Liquid phase selective oxidation of cyclohexane using gamma alumina doped manganese catalysts and ozone: an insight into reaction mechanism

Fourier transform-infrared spectroscopy (FT-IR)

The functional groups present in catalysts were identified using transmission FT-IR spectroscopy. The analysis was carried out using a Bruker Tensor 27 FT-IR spectrometer with a standard attenuated total reflection (ATR) cell. Ethanol was used to clean the surface prior to analysis. The pressure was adjusted to 90 Gauge for proper contact between the surfaces. The mid-IR region for catalysts scan was kept between 450 and 4000 cm⁻¹ range.

X-ray diffraction

The crystal lattice and phases of the prepared catalysts were examined using X-ray diffraction spectroscopy. Typically, the Bruker AXS D8 diffractometer with monochromatic Cu Kα (λ = 1.5406 Å) incident radiation at 40 kV and 40 mA at room temperature was used. The diffractometer utilized the CuKα as the radiation source of wavelength = 1.5406 nm. Scan speed was set to be 0.3/min over 10°–90° scan range. Prior to each analysis the sample holder was properly cleaned with ethanol solvent and dried.

Scanning electron microscopy – energy dispersive X-ray spectroscopy (SEM-EDX)

Scanning electron spectroscopy (SEM) was used to study the morphology of the catalysts. The instrument used was Carl Zeiss FE-SEM Sigma VP-03-67 with operation conditions of acceleration voltage of 20 kV over the working distance of 6–9 mm. All the samples were ground into fine powder with mortar and pestle prior to the analysis. The analysis was completed by placing a small amount of the powdered sample on a piece of two-way carbon tape and mounted it on a sample holder (stub). The elemental composition of the sample was carried out on an Oxford instruments X-MaxN 50 model 54-XMX1003 EXD analyser.

Transmission electron microscopy (TEM)

Particle size and morphology of the calcined catalysts were studied by TEM analysis. The analysis was conducted using JOEL JEM-1010 electron microscope with an acceleration voltage of 100 kV. The images were captured with Meagaview III camera and analysed with iTEM imaging software. The sample preparation involved dispersing the catalyst powder into toluene for 20 min. The analysis was conducted by placing a drop of the catalyst sample on a copper grid coated with formvar with mesh size of 150 and allowed to dry at room temperature before TEM images were captured.

Inductively coupled plasma-optical emission spectroscopy (ICP-OES)

The metal content of the catalysts was quantified using Agilent 700 series ICP-OES with a 710 ICP-OES detector instrument. A multi element standard solution was diluted to 25 ppm, 50 ppm, 75 ppm and used for calibration of the instrument. Prior, to the analysis the adequate amount of the catalysts samples were digested in 3.0 mL HCl, 3.0 mL HNO₃ and 4.0 mL HF at 60 °C for 1 h until digestion was complete. The sample preparation was completed by subsequent dilution of the resultant solution with distilled water up to the 100 mL mark of the volumetric flask to make 25 ppm, 50 ppm and 75 ppm. The content of the metals was extrapolated from the calibration curve established from the standard solution.

Brunauer–Emmet–Teller (BET)

The surface area analysis of the catalysts was conducted with Brunauer Emmet and Teller (BET) surface area analyser. The catalysts (0.2 g) were initially degassed at 200 °C in Micrometrics flow prep 060 instrument overnight under nitrogen flow. Subsequently, the analyses were done in an automated single and then multiple point Micrometrics Tristar 3000 BET surface area analyser under liquid nitrogen flow.

X-ray photoelectron spectroscopy

X-ray photoelectron spectroscopy (XPS) analysis were perfumed to evaluate near surface state of the catalysts using an AXIS Ultra DLD equipped with charge neutralizer. The centre utilized for the scans was 520 eV with a 1205 eV width, dwell time at 100 ms and steps at 1 eV. The high-resolution scans were obtained using 80 eV pass energy in slot mode. The fitting of the curves was performed with Gaussian-Lorentzian peak shape post linear background collection.

Electron paramagnetic resonance

Electron paramagnetic resonance (EPR) spectroscopy measurements were carried out using the Bruker EMX Plus EPR spectrometer, specifications: model number: EMP-9.5/12B/P. set at 0.632 mW for the microwave power, frequency 9.714 GHz, resolution 2048 points, at a centre field of 3500 G and 200 G for the sweep width and time constant of 5.12 for the hydroxide radicals determination, using DMPO as the quencher.

Fourier-transform infrared (FT-IR) spectroscopy

The FT-IR spectra of the γ-Al₂O₃ support and Mn/γ-Al₂O₃ catalysts are displayed in Fig. 2. The FT-IR spectra exhibit a broad band at 3500 cm⁻¹ which is due to the stretching vibrational frequency of OH groups of the adsorbed H₂O and Lewis acid sites of the γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts. Another band can be observed at 1600 cm⁻¹ which can be assigned to the bending vibrations of OH group [16, 17]. The existence of both the Al–O bond Mn associated bond is trivial as there were no peaks corresponding to Al–O and Mn metal observed.

Fig. 2:

FT-IR spectra of (a) γ-Al₂O₃, (b) 2.5% Mn/γ-Al₂O₃, (c) 5% Mn/γ-Al₂O₃, (d) 7.5% Mn/γ-Al₂O₃, (e) 10% Mn/γ-Al₂O₃, (f) 12.5% Mn/γ-Al₂O₃ and (g) 15% Mn/γ-Al₂O₃ catalysts.

X-ray diffraction (XRD)

The X-ray diffraction spectra of γ-Al₂O₃ support and M/γ-Al₂O₃ catalysts are displayed in Fig. 3. The γ-Al₂O₃ displays diffraction peaks at 2θ = 19.89°, 24.88°, 30.49°, 35.02°, 38.95°, 42.08°, 47.36°, 54.39°, 58.17°, 68.35° and 72.91° corresponding to (101), (111), (220), (221), (331), (222), (400), (422), (511), (440) and (444) hkl of cubic γ-Al₂O₃ [18], [19], [20]. There are no X-ray diffraction peaks corresponding to Mn species observed in 2.5%–10% Mn/γ-Al₂O₃ catalysts. This suggests that the Mn is well incorporated into crystalline structure of the γ-Al₂O₃. However, significant alterations in crystalline structure of γ-Al₂O₃ were observed upon Mn loading. For example, the incorporation of the Mn was followed by broadening of the diffraction peaks corresponding to γ-Al₂O₃ and the disappearance of the (111) and (101) peak. This has been reported before in Fe doped γ-Al₂O₃ [21]. The XRD spectra of 12.5% and 15% displays additional peaks at 2θ = 17.24°, 20.43°, 31.62°, and 43.5° corresponding to (200), (111), (311), and (112) hkl of the MnO phase [22].

Fig. 3:

XRD spectra of (a) γ-Al₂O₃, (b) 2.5% Mn/γ-Al₂O₃, (c) 5% Mn/γ-Al₂O₃, (d) 7.5% Mn/γ-Al₂O₃, (e) 10% Mn/γ-Al₂O₃, (f) 12.5% Mn/γ-Al₂O₃ and (g) 15% Mn/γ-Al₂O₃ catalysts.

Scanning electron microscopy (SEM)

Figures 4 and S1 (Supplementary information) display the SEM images of the γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts. The SEM image of pristine γ-Al₂O₃ support display properties of the fluffy powdered material with poorly defined structure, with rough morphology. The incorporation of the Mn metal resulted to a slight change in morphology from rough to smooth, highly aggregated surface. The EDX spectra are displayed in Fig. S2. The EDX spectra confirm the existence of the Mn metal (Table 1) and the elemental composition is in correlation with the anticipated composition. The results of the percentages of the Mn in γ-Al₂O₃ were also confirmed with ICP-OES (Table 1). Therefore, the Mn percentage reports of both EDX and ICP-OES display high accuracy toward anticipated percentage reports.

Fig. 4:

SEM images of (a) γ-Al₂O₃, (b) 2.5% Mn/γ-Al₂O₃, (c) 5% Mn/γ-Al₂O₃ and (d) 7.5% Mn/γ-Al₂O₃ catalysts.

Transmission electron microscopy (TEM)

The TEM images of γ-Al₂O₃, Mn/γ-Al₂O₃ catalysts and their corresponding particle diameter histograms are displayed in Figs. 5 and S3 (SI). The TEM image of γ-Al₂O₃ displays properties of a highly aggregated, spherical shaped material with uniform particle diameter of 12.8 nm in average. Upon incorporation of Mn to γ-Al₂O₃, an alteration in particle shape from spherical to rod shaped particle was perceived. This is due to the elongation of the γ-Al₂O₃ particles, which suggests that Mn/γ-Al₂O₃ catalysts are new materials with unique properties. In addition, a decrease in particle diameter was also perceived, which became more perceptive with increase in Mn loading. This is also in correlation with the XRD results in which Mn loading resulted to broadening of the γ-Al₂O₃ peaks suggesting a decrease in particle sizes. The average particle diameters of 2.5%, 5%, 7.5%, 10%, 12.5% and 15% Mn/γ-Al₂O₃ were 10.6 nm, 9.8 nm, 7.7 nm, 7.2 nm, 6.3 nm and 5.7 nm, respectively.

Fig. 5:

TEM images of (a) γ-Al₂O₃, (c) 2.5% Mn/γ-Al₂O₃, (e) 5% Mn/γ-Al₂O₃ and (g) 7.5% Mn/γ-Al₂O₃ and particles diameter histograms, (b) γ-Al₂O₃, (d) 2.5% Mn/γ-Al₂O₃, (f) 5% Mn/γ-Al₂O₃ and (h) 5% Mn/γ-Al₂O₃ catalysts.

X-ray photoelectron spectroscopy

The XPS spectra of the 2.5% Mn/γ-Al₂O₃ catalyst is shown in Fig. 6. The survey spectrum (Fig. 6a) shows the presence of Al, O and Mn at 74 eV, 535 eV and 650 eV, respectively. The high resolution spectrum of Al 2p (Fig. 6b) displays four distinctive peaks at 74.5 eV, 74.1 eV, 72.7 eV and 70.1 eV corresponding to Al₂O₃, AlO, Al₂O₃ and Al(OH)₃, respectively. The presence of the multiple oxidation states of Al element suggests that the surface bonding is not uniform due to Mn doping which often results to surface defects. Even though the XRD analysis showed the occurrence of MnO only, the XPS spectrum of the Mn 2p (Fig. 5c) displays various oxidation states of Mn (MnO, Mn₂O₃ and MnO₂) which also resulted from alteration in bonding upon incorporation to γ-Al₂O₃ support. The XPS spectrum of O 1s displays three peaks at 530.2 eV, 531.1 eV and 531.9 eV emanating from the crystal lattice, surface hydroxide and oxygen defect sites with low oxygen coordination [23, 24], respectively.

Fig. 6:

XPS spectra of (a) full range 2.5%/Mnγ-Al₂O₃, (b) Al 2p, (c) Mn 2p and (d) O 1s of 2.5% Mn/γ-Al₂O₃ catalysts.

Brunauer–Emmet–Teller (BET) surface analysis

The surface area is a crucial factor in heterogeneous catalysis. The BET surface analysis was performed to investigate surface areas, pore volumes and pore sizes of the as synthesized Mn/γ-Al₂O₃ catalysts. The large surface area provides more active sites which subsequently enhances absorption of the cyclohexane to the surface of the catalyst [25]. The BET isotherms of Mn/γ-Al₂O₃ catalysts are shown in Figs. 7 and S4. The isotherms display characteristics of the type IV BET isotherm with a hysteresis loop, proving the existence of mesopores on the Mn/γ-Al₂O₃ catalysts [26]. The surface areas, pore volumes and pore sizes of the Mn/γ-Al₂O₃ are listed in Table 1. The results show that the incorporation of the Mn to γ-Al₂O₃ resulted to a decrease in surface area and pore volume and an increase in pore sizes. The effect is more perceptive with increasing Mn ratio to γ-Al₂O₃. This observation suggests that the Mn is deposited to the pores of the γ-Al₂O₃ support.

Fig. 7:

BET isotherms of (a) γ-Al₂O₃, (c) 2.5% Mn/γ-Al₂O₃, (e) 5% Mn/γ-Al₂O₃ and (g) 7.5% Mn/γ-Al₂O₃ catalysts.

Percentage conversion

The liquid oxidation of cyclohexane was investigated using various percentages of Mn/γ-Al₂O₃ and ozone 20 ± 1 °C and 1 atm of temperature and pressure, respectively. The GC results of the cyclohexane before oxidation, reaction mixture after 30 min oxidation and reaction mixture after 1 h oxidation are shown in Fig. 9. The results show that there is only one observed peak before oxidation corresponding to cyclohexane and are three peaks appearing after 30 min oxidation emanating from cyclohexane, cyclohexanol and cyclohexanone, whereas there are only two peaks observed after 1 h attributed to cyclohexane and cyclohexanone. The GC was equipped with MS to qualify the GC peaks and the MS spectra of cyclohexanol, cyclohexanone and cyclohexanone are displayed in Fig. S5. The reaction mixture was further characterized with FT-IR. Figure S6 displays FT-IR spectra of cyclohexane before oxidation, reaction mixture after 30 min oxidation and reaction mixture after 1 h oxidation. There, is no C=O and O–H peaks observed before oxidation. However, after 30 min both C=O and O–H peaks can be observed, whereas after 1 h only C=O peak is observed which is correlation with the GC-MS results. Figure 8 and Table 2 summarizes percentage conversions and selectivities of cyclohexane to cyclohexanol and cyclohexanone after 1 h of reaction time. The catalysts and the cyclohexane were left in the dark for 1 h to reach adsorption-desorption equilibrium. Prior to studying catalysed reactions, the blank oxidation was performed. The comparison of the percentage conversions between blank and pristine γ-Al₂O₃ catalysed reactions suggest that the pristine γ-Al₂O₃ is catalytically active in oxidation of cyclohexane. The percentage conversion obtained in blank oxidation was 9%, whereas 16% was recorded when γ-Al₂O₃ was used after 1 h. This was anticipated since γ-Al₂O₃ has been reported in literature to exhibit catalytic activity toward various processes [27, 28]. The percentage conversions obtained from the Mn doped γ-Al₂O₃ catalysed reactions suggest that the Mn functionalised γ-Al₂O₃ catalysts are more catalytically active than their unfunctionalized γ-Al₂O₃ counterpart.

Fig. 8:

Plots of (a) percentage conversion, (b) products selectivities of the blank, γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysed reactions.

Fig. 9:

Gas chromatograms of cyclohexane, (a) before oxidation, (b) after 30 min oxidation and (c) after 1 h oxidation using 2.5% Mn/γ-Al₂O₃.

The percentage conversions obtained using 2.5%, 5%, 7.5%, 10%, 12.5%, 15% Mn/γ-Al₂O₃ were 33%, 24%, 17%, 15%, 13%, 11%, respectively. The results show that 2.5% Mn/γ-Al₂O₃ exhibited higher catalytic activity than its analogous. The decrease in catalytic activity with increasing Mn loading in γ-Al₂O₃ can be explained by surface area effect. The BET results revealed that an increase in Mn loading resulted to a decrease in surface area of the catalysts. A decrease in surface area results to a limited number of the active sites, hence lower catalytic activity.

Percentage selectivity

Typically, the oxidation of cyclohexane proceeds through the free radical chain to yield the mixture of cyclohexanol and cyclohexanone (KA oil) and other side products such as hexanedioic acid, cyclohexyl hydroperoxide and dicyclohexyl adipate [29, 30]. The summary of the percentage selectivities of both uncatalysed (blank) and catalysed oxidation reactions obtained from GC-MS is shown in Fig. 8 and Table 2. The evaluation of the results shows that cyclohexanol and cyclohexanone were the only oxidation products identified with cyclohexanone being the major product after 30 min of oxidation. However, after 1 h only cyclohexanone was identified as the reaction product. This could be attributed to the ambient reaction constitutes and the short oxidation time applied in this study rather than the catalysts used. This is because the similar results were also observed for blank oxidation. In addition, from the results obtained it can be concluded that the formed cyclohexanol is further oxidised to cyclohexanone. This is because cyclohexanol was not identified after 1 h of oxidation. This is in good agreement with the literature [29], [30], [31]. The percentage selectivity obtained from blank oxidation were 14% and 86% to cyclohexanol and cyclohexanone, respectively at pH 3 after 30 min. The selectivities obtained using 2.5%, 5%, 7.5%, 10%, 12.5%, 15% Mn/γ-Al₂O₃ were 14%, 7%, 9%, 11%, 6%, 8%, 10% and 13% towards cyclohexanol and 86%, 93%, 91%, 89%, 94, 93%, 90% and 87% to cyclohexanone, respectively after 30 min oxidation. After 1 h, only cyclohexane and cyclohexanone were identified in GC-MS (Fig. 8) and FT-IR. The evaluation of the selectivity results shows that there is no observable pattern in percentage selectivity to both cyclohexanol and cyclohexanone with respect to Mn loading after 30 min, suggesting that Mn has no barring effect on the percentage selectivity.

Electron paramagnetic resonance studies

The hydroxide and carbon centred radicals quenching experiments were performed using electron paramagnetic resonance (EPR). The 5,5-dimethyl-1-pyrroline N-oxide (DMPO) and α-phenyl-tert-butyl nitrone (PBN) in DMSO were used as the ^•OH and carbon radicals trapping agents, respectively. The cyclohexane was ozonized in the presence of 2 mL of 0.1 M of DMPO and PBN separately for 5 min in the absence and the presence of the 0.2 g of the catalysts followed by analysing the reaction mixture for ^•OH signal using EPR. Figure 10 displays the EPR results of the blank ozonation, γ-Al₂O₃, 2.5% Mn/γ-Al₂O_3, 5% Mn/γ-Al₂O_3, 7.5% Mn/γ-Al₂O_3, 10% Mn/γ-Al₂O_3, 12.5% Mn/γ-Al₂O₃ and 15% Mn/γ-Al₂O₃. In the presence of DMPO, the results showed that the presence the catalysts play a profound role in the production of the ^•OH species. Typically, there were no ^•OH species generated in the absence of the catalysts. However, in the presence of pristine γ-Al₂O₃ weak ^•OH signals were observed. When 2.5% Mn/γ-Al₂O₃ catalyst was used instead, the ^•OH peak signal intensity was significantly induced, followed by gradually decrease with increasing Mn doping from 5% to 15%. These observations show that the pristine γ-Al₂O₃ and the Mn/γ-Al₂O₃ catalyses the decomposition of ozone to ^•OH species, however the 2.5% Mn/γ-Al₂O₃ catalyst is more catalytically active towards ozone decomposition to ^•OH species than its analogous.

Fig. 10:

EPR spectra of the blank ozonation, γ-Al₂O₃ and Mn/γ-Al₂O₃, catalysed reactions in the presence of (a) DMPO and (b) PBN.

The higher catalytic ozone decomposition aptitude of the 2.5% Mn/γ-Al₂O₃ catalyst than the γ-Al₂O₃ support can be visualised from the metal-support interaction synergy. The metal-supported catalysts possess synergistic effect which results to the development of an interface between the support and the transition metal. The interface possesses unique properties compared and often changes the properties of both the metal and the support material. Therefore, the electron transfer between the metal and the support occurs. This effect which promotes a decrease in energy required for the formation of oxygen vacancy, which subsequently promotes the formation of the oxygen vacancies. This was also confirmed by multiple oxidation states of the Mn in 2.5% Mn/γ-Al₂O₃ observed from the XPS results. This phenomenon enhances the Mars van Krevelen reaction mechanisms, catalytic robustness and reverse oxygen spill over of the catalyst. The higher catalytic ozone decomposition aptitude of the 2.5% Mn/γ-Al₂O₃ catalyst than its Mn/γ-Al₂O₃ analogous is due to its higher surface area elucidated with BET analysis.

Radical scavengers

In addition to EPR results, radical trapping studies were performed using tert-butanol and tribromochloromethane (CBr₃Cl) as hydroxide and carbon radicals’ quenchers, respectively. The tert-butanol is a strong hydroxide radical scavenger, which reacts with ^•OH to highly inert intermediates therefore terminating the ^•OH radicals chain reactions. Hence, in this study the generation of the ^•OH radicals in oxidation of cyclohexane was evaluated by addition of tert-butanol (0.1 mol L⁻¹, 1 mL). The results reveal that the addition of tert-butanol in catalytic oxidation of cyclohexane in the absence of the catalysts (blank) did not have much effect on oxidation efficiency (Fig. 11 and Table 3). The percentage conversions obtained in the absence of the catalysts were 5% and 8% after 30 min and 1 h, respectively. Nonetheless, when tert-butanol was added in the presence γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts, a significant decrease in cyclohexane percentage conversion was observed. Typically, the percentage conversion of 12% was obtained using 2.5% Mn/γ-Al₂O₃ catalyst upon addition of tert-butanol, which is lower than 33% obtained in the absence of tert-butanol. This phenomenon suggest that the ^•OH radicals play an important role in oxidation of cyclohexane. In addition, the results reveal that cyclohexane oxidation in the absence of the catalyst is not instigated by ^•OH radicals, since the addition of tert-butanol did not affect the oxidation efficiency. This suggests that in the absence of the catalyst the oxidation transpires through 1,3 dipolar insertion of the ozone to cyclohexane. This has been reported in literature [32, 33].

Fig. 11:

Plots of (a) and (c) percentage conversion, (b) and (d) products selectivevities of the blank, γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysed reactions.

The addition of the CBr₃Cl (0.1 mol L⁻¹, 1 mL) as the carbon centred radical scavenger revealed that the oxidation of cyclohexane does not encompass the production of the carbon centred radical. This is because there was no observable effect in oxidation efficiency upon the addition of the CBr₃Cl. The results are in correspondence with the EPR results, since there were no EPR signals observed for carbon centred radicals. Nonetheless, when the CBr₃Cl was added in γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysed reactions, a noticeable decrease in percentage conversions were perceived, suggesting that the carbon centred radicals were formed in γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysed cyclohexane reactions. This was also perceived in EPR studies.