The catalytic oxidation of olefins in the presence of molecular O₂ is highly attractive from economic and environmental viewpoints. Particularly, the partial oxidation of olefins to epoxides and aldehydes is an industrially important reaction because epoxides and aldehydes are widely used as the precursors and intermediates for the syntheses of various fine chemicals and useful chemical products in the pharmaceutical, resin, and paint industries ^{[1, 2, 3, 4]}.

In principle, the direct oxidation of olefins with O₂ is highly unfavorable because the ground state of O₂ is a triplet, whereas the ground states of olefins are singlets. Accordingly, the oxidation of olefins with O₂ has generally been performed in the co-presence of reductants such as H₂, alcohols, and aldehydes. However, Tang et al. ^{[5, 6]} first demonstrated that Co²⁺-exchanged faujasite-type zeolites could effectively catalyze the oxidation of styrene with O₂ in the absence of reductants. Consequently, subsequent studies focused on the catalytic activities of various Co-exchanged and Co-containing microporous and mesoporous materials such as alkali and alkaline metal-modified Co-X ^[7], Co-Y ^[8], Co-SSZ-51 ^[9], Co-SBA-15 ^[10], Co-ZSM-5 ^{[11, 12, 13]}, Co-SAPO-5 and Co-SAPO-34 ^[14], Co-MOR, Co-5A ^[15], and Co-coordinated organic-inorganic hybrid catalysts ^[16] toward the oxidation of olefins with O₂ or air. Among them, Co-X, Co-Y, and Co-SBA-15 ^{[5, 6, 7, 8, 10]} were effective in the epoxidation of olefins with molecular O₂. Co-SSZ-51 was also active, but required higher O₂ pressures (> 4.14 bar) ^[9]. Conversely, other catalysts (Co-ZSM-5, Co-SAPO-5, Co-SAPO-34, Co-MOR, Co-5A, and Co-coordinated organic-inorganic hybrid material) required substantial amounts of initiators, such as tert-butyl hydroperoxide or cumyl hydroperoxide, to achieve high conversions ^{[11, 12, 13, 14, 15, 16]}. Thus, the development of more active catalysts that do not require initiators in the target catalytic process is of interest.

ETS-10 is a unique titanosilicate that constitutes regularly spaced TiO₃²⁻ quantum wires with diameters (d) of ~0.67 nm (Fig. 1(a)). Furthermore, the structure comprises corner- sharing SiO₄ tetrahedra and TiO₆ octahedra linked via bridging oxygen atoms ^{[17, 18, 19]}. Formation of the structure can be described as an intergrowth of two hypothetical ordered Polymorph A (not shown) and Polymorph B (Figs. 1(b),(c)). Polymorph A has identical projection along the (100) and (010) directions to that of Polymorph B along the (110) and (11(_)0) directions. In Polymorph B, the TiO₃²⁻ quantum wires run along the (110) and (11(_)0) directions in the crystal (Figs. 1(b),(c)). Each TiO₃²⁻ quantum wire is surrounded by nanoporous silica (Fig. 1(d)), with a pore dimension of 8 × 5 Ų. The pore structure of ETS-10 contains 12 rings in all three dimensions. Only a few microporous zeolites with a three-dimensional 12-ring pore system are known, and in this aspect ETS-10 has excellent diffusion characteristics. It is important to point out that the structure is inherently an intergrowth structure, consisting of randomly stacked layers. The basic unit is Si₄₀Ti₈O₁₀₄¹⁶⁻, which is counterbalanced by 16 monovalent (Na⁺ and K⁺) cations. These counter cations can be exchanged with various other monovalent, divalent, or trivalent cations through ion exchange. AM-6 is isostructural to ETS-10, featuring VO₃²⁻ quantum wires instead of TiO₃²⁻ quantum wires ^{[20, 21]}.

Fig. 1. (a) Schematic illustration of the structure of ETS-10 or AM-6 imbedded with a three dimensional networks of SiO2 channels (blue) and TiO32- or VO32- molecular wires (red) in case of polymorph B. (b) View along the (110) axis. (c) View along the [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] axis. (d) A single TiO32- or VO32- molecular wire (pink) surrounded by silica (gold) and a single TiO32- or VO32- molecular wire.

To date, only Co-exchanged aluminosilicates have been examined as olefin oxidation catalysts. Therefore, expanding this research area by examining the catalytic activities of titanosilicate (ETS-10), vanadosilicate (AM-6), and their Co-exchanged forms is of interest to explore the role of the titanate and vanadate quantum wires within the structures. Herein, we report the catalytic performances of ETS-10, AM-6, and their Co²⁺-exchanged forms featuring varying degrees of Co²⁺ exchange toward the oxidation of styrene with O₂ to styrene epoxide (E) and benzaldehyde (B).

Na₂SiO₃ (17%-19% Na₂O, and 35%-38% SiO₂, Kanto), titanium isopropoxide (98%, Junsei), V₂O₅ (99%, Aldrich), H₂SO₄ (95%, Duksan), NaOH, (93%-100%, Samchun), KF (95%, Samchun), Co(CH₃COO)₂·4H₂O (Junsei), styrene (99%, Sigma-Aldrich), styrene oxide (97%, Aldrich), benzaldehyde (99%, Sigma-Aldrich), ethanol (95%, SK), and N,N- dimethylformamide (DMF, 99%, Samchun) were purchased and used without further purification.

A Si source solution was first prepared by dissolving Na₂SiO₃ (18.4 g) in H₂O (60 g), to which NaOH aqueous solution (2.4 g NaOH and 20 g H₂O) was added with vigorous stirring, and the mixture was stirred for 2 h. For the preparation of the Ti source solution, titanium isopropoxide (5.7 g), H₂SO₄ (4.5 g), and H₂O (35 g) were mixed together and boiled at 100 °C for 90 min, and allowed to cool to room temperature. The Ti source solution was added dropwise to the Si source solution, and the mixture was stirred for 1 h. A dilute KF solution (1.2 g KF and 15 g H₂O) was added to the above mixture. The mixture was aged for 16 h at room temperature and transferred to a Teflon-lined autoclave, and heated at 200 °C for 22 h under static conditions. After cooling the autoclave to room temperature, the crystals were collected by centrifugation and washed with copious amounts of distilled deionized water.

To prepare the Si source solution, a NaOH solution (3 g NaOH and 20 g H₂O) was added to the sodium silicate solution composed of Na₂SiO₃ (12.2 g) and H₂O (50 g).

To prepare the V source solution, a required amount of H₂SO₄ (4.5 g) was added to a 50-mL round-bottom flask containing H₂O (10 g). Subsequently, V₂O₅ (1.4 g) and ethanol (4 g) were sequentially added to the flask. The heterogeneous mixture was refluxed for 25 min.

Subsequently, the greenish-yellow V source solution was added dropwise to the Si source solution. A dilute KF solution (2 g KF and 10 g H₂O) was added to the above mixture. After aging the mixture for 15 h at room temperature, the resulting gel was transferred to a 50-mL Teflon-lined autoclave, which was placed in a preheated oven at 220 °C for 16 h under static conditions. The precipitated crystals were collected, washed, and dried at 100 °C for 1 h.

Co²⁺-exchanged ETS-10 powders with varying ion exchange degrees of 0, 9%, 26%, 68%, and 81% were prepared. Pristine ETS-10 is denoted as Co-E10-0. For the preparation of Co-E10-9, Co-E10-26, and Co-E10-68, pristine ETS-10 (1 g) was mixed with Co(CH₃CO₂)₂ solution(50 mL) at varying concentrations of 4, 10, and 50 mmol/L, respectively. The ion exchange was conducted at 50 °C for 1 h. For the preparation of Co-E10-81, the ion-exchange was conducted thrice with 100 mL of 50 mmol/L Co(CH₃CO₂)₂ solution. All samples were washed with copious amounts of water and dried in an oven at 90 °C for 2 h.

Co²⁺-exchanged AM-6 with varying ion exchange degrees of 0, 8%, 23%, 63%, and 79% were prepared. Pristine AM-6 is denoted as Co-A6-0. For the preparation of Co-A6-8, Co-A6-23, and Co-A6-63, pristine AM-6 (1 g) was mixed with Co(CH₃CO₂)₂ solution (50 mL) at varying concentrations of 4, 10, and 50 mmol/L, respectively. The ion exchange was conducted at 50 °C for 1 h. For the preparation of Co-A6-79, the ion-exchange was conducted thrice with 100 mL Co(CH₃CO₂)₂ solution (50 mmol/L). All samples were washed with copious amounts of water and dried in an oven at 90 °C for 2 h.

Scanning electron microscopy (SEM) images were obtained using a field-emission scanning electron microscope (Hitachi S-4300) operating at an acceleration voltage of 20 kV. Powder X-ray diffraction (XRD) patterns were recorded on a Rigaku D/MAX-2500/pc diffractometer using Cu K_α (λ = 0.154056 nm) radiation operating at 40 kV and 200 mA. The diffractograms were recorded in the 2θ range of 3°-50° with a step width of 0.02° and 1°/min in continuous mode. The BET surface areas of the samples were determined using N₂ adsorption-desorption performed at −196 °C on a BELSORP-MAX analyzer. The samples were evacuated at 200 °C for 10 h prior to analysis. Raman spectra of the samples were recorded on an in-house-built setup equipped with Ar⁺ ion laser (Spectra-Physics Stabilite 2017) as an excitation beam source, a spectrometer (Horiba JobinYvon TRIAX 550), and a CCD detector (Horiba JobinYvon Symphony) cooled at −196 °C. The wavelength of the excitation beam was 514.5 nm. The diffuse reflectance UV-Vis spectra of the samples were recorded on a Varian Cary 5000 UV-Vis-NIR spectrophotometer equipped with an integrating sphere. Barium sulfate was used as the reference. The diffuse reflectance spectra were converted using the Kubelka-Munk (K-M) function. Elemental analyses of the samples for Na⁺, K⁺, Co²⁺, Si, Ti, and V were conducted via energy-dispersive X-ray (EDX) spectroscopy analysis of the samples using an X-Max EDS-detector from Oxford instruments attached to a scanning electron microscope (JEOL-JSM-7600F).

Co-E10-n and Co-A6-m powders were vacuum dried at 150 °C for 10 h. The dehydrated powders were transferred to a glove box charged with high-purity Ar. To examine the catalytic activity of the powders, each dehydrated powder sample (100 mg) was transferred to a round-bottom flask. The reactions were carried out in a 25-mL round-bottom flask equipped with a water-cooled condenser and an oxygen inlet to introduce oxygen. The reaction mixture was continuously stirred using a magnetic stirrer. The temperature of the reactor was maintained at 95 °C. In a typical reaction, the dehydrated catalyst powder (100 mg, 0.22 mmol) was first dispersed in DMF (20 mL) by sonication. Subsequently, styrene (500 mg, 5 mmol) was introduced into the reaction system. The reaction was carried out at 95 °C with continuous bubbling of O₂ (4 mL/min) into the reactor. After 4 h of reaction, the catalyst was removed from the reaction mixture by centrifugation and the products in the supernatant solution were analyzed by gas chromatography mass spectrometry. A reaction temperature of 95 °C and a reaction time of 4 h were chosen because conversion saturation could be obtained under these reaction conditions. The conversion is based on the amount of styrene consumed in the reaction mixture.

The crystallinity of the Co-E10-n and Co-A6-m powders was analyzed by XRD, and the patterns are shown in Fig. 2. All samples displayed characteristic diffraction patterns of ETS-10-type phase. In general, the Co²⁺ ion exchange did not affect the crystallinity of the catalysts except for the gradual slight decreases in the intensity of the diffraction peaks at 2θ = 12.16° and 28.34° and slight increases in the intensity of the peaks at 2θ = 12.96°, 17.88°, and 20.06° with increasing degrees of ion exchange. Importantly, the Co²⁺ ion exchange did not induce any collapse of the Co-E10-n and Co-A6-m structures.

Fig. 2.XRD patterns of Co-E10-n (a) and Co-A6-m (b). Pristine ETS-10 and AM-6 denoted as Co-E10-0 and Co-A6-0, respectively.

The SEM images of the Co-E10-n and Co-A6-m samples are shown in Fig. 3. As observed, the crystals adopted a truncated tetragonal bipyramidal structure and were rather uniform in size. The surfaces were very smooth, and secondary phases or unreacted gel were absent in the samples. The lengths, L, in the (110) and (001) directions were 200-300 nm for Co-E10-n and 150-250 nm for Co-A6-m.

Fig. 3.SEM images of Co-E10-n with n = 0, pristine ETS-10 (a), 9 (b), 26 (c), 68 (d), and 81 (e) and of Co-A6-m with m = 0, pristine AM-6 (f), 8 (g), 23 (h), 63 (i), and 79 (j). Scale bar = 300 nm.

The pore structures of the Co-E10-n and Co-A6-m samples were studied by N₂ sorption. The N₂ sorption isotherms of the samples were measured within a relative pressure (p/p₀) range of 10⁻⁶ to 0.998; the isotherms are shown in Fig. 4. As observed, all samples displayed Type I isotherms according to the classification of BET, which is typical of microporous materials. The BET surface areas of the Co-E10-n samples (n = 0, 9, 26, 68, 81) were respectively 412, 410, 391, 358, and 337 m²/g and those of the Co-A6-m samples (m = 0, 8, 23, 63, and 79) were respectively 397, 386, 373, 349, and 333 m²/g. Thus, as deduced, as the degree of Co²⁺-ion exchange increased, the surface area gradually decreased. This trend indicated the onset of a slight structure collapse as the degree of Co²⁺ ion exchange increased, though the samples featured comparable XRD patterns and SEM results with no noticeable differences regardless of the ion exchange degree. Additionally, Co-A6-0 (pristine AM-6) displayed a similar surface area to that of previously reported high-quality AM-6 ^{[21, 22]}, thereby indicating that the crystallinity of Co-A6-0 is very high. The relative crystallinity, BET surface areas, and other structural information of the Co-E10-n and Co-A6-m samples are summarized in Table 1.

Fig. 4.N2 adsorption-desorption isotherms of Co-E10-n (a) and Co-A6-m (b).

Table 1 Chemical compositions and physical properties of catalysts used in this study. Catalyst Composition Degree of Co2+ exchange (%) Weight % of Co2+ Crystallinity (%) Surface area (m2/g) Raman band width (cm-1) Co-E10-0 Na1.6K0.40TiSi5O13 0 0.0 100 412 29 Co-E10-9 Co0.09Na1.43K0.40TiSi5O13 9 1.1 96 410 30 Co-E10-26 Co0.26Na1.10K0.38TiSi5O13 26 3.4 92 391 31 Co-E10-68 Co0.68Na0.35K0.29TiSi5O13 68 8.8 87 358 43 Co-E10-81 Co0.81Na0.17K0.21TiSi5O13 81 10.5 83 337 50 Co-A6-0 Na1.6K0.40VSi5O13 0 0.0 100 397 10 Co-A6-8 Co0.08Na1.45K0.39VSi5O13 8 1.0 94 386 11 Co-A6-23 Co0.23Na1.21K0.33VSi5O13 23 3.0 90 373 14 Co-A6-63 Co0.63Na0.43K0.31VSi5O13 63 8.1 89 349 27 Co-A6-79 Co0.79Na0.18K0.24VSi5O13 79 10.2 81 333 38 Pristine ETS-10 and AM-6 denoted as Co-E10-0 and Co-A6-0, respectively.

Table 1 Chemical compositions and physical properties of catalysts used in this study.

The Raman spectra of the Co-E10-n and Co-A6-m samples are shown in Fig. 5. The Raman spectra of the longitudinal (along the wire) vibration modes of the -Ti-O-Ti-O- and -V-O-V-O- chains provide highly useful information on the quality and local environment of the TiO₃²⁻ and VO₃²⁻ quantum wires in Co-E10-n and Co-A6-m, respectively. The ν_max value is related to the length of the quantum wires, and decreased as the length increased ^[22]. The high-quality ETS-10 crystals displayed ν_max in the 721-724 cm⁻¹ region ^{[23, 24]}. The intensity of the longitudinal vibration is related to the number of quantum wires and increased as the number increased. The bandwidth (fwhm) is related to the homogeneity of the length of the quantum wires (l) and decreased as the homogeneity of l increased. Reported values of the bandwidth have been observed between 23 and 120 cm⁻¹.

Fig. 5. Raman spectra of Co-E10-n (a) and Co-A6-m (b).

The obtained ν_max value at 724 cm⁻¹ for Co-E10-n (n = 0, 9, 26, 68) indicated that the average length of the titanate quantum wires in the Co-E10-n crystals remained unchanged regardless of the Co²⁺ exchange degree up to 68%. In contrast, Co-E10-81 displayed a shift in ν_max to a lower wavenumber region, indicating that the average length of the titanate quantum wires in Co-E10-81 was shorter than that in Co-E10-n (n = 0, 9, 26, 68). The bandwidths of Co-E10-n (n = 0, 9, 26, 68, 81) were 29, 30, 31, 43, and 50 cm⁻¹, respectively. This finding indicates that though the length homogeneity of the titanate quantum wires in pristine ETS-10 (Co-E10-0) was high, it gradually decreased with increasing degrees in Co²⁺ ion exchange. Thus, the introduction of Co²⁺ led to disconnection of the titanate quantum wires.

The high-quality AM-6 displayed a Raman shift, corresponding to the longitudinal vibration of the vanadate quantum wires, at 870 cm⁻¹ with a bandwidth of 10 cm^{−1 [20, 21]}. The ν_max and fwhm of Co-A6-0 (pristine AM-6) were 867 and 10 cm⁻¹, respectively, which was indicative of the very high crystallinity of Co-A6-0. The ν_max value of Co-A6-n (n = 8, 23, 63) was also 870 nm. This finding indicated that the average length of the vanadate quantum wires remained unchanged regardless of the degree of Co²⁺ exchange up to 63%. In contrast, Co-A6-79 displayed a ν_max shift to a lower wavenumber region, thereby indicating that the average length of the vanadate quantum wires in this sample was shorter than that in Co-A6-n (n = 0, 8, 23, 63). The bandwidths of Co-A6-n (n = 0, 8, 23, 63, 79) were 10, 11, 14, 27, and 38 cm⁻¹, respectively. This finding indicated that though the length homogeneity of the vanadate quantum wires in pristine AM-6 (Co-A6-0) was high, it gradually decreased with increasing degrees of Co²⁺ ion exchange. Likewise, the introduction of Co²⁺ led to disconnection of the vanadate quantum wires.

Pristine ETS-10 (Co-E10-0) was colorless unlike the corresponding Co²⁺-exchanged samples (n = 9, 26, 68, 81) that were light pink. The UV-visible spectra of the Co-E10-n samples are shown in Figs. 6(a)-(c) in the spectral regions of 200-800, 200-400, and 300-800 nm. The UV-visible spectrum of Co-E10-0 displayed absorption bands below 340 nm, and these of Co-E10-n (n = 9, 26, 68, 81) displayed additional bands at ~350 and ~520 nm. These new bands were obtained owing to the incorporated Co²⁺ ions. The intensity of the bands increased with increasing degrees of Co²⁺ exchange. Regarding the visible band, the absorption band blue shifted with increasing degrees of Co²⁺ exchange, indicating that it has the framework oxygen to Co²⁺ charge transfer nature.

Fig. 6. Diffuse reflectance UV-Vis spectra of Co-E10-n in the spectral regions of 200-800 (a), 200-400 (b), and 300-800 nm (c) and Co-A6-m in the spectral regions of 200-1500 (d), 200-400 (e), and 300-800 nm (f), respectively.

Pristine AM-6 (Co-A6-0) was pale yellow, and the corresponding Co²⁺-exchanged samples (m = 8, 23, 63, 79) were light gray. The UV-visible spectra of the Co-A6-m samples are shown in Figs. 6(d)-(f) in the spectral regions of 200-1500, 200-400, and 300-800 nm. The UV-visible spectrum of pristine AM-6 (Co-A6-0) displayed absorption maximums (λ_max) at 238 (0.98), 260 (1), 558 (0.036), and 1113 (0.061) nm (the relative intensity of the bands are shown in parentheses). The spectra of Co-A6-m (m = 8, 23, 63, 79) displayed additional absorption bands at ~350 and ~520 nm owing to the incorporation of Co²⁺.

The catalytic activities of Co-E10-n and Co-A6-m were assessed toward the oxidation of styrene with molecular oxygen (Fig. 7, Table 2). In both cases, styrene epoxide (E) and benzaldehyde (B) were obtained as the major products with trace amounts of phenylacetaldehyde. The overall rate of conversion was between 5.4 and 12.4 mol·g⁻¹·h⁻¹. Interestingly, even in the absence of Co²⁺, both Co-E10-0 (pristine ETS-10) and Co-A6-0 (pristine AM-6) were catalytically active toward the oxidation of styrene to E and B with molecular oxygen. This finding suggested that both Ti and V could also catalyze the reaction. However, the activity increased upon increasing addition of Co²⁺ content.

Fig. 7. Plots of the conversion and the major products with respect to the degree of Co2+ ion in Co-E10-n (a) and Co-A6-m (b). Plots of the E/B ratio with respect to the degree of Co2+ ion in Co-E10-n and Co-A6-m (c). Comparison of the TOF (Co-1·h-1) of various catalysts for styrene to styrene epoxide conversion (d).

Table 2 Oxidation of styrene with air over Co-E10-n and Co-A6-m catalysts. Catalyst Conversion (%) Rate (mmol·g-1·h-1) Selectivity (%) E/B TOF (Co-1·h-1) E B E B Blank <2.5 — 45 65 0.69 — — Co-E10-0 49 6.1 68 32 2.1 — — Co-E10-9 93 11.6 59 41 1.4 36.3 25.2 Co-E10-26 100 12.5 24 76 0.3 5.2 16.5 Co-E10-68 99 12.4 12 88 0.1 1.0 7.3 Co-E10-81 98 12.3 9 91 0.1 0.6 6.3 Co-A6-0 43 5.4 56 44 1.3 — — Co-A6-8 73 9.1 29 71 0.4 15.0 36.7 Co-A6-23 87 10.9 15 85 0.2 3.2 18.2 Co-A6-63 99 12.4 9 91 0.1 0.8 8.2 Co-A6-79 99 12.4 7 93 0.1 0.5 6.7 Reaction conditions: dehydrated catalyst 100 mg, 95 °C, styrene 5 mmol, DMF 20 mL, O2 ~4 mL/min, 4 h. Pristine ETS-10 and AM-6 denoted as Co-E10-0 and Co-A6-0, respectively.

Table 2 Oxidation of styrene with air over Co-E10-n and Co-A6-m catalysts.

Comparison between the two sets of reaction profiles (Figs. 7(a) and (b)) revealed that the Co-E10-n catalysts were more active than the Co-A6-m catalysts. The higher catalytic activity was likely due to the slightly larger pore size of ETS-10 relative to that of AM-6. Furthermore, as observed, a Co²⁺ exchange level of 26% was required to reach 100% conversion using the ETS-10 catalyst system. Contrarily, a Co²⁺ exchange level of 63% was required to achieve 100% conversion using the AM-6 catalyst system. In both catalyst systems (Co-E10-n and Co-A6-m), the reaction rates increased, while the E/B ratio decreased sharply with increasing degrees of Co²⁺ ion exchange (Fig. 7(c)). The highest E/B ratio (2.1) was obtained using Co-E10-0 that was lower than those observed using other catalysts (SrCoX18, BaCoX19, Co-SBA-15-20-5.2, Co-Beta, and Co-SAPO-34), of which the E/B ratios were in the range of 3.3-5.7 (Table 3). Accordingly, we could deduce that the Co-E10-n and Co-A6-m catalysts were more suited for application in the selective production of B over that of E from styrene and molecular oxygen. In particular, the E/B ratio became very small (0.1) when the highly Co²⁺-exchanged ETS-10 and AM-6 (Co-E10-68, Co-E10-81, Co-A6-63, and Co-A6-79) were used as the catalysts. Nevertheless, it is interesting to note that in the case of Co-E10-9, the turnover frequency per Co atom per hour was 36.3, which was higher than those obtained from the other catalysts studied (Table 3 and Fig. 7(d)).

Table 3 Catalytic performances of the catalysts extracted from the literature. Catalyst Conversion (%) Rate (mmol·g-1·h-1) Selectivity (%) E/B TOF (Co-1·h-1) Ref. E B E B Co2+-X (8.8 wt%) 45 5.6 65.0 35.0 1.9 4.0 0.04 [5] SrCoX18 100 12.5 85.0 15.0 5.7 20.0 0.03 [7] BaCoX19 100 12.5 83.0 17.0 4.9 26.0 0.04 [7] Co-SBA-15-20-5.2 92.4 4.1 63.2 15.6 4.1 3.0 0.79 [10] Co-Beta 35.5 2.1 72.8 21.9 3.3 15.3 3.85 [13] Co-SAPO-34 37.8 1.9 76.7 17.5 4.4 3.6 0.81 [14] Co-MOR 10.3 0.6 66.7 33.3 2.0 5.0 1.89 [15]

Table 3 Catalytic performances of the catalysts extracted from the literature.

The catalytic activity of ETS-10 exchanged with other transition metal ions, such as Cr³⁺, Mn²⁺, Fe³⁺, Ni²⁺, Cu²⁺, and Zn²⁺, was also examined. However, the resulting activities were considerably lower than that of the Co-E10-n catalysts, with Cr³⁺-ETS-10 displaying a conversion of 11% and E selectivity of 55%.

In the subsequent studies, Co-E10-9 and Co-A6-79 were chosen as the model catalysts. Following catalysis, the catalysts were recovered from the corresponding reaction mixtures by centrifugation and washed sequentially with copious amounts of DMF and water. The catalysts were then dried in an oven at 90 °C for 1 h, and finally vacuum dried at 150 °C for 10 h. The dried catalysts were reused in subsequent catalysis reactions. As noted, the conversion and selectivity remained unchanged after four catalytic cycles (Fig. 8), thereby indicating that Co-E10-9 and Co-A6-79 are highly stable and can be used efficiently with no loss in catalytic activity.

Fig. 8. Recycle test of the catalysts in styrene conversion with molecular oxygen on Co-E10-9 (a) and Co-A6-79 (b). Reaction conditions: same as Table 2.

Based on the current study and other literature reports ^{[7, 10]}, a plausible reaction mechanism for the epoxidation of styrene with molecular oxygen over Co²⁺-containing catalysts is shown in Scheme 1. It is expected that DMF molecules coordinate to the extraframework Co²⁺ cations (ZO)Co²⁺ present in the zeolite pores through the oxygen atom of DMF resulting in the formation of (DMF,ZO)Co²⁺. The Co²⁺ cation in (DMF,ZO)Co²⁺ is electron rich because it is coordinated to the negatively charged framework oxygen atoms. Consequently, (DMF,ZO)Co²⁺ attracts O₂. Subsequently, an oxygen molecule (O₂) coordinates to the Co²⁺ ion, resulting in the formation of a cobalt superoxo complex (DMF,ZO)Co³⁺-OO• (I). A styrene molecule coordinates to I through the vinyl group to form (DMF, ZO, Styrene)Co³⁺-OO• (II). Through migratory insertion, II rearranges to an intermediate III. Then, III disintegrates to the cyclic peroxide IV and (DMF,ZO)Co²⁺. The cyclic peroxide IV then undergoes two different types of reactions: (1) thermal decomposition to benzaldehyde and formaldehyde and (2) transfer of an oxygen atom to another styrene molecule to form two styrene epoxide molecules.

Scheme 1. A possible reaction mechanism for the oxidation of styrene into epoxide and benzaldehyde with air as the oxidant. ZO denotes E10-n and A6-m.

Pristine ETS-10 (Co-E10-0) and AM-6 (Co-A6-0) displayed catalytic activities toward the oxidation of styrene with molecular O₂ to styrene epoxide and benzaldehyde in DMF. The corresponding Co²⁺-exchanged powders showed higher catalytic activities, and the activity increased with increasing degrees of ion exchange. Contrary to the conversion, the E/B ratio decreased with increasing n or m reaching 0.1. Therefore, highly Co²⁺-exchanged ETS-10 and AM-6 (Co-E10-68, Co-E10-81, Co-A6-63, Co-A6-79) are very good catalysts for the selective production of B from styrene and molecular O₂. The highest turnover frequency obtained was 36.3 Co⁻¹·h⁻¹, which was the highest value obtained among the different Co²⁺-exchanged zeolites and mesoporous silica studied in this work.

Acknowledgments

This work was supported by the Korea Center for Artificial Photosynthesis, funded by the Ministry of Science, ICT and Future Planning through the National Research Foundation of Korea [no. 2009-0093886 and no. 2012R1A2A3A01009806]. We also thank J. Y. Lee for the help in drawing Fig. 1.