Introduction

Transition-metal-catalyzed inert C–H bond direct functionalization has become a powerful tool in modern synthetic chemistry1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17. One of the remained challenges for the widespread application of C–H bond functionalization is how to control the regioselectivity and stereoselectivity. Up to now, extensive attention has been paid on this field of research and significant results have been achieved18,19,20,21,22. Recently, we developed a Pd(II)-catalyzed C–H direct arylation of ferrocenes with thioketones as effective directing groups23. The importance of planar chiral ferrocenes24,25,26,27 prompted us to explore a corresponding enantioselective arylation reaction. In this regard, Pd(II)/monoprotected amino acid (MPAA) catalytic system28,29,30,31,32,33,34,35,36, established by Yu and coworkers, was firstly examined as our initial attempts for this asymmetric C–H bond arylation. The desired product could be obtained but without any enantioselectivity. Chiral phosphate anion as a counteranion with palladium has also been demonstrated as an effective catalytic system in a number of asymmetric C–H bond functionalization reactions37,38,39,40,41. Unfortunately, no positive results on enantioselective control were obtained when various chiral phosphoric acids (CPAs) were introduced (Fig. 1a). Therefore, the development of enantioselective C–H arylation reaction to provide structurally diverse planar chiral ferrocenes is highly desirable42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61.

Fig. 1
figure 1

Thiocarbonyl-chelation-assisted C–H bond arylations. a Failed attempts via palladium(II) catalyzed direct C–H arylation. b Enantioselective C–H arylation of ferrocenes enabled by rhodium(I) and monodentate phosphonite

Recently, Rh(I)-catalyzed enantioselective C–H functionalizations progressed rapidly. In 1997, Murai and coworkers reported Rh(I)-catalyzed asymmetric C–H/olefin coupling reaction with modest enantioselectivity62. Later, Ellman, Bergman and coworkers achieved a remarkable progress on Rh(I)-catalyzed enantioselective intramolecular C–H alkylation reaction, which gave cyclic products with up to 96% ee63. Several elegant Rh(I)-catalyzed asymmetric intramolecular dehydrogenative silylations have been demonstrated by the groups of Takai64,65, He66, and Shibata67, respectively. Notably, Glorius and coworkers have made breakthroughs that the combination of a rhodium(I) precatalyst with chiral N-heterocyclic carbene (NHC) or monodentate phosphonite ligand enabled highly enantioselective Csp3–H arylations68,69.

Inspired by these pioneering results, we explore Rh(I) catalyzed enantioselective arylation of thiocarbonylferrocenes (Fig. 1b) after achieving the synthesis of axially chiral heterobiaryls via isoquinoline directed C-H functionalization70. In the presence of Rh(I) catalyst derived from a chiral phosphonite, direct arylation of thioketone substituted ferrocenes with readily available aryl iodides proceeds in excellent enantioselectivity. Herein, we report the results of this study.

Results

Optimization of reaction conditions

We began our initial attempts by utilizing chiral diphosphines L1-2 (BINAP, DIOP) as the ligand. However, no product was observed when thiocarbonylferrocene 1a was treated with 1.1 equiv of iodobenzene 2a in the presence of 5 mol% [Rh(C2H4)2Cl]2 and 3.0 equiv of LiOtBu in THF at 80 °C (Table 1, entries 1–2). When Feringa ligand L3 was used, to our delight, planar chiral product (3a) was obtained in 19% yield with 88% ee (Table 1, entry 3). Stereochemical induction is predominantly determined by the BINOL backbone of the phosphoramidite ligand by comparing the results of that of diastereomeric ligand L4 (Table 1, entry 4, 15% yield, −71% ee). The utilization of TADDOL-derived phosphoramidite L5 could further increase the yield to 29% (Table 1, entry 5). Notably, 3a was isolated in 46% yield and 90% ee when TADDOL-derived phosphonite L7, introduced by the Glorius group69, was used (Table 1, entry 7). Promoted by these encouraging results, we further investigated other TADDOL-derived phosphonite ligands (Table 1, entries 8–13). It was found that the chiral backbone assembled with 2-naphthyl groups delivered 3a in 26% yield with 91% ee (Table 1, entry 8). Further evaluation of substituents [Ar = 3,5-(CF3)2C6H3 and 3,5-(tBu)2C6H3] on the TADDOL backbone revealed that both ligands could promote the reaction (33% and 44% yield, respectively) albeit with decrease of enantioselectivity (Table 1, entries 9–10). Ligand L11 bearing a tbutyl group could not promote this reaction (Table 1, entry 11). On the other hand, varying the aromatic moiety on the P atom afforded product 3a with comparative results (45–46% yields, 90% ee) as those of L7 (Table 1, entries 12–13). Thus, the easily accessible ligand L7 was chosen as the optimal one. The addition of molecular sieves was found to increase the yield of 3a (Table 1, entry 14, 59% yield with 92% ee). Increasing the amount of iodobenzene 2a to 1.3 equivalents also has a positive influence on the reaction outcome (Table 1, entry 15, 78% yield, 94% ee). Other molecular sieves did not give a noticeable improvement of the yield. Various solvents were next screened, and it was found that the reaction in dioxane afforded 3a in 76% yield with 97% ee (Table 1, entry 18, see the Supplementary Table 1 for complete optimization). The ratio between Rh(I) and ligand had a profound effect on the results. The use of 20 mol% of L7 ([Rh]/L = 1/2) provided 3a without erosion in term of yield and enantioselectivity (Table 1, entry 17, 75% yield, 97% ee). The loading of L7 could be further reduced to 15 mol % (Table 1, entry 18, [Rh]/L = 1/1.5), 76% 97% ee). However, both yield and enantioselectivity were decreased when 10 mol% of L7 ([Rh]/L = 1/1) was used (Table 1, entry 19).

Table 1 Optimization of the reaction conditions[a]

Substrate scope

Under the above optimized reaction conditions [1a (0.2 mmol), 2a (0.26 mmol), [Rh(C2H4)2Cl]2 (5 mol%), L7 (15 mol%), LiOtBu (0.6 mmol) and 3 Å MS (100 mg) in dioxane at 80 °C], C–H arylation reactions of thiocarbonylferrocene 1a with various aryl iodides were investigated (Table 2). Notably, the absolute configuration of product 3a (99% ee after recrystallization) was determined by X-ray crystallographic diffraction as Sp (see the Supplementary Table 2 for details). Both electron-donating groups (such as methyl, methoxy), halogen groups (F, Cl, and Br) and electron-withdrawing groups (cyano and trifluoromethyl) on the phenyl ring were all well tolerated, and the desired products were delivered in good yields with good to excellent enantioselectivity (3b–h, 58–82% yields, 89–95% ee). Delightedly, the reaction of 1-bromo-4-iodobenzene 2f afforded product 3f in 68% yield with 90% ee, leaving the Br atom intact. Meta- and ortho-substituted aryl iodides 1i, 1j, and 1k also proceeded smoothly to give the arylated products 3i–k in 70–80% yields with 91–97% ee. 2-Iodonaphthalene was found to be a suitable substrate, and 3l was obtained in 62% yield with 97% ee. Notably, heteroaryl iodides such as benzofuranyl, thienyl and pyridyl iodides were found to be effective coupling partners, providing the desired products 3m–q in 53–75% yields with 74–94% ee. However, 30 mol% of L7 was required to obtain reasonable yields for thienyl and pyridyl iodides. This is likely due to their relatively strong coordination with rhodium. This asymmetric C–H arylation reaction is also compatible with other thiocarbonylferrocenes such as 1r and 1s, leading to the formation of 3r and 3s in 65–76% yields with 99% ee.

Table 2 Scope of asymmetric C–H arylation[a,c,d]

To demonstrate the potential utility of such an asymmetric C–H bond arylation reaction, a gram-scale reaction was performed. Thioketone 1a (5 mmol) could be efficiently converted into 3a (1.56 grams) with 96% ee in the presence of 2.5 mol% [Rh(C2H4)2Cl]2 and 7.5 mol% L7 (Fig. 2a). Furthermore, planar chiral ferrocenyl ketone 4 was obtained in 93% yield without erosion of enantioselectivity by the treatment with Ag2CO3 under mild conditions (Fig. 2b).

Fig. 2
figure 2

Gram-scale reaction and transformation. a Arylation of 1a on a 1.56 g scale. b Oxidation of thioketone 3a

Mechanistic studies

To gain insights into this enantioselective C–H arylation reaction of ferrocenes, several control and deuteration experiments were carried out. No desired arylation product was observed when ketone substituted ferrocene 1a’ was subjected to the standard conditions, which indicated the unique role of thioketone group in the C–H bond activation process (Fig. 3a). 3a was obtained in 25% yield with 95% ee when bromobenzene was used as a coupling partner (Fig. 3b). The competition experiment between 4-iodoanisole 2c and 4-iodobenzonitrile 2g revealed that electron-deficient substrate was more reactive than the electron-rich one (Fig. 3c). Significant deuteration was observed for the recovered starting material 1a when the C–H arylation reaction was performed in the presence of 20 equiv of CD3OD (Fig. 3d). More importantly, we attempted this H/D exchange experiment in the absence of iodobenzene, which led to the recovered 1a with 70% deuterium (Fig. 3e). These results indicate that the C–H cleavage of thioketone substituted ferrocene is reversible and may be not the rate-determining step.

Fig. 3
figure 3

Control experiments. a Arylation of carbonyl compound 1a′. b Bromobenzene was used as a coupling partner. c Competition experiment between 4-iodoanisole 2c and 4-iodobenzonitrile 2g. d H/D exchange experiment of 1a and 2a. e H/D exchange experiment of 1a

Discussion

In conclusion, we have developed an efficient thioketone-directed Rh(I)-catalyzed enantioselective C–H bond arylation of ferrocenes. In the presence of Rh(I) catalyst derived from [Rh(C2H4)2Cl]2 and a TADDOL-based chiral phosphonite ligand, various aryl iodides reacted with ferrocenyl thioketones smoothly in good yields. The reactions displayed a broad substrate scope, and the planar chiral ferrocenes were obtained in excellent enantioselectivity. Heteroaryl iodides also work well in this catalytic system, affording heteroaryl substituted ferrocenes, which are difficult to be obtained by the previously known methods. Further mechanistic studies and applications of the products are ongoing in our laboratory.

Methods

Representative procedure

3 Å MS (100 mg) was added to a dry Schlenk tube (25 mL). The flask was evacuated and backfilled with argon for 3 times. Then, LiOtBu (0.6 mmol, 48.1 mg), L7 (0.03 mmol, 17.2 mg), [Rh(C2H4)2Cl]2 (0.01 mmol, 3.9 mg), and 1 (0.2 mmol) were added to the Schlenk tube. The flask was evacuated and backfilled with argon for 3 times again, and followed by addition of dioxane (1.5 mL). Aryl iodide 2 (0.26 mmol, 1.3 equiv.) was added at last. The mixture was stirred at 80 °C. After the reaction was complete (monitored by TLC), the mixture was cooled to room temperature. The mixture was diluted with petroleum ether (~10 mL). Then, silica gel was added and the solvent was evaporated under reduced pressure. The product was isolated by sicila gel column chromatography (petroleum ether or petroleum ether/ethyl acetate = 20/1).