Introduction

With the expanding importance of chiral molecules in various fields of chemistry, ranging from the pharmaceutical industry to agrochemistry, fragrances, and material science, the development of sustainable and highly selective methodologies allowing to rapidly access enantiopure complex molecules is amongst the most vivid fields of organic synthesis and catalysis1,2,3. While aiming for more sustainable synthesis, the expansion of the C–H activation field rapidly affording molecular complexity from simple, non-prefunctionalized substrates is certainly a key achievement4. However, the majority of stereoselective C–H functionalization strongly rely on the use of noble metal-based catalysts5,6,7,8,9,10, including palladium, iridium, or rhodium. In clear contrast, such transformation catalyzed by 3d-metals11,12, and in particular from the most abundant, low cost, and non-toxic iron13,14,15,16 (Fig. 1a), continues to be challenging. Indeed, various oxidation states of iron, combined with a diversity of reaction scenarios conceivable in a presence of such a catalyst and difficulty in isolating well-defined iron-complexes17,18,19,20, renders the chiral iron-catalysis still in its infancy.

Fig. 1: Toward new sustainable assembly of chiral indoles bearing two chiral elements by iron-catalyzed stereoselective C–H alkylation.
figure 1

a Iron as catalyst of choice for catalytic transformations. b Different approaches for iron-catalyzed asymmetric transformation. c Iron-catalyzed enantioselective C−H alkylation. d Examples of bioactive C−N axially chiral compounds. e This work: iron-catalyzed diastereo- and enantioselective C−H alkylation.

Full size image

Although a few asymmetric cross-coupling reactions catalyzed by chiral iron-phosphine21,22 and iron-oxazoline23,24 complexes have been achieved, the need to use prefunctionalized substrates and equivalent organometallic reagents limits their applications25,26,27 (Fig. 1b). And yet, while considering the complementary reactivity of iron-complexes and noble 4d- and 5d-metal catalyst, combined with often high reactivity of such complexes translating into mild reaction conditions28,29,30, stereoselective iron-catalyzed C–H activation holds great promise to expand the diversity of easily accessible enantiopure molecules. Building upon previous reports31, our group developed a highly enantioselective C–H alkylation of indoles with vinylferrocenes and electron-rich styrenes enabled by a newly designed chiral N-heterocyclic carbene (NHC)32 catalyst15 (Fig. 1c). However, it is worth noting that there has been limited subsequent research on iron-catalyzed asymmetric C–H activation.

Over the decades, C-stereogenic molecules have focused major scientific interest to escape from flatness6. However, the diversity of chiral molecules spreads far beyond, englobing Si-, P- and S-stereogenic molecules33,34, or atropisomeric biaryls, heterobiaryls and styrenes35,36,37,38. Besides, restricted rotation around an N–Ar bond has also been attracting considerable attention from the scientific community39,40, progressively evolving from purely fundamental curiosity to an appealing tool for the design of highly selective drug candidates41,42,43,44 (Fig. 1d). Synthetic routes affording such compounds are yet rare and almost exclusively based on the use of noble metals45,46,47,48,49,50,51,52. Moreover, further increasing molecular complexity and the three-dimensional structure of the molecules by introducing simultaneously two chirality elements via a one-step process remains challenging53,54. In particular, the possibility of simultaneously controlling C–N atropoisomerism and C-central chirality has rarely been reported55,56,57,58. Despite the originality of these examples, the need for expensive and rare rhodium, iridium or palladium-catalysts seriously limits the synthetic potential of these transformations.

Inspired by previous research15,31,59, we were wondering if a more challenging C–N atroposelective transformation by iron catalysis could also be designed. Indeed, the introduction of a suitable, sufficiently sterically congested aromatic substituent on the N-atom could translate into the generation of an atropisomeric compound via the introduction of a substituent at the C2-position. Moreover, the possibility of generating an additional element of chirality by trapping the chiral metalacyclic intermediate with a prochiral coupling partner, such as olefin, appears as an ultimate challenge. Such an unprecedented transformation would involve not only the stereoselective olefin insertion but also simultaneously impose the configuration of the N–Ar bond thus promoting the formation of the indole product with two chiral elements.

Herein, we report an original, iron–NHC complex catalyzed asymmetric transformation delivering substituted indoles bearing both, C–N axial and C-central chirality, which is a unique and more challenging achievement to realize simultaneous stereocontrol of two chiral elements within one elementary step60 compared to single chiral center construction15 (Fig. 1e). The salient features of this reaction are: 1) the first example of iron-catalyzed C–N atroposelective reaction; 2) the first example of the use of iron-catalyst to build up complex molecules bearing both, axial and central chirality; 3) synthesis of various  substituted indoles exhibiting complex three-dimensional structure. In addition to these synthetic values, detailed experimental and theoretical mechanistic studies allowed shedding light on the asymmetric iron-catalysis from a broader perspective.

Results and discussion

Optimization of reaction conditions

We initiated our studies into the stereoselective C–H alkylation of indole derivative 1a with styrene 2a using the Fe/NHC catalytic system (Fig. 2). Pleasingly, the desired alkylation product 3 was obtained in 51% yield with >95:5 diastereomeric ratio (d.r.) and 67% enantiomeric excess (e.e.) when using Fe(acac)3 as the metal catalyst and L1 as chiral ligand. Subsequently, the structure of the chiral NHC ligand was further probed. The perfect diastereocontrol have been achieved through careful ligand design. The introduction of a methoxy or phenoxy group at the para-position of the aromatic ring (L2, L3) had a minor effect on the stereoselectivity of the alkylation reaction, while the presence of an additional methoxy group at the meta-position (L4) translated into a severe decrease in both reaction efficiency and stereoselectivity. In contrast, a remarkably 95% e.e. was realized when a more electron-donating and less sterically demanding methoxy motif is present at the ortho-position of the ligand (L5). Further investigations into the role of the ortho-substituent of the aromatic ring revealed that the benzyloxy-substituted ligand (L6) is inefficient while isopropoxy-derived L7 furnished the desired product in the same 95% e.e., but a lower yield. The substitution pattern at the meta-position had a dramatic impact on the stereoinduction. While a tert-butyl derived ligand (L8) performed equally well as the adamantyl congener (L5), a significant drop in the stereoinduction to 75% e.e. was observed with a methyl-substituted L9.

Fig. 2: Condition optimization for the iron-catalyzed asymmetric C–H alkylation.
figure 2

Reaction conditions: 1a (0.1 mmol), 2a (0.15 mmol), Fe(acac)3 (10 mol%), L* (20 mol%), CyMgBr (1 M in THF, 0.11 mmol) and TMEDA (0.2 mmol) were stirred in THF (0.2 mL) at 45 °C for 72 h under N2, then added HCl aq. (1 M, 1.0 mL) and stirred for 2 h. The yield was determined by 1H NMR spectroscopy using 1,3,5-trimethoxybenzene as the internal standard. The diastereomeric ratio (d.r.) was determined by 1H NMR spectroscopy. The enantiomeric excess (e.e.) was determined by HPLC. a1a (0.1 mmol), 2a (0.15 mmol), Fe(acac)3 (10 mol%), L5 (10 mol%), CyMgBr (1 M in THF, 50 mol%) and TMEDA (0.2 mmol) were stirred in Et2O (0.2 mL) at room temperature for 72 h under N2. TMEDA, N,N,N’,N’-tetramethylethylenediamine; THF, tetrahydrofuran.

Full size image

Next, the reaction conditions were further optimized. The tuning of the reaction parameters, including solvent and temperature illustrated that high yield and enantioselectivity could be obtained when the reaction was performed in Et2O at a relatively low temperature. Remarkably, the optimal results in terms of isolated yield (90%), diastereo- (>95:5) and enantioselectivity (97%) were realized using a reduced amount of the chiral NHC ligand (10 mol%), reflecting the efficiency and potential of this new catalytic system (see Supplementary Table 2 for details).

Substrate scope

To delineate the robustness of the diastereo- and enantioselective iron catalysis, indoles bearing a variety of N-aryl substituents were explored (Fig. 3a). Electronic changes and the steric hindrance of the ortho-substituents of the N-aromatic ring had a minor impact on the stereoselectivity of this reaction (3–9, all >95:5 d.r., up to 98% e.e.) and the atropostability of the products was guaranteed even for the less hindered methyl-substituted product 4. Further studies showed that both electron-withdrawing and electron-donating groups on the indole ring were well tolerated, delivering the desired products 10–17 in moderate to good yields with excellent diastereo- and enantioselectivities (up to >95:5 d.r., 98% e.e.) (Fig. 3b). Notably, azaindole 1p was also a potent substrate, furnishing product 18 in a good yield with a high level of stereoselectivity (>95:5 d.r., 91% e.e.). The reaction scope regarding the alkene was also very broad (Fig. 3c). A variety of differently substituted alkenes bearing electron-withdrawing or electron-donating substituents on the para-, meta- or ortho-position of aromatic ring furnished the desired products 19–31 in excellent stereoselectivities (all >95:5 d.r., 94-99% e.e.). 2-Vinylnaphthalene, 1-vinylnaphthalene, and vinylferrocene were compatible and afforded the target products 32–34 in good yields and enantioselectivities (90-99% e.e.). The absolute configuration of the alkylation products was confirmed unambiguously by a single-crystal X-ray diffraction analysis of indole product 3.

Fig. 3: Substrate scope and synthetic transformations for the iron-catalyzed asymmetric C–H alkylation.
figure 3

a N-substituent pattern. b Robustness with respect to indoles. c Scope of alkenes. d Scale-up and late-stage transformations. Reaction conditions: 1 (0.1 mmol), 2 (0.15 mmol), Fe(acac)3 (10 mol%), L5 (10 mol%), CyMgBr (1 M in THF, 50 mol%) and TMEDA (0.2 mmol) were stirred in Et2O (0.2 mL) at room temperature for 72 h under N2, then added HCl aq. (1 M, 1.0 mL) and stirred for 2 h. Yields are those of the isolated products. The diastereomeric ratio (d.r.) was determined by 1H NMR spectroscopy. The enantiomeric excess (e.e.) was determined by HPLC.

Full size image

Scale-up and late-stage transformations

A gram-scale reaction of indole derivative 1a and styrene 2a was successfully performed, delivering the desired product 3 in 83% yield without deterioration of the stereoselectivity (Fig. 3d), thereby demonstrating the utility of the stereoselective iron catalysis in preparative-scale organic synthesis. The presence of the formyl group on the product provided also important opportunities for expanding the molecular diversity. Accordingly, the corresponding alcohol 35, olefin 36, and amino 37 were accessed without loss of optical purity, as shown in Fig. 3d.

Mechanistic studies

To gain insights into the reaction mechanism, a series of experiments were carried out (Fig. 4). First, the kinetic isotope effect (KIE) study suggested that the C–H cleavage is facile (Fig. 4a). The reaction of the C2-deuterated indole substrate [D]1-1a with styrene 2a afforded the product [D]1-3 together with 12% of the unreacted substrate, isolated as a 75:25 mixture of [D]1-39 and 39. As to the alkylation product [D]1-3, the significant transfer of the C2-deuterium from the indole into the methyl position (81% D) indicates that the C−H activation occurs by ligand-to-ligand hydrogen transfer (LLHT) or C−H oxidative addition (Fig. 4b). In addition, a competition experiment between two different substrates ([D]1-1a and 1k) suggested that there was no deuterium scrambling distribution (Fig. 4c). Non-linear effect studies revealed a linear relation between the enantiopurity of the ligand and the e.e. of the product formed (Fig. 4d). This result is suggestive of only one NHC ligand being involved in the control of the reaction stereoselectivity.

Fig. 4: Mechanistic studies.
figure 4

a Kinetic isotope effect. b Deuterium-labeling experiment. c Deuterium scrambling experiment. d Non-linear effect studies.

Full size image

Further studies were pursued to 1) define the potential reactive iron-NHC species responsible for C–H activation and 2) obtain molecular-level insight into the role of the TMEDA additive in catalysis. For the latter, a recent report demonstrated the generation of a low-valent Fe-styrene species under similar reaction conditions which could be a precursor to a low-valent Fe-NHC complex for catalysis61. Consistent with this hypothesis, the synthesis of the Fe(0) species [Fe(Cy)(η2-styrene)3][MgCl(THF)5] (A) with the cyclohexyl Grignard reagent (CyMgCl) (Fig. 5a) and subsequent reaction with ligand L5 in the presence of excess CyMgCl led to the formation of a new iron species as identified by in situ freeze-trapped 80 K 57Fe Mössbauer spectroscopy (see Supplementary Information for details). This species was postulated to be a low-valent Fe(L5)(η2-styrene)2 complex (B) as related compounds have been reported in the literature62,63. These studies were further extended to evaluation of the in situ iron speciation during catalysis, where freeze-trapped 57Fe Mössbauer spectroscopy at 360 minutes into the C–H alkylation reaction revealed the presence of a single major iron species with parameters similar to B (Fig. 5b).

Fig. 5: Active catalyst studies.
figure 5

a Reaction pathway for generating A and B*, respectively. 80 K 57Fe Mössbauer spectra of their solid crystals with parameters δ = 0.37 mm/s and |ΔEQ | = 1.16 mm/s (A) and δ = 0.39 mm/s and |ΔEQ | = 2.92 mm/s (B*), and their crystal structures with selected bond lengths and angles. b 80 K 57Fe Mössbauer spectrum of the catalytic reaction with chiral NHC L5 at t = 360 min. c Reaction of in situ generated B* with indole substrate 1a to generate C−H alkylated product.

Full size image

While B was not readily amenable to crystallographic analysis, utilizing the achiral surrogate SIMes•HCl (possessing a saturated backbone like L5) enabled access to the S = 1 iron(0) complex Fe(SIMes)(η2-styrene)2 (B*) (Fig. 5a) which, combined with Mössbauer parameter calculations (see Supplementary Information for details), further supported the assignment of B as Fe(L5)(η2-styrene)2. To evaluate the catalytic relevance of such low coordinate Fe-NHC species, the reaction of B* with indole substrate 1a resulted in the generation of the C–H alkylated product (rac)-3 (Fig. 5c). Overall, these results support Fe(L)(η2-styrene)2 species as the likely active iron species in the current system, accessed from the low-valent catalytic precursor [Fe(Cy)(η2-styrene)3]-.

Computational studies

To understand the mechanistic details and to account for the origins of stereoselectivity, we furthermore performed density functional theory (DFT) calculations64. The competing reaction pathways for the iron-catalyzed asymmetric C−H alkylation for (Ra,S)-conformer are shown in Fig. 6a. Initiating from the triplet iron(0) complex int1 ligated by chiral NHC and two styrenes, whose reactivity already had been verified experimentally (Fig. 5c), ligand exchange occurs forming substrate coordinated intermediate int2. Classic oxidative addition of the indole C−H bond to iron center via triplet three-membered transition state TS3 is viable, generating quintet iron(II) hydride int4. Subsequent olefin insertion into Fe−H bond is facile leading to quintet alkyl iron(II) intermediate int6. However, the competing LLHT pathway from int2 directly to int6 is more favorable by 4.7 kcal/mol than the oxidative addition with sequential olefin insertion into Fe−H bond. We also considered the possibility of olefin insertion into the Fe−C bond via four-membered transition state TS7, while this insertion is proven to be unfeasible. Therefore, LLHT is most likely the preferred pathway for the generation of the alkyl iron(II) intermediate int6, where free rotation over the C−N axis is still achievable (see Supplementary Fig. 19). Int6 then undergoes irreversible reductive elimination which is the rate- and enantio-determining step to release the product and meanwhile regenerate the catalytic active iron(0) species int1. It is worth mentioning that only the most stable spin state of each species is presented in Fig. 6a, for the free energy profile of all possible spin states see Supplementary Fig. 18.

Fig. 6: DFT calculations on the reaction mechanism and origins of the stereoselectivity.
figure 6

a DFT-computed free energy profile of the competing reaction pathways for iron-catalyzed asymmetric C−H alkylation for major product. b Optimized structures and relative energies of reductive elimination transition states (trivial hydrogens are omitted for clarity).

Full size image

Based on these experimental and computational findings, we became interested in the controlling factors of stereoselectivity. The four reductive elimination transition states are shown in Fig. 6b with respective optimized structures and energies. TS10(Ra,S) is at least 3.5 kcal/mol more favorable than the other three transition states TS10(Ra,R), TS10(Sa,S) and TS10(Sa,R). The coordination of methoxy group to the iron center stabilizes the transition state TS10(Ra,S) for major product. Moreover, noncovalent interactions play an important role in the stereoselectivity control. The CH–π interaction in TS10(Ra,S) and TS10(Ra,R) between the methoxyl group of the NHC ligand and indole determines the axial chirality. This CH–π interaction is confirmed by the independent gradient model (IGM) analysis65. The same CH–π interaction is not present in TS10(Sa,S) or TS10(Sa,R). On the contrary, the steric hindrance between the methoxy group and the isopropyl group of the substrate is instead dominant. Another stereoselectivity-controlling factor is π-π stacking interaction in transition state TS10(Ra,S) between the indole and phenyl group, which determines the central chirality. The energy decomposition analysis of TS10(Ra,S) also reflects the impact of the noncovalent interactions (Supplementary Fig. 20). The coordination in combination with noncovalent interaction in transition state TS10(Ra,S) highlights the key role of methoxy group of chiral NHC.

A highly efficient selective C−H alkylation of indoles with aryl alkenes was achieved by sustainable iron catalysis, leading to rare atropoenriched and enantioenriched substituted indoles with high structural diversity. This iron-catalyzed asymmetric C−H alkylation was viable at room temperature, demonstrating the great potential of iron catalysis in the field of stereoselective C−H activation. Detailed mechanistic studies by experiment, Mössbauer spectroscopy and computation revealed an iron(0) complex as catalytic active species and this iron(0) complex undergoes a LLHT process instead of classic C−H bond oxidative addition. Computational studies revealed that the involvement of the methoxy group in the side arm of the chiral NHC ligand is vital for providing both covalent and noncovalent interaction in the rate- and enantio-determining transition state. This insight emphasizes the necessity for the de novo design of chiral ligands in asymmetric synthesis. We envisioned that the present approach and the mechanistic findings will promote the development in related challenging iron-catalyzed C−H functionalization constructing multiple chiral centers.

Methods

General procedure for iron-catalyzed stereoselective C–H alkylation

To a flame-dried and N2-purged Schlenk tube were added indole substrate (0.1 mmol), Fe(acac)3 (0.01 mmol, 10 mol%) and chiral NHC ligand L5 (0.01 mmol, 10 mol%). The Schlenk tube was then sealed, purged and backfilled with N2 three times. Et2O (0.2 mL), TMEDA (0.2 mmol) and alkene substrate (0.15 mmol) were added via syringe. CyMgBr (1 M in THF, 0.05 mmol, 0.05 mL) was then added dropwise and the resulting mixture was stirred at room temperature for 72 hours. Then, the reaction mixture was diluted with THF (2.0 mL) and quenched with HCl aqueous solution (1 M, 1.0 mL). The resulting mixture was stirred at room temperature for 2 hours. The phases were then separated, the aqueous layer was extracted with ethyl acetate (5.0 mL ×3). The combined organic layer was washed with brine, dried over Na2SO4, filtered and concentrated in vacuo. The diastereomeric ratio was determined by 1H NMR analysis of the crude reaction mixture. The residue was purified by column chromatography on silica gel (n-hexane: ethyl acetate = 10:1) to afford the desired product.