Amination of Phenols and Halophenols via Pyridinium–Iridium Dual Photocatalysis

Abstract

In this study, we present a photochemical method for the amination of phenols (C─H) and halophenols (S_NAr), facilitated by dual catalytic pathways involving both Ir(III) photocatalysis and phenol–pyridinium electron donor–acceptor complexation. By incorporating a pyridinium additive, we achieved efficient C─N coupling between phenols and diverse aromatic nitrogen nucleophiles, delivering high yields (up to 99%) across a wide range of substrates, including pharmaceuticals and natural products. We investigate reaction selectivity and substrate compatibility/limitations through a combination of experimental and computational techniques. Moreover, we highlight the synthetic versatility of the amination products through various late-stage functionalizations including the grafting of two different heteroarenes onto one phenol scaffold.

Graphical Abstract

INTRODUCTION

The construction of C─N bonds has been pivotal in the synthetic development of pharmaceuticals, agrochemicals, and materials chemistry.^1,2 The assembly of new C(sp²)─N bonds with amines and azoles typically requires transition-metal catalysts, high temperatures, and the prefunctionalization of the carbon fragment, often as boronic acids, aryl halides, and other electrophilic groups.^3,4

Photoredox,^5,6 electrochemistry, and hypervalent iodine methods have allowed for the diverse functionalization of electron-rich arenes (aryl ethers, etc.), often through radical cation and radical–radical mechanisms (Figure 1A).^7-18 These methods have allowed for the functionalization of unactivated C─H, C─O, and C─X bonds to install amines and azoles. While demonstrating an extensive arene scope, these works are not compatible for the activation of phenolic substrates.

Figure 1.

(A) Arene radical cation generation. (B) Previous synthetic approaches for amination of phenols. (C) Previous work on pyridination of phenols. (D) This work.

The utilization of free phenols in organic synthesis has allowed for diverse C─H functionalization with excellent control of regio- and chemoselectivity. These methods leverage the innate reactivity of the free phenols, specifically their facile oxidation.^19-21 With electron-rich phenols, several research groups have reported the formation of C─N bonds through coupling of diarylamines and azoles (Figure 1B).^22-37 These reactions are typically proposed to proceed through a radical–radical coupling mechanism and employ strong oxidants that act on both the phenol and amine including photochemical examples.^22-24 The majority of these methods afford ortho-C─H-functionalized phenolic products, although a few para-functionalized products are present. Of note, the C─N coupling of indoles has not been achieved with any of the prior methods and photochemical methods have been restricted to diarylamines.

The recent emergence of electron donor–acceptor (EDA) photocatalysis has also enabled chemistries involving single-electron transfer (SET) events with low-energy visible light and without the need for photoredox catalysts.^38,39 This avenue of radical generation has largely been biased toward activation of an electron-poor component (acceptor) where an electron-rich component (donor) that acts as a catalyst/reagent is optimized. Recently, we disclosed the photochemical C─H and S_NAr pyridination of electron-rich phenols to furnish a variety of phenol–pyridinium salts (Figure 1C).⁴⁰ The reaction proceeded via EDA complexation between the phenol and the protonated pyridine (pyridinium) to form reactive phenoxyl radical cations, which are extremely electrophilic, upon light irradiation. In this instance, the pyridine reacts from within the EDA complex to form the C─N bond. The addition of an Ir^(III) photocatalyst enhanced the yields, aiding in the phenol oxidation.

From this chemistry, we postulated that it would be possible to design a pyridinium acceptor that would not act as a nucleophile, allowing an external nucleophile to react (Figure 1D).⁴¹ Such a pyridine could be employed as a substoichiometric catalyst if it satisfies several criteria: (1) is sufficiently basic to protonate, (2) is less nucleophilic than the exogenous nucleophile, (3) is sufficiently electron-poor to participate in EDA complexation, and (4) is not subject to direct reduction by the photocatalyst.

RESULTS AND DISCUSSION

Reaction Design.

Using our prior pyridination conditions,⁴⁰ the initial reaction discovery focused on the coupling of 2,4-dimethoxyphenol (1a) and pyrazole (2a) using Ir(ppy)₃ in HFIP (Scheme 1). Our efforts started here as the Nicewicz reaction conditions with this phenol using acridinium photocatalyst dit-BuMesAcrBF₄ yielded no product.⁸ We then screened several pyridine (with acid) and alkylpyridinium additives of varying steric and electronic nature (Scheme 1A). C─H amination product 3a was obtained in the highest LC assay yield (AY) using 2-bromo-5-(trifluoromethyl)pyridine (pyr1) and TfOH (19%). Pyr1 was strategically selected to fulfill the criteria discussed above. Namely, pyridines bearing C2 substituents were unreactive as nucleophiles in our pyridination work.⁴⁰ The CF₃ group ensures that pyr1 is a good acceptor for the phenols while not being too electron-poor as to be deactivated by direct reduction with the photocatalyst or to preclude protonation. Phenol homocoupling product 3aa was observed as a major reaction byproduct, as is commonly seen in photochemical phenol oxidation.^42,43

Scheme 1. Reaction Discovery and Optimization.

^aScreen of pyridinium additives. AY is relative LC assay yields of 3a and 3aa with 4,4′-di-tert-butylbiphenyl internal standard (254 nm). *No TfOH was used with pyr4. Yield shown in parentheses is the isolated yield after column chromatography. ^bScreen of acid additives. ^cSensitivity analysis of reaction based on percent difference from optimized condition for each respective change.

A range of solvents, acids, oxidants, photocatalysts, light sources, stoichiometries, and concentrations were analyzed for the desired reaction (see the Supporting Information for full optimization). The acid screen shown in Scheme 1B illustrates the importance of pyridine protonation for reaction success with TfOH outperforming the other strong acids, for which counterion stability is likely important. An excess of pyrazole (2a) and catalytic pyridinium (30 mol %) in dilute conditions (0.08 M) achieved the highest yield (52% AY and 51% isolated yield) and greatest selectivity for amination: homocoupling (15.9:1) (Scheme 1, entry 5). Removing the photocatalyst, Ir(ppy)₃, from the reaction mixture lowered the yield (35%), but not drastically, indicating that the reaction can be promoted by the pyridinium itself (entry 7). In contrast, the use of only Ir(ppy)₃, without any pyridinium additive, resulted in no desired product (entry 10). As this substrate is well within the oxidation window of Ir(ppy)₃ (E_ox = +0.31 V vs SCE),⁵ it appears that back electron transfer (BET) prevents the forward reaction.^44,45 The combination of TfOH and Ir(ppy)₃ led to some product formation (33%), suggesting that acid aids in phenoxyl stabilization or slows/prevents BET of 1a (entry 12). UV–vis analysis between 1a and 2a in the presence and absence of TfOH indicated no EDA complexation. Together, these results support the phenol–pyridinium EDA complex as a key intermediate that can both undergo productive direct excitation with light and stabilize the phenoxyl radical cations. Overall, the reaction sensitivity plot summarizes the significant dependence on the presence of pyridinium and an O₂ atmosphere to achieve optimal reactivity (Scheme 1C).

C─H Amination Phenol Scope.

With the optimized C─H amination conditions, we explored the scope using a variety of electron-rich phenols (Scheme 2). ortho-C─H amination was facilitated in moderate yields (up to 63% yield) with good functional group tolerance (aldehydes, alkynes, esters, and fluoride) for 11 substrates (3a–j). Electron-rich (alkoxy and amine) substituents were needed at the para-position for sufficient yields. The amination of 2-methoxyhydroquinone (1h) resulted in a mixture of 3h (31%) and 3ha (53%) with a 84% yield overall. We hypothesize that a 1,5-hydrogen atom transfer (hydrogen atom transfer) with the phenoxyl radical and adjacent methoxy group resulted in the ketal cyclization product 3ha.⁴⁶ Amination of a phenolic analogue of the selective serotonin reuptake inhibitor fluoxetine (1j) proceeded in a 43% yield after only 4 h. Electron-neutral phenols were found to be largely unreactive to the C─H amination, potentially due to a lack of sufficiently electrophilic sites in the radical cation¹¹ (see Supporting Information for substrate limitations).

Scheme 2. Phenol Scope of C─H Amination^c.

^aObtained as a mixture of separable isomers. ^bFour h reaction time. ^cYields shown are isolated yields after column chromatography. Obtained as a mixture of separable isomers.

S_NAr Amination Phenol Scope.

Recently, Knowles and co-workers demonstrated the electron-withdrawing capabilities of the phenoxyl radical through Fe- and base-mediated S_NAr of halophenols with benzoates/carboxylates.⁴⁷ Building upon our successful photocatalytic S_NAr pyridination involving 4-(trifluoromethyl)pyridine and 2,6-dimethyl-4-fluorophenol (1l),⁴⁰ the amination of a range of halophenols (1k–u) was explored via activation of C─X bonds (Scheme 3). Using the optimized C─H amination conditions, ortho- and para-S_NAr amination with pyrazole 2a occurred in excellent yields (up to 99%). For the 2,6-dimethylphenol backbone, amination occurred for 4-chloro (1k), 4-fluoro (1l), 4-bromo (1m), and 4-iodo (1n) derivatives (in this order of yield). Without the pyridinium additive (pyr1 and TfOH), a reduced isolated yield of 52% was observed for fluorophenol 1l. The pyridinium yield enhancement for S_NAr was not as profound as it was for C─H amination (0% without vs 51% with pyridinium) as BET is likely not as detrimental for these halogenated phenols.⁴⁴ The S_NAr of 1l occurred in a comparable yield with the removal of Ir(ppy)₃ (62% isolated yield), highlighting the novel organocatalytic capabilities of pyr1 and acid. This suggests that the photocatalyst may not be necessary for adequate S_NAr reactivity.

Scheme 3. Phenol Scope of S_NAr Amination^d.

^a2.0 equiv 2a. ^b4CzIPN instead of Ir(ppy)₃. ^cMesAcrBF₄ instead of Ir(ppy)₃ and TFE instead of HFIP. ^dYields shown are isolated yields after column chromatography.

Regioselective C4 amination product 3q was obtained in a 31% yield and the disubstitution product 3qa was isolated in a 26% yield when using organophotocatalyst 4CzIPN (E_ox = +1.43 V vs SCE).⁴⁸ Despite the poor performance of 2-bromophenol 1p, ortho-S_NAr could be achieved with other phenols such as 2-chlorophenol 1r in a 67% yield. A bromo catecholamine analogue 1s and plant alkaloid 1t were selectively aminated in 56% and 18% yields, respectively.

The S_NAr proved to be scalable; for the amination of 1u, a 99% yield was obtained on a 0.25 mmol scale and an 89% yield on a 2.5 mmol scale (using only one Kessil lamp). The amination product 3u was then subjected to sequential C─H amination to furnish tetra-substituted 3v in 33% yield using MesAcrBF₄ (E_ox = +2.06 V vs SCE)⁵ in TFE. High-throughput experimentation was utilized to optimize the reaction conditions for this particular reaction as poor conversion was seen with Ir(ppy)₃ (see the Supporting Information).

Nucleophile Scope.

While pyrazole (2a) served as the primary nucleophile during the phenol exploration, the diverse chemical landscape and complexity inherent to the nucleophile was also surveyed. We achieved mild C─N functionalization using various aromatic nitrogen nucleophiles for both C─H and S_NAr amination, resulting in satisfactory yields of up to 70% (Scheme 4). As previously observed, the S_NAr was typically higher yielding than the C─H amination for the same nucleophile (3a/k and 3w/x). Notably, indoles have not previously been used in an oxidative C─N coupling reaction with free phenols.

Scheme 4. Scope of Nucleophiles in Amination^d.

^aObtained as a mixture of inseparable regioisomers. ^b2.0 equiv 2ad. ^c1.1 equiv 2ag. ^dSame conditions as Scheme 3. Yields shown are isolated yields after column chromatography.

The mild nature of the reaction conditions enabled the use of natural product and pharmaceutical substrates. Of note, the amination of N,N-dimethyltryptamine (2ac) and melatonin (2ad) proceeded in 47% and 25% yields, respectively. A novel UV-327 analogue (3z) was obtained as a mixture of N1/N2 regioisomers (1:1.5) in a 43% overall yield. Remarkably, the antimalarial drug amodiaquine (2ag), which contains a 2,4-substituted phenol, underwent C─N coupling in a 23% yield with only 1.1 equiv of the nucleophile.

This transformation was unsuccessful with aliphatic amines or amides (for successful use of the nucleophiles in nonphenolic arene couplings, see refs 7-18). Other heteroatom-based nucleophiles (thiols, alcohols, carboxylates, etc.) also did not lead to any desired products (see Supporting Information).

Reaction Mechanism.

To probe the reaction selectivity for C─H and S_NAr amination, intermolecular competition experiments were performed using phenols 1l and 1a with nitrogen nucleophiles 2a (pyrazole) and 2ac (indole) (Scheme 5A). With a pyrazole (2a) nucleophile, the S_NAr predominates with a 57% AY of 3w and a <1% AY of the C─H amination product 3a. Conversely, an indole nucleophile (2ac) gives rise to no S_NAr amination (3aca) with only C─H amination product 3ac being observed in 29% AY.

Scheme 5. Experimental andComputational Mechanism Investigation.

^aAY is relative LC assay yields of with 4,4′-di-tert-butylbiphenyl internal standard (254 nm). Free energies were computed using UM06-2X/6-311++G(d,p)-CPCM(HFIP)//UB3LYP/6-31+G(d,p)-CPCM(HFIP). ^bChange in site electron density (q) for highlighted positions from neutral to radical cation species. Electron density values were obtained with NBO formalism using 6-31+G(d,p)-CPCM(HFIP). ^cOxidation potentials measured with degassed HFIP, vs Ag/AgCl. (See Supporting Information for more details.)

To understand this selectivity, DFT studies were conducted using UM06-2X/6-311++G(d,p)-CPCM(HFIP)//UB3LYP/6-31+G(d,p)-CPCM(HFIP). A comprehensive experimental and computational study by Nicewicz and co-workers on a similar arene amination reaction determined that nucleophilic addition to the arene radical cation was the rate-determining step.¹⁵ With this in mind, we calculated the transition-state energies for addition of each nucleophile to the respective phenol radical cations. The barriers matched the experimental observations, with the observed major products having lower transition-state energies (>6 kcal/mol). For azoles, the neutral adjacent nitrogen can serve as the nucleophilic center, but for indoles, the π system must act as the nucleophile. Addition with nucleophiles of the latter type favors C─H bond activation, whereas the former favors S_NAr reaction.

Also inspired from work by Nicewicz et al.,¹¹ we calculated the natural population analysis (NPA) values or site electron densities for the reactive sites (ortho) of neutral phenols vs their radical cations (Scheme 5B). One would anticipate a shift toward greater positivity in these values as SET of a free phenol results in the formation of the electrophilic radical cation. We observed a good correlation (R² = 0.83) between the amination yield and the increase in the NPA value (Δq) from the neutral phenol to radical cation. When Δq exceeds 0.023, reactivity is seen. To forecast the reactivity of unidentified phenols, one could compute this value to determine if it is above this critical threshold.

When exploring the reaction scope, we observed phenoxazine dimerization (3ak) in a 61% AY when attempting the amination reaction with phenoxazine 2ak and 1a (Scheme 5C).⁴⁹ Utilizing cyclic voltammetry in HFIP, we determined the oxidation potentials (E_1/2) for the two substrates and noted a significant disparity, with 2ak exhibiting a notably lower potential (0.067 V) compared to 1a (0.39 V). This transformation therefore likely occurs through aminyl radical formation followed by subsequent radical–radical coupling. We observed no C─H amination product with phenoxazine 2ak, indicating that a radical–radical coupling mechanism is improbable for our reaction. This sets the current work apart from previous phenol aminations and suggests the probable involvement of the phenoxyl radical cation pathway (Scheme 6).

Scheme 6.

Proposed Mechanism and UV–Vis Analysis of EDA Complexation

In the proposed mechanism depicted in Scheme 6A, the catalytic formation of phenol–pyridinium EDA complex D leads to SET to form complex E•F. The stabilization of the phenoxyl radical cation F by the pyridinyl radical E prevents/slows down BET of F. Simultaneously, a parallel photocatalytic cycle generates an additional supply of F, thus providing two sources of phenoxyl radical cations. Addition of nucleophile G to E•F or F leads to H. Subsequent proton loss and rearomatization culminates in the formation of the amination product J. In Scheme 6B, UV–vis analysis demonstrated that an EDA complex between phenol 1a and pyr1 in the presence of TfOH can be directly excited with the wavelength used in our conditions (427 nm).

Utilization of Amination Products.

To highlight the synthetic utility of the obtained amination products and the potential of the phenol functional group for further reactivity, we explored late-stage modifications of phenol 3u (Scheme 7). From a 2.5 mmol scale reaction, we obtained ~500 mg of 3u to use for additional transformations. In Scheme 7A, we utilized recent conditions from Fier and co-workers at Merck using a transition-metal-free approach to turn phenols into anilines with sequential S_NAr.⁵⁰ Fluoro aniline 4a was formed in a 19% yield (70% brsm) using trifluoroethylamine as a nucleophile. We leveraged the nucleophilic potential of the phenol upon deprotonation (phenolate) in a conventional S_NAr reaction with a (hetero)aryl chloride to form aryl ether 4b in a 98% yield (Scheme 7B).

Scheme 7. Late-Stage Diversification of Phenol 3u^a.

^aReaction conditions: (A) i) K₃PO₄ (7.0 equiv), 2,2,2-trifluoroethylamine hydrochloride (1.3 equiv), 5,6-dichloropyrazine-2,3-dicarbonitrile(1.2 equiv), 1,4-dioxane, 50 °C, 2 h ii) 3u (0.25 mmol), DMSO, 100 °C, 16 h iii) Zn (10 equiv), AcOH, 80 °C, 30 min (B) 3u (0.25mmol), K₂CO₃ (3.0 equiv), 2-chloropyrazine (1.0 equiv), DMF, 110 °C, 16 h (C) 3u (0.26 mmol), PhB(OH)₂ (1.5 equiv), Pd(PPh₃)₄ (10 mol %),K₂CO₃ (3.0 equiv), AISF (1.2 equiv), 5:1 1,4-dioxane:H₂O, 100 °C, 1 h.

Finally, researchers at Pfizer have pioneered the use of [4-(acetylamino)phenyl]imidodisulfuryl difluoride (AISF) for in situ activation of phenols in a variety of cross-coupling reactions.⁵¹ We successfully facilitated a net C─C bond formation employing this reagent under typical Suzuki–Miyaura conditions, yielding biaryl 4c in a one-pot manner with a satisfactory 58% yield (Scheme 7C). Overall, the phenol functional group can be seen as a directing/activating group to install three groups sequentially to rapidly increase molecular complexity. Since this transformation is applicable to both ortho- and para-amination, a large diversity of regiochemical space can be scanned.

CONCLUDING REMARKS

In summary, we have developed a mild method utilizing pyridinium EDA complexation and photocatalytic cycles for the formation of C─N bonds between phenols/halophenols and a diverse range of nitrogen nucleophiles (azoles, indoles, and diarylamines). Selective C─H amination ensues for electron-rich phenols, while S_NAr is seen with halogenated phenols. Selective sequenced reactions are also possible as amines that react via an sp² lone pair undergo S_NAr reactions, while amines that react via a π system undergo oxidative C─H amination. Calculations and mechanistic experiments support a reaction pathway featuring oxidation to a phenoxyl radical cation followed by nucleophilic addition. The pyridinium catalyst aids phenol oxidation by formation of an EDA complex and also stabilizes the radical cation preventing BET.

The versatility of the resulting products, as well as phenols in general, is showcased through their applications in late-stage functionalizations. With this work, up to three hetero(arene) bonds can be forged from phenol starting materials to rapidly explore chemical space. Future aspirations involve the development of an efficient one-pot system for halogenation–amination, targeting difficult-to-activate C(sp²)─H bonds.

Data Availability Statement

The data underlying this study are available in the published article and its Supporting Information.