Copper-Catalyzed Allylic Amination of Alkenes Using O-Acylhydroxylamines: A Direct Entry to Diverse N-Alkyl Allylamines

Abstract

We report a copper-catalyzed direct allylic amination of alkenes using readily available O-benzyolhydroxylamines as the alkylamine precursors and internal oxidant. A range of primary and secondary alkylamines can be installed onto diversely substituted alkenes for rapid construction of N-alkyl allylamines. Mechanistic studies support that the reaction engages an initial electrophilic amination to alkenes with anti-Markovnikov selectivity and subsequently a regioselective oxidative elimination to furnish the double bond transposition. In the electrophilic amination step, the use of strong Brønsted acid is critical for generating the key aminium radical cation (ARC) species.

Graphical Abstract

Allylamines are essential building blocks in organic chemistry^1–2 and also feature prevalently in the structure of natural products and small molecule therapeutics (Scheme 1A).^3–7 Effective synthesis of this class of compounds is therefore fundamentally important. Among different strategies to access allylamines, oxidative allylic amination of simple alkenes represents a direct, powerful entry from feedstock materials. Recent developments include metal-free hetero-ene reactions,^8–10 as well as transition-metal catalyzed nitrene insertion, C-H activation and aza-Wacker reactions.^11–28 These methods allow for preparing allylic anilines, sulfonamides, carbamates and imines, yet they are incompatible with the implementation of electron-rich alkylamines.

Scheme 1.

Intermolecular allylic amination of alkenes toward the synthesis of N-alkyl allylamines.

Allylic amination of alkenes to directly prepare aliphatic allylamines has been challenging until recent breakthroughs. Using Pd-catalyzed allylic C-H activation, White,²⁹ Jiang,^30–31 and Gevorgyan³² have independently realized the allylic amination of alkenes in the presence of an oxidant (Scheme 1B). These nucleophilic amination protocols use convenient alkylamines yet have been limited to terminal or 1,2-disubstitued alkenes for the synthesis of acyclic linear alkylamines.^33–34 By an umpolung strategy, electrophilic amination of alkenes has been explored toward the synthesis of N-alkyl allylamines (Scheme 1C). This approach also eliminates the need of external oxidant in the nucleophilic amination strategy. The Ritter group developed the first photocatalyzed allylic amination of diverse alkenes for the synthesis of secondary alkylamines using iminothianthrenes.³⁵ Using photo and cobalt dual catalysis, Qi and Lei have achieved the allylic amination of multi-substituted alkenes using secondary alkylamines as precursors for electrophilic aminium radical cations.³⁶ Yet this reaction was constrained by the formation of tertiary alkyl radical intermediates, and thus unsuitable for monosubstituted and 1,2-disubstituted unactivated olefins. Despite the above significant advances in allylic amination of alkenes, direct construction of N-alkyl allylamines is restricted to certain substitution patterns on alkenes or specific alkylamines.

In our previous studies of alkene difunctionalization,^37–40 we found that O-acylhydroxylamines are capable of installing alkylamines across a double bond with anti-Markovnikov selectivity, via an electrophilic addition of aminium radical cations (ARCs) to alkenes followed by nucleophilic coupling of the intermediary radical species. We envision such an electrophilic amination approach could be tailored toward the elimination pathway for the synthesis of N-alkyl allylamines. Herein we report an effective allylic amination of alkenes achieved by a simple copper catalyst in an electrophilic amination approach (Scheme 1D). This transformation operates on diversely substituted alkenes and allows for regioselective installation of primary and secondary alkylamino groups using readily available O-benzoylhydroxylamines.

Our studies were commenced in the search of an effective catalytic system for the oxidative allylic amination of model substrates 1a and 1b using BzO-hydroxymorpholine 2a (Table 1). Our standard conditions were established on the allylic amination reaction affording 3-morpholinocycloheptene 3a using Cu(MeCN)₄PF₆ (5 mol %) catalyst and benzenesulfonic acid (PhSO₃H) additive (entry 1, see SI for more details). With 10 mol % copper catalyst, the formation of 3a was comparable in 68% yield (entry 2), while no product was formed in the absence of copper catalyst (entry 3).

Table 1.

Condition for allylic amination^a

entry	1	variation	3a/3b(%)a
1	1a	None	69
2		Cu(MeCN)4PF6 (10 mol %)	68
3		without Cu(MeCN)4PF6	ND
4	1b	Cu(MeCN)4PF6 (10 mol %)	72 (72)b
5c		CuX (X = OAc, OTf, Cl, CN, etc.)	<40
6c		CuX2 (X = OAc, OTf, acac, TFA, Cl)	<39
7c		no PhSO3H	ND
8c		TsOH•H2O instead PhSO3H	72
9c		CSA instead PhSO3H	67
10c		TFA instead PhSO3H	52
11c		other acids (e.g., BzOH, PPTS)	trace

^aStandard conditions: 1 (0.2 mmol, 1.0 equiv), 2a (2.0 equiv), Cu(MeCN)₄PF₆ (5 mol %), PhSO₃H (2.4 equiv), DCE (1 mL), 80 °C. Yields determined by¹H-NMR of the crude mixture using CH₂Br₂ as internal standard.

^bIsolated yields in parentheses. ND = Not detected.

^cwith Cu(MeCN)₄PF₆ (10 mol %)

To differentiate if the allylamine product was formed by the electrophilic amination pathway as we hypothesized or alternatively by direct allylic C-H amination, we investigated the reaction of diethyl cyclohept-4-ene-1,1-dicarboxylate 1b. Under standard conditions using 10 mol % Cu(MeCN)₄PF₆, the desired product 3b was isolated in 72% yield (entry 4), with no detection of 3b’ which would be formed from direct C-H amination pathway employed in previous work.²⁶ Other copper (I) and copper (II) catalysts were capable of promoting the allylic amination, yet less efficiently (entries 5 and 6). Consistent with our hypothesis involving the in-situ formation of aminium radical cations (ARC), the use of a strong Brønsted acid (PhSO₃H) is required in this reaction (entry 7). Other sulfonic acids were found comparably effective to PhSO₃H, such as TsOH•H₂O and camphorsulfonic acid (CSA) (entries 8 and 9). TFA was capable to promote allylic amination though less efficiently (entry 10). Other Brønsted acids examined were ineffective (entry 11), including TfOH, benzoic acid, and pyridinium p-toluenesulfonate (PPTS). Variation of other parameters such as solvents, equivalent, temperatures, and concentrations decreased the reaction efficacy (see SI).

To examine the generality of this oxidative allylic amination, we investigated the scope of both alkenes and amines of this transformation (Table 2). Using BzO-hydroxypiperidine, we examined the amination reactions on a diverse set of alkenes bearing different substitution patterns. We first investigated the allylic amination reactions of endocyclic alkenes, known to be challenging in previous methods.^{29–31, 36} These reactions all delivered desired N, N-dialkyl allylamines, including simple 5, 6, 7 and 8-membered cycloalkenes (4a-4f). This method was also applicable to acyclic internal alkenes. Starting from either cis- and trans-4-octene, allylamine product 4g was formed comparably, suggesting little influence of E- vs Z-alkene substrates in this reaction. For unsymmetrically 1,2-disubstituted alkenes, the amination occurred selectively at the more accessible position (i.e., 4h vs 4h’ and 4i vs 4i’). For tri-substituted alkenes, selective allylic amination also occurred at the less sterically hindered site, for example, forming 4j and 4k, in 58% and 74% yield respectively. Yet the tetra-substituted alkene was too sterically hindered to be effective in this reaction. We also examined the reaction of terminal alkenes. The allylic amination reactions proceeded in an anti-Markovnikov selective manner with the piperidine group installed at the terminal position of the original double bond (4l-4u). Both E- and Z-isomers were observed, with a slight preference to the E-isomers.⁴¹ These examples have also demonstrated a good tolerance of the method with different functional groups on the alkenes, such as aryl bromide (4n), indolyl (4p), phthalimidyl (4q), trichloroethoxycarbonyl (Troc)⁴² (4r), and cyclopropane (4s).

Table 2.

Scope of alkenes and amines^a

^aStandard conditions: 1 (1.0 equiv, 0.2 mmol), 2 (2.0 equiv), Cu(MeCN)₄PF₆ (5 mol %), PhSO₃H (2.4 equiv), DCE (1 mL), 80 °C. Isolated yields shown. Diastereomeric ratio (d.r.) and ratios of E/Z isomers determined by ¹H-NMR. Only E-isomers shown.

^bYields determined by ¹H-NMR of the crude mixture using CH₂Br₂ as internal standard.

^cCu(MeCN)₄PF₆ (10 mol %).

^dReactions run with 1 (5.0 equiv), 2 (0.2 mmol, 1.0 equiv), Cu(MeCN)₄PF₆ (5 mol %), PhSO₃H (1.2 equiv).

We next examined a representative set of cyclic and acyclic alkylamine precursors on the generality of this transformation for the installation of different alkylamines (Table 2). The reactions of cycloheptene using 6-membered cyclic alkylamines all smoothly provided desired products, such as 4-ethyl ester-piperidine (5a), 3-methyl-pi-peridine (5b), N-Bz-piperazine (5c), thiomorpholine (5d), and 1,1-dioxothiomorpholine (5e). 7-Membered cyclic amines were also effective such as azepane (5f) and N-benzoyldiazepane (5g). The installation of different cyclic amines was also successful in the reactions of acyclic alkenes such as allylbenzene (5h-5j) and vinylcyclohexane (5k). This reaction also allowed for the installation of acyclic alkylamines, such as N-methylphenethylamine (5k), and diethylamine (5l). Even primary alkylamines were readily installed onto different alkenes, such as 5m-5o. Overall, these examples demonstrated the general effectiveness of this oxidative allylic amination on a variety of alkenes and aliphatic alkylamines. Diverse functional groups are well tolerated, ranging from ester, carbamate, imide, amide, to bromide, sulfonate, and thiol groups.

To explore synthetic potential of this allylic amination method, the synthesis of allylamine 4q was performed on 1-mmol and 2.5-mmol scale reactions, in 77% yield and 66% yield, respectively (Scheme 2). The reaction was efficient when either alkene 1q or amine precursor 2b was used as the limiting reagent, offering flexibility in its practical application. This method also has proven effective for the allylic amination of 2,3,6,7-tetrahydro-azepine 1w to form allylamine 6. Subsequent dihydroxylation reaction readily furnished richly functionalized azepane product 7, with only single diastereomer observed (d.r. > 20:1). The relative anti-stereochemistry of diol opposite to piperidine ring has been confirmed by the X-ray analysis of derived di-acetate 8.

Scheme 2. Synthetic applications.

^aStandard conditions: 1 (1.0 equiv), 2b (2.0 equiv), Cu(MecN)₄PF₆ (5 mol %), PhSO₃H (2.4 equiv), DCE (1 mL), 80 °C. ^bRun with 1 (5.0 equiv), 2b (1.0 equiv), Cu(MecN)₄PF₆ (5 mol %), PhSO₃H (1.2 equiv). Isolated yields shown.

We sought to gain further mechanistic insights into this oxidative allylic amination reaction. Given our working hypothesis that the reaction was initiated by radical addition of aminium radical cations (ARCs) to the alkene, we first conducted control experiments to probe the involvement of the radical pathway in the amination reaction of using allylbenzene (Scheme 3A). The reaction in the presence of (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) as a radical scavenger resulted in the formation TEMPO-trapped amino oxygenated product (9), suggesting the initial formation of the C-N bond and the formation of alkyl radical intermediates under the reaction conditions. In the control reaction with the addition of butylated hydroxytoluene (BHT), several BHT-trapped products (10–12) were observed, further supporting the formation of both nitrogen- and carbon-radical species in this reaction.

Scheme 3. Studies on the reaction mechanism and hypothesis^a.

^aStandard conditions: alkene (1.0 equiv), 2 (2.0 equiv), PhSO₃H (2.4 equiv), DCE (0.2 M), 80 °C. Isolated yields shown.

Control experiments were next performed to probe the nature of the desaturation step from the carbon radical intermediate (Scheme 3B). Based on Kochi’s seminal studies,^43–46 there are two plausible pathways for copper-catalyzed radical oxidative elimination: (1) recombination of the alkyl radical and Cu^(II) to form an alkylcopper^(III) intermediate, followed by a β-oxidative elimination, and (2) Cu^(II) promoted oxidation of the alkyl radical to a carbocation followed by a subsequent β-deprotonation. If elimination occurs through β-deprotonation, abstraction of the more acidic proton should be preferred in the desaturation. Yet, the regioselectivity observed in our reaction of ethyl 3-methyl-3-butenoate was to the contrary of this hypothesis. The selective formation of allylamine 4w by the preferential elimination of more hydridic hydrogen atom (i.e., Ha) suggests that the desaturation step in this allylic amination more likely involves an organocopper(III) species, while the formation of the carbocation could be impeded by a vicinal aminium moiety upon the installation of an alkylamine. Furthermore, we investigated the parallel kinetic isotope effect (KIE) of this oxidative elimination step. The insignificant KIE (K_H/K_D = 1.05) observed in the reactions of allylbenzene 1m and deuterium-labeled 1m-d₂ suggests that the elimination step is rapid and not the rate-limiting step in this transformation. These observations are consistent with the oxidative elimination mechanism proposed in Kochi’s work and recent studies on radical oxidative elimination by copper catalysis.^47–51

Based on our studies and literature precedents, a working hypothesis for this oxidative allylic amination reaction is proposed (Scheme 3C). The reaction is initiated by a copper-catalyzed oxidative cleavage of O-benzoylhydroxylamine, either via oxidative addition or single electron transfer process, generating aminium radical cation (ARC) species in the presence of a strong Brønsted acid (i.e., PhSO₃H). Subsequently, the ARC-promoted electrophilic amination occurs regioselectively at the less substituted position of double bond, resulting in the formation of carbon radical intermediates. The following restoration of the double bond occurs through a β-oxidative elimination, likely via an alkylcopper(III)-complex.

In conclusion, we have developed a copper-catalyzed oxidative amination for the synthesis of N,N-dialkyl, and N-alkyl allylamines from diversely substituted alkenes. The mechanistic studies suggest that this reaction is initiated by the formation of aminium radical cation (ARC) species from O-acylhydroxyalkylamine in the presence of copper catalyst and a strong Brønsted acid. The selective allylic amination process engages an anti-Markovnikov addition of aminium radical cations to olefins followed by a site-selective β-oxidative elimination.