Mesoporous Co 3 O 4 as a New Catalyst for Allylic Oxidation of Cyclohexene

Mesoporous cobalt oxide was investigated for the liquid phase oxidation of cyclohexene using tertiobutylhydroperoxide (TBHP) as an oxidant. The results were compared with several series of supported cobalt catalysts to study the influence of the cobalt loading and solvents on the overall conversion and selectivity. Mesoporous cobalt was synthesized through the nanocasting route using siliceous SBA-15 mesoporous material as a hard template and cobalt nitrate as the cobalt oxide precursor. Supported cobalt oxide catalysts (Co/MxOy) were synthesized by the impregnation method using two loadings (1 and 5 wt.%) and Al2O3, TiO2, and ZrO2 as supports. Samples were characterised by means: elemental analysis, X-ray powder Diffraction (XRD), BET (surface area), UV-Vis DR Spectroscopy, and Transmission Electron Microscopy (TEM). The results obtained showed that the cobalt oxide retains the mesoporous structure of SBA-15, and in all Co/MxOy, crystalline Co3O4, and CoO phases are observed. The mesoporous cobalt oxide is more active than the supported cobalt catalysts in the allylic oxidation of cyclohexene, with a conversion of 78 % of cyclohexene and 43.3 % selectivity toward 2-cyclohexene-1-ol. The highest activity of mesoporous cobalt oxide could be ascribed to its largest surface area. Furthermore, Co3O4 has both Lewis and Brönsted acidic sites whereas Co/MxOy has only Lewis acidic sites, which could also explain its superior catalytic activity. Moreover, mesoporous cobalt oxide was more stable than supported cobalt catalysts. Therefore, this catalyst is promising for allylic oxidation of alkenes. Copyright © 2018 BCREC Group. All rights reserved


Introduction
In the petroleum industry a large amounts of alkane and alkene are produced as by-products which can be oxidized to oxygenated derivatives and converted into value-added products, such as: alcohols, ketones, aldehydes, and carboxylic acids [1].
During the allylic oxidation process, the reaction occurs on the C-H key in allyl of alkene to produce an alcohol and ketone and the C=C bond is retained [2,3].When the reaction take place at C=C key, peroxide is generates.The oxidation of cylohexene has been the research topic of a large number of investigations because of the importance of the obtained prod-ucts, like the cyclohexene oxide, 2-cyclohexene-1-one [4,5] and 2-cyclohexene-1-ol [5][6][7][8].The 2-cyclohexen-1-ol (enol) can be converted to phenol without co-products.The 2-cyclohexen-1-one (enone) is extensively utilized in the synthesis of drugs, pesticides and adipic acid which is utilized in nylon-6 manufacture.The synthesis of fine chemicals uses cyclohexene oxide as a key intermediate owing to its high reactivity which is ascribed to the opening of the highly strained three-membered ring [9,10].
Although the homogeneous catalysts have a high activity in the oxidation reactions, however they have significant disadvantages in the recovery of the catalyst and the isolation of the product.In contrast, heterogeneous catalysts systems exhibit high activity, stability, and reusability.Various cobalt-substituted mesoporous materials have been investigated as catalysts for selective liquid phase oxidation [5,[11][12][13][14].
Zhen et al. [15] reported that cobalt oxide in the form of nanocrystalline particles is a crucial factor for attaining high activity for combustion of VOCs, CO oxidation, and NO2 decomposition [15].Hence, the synthesis of nanostructured cobalt oxides especially with ordered mesoporous structure has been investigated [15,16].Mesoporous Co3O4 has exhibited a higher catalytic activity in the combustion of VOCs than that obtained by a conventional Co3O4.To the best of our knowledge this is the first report investigating mesoporous Co3O4 as a catalyst for selective liquid phase oxidation.
In this paper, mesoporous Co3O4 was prepared via the nanocasting route using mesoporous SBA-15 silica as a mold and cobalt nitrate as a precursor through a solid-liquid route [17].Mesoporous cobalt oxide was investigated for the liquid phase oxidation of cyclohexene using tertiobutylhydroperoxide (TBHP) as an oxidant.The results were compared with several series of supported cobalt catalysts to study the influence of the cobalt loading and solvents on the overall conversion and selectivity.

Mesoporous cobalt oxide (Co3O4)
Synthesis of the mesoporous silica SBA-15 was conducted in a way similar to that used by Zhao et al. [18].The mesoporous cobalt oxide was obtained using the method given below, i.e. 1 mmole of the precursor Co(NO3)x.yH2Owas combined with 0.15 g of SBA-15 and the mix was then crushed for a short moment in an agate mortar and pestle.Next, the admixture was placed in a melting pot that was put in a muffle furnace, where the temperature was raised from ambient temperature to 400 °C, at a rate of 1 °C/min; it was then maintained at the final value for the period of 5 hours.After that, the sample was cooled down back to ambient temperature.A 10 %wt.HF aqueous solution was used to remove the silica template.Centrifugation was used to recuperate the mesoporous cobalt oxide which was then washed with distilled water trees times.

Supported Co/MxOy
The supports, such as: Al2O3 (Oxid C Degussa), TiO2 (Titandioxid Degussa), and ZrO2 (Aldrich), are all available on the market; they are in the form of a nanopowder (nanoparticles in the range from 10 to 50 nm).A preliminary treatment is required before the metal precursors are impregnated.
The chosen supports were mixed with water (an amount of 200 mL of H2O with 100 g of support) in order to obtain a paste.This mixture was left to dry overnight at the temperature of 120 °C, and was then sifted in order to hold back only the particles with a diameter that is between 1 and 2.5 Å.After that, the support was calcined at the temperature of 400 °C under oxidative flow (20 % O2, 80 % Ar, 60 mL/min) for a period of 4 hours.These oxides were next impregnated with an aqueous solution of Co(NO3)2.6H2O (Fluka) in order to get 1 and 5 %wt. of cobalt catalyst.First, the solvent was evaporated, next the solids were dried at the temperature of 120 °C overnight, and then they were calcined at 400 °C.

Catalysts Characterization
Instrumental analyses have been employed to investigate the physicochemical characteristics of the prepared catalysts.The metal content was checked by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) using an OPTIMA 2000 DV spectrometer.The surface areas and pore sizes were measured by nitrogen adsorption at 77 K on a Micrometrics Tristar 3000 apparatus after heat pre-treatment under vacuum for 6 hours at 250 °C.The acidity contents were determined by Fourier Transform Infra Red (FTIR) analysis of chemisorbed pyridine using a Nicolet Magna IR spectrometer.Details on the experimental procedure have been described in previous works [4,19].
Wide angle powder X-ray Diffraction (XRD) patterns were collected on a Bruker D5005 dif-fractometer, using a Cu-Kα radiation (λ = 1.54 Å).For small-angle analysis, the XRD patterns were collected on a Bruker AXS D5005 X-ray source; the signal was recorded for 2θ comprised between 0.7 and 5 °.DRS spectra of the samples studied were recorded in the wavelength range 200-800 nm at room temperature, using a UV-Vis spectrophotometer (Perkin Elmer) with KBr as the reference.The size and dispersion of the Co particles were observed using Transmission Electron Microscopy (TEM) in conjunction with energy dispersive X-ray Spectroscopy (EDXS).Micrographs are obtained on a JEOL 2100 instrument (operated at 2200 KV with a LaB6 source and equipped with a Gatan Ultra Scan camera).

Catalytic Reactions
Tertiobutylhydroperoxide TBHP (Aldrich, 70 wt.% in H2O) was used as an oxidant for the catalytic oxidation of cyclohexene in a two-neck glass round bottom flask fitted with a magnetic stirrer and a reflux condenser.Firstly, the solvent and TBHP were mixed and well agitated to achieve a phase transfer from water to organic phase.
Normally, 20 mL of solvent and 5.5 mL of oxidant (TBHP) were blended into a closed erlenmeyer flask and magnetically agitated for a period of 24 hours.Next, the organic phase contained was isolated from the aqueous phase.The concentration of the TBHP that remained in the aqueous phase was determined by iodemetric titration in order to control the phase transfer.The quantity of TBHP that remained in the aqueous phase was less than 10 % of the initial amount.After that, the TBHP-solvent mixture, 3 mL of cyclohexene, and 0.1 g of catalyst were admixed in the magnetic stirrer-glass reactor, at 343 K and stirred for a period of 6 hours.
Identification of the reaction products was done by comparing them with authentic ones.The course of the reactions was monitored by gas chromatography (GC), using a YL6500 GC system equipped with an Agilent HP-FFAP capillary column and a flame ionization detector (FID).The remaining TBHP was decomposed before the GC analysis by incorporating an excess of triphenylphosphine (Aldrich).Moreover, an iodometric titraction was carried out at the end of the reaction (after 6 h) by analysing the organic phase, in order to control the remaining TBHP.The catalytic performances were reported in terms of cyclohexene conversion, selectivity towards products and turnover frequency.They are calculated following the Equations (1-2). (1) (2)

N2 Adsorption-desorption isotherms
N2 adsorption-desorption isotherms and pore size distribution of Co3O4 are presented in Figure 1.The textural properties of Co3O4 and supported cobalt oxide catalysts are listed in Table 1. Figure 1 shows that Co3O4 has a jump at P/P 0 = 0.8-1.0 which is typical for mesoporous solids indicating that the mesoporous structure is retained in the nanocasting process via a solid-liquid route [20,21].As shown in Table 1, mesoporous Co3O4 exhibits a larger surface area than those of supported catalysts.
It is important to note that the surface area of mesoporous Co3O4 synthesized in this work by the nanocasting pathway via a solid-liquid route is much higher than those of mesoporous Co3O4 synthesized by the nanocasting pathway via the solvent evaporation method [22,23].The lower surface area of mesoporous Co3O4 as compared to SBA15 silica can partly be ascribed to the difference between the densities of the two materials, as reported for mesoporous MgO synthesized by the nanocasting pathway [24].Bulk Co3O4 has a density of 6.11 g/cm³ while that of amorphous SiO2 is 2.2 g.cm -3 .It is reasonable to suppose analogous differences in the densities for the pore walls of the two mesoporous materials; therefore, the surface areas of mesoporous Co3O4 and SBA-15 silica are closer to each other when normalized with respect to the density.The lower surface area of mesoporous Co3O4 is also attributed to the partial loss of structural order, which is in agreement with the XRD results.

Acid properties of catalysts
Pyridine adsorption is still one of the most frequently used methods for evaluating the acidity of solids.The infrared spectrum in the region between 1400 and 1700 cm -1 has been determined for pyridine adsorbed on acidic solids.FTIR spectra of pyridine adsorbed on 5 wt% Co/Al2O3, 5 wt% Co/ZrO2, and 5 wt% Co/TiO2 (not shown) showed only the 1455 cm -1 band which is typical of Lewis sites.No characteristic bands of pyridine adsorption on Brönsted acid sites were seen (formation of the pyridinium ion leading to two adsorption bands at 1545 cm -1 and 1637 cm -1 ) [25,26].
The FT-IR spectrum of mesoporous Co3O4 showed both Lewis acid site bands and Brönsted acid site bands at 1446 and 1554 cm -1 , respectively.The data summarized in Table 1 indicate that mesoporous Co3O4 possesses significant Brönsted and Lewis acidity.The Co/Al2O3 sample was found to be more acidic (150 µmol.g -1 at 150 °C) than Co/ZrO2 (52 µmol.g -1 at 150 °C) and Co/TiO2 (6 µmol.g -1 at 150 °C).The cobalt loading measured by ICP-AES is 1 or 5 wt% in all supported materials.

XRD characterizations
The XRD patterns of the samples are presented in Figure 2. The low-angle XRD pattern of Co3O4 (Figure 2a) shows one diffraction corresponding to (100) reflection which can be related with the hexagonal symmetry characteristic of mesoporous SBA-15 [27] showing that the mesoporous structure is retained in the nanocasting process via a solid-liquid route.  is worth noting a certain degree of broadening of the reflections which indicate some loss in structural order.Figure 2b shows the wide-angles XRD patterns; on the Co3O4 pattern the diffraction peaks at 2θ = 19.08°,31.38°,36.88°,44.88°, 59.38°, and 65.28° are ascribed to Co3O4 in the cubic phase with Fd3m space group (JCPDS no.42-1467) [28].On all Co/ MxOy patterns (Figure 3), crystalline Co3O4 and CoO phases are observed and metallic Co is not detected.In addition, the diffraction peaks at 2θ = 34.5°and 42.4° could be attributed to CoO in the cubic form (JCPDS no.71-1178) [29,30].For 5 % Co/Al2O3, the diffraction peaks at 2θ = 37°, 45°, and 66° are attributed to g Al2O3 [31] and for 5 wt% Co/ZrO2, the diffraction peaks at 2θ = 28.17°, and 31.47° are ascribed to the presence of ZrO2 in the monoclinic phase (JCPDS no.37-1484), and the peak at 30.12° indicates the presence of zirconia in the face centered cubic form (JCPDS no.49-1642).On Co/TiO2 pattern, the simultaneous presence of rutile (JCPDS no.21-1276) and anatase (JCPDS no.21-1272) phases are detected indicating that TiO2 is constituted of a mixture of titanium oxide in the quadratic and centered quadratic phases.

DR-UV-Vis analysis
Diffuse-reflectance UV-Vis spectroscopy is a sensitive tool that is frequently used to detect the environment of cobalt.Mesoporous cobalt oxide spectrum is characterized by the presence of six bands (Figure 4).The bands centered at about 410 nm and 720 nm are assigned to O 2--Co 2+ and O 2--Co 3+ charge transfer processes, respectively [32][33][34], indicating the formation of Co3O4 which in agreement with the XRD results.The peak intensity at 400 nm on mesoporous Co3O4 was higher than that at 700 nm, which indicate the importance of Co 2+ species on this catalyst.The bands at 320 nm, 520 nm, and 620 nm are ascribed to the presence of Co 2+ in the tetrahedral coordination [35,36].The absorption band at 248 nm can be assigned to NO3 -ions present inside the pores [37].
The UV-vis-spectra of the catalysts supported cobalt Co/MxOy (Figure 5) exhibit strong band at ~240 nm originating from the oxygento-metal charge transfer transitions [38].The spectrums show a triplet at 320, 500, and 600 nm, characteristic of the presence of Co 2+ ions in tetrahedral coordination.Two others bands at 400 and 670 nm could be assigned to octahedral Co 3+ ion similarly to that in Co3O4 [34,.

TEM characterizations
Figures 6 and 7 present TEM images of the cobalt based catalysts.Figure 6 shows, in accordance with the XRD and N2 adsorptiondesorption results, the mesoscopic order of Co3O4, indicating that cobalt oxide retains the SBA-15 mesostructure.For 5 % Co/Al2O3 (Figure 7a) and 5 % Co/ZrO2 (Figure 7b) of TEM analysis reveals an intensive agglomeration of cobalt particles on these samples and, then, a true reliable distribution cannot be given.The average size of these agglomerations is between 8 nm and 10 nm.

Oxidation of Cyclohexene
The cyclohexene oxidation with tert-butyl hydroperoxide (TBHP) as oxidant was used to assess the catalytic activity of the as-prepared  catalysts.It should be noted that Al2O3, ZrO2, and TiO2 supports are inactive in this reaction.The cyclohexene oxidation can give six products: cyclohexene epoxide (Epo); 2-cyclohexene-1-ol (enol); 2-cyclohexene-1-one (enone); cyclohexanol (ol); cyclohexanone (one); and cyclohexene-1,2-diol.A possible mechanism of cyclohexene oxidation has been proposed; it involves two main paths.The first one is a direct epoxidation, in the presence of redox catalysts, leading to the epoxide, which may be transformed into a diol in the presence of water and an acidic medium.The second one is an allylic oxidation which, in the presence of Lewis acid sites and redox centers, may give 2-cyclohexene-1-ol (enol), 2-cyclohexene-1-one (enone), cyclohexanone, and cyclohexanol [43] as illustrated in the following catalytic scheme (Figure 8).
Our catalysts drive the reaction selectivity towards the allylic oxidation due to their strong acidity [4,25,44].The orientation of the reaction towards the formation of 2-cyclohexene-1-ol as majority product indicates the preferential attack of the activated C-H bond over the C=C bond.It is important to mention that, the 2-cyclohexene-1-one, was only detected in presence of Co3O4 mesoporous oxide (Table 2).
From the results shown in Table 2 we note: The mesoporous cobalt oxide has a better activity with 78 % of conversion compared to Co/MxOy catalysts.Co3O4 exhibits a larger surface area than those of supported catalysts (Table 1) which could explain the highest catalytic activity of this catalyst.Additionally, Co3O4 has both Lewis and Brönsted acidic sites whereas Co/MxOy have only Lewis acidic sites (Table 1), which could also explain its superior catalytic activity, so it can be suggested that the presence of Brönsted acidic sites improves the catalytic activity in the cyclohexene oxidation.
These results could be explained in terms of the Lewis acidity of the catalysts.It is well established in the literature that the acidity jointly with the redox properties of the metal determine the selectivity of the catalyst towards oxidation products [45].Moreover, the contribution of Lewis acid sites in the oxidation of cyclohexene has many times been stressed in the literature [45][46][47]    The results reveal (Table 1) that all Co/MxOy systems seem to have the following relative order of Lewis acidity: Co/Al2O3 > Co/ZrO2 > Co/TiO2.This Lewis order of acidity is in agreement with the catalyic activity of our samples and also with other results previously reported in the literature [48].On other hand, TBHP as oxidant promotes the allylic oxidation pathway and minimizes the epoxidation, particularly under the highly acidic properties of catalysts [49].
It is well known that the nature of solvents used has a major impact on the reaction kinetics and product selectivity in the oxidation of cyclohexene.The oxidation reaction was carried out in polar protic solvent (acetic acid) in presence of Co3O4 and 5 %Co/Al2O3 the most acidic catalyst among the 5 %Co/MxOy catalysts (Table 2).It has been found that 2-cyclohexene-1-ol was formed as the major product with high conversion (78 %) and selectivity (43.3 %) over mesoporous Co3O4.Obviously, the catalyst surface prefers to adsorb the more polar molecules, such as ketone and alcohol molecules, which are able to block the active sites of the catalyst.Then, the polar solvent could wash out the other products from the surface and consequently the oxidant and cyclohexene would have more chance to reach the active sites [50,51].Moreover, the acetic acid acts as an oxidizing agent (formation of the peracetic acid complex) [19].

Reusability of the Catalyst
The catalytic stability of mesoporous Co3O4 and 5 % Co/Al2O3 in cyclohexene oxidation is studied (Table 3).The catalyst was reused for a new run after it was separated from the reaction mixture through filtration, washed with water and dried at 100 °C.
For mesoporous Co3O4, the activity of the refreshed catalyst after the first cycle remains the same as that of fresh catalyst and decreases of 21.8 % and 34.6 % for the 3 rd and 4 rd runs, respectively.It is also noted that the selectivity changes slightly, the appearance of the diol after the first cycle is probably due to the presence of the water framework, adsorbed during the course of the reaction.In contrast, 5 %Co/Al2O3 loses its activity under recycling.The deactivation of the catalyst is probably related to frequent inhibition processes found in oxidation reactions such, as the adsorption of oxygenated species on the catalyst surfaces [3].
Leaching of the metal into solution was checked by AAS analyses of the supernatant solution (Table 3).We found that a negligible amount of cobalt was leached into solution when mesoporous Co3O4 was used as catalyst, so it is not the likely reason for the catalyst deactivation.In contrast, a higher metal leaching (3.61 %.) was observed for Co/Al2O3.
In order to confirm that mesoporous Co3O4 acts truly as a heterogeneous catalyst, it was separated by filtration from the reaction mixture and the reaction was allowed to continue on the remaining filtrate solution for another 2 h under the same reaction conditions; no further conversion of cyclohexene was observed, showing that the oxidation of cyclohexene over mesoporous Co3O4 is purely heterogeneous.
In order to investigate the causes for the decrease in the activity during the recycling experiment, the spent Co3O4 catalysts were characterized by XRD and N2 physisorption.9 that the XRD patterns of fresh and spent catalysts are similar, indicating that the ordered structure of Co3O4 is not affected during the recycling experiment, showing that the catalytic system can be reused without loss of catalyst's structure.
In contrast, the textural properties of Co3O4 catalysts are affected during the recycling experiment (Table 4), the BET surface area and the pore volume decrease, which is attributed to clogging catalyst pores by carbonaceous species that makes them inaccessible for nitrogen adsorption.These carbonaceous species, formed inside the catalyst porosity through secondary reaction between the reaction products, block the pore access of the catalyst, leading to a deactivation during the reuse.The spent catalyst can be easily regenerated by calcination in air at 500 °C.

Conclusions
This study demonstrates that mesoporous Co3O4 is more active than supported cobalt oxide catalysts for allylic oxidation of cyclohexene with TBHP as an oxidant.The superior catalytic activity of mesoporous Co3O4 could be explained by its larger surface area compared to those of supported materials (Co/MxOy).Indeed, a high surface area often leads to a greater activity due to an increased dispersion of active sites.
Additionally, Co3O4 has both Lewis and Brönsted acidic sites whereas Co/MxOy have only Lewis acidic sites, which could also explain its higher catalytic activity.Indeed, a high surface area often leads to a greater activity due to an increased dispersion of active sites.Moreover, mesoporous cobalt oxide is more stable than supported cobalt catalysts and it can be regarded as self-supported nanoscale catalysts; it does not need any support, as required by normal nanoparticle catalysts.Therefore, this catalyst is promising for allylic oxidation of alkenes.

Figure 8 .
Figure 8. Different pathways for cyclohexene oxidation reaction

Table 1 .
Textural properties of supports, Lewis, and Brönsted acidity and metal content of catalyst .

Table 2 .
Oxidation of cyclohexene with different catalysts

Table 3 .
Recycling results of the catalystsIt can be seen from Figure

Table 4 .
Textural properties of fresh and used catalystsFigure 9. XRD patterns of fresh and used catalysts ndrun Intensity (a.u.)