Atropoisomeric compounds are privileged scaffolds in pharmaceuticals, agrochemicals, bioactive natural products, and advanced materials1,2,3. They also serve as pivotal ligands and catalysts in organic chemistry4,5. In addition to the atropisomeric biaryls bearing C–C stereogenic axes, C–N axially chiral compounds have recently showcased their potential of being a novel type of axially chiral scaffolds that might have widespread application (Fig. 1A)6,7,8. However, catalytic asymmetric preparation of the latter has only gained limited success when compared with their C–C axially chiral siblings9. A fundamental challenge is the higher structural vulnerability of C–N chiral scaffolds originated from the increased degree of rotational freedom (Fig. 1B)10,11,12,13,14, which has greatly limited the application of many synthetic tools. Ever since the seminal independent reports by Taguchi15 and Curran16, several strategies have been developed for the catalytic asymmetric preparation of such compounds, namely ring formation, C–N axes construction, desymmetrization and rotational blocking of the C–N axes11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29. However, the utility of these elegant protocols has been inhibited by either substrate specificity or catalyst complexity. As a result, the development of novel and broadly applicable synthetic tools remains an urgent research topic.

Fig. 1: Preparation of C–N axially chiral compounds via asymmetric Catellani approaches.
figure 1

A Selected application of C–N axially chiral compounds. B Comparison between chiral C–C and C–N axes. C Stereocontrol strategies in asymmetric Catellani reaction and their applications in generating C–N axial chirality. D This work: chiral dinitrogen ligand enabled asymmetric Catellani reaction toward C–N axially chiral scaffolds. i = ipso, o = ortho.

The Catellani reaction, a type of palladium/norbornene (NBE) cooperative catalysis, has been recognized as a modular and highly efficient route toward highly functionalized molecules30,31,32,33,34,35,36. The reversible insertion of NBE has allowed the one-pot preparation of such molecules through functionalization of otherwise inaccessible ortho-C–H bond of iodoarenes with diverse elelctrophiles, followed by the ipso-coupling with various terminating reagents. Despite the powerfulness of this synthetic tool, controlling enantioselectivity in such a mechanistically complicated process hasn’t met any success until very recently, by using chiral NBE (previously designed for functional group directed remote meta-C–H activation37)38,39,40,41,42,43,44,45, chiral amine46,47 and chiral phosphine ligands48,49 as the chiral source, respectively. However, the assembly of C–N axially chiral scaffolds through asymmetric Catellani approaches remains underdeveloped (Fig. 1C). In 2021, the Zhou group50,51 established the generation of C–N axial chirality using chiral NBE strategy, a powerful asymmetric Catellani platform pioneered by Dong38 and Zhou39,41,42,43,44. Despite the elegance of this method, the requirement of using high loading of less facilely accessible chiral NBE (typically 50 mol%) and bulky 2,6-disubstituted aryl bromides as coupling partners has significantly hampered the synthetic utility of this strategy. Considering the high loading of NBE in many of the Catellani reactions, sometimes even in super-stoichiometric amount32,33,34,35,36,45, we postulated that the former could be addressed through the employment of an inexpensive NBE. However, the latter, originated from the indirect stereocontrol mode of chiral NBE strategy (generating a C–C chiral axis before the release of chiral NBE, followed by chirality transferring C–N formation)50, could only be solved through tuning the chiral induction strategy. While chiral amine strategy developed by Zhou46 and Gong47 is specific for carbonyl coupling partners, the chiral ligand strategy pioneered by Gu48,49 has only been applied to C–C formation involved asymmetric Catellani reactions in less satisfactory enantioselectivities52. In accordance with our continuous efforts on constructing C–N axially chiral scaffolds12,26, we intended to explore a practical asymmetric Catellani route toward such compounds complementary to chiral NBE strategy. Inspired by Gu’s pioneering work48,49 and the power of chiral ligands in related Pd-catalysis such as Pd(0) catalyzed asymmetric C–H activation53,54,55, we turned our attention to the potentially general chiral ligand strategy. We envisioned that the reactivity toward C–N forming Catellani reaction, along with the enhancement of enantioselectivity could be achieved through judicious choice of chiral ligands. Importantly, the strategy holds great promise to reduce the loading of the chiral source when proper ligand sphere was identified.

Herein, we report a chiral dinitrogen ligand enabled asymmetric Catellani reaction, which allows the efficient and modular assembly of C–N axially chiral scaffolds from readily available starting materials (Fig. 1D). The chiral source (BiIM ligand), used in decreased loading when compared with chiral NBE strategy (20 mol% vs 50 mol%), could be readily prepared from cost-effective chiral amino alcohol, oxalate and aniline with a single-step column chromatography56,57. Less bulky mono-ortho-substituted bromoarenes, a broad range of coupling partners challenging to control the enantioselectivity with chiral NBE strategy44, is compatible with our method. This is ascribed to the replacement of the indirect stereocontrol mode in chiral NBE strategy with a direct enantiodeterming C–N forming termination. To note, plain NBE, an inexpensive feedstock rarely been adopted in asymmetric Cetellani reactions, has been used as the mediator. The protocol is also featured with high reactivity, good enantioselectivity, broad substrate scope, ease of operation and scale-up. The method has further enabled two-fold asymmetric Catellani reaction, affording chiral optoelectronic material candidates featuring two C–N chiral axes. Facile transformation of a product allows the stereospecific preparation of C–N axially chiral carboxylic acids (CCA), a potentially good chiral ligand for asymmetric C–H activation. This work significantly enriched the application of chiral ligand strategy, unveiling its potential of being a generally applicable asymmetric Catellani reaction platform complementary to chiral NBE strategy.


Optimization of the reaction conditions

Initially, we selected 1-iodonaphthalene (1a) as the model substrate and 2-bromo-N-(2-(tert-butyl)phenyl)benzamide (2a) as the coupling partner. We commenced our investigation by screening various chiral ligands in the presence of Pd2(dba)3, plain NBE mediator, Cs2CO3 in toluene (Fig. 2). Although two chiral sulfonamide phosphine ligands (L9-L10) provided product 3 with noticeable enantioselectivity (13% and 32% ee), other monodentate or bidentate chiral phosphine ligands (L1-L8) only resulted in negligible enantioselectivity. In accordance with our hypothesis, these results highlight the challenge of using chiral phosphine ligands to construct of C–N chiral axes. We were pleased to see that chiral bidentate N,N-ligands forming 5-membered palladacycle upon chelating to a Pd-center, such as pyridinyl oxazoline ligands (L11-L15), BOX (L16-L18) and N,N’-aryl substituted biimidazoline (BiIM) ligands (L20-L21), afforded the desired product in promising enantioselectivity. Surprisingly, no enantioselectivity was observed with a BOX ligand that forms 6-membered palladacycle (L19). Upon further modulating the electronic properties and the bite angle of BiIM ligands by varying the N,N’-substituents, we finally figured out L25 bearing 3-trifluoromethylphenyl groups on both sp3 hybridized nitrogen atoms as the optimal ligand, affording the enantioenriched product in 78% yield and 60% ee. Through examining other reaction parameters (Supplementary Table 1 in the Supplementary Information), the optimal conditions were established when Cs2CO3 was replaced with Ag2SO4 and 200 μL of water and 100 mg of 4 Å molecular seivies were employed as additives, affording product 3 in 95% yield and 91% ee. The pivotal role of the catalyst, ligand and other reagents was confirmed by control experiments (Supplementary Table 3 in the Supplementary Information). We postulate that water improves the enantioselectivity by forming hydrogen bonding with the ligand and influencing the bite angle58, whereas the silver salt serves as halide scavenger to improve the reactivity59. By heating product 3 in isopropanol at 150 oC and monitor the ee values, the rotational barrier was measured to be 34.7 kcal/mol, suggesting the remarkably high stability of this product even at 150 oC (t1/2(150 oC) = 8.48 hours, see Page 95–96 in the Supplementary Information for details).

Fig. 2: Evaluation of chiral ligands.
figure 2

aReactions were carried out in 0.1 mmol scale with respect to amide 2. Conditions: 1a (1.5 equiv), 2a (0.10 mmol), Pd2(dba)3 (5 mol%), Ligand (20 mol%), NBE (1.5 equiv), Cs2CO3 (2.0 equiv) in toluene (0.05 M) at 80 °C for 36 h under N2. bCarried out under modified conditions: 1a (1.5 equiv), 2a (0.10 mmol), Pd2(dba)3 (5 mol%), L25 (20 mol%), NBE (1.5 equiv), Ag2SO4 (2.0 equiv), H2O (200 μL), 4 ÅMS (100 mg) in toluene (0.05 M) at 80 °C for 36 h under N2. Yields were isolated yields and the ee’s were determined by chiral HPLC analysis. Nap = naphthyl.

Substrate Scope

With the optimized condition in hand, we next investigated the scope of aryl iodides (Fig. 3). A series of ortho-substituted aryl iodides, such as 1-iodonaphthalene derivatives (4–8), 2-methyl iodobenzene (9–10), and 2-iodo-1,1’-biphenyl (11), are compatible with this method. Heteroarenes like dibenzo[b,d]furan (12) and the strongly coordinating quinoline moiety (13–14) are also viable substrates, affording the desired products in moderate to good yields and high enantioselectivities. The compatibility with diverse functional groups were tested on the 1-iodonaphthalene skeleton. In addition to the direct attachment of alkyl groups (4–5), ether (6) and halides (7–8), both strongly electron-withdrawing and electron-donating groups are well tolerated when linked onto 1-iodonaphthalene through an aryl bridge (15–28). The direct attachment of heteroarenes, such as pyridine (29), thiophene (30–32), and furan (33–34), were also well tolerated. Enantioenriched product 35 (90% ee) was obtained as a mixture of rotamers in 1:1.2 ratio (detected by 1H NMR), and the ratio disappeared after cleavage of Boc group. The yield of 35 was only moderate, perhaps due to the bulkiness of the protecting group.

Fig. 3: Scope of Aryl Iodides.
figure 3

aReactions were carried out in 0.1 mmol scale with respect to amide 2. Conditions: 1 (1.5 equiv), 2a (0.10 mmol), Pd2(dba)3 (5 mol%), L25 (20 mol%), NBE (1.5 equiv), Ag2SO4 (2.0 equiv), H2O (200 μL), 4 ÅMS (100 mg) in toluene (0.05 M) at 80 °C for 36 h under N2. bDichloroethane instead of toluene was used as solvent. cThe dr was observed through 1H NMR, which disappeared after Boc-deprotection. The ee’s were determined by chiral HPLC analysis. Ac = acetyl; TBS = tert-butyldimethylsilyl.

Next, we further investigated the scope of the coupling partners (Fig. 4). Ortho-bromo benzamides bearing an electron-withdrawing and electron-donating groups on the benzoyl moiety are well-tolerated, affording the products in moderate to good yield and high enantioselectivity (36–39). Additionally, disubstituted bromo benzamides are also compatible (41–42, 47–48). Substitution on the aniline moiety was also allowed, with alkynyl (44), aryl (46) and alkyl (45) groups well tolerated. Notably, an additional bromine group para to the amine group (43) also afforded the target product in moderate enantioselectivity, regardless of its sterically more favourable nature over the bromine ortho to the benzoyl moiety. The ortho tert-butyl group at the aniline motif could be replaced by bulky alkoxyl groups (49–50), but the enantioselectivity is diminished when the coordinative hydroxyl group is not protected (51).

Fig. 4: Scope of Amides.
figure 4

aReactions were carried out in 0.1 mmol scale with respect to amide 2. Conditions: 1 (1.5 equiv), 2a (0.10 mmol), Pd2(dba)3 (5 mol%), L25 (20 mol%), NBE (1.5 equiv), Ag2SO4 (2.0 equiv), H2O (200 μL), 4 ÅMS (100 mg) in toluene (0.05 M) at 80 °C for 36 h under N2. bDichloroethane instead of toluene was used as solvent. The ee’s were determined by chiral HPLC analysis.

Synthetic Application

Having investigated the scope of this protocol, a series of experiments were carried out to illustrate its synthetic utility (Fig. 5). Recently, the preparation of molecules possessing multiple chiral axes has attracted significant interest among organic chemistry community60,61,62,63,64, for the unique role of these topologically complex scaffolds in material science and catalysis. In addition, the quick expansion of conjugate system during the reaction process further encouraged us to wonder whether products bearing more extended π-conjugate system were accessible. To our delight, the 1,5-diiodonaphthalene substrate underwent two-fold asymmetric Catellani-type reaction smoothly and afforded products bearing two chiral C–N axes and a larger conjugate system (52–54). In accordance with the Horeau principle65, ee of the chiral products amplified after two enantioselective operation (obtained in 94% to 99% ee). It appears that the ratio of meso-product raised as the size of bulky ortho group at the aniline motif increases (ranging from 10.8:1 to 1.4:1), likely due to the decreased repulsion within the meso-product, whose two bulky groups are located at different sides of the π-system. A gram-scale preparation of product 49 was carried out with equally high yield and ee (81% yield, 91% ee for gram-scale vs 83% yield and 92% ee for 0.1 mmol scale). Subsequent transformations, such as cleavage of silyl group (51), protection of the free alcohol (55, CCDC 2249376), and oxidation of the alcohol to aldehyde (56), were all carried out with full retention of configuration. The enantioenriched C–N axially chiral aldehyde (56) could be further transformed into the corresponding oxime (57) and carboxylic acid (58) with high fidelity of enantiopurity. Considering the versatility of chiral carboxylic acid (CCA) ligands in asymmetric C–H activation, we subjected the obtained C–N axially chiral CCA (58) in a Ru(II)-catalysed enantioselective C–H annulation reaction66,67. To our delight, the reaction of sulfoximine 59 with sulfoxonium ylide 60 was carried out in high yield with moderate enantioselectivity under non-optimal conditions (96%, 55% ee). This application showcased the utility of our protocol in synthetic chemistry.

Fig. 5: Synthetic application.
figure 5

A Preparation of products with two C–N stereogenic axes. B Gram-scale experiment and follow-up transformations. C Application of CCA 58 as chiral ligand in Ru-catalysed asymmetric C–H activation. Ms = methanesulfonyl.

After obtaining an array of C–N axially chiral scaffolds bearing extended rigid-planar π-conjugate systems, we set out to examine the photophysical properties of selected products (Fig. 6). Photo luminescence quantum yields (PLQY) of these compounds are remarkably high, ranging from 40% to 97% (Fig. 6A). The fluorescence of these products ranges from bright blue to violet under UV light irradiation (Fig. 6B). Not surprisingly, the incorporation of an additional C–N stereogenic axis and more extended π-conjugate system has led to a red-shift in their absorption maxima (Fig. 6C, λabs = 299 nm for 52, 53 and 54 vs λabs = 280 nm for 21 and 33). The wavelength of emission maxima for 52, 53 and 54em = 420 nm) is shorter than 21em = 445 nm) and 33em = 469 nm) (Fig. 6D). We further investigated the chiroptical properties of products 52, 53, 54 and their enantiomers by measuring their circular polarized luminescence (CPL) spectroscopies (Fig. 6E). Interestingly, all the selected three compounds are CPL-active, and each pair of enantiomers displayed clear mirror images. The dissymmetry factors (glum) of both enantiomers of 52, 53, 54 around emission maxima (average value between 420–450 nm) range from −9.8 × 10−5 to 2.2 × 10−4 (Fig. 6F), demonstrating the potential application of these easily accessible chiral compounds in developing interesting chiroptical devices68.

Fig. 6: Photophysical Property Investigations.
figure 6

A Structure and PLQY information of selected compounds for photophysical property characterization. B Fluorescence images of selected silicon-bridged heterocycles (λex = 365 nm). C Absorption spectra of selected compounds in DCM (20 μM). D Emission spectra of selected compounds in DCM (20 μM). E CPL spectra of compounds in DCM (1.0 mM) at room temperature, excited at 300 nm. F glum values-wavelength curves.


We have developed a chiral dinitrogen ligand enabled asymmetric Catellani reaction that allows the efficient and modular assembly of C–N axially chiral scaffolds. Key to success of this methods is the utilization of an easily prepared and diversified chiral biimidazoline (BiIM) ligand that possesses a controllable bite angle as well as variable electronic and steric properties. The method features broad substrate scope, good reactivity, high enantioselectivity and ease of scale-up. A series of transformations around one of the products (49) were all carried out with full retention of configuration. A C–N axially chiral carboxylic acid (CCA 58) derived from 49 has showcased its potential as the chiral ligand for Ru-catalysed asymmetric C–H activation. Two-fold Catellani reaction with a diiodoarene afforded several CPL-active compounds bearing two C–N chiral axes in high enantioselectivity, which are potential candidates for chiroptical materials. The successful enantiocontrol in the construction of C–N axial chirality with chiral dinitrogen ligand strongly suggest that chiral ligand strategy has the potential to become a generally applicable asymmetric Catellani platform complementary to chiral NBE strategy. Further application of this system in other Catellani-type reactions is ongoing in our lab.


General procedure for BiIM ligand enabled asymmetric Catellani reaction

To an oven-dried 10 mL Schlenk tube were added substrate 1 (0.15 mmol), amide 2 (0.1 mmol), Pd2(dba)3 (4.9 mg, 0.005 mmol), L25 (10.2 mg, 0.020 mmol), Ag2SO4 (62.3 mg, 0.2 mmol), Toluene or 1,2-dichloroethane (2.0 mL), H2O (0.2 mL), 4ÅMS (100 mg), NBE (15 mg, 0.15 mmol). The mixture was stirred for 36 h at 80 °C. The resulting mixture was quenched by filtered through a celite pad and concentrated in vacuo. The residue was purified by preparative TLC to afford the product.