Photoinduced Palladium-Catalyzed Carbofunctionalization of Conjugated Dienes Proceeding via Radical-Polar Crossover Scenario: 1,2-Aminoalkylation and Beyond

Abstract

A photoinduced palladium-catalyzed 1,2-carbofunctionalization of conjugated dienes has been developed. This mild modular approach, which does not require employment of exogeneous photosensitizers and external oxidants, allows for efficient and highly regio- and stereoselective synthesis of a broad range of allylic amines from readily available 1,3-dienes, alkyl iodides, and amines. Employment of O- and C-nucleophiles toward oxyalkylation and dialkylation products was also demonstrated. A putative π-allyl palladium radical-polar crossover path is proposed as a key event in this three-component coupling process. The utility of this protocol is highlighted by its application for derivatization of several amine-containing drugs.

1,3-Dienes are readily available versatile building blocks, which enjoy numerous applications in synthesis,¹ material science,² and medicinal chemistry.³ Difunctionalization of 1,3-dienes enables rapid access to complex molecules in one step and is therefore highly sought after.⁴ In particular, 1,2-carboamination of dienes represents an efficient synthetic route to allylic amines, which are valuable synthons,⁵ as well as important motifs in bioactive molecules (Scheme 1a). However, only isolated examples have been reported to date, with the vast majority being two-component annulative reactions proceeding at elevated temperatures (Scheme 1b).⁶ As a result, not only the choice of substrates but also the diversity of obtained products is limited. A multicomponent reaction (MCR) in which the carbon and nitrogen functionalities are introduced independently would therefore be an appealing method to address current limitations and allow for expedient synthesis of structurally diverse allylic amines. Herein, we report a light-induced palladium-catalyzed three-component coupling⁷ of 1,3-dienes, alkyl iodides and amines that occurs under mild conditions (Scheme 1c). This transformation is proposed to involve a putative π-allyl hybrid palladium radical intermediate A. The latter, upon radical-polar crossover step, is converted into a “traditional” π-allyl palladium intermediate B,⁸ which is then trapped with an amine to deliver the reaction products.

Scheme 1.

Background and Proposed 1,2-Carboamination of 1,3-Dienes

In recent years, the photoinduced palladium catalysis has become a highly emerging area of research.⁹ It has been established by our group and others that hybrid palladium radical species can be generated via cleavage of C(sp³)– and C(sp²)–X (X = Cl, Br, I, CO₂NPhth, OTf) bonds in the presence of photoexcited palladium complexes. Formation of these species was crucial for achieving various processes, including remote desaturation,¹⁰ alkyl Heck reactions,¹¹ and other transformations.¹² Keeping in mind a facile addition of hybrid alkyl radical palladium species across the double bond,¹¹ we envisioned that analogous addition at conjugated diene would also be feasible, thereby leading to a putative hybrid π-allyl palladium radical species A that would exist in equilibrium with the classical π-allyl complex B. The anticipated aminoalkylation product would be delivered upon nucleophilic attack by amine at the latter (Scheme 1c).

To this end, we commenced our study with examining the reaction of 1-phenylbutadiene (1), (trimethylsilyl)methyl iodide (2), and N-methylbenzylamine (3). Expectedly, the major challenge was to suppress the expected side reactions, including Heck reaction (5), hydroamination (6), and amine alkylation (7) processes. After extensive screening of reaction parameters,¹³ we were delighted to obtain 4a in 80% isolated yield (Table 1, entry 1). Similarly to our previous studies,^11a the combination of Pd salt with Xantphos ligand proved to be the most efficient catalytic system. This reaction appeared to be insensitive to the stereochemistry of diene, as employment of a 1:1.2 mixture of (E)/(Z) isomers of 1 did not lead to a significant change of the yield (entry 2). Control experiments indicated that both visible light irradiation and palladium catalyst were necessary (entries 3, 4). Without silver triflimide (entry 5), the reaction was significantly less efficient due to the formation of substantial amounts of radical dimerization products (see Scheme 4b and discussion, therein). Likewise, the reaction efficiency was substantially lower in the absence of the base (entry 6). Employment of a bromo analog of 2 was much less efficient (24%), whereas the chloro counterpart did not react under these conditions at all.

Table 1.

Optimization of Reaction Conditions^a

Entry	Deviation from standard conditions	4:5:6:7b	Yield of 4ac
1	None	98:1:0.5:0.5	89% (80%)
2	(E)/(Z)-1 (1:1.2)	96:1:1:2	83% (75%)
3	45 or 80 °C (dark)	8:0:0:92	Trace
4	Without PdCl2	4:0:0:96	Trace
5	Without AgNTf2	91:5:2:2	48%d
6	Without Cs2CO3	62:29:7:2	44%

^a0.2 mmol scale; (E)-1:2:3 = 1:1.3:1.3.

^bDetermined by GC/MS using pentadecane as internal standard; 4a:4b > 20:1.

^cDetermined by ¹H NMR using dibromomethane as internal standard; isolated yields given in parentheses.

^dFormation of substantial amounts of radical dimerization products was observed.

Scheme 4.

Mechanistic Studies

With the optimized conditions in hand, we examined the generality of this methodology (Table 2). First, the diene scope was tested. It was found that 1-aryl and hetaryl substituted dienes are all capable partners for this transformation. Thus, electron-rich dienes reacted smoothly to give the corresponding allyic amines in good yields (8–12). Notably, in the case of a phenol-containing moiety (9), the formation of a potential competing carboalkoxylation product was not observed, thus indicating high chemoselectivity of this reaction. Electron-deficient aryl groups were also tolerated, thus delivering the expected products in moderate to good yields (13–15). Alkyne and halide functionalities were also compatible with this transformation (16–19). Aminoalkylation of dienes possessing substituents at meta- and ortho-positions of the arene proceeded uneventfully producing the reaction products 20–24. Ferrocenyl and heteroaryl groups were also amenable to this catalytic process, affording the corresponding products in moderate to good yields (25–30). Alkyl-substituted dienes were also reactive under these reaction conditions. Selective 1,2-addition was observed with a secondary alkyl substituent (31), whereas mixtures of 1,2- and 1,4-adducts were isolated for the primary counterparts (32–33), probably due to steric reasons. In addition, this MCR can be performed with trienes to obtain the corresponding dienamines 34–35. A lower yield from diene ester (36) was obtained, presumably due to the electronic mismatch between the silylmethyl radical and electron-deficient diene ester. Importantly, 1,2-disubstituted butadiene also proved to be a viable substrate to produce trisubstituted alkenyl amine 37. A higher catalyst loading, however, was required to achieve a reasonable yield for this substrate.

Table 2.

Scope of Three-Component 1,2-Carboamination of 1,3-Dienes^a

^a0.4 mmol scale; isolated yields.

^bNMR yield.

^c0.2 mmol scale.

^d10 mol % PdCl₂, 20 mol % Xantphos and 20 mol % AgNTf₂ were used.

^e10 mol % Pd(OAc)₂, 20 mol % Xantphos and 20 mol % AgNTf₂ were used.

^fTHF/MeOH = 1:1 (0.25 M).

^gTHF/DCM = 2:1 (0.33 M).

Next, the scope of alkyl iodides has been evaluated. Primary alkyl iodides bearing α-silyl or germyl substituents reacted efficiently to furnish the coupling products 38–40 in reasonable to good yields. Reactions employing primary alkyl iodides gave diminished yields (41–44), presumably due to the stronger C–I bond and thus lower reactivity,¹⁴ as well as due to the competing alkylation reaction of the amine. In contrast, different secondary alkyl iodides reacted well producing the corresponding allylic amines 45–50 in good yields. Due to the steric reasons, employment of bulky tert-butyl iodide, however, produced a nearly equimolar mixture of 1,2- and 1,4-adducts (51). This reaction was found to be very general with the respect to the secondary amine used. Thus, benzylamines readily participated in this reaction, delivering the allylic amines 52–55 in moderate to good yields. An amine containing a picolyl moiety (56) was also found to be effective. Aliphatic acyclic amines also reacted, though with somewhat lower efficiency (57–58). Reactions with various amine heterocycles, including pyrrolidine, piperidine, piperazine, and morpholine, proceeded smoothly to furnish the target allylic amines in good yields (59–64), presumably due to higher nucleophilicity¹⁵ in comparison to that of their acylic counterparts. Remarkably, reactions of less nucleophilic N-Me-aniline and indoline were successful, as well (65–66). Importantly, employment of allylic amines led to the corresponding di- and triallylamines, albeit in low yields (67–68). Notably, N-methyl-1-naphthyl amine underwent smooth reaction to give 69, which represents an alkylated analogue of the drug naftifine. The scope of this transformation was also tested in a more complex setting. Hence, a secondary amine moiety of ciprofloxacin and desloratadine drugs was successfully converted into allylic amine functionality (70–71).

Importantly, 1,3-dienes containing substituents at C2 and/or C3 positions were also proven competent reaction partners. However, in these cases, 1,4-aminoalkylation products 73 and 75 were obtained selectively (Scheme 2).¹⁶

Scheme 2.

Examples of 1,4-Carboamination of 1,3-Dienes^a

^a0.4 mmol scale; isolated yields.

Moreover, the use of O- and C-nucleophiles, such as phenol (76) and malonate derivatives (78), respectively, led to the 1,2-carbofunctionalization products in reasonable yields (Scheme 3).

Scheme 3.

Employment of O- and C-Nucleophiles^a

^a0.2 mmol scale; isolated yields.

Naturally, we were eager to elucidate the mechanism of this novel transformation. We believe that under these reaction conditions, an interplay between hybrid radical and ionic palladium intermediates takes place. The involvement of radical intermediates was unambiguously supported by the following experiments. Thus, initially formed alkyl radicals from alkyl iodide 2 were trapped by 2,2,6,6-tetramethyl-piperidine-N-oxyl radical (TEMPO) in 22% yield (Scheme 4a, 80^11a). Next, the intermediacy of the allylic radicals was uncovered in the reaction performed without silver triflimide. In that case, the reaction produced a substantial amount of allylic radical dimerization¹⁷ products 81 (Scheme 4b, Table 1, entry 5). A beneficial effect of accessing more electrophilic π-allyl palladium entities in reactions with nucleophiles by employing weakly coordinating anions is well documented.¹⁸ Accordingly, we presumed that employment of triflimide would not only render the Pd-complex more electrophilic but also potentially shift the proposed equilibrium between π-allyl radical- and π-allyl ionic Pd entities to the right (cf. Scheme 1c, A ⇆ B). To validate this assumption, we synthesized model π-allyl palladium complexes bearing a Xantphos ligand¹⁹ with iodide (82) and triflimide (83) counterions and subjected them to the reaction with TEMPO under blue light irradiation (Scheme 4c). In both cases,²⁰ the product of allyl radical trapping (84)^17c,21 was observed in good yields; however, the reaction of 83 occurred at a slower rate as compared to that of complex 82.¹³ Control experiments under the thermal conditions showed that the reaction of 82 with TEMPO still proceeded albeit much slower, while complex 83 was completely nonreactive under such conditions. These results support our hypothesis that a triflimide anion stabilizes the π-allyl Pd complex against homolysis upon irradiation with blue light or thermal conditions, thereby suppressing the undesired radical dimerization process.

Based on the results of the aforementioned studies, the following reaction mechanism is proposed (Scheme 5). First, photoexcited L_nPd⁰ undergoes a single electron transfer (SET) process with alkyl iodide to produce hybrid alkyl palladium radical species C,¹¹ which adds to the terminal position of diene to produce radical species A that exists in equilibrium with π-allyl complex B. In the presence of a triflimide anion, this equilibrium is shifted toward complex B. A subsequent attack of the amine at the latter delivers the carbofunctionalization product and regenerates the palladium catalyst.²²

Scheme 5.

Proposed Mechanism

In conclusion, the first visible-light-induced protocol for the intermolecular aminoalkylation, oxyalkylation, and dialkylation of 1,3-dienes has been developed. This mild multicomponent coupling reaction, which utilizes readily available reaction partners with broad functional group compatibility, does not require employment of exogeneous photosensitizers or external oxidants. It can be used for the late-stage derivatization of complex molecules and drugs. Preliminary mechanistic studies suggest that the radical-polar crossover path is crucial for the success of this reaction. It is expected that this modular, general, and mild method for synthesis of densely functionalized alkenes will find use in organic synthesis and drug discovery.