon thus enhancing the efficiency of the chiral induction. To our delight, the reaction of phosphonium salts 1 and silyborane 4 as the silicon source under similar conditions as the enantioselective alkynylation of C–P bond provides the desired products 5 in high yields and excellent enantioselectivities (Fig. 3). A wide range of phosphonium salts 1 with different substituents were tolerated affording the corresponding products in high yields and excellent enantioselectivities of up to 96% ee (5a–5o). The introduction of substituents on the aromatic ring of naphthalene was also applicable providing the desired products in high ees (5p–5s). In addition, the hindered biphenyl phosphonium salts were also tolerated providing the desired products in good yields and high ee values (5t–5u). The structure of the product was confirmed by single-crystal X-ray diffraction analysis, the stereogenic axis of 5o was determined as R, and the chirality at the P-atom is the R configuration (Figs. 3, 5o, CCDC 2227154).

Reaction scope of enantioselective borylation of C–P bond

Arylboronic esters play important roles in a variety of fields ranging from organic synthesis to drug discovery and material science39,40,41. Considering the generality of enantioselective alkynylation and silylation of C–P bond, we set our sights on develo** the enantioselective borylation of biaryl phosphonium salts 1 leading to the ortho-borylated MOPs 6 with C–B bond as a versatile handle for further synthetic elaborations to construct structurally diverse MOPs containing both P- and axial chirality. To our delight, the reaction of phosphonium salts 1 and diboron esters B2neop2 as the boron source under the modified conditions provide the desired products 6 in moderate to high yields and excellent enantioselectivities (Fig. 4). In addition, the hindered biphenyl phosphonium salt was also tolerated providing the desired product 6g in 41% yield and 96% ee. It should be noted that the enantioselective borylation of the C–P bond was much more sensitive to the space steric of the substitutes on the phosphorous atom. For example, the borylated product 6f was only afforded in 38% yield but with high enantioselectivity of 86% ee.

Fig. 4: Substrate scope of enantioselective borylation of C-P bond.
figure 4

Reaction conditions: 1 (0.10 mmol), B2neop2 (2.0 equiv), [Pd(Allyl)Cl]2 (5 mol%), L5 (22 mol%), CuCl (0.75 equiv), Cs2CO3 (2.0 equiv), p-xylene (1.0 mL) for 12 h, then S8 (5 equiv) or BH3·SMe2 (2 equiv, 10 M in Me2S), unless otherwise stated. Isolated yields were reported. ee values of the major isomers are shown and determined by chiral HPLC analysis. dr values were determined by 1H NMR analysis of the crude reaction mixtures; athe reaction was conducted at 35 °C for 36 h; b0.10 mmol scale.

Full size image

Reaction scope of enantioselective reduction of C–P bond with water as a Hydride Source

Diboron–H2O has been recently used as a hydride source in Pd-catalyzed transfer hydrogenation of alkenes and alkynes by Stokes and coworkers42. Zhu and coworkers reported the elegant palladium-catalyzed enantioselective reductive Heck reactions via the reduction of C(sp3)–Pd bond with water as a hydride source assisted by diboron43. However, to the best of our knowledge, the enantioselective reduction of C(sp2)–Pd bond via the cleavage of C–P bond with water as a hydride source has never been achieved. Water is without doubt the most environmentally benign and cost-efficient hydride source. We became interested in develo** an enantioselective reductive C–P bond cleavage of the phosphonium salt 1 thus providing the corresponding MOPs 7 containing both P- and axially chirality. To our delight, the reaction underwent smoothly to deliver the desired product 7a in 91% yield and 91% ee under simple modified reaction conditions with TBME instead of 2-Me-THF as the solvent at 60 °C for 12 h. Then a wide range of phosphonium salts 1 with different substituents were tested affording the corresponding products in high yields and excellent enantioselectivities of up to 96% ee (Fig. 5). The introduction of substituents on the aromatic ring of naphthalene was also applicable providing the desired products in high ees (7n–7s). In addition, the hindered biphenyl phosphonium salts were also tolerated providing the desired products in good yields and high ee values (7t–7u). The stereogenic axis of 7h was determined as R, and the chirality at the P-atom is the R configuration by X-ray diffraction analysis (CCDC 2227155).

Fig. 5: Substrate scope of enantioselective reduction of C–P bond.
figure 5

Reaction conditions: 1 (0.20 mmol), B2pin2 (1.2 equiv), H2O (4.0 equiv), [Pd(Allyl)Cl]2 (5.0 mol%), L6 (22 mol%), CuCl (0.75 equiv), Cs2CO3 (2.0 equiv), TBME (2.0 mL) at 60 °C for 12 h, then S8 (5 equiv) or BH3·SMe2 (2 equiv, 10 M in Me2S), unless otherwise stated. Isolated yields were reported. ee values of the major isomers are shown and determined by chiral HPLC analysis. dr values were determined by 1H NMR analysis of the crude reaction mixtures; athe reaction was conducted at 30 °C for 24 h; bthe reaction was conducted at 45 °C for 24 h; c0.10 mmol scale.

Full size image

Synthetic utility and control experiments

To demonstrate the synthetic potential of this strategy, we readily converted the alkyne moiety in the enantio-enriched product 3a to E-alkenyl group in 8 in one reduction step (Fig. 6a), which can offer an opportunity to access to P, alkene-bidentate ligand in the transition-metal catalyzed reactions. In addition, the formed C–B bond can be readily converted into C–O bond by oxidation of 6a with NaBO3·4H2O providing the hydroxylation product 9 in 49% yield and 96% ee (Fig. 6b). Furthermore, the formed C–B bond can also be readily converted into C–H bond providing 7i in 88% yield (Fig. 6c), thus the absolute configuration of boronated products 6 was confirmed as the same with that of products 7. The stereogenic axis of 6 was determined as R, and the chirality at the P-atom is the R configuration.

Fig. 6: Synthetic utility and control experiments.
figure 6

a Reduction of the alkyne moiety to E-alkenyl group; b Oxidation of the C–B bond to C–O bond; c Converted the C–B bond to C–H bond; d Preparation of the D-substituent product; e The non-linear effect studies of the enantioselective alkynylation of C–P bond with terminal alkynes. f Control study of the protonation of C–B bond; g Control study of the reduction of C–P bond without B2pin2.

Full size image

For the enantioselective alkynylation of the C–P bond with terminal alkynes, a linear correlation between the ee values of the product 3a and phosphoramidite L6 was observed, suggesting that monomeric PdL* (L* = phosphoramidite L6) might be involved in the enantio-determining steps (Fig. 6e). Treatment of 1a using HCOONa as a hydride donor under the standard conditions afforded 7i in 22% 1H NMR yield with 67% ee (for details, see the Supplementary Information). 7i-D bearing a D substituent was afforded with 90% D-incorporation using D2O instead of H2O under standard conditions (Fig. 6d), showing that H2O is the hydride source of this reaction. In addition, 7i was not observed when 6a was treated with H2O for 12 h under the standard reaction conditions, suggesting that protonolysis of 6a is not responsible for the formation of 7i (Fig. 6f). Furthermore, the reaction of 1a with water in the absence of B2pin2 under standard conditions failed to afford compound 7i, suggesting that B2pin2 plays an important role in the enantioselective reduction of C–P bond (Fig. 6g).

The chiral phosphines L7–L11 were synthesized via the deprotection of the phosphine sulfides or boranes treated with Raney Ni or DABCO respectively (see Supplementary Information for details)34. With the rapidly constructed ligand library in hand, we next tested it in the asymmetric construction of functionalized bicyclic imides via [3 + 2] annulation of MBH carbonate (Table 2). For the reaction between the MBH carbonate 10 and N-methylmaleimide 11, MOP was not effective and no desired product was observed (Table 2, entry 1). To our delight, with the phosphine L11 as the catalyst, the functionalized bicyclic imides 1244 was obtained in 94% ee and 93% yield. This example proved the potential of the structurally diverse chiral monodentate biaryl phosphines in asymmetric catalytic reactions.

Table 2 Testing the developed ligand library in [3 + 2] annulation of MBH carbonate
Full size table

Based on the above control experiments, and the previous reports32,33,34,43, a possible reaction mechanism was proposed to explain the reaction processes. As shown in Fig. 7, phosphonium salts (Sa)-1a and (Ra)-1a were in rapid equilibrium with each other via the rotation of the C–C single bond45. The preferentially oxidative addition of Pd0 with substrate (R)−1a cleaved C–P bond a to form enantio-enriched biaryl intermediate A. the C-P bond b was in high steric hindrance due to the naphthyl and tBu groups, that’s why the reaction shows high dr value. Subsequent transmetallation of intermediate A with the terminal alkynes, R3Si-Bpin or diboron esters generated the PdII species B (Fig. 7(I)). Reductive elimination of B furnished the chiral phosphine product, which would give the final product 3, 5, or 6 after being treated with S8 or BH3 with concurrent regeneration of Pd0L* catalyst. In addition, as shown in Fig. 7(II), The reaction of PdII intermediate A and water afforded complex C. The transmetallation of complex C with the B2pin2 provided complex D. The reaction of complex D and water would furnish complex E. 1,4-D migration from E provided complex F. Finally, reductive elimination of F would deliver the desired product 7 after treated with S8 or BH3.

Fig. 7: Proposed reaction mechanism for the enantioselective cross-coupling and reduction of C–P bond.
figure 7

Cycle I: Phosphonium salts (Sa)-1a and (Ra)-1a are in rapid equilibrium with each other via the rotation of the C–C single bond. The oxidative addition of Pd0 with substrate (R)-1a cleaves C–P bond a to form enantio-enriched biaryl intermediate A. Subsequent transmetallation of intermediate A with the terminal alkynes, R3Si-Bpin or diboron esters form the PdII species B. Reductive elimination of B furnishes the chiral phosphine product with concurrent regeneration of Pd0L* catalyst. Cycle II: The reaction of PdII intermediate A and water affords complex C. The transmetallation of complex C with the B2pin2 provides complex D. The reaction of complex D and water furnishes complex E. 1,4-D migration from E provides complex F. Reductive elimination of F delivers the desired product 7 with concurrent regeneration of Pd0L* catalyst.

Full size image

Discussion

In summary, we report an enantioselective and diastereoselective cleavage of the C–P bond for the synthesis of structurally diverse monodentate biaryl phosphines containing both P- and axial chirality which are expected to be useful chiral ligands or organocatalysts for enantioselective catalysis. These reactions feature broad substrate scope and mild conditions leading to the formation of C(sp2)–C(sp) bond, C(sp2)–Si bond, C(sp2)–B bond, and C(sp2)–H bond. A series of structurally diverse monodentate biaryl phosphines are obtained in excellent yields with high diastereo-and enantioselectivities. The D-substituted monodentate biaryl phosphines can be easily synthesized using D2O as a D-donor with a high level of D-incorporation. The P- and axially chiral monodentate biaryl phosphines can serve as excellent catalysts in asymmetric [3 + 2] annulation of MBH carbonate affording the chiral functionalized bicyclic imide, which proves the potential of the structurally diverse chiral monodentate biaryl phosphines in asymmetric catalytic reactions.

Methods

Materials

Unless otherwise noted, materials were purchased from commercial suppliers and used without further purification. All the solvents were treated according to general methods. Flash column chromatography was performed using 200–300 mesh silica gel. See Supplementary Methods for experimental details.

Procedure for enantioselective alkynylation of C–P bond with terminal alkynes

To a 10 mL Schlenk tube containing anhydrous Cs2CO3 (0.4 mmol, 2 equiv) was added phosphonium salts 1 (0.2 mmol), CuBr (0.2 mmol, 1.0 equiv), [Pd(allyl)Cl]2 (0.005 mmol, 2.5 mol%), chiral phosphine ligand L6 (0.022 mmol, 11 mol%), 2-Me-THF (2 mL, 0.1 M), and terminal alkyne (0.4 mmol, 2 equiv) sequentially under nitrogen. The Schlenk tube was then sealed and stirred for 36 h at 45 °C. The reaction mixture was cooled to rt, S8 (5 equiv, 1 mmol) or BH3•SMe2 (1 equiv, 10 M in Me2S) was added and stirred for 2 h at rt. The reaction mixture was then filtered through a pad of celite eluenting with CH2Cl2/EtOAc (20 mL). The filtrate was concentrated, and the residue was purified by silica gel chromatography to afford the corresponding product.

Procedure for enantioselective silylation of C-P bond

To a 10 mL Schlenk tube containing anhydrous Cs2CO3 (0.4 mmol, 2 equiv) was added phosphonium salts 1 (0.2 mmol), CuCl (0.15 mmol, 0.75 equiv), [Pd(allyl)Cl]2 (0.005 mmol, 2.5 mol%), chiral phosphine ligand L6 (0.022 mmol, 11 mol%), 2-Me-THF (2 mL, 0.1 M), and R3Si-Bpin (0.4 mmol, 2 equiv) sequentially under nitrogen. The Schlenk tube was then sealed and stirred for 12 h at 60 °C. The reaction mixture was cooled to rt, S8 (5 equiv, 1 mmol) or BH3•SMe2 (2 equiv, 10 M in Me2S) was added and stirred for 2 h at rt. The reaction mixture was then filtered through a pad of celite eluenting with CH2Cl2/EtOAc (20 mL). The filtrate was concentrated, and the residue was purified by silica gel chromatography to afford the corresponding product.

Procedure for enantioselective borylation of C–P bond

To a 10 mL Schlenk tube containing anhydrous Cs2CO3 (0.2 mmol, 2 equiv) was added phosphonium salts 1 (0.1 mmol), B2(OR)2 (0.2 mmol, 2 equiv), CuCl (0.075 mmol, 0.75 equiv), [Pd(allyl)Cl]2 (0.005 mmol, 5 mol%), chiral phosphine ligand L5 (0.022 mmol, 22 mol%), and p-xylene (1 mL, 0.1 M) sequentially under nitrogen. The Schlenk tube was then sealed and stirred for 36 h at 35 °C. The reaction mixture was cooled to rt, S8 (5 equiv, 1 mmol) or BH3•SMe2 (2 equiv, 10 M in Me2S) was added and stirred for 2 h at rt. The reaction mixture was then filtered through a pad of celite eluenting with CH2Cl2/EtOAc (20 mL). The filtrate was concentrated, and the residue was purified by preparative TLC.

Procedure for enantioselective reduction of C–P bond with water as a Hydride Source

To a 10 mL Schlenk tube containing anhydrous Cs2CO3 (0.4 mmol, 2 equiv) was added phosphonium salts 1 (0.2 mmol), B2pin2 (0.24 mmol, 1.2 equiv), CuCl (0.15 mmol, 0.75 equiv), [Pd(allyl)Cl]2 (0.01 mmol, 5 mol%), chiral phosphine ligand L6 (0.044 mmol, 22 mol%), TBME (4 mL, 0.05 M), and H2O (14 μL, 4 equiv) sequentially under nitrogen. The Schlenk tube was then sealed and stirred for 12 h at 60 °C. The reaction mixture was cooled to rt, S8 (5 equiv, 1 mmol) or BH3•SMe2 (2 equiv, 10 M in Me2S) was added and stirred for 2 h at r.t. The reaction mixture was then filtered through a pad of celite eluenting with CH2Cl2/EtOAc (20 mL). The filtrate was concentrated, and the residue was purified by silica gel chromatography to afford the corresponding product.