Selective C-C coupling reaction of dimethylphenol to tetramethyldiphenoquinone using molecular oxygen catalyzed by Cu complexes immobilized in nanospaces of structurally-ordered materials.

Two high-performance Cu catalysts were successfully developed by immobilization of Cu ions in the nanospaces of poly(propylene imine) (PPI) dendrimer and magadiite for the selective C-C coupling of 2,6-dimethylphenol (DMP) to 3,3',5,5'-tetramethyldiphenoquinone (DPQ) with O2 as a green oxidant. The PPI dendrimer encapsulated Cu ions in the internal nanovoids to form adjacent Cu species, which exhibited significantly high catalytic activity for the regioselective coupling reaction of DMP compared to previously reported enzyme and metal complex catalysts. The magadiite-immobilized Cu complex acted as a selective heterogeneous catalyst for the oxidative C-C coupling of DMP to DPQ. This heterogeneous catalyst was recoverable from the reaction mixture by simple filtration, reusable without loss of efficiency, and applicable to a continuous flow reactor system. Detailed characterization using ultraviolet-visible (UV-vis), Fourier transform infrared (FTIR), electronic spin resonance (ESR), and X-ray absorption fine structure (XAFS) spectroscopies and the reaction mechanism investigation revealed that the high catalytic performances of these Cu catalysts were ascribed to the adjacent Cu species generated within the nanospaces of the PPI dendrimer and magadiite.

1. Introduction

The oxidative C–C coupling reaction of 2,6-dimethylphenol (DMP) has attracted interest as a synthetic method for 3,3',5,5'-tetramethyldiphenoquinone (DPQ), which is a useful building block of aryl epoxy resins [1] and acceptor-doped polymers [2], and an activator for redistribution of poly(2,6-dimethyl phenylene ether) (PPE) [3]. Traditionally, the oxidative C–C coupling of DMP to DPQ has been carried out using excess amounts of stoichiometric metal oxidants such as Mn(OAc)3 [4], FeCl3·6H2O [5], Co(OAc)3 [6], and Ph3Bi(OAc)2 [7]. However, these methods suffer from the production of a large quantity of harmful metal wastes.

With respect to environmental impact and atom efficiency, one of the most promising methods is the catalytic oxidative coupling of DMP to DPQ with O2 as an oxidant because only H2O is formed as the co-product. In this context, various transition metal catalysts have been applied to the aerobic coupling reaction of DMP. However, the control of the selectivity for C–C coupling has been difficult, resulting in the occurrence of side reactions such as C–O coupling of DMP to PPE and oxidation of DMP to dimethylquinone (DMQ) [8,9,10,11,12,13,14,15]. Up to now, few successful catalyst systems for selective C–C coupling of DMP to DPQ using O2 have been reported. Bhalerao et al. reported the first demonstration of the regioselective coupling of DMP to DPQ by using enzyme mushroom tyrosinase in 1990. Mushroom tyrosinase, which has a dinuclear copper species as an active center, promotes the selective C–C coupling in a phosphate buffer solution at room temperature [16]. Since this report, several homogeneous catalysts have been developed [17,18,19]. For example, Mukherjee et al. reported a biomimetic copper catalyst [(α,α'-bis[(N-methyl-2-pyridyl)ethylamino]-2-fluoro-m-xylene)Cu2(MeCN)2][ClO4]2 [17], and Liu et al. developed a binuclear copper catalyst Cu2(bpnp)(μ-OH)(TFA)3 (bpnp = bis(2-pyridyl)-1,8-naphthyridine) [18]. The heteropolyanion H5PV2Mo10O40 was also found to catalyze the C–C coupling of DMP to DPQ by Lissel et al. [19]. However, these catalyst systems have not yet addressed problems with low activity, and/or difficulty in catalyst separation and reuse. Consequently, the development of efficient and reusable catalytic systems for the selective C–C coupling of DMP to DPQ is still in high demand.

A dendrimer is a spherical macromolecule having a regularly branched framework from a core to termini. One of the unique features, different from linear and branched polymers, is an internal nanovoid confined by a regularly branched structure, which enables accommodation of various organic/inorganic molecules within the void space [20,21]. The internal nanovoid serves as a platform for the synthesis of unique metal species including metal nanoclusters [22,23] and the encapsulated metal species are applicable as catalysts for organic reactions [24]. Our research group has synthesized various dendrimers encapsulating mononuclear metal complexes [25], metal nanoparticles [26], and subnano metal clusters [27,28] and explored their unique catalyses for selective molecular transformations. In the course of our studies on these dendrimer-encapsulated metal catalysts, a poly(propylene imine) (PPI) dendrimer-immobilized Cu complex catalyst was developed for the regioselective coupling reaction of DMP to DPQ with O2 as an oxidant [29]. The PPI dendrimer encapsulated Cu ions to generate unique adjacent Cu species within the nanovoids, exhibiting a higher catalytic activity compared to previously reported catalysts.

This article describes a comprehensive study of the development of the above dendrimer-encapsulated Cu complex catalyst. Furthermore, in order to realize a more efficient catalyst system with easy handling and high durability, we newly devise a Cu2+-exchanged magadiite (Cu2+-magadiite) as a heterogeneous catalyst for the selective oxidative coupling reaction of DMP to DPQ. Cu2+-magadiite is easily separable from the reaction mixture by filtration and reusable without loss of activity or selectivity. Cu2+-magadiite also can be applied to a continuous flow reactor system, achieving a multigram-scale synthesis of DPQ. The detailed characterizations using UV-vis, FTIR, XRD, XAFS, and ESR measurements of these dendrimer-encapsulated and magadiite-immobilized Cu catalysts reveal that the adjacent Cu species generated within the nanospaces of dendrimer and magadiite play a key role in the efficient oxidative coupling of DMP to DPQ.

2. Results

2.1. Preparation of Immobilized Cu Catalysts

2.1.1. PPI Dendrimer-Encapsulated Cu Catalyst

A PPI dendrimer possesses tertiary amino groups at branching points and primary amino groups at termini, which act as metal coordination sites [30] and base sites [31]. The modification of the terminal amino groups with bulky triethoxybenzamide groups creates confined internal nanovoids within the PPI dendrimer, which can encapsulate metal ions. The triethoxybenzamide-terminated fourth-generation PPI dendrimer (G4-TEBA) was synthesized through the reaction between NH2-terminated G4-PPI dendrimer (G4-NH2) with 3,4,5-triethoxybenzoyl chloride [26]. Next, a series of G4-TEBA-encapsulated Cu2+ complexes G4-Cu2+n (n denotes the number of Cu2+ ions in the G4-TEBA, n = 2, 8, 12, 16, and 24) were prepared by stirring a mixture of CuCl2 and G4-TEBA (CuCl2 to G4-TEBA molar ratio was n:1) in CH3CN/CHCl3 (1:2 v/v) under an Ar atmosphere [32]. The resulting solution was evaporated to afford G4-Cu2+n as brown waxy solids.

2.1.2. Magadiite-Immobilized Cu Catalyst

Magadiite (Na2Si14O29·nH2O) is a layered silicate consisting of negatively charged silicate layers and Na+ cations in the interlayer nanospace, and has characteristics such as easy availability by a simple hydrothermal procedure [33] and high cation-exchange capacity (180 meq/100 g) [34]. The accessibility and cation-exchange ability of magadiite have led to several applications, for example, adsorbents of metal species [35] and building units of photofunctional materials [36].

Magadiite was synthesized by a hydrothermal reaction of SiO2, NaOH, and H2O [33]. The magadiite-immobilized Cu complex was prepared by a cation exchange reaction (see Experimental Section). In brief, magadiite was added to the methanol solution containing Cu(ClO4)2·6H2O and N,N,N',N'-tetramethylethylenediamine (TMEDA) (molar ratio of TMEDA to copper was 1:1), and then the resulting mixture was stirred at 313 K. The obtained solid was filtered, washed, and dried, affording Cu2+-magadiite as a light blue powder.

2.2. Characterization of Immobilized Cu Catalysts

2.2.1. PPI Dendrimer-Encapsulated Cu Catalyst

A series of G4-Cu2+n was characterized by UV-vis, FTIR, ESR, and XAFS analyses. G4-Cu2+n exhibited a d-d transition band at 819 nm, whereas Cu2+ ions in the absence of G4-TEBA showed a broad absorption peak centered at 860 nm [32]. The absorption intensity at 819 nm increased upon increasing the amount of Cu ions (ε ≈ 130 M−1cm−1), indicating the encapsulation of Cu2+ ions to G4-TEBA [32]. Based on the titration of G4-TEBA with CuCl2 in CHCl3/CH3CN at 819 nm, the maximum number of Cu2+ ions encapsulated within a G4-TEBA was estimated to be 30 [30,32]. Similar phenomena and the determination of the maximum number of Cu2+ ions in a dendrimer based on spectrophotometric titration were reported by Crooks et al., where the PPI dendrimer encapsulated Cu2+ ions in CHCl3/MeOH solvent [30]. Additionally, the absorption band attributed to the LMCT of N- and Cl-Cu was observed at approximately 300 nm [32,37,38].

FTIR analysis was carried out to gain structural information on the encapsulated Cu species of G4-Cu2+n [32]. It was reported that the treatment of Cu mono- and di-nuclear complexes with NaN3 gave terminal Cu azide complexes and bridged Cu azide complexes, exhibiting a single peak around 2040 cm−1 and two peaks around 2070 cm−1 derived from the N3 asymmetric stretch (νas(N3)), respectively [39]. When treating G4-Cu2+n (n = 2 or 8) with NaN3, the FTIR spectrum exhibited a peak around 2048 cm−1 derived from the νas(N3) of the terminal mononuclear Cu azide complex. As n increased above 12 (n = 12, 16, and 24), the FTIR spectra of the Cu-azide complexes exhibited two peaks around 2055 and 2080 cm−1. Both peaks were attributed to the νas(N3) of μ-1,1-binuclear Cu azide complex. The above FTIR measurements indicated that adjacent Cu species were generated within G4-Cu2+n (n ≥ 12). As a control experiment, we synthesized a branched polyethyleneimine modified with TEBA groups (PEI-TEBA) as an irregularly branched polyamine [32], and prepared the NaN3-treated PEI-Cu2+ complex (the molar ratio of tertiary amino groups to Cu2+ was adjusted to that in G4-Cu2+12). The FTIR spectrum showed a peak corresponding to the νas(N3) of terminal mononuclear Cu azide complex at 2048 cm−1, indicating the formation of monomeric Cu2+ species in PEI-Cu2+. Consequently, the internal nanovoid confined by the regularly branched structure of the dendrimer is essential to form adjacent Cu species.

The generation of adjacent Cu species within G4-Cu2+n (n ≥ 12) could also be supported by ESR measurements. The ESR spectrum of G4-Cu2+2 in CHCl3 exhibited a mononuclear Cu2+ signal around 320 mT (Figure 1). The intensity of the Cu2+ signal of G4-Cu2+n decreased with increasing n from 2 to 12, and then the spectra of G4-Cu2+n (n ≥ 12) were almost silent. It is known that the spin–spin interaction between closely related Cu2+ ions decreases the intensity of the Cu2+ signal [40], revealing the main formation of adjacent Cu2+ species within G4-Cu2+n (n ≥ 12).

Figure 1

ESR spectra of G4-Cu2+n (n = 2, 8, 12, 16, and 24) in CHCl3 recorded with the following parameters: temperature: 298 K; power: 10.0 mW; modulation amplitude: 0.5 G; modulation frequency: 100 kHz.

To gain insight into the coordination environment of the Cu2+ species of G4-Cu2+12, XAFS analysis was carried out [32]. The Cu K-edge X-ray absorption near edge structure (XANES) spectrum of G4-Cu2+12 exhibited an edge peak at 8982 eV. This peak might be ascribed to the shake-down transition involving a 1s→4p transition with simultaneous LMCT [41]. Fourier transforms (FT) of the k3-weighted extended X-ray absorption fine structure (EXAFS) spectrum of G4-Cu2+12 showed a scattering peak at 1.8 Å corresponding to the Cu-N and Cu-Cl shells based on comparison with those of CuCl2 [42] and Cu(ImH)4SO4 [43]. Curve-fitting analysis revealed that the Cu species were surrounded by two N and two Cl atoms.

From the combined results of UV-vis, FTIR, ESR, and XAFS analyses, a possible structure of Cu species within the nanovoids of G4-Cu2+12 is shown in Figure 2. CuCl2 species are encapsulated through the coordination to two amino groups of the branch units of G4-TEBA to form adjacent Cu centers.

Figure 2

Proposed structure of Cu species within G4-Cu2+12.

2.2.2. Magadiite-Immobilized Cu Catalyst

The XRD analysis revealed that the basal spacing of Cu2+-magadiite (d001 = 13.9 Å) decreased from that of the parent magadiite (d001 = 15.1 Å). The d001 value of Cu2+-magadiite was larger than those of proton-exchanged magadiite (H+-magadiite) (d001 = 11.5 Å) [44] and dehydrated magadiite (d001 = 12.8 Å) [45], confirming that Cu species were incorporated within the magadiite interlayer. The interlayer gallery of Cu2+-magadiite was estimated as 2.4 Å based on the subtraction of layer thickness of H+-magadiite [46]. The elemental analysis showed C, 1.48; H, 2.22; N, 0.60; Na, 1.49; Cu, 1.39; Si, 36.1%, suggesting that the stoichiometric amount of TMEDA to Cu2+ was 0.975 and Cu2+ species were incorporated into magadiite through a cation exchange reaction with interlayer Na+ cations.

The UV-vis spectrum of Cu2+-magadiite showed absorption bands around 690 and 280 nm corresponding to the d-d transition band and LMCT band, respectively (Figure S1) [47]. The absorbance around 690 nm indicated the presence of Cu2+ species coordinated by two N atoms [48]. The absorbance around 280 nm was deconvoluted into two peaks at 270 and 340 nm (Figure S2), attributed to N- and bridging OH−-Cu LMCT, respectively [47,48].

The Cu K-edge XANES spectrum of Cu2+-magadiite showed an absorption edge at 8980 eV, which was assigned to the formation of Cu2+ species (Figure S3) [47]. The FT of the k3-weighted EXAFS spectrum exhibited two scattering peaks around 1.5 and 2.9 Å (Figure 3). These peaks were also observed in the spectrum of a dinuclear Cu complex [Cu(OH)TMEDA]2Cl2 [39], while the scattering peaks around 2.9 Å were absent in that of the magadiite-immobilized Cu(ethylenediamine)2 mononuclear complex (Cu2+(mono)-magadiite) [49,50]. Thus, the above two peaks around 1.5 and 2.9 Å in the spectrum of Cu2+-magadiite were ascribed to Cu-N/O and Cu-Cu shells, respectively. The inverse FT of these peaks was well fitted to Cu-N/O and Cu-Cu shells with coordination numbers of 4.4 and 0.9, respectively (Figure 3 and Table 1). These results indicate that dinuclear Cu2+ species are generated and coordinated by nitrogen/oxygen ligands such as TMEDA and bridging OH− groups [48].

Figure 3

FT of k3-weighted Cu K-edge EXAFS spectra of (a) Cu2+-magadiite, (b) Cu2+-magadiite (used), (c) [Cu(OH)TMEDA]2Cl2, (d) Cu2+(mono)-magadiite, (e) CuO, and (f) Cu2O.

Table 1

Results of curve-fitting analysis of Cu K-edge EXAFS data for Cu2+-magadiite a.

Sample	Shell	CN b	R c [Å]	σ2 d [Å2]
[Cu(OH)TMEDA]2Cl2	Cu-O/N	4.0	2.02	-
	Cu-Cu	1.0	2.99	-
Cu2+-magadiite (fresh)	Cu-O/N	4.4	1.99	0.0016
	Cu-Cu	0.9	2.97	0.0042
Cu2+-magadiite (used)	Cu-O/N	4.5	1.98	0.0060
	Cu-Cu	1.1	2.93	0.0022

a The region of 1.0–3.3 Å in FT of samples was inversely transformed. b Coordination number. c Interatomic distance. d Debye-Waller factor.

In ESR measurements, the intensity of the Cu2+ signal of Cu2+-magadiite was much lower than that of Cu2+(mono)-magadiite (Figure 4). This result was ascribed to the spin–spin interaction of closely related Cu2+ species [40]. The ESR spectrum of Cu2+-magadiite yielded a g-value of 2.263 and hyperfine splitting A of 165 cm−1. These two values indicate the presence of Cu2+ species in 2N2O coordination environments [51].

Figure 4

ESR spectra of magadiite-immobilized Cu2+ complexes recorded with the following parameters: temperature: 298 K; power: 10.0 mW; modulation amplitude: 0.5 G; modulation frequency: 100 kHz.

Consequently, we propose the structure of Cu species in Cu2+-magadiite (Figure 5). Adjacent Cu species (Cu···Cu ≈ 2.9 Å) coordinated by TMEDA and oxygen ligands such as bridging OH− are immobilized in the interlayer nanospace of magadiite.

Figure 5

Proposed structure of Cu species in Cu2+-magadiite.

2.3. Oxidative Coupling of DMP Using Immobilized Cu Complex Catalysts

2.3.1. G4-Cu2+n-Catalyzed Oxidative Coupling of DMP

The oxidative coupling reactions of DMP (1a) using various G4-Cu2+n (n = 2–24) were examined at 323 K under 1 atm of O2 in CHCl3 (Table 2). The amount of encapsulated Cu2+ ions to G4-TEBA (n) strongly affected the activity and selectivity. Among the tested G4-Cu2+n, G4-Cu2+12 exhibited the highest activity and selectivity for the C–C coupling, affording the desired product DPQ (2a) and 3,3',5,5'-tetramethylbiphenol (TMBP, 3a), which is known as the reaction intermediate for 2a [19], in 55% and 10% yields, respectively, where the selectivity for the C–C coupling reached 97% (entry 3). Prolonging the reaction time allowed complete conversion of 1a giving 2a in 97% yield (entry 6). When decreasing n from 12 to 2, both the conversion of 1a and the selectivity for 2a decreased to 9% and 44%, respectively (entries 1 and 3). The increase of n from 12 to 24 diminished the conversion of 1a to 16% while maintaining the high selectivity for the C–C coupling (entries 3 and 5).

Table 2

Oxidative coupling of DMP using various Cu-amine catalysts.

Entry	Catalyst	Solvent	Time [h]	Conv. a [%]	Sel. to	Yield a [%]
					C-C b [%]	2a	3a	4a
1	G4-Cu2+2	CHCl3	6	9	44	4	0	4
2	G4-Cu2+8	CHCl3	6	25	68	8	9	7
3	G4-Cu2+12	CHCl3	6	67	97	55	10	2
4	G4-Cu2+16	CHCl3	6	34	88	15	15	4
5	G4-Cu2+24	CHCl3	6	16	87	7	7	2
6	G4-Cu2+12	CHCl3	18	>99	97	97	trace	2
7	G4-Cu2+12c	CHCl3	6	29	96	22	6	1
8	CuCl2-TEA	CHCl3	6	9	44	1	3	5
9	CuCl2-TMPDA	CHCl3	6	98	46	41	4	52
10	PEI-Cu2+	CHCl3	6	11	63	3	4	1
11	G4-Cu2+12	MeOH	6	97	46	32	12	47
12	G4-Cu2+12	CH3CN	6	9	33	0	3	0
13	G4-Cu2+12	TFT	18	>99	96	96	trace	2
14 d	G4-Cu2+12	CHCl3	24	>99	97	97 e	trace	2

a Determined by 1H NMR standard technique. b Calculated from the ratio of yield of (2a + 3a) to conv. of 1a. c G4-Cu2+12 was treated with HCl(aq) (0.01 N, 0.25 mL) before the reaction. d Reaction conditions: G4-Cu2+12 (Cu: 18 µmol), 1a (1.10 g, 9 mmol), CHCl3 (16 mL), 323 K, 24 h, O2 (5 atm). e Isolated yield.

To prove the high efficiency of G4-Cu2+12, several tertiary amine compounds were used instead of G4-TEBA [52]. The low-molecular-weight amines such as triethylamine (TEA) and N,N,N',N'-tetramethyl-1,3-propanediamine (TMPDA) could not promote the selective oxidative coupling of 1a, resulting in the formation of PPE (4a) as a main product (entries 8 and 9). In addition, the use of PEI-Cu2+ gave only 3% yield of 2a with 63% selectivity (entry 10), demonstrating that G4-Cu2+12 acts as a unique and efficient catalyst for the oxidative C–C coupling of DMP to DPQ.

The choice of solvent significantly influenced the catalytic activity and selectivity of G4-Cu2+12 for the oxidative coupling reaction of DMP to DPQ. Among the solvents tested, CHCl3 gave the highest yield of 2a (97% yield). Although the high conversion of 1a was achieved in MeOH, the selectivity for the C–C coupling was only 46% (entry 11). CH3CN was not effective, resulting in the formation of a slight amount of 3a (entry 12). It is noteworthy that under the reaction conditions using G4-Cu2+12 in α,α,α-trifluorotoluene (TFT) solvent, 2a was obtained in 96% yield (as determined by 1H-NMR analysis) and existed as a solid after the reaction (entry 13). Thus, 2a was easily obtained from the reaction mixture by simple filtration. The recovered TFT solution containing G4-Cu2+12 was reusable with retention of its efficiency during the recycling experiments (in 93% yield) [47].

Under the optimized reaction conditions, the G4-Cu2+12-catalyzed oxidative coupling of other phenol derivatives was investigated. G4-Cu2+12 efficiently catalyzed the C–C selective coupling reaction of 2,6-substituted phenols including 2,6-diisopropylphenol (1b) and 2-tert-butyl-6-methylphenol (1c) to afford corresponding diphenoquinone derivatives (2b and 2c) in 97% yields for 24 h (Scheme 1). These results differ from those obtained with mushroom tyrosinase, where both 1b and 1c were not fully converted even after extending the reaction time [16]. The coupling reaction of unsubstituted phenol or phenol derivatives with electron-withdrawing chlorine atoms did not proceed, which is similar to the results obtained by other Cu catalyst systems [16,18].

Scheme 1

G4-Cu2+12-catalyzed oxidative coupling of phenol derivatives.

A gram-scale reaction of 1a was carried out using G4-Cu2+12. 1.10 g of 1a was selectively converted to 1.04 g (97% isolated yield) of 2a (Table 2, entry 14) [32]. In this case, G4-Cu2+12 exhibited both high turnover number (TON) and turnover frequency (TOF), reaching 485 and 20.2 h−1, respectively. These values were much higher than those of reported catalyst systems: mushroom tyroinase (TON, 48; TOF, 5.33 h−1) [16] [(α,α'-bis[(N-methyl-2-pyridyl)ethylamino]-2-fluoro-m-xylene)Cu2(MeCN)2] [ClO4]2 (TON, 8.6; TOF, 1.43 h−1) [17], [Cu2(bpnp)(μ-OH)(TFA)3] (TON, 100; TON, 4.16 h−1) [18], and H5PV2Mo10O40 (TON, 40; TOF, 10 h−1) [19].

2.3.2. Heterogeneous Oxidative Coupling Reaction of DMP Using Cu2+-Magadiite

Although heterogeneous catalysts are superior to homogeneous catalysts in terms of handling, separation, and reuse, a heterogeneous catalyst for the oxidative C–C coupling of DMP to DPQ has never been reported. Therefore, we focused on development of the solid support-immobilized Cu complex catalyst for the selective C–C coupling of DMP to DPQ.

The catalytic activities of a series of the Cu complexes immobilized to solid supports were investigated in the oxidative coupling of 1a. The results are summarized in Table 3. Cu2+-magadiite efficiently promoted the oxidative coupling reaction under 1 atm of O2 at 328 K, affording a 67% yield of 2a for 12 h (entry 1). When prolonging the reaction time to 18 h, the yield of 2a reached 95% with the formation of a 4% yield of 4a (entry 2). The use of Cu2+-SiO2 decreased the selectivity for 2a from 95% to 60% (entry 6 vs. entry 1). Cu2+-mordenite exhibited low activity and selectivity to give 2a in 27% yield with 53% selectivity (entry 7). In addition, Cu2+(mono)-magadiite having monomeric Cu2+ species gave poor yields of 2a (entry 8).

Table 3

Oxidative coupling of DMP catalyzed by Cu complexes immobilized to solid supports. a

Entry	Catalyst	Time [h]	Conv. b [%]	Yield b [%]
				2a	4a
1	Cu2+-magadiite	12	75	67	3
2	Cu2+-magadiite	18	>99	95	4
3 c	Cu2+-magadiite	18	>99	94	4
4 d	Cu2+-magadiite	18	>99	94	4
5 e	Cu2+-magadiite	48	>99	95	3
6	Cu2+-SiO2	12	>99	60	3
7	Cu2+-mordenite	12	51	27	2
8	Cu2+(mono)-magadiite	12	15	2	2

a Reaction conditions: catalyst (Cu: 17.5 µmol), 1a (0.4 mmol), CHCl3 (3.5 mL), MeOH (0.5 mL), 328 K, O2 (1 atm). b Determined by 1H NMR standard technique. c 1st reuse. d 2nd reuse. e Cu2+-magadiite (Cu: 0.58 µmol), 1a (0.4 mmol), CHCl3 (3.5 mL), MeOH (0.1 mL), 353 K, O2 (10 atm).

Cu2+-magadiite was also applicable to the oxidative coupling reaction of other phenol derivatives having electron-donating alkyl groups such as 1b and 1c to afford the corresponding diphenoquinones 2b and 2c in 97% yields for 24 h. As in the case of G4-Cu2+12, the oxidative coupling of phenol or 2,6-dichlorophenol did not occur using Cu2+-magadiite.

One of the significant advantages of inorganic solid supports over organic polymer supports is their high durability. After the oxidative coupling reaction of 1a, Cu2+-magadiite could be recovered by simple filtration and reused without any loss of activity or selectivity (Table 3, entry 2 vs. entries 3 and 4). Cu2+-magadiite catalyzed the selective coupling of 1a even under a higher temperature of 353 K to afford 2a in 95% yield with high TON of 655 (Table 3, entry 5). This TON value is the highest among previously reported catalysts. These results demonstrated that Cu2+-magadiite represents a durable heterogeneous catalyst for the oxidative coupling of DMP to DPQ. Furthermore, Cu2+-magadiite was applied to a continuous flow reactor system [53,54]. Cu2+-magadiite (0.8 g) was placed in a tubular stainless-steel reactor, and 2.44 g of DMP dissolved in CHCl3/EtOH (500 mL, 4/1 v/v) and O2 were passed through it, successfully giving 2.20 g of DPQ (92% isolated yield) (Figure 6). These experiments demonstrate that the development of heterogeneous catalysts for the selective oxidative C–C coupling of DMP offers a more efficient and practical synthetic method of DPQ from DMP.

Figure 6

A continuous flow reaction system of Cu2+-magadiite-catalyzed oxidative coupling of DMP to DPQ.

3. Discussion

In the case of G4-Cu2+n, the selectivity for the C–C coupling (Section 2.3.1., Table 2) was closely correlated with the structure of the Cu2+ species within G4-Cu2+n (Section 2.2.1.). G4-Cu2+n (n ≤ 8) having mononuclear Cu2+ species gave poor selectivity, whereas G4-Cu2+n (n ≥ 12) having adjacent Cu2+ species exhibited high selectivity. These results demonstrate that the adjacent Cu2+ species within the dendrimer are the active species for the selective C–C coupling of DMP. The higher activity of G4-Cu2+12 than those of G4-Cu2+n (n = 16 and 24) might be due to the presence of non-coordinated tertiary amino groups in G4-Cu2+12. The free amino groups in G4-Cu2+12 would act as base sites to promote the facile formation of Cu2+-phenolate species by trapping the accompanying HCl [55], thus delivering the superior activity of G4-Cu2+12. To investigate the accelerating effect of the free amino groups, G4-Cu2+12 (Cu: 5 μmol) treated with HCl(aq) (0.01 N, 0.25 mL) was used for the coupling reaction for 6 h. The conversion of 1a resulted in only 29% (Table 2, entry 7), indicating that the basic sites of free amino groups of the dendrimer play a key role in the efficient C–C coupling reaction. The high efficiency of G4-Cu2+12 is ascribed to the concerted catalysis between the adjacent Cu species and the free amino groups. Furthermore, Liu et al. reported that a dinuclear copper phenolate species is a key intermediate for the selective C–C coupling in the oxidative coupling of DMP to DPQ [18]. From the above facts, we propose the reaction mechanism of the G4-Cu2+12-catalyzed oxidative C–C coupling of 1a as shown in Figure 7.

Figure 7

Proposed reaction mechanism of efficient oxidative coupling of DMP to DPQ using G4-Cu2+12.

Initially, a ligand exchange reaction between a Cl species of Cu2+ and 1a occurs to give a Cu2+-phenolate anion species associated with the formation of HCl, which is trapped by a tertiary amino group of the dendrimer. The two Cu2+-phenolate anions within the nanovoids are suitably oriented for the C–C selective coupling reaction through the one-electron oxidation of the phenolate anion, giving 3a. The two generated Cu+ species are oxidized by O2 to regenerate the Cu2+ species. Successive oxidation of 3a yields the final product 2a.

The proposed reaction mechanism is well supported by the following control experiments. In situ UV-vis analysis was carried out during the reaction of equimolar amounts of G4-Cu2+12 with 1a under an Ar atmosphere for 24 h to afford 2a and 3a (Scheme S1) [47]. The d-d transition adsorption band derived from the Cu2+ species around 740 nm gradually decreased and finally disappeared at 24 h (Figure S4) [47]. These results support the one-electron transfer from 1a or 3a to the Cu2+ species, followed by the formation of the C–C coupling products. In addition, 3a is confirmed as a reaction intermediate of 2a by quantitative production of 2a from 3a [32].

The main roles of the regularly arranged tertiary amino groups of the nanovoid are (1) coordination to Cu ions to generate unique adjacent Cu species, (2) facile promotion of the ligand exchange of Cu-Cl with DMP by the base sites trapping of HCl, and (3) accumulation of both adjacent Cu species and basic sites within the confined nanovoids. The concerted catalysis between the adjacent Cu species and the base sites within G4-Cu2+12 enabled the efficient coupling of DMP to DPQ.

In the case of the magadiite-immobilized Cu catalysts, Cu2+-magadiite exhibited much higher activity and selectivity compared to Cu2+(mono)-magadiite (Section 2.3.2., Table 3, entries 1 and 8), showing that the active species of oxidative C–C coupling of DMP is the dinuclear Cu species generated in the layer nanospace of magadiite. Bridging OH− groups coordinating the adjacent Cu species might deprotonate DMP [40,48] to promote the formation of two Cu2+-phenolate species, giving the C–C coupling product through a similar pathway to that in G4-Cu2+12. To clarify whether DMP can access the active dinuclear Cu species in the interlayer, the nanospace of magadiite was estimated by XRD analysis. When Cu2+-magadiite was soaked into CHCl3/MeOH solvent, the basal spacing of Cu2+-magadiite increased from 13.9 to 14.8 Å, meaning that the interlayer gallery of Cu2+-magadiite expanded from 2.4 to 3.2 Å. This value is larger than the minimum thickness of DMP and DPQ (molecular size of DMP: ca. 1.8 × 6.7 × 7.0. Å; DPQ: ca. 1.8 × 6.7 × 9.3. Å) [56], confirming that DMP can react with dinuclear Cu species within the interlayer nanospace and the product DPQ can be removed. As described in Section 2.3.2., Cu2+-magadiite maintained its high activity and selectivity during the reuse experiments (Table 3, entries 3 and 4). The Cu content in the used catalyst was the same as in the fresh one evidenced by ICP-AES analysis (Cu: 1.39 wt %). The FT of the k3-weighted EXAFS spectrum and the curve-fitting result of used Cu2+-magadiite were similar to those of the fresh one (Figure 3 and Table 1). These data show that the dinuclear Cu species remained unchanged after the coupling reaction. The dinuclear Cu2+ species in the nanospace is stabilized through the electrostatic interaction between the Cu2+ species and the layered silicate anion of magadiite, providing high durability of Cu2+-magadiite. The prominent catalysis of Cu2+-magadiite could be ascribed to the dinuclear Cu species, which were generated and stabilized within the two-dimensional interlayer nanospace.

4. Experimental Section

4.1. Preparation of G4-Cu2+n

G4-NH2 was synthesized from the NH2-terminated third-generation PPI dendrimers (G3-NH2) by the divergent method [57]. Surface modification of G4-NH2 with 3,4,5-triethoxybenzamide groups was carried out using a previously reported procedure to afford G4-TEBA [26]. The mixture of CuCl2 (5 µmol) and G4-TEBA (0.42 µmol in the case of preparation of G4-Cu2+12) was stirred for 2 h in 1.5 mL of CHCl3/CH3CN (2:1 v/v) under an Ar atmosphere at 298 K [29]. The resulting solution was evaporated to give G4-Cu2+12 as a brown, waxy solid.

4.2. Oxidative Coupling of DMP Using G4-Cu2+n

1a (0.5 mmol) was added to 4 mL of a CHCl3 solution of G4-Cu2+n (Cu: 5 µmol) and the mixture was stirred under O2 (1 atm) at 333 K in a Schlenk flask. Following completion of the reaction, hexadecane was added as an internal standard and the mixture was treated with aqueous 6 N HCl, after which the organic phase was separated and characterized by 1H NMR.

4.3. Preparation of Cu2+-Magadiite

Magadiite was synthesized by hydrothermal reaction from SiO2 (Wakogel Q-69), NaOH, and deionized water according to the previously reported paper [33]. Next, Cu(ClO4)2·6H2O (100 µmol) and TMEDA (100 µmol) were dissolved in MeOH (10 mL) at 298 K, and magadiite (0.4 g) was added into the Cu-TMEDA solution. The resulting mixture was further stirred for 6 h at 313 K. After the reaction, the obtained solid was filtered, washed with MeOH (100 mL), and dried to afford a light blue powder.

4.4. Oxidative Coupling of DMP Using Cu2+-Magadiite

1a (0.5 mmol) and Cu2+-magadiite (0.08 g, Cu: 17.5 µmol) were placed in a Schlenk flask, and then 4 mL of CHCl3/MeOH (7:1 v/v) solvent was added. The reaction mixture was heated at 328 K for appropriate times under O2 (1 atm). After the reaction, hexadecane was added as an internal standard and the mixture was filtered and subjected to 1H NMR analysis.

4.5. Cu2+-Magadiite-Catalyzed Oxidative Coupling of DMP Using Continuous Flow Reactor

A stainless steel column (inner diameter: 5 mm; length: 150 mm) was filled with Cu2+-magadiite (0.8 g) and SiO2 (Wakogel C-400HG, 0.15 g). A solution of 1a in CHCl3/EtOH (2.44 g: 20 mmol in 500 mL of CHCl3/EtOH (v/v 4:1), 0.2 mL/min) and O2 (60 mL/min) were passed through the column at 353 K for 42 h. After the reaction was completed, the solvent was evaporated. The obtained supernatant was passed through a silica gel column (Wakogel C-200). Byproducts were first removed by elution with hexane/ethyl acetate (v/v = 3/2) and 2a was subsequently eluted with hexane/CH2Cl2 (v/v = 1/3). The evaporation of the solvent gave 2a in 92% yield (2.20 g).

5. Conclusions

We developed efficient Cu complex catalysts by using the nanospaces of PPI dendrimers or magadiite for the selective oxidative C–C coupling of DMP to DPQ. The PPI dendrimer accommodated Cu ions into the internal nanovoids to form adjacent Cu species, which promoted the highly efficient regioselective coupling reaction. We also devised an immobilized dinuclear Cu complex catalyst by incorporation of Cu complexes into the interlayer nanospace of magadiite, and successfully demonstrated the selective oxidative coupling reaction of DMP to DPQ. This is the first example of a selective heterogeneous catalyst for the oxidative C–C coupling of DMP to DPQ. The magadiite-immobilized Cu catalyst has significant advantages such as (1) recoverability of catalyst, (2) catalyst reusability, (3) high durability, and (4) applicability to a flow reactor system, providing the highly practical and green synthesis of DPQ from DMP and O2. These strong catalytic performances of PPI dendrimer-encapsulated and magadiite-immobilized Cu complex catalysts for the selective oxidative C–C coupling of DMP to DPQ are ascribed to the adjacent Cu species generated within the nanospaces of the PPI dendrimer and magadiite. We believe that the regulated nanospaces of structurally ordered materials would serve as a platform for preparation of multinuclear metal species to provide strong catalytic performance.