Nickel catalyzed C-N coupling of haloarenes with B₂N₄ reagents

Abstract

Carbon-heteroatom bond (especially for C-N bond) formation through nickel catalysis has seen significant development. Well-established Ni(0)/Ni(II) redox cycle and photoinduced Ni(I)/Ni(III) redox cycle have been the dominant mechanisms. We report a thermally driven Ni-catalyzed method for C-N bond formation between haloarenes and B₂N₄ reagents, yielding N,N-dialkylaniline derivatives in good to excellent yields with broad functional group tolerance under base-free conditions. The catalytic protocol is useful for base-sensitive structures and late-stage modifications of complex molecules. Detailed mechanistic studies and density functional theory (DFT) calculations indicate that a Ni(I)/Ni(III) redox cycle is preferred in the C-N coupling process, and B₂N₄ reagent serves both as a single electron transfer donor and a N,N-dialkylation source.

Introduction

N,N-dialkylanilines are key scaffolds in bioactive compounds^1,2,3 and important building blocks for organic synthesis^4,5,6,7,8,9. The development of more efficient and sustainable N-alkylation processes has continuously attracted the attention of chemists^10,11,12,13. While direct N-alkylation with alkyl halide (e.g., toxic methyl iodide)¹⁴ and reductive amination^15,16 remains in use, transition-metal (Pd/Ni etc.) promoted C-X amination reactions utilizing dialkylamines or their equivalences through well-established M(0)/M(II) redox cycle^{17,18,19,20,21,22} have emerged as effective alternatives (Fig. 1a). Productive catalysis through the M(0)/M(II) cycle has been achieved through elegant ligand designs based on phosphorous^23,24,25 or N-heterocyclic carbenes (NHC)^26,27,28,29. In recent years, Ni(I)/Ni(III)^{30,31,32,33,34} redox cycle has been confirmed to be more efficient and versatile since high-valent Ni(III) species lead to a faster and energetically downhill C-N reductive elimination (RE)²⁸. Recent developments in photoredox catalysis^{35,36,37,38,39} (Fig. 1b) have demonstrated that C-N bond formation can be achieved by using simple ligands (even without ligands) under exceptionally mild reaction conditions. MacMillan^40,41 and Buchwald⁴² developed a nickel/photochemical C-N bond formation methodology without complex ligands between haloarenes and amines. Miyake⁴³ also demonstrated a photocatalyst-free C-N cross coupling of electron deficient haloarenes via ultraviolet light photoexcitation of Nickel-amine complexes. In the absence of additional photocatalysts, both Xue^44,45 and our group⁴⁶ have reported the strategies for C-N bonds construction via photo-induced nickel catalysis. Most recently, Ritter⁴⁷ reported a Ni(I)/Ni(III)-photo-catalyzed C-heteroatom bond formation of pre-functionalized arylthianthrenium salts based on simple Ni(II) salt.

Fig. 1: Transition-metal catalyzed C-N bond formation of haloarenes.

a transition metal catalyzed amination with haloarenes. b Ni(I)/Ni(III) redox cycles promoted by single electron reductant. c transition metal catalyzed borylation utilizing B₂O₄ type regents. d Ni(I)/Ni(III) cycle promoted C-N coupling of B₂N₄. LG leaving group, FG functional group.

While these approaches have found wide use for C-N bond formation, they are primarily hampered by their inability to efficiently react with electron rich haloarenes and base sensitive substrates, since most of these reactions require the consumption of super-stoichiometric amounts of exogenous bases as sacrificial reagents. Still remains undeveloped, there are limited reports of base-free thermally sustained Ni(I)/Ni(III) coupling strategies, utilizing external single-electron reductants such as manganese and zinc^48,49,50,51.

While B₂O₄ type diboron reagents, such as tetrahydroxydiboron (B₂(OH)₄), are applied in the Suzuki-Miyaura coupling reaction^52,53 for C-C bond construction, bis(pinacolato)diboron (B₂pin₂) and B₂pin₂ derivatives (B₂cat₂, B₂nep₂ etc.), are often used in C-B bond formation through Miyaura borylation reaction (Fig. 1c)^54,55. Diboron compounds are common and useful single electron transfer (SET) reagents^56,57 in photoredox catalysis, but they are less used as SET reagents in transition-metal catalysis. While seminal reports^58,59,60 in the field of C-C/C-X bonds formation utilizing B₂O₄ type diboron reagents have been accomplished, in sharp contrast, B₂N₄-type diboron reagents, such as tetrakis(dimethylamino)diboron (B₂(NMe₂)₄), have not been found to be useful for transition-metal-catalyzed C-N bond construction. Could a B₂N₄ type diboron reagent be used both as a SET reagent and a “N” source by the Ni(I)/Ni(III) redox cycle without photoexcitation under base-free and mild reaction conditions? Herein, we report such a case, as shown in (Fig. 1d). Remarkably, nickel offers distinct advantages over palladium, such as earth-abundant metal catalyst and a reduced propensity for β-hydride (or β-hydrogen) elimination⁶¹ when working with alkyl fragments.

N,N-dimethyl anilines are most valuable intermediates among N,N-dialkylanilines used for the preparation of bioactive molecules, polyester resins and dyes. Direct N,N-dimethylation of haloarenes, especially electron rich haloarenes, with dimethylamine is more challenging because of its low boiling point (~7 °C). Several protocols^62,63,64,65 for the dimethylamination of haloarenes using the equivalents of dimethylamine have been reported as well. However, the pre-functionalization of dimethylamine, the requirement for strong base and complex ancillary ligands are necessary, making it nontrivial for high site selectivity. In this work, our photo-free “two-in-one” protocol makes it complementary to other C-N bond formation strategies (especially for syntheses of N,N-dimethyl anilines), whose effectiveness is exemplified through working with base-sensitive partners and late-stage modifications of a series of bioactive complex molecules.

Results and discussion

Reaction design and optimization

Building on this basis, we set out to examine the possibility of base-free, complex ligand-free Ni-catalyzed C-N coupling of haloarenes with commercially available B₂(NMe₂)₄, aiming to establish a protocol that displays an effective N,N-dimethyl aniline synthesis. In our initial study, the coupling reaction of 4-bromophenoxybenzene (1a) with B₂(NMe₂)₄ was selected as a model system for reaction development, as shown in (Table 1). Various transition metal catalysts were systematically evaluated for their efficacy in our study (Table S1). These experiments demonstrated that the N,N-dimethyl aniline product 2a formed with 94% isolated yield in the presence of 10 mol % of Ni(acac)₂ catalyst. Copper and palladium catalysts showed no reactivity for this reaction (entries 1-2). The yield of 2a was slightly reduced with NiCl₂•6H₂O as catalyst (entry 3). In contrast, Ni(0) catalysts, for example Ni(COD)₂, did not show activity (entry 4). Thus, Ni(II) is indispensable for optimizing reaction efficiency in this C-N cross coupling reaction. Subsequent studies revealed DMF to be the optimal solvent for the coupling reaction (entries 5-7). Lower reaction temperature and reduced loading of B₂(NMe₂)₄ significantly diminished the yield of 2a (entries 8-9). Control experiments revealed that no reaction occurred without Ni(II) catalyst (entry 10).

Table 1 Optimization of reaction conditions

Scope of substrates

With the optimal condition in hand, we first investigated the scope and limitations of our Ni-catalyzed C-N coupling reaction with respect to the para-substituted bromoarenes (Fig. 2A). This method showed excellent functional group tolerance, both electron-donating and electron-withdrawing substituents could be well accommodated, including alkyl (2c), alkenyl (2u), alkynyl (2v), thioether (2e), free hydroxyl and amino (2s, 2t), cyclopropyl (2w), (hetero)aryl (2o-2r). Carbonyl functionalities such as esters (2l), amides (2m-2n) and ketones (2w-2x) that might be susceptible to acylation or condensation reaction were not affected when B₂(NMe₂)₄ was used. Nitrogen-, oxygen-, sulfur-, boron-, silicon-containing haloarenes were also prepared in high yields.

Fig. 2: Synthetic scope of nickel catalyzed N,N-dimethylation of bromoarenes.

A The functional group tolerance of para-substituted haloarenes; B Variation of haloarene coupling partner; C Variation of heterohaloarene coupling partner. Reaction condition: substrates (0.2 mmol, 1.0 equiv.), Ni(acac)₂ (0.02 mmol, 10 mol%), B₂(NMe₂)₄ (0.2 mmol, 1.0 equiv.), DMF (0.5 mL), stirring at 80 °C for 24 h. Yields are reported for material obtained following purification and isolation.

We next extended the protocol to the synthesis of poly-substituted aniline derivatives, bearing in mind their importance in medicinal chemistry (Fig. 2B). Bromoarenes bearing electron-donating and electron withdrawing groups at ortho and meta position were all compatible to give the corresponding anilines in good to excellent yields (2y-2ac), bulky ortho-OPh-bromoarenes (2aa), albeit with slightly diminished yields. Di-substituted bromoarenes with electron-donating and electron-withdrawing groups, such as 3,4- 2,4- 2,5 and 3,5-disubstituted bromoarenes could give good to high yields (2ad-2al), showcasing the compatibility of our method with a broad array of functionality. Naphthyl bromides (2am-2ap) were also good partners to afford desired products. Given the critical role of heterocyclic motifs in drug development, we extended this methodology to diverse heteroaryl bromides, spanning benzothiophene, benzofuran, indole, pyridine, coumarin, quinoxaline, (iso)quinoline, pyrazine, and pyrimidine scaffolds. All of these substrates yielded the expected products in good to excellent yields (2aq-2bc, Fig. 2C).

Due to the mild and base-free reaction conditions, we envisioned that our strategy might be applicable to substrates that would readily undergo undesired racemization, elimination and decomposition reactions in the presence of strong bases. Chiral substituents at the α-carbonyl position are highly susceptible to epimerization, while the presence of strong bases can further induce decomposition. Armed with our standard conditions, no ee erosion was observed as exemplified by the enantio-retention experiment utilizing 4-Br-L-phenylalanine as model substrate (3a, Fig. 3A). In contrast, only decomposition product were observed under classic Buchwald-Hartwig C-N coupling conditions (Fig. S2). In addition, elimination-prone coupling partners bearing alkyl chloride (3b) were suitable coupling partners. The alkyl halides would valuable for subsequent synthetic applications. Furthermore, heteroaryl rings (Fig. 3B), such as oxazole (3c), benzothiazole (3d), indazole (3e) and imidazole (metronidazole derivative 3f), which would be decomposed under strong basic conditions⁶⁶, could be well tolerated in our system. This result inspired us to pursue the three-step synthesis of PET (the positron emission computed tomography) tracer analog 3g for Alzheimer’s disease (AD) in 66% overall yield.

Fig. 3: Application of mild and base-free conditions in the functionalization of substrates prone to decomposition and racemization.

A Enantioretention experiment. B Base-sensitive coupling partners. C Synthesis of base-sensitive AD PET tracer analog. Condition A: 2-amino-6-fluorobenzothiazole (1.0 equiv.), ethylene glycol (4 equiv.), KOH (6.0 M, 30 equiv.), reflux at 140 °C for 24 h; condition B: bis(2-amino-5-fluorophenyl)disulfide (1.0 equiv.), 4-bromobenzaldehyde (1.02 equiv.), sodium metabisulfite (1.02 equiv.), DMSO (0.17 M), reaction at 120 °C for 2 h; condition C: 2-(4-Bromophenyl)−6-fluorobenzothiazole (0.2 mmol), B₂(NMe₂)₄ (0.2 mmol), Ni(acac)₂ (10 mol%), DMF (0.5 mL), reaction for 24 h under 80 °C. AD Alzheimer’s disease, PET the positron emission computed tomography. See SI for full experimental details and conditions.

Different B₂N₄ reagents^67,68 were next used to explore the applicability of these reagents in forming diversified amino products (Fig. 4). B₂N₄ reagents derived from pyrrolidine and piperidine could be utilized to construct corresponding N,N-dialkylanilines (4a-4d). For example, electron-rich and electron-deficient aryl bromides were well tolerated when reacting with tetrakis(pyrrolidino)diborane under standard conditions (4a-4c). Furthermore, more valuable tertiary aniline products (4e-4g) were successfully prepared with good to excellent yields using corresponding B₂N₄ reagents derived from several secondary amines. Unfortunately, aliphatic primary amine-derived diboron species didn’t work under standard conditions, most likely due to the undesired hydrolysis of diboron species. For example, only benzyl amine was detected when utilizing B₂(BnNH)₄ as B₂N₄ reagent. To our delight, pharmaceutical-related diarylamines could be prepared from B₂(NHPh)₄ with (hetero)aryl bromides (4h-4j). Of note is that more sterically hindered B₂N₄ reagents were also good partners to afford diarylamine product (4k).

Fig. 4: C-N bond formations with diverse B₂N₄ reagents.

Reaction condition: substrates (0.2 mmol, 1.0 equiv.), Ni(acac)₂ (0.02 mmol, 10 mol%), B₂(NMe₂)₄ (0.2 mmol, 1.0 equiv.), DMF (0.5 mL), stirring at 80 °C for 24 h. See SI for full experimental details and conditions.

The broad functional group tolerance for this dimethylamination of haloarenes encouraged us to test this strategy for late-stage functionalization of complex molecules, such as natural products and active pharmaceutical ingredients (APIs). As shown in Fig. 5, N,N-dimethyl anilines derived from gallic acid (5a), thymol (5c-5d), estradiol (5e), tyramine (5i), as well as a variety of APIs (5b, 5f-5h, 5j-5l) bearing polar groups, basic heteroatoms, and heterocyclic frameworks were readily synthesized, confirming versatility of our method. Furthermore, a wide variety of powerful synthetic transformations of the resulting N,N-dimethyl complexes were also demonstrated using N,N-dimethyl aniline derivatives (Fig. S5). These valuable transformations revealed the possibility of global derivatization around the core N,N-dimethylaniline structure, and such derivatization can turn N,N-dimethylaniline into a versatile synthetic building block in organic synthesis.

Fig. 5: Late-stage N,N-dimethylation of complex molecules.

Reaction condition: substrates (0.2 mmol, 1.0 equiv.), Ni(acac)₂ (0.02 mmol, 10 mol%), B₂(NMe₂)₄ (0.2 mmol, 1.0 equiv.), DMF (0.5 mL), stirring at 80 °C for 24 h. See SI for full experimental details and conditions.

Mechanistic studies

The mechanism of this Ni-catalyzed C-N coupling of haloarenes with B₂N₄ type reagents was then investigated. A series of probing experiments were conducted to investigate the reaction pathway. No desired products were observed with other potential intermediates such as dimethylamine, N,N-dimethylformamide and tris(dimethylamino)borane (Fig. S7). In the radical trapping experiments (Fig. 6A), when phenyl tert-butyl nitrone (PBN) was introduced into the experiments as a spin-trapping reagent, radical EPR signals were observed (Fig. 6Aa). A g-factor of 2.005 indicates that an organic radical species, whose adduct with PBN was detected by high-resolution mass spectrometry (HRMS [M + K⁺]: 513.3418), was generated from mixture of Ni(acac)₂ and B₂N₄ reagent after heating. In fact, radical intermediate A1 was identified by oxygen free high-resolution mass spectrometry (HRMS [M + H⁺]: 298.2716), confirming the single electron transfer progress between Ni(II) and B₂N₄ reagent. Notably, the C-N coupling reaction was completely suppressed using TEMPO as a radical inhibitor, and the adduct of TEMPO-acac was detected by HRMS (Fig. 6Ab; [M + H⁺]: 256.1914). We speculated that TEMPO abstracted acac radical⁶⁹ from key radical cage Ni(acac), releasing non-catalytic nickel(0). Additionally, no obvious inhibition of C-N bond formation was observed using 1,1-diphenylethylene as a neutral radical scavenger, indicating that no radical species was involved in the Ni redox cycle (Fig. 6Ac). In this trapping experiment, a key radical intermediate was captured by 1,1-diphenylethylene (HRMS [M + H⁺]: 479.3736).

Fig. 6: Mechanistic studies of nickel catalyzed C-N coupling of haloarene with B₂(NMe₂)₄.

A Radical trapping experiments: (Aa) PBN trapping experiment; (Ab) TEMPO trapping experiment; (Ac) 1,1-diphenylethylene trapping experiment. PBN phenyl tert-butyl nitrone; TEMPO 2,2,6,6-Tetramethyl-1-piperinedinyloxy; EPR electron paramagnetic resonance. B Ni(I)/Ni(III) pathway confirmation. C Ni(I) confirmation through comproportionation reaction (C1) and standard reaction condition (C2) by electron paramagnetic resonance (EPR) experiments in DMF solvent. D Key intermediates detected by HRMS. See SI for full experimental details and conditions.

Next, we performed several experiments to confirm whether the dominant productive pathway involves a Ni(0)/Ni(II) or Ni(I)/Ni(III) redox cycle. As shown in Fig. 6B, when utilizing Ni(COD)₂ as the Ni(0) catalyst, instead of C-N coupling product 2a, homo-coupling product 2a’ (room temperature or 80 °C) from 4-bromophenoxybenzene was isolated as the single product through the well-established M(0)/M(II) redox cycle (i.e., Negishi coupling⁷⁰, Suzuki-Miyaura coupling^51,52), revealing that Ni(0) was not involved in the Ni redox cycle. Proceeded at room temperature, no desired C-N coupling product 2a was detected no matter Ni(0) or Ni(II) was used. Well-defined examples of comproportionation reactions⁷¹ of Ni(0) and Ni(II) to generate Ni(I) species have been reported since Heimbach’s pioneering work⁷². Based on these general comproportionation strategies of Ni(0) and Ni(II), after 4 h reaction of Ni(0) and Ni(II) (Ni(COD)₂: Ni(acac)₂ = 20 mol%: 20 mol%) at room temperature, the Ni mixture was added into the reaction system, the desired C-N coupling product was detected with 5% yield (80% yield at 80 °C), indicating that an active Ni(I) species was formed through comproportionation reaction between Ni(0) and Ni(II). The formation of the Ni(I) species was evidenced by electron paramagnetic resonance (EPR) spectroscopy, which revealed a Ni(I) species with g_iso = 2.366 from the comproportionation reaction (Fig. 6C, C1). We also conducted EPR measurements on the reaction mixture obtained under standard reaction conditions after 10 h. The unpaired electron in Ni(I) was directly observed by EPR (g_iso = 2.349 at 100 K Fig. 6C, C2). While the EPR of C1 is clean, that of C2 is likely affected somewhat by the presence of other organic radicals, for example, A1. Perhaps more straightforwardly, radical intermediate A1 and related fragments A2-A4 (m/z of [M + H⁺]: 298.2716 for A1; m/z of [M+Na⁺]: 256.0756 for A2; m/z of [M + H⁺]: 199.1612 for A3; m/z of [M + H⁺]: 101.1249 for A4; m/z of [M+Na⁺]: 290.9448 for A5) and key Ni(III) intermediates (m/z of [M-acac] ⁺: 305.9188 for B; m/z of [M + H⁺]: 371.1028 for C) were also identified by oxygen free high-resolution mass spectrometry (see supporting information for detailed oxygen free HRMS experiments) after 1 h under standard reaction conditions (Fig. 6D), which supports the potential energy surface obtained from the following density functional theory (DFT) calculations, indicating a Ni(I)/Ni(III) redox cycle was involved in the nickel catalyzed C-N coupling of haloarene.

Computational studies were performed to gain an in-depth understanding of the reaction mechanism (Fig. 7, Supplementary Computational Details). Ni(acac)₂, B₂(NMe₂)₄, and PhBr were selected as model substrates for this theoretical study. Based on the results of mass spectrometry experiments, we conducted DFT calculations to evaluate the feasibility of forming the B₂(NMe₂)₄(acac) radical species A1 and Ni^I(acac) via single electron transfer (SET) process between Ni(acac)₂ and B₂(NMe₂)₄. Additionally, we also carried out a comprehensive assessment of the solvents’ role in our reaction (Fig. S26). The DFT results indicate that the Gibbs free energy (ΔG) for this reaction is 6.4 kcal/mol, suggesting that the reaction is thermodynamically feasible. After the SET process, DMF binds to Ni^I(acac) to form intermediate Int1, and then oxidative addition occurs via transition state TS1, resulting in the Ni(III) intermediate Int2 with an activation barrier of 13.1 kcal/mol. Next, the Ni(III) intermediate Int2 coordinates with B₂(NMe₂)₄ to form Int3. Due to boron’s metalloid property, Int3 undergoes a transmetallation-like four-membered ring transition state TS2 (25.1 kcal/mol) to form Int4. In fact, intermediate C (m/z of [M + H⁺]: 371.1028 in Fig. 6D), derivative of Int4, has been detected by HRMS. Int4 then produces the Ni(I) intermediate Int5 via the reductive elimination transition state TS3 (20.3 kcal/mol). Finally, another molecule of DMF replaces PhNMe₂ in Int5 to yield the aniline product and regenerate Int1 for the next catalytic cycle.

Fig. 7: Energy profile of reduction of Ni(II) species and subsequent Ni(I)/Ni(III) catalytic cycle.

A1: B₂(NMe₂)₄(acac) radical species; A2: B₂(NMe₂)₃Br; int intermediate, TS transition state.

Based on these findings and previous reports^{40,41,42,43,44,45,46,47,73,74,75,76,77,78,79,80}, we propose the following mechanism (Fig. 8). First, single electron transfer progress between Ni(acac)₂ and B₂(NMe₂)₄ (A) generates B₂(NMe₂)₄ radical species A1 and Ni^I(acac). Next, Ni^I(acac) undergoes oxidative addition with the aryl bromide 1a, leading to the formation of Ni(III) species B. Subsequent amination of B with B₂(NMe₂)₄ delivers C and A2. Feasible reductive elimination of C enables the formation of product 2a and regenerates Ni(I) catalyst.

Fig. 8: Proposed reaction mechanism of nickel catalyzed C-N coupling of haloarene with B₂(NMe₂)₄.

A: B₂(NMe₂)₄; A1: B₂(NMe₂)₄(acac) radical species; A2: B₂(NMe₂)₃Br; A3: B(NMe₂)₂(acac); A4: HB(NMe₂)₂; A5: B₂(NMe₂)₂Br₂; HAT hydrogen atom transfer, OA oxidative addition, RE reductive elimination.

Additionally, A2 is also responsible for 2a synthesis through Ni(I)/Ni(III) redox cycle (green line). 85% yield of 2a can be isolated with 0.7 equivalent of B₂(NMe₂)₄, indicating that, in addition to B₂(NMe₂)₄, other dimethylamine sources such as A2 participate in the C-N bond formation reaction. The by-product (A5) of A2 has been captured by HRMS, showcasing the fact that A2 is one of the dimethylamine sources.

In conclusion, we have developed an efficient C-N cross coupling methodology for N,N-dialkylation of haloarenes with B₂N₄ reagents by complex ligand-free and base-free nickel catalysis without an external reductant. Compared with the traditional N,N-dialkylaniline synthetic strategies, this protocol benefits from high functional group tolerances, reduced synthetic steps and feasible base sensitive aniline derivatives syntheses. The protocol’s excellent functional group tolerance enables the functionalization of a variety of biologically relevant compounds, representing a valuable potential industrial application of the simple nickel catalyst system. A detailed mechanistic investigation and DFT calculations give a reasonable explanation for the Ni(I)/Ni(III) redox pathway.

Methods

General procedure for nickel catalyzed C-N coupling of haloarenes with B₂N₄ reagents

An oven-dried 4 mL vial was charged with a magnetic stir bar, aryl bromides (0.2 mmol, 1.0 equiv.), Ni(acac)₂ (0.02 mmol, 10 mol%), B₂N₄ (0.2 mmol, 1.0 equiv.), DMF (0.5 mL) in the glove box. The vial was sealed with a plastic cap and then stirring was achieved by placing the assembled reactor at 80 °C on IKA C-MAG HS 7 control magnetic stir bars for 24 h. After reaction completion, the reaction was quenched with H₂O and diluted with EtOAc. The resulting mixture was separated and extracted with EtOAc (three times). The combined organic layer was dried over anhydrous Na₂SO₄, filtered and concentrated in vacuo. The reaction mixture was purified by fresh silica gel chromatography to afford the desired product.