Direct synthesis of branched amines enabled by dual-catalyzed allylic C─H amination of alkenes with amines

Abstract

Direct conversion of hydrocarbons into amines represents an important and atom-economic goal in chemistry for decades. However, intermolecular cross-coupling of terminal alkenes with amines to form branched amines remains extremely challenging. Here, a visible-light and Co-dual catalyzed direct allylic C─H amination of alkenes with free amines to afford branched amines has been developed. Notably, challenging aliphatic amines with strong coordinating effect can be directly used as C─N coupling partner to couple with allylic C─H bond to form advanced amines with molecular complexity. Moreover, the reaction proceeds with exclusive regio- and chemoselectivity at more steric hinder position to deliver primary, secondary, and tertiary aliphatic amines with diverse substitution patterns that are difficult to access otherwise.

A direct amination of more sterically hindered allylic C─H bond with free amines has been developed.

INTRODUCTION

Amines are prevalent motifs in natural products and pharmaceuticals and serve as common precursors for constructing complex organic molecules and functional materials (1, 2). The allylamine derivatives are especially attractive synthetic targets as they are both versatile synthetic intermediates and valuable targets (Fig. 1A) (3–11). Among the strategies of C─N bond formation, direct conversion the C─H bonds of hydrocarbons into C─N bond represents one of the most attractive goals for decades (12–20). Particularly, intermolecular cross-coupling of terminal olefins with nitrogen source to form complex amines remains a major challenge in chemical synthesis. Over the past decades, great efforts have been devoted to the coupling of alkenes with nitrogen sources via C─H functionalizations (Fig. 1B). However, direct conversion of hydrocarbons into amines represents an attractive yet challenging goal. One of the major reasons that attributes to highly coordinating amines is basic and good nucleophiles in non-directed, electrophilic metal-catalyzed aminations, tending to bind to metal centers and thereby inhibiting the desired catalytic process. Thus, most protocols using nitrogen pronucleophiles are usually limited to nitrogen sources with one or more electron-withdrawing groups covalently bonded to nitrogen, such as sulfonamides and carbamates (21–32). To address these limitations, White and co-workers (33) and Jiang and co-workers (34) demonstrated elegant examples enabled by ligated Pd-catalyzed amination of alkenes with secondary aliphatic amines to give linear tertiary allylic amines. The renaissance of photocatalysis and electrochemistry offers further opportunity for the amination reaction of alkenes under milder conditions through radical pathways. In 2020, Ritter group developed the amination of aliphatic alkenes enabled by radical addition of aminium radicals to alkenes using a thianthrene-based aminating reagent to construct C─N bond followed by elimination to give allylic secondary amines (35). In 2022, Lei group reported an elegant photoredox/cobaloxime catalyzed amination of 1,1-disubstituted alkenes with secondary alkyl amines to afford allylic tertiary amines (36). However, large excess amount of alkenes (more than 20 equivalents) is typically required to achieve satisfactory results. In 2021, Wickens and co-workers developed an electrochemical enabled amination of 1-substituted aliphatic alkenes with secondary aliphatic amines to deliver linear tertiary amines (37). On contrast, the direct amination at allylic position of alkenes without isomerization to afford branched amines remains extremely challenging and scarce. To date, there are only a few examples on direct allylic C─H insertion of metal-nitrenoid species with branched selectivity (Fig. 1C). However, the nitrogen sources for nitrenoid formation are generally limited to azides (38–41), sulfonyl amides (42–44), and dioxazolones (45–48) as mentioned before. Moreover, selectivity issues between aziridination and C─H amination impose additional challenge on the reaction of transition metal–catalyzed nitrene transfer with alkenes. In 2022, Gevorgyan and co-workers reported a seminal work on blue light–induced Pd-catalyzed direct allylic C─H amination of alkenes with primary and secondary aliphatic amines (49). The key to success of this reaction is to used bulky aryl halides as hydrogen atom transfer reagent and oxidant to achieve allylic radical intermediates. However, the regioselectivity is dominated by thermodynamic control among all the reported cases. No example of C─H amination at more steric position with kinetic cotrol is reported. Here, we disclose a direct and general amination of allylic C─H bond of alkenes with free amines to afford branched amines at room temperature (Fig. 1D). The combination of visible-light catalysis and cobalt catalysis allows for the allylic amination of terminal alkenes with primary and secondary aliphatic amines and even more challenging ammonium salts, affording branched allylic primary, secondary, and tertiary amines.

Fig. 1. Strategies for the synthesis of allylic amines from alkenes.

(A) Selected branched allylic amines in pharmaceuticals and bioactive molecules. (B) Representative methods for allylic C─H amination to linear amines. (C) Allylic C─H amination of alkenes via metal nitrenoid to branched amines. (D) Photocatalytic allylic C─H amination with amines to branched amines (this work).

RESULTS AND DISCUSSION

To test the feasibility of this proposal, optimization studies were conducted using α-decyl styrene (1a) and amantadine (2a, AdNH₂) as model substrates. After evaluation of different dimensions of parameters, we defined the standard condition as using Mes-3,6-tBu₂-Acr-Ph⁺BF₄⁻ (1 mol %) as photocatalyst (PC), Co (dmgH)(dmgH₂)Br₂ (10 mol %) as dehydrogenation reagent, Na₂CO₃ (50 mol %) as base, and NaI (2.0 equivalent) as additive in 1,2-dichloroethane (DCE; 0.05 M) under the irradiation of blue light-emitting diodes (LEDs) at room temperature for 18 hours, affording the direct C─H amination of allylic C─H bond to give branched allylic amine (3a) in 81% yield (Table 1, entry 1). The investigation of PC revealed that PCs with high oxidative potential are necessary to catalyze the reaction (Table 1, entries 2 to 5). Optimization of the cobaloxime catalyst indicated that different cobaloxime complexes could mediate the reaction to deliver the desired product 3a, albeit with lower efficiencies (Table 1, entries 6 to 10). Moreover, we found that the selection of solvent is crucial for the success of the desired reaction and preferred the chlorinated solvents, affording the optimal result in DCE (Table 1, entries 11 to 14). Control experiments suggested that the reaction proceeded smoothly in the absence of Na₂CO₃ or NaI, albeit affording 3a in lower yields (Table 1, entries 15 and 16). The presence of light, PC, and cobaloxime catalyst are all required for the reaction. No desired C─H amination product 3a could be generated without light or PC or cobaloxime catalyst (Table 1, entries 17 to 19).

Table 1. Optimization of the reaction conditions*.

PC, photocatalyst. DCM, dichloromethane; MeCN, acetonitrile; THF, tetrahydrofuran; NMR, nuclear magnetic resonance.

Entry	Variation from “standard condition”	Yield of 3a
1	None	85% (81%)†
2	[Mes-Acr-Me+]ClO4−	49%
3	[Mes-Acr-Ph+]BF4−	40%
4	[XylF-3,6-tBu2-Acr-Ph+]BF4−	60%
5	Ir(dF(CF3)ppy)2(dtbbpy)PF6	0%
6	Co(dmgH)(dmgH2)I2	71%
7	Co(dmgH)(dmgH2)Cl2	66%
8	Co(dmgH)(dmgH2)DMAPI	78%
9	Co(dmgH)(dmgH2)DMAPBr	76%
10	Co(dmgH)(dmgH2)DMAPCl	76%
11	DCM	27%
12	MeCN	0%
13	Dioxane	0%
14	THF	0%
15	No Na2CO3	56%
16	No NaI	47%
17	No light	0%
18	No PC	0%
19	No [Co]	0%

*Reaction was conducted using 1a (0.25 mmol) and 2a (0.1 mmol) under indicated conditions. ¹H NMR yield of the reaction using mesitylene as internal standard.

†Isolated yield.

With the optimized reaction conditions, we turned to test the scope of the direct amination of allylic C─H bond with free amines (Fig. 2). First, the scope of alkenes was investigated. Styrenes with both electron-donating and electron-withdrawing groups on the aromatic ring are good substrates for the allylic C─H amination reaction, affording the desired allylic amines (3b to 3h) in 52 to 74% yields. Moreover, para-, meta-, or ortho-substituted styrenes are all well tolerated in the reaction. A variety of α-substituted styrenes with different alkyl chains are compatible in the reaction to give corresponding C─H amination products (3i to 3s) in 51 to 85% yields. Notably, cyclopropylmethyl (3m), benzyl (3n), p-sulfonamide (3p), phthalimide (3q), ester (3r), and alkoxy (3s) were well tolerated in the C─H amination reaction. Impressively, α-tertiary amines (3t) could be accessed in 59% yield. Unfortunately, internal alkenes proved unsuccessful in the direct allylic C─H amination reaction under the reaction conditions. Next, the scope of amine source was evaluated under standard conditions. Primary amines were found to be suitable for this C─H amination reaction, providing a variety of allylic secondary amines in good yields (3u to 3ad). It deserves mentioning that more nucleophilic secondary amines were selectively obtained from primary amines using this protocol. Primary amines with diverse steric effects, such as α-tertiary, secondary, and linear primary amines, are all well tolerated in the reaction, furnishing corresponding secondary amines under mild conditions, circumventing over-allylation reactions under traditional conditions. Notably, unprotected amino alcohols underwent exclusively mono-allylation on nitrogen to give 3ae in 52% yield. Electron-deficient sulfonamides were also tolerated, delivering allylic C─H nitrogenation products (3af and 3ag) in 51 and 65% yields. Secondary amines are also compatible in the C─H amination reaction, delivering allylic tertiary amines (3ah to 3ap) in moderate to good yields. Piperidines bearing various functional groups such as esters (3ah), ketals (3ai), and methoxyl groups (3aj) could be converted into the corresponding tertiary amine products. Moreover, a wide range of substituted piperazines with esters (3ak), ketones with acidic protons (3al), and sulfones with acidic protons (3am) are well tolerated. In addition, another widespread N-heterocyclic substructure morpholine (3an) can also be efficiently converted to the corresponding allylic morpholine product in 81% yield. The structure of the allylic C─H amination products with primary and secondary amines is confirmed by the x-ray diffraction analysis of 3b and 3an, respectively. Acyclic secondary amines could be coupled with allylic C─H bond to give corresponding tertiary amines (3ao and 3ap) in synthetic useful yields. N-heteroaryl amines, such as pyrazole and 1H-benzotriazole, were also tolerated, affording allylic C─H amination products (3aq and 3ar) in 63% and 54% yields, respectively. It deserves mentioning that allylic primary amines (3as to 3au) could be accessible from alkenes with inorganic ammoniums using this visible-light and cobalt-catalyzed allylic C─H amination.

Fig. 2. Scope of alkenes and amines.

See Table 1 (entry 1) for reaction conditions. (*) Reaction was conducted using alkenes (0.10 mmol), amines (0.25 mmol) under indicated conditions. (†) La(OTf)₃ (50 mol %), Co(dmgH)(dmgH₂)Cl₂ (1 mol %), Mes-Acr-Me⁺ClO₄⁻ (2 mol %), DCE/CH₃CN = 1.9 ml/0.2 ml, room temperature (rt), 18 hours.

To showcase the practicability of this photocatalyzed direct amination of allylic C─H bond of alkenes with free amines, the reaction of 1v with 2a was performed on 1.0-mmol scale, giving the desired branched allylic amine 4 in 85% isolated yield (Fig. 3A). Moreover, the additional synthetic value of allylic amines as building blocks was documented by the conversion of 4 to the α-aminated ketone 5 in 73% yield and γ-amino alcohol 6 in 68% yield, respectively (Fig. 3, B and C) (50, 51). To further probe the mechanism of the reaction, a series of control experiments were performed. First, a radical clock reaction was performed using 7 with 2a under standard conditions (Fig. 3D). An allylic C─H amination product 8 with ring opening of cyclopropane was isolated in 12% yield, which may undergo the radical enabled ring-opening of cyclopropane in intermediate 8-1 to intermediate 8-2, followed by intramolecular radical cyclization with tethered arenes. To further demonstrate the dehydrogenation evolution, H₂ gas was detected by a gas chromatography–thermal conductivity detector for the reaction of alkene 1a in presence of amine 2a under standard conditions (Fig. 3E and fig. S2). Then, an alkene isomerization experiment of 1a in the absence of 2a under standard conditions was carried out (Fig. 3F). 1a was consumed and corresponding internal alkene 9 was formed in 1 hour. The results demonstrated alkene isomerization was the key to the selectivity of the C─H amination reaction of alkenes and may occur before C─N bond forming process. To further prove this hypothesis, the time course of the reaction using 1a and 2a as substrate under standard conditions was conducted (Fig. 3G). Substrate 1a was quickly consumed, yet no substantial amount of desired allylic C─H amination product 3a was formed in first 1 hour. Instead, internal alkene 9 was generated in quantitative yield. Then, the formation of 3a was detected along with the consumption of 9. The results further indicated that internal alkene 9 would be likely to server as an intermediate for the visible-light and Co-catalyzed allylic C─H amination reaction, suggesting the isomerization of alkenes to more stable internal alkenes before amination reaction. In addition, luminescence quenching experiments were conducted with α-ethyl styrene 1i and 2a under standard conditions (Fig. 3H). The experiments disclosed that the quenching effect of internal alkene 10 to the excited state of Mes-3,6-tBu₂-Acr-Ph⁺BF₄⁻ is much superior to that of 1i and 2a, which further demonstrated that the photocatalytic cycle started predominantly by the reductive quenching of excited state of PC with alkene 10.

Fig. 3. Synthetic applications and mechanistic investigations.

(A) Reaction on 1.0-mmol scale. (B) Synthesis of an α-aminated ketone. (C) Synthesis of a γ-amino alcohol. (D) Radical probe experiment. (E) Detection of H₂ for the dehydrogenation evolution. (F) Alkene isomerization experiment. (G) Time-course study. (H) The plots of luminescence quenching experiments. I₀ and I represent the fluorecence intensity of the photocatalyst in absence and presence of the quencher, respectively; E and Z represent two diastereomers of 10.

On the basis of the control experiments and mechanistic studies as well as literature precedence (52–59), a plausible mechanism for the visible-light and Co-catalyzed direct allylic C─H amination of alkenes with amines is proposed in Fig. 4. First, the alkenes 1 isomerized to form corresponding internal alkenes A with assistance of visible light and cobalt under standard conditions (60, 61). Next, excited state of PC* from visible light irradiation of the PC was quenched by A to generate reduced photocatalyst (PC^•−) species along with the formation of the radical cationic intermediate B. Next, the radical cationic intermediate B was trapped by nitrogen source to generate a carbon-centered radical intermediate C by forming C─N bond. [Co^III] species interacted with PC^•− via single electron transfer (SET) to regenerate the PC and the [Co^II] species, which could trap the carbon-centered radical intermediate C followed by β-hydrogen elimination to furnish the final products 3 and [Co^III]-H species (62, 63). Quench of [Co^III]-H with proton regenerated [Co^III] species by release of hydrogen gas to close the catalytic cycle. At this stage, another pathway via aminium radical cation generation cannot be excluded. Although the quenching effect of excited state of PC by secondary amines is slightly weaker than alkenes (figs. S9 and S10). However, considering the molar ratio of alkene and amine of 2.5:1, we believe that the proposed mechanism in Fig. 4 is reasonable.

Fig. 4. Proposed mechanism for the reaction.

In summary, a visible-light and cobalt-catalyzed direct allylic C─H amination of alkenes with free amines to afford branched allylic amines at room temperature has been developed. The reaction proceeded chemoselectively at branched allylic position to undergo mono-amination to forge C─N bond of allylic C─H bond. The reaction features photocatalyzed direct C─H amination of hydrocarbons under mild conditions. Secondary and primary free amines as well as ammonium salts are suitable nitrogen sources for this reaction, undergoing controlled onefold C─H amination process, providing the first general protocol for the synthesis of primary, secondary, and tertiary allylic amines. Notably, sterically congested α-branched and α-tertiay amines are easily synthesized, which are difficult access otherwise. Mechanistic investigations reveal that the reaction undergoes a nontrivial dual-catalyzed regiospecific isomerization, C─N bond-formation, and elimination process.

MATERIALS AND METHODS

General experimental procedures

All reactions were performed under nitrogen atmosphere, in flame dried glassware with magnetic stirring using standard Schlenk techniques, unless otherwise mentioned. All the solvents were distilled over calcium hydride and were purged with nitrogen before use. All other reagents were purchased from commercial sources and were used without purification, unless otherwise noted. Thin-layer chromatography was performed on Merck 60 F254 silica gel and visualized with a ultraviolet lamp (254 nm) or with a potassium permanganate solution. ¹H nuclear magnetic resonance (NMR) and ¹³C NMR spectra were recorded at room temperature unless otherwise required on a Bruker Avance 400 MHz or a Bruker Avance 600-MHz spectrometer. High-resolution electrospray ionization and electronic impact mass spectrometry was performed on a Q-Exactive Liquid Chromatography–Quadrupole Orbital Well Mass Spectrometry and an Orbitrap Fusion high-resolution liquid-mass spectrometer.

General method for the synthesis of alkenes

To a suspension of methyltriphenyl phosphonium bromide (12.0 mmol, 1.2 equivalent) in tetrahydrofuran (30 ml) at 0°C was added n-BuLi (12.0 mmol, 1.2 equivalent) dropwise. The reaction was stirred at 0°C for 1 hour before corresponding ketone (10.0 mmol, 1.0 equivalent) was added dropwise (if liquid) or as single portion (if solid), and the resulting mixture was warmed to room temperature and stirred for overnight. Upon completion, the reaction mixture was quenched by addition of saturated aqueous ammonium chloride and extracted by diethyl ether (30.0 ml × 3). The combined organic layer was washed with saturated brine, dried with anhydrous sodium sulfate, and concentrated under vacuum. The final styrene derivatives were purified by column chromatography using silica gel with hexanes/ethyl acetate eluents. The spectroscopic characterization for the alkenes matched those reported in the literature (64–76).

General method for the synthesis of branched allylic amines

Under N₂ atmosphere, an oven-dried Schlenk-tube equipped with a magnetic stir bar was charged with Mes-3,6-tBu₂-Acr-Ph⁺BF₄⁻ (0.6 mg, 1.0 μmol, 1.0 mol %), Co(dmgH)(dmgH₂)Br₂ (4.5 mg, 0.01 mmol, 10 mol %), nitrogen source (0.1 mmol, 1.0 equivalent) (if solid), NaI (30.0 mg, 0.2 mmol, 2.0 equivalent), and Na₂CO₃ (5.3 mg, 0.05 mmol, 50 mol %). Then, DCE (2.0 ml), alkene (0.25 mmol), and nitrogen source (0.1 mmol, 1.0 equivalent) (if liquid) were added consecutively via syringe. The tube was sealed with a Teflon-coated septum cap and stirred at ambient temperature under irradiation with 50-W blue LEDs for 18 hours. Upon completion, the resulting mixture was diluted with water (4.0 ml) and extracted with dichloromethane (4.0 ml). The combined organic phase was concentrated under vacuum. The crude mixture was analyzed by ¹H NMR with mesitylene as internal standard to determine the conversion and was directly purified by column chromatography on silica gel with hexanes (100.0 ml) and acetone (10.0 ml) as eluent to give the corresponding product.

Supplementary Materials

This PDF file includes:

Materials and Methods

Supplementary Text

Figs. S1 to S11

Tables S1 to S13