Room-Temperature Decarboxylative Amination of Electron-Deficient (Hetero)Aromatic Carboxylic Acids

Abstract

Decarboxylative amination reactions offer the opportunity to access (hetero)aromatic amines from readily available carboxylic acid starting materials. However, the reliance of the Curtius rearrangement on organoazides introduces safety concerns, while more recent transition-metal-mediated approaches suffer from high reaction temperatures and limited substrate scopes. Here, we demonstrate a room-temperature decarboxylative amination to generate a wide array of primary (hetero)aromatic amines mediated by N-fluorobenzesulfonimide (NFSI). The broad functional group tolerance of this reaction enables efficient amination of biologically relevant small molecules. Mechanistic studies suggest that NFSI provides access to an activated sulfonimide intermediate, a functional equivalent to acyl azide intermediates, while avoiding the hazards associated with organoazides.

Aromatic amines are prevalent substructures in a variety of pharmaceuticals,¹ agrochemicals² and functional materials.³ Primary anilines in particular, are valuable not only as key features in biologically active structures, but also because they are important intermediates in heterocycle syntheses.⁴ Traditional aniline syntheses, relying on an aromatic nitration-reduction sequence, typically generate large amounts of acid waste and yield mixtures of isomeric nitroarene products (Scheme 1-Ia).⁵ Similarly, S_NAr reactions that generate aryl amines from aryl halides⁶ or arenes⁷ require high reaction temperatures and strongly electron-withdrawing directing groups to achieve reactivity and regioselectivity (Scheme 1-Ib). More recently, the direct amination of aryl boronic acids has been achieved with hydroxylamine-derived reagents in the presence of strong base (Scheme 1-Ic, Scheme S1).⁸ Alternatively, Curtius rearrangement approaches generate the primary anilines regiospecifically from carboxylic acids following decomposition of an acyl azide intermediate (Scheme 1-Id).^9,10 The implementation of these decarboxylation reactions in pharmaceutical syntheses¹¹ highlights the importance and utility of these transformations, while the requirement for specialized equipment and flow-based methods¹² underscores the safety concerns associated with organic azides¹³ and the need for less-hazardous alternatives.¹⁴

Scheme 1.

Representative Conditions and Drawbacks in Common Routes to Primary Anilines.

An alternative approach to access primary anilines relies on the transition-metal-catalyzed C-H and C-X amination reactions of arenes (Scheme I-IIa)¹⁵ and aryl halides (Scheme 1-IIb) using ammonia.^16,17 Unfortunately, challenges associated with the use of ammonia have hindered the widespread implementation of such methods.¹⁷ Although Pd-, Cu-, and Ni-catalyst systems^{18, 19, 20} have been developed to overcome these challenges, some drawbacks still remain. For example, the direct amination of C-H bonds with ammonia²¹ typically rely on stoichiometric loadings of metal mediators and directing groups to achieve regioselective amination (Scheme 1-IIa). The complementary aminations of C-X bonds often require the use of super stoichiometric loadings of strong base to facilitate the deprotonation of ammonia as well as sterically demanding phosphine ligands to suppress over amination (Scheme 1-IIb).²² The corresponding Cu-catalyzed amination of C-B bonds has been demonstrated, although with a modest substrate scope (Scheme 1-IIc).²³ While these methods are promising advances toward the use of ammonia as an aminating agent, the requirements for stoichiometric metal loadings or specialized ligands have limited their synthetic utility.

Transition-metal catalyzed decarboxylative amination is an attractive alternative for the generation of amines from carboxylic acids (Scheme 1-IId).²⁴ Unfortunately, these reactions have so far been ineffective for the generation of primary anilines and typically require the more acidic amide and aniline coupling partners. Furthermore, the decarboxylative amination reactions of (hetero)aromatic carboxylic acids require stoichiometric oxidants under high-temperature conditions and are also limited to activated benzoic acids, such as those bearing ortho-nitro or -sulfonyl groups.²⁵

Here, we demonstrate that N-fluorobenzenesulfonimide (NFSI)^26,27 enables the efficient decarboxylative amination of a wide scope of (hetero)aromatic carboxylic acids to the corresponding primary amines at room-temperature under metal-free conditions (Scheme 2a). The use of the commercially available electrophilic aminating agent provides access to a key N-fluorosulfonimide intermediate that is functionally analogous to the classic acyl azide intermediate (Scheme 2b), while bypassing the hazards associated with organoazides.

Scheme 2.

(a) NFSI-Mediated Decarboxylative Amination and (b) the Key Reactive Intermediate

We began our study by evaluating the amination of 2-nitrobenzoic acid (1a) with NFSI in the presence of base at room temperature. Elevated temperatures did not afford improved yields, in contrast to the typical decarboxylative amination reactions described above (Table S1). Broad screening of the reaction conditions by high-throughput methods revealed the interplay of base and solvent (Tables S6–S7). Selected bases and solvents were employed in larger-scale reactions for further optimization and demonstration of reproducibility (Figure 1). The reaction is highly dependent on the solvent with efficient amination occurring primarily in amide-based solvents such as N,N-dimethylformamide (DMF, 99%), N,N-dimethylacetamide (DMA, 78%), and N-methylpyrrolidinone (NMP, 71%). Acetonitrile was also effective, providing 2-nitroaniline in 65% yield, however other polar aprotic solvents, such as dimethylsulfoxide (DMSO), were ineffective (<5% yield) (Figure 1, Table S5). The base is crucial for the reaction to proceed; no 2-nitroaniline was observed in the absence of base. In general, carbonate bases were more effective than other inorganic bases explored (Table S6, Figure S1), with cesium carbonate and potassium carbonate providing 99% and 96% yields, respectively (Table S4). The yields of 2-nitroaniline decreased when the base loading was reduced from 2 equivalents to 1 equivalent (64% and 53% for Cs₂CO₃ and K₂CO₃, respectively, Table S4). A variety of organic bases were also explored with most showing low yields of 2-nitroaniline (Table S6). DMAP, however, provided 70% yield under otherwise standard conditions (Table S4). Thus, the final optimized reaction conditions employ a combination of NFSI (2.5 equiv) and Cs₂CO₃ (2 equiv) in DMF at room temperature to generate 2-nitroaniline in quantitative yield.

Figure 1.

Optimization of the solvent and base in the decarboxylative amination reaction. Standard reaction conditions: 1a (0.1 mmol), NFSI (0.25 mmol) and base (0.2 mmol Cs₂CO₃) in solvent (1.6 mL DMF) under N₂ for 3 h, unless otherwise noted. Yields obtained from 1H NMR spectroscopy with methyl 3,5-dinitrobenzoate as internal standard. Optimized conditions in DMF with 2 equiv Cs₂CO₃ indicated with solid-filled bars.

We next explored the scope and limitations of the reaction (Figure 2). A large array of benzoic acids was surveyed on a small scale and the products selected for isolation were chosen to highlight the diversity in scope as well as to acknowledge limitations in the method (Scheme S2).²⁸ Due to the volatility of some products, we have included ¹H NMR spectroscopy yields to demonstrate the efficiency of the reaction, independent of isolation. Substitution in the ortho-positions was well-tolerated and high yields were obtained for anilines containing both electron-deficient (2a-2f,) and electron-rich substituents (2g, 2h). The efficient access to ortho-substituted anilines makes this method complementary to Buchwald-Hartwig aminations which often require catalyst alterations or increased loadings for the successful amination of ortho-substituted aryl halides.²⁹ Simple meta- and para-substituted benzoic acids (2n, 2o) also provided the corresponding anilines, although in lower yields suggesting an important role for ortho-substitution. In general, electron-withdrawing groups are necessary for obtaining high yields, consistent with the observed Hammett correlation (r = 0.6, Figure S3). A large variety of functional groups could be tolerated using this methodology. In particular, anilines bearing halogens (2e, 2i, 2l, 2m), nitriles (2c), ethers (2f, 2l) and esters (2u) could all be accessed in good yields. The successful generation of 2c and 2u is particularly important given that nitrile and ester groups are often problematic under traditional Buchwald-Hartwig amination conditions using strong bases (NaOtBu, KOH, etc). Similarly, ethers, halides and nitro groups can pose challenges in some Curtius rearrangement reactions,³⁰ highlighting the complementarity of this new protocol.

Figure 2.

Scope and functional group tolerance of the decarboxylative amination of (hetero)aromatic acids. Reaction conditions: 1 (0.30 mmol), NFSI (0.75 mmol) and Cs₂CO₃ (0.60 mmol) in DMF (5 mL) under N₂. ¹H NMR yields given with isolated yields in parentheses. ^bIsolated as the Boc-protected aniline. ^cNMR yield could not be obtained.

Various di- and trisubstituted benzoic acids also underwent successful decarboxylation to form the desired products in good to high yields (2i-2m, 2p-2r). The higher yields of 2p (58%) and 2q (65%) relative to 2n (18%) and 2o (38%) reveal the steric role of the ortho-substituent. However, decarboxylation of 2,6-dinitrobenzoic acid underwent decarboxylative amination to form the product only in low yield (2k, 14%). Similarly, 2,6-dimethylbenzoic acid proved to be unreactive (Scheme S2), suggesting that the challenge in these cases may arise from a steric limitation.

Given the importance of heteroaryl amines in biologically active structures^1,2 and the challenges associated with amination of heteroaromatic halides,²⁹ we explored the corresponding amination of heteroaromatic carboxylic acids. The amination reaction was successfully extended to substituted pyridines (2s-2z), quinolines (2aa, 2ab), pyrazines (2ac) and triazoles (2ad). As observed with the benzoic acids, substrates bearing electron-withdrawing substituents typically afforded higher product yields than the unsubstituted heterocycles (Scheme S2). The preservation of the bromine in 2t and 2z further highlights the orthogonality of this method with respect to traditional cross-coupling protocols for amination. Overall, the successful decarboxylative amination of a wide variety of substituted (hetero)aromatic carboxylic acids highlights the utility of this system as it overcomes the common ortho-substituent limitation of transition-metal-catalyzed decarboxylative coupling reactions.^{24, 25, 31}

The functional group tolerance was further evaluated with the inclusion of small molecule additives bearing a variety of functional groups.³² In general, the reaction is tolerant of most functional groups surveyed with high yields of 2-nitroaniline and high recoveries of the additive observed for alkyl and aryl halides, ketones, esters and amides, as well as alkynes, alkenes, nitriles and epoxides. An array of heteroarenes were also compatible with the reaction conditions. Imidazole and indole presented as the only incompatible heteroarenes of those explored. Similarly, other nucleophilic additives such as amines, thiols and alcohols were incompatible, likely due to side reactions with the NFSI or the proposed isocyanate intermediate (vide infra). The less nucleophilic phenols, however, were tolerated.

The efficient amination of (hetero)aromatic carboxylic acids and the broad functional group tolerance were highlighted with the decarboxylative amination of three biologically-relevant carboxylic acids. A flavoxate derivative, probenecid, and acifluorofen underwent successful amination yielding the corresponding amines 2ae (57%), 2af (32%), and 2ag (71%) in good yields (Figure 3). Finally, the isolation of both 2b and 2x on a 1-gram scale demonstrated the reaction to be amenable to larger scales (Figure 2).

Figure 3.

Primary amines generated from the decarboxylative amination of biologically active carboxylic acids. 1H NMR yields given with isolated yields in parentheses.

To gain insight into the reaction mechanism, we conducted a series of control reactions. When nitrobenzene was included in the reaction mixture in place of 2-nitrobenzoic acid, no amination product was observed indicating that the decarboxylative amination reaction does not proceed through a protodecarboxylation step followed by NFSI-mediated amination of the resulting nitrobenzene (Scheme 3a).³³ Because NFSI is known to form nitrogen-centered radicals,³⁴ the possible formation of radical intermediates was probed. When the standard amination reaction was conducted in the presence of either 1,1-diphenylethylene or 9,10-dihydroanthracene as radical trapping agents, 2a was formed in high yields (>95%) and no trapped radical species were observed (Scheme 3b) suggesting against the intermediacy of trappable radicals in these reactions.

Scheme 3.

Control Reactions.

Instead, we favor a pathway involving an isocyanate intermediate,¹¹ which upon hydrolysis forms the primary amine (Scheme 4a). The reaction begins with nucleophilic attack by the carboxylate on the sulfonyl to generate the sulfonate ester A. The resulting N-fluorosulfonamide anion attacks the carbonyl of A to provide the N-fluoroacylsulfonimide B, which undergoes a Curtius-like rearrangement to give the isocyanate intermediate C. The aniline is then generated in situ via hydrolysis. This pathway is consistent with the observed formation of sulfonic acid and benzene sulfonyl fluoride byproducts, as well as the control experiments described above. The importance of ortho-substitution in this decarboxylative amination reaction is also reminiscent of related observations in the Curtius rearrangement of isocyanates.³⁵

Scheme 4.

Proposed Reaction Pathway and Experimental Support of the Pathway.

The reaction of 2-nitrophenyl isocyanate (3a) with NFSI under the standard reaction conditions led to formation of the corresponding aniline 2a in moderate yield (28%, Scheme 4b). Higher yields of 2a were obtained in the presence of sulfonic acid, the observed reaction byproduct (65%, Scheme 4b). Although these results do not require isocyanate to be an intermediate, it confirms that isocyanate intermediates are viable under these conditions. Additionally, conducting the standard decarboxylative amination in the presence of methanol provided the corresponding carbamate (43%) in addition to the aniline product (29%) providing support for an isocyanate intermediate (Scheme 4c). To further interrogate this pathway, we synthesized proposed intermediates B (R = F (4a), NO₂ (4b)) and probed their reactivity.³⁶ Both compounds led to formation of the corresponding isocyanate during the low temperature synthesis of B (3a, 40%; 3b, 29% Scheme 4d, SI). Additionally, the presence of 4a in the amination of 2-fluorobenzoic acid was confirmed by ¹⁹F NMR spectroscopy, supporting a pathway involving generation and rearrangement of a N-fluoroacylsulfonimide intermediate.

In summary, we have developed a new decarboxylative amination protocol which provides access to primary amine products under room temperature conditions from the corresponding carboxylic acid starting materials. The mild reaction conditions arise from a new NFSI-mediated decarboxylation step, which provides convenient and safe access to the well-known Curtius-like pathways. This approach also avoids the challenges associated with accessing primary amines from ammonia. This strategy was applied to a large series of (hetero)aromatic carboxylic acids demonstrating its potential as a valuable alternative to traditional amination reactions. Further exploration of the scope and mechanism of this reaction is underway.

Supplementary Material

The Supporting Information is available free of charge on the ACS Publications website.

Experimental details, characterization data, ¹H, ¹³C NMR and ¹⁹F NMR spectra (PDF)

Data Availability Statement

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