Cation Radical-Accelerated Nucleophilic Aromatic Substitution for Amination of Alkoxyarenes

Abstract

Nucleophilic aromatic substitution (S_NAr) is a common method for arene functionalization; however, reactions of this type are typically limited to electron-deficient aromatic halides. Herein, we describe a mild, metal-free, cation-radical accelerated nucleophilic aromatic substitution (CRA-S_NAr) using a potent, highly oxidizing acridinium photoredox catalyst. Selective substitution of arene C-O bonds on a wide array of aryl ether substrates was shown with a variety of primary amine nucleophiles. Mechanistic evidence is also presented that supports the proposed CRA-S_NAr pathway.

Graphical Abstract

Anilines represent an important functional group in pharmaceutical, agrochemical, and materials industries. These functionalities are present in a wide variety of bioactive compounds.¹ Additionally, the prevalence of anilines in materials industries is ubiquitous, representing a common structural motif found in organic dyes and conductive polyaniline polymers.¹ Due to their pervasiveness in industry and their synthetic utility, the development of new aryl C-N bond forming reactions is an important area of research.

Historically, benzenoid amination has been accomplished via aromatic nitration to provide primary anilines.^2,3 However, these reactions require harsh conditions restricting functional group tolerance and overall utility. Nucleophilic aromatic substitution (S_NAr), however, can be conducted under more mild conditions. Classical S_NAr reactions are widely used for installing functional groups onto electron-poor halogenated benzenoids and heteroaromatic systems and represent some of the most used transformations in medicinal chemistry.^3,4 Proceeding through either a Meisenheimer complex, or via a concerted S_N2-type mechanism, S_NAr reactions require strongly electron-withdrawing groups to be placed in the ortho or para positions to a particular nucleofuge, typically fluorine, to overcome the high energetic demand.^3,5 The high energetic demand of S_NAr reactions typically limit this transformation to electron-poor arenes, leaving S_NAr of electron-rich arenes as an area to still be explored.

Transition metal cross-coupling reactions such as Buch-wald-Hartwig and Ullmann couplings provide an alternative avenue toward rapid access to aniline products.^6,7 The requirement of halogenated or pseudohalogenated cross-coupling partners can prove to be problematic in late stage functionalization, however, as these functional groups can be difficult to carry through a synthesis due to their synthetic lability over various other synthetic methods.⁸

In order to address substrate limitations, several reactions have been developed that utilzie aryl ether electrophiles as cross-coupling partners (Scheme 1A). These methods all benefit from the commercial availability of a large number of phenol and aryl ether starting materials stemming from their prevalence in biomass, such as lignin.⁹ Aryl ethers are also less synthetically labile than halides and can more efficiently be carried through a synthesis without deleterious side reactivity.⁸

Scheme 1.

Aryl Ether Amination Strategies

One such example of a transition metal mediated cross-coupling reaction using aryl methyl ethers was developed by Chatani and Tobisu.¹⁰ This method utilizes a catalytic quantity of NHC ligated nickel complexes in the presence of stoichiometric NaOtBu to affect formal aryl methyl ether amination. However, this method is limited to heteroaryl and π-extended aryl ethers and is only effective with secondary amine cross-coupling partners.

Biswas and co-workers have developed a metal-free approach to aryl methyl ether functionalization by means of triflic acid promoted oxocarbenium ion-catalyzed S_NAr (Scheme 1b).¹¹ Conversely, base-promoted S_NAr reactions using aryl methyl ethers have been developed as well, using strong bases such as stoichiometric n-butyl lithium, NaH-LiI composites, and phosphazene superbase to generate reactive amine anions (Scheme 1c).^12–15 While noteworthy, the substrate compatibility can be limited, demonstrating a need for more mild reactions to access S_NAr pathways.

Our group has developed a research program that utilizes the reactivity of aryl cation radicals generated by highly potent visible-light activated acridinium photoredox catalysts. Arene cation radicals of aryl ethers have shown to be associated with a significant increase in electrophilicity of the arene in the para, ortho, and ipso positions, allowing for C-H functionalization using azole, cyanide, primary amine, and fluoride nucleophiles under oxidative conditions.^16–19 By modifying the reaction conditions to a redox neutral system, our group has been able to show substitution at the ipso position using azoles, ammonia, cyanide, and fluorine nucleophiles.^20,21 Herein, we describe the expansion of this methodology to primary amine nucleophiles (Scheme 1d).

Beginning with conditions previously reported by our group for CRA-S_NAr, optimization commenced with 4-(tert)butyl-2-chloroanisole and benzylamine as the amine coupling partner using a catalytic amount of acridinium salt di-tBu-Mes-Acr (E*_red = +2.15 V vs SCE). For significant conversion to 1 to occur, 2 equiv of the primary amine were required, while 1,2-dichloroethane proved to be the optimal solvent. Solvent and concentration optimization studies indicated that a 0.1 M DCE solvent system afforded the desired product in 52% isolated yield (Table 1). Screening other acridinium photoredox catalysts reveled that the (tert)butyl groups were necessary for optimal aniline formation to occur. Additionally, the reaction was run with an acridinium catalyst with a less oxidizing excited state (hexa-OMe-Mes-Acr E*_red = +1.65 V vs SCE) and no reaction was observed (see Supporting Information for details). The mass balance for this reaction is excellent, with the reaction mixture consisting of primarily product and the remaining aryl ether starting material. Oxidative dimerization of benzylamine to N-benzyl phenyl-methanimine was also observed as byproduct in 5% yield.

Table 1.

Primary Amine CRA-S_NAr Optimization^a

entry	solvent	ratio A/B	yield 1 (%)
1	1:1 DCE/TFE	1:1	23
2	2:1 DCE/TFE	1:1	28
3	1:2 DCE/TFE	1:1	17
4	TFE	1:1	13
5	MeCN	1:1	11
6	DCE	1:1	50
7	DCE	2:1	50
8	DCE	1:2	60 (52)
9	DCE	1:4	55

^aReactions run on 0.1 mmol scale. Yields reported are NMR yields with isolated yields in parentheses. NMR yields reported referenced to a (Me₃Si)₂O internal standard.

With optimized conditions in hand, the substrate scope was explored (Figure 1). Several 2-chloro monomethoxy benzenoids (1–3)afforded the desired secondary amine products in moderate yields, using benzylic or heterobenzylic amine coupling partners. Alternative ether leaving groups were examined, and both benzyloxy and biaryl ether 4 and 5 afforded the resultant aniline in moderate yield. Electron-poor leaving groups such as acetoxy, triflate, and tosylate were also examined, but only returned starting material was observed.

Figure 1.

Primary amine CRA-S_NAr scope. ^a2-Picolylamine used in cases where purification proved difficult with other amine nucleophiles. ^b64% yield on 1.0 mmol scale using continuous flow setup. ^cFrom (R)-1-phenethan-1-amine. ^dAmine source was HBr salt. Reaction run in a 4:1 DCE/ aqueous pH 8 phosphate buffer solution. ^eAmine source was HCl salt. Reaction run in a 4:1 DCE/aqueous pH 8 phosphate buffer solution.

Next, 1,2-dimethoxy-veratrole derived substrates were tested. Amination of this class of arenes worked well overall, giving synthetically useful yields ranging from 25% to 77% yield (6–18). This class of 1,2-dimethoxy substrates bearing an electron-withdrawing group para- to the nucleofuge was particularly well-suited for CRA-S_NAr giving the best yields. This class of substrates may show increased yields due to captodative-like stabilization of the radical intermediate formed upon addition of the nucleophile to the arene cation radical.²¹ In all cases only one amination event was observed, with the other methoxy group still intact.

Importantly, various functional groups were well-tolerated including ketones, 7 and 13–17, generating the corresponding products in 50% and to 77% yield. Benzoate and benzonitrile substrates 8 and 9 also underwent the desired transformation in synthetically useful yields of 71% and 62%. 4-Chlorovera-trole 11 and aryl triflate 12 also underwent MeO-S_NAr with benzylamine, albeit with more moderate yields.

For all 3,4-dimethoxy arenes, preference for S_NAr at C4 was observed, with regioisomeric ratios ranging from 2:1–7:1 C4/ C3. This is in accordance with our previous ipso-substitution work^20,21 and follows computation trends identified using natural population analysis (NPA) of electron density of the ground and cation radical states of the arene, with the major site of substitution having the largest difference in positive charge density of the NPA values.^21,22 Additionally, the identity of the alkoxy group can improve regiocontrol in this substrate class, increasing the C4/C3 ratio from 3.3:1 to ~9:1 in favor of a methoxy leaving group over a benzyloxy (13–16).

Trimethoxy arenes also proved to be viable substrates, and excellent regioselectivity was observed providing a single isomer at the 2 (or 4) position, following the same regioselectivity trends as dimethoxy arenes (19–22).²¹ Additionally, heterocyclic compounds were also tolerated as demonstrated by quinoline (23) and isoquinoline (24), albeit with lower yields but with complete regiocontrol and, in the case of the Cl-substituted quinolone, complete chemo-selectivity.

The reaction also proceeded efficiently with a variety of primary amines. Halogenated benzyl amines worked effectively as nucleophiles (25 and 26) providing aniline products with a useful functional handle for downstream cross-coupling reactions. Unfortunately, attempts to employ secondary amines were unsuccessful in this transformation.

Employing (R)-1-phenylethan-1-amine as the nucleophile resulted in similar reactivity to the other benzylamine nucleophiles, affording 61% of the desired aniline (27). Unfortunately, this reaction proceeded with complete racemization, most likely due to oxidation of the amine or aniline product followed by formation of an α-amino radical.^23,24 Based on this result it is somewhat surprising that a cyclopropyl ring is tolerated (28), affording 46% of the desired aniline, suggesting that α-amino radical formation in this substrate is less favorable.

Next, other more oxidizable amines were employed as nucleophiles. Propyl (29), allyl (30), and propargyl (31) amine gave 65%, 71%, and 50%, yields respectively. These primary amine nucleophiles are more oxidizable than both benzylamine (+1.80 V vs SCE) and 3,4-dimethoxy-acetophenone (+1.60 V vs SCE), demonstrating that the reaction can proceed even with competitive oxidation of the primary amine. Additionally, Stern-Volmer analysis demonstrated that the rates of photoinduced electron transfer (k_q) kinetically favor oxidation of the arene (Figure 2a).

Figure 2.

(a) Catalyst fluorescence quenching data. (b) Stern-Volmer quenching analysis. (c) Proposed mechanism.

Phenethylamine, a prominent structural motif in many psychoactive compounds, was also effective albeit in more moderate yields affording the corresponding aniline (32) in 46% yield.²⁵ Aniline formation using 3-bromopropyl amine (33), ethanolamine (34), glycine (35), and aminoacetaldehyde dimethyl acetal (36) was also observed, demonstrating functional group compatibility. Lastly, heterocycle-containing primary amines (pyridine, pyrazine, and furan) were competent nucleophiles, affording products 37–39 in moderate to good yield.

Time-resolved Stern-Volmer quenching analysis was further utilized to compare the rates of quenching of the aniline products to those of the reactants (Figure 2b). These studies revealed that the rate of quenching of the aniline products is significantly faster than that of both the substrate and the primary amine reactants. In particular, 7 quenched slightly faster than the diffusion limit of DCE, which has been shown to be possible in cases where the Marcus inverted region is suppressed.²⁶ The oxidation potentials of the aniline products are significantly lower than that of the parent substrate as well, demonstrating that quenching of the catalyst by these aniline compounds is both thermodynamically and kinetically favored to oxidation of the substrate. This may lead to product inhibition of this reaction and offers some explanation as to why complete conversion to product is not observed in this system.

Though the mechanism for this transformation is still under investigation, we believe the reaction proceeds via a similar pathway akin to our previous report (Figure 2c).²⁰ Following excitation of di-tBu-Mes-Acr by visible light, the catalyst can be reductively quenched by the arene substrate, resulting in a single electron transfer. While the excited state acridinium catalyst can be quenched by both the arene substrate and the primary amine nucleophile, it is believed that aniline formation only occurs via kinetically favored oxidation the arene coupling partner to form an aryl cation radical. The primary amine nucleophile then adds to the methoxy-bearing carbon of the arene cation radical resulting in a distonic cation radical. Whether or not this is a discrete Meisenheimer-like intermediate or a concerted S_N2-like pathway is still under investigation. The acridine radical is oxidized by the S_NAr intermediate and extrusion of methanol occurs to give the final adduct, completing the catalytic cycle.

In conclusion, we have developed a photoredox-catalyzed procedure for S_NAr of simple methoxy aromatics using primary amine nucleophiles. These mild, metal-free conditions rely upon a highly oxidizing acridinium photoredox catalyst to generate reactive arene cation radicals. Mono-, di-, and trimethoxy arene starting materials, accessible from lignin derived feedstock chemicals, underwent S_NAr amination to yield aniline products in good to moderate yields and on a complementary set of substrates to both traditional S_NAr and transition metal cross-coupling reactions.