|Home | About | Journals | Submit | Contact Us | Français|
Benzene was hydroxylated with hydrogen peroxide (H2O2) in the presence of catalytic amounts of copper complexes in acetone to yield phenol at 298 K. At higher temperature, phenol was further hydroxylated with H2O2 by catalysis of copper complexes to yield p-benzoquinone. The kinetic study revealed that the rate was proportional to concentrations of benzene and H2O2, but to the square root of concentration of a copper(II) complex ([Cu(tmpa)]2+: tmpa = tris(2-pyridylmethyl)amine). The addition of a spin trapping reagent resulted in formation of a spin adduct of hydroperoxyl radical (HO2•), as observed by EPR spectroscopy, inhibiting phenol formation. HO2• produced by the reaction of [Cu(tmpa)]2+ with H2O2 acts as a chain carrier for the radical chain reactions for formation of phenol. When [Cu(tmpa)]2+ was incorporated into mesoporous silica-alumina (Al-MCM-41) by a cation exchange reaction, the selectivity to production of phenol was much enhanced by prevention of hydroxylation of phenol, which was not adsorbed to Al-MCM-41. The high durability with turnover number of 4320 for the hydroxylation of benzene to phenol with H2O2 was achieved using [Cu(tmpa)]2+ incorporated into Al-MCM-41 as an efficient and selective catalyst.
Phenol is one of the most important basic chemicals, widely used as sources for resin, fibres, and other various organic materials, being produced in industry by the three-step cumene process, which requires high energy consumption due to high-pressure conditions and produces acetone as a coproduct via cumene hydroperoxide.1,2 Thus, much effort has been devoted to develop the one-step hydroxylation of benzene to phenol using different oxidants, such as N2O,3 O2 with reductants4–7 and H2O2.8–22 Among these oxidants, H2O2 has a clear advantage from the viewpoint of an environmentally benign green process and economical efficiency because of relatively mild reaction conditions for catalytic hydroxylation of benzene.8–23 However, selective hydroxylation of benzene to phenol with H2O2 has been difficult because phenol is normally easier to be oxidised to yield over-oxidised products such as p-benzoquinone.8–21 Although selective hydroxylation of benzene to phenol without over-oxidation has been achieved in some cases, the catalytic mechanism of hydroxylation of benzene to phenol has yet to be clarified,24–30 except for some cases of photocatalytic hydroxylation of benzene to phenol.28–31 Extensive efforts have been devoted to enhance the product selectivity and catalytic activity with heterogeneous catalysts including metal complex catalysts incorporated into a mesoporous material. For instance, it has been reported that a manganese complex incorporated into a mesoporous material Al-MCM-41 exhibits a higher performance compared to that of the complex in a homogeneous solution because of the stabilisation of catalytically active species and prevention of phenol over-oxidation due to the highly acidic nature in Al-MCM-41.26
We report herein the one-step selective hydroxylation of benzene to phenol with H2O2 catalysed by a copper(II) complex ([Cu(tmpa)]2+: tmpa = tris(2-pyridylmethyl)amine) incorporated into mesoporous silica-alumina in acetone at 298 K. The mechanism is clarified by a kinetic study and by detection of a reactive radical intermediate using a spin trap. The incorporation of the [Cu(tmpa)]2+ into mesoporous silica-alumina resulted in significant improvement of the selectivity and the durability of the catalyst.
A series of Cu(II) complexes, [CuII(tmpa)(CH3CN)](ClO4)2, [CuII2(N5)(H2O)2](NO3)4 (N5 = − (CH2)5− linked bis[2-(2-pyridyl)ethyl]amine) and [CuII(tepa)(ClO4)]ClO4 (tepa = tris(2-pyridylethyl)amine), were all found to catalyse hydroxylation of benzene with H2O2 to produce phenol (Figure 1).32 The time profiles for these reactions using a catalytic amount of Cu(II) complexes in acetone at 298 K are shown in Figure 2. The most reactive Cu(II) complex was [CuII(tmpa)(CH3CN)](ClO4)2, which has the most negative one-electron reduction potential (Ered = −0.01 V vs. SCE)29 as compared with those of [CuII2(N5)(H2O)2](NO3)4 (Ered vs. SCE = 0.27 V) and [CuII(tepa)(ClO4)]ClO4 (Ered vs. SCE = 0.44 V)31 as shown in Figure S1 in Electronic Supplementary Information (ESI).17 It should be noted that CuII(ClO4)2 without any ligand exhibited virtually no catalytic activity. Thus, the tmpa ligand plays an essential role in the catalytic activity.
When acetone was replaced by strongly coordinating solvents such as DMF and DMSO, no catalytic hydroxylation of benzene with H2O2 occurred using [CuII(tmpa)(CH3CN)](ClO4) (1) as a catalyst (see Figure S2 in ESI). Thus, a coordination site is required to exhibit the catalytic activity. In acetonitrile, 1-catalysed hydroxylation of benzene with H2O2 occurred to produce phenol. However, p-benzoquinone was also formed as a byproduct, which may be produced via oxidation of hydroquinone. In MeOH, the catalytic rate of hydroxylation of benzene with H2O2 was slower than those in acetone and acetonitrile. Thus, we have chosen acetone as a suitable solvent for mechanistic studies.
The products in the reaction of benzene with H2O2 and a Cu(II) complex catalyst 1 were examined by GC-MS measurements. When a large concentration of benzene (2.1 M) was used for the 1-catalysed reaction of benzene with H2O2, phenol was selectively produced without formation of p-benzoquinone at the initial stage of the reaction (black circle in Figure 3). When benzene was replaced by phenol, however, p-benzoquinone was produced with a comparable rate as the case for phenol production from benzene (red cross in Figure 3).
The stoichiometry of the catalytic hydroxylation of benzene with H2O2 is given by eqn (1). The amount of reacted H2O2 was determined by the titration with Ti–TPyP (oxo[5,10,15,20-tetra(4-pyridyl)porphyrinato]titanium(IV);27 see Experimental section and Figure S3 in ESI). The ratio of the reacted H2O2 to phenol was determined to be nearly 3:1, indicating that H2O2 decomposed in competition with the catalytic hydroxylation of benzene to phenol.
The dependencies of the initial rate of phenol production on the concentrations of 1, H2O2 and benzene were examined to determine a kinetic formulation (Figure 4). The initial rates of phenol production were determined from the time profiles of production of phenol in the catalytic hydroxylation of benzene with H2O2 (Figures S4a–c in ESI). The initial rate of phenol production was proportional to the concentrations of H2O2 and benzene as shown in Figure 4a and Figure 4b, respectively. The dependence of the initial rate of phenol production on concentration of 1 was not linear (Figure 4c), but instead was proportional to square root of the concentration of 1. Thus, the kinetic equation for the catalytic hydroxylation of benzene with H2O2 is given by eqn (2),
where kcat is the catalytic rate constant. The dependence of the rate on square root of concentration of 1 usually suggests that the catalyst may dissociate into two species, which are in equilibrium with 1, and that the concentration of the two catalytically active species is proportional to square root of concentration of 1 when the equilibrium lies to the far left-hand side. However, the Cu(II) complex (1) is a monomer complex which has no way to dissociate into two species. The unusual kinetic formation in eqn (1) in terms of the dependence of the rate on 1 provides valuable insight into the catalytic mechanism as discussed later.
The trapping of a reactive free radical species with a diamagnetic spin trap to generate a persistent spin adduct which can be characterised by its EPR spectrum constitutes the well known spin trapping technique.28,33–46 5,5-Dimethyl-1-pyrroline N-oxide (DMPO in Figure 5a) is one of the most popular spin traps, and EPR features of its spin adducts are well established. For example, the EPR spectrum of the HO• radial adduct (DMPO–OH) exhibits signals with a relative peak ratio of 1:2:2:1 with the hyperfine coupling constants of aN = aH = 14.9 G, whereas the EPR spectrum of the HO2• radical adduct affords more lines with hyperfine coupling constant of aN (13–14 G) which is significantly larger than aH (9–11 G).28,37–41,47 Carbon-centred radical adducts afford a larger aH value than the aN value.28,42 Thus, the type of radical species can be readily distinguished based on the hyperfine splitting pattern of the DMPO spin adducts.
Figure 5 shows an EPR spectrum observed during the reaction of benzene with H2O2 in the presence of 1 and DMPO. The hyperfine splitting pattern of the observed radical clearly indicates that HO2• is produced in the reaction of benzene with H2O2 and 1 is trapped by DMPO to produce DMPO–OOH, which has the g value of 2.0060 and aN = 13.1 G and aH = 10.8 G as shown in Figure 5, where the observed spectrum (part a) agrees with the computer simulation spectrum (part b). It was confirmed that the addition of DMPO to an acetone solution of benzene and H2O2 resulted in no observation of an EPR signal. In summary, when 1 was added to the acetone solution of benzene, H2O2 and DMPO, a strong EPR signal assignable to DMPO–OOH was observed.
The concentration of DMPO–OOH was determined by comparing the double integration value of the EPR signal of DMPO–OOH with that of a stable reference radical (2,2-diphenyl-1-picrylhydrazyl radical, see Experimental Section). The concentration of DMPO–OOH increased with the reaction time as shown in Figure S5 (ESI). More importantly, the addition of a catalytic amount of DMPO to the reaction solution of benzene with H2O2 and 1 resulted in nearly complete inhibition of the reaction as shown in Figure 6. This indicates that HO2• produced in the reaction of 1 with H2O2 acts as a chain carrier in the catalytic hydroxylation of benzene, which proceeds via radical chain reactions (vide infra). The chain length of the propagation step in Scheme 1 is evaluated as the ratio of the rate of formation of phenol (1.08 × 10−7 M s−1 determined at 5 h) to twice of DMPO–OOH (9.62 × 10−11 M s−1 determined at 6 h), which is 1120, because two HO2• radicals are produced in the initiation step (see Scheme 1).
A proposed radical chain mechanism48–50 of the 1-catalysed hydroxylation of benzene with H2O2, which agrees with the kinetic formulation [eqn (2)], is shown in Scheme 1. The catalytic reaction is started by the Fenton-like reaction,43,44,51 in which H2O2 is reduced by [CuII(tmpa)]2+ to produce HO• radical and the hydroxide adduct ([CuIII(OH)(tmpa)]2+). In competition with the fast back reaction between HO• and [CuIII(OH)(tmpa)]2+, HO• reacts with H2O2 to produce HO2• and [CuIII(OH)(tmpa)]2+ also reacts with H2O2 to produce HO2•, accompanied by regeneration of [CuII(tmpa)]2+. HO2• reacts with benzene to produce the HO2• adduct of benzene as the rate-determining step (RDS). The subsequent fast hydrogen abstraction from H2O2 by the HO2• adduct produces phenol and H2O, accompanied by regeneration of HO2•, thus constituting the radical chain reactions. The termination step is the disproportionation reaction of two HO2• to produce H2O2 and O2. Once HO2• is trapped by DMPO to produce DMPO–OOH, the radical chain is stopped to inhibit the benzene hydroxylation.
reaction of HO2• with benzene. The rate of formation and decay of HO2• is given by eqn (4), where k2 is the rate constant
of the reaction of the caged radical pair ([CuIII(OH)(tmpa)]2+ HO•) with H2O2 to produce HO2• and kt is the rate constant of disproportionation of HO2•. The rate of decay of HO2• by the reaction with benzene is the same as the formation rate of HO2• by the reaction of HO2• adduct of benzene with H2O2. The rate of formation and decay of the caged radical pair ([CuIII(OH)(tmpa)]2+ HO•) is given by eqn (5), where k1 is the
rate constant of the Fenton-like reduction of H2O2 by 1 and k−1 is the rate constant of the back reaction. Under the steady-state conditions (d[HO2•]/dt = 0), [HO2•] is given from eqn (4) by eqn (6), whereas [([CuIII(OH)(tmpa)]2+ HO•)] is given by eqn
formulation is derived as given by eqn (9), which agrees with
the experimental observations (eqn (2)). Such agreement including the unusual dependence of the rate on  strongly supports the validity of Scheme 1. The feasibility of the chain reactions in Scheme 1 is also supported by DFT calculations (see Figure S6, Table S1 and S2 in ESI).52 The rate-determining propagation step of the reaction of HO2• with benzene to produce the HO2• adduct of benzene is calculated to be uphill by 11.1 kcal mol−1, whereas the reaction of the HO2• adduct of benzene with HO2• to produce phenol and H2O is downhill by 72.3 kcal mol−1. The formation of the intermediate prior to phenol and H2O is uphill by 5.6 kcal mol−1. Thus, the intermediate is given in parenthesis in Scheme 1.
Mesoporous silica and silica-alumina have often been used to improve the catalytic performance of metal complexes in a role as catalyst supports whose pores provide broad reaction fields with efficient incorporation.14–15,21,26,53–57 To improve the reaction selectivity of the benzene hydroxylation by H2O2, the catalyst 1 was incorporated into a mesoporous silica-alumina Al-MCM-4144,53a,54–61 (the Brunauer-Emmett-Teller (BET) surface are: 1200 m2 g−1, the external surface area: 32 m2 g−1, the pore diameter: 3.5 nm) to prepare [CuII(tmpa)]2+ @Al-MCM-41 (see Experimental section).26,53a,62 [CuII(tmpa)]2+ @Al-MCM-41 was characterised by a UV-vis diffuse reflectance spectrum (Figure S7 in ESI) and the amount of incorporated 1 was determined to be 2.50 × 10−5 mol g−1 by the absorption spectral change of the mother liquid according to the literature (Figure S8 in ESI).53a The much larger BET surface area compared to the external surface suggests that most of the incorporated [Cu(tmpa)]2+ is placed inside the mesopore which is large enough (the diameter: 3.5 nm) for incorporation of [Cu(tmpa)]2+ species smaller than 1.2 nm.
When [CuII(tmpa)]2+ @Al-MCM-41 was used as a catalyst for the hydroxylation of benzene with H2O2, benzene was oxidised to phenol as the case of the homogeneous system (Figure 7a).63 Somewhat slower phenol formation with [CuII(tmpa)]2+ @Al-MCM-41 compared to that with [CuII(tmpa)]2+ can be explained by the smaller diffusion rate in the pore. The unchanged blue colour of [CuII(tmpa)]2+ @Al-MCM-41 after the reaction suggests that [CuII(tmpa)]2+ cations are encapsulated in Al-MCM-41 without major leaching. The selectivity to produce phenol was significantly improved as p-benzoquinone was hardly produced from phenol, as shown in Figure 7b. It was confirmed that no phenol was produced using Al-MCM-41 without 1, indicating that the reaction was catalysed by [CuII(tmpa)]2+ incorporated into Al-MCM-41 (Figure 7a). It was found that benzene does adsorb to Al-MCM-41, whereas phenol was hardly adsorbed (Figures S9 and S10 in the ESI). The difference in the adsorption behaviours can be explained by the solvophobic surface present in Al-MCM-41 compared to the environment in solution.64 This selective adsorption of benzene and lack of adsorption of phenol result in the selective hydroxylation of benzene with H2O2 catalysed by the [CuII(tmpa)]2+ site in Al-MCM-41 to produce only phenol, which is desorbed from Al-MCM-41. The solution was not coloured after the reaction, suggesting that leaching of Cu2+ species is not the major reason for the saturation of phenol formation. This result shows the potential for recyclability of [CuII(tmpa)]2+ @Al-MCM-41 after recovery and washing with acetonitrile.
The reaction of 1 mM benzene with 2.1 M H2O2 and 200 µM of [CuII(tmpa)]2+ @Al-MCM-41 resulted in formation of phenol in 17% yield with 100% selectivity after 118 h (Figure S11 in ESI). The durability of the catalyst [CuII(tmpa)]2+ @Al-MCM-41 was also examined by the reaction of 2.1 M benzene with 2.1M H2O2 and 1 µM of the catalyst. The turnover number was determined to be 4320 after 118 h, demonstrating a high durability for this catalyst (Figure 8).
In conclusion, one-step hydroxylation of benzene with H2O2 to produce phenol with a high turnover number (TON = 4320) and high selectivity (>99%) has been achieved by using a copper(II) complex ([Cu(tmpa)]2+) incorporated into mesoporous silica-alumina (Al-MCM-41) in acetone at 298 K. The catalytic hydroxylation of benzene with H2O2 proceeds via a radical chain mechanism in which HO2• produced via the Fenton-like reduction of H2O2 with [Cu(tmpa)]2+ acts as the chain carrier radical. Spin trapping of the chain carrier radical resulted in the complete inhibition of the catalytic benzene hydroxylation. The present study on the selective catalytic hydroxylation of benzene to phenol using metal complex-incorporated mesoporous silica-alumina as a catalyst provides a general strategy that may be applied to other selective catalytic reactions.
Chemicals were purchased from commercial sources and used without further purification, unless otherwise noted. Tetraethyl orthosilicate, cetyltrimethylammonium bromide, sodium aluminate, sodium hydroxide, an aqueous solution of hydrogen peroxide (30 wt%) and perchloric acid (60%) were purchased from Wako Pure Chemical Industries. Tris(2-pyridylmethyl)amine and oxo[5,10,15,20-tetra(4-pyridyl)porphyrinato]titanium(IV) (Ti–TPyP) were supplied from Tokyo Chemical Industry. Copper(II) perchlorate hexahydrate were delivered by Sigma-Aldrich. Acetone and acetonitrile were purchased from Nacalai tesque as a spectral grade and used as received. Ultra-Pure water was provided by a water purification system, Millipore Direct-Q3 UV, where the electronic conductance was 18.2 MΩ cm. Benzene was purchased from Wako Pure Chemical Industries and purified by washing with sulphuric acid and water and following distillation. [Cu(tmpa)(CH3CN)](ClO4)2, [Cu(tepa)(ClO4)]ClO4, [CuII2(N5)(H2O)2](NO3)4 and Al-MCM-41 were synthesised by literature methods.53a,58–61
Al-MCM-41 was synthesised by a reported method.61 Cetyltrimethylammonium bromide (18.3 g, 50.4 mmol) was dissolved in an aqueous solution (1.0 l) of NaOH (8.39 g, 210 mmol). Tetraethylorthosilicate (87.5 g, 94.0 mL, 420 mmol) was added dropwise to the aqueous solution for 3 h at 35–40°C with stirring in water bath and further stirred for 30 min. Then, sodium aluminate (0.484 g, 5.90 mmol) was added to the reaction vessel and stirred for 4 h at 298 K. The white precipitate obtained was filtered under reduced pressure and washed with distilled water. After drying at 333 K in an oven for 10 h, the solid obtained was calcined at 873 K for 6 h with 1 K min−1 of a rate of temperature increasing. The Si/Al ratio of this solid is 60,65 which was calculated from the amount of the precursor materials. The Brunauer-Emmett-Teller (BET) surface area and the external surface area of Al-MCM-41 were determined to be 1200 m2 g−1 and 32 m2 g−1, respectively, by the N2 isotherm and the t-plot. The pore diameter was also determined to be 3.5 nm by the powder XRD pattern. Al-MCM-41 with the ratio of Si/Al= 20 was prepared by the same method with the corresponding amount of sodium aluminate.
[CuII(tmpa)]2+ @Al-MCM-41 was prepared by a cation exchange method in an acetonitrile solution according to the literature method.26 Al-MCM-41 (300 mg) was suspended in an acetonitrile solution (15 mL) of [CuII(tmpa)(CH3CN)](ClO4)2 (16.5 mg, 27.8 µmol, [[CuII(tmpa)]2+]: 1.85 mM) and stirred for 2 h at 298 K. The suspension was centrifuged to collect solid, which was washed with acetonitrile (5.0 mL × 2) and dried in vacuo. Incorporation of [CuII(tmpa)]2+ was confirmed by the diffuse reflectance spectrum of the solid obtained resembling the absorption spectrum of [CuII(tmpa)]2+ in acetonitrile (Figure S7),26,53a which could also be seen by clear colour change of the solid from white to blue. The amount of incorporated [CuII(tmpa)]2+ was determined to be 2.50 × 10−5 mol g−1 by the decrease in the UV-vis absorption band around 860 nm due to [CuII(tmpa)]2+ in the mother liquid according to the literature (Figure S8).53a
A typical procedure for catalytic hydroxylation is as follows: 4.75 mL of an acetone solution containing benzene (2.1 M) and 30 wt% aqueous H2O2 (2.1 M) was added to [CuII(tmpa)(CH3CN)](ClO4)2 (0.31 µmol) and vigorously stirred at 298 K. The sample solutions for GC-MS measurements were prepared from the reaction solution by dilution with acetonitrile as needed to analyse products. The Al-MCM-41 and [CuII(tmpa)]2+ @Al-MCM-41 were removed by filtration before the measurement. GC-MS measurements were performed with a Shimadzu QP-2010 Ultra instrument.
Cyclic voltammetry (CV) measurements were performed with an ALS630B electrochemical analyser in deaerated acetonitrile containing 0.1 M Bu4NPF6 (TBAPF6) as a supporting electrolyte at 298 K. The platinum working electrode (BAS, surface i.d. 1.6 mm) was polished with BAS polishing alumina suspension and rinsed with milli-Q ultra-pure water, acetone and acetonitrile before use. The counter electrode was a platinum wire (0.5 mm dia.). The measured potentials were recorded with respect to an Ag/AgNO3 (0.01 M) reference electrode. The values of redox potentials (vs. Ag/AgNO3) are converted into those vs. SCE by addition of 0.29 V.66
The concentration of hydrogen peroxide was determined using the Ti–TPyP reagent according to the literature as follows27: aqueous solutions of 50 µM Ti–TPyP with 50 mM hydrochloric acid and 5 M perchloric acid were prepared, respectively. Sample solutions were prepared by 5000 times dilution of 50 µl of the reaction mixture (the concentrations are [C6H6] = 2.1 M, [H2O2] = 0.6 M and [[CuII(tmpa)]2+] = 200 µM in acetone (4.75 mL)) with acetonitrile. The sample solution (50 µl) was added to the mixed solution of Ti–TPyP (250 µl) and perchloric acid (250 µl), shaken and diluted with H2O (1950 µl). The concentration of hydrogen peroxide was calculated from the absorbance of the resulting peroxo complex at 435 nm observed by UV-vis measurements (Hewlett-Packard 8453 diode array spectrophotometer).
A typical procedure for a spin trap experiment is as follows: 4.75 mL of an acetone solution containing benzene (2.1 M), 30 wt% aqueous H2O2 (0.1 M), 1 (0.31 µmol, 67 µM), and DMPO (0.2 M) was vigorously stirred at 298 K. The sample for EPR was prepared by Ar bubbling (15 min) with a Teflon tube of 200 µl of the reaction solution in a quartz EPR tube (2.0 mm i.d.). EPR spectra of solutions were taken on a JEOL X-band spectrometer (JES-RE1XE) at 298 K. The g value was calibrated using an Mn2+ marker. The EPR spectra were recorded under non-saturating microwave power conditions. The magnitude of modulation was chosen to optimise the resolution and the signal-to-noise (S/N) ratio of the observed spectra. The double integrated first derivative of a stable radical 2,2-diphenyl-1-picrylhydrazyl radical (10 µM) was obtained from a EPR spectrum in a mixed solvent benzene/acetone/H2O (1:1:5, v/v/v) similar to the reaction conditions as a reference. The spin concentrations of DMPO–OOH signals in the reaction solutions observed by EPR were calculated by comparing with the value obtained from DPPH. The GC-MS (Shimadzu QP-2010 Ultra) measurements were performed as needed to examine the relation between the spin trap and the hydroxylation.
UV-vis absorption spectra were recorded on a Hewlett-Packard 8453 diode array spectrophotometer for UV-vis at 298 K. UV-vis diffuse reflectance spectra were recorded on a JASCO V-670 spectrophotometer equipped with an SIN-768 attachment.
Density functional theory (DFT) calculations were performed on a 32CPU workstation (PQS, Quantum Cube QS8-2400C-064). Geometry optimisations were carried out using the B3LYP/6-31++G(d) level of theory for compounds and intermediates relating to the reaction as implemented in the Gaussian 09 program Revision A.02.52
This work was supported by an Advanced Low Carbon Technology Research and Development (ALCA) and SENTAN programs from Japan Science Technology Agency (JST) (to S.F.) of the Ministry of Education, Culture, Sports, Science and Technology, Japan. Support from the USA National Institutes of Health (NIH) for K.D.K. is also acknowledged.
†Electronic Supplementary Informa)on (ESI) available: Cyclic voltammogram (Figure S1), time profiles of phenol, p-benzoquinone, H2O2, spin or benzene (Figures S2 – S5 and S9 – S10), DFT results (Figure S6 and Tables S1 – S2), UV-vis absorption spectra (Figure S7) and UV-vis DRS (Figure S8). See DOI: 10.1039/x0xx00000x