Structure–activity correlation in aerobic cyclohexene oxidation and peroxide decomposition over Co_xFe_3−xO₄ spinel oxides

Abstract

Nanoparticulate Co_xFe_3−xO₄ (0 ≤ x ≤ 3) catalysts were prepared by spray-flame synthesis and applied in liquid-phase cyclohexene oxidation with O₂ as oxidant. The catalysts were characterized in detail using N₂ physisorption, XRD, TEM, XPS, FTIR, Raman, and Mössbauer spectroscopy. A volcano plot was obtained for the catalytic activity in cyclohexene oxidation as a function of the Co content with a maximum at x = 1. Thus, CoFe₂O₄ achieved the highest degree of cyclohexene conversion and the fastest decomposition rate of the key intermediate 2-cyclohexene-1-hydroperoxide. Kinetic studies and a stability test were performed over CoFe₂O₄, showing that cyclohexene oxidation follows first-order kinetics with an apparent activation energy of 58 kJ mol⁻¹. The catalytic hydroperoxide decomposition during cyclohexene oxidation was further investigated using H₂O₂ and tert-butyl hydroperoxide as simpler surrogates resulting in similar volcano-type correlations. The increase in catalytic activity with increasing Fe content with a maximum at x = 1 is ascribed to the increasing concentration of octahedrally coordinated Co²⁺ cations in the spinel structure leading to the presence of coordinatively unsaturated Co_3c²⁺ surface sites, which are identified to be the most active sites for 2-cyclohexene-1-hydroperoxide decomposition in cyclohexene oxidation.

---

Introduction

The oxidation of olefins is a key technology in the chemical industry, as the formed products are important building blocks for the synthesis of fine and bulk chemicals, comprising agrochemicals, pharmaceuticals, fragrances, and polymers.^1,2 In particular, the oxidation of cyclohexene provides access to valuable oxygenated products, such as 2-cyclohexene-1-one, 2-cyclohexene-1-ol, cyclohexane-1,2-diol, and cyclohexene oxide.^2,3 Industrial processes are mainly conducted in the gas phase under harsh reaction conditions and operated at low conversion to retain high selectivity and to suppress total oxidation. In contrast, liquid-phase reactions can be carried out under milder reaction conditions leading to higher selectivities. Nevertheless, aerobic cyclohexene oxidation is subject to a complex reaction network because of the two active centres in the cyclohexene molecule (Scheme 1). When allylic oxidation takes place at the C–H bond, cyclohexene is oxidized to 2-cyclohexene-1-hydroperoxide, which can further react to 2-cyclohexene-1-ol, 2-cyclohexene-1-one and 7-oxabicyclo[4.1.0]-heptane-2-one. On the other hand, epoxidation can occur at the olefinic C=C bond, yielding cyclohexene oxide and cyclohexane-1,2-diol. Moreover, both reaction pathways cannot be clearly separated, because 2-cyclohexene-1-hydroperoxide formed within the allylic pathway can react with another cyclohexene molecule resulting in the formation of cyclohexene oxide, which is typically referred to the epoxidation pathway. Early studies identified 2-cyclohexene-1-hydroperoxide to be an unstable reaction intermediate of cyclohexene oxidation. Already in 1939, Criegee et al. detected 2-cyclohexene-1-hydroperoxide to be the main product of cyclohexene autoxidation and clarified its structure.⁴ Sridhar and coworkers investigated the uncatalyzed decomposition of 2-cyclohexene-1-hydroperoxide and found 2-cyclohexene-1-ol and 2-cyclohexene-1-one to be the main products even in the absence of O₂.⁵ It can therefore be deduced that the aerobic oxidation of cyclohexene at least partially occurs autocatalytically forming 2-cyclohexene-1-hydroperoxide with high selectivity.

Scheme 1 Reaction network of cyclohexene oxidation.

The autoxidation of cyclohexene renders the decomposition of the hydroperoxide intermediate to be the key step for selective cyclohexene oxidation. Thus, the use of a heterogeneous catalyst offers the possibility to selectively decompose 2-cyclohexene-1-hydroperoxide into one of the desired products, enabling the formation of high-value chemicals in high yields.

For this purpose, highly active catalysts are needed. Transition metal oxides have raised much interest because of their low cost and their potential to replace conventional noble metal catalysts. Especially, oxides with spinel structure were found to be highly stable under extreme reaction conditions and flexible with regard to the variety of metal ions with different valencies incorporated into their lattices.⁶ Spinel oxides consist of a close-packed cubic lattice of oxygen anions with the general formula AB₂O₄, where A can be a group IIA metal or a transition metal cation with the oxidation state 2+ and B is a group IIIA metal or transition metal cation in the 3+ oxidation state. In a normal spinel structure like ideal Co₃O₄, the A²⁺ ions occupy one-eighth of the tetrahedral holes and the 3+ species occupy half of the octahedral voids. In contrast, within an inverse spinel such as ideal CoFe₂O₄, the A²⁺ ions and half of the B³⁺ ions exchanged their positions.⁷

In addition to the occupation of octahedral and tetrahedral sites within the spinel structure, the particle size and the resulting facets of the particles also play an important role. While it is generally accepted that conversion increases with increasing specific surface area, our recent studies show that larger Co₃O₄ particles with consequently smaller surface areas can nevertheless lead to higher conversion due to the exposure of catalytically more active facets with, for example, multiple coordinatively unsaturated Co_cus³⁺ sites.⁸

Because of their high stability, their large variety of materials and also semiconducting properties, spinel oxides are applied in many fields, such as Li-ion batteries,⁹ photodegradation of pollutants,¹⁰ water oxidation,¹¹ and electrochemical oxygen evolution reaction.¹² In addition, Co-based transition metal oxides have already been applied in liquid-phase oxidation reactions in our previous studies.^13–17

Spinel oxides can be synthesized by several methods, such as mixed powder synthesis,⁷ co-precipitation,¹⁸ sol–gel synthesis,¹⁹ hydrothermal synthesis,²⁰ and spray and freeze drying.²¹ Among these methods, the continuous and scalable spray-flame synthesis is a suitable method providing high specific surface areas, small nanoparticle sizes, and well-controlled catalyst compositions.^13,22

In the present work, we report on the liquid-phase oxidation of cyclohexene in acetonitrile with molecular O₂ over a series of spray-flame synthesized Co_xFe_3−xO₄ spinel nanoparticles under mild conditions. To date, to the best of our knowledge, no comparable study has been reported on the liquid-phase cyclohexene oxidation over mixed Co–Fe spinel oxides. We aim at understanding the effect of Fe substitution into the Co₃O₄ spinel structure on the reaction mechanism of cyclohexene oxidation. The Co_xFe_3−xO₄ nanoparticulate materials were characterized in detail, and their catalytic properties were systematically elaborated. Kinetic investigations, a reusability test and the influence of H₂O were studied over the best-performing catalyst CoFe₂O₄. The decomposition of H₂O₂ and tert-butyl hydroperoxide (TBHP) was carried out to mimic the decomposition of the key intermediate 2-cyclohexene-1-hydroperoxide to correlate the rates of cyclohexene oxidation and peroxide decomposition and, thereby, identify the active sites.

Results and discussion

Catalyst characterization

A series of Co_xFe_3−xO₄ (x = 0, 0.5, 1, 1.5, 2, 2.5, 3) nanoparticulate materials was synthesized by spray-flame synthesis using dissolved metal nitrates as precursors with a concentration of 0.2 M. The detailed synthesis procedure can be found elsewhere.^13,23 After synthesis, the samples were heat-treated at 250 °C for 2 h in air to remove adsorbed organic and water residuals from the particles surface.

The N₂ physisorption-derived specific surface areas determined by applying the BET equation were high, ranging between 87 and 151 m² g⁻¹ (Table 1). Assuming monodisperse and spherical particles, average particle sizes of 8.3 to 12.0 nm were calculated (Table 1). The count mean diameters based on TEM images resulted in particle sizes between 8.4 and 11.1 nm with homogeneous particle size distributions, which are in agreement with the BET-based particle sizes (Table 1). Nevertheless, according to the analyzed TEM images, besides spherical particles, a reasonable number of faceted particles were identified (Fig. S1†). All catalysts show a single crystalline nanostructure with high crystallinity but crystallographically disordered structures (Fig. 1 and S1†). The EDX elemental mapping identified a homogeneous distribution of Co and Fe for all samples (Fig. S3†) and on average, the experimental Co/Fe ratios were similar to the desired nominal ones summarized in Table 1.

Table 1 Specific surface areas, particle sizes determined by N₂ physisorption and TEM, bulk and surface Co/(Co + Fe) ratios based on EDX and XPS analyses of the Co_xFe_3−xO₄ samples

Sample	Specific surface area	Particle size^a	Particle size^b	Co/(Co + Fe)^c	x(CoxFe3−xO4)^c	Co/(Co + Fe)^d	x(CoxFe3−xO4)^d
	[m^2 g^−1]	[nm]	[nm]
a Estimated from the specific surface areas assuming spherical particles. b Determined by TEM. c Determined by EDX. d Determined by XPS.
Co3O4	118	8.3	11.1	1	3	1	3
Co2.5Fe0.5O4	87	12.0	8.8	0.82	2.47	0.84	2.51
Co2Fe1O4	128	8.3	8.8	0.67	2.01	0.68	2.04
Co1.5Fe1.5O4	151	7.5	8.4	0.51	1.54	0.52	1.55
Co1Fe2O4	129	8.8	8.7	0.33	0.99	0.36	1.07
Co0.5Fe2.5O4	122	9.4	9.9	0.18	0.53	0.18	0.55
γ-Fe2O3	111	10.5	9.2	0	0	0	0

---

Fig. 1 TEM and HAADF-STEM images of CoFe₂O₄ with lattice spacings.

The HAADF-TEM images of Co₃O₄ and CoFe₂O₄ were used to determine the lattice spacings of the spinel nanoparticles. For Co₃O₄, a distance of d = 0.234 nm was recorded in agreement with the theoretical lattice spacing of d₁₀₀ = 0.24 nm (Fig. S2†).²⁴ For CoFe₂O₄, a significantly higher number of hexagonal shaped particles was found, indicating (111) surfaces.²⁵ The determined lattice spacings of 0.509, 0.476 and 0.480 nm agree with the theoretical value of d₁₁₁ = 0.484 nm, suggesting that (111) surfaces are dominant for CoFe₂O₄.²⁶ Whereas (100)-oriented surfaces expose coordinatively unsaturated metal cations, which are fivefold coordinated at the octahedral sites of the spinel (M^O_5c), those octahedrally coordinated metal cations are only threefold coordinated at the (111) surfaces (M^O_5c).²⁵ Metal cations with a lower coordination number are assumed to be more catalytically active.

XRD patterns were recorded for the Co_xFe_3−xO₄ catalysts (Fig. 2). A cubic spinel structure was identified for all samples. For Co₃O₄, the pattern fits better to the reference pattern of a cobalt-deficient spinel phase than to stoichiometric Co₃O₄. With increasing Fe content, the reflections start shifting to lower 2θ angles as expected. Based on the XRD patterns, the iron oxide sample (x = 0) cannot be clearly identified, because the magnetite (Fe₃O₄) and maghemite (γ-Fe₂O₃) phases have both almost identical lattice parameters. Crystal structure analysis was performed by Rietveld refinement resulting in lattice parameters, which amount to 8.08, 8.38 and 8.36 Å, well-fitting to the literature-reported values of 8.08,²⁷ 8.38,²⁸ and 8.35 Å (ref. 29) for Co₃O₄, CoFe₂O₄ and γ-Fe₂O₃, respectively. The reported lattice constant of Fe₃O₄ is 8.39 Å,²⁹ indicating the presence of γ-Fe₂O₃ instead of Fe₃O₄ (Fig. S4†).

Fig. 2 XRD patterns of the Co_xFe_3−xO₄ (x = 0, 0.5, 1, 1.5, 2, 2.5, 3) samples and corresponding reference patterns taken from the ICSD database (Co₃O₄: ICSD 36256, CoFe₂O₄: ICSD 109044, Fe₃O₄: ICSD 26410, γ-Fe₂O₃: ICSD 172906).

Raman spectroscopy was applied to investigate the electronic structure and the cation occupation of the tetrahedral and octahedral sites in the spinel structure of the Co_xFe_3−xO₄ samples (Fig. 3). The recorded spectra clearly show five bands, which can be assigned to A_1g, E_g and three F_2g phonon modes originating from lattice vibrations of the cubic spinel structure. With increasing Fe amount in the samples from Co₃O₄ to Co_1.5Fe_1.5O₄, a redshift of the A_1g mode was observed corresponding to the symmetric Co³⁺–O stretching vibration of the octahedral sites, which indicates a lower occupancy of Co³⁺ ions at octahedral sites and a starting phase transition towards an inverse spinel.^30,31 The intensity of the F_2g(3) mode at 195 cm⁻¹ also decreased, indicating the lower occupation of tetrahedral voids by Co²⁺ ions.³² Similarly, the intensity of the E_g mode at 481 cm⁻¹, corresponding to different motions of the AO₄-tetrahedral unit,³¹ also decreased. The F_2g(3) mode at 475 cm⁻¹ of Co_0.5Fe_2.5O₄ and CoFe₂O₄ corresponds to the vibration of Co²⁺ ions in octahedral voids,³³ while the A_1g(1) mode (688 cm⁻¹) of both samples corresponds to the symmetric stretching of oxygen atoms with respect to Fe³⁺ cations in tetrahedral voids.³⁴ The Raman spectrum of the iron oxide sample indicates the formation of the Fe₃O₄ and the γ-Fe₂O₃ phase,²⁸ which is a Fe²⁺-vacant spinel structure. It is deduced that the samples CoFe₂O₄, Co_0.5Fe_2.5O₄ and γ-Fe₂O₃ have a predominantly inverse spinel structure.

Fig. 3 Raman spectra of the Co_xFe_3−xO₄ (x = 0, 0.5, 1, 1.5, 2, 2.5, 3) samples.

FTIR spectra were recorded in the frequency range between 800 and 400 cm⁻¹, reflecting fundamental vibrations of metal–oxygen bonds in crystalline solids (Fig. S5†). The two bands for normal Co₃O₄ at 656 and 552 cm⁻¹ fit to previously reported values corresponding to Co²⁺ in tetrahedral and octahedral voids, respectively.³⁵ The band at 541 cm⁻¹ for inverse γ-Fe₂O₃ is assigned to tetrahedrally coordinated Fe³⁺, while the band at 381 cm⁻¹ originates from octahedral Fe³⁺.^36,37 The shifting and vanishing of the recorded signals shows the transition from normal to inverse spinel with increasing Fe amount incorporated into the Co₃O₄ spinel structure.

The Co_xFe_3−xO₄ samples were further characterized using X-ray photoelectron spectroscopy (XPS) to examine the oxidation states and the surface composition of the catalysts. The Co 2p region in Fig. 4 shows the Co 2p_1/2 and Co 2p_3/2 signals at 794.8 and 779.7 eV, respectively.³⁸ The intense shake-up satellite peak at 787 eV can be assigned to Co²⁺, which was increasing with increasing Fe amount. Co³⁺only exhibits a weak satellite peak at 790 eV indicating a decrease in Co³⁺ concentration with increasing Fe amount.¹⁴ For the samples Co_0.5Fe_2.5O₄ and CoFe₂O₄, only Co²⁺ was identified as expected, whereas the other samples contain mixed Co²⁺/Co³⁺ oxidation states. For Co₃O₄, a value of 60% Co²⁺ was determined with respect to the total Co amount on the catalysts' surface. Fig. 4 displays the Fe 2p_1/2 and 2p_3/2 peaks at 724.1 and 710.7 eV, respectively. The peak intensities gradually increased with increasing Fe amount in the Co_xFe_3−xO₄ samples. The positions of the 2p_3/2 peak and the satellite peak at 718.4 eV indicate the presence of Fe³⁺ for all samples. For Fe²⁺, both peaks would be shifted to lower binding energies.^14,39 The C 1s region shows a distinct peak at 284.5 eV, which corresponds to C–C bonds and was used as a reference for calibration. The O 1s peak at 529.5 eV can be fitted by contributions of bulk and surface O (Fig. S6†). The surface composition of the Co_xFe_3−xO₄ samples is summarized in Table S1.† The calculated surface Co/Fe ratios based on the XPS measurements agree with the bulk Co/Fe ratio determined by EDX, indicating similar bulk and surface compositions for all samples (Table 1).

Fig. 4 XP spectra of the Co 2p and Fe 2p of the Co_xFe_3−xO₄ (x = 0, 0.5, 1, 1.5, 2, 2.5, 3) samples.

Mössbauer spectroscopy measurements at 4.3 K were performed to additionally identify and verify the oxidation state of Fe in the Co_xFe_3−xO₄ samples (Fig. S7†) with the applied methodology for evaluation of experimental spectra being illustrated in the ESI† in detail. A magnetically ordered sextet structure was recorded, which can be reproduced sufficiently for most of the samples by two subspectra representing the spinel tetrahedral A- and octahedral B-site. The sites were identified by their characteristic isomer shifts of δ ≈ 0.37 mm s⁻¹ for the A-site and δ ≈ 0.47 mm s⁻¹ for the B-site relative to α-Fe at room temperature.^40,41 As no third subspectrum, which could be assigned to Fe²⁺ ions, was observed for any sample, only Fe³⁺ was identified in agreement with the XPS results. While for x ≥ 1 no Fe²⁺ is to be expected due to the presence of Co²⁺, the absence of Fe²⁺ for x = 0 indicates that this sample is oxidized to maghemite (γ-Fe₂O₃). Upon rising the Co-fraction x, a gradual change in the structure of the Mössbauer spectra can be observed, resulting in lower average hyperfine magnetic fields B_hf visible in Fig. S7A,† as well as increasing spin canting due to the enhanced magnetic anisotropy of CoFe₂O₄ and the trend towards antiferromagnetic ordering of Co₃O₄ for higher values of x (Fig. S7B†). Fe ion site occupation can be extracted directly from the relative A- and B-site subspectral intensities and thereby, the distribution of Co²⁺ ions on the different lattice positions can be inferred up to x = 3 when assuming that Co³⁺ is placed on octahedral positions exclusively, as illustrated in the ESI.† Following this approach, inversion parameters in the range of ca. 0.75 were found for x ≲ 1, pointing towards the tendency of Co²⁺ to be placed preferably on octahedral sites, corresponding more closely to the inverse spinel structure. For x > 1 the inversion parameter gradually decreases, as Co³⁺ is placed on B-site positions, reaching a minimum value of ca. 0.25 for x = 2.5 when approaching the normal spinel structure of Co₃O₄.⁴² Here, one has to consider that for x ≳ 1.5, due to the increasing spin frustration, applying the external field no longer yields better subspectral resolution. Still, even from zero field spectra, general trends in site occupancy can be extracted to some degree, displaying an evolution in the inversion parameter consistent with results from Raman spectroscopy.

In summary, we successfully synthesized a series of Co_xFe_3−xO₄ catalysts with chemical compositions close to the nominal ones, high specific surface areas ranging from 87 to 151 m² g⁻¹, and small particle sizes between 8.3 and 12.0 nm. XRD measurements identified phase-pure samples of a cubic spinel structure. For x = 0, Rietveld refinement and Raman spectroscopy indicated the presence of the Fe²⁺-vacant structure γ-Fe₂O₃. XPS and Mössbauer measurements both demonstrated the absence of Fe²⁺ species in all samples. Raman and Mössbauer spectroscopy showed the tendency of Co₃O₄, Co_2.5Fe_0.5O₄ and Co₂FeO₄ towards the normal spinel phase, whereas CoFe₂O₄, Co_0.5Fe_2.5O₄ and γ-Fe₂O₃ were found to have a predominantly inverse spinel structure. The lattice spacing of CoFe₂O₄ indicates a dominant (111) surface, exposing coordinatively unsaturated M^O_5c metal sites.

Cyclohexene oxidation over Co_xFe_3−xO₄

The catalytic properties of the Co_xFe_3−xO₄ nanoparticulate spinel catalysts in cyclohexene oxidation were studied under mild reaction conditions in acetonitrile using molecular O₂ as oxidizing agent. Initial experiments on the variation of the stirring speed showed that a speed of agitation of 600 rpm is suitable to exclude external mass transfer limitation (Fig. S8†).

After 6 h, all catalysts revealed a moderate to high catalytic activity with degrees of conversion ranging from 44 to 69% (Fig. S9†), which is comparable to the degree of conversion (75%) over spray-flame synthesized LaCoO₃ perovskite oxides.^13,16 The comparison of Co₃O₄ and γ-Fe₂O₃ clearly shows the superior activity of the Co-containing catalyst after 6 h. In cyclohexene oxidation, Co₃O₄ achieved a significantly enhanced conversion as well as a lower selectivity to 2-cyclohexene-1-hydroperoxide. These results indicate a faster hydroperoxide decomposition, leading to an increase in 2-cyclohexene-1-one selectivity. In contrast, cyclohexene oxidation over γ-Fe₂O₃ was characterized by a slow hydroperoxide decomposition, so that the ketone selectivity is lowered by 12% in comparison to Co₃O₄.

The comparison of the Co_xFe_3−xO₄ catalysts containing both Co and Fe clearly shows a positive effect of Fe substitution on the initial catalytic activity (Fig. 5). With increasing Fe amount up to CoFe₂O₄ (x = 1), the degree of conversion gradually increased from 14.3 to 23.0% after 0.5 h, whereas a blank reaction resulted in only 4.8% conversion (Fig. S10†). The increase in conversion was accompanied by a decreased hydroperoxide selectivity by 24.5%, indicating a faster decomposition, which results in higher product yields. However, the product selectivity was not affected, leading to the assumption that the catalyst is mainly involved in the formation and decomposition of the intermediate 2-cyclohexene-1-hydroperoxide. When the Fe content is further increased beyond CoFe₂O₄ (x < 1), the catalytic activity dropped as shown by the low degree of conversion of 14.4%, and the hydroperoxide selectivity became the highest within the whole catalyst series at 75.5%, identifying γ-Fe₂O₃ to be least active for cyclohexene oxidation.

Fig. 5 The influence of Fe substitution into Co_xFe_3−xO₄ spinel catalysts on cyclohexene conversion, product selectivity and initial reaction rate. Reaction conditions: 20 mmol cyclohexene, 30 mL acetonitrile, 50 mg catalyst, 80 °C, 10 bar O₂, 600 rpm, 0.5 h.

The correlation of specific surface areas and particle sizes of the investigated catalysts with their catalytic activity excludes strong microstructural effects (Fig. S11†).

The comparison with the catalytic activity of spray-flame-synthesized LaCoO₃ nanoparticles for cyclohexene oxidation we presented in a previous study indicates a comparable catalytic activity of the herein presented CoFe₂O₄ nanoparticles but a lower selectivity to 2-cyclohexene-1-hydroperoxide.¹³ Compared to other Co-based catalysts presented in literature reports, our CoFe₂O₄ achieves significantly higher degrees of conversion at similar or shorter reaction times (Table S2†).

The CoFe₂O₄ catalyst was chosen for further kinetic investigations, as it had the highest catalytic activity in cyclohexene oxidation with molecular O₂. As expected, a strong influence of temperature on cyclohexene oxidation was observed (Fig. S12†). Cyclohexene conversion was limited to 38.7% at 60 °C, whereas at 100 °C nearly full conversion of 91.4% was recorded after 6 h. Additionally, hydroperoxide decomposition strongly increased at higher temperatures, resulting in its complete decomposition at 100 °C. Thus, product selectivity is enhanced and 2-cyclohexene-1-one becomes the main product of cyclohexene oxidation with a selectivity of 28.2%. Meanwhile, 2-cyclohexene-1-ol selectivity decreased, indicating the further oxidation of 2-cyclohexene-1-ol to the corresponding ketone. This led to strongly increasing ketone/alcohol ratios, which amount to 1.9 at 60 °C and 7.6 at 100 °C. However, the product distribution within the epoxidation products was not that strongly affected by temperature, as the cyclohexene-1,2-diol/cyclohexene oxide ratio only slightly decreased from 3.2 to 2.5. Nevertheless, higher temperatures shifted cyclohexene oxidation to the epoxidation pathway, as the ratio of allylic/epoxidation products decreased from 3.2 to 1.3. The catalytic data at different temperatures were used for the Arrhenius analysis. The degrees of cyclohexene conversion at different temperatures as a function of time are shown in Fig. 6A. Based on previous studies, cyclohexene oxidation over transition metal oxide catalysts is assumed to follow first-order reaction kinetics.^13,16 The linearized plots based on first-order reaction kinetics were well suited for linear regression (R² > 0.98), suggesting that cyclohexene oxidation over CoFe₂O₄ follows first-order reaction kinetics (Fig. S12D†).

Fig. 6 Kinetic investigations of cyclohexene oxidation over CoFe₂O₄. (A) Cyclohexene conversion as a function of time at different temperatures, and (B) the Arrhenius plot based on first-order kinetics. The rate constants k were determined according to Fig. S12D.† Reaction conditions: 20 mmol cyclohexene, 30 mL acetonitrile, 50 mg catalyst, 10 bar O₂, 600 rpm, 6 h.

The Arrhenius analysis resulted in an apparent activation energy of 58.1 kJ mol⁻¹ (Fig. 6B). A reusability test was carried out to investigate the stability of the CoFe₂O₄ catalyst in terms of cyclohexene conversion and product selectivity (Fig. S13†). Four oxidation runs were performed, resulting in high and essentially equal degrees of cyclohexene conversion between 66.9 and 68.0%, which is in the range of experimental error and identifies CoFe₂O₄ to be a highly reusable catalyst under these mild oxidation conditions. Only a slight decrease in hydroperoxide decomposition by 5% was observed after the second run, resulting in a lower selectivity to 2-cyclohexene-1-one, indicating that the catalyst plays a crucial role in hydroperoxide decomposition. After the reusability test, a similar XRD pattern of phase-pure spinel structure of high crystallinity (Fig. S14†) and similar TEM images (Fig. S15†) were recorded, confirming the high stability of CoFe₂O₄ during cyclohexene oxidation.

As water is produced as a by-product in several reaction steps of cyclohexene oxidation, its influence on the catalytic activity was investigated by adding water to the initial reaction mixture (Fig. S16†). The initial reaction rate of cyclohexene oxidation was clearly suppressed. Nevertheless, cyclohexene conversion was only lowered by 5% after 6 h, and the hydroperoxide selectivity increased at low reaction times but reached a comparable value after 6 h. This was observed for almost all reaction products, indicating that H₂O may block the active sites of CoFe₂O₄ at the beginning of the reaction, but the adsorption of the reactant is favored.

Mimicking 2-cyclohexene-1-hydroperoxide decomposition

Cyclohexene oxidation underlies an enormously complex reaction network, in which 2-cyclohexene-1-hydroperoxide plays a key role. The decomposition of 2-cyclohexene-1-hydroperoxide can lead to the formation of several reaction products (Scheme 1). Thus, the heterogeneously catalyzed decomposition of the hydroperoxide intermediate is the main target for selective cyclohexene oxidation. Within cyclohexene oxidation reactions, the hydroperoxide decomposition rate may be influenced by many factors, as the reaction mechanism and several correlations within the reaction network have not been fully clarified yet. Therefore, we implemented less complex peroxide decomposition reactions to mimic the hydroperoxide decomposition kinetics during cyclohexene oxidation over the Co_xFe_3−xO₄ catalysts using H₂O₂ and TBHP as surrogates (Fig. S17†). Fig. S18† shows the evolved O₂ volume as a function of time for both reactions. To avoid microstructural effects, the initial reaction rates normalized to the surface areas of the catalysts were calculated (Fig. 7). For H₂O₂ and TBHP decomposition, CoFe₂O₄ exhibited the highest reaction rates, whereas pure Co₃O₄ and γ-Fe₂O₃ showed a low catalytic activity, which is in agreement with the catalytic results of cyclohexene oxidation. Thus, these results indicate a synergistic effect of Co and Fe in Co_xFe_3−xO₄ nanoparticles, which improves the decomposition rates of hydroperoxides such as 2-cyclohexene-1-hydroperoxide, TBHP and H₂O₂, resulting in the following activity ranking of the catalysts:

Fig. 7 (A) Volcano plots as a function of Co content for cyclohexene conversion and selectivity to 2-cyclohexene-1-hydroperoxide after 0.5 h, (B) initial reaction rates of H₂O₂ decomposition after 3 min and (C) initial reaction rates of TBHP decomposition after 5 min normalized to the surfaces areas of the Co_xFe_3−xO₄ catalysts.

CHHP:	Co1>	Co1.5>	Co2>	Co2.5>	Co3>	Co0.5>	Co0
H2O2:	Co1>	Co1.5>	Co2>	Co2.5>	Co3>	Co0.5>	Co0
TBHP:	Co1>	Co1.5≈	Co0.5>	Co2>	Co0>	Co2.5>	Co3

The Co_xFe_3−xO₄ catalysts show an equal order of activity for 2-cyclohexene-1-hydroperoxide and H₂O₂ decomposition with CoFe₂O₄ as the most active sample. Both reactions exhibited a drastic decrease in activity over the samples containing a higher Fe amount than CoFe₂O₄ (x < 1). This activity drop was less pronounced in TBHP decomposition, resulting in a slightly different activity order, but the overall trend on TBHP decomposition is very similar to 2-cyclohexene-1-hydroperoxide and H₂O₂ decomposition (Fig. 7).

The rate constants for H₂O₂ decomposition over the Co_xFe_3−xO₄ series are summarized in Table S3,† again identifying CoFe₂O₄ to be the most active catalyst with a rate constant of k = 0.679 s⁻¹. The comparison with CoFe₂O₄ spinel samples from literature reports clearly shows a superior catalytic activity of the spray-flame synthesized CoFe₂O₄ sample (Table S4†). Cota et al.⁴³ and Goldstein et al.,⁴⁴ who both had a leading role in the research on peroxide decomposition in the last century, prepared CoFe₂O₄ samples enabling surface-area-normalized rate constants of k_s = 0.071 s⁻¹ m⁻² and k_s = 0.120 s⁻¹ m⁻² in alkaline solution, respectively, which are significantly lower compared with our catalyst (k_s = 0.527 s⁻¹ m⁻²). Onuchukwu et al.⁴⁵ also studied H₂O₂ decomposition in alkaline solution over CoFe₂O₄ synthesized by co-precipitation and obtained a mass-normalized rate constant of k_m = 14 s⁻¹ g⁻¹, which is one fifth of the spray-flame synthesized CoFe₂O₄ (k_m = 67.9 s⁻¹ g⁻¹). Mimani et al.⁴⁶ synthesized CoFe₂O₄ by a low-temperature decomposition approach and obtained a reaction rate constant of k_s = 0.232 s⁻¹ m⁻², while CoFe₂O₄ materials synthesized by co-precipitation and sol–gel method led to significantly lower reaction rate constants of k_m = 0.037 s⁻¹ g⁻¹ and k_s = 0.04 s⁻¹ m⁻², respectively.^47,48

Interestingly, the H₂O₂ decomposition experiments using Co₃O₄ and CoFe₂O₄ synthesized by a precursor decomposition approach and direct co-precipitation revealed Co₃O₄ to be more active than CoFe₂O₄.¹⁷ This observation can be attributed to the platelet-like structure and the significantly smaller specific surface areas (∼30 m² g⁻¹) of these catalysts. Furthermore, unlike for the herein presented Co₃O₄, the surfaces of the Co₃O₄ sample in literature synthesized by co-precipitation were (111)-oriented.

Earlier studies on H₂O₂ decomposition over Co- and Fe-based spinels showed that it is not the incorporation of Fe, but rather the change of cation distribution within the spinel structure playing a crucial role in the high catalytic activity of Co–Fe mixed oxide catalysts. This assumption is supported by the low activity over γ-Fe₂O₃. Table S5† shows the theoretical cation valency distribution and spinel type of the prepared samples based on the crystal field stabilization energy of the transition metal ions.⁴⁹ With increasing Fe amount incorporated in the spinel structure, a change from normal to inverse spinel and an increase of Co²⁺ ions located at octahedral voids is predicted. This was corroborated by our experimental results from Raman and Mössbauer spectroscopy, showing a clear trend towards a predominantly inverse structure with increasing Fe content.

Anthony and Onuchukwu⁵⁰ investigated H₂O₂ decomposition over two series of Ni_xFe_3−xO₄ and Cu_xFe_3−xO₄ catalysts. The authors highlighted the importance of the electron exchange by redox couples and were not able to demonstrate any significant influence of Fe present in the spinel samples. Similarly, Onuchukwu and Zuru⁴⁵ prepared a Co_xFe_3−xO₄ spinel series by co-precipitation and investigated its catalytic activity in H₂O₂ decomposition. They found Co_1.5Fe_1.5O₄ to be the best catalyst and explained this observation by the presence of each 0.5 mol Co²⁺ and Co³⁺ per formula unit in the octahedral voids of the spinel, enabling an optimal electron exchange between the redox couple within the octahedral sites.⁴⁵ However, H₂O₂ decomposition was carried out in a highly alkaline environment. The present study identifies CoFe₂O₄ to be the most active catalyst in acetonitrile, which does not contain the Co²⁺/Co³⁺ redox couple at the octahedral sites. Mimani and Patil⁴⁶ found cobaltites to significantly better decompose H₂O₂ than ferrites, identifying the importance of Co species for H₂O₂ decomposition.

Goldstein and Tseung⁵¹ also investigated the complete series of Co_xFe_3−xO₄ in H₂O₂ decomposition with 0 ≤ x ≤ 3 under basic conditions and proposed a ranking of redox couples:

[Co²⁺–Co³⁺]_O > [Co²⁺–Fe³⁺]_O > (Co²⁺)_T ≫ [Fe²⁺–Fe³⁺]_O > (Fe²⁺)_T

Based on these results, the superior activity of Co₃O₄ compared with Fe₃O₄ becomes clear, as Co²⁺ ions in tetrahedral voids are much more active than Fe²⁺ ions in octahedral voids. [Co²⁺]_O exhibited the highest activity in H₂O₂ decomposition. Moreover, a higher activity of [Co²⁺]_O was found when the adjacent octahedral site ion was Co³⁺ instead of Fe³⁺. Although 0.5 mol redox couple of Co²⁺/Co³⁺ is present in the octahedral voids in Co_1.5Fe_1.5O₄, 1 mol [Co²⁺]_O per formula unit is present in the spinel structure of CoFe₂O₄ resulting in the highest concentration of [Co²⁺]_O in the whole Co_xFe_3−xO₄ series justifying its outstanding catalytic activity. The samples x = 1.5 to x = 3 all have the same amount of 1 mol Co²⁺ ions per formula unit present in the spinel structure, but 0.5 mol Co²⁺ is present in octahedral voids for x = 1.5. Thus, this sample should have a higher activity, whereas the samples x = 2, x = 2.5 and x = 3 should exhibit similar activities (Table S5†). The comparison of Co₃O₄ (x = 3) containing 1 mol (Co²⁺)_T with Co_0.5Fe_2.5O₄ (x = 0.5) exhibiting 0.5 mol [Co²⁺]_O suggests Co₃O₄ to be more active than Co_0.5Fe_2.5O₄, whereas Fe₃O₄ (x = 0) is expected to show the lowest activity in terms of H₂O₂ decomposition as it contains no Co²⁺ ions. Based on these considerations, Goldstein and Tseung⁴⁴ proposed the following activity ranking for the Co_xFe_3−xO₄ samples:

Co₁ > Co_1.5 > Co₂ = Co_2.5 = Co₃ > Co_0.5 > Co₀

This order is fully consistent with the obtained order of the rate constants as a function of the catalyst composition for H₂O₂ and 2-cyclohexene-1-hydroperoxide decomposition (Fig. 7). For TBHP decomposition, the activity order based on the produced O₂ volume is consistent, too, but the area-normalized rate clearly shows deviating results suggesting additional influences on TBHP decomposition besides the catalyst composition. Thus, the high catalytic activity of CoFe₂O₄ can be presumably referred to the high concentration of octahedrally coordinated Co²⁺ cations, so that the incorporation of Fe leads to a beneficial change of the cation distribution within the spinel structure, while the Fe cations themselves might play a minor role for the catalytic activity.

The reaction mechanism of H₂O₂ decomposition by ferrous and ferric salts was investigated by Haber and Weiss in 1932 and extended by Barb in 1949.^52,53 They proposed H₂O₂ decomposition to be a radical chain reaction mechanism involving ˙OH and HO₂˙ radicals as well as HO⁻ and HO₂⁻ anions. Fe²⁺ is proposed to be the active site for H₂O₂ decomposition being oxidized to Fe³⁺ and forming OH⁻ anions and ˙OH radicals, which initiate the radical chain reaction. The inferior activity of Fe³⁺ in H₂O₂ decomposition is confirmed by the lowest catalytic performance of γ-Fe₂O₃, which is a Fe²⁺-free structure. Similar to Fe²⁺ in the Haber–Weiss mechanism presented in Scheme S1,† we propose Co²⁺ to initiate a comparable reaction mechanism of H₂O₂ decomposition with a superior catalytic activity compared with Fe²⁺ and propose an analogous reaction mechanism for H₂O₂ decomposition over Co²⁺ (Scheme 2).

Scheme 2 Proposed reaction mechanism for H₂O₂ decomposition over Co²⁺ based on the Haber–Weiss mechanism.^52,53

The presence of divalent Co cations at octahedral sites of the spinel structure is thus essential for the decomposition of peroxides. This observation is directly connected to the surface termination of particles, as different facets preferentially expose different cations. Sojka et al.²⁵ investigated the surface structure of spinel nanocrystals by DFT + U calculations and TEM and found the (111) facet to be the most prominent facet with 67.5% of the overall surface area of truncated octahedral CoFe₂O₄ nanoparticles. This correlates with our findings, since TEM images of CoFe₂O₄ indicate a higher number of truncated octahedral particles with hexagonal and square surface planes (Fig. 1). Moreover, they calculated the surface composition of the (100), (110) and (111) facets by a slab model demonstrating the octahedral cations to be most pronounced on the surface of the (111) facet. For this reason, three-fold coordinated, coordinatively unsaturated Co_3c²⁺ species in octahedral sites are highly exposed at CoFe₂O₄ nanoparticles acting as most active sites for peroxide decomposition reactions. In contrast, Co₃O₄ particles were reported to predominantly expose the (100) surface, which was also indicated for Co₃O₄ in this study. In cubic Co₃O₄ nanoparticles, five-fold coordinated divalent cations are exposed at the (100) facets, which can be regarded as significantly less active, since more strongly unsaturated cations allow higher catalytic activity. Overall, the coordinatively unsaturated Co_3c²⁺ cations at octahedral sites on the (111) terminated surface of CoFe₂O₄ can be assumed to be the most active sites for peroxide decomposition.

Conclusions

A series of Co_xFe_3−xO₄ nanoparticulate catalysts with x = 3, 2.5, 2, 1.5, 1, 0.5 and 0 was prepared by spray-flame synthesis and characterized in-depth by N₂ physisorption, XRD, TEM, XPS, FT-IR, Raman and Mössbauer spectroscopy. The Co_xFe_3−xO₄ samples were found to have chemical compositions close to the nominal ones, high specific surface areas ranging from 87 to 151 m² g⁻¹ and small particle sizes between 8.3 and 12.0 nm. The cation distribution was evaluated by Mössbauer spectroscopy, being consistent with trends also seen in theoretical estimates based on the crystal field stabilization energy. The application of Co_xFe_3−xO₄ catalysts in cyclohexene oxidation resulted in a volcano plot with the highest catalytic activity at x = 1 (CoFe₂O₄). Cyclohexene conversion gradually increased with decreasing Co amount from x = 3 to x = 1. Simultaneously, the selectivity to the key intermediate 2-cyclohexene-1-hydroperoxide decreased, indicating its faster decomposition. Beyond this Co amount (x < 1), the catalytic activity strongly decreased again. The product selectivity was poorly affected by the Co and Fe amount incorporated into the Co_xFe_3−xO₄ catalysts, identifying the catalysts to mainly accelerate the decomposition of the hydroperoxide. Kinetic investigations were performed using the most active catalyst CoFe₂O₄ revealing first-order reaction kinetics with an apparent activation energy of 58 kJ mol⁻¹. A reusability test confirmed the excellent stability of the CoFe₂O₄ nanoparticles in cyclohexene oxidation. In addition, we were able to decouple 2-cyclohexene-1-hydroperoxide decomposition from the complex reaction network of cyclohexene oxidation by investigating the rates of H₂O₂ and TBHP decomposition. A very good agreement of the decomposition rates of 2-cyclohexene-1-hydroperoxide, H₂O₂ and TBHP was found, identifying CoFe₂O₄ as the most active catalyst for all three reactants, suggesting that 2-cyclohexene-1-hydroperoxide is also decomposed following the Haber–Weiss mechanism with octahedral Co_3c²⁺ exposed at the (111) facets as the most active sites.

Experimental

Materials

Co_xFe_3−xO₄ nanoparticles were synthesized using the metal precursors of cobalt(II) nitrate (Co(NO₃)₂·6H₂O, >99.0%, Honeywell) and iron(III) nitrate (Fe(NO₃)₃·6H₂O, >99.9%, Sigma-Aldrich). As solvents, ethanol (>99.9%, VWR) and 2-ethylhexanoic acid (>99%, Alfa Aesar) were used.

For cyclohexene oxidation, cyclohexene (99%), 2-cyclohexene-1-one (98%), 2-cyclohexene-1-ol (95%), cyclohexene oxide (98%), 7-oxabicyclo[4.1.0]heptane-2-one (98%), cyclohexane-1,2-diol (98%) and 1,2-dichlorobenzene (99%) were purchased from Sigma-Aldrich. Acetonitrile in analytical reagent grade was bought from Fisher Chemicals. All reagents were employed without further purification.

Catalyst synthesis

The synthesis of spinel nanoparticles was performed in a spray-flame reactor described previously in detail.^54,55 Metal nitrates were used as precursors and dissolved in a mixture of 35 vol% ethanol and 65 vol% 2-ethylhexanoic acid. The precursor solutions contained a total metal-ion concentration of 0.2 M. The solutions were injected via syringe pumps to a capillary of an external-mixing atomizing nozzle at a constant flow rate of 2 mL min⁻¹. Fine spray was formed by the interaction of the liquid flow with the dispersion O₂ gas (6 slm, Air Liquide, technical). The spray is ignited by a continuously burning premixed pilot flame of methane (3 slm, Air Liquide, N25, 99.5%) and oxygen (5 slm). The pilot flame itself is surrounded by a sheath-gas flow (140 slm, compressed air) to stabilize the flame and shield it from the reactor walls. Downstream the reaction chamber, a quenching gas flow of compressed air was added to control the sintering of the particles and the off-gas temperature at values lower than 130 °C in order to prevent the damage of the filter membrane which is used to collect the particles during the synthesis process.

Methods

X-ray diffraction patterns were recorded in a PANalytical X'Pert PRO operated with Cu-K_α radiation (0.15406 nm, 40 kV, 40 mA). The diffraction intensity was recorded at 2θ =10–60° with a step size of 0.05°.

N₂ physisorption measurements were performed at 77 K in a Quantachrome Nova2000. The specific surface areas were determined from the adsorption isotherm using the BET method.

High-resolution and high-angle annular dark-field scanning transmission electron microscopy (HR-TEM, HAADF-STEM), and energy dispersive X-ray spectroscopy (EDX) measurements were performed at a probe-side aberration-corrected JEM-2200FS (JEOL, Akishima, Japan) with an acceleration voltage of 200 kV.

Fourier-transform infrared spectroscopy was carried out in a Vertex 80 from Bruker using a KBr beam splitter, a DigiTect DLaTGS detector, and a mid-infrared light source from 400 to 4000 cm⁻¹ in combination with an ATR sample holder.

Raman spectra were recorded using a Renishaw InVia confocal Raman microscope with a 633 nm laser operating at 1% of the total laser power (15 mW).

X-ray photoelectron spectra were obtained using a ULVAC-PHI device (Versaprobe II) using a XPS twin anode with Mg Kα & Al Kα, a spot size of 100 μm, and an energetic resolution of ∼0.5 eV. The calibration of the spectra was done using the C1s adventitious carbon C–C binding energy at 284.8 eV. The obtained data were fitted based on fixed positions, fixed contributions of the specific peaks to the overall signal as well as fixed FWHM. Using this procedure, a satisfactory overall fit was obtained in which nearly no area was left unfitted.

Mössbauer spectra were recorded in transmission geometry at 4.3 K with and without applying an external magnetic field, using a ⁵⁷Co(Rh) radiation source, liquid helium (magnet) cryostats, 20–35 mg of each nanoparticle powder and a velocity transducer operating at constant acceleration.

Catalytic oxidation reactions

Oxidation reactions were carried out in a 100 mL autoclave reactor equipped with a Teflon liner (Parr Instruments). 50 mg catalyst were dispersed in 30 mL acetonitrile. 20 mmol cyclohexene and 4.5 mmol 1,2-dichlorobenzene as the internal standard for GC analysis were added. The autoclave was purged with oxygen for three times and pressurized to 10 bar. Subsequently, the reaction mixture was heated to 80 °C. The reaction was initiated by switching on the stirrer to 600 rpm. Samples were taken through an online sampling system after 0.5, 1, 2, 4, and 6 h, and analyzed by GC. The catalyst was separated by centrifugation.

To test the reusability of the catalyst, four reaction runs were carried out under standard conditions. After each run, the catalyst was separated by centrifugation, washed three times with 5 mL acetonitrile, and dried overnight at room temperature.

For the investigation of cyclohexene oxidation in the presence of H₂O, 3 mL H₂O were added to the initial standard solution whereas all other reaction conditions remained unchanged.

Conversion (X), selectivity (S), the area-related reaction rate (r) normalized to the mass (m_Cat) and the specific surface area (S_BET) of the catalyst were calculated, in which n_{Cy, 0} and n_P,0 denote the initial molar amount of cyclohexene and the respective product at t₀, respectively, and n_Cy and n_P denote the molar amount of cyclohexene and the respective product at a defined time t. The carbon balance was higher than 97% in all experiments.

Gas chromatography

Gas chromatography analysis was carried out in a 7820 A GC from Agilent Technologies. It was equipped with an Agilent DB-XLB column (30 m × 180 μm × 0.18 μm) and an FID detector. The injection volume was set to 0.5 μL with a split ratio of 75 : 1, a split flow of 30 mL min⁻¹, and an inlet temperature of 260 °C. The column was first kept at 80 °C for 5 min. Subsequently, the oven was heated to 170 °C with a rate of 15 °C min⁻¹. Afterwards, it was heated with a ramp of 30 °C min⁻¹ up to 300 °C to avoid deposits of the PPh₃ in the column. The end temperature was kept for 1 min.

TBHP decomposition

TBHP decomposition was carried out in a peroxide decomposition setup (Gasmess-5, Fig. S17†). 50 mg catalyst were dispersed in 30 mL acetonitrile. The solution was heated to 60 °C and stirred at 600 rpm. 500 μL TBHP were added and the gas volume measurement was started immediately. The decomposition reaction was run for 6 h.

H₂O₂ decomposition

H₂O₂ decomposition was carried out in a peroxide decomposition setup (Gassmess-5, Fig. S17†). 10 mg catalyst were dispersed in 30 mL acetonitrile. The solution was kept at 30 °C and stirred at 600 rpm. 80 μL H₂O₂ (30 wt%) were added and the gas volume measurement was started immediately. The decomposition reaction was run for 30 min.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation, TRR 247 “Heterogeneous Oxidation Catalysis in the Liquid Phase”, project number: 388390466). The work is supported by the “Center for Solvation Science ZEMOS” funded by the German Federal Ministry of Education and Research BMBF and by the Ministry of Culture and Research of Nord Rhine-Westphalia. We thank Dr. Hagemann and Dr. Heidelmann (Interdisciplinary Center for Analytics on the Nanoscale (ICAN), University of Duisburg-Essen) for XPS and TEM measurements, respectively.

References

1. I. M. Denekamp, M. Antens, T. K. Slot and G. Rothenberg, Selective Catalytic Oxidation of Cyclohexene with Molecular Oxygen: Radical versus Nonradical Pathways, ChemCatChem, 2018, 10, 1035–1041.
2. Y. Cao, H. Yu, F. Peng and H. Wang, Selective Allylic Oxidation of Cyclohexene Catalyzed by Nitrogen-Doped Carbon Nanotubes, ACS Catal., 2014, 4, 1617–1625.
3. B. G. Rao, P. Sudarsanam, P. R. Nallappareddy, M. Y. Reddy, T. V. Rao and B. M. Reddy, Selective Allylic Oxidation of Cyclohexene Catalyzed by Nanostructured Ce-Sm-Si Materials, Catal. Commun., 2017, 101, 57–61.
4. R. Criegee, H. Pilz and H. Flygare, Zur Kenntnis der Olefinperoxyde, Berichte der deutschen chemischen Gesellschaft (A and B Series), 1939, 72, 1799–1804.
5. S. M. Mahajani, M. M. Sharma and T. Sridhar, Uncatalysed Oxidation of Cyclohexene, Chem. Eng. Sci., 1999, 54, 3967–3976.
6. V. S. Kirankumar and S. Sumathi, A Review on Photodegradation of Organic Pollutants Using Spinel Oxide, Mater. Today Chem., 2020, 18, 100355.
7. P. Cousin and R. A. Ross, Preparation of Mixed Oxides: A Review, Mater. Sci. Eng., A, 1990, 130, 119–125.
8. T. Falk, S. Anke, H. Hajiyani, S. Saddeler, S. Schulz, R. Pentcheva, B. Peng and M. Muhler, Influence of the Particle Size on Selective 2-Propanol Gas-Phase Oxidation over Co3O4 Nanospheres, Catal. Sci. Technol., 2021, 11, 7552–7562.
9. S. Patoux, L. Daniel, C. Bourbon, H. Lignier, C. Pagano, F. Le Cras, S. Jouanneau and S. Martinet, High Voltage Spinel Oxides for Li-Ion Batteries: From the Material Research to the Application, J. Power Sources, 2009, 189, 344–352.
10. C. G. Anchieta, D. Sallet, E. L. Foletto, S. S. Da Silva, O. Chiavone-Filho and C. A. O. do Nascimento, Synthesis of Ternary Zinc Spinel Oxides and their Application in the Photodegradation of Organic Pollutant, Ceram. Int., 2014, 40, 4173–4178.
11. Y. Sun, H. Liao, J. Wang, B. Chen, S. Sun, S. J. H. Ong, S. Xi, C. Diao, Y. Du and J.-O. Wang, Covalency Competition Dominates the Water Oxidation Structure–Activity Relationship on Spinel Oxides, Nat. Catal., 2020, 3, 554–563.
12. S. Pal, U. P. Azad, A. K. Singh, D. Kumar and R. Prakash, Studies on Some Spinel Oxides Based Electrocatalysts for Oxygen Evolution and Capacitive Applications, Electrochim. Acta, 2019, 320, 134584.
13. J. Büker, B. Alkan, Q. Fu, W. Xia, J. Schulwitz, D. Waffel, T. Falk, C. Schulz, H. Wiggers and M. Muhler, Selective Cyclohexene Oxidation with O2, H2O2 and Tert-Butyl Hydroperoxide over Spray-Flame Synthesized LaCo1−xFexO3 Nanoparticles, Catal. Sci. Technol., 2020, 10, 5196–5206.
14. T. Falk, E. Budiyanto, M. Dreyer, C. Pflieger, D. Waffel, J. Büker, C. Weidenthaler, K. F. Ortega, M. Behrens, H. Tüysüz, M. Muhler and B. Peng, Identification of Active Sites in the Catalytic Oxidation of 2-Propanol over Co1+xFe2-xO4 Spinel Oxides at Solid/Liquid and Solid/Gas Interfaces, ChemCatChem, 2021, 13, 2942–2951.
15. D. Waffel, E. Budiyanto, T. Porske, J. Büker, T. Falk, Q. Fu, S. Schmidt, H. Tüysüz, M. Muhler and B. Peng, Investigation of Synergistic Effects between Co and Fe in Co3-xFexO4 Spinel Catalysts for the Liquid-Phase Oxidation of Aromatic Alcohols and Styrene, Mol. Catal., 2020, 498, 111251.
16. J. Büker, B. Alkan, S. Chabbra, N. Kochetov, T. Falk, A. Schnegg, C. Schulz, H. Wiggers, M. Muhler and B. Peng, Liquid-Phase Cyclohexene Oxidation with O2 over Spray-Flame-Synthesized La1-xSrxCoO3 Perovskite Nanoparticles, Chem. – Eur. J., 2021, 27, 16912–16923.
17. A. Rabe, J. Büker, S. Salamon, A. Koul, U. Hagemann, J. Landers, K. Friedel Ortega, B. Peng, M. Muhler, H. Wende, W. Schuhmann and M. Behrens, The Roles of Composition and Mesostructure of Cobalt-based Spinel Catalysts in Oxygen Evolution Reactions, Chem. – Eur. J., 2021, 27, 17038–17048.
18. Y.-I. Jang, H. Wang and Y.-M. Chiang, Room-Temperature Synthesis of Monodisperse Mixed Spinel (CoxMn1–x)3O4 Powder by a Coprecipitation Method, J. Mater. Chem., 1998, 8, 2761–2764.
19. T. H. Dolla, K. Pruessner, D. G. Billing, C. Sheppard, A. Prinsloo, E. Carleschi, B. Doyle and P. Ndungu, Sol-Gel Synthesis of MnxNi1-xCo2O4 Spinel Phase Materials: Structural, Electronic, and Magnetic Properties, J. Alloys Compd., 2018, 742, 78–89.
20. Z. Chen, E. Shi, W. Li, Y. Zheng and W. Zhong, Hydrothermal Synthesis and Optical Property of Nano-Sized CoAl2O4 Pigment, Mater. Lett., 2002, 55, 281–284.
21. M. Seyedahmadian, S. Houshyarazar and A. Amirshaghaghi, Synthesis and Characterization of Nanosized of Spinel LiMn2O4 via Sol-Gel and Freeze Drying Methods, Bull. Korean Chem. Soc., 2013, 34, 622–628.
22. B. Alkan, D. Medina, J. Landers, M. Heidelmann, U. Hagemann, S. Salamon, C. Andronescu, H. Wende, C. Schulz and W. Schuhmann, Spray-Flame-Prepared LaCo1-xFexO3 Perovskite Nanoparticles as Active OER Catalysts: Influence of Fe Content and Low-Temperature Heating, ChemElectroChem, 2020, 7, 2564–2574.
23. S. Angel, J. Neises, M. Dreyer, K. Friedel Ortega, M. Behrens, Y. Wang, H. Arandiyan, C. Schulz and H. Wiggers, Spray-Flame Synthesis of La(Fe,Co)O3 Nano-Perovskites from Metal Nitrates, AIChE J., 2020, 66, e16748.
24. B. Varghese, C. H. Teo, Y. Zhu, M. V. Reddy, B. V. R. Chowdari, A. T. S. Wee, V. B. Tan, C. T. Lim and C.-H. Sow, Co3O4 Nanostructures with Different Morphologies and their Field-Emission Properties, Adv. Funct. Mater., 2007, 17, 1932–1939.
25. F. Zasada, J. Grybos, P. Indyka, W. Piskorz, J. Kaczmarczyk and Z. Sojka, Surface Structure and Morphology of M [CoM']O4 (M = Mg, Zn, Fe, Co and M′ = Ni, Al, Mn, Co) Spinel Nanocrystals: DFT+ U and TEM Screening Investigations, J. Phys. Chem. C, 2014, 118, 19085–19097.
26. S. Wang, Y. Zhao, H. Xue, J. Xie, C. Feng, H. Li, D. Shi, S. Muhammad and Q. Jiao, Preparation of flower-like CoFe2O4@ graphene composites and their microwave absorbing properties, Mater. Lett., 2018, 223, 186–189.
27. R. Al-Tuwirqi, A. A. Al-Ghamdi, N. A. Aal, A. Umar and W. E. Mahmoud, Facile Synthesis and Optical Properties of Co3O4 Nanostructures by the Microwave Route, Superlattices Microstruct., 2011, 49, 416–421.
28. L. J. Cote, A. S. Teja, A. P. Wilkinson and Z. J. Zhang, Continuous Hydrothermal Synthesis of CoFe2O4 Nanoparticles, Fluid Phase Equilib., 2003, 210, 307–317.
29. G. Gnanaprakash, S. Ayyappan, T. Jayakumar, J. Philip and B. Raj, Magnetic Nanoparticles with Enhanced γ-Fe2O3 to α-Fe2O3 Phase Transition Temperature, Nanotechnology, 2006, 17, 5851.
30. A. Diallo, A. C. Beye, T. B. Doyle, E. Park and M. Maaza, Green Synthesis of Co3O4 Nanoparticles via Aspalathus Linearis: Physical Properties, Green Chem. Lett. Rev., 2015, 8, 30–36.
31. B. Rivas-Murias and V. Salgueiriño, Thermodynamic CoO-Co3O4 Crossover Using Raman Spectroscopy in Magnetic Octahedron-Shaped Nanocrystals, J. Raman Spectrosc., 2017, 48, 837–841.
32. Y. Zheng, Y. Yu, H. Zhou, W. Huang and Z. Pu, Combustion of Lean Methane over Co3O4 Catalysts Prepared with Different Cobalt Precursors, RSC Adv., 2020, 10, 4490–4498.
33. K. Gandha, K. Elkins, N. Poudyal and J. Ping Liu, Synthesis and Characterization of CoFe2O4 Nanoparticles with High Coercivity, J. Appl. Phys., 2015, 117, 17A736.
34. V. Bartůněk, D. Sedmidubský, Š. Huber, M. Švecová, P. Ulbrich and O. Jankovský, Synthesis and Properties of Nanosized Stoichiometric Cobalt Ferrite Spinel, Materials, 2018, 11, 1241.
35. R. N. Bhowmik and N. Naresh, Structure, ac Conductivity and Complex Impedance Study of Co3O4 and Fe3O4 Mixed Spinel Ferrites, International Journal of Engineering, Science and Technology, 2010, 2, 40–52.
36. R. Sagayaraj, S. Aravazhi and G. Chandrasekaran, Tuning of Ferrites (CoxFe3-xO4) Nanoparticles by Co-Precipitation Technique, Appl. Sci., 2019, 1, 1–11.
37. A. M. Wahba and M. B. Mohamed, Structural and Magnetic Characterization and Cation Distribution of Nanocrystalline CoxFe3−xO4 Ferrites, J. Magn. Magn. Mater., 2015, 378, 246–252.
38. M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. Lau, A. R. Gerson and R. S. Smart, Resolving Surface Chemical States in XPS Analysis of First Row Transition Metals, Oxides and Hydroxides: Cr, Mn, Fe, Co and Ni, Appl. Surf. Sci., 2011, 257, 2717–2730.
39. H. Liu, G. Wei, Z. Xu, P. Liu and Y. Li, Quantitative Analysis of Fe and Co in Co-Substituted Magnetite Using XPS: The Application of Non-Linear Least Squares Fitting (NLLSF), Appl. Surf. Sci., 2016, 389, 438–446.
40. H. Le Trong, L. Presmanes, E. de Grave, A. Barnabé, C. Bonningue and P. Tailhades, Mössbauer Characterisations and Magnetic Properties of Iron Cobaltites CoxFe3−xO4 (1≤ x≤ 2.46) Before and After Spinodal Decomposition, J. Magn. Magn. Mater., 2013, 334, 66–73.
41. P. J. Murray and J. W. Linnett, Mössbauer studies in the spinel system CoxFe3− xO4, J. Phys. Chem. Solids, 1976, 37, 619–624.
42. G. M. Bancroft, T. K. Sham, C. Riddle, T. E. Smith and A. Turek, Ferric/Ferrous-Iron Ratios in Bulk Rock Samples by Mössbauer Spectroscopy—The Determination of Standard Rock Samples G-2, GA, W-1 and mica-Fe, Chem. Geol., 1977, 19, 277–284.
43. H. M. Cota, T. Katan, M. Chin and F. J. Schoenweis, Decomposition of Dilute Hydrogen Peroxide in Alkaline Solutions, Nature, 1964, 203, 1281.
44. J. R. Goldstein and A. C. Tseung, The Kinetics of Hydrogen Peroxide Decomposition Catalyzed by Cobalt-Iron Oxides, J. Catal., 1974, 32, 452–465.
45. A. I. Onuchukwu and A. B. Zuru, The Cobalt-Ferrite, CoxFe3−xO4, Activity in the Catalytic Chemical Decomposition of Hydrogen Peroxide, Mater. Chem. Phys., 1986, 15, 131–138.
46. T. Mimani, P. Ravindranathan and K. C. Patil, Catalytic Decomposition of Hydrogen Peroxide on Fine Particle Ferrites and Cobaltites, Springer, 1987, vol. 4.
47. P. Lahiri and S. K. Sengupta, Spinel Ferrites as Catalysts: A Study on Catalytic Effect of Coprecipitated Ferrites on Hydrogen Peroxide Decomposition, Can. J. Chem., 1991, 69, 33–36.
48. T. Tatarchuk, A. Shyichuk, I. Trawczyńska, I. Yaremiy, A. T. Pędziwiatr, P. Kurzydło, B. F. Bogacz and R. Gargula, Spinel cobalt (II) ferrite-chromites as catalysts for H2O2 decomposition: Synthesis, morphology, cation distribution and antistructure model of active centers formation, Ceram. Int., 2020, 46, 27517–27530.
49. U. Müller, Anorganische Strukturchemie, Vieweg+Teubner Verlag/GWV Fachverlage GmbH Wiesbaden, Wiesbaden, 6th edn, 2008.
50. A. I. Onuchukwu, Kinetics of the Decomposition of Hydrogen Peroxide Catalysed by Copper and Nickel Ferrites, J. Chem. Soc., Faraday Trans. 1, 1984, 80, 1447–1456.
51. A. C. Tseung and J. R. Goldstein, The Preparation and Characterization of Ultrafine Cobalt-Iron Oxides, J. Mater. Sci., 1972, 7, 1383–1390.
52. F. Haber and J. Weiss, The Catalytic Decomposition of Hydrogen Peroxide by Iron Salts, Proc. R. Soc. London, Ser. A, 1934, 147, 332–351.
53. W. G. Barb, J. H. Baxendale, P. George and K. R. Hargrave, Reactions of Ferrous and Ferric Ions with Hydrogen Peroxide. Part I.—The Ferrous Ion Reaction, Trans. Faraday Soc., 1951, 47, 462–500.
54. S. Angel, F. Schneider, S. Apazeller, W. Kaziur-Cegla, T. C. Schmidt, C. Schulz and H. Wiggers, Spray-Flame Synthesis of LaMO3 (M= Mn, Fe, Co) Perovskite Nanomaterials: Effect of Spray Droplet Size and Esterification on Particle Size Distribution, Proc. Combust. Inst., 2021, 38, 1279–1287.
55. S. Hardt, I. Wlokas, C. Schulz and H. Wiggers, Impact of Ambient Pressure on Titania Nanoparticle Formation During Spray-Flame Synthesis, J. Nanosci. Nanotechnol., 2015, 15, 9449–9456.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2cy00505k

---