Nickel-Catalyzed α‑Arylation and Alkenylation of α‑Trifluoromethyl Amines via 1,2-HAT Assisted by DMSO

Abstract

α-Trifluoromethyl amines are privileged scaffolds in pharmaceuticals and bioactive compounds, serving as metabolically stable amide bioisosteres. However, conventional synthetic approaches often suffer from multistep reactions, limited functional group tolerance, and instability of imine precursors. To overcome these limitations, we report a nickel-catalyzed cross-electrophilic coupling constructing diverse α-trifluoromethylamines via a mechanistically distinct pathway involving solvent-enabled 1,2-hydrogen atom transfer (HAT). In this process, the N–O bond is activated by Ni⁽⁰⁾-promoted SET to form transient N-centered radicals, which undergo DMSO-promoted 1,2-HAT to generate nucleophilic C­(sp³)-centered radicals. These radicals then participate in nickel-catalyzed coupling with a wide range of commercially available aryl halides and vinyl bromides. This strategy suppresses competing C–N bond formation and offers several advantages: the use of bench-stable radical precursors, high chemoselectivity, broad substrate scope, mild reaction conditions, and applicability to the late-stage functionalization of natural products and drug derivatives. Furthermore, the method relies exclusively on earth-abundant nickel and commercially available feedstocks. By leveraging polar solvent-assisted HAT to circumvent traditional intermediates, this methodology enables efficient and scalable synthesis of high-value α-aryl- or alkenyl-substituted α-trifluoromethyl amines.

1. Introduction

Amines play a crucial role in synthetic chemistry, medical chemistry, agrochemistry, and materials science. Specifically, α-trifluoromethylamines represent privileged structural motifs prevalent in numerous bioactive compounds and pharmaceuticals (Scheme A). The tunable basicity of the amine and the unique geometry of this structure make it an exceptionally valuable scaffold in the medicinal chemist’s toolkit. These features allow α-CF₃ amines to act as bioisosteres for amide groups, offering notably greater resistance to proteolytic cleavage (Scheme B). Consequently, ongoing efforts focus on developing efficient methods for the synthesis of trifluoromethylated amine derivatives. Current approaches for preparing these compounds are dominated by three main strategies: reductive amination of trifluoromethyl ketones, nucleophilic addition to trifluoromethyl iminium species, and the reaction of trifluoromethyl anions with functionalized imines. Additional methods such as modifying trifluoroacetamides and the hydroamination of trifluoromethyl alkenes have also been investigated (Scheme C). However, these approaches typically suffer from lengthy multistep syntheses, limited tolerance of functional groups, and dependence on heavily prefunctionalized starting materials, restricting their wider application. Therefore, a crucial unmet need for innovative syntheses of α-trifluoromethylamines that circumvent these limitations, though achieving these, poses significant challenges.

1. (A–E) Importance of α-Trifluoromethyl Amines and the Synthetic Methods.

A resurgence in classical approaches has leveraged oxygen- and nitrogen-centered radicals to promote favorable 1,5-hydrogen atom transfer (1,5-HAT) reactions. The resulting translocated radicals can rapidly couple with various partners to form C–N, C–C, C–O, or C–S bonds, endowing this strategy with considerable synthetic versatility. The high fidelity of 1,5-HAT stems from the thermodynamic stability of its chairlike, six-membered transition state, which generally outcompetes HAT processes at other positions. Emerging evidence indicates that the typically favored intramolecular 1,5-HAT may be kinetically surpassed by an operative 1,2-HAT pathway. Recently, an innovative strategy for C–C bond construction via a net 1,2-HAT process mediated by nitrogen-centered radicals has been developed. Often in these processes, nitrogen-containing reagents undergo single-electron oxidation or reduction to generate nitrogen radicals, which subsequently engage in proton transfer to form carbon-centered radicals mediated by a base. These intermediates then participate in radical–radical coupling or olefin addition to forge new C–C bonds.

Cross-electrophile coupling enables direct bond formation between electrophilic partners under reductive conditions. This approach capitalizes on the broader accessibility and superior stability of electrophiles compared to nucleophiles, thereby streamlining synthetic routes and reducing costs. Cross-electrophile coupling has become an important strategy for constructing valuable bonds (e.g., C–C, C–N, and C–O) and supports the modular assembly of complex molecular architectures. Radical nickel catalysis has recently enabled powerful new routes to fluorinated alcohols. Bandini et al. developed cross-electrophilic coupling that generates α-aryl-α-trifluoromethyl alcohols without toxic trifluoroacetaldehyde, using 1,2-HAT to form a reactive α-hydroxy radical. Further, Qing's group reported a light-driven protocol, which achieves similar bond formation through photoinduced charge transfer between a Hantzsch ester and phthalimido trifluoroethanol. Monteith and Rousseaux extended these strategies to a dual Ni/photoredox system, coupling N-alkoxyphthalimides with aryl halides to afford mandelic acid derivatives. Recently, Wang et al. developed an asymmetric Ni/photoredox trifluoroalkylation, affording enantioenriched α-trifluoromethyl acylations through a 1,2-HAT mechanism. Shen and co-workers further demonstrated a photocatalytic radical Brook rearrangement followed by Ni-catalyzed cross-coupling with aryl/heteroaryl bromides, enabling access to chiral α-aryl-α-fluoroalkyl alcohols (Scheme D).

In contrast, achieving 1,2-HAT from N-centered radicals remains a significant challenge and must be explored in cross-electrophile coupling. Notably, the use of a polar solvent to promote the 1,2-HAT of N-radicals under reductive cross-electrophile coupling conditions has no precedent in the literature. Building on prior reports, we hypothesized that a formal 1,2-HAT of amidyl radicals could be feasible with the aid of a suitable solvent under reductive conditions, generating a nucleophilic α-amino-α-trifluoromethyl radical (Scheme E). Herein, we demonstrate advantages over conventional approaches for synthesizing α-trifluoromethyl amines, the key features including (1) use of a bench-stable radical precursor, ensuring safe and practical handling; (2) high chemoselectivity under mild reaction conditions that preserve sensitive functional groups and are compatible with a broad range of commercial and economical aryl electrophiles; (3) a DMSO (dimethyl sulfoxide)-assisted 1,2-HAT mechanism that enables the efficient arylation and alkenylation; (4) Ni⁽⁰⁾-catalyzed N–O bond activation; (5) late-stage functionalization of complex natural products and drug derivatives; (6) DFT calculations supporting a mechanism wherein the radical generated via 1,2-HAT is intercepted by a Ni^(II) species, with crossover pathways being operative in the cross-coupling process.

2. Results and Discussion

2.1. Reaction Exploration

We initiated our investigation into the cross-electrophile synthesis of α-aryl-α-trifluoromethyl amines via 1,2-HAT using tert-butyl [(2,4,5-trifluorobenzoyl)­oxy]­(2,2,2-trifluoroethyl) carbamate (1a) as the radical precursor and iodobenzene (2a) as the model electrophile (Table ). After comprehensive optimization, the Boc-protected amine product (3) was isolated in 82% yield under the following standard conditions: NiCl₂ (10 mol %), dtbpy (12 mol %), MgCl₂ (2.0 equiv), DMSO as a solvent, and Mn as a reductant, at 50 °C for 12 h under a N₂ atmosphere (entry 1). It is noteworthy that replacing NiCl₂ with NiBr₂ under otherwise identical conditions resulted in a comparatively lower yield of 54%. Alternatively, CoCl₂ showed a poor yield. In contrast, alternative catalysts commonly used in cross-electrophile coupling, such as FeCl₃ and CrCl₂, proved ineffective and did not yield any product (entry 2). Subsequently, we evaluated a library of chelating ligands in the model transformation, including bidentate bipyridines (L1–L5), phenanthrolines (L7 and L8), tridentate terpyridine (L6 and L13), and phosphine-based ligands (L9–L12). While all phenanthroline ligands proved ineffective like 1,10-phen afforded the product in 12%, bipyridyl ligands exhibited variable efficacy. For instance, bipyridine (L1) afforded product 3 in a 21% yield. Phosphine-based ligands (e.g., dppp) and terpyridine (L7) yielded no detectable product under the reaction conditions (entry 3). Notably, replacing manganese with zinc severely compromised reaction efficiency (entry 4). The additive (MgCl₂) is crucial for achieving higher yields, while reducing its amount to 0.5 equiv resulted in a lower yield (entry 5). Alternative additives like MgBr₂ and LiBr also afforded lower yields (entries 6–7). Notably, other common polar solvents such as DMA and DMF, when used as a solvent with 3.0 equiv of DMSO as an additive, afforded only trace amounts of the product (entry 8). A DMSO/THF (1:1) solvent system reduced the yield to 16% (entry 9). Reducing the temperature to ambient or diluting the reaction concentration further suppressed the yields (entries 10–12). Control experiments omitting the catalyst, ligand, or reductant resulted in no product formation (entry 13), and the system yielded only 18% in the absence of MgCl₂ (entry 14). It likely stabilized reactive intermediates and facilitated substrate activation. Moreover, replacing aryl iodide with aryl bromide afforded products in slightly lower yields (entry 15). The optimization for triflate electrophiles identified ligand L8 and substoichiometric LiBr (0.5 equiv) as key choice for effective coupling (entry 16). For reactions involving aryl triflates, substoichiometric LiBr was found to be effective, presumably by promoting halide exchange and increasing the electrophilicity of the substrate. Finally, electronic perturbation studies on radical precursors (1a–1d) revealed that electron-withdrawing substituents significantly enhance the reaction efficiency.

1. Optimization of the Reaction Conditions .

^aReaction conditions: 1a (0.2 mmol), 2a (0.4 mmol), catalyst (10 mol %), ligand (12 mol %), Mn (3.0 equiv), DMSO (0.5 mL), MgCl₂ (2.0 equiv) at 50 °C for 12 h under N₂. GC yields.

2.2. Reaction Scope

We next evaluated the substrate scope of this nickel-catalyzed 1,2-HAT-mediated cross-electrophile coupling. The reaction accommodated diverse (hetero)­aryl electrophiles, including iodides, bromides, and triflates as well as vinyl bromides, demonstrating broad applicability of this strategy (Scheme ). The reaction demonstrates exceptional functional group tolerance across electronically diverse systems, accommodating electron-neutral (3 and 5), electron-rich (4, 7, 11, and 13), and electron-deficient (8, 9, and 10) aryl substituents. Crucially, sensitive groups remain intact under optimized conditions, including polar groups (cyano: 18 and 19; unprotected primary amino: 8) and unsaturated functionalities (alkene: 12; TMS-protected alkyne: 17). This broad compatibility underscores the operational practicality of this α-amino-α-trifluoromethylation platform. Notably, the reaction tolerated the −BPin functionality (20), enabling downstream diversification. Exceptional chemoselectivity was evidenced by compatibility with halo groups including chloro, bromo, and iodo groups (21–23) and even disubstituted halides (24). Moreover, this methodology also tolerated disubstituted (7) and trisubstituted (15) arenes. Interestingly, biologically active heterocycles including indole (25), pyridine (26), indoline (27), quinoline (29), morpholine (30), and benzothiophene (27) also afforded products in moderate yields. However, ortho-substituted electrophiles are incompatible with this reaction, affording only trace amounts of the product (31), likely due to steric hindrance.

2. Substrate Scope of Aryl Halides, Triflates, and Vinyl Bromides.

^a Reaction conditions: 1a (0.2 mmol), 2 (0.4 mmol), NiCl₂ (10 mol %), dtbpy (12 mol %), Mn (3.0 equiv), DMSO (0.5 mL), MgCl₂ (2.0 equiv) at 50 °C for 12 h under N₂; isolated yields.

^b NiCl₂ (10 mol %), L8 (12 mol %), Mn (3.0 equiv), DMSO (0.5 mL), 4 Å MS (0.2 g), LiBr (0.5 equiv), at 50 °C for 12 h under N₂.

^c ArX (1.0 equiv).

Collectively, these sustainable electrophile streams enable the efficient construction of diverse α-aryl-α-trifluoromethyl amines through nickel-catalyzed cross-electrophile coupling. To evaluate the robustness of the developed methodology, vinyl bromides were synthesized from aldehyde feedstocks, and the substrate scope was systematically explored. The results demonstrated that electron-neutral (32, 34, and 35), electron-donating (33 and 37–39), and electron-withdrawing (36, 40, and 41) substituents were all well-tolerated under the standard reaction conditions, affording the desired products in good to excellent yields. Furthermore, we evaluated diverse fluoroalkyl groups (42–48) bearing varied substituents, all demonstrating compatibility under the standard reaction conditions.

Capitalizing on the broad functional group tolerance of this method, we applied this nickel-catalyzed 1,2-HAT platform to the late-stage modification of medicinally significant scaffolds (Scheme ). This nickel-catalyzed 1,2-HAT platform successfully achieves direct late-stage α-amino-α-trifluoromethylation of structurally complex molecules using diverse electrophilic partners including aryl iodides, bromides, and vinyl bromides. The methodology demonstrates exceptional versatility through its application to pharmaceutical agents such as nonsteroidal anti-inflammatory drugs (NSAIDs) mefenamic acid (50 and 57) and naproxen (54), the antidiabetic drug repaglinide (56), and natural products including menthol (51), vitamin E (55), and diacetone-β-d-fructose (53). Furthermore, a vinyl bromide-containing derivative of HSV-1 inhibitor gallic aldehyde (52) underwent efficient functionalization, confirming broad applicability. Crucially, the strategy enables direct diversification of the lapatinib intermediate (49) converted into the corresponding α-aryl-α-trifluoromethyl amines. Moreover, deprotected α-aryl-α-trifluoromethyl amine (59) undergoes amidation with retinoic acid (vitamin A) to afford α-aryl-α-trifluoromethyl amide derivative (58). The successful execution of these transformations under mild conditions, coupled with their excellent functional group tolerance, underscores the method’s unique capacity for the modular modification of advanced molecular architectures.

3. Late-Sage α-Arylation or Alkenylation of α-Trifluoromethyl Amines.

2.3. Synthetic Applications

To explore the synthetic potential and practical applicability of the synthesized products, we performed further transformations on α-aryl-α-trifluoromethyl amines (Scheme ). Initially, gram-scale reactions were performed for two representative substrates, compounds 3 and 21, which furnished the desired products in 68% and 59% yields, respectively. In addition, deprotection of compound 21 afforded free amine 59 in 88% yield. The α-aryl-α-trifluoromethyl amines proved to be versatile intermediates for a range of nucleophilic substitution reactions. Methylation with CD₃OTs efficiently introduced a deuterated methyl group at the nitrogen atom, affording the N-CD₃ derivative (60) in an excellent yield. Similarly, allylation with allyl bromide (63) furnished the N-allyl derivative (64) in 62% yield, while carbamoylation with benzyl alcohol (61) under Ti­(O ⁱ Pr)₄-mediated conditions delivered compound 62 in 62% yield. The deprotected α-CF₃ amine (59) undergoes a condensation reaction with 3-pyridine aldehyde (65) to furnish the corresponding trifluoromethyl containing imine (66) in good yield (72%), demonstrating the broad synthetic utility of this scaffold. Moreover, the α-CF₃ amine (59) underwent a ligand-free Suzuki cross-coupling reaction with boronic acid (67) under solvent-free conditions, affording the coupled product (68) in 89% yield. Notably, compound 68 could serve as a key synthetic precursor to commercially available drug odanacatib (69).

4. Gram-Scale Synthesis and Modification of α-Aryl-α-trifluoromethyl Amines.

2.4. Mechanistic Studies and Proposed Mechanism

To elucidate the reaction mechanism of this protocol, a series of mechanistic studies were conducted. First, radical trapping experiments were performed (Scheme A). Definitive evidence for radical intermediates was obtained through parallel experimental observations; significant yield suppression occurred with 5.0 equiv of TEMPO (trace), while HRMS analysis unambiguously detected the corresponding TEMPO-radical adduct (70). The use of Ni­(COD)₂ as a preformed Ni⁽⁰⁾ source afforded product 3 in 61% yield, supporting the potential involvement of Ni⁽⁰⁾ species in the catalytic cycle (Scheme B). Further, stoichiometric experiments highlighted the essential role of the reductant. The yield increased markedly from 23% to 48% by the introduction of Mn as a reductant (Scheme C). The generation of desired products without Mn suggests that Ni⁽⁰⁾ participates in single-electron transfer (SET) activation of the N–O bond to give a radical, while the reaction without Mn might proceed through the pathway of comproportionation of the Ni intermediate. To further probe the catalytic cycle, the Ni^(II) complex (71) was employed for the stoichiometric reaction with 1a, affording product 9 in 28% yield (Scheme D). Subsequently, a competitive experiment between bromobenzene and stoichiometric 71 revealed the preferential formation of 9 (26% yield) over 3 (12% yield) (Scheme E). These results support the formation of a Ni^(II) intermediate during the reaction pathway. Finally, to clarify the role of DMSO and rule out its function as a hydrogen atom donor, the reaction was performed in deuterated DMSO. The absence of deuterium incorporation into the product confirmed that no hydrogen exchange occurs between the solvent and the in situ generated radicals (Scheme F). Kinetic analysis revealed a zero-order dependence on the concentration of electrophile 2a and a first-order dependence on both radical precursor 1a and a nickel catalyst. These findings suggest that the interaction between the nickel catalyst and 1a is involved in the rate-determining step. This interpretation was further supported by Hammett studies, which yielded a positive ρ-value, indicating that electron-withdrawing groups on the substrate accelerate the reaction, while electron-donating groups decelerate it. Taken together, the rate-determining step likely involves the interaction between the radical and the catalyst, potentially through the formation of a radical–catalyst complex. The positive ρ-value is consistent with a transition state that is stabilized by electron-withdrawing substituents on the electrophile (Scheme G).

5. (A–G) Mechanistic Studies.

The above-mentioned results were further supported by detailed density functional theory (DFT) studies (Figure ). Initially, compound 1a undergoes SET mediated by a commercially available nickel catalyst. Calculations indicate that SET is challenging with Ni^(I) but favorable with Ni⁽⁰⁾ (see Scheme S1 in the Supporting Information), leading to the formation of intermediate 1-A with a ΔG of 1.0 kcal/mol. Intermediate 1-A then undergoes facile N–O bond cleavage via TS-3, with a ΔG of 5.9 kcal/mol, to generate N-centered radical 1-B, which is stabilized with a ΔG of −42.9 kcal/mol. It indicated that the facile Ni⁽⁰⁾-mediated SET and formation of a transient N-centered radical account for the observed rate dependencies. Importantly, the 1,2-HAT takes place through the three-membered transition state TS-4 with a ΔG ^‡ of −22.5 kcal/mol assisted by the polar solvent DMSO. Alternatively, in the absence of DMSO, the corresponding pathway TS-5 with a ΔG ^‡ of −3.9 is highly unfavorable or inaccessible (Figure A). Furthermore, the catalytic operative cycle starts by the reduction of the commercially available Ni^(II) complex using a reductant that may convert it to Ni⁽⁰⁾ or Ni^(I), which may undergo oxidative addition. The studies showed that Ni⁽⁰⁾ is less likely to undergo oxidative addition than Ni^(I) (see Scheme S3 in the Supporting Information). So, the operative cycle starts from the in situ generated active catalyst Ni^(I) Cat-1. The iodine atom in iodobenzene (2a) coordinates with Cat-1 to form intermediate IM-1. This intermediate then undergoes oxidative addition via three-membered ring transition state TS-1, leading to the Ni^(III) intermediate IM-2. This triplet oxidative addition pathway is highly favorable, exhibiting both kinetic favorability with a ΔG ^‡ of 10.0 kcal/mol (Cat-1 → TS-1) and a ΔG of −2.9 kcal/mol (TS-1 → IM-2). The intermediate IM-2 is subsequently reduced by Mn powder to yield Ni^(II) intermediate IM-3. During this process, either halogen atom (Cl or I) of IM-2 can undergo reduction. The Ni^(II) intermediates IM-3 and IM-3′ are formed by the removal of Cl or I, respectively, with corresponding energies of −53.1 and −56.6 kcal/mol. Subsequently, intermediate IM-3 undergoes C­(sp³) radical addition to yield IM-4 and IM-4′, with calculated energies of −56.0 and – 51.0 kcal/mol, respectively. The lower energy of IM-4 indicates that this radical insertion pathway is more favorable and represents the preferred route. Finally, IM-4 undergoes reductive elimination via a three-membered transition state (TS-2) with an activation energy of −38.6 kcal/mol, leading to the formation of product 3 and regeneration of the active Ni^(I) catalyst (Cat-2), which is available for the subsequent catalytic cycle (Figure B). Overall, kinetic analysis and Hammett studies indicate that the overall reaction rate is governed by a radical rebound step, corresponding to the capture of the carbon-centered radical by the nickel catalyst. This conclusion is supported by the observed sensitivity of the reaction rate to electronic substituent effects. DFT calculations were carried out to examine the intrinsic energetics of individual elementary steps along the proposed catalytic pathway. While reductive elimination from the Ni^(III) intermediate IM-4 is associated with the highest computed activation barrier, this transformation occurs subsequent to the turnover-limiting radical rebound step and proceeds from a thermodynamically deep minimum. As such, reductive elimination does not control the overall reaction rate under steady-state catalytic conditions. Together, the experimental and computational data provide a coherent mechanistic picture in which radical rebound is rate-determining, while reductive elimination is kinetically accessible once the Ni^(III) intermediate is formed.

1.

DFT calculations. (A) Ni⁽⁰⁾-promoted SET for radical generation; (B) model DFT studies on the operative catalytic cycle.

Based on experimental findings, DFT calculations, and prior literature, we proposed a reaction mechanism (Scheme ). The cycle begins with the reduction of the Ni^(II) precatalyst (A) by manganese powder, facilitated by ligand coordination, to form an electron-rich Ni⁽⁰⁾ complex (B). Species B then undergoes single-electron transfer (SET) to radical precursor 1a, cleaving the N–O bond and generating an N-centered radical (H) alongside a Ni^(I) intermediate (C). Oxidative addition of an aryl halide 2 to C yields a Ni^(III)–aryl species (D), which is reduced by Mn to afford the Ni^(II)–aryl intermediate (E). Simultaneously, the N-centered radical (H) undergoes a rapid, DMSO-promoted 1,2-hydrogen atom transfer (HAT) to give the nucleophilic C-centered radical (I). This radical is captured by the Ni^(II)–aryl complex (E), forming the Ni^(III) intermediate (F). Subsequent reductive elimination releases the final α-aryl-α-trifluoromethylamine product and generates the Ni^(I) species (G). Finally, reduction of G by Mn closes the cycle by reforming the active Ni⁽⁰⁾ catalyst (B).

6. Proposed Catalytic Cycle.

3. Conclusions

This work establishes a nickel-catalyzed reductive α-arylation and alkenylation of α-trifluoromethyl amines through a solvent-gated radical translocation mechanism. Key to this strategy is DMSO-promoted 1,2-HAT from transient N-centered radicals to C-centered radicals, which couple with diverse abundant and commercial aryl/vinyl electrophiles (triflates, bromides, iodides, and vinyl bromides) under mild conditions, demonstrating exceptional chemoselectivity and functional group tolerance (esters, amines, Bpin, alkenes, electron-donating/withdrawing/neutral groups, and heterocycles). The method also enables efficient late-stage functionalization of complex drug derivatives and natural products, and the successful diversification of products highlights their synthetic utility. By circumventing traditional imine intermediates, this approach offers a scalable route to high-value α-trifluoromethyl amine scaffolds from readily available feedstocks. Mechanistic studies, including DFT calculations, support the proposed solvent-enabled HAT pathway. Overall, this work enables a general approach to C­(sp³)–C­(sp²) bond formation through the merger of radical translocation and cross-electrophile coupling, paving the way for the synthesis of privileged structural motifs.

The Supporting Information is available free of charge at ACS Publications Web site. The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c01732.

• Materials and methods; experimental procedures; optimization studies; ¹H NMR, ¹³C NMR, ¹⁹F NMR, and MS data (PDF)

The authors declare no competing financial interest.