Mechanistic Insights into the FeCl3-Catalyzed Oxidative Cross-Coupling of Phenols with 2-Aminonaphthalenes

The selective FeCl3-catalyzed oxidative cross-coupling reaction between phenols and primary, secondary, and tertiary 2-aminonaphthalene derivatives was investigated. The generality of this scalable method provides a sustainable alternative for preparing N,O-biaryl compounds that are widely used as ligands and catalysts. Based on a comprehensive kinetic investigation, a catalytic cycle involving a ternary complex that binds to both the coupling partners and the oxidant during the key oxidative coupling step is postulated. Furthermore, the studies showed that the reaction is regulated by off-cycle acid–base and ligand exchange processes.


■ INTRODUCTION
Iron-catalyzed oxidative phenol coupling reactions 1 bring together phenols with unfunctionalized C−H nucleophiles such as 1,3-dicarbonyl compounds, 2 conjugated alkenes, 3 arenes, polyaromatic hydrocarbons (PAHs), 2b,4 and a second phenolic coupling partner. 5 This method is considered to be highly attractive in terms of step-and atom-economy for assembling new phenolic architectures. 6 As part of our group research program, we aimed to extend this reaction for the coupling of anilines to afford N,O-biaryl compounds that are widely used in asymmetric transformations. 7 Anilines and phenols share some of the same properties: they are both electron-rich cyclic π-systems that are prone to oxidation, generating a highly reactive electrophilic radical species. 8 Therefore, the development of selective oxidative crosscoupling reactions between phenols and unprotected anilines is a challenging task that has rarely been achieved. 9 In an early work, Kocǒvskýstudied the reaction between 2naphthol 1a and 2-aminonaphthalene 2a using a stoichiometric amount of a redox copper amine complex, 9k,l,n affording (±)-2amino-2′-hydroxy-1,1′-binaphthyl 3 (NOBIN, Scheme 1A). Recently, the Shindo group has developed aerobic oxidative cross-coupling conditions based on a heterogeneous Rh/C catalyst for the reaction between tertiary N,N-dialkylamino-2naphthalenes and different nucleophiles, such as N,Ndialkylanilines, arenes, and phenols (Scheme 1A). 9f,10 Lately, our group has developed an M[TPP]Cl (M = Fe or Mn, TPP = 5, 10,15,20-tetraphenyl-21H,23H-porphine)-catalyzed paraselective oxidative amination of phenols by primary and secondary anilines (Scheme 1B). 11 We have demonstrated that, depending on the identity of the phenolic para-R group, the products of this coupling are either benzoquinone anils (when R = H or OMe) or N,O-biaryl compounds (when R = alkyl). In a previous paper, we developed a two-step synthesis of optically pure NOBIN derivatives. The practical method is based on a stereoselective FeCl 3 -catalyzed oxidative crosscoupling between 2-naphthols (e.g., 1a, 1.5 equiv) and 2aminonaphthalenes with a labile chiral auxiliary group (such as 2c, 1 equiv, Scheme 1C), 12 affording a mixture of two separable NOBIN diastereoisomers [e.g., (R a ,S)-5 and (S a ,S)-5]. A simple hydrogenolysis of the auxiliary group (H 2 , Pd/C) offers a direct entry to the desirable (R a )-3 and (S a )-3 NOBINs in excellent chemical yields. Intrigued by the high degree of cross-coupling selectivity and the excellent yields imparted by the FeCl 3 /TFA/t-BuOOt-Bu catalytic system, we were interested in probing the underlying mechanism and studying the generality of this method for the preparation of N,O-biaryl compounds (Scheme 1D).
The general mechanistic line for the oxidative coupling of phenols by iron catalysts involves three key steps: (1) the formation of high-valent iron-phenolate complexes, (2) the generation of a ligated phenoxyl radical intermediate, and (3) coupling with a π-nucleophile or a radical species. 1a Recent mechanistic studies by our group for the FeCl 3 -catalyzed oxidative homo-and cross-coupling reaction of phenols revealed a zero-order dependence on the [phenol]. 5b Based on these results, it was suggested that a multicoordinated iron catalyst mediates an inner-sphere oxidative radical−anion coupling between two neighboring ligands ( Figure 1A). 5b However, the partial order for the phenolic component is no longer zero when the catalyst has a limited number of vacant sites, as exemplified by Katsuki [(Fe[μ-OH][salen]) 2 catalyst] 13 and Pappo [Fe[phosphate] 3 catalyst] ( Figure 1B). 5c Furthermore, when the reaction is mediated by Fe [TPP]Cl, which has only a single axial position available for binding, the coupling takes place between a ligated phenoxyl radical and a liberated phenoxyl 5a or an anilino radical 11 by an outer-sphere radical−radical coupling mechanism ( Figure 1C). These studies show that the coupling mechanism changes as a function of the iron coordination sphere. Therefore, the selectivity and the efficiency of the oxidative coupling are expected to be affected by the relative binding strengths of the two coupling partners to the redox iron complex.
Herein, we report that FeCl 3 is an efficient catalyst for the oxidative coupling between readily oxidized phenols and primary, secondary, and tertiary 2-aminonaphthalene derivatives. The selective conditions were successfully applied for the synthesis of a long list of novel N,O-biaryl compounds that are needed as ligands in catalysis. Our comprehensive mechanistic studies support the existence of an inner-sphere coupling mechanism between a phenoxyl radical and a 2-aminonaphthalene ligand. Furthermore, initial rate kinetic experiments uncovered (a) the involvement of a ternary complex that binds to both the coupling partners and the oxidant during the key oxidative coupling step and (b) the existence of two off-cycle acid−base and ligand exchange processes that regulate the reaction rate.
Mechanistic Studies. With the aim to elucidate a detailed catalytic cycle that will rationalize the observed reactivity and selectivity, a set of kinetic experiments were performed. The oxidative cross-coupling between 2,6-dimethylphenol (1b) and   Figure  4A and Figure S1 in the Supporting Information) and close to a zero-order dependency at a high concentrations range (0.1−0.3 M); (ii) a positive rate dependence for t-BuOOt-Bu is found when the experiment was performed at a high level of [1b] (0.25 M, Figure 5A); and (iii) first order in the catalyst was observed for FeCl 3 ( Figure 6A). The reaction does not take place in the absence of the redox catalyst, as confirmed by the zero intercept and our control experiments. The existence of free radical mechanisms was ruled out since the addition of butylated hydroxytoluene (BHT) to the reaction mixture had no effect on the coupling yield (see the Supporting Information). Furthermore, these kinetic results strengthen the premise that t-BuOOt-Bu binds to the iron catalyst prior to the slow oxidative coupling step, 2a,5c ruling out its action as a terminal oxidant that regenerates Fe(III) from Fe(II) after the coupling step.
The initial rate kinetic study indicated that 1b, t-BuOOt-Bu, and 2f, which showed negative order dependency (vide infra), are bound to the iron prior to the irreversible oxidative coupling step. Therefore, a kinetic behavior that characterizes a ternary enzyme (E, Figure 7) was considered. The action of ternary enzymes has been comprehensively studied by Cleland 15 and others. 16 These studies indicate that the mechanistic scheme of Bi-substrate enzyme-catalyzed reactions is characterized by a reversible binding of two substrates (A and B) to the enzyme prior to the slow step (E·A·B → products). The formation of E·A·B by the sequential binding of A and B can take place either via "random" or "ordered" sequential mechanisms. In a random mechanism, E·A·B is obtained from both E·A and E·B; i.e., the dissociation Figure 2. Scope of the oxidative coupling between 2-naphthols and 2aminonaphthalene derivatives. Reaction conditions: 2-naphthol (1.5 equiv), 2-aminonaphthalene (1 equiv), FeCl 3 (10 mol %), t-BuOOt-Bu (1.5 equiv), TFA (1.25 equiv), HFIP (0.5 M), room temperature, and 24 h. a The reaction was performed without TFA. b The reaction was performed with 3 equiv of 2-naphthol 1a, 4.5 equiv of t-BuOOt-Bu, and 3.75 equiv of TFA in total. Figure 3. Scope of the oxidative coupling between substituted phenols and 2-aminonaphthalene derivatives. Reaction conditions: phenol (1.5 equiv), 2-aminonaphthalene (1 equiv), FeCl 3 (10 mol %), t-BuOOt-Bu (1.5 equiv), TFA (1.25 equiv), HFIP (0.5 M), room temperature, and 24 h. a The reaction was performed with 3 equiv of t-BuOOt-Bu and 2.5 equiv of TFA in total. b The reaction was performed on a 2 mmol scale. c The conditions were similar, except for phenol (1 equiv).
The Journal of Organic Chemistry pubs.acs.org/joc Article constants of substrates to the free enzyme (K iA for A and K iB for B) and from E·A·B to enzymes E·A and E·B (K A and K B , respectively) are equal (K iA = K A and K iB = K B ). 17 However, in an ordered sequential mechanism E·A·B is obtained solely from E·A if substrate A binds preferentially to the free enzyme E (K iA < K A and K iB > K B ) or solely from E·B if binding of substrate A to this complex occurs in higher affinity (K iA > K A and K iB < K B ). 18 The order in which phenol 1b and t-BuOOt-Bu bind to the iron catalyst (assuming that E = [Fe](2f) m ] was determined by performing a set of double-reciprocal analysis experiments. 19 First, phenol 1b (assigned as substrate A) was varied at fixed concentrations of t-BuOOt-Bu (0.05, 0.15, and 0.25 M, Figure  4B; see also Figure S1 in the Supporting Information) and then t-BuOOt-Bu (assigned as substrate B) was varied at fixed [phenol 1b] values (0.15, 0.25, and 0.30 M, Figure 5B; see also Figure S2 in the Supporting Information). The Lineweaver−Burk plot for the phenol ( Figure 4B) shows linear lines that intersect above the horizontal axis, whereas the position of the crossover point for the peroxide's linear lines ( Figure 5B) is below the x-axis. According to Frieden analysis, 19 these results indicate that K iA > K A and K iB < K B (Figure 7, Eq. 3), 16,17,19,20 suggesting that [Fe]·(2f) m ·(1b)·(t-BuOOt-Bu) III is formed from [Fe]·(2f) m I by a sequential binding of the peroxide (step A, Scheme 2) and the phenol (Step B).
The dependence of 2-aminonaphthalene 2f on the reaction velocity was examined ( Figure 6B). The negative relationship between the reaction rate and [2f] indicates the presence of a competitive off-cycle equilibrium. 21 It is suggested that the association of the peroxide and the phenol to complex I (step A, Scheme 2) is suppressed by the competitive binding of 2f, affording [Fe]·(2f) m+1 (V, off-cycle step, Scheme 2). Consequently, the rate of the coupling decelerates as [2f] increases. These results also support the assumption that 2-  The Journal of Organic Chemistry pubs.acs.org/joc Article aminonaphthalene 2f serves as a strong N-ligand that coordinates to the iron in preference to phenol 1b and t-BuOOt-Bu.
Product inhibition experiments offer useful inputs when deciding the kinetic mechanism of a ternary complex (E·Q ⇄ E + Q, Figure 7). The experiments were performed by monitoring the formation rate of product 30 in the presence of increasing concentrations of NOBIN 6 or t-BuOH. NOBIN 6 was chosen for practical reasons associated with the fact that NOBINs 30 and 6 have different retention times in the HPLC. Figure 8 shows that although t-BuOH acts as a weak inhibitor, at saturating values of NOBIN 6, the catalyst's activity approaches zero. Unsurprisingly, these results indicate that the coupling product acts as a competitive ligand. It is expected that the concentration of complex IV ([Fe]·(2f) m−1 (30), Scheme 2) builds up as the reaction proceeds; consequently, the velocity of the coupling decreases.
Based on these kinetic results, a detailed mechanistic scheme was postulated and is presented in Scheme 2. The catalytic cycle begins with the reversible binding of the peroxide and the phenol to [Fe]·(2f) m (I), affording complex III ([Fe]·(2f) m · (1b)·(t-BuOOt-Bu)), steps A and B). The homolytic cleavage of the peroxide bond by the iron, followed by an inner-sphere coupling between a phenoxyl radical and a neighboring 2aminonaphthalene ligand, will afford complex IV and two molecules of t-BuOH (step C). 2d,e,5b, 14 The catalytic cycle is terminated by a reversible ligand exchange process that involves the liberation of N,O-biaryl product 30, along with the binding of 2-aminonaphthalene 2f (step D).   The Journal of Organic Chemistry pubs.acs.org/joc Article Our study implies that the reaction kinetics is strongly influenced by the relative binding strength of the substrates (2aminonaphthalene, phenol, and peroxide) and the coupling product to the iron. As mentioned previously, the addition of TFA is mandatory when secondary and tertiary 2-amino-naphthalenes are being reacted. It is expected that TFA, which forms an acid−base adduct with the latter coupling partners, interferes in the net of ligand exchange processes. To clarify the role of the acid, we performed additional sets of kinetic experiments.
The dependence of the initial rate on [TFA] ( Figure 6C) revealed that, although no reaction occurs in the absence of the acid, the maximum reactivity is achieved when [2f] and [TFA] are equalized (ca. a 1:1 ratio). However, as the acid concentration increases, the reaction velocity diminishes. These results can be rationalized by the existence of ligandto-metal exchange and acid−base net reactions (Scheme 2). It is suggested that the entire catalytic process is regulated by TFA, which forms an acid−base adduct with 2f. Accordingly, as the acid concentration increases, the concentration of free 2aminonaphthalene 2f drops (step E). Consequently, the offcycle equilibrium inclines toward complex I and the rate accelerates (0.05 M < [TFA] < 0.12 M). On the other hand, at high concentrations of TFA ([TFA] > 0.12 M), the concentration of 2f diminishes. Consequently, the catalytic cycle termination step (IV → I, step D), which includes the reversible ligand exchange of the N,O-biaryl product 30 with 2f, is discouraged, and the reaction rate declines.
The strength of the TFA-based adduct depends on the basicity of the 2-aminonapthalene molecule. Therefore, different amounts of acid should be used to regulate the coupling of primary, secondary, or tertiary 2-aminonaphthalenes. To support this claim, a set of competitive experiments that studied the coupling of 2-aminonaphthalenes 2a, 2d, or 2e (1 equiv) and 2-naphthol (1a, 1.5 equiv) either with or without 2 equiv of TFA were performed (Figure 9). The results show that the addition of TFA to the reaction of 2a, which is a weaker base in comparison with 2d and 2e, negatively affects the reaction rate ( Figure 9A). On the other  The Journal of Organic Chemistry pubs.acs.org/joc Article hand, the reaction of 2d in the presence of TFA resulted in a significant improvement in the reactivity ( Figure 9B) and cross-coupling selectivity (see Figure S3 in the Supporting Information). Finally, tertiary 2-aminonaphthalene 2e exhibited only a mild improvement in the rate upon the addition of TFA ( Figure 9C). This is probably because 2 equiv of TFA are insufficient to regulate the inhibiting off-cycle process. Indeed, almost twice the amount of TFA (3.75 equiv) is needed to ensure efficient cross-coupling, affording NOBIN 7 in 97% yield ( Figure 2). Ultimately, the coupling of primary 2aminonaphthalene takes place at a high efficiency without TFA (see the inserted table, Figure 9), whereas the successful coupling of secondary and tertiary 2-aminonaphthalenes relies on the addition of TFA (1.25 equiv and 3.75 equiv, respectively).
The changes in the catalytic activity at high concentrations of TFA may also be attributed to the generation of iron trifluoroacetate complexes [Fe(CF 3 CO 2 ) n (Cl) m ]. To examine this hypothesis, the Fe(CF 3 CO 2 ) 3 complex 22 was prepared and used as a catalyst (10 mol %) in the coupling between 1b and 2f ( Figure 10). Fe(CF 3 CO 2 ) 3 exhibited almost no catalytic activity. However, the reactivity was enhanced with the addition of 10 mol % of tetrabutylammonium chloride (TBAC). Interestingly, almost a complete recovery of the catalytic activity (in comparison to FeCl 3 ) was achieved with 1:2 and 1:3 iron to chloride ratios. These results suggest that the chloride anions play a key role during the reaction.

■ CONCLUSIONS
In conclusion, the FeCl 3 -catalyzed oxidative phenol coupling reaction was applied to combine readily oxidized phenols with primary, secondary, and tertiary 2-aminonaphthalenes. This sustainable and practical method enables a highly selective and efficient synthesis of N,O-biaryl compounds that are not readily available by other means.
Our mechanistic data, which include control experiments and comprehensive kinetic studies, revealed the existence of a catalytic cycle that involves the formation of a ternary iron complex [Fe]·(2f)·(1b)·(t-BuOOt-Bu) (III) from [Fe]·(2f) m (I) by the sequential binding of peroxide and phenol. The irreversible rate-determining oxidative coupling step comprises the conversion of complex III to IV ([Fe]·(2f) m-1 (30)) and the liberation of two molecules of t-BuOH. In this transformation, a reaction between an iron-bound phenoxyl radical and a neighboring 2-aminonaphthalene ligand takes place. The velocity of the reaction is regulated by a net of acid−base and ligand exchange processes. The reaction rate is highly sensitive to changes in the concentrations of the substrates (2aminonaphthalene, phenol, and peroxide), the acid (TFA), and the N,O-biaryl product. Furthermore, the chloride anions have a strong effect on the reaction efficiency. Finally, this study is a part of our laboratory ongoing research that aims to develop selective oxidative cross-coupling reactions for the coupling of anilines by first-row metal catalysts.

■ EXPERIMENTAL SECTION
General Methods. All reagents were of reagent-grade quality, purchased commercially from Sigma-Aldrich, Alfa-Aesar, or Fluka, and used without further purification. FeCl 3 (anhydrous 98%) was purchased from Strem Chemicals. Purification by column chromatography was performed on Merck chromatographic silica gel (40−63 μm). Thin-layer chromatography (TLC) analyses were performed using Merck silica gel glass plates 60 F254. NMR spectra were recorded on Bruker DPX400 or DMX500 instruments; chemical shifts are relative to Me 4 Si as the internal standard or to the residual solvent peak. High-resolution mass spectrometry (HRMS) data were obtained using an LTQ Orbitrap XL ETD (Thermo Fisher Scientific, Germany and USA) high-resolution mass spectrometer. The reactions in the microwave were performed using a CEM Discover SP microwave synthesizer. IR spectra were recorded on a JASCO FT/IR-460 Plus FT-IR instrument. HPLC analysis was carried out on an Agilent 1260 instrument equipped with a G4212-60008 photodiode array detector and an Agilent reverse phase ZORBAX Eclipse plus C18 3.5 μm column (4.6 × 100 mm).
General Procedures for the Synthesis of N-Alkyl-2-aminonaphthalenes. Method A. A mixture of 2-naphthol derivative (1 equiv) and alkyl/arylamine (5 equiv) was irradiated in a microwave for 20 h (sealed reaction vessel, temperature of 275°C was monitored by using an external surface sensor and a power of 200 W). The volatiles were removed under reduced pressure, and the crude residue was further purified by silica-gel column chromatography (silica gel 40−63 μm). This method was used for the preparation of 2aminonaphthalene derivatives 2d, 2e, 2i, 2j, 2n, and 2o.
Method B. 23 A mixture of 2-aminonaphthalene (1 equiv) and benzaldehyde (1.1 equiv) was stirred in methanol (0.17 M) for 1 h, and then NaBH 4 (1.5 equiv) was added. The reaction was stirred for 20 min, and the volatiles were removed under reduced pressure. NaOH (1 M, 30 mL) was added and extracted with diethyl ether (3 × 20 mL). The combined organic phase was dried over MgSO 4 and evaporated under reduced pressure. The crude residue was further purified by silica-gel column chromatography (silica gel 40−63 μm). This method was used for the preparation of 2-aminonaphthalene derivatives 2f, 2g, 2h, 2k, 2l, 2 m, and 2p.