2.1 Instruments and reagents

Infrared spectra were recorded on KBr pellets using a Unicam Matson 1000 FT-IR. A Bruker FT-NMR 500 (500 MHz) spectrometer was utilized to obtain the NMR spectra. A Varian AA240 atomic absorption spectrometer was used for molybdenum determination. The high-power ultrasonic cleaning unit Bandelin Super Sonorex RK-100H with ultrasonic peak output 320 W and HF power 80 W_eff has been used. The melting points were measured on an Electro thermal 9100 apparatus. X-ray powder diffraction (XRD) was carried out on a Philips X’Pert diffractometer using the Cu Kα radiation. Scanning electron microscopy (SEM) images were obtained on a Hitachi S-1460 field emission scanning electron microscope using an ACC voltage of 15 kV. Transmission electron microscopy (TEM) image were obtained with a Zeiss EM-900 transmission electron microscopy device using an ACC voltage of 80 kV. The oxidation products were analyzed with a gas chromatograph (Shimadzu, GC-14B) equipped with a SAB-5 capillary column (phenyl methyl siloxane 30 m × 320 mm × 0.25 mm) and a flame ionization detector.

2,2′-Dimethylpropylenediamine, 5-bromo-2-hydroxybenzaldehyde, molybdenyl acetylacetonate and hydrogen peroxide (solution 27% in water) were used as received from commercial suppliers. The solvents were dried and distilled by standard methods before use. Other chemicals were purchased from Merck or Fluka chemical companies.

2.2 Preparation of the Schiff base ligand (H₂L)

The Schiff base ligand was prepared by the reported methods [41,42].

2.3 Preparation of molybdenum(VI) complex

The molybdenum complex (Fig. 1) was prepared as follows: the Schiff base ligand, H₂L (0.468 g, 1 mmol) was dissolved in 20 mL of ethanol. An ethanolic solution of molybdenyl acetylacetonate (0.326 g, 1 mmol) was added and the reaction mixture was refluxed for 1 h. The light orange solution was concentrated to yield orange powders. The products were washed with warm ethanol. Mo(O)₂L: Yield: 83%, 0.493 g, D.p.: > 295, Selected FT-IR data, υ (cm⁻¹): 2922 (CH), 1614 (CN), 1520 (CC), 826 and 913 (Mo–O). ¹H NMR (δ): 0.75, 1.17 (s, 6H, NCH₂C(CH₃)₂CH₂N), 1.18 (s, 2H, NCH), 3.38–4.63 (4H, NCH₂C(CH₃)₂CH₂N), 6.54–7.60 (m, 6H, ArH), 8.04–8.12 (s, 2H, NCH).

2.4 Preparation of the (α-Fe₂O₃)–MCM-41–Mo(O)₂L

Fe₃O₄ magnetic nanoparticles were prepared by the chemical co-precipitation method reported in the literature [10,43]. Briefly, a solution with molar composition of 3.2 FeCl₃:1.6 FeCl₂:1 CTABr:39 NH₄OH:2300 H₂O was used for the preparation of Fe₃O₄ nanoparticles at room temperature under N₂ atmosphere. The magnetic MCM-41 was prepared by adding 20 mL of the Fe₃O₄ colloid after sonication to a 1 L solution with molar composition 292 NH₄OH:1 CTABr:2773 H₂O under vigorous mixing and sonication. Then, sodium silicate (16 mL) was added, and the mixture was allowed to react at room temperature for 24 h under consecutive mixed conditions. The magnetic MCM-41 was filtered and washed. The surfactant template was then removed from the synthesized material by calcination at 450 °C for 4 h and so (Fe₃O₄)–MCM-41 was converted to (α-Fe₂O₃)–MCM-41[10]. (α-Fe₂O₃)–MCM-41 (3 g) was suspended in 60 mL of chloroform with sonication. To this mixture, an excess amount of Mo(O)₂L was added, and the resulting mixture was stirred for 12 h with a mild magnetic agitator at room temperature, then filtered to obtain a brick-red precipitate, which was washed with chloroform as a Mo(O)₂L solvent, and finally dried in air. The preparation of (α-Fe₂O₃)–MCM-41–Mo(O)₂L is shown in Fig. 2.

2.5 General heterogeneous green oxidation procedure

Catalytic experiments were carried out in a 10-mL glass flask fitted with a water condenser. In a typical procedure, 0.5 mmol cyclooctene, 0.015 g of (α-Fe₂O₃)–MCM-41–Mo(O)₂L and 2 mL (20 mmol) of H₂O₂ were added in 3 mL of ethanol. The reaction mixture was refluxed for 8 h. The reaction products were monitored at periodic time intervals using gas chromatography (GC). The oxidation products were identified by comparison with authentic samples (retention times in GC).

3.1 Characterization of the ligand and of the dioxomolybdenum(VI) complex

3.2 Characterization of (α-Fe₂O₃)–MCM-41–Mo(O)₂L

3.2.1 Physicochemical characterization

Immobilization of the dioxomolybdenum complex on the surface of silica-coated magnetite nanoparticles is proposed to occur by hydrogen bond formation between the MoO and PhO groups of the complex as hydrogen bond acceptors, and the OH groups of the support as hydrogen bond donors. Also, the weak coordination of the OH groups to the metal center as a hard acid may be also involved. However, with respect to the negligible catalyst leaching, the formation of strong hydrogen bonds seems to be a more plausible explanation. A broad peak in the region 1050–1200 cm⁻¹ and two peaks in 804 and 950 cm⁻¹, respectively, are assigned to the stretching vibrations of SiOSi (mesoporous framework). The peak at 1627 cm⁻¹ belongs to the FeO bond of α-Fe₂O₃. By entering the complex within the pores of α-Fe₂O₃–MCM-41, no change in the IR spectrum was observed. Peaks at 2921 cm⁻¹ and 1458–1514 cm⁻¹ can be related to the CH stretching vibrations and CC bonds, respectively (Fig. 3).

The molybdenum content of the catalyst was determined by atomic absorption spectroscopy. Accordingly, the metal content of the catalysts was found to be ca. 0.0023 wt %. Scanning electron micrographs (SEM) were recorded to obtain the shape and diameter of the catalyst (Fig. 4a). The results show the spherical morphology of the catalyst; particles were well distributed and the average particle final size was 80–90 nm for (α-Fe₂O₃)–MCM-41–Mo(O)₂L.

Fig. 4b shows the transmission electron microscopy (TEM) images of (α-Fe₂O₃)–MCM-41–Mo(O)₂L. The TEM micrograph of the prepared catalyst shows an ordered hexagonal pore system with embedded α-Fe₂O₃ nanoparticles. As it can be seen, the products are formed from particles with a mean diameter of 80–90 nm ((α-Fe₂O₃)–MCM-41–Mo(O)₂L), which is in agreement with other results.

X-ray diffraction analysis was performed in order to investigate the effect of the incorporation of the molybdenum complex on the texture properties of the materials. In the XRD pattern of (α-Fe₂O₃)–MCM-41 (Fig. 5), a strong peak at 2θ = 2.75 and two weaker peaks at higher reflection angles of the hexagonal unit cell are observed. The XRD pattern of the supported molybdenum catalyst (α-Fe₂O₃)–MCM-41–Mo(O)₂L is shown in Fig. 5. All the peaks are weaker than the former one, and the strongest peak shifted to a lower angle (2θ = 2.65), which is probably due to the incorporation of the molybdenum complex into (α-Fe₂O₃)–MCM-41.

3.3 Heterogeneous catalytic green epoxidation of olefins

The catalytic performance of (α-Fe₂O₃)–MCM-41–Mo(O)₂L was investigated in the epoxidation of olefins with hydrogen peroxide (H₂O₂) and tert-butyl hydroperoxide (TBHP) as oxygen donors. A series of blank experiments (Table 2) revealed that the presence of both catalyst and oxidant is essential for an effective catalytic reaction. In the absence of the catalyst (entry 1 in Table 2) or oxidant (entry 2 in Table 2), the reactions did not proceed under reflux. Also, in the presence of (α-Fe₂O₃)–MCM-41 as a catalyst and TBHP as oxidant, no product was formed (entry 3 in Table 2).

In order to find the suitable reaction conditions, the effect of various reaction parameters that may affect the conversion and selectivity of the reaction were studied. Solvent, catalyst concentration, the nature and the concentration of oxidant are the factors that have been evaluated. The influence of different solvents on the oxidation of cyclooctene was studied using nanoparticles of (α-Fe₂O₃)–MCM-41–Mo(O)₂L as the catalysts; the results are presented in Fig. 6. Chloroform, acetonitrile, methanol and ethanol were used as solvents and the highest conversion (100% with H₂O₂) was obtained in ethanol. According to the data in Fig. 6, the conversions are comparable in methanol and acetonitrile. Aqueous hydrogen peroxide has a low solubility in dichloromethane, but is soluble in the other solvents. Therefore, solvents with higher dielectric constants, i.e. methanol, ethanol and acetonitrile are much better than dichloromethane. As it may be seen from Scheme 1, the presence of unoccupied coordination sites on the metal center of the catalyst is crucial for its catalytic performance [49]. On the other hand, methanol and acetonitrile are of smaller size compared to ethanol; their coordination to the metal center is expected to be more facile than with ethanol. Accordingly, the lower catalytic activity of the title catalyst observed in acetonitrile and methanol, compared to that in ethanol, seems to be due to the greater competition between the peroxide and the former for coordination sites on the metal center in comparison with the latter. Also, the higher conversions in ethanol relative to methanol are possibly due to the higher boiling point of the former.

In order to investigate the effect of the oxidizing agent in the oxidation reaction, TBHP and hydrogen peroxide were used. In the presence of TBHP, the reaction requires 3 h for competition, while the reaction was completed after 8 h with eco-friendly hydrogen peroxide.

Different catalyst concentrations have been used in the oxidation of cyclooctene (Fig. 7). It was observed that the oxidation of cyclooctene required 0.015 g of catalyst for completion. The highest conversion was obtained with 0.15 g of catalyst and further increments of catalyst hardly improved conversion. This may be due to the higher degradation of the oxidant at higher concentration of catalysts [49].

Different oxidant (H₂O₂) concentrations and addition methods have been studied in the oxidation of cyclooctene (Fig. 8). It was observed that oxidation of cyclooctene required 20 mmol of H₂O₂ for completion. Gradual or immediate addition of H₂O₂ to the reaction mixture led to nearly similar results, though the latter gave relatively higher conversion. The result shows that there is little disproportionation of hydrogen peroxide in the used concentration.

In order to establish the general applicability of this catalytic system, under optimized conditions, the oxidation of different olefins was performed in the presence of a catalytic amount of (α-Fe₂O₃)–MCM-41–Mo(O)₂L and the results are presented in Fig. 9. The oxidation of cyclooctene, cyclohexene, 1-methylcyclohexene, 1-octene and indene led to the corresponding epoxide as the sole product, while the oxidation of styrene led to the formation of benzaldehyde (Fig. 9, column 4), and that of substituted styrenes giving the substituted benzaldehydes (Fig. 9, columns 5–7); also, in the oxidation process of α-methyl styrene, acetophenone (Fig. 10, column 8) was detected as the sole product [50]. The production of benzaldehyde in the oxidation of styrene is due to the over-oxidation of styrene oxide with H₂O₂. The lower catalytic activity of α-methylstyrene with respect to styrene seems to be due to the steric hindrance of the methyl substituent of the former [51]. It should be noted that if electronic effects due to the methyl substituents were important, then, α-methylstyrene as a more electron-rich olefin would be expected to show higher reactivity relative to styrene. Epoxidation of the terminal alkene (oct-1-ene) proceeds in fair yield (73%).

In the proposed catalytic cycle, the oxidant will be activated by coordination to the molybdenum center and will form a hepta-coordinated molybdenum intermediate (I, Scheme 1) [52–56]. Then, olefin as a nucleophile will attack the oxygen atom of the coordinated hydrogen peroxide.

3.4 Catalyst separate, reuse and stability

One of the most attractive features of heterogeneous catalysts is their separation property. Catalyst recovery and reuse are the two most important features for many catalytic processes, and most heterogeneous systems require a filtration or centrifugation step and/or a tedious work up of the final reaction mixture to recover the catalyst. Magnetically supported catalysts can be recovered with an external magnet (Fig. 10). The homogeneous molybdenum complex cannot be recovered even once; in contrast, the heterogeneous catalysts can be magnetically separated and reused multiple times without significant loss of catalytic activity. The stability of the magnetic nanocatalyst was monitored using a multiple sequential epoxidation reaction. For each one of the repeated reactions, the catalyst was separated, washed with chloroform and methanol and dried before being used. The catalyst was consecutively reused six times (Fig. 11) without detectable catalyst leaching or significant loss of activity. The results indicated that the catalyst, after having been reused several times, showed no detectable leaching or change in its IR spectra.